Economic
zhuodahai
MicroeconomicsTheoryandApplications12th.pdf
Microeconomics: Theory & Applications
Twelfth Edition
Edgar K. Browning Texas A&M University
Mark A. Zupan University of Rochester
Vice President & Executive Publisher George Hoffman
Executive Editor Joel Hollenbeck
Sponsoring Editor Marian Provenzano
Project Editor Brian Baker
Associate Editor Christina Volpe
Marketing Manager Puja Katariwala
Design Director Harry Nolan
Senior Designer Maureen Eide
Assistant Editor Courtney Luzzi
Senior Editorial Assistant Jacqueline Hughes
Editorial Assistant Tai Harris
Cover Photo Credit © hddigital /iStockphoto
This book was set in 10/12 STIXGeneral by Laserwords and printed and bound by Quad Graphics
Versailles. The cover was printed by Quad Graphics Versailles.
Copyright © 2015, 2012, 2008 John Wiley & Sons, Inc. All rights reserved. No part of this
publication may be reproduced, stored in a retrieval system or transmitted in any form or by any
means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted
under Sections 107 or 108 of the 1976 United States Copyright Act, without either the prior written
permission of the Publisher, or authorization through payment of the appropriate per-copy fee to
the Copyright Clearance Center, Inc. 222 Rosewood Drive, Danvers, MA 01923, website www.
copyright.com. Requests to the Publisher for permission should be addressed to the Permissions
Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030-5774, (201)748-6011,
fax (201)748-6008, website http://www.wiley.com/go/permissions.
Founded in 1807, John Wiley & Sons, Inc. has been a valued source of knowledge and
understanding for more than 200 years, helping people around the world meet their needs and
fulfill their aspirations. Our company is built on a foundation of principles that include
responsibility to the communities we serve and where we live and work. In 2008, we launched a
Corporate Citizenship Initiative, a global effort to address the environmental, social, economic,
and ethical challenges we face in our business. Among the issues we are addressing are carbon
impact, paper specifications and procurement, ethical conduct within our business and among
our vendors, and community and charitable support. For more information, please visit our
website: www.wiley.com/go/citizenship.
Evaluation copies are provided to qualified academics and professionals for review purposes
only, for use in their courses during the next academic year. These copies are licensed and may
not be sold or transferred to a third party. Upon completion of the review period, please return
the evaluation copy to Wiley. Return instructions and a free of charge return shipping label are
available at www.wiley.com/go/returnlabel. Outside of the United States, please contact your local
representative.
ISBN-13 978-1118-75887-8
ISBN-10 1118-75887-0
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1
iii
FM.INDD 07:39:38:PM 08/08/2014 PAGE IIITrim Size: 203.2 mm X 254 mm
Preface
According to certain labor unions, traditional retailers, and community groups, Walmart imposes significant costs on society. Among the asserted costs are the destruction of jobs
in competing stores, driving of employees toward public welfare systems by paying lower
wages and providing limited health care coverage, and the fostering of urban sprawl. Arrayed
against these claimed costs are the benefits generated by Walmart through the employment
of a large number of workers (Walmart now is the largest private-sector employer in the
United States) and the promotion of lower retail prices for consumers.
How can one assess the validity of the claims made by Walmart’s critics? Moreover, is
the combined magnitude of costs associated with Walmart sizable enough to outweigh the
benefits generated by the retailing giant? A thorough knowledge of microeconomics can
help answer topical questions such as these and, more broadly, gives students an under-
standing of how markets operate and allows them to see the world through the eyes of an
economist.
Our intention with this edition of the text is to give students the fundamental tools of
analysis and to show how the tools can be used to explain and predict market phenomena.
To this end, we present basic microeconomic principles in a clear, thorough way,
using numerous applications to illustrate the use of theory and to reinforce students’
understanding of it.
We believe that microeconomics is the most important course in the undergraduate
economics curriculum. We also believe that understanding microeconomics provides an
essential foundation to any bachelor’s or master’s degree business student. As a result,
our text is written so that both economics and business students will learn microeconomic
theory and how to use it correctly.
Organization and Content
The twelth edition of Microeconomics: Theory and Applications continues to reflect our belief that it is better for students to be exposed to thorough coverage of fundamental
microeconomic concepts and techniques than to skim through a superficial treatment of
a great number of topics, many of which they will never encounter again. The enthu-
siastic reception given the first 11 editions suggests that a large number of instructors
also share this view. Apart from the emphasis on the core principles of microeconomics
and how to use them, the text is by and large conventional in structure and organiza-
tion except for one feature: Four chapters are devoted exclusively to applications. These
are Chapter 5, “Using Consumer Choice Theory”; Chapter 10, “Using the Competitive
Model”; Chapter 15, “Using Noncompetitive Market Models”; and Chapter 18, “Using
Input Market Analysis.”
A distinguishing feature of the text is the attention we give to input market analysis. Tradi-
tionally, this has been a weak area in most microeconomics texts, with seldom more than two
chapters, and frequently only one, on the subject. Yet in a fundamental quantitative sense,
input markets and product markets are of equal importance, because the sum of incomes gen-
erated in input markets (national income) equals total outlays on goods and services (national
product). Moreover, public policy issues relating to input markets have become increasingly
iv Preface
FM.INDD 07:39:38:PM 08/08/2014 PAGE IVTrim Size: 203.2 mm X 254 mm
important, as suggested by the recent attention given to managerial compensation, income
distribution, welfare programs, discrimination, comparable worth, interest rates and invest-
ment, Social Security, and minimum wage legislation. Consequently, we devote three chap-
ters to the subject of input market analysis (Chapters 16 through 18).
Because not all microeconomics courses are taught the same way, the text is designed
to give instructors great flexibility in adapting the book to their requirements. For example,
in a short course emphasizing the theoretical underpinnings of partial equilibrium analysis,
the instructor might cover only Chapters 1 through 4, 7 through 11, 16, and 17. A longer,
more theoretically oriented course could include all chapters except that most instructors
will steer a middle course and select three or four applications from each of these chapters
(the way we normally use the material). In addition, instructors can either assign the appli-
cations as they appear in the text—following the development of the theory—or integrate
them into their presentations of the theory chapters.
Applications
We believe that a large dose of applications is an essential ingredient in any microeco-
nomics course. Although economists know that microeconomics is important and often
exciting, students occasionally need to be convinced that this is so. Applications serve this
purpose. In addition, they enliven the subject for students and help them better appreciate
the theory. Time permitting, the more applications covered, the better prepared students
will be to use the theory on their own.
Each of the four applications chapters (Chapters 5, 10, 15, and 18) contains four
to six longer applications that use and reinforce the graphical and logical techniques
developed in the theory chapters. In Chapter 10, for example, the competitive model
is employed to analyze taxicab licensing, airline regulation, and international trade. In
Chapter 18, “Using Input Market Analysis,” the theory is applied to discrimination,
the incidence of the Social Security payroll tax, the effects of the National Collegiate
Athletic Association on college athletes, and the benefits and costs of immigration.
Applications are not relegated exclusively to the four applications chapters; all other
chapters contain several shorter applications. We feel, however, that it is appropriate to
use more applications in some areas than in others. For example, it seems a misallocation
of limited textbook space to include as many applications for general equilibrium theory as
for the competitive and monopoly models. Not only are the applications in the latter two
areas likely to be more interesting to students, they are also likely to provide more useful
background for students’ later work.
Changes in the Twelfth Edition
Based on comments from users and reviewers of the eleventh edition, as well as our own
desire to further improve the text, we have revised it in three important ways. The principal
aim of our revisions is to enhance the hallmarks of the text, namely: the wealth of real-
world illustrations of microeconomic theory at work; clear and engaging exposition; and a
commitment to coverage of cutting-edge concepts.
Only the Best Applications When asked to identify the strengths of this text, reviewers and users overwhelmingly cite
the applications—the four chapters devoted to longer illustrations of microeconomic theory
Preface v
FM.INDD 07:39:38:PM 08/08/2014 PAGE VTrim Size: 203.2 mm X 254 mm
at work as well as the 100-plus shorter applications sprinkled throughout the other chapters.
To continue building on this hallmark, we rely on a systematic rating system whereby we
ask reviewers to evaluate each application. On the basis of their responses, we have added
20 new applications in this edition. The topical issues these address include the rise and
fall of cigarette consumption in the United States; the growth of premium fast food options
such as Chipotle; promoting efficiency in gift card giving; moral hazard when it comes
to the operation of taxi cabs in New York City; the differing fortunes of college athletes
and coaches; why Canadians are flying south of the border; why holiday home prices in
Switzerland are soaring; privatization and productivity in China; and monopolistic compe-
tition in the refractive eye surgery marketplace.
We have retained (and whenever possible, enhanced) the top 80 percent of the
applications from the previous edition. These applications cover topics such as monopsony
in Major League Baseball; whether cell phone use while driving should be banned;
trash pricing and recycling; the demand for and supply of school choice; the economic
and accounting costs of the Sarbanes–Oxley (SOX) Act intended to enhance corporate
governance; the hidden cost of our Social Security system; why price ceilings are proving
deadly to individuals seeking an organ transplant; the returns to investing in a BA and an
MBA; and compensating wage differentials for “glowboys”—individuals who fix steel
pipes in aging nuclear power plants.
By culling the cream of the applications from the preceding edition and adding
numerous interesting demonstrations of the way microeconomic theory can be used to
explain and predict real-world phenomena, we’ve made the book’s best-regarded feature—
its applications—stronger than ever in this revision.
Clear and Engaging Exposition A second key feature of the text consistently noted by adopters and reviewers is its
clarity of exposition. To strengthen this feature even further, we have looked carefully
at each chapter—applying Occam’s razor to make our explanations as straightforward
as possible. We have also sought to relegate all optional materials (e.g., mathematical
appendices) to the accompanying book companion site. One telling manifestation of the
care that has been applied to focusing on the essentials is that the actual text is approxi-
mately a full pound lighter than competing texts, notwithstanding our thoroughness in
the coverage of key topics. The expositional clarity translates into an important benefit
for students of microeconomics—superior comprehension.
To make the text as clear as possible and more engaging, we have paid particular
attention to the illustrations and how they teach economic concepts. Effective graphs can
truly be worth thousands of words. From the layout of graphs and tables to the number of
subsections a chapter is broken down into, we have sought to respond to comments from
reviewers and adopters as to the best way to showcase the content and thereby promote
positive learning outcomes.
A Commitment to the Cutting Edge Some key themes in business and economics education today are globalization, ethics/
integrity, sustainability, and the appropriate role of government in society. Wherever pos-
sible, we have sought to show how economics can contribute to students’ understanding
of these topics, often in unexpected ways. For example, Chapter 14 provides a framework
for understanding when ethical leadership is more likely to emerge through the context of
a prisoner’s dilemma game. The extent to which such a game is one-shot versus repeated
and indefinitely lived allows us to predict the settings in which prospective leaders may
be more likely to behave unethically. The perspective also allows us to grasp why market
settings and capitalism, through the promotion of repeated and indefinitely lived settings,
may encourage greater integrity.
vi Preface
FM.INDD 07:39:38:PM 08/08/2014 PAGE VITrim Size: 203.2 mm X 254 mm
The text and accompanying applications of Chapters 5, 7, and 20 explore a variety of
aspects related to pollution and sustainability. Perhaps surprisingly to many students (but
not to their economics instructors), this coverage illustrates how incentives and markets can
actually be used to promote environmentally beneficial activities. Chapter 20, for instance,
shows how the absence of liability caps would have encouraged British Petroleum to
exercise greater care in its drilling operations and thereby perhaps averted the disaster that
occurred in the Gulf of Mexico during the spring of 2010.
With regard to the appropriate role of government in society, Chapter 20 explores
the extent to which government prolonged the Great Depression. This is a timely topic,
given the recent debate over the appropriate government response to the major economic
downturn of 2007–2010. Application 2.1 delves into extending unemployment insurance
benefits and the impact on the unemployment level. Application 5.2 discusses school
choice. A longer application in Chapter 5 provides a detailed analysis of certain important
effects of the recent overhaul of health care policy in the United States (i.e., ObamaCare).
On the topic of globalization, Application 3.2 explains how Kraft successfully revised
the Oreo in order to account for different consumer preferences in China. A full section in
Chapter 10 deals with the net benefits of trade while Application 7.3 analyzes how returns
to scale explain cross-country trade flows.
Students also have access to Excel-based tutorials relating to 16 of the key microeco-
nomic concepts covered by our text through the companion Web site www.wiley.com/ college/browning. These concepts are typically covered in an intermediate microeconomics one-semester course, and they enable students to manipulate the graphical presentations so
that they can actually see the concepts in action as they change the input values associated
with each tutorial.
Pedagogical Aids
Several other in-text pedagogical aids help students to structure and retain information.
Learning Objectives Each chapter begins with a list of key learning objectives. These offer a preview of the
chapter content and help structure study and review.
Glossary A running glossary has been added in the margins of the text as a way to cement students’
understanding of key concepts and terms. A complete glossary is also included at the end
of the book.
Graphs We have paid careful attention to the graphs used in the text. Unusually thorough explana-
tions of graphs are given. Furthermore, the explanatory captions and liberal use of color
will help students follow the text discussion and understand graphical analysis.
End-of-Chapter Aids A summary at the end of each chapter highlights the important points of the chapter to help
students review their knowledge of the basic material. More than 450 review questions and
problems test students on chapter material and require them to solve analytical exercises.
Answers to questions and problems with asterisks are provided on the Book Companion Site.
Preface vii
FM.INDD 07:39:38:PM 08/08/2014 PAGE VIITrim Size: 203.2 mm X 254 mm
Chapter Appendices Starred appendices for chapters 2,3,4,6,7,8,9,11,12 and 16 are available on the book com-
panion site.
Teaching and Learning Resources
An Instructor’s Manual, written by Brian is the only author for the IM and he is still at the univ. of tampa Brian Kench, of University of Tampa, accompanies the text. Each chapter
in the manual features a chapter outline, general comments on the chapter, specific sec-
tion-by-section comments, and suggestions that may help in developing lectures and class
discussion topics. The appendix in the Instructor’s Manual contains the answers to those
questions and problems in the text that are not already answered at the end of the text.
Lecture Slides in PowerPoint, prepared by Della Lee Sue of Marist College, provides notes for all chapters with enlarged versions of all the figures contained in the text. This set
can be used to create overhead transparencies for viewing in the classroom, or they can be
copied and used as handouts for students.
The Test Bank prepared by Kenneth Slaysman of York College of Pennsylvania, con- tains 1,500 multiple-choice and short answer questions with answers. This material is also
available electronically through Respondus, enabling instructors to create and manage
exams that can be printed or published directly to their LMS.
A study guide prepared by Lori B. Anderson is also available. Each chapter features an
in-depth section-by-section analysis, a key concepts review list, and a variety of practice
and discussion questions.
A dedicated Web site with extensive resources for both students and professors
(www.wiley.com/college/browning) is also available as are videos providing additional helpful learning materials from a microeconomics course taught in 2013–2014 at the
University of Rochester, Simon Business School by Mark Zupan through Coursera, top
Massively Open Online Course provider.
The Wiley E-Text: Powered by VitalSource gives students anytime, anywhere, access to the best economics content when and where they study: on their desktop, laptop, tablet,
or smartphone. Students can search across content, highlight, and take notes that they can
share with teachers and classmates.
Wiley’s E-Text for Microeconomics: Theory and Applications, 12th edition takes learning from traditional to cutting edge by integrating inline interactive multimedia with
market–leading content. This exciting new learning model brings textbook pages to life-no
longer just a static e-book, the E-Text enriches the study experience with dynamic features:
Interactive Tables and Graphs allow students to access additional rich layers and “hot areas” of explanation by manipulating slider controls or clicking on embedded
“hotspots” incorporated into select tables and graphs
Embedded Practice Quizzes allow students to practice as they read and thereby receive instant feedback on their progress
Audio-Enhanced Graphics provide further explanations for key graphs in the form of short audio clips.
viii Preface
FM.INDD 07:39:38:PM 08/08/2014 PAGE VIIITrim Size: 203.2 mm X 254 mm
Acknowledgments
We have been fortunate to have had the assistance of many able economists in the prepara-
tion of this book. Those who have worked at various stages in the development of this edi-
tion and the 11 editions that preceded it include:
David Anderson, Centre College Gary Anderson, California State University,
Northridge Peter Aranson, Emory University, deceased Michael Arnold, University of Delaware Lee Badgett, University of Massachusetts Jeff Baldani, Colgate University Hamid Bastin, Shippensburg University Marco Battaglini, Princeton University Doug Berg, Sam Houston State University Anu Bhayana, California State University,
Fullerton Tibor Besedes, Georgia Institute of
Technology David Black, University of Delaware David Blau, University of North Carolina Larry Blume, University of Michigan David Boyd, Denison University Wayne Boyet, University of Mississippi Charles Breeden, Marquette University Bruce Brown, California State Polytechnic
University, Pomona Jack Bucco, Austin Comminity College Stuart Burness, University of New Mexico Richard Butler, Trinity University Bruce Caldwell, University of North Caro-
lina, Greensboro Charles A. Capone, Baylor University Richard Caves, Harvard University David Chaplin, State of Wyoming Joni Charles, Texas State University Basanta Chaudhuri, Rutgers University Whewon Cho, Tennessee Technology
University Dennis Coates, Ithaca College Alvin Cohen, Lehigh University Patrice Karr Cohen, University of Mississippi Basil Coley, North Carolina, A&T State
University Darius Conger, Central Michigan University David Conn, University of Kansas Robert Connolly, University of North Caro-
lina, Greensboro Peter Crabb, Northwest Nazarene University Erik Craft, University of Richmond Jerry L. Crawford, Arkansas State University Tom Creahan, Morehead State University Keith J. Crocker, University of Michigan Steven Cuellar, Sonoma State University Manabendra Dasgupta, University of Ala-
bama, Birmingham Carl Davidson, Michigan State University Dennis Debrecht, Carroll College
Eddie Dekel, Northwestern University Cliff Dobitz, North Dakota State University Asif Dowla, St. Mary’s College of Maryland Rich Eastin, University of Southern
California Robert R. Ebert, Baldwin-Wallace College Karl Einolf, Mount Saint Mary’s University Robert Ekelund, Auburn University David Emmons, Wayne State University Maxim Engers, University of Virginia William J. Field, DePauw University Richard E. French, Deree College Yee-Tien Fu, Stanford University Richard Gaddis, Oklahoma Wesleyan
University Ian Gale, University of Wisconsin Javier F. Garcia III, Ashland University Carlos Garriga, Florida State University David Gay, University of Arkansas Charles Geiss, University of Missouri Soumen Ghosh, Tennessee State University Amy Gibson, University of South Alabama James Giordano, Villanova University John Goddeeris, Michigan State University Robert Goldfarb, George Washington
University Michael Gootzeit, University of Memphis Lawrence H. Goulder, Stanford University Warren Gramm, Washington State University Chiara Gratton-Lavoie, California State Uni-
versity, Fullerton Stuart Greenfield, St. Edward’s University Thomas Gresik, University of Notre Dame James M. Griffin, Texas A&M University Elias C. Grivoyannis, Yeshiva University Timothy Gronberg, Texas A&M University Joseph Haimowitz, Avila University Shawkat Hammoudeh, Drexel University Robert G. Hansen, Amos Tuck School, Dart-
mouth College John Harford, Cleveland State University Mehdi Haririan, Bloomsburg University Janice Hauge, University of North Texas Jack Henderson, Covenant Christian College Philip Hersch, Wichita State University Glen Hueckel, Purdue University Thomas Hiestand, Concordia College Barry Hirsch, Georgia State Don Holley, Boise State University James Holcomb, University of Texas, El Paso W. L. Holmes, Temple University Gary Hoover, University of Alabama Joseph Hughes, Rutgers University Joseph Hunt, Shippensburg University
E. Bruce Hutchison, University of Tennessee, Chattanooga
Jeanette Iwasa, California State University, Fresno
Joseph Jadlow, Oklahoma State University Harvey James, University of Missouri Sumit Joshi, George Washington University Folke Kafka, University of Pittsburgh Joseph P. Kalt, Harvard University Arthur Kartman, San Diego State University Mary Kassis, State University of West
Georgia Amoz Katz, Virginia Technology University Larry Kenny, University of Florida Philip King, San Francisco State University Edward Kittrell, Northern Illinois University Janet Koscianski, Shippensburg University Yasar Kuyuk, Manhattan College Leonard Lardero, University of Rhode Island Daniel Y. Lee, Shippensburg University Sang H. Lee, Southeastern Louisiana
University Tom Lee, California State University,
Northridge Robert J. Lemke, Lake Forest College Gary Lemon, DePauw University Armando Levy, North Carolina State
University Benjamin Liebman, Saint Joseph’s University Al Link, Auburn University Christine Loucks, Boise State University Leonard Loyd, University of Houston R. Ashley Lyman, University of Idaho Mark Machina, University of California, San
Diego Robert Main, Butler University Robert Maness, Texas A&M University Chris Manner, Lambuth University Charles Mason, University of Wyoming James McClure, Ball State University William McEachern, University of
Connecticut Haririan Mehdi, Bloomsburg University Khalid Mehtabdin, The College of Saint Rose John Merrifield, The University of Texas,
San Antonio Robert Michaels, California State University,
Fullerton D. E. Mills, University of Virginia Aparna Mitra, University of Oklahoma Robby Moore, Occidental College Roger Morefield, University of Saint Thomas Bob Mueller, Covenant Christian College John Nader, Grand Valley State University
Preface ix
FM.INDD 07:39:38:PM 08/08/2014 PAGE IXTrim Size: 203.2 mm X 254 mm
Aimee Narcisenfeld, George Washington University
Edd Noell, Westmont College William Novshek, Stanford University Yuka Ohno, Rice University Richard E. Olsen, Washburn University William O’Neil, Colby College Patrick B. O’Neill, University of North
Dakota Z. Edward O’Relley, North Dakota State
University Lydia Ortega, San Jose State University H. Craig Petersen, Utah State University Jeffrey Pliskin, Hamilton College Bruce Prengruber, Washington State Univer-
sity, Vancouver Ed Price, Oklahoma State University Rati Ram, Illinois State University Michael Ransom, Brigham Young University Michael Reclam, Virginia Military Institute John Riley, University of California, Los
Angeles H. David Robison, LaSalle University Michael Rogers, Albany State University
Eric Schansberg, Indiana University, New Albany
Ute Schumacher, Lafayette College Peter Schuhmann, University of North Caro-
lina, Wilmington Radwan Shaban, Georgia Institute of
Technology Stephen Shmanske, California State Univer-
sity, Hayward Arvind Singh, University of Arizona David Sisk, San Francisco State University Gene Smiley, Marquette University Peter F. Smith, Southeastern University Scott Smith, George Mason University William Doyle Smith, University of Texas,
El Paso Hubert O. Sprayberry, Howard Payne
University Farley Ordovensky Staniec, University of the
Pacific Stanley Stephenson, Pennsylvania State
University Douglas Steward, San Diego State University Eric Stout, Brandeis University
Della Lee Sue, Marist College Vasant Sukhatme, Macalester College Shirley Svorny, California State University,
Northridge Thom Swanke, Chadron State College Wayne Talley, Old Dominion University Bryan Taylor, California State University,
Los Angeles Roger Trenary, Kansas State University Roy Van Til, Bentley College Michele Villinski, DePauw University Nicholas Vonortas, George Washington
University Charles Waldauer, Widener University Don Waldman, Colgate University Doug Walker, Louisiana State University Donald Wells, University of Arizona Tara Westerhold, Western Illinois
University Mark White, College of Staten Island/CUNY Arlington Williams, Indiana University Andrew Yuengert, Pepperdine University Huizhong Zhou, Western Michigan
University
These reviewers were generous with their time, and their comments have greatly
enhanced the quality of the book. To them we extend our gratitude and our hope that the
final product meets with their approval.
Special mention should be made of the late Jacquelene M. Browning, who was the co-
author of the first four editions. Her pedagogical skills, together with her insistence that the
text be one from which students could learn effectively, continue to have a profound influ-
ence in the present edition.
Edie Trimble provided outstanding research and editorial assistance. We would also
like to thank the people at John Wiley who made important contributions to this edition,
in particular, Vice President and Executive Publisher, George Hoffman, Executive Editor,
Joel Hollenbeck, Sponsoring Editor, Marian Provenzano, Project Editor, Brian Baker, Edi-
torial Assistants Courtney Luzzi and Jacqueline Hughes, Senior Production Editor, Anna
Melhorn, Senior Designer, Maureen Eide, Associate Editor, and Christina Volpe.
This book is dedicated to our family and friends, without whose unflagging encourage-
ment and support our vision for the book would have never become a reality.
Edgar K. Browning Mark A. Zupan
x Table of Contents
FM.INDD 07:39:38:PM 08/08/2014 PAGE XTrim Size: 203.2 mm X 254 mm
x
Table of Contents
Preface iii
Acknowledgments viii
Chapter 1: An Introduction to Microeconomics 1
1.1 The Scope of Microeconomic Theory .......... 2 1.2 The Nature and Role of Theory ................... 2
What Is a Good Theory? 2
1.3 Positive versus Normative Analysis ............. 3 1.4 Market Analysis and Real versus
Nominal Prices ............................................... 4 Application 1.1 Real Versus Nominal Presidential Salaries ..............................................................5
1.5 Basic Assumptions about Market Participants .................................................... 5
1.6 Opportunity Cost ........................................... 6 Application 1.2 Why the King Left Cleveland in 2010, and Can Benefits in Sports Be Measured? ...... 7 Economic versus Accounting Costs 7
Application 1.3 The Accounting and Economic Costs of SOX ...................................................... 8 Sunk Costs 8
Application 1.4 Why It Was Profitable to Demolish a Profitable Hong Kong Hotel ............... 9
1.7 Production Possibility Frontier .................... 9 Constant versus Increasing per-unit
Opportunity Costs 10
Chapter 2: Supply and Demand 13
2.1 Demand and Supply Curves ....................... 14 The Demand Curve 14
Application 2.1 The Law of Demand at Work for Non-work......................................................... 15 Determinants of Demand Other Than
Price 16
Shifts in versus Movements along a
Demand Curve 17
Application 2.2 The Rise and Fall of Cigarette Consumption in the United States ...................... 18
Application 2.3 The Occasional Interplay Between Price and Non-Price Factors in Determining Quantity Demanded .........................................18 The Supply Curve 19
Shifts in versus Movements along a Supply
Curve 20
2.2 Determination of Equilibrium Price and Quantity ........................................................ 21
2.3 Adjustment to Changes in Demand or Supply ........................................................... 22 Application 2.4 Why Holiday Home Prices in Switzerland are Soaring .................................... 23 Using the Supply–Demand Model to Explain
Market Outcomes 24
2.4 Government Intervention in Markets: Price Controls .............................................. 25 Rent Control 25
Who Loses, Who Benefits? 27
Black Markets 28
Application 2.5 Health Care Reform and Price Controls ................................................... 29 Application 2.6 Price Ceilings Can Be Deadly for Buyers ............................................. 30
2.5 Elasticities ..................................................... 30 Price Elasticity of Demand 30
Calculating Price Elasticity of Demand 32
Application 2.7 Demand Elasticity and Cable Television Pricing ..................................... 34 Demand Elasticities Vary among Goods 34
The Estimation of Demand Elasticities 35
Application 2.8 Why Canadians Are Flying South of the Border ...........................................36 Three Other Elasticities 36
Application 2.9 Price Elasticities of Supply and Demand and Short-Run Oil Price Gyrations ............................................ 38
Chapter 3: The Theory of Consumer Choice 42
3.1 Consumer Preferences ................................ 43 Consumer Preferences Graphed as
Indifference Curves 44
Curvature of Indifference Curves 45
Table of Contents xi
FM.INDD 07:39:38:PM 08/08/2014 PAGE XITrim Size: 203.2 mm X 254 mm
A Graphical Examination of a Tax-Plus-
Rebate Program 83
4.3 Income and Substitution Effects: Inferior Goods .............................................. 85 A Hypothetical Example of a Giffen Good 87
The Giffen Good Case: How Likely? 87
4.4 From Individual to Market Demand ......... 88 Application 4.3 Aggregating Demand Curves for ITO ............................................................. 88
4.5 Consumer Surplus ....................................... 89 The Uses of Consumer Surplus 92
Application 4.4 The Consumer Surplus Associated with Free TV .................................... 93 Consumer Surplus and Indifference
Curves 93
Application 4.5 The Benefits of Health Improvements .................................................. 94
4.6 Price Elasticity and the Price–Consumption Curve ............................................................... 95
4.7 Network Effects ............................................ 96 The Bandwagon Effect 97
The Snob Effect 98
Application 4.6 Network Effects and the Diffusion of Communications Technologies and Computer Hardware and Software .................................... 99
4.8 The Basics of Demand Estimation ........... 100 Experimentation 100
Surveys 101
Regression Analysis 101
Chapter 5: Using Consumer Choice Theory 107
5.1 Excise Subsidies, Health Care, and Consumer Welfare ..................................... 108 The Relative Effectiveness of a Lump-Sum
Transfer 109
Using the Consumer Surplus Approach 110
Application 5.1 The Price Sensitivity of Health Care Consumers .................................. 111
5.2 Subsidizing Health Insurance: ObamaCare ................................................ 112 The Basics of ObamaCare 112
The Subsidy’s Effect on the Budget Line 112
Bringing in Preferences 113
The Costs and Benefits of the Subsidy 114
Other Possible Outcomes 114
Can a Subsidy Harm the Recipient? 115
One Other Option 115
Application 3.1 Diminishing MRS and Newspaper Retailing ........................................ 48 Individuals Have Different Preferences 48
Application 3.2 Oreos in the Orient ..................... 48 Graphing Economic Bads and Economic
Neuters 49
Perfect Substitutes and Complements 51
3.2 The Budget Constraint ................................ 52 Geometry of the Budget Line 53
Shifts in Budget Lines 54
3.3 The Consumer’s Choice .............................. 56 A Corner Solution 57
The Composite-Good Convention 58
Application 3.3 Premium Fast Food: Why Chipotle Is One Hot Pepper of a Stock ................ 59
3.4 Changes in Income and Consumption Choices .......................................................... 60 Normal Goods 60
Inferior Goods 62
The Food Stamp Program 64
Application 3.4 The Allocation of Commencement Tickets ............................................................. 65
3.5 Are People Selfish? ...................................... 66 Application 3.5 Is Altruism a Normal Good? ........ 68
3.6 The Utility Approach to Consumer Choice ........................................................... 69 The Consumer’s Optimal Choice 70
Relationship to Indifference Curves 71
Chapter 4: Individual and Market Demand 75
4.1 Price Changes and Consumption Choices ... 76 The Consumer’s Demand Curve 77
Some Remarks about the Demand Curve 77
Application 4.1 Using Price to Deter Youth Alcohol Abuse, Traffic Fatalities, and Campus Violence .............................................. 78 Do Demand Curves Always Slope
Downward? 79
Application 4.2 Why the Flight to Poultry and away from Red Meat by U.S. Consumers? ........... 80
4.2 Income and Substitution Effects of a Price Change ................................................ 80 Income and Substitution Effects Illustrated:
The Normal-Good Case 81
The Income and Substitution Effects
Associated with a Gasoline Tax-Plus-
Rebate Program 83
xii Table of Contents
FM.INDD 07:39:38:PM 08/08/2014 PAGE XIITrim Size: 203.2 mm X 254 mm
5.3 Public Schools and the Voucher Proposal ...................................................... 116 Using Consumer Choice Theory to Analyze
Voucher Proposals 118
Application 5.2 The Demand for and Supply of School Choice ................................................ 119
5.4 Paying for Garbage ................................... 121 Does Everyone Benefit? 122
Application 5.3 Trash Pricing and Recycling ...... 123 5.5 The Consumer’s Choice to Save or
Borrow ........................................................ 124 A Change in Endowment 125
Application 5.4 Social Security and Saving ....... 127 Changes in the Interest Rate 128
The Case of a Higher Interest Rate Leading to
Less Saving 128
5.6 Investor Choice .......................................... 130 Application 5.5 Entrepreneurs and Their Risk–Return Preferences .................................. 131 Investor Preferences toward Risk: Risk
Aversion 131
Investor Preferences toward Risk: Risk
Neutral and Risk Loving 134
Application 5.6 Risk Aversion While Standing in Line ............................................................ 135 Minimizing Exposure to Risk 135
Chapter 6: Exchange, Efficiency, and Prices 140
6.1 Two-Person Exchange ............................... 141 The Edgeworth Exchange Box Diagram 142
The Edgeworth Exchange Box with
Indifference Curves 143
Application 6.1 The Benefits of Exchange and eBay ....................................................... 145
6.2 Efficiency in the Distribution of Goods .......................................................... 147 Application 6.2 Promoting Efficiency in Gift Card Giving .................................................... 149 Efficiency and Equity 149
6.3 Competitive Equilibrium and Efficient Distribution ................................................ 150 Application 6.3 Should Ticket Scalpers Be Disparaged or Deified? ................................... 153
6.4 Price and Nonprice Rationing and Efficiency .................................................... 154 Application 6.4 The Benefits and Costs of Rationing by Waiting ...................................... 155
Chapter 7: Production 159
7.1 Relating Output to Inputs ......................... 160 The Production Function 160
7.2 Production When Only One Input Is Variable: The Short Run ...................... 160 Total, Average, and Marginal Product
Curves 162
The Relationship between Average and
Marginal Product Curves 162
The Geometry of Product Curves 163
Application 7.1 What the Marginal–Average Relationship Means for Your Grade Point Average (GPA)................................................. 164 The Law of Diminishing Marginal
Returns 165
Application 7.2 The Law of Diminishing Marginal Returns, Caffeine Intake, and Exam Performance .................................................. 166
7.3 Production When All Inputs Are Variable: The Long Run ...................... 166 Production Isoquants 167
Four Characteristics of Isoquants 168
Marginal Rate of Technical Substitution
(MRTS) 168 MRTS and the Marginal Products of Inputs 169 Using MRTS: Speed Limits and Gasoline Consumption 170
7.4 Returns to Scale ......................................... 171 Factors Giving Rise to Increasing Returns 171
Factors Giving Rise to Decreasing Returns 173
Application 7.3 Returns to Scale and Cross- Country Trade Flows ....................................... 173
7.5 Functional Forms and Empirical Estimation of Production Functions ........ 174 Linear Forms of Production Functions 175
Multiplicative Forms of Production Functions:
Cobb–Douglas as an Example 176
Exponents and What They Indicate in
Cobb–Douglas Production Functions 177
Chapter 8: The Cost of Production 181
8.1 The Nature of Cost .................................... 182 8.2 Short-Run Cost of Production .................. 182
Measures of Short-Run Cost: Total Fixed and
Variable Costs 183
Fixed versus Sunk Costs 183
Table of Contents xiii
FM.INDD 07:39:38:PM 08/08/2014 PAGE XIIITrim Size: 203.2 mm X 254 mm
Five Other Measures of Short-Run Cost 184
Behind Cost Relationships 184
8.3 Short-Run Cost Curves ............................. 185 Marginal Cost 186
Average Cost 187
Application 8.1 The Effect of Walmart on Retailing Productivity, Costs, and Prices ............ 188 Marginal–Average Relationships 189
The Geometry of Cost Curves 189
8.4 Long-Run Cost of Production .................. 191 Isocost Lines 191
Least Costly Input Combinations 192
Interpreting the Tangency Points 192
Application 8.2 American Airlines and Cost Minimization .................................................. 194 The Expansion Path 194
Is Production Cost Minimized? 195
Application 8.3 Privatization and Productivity in China ...................................... 195
8.5 Input Price Changes and Cost Curves ..... 196 Application 8.4 The Economics of Raising and Razing Buildings ............................................. 197 Application 8.5 Applying the Golden Rule of Cost Minimization to the Baseball Diamond ..... 198
8.6 Long-Run Cost Curves .............................. 199 The Long Run and Short Run Revisited 200
8.7 Learning by Doing ..................................... 202 The Advantages of Learning by Doing to
Pioneering Firms 203
8.8 Importance of Cost Curves to Market Structure ..................................................... 203 Application 8.6 The Decline and Rise of Breweries in the United States ..........................205
8.9 Using Cost Curves: Controlling Pollution ...................................................... 205
8.10 Economies of Scope ................................... 208 8.11 Estimating Cost Functions ........................ 209
Chapter 9: Profit Maximization in Perfectly Competitive Markets 213
9.1 The Assumptions of Perfect Competition ................................................ 214 The Four Conditions Characterizing Perfect
Competition 214
9.2 Profit Maximization ................................... 215 Application 9.1 Aligning Managerial Actions with Shareholder Interests: Lessons from the Recession of 2007–2009 ............................... 216
9.3 The Demand Curve for a Competitive Firm ....................................... 217
9.4 Short-Run Profit Maximization ............... 218 Short-Run Profit Maximization Using per-Unit
Curves 220
Operating at a Loss in the Short Run 221
9.5 The Perfectly Competitive Firm’s Short-Run Supply Curve ............................................. 223 Output Response to a Change in Input
Prices 224
Application 9.2 Why Firms That Fatten Cattle are Seeing Their Own Numbers Thinned ................. 224
9.6 The Short-Run Industry Supply Curve ... 225 Price and Output Determination in the Short
Run 227
9.7 Long-Run Competitive Equilibrium ........ 227 Zero Economic Profit 228
Zero Profit When Firms’ Cost Curves Differ? 230
9.8 The Long-Run Industry Supply Curve ... 231 Constant-Cost Industry 231
Increasing-Cost Industry 233
Decreasing-Cost Industry 235
Application 9.3 The Bidding War for Business School Professors ........................................... 236 Comments on the Long-Run Supply Curve 237
Application 9.4 Cashing In on Corn .................. 238 9.9 When Does the Competitive Model
Apply? ......................................................... 239
Chapter 10: Using the Competitive Model 243
10.1 The Evaluation of Gains and Losses ........ 243 Producer Surplus 244
Consumer Surplus, Producer Surplus, and
Efficient Output 245
The Deadweight Loss of a Price Ceiling 247
10.2 Excise Taxation .......................................... 249 The Short-Run Effects of an Excise Tax 250
The Long-Run Effects of an Excise Tax 251
Who Bears the Burden of the Tax? 252
Tax Incidence: The Effect of Elasticity of
Supply 252
Tax Incidence: The Effect of Elasticity of
Demand 253
Application 10.1 Relative Ability to “Run” and Tax Incidence ................................................. 254 The Deadweight Loss of Excise
Taxation 254
xiv Table of Contents
FM.INDD 07:39:38:PM 08/08/2014 PAGE XIVTrim Size: 203.2 mm X 254 mm
Application 10.2 The Long and the Short (Run) of the Deadweight Loss of Rent Control ............ 256
10.3 Airline Regulation and Deregulation ....... 257 What Happened to the Profits? 258
After Deregulation 259
Application 10.3 The Contestability of Airline Markets ......................................................... 260 The Push for Reregulation 261
10.4 City Taxicab Markets ................................ 262 The Illegal Market 264
Application 10.4 Why New York City Cab Drivers Are Poor and Drive So Fast ................... 265
10.5 Consumer and Producer Surplus, and the Net Gains from Trade ......................... 266 The Gains from International Trade 268
The Link between Imports and Exports 269
Application 10.5 Protecting Steel Jobs Steals Jobs ............................................................... 270
10.6 Government Intervention in Markets: Quantity Controls ...................................... 270 Sugar Policy: A Sweet Deal 271
Application 10.6 Why Sugar Import Quotas Were Job Losers with Respect to LifeSavers ....... 273 Quotas and Their Foreign Producer
Consequences 273
Chapter 11: Monopoly 277
11.1 The Monopolist’s Demand and Marginal Revenue Curves ......................................... 278
11.2 Profit-Maximizing Output of a Monopoly .................................................... 280 Graphical Analysis 281
The Monopoly Price and Its Relationship to
Elasticity of Demand 282
Application 11.1 Demand Elasticity and Parking at Jack in the Box ................................ 285
11.3 Further Implications of Monopoly Analysis ....................................................... 285
11.4 The Measurement and Sources of Monopoly Power ........................................ 288 Measuring Monopoly Power 289
The Sources of Monopoly Power 289
Barriers to Entry 290
Application 11.2 The Effect of State Licensing on One of the World’s Oldest Professions ...........292 Strategic Behavior by Firms: Incumbents and
Potential Entrants 292
Application 11.3 March Monopoly Madness ....... 293
11.5 The Efficiency Effects of Monopoly .......... 295 A Dynamic View of Monopoly and Its
Efficiency Implications 296
Application 11.4 Static Versus Dynamic Perspectives on Monopoly Control of Government................................................... 298
11.6 Public Policy toward Monopoly ............... 298 Regulation of Price 299
Application 11.5 What Not to Say to a Rival on the Telephone .................................... 300 Application 11.6 Static versus Dynamic Views of Monopoly and the Microsoft Antitrust Case ...... 300 Application 11.7 Efficiency and the Regulation of Pharmaceutical Drug Prices in the European Union and the United States ........................... 303
Chapter 12: Product Pricing with Monopoly Power 306
12.1 Price Discrimination .................................. 307 First-Degree Price Discrimination 307
Implementing First-Degree Price
Discrimination 308
Second-Degree Price Discrimination 309
Third-Degree Price Discrimination 309
Application 12.1 Giving Frequent Shoppers the Second Degree ............................................... 310 Application 12.2 The Third Degree by Car Dealers .......................................................... 311
12.2 Three Necessary Conditions for Price Discrimination ............................................ 311 Application 12.3 Gray Hairs and Gray Markets .. 312
12.3 Price and Output Determination with Price Discrimination .................................. 313 Price Determination 313
Output Determination 314
Application 12.4 The Cost of Being Earnest When It Comes to Applying to Colleges ......................... 316
12.4 Intertemporal Price Discrimination and Peak-Load Pricing ..................................... 316 Application 12.5 Yield Management by Airlines .... 318 Peak-Load Pricing 319
The Advantages of Peak-Load Pricing 320
Application 12.6 Using Peak-Load Pricing to Combat Traffic Congestion .............................. 321
12.5 Two-Part Tariffs ........................................ 322 Many Consumers, Different Demands 323
Why the Price Will Usually Be Lower
Than the Monopoly Price 325
Table of Contents xv
FM.INDD 07:39:38:PM 08/08/2014 PAGE XVTrim Size: 203.2 mm X 254 mm
Application 12.7 The Costs of Engaging in Price Discrimination ........................................ 326
Chapter 13: Monopolistic Competition and Oligopoly 330
13.1 Price and Output under Monopolistic Competition ................................................ 331 Determination of Market Equilibrium 331
Monopolistic Competition and Efficiency 333
Is Government Intervention Warranted? 335
Application 13.1 Monopolistic Competition: The Eyes Have It (When It Comes to Refractive Surgery) ........................................... 335
13.2 Oligopoly and the Cournot Model ........... 336 The Cournot Model 337
Reaction Curves 338
Evaluation of the Cournot Model 339
13.3 Other Oligopoly Models ............................ 340 The Stackelberg Model 340
The Dominant Firm Model 342
The Elasticity of a Dominant Firm’s Demand
Curve 344
Application 13.2 The Dynamics of the Dominant Firm Model in Pharmaceutical Markets ......................................................... 345
13.4 Cartels and Collusion ................................ 345 Cartelization of a Competitive Industry 346
Application 13.3 Does the Internet Promote Competition or Cartelization? .......................... 347 Why Cartels Fail 347
Application 13.4 The Difficulty of Controlling Cheating ........................................................ 349 Application 13.5 The Rolex “Cartel” .................. 349 Oligopolies and Collusion 350
The Case of OPEC 351
The Reasons for OPEC’s Early Success 352
The Rest of the OPEC Story 352
Chapter 14: Game Theory and the Economics of Information 357
14.1 Game Theory ............................................. 358 Determination of Equilibrium 358
Application 14.1 Dominant Strategies in Baseball ........................................................ 359 Nash Equilibrium 360
14.2 The Prisoner’s Dilemma Game ................ 361
Application 14.2 The Congressional Prisoner’s Dilemma ......................................... 362 The Prisoner’s Dilemma and Cheating by
Cartel Members 363
A Prisoner’s Dilemma Game You May Play 365
14.3 Repeated Games ........................................ 366 Application 14.3 Cooperation in the Trenches of World War I .................................. 368 Do Oligopolistic Firms Always Collude? 368
Game Theory and Oligopoly: A Summary 369
14.4 Asymmetric Information .......................... 369 The “Lemons” Model 370
Market Responses to Asymmetric
Information 371
The Relevance of the Lemons Model 372
Is There a Lemons Problem in Used Car
Markets? 372
Application 14.4 Job Market Signaling ............... 373 14.5 Adverse Selection and Moral
Hazard ........................................................ 373 Adverse Selection 374
Market Responses to Adverse Selection 374
Application 14.5 Adverse Selection and the American Red Cross ........................................ 375 Moral Hazard 375
Application 14.6 Moral Hazard and Subprime Home Mortgages ............................. 376 Market Responses to Moral Hazard 377
14.6 Limited Price Information ........................ 377 Application 14.7 Moral Hazard on the Streets of New York City ................................. 377 The Effect of Search Intensity on Price
Dispersion 378
14.7 Advertising ................................................. 379 Advertising as Information 379
Advertising and Its Effects on Products’ Prices
and Qualities 380
Advertising, the Full Price of a Product, and
Market Efficiency 381
Application 14.8 The Effectiveness of Internet Advertising: The Case of Online Dating ............ 382
Chapter 15: Using Noncompetitive Market Models 385
15.1 The Size of the Deadweight Loss of Monopoly .................................................... 385 Why Are the Estimates of the Deadweight
Loss Not Large? 387
xvi Table of Contents
FM.INDD 07:39:38:PM 08/08/2014 PAGE XVITrim Size: 203.2 mm X 254 mm
Other Possible Deadweight Losses of
Monopoly 388
15.2 Do Monopolies Suppress Inventions? ...... 389 The Effect of Inventions on Output 390
Application 15.1 The Cost of Not Cannibalizing ............................................... 391
15.3 Natural Monopoly ..................................... 392 Regulation of Natural Monopoly: Theory 393
Regulation of Natural Monopoly: Practice 394
Application 15.2 Regulating Natural Monopoly through Public Ownership: The Case of USPS ......................................... 395
15.4 More on Game Theory: Iterated Dominance and Commitment ................... 396 Iterated Dominance 396
Commitment 398
Application 15.3 Why It May Be Wise to Burn Your Ships ...................................................... 399
Chapter 16: Employment and Pricing of Inputs 402
16.1 The Input Demand Curve of a Competitive Firm ............................................................ 403 The Firm’s Demand Curve: One Variable
Input 403
The Firm’s Demand Curve: All Inputs
Variable 405
The Firm’s Demand Curve: An Alternative
Approach 406
16.2 Industry and Market Demand Curves for an Input ................................................ 408 A Competitive Industry’s Demand Curve
for an Input 408
The Elasticity of an Industry’s Demand Curve
for an Input 409
Application 16.1 Explaining Sky-High Pilot Salaries under Airline Regulation .................................. 410 The Market Demand Curve for an Input 411
16.3 The Supply of Inputs ................................. 411 16.4 Industry Determination of Price and
Employment of Inputs ............................... 413 Process of Input Price Equalization across
Industries 414
16.5 Input Price Determination in a Multi- Industry Market ........................................ 416
16.6 Input Demand and Employment by an Output Market Monopoly ........................ 418
16.7 Monopsony in Input Markets ................... 420 Application 16.2 Major League Monopsony ....... 421
Chapter 17: Wages, Rent, Interest, and Profit 425
17.1 The Income–Leisure Choice of the Worker ........................................................ 425 Is This Model Plausible? 427
17.2 The Supply of Hours of Work .................. 428 Is a Backward-Bending Labor Supply Curve
Possible? 429
The Market Supply Curve 430
Application 17.1 An Example of a Backward- Bending Labor Supply Curve: The Work Effort Choices of Dentists Versus Physicians ............... 431 Application 17.2 Why Do Americans Work More Than Europeans? ........................... 432
17.3 The General Level of Wage Rates ............ 433 Application 17.3 The Malaise of the 1970s ........ 434
17.4 Why Wages Differ ...................................... 435 Compensating Wage Differentials 436
Differences in Human Capital Investment 437
Application 17.4 Twelve Hours’ Pay for Ten Minutes’ Work .......................................... 437 Application 17.5 The Returns to Investing in a BA and an MBA .............................................. 438 Differences in Ability 438
17.5 Economic Rent ........................................... 439 17.6 Monopoly Power in Input Markets: The
Case of Unions ............................................ 441 Application 17.6 The Decline and Rise of Unions and Their Impact on State and Local Government Budgets ...................................... 443 Some Alternative Views of Unions and an
Assessment of the Impact of Unions on Worker
Productivity 444
17.7 Borrowing, Lending, and the Interest Rate ............................................................. 444
17.8 Investment and the Marginal Productivity of Capital .................................................... 446 The Investment Demand Curve 447
17.9 Saving, Investment, and the Interest Rate ............................................................. 448 Equalization of Rates of Return 450
17.10 Why Interest Rates Differ ......................... 450
Chapter 18: Using Input Market Analysis 453
18.1 The Minimum Wage .................................. 454 Further Considerations 455
Does the Minimum Wage Harm the Poor? 456
Table of Contents xvii
FM.INDD 07:39:38:PM 08/08/2014 PAGE XVIITrim Size: 203.2 mm X 254 mm
The Minimum Wage: An Example of an
Efficiency Wage? 458
Application 18.1 The Disemployment Effect of the 1990–1991 Minimum Wage Hike ............. 459
18.2 Who Really Pays for Social Security? ..... 460 But Do Workers Bear All the Burden? 461
18.3 The Hidden Cost of Social Security ......... 463 The Rest of the Story 464
Application 18.2 Other Hidden Costs of PAYGO Social Security .................................... 465 The Effect on Labor Markets 466
18.4 The NCAA Cartel ...................................... 467 An Input Buyers’ Cartel 467
The NCAA as a Cartel of Buyers 469
Eliminate the Cartel Restrictions on Pay? 470
Application 18.3 The Differing Fortunes of College Athletes and Coaches .......................... 471
18.5 Discrimination in Employment ................ 472 What Causes Average Wage Rates to Differ? 474
18.6 The Benefits and Costs of Immigration ... 476 More on Gains and Losses 478
Chapter 19: General Equilibrium Analysis and Economic Efficiency 482
19.1 Partial and General Equilibrium Analysis Compared ................................... 483 The Mutual Interdependence of Markets
Illustrated 483
When Should General Equilibrium
Analysis Be Used? 485
19.2 Economic Efficiency ................................... 486 Efficiency as a Goal for Economic
Performance 487
19.3 Conditions for Economic Efficiency ......... 488 19.4 Efficiency in Production ............................ 489
The Edgeworth Production Box 489
The Production Contract Curve and Efficiency
in Production 491
General Equilibrium in Competitive Input
Markets 491
19.5 The Production Possibility Frontier and Efficiency in Output ................................... 492 Efficiency in Output 494
An Economy’s PPF and the Gains from
International Trade 495
19.6 Competitive Markets and Economic Efficiency .................................................... 497
The Role of Information 498
Application 19.1 Can Centralized Planning Promote Efficiency? ........................................ 499
19.7 The Causes of Economic Inefficiency ....... 500 Market Power 500
Application 19.2 How Government Prolonged the Great Depression ....................... 501 Imperfect Information 502
Application 19.3 Deterring Cigarette Smoking .... 502 Externalities/Public Goods 503
Chapter 20: Public Goods and Externalities 505
20.1 What Are Public Goods? .......................... 506 The Free-Rider Problem 507
Application 20.1 An Online Horror Tale ............ 507 Free Riding and Group Size 507
20.2 Efficiency in the Provision of a Public Good ............................................................ 508 Efficiency in Production and Distribution 510
Application 20.2 The Lowdown on Why Lojack Installations Are Lower Than the Efficient Output .............................................. 510 Patents 511
20.3 Externalities ............................................... 512 Externalities and Efficiency 512
External Costs 513
Application 20.3 Traffic Externalities: Their Causes and Some Potential Cures .................... 514 Application 20.4 Liability Caps and the British Petroleum Gulf Oil Disaster .............. 515 External Benefits 516
Application 20.5 Should Cell Phone Use While Driving Be Banned? ......................... 517
20.4 Externalities and Property Rights ........... 518 The Coase Theorem 519
Application 20.6 Making Telemarketers Pay .......................................... 520
20.5 Controlling Pollution, Revisited ............... 523 The Market for Los Angeles Smog 523
Answers to Selected Problems ..................... 528
Glossary .................................................................. 536
Index ........................................................................ 543
xviii
FM.INDD 07:39:38:PM 08/08/2014 PAGE XVIIITrim Size: 203.2 mm X 254 mm
5.3 Trash Pricing and Recycling 5.4 Social Security and Saving 5.5 Entrepreneurs and Their Risk–Return
Preferences
5.6 Risk Aversion While Standing in Line
6.1 The Benefits of Exchange and eBay 6.2 Promoting Efficiency in Gift Card Giving 6.3 Should Ticket Scalpers Be Disparaged or
Deified?
6.4 The Benefits and Costs of Rationing by Waiting
7.1 What the Marginal–Average Relationship Means for Your Grade Point Average (GPA)
7.2 The Law of Diminishing Marginal Returns, Caffeine Intake, and Exam Performance
7.3 Returns to Scale and Cross-Country Trade Flows
8.1 The Effect of Walmart on Retailing Productivity, Costs, and Prices
8.2 American Airlines and Cost Minimization 8.3 Privatization and Productivity in China 8.4 The Economics of Raising and Razing
Buildings
8.5 Applying the Golden Rule of Cost Minimization to the Baseball Diamond
8.6 The Decline and Rise of Breweries in the U.S.
9.1 Aligning Managerial Incentives with Shareholder Interests: Lessons from the
Recession of 2007–2009
9.2 Why Firms That Fatten Cattle are Seeing Their Own Numbers Thinned
9.3 The Bidding War for Business School Professors
9.4 Cashing In on Corn
10.1 Relative Ability to “Run” and Tax Incidence 10.2 The Long and the Short (Run) of the
Deadweight Loss of Rent Control
10.3 The Contestability of Airline Markets 10.4 Why New York City Cab Drivers Are Poor
and Drive So Fast
1.1 Real Versus Nominal Presidential Salaries 1.2 Why the King Left Cleveland, and Can
Benefits in Sports Be Measured?
1.3 The Accounting and Economic Costs of SOX 1.4 Why It Was Profitable to Demolish a
Profitable Hong Kong Hotel
2.1 The Law of Demand at Work for Non-work 2.2 The Rise and Fall of Cigarette Consumption
in the United States
2.3 The Occasional Interplay Between Price and Non-Price Factors in Determining Quantity
Demanded
2.4 Why Holiday Home Prices in Switzerland are Soaring
2.5 Health Care Reform and Price Controls 2.6 Price Ceilings Can Be Deadly for Buyers 2.7 Demand Elasticity and Cable Television
Pricing
2.8 Why Canadians Are Flying South of the Border 2.9 Price Elasticities of Supply and Demand and
Short-Run Oil Price Gyrations
3.1 Diminishing MRS and Newspaper Retailing 3.2 Oreos in the Orient 3.3 Premium Fast Food: Why Chipotle Is One
Hot Pepper of a Stock
3.4 The Allocation of Commencement Tickets 3.5 Is Altruism a Normal Good?
4.1 Using Price to Deter Youth Alcohol Abuse, Traffic Fatalities, and Campus Violence
4.2 Why the Flight to Poultry and away from Red Meat by U.S. Consumers?
4.3 Aggregating Demand Curves for ITO 4.4 The Consumer Surplus Associated with
Free TV
4.5 The Benefits of Health Improvements 4.6 Network Effects and the Diffusion of
Communications Technologies and Computer
Hardware and Software
5.1 The Price Sensitivity of Health Care Consumers 5.2 The Demand for and Supply of School Choice
Applications
Appl icat ions xix
FM.INDD 07:39:38:PM 08/08/2014 PAGE XIXTrim Size: 203.2 mm X 254 mm
10.5 Protecting Steel Jobs Steals Jobs 10.6 Why Sugar Import Quotas Were Job Losers
with Respect to LifeSavers
11.1 Demand Elasticity and Parking at Jack in the Box
11.2 The Effect of State Licensing on One of the World’s Oldest Professions
11.3 March Monopoly Madness 11.4 Static Versus Dynamic Perspectives on
Monopoly Control of Government
11.5 What Not to Say to a Rival on the Telephone 11.6 Static versus Dynamic Views of Monopoly
and the Microsoft Antitrust Case
11.7 Efficiency and the Regulation of Pharmaceutical Drug Prices in the European
Union and the United States
12.1 Giving Frequent Shoppers the Second Degree
12.2 The Third Degree by Car Dealers 12.3 Gray Hairs and Gray Markets 12.4 The Cost of Being Earnest When It Comes to
Applying to Colleges
12.5 Yield Management by Airlines 12.6 Using Peak-Load Pricing to Combat Traffic
Congestion
12.7 The Costs of Engaging in Price Discrimination
13.1 Monopolistic Competition: The Eyes Have It (When It Comes to Refractive Surgery)
13.2 The Dynamics of the Dominant Firm Model in Pharmaceutical Markets
13.3 Does the Internet Promote Competition or Cartelization?
13.4 The Difficulty of Controlling Cheating 13.5 The Rolex “Cartel”
14.1 Dominant Strategies in Baseball 14.2 The Congressional Prisoner’s Dilemma 14.3 Cooperation in the Trenches of World War I 14.4 Job Market Signaling 14.5 Adverse Selection and the American
Red Cross
14.6 Moral Hazard and Subprime Home Mortgages
14.7 Moral Hazard on the Streets of New York City 14.8 The Effectiveness of Internet Advertising:
The Case of Online Dating
15.1 The Cost of Not Cannibalizing 15.2 Regulating Natural Monopoly through Public
Ownership: The Case of the USPS
15.3 Why It May Be Wise to Burn Your Ships
16.1 Explaining Sky-High Pilot Salaries under Airline Regulation
16.2 Major League Monopsony
17.1 An Example of a Backward-Bending Labor Supply Curve: The Work Effort Choices of
Dentists versus Physicians
17.2 Why Do Americans Work More Than Europeans?
17.3 The Malaise of the 1970s 17.4 Twelve Hours’ Pay for Ten Minutes’ Work 17.5 The Returns to Investing in a BA and an MBA 17.6 The Decline and Rise of Unions and Their
Impact on State and Local Government
Budgets
18.1 The Disemployment Effect of the 1990–1991 Minimum Wage Hike
18.2 Other Hidden Costs of PAYGO Social Security
18.3 The Differing Fortunes of College Athletes and Coaches
19.1 Can Centralized Planning Promote Efficiency? 19.2 How Government Prolonged the Great
Depression
19.3 Deterring Cigarette Smoking
20.1 An Online Horror Tale 20.2 The Lowdown on Why Lojack Installations
Are Lower Than the Efficient Output
20.3 Traffic Externalities: Their Causes and Some Potential Cures
20.4 Liability Caps and the British Petroleum Gulf Oil Disaster
20.5 Should Cell Phone Use While Driving Be Banned?
20.6 Making Telemarketers Pay
1
C01.INDD 10:43:41:AM 08/06/2014 PAGE 1Trim Size: 203.2 mm X 254 mm
Why have health-care costs been rising? Will policies intended to provide universal health care coverage brake such cost increases in the future? If the government requires
employers to provide Social Security/pension support and health care for employees, who
bears the cost of such a mandate? How did mortgage-backed securities and incentive sys-
tems contribute to the recent financial market meltdown? Will financial market regulations
passed by the U.S. government mitigate the likelihood of such downturns in the future?
Why are Americans getting fatter? Will issuing firms tradable permits to pollute be an
effective way to deal with global warming? Should apparent monopolies such as Micro-
soft, Google, and Apple be praised for their efficiency and profitability, or should they
be subject to antitrust prosecution and broken up? What can be done to prevent future
oil spill disasters such as the one involving British Petroleum in the Gulf of Mexico in
2010? Does government-provided unemployment insurance increase or decrease unem-
ployment? Does the Internet promote competition or cartelization? Why do dry cleaners
charge more to launder women’s blouses than men’s shirts? Should the minimum wage in
the United States be raised from its present $7.25 per hour level? Are bans on cell phone
calls by drivers warranted?
As these questions suggest, there are many interesting issues that microeconomic theory
can help us understand. This text presents the analytical techniques of microeconomics and
shows how to apply them to explain and predict real-world phenomena.
CHAPTER
Learning Objectives
Convey the scope of microeconomic theory. Explain why theory, is essential to understanding and predicting real-world outcomes. Distinguish between positive and normative analyses. Differentiate between real and nominal prices. Describe the basic assumptions economists make about market participants. Introduce the concept of opportunity cost and explain how economic costs differ from accounting costs. Show how a production possibility frontier graphically depicts the basic assumptions economists make about market actors as well as the concept of opportunity cost.
Memorable Quote “Don’t measure yourself by what you have accomplished, but rather by what you should have accomplished with your ability.”
—John Wooden, UCLA basketball coaching legend with an opportunity-cost-based perspective
1 An Introduction to Microeconomics
2 An Introduct ion to Microeconomics
C01.INDD 10:43:41:AM 08/06/2014 PAGE 2Trim Size: 203.2 mm X 254 mm
This chapter introduces microeconomic theory by first discussing its nature and the role
of theory in general. The remainder of the chapter covers the basic assumptions economists
make about market participants and the key concept of opportunity cost.
1.1 The Scope of Microeconomic Theory The prefix micro- in microeconomics comes from the Greek word mikros, meaning small. It contrasts with macroeconomics, the other branch of economic theory. Macroeconomics deals primarily with aggregates, such as the total amount of goods and services produced
by society and the absolute level of prices, while microeconomics analyzes the behavior of “small” units: consumers, workers, savers, business managers, firms, individual industries
and markets, and so on. Microeconomics, however, is not limited to “small” issues. Instead,
it reflects the fact that many “big” issues can best be understood by recognizing that they
are composed of numerous smaller parts. Just as much of our knowledge of chemistry and
physics is built on the study of molecules, atoms, and subatomic particles, much of our
knowledge of economics is based on the study of individual behavior.
Individuals are the fundamental decision makers in any society. Their decisions, in
aggregate, define a society’s economic environment. Consumers decide how much of vari-
ous goods to purchase, workers decide what jobs to take, and business owners decide how
many workers to hire and how much output to produce. Microeconomics encompasses the
factors that influence these choices and the way these innumerable small decisions merge to
determine the workings of the entire economy. Because prices have important effects on
these individual decisions, microeconomics is frequently called price theory.
1.2 The Nature and Role of Theory In disciplines from physics to political science, using a theory to make sense of a complex
reality is essential. Facts do not always “speak for themselves.” In economics, facts may
describe a historical episode, but facts can never explain why the episode occurred or how
things would have been different had, for example, the government pursued another policy.
Moreover, facts can never demonstrate how, for instance, a change in agricultural price
supports will affect agricultural production next year. For purposes of explanation or pre-
diction, we must employ a theory that shows how facts are related to one another.
Theory in economics, as in other sciences, is based on certain assumptions. For exam-
ple, economists assume that firms strive to maximize profit. Based on this assumption, the
economic theory of the firm explains what mix of steel and plastic firms such as Toyota
and General Motors (GM) employ in production as well as how many cars and trucks they
produce. Theory also explains how Toyota’s and GM’s desired input mixes and final output
levels are affected by changes in, say, the price of steel or the price received per car sold.
Economic theory can be used to predict as well as to explain real-world outcomes. For
instance, the basic supply–demand model (discussed in Chapter 2) can explain the effects
observed in cities that have enacted rent control laws. It can also predict the effects should
the federal government impose similar price ceilings on health-care services.
What Is a Good Theory? How do we know if a theory, whether it be in economics, physics, or political science, is
a “good” theory? Basically, a theory is considered to be valid and useful if it successfully explains and predicts the phenomena that it is intended to explain and predict. In keeping
macroeconomics the study of aggregate economic factors
microeconomics the study of the behavior of small economic units such as consumers and firms
price theory another term for microeconomics
Posit ive versus Normative Analys is 3
C01.INDD 10:43:41:AM 08/06/2014 PAGE 3Trim Size: 203.2 mm X 254 mm
with this litmus test, theories are continually stacked up against real-world data. Depending
on how well a theory matches the data, the theory is maintained, refined, or sometimes even
discarded (perhaps in favor of a competing explanation). The continual process of testing the-
ories against real-world data is critical to the advancement of any science, not just economics.
In testing a theory, it is important to note that imperfection tends to be the norm. That
is, “good” theories typically do not explain the observed data perfectly, nor are the assump-
tions on which they are based entirely realistic. For example, consider the calorie theory, one accepted by millions of people. The calorie theory holds that a person’s weight depends
on the number of calories consumed per day: the more calories ingested, the heavier the
person will be.
The calorie theory predicts that to lose weight, a person should cut his or her calorie
intake. Is this a valid and useful theory? Consider two criticisms: first, the calorie theory
is based on assumptions that are not completely realistic. That is, no one has ever seen a
calorie, much less observed the human body convert it into weight. Second, the theory is
not perfect. Reducing your calorie intake will not necessarily make you thin. Other factors,
besides calories, influence a person’s weight: heredity, exercise, metabolism, ratio of fat to
protein consumption, and so on.
Does this mean that people who count calories are wrong? Not at all. In fact, the calorie
theory is quite useful for millions of weight watchers around the world. For them, the gen-
eral relationship between calories and weight tends to hold and becomes even stronger once
the calorie theory is refined to account for other factors such as heredity, exercise, metabo-
lism, and so forth.
Such is the case with economics. While firms may not appear to maximize profit (think
about Amazon.com or Biogen), and refinements accounting for special features of particu-
lar markets may be necessary (long-run versus short-run profitability in industries where
firms must make substantial up-front research and development investments), the economic
theory of the firm based on the assumption of profit maximization successfully explains and
predicts a wide range of real-world phenomena. Thus the theory is useful to both business
managers and public policymakers.
1.3 Positive versus Normative Analysis Economic theory is a tool for understanding relationships in the economy. While it can
explain the behavior of market actors, it cannot determine which public policies are desir-
able and which are not. Economics can help us evaluate the results of public policies, but it
never, by itself, demonstrates whether the results are good or bad.
Consider the federal minimum wage—first set in 1938 at $0.25 per hour and periodically
increased over the years (to $7.25 per hour by July 2009). Evaluating the desirability of this
policy requires three steps. First, one must determine the qualitative effects of the policy.
For example, how does it affect the employment of workers by firms? Does it increase
or decrease employment? Second, one must determine the magnitude of the effects. If the
minimum wage leads to less employment, how much less? How many workers lose their
jobs and how many retain their jobs at the higher wage rate? Finally, a judgment needs to
be made as to whether the policy’s effects are desirable. Does the benefit to workers who
remain employed outweigh the costs to those workers whose jobs are cut?
The first step involves identifying the qualitative nature of a policy’s consequences. This
step is in the realm of positive analysis, assessing the expected, objective outcomes. The distinguishing feature of positive analysis is that it deals with propositions that can be tested
with respect to both their underlying logic and the empirical evidence. It deals with what is,
or what might be, without deciding whether something is right or wrong, good or bad. Posi-
tive analysis is scientific because it draws on accepted rules of logic and evidence, of both a
positive analysis assessment of expected, objective outcomes
4 An Introduct ion to Microeconomics
C01.INDD 10:43:41:AM 08/06/2014 PAGE 4Trim Size: 203.2 mm X 254 mm
qualitative and quantitative nature, that can be used to determine the truth or falsity of state-
ments. Microeconomic theory is a form of positive analysis; it can be used, for example, to
make the qualitative prediction that a minimum wage law will reduce employment.
If we want to resolve the question of desirability, however, identifying the qualitative
nature of the effects is not sufficient. We also need some idea of the size of the effects.
It may matter a great deal whether the minimum wage causes 1 percent or 25 percent of
unskilled workers to lose their jobs. Note that this step still involves positive analysis, but
in quantitative rather than qualitative terms.
Knowing the consequences, both qualitative and quantitative, is still not sufficient to
determine whether a policy is desirable. A final step is necessary: we must decide
whether the consequences themselves are, on balance, desirable. To make this evalua-
tion, each person must make a normative analysis, or value judgment. By nature, such a judgment is nonscientific. It cannot be proved right or wrong by facts, evidence, or logic.
It stems from the value system of the person making the judgment. For example, a belief
that it is desirable to raise the wages of the lowest-paid workers, even at the expense of
others, falls into this category. People may agree that a particular policy has this effect,
but some may hold that the outcome is desirable and others that it is not. Their value
judgments differ.
Microeconomic theory cannot demonstrate that a particular set of economic institutions
or policies is desirable—and neither, for that matter, can any other scientific branch of
knowledge. A belief that something is desirable requires a nonscientific judgment of what
constitutes desirability, and that value judgment is the domain of normative analysis. None- theless, microeconomic theory can assist each of us in reaching such normative judgments
by helping us determine the likely outcomes. In other words, microeconomics helps us take
the first two of the three steps necessary to evaluate real-world phenomena.
1.4 Market Analysis and Real versus Nominal Prices Most of microeconomics involves the study of how individual markets function.
Markets involve the interplay of all potential buyers and sellers of a particular commod- ity or service. Most economic issues concern the way particular markets function. For
example, an economist’s wages are likely to be higher than those of a gas station atten-
dant but lower than those of a doctor. This situation reflects the workings of the three
labor markets.
To analyze markets, we concentrate on factors having the greatest influence on the deci-
sions of buyers and sellers. Prices receive special attention. Prices result from market trans-
actions, but they also strongly influence the behavior of buyers and sellers in every market.
In microeconomics, the term price always refers to the relative or real price of an item. The nominal price, or absolute price, by itself does not tell us how costly an item really is. Is a 10-cent cup of coffee expensive? In 1900 it would have been outrageously expensive;
today it would be a bargain. The problem with nominal prices is that a dollar is an elastic
yardstick. The real price of a good reflects its nominal price adjusted for the changing value of money. Table 1.1 clarifies the distinction between real and nominal prices. Between 1983 and 2013, the price level, or average price of goods and services, rose by 135 percent
according to the consumer price index (CPI). This can be determined by the facts that the
CPI for all items was 235 in 2013 and 100 in the base year, 1983, so it rose by:
( ) /235 100 100 135− = percent.
The CPI measures the change in nominal prices. Table 1.1 indicates that the nominal prices
of some goods, such as college tuition, rose by much more than the average 135 percent,
and the prices of others, like telephone services, rose less.
normative analysis a nonscientific value judgment
markets the interplay of all potential buyers and sellers of a particular commodity or service
nominal price the absolute price, not adjusted for the changing value of money
real price the nominal price adjusted for the changing value of money
Basic Assumptions about Market Part ic ipants 5
C01.INDD 10:43:41:AM 08/06/2014 PAGE 5Trim Size: 203.2 mm X 254 mm
The last column in Table 1.1 lists the change in each item’s price compared with the
change in the average of all prices. Although the nominal price of gasoline rose by 210 per- cent, the overall price level rose by 135 percent over the same period, so the real price of gasoline rose by only 32 percent:
( ) / .310 235 235 32− = percent
No matter how the nominal price changed between 1983 and 2013, an economist would say
that the prices of the first five individual items rose while the prices of the last three fell. The
term price always refers to a real price. The prices we use in discussion and in various dia- grams refer to real prices, unless otherwise noted. But these prices are generally measured in
dollar units. This practice is legitimate as long as we are using dollars of constant purchasing power—which is the same as measuring each price in comparison with the general price level.
APPLICATION 1.1
President Barack Obama was paid a government salary of $400,000 in 2013, or 16 times what President Abra- ham Lincoln earned 150 years prior. Yet, expressed in con- stant dollars, Lincoln’s salary of $25,000 in 1863 equates to
Real Versus Nominal Presidential Salaries
$652,778 in constant 2013 dollars (the CPI was 9 for 1863 versus 235 in 2013). While the salary paid to U.S. presidents has grown markedly in nominal terms over the last century and a half, it has declined in real terms.
1.5 Basic Assumptions about Market Participants Economists make three basic assumptions about buyers and sellers. Let us address these in
turn: goal orientation, rationality, and scarcity. First, market participants are presumed to be goal oriented—that is, interested in fulfilling their own personal goals. For example, the Sultan of Brunei may desire an opulent personal jet and advanced medical care for his
country’s people. Maverick entrepreneur Richard Branson has longed to circumnavigate
the globe in a hot air balloon while launching and growing successful ventures such as
Virgin Records and Virgin Atlantic Airways. The late film star Marilyn Monroe hoped for
ever greater success on the screen and stage, an Academy Award, and children of her own.
goal-oriented behavior the behavior of market participants interested in fulfilling their own personal goals
Nominal and Real Price Changes, 1983–2013 Index of Nominal Prices in 2013 (1983 = 100)
Change in Real Prices, 1983 to 2013
All items 235 —
Tobacco and smoking products 869 +270% College tuition 636 +171% Medical care 423 +80% Gasoline 310 +32% Residential rent 267 +14% Women’s and girls’ apparel 113 −52% Telephone services 89 −62% Personal computers and peripheral services 28 −75%
Source: U.S. Department of Labor, CPI Detailed Report, May 2013.
Table 1.1
6 An Introduct ion to Microeconomics
C01.INDD 10:43:41:AM 08/06/2014 PAGE 6Trim Size: 203.2 mm X 254 mm
The assumption of goal-oriented behavior often is taken to indicate that individuals are
self-interested. This assumption, however, does not imply that market participants care
solely about their own pocketbooks. As economists use this term, the behavior of Nobel
Peace Prize winner Mother Teresa could accurately be described as goal oriented. Although
Mother Teresa’s actions clearly indicated that she had little interest in worldly possessions,
they did reflect her own personal desire to help the poor of Calcutta. The assumption of
goal-oriented behavior does not rule out altruistic goals.
The second assumption economists make about market participants is that they engage
in rational behavior. For example, we presume that Toyota’s decision to build a factory in the United States is the outcome of a careful, deliberative process that weighs the expected
benefits and costs. We presume an individual buys a new home based on knowledge of its
market value and an honest appraisal of what he or she can afford.
The third, and most important, assumption made by economists about market partici-
pants is that they confront scarce resources. For example, there is simply not enough time, money, or other resources for the typical consumer to satisfy all of his or her desires.
Human beings have relatively limitless desires, and no matter how wealthy they become,
resources will never be plentiful enough to ensure that all their desires can be fulfilled.
If individuals rationally pursue their goals but have limited resources, choices must be
made. Specifically, one must decide which goal to pursue and how far to pursue it. Micro-
economics explores this process of making choices subject to resource constraints.
1.6 Opportunity Cost Whenever you pursue one goal, you limit the extent to which your other goals can be satis-
fied with your scarce resources. For example, suppose that after getting your bachelor’s
degree and working for a few years, you enroll in a full-time, two-year MBA program.
What would the cost of this choice be? You would incur some explicit costs, such as tuition, books, and parking. The dollars spent on such items could have been devoted to the
pursuit of other goals. You would also face implicit costs associated with your own use of time and other resources in the pursuit of a particular activity versus alternatives. For exam-
ple, instead of going to business school, you could have continued working and making
$40,000 per year. The $40,000 in annual forgone wages would be an implicit cost associ-
ated with pursuing an MBA. In other words, the time and effort devoted to pursuing the
MBA instead could have been used to generate $40,000 in each of the two years that you
attended graduate business school.
To understand why implicit costs matter, assume that, relative to the option of remaining
at work, the MBA entails explicit costs (such as tuition) of $70,000 and will increase your
postgraduate lifetime earnings by $60,000. In this case, you likely would not leave your job
to pursue the MBA. The $60,000 increase in postgraduate earnings would be outweighed
by the combined $70,000 in explicit costs and $80,000 in implicit costs of two years’ lost
wages.
The concepts of explicit and implicit costs also apply to the production side of a market.
For firms making production decisions, explicit costs are those that are usually counted as
costs in conventional accounting statements. They include payroll, raw materials, insur-
ance, electricity, interest on debt, and so on. Implicit costs reflect the fact that a firm’s
resources can be allocated to other uses—Time Warner, for example, can reallocate its
resources from magazine publishing to the production of interactive video products.
The sum of the explicit and implicit costs associated with using some resource in a particu-
lar way is defined to be the resource’s economic cost or opportunity cost. The concept of opportunity cost forces us to recognize that costs are not just money payments but also sacri-
ficed alternatives. Where more than two uses for a resource exist and the resource can be
rational behavior the behavior of market participants based on a careful, deliberative process that weighs expected benefits and costs
scarce resources insufficient time, money, or other resources for individuals to satisfy all their desires
explicit costs money used in the pursuit of a goal that could otherwise have been spent on an alternative objective
implicit costs costs associated with an individual’s use of his or her own time and other resources in the pursuit of a particular activity versus alternatives
economic cost or opportunity cost the sum of explicit and implicit costs
Opportunity Cost 7
C01.INDD 10:43:41:AM 08/06/2014 PAGE 7Trim Size: 203.2 mm X 254 mm
devoted to only one use at a time, the opportunity cost of using the resource in a particular way
is the value of the resource in its best alternative use. So, if your options are business school,
continuing to work in your current job for $40,000 per year, and switching to a similarly
demanding job that only pays $30,000 per year, you would take into account only the implicit
cost of giving up your current job in determining the opportunity cost of pursuing the MBA.
APPLICATION 1.2
In 2010, basketball star LeBron James (dubbed “the King”) had to decide between staying with the Cleveland Cavaliers in Ohio and switching to one of several other teams bidding for his services. LeBron’s decision to join the Miami Heat required an analysis similar to the one we have just spelled out for enrolling in a full-time MBA program. LeBron had to take into account the explicit costs involved with playing for a particular team as well as the implicit costs associated with forgone alternatives.
Not all the costs are obvious. For example, while the Cavaliers offered a more generous salary package staying in Ohio would have also involved paying greater state and local income taxes on salary as well as endorsement income. Indeed, Richard Vedder, an economist at Ohio University, estimated that the net present value of income tax savings just on salary to LeBron from living in Miami as opposed to his home town of Akron, Ohio, was $6–8 million. The com- bined state and local income tax rate in Akron is 7 percent versus 0 percent for Miami, Florida. The rates are even high in New York City (12.85 percent) and New Jersey (9 percent)— the locations of two other teams that were actively in the running for LeBron’s services.
Estimating the explicit benefits to LeBron associated with various teams also is no easy matter. How can one put a value, for example, on playing with two particular friends (Dwayne Wade and Chris Bosh) who were already stars on the Miami Heat squad? How about calculating the ben- efit associated with a greater likelihood of winning the NBA championship through playing with the Heat versus the Cavaliers?
Economics is predicated on the assumption that people take such benefits and costs into account when making decisions. And, while costs and benefits may be difficult
Why the King Left Cleveland in 2010, and Can Benefits in Sports Be Measured?
to measure, and can change over time (note that LeBron opted to return to Cleveland in 2014 an account of the value he placed on being back home), economics presumes that individuals strive to place values on these costs and benefits in seeking to make rational choices.
For another example from the sports world, consider whether public funds should be devoted to building a new stadium. This was the case faced by Minnesota officials when their professional football team, the Vikings, threat- ened to leave without a partial public subsidy for a new $870 million stadium to replace the existing Metrodome in downtown Minneapolis. In addition to more direct ben- efits that a sports franchise provides a locality, policymakers had to estimate whether the stadium subsidy could at least partly be justified by the civic pride and purpose generated by the Vikings for Minnesotans.
Economists Aju Fenn and John Crooker estimated that, as of 2002, the average Minnesotan derived a benefit on the order of $530 annually from the joy and pride of hav- ing the Vikings in Minnesota.1 Although some other econo- mists question the estimates due to the extent to which suryvey-to-based evidence accurately reflects individual actual willingness to spend money, the Fenn and Crooker study does indicate that while the benefits of fandom in sports are not priceless, they are also nonzero and need to be incorporated into decisions such as whether to provide public funds for a new local stadium.
1Aju J. Fenn and John R. Crooker, “Estimating Local Welfare Gener- ated by an NFL Team Under Credible Threat of Relocation,” Southern Economic Journal, 76, No. 1 (July 2009), pp. 198–223.
Economic versus Accounting Costs Because opportunity costs are not always readily apparent (especially their implicit compo-
nents), they often are not accurately reflected in companies’ net income statements. For
example, consider a family-run grocery store in downtown Tokyo whose owners acquired
the property several generations ago for almost nothing. From an accounting perspective,
the grocery store may appear to be generating positive net income: revenue exceeds the sum
of accounting costs comprising payroll, electricity, insurance, wholesale grocery costs, and so on. Still, the grocery store may be losing money from an economic perspective once the
accounting costs costs reported in companies’ net income statements generated by accountants
8 An Introduct ion to Microeconomics
C01.INDD 10:43:41:AM 08/06/2014 PAGE 8Trim Size: 203.2 mm X 254 mm
opportunity cost of the land on which it sits is taken into account. That is, the land could be
sold or rented to someone else. This choice will generate payments that are sacrificed when
the family uses the land to run a grocery. These forgone earnings represent an opportunity
cost—and this cost can be significant. For example, the value of just the Imperial Palace
grounds situated in the heart of Tokyo has been estimated to exceed, in certain years, the
total value of real estate in the state of California.
APPLICATION 1.3
In the wake of several high-profile corporate scandals such as Enron, Tyco, Arthur Andersen, and Adelphia, the U.S. Congress passed the Sarbanes-Oxley Act (SOX) in 2002. Signed into law soon thereafter by President Bush, SOX was intended to significantly enhance corporate governance by changing the rigor with which publicly traded companies reported their finances, communicated with shareholders, and governed themselves.
While well intentioned, the accounting costs to corpo- rations from complying with SOX have been substantial. Based on surveys of publicly traded firms by Korn/Ferry, Foley & Larder, and A.R.C. Morgan, the annual accounting costs of complying with SOX appear to range from $11 bil- lion to $26 billion.
These accounting costs, however, pale in comparison to the economic costs of SOX, which have been estimated to be as high as $140 billion annually.2 Among these harder- to-measure but much more substantial costs are the fact that SOX diverts the attention of senior management from
The Accounting and Economic Costs of SOX
doing business. As the chief accounting officer of General Motors has noted: “The real cost isn’t the incremental dollars, it is having people that should be focused on the business focused instead on complying with the details of the [SOX] rules.” SOX also exposes executives to greater litigation risks and stiffer penalties. As a result, CEOs are less likely to take riskier, entrepreneurial actions, consequently changing their business strategies and potentially reducing the value of their firms—and altering the future of the U.S. economy.
To gauge the economic costs associated with SOX, Pro- fessor Ivy Zhang of the University of Minnesota employed what is known as an “event-study” analysis focusing on the legislative events or “news” leading to the passage of SOX in July 2002. The analysis examined broad movements in the stock market and used well-known statistical tech- niques to isolate the impact of a particular factor—in this case, the Sarbanes-Oxley legislation. The analysis indicated that, holding constant other factors, the cumulative abnor- mal return of the stock market stemming from the legisla- tive events leading to the passage of SOX was significantly negative and translated into a loss of over 12 percent ($1.4 trillion) of the total market value ($11.3 trillion) of publicly traded firms in the United States.
2Ivy Zhang, “Economic Consequences of the Sarbanes-Oxley Act of 2002,” Journal of Accounting and Economics, 44, Nos. 1–2 (Sep- tember 2007), pp. 74–115.
Sunk Costs Although opportunity costs may not be readily apparent, they should always be taken into
account when making economic decisions. The opposite is the case for sunk costs—costs that have already been incurred and are beyond recovery. Even though sunk costs are usu-
ally quite apparent, they need to be ignored when making economic decisions.
Consider the case of the Miami Heat, who negotiated a 6-year, $110 million contract
with LeBron James in 2010. The contract involves a signing bonus of $20 million, plus
annual payments averaging $15 million, should the Heat exercise their option of playing
James. But suppose that after the contract is signed, a player comparable in talent to James
offers his services to the Heat for $96 million for the next 6 years—annual payments of $16
million. What should the Heat management do? The answer is, stick with James. Once the
$20 million signing bonus has been paid to James, it is a sunk cost. The opportunity cost of
exercising the James option is thus $90 million (the remaining amount that must be paid to
James) versus the $96 million it would cost to hire the rival player.
sunk costs costs that have already been incurred and are beyond recovery
Product ion Poss ibi l i ty Front ier 9
C01.INDD 10:43:41:AM 08/06/2014 PAGE 9Trim Size: 203.2 mm X 254 mm
APPLICATION 1.4
In June 1995, the 26-story Hong Kong Hilton, the first five- star hotel in the central business district of Hong Kong, was smashed to rubble. The hotel was demolished despite the facts that accounting statements showed $25 million in profit being earned on $58 million in revenue in 1994; $16 million had recently been spent to rebuild the hotel’s lobby—more than the hotel cost to build in 1963; and the owner of the
Why It Was Profitable to Demolish a Profitable Hong Kong Hotel
hotel had to pay $125 million to Hilton’s parent company to break the last 20 years of the hotel’s management contract. Why did the demolition make sense? With the astronomi- cally high rental prices for office space in Hong Kong, prop- erty consultants estimated that an extra $70 million in rental income per year could be earned by constructing an office tower on the site historically occupied by the hotel.
1.7 Production Possibility Frontier We can display in graphical form the basic economic assumptions about market actors as
well as the concept of opportunity cost. Specifically, a production possibility frontier (PPF) depicts all the different combinations of goods that a rational actor with certain per- sonal goals can attain with a fixed amount of resources. For example, suppose you are pres-
ident of a university. By effectively employing the resources on your campus, such as the
faculty and staff, classrooms, libraries, laboratories, dorms, cyclotron, and so on, you can
produce two possible services: research and teaching.
Based on the resources at your disposal, assume that the different combinations of
research and teaching that your university can produce each year are represented by the
PPF depicted in Figure 1.1. At one extreme, if your university were devoted solely to research, you could produce 1,000 units of research and 0 units of teaching (point A) with your limited resources. At the other extreme, if classroom instruction were the overriding
objective, your university could produce a maximum of 500 units of teaching and 0 units of
research (point Z). Of course, you need not be at either of the two extremes on your PPF. You also have the option of producing a mix of 500 units of research and 250 teaching
units (point E) or, for that matter, any point lying on or inside (such as Y) the straight-line segment that we have drawn connecting endpoints A and Z of the PPF shown in Figure 1.1.
production possibility frontier (PPF) a depiction of all the different combinations of goods that a rational actor with certain personal goals can attain with a fixed amount of resources
A Production Possibility Frontier (PPF) A PPF depicts the three basic assumptions made by economists about market participants (goal-oriented behavior, scarce resources, and rationality) as well as the concept of opportunity cost. With a nonsatiable desire for both research (R) and teaching (T), a university president would prefer to be as far to the northeast as possible on the graph. Scarce resources limit the president to any combination on or below the PPF boundary AZ. Rational behavior implies that the president will choose to be on the boundary as opposed to below it. Opportunity cost is reflected by the slope of the PPF.
1,000
900
800
500
50 100
250 500 Teaching (T)0
A
B
C
E
1T Z
U
Research (R)
Y
2R
Production possibility frontier (PPF)
Figure 1.1
10 An Introduct ion to Microeconomics
C01.INDD 10:43:41:AM 08/06/2014 PAGE 10Trim Size: 203.2 mm X 254 mm
How does a PPF such as the one shown in Figure 1.1 reflect the three basic assumptions that we have made about market actors? Your desire as university president to encourage
both research and teaching, if unlimited, would imply that you would like to be as far to
the northeast as possible in the graph—generating an infinite amount of both research and
teaching. But scarce resources, represented by the boundary, segment AZ, of your PPF, keep you inside the PPF. Points such as U, beyond your PPF’s boundary, thus are unattain- able. Finally, rational behavior presumes that you will choose a point on the PPF boundary rather than inside the boundary. Why choose a mix of output involving 500 research and 50
teaching units (point Y), when the same resources at your disposal can get you more of both research and teaching (such as at point C)? Rational behavior implies that you will select a point on the boundary (segment AZ) of your PPF.
Although the three basic economic assumptions about market actors imply that you will
be on the boundary of your PPF, you must still choose a specific combination of research and teaching on the boundary. The three basic assumptions (positive analysis) can place
you only on the PPF boundary. Once on the boundary, a value judgment (normative analy- sis) is necessary to determine the “best” point for you along the boundary. For example, if
you believe that a good university should focus solely on research, you will probably opt
for point A on the PPF boundary. If, instead, you believe that a university is distinguished exclusively by the quality of its classroom instruction, point Z is likely to be your selection.
Figure 1.1 also depicts the concept of opportunity cost. Consider the movement between
points A and B on the boundary of the PPF. If you start off at point A with 0 teaching units and move to point B with 50 teaching units, 100 research units will have to be given up (from 1,000 research units at A to 900 at B). Thus, if the three basic economic assumptions about market participants hold and you are forced to move along the PPF boundary, you confront a trade-off of 100 research units lost per 50 teaching units gained between points
A and B. This trade-off is the opportunity cost of using your resources to increase the num- ber of teaching units from 0 to 50. By expanding output of teaching from 0 to 50, you are
implicitly giving up 100 units of research. The opportunity cost of each additional teaching
unit gained between 0 and 50 teaching units is 2 research units. And, since the boundary
of the PPF as we have drawn it is a straight line, this per-teaching-unit opportunity cost is constant over the entire AZ boundary of the depicted PPF.
Constant versus Increasing per-unit Opportunity Costs Constant per-unit opportunity costs occur only if the PPF boundary is a straight line. But the more typical PPF has a concave-shaped boundary bowed out from the origin, as in Figure 1.2. With a concave-shaped PPF, the slope of the boundary AZ becomes steeper
Teaching
A B
Z
Y 100
1,000 999
Research
0 1 499
500
The Typical-Case PPF: Concave to the Origin With a concave PPF, the per-unit opportunity cost of an additional unit of teaching increases with the more teaching one produces. A one-teaching-unit increase between points A and B along the boundary of the PPF is associated with an opportunity cost equal to 1 research unit. A one-teaching- unit increase between points Y and Z carries with it an opportunity cost of 100 research units.
Figure 1.2
Review Quest ions and Problems 11
C01.INDD 10:43:41:AM 08/06/2014 PAGE 11Trim Size: 203.2 mm X 254 mm
(that is, more negative) as one moves from point A, where the university is producing just research and no teaching, to point Z, where the reverse is true. The per-unit opportunity cost of producing additional teaching units, in terms of the research that must be given up,
grows with the total output of teaching. For example, if you are producing 0 teaching units
and want to expand output to 1 teaching unit, moving from point A to B in Figure 1.2, you would have to drop 1 research unit. The first teaching unit produced would have an oppor-
tunity cost of 1 research unit. When you are producing 499 teaching units, however, and
want to expand teaching output by the same additional unit, the opportunity cost in terms
of research that must be given up is much higher—100 research units must be given up to
move from point Y to Z. Why does the per-unit opportunity cost associated with expanding the output of any
particular commodity typically increase in this way? The reason stems from differences
in relative productivity across resources as related to various commodities. For example,
some faculty are relatively better researchers, and others do better in the classroom. When
a university is at point A, the per-unit opportunity cost of increasing teaching output is fairly low. There are bound to be some faculty who are not prolific researchers, but who
are adept at teaching. Reallocating such faculty to increase teaching output will not involve
much of a per-unit loss in terms of research (one research unit between points A and B). By contrast, when one is already producing 499 units of teaching and contemplating
increasing production to 500 (that is, moving from point Y to point Z), one has to move more talented researchers full time into the classroom, a far more costly undertaking in
terms of forgone research.
SUMMARY
Microeconomics is the branch of economics that
studies the behavior of individual economic units such as
consumers and business firms.
Microeconomics considers how the decisions of
individuals and firms are coordinated through interac-
tions in markets.
Economists assume that market participants are
goal oriented, rational, and constrained by scarce
resources.
Because of scarce resources, market participants can’t
fulfill their desires to the extent they would like, and
choices must be made.
Whenever one alternative is chosen, an opportunity
cost is involved.
A production possibility frontier (PPF) allows us to graphically depict the basic economic assumptions
about market participants, as well as the concept of
opportunity cost.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
1.1 Say that the citizens of Tucson, Arizona, vote to increase the minimum wage within the city to $10 per hour above the
prevailing federal minimum wage of $7.25. Explain how one
would use positive and normative analysis to evaluate the
desirability of this proposed policy.
1.2 What economic forces might explain why the relative price of telephone services fell while the relative price of medical
care increased between 1983 and 2013?
1.3 Explain why it is important to look at a good’s real price as opposed to its nominal, or absolute, price.
*1.4 The RAND (short for “research and development”) Cor- poration is a think tank located on 15 prime acres of seaside
property in the center of Santa Monica, California. RAND
purchased the land for its offices from the city in 1952 for
$250,000. “Given that the money RAND paid for its land
in 1952 can be treated as a sunk cost, the cost of the land to
RAND is zero and RAND would thus be foolish to consider
purchasing a new site in Las Vegas and relocating there.” Is
this statement true, false or uncertain? Explain.
12 An Introduct ion to Microeconomics
C01.INDD 10:43:41:AM 08/06/2014 PAGE 12Trim Size: 203.2 mm X 254 mm
1.5 Tokyo’s streets are characterized by a plethora of vend- ing machines—dispensing everything from soft drinks, candy
bars, and cigarettes to magazines, personal toiletries, and beer.
Unlike in other major cities such as New York and London,
virtually every downtown street corner seems to have at least
one vending machine. Relying on the concept of opportunity
cost, explain why vending machines are so prevalent in Tokyo
versus more traditional purveying mechanisms such as news-
stands and grocery stores.
1.6 Reconsider the example in the text of whether you should pursue an MBA. Suppose that prior to making a decision you
are robbed of $5,000. Should this theft affect your decision?
How will the theft affect the accounting and economic costs
associated with pursuing the MBA?
1.7 Some people have argued that the United States cannot afford a volunteer army in which wages are high enough to
attract competent enlistees. Instead, they suggest paying lower
wages and drafting the required number of recruits. Would
such a policy change lower the accounting and economic costs
to the government for maintaining an army? Explain.
1.8 A university produces two commodities: research and teaching. The resources the university uses include faculty and
staff, libraries, classrooms, and so on. The following schedule
indicates some points on the university’s PPF:
A B C D E F G Research 900 750 600 450 300 150 0
Teaching 0 20 45 75 110 150 200
a. Does research production by the university exhibit increas- ing, constant or decreasing per-unit opportunity costs?
b. Graph the university’s PPF (assuming that straight-line segments connect the points specified above). Indicate
which areas of the graph correspond to unattainable produc-
tion points, production points that make the most effective
use of the university’s resources, and points where there are
unemployed resources.
c. Suppose that the university is at point B but would like to alter production to point C. What would be the per-teaching- unit opportunity cost of producing the extra teaching units?
d. Suppose that the university is at point C but would like to alter production to point B. What would be the per-research- unit opportunity cost of producing the extra research?
e. What will happen to the university’s PPF if the main labo- ratory burns down (assume that the laboratory is not used to
produce teaching but is used solely to produce research)?
Graph the new PPF. f. What will happen to the PPF if all of the campus resources
are cut in half? Graph the new PPF. g. Suppose the university is at point F. The university pres-
ident proposes to move the school to point B; she claims
that B is a more desirable choice since the total output is 750 + 20 = 770 total units of output at B versus 150 + 150 = 300 output units at F. Is the president correct?
1.9 “Motorola and other backers sank more than $5 billion in the 1990s into the development of Iridium, a satellite com-
munications system to connect wireless telephone users any-
where on earth. Although the number of subscribers signing
up for the service has fallen significantly below projections
and operating costs vastly exceed revenues, this is no reason
for investors to back out of the venture. Indeed, it would be
foolish to quit now, given the large amount of money that has
already been invested in Iridium.” True, false, or uncertain?
Explain.
1.10 A study by Professor Gerald Scully of the University of Texas finds that government-sponsored killing of its own peo-
ple—an act that claimed the lives of 170 million people in the
twentieth century (7.3 percent of the total population and four
times as many individuals as killed by international and civil
wars combined)—is less likely to occur the more productive
is a country’s populace (as measured by real gross domestic
product per capita). Explain why this is consistent with the
concept of opportunity cost.
1.11 According to a recent study, the average adult American living in an urban area spends 41 hours per year sitting in
traffic. The estimated cost of the gridlock, according to the
study, is substantial: $78 billion in burned gasoline (explicit
costs) and wasted time (implicit costs). How should the esti-
mated opportunity costs with the growing traffic gridlock be
adjusted if the study has not accounted for the fact that motor-
ists can make cellular phone calls while traveling on freeways?
Explain.
1.12 In 2002, Pat Tillman turned down a contract offer of $3.6 million over 3 years from the Arizona Cardinals to enlist in
the U.S. Army in the wake of the September 11 attacks on the
United States. Tillman was killed by friendly fire in Afghanistan
in 2004. From an economic perspective, did Tillman make a
rational decision by not signing with the Cardinals in 2002?
Explain.
1.13 By any absolute or relative measure, mail order and online shopping has increased dramatically over the past few decaders.
Explain why this might be the case on account of the growth of
two-wage-earner families.
1.14 One of the questions prospective full-time MBA program applicants confront is whether they should pursue the degree
soon after college or wait four to six years. Assume that 10 years
after college, a full-time MBA graduate earns as much whether
she pursued the business degree right after college or waited four
to six years. From an opportunity cost perspective, why might
enrolling for full-time MBA studies earlier rather than later in
one’s career be advantageous? Explain.
13
Supply and Demand2CHAPTER
What will government controls on the rates that doctors charge for medical services do to the availability of services? As Skype and other forms of videoconferencing become
more common, what will happen to the demand for air travel? If an Internet service pro-
vider raises its rates, will total revenues and profits also increase? Will a hike in the gov-
ernment sales tax on cigarettes have an appreciable effect on teen smoking? On cigarette
company profits?
A grounding in the basics of supply and demand can help us address these and many
other real-world questions. The supply–demand model reviewed in this chapter indicates
how the competitive interaction of sellers and buyers determines a good’s market price
and quantity. In addition, the model indicates how the market price and quantity of a good
respond to changes in other economic variables such as input costs, technology, consumer
preferences, and the prices of other goods. Furthermore, the supply–demand model can be
used to analyze the effects of various forms of government intervention in markets. Price
controls are the form that we will analyze in this chapter. Finally, we will examine how
markets operate from a quantitative as well as a qualitative perspective. In the business world especially, we often need a quantitative answer to the question of how a change in
one economic variable such as consumer income, price, price of another good or price of an
input affects the quantity demanded or supplied of a particular good.
Learning Objectives
Understand how the behavior of buyers and sellers can be characterized through demand and supply curves. Explain how equilibrium price and quantity are determined in a market for a good or service. Analyze how a market equilibrium is affected by changes in demand or supply. Explore the effects of government intervention in markets and how a price ceiling impacts price, quantity supplied, quantity demanded, and the welfare of buyers and sellers. Show how elasticities provide a quantitative measure of the responsiveness of quantity demanded or supplied to a change in some other variable such as price or income. *Explain the mathematics associated with elasticities.
Memorable Quote “If you think health care is expensive now, wait until you see what it costs when it’s free.”
—P.J. O’Rourke, American political satirist
14 Supply and Demand
2.1 Demand and Supply Curves Markets are composed of buyers and sellers. Our analysis of the behavior of buyers relies
on demand curves; supply curves depict the behavior of sellers. Let’s begin with the buyer,
or demand, side of the market.
The Demand Curve The amount of a good that a consumer or a group of consumers wishes to purchase
depends on many factors: income, age, occupation, education, experience, buyer prefer-
ences, taxes, subsidies, expectations, and so on. It also depends on the price of the good.
According to the law of demand, the lower the price of a good, the larger the quantity consumers wish to purchase. To this law we must add an important condition. The rela- tionship will hold only if the other factors affecting consumption, such as income and pref-
erences, do not change at the same time that the good’s price changes. The assumption that
all other factors remain constant is an important one to keep in mind when examining
many relationships in economics.
Figure 2.1 shows a hypothetical market demand curve for digital video/versatile disc
(DVD) players. At each possible price, the curve identifies the total quantity desired by
consumers. So, at a per-unit price of $40, the quantity demanded will be 400,000, while at a per-unit price of $30, the quantity demanded will be 550,000. Note that we do not say that demand is higher at the lower price, only that the quantity demanded is. When economists use the term demand by itself (as in demand and supply), we are referring to the entire rela- tionship, the demand curve. Quantity demanded, however, refers to one particular quantity on the demand curve.
The negative slope of the demand curve—higher prices associated with lower quan- tities—is the graphical representation of the law of demand. Economists believe that the demand curves for all, or virtually all, goods and inputs slope downward. As a con-
sequence, the proposition that demand curves have negative slopes has been elevated in
economic jargon to the position of a “law.” It is probably the most universally valid propo-
sition in economics.
law of demand the economic principle that says the lower the price of a good, the larger the quantity consumers wish to purchase
A Demand Curve The demand curve D shows the quantity of DVD players that consumers will purchase at alternative prices. Its negative slope reflects the law of demand: more DVD players are purchased at a lower price.
$40
$30
400,000 550,000 Quantity of DVD players (per month)
0
D
Price per DVD player
Figure 2.1
Demand and Supply Curves 15
A demand curve for a product pertains to a particular time period. For example, the
demand curve in Figure 2.1 may refer to consumer buying behavior for July of this year.
Another demand curve may be relevant for a different time period. In addition, the infor-
mation conveyed by the demand curve refers to alternative possibilities for the same time
period. If the per-unit price is $40 in July of this year, consumers will purchase 400,000
DVD players; if, instead, it is $30 for the same time period, consumers will purchase
550,000 DVD players.
Although economists usually interpret the curve as showing the quantities purchased at various prices, they sometimes use an equivalent interpretation. The demand curve also identifies the price that consumers will pay for various quantities. If the demand curve in Figure 2.1 is correct and 400,000 DVD players are placed on the market, consumers will
be willing to pay $40 per unit up to the marginal, 400,000th player. (400,001 DVD play-
ers will not be purchased, according to Figure 2.1, if the per-unit price is $40.) If the larger
quantity, 550,000, is offered on the market, consumers will purchase that total quantity only
at the lower per-unit price of $30. That a larger quantity can be sold only at a lower price is
another, and equivalent, way of stating the law of demand.
A final point about demand curves: their negative slope is not due only to the presence of
more consumers at lower prices. For some goods, like water, the number of consumers will
be the same regardless of price, though the amount they use may vary. Increased water con-
sumption when the price is lower reflects greater consumption per person, not more people
consuming water. At the other extreme are goods for which more consumption at lower
prices results mainly from new consumers entering the market. Personal computers might be
APPLICATION 2.1
The law of demand applies to all goods and services, including the amount of time people remain unemployed. If the cost of remaining unemployed goes down, we would expect unemployed individuals to take longer to find another job. This will increase the nation’s unemployment rate.
But what could cause the cost of remaining unemployed to go down? The answer, surprisingly, is the U.S. govern- ment’s own program designed to deal with unemployment, unemployment insurance (UI). UI programs (they vary by state) typically pay unemployed persons a benefit equal to half of their previous wage earnings. What this means is that the cost of staying unemployed is reduced by half by UI since that cost is what you sacrifice by not working, and UI gives you half that amount. The predictable result is that people will stay unemployed for longer periods of time when they are subsidized in this way, and that adds to the unemployment rate.
Empirical work by economists confirms the prediction that UI increases the duration of unemployment and hence the unemployment rate. One bit of evidence is particularly compelling. It turns out that about five to seven percent of those unemployed find a job and return to work each week. At least this is true for the first 25 weeks of unemployment; in
The Law of Demand at Work for Non-work
week 26 the percentage finding a job triples to 16.5 percent.1
What is special about week 26? It turns out that UI tradition- ally has paid benefits for only 26 weeks. Apparently, a sizable number of unemployed persons choose to stay unemployed as long as they can collect unemployment benefits.
Historically, UI has been estimated to increase the nation’s unemployment rate by roughly one percentage point. The impact has grown larger in recent years as federal policymakers, have extended unemployment benefits to 99 weeks from the standard 26 weeks. While this was done out of compassion in the wake of the severe macroeco- nomic downturn that began in 2007, it has been estimated that the unemployment rate would have been 2.7 to 1.5 percent lower (6.8 to 8.0 percent versus 9.5 percent in June 2010) were it not for the extension of unemployment ben- efits. The estimates mean that a large share of preventable unemployment is caused by UI. It’s just the law of demand at work—in this case, for non-work.
1This application is based on Jonathan Gruber, Public Finance and Public Policy (New York: Worth Publishers, 2005), pp. 374–375; “Stimulating Unemployment,” Wall Street Journal, July 20, 2010, p. A16; and Robert Barro, “The Folly of Subsidizing Unemployment,” Wall Street Journal, August 30, 2010, p. A15.
16 Supply and Demand
such a good. Most goods fall between these extremes—more consumers entering the market,
and more consumption per consumer occurring at lower prices. The downward-sloping
demand for DVD players is probably due mostly to additional families buying DVD players
at lower prices, but some families may purchase more than one for their own usage or give
them away as gifts.
Determinants of Demand Other Than Price As mentioned earlier, many factors influence consumer purchases. A demand curve focuses
just on the effect of changes in a product’s own per-unit price, with other factors held con-
stant. For example, consumers’ incomes are taken to be invariant at all points on a particu-
lar demand curve. Now let’s consider the other factors, besides the good’s price, that might
affect the consumption of it. First are the incomes of consumers. Income level is almost certain to affect the amount of goods consumers will purchase; usually, they wish to pur-
chase more when income rises. The term normal goods refers to those for which an increase in income leads to greater consumption. Studies indicate that DVD players fall
into this category. There are, however, certain goods, called inferior goods, whose con- sumption falls when income rises. Examples of this latter, more rare, category might
include hamburger and public transportation. If you won $10 million in the state lottery,
would you continue to take the bus?
In addition to incomes, the prices of related goods also affect the quantity of DVD player purchases. Related goods fall into two distinct groups: complements and substitutes. Two goods are complements if they tend to be consumed together, so consumption of both goods tends to rise or fall simultaneously. Examples of complements to DVD players
include television sets, DVDs, and popcorn. After all, trying to use a DVD player without
discs is difficult. And popcorn, for many consumers, increases the pleasure of using a DVD
player. If two goods are complements, an increase in the price of one decreases the demand for the other, and vice versa. If the price of television sets rises, DVD player consumption will decrease (the demand curve for DVD players shifts in).
Substitutes, on the other hand, are goods that can replace one another in consumption. Their consumption is frequently an “either–or” choice since they serve similar purposes,
and one or the other may be chosen. For instance, live theater and movies on demand on
cable television, are, in the eyes of many consumers, substitutes for DVD players. If two
goods are substitutes, an increase in the price of one increases demand for the other. If the
price of live theater rises, DVD player consumption will increase (the demand curve shifts
out); DVD players are substituted for theater when theater becomes more expensive.
It is not always readily apparent whether two goods are complements or substitutes for
one another. For example, one might think that personal computers and word processing
programs would be substitutes for paper since virtual text can be used instead of hard
copy. If anything, however, personal computers and word processing programs initially
proved to be complements to paper. That is, paper usage increased with the advent of
virtual text. This appears to be the case because as technology made it easier to create,
revise, and access text, individuals had more information that they could print in hard
copy form.
Consumers’ tastes or preferences also affect consumption. By tastes or preferences we mean the subjective feelings of consumers about the desirability of different goods. Should
consumers decide that outdoor exercise is more appealing than watching DVDs—a change
in tastes—the purchases of players would drop off. A real-world example concerns Minne-
sota, where video-based entertainment is less popular than it is, on average, in other states.
The phenomenon appears to reflect the fact that Minnesota was settled disproportionately
by Scandinavians, who do not regard video-based entertainment as highly as do other eth-
nic groups in America.
normal goods those goods for which an increase in income leads to greater consumption
inferior goods those goods whose consumption falls when income rises
complements two goods that tend to be consumed together, so consumption of both tends to rise or fall simultaneously
substitutes goods that can replace one another in consumption
tastes or preferences the feelings of consumers about the desirability of different goods
Demand and Supply Curves 17
Because tastes are a harder-to-quantify factor than price or income, it may be tempting
to omit them from an analysis of demand for a product. To do so, however, can lead to
incomplete explanations and inappropriate conclusions. For example, the law of demand
applies to the market for beef in India: namely, if the price of beef rises in India, the quan-
tity of beef demanded will fall. Price, by itself, however, cannot explain why per capita
beef consumption is so low in India relative to the United States. To understand this
requires knowledge of factors such as religious beliefs and cultural taboos that influence
Indian consumer preferences regarding beef.
The preceding factors (incomes, prices of related goods, and tastes) are not the only
influences on demand beyond the price of the good in question. They are, however, almost
certain to be significant for virtually all goods. Thus, we will concentrate on them in the
remainder of our analysis.
Shifts in versus Movements along a Demand Curve In drawing a demand curve, we assume that incomes, the prices of related goods, and pref- erences are the same at all points on the curve. The purpose of holding them constant is not to deny that they change but to identify the independent influence of the good’s own
price on consumer purchases. If incomes, prices of related goods or preferences do change,
the entire demand curve shifts. Figure 2.2 illustrates such a shift. (Note that we begin to
use shorthand terms on the axes, price and quantity. It should be understood that we mean price per unit and quantity per time period.) Demand curve D, for example, reflects condi- tions when consumers’ annual incomes average $50,000 and the other factors are held con-
stant. If consumers’ average annual incomes rise to $70,000 and DVD players are a normal
good, consumers will wish to purchase more players at every price than they did before.
The change in income produces a shift in the demand curve from D to D′. An increase in income increases demand for the good (in the normal good case), meaning that the entire demand curve shifts outward or rightward. A decrease in demand refers to an inward, or leftward, shift in the demand curve toward the origin. If consumers’ preferences shift from
being a couch potato and watching DVDs to exercising outdoors, the demand for DVD
players will decrease.
An Increase in Demand For a demand curve, other influences besides the price of the good being examined (consumers’ incomes, consumers’ preferences, prices of related goods, and so on) are held constant at all points along the curve. Changes in these underlying factors normally cause the demand curve to shift. Here, an increase in consumers’ incomes causes the demand curve to shift from D to D′ because DVD players are assumed to be a normal good.
$40
$30
400,000
550,000
700,000
900,000
Quantity of DVD players
0
A
B
D D′
PriceFigure 2.2
18 Supply and Demand
APPLICATION 2.2
During the first 65 years of the 20th century, annual cig- arette consumption per adult in the United States increased steadily from 54 in 1900 to 4,259 in 1965. Thereafter, the opposite has been the case as annual cigarette consump- tion per adult declined to below 1,700 by 2006. What accounts for this rise and fall?
Both non-price and price factors have influenced ciga- rette consumption. For example, during the earlier half of the 20th century, the advent of complements such as safety matches and advertising campaigns that informed/ influenced consumer preferences (e.g., the mass market- ing of Camel brand cigarettes by the R.J. Reynolds Tobacco Company) shifted the demand curve for cigarettes outward. In addition, the invention of machinery to mass-produce cigarettes lowered the real price of cigarettes and resulted in a movement downward along a given demand curve for cigarettes, thereby increasing quantity demanded.
Similarly, the decline of per-capita cigarette consump- tion post 1965 in the United States has been influenced by both non-price and price factors. The publication of studies linking cigarettes to lung cancer and other diseases
The Rise and Fall of Cigarette Consumption in the United States
negatively influenced consumer preferences and shifted the demand curve for cigarettes inward. Rising real prices, due to increased federal and state taxes and legal settle- ments between plaintiffs and cigarette companies regard- ing the adverse health consequences of their products, resulted in a movement upward along a given demand curve for cigarettes, thereby reducing quantity demanded.
While considerable debate is ongoing regarding the rela- tive effectiveness of non-price versus price factors to fur- ther impact cigarette consumption (and for that matter, other goods such as alcohol as well as sugary soft drinks that have been found to contribute to obesity), one thing that is clear is that price, and thereby the law of demand, cannot be ignored in terms of the significance of its poten- tial impact on the quantity demanded of a good. For exam- ple, a Congressional Budget Office study in 2012 estimated that a $1-per-pack tax increase on cigarettes would prompt 2.8 million adult smokers to quit by 2021, save 20,000 lives over the same period, reduce health care costs, and generate at least $80 billion in additional revenues for the federal government.
APPLICATION 2.3
Behavioral economics is an emerging sub-discipline that seeks to incorporate sociological and psychological factors to improve the predictive power of the classic microeconomic model. One of the insights drawn from this sub-discipline
is that there occasionally can be a more complex interplay between the price and non-price factors that affect the quantity demanded of a particular good or service. Michael Sandel, a noted American political philosopher,
points out how focusing on the influence price has on con- sumption both ignores the important role non-price factors, such as preferences or norms, can play and may miss the subtle and counterproductive influence that greater reliance on price incentives might have in certain cases.2
Sandel points to a study of some Israeli child-care cen- ters that were trying to develop a way to deal with parents
The Occasional Interplay Between Price and Non- Price Factors in Determining Quantity Demanded
that showed up late to pick up their children. When the parents were late, a teacher had to remain on site. To deal with such tardiness, the centers began levying a fine for late pickups. Counterintuitively, late pickups increased when the price associated with tardiness rose.
Why? Beyond the inverse relationship between a higher price and the quantity demanded of tardiness predicted by the law of demand, the advent of fines also appeared to erode the guilt parents otherwise felt from showing up late. By reducing preferences for remaining free of guilt, the indirect effect of a higher fine on tardiness ended up more than counteracting the intended direct and negative effect of fines on tardiness through the law of demand.
Sandel notes that an analogous interplay comes into play when trying to encourage children to do well at school or to be courteous to others. Studies indicate that paying kids to get good grades or to write thank you notes doesn’t always produce the desired results. Account has to be made for the role intrinsic preferences for learning/civility play and how best to inculcate such preferences. Mere reliance on incentives can in certain cases lead to unintended out- comes if those incentives diminish the attention paid to developing intrinsic preferences.
2This application draws on: Michael Sandel, What Money Can’t Buy (New York: Farrar, Straus and Giroux, 2012); Daniel Pink, Drive: The Surprising Truth About What Motivates Us (New York: Riverhead Books, 2009); and Uri Gneezy and Aldo Rustichini, “A Fine is a Price,” Journal of Legal Studies, 29 (January 2000), pp. 1-17.
Behavioral economics The study of the effects of psychological, social, cognitive, and emotional factors on economic decisionmaking and outcomes.
Demand and Supply Curves 19
To use demand curves correctly, we must distinguish clearly between situations that involve
a movement along a given demand curve and those that involve a shift in demand. A move- ment along a given demand curve occurs when the quantity demanded changes in response to
a change in price of a particular good while the other factors affecting consumption are held
constant. This is not a change in the demand curve. An example would be the movement from
point A to point B along demand curve D in Figure 2.2. A shift in demand, a movement of the curve itself, occurs when there is a change in factors besides price such as income, preferences,
and the price of related goods, affecting the quantity demanded at each possible price.
An example is the movement of the entire demand curve from D to D′ in Figure 2.2.
The Supply Curve On the supply side of a market, we are interested in the amount of a good that business
firms will produce. According to the law of supply, the higher the price of a good, the larger the quantity firms want to produce. As with the law of demand, this relationship
will necessarily hold only if other factors that affect firms’ decisions remain constant when
the price of the good changes. The amount firms offer for sale depends on many factors,
including the technological know-how concerning production of the good, the cost and
productivity of relevant inputs, expectations, employee–management relations, the goals of
firms’ owners, the presence of any government taxes or subsidies, and so on. The price of
the good is also important because it is the reward producers receive for their efforts. The
supply curve summarizes the effect of price on the quantity that firms produce.
Figure 2.3 shows a hypothetical market supply curve for DVD players. For each possible
price the supply curve identifies the sum of the quantities offered for sale by the separate firms. Because all the firms that produce a particular product constitute the industry, this curve is
generally called the industry, or market, supply curve. It shows, for example, that at a price
of $20 per DVD player, the quantity supplied will be 300,000, whereas at a price of $30, the quantity supplied will be 600,000. Note that we do not say that supply is greater at the higher price, only that the quantity supplied is. The term supply by itself refers to the entire supply curve, while quantity supplied refers to one particular quantity on the curve. This parallels the terminology used for the demand curve.
Supply curves for most goods slope upward. Basically, this upward slope reflects the
fact that per-unit opportunity costs rise when more units are produced, so a higher price is
necessary to elicit a greater output.3
movement along a given demand curve a change in quantity demanded that occurs in response to a change in price, other factors holding constant
shift of a demand curve a change in the demand curve itself that occurs with a change in factors, besides price (such as income, the price of related goods, and preferences) that affects the quantity demanded at each possible price
law of supply the economic principle that says the higher the price of a good, the larger the quantity firms want to produce
3In Chapter 9, however, we will see that supply curves for some products may be horizontal. Upward-sloping supply curves, however, are thought to be the most common shape, and we draw them this way here.
A Supply Curve The supply curve S shows the quantity of DVD players firms will be willing to produce at alternative prices. It generally slopes upward, indicating that a higher price will result in increased output.
$30
$20
300,000 600,000 Quantity of DVD players
0
Price SFigure 2.3
20 Supply and Demand
Like the demand curve, the supply curve pertains to a particular period of time. In addi-
tion, different points on the supply curve refer to alternative possibilities for the same
period of time. Finally, the number of firms producing the good may vary along the supply
curve. At low prices some firms may halt production and leave the industry; at high prices
new firms may enter the industry.
Shifts in versus Movements along a Supply Curve The supply curve shows the influence of price on quantity supplied when other factors that
also influence output are held constant. When any of the other factors change, the entire
supply curve shifts.
Beyond a good’s own price, two determinants of quantity supplied deserve special
emphasis. First is the state of technological knowledge concerning the various ways a good can be manufactured. Second are the conditions of supply of inputs, like labor and energy, that are used to produce the good. Supply conditions for inputs relate to the prices that must
be paid for their use. Other factors may be important in particular cases—for example, gov-
ernment in the case of health care, weather in the case of agriculture, and an organization’s
goals in the case of nonprofit institutions like the Red Cross—but technology and input
supply conditions influence all output markets.
In drawing a supply curve, we assume that factors such as technological knowledge and
input supply conditions do not vary along the curve. The supply curve shows how variation
in price alone affects output. If technology or input supply conditions do change, the supply
curve shifts. For instance, if a new technology allows manufacturers to produce DVD play-
ers at a lower cost, the supply curve shifts to the right, as illustrated in Figure 2.4. Because
producers’ costs are now lower due to the technological advance, individual producers will
want to produce more at any price. After the technological advance, the quantity supplied
is greater at each possible price, as shown by the shift in the supply curve from S to S′. The rightward shift in the supply curve reflects an increase in supply. If there is an increase in supply, each quantity will be available at a lower price than before. For example, before
the technological advance, 300,000 DVD players would have been produced only if the
price were at least $20; afterward, 300,000 would be produced at a price of $15.
Just as with demand curves, we must distinguish a movement along a given supply curve from a shift in supply. A movement along a supply curve occurs when the quan- tity supplied varies in response to a change in the good’s selling price while the other
factors that affect output hold constant. A shift in the supply curve occurs when the other
factors that affect output change.
movement along a given supply curve a change in quantity supplied that occurs in response to a change in the good’s selling price, other factors holding constant
shift of a supply curve a change in the supply curve itself that occurs when the other factors, besides price, that affect output change
An Increase in Supply For a supply curve, technology and input supply conditions are held constant at all points along the curve. Changes in these underlying factors normally cause the supply curve to shift. Here, a technological change causes the supply curve to shift from S to S′.
$15
$20
300,000 500,000 Quantity of DVD players
0
Price S
S′
Figure 2.4
Determinat ion of Equi l ibr ium Pr ice and Quant ity 21
2.2 Determination of Equilibrium Price and Quantity The demand curve shows what consumers wish to purchase at various prices, and the sup-
ply curve shows what producers wish to sell. When the two are put together, we see that
there is only one price at which the quantity consumers wish to purchase exactly equals the
quantity firms wish to sell. In Figure 2.5, that price is $30, where consumers wish to pur-
chase 475,000 DVD players and firms wish to sell the same quantity. It is identified by the
point of intersection between the supply and demand curves.
The intersection identifies the equilibrium price and quantity in the market. Upon reaching equilibrium, price and quantity will remain there. Of course, if the supply or the
demand curve shifts, the equilibrium point will change, too. A basic assumption of micro-
economic theory is that the independent actions of buyers and sellers tend to move the
market toward equilibrium. We can see how this happens if we first imagine that the price
is not at its equilibrium level. Suppose, for example, that the price is $20 in Figure 2.5. At
$20, the demand curve indicates that consumers want 600,000 DVD players, but the sup-
ply curve shows that firms will produce only 200,000 DVD players. This situation is a
disequilibrium; the quantity demanded exceeds the quantity supplied, so the plans of buy- ers and sellers are inconsistent. The excess of the amount consumers want over what firms
will sell—in this case 400,000 DVD players—is called the excess demand (XD), or short- age, at the price of $20.
How will the people involved—both consumers and business managers—react in this situ-
ation? Consumers will be frustrated by not getting as much as they wish and will be willing
to pay a higher price to obtain more DVD players. Business managers will see that quantity
demanded is greater than quantity supplied and will be prompted to hike their selling price.
Consequently, whenever there is a shortage at some price, market forces—defined as the
behavior of buyers and sellers in the market—tend to produce a higher price. In this example
the price rises to $30. As the price rises, quantity demanded falls below 600,000 (a movement
along the demand curve), and quantity supplied increases beyond 200,000 (a movement along
the supply curve). The process continues until quantity demanded equals quantity supplied at
a price of $30.
equilibrium a situation in which quantity demanded equals quantity supplied at the prevailing price
disequilibrium a situation in which the quantity demanded and the quantity supplied are not in balance
shortage excess demand for a good
Determination of the Equilibrium Price and Quantity The intersection of the supply and demand curves identifies the equilibrium price and quantity. Here, at the price of $30, the quantity demanded by consumers exactly equals the quantity supplied by firms. Market forces tend to produce this outcome.
$20
$40
$30
200,000 475,000
600,000
Quantity of DVD players
0
Price
S
D
XD
XS
Figure 2.5
22 Supply and Demand
Alternatively, if for some reason the price is above $30, the quantity firms wish to sell
will be greater than the quantity consumers are willing to buy. An excess supply (XS), or surplus, will exist at a higher-than-equilibrium price. Unsold goods pile up. In this case market forces exert downward pressure on price: firms cut prices rather than accumulate
unwanted inventories and consumers realize that they do not have to pay as high a price for
the good.
Therefore, at any price other than the equilibrium price, market forces will tend to cause
price and quantity to change in the direction of their equilibrium values. The equilibrium
position itself will change whenever demand or supply curves shift, so actual markets may,
in effect, be pursuing a moving target as they continually adjust toward equilibrium.
2.3 Adjustment to Changes in Demand or Supply The most common application of the supply and demand model is to explain or predict
how a change in market conditions affects equilibrium price and output. In Figure 2.6a, we
see an increase in demand but no change in supply. Demand might increase, for instance,
following a report from the Surgeon General’s office that watching at least two hours of
DVDs per day reduces the risk of heart attack. This report would shift the demand curve for
DVD players to the right but leave the supply curve unaffected.
Before demand increases, the equilibrium price and quantity are $30 and 475,000 DVD
players, respectively. When the demand curve shifts to D′, a shortage will temporarily exist at the original price of $30—quantity demanded (650,000) will exceed quantity sup-
plied (475,000). As a consequence, there will be upward pressure on price. Price will rise,
quantity supplied will increase, and quantity demanded will decline until a new equilib-
rium price and quantity of $40 and 600,000 DVD players, respectively, are determined, as
surplus excess supply of a good
Market Adjustments to Changes in Demand and Supply (a) An increase in demand from D to D′, with supply unchanged, leads to a higher equilibrium price and output. (b) An increase in supply from S to S′, with demand unchanged, leads to a lower equilibrium price and a higher equilibrium output.
$40
$30
475,000 650,000
600,000
Quantity of DVD players
0
Price
S
D′
D
$30
$25
475,000 625,000
550,000
(b)(a)
Quantity of DVD players
0
Price
S
S′
D
Figure 2.6
Adjustment to Changes in Demand or Supply 23
indicated by the intersection of D′ and S. Note that the higher output is not described as an increase in supply; only the quantity supplied has increased.
Figure 2.6b shows the effects of an increase in supply when there is no change in
demand. Suppose that a technological advance reduces DVD player production costs.
This will cause the supply curve of DVD players to shift rightward. At the initial price of
$30 firms would now wish to sell more players than consumers would be willing to buy,
and a temporary surplus would result. Price would fall, and a new equilibrium with a price
of $25 and an output of 550,000 would be established. Note that the greater purchases by
consumers at the new equilibrium are not described as an increase in demand. Demand has
not increased; quantity demanded has increased because of a lower price.
Do real-world markets respond to the forces of supply and demand in the way suggested
by the theory? Over time, economists have accumulated a great deal of evidence indicating
that they do.
Some recent confirmations of the theory come from such disparate sources as online
shopping and variable-price parking meters, soft-drink vending machines and baseball
ticket pricing, as well as growing obesity in the United States. For example, the next
generation of parking meters and soft-drink vending machines is being armed, through
computer chips, to adjust the price charged in response to changes in demand—a higher
price when traffic is more congested or when the weather is hotter and, consequently,
consumer thirsts are greater. In 2014, the Boston Red Sox began charging higher ticket
prices for more popular baseball games–either based on the time of day/week or the
opposing team. For another example, the leading explanation about why Americans are
getting fatter is the decrease in the real price of food. The reduction in the mass prepara-
tion of food over the past four decades has shifted the food supply curve rightward and
thereby both lowered the price of food and led to an increase in caloric intake.
The Web has also increased the speed at which the market equilibrium adjusts to
changes in supply and demand across various sectors of the economy. In contrast to tradi-
tional retail models, where fixed prices for relatively long periods of time are the norm, the
Internet allows for quicker updating of prices and customer feedback. By bringing informa-
tion about supply and demand together at the moment of sale, online shopping increases
the alacrity with which equilibrium prices and quantities adjust, in predictable ways, to a
change in demand or supply.
APPLICATION 2.4
As of 2013, there were 513,000 vacation homes in Switzerland with over 400,000 of these being built in the preceding four decades.4 Concerned with over- development and the presence of too many “cold beds,” as the part-time dwellings are known, in certain popular vacation spots, Swiss voters passed a law in 2012 ban- ning the construction of any more holiday homes in towns where they account for more than 20 percent of the housing. Said, one vacationer who shares such a home in Grindelwald that his family and friends use throughout
Why Holiday Home Prices in Switzerland are Soaring
the year “they turn the villages into ghost towns . . . we see the permanently closed shutters without a single geranium outside the windows.”
The effect of the ban on the holiday home market has been predictable. The ban has arrested further out- ward movements of the supply curve of such homes and equilibrium prices have risen dramatically—by 10–15 per- cent in the year following the vote. Over time, the ban is likely to further the price gap between vacation homes in Switzerland and other Alpine locales. As of 2013, the pur- chase price of Swiss vacation homes in St. Moritz averaged $1,830 per square foot versus $650 in nearby Chamonix, France, and $372 in the Austrian getaway of St. Anton.
4This application is based on: “The Swiss Try to Stem a Chalet Glut,” Bloomberg Businessweek, July 8–14, 2013, pp. 38–39.
24 Supply and Demand
Using the Supply–Demand Model to Explain Market Outcomes We have focused so far on how the supply–demand model can be employed to predict market outcomes. An increase in the price of gasoline, for example, can be forecast to shift
the demand curve for cars leftward (since gasoline and cars are complements) and thus
to reduce the equilibrium price and quantity in the market for cars. However, the supply–
demand model can also explain market outcomes. For example, why has the equilibrium price for, and per capita consumption of, medical care increased so dramatically over the
last fifty years? Why is the price of gasoline so high and per capita consumption so much
lower in Western Europe than the United States?
We can use the supply–demand model “in reverse” to explain such puzzling market out-
comes in a fairly straightforward way. The first step involves determining how the equilib-
rium in a market has changed. Suppose, for example, that, as shown in Figure 2.7, we start
off with an initial equilibrium price and quantity (P* and Q*) determined by the intersec- tion of supply and demand curves (S and D). The initial equilibrium can be altered in four ways, represented by four quadrants.
Once we have determined what quadrant the new equilibrium is in, we can see whether
demand or supply has produced the new market outcome. For example, if the new equi-
librium is to the northeast of the initial equilibrium, we know that the demand curve has
altered the equilibrium: the demand curve shifts to the right to produce a new equilibrium
located northeast of the initial one. After we know which curve has produced the new
equilibrium, we can attempt to isolate the factor that produced the observed change in
market outcome.
Take the case of gasoline in Western Europe. As anybody who has traveled there
knows, the per-unit price of gasoline is usually two to three times as high as it is in the
United States. Moreover, per capita consumption of gasoline is significantly lower in
Western Europe than in the United States. Why is this the case? In Figure 2.7, the Western
European per capita equilibrium for gasoline is located to the northwest of the equilibrium
in the United States. Its location to the northwest means that the supply curve must be
explaining the Western European market outcome. Specifically, the supply curve has to
shift to the left to induce a movement in the gasoline market equilibrium to the northwest.
Now that we have isolated the supply curve as playing the dominant explanatory role, we
can focus on what determinant of supply might be producing such a leftward shift in the
P*
0
lower equilibrium price and quantity
higher equilibrium price and lower equilibrium quantity
higher equilibrium price and quantity
higher equilibrium quantity and lower equilibrium price
QuantityQ*
S
D
Price
Using the Supply–Demand Model to Explain Market Outcomes The supply–demand model can be used “in reverse” to explain market outcomes. By determining where the new equilibrium is relative to the initial equilibrium, it is possible to determine whether the demand or supply curve has produced the new market outcome.
Figure 2.7
Government Intervent ion in Markets : Pr ice Controls 25
supply curve. The culprit turns out to be the taxes levied by Western European govern-
ments on the sale of gasoline—taxes that significantly raise the cost of supplying gasoline
to the market.
The supply–demand model also can be applied, in reverse, to explain the market for
medical care. By most measures, per capita consumption and the per-unit price (adjusted
for inflation) of medical care have increased over the past 70 years in the United States.
What is behind this phenomenon? We are clearly to the northeast, in terms of Figure 2.7,
of the equilibrium that prevailed at the end of World War II. The demand curve, by shifting
to the right, must be playing the explanatory role. One likely reason the demand curve has
shifted is growth in consumers’ incomes and the fact that medical care is a normal good. As
people get richer, they appear to be spending more money on medical care for themselves
and their families. As we shall see in later chapters, however, the full explanation is a bit
more complicated.
2.4 Government Intervention in Markets: Price Controls Markets can be thought of as self-adjusting mechanisms; they automatically adjust to
any change affecting the behavior of buyers and sellers in the market. But for this mecha-
nism to operate, the price must be free to move in response to the interplay of supply and
demand. When the government steps in to regulate prices, the market does not function in
the same way. We can use the supply–demand framework to analyze this form of govern-
ment intervention.
Policymakers may believe that market-determined prices are either too high or too low.
In the former case, they may impose a legislated maximum price, or price ceiling. Under a legislated price ceiling it is illegal to charge a price higher than the ceiling. In the latter
case, a minimum price, or price floor, may be legislated. In this section, we’ll analyze price ceilings and show how their economic effects may be directly contrary to the stated objec-
tives of the policymakers who impose them.
Price ceilings are not uncommon. In the twentieth century, broad-ranging price con-
trols were established at the federal level during several major crises, including World War
II, the Korean War, and the Vietnam War. Other examples of price ceilings on specific
items include rent control, caps on automobile insurance rates in some states, federal con-
straints on the prices that may be charged for human body organs for transplant, and limits
on reimbursement for medical services imposed by Medicare/Medicaid. We will examine
rent control in detail, but keep in mind that the effects of government intervention can be
generalized to other markets where price ceilings prevent competitive market forces from
determining the equilibrium.
Rent Control During World War II, many local governments in the United States applied price ceilings
to rental housing units, a policy generally referred to as rent control. New York City was the only major city to continue rent control after World War II, and it still uses it today.
Apart from New York, relatively few cities experimented with rent control until the 1970s,
when an increasing number of cities adopted the practice. By 1990, more than 200 cities
were using some form of rent control, including Los Angeles, Washington, Boston, and
Newark (although over the past two decades, the state legislature of Massachusetts has
voted to end local rent control, while California has limited the extent to which local cities
can intervene with regards to the setting of rents).
price ceiling a legislated maximum price for a good
price floor a legislated minimum price for a good
rent control price ceilings applied to rental housing units
26 Supply and Demand
We can examine the effects of rent control with the aid of Figure 2.8, which shows
the supply and demand curves for rental housing units in a particular city. As shown in
Figure 2.8, in the absence of rent control, the equilibrium monthly price is P, or $1,000, and Q is the equilibrium quantity. Rent control imposes a maximum price on a rental unit below the equilibrium level. Suppose that the price is not allowed to rise above Pc, or $750. The first question to answer is whether the law can be effectively enforced. Because tenants are
willing to pay a higher price, landlords have an incentive to extract side payments from ten-
ants. Such side payments might include a nonrefundable key deposit; purchase of a parking
space or furniture as a condition for renting; or payment of one’s own repairs. All these
practices have been observed to occur under rent control. In this way, a landlord might
charge the regulation $750 in explicit rent but receive enough extra in side payments to get
the effective price closer to the market price.
Let’s assume that these methods of circumventing the law are not allowed. (They are, in
fact, illegal under most rent control laws.) The price that consumers must pay is lower, and
at a lower price the quantity demanded is greater, namely, Q2. At the lower price, however, the quantity supplied falls to Q1, because investment in this market becomes less profitable. Fewer new rental units are constructed. New York City, for example, has the lowest rental
vacancy rate in the country. And the number of new rental housing units (per capita) added
to the city’s total supply of such units is the lowest in the nation. Moreover, owners allow
existing units to deteriorate more rapidly by spending less on maintenance for those units.
One study found repair expenditures in New York City on rent-controlled apartments to be
only half as large as the repair expenditures made on comparable apartments not subject to
rent control.5
The result of rent control, like that of any other price ceiling applied in a competitive
market, is a shortage. The quantity that potential tenants would like to rent, Q2, is greater than the quantity available, Q1, and the excess of quantity demanded over quantity supplied,
5G. Sternlieb, The Urban Housing Dilemma (New York: New York Housing and Development Administra- tion, 1972), p. 202.
Rent Control With a legal maximum rent of $750 set below the market equilibrium level of $1,000, quantity supplied falls from Q to Q1, and quantity demanded rises to Q2. The difference, Q2 minus Q1, is the excess demand, or shortage, created by the rent control policy.
Rental housing services
0 Q1 Q Q2
$1,500 = P1 S
D
$1,000 = P
$750 = Pc
Monthly rent per unitFigure 2.8
Government Intervent ion in Markets : Pr ice Controls 27
Q1Q2, measures the shortage. Because only Q1 units are actually available, the marginal value of housing units to consumers must be at least $1,500 (the height of the demand
curve at Q1). Price, however, cannot legally rise above $750 per month, so producers have no incentive to increase quantity beyond Q1.
Rent control has effects on other markets, too. Not all people who wish to rent are able
to do so, so they must make other living arrangements. Apart from living in a different
community, the major alternative is some form of owner-occupied housing. Therefore, the
demand for such housing will increase as frustrated apartment hunters turn to home own-
ership. At the same time, the owners of rent-controlled apartments have an incentive to
convert their rental units into owner-occupied units, or condominiums, and sell them to
tenants. Typically, communities with rent controls resort to limitations on condominium
conversion to prevent the supply of rental units from drying up completely. Because land-
lords also have an incentive to convert rental units to commercial units (such as stores and
business offices) that are not subject to rent control, most communities with rent control
also outlaw this practice.
Who Loses, Who Benefits? Those obviously harmed by rent control are the owners of rental units at the time the policy
is implemented. They have invested in the construction or purchase of units in the expecta-
tion of being able to charge $1,000 per month (in our example), but they find their return
reduced by law. Although landlords are often depicted as wealthy and easily able to bear
the losses imposed by rent control, this is often not the case. While the evidence is sketchy,
landlords often have incomes that are no greater than those of their tenants. For example, a
1988 survey of New York landlords found that 30 percent had incomes below $20,000, and
half had incomes below $40,000.6
Now consider who benefits from rent control. The intended beneficiaries are clear. In
virtually all cases, proponents of rent control expressly seek to benefit tenants. Economists,
however, are skeptical about the degree to which this benefit actually occurs. Indeed, some
economists believe that tenants, on average, are worse off under rent control. While lower
rents by themselves are good for tenants lucky enough to get a rent-controlled apartment,
other changes in the market are not so advantageous.
First, the lower rental price is necessarily accompanied by a lower quantity (quantity
falls from Q to Q1 in Figure 2.8). A lower price is good for tenants, but fewer rental units are not, and the net effect of the two is uncertain. It is conceivable that all tenants are made
worse off. To see this, suppose that the quantity of available rental units falls to zero under
rent control. A lower price does tenants little good if they cannot find housing. Moreover,
all tenants likely will not be affected in the same way. Some tenants may find what they
want at the lower rents, while others may not. For example, while former New York Mayor
Ed Koch had a rent-controlled apartment in Greenwich Village in 1989 for $352, a recent
college graduate was paying $300 to live in a pantry in another person’s apartment.7 (Pan-
tries are not subject to rent control.) The biggest losers among the tenants are, of course, the
“potential tenants” who are unable to find rental apartments and must either purchase hous-
ing or live elsewhere.
Another disadvantage to tenants of rent control involves its impact on quality. Just as a
quantity reduction acts to the detriment of tenants, so does the decrease in quality. Landlords
have an incentive to lower costs by reducing maintenance. Normally, they would not do this
because they would lose tenants, but because of the rent-control-created shortage they can
reduce the value of rental units without driving tenants away. Consequently, tenants get a
lower-quality product for the lower price.
6Irving Welfeld, Where We Live (New York: Simon & Schuster, 1988), p. 146. 7William Tucker, “It’s a Rotten Life,” Reason (February 1989), p. 23.
28 Supply and Demand
Quality is also likely to suffer when rent control laws have provisions permitting land-
lords to raise the rent on a unit that becomes vacant. Because such provisions give land-
lords an incentive to evict tenants, the law typically also has strong anti-eviction provisions.
In turn, anti-eviction provisions give landlords the incentive to make tenants as unhappy as
possible so they will choose to leave, and one way to do that is to let the rental unit dete-
riorate. The normal incentive of landlords—to provide a quality unit so tenants will stay a
long time—is turned upside down by rent control.
Under rent control, nonprice rationing becomes more prevalent. Since price is not
allowed to ration the available quantity among competing consumers, quantity supplied
does not equal quantity demanded, and some other way of determining who gets the good
and who doesn’t must arise. Nonprice rationing can take many forms and works to the
disadvantage of some tenants. Because there are many more potential tenants than apart-
ments, landlords can be highly selective. For example, they are likely to favor tenants
without children or pets (children and pets increase maintenance and repair costs) and ten-
ants with histories of steady employment at good wage rates (who can be counted on to
pay on time).8 Minorities may not fare well under rent control if landlords have prejudices
against them and choose to indulge such prejudices in selecting tenants.
Another form of nonprice rationing is rationing on a first-come, first-served basis.
Because of the lack of available units, potential tenants incur the cost of waiting in line (or
in a pantry) and searching for that rare commodity, a vacant rent-controlled unit. This fac-
tor, of course, adds to the true cost of rental housing since not only the rent but also the cost
of time spent waiting and searching must be paid by prospective tenants.
A more obvious form of nonprice rationing is the payment of bribes to secure a rent-
controlled unit. Because a rent-controlled apartment is often worth more to tenants than
the listed rent, they are willing to pay a “finder’s fee” to anyone securing a rental unit
for them.
Black Markets Under rent controls, black markets may emerge, with units renting for more than $750. This occurs because prospective tenants are willing to pay more than the legal price for an
apartment. In Figure 2.8, when the quantity Q1 is legally available, tenants are willing to pay as much as P1—the height of the demand curve at that quantity—for an apartment. Landlords benefit by renting a unit at more than $750, so there is room for transactions that
benefit both consumers and producers—which is why black market exchanges are likely to
occur. The extent of black market activities will depend on the penalties the government
applies to this behavior and how rigorously the penalties are enforced.
While most local governments vigorously police against black market exchanges
between landlords and initial tenants, they often look the other way when it comes to
the practice of subletting. Subletting occurs when the initial tenant rents the apartment to a secondary tenant (generally without the landlord’s knowledge). New York City does
not actively police against this practice, and as a result most subletting rental rates are
substantially greater than the official rent control rates for the same apartments. Since
subletting is essentially a black market in rental units, to the extent that cities condone
the practice, the rental rates paid by tenants (at least secondary tenants) can end up higher
than they would be were there no rent control—P1 versus P in Figure 2.8. In 2011, the average monthly rent in New York City was $1,160 for rent-regulated apartments, the
overall average rent for available apartments was over $3,000, and the black market rate
was above $4,000.9
black market an illegal market for a good
8Landlords in rent-controlled cities often screen for and attempt to avoid renting to law students. It seems that such students are prone to practice their future trade on landlords. 9http://www.businessinsider.com/the-8-reasons-why-new-york-rents-are-so-ridiculously-high-2013-7.
Government Intervent ion in Markets : Pr ice Controls 29
Rent controls also involve administrative costs. For example, in Santa Monica, Califor-
nia, the Rent Control Board’s annual budget exceeded $5 million in certain years. To pay
for this budget, each tenant was levied an annual fee of $132. Furthermore, since property
taxes are the most important source of revenue for most cities, and rent control lowers the
market values of rental properties, a city’s tax base is eroded. This loss can be substantial.
It has been estimated, for example, that property tax revenues fell by 10 to 20 percent in
Cambridge, Massachusetts, as a result of rent controls.11 This means that residents, including
tenants, either receive fewer services provided by the city or have to pay higher income taxes.
For all these reasons, the benefits to tenants from rent control are likely to be a good deal
smaller than they would appear on the surface. Generalizing the outcome for tenants as a
group is difficult because particular tenants are likely to be affected in different ways. Prob-
ably some tenants benefit, especially those occupying rental units at the time rent control
takes effect, but the deterioration of the quality of the units diminishes even their benefits.
Other tenants are almost certainly worse off because of the side effects that accompany the
lower rents.
APPLICATION 2.5
A key element in the debate over the U.S. health-care reform legislation passed in 2010 was the extent to which policymakers would have the ability to limit the prices charged by insurance companies and/or health care pro- viders.10 During the debate on the legislation, for example, President Obama proposed giving the federal government the power to regulate insurance premiums.
As our analysis has shown, price ceilings create short- ages as producers are incented to cut back on the quantity supplied—in the case of health-care insurers, either by dropping their most expensive customers and/or cutting back on their reimbursement rates to hospitals and physi- cians. Had the price ceilings remained in the final legislation, the results would have been as predictable as for the case of rent control.
For example, Medicare reimburses only 80 cents per dollar of expenses, and Medicaid pays even less. As a con- sequence, over one-third of all physicians have closed their practice to Medicaid patients, and 12 percent do not service individuals covered by Medicare.
Were insurers to clamp down more tightly on reim- bursements to physicians and hospitals, the expected out- comes would include some hospitals shutting down and a greater number of physicians opting to retire early, pursue
Health Care Reform and Price Controls
alternative career choices, or switch to providing “concierge- type” services to a select group of privately paying patients.
Britain and Canada impose strict limits on reimburse- ments to hospitals and physicians as part of their universal “free” health care. The following are some of the effects in these countries: 750,000 Britons waiting to be admit- ted to hospitals each year, with 50,000 surgeries being postponed annually because the person on the waiting list has become too sick for the surgery to proceed; almost 800,000 people on the waiting list for care in Canada annually, many in chronic pain; routine “hallway medicine” for the patients who are admitted to Canadian hospitals but are either waiting for attention or recuperating from opera- tions on hallway stretchers with the stretcher locations, in many cases, having permanent numbers; and an estimated 20 percent of heart attack patients in Vancouver hospitals who should be treated in 15 minutes having to wait at least 1 hour.
There is no comparable crisis in dental and veterinary care in either Britain or Canada because those sectors still operate without government price controls. As Michael Bliss, a medical historian, notes, “So we have the absurdity in Canada that you can get faster care for your gum disease than your cancer, and probably more attentive care for your dog than your grandmother.” Indeed, in the Canadian prov- ince of Ontario, one man on a lengthy waiting list for mag- netic resonance imaging (MRI) tests reserved a session for himself at a private animal hospital with the same machine. He registered under the name Fido.
10This application is based on “Full Hospitals Make Canadians Wait and Look South,” New York Times, January 16, 2000, p. 3; and Michael D. Tanner, “Price Controls by Any Other Name,” New York Post, February 23, 2010.
11Peter Navarro, “Rent Control in Cambridge, Massachusetts,” The Public Interest, 78 (Winter 1985), pp. 83–100.
30 Supply and Demand
2.5 Elasticities We have so far focused on specifying qualitative relationships between determinants of sup-
ply and demand and the actual quantity demanded and quantity supplied of a good. Although
qualitative relationships provide meaningful information, they cannot measure the impact
produced by a change in a particular determinant on the quantity demanded or supplied of a
commodity. In the business world especially, we need to know the quantitative impact of a
change in one determinant such as price, income or the price of inputs on the quantity
demanded or quantity supplied of a commodity. Quantitative impacts are also often impor-
tant in the public policy arena. For instance, when considering an increase in the sales tax on
cigarettes, government decisionmakers may be concerned about the magnitude of the effect
of the tax on the quantity of cigarettes demanded by smokers (perhaps, in particular, by teen-
age smokers). Elasticities measure the magnitude of the responsiveness of any variable (such as quantity demanded and quantity supplied) to a change in particular determinants.
Price Elasticity of Demand Even though we assume that all market demand curves have negative slopes (implying that
at a lower price a greater quantity will be purchased), the degree of responsiveness varies
widely from one commodity to another. A reduction in the price of cigarettes may lead
to an infinitesimal increase in purchases, while a reduction in airplane fares may produce
a veritable explosion in air travel. The law of demand tells us to expect some increase in quantity demanded, but not how much.
The price elasticity of demand is a measure of how sensitive quantity demanded is to a change in a product’s price. It can be defined as the percentage change in quantity demanded divided by the percentage change in price. The ratio will always be negative for any downward-sloping demand curve. For example, if a 10 percent price increase
brings about a 20 percent reduction in quantity demanded, the price elasticity of demand
elasticities measures of the magnitude of the responsiveness of any variable (such as quantity demanded or supplied) to a change in particular determinants
price elasticity of demand a measure of how sensitive quantity demanded is to a change in a product’s price
APPLICATION 2.6
Roughly 100,000 people are currently waiting for organ transplants in the United States.12 More than 6,000 of these individuals will die annually, at least in part because of a price ceiling of zero imposed on potential organ suppliers. Under the existing U.S. policy, altruistic organ donations are encouraged, whereas it is a criminal offense to sell organs, either during life or after death.
To deal with the increasingly acute shortage of organs for transplant, there has been some hope in recent years
Price Ceilings Can Be Deadly for Buyers
that the American Medical Association will launch an experimental program permitting the payment of small sums of money ($300 to $3,000) for cadaveric organs. Professor Richard Epstein of NYU’s Law School has been the leading advocate for such an experimental program arguing that it is time to “enact, not repeal, the laws of supply and demand.”
Professor Gary Becker of the University of Chicago’s Booth School of Business estimates that not only would many lives be saved each year through permitting a mar- ket for organs but the price would also decline substantially to $15,000 per kidney transplant and $35,000 per liver transplant. Currently, with organ sales being prohibited, a kidney transplant costs in the range of $100,000, while the going rate for a liver transplant is $175,000.
12This application is based on Gary S. Becker, “Should the Purchase and Sale of Organs for Transplant Surgery be Permitted?” Capital Ideas, University of Chicago Graduate School of Business, April 2006, pp. 4–7; and Richard A. Epstein, “The Market Has a Heart,” Wall Street Journal, February 21, 2002, p. A18.
Elast ic i t ies 31
is −20 percent/+10 percent, or −2.0. Economists usually drop the minus sign on the under- standing that price and quantity demanded always move in different directions and simply
refer to the elasticity as being, in this case, 2.0.
Price elasticity of demand provides a quantitative measure of the price responsiveness of
quantity demanded along a demand curve. The higher the numerical value of the elasticity,
the larger the effect of a price change on quantity. If the elasticity is only 0.2, then a 10 per-
cent price increase will reduce quantity demanded by just 2 percent:
2 percent 1 percent = 2./ .0 0
Alternatively, if the elasticity is 4.0, a 10 percent rise will reduce quantity demanded by
40 percent:
4 percent 1 percent = 40 0 0/ . .
If the price elasticity of demand exceeds 1.0, then demand is said to be elastic. Elasticity is greater than 1.0 whenever the percentage change in quantity demanded is greater than the
percentage change in price, implying that the quantity demanded is relatively responsive to
a price change. If the price elasticity of demand is less than 1.0, then demand is said to be
inelastic. Elasticity is less than 1.0 whenever the percentage change in quantity demanded is less than the percentage change in price, implying that quantity demanded is relatively
unresponsive to a price change. When the price elasticity of demand is equal to 1.0, then
demand is said to be unit elastic, or of unitary elasticity. Unitary elasticity occurs when- ever the percentage changes in price and quantity demanded are equal.
Whether demand is elastic, unit elastic or inelastic determines how a price change will
affect total expenditure on the product. Total expenditure equals price times quantity, or
P × Q. A change in price affects these terms in offsetting ways. A higher price increases the P term but reduces the Q term (quantity demanded is lower at a higher price). The net effect on total expenditure, therefore, depends on the relative size of the two changes. Put
differently, the net effect on total expenditure depends on how responsive quantity is to the
price change; it depends on the price elasticity of demand. If a 10 percent increase in price
reduces quantity by 10 percent (the unit elastic case), then total expenditure, P × Q, remains unchanged. If a 10 percent increase in price reduces quantity by more than 10 percent
(the elastic demand case), then total expenditure will fall because of the sharper reduction
in quantity purchased. Finally, if a 10 percent increase in price reduces quantity by less
than 10 percent (the inelastic demand case), then total expenditure will rise.
Figure 2.9a depicts a case where a small change in the per-gallon price of gasoline has
a large effect on quantity purchased. Demand is elastic in this case since the percentage
change in quantity demanded exceeds the percentage change in price. If, for example, price
falls from $3.00 to $2.75 per gallon, quantity increases sharply from 100 to 200 gallons.
The price reduction increases total expenditure on gasoline from $300 ($3.00 per gallon
multiplied by 100 gallons) to $550 ($2.75 multiplied by 200 gallons). Conversely, if the
price rises from $2.75 to $3.00, total expenditure falls from $550 to $300. Thus, we see
graphically how a price change affects total expenditure when demand is elastic.
Figure 2.9b examines the relationship between price and total expenditure if the demand
for gasoline is very inelastic: a change in price has little effect on quantity. When the price
falls from $3.00 to $1.50 per gallon, total expenditure falls from $300 to $180. When price
rises from $1.50 to $3.00, total expenditure rises. Figure 2.9c shows the intermediate case
of unit elasticity.13 In this case total expenditure remains unchanged when price varies.
elastic the situation in which price elasticity of demand exceeds 1.0 or unity
inelastic the situation in which price elasticity of demand is less than 1.0 or unity
unit elastic the situation in which price elasticity of demand equals 1.0 or unity
13Because the product of price and quantity is unchanged at all points along a demand curve with unit elas- ticity, such a curve must satisfy the equation P × Q = k (a constant). This equation describes a rectangular hyperbola.
32 Supply and Demand
Consumers purchase 100 gallons at a price of $3.00 per gallon (total expenditure of $300)
and 200 gallons at a price of $1.50 per gallon (total expenditure still $300).
In short, when demand is elastic (elasticity greater than 1.0), price and total expenditure
move in opposite directions. When demand is inelastic (elasticity less than 1.0), price and
total expenditure move in the same direction. And when demand is unit elastic, total expen-
diture remains constant when the price varies.
Calculating Price Elasticity of Demand Calculating price elasticity of demand from a pair of price–quantity points is frequently
necessary. Suppose that we are given the following price–quantity values for gasoline
(where quantity demanded is measured in gallons):
P P Q Q
1 2
d1 d2
3. 2 97
1 1 5.
= = = =
$ $ .
, ,
00
000 00
Our definition of price elasticity of demand is the percentage change in quantity demanded
divided by the percentage change in price. This relationship is expressed as a formula, let-
ting η (the Greek letter eta) stand for price elasticity of demand:
η = ( / ) ( / )
. ΔQ Q
P P d d
Δ
Here, ΔQ Qd d/ is the percentage change in quantity demanded, and ∆P/P is the percentage change in price.14 In applying this formula—called the point elasticity formula—we encounter an ambiguity. While ∆Qd and ∆P are unambiguously determined (a 5 gallon change in quantity and a $0.03 change in the per-gallon price), what values should be used
for Qd and P? If we enter the values for P1 and Qd1 into the formula, we obtain:
( / )
( / )
( / , )
($ . / $ . ) . .
Δ Δ Q Q P P d d1
1
5 1 000
0 03 3 00 0 50= =
point elasticity formula
= ( / ) ( / )
Δ Δ Q Q P P d d
14The Greek letter ∆ (delta), as in ∆Qd, simply means “change in.”
Price Elasticity of Demand and Total Expenditure (a) If demand is elastic, a lower price increases total expenditure. (b) If demand is inelastic, a lower price decreases total expenditure. (c) If demand is unit elastic, a lower price leaves total expenditure unchanged.
$3.00 $2.75
0
D
100 200 Gallons of gasoline
Price
$3.00
$1.50
0
D
100 120 Gallons of gasoline
Price
$3.00
$1.50
0
D
100
(a) (b) (c)
200 Gallons of gasoline
Price
Figure 2.9
Elast ic i t ies 33
Alternatively, if we use P2 and Qd2, we obtain:
( / )
( / )
( / , )
($ . / $ . ) . .
Δ Δ Q Q P P d d2
2
5 1 005
0 03 2 97 0 49= =
Because we are dealing with small changes in this case, which values we choose makes
little quantitative difference. There is, however, a slight difference, and it reflects the fact
that the percentage change between two prices depends on the direction of the change. If
price falls from $3.00 to $1.50 per gallon, this is referred to as a 50 percent decrease (a
$1.50 change in price divided by the initial price, $3.00). Alternatively, if the price rises from $1.50 to $3.00 per gallon, this is a 100 percent increase (a $1.50 change in price
divided by the initial price, $1.50). Don’t be sidetracked by this arithmetical obscurity. The important point is that some base Qd and P must be employed in the formula, but for small changes in Qd and P, which base is chosen makes no significant difference to the results.
There is a substantial difference, however, when a large change in price and quantity is
involved. Suppose, for example, that we have the following values:
P P Q Q
1 2
d1 d2
3.
1
= = = =
$ $ .
, , .
00 1 50
000 2 000
By inspection we see that total expenditure is $3,000 (quantity demanded is once again
being measured in gallons) at both prices, so we know that demand is unit elastic. Surpris-
ingly, though, it now makes a great deal of difference what base values of P and Qd we use if we try to apply the point elasticity formula:
( / ) / ( / ) ( , / , ) / ($ . / $ . ) .
(
Δ Δ Δ
Q Q P P Q
d d1
d
and1 1 000 1 000 1 50 3 00 2 0= = // ) / ( / ) ( , / , ) / ($ . / $ . ) . .Q P Pd2 Δ 2 1 000 2 000 1 50 1 50 0 5= =
According to one calculation, price elasticity of demand is 2.0; according to the other it is
0.5. Both are wrong, and the true value, unity, lies between these estimates. The basic
problem in this case is that the elasticity of demand tends to vary from one point (one P, Qd combination) to another on the demand curve, and for a large change in price and quantity we need an average value over the entire range. Consequently, when we
deal with large changes in price and quantity, we should use the following arc elasticity formula:
η = ( ) +
⎡
⎣ ⎢
⎤
⎦ ⎥
( ) + ⎡
⎣ ⎢
⎤
⎦ ⎥
Δ
Δ
Q Q Q
P P P
d
d1 d2 1
2
1 2 1 2
( )
( )
Note that this formula differs from the point elasticity formula only in using the average
of the two quantities, (1/2)(Qd1 + Qd2), and the average of the two prices, (1/2)(P1 + P2). Applying this formula to the preceding figures yields the true value of the elasticity over
the entire range of prices considered:
Δ
Δ
Q Q Q
P P P
d
d1 d2 1
2
1 2
1 2
1 2
1 000
1 000( ) + ⎡
⎣ ⎢
⎤
⎦ ⎥
( ) + ⎡
⎣ ⎢
⎤
⎦ ⎥
= ( ) +( )
( )
,
( , 22 000
1 50
1 50 3 00
1 0
1 2
, )
$ .
($ . $ . )
. .
⎡
⎣ ⎢
⎤
⎦ ⎥
( ) + ⎡
⎣ ⎢
⎤
⎦ ⎥
=
arc elasticity formula
= ( ) +
⎡
⎣ ⎢ ⎢
⎤
⎦ ⎥ ⎥
( ) + ⎡
⎣ ⎢ ⎢
⎤
⎦ ⎥ ⎥
Δ
Δ
Q Q Q
P
P P
d
d1 d212
1 2 1 2
( )
( )
34 Supply and Demand
Thus, we have two formulas. The first works well when small changes in P and Qd are involved because, in that case, which P and Qd are used makes little difference. The second formula avoids the problem of having to pick one specific point by using the average values
of price and quantity demanded and should be used with large changes in price and quan-
tity demanded.
Demand Elasticities Vary among Goods We can never know why people respond exactly as they do. Nonetheless, two general fac-
tors seem to have a pronounced effect on the elasticity of demand for a particular product.
The first, and most important, factor is the availability and closeness of substitutes. The more substitutes there are for a product, and the better the substitutes, the more elastic will be the demand for the product. When there are good substitutes, a change in the price of the product will lead to considerable substitution among products by consumers. The demand
for margarine, for example, would probably be quite elastic, because when its price rises,
many people would switch to butter. Remember that when we evaluate the elasticity of
demand for margarine, we assume the price of butter to be unchanged. Thus, a higher price
of margarine (with an unchanged price of butter) leads people to shift from margarine to
butter, because they are close substitutes.
The degree to which a good has close substitutes depends in part on how specifically it
is defined. A narrowly defined good will frequently have close substitutes, and elasticity
will tend to be higher. For example, the elasticity of demand for Dial soap (a very narrowly
defined good with many substitutes) will be greater than the elasticity of demand for soap.
Along these lines, the demand for any particular brand of some product (such as Cheerios
cereal) will be more elastic than the demand for all brands taken together (cereal).
APPLICATION 2.7
Cable systems historically have offered two tiers of ser- vice: a “basic” tier including programming such as CNN, ESPN, and MTV; and “pay” tiers featuring movie channels such as HBO and Showtime. Subscribers purchase the basic tier as a package. Pay tiers are available separately for an additional fee per channel.
While pay tiers have never been subject to govern- ment rate control, basic tiers were regulated prior to 1987. In 1987, basic rates were deregulated, and cable operators had to determine what price they should charge. Elasticity of demand considerations helped them decide.
As we saw earlier in this section, whether demand is inelastic, unit elastic or elastic determines how a price change will affect the total expenditure on a product. Calculating basic tier demand elasticity was thus vital to operators inter- ested in finding out whether the rates they were charging under regulation were too low in terms of maximizing profit.
Studies conducted by cable operators found that basic tier demand elasticity at the time of rate deregulation in 1987 was less than unity—between 0.1 and 0.5. The esti- mated elasticity indicated that profits would increase
Demand Elasticity and Cable Television Pricing
if basic rates were raised. Two factors are at work here. First, if a firm is operating along an inelastic portion of its demand curve, the total revenue earned from subscribers will increase if the rate is raised. The effect on revenue of a decrease in the number of subscribers will be more than compensated for by the higher rate earned per remaining subscriber when demand is inelastic. Second, at the higher rate, fewer subscribers will be served, and the total cost of providing service will be lower. Between the increase in total revenue and the decrease in total cost, the profit earned on basic service will increase (profit is the difference between total revenue and total cost). In sum, operators stood to lose some subscribers but to more than make up for that from the ones they kept.
Relying on estimates that demand elasticity was less than unity, operators began to raise rates following dereg- ulation. Average basic rates increased 75 percent between 1987 and 1991—an amount significantly greater than the inflation rate, even after adjustments made for improve- ments in programming quality. Operators’ profits on basic service increased along with the rates.
Elast ic i t ies 35
In addition to the number and the quality of substitutes, a second factor that can be impor-
tant in determining elasticity of demand is the time period over which consumers adjust to
a price change. The longer the time period involved, the fuller is the adjustment consumers can make. In part, this reflects the fact that it takes time for consumers to learn about a price change, but there are other reasons, too. Consider an increase in the price of electricity. In
the month following the price increase, people can cut back their use somewhat by switch-
ing lights off more conscientiously, turning down thermostats (if electric heating is used)
or turning off air conditioners. The number of ways people can economize on electricity,
however, is greater when we consider what they can do over a longer period. Over a year,
for example, they can substitute lower-wattage light bulbs for existing light bulbs, convert
electric furnaces to oil or gas furnaces, insulate their houses, use portable kerosene heaters,
buy appliances that require less electricity, and so on. In short, demand will be more elastic
the longer the time period over which consumers can adjust and, in essence, find substitutes.
For many goods, consumers will not require much time to make a full adjustment to a
change in price. In these cases, the long-run and short-run responses will not differ substan-
tially. Changes in the prices of electricity and gasoline will necessitate major alterations in
the consumption of very durable goods (houses, appliances, and cars) before consumers
have fully adjusted, and those alterations take time. But for most goods (beer, shoes, wrist-
watches, meat, televisions, compact discs, and so on) we would not expect the short- and
long-run elasticities of demand to be much different.
The Estimation of Demand Elasticities In practice, estimating elasticities of demand is problematic because elasticity of demand
refers to a given demand curve, but the demand curve itself is likely to shift over time. Econ- omists have developed some sophisticated techniques (briefly overviewed in Chapter 4)
to deal with this problem and to permit estimation of demand elasticities. Table 2.1 lists
some selected estimates of elasticities of demand for a variety of products.
As Table 2.1 indicates, the estimates of demand elasticities differ widely among goods. Not
surprisingly, the demand for cigarettes is inelastic (an elasticity of 0.35). Although people com-
monly think of medical care consumption as almost totally unresponsive to price, the elasticity
of demand for physicians’ services is 0.6. While this demand is inelastic, it does imply that a
50 percent increase in price would reduce consumption by fully 30 percent. Some products,
such as air travel and automobiles, are apparently in highly elastic demand (in the long run).
Selected Estimates of Demand Elasticities
Short Run Long Run Cigarettes — 0.35
Water — 0.4
Beer — 0.8
Physicians’ services 0.6 —
Gasoline 0.2 0.5–1.5
Automobiles — 1.5
Chevrolets — 4.0
Electricity (household utility) 0.1 1.9
Air travel 0.1 2.4
Sources: Hendrik S. Houthakker and Lester D. Taylor, Consumer Demand in the United States, 1929–1970 (Cambridge, MA: Harvard University Press, 1966 and 1970 editions); Kenneth G. Elzinga, “The Beer Industry,” in The Structure of American Industry, edited by Walter Adams (New York: Macmillan, 1977); and James L. Sweeney, “The Response of Energy Demand to Higher Prices: What Have We Learned?” American Economic Review, 74, No. 2 (May 1984), pp. 31–37.
Table 2.1
36 Supply and Demand
One implication of these estimates (and many others that could be cited) should not be
missed: they all support the law of demand. Consumers purchase more at a lower price,
other things being equal.
APPLICATION 2.8
As many as 5 million Canadians annually drive south to the United States to catch a flight to a location that often also can be reached with a flight originating from a Canadian airport.15 The growing cross-border migration is evidenced by the fact that American cities from Bellingham, Washington to Buffalo New York are expanding airport capacity so as to accommodate rising traffic from Canada. In 2013, for example, Canadians accounted for 47 percent of all passengers at Buffalo Niagara International Airport, up from 28 percent in 2006. Plattsburgh airport, which was an Air Force based during the Cold War and is in a remote part of upstate New York, now touts itself as “Montreal’s U.S. airport” and reports 155,000 annual passenger boardings, 83 percent of whom are Canadian.
Why the trend? For one thing, due to the greater over- all competitiveness of the market, fares charged by airlines on flights originating in the United States are generally 30 percent cheaper than those originating in Canada. Low-cost carriers such as Southwest Airlines, Allegiant, Spirit, and Jet Blue do not operate in Canada. The result of such extra
Why Canadians Are Flying South of the Border
competitive pressure is that a ticket to an American desti- nation from Canada costing $200 would cost $140 from an American location.
In addition, the Canadian government has been levying increasingly higher fees and taxes on top of airlines’ pub- lished fares. As of 2013, $81.13 in taxes and fees would be imposed on a $200 ticket to an American destination from Canada versus $32.42 on the average published fare of $140 from an American airport.
The upshot of the foregoing is that the price elastic- ity of demand faced by Canadian airlines will be larger the more consumers in Canada can avail themselves of substi- tute air routings south of the border. Moreover, the extent to which taxes and fees reduce the quantity demanded of air travel from Canadian airports also will be greater the better are the air travel substitutes available south of the border. When attempting to raise taxes and fees, therefore, policymakers who fail to take into account the extent of cross-border substitutes will underestimate the impact of taxes and fees on quantity demanded as well as overestimate the revenue to be raised from higher taxes and fees.15“Travelers From the North,” New York Times, July 16, 2013, p. B6.
Three Other Elasticities Price elasticity of demand is the most important elasticity concept in economics, but we
can define elasticity in general as a measure of the response of any variable to the change in
some other variable. Two other common elasticity measures that relate to consumer behav-
ior are the income elasticity of demand and the cross-price elasticity of demand. On the
supply side, the most important elasticity is price elasticity of supply. The price elasticity of
supply measures the responsiveness of quantity supplied to price.
Income Elasticity The income elasticity of demand measures how responsive consumption of some item is to a change in income (I), assuming that the price of the good itself remains unchanged. We
define income elasticity as the percentage change in consumption of a good Qd divided by the percentage change in income or (in point elasticity form):
Income elasticity of demand for a good
d d= ( )Δ
Δ Q Q
I I /
/ .
For example, if income rises by 10 percent, and a consumer increases purchases of gasoline
by 5 percent, then the income elasticity of gasoline is 0.5. Note that the algebraic sign of
income elasticity of demand a measure of how responsive consumption of some item is to a change in income, assuming the price of the good itself remains unchanged
Elast ic i t ies 37
this elasticity distinguishes between normal and inferior goods. Whenever income elastic-
ity is positive, consumption of the good rises with income, so the good must be normal.
Whenever income elasticity is negative, consumption of the good falls when income rises,
and the good must be inferior. A unitary income elasticity means the consumer continues to
spend the same percentage of income on the good when income rises.
Cross-Price Elasticity of Demand The cross-price elasticity of demand measures how responsive consumption of one good is to a change in the price of a related good. We define cross-price elasticity as the percent- age change in consumption of one good, X, divided by the percentage change in the price of a different good, Y, or (in point elasticity form):
Cross-price elasticity of demand for with respect to theX price of d dY Q Q
P P X X
Y Y = ( / )
( / ) .
Δ Δ
For example, if the price of BMW automobiles rises by 10 percent, and the quantity of
Mercedes cars purchased increases by 5 percent, then the cross-price elasticity of demand
for Mercedes cars with respect to the price of BMWs is 0.5. Note that cross-price elastic-
ity will be positive when the goods are substitutes (as are BMW and Mercedes cars) and
negative when the goods are complements (for example, gasoline and cars). Indeed, the
major use of this elasticity is to measure the strength of the complementary or substitute
relationship between goods. The concept of cross-price elasticity is widely used in antitrust
cases. How a market is defined and how competitive it is depend on the availability of sub-
stitutes. One way to ascertain substitutability is with a measure of the cross-price elasticity
of demand.
Elasticity of Supply The price elasticity of supply, or elasticity of supply, is a measure of the responsiveness of the quantity supplied of a commodity to a change in the commodity’s own price. It is
defined as the percentage change in quantity supplied, Qs, divided by the percentage change in price. Using the Greek letter ε (epsilon) to represent price elasticity of supply, we can express it as (in point elasticity form):
ε = ( / )
( / ) .
Δ Δ Q Q P P
s s
Any upward-sloping supply curve—the increasing per-unit cost case—has a positive elas-
ticity of supply because price and quantity supplied move in the same direction. If per-unit
production costs are constant, the supply curve is horizontal, and the price elasticity of sup-
ply is infinity. For example, the supply of dimes in terms of nickels is a horizontal curve
with a height of 2 at most banks. (Banks are willing to provide you an additional dime so
long as you give them 2 nickels per dime.) As the price of dimes rises from 1.99 nickels
(a rate at which banks would be unwilling to sell you any dimes in exchange for nickels)
to 2 nickels (a rate at which banks would become willing to supply you quite a few dimes
in exchange for nickels), the percentage change in quantity supplied (from zero to a lot
of dimes) is infinite relative to the percentage change in price (from 1.99 to 2 nickels per
dime).
At the opposite extreme, if supply is entirely unresponsive to price, the supply curve is
vertical and the elasticity of supply is equal to zero. For example, no matter how high the
price gets, it is impossible to produce more original Picasso paintings (although several
imposters have attempted to copy the dead artist’s style and pass off the result as Picasso
originals). The responsiveness of the quantity of Picasso paintings supplied to increases (or
decreases) in the price of Picasso paintings is thus zero.
cross-price elasticity of demand a measure of how responsive consumption of one good is to a change in the price of a related good
price elasticity of supply a measure of the responsiveness of the quantity supplied of a commodity to a change in the commodity’s own price
38 Supply and Demand
APPLICATION 2.9
Small fluctuations in oil supply or demand cause large gyrations in prices over the short run. For example, after the price of oil fell to below $12/barrel in 1998, the three largest exporters of oil to the United States—Saudi Arabia, Venezuela, and Mexico—reached an agreement, known as the Riyadh Pact, to reduce their collective output by 1.5 to 2.0 million barrels a day, about 2 to 3 percent of the 73-million-barrel-per-day production at the time. The price of oil shot above $30 in the wake of the pact—an increase of 150 percent in response to the 2 to 3 percent supply cutback. Similarly, between April and June 2009, unrest in Nigeria resulted in that country’s oil production declining by 0.84 million barrels per day (or 1 percent of the prevailing world supply). The relatively small supply decrease caused a 75 percent increase in the price of oil— from $40 to over $70 per barrel.
Why do small oil supply/demand shocks cause such large gyrations in prices? The answer has to do with the short-run price elasticities of oil supply and demand, which are both relatively small and close to zero (see Table 2.1, for example, for the estimated price elasticity of demand for gasoline). This is because in the short run, consumers can- not readily adjust their energy usage patterns. Commut- ing patterns, motor vehicle options, homes and heating oil choices, and production processes are hard to alter quickly.
Price Elasticities of Supply and Demand and Short-Run Oil Price Gyrations
As a result, a relatively small oil supply cutback necessitates a significantly higher price in order to ensure the requisite reduction in quantity demanded.
Similarly, the short-run price elasticity of supply is also quite small because the quantity supplied of oil does not increase markedly over the near term in response to even relatively large increases in price. It takes time to identify new sources of crude oil and/or develop the means to extract new sources and bring them to market. For example, wells need to be dug and pipelines constructed. A small cutback in supply or increase in demand thus requires a large increase in price in order to equilibrate the market.
An analogy from airline travel might help drive home the point. Think about when a flight is overbooked even by as little as one or two seats. The airline might end up offer- ing extra travel vouchers that are at least 100 percent the amount of the original fare to the passenger(s) willing to be rebooked, at no cost, on a later flight. A small decrease in available seats leads to such a relatively large increase in price because the own-price elasticities of demand and supply are both small. Travelers cannot readily switch to the later flight (sometimes on the following day) due to professional and/or personal commitments, while pro- ducers cannot easily produce extra, substitute seats over the short run.
Finally, as in the case of elasticity of demand, when the ratio of the percentage change in
quantity supplied to the percentage change in price is greater than unity, we say that supply
is elastic. When supply is elastic, an increase in price produces a more than proportionate increase in quantity supplied. When the elasticity of supply is less than unity, supply is
inelastic and a higher price produces a less than proportionate increase in quantity supplied. When the ratio equals unity, supply is unit elastic and a higher price produces a proportion- ate increase in quantity supplied.
SUMMARY
Most economic issues involve the workings of indi-
vidual markets.
In the supply–demand model, we analyze the behav-
ior of buyers by using the demand curve.
The demand curve shows how much people will pur-
chase at different prices when other factors that affect
purchases are held constant. The demand curve slopes
downward, reflecting the law of demand.
Analysis of the seller side of the market relies on the
supply curve, which shows the amount that firms will
offer for sale at different prices, other factors being con-
stant. The supply curve typically slopes upward.
The intersection of the demand and supply curves,
reflecting the behavior of buyers and sellers, identifies
the equilibrium price and quantity.
A shift in the supply or demand curve produces a
change in the equilibrium price and quantity.
For the market mechanism to operate, price must
be free to adjust to any change affecting the behav-
ior of buyers and sellers in the market. Thus, when the
Review Quest ions and Problems 39
government steps in to regulate prices, the market does
not function in the same way.
A government-imposed price ceiling results in a
shortage and may lead to diminution in product qual-
ity, nonprice rationing, black markets, administrative
costs, and increased demand for and supply of substitute
goods. Sellers are clearly harmed by the imposition of a
price ceiling, and the effect on buyers as a group may not
be beneficial.
Elasticities provide a quantitative measure of the
magnitude of the responsiveness of quantity demanded
or supplied to a change in some other variable.
The most important elasticity in economics is price
elasticity of demand, which measures how responsive
the quantity demanded of a commodity is to a change in
the commodity’s own price. It is measured by the per-
centage change in quantity demanded divided by the per-
centage change in price.
When price elasticity exceeds unity, demand is elas-
tic, and a lower price expands purchases so sharply that
total expenditure rises.
When price elasticity is less than unity, demand is
inelastic, and a lower price leads to a reduction in total
expenditure.
When price elasticity equals unity, demand is unit elas-
tic, and total expenditure is unchanged at a lower price.
Three other important elasticities are the income
elasticity of demand, cross-price elasticity of demand,
and price elasticity of supply. They are constructed in
a manner analogous to that employed to construct price
elasticity of demand and measure, respectively, the
responsiveness of quantity demanded to income, the
responsiveness of quantity demanded of one good to the
price of a related good, and the responsiveness of the
quantity supplied of a commodity to the commodity’s
own price.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
*2.1 A newspaper article points out that the price of economics textbooks is up 10 percent this year over last year, and yet the
number of textbooks sold is higher this year. The article claims
that these figures show that the law of demand does not apply
to textbooks. Is there a flaw in this argument?
2.2 A demand curve is drawn holding “other things constant.” What does the “other things constant” provision mean, and why
is it important to a correct interpretation of the law of demand?
2.3 “Because demand curves and supply curves are always shifting, markets can never attain an equilibrium.” Does this
imply that the concept of equilibrium is not useful?
2.4 If we know that the (real) price of tennis rackets is higher now than last year, can we conclude that the demand curve
shifted out over the year? Explain.
2.5 The supply and demand schedules for apples are as follows:
Demand Supply
Price per Pound
Quantity Demanded per
Year Price per
Pound
Quantity Supplied per
Year $0.90 100,000 $0.60 100,000
0.80 110,000 0.70 120,000
0.70 120,000 0.80 140,000
0.60 135,000 0.90 150,000
Use graphs to answer the following questions. a. What is the market equilibrium price and quantity?
b. The government agrees to purchase as many pounds of apples as growers will sell at a price of $0.80 per pound.
How much will the government purchase, how much will
consumers purchase, and how much will be produced?
c. Suppose the government policy in part b remains in effect, but consumer demand increases by 10 percent (consumers
will purchase 10 percent more at each price than they did
before). What will be the effect on total apple output, pur-
chases by consumers, purchases by the government, and the
price of apples?
2.6 In economics, what do we mean by the term shortage? In unregulated competitive markets, are there ever shortages?
Have you ever heard noneconomists use the term shortage with a different meaning?
2.7 “A decrease in supply will lead to an increase in the price, which decreases demand, thereby lowering price. Thus, a
decrease in supply has no effect on the price of a good.” Evalu-
ate this statement.
*2.8 Consider the market for taxi service in a city. Explain, by using supply and demand curves, how each of the following
actions will affect the market. (Consider each case separately.)
a. Bus drivers go on strike. b. Bus fares increase after a strike by bus drivers. c. Taxi drivers must pass a competency test, and one-third fail. d. Gasoline prices increase. e. Half the downtown parking lots are converted to office
buildings.
f. The population of the city increases. 2.9 As a regional manager for American Airlines, you have recently undertaken a survey of economy-class load factors
(the percentage of economy-class seats that are filled with
40 Supply and Demand
paying customers) on the Chicago–Columbus, Ohio, route that
you service. The survey was conducted over five successive
months. For each month, data collected include the one-way
fare you charge per economy seat, the price charged by rival
United Airlines, the average (monthly) per capita income in
the combined Chicago–Columbus market, and the average
economy-class load factor for both American and United
Airlines. Assume that all other factors (the price charged by
Southwest Airlines, the number of flights, the size of planes
flown, and so on) have remained constant.
Month AA
Price UA
Price Income
AA Load
Factor
UA Load
Factor 1 $110 $112 $2,000 65 60
2 109 110 1,900 62 63
3 110 112 2,100 70 66
4 109 111 1,900 70 61
5 108 110 1,900 68 59
a. On the Chicago–Columbus route, identify the arc price elasticity of demand for American economy seats, the arc
income elasticity of demand for American economy seats,
and the arc cross-price elasticity of demand for American
economy seats with respect to United prices.
b. Based on the data that you have collected, is United a substitute or complement for American in the Chicago–
Columbus market? Explain.
c. Are American’s economy seats a normal or inferior good in the Chicago–Columbus market? Explain.
d. Would the estimated demand elasticity for your product be larger or smaller if consumers had been given more time to
respond to any price change (for example, one year versus
one month)?
e. Compared with the price elasticity of demand for United and American, is the demand elasticity for economy seats
in general in the Chicago–Columbus market (regardless of
which airline provides them) larger or smaller? Explain.
*2.10 Suppose that the demand elasticity for cigarettes is equal to 2.0. If the demand elasticity for Camel cigarettes is equal to
6.0, must there be at least some cigarette brands with a demand
elasticity less than 2.0?
2.11 Suppose that the typical economics student is interested in consuming (and spends all her money on) only two com-
modities: economics study guides and horror movie passes.
An unlimited supply of horror movie passes is available
at a price of $5 per pass, while an equally unlimited supply
of study guides covering endlessly different (and interest-
ing) nuances in economics can be purchased for $30 each.
The student currently purchases 20 horror movie passes and
10 study guides per semester. If the typical student’s price
and income elasticities of demand for horror movie passes
are both unity, what is the student’s cross-price elasticity of
demand for study guides with respect to changes in the price
of horror movie passes?
2.12 Suppose that the demand curve for corn is downward- sloping but that the supply curve is perfectly price inelastic at a
quantity of Q* once the corn is harvested. Furthermore, assume that the equilibrium price is $5 per bushel.
a. If the U.S. government decides to enter the market for corn and purchase enough so that the price doubles to $10 per
bushel (assume that the corn is given away to Russia), indi-
cate on a supply–demand diagram the amount spent on corn
by private American consumers and the amount spent on
corn by the U.S. government. If the demand for corn by
private American consumers is price inelastic (suppose that
demand elasticity is equal to 0.5), which amount is larger—
that spent by the government or that spent by private con-
sumers? Explain your answer.
b. An alternative method for helping corn farmers is to have the government pay them a subsidy of $5 per bushel of
corn. Show graphically the amount spent on corn by private
consumers under this proposal as well as the amount spent
by the U.S. government. Again assuming that the demand
elasticity for corn is 0.5, is the cost to the government of the
subsidy program greater or less than the cost of the program
described in part a? Explain your answer.
2.13 Heidi spends all her income on going to the movies, regardless of her income level or the price of movie passes.
What is her income elasticity of demand for movie passes?
What is her price elasticity of demand for movie passes?
*2.14 Assume that the demand for crack cocaine is inelastic and that users get the funds to pay for crack cocaine by stealing.
Suppose that the government increases penalties on suppliers of
crack cocaine and thereby reduces supply. What will happen to
the amount of crime committed by crack cocaine users?
2.15 If the price of gasoline is $4.00 per gallon and the price elasticity of demand is 0.4, how much will a 10 percent reduc-
tion in the quantity placed on the market increase the price?
Will total spending on gasoline rise? If so, by what percentage?
2.16 Given two parallel, downward-sloping, linear demand curves, is the demand elasticity the same at any given price?
Given two downward-sloping, linear demand curves, with one
showing consumption to be 50 percent greater than the other
demand curve at each price, is the demand elasticity the same
at any given price?
2.17 A few years ago, the bidding for a human kidney offered on the Internet auction site eBay hit $5.7 million before the
company put a stop to it citing a federal law that makes it
illegal to sell one’s own organs. Using a supply and demand
graph, explain this bidding phenomenon given that the federal
law mandates a price of zero for the sale of body organs.
2.18 Some opponents of the death penalty are opposed to exe- cuting individuals who have been convicted of murder because
they believe that murder is an irrational act and that raising the
price of murder through capital punishment thus will not have
a deterrent effect on prospective murderers. If these opponents
are correct in their view of murder being an irrational act, depict
Review Quest ions and Problems 41
what the demand curve for murder looks like. What is the price
elasticity of demand for murder according to this view?
2.19 In the antitrust case brought by the Justice Department against Microsoft, explain why the cross-price elasticity of
demand between rival operating systems such as Linux and
Microsoft’s own MS-DOS system might have been of inter-
est in determining Microsoft’s ability to control the prevailing
price in the operating system market.
2.20 If a cable operator estimates basic tier demand elastic- ity to equal unity (Application 2.7), will profit be increased by
raising rates? Explain. How about if the basic tier demand elas-
ticity is estimated to exceed unity? Explain.
2.21 Consider the case of shopping for a Valentine’s Day pres- ent. What is likely to be your price elasticity of demand the less
time you leave to shop for the present in advance of Valen-
tine’s Day, everything else being equal? Explain.
2.22 In a recent year, close to 30 percent of all flights into and out of New York City’s John F. Kennedy International Airport
experienced delays. Explain how this phenomenon is related to
a federal decision to remove the limit, in 2007, on the num-
ber of arrivals and departures between 3 P.M. and 8 P.M. and
the unwillingness to rely on prices to ensure that the quantity
supplied of runway capacity equals quantity demanded at any
given point in time. Are passengers better off under a system in
which airlines are charged landing/takeoff fees at Kennedy that
are often below equilibrium levels?
2.23 A show by U2 in the New York City area was an instant sellout due to major purchasing of tickets by Ticketmaster. Just
as quickly, however, thousands of ticket listings started appear-
ing on TicketsNow.com, a Ticketmaster subsidiary, where fans
and brokers flip tickets, often at prices far above face value. For
example, one $253 face-value ticket for a seat near the stage
was listed for $10,000. The state of New York’s anti-scalping
laws were suspended in 2007. Should they be reinstated to pre-
vent such “consumer rip-offs,” as they have been termed by
certain policymakers who also believe that there is no benefit
to consumers from allowing tickets to most large events to be
resold at any price?
2.24 With regard to Application 2.4, what would you predict to be the effect of the ban on vacation home building, all else
equal, on hotel room rates at Swiss Alpine locations? How
about on the price of vacation homes in nearby French and
Austrian Alpine locations? Explain.
2.25 In August 2013, after contentious debate in Congress, President Obama signed into law a bill restoring lower inter-
est rates for college student loans. President Obama pledged
that the hard-fought compromise with legislative representa-
tives would be just the first step in a broader, concerted fight to
reign in the cost of a college education. Evaluate the validity of
the President’s statement and whether government-subsidized
lower interest rates for student loans reduce or increase the cost
of college education.
42
C03.INDD 10:51:15:AM 08/06/2014 PAGE 42Trim Size: 203.2 mm X 254 mm
The Theory of Consumer Choice3CHAPTER
Consumers spend over $11 trillion annually in the United States. These outlays reflect countless decisions by consumers to buy or not to buy various goods. Why do consumers
purchase some things and not others? How do incomes, prices, and tastes affect consump-
tion decisions? In this chapter, we develop the fundamentals of the theory economists use
to explain how these factors interact to determine consumption choices.
One use of consumer choice theory is to explain why demand curves slope downward. But
if the theory of consumer behavior provided nothing more than a justification for drawing
demand curves with negative slopes, it would hardly be worth discussing. The basic prin-
ciples of the theory, however, have far broader applications. For example, in business, the
theory yields information for: car companies worried about the extent to which consumers
value safety versus fuel mileage; railroad and bus firms facing rising consumer incomes;
financial managers concerned about how best to structure clients’ portfolios; and suppliers
wondering how minutes of phone service sold and profits will be affected by billing customers
a constant amount per minute versus offering a flat monthly service fee irrespective of usage.
In the public policy arena, consumer choice theory can assist in the design of programs to
promote health care and encourage recycling. Furthermore, it can shed light on the school
choice debate and whether vouchers that can be used to help underwrite the education of a
child at a school of a family’s choosing enhance household well-being relative to the histori-
cal model of public provision of a particular school in each family’s district.
Learning Objectives
Develop an approach for analyzing consumer preferences. Explain how a consumer’s income and the prices that must be paid for various goods limit consumption choices. Describe how the market basket chosen by a consumer reflects both the consumer’s prefer- ences and the budget constraints imposed on the consumer by income and the prices that must be paid for various goods. Determine how changes in income affect consumption choices. Explain how altruism can be explained by the theory of consumer choice. Relate the utility approach to the indifference curve method of analyzing consumer choice. *Explain the mathematics behind consumer choice.
Memorable Quote “Too much of a good thing is wonderful.”
—Mae West, movie actor, comedian, and writer
Consumer Preferences 43
C03.INDD 10:51:15:AM 08/06/2014 PAGE 43Trim Size: 203.2 mm X 254 mm
Economists also have extended consumer choice theory to individuals’ decisions con-
cerning labor supply, saving and investment, charitable contributions, voting, and even
marriage. Indeed, some believe it provides the basis for a general theory of all human
choices, not just consumer choices among goods in the marketplace. Several applications
will be examined later in Chapter 5, but first we develop the theory fully as it pertains to the
simple choices, among goods, made by a consumer.
The basic model focuses on two important factors influencing consumer behavior. First
is the consumers’ preferences, or tastes, over various combinations of goods. Second is the
ability of consumers to acquire goods as determined by income and the prices of the goods.
3.1 Consumer Preferences Everyday observation tells us that consumers differ widely in their preferences: some like
liver, others despise it; some smoke cigarettes, others avoid cigarette smoke like the plague;
some want a different pair of shoes for every occasion, others wear running shoes every-
where. Given such diversity in preferences about goods, how should we incorporate the
influence they have on consumer choices? To deal with this problem, economists base their
analysis on some general propositions about consumer behavior that are widely believed to
be true. These propositions do not explain why people have the exact tastes they do; they
only identify some characteristics shared by the preferences of virtually everyone.
Economists make three assumptions about the typical consumer’s preferences. First, we
assume that preferences are complete in the sense that a consumer can rank (in order of pref- erence) all market baskets. In other words, between a McDonald’s Big Mac and a Burger
King Whopper hamburger, the consumer prefers the Big Mac to the Whopper, prefers the
Whopper to the Big Mac, or is indifferent between the two. We say a consumer is indifferent between two options when both are equally satisfactory. Importantly, this preference ranking
reflects the relative desirability of the options themselves and ignores their cost. For example,
it is not inconsistent for a consumer to prefer a Mercedes to a Chevrolet automobile but to
buy the Chevrolet. A purchase decision reflects both the preference ranking and the consum-
er’s ability to acquire goods, which is determined by income and the prices of the goods; the
consumer purchases the Chevrolet because its lower purchase price makes it more attractive
when both cost and the intrinsic merits of the vehicles are considered. Preferences and bud-
gets both influence consumer choice, but for the moment we will focus only on preferences.
Second, we assume that preferences are transitive. Transitivity means that if a consumer prefers market basket A to B, and B to C, then the consumer prefers A to C. For example, if Tyra Banks likes Pepsi better than Coke and Coke better than 7-Up, then logically she likes
Pepsi better than 7-Up. In a sense this condition simply requires that people have rational or
consistent preferences.
Third, a consumer is presumed to prefer more of any good to less. For example, given a
choice between one vacation in Tahiti and two vacations in Tahiti, a consumer will prefer
the latter provided that the choices are otherwise identical. This characteristic is termed
nonsatiation and is expressed as “more is preferred to less.” Are the preceding three assumptions about preferences valid? In general, yes, although
there are exceptions. For example, the assumption of transitivity is violated in the case of
individuals with a schizophrenic disorder and has been found to be less likely to hold the
younger the consumer. Researchers attribute the latter phenomenon either to a willingness
to experiment in one’s formative years or to the fact that being able to rank order prefer-
ences in a consistent manner is an acquired skill.
The assumption that more is better is also not universally true. Hot dogs might be
appealing to most individuals, but more may not always be preferred to less if they have
to be consumed all at once. That is, two hot dogs may be preferred to one hot dog, but fifty
44 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 44Trim Size: 203.2 mm X 254 mm
hot dogs are less appealing than two hot dogs—even to the heartiest eaters—if the hot dogs
have to be eaten in one fell swoop.
Moreover, there are other goods such as pollution and liver (for some people), where
less is preferred to more over all possible ranges of consumption. We call such commodi-
ties economic “bads” to distinguish them from the more frequently encountered economic “goods.” An economic “good” is one for which more is better than less; in effect it is a desirable commodity in the consumer’s view.
Notwithstanding the exceptions, completeness, transitivity, and nonsatiation appear
to be reasonable and robust characteristics of consumer preferences. We start with these
assumptions as a basis and show that a versatile theory can be developed without having to
resort to more and stronger assumptions. Along the way, we point out how the theory can
accommodate exceptions to the basic assumptions, such as economic “bads.”
Consumer Preferences Graphed As Indifference Curves A consumer’s preferences across various market baskets or combinations of goods can be shown in a diagram with indifference curves. An indifference curve plots all the market baskets that a consumer views as being equally satisfactory. In other words, it identifies the
various combinations of goods among which the consumer is indifferent. Figure 3.1 shows
an indifference curve, U1, for a student-consumer interested in two goods: movie passes (M) and compact discs (C). The student is equally satisfied with 10M plus 4C (basket A) or 5M plus 12C (basket B)—or any other combination of the two goods along U1.
From our basic assumptions, we can deduce several characteristics that indifference
curves must have. First, an indifference curve must slope downward if the consumer views
the goods as desirable. To see this, start with point A on U1 in Figure 3.1. If we change the composition of the market basket so that it contains more compact discs but the same
amount of movie passes (so the new basket is at a point such as D), the student will be better off—more compact discs are preferred to less. Note, though, that the consumer will no
longer be on U1, the original indifference curve. If we are required to keep the consumer indifferent between alternative combinations of movie passes and compact discs, we must
find a market basket that contains more compact discs but fewer movie passes. Market bas- kets that are equally satisfactory must contain more of one good and less of the other; in
other words, the curve must have a negative slope.
A second characteristic of indifference curves is that a consumer prefers a market basket
lying above (to the northeast of ) a given indifference curve to every basket on the indiffer-
ence curve. (Similarly, the consumer regards a basket below the indifference curve as less
economic “bads” commodities of which less is preferred to more over all possible ranges of consumption
economic “goods” commodities of which more is better than less
market baskets combinations of goods
indifference curve a plot of all the market baskets the consumer views as being equally satisfactory
An Indifference Curve The indifference curve, U1, shows all the combinations of movie passes and compact discs that the consumer considers equally satisfactory. The consumer prefers any market basket lying above U1 (like point E) to all market baskets on U1, and any market basket on U1 is preferred to any market basket lying below U1.
Movie passes (M)
0 4 12
10
5
Compact discs (C)
A
E
D (12, 10)
B U1
Figure 3.1
Consumer Preferences 45
C03.INDD 10:51:15:AM 08/06/2014 PAGE 45Trim Size: 203.2 mm X 254 mm
desirable than any on the indifference curve.) In Figure 3.1, pick any point above U1—for instance, E. There must be a point on U1 that has less of both goods than E—point A, for example. Basket E will clearly be preferred to A because it contains more of both goods, and more is preferred to less. Because A is equally preferred to all points on U1, point E must also be preferred to all points on U1, from the transitivity assumption. Similar reasoning implies that every basket on U1 is preferred to any basket lying below the curve.
So far we have examined only one indifference curve. To show a consumer’s entire prefer-
ence ranking, we need a set of indifference curves, or an indifference map. Figure 3.2 shows three of the consumer’s indifference curves. Because more is preferred to less, the consumer
prefers higher indifference curves. Every market basket on U3, for example, is preferred to every basket on U2. Likewise, every basket on U2 is preferred to every basket on U1.
A set of indifference curves represents an ordinal ranking. An ordinal ranking arrays market baskets in a certain order, such as most preferred, second-most preferred, and third-most
preferred. It shows order of preference but does not indicate by how much one basket is pre-
ferred to another. There is simply no way to measure how much better off the consumer is
on U2 compared with U1. Fortunately, we do not need this information to explain consumer choices when using indifference curves: knowing how consumers rank market baskets is suf-
ficient. The numbers used to label the indifference curves measure nothing; they are simply a
means of distinguishing more-preferred from less-preferred market baskets.
Now that we have described how preferences can be represented by a set of indifference
curves, a third characteristic of these curves can be stated: two indifference curves cannot inter-
sect. We can see this by incorrectly assuming that two curves intersect and then noting that this
proposition violates our basic assumptions. In Figure 3.3, two indifference curves have been
drawn to intersect. Consider three points: the intersection point E and two other points such that one (B) has more of both goods than the other (A). Now, because B and E lie on U2, they are equally preferred. Also, because E and A lie on U1, they are equally preferred. Thus, B is equal to E, and E is equal to A, so by transitivity B should equal A. However, because B has more of both goods than A, B must be preferred to A (more is preferred to less). We arrive at a contradic- tion: B cannot be equal to A and preferred to A simultaneously. Intersecting indifference curves violate our transitivity and nonsatiation assumptions. In short, they don’t make sense.
Curvature of Indifference Curves We have discussed three features of indifference curves: they slope downward, higher
curves are preferred to lower ones, and they cannot intersect. These features are implied by
the assumptions about consumer preferences made earlier. Convexity is a fourth feature of
indifference map a set of indifference curves that shows the consumer’s entire preference ranking
An Indifference Map A set of indifference curves, or an indifference map, indicates how a consumer ranks all possible market baskets. Market baskets lying on indifference curves farther from the origin are preferred to those on curves closer to the origin.
U1
U2
U3
Movie passes
0 Compact discs
Figure 3.2
46 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 46Trim Size: 203.2 mm X 254 mm
indifference curves, but because we cannot logically deduce it from the basic assumptions
about preferences, further explanation is required.
So far we have seen indifference curves that are convex to the origin; that is, they bow
inward toward the origin, so the slope of the curve becomes flatter as you move down the
curve. To explain why indifference curves have this shape we introduce the concept of
marginal rate of substitution, or MRS. A student’s marginal rate of substitution between, for example, compact discs and movie passes (MRSCM) is the maximum number of movie passes the consumer is willing to give up to obtain an additional compact disc. Because it is a measure of the willingness to trade one good for another, the MRS depends on the ini- tial quantities held: holding the number of movie passes constant, the student’s willingness
to exchange movie passes for compact discs will likely differ if the student has 10 compact
discs rather than five. Thus, a consumer’s MRS is not a fixed number but will vary with how many of each good the consumer has.
The marginal rate of substitution is related to the slope of the consumer’s indifference
curves. In fact, the slope of the indifference curve (multiplied by −1) is equal to the MRS. For example, say a market basket contains 15 movie passes and five compact discs. Let’s
assume that a student is willing to trade a maximum of four movie passes for one more
compact disc. In other words, the MRS at this point is 4M per 1C. What happens to the stu- dent’s well-being if 4M are lost and 1C is gained, so that the market basket contains 11M and 6C? The student will be no better off and no worse off than before. This is so because we have taken away the maximum number of movie passes (4) that the student was willing
to give up for another compact disc. In other words, if the student’s MRS is four movie passes per compact disc, and we take away four movie passes and add one compact disc,
the new market basket will be preferred equally to the original one. Both market baskets lie
on the same indifference curve.
This relationship is illustrated in Figure 3.4. Market baskets A (15M and 5C) and B (11M and 6C) are both on indifference curve U1. Note that the curve’s slope between points A and B is −4M/1C. The slope—or, more precisely, −1 times this slope—measures the consumer’s MRS. Purely for ease of communication, we define the MRS as a positive number so that the slope of the indifference curve, which is negative, must be multiplied by −1. Don’t let this definitional complication confuse you: the MRS and the indifference curve slope are identical concepts, both measuring the willingness of a consumer to substitute one good for
another. The indifference curve slope shows how many movie passes can be exchanged for
a compact disc without changing the consumer’s well-being—which is precisely what the
marginal rate of substitution (MRS) a measure of a consumer’s willingness to trade one good for another
Why Intersecting Indifference Curves Are Inconsistent Intersecting indifference curves are inconsistent with rational choice; they violate the assumptions of nonsatiation and transitive preferences.
Movie passes
0 Compact discs
E (3, 13)
B (6, 11)
A (5, 10)
U1
U2
Figure 3.3
Consumer Preferences 47
C03.INDD 10:51:15:AM 08/06/2014 PAGE 47Trim Size: 203.2 mm X 254 mm
MRS measures.1 Because an indifference curve’s slope and the MRS measure the same thing, drawing an indifference curve as convex (that is, with a flatter slope as we move down the curve)
means that the MRS declines as we move down the curve. In Figure 3.4, the MRS declines from 4M per 1C at basket A to 3M per 1C at B, and so on. To justify drawing indifference curves as convex, we need to explain why the MRS can be expected to decline as we move down the curve.
A diminishing MRS means that as more and more of one good is consumed along an indifference curve, the consumer is willing to give up less and less of some other good to
obtain still more of the first good. Look at point F in Figure 3.4. Here the student has a large number of compact discs and very few movie passes; in comparison with points such
as A farther up the curve, compact discs are relatively plentiful and movie passes are rela- tively scarce. Under these circumstances the student will probably be unwilling to exchange
many movie passes (already scarce) for more compact discs (already plentiful). Thus, it
seems reasonable to suppose that the MRS is lower at F than at A. At A, movie passes are more plentiful and compact discs more scarce, so we might anticipate that the student will
place a higher value on compact discs—that is, be willing to sacrifice a larger amount of
movie passes to obtain an additional compact disc. In other words, the assumption of a declining MRS embodies the belief that the relative amounts of goods are related systemati- cally to the consumer’s views about their relative importance. In particular, the more scarce one good is relative to another, the greater its relative value in terms of the other good.
This discussion is merely an appeal to the intuitive plausibility of convex indifference
curves; it is not proof. Thinking along these lines, however, has convinced many people
that indifference curves generally reflect a declining MRS. We therefore assume that indif- ference curves are convex to the origin, which implies that the consumer’s MRS declines as we move down any one of these curves.
diminishing MRS a consumer’s willingness to give up less and less of some other good to obtain still more of the first good
1The slope of the indifference curve at point A is not exactly –4M/1C. A curved line has a different slope at each point on it; its slope is measured by the slope of a straight line drawn tangent to the curve. As shown in the figure here, the slope at point A is ΔM/ΔC. Identifying the slope as we do in Figure 3.4 is an approximation to the correct measure, but for small movements along the curve the two measures are approximately equal.
A ÊM
ÊC
Diminishing MRS along an Indifference Curve Indifference curves are convex toward the origin, implying that the slope of each curve becomes flatter as we move down the curve. The (absolute value of the) slope of U1 at A is 4M/1C; at E it is 1M/1C. The indifference curve ’s slope measures the consumer’s marginal rate of substitution between goods. The convexity of indifference curves thus embodies the assumption of diminishing MRS along an indifference curve.
U1
U2
Movie passes
0 Compact discs
–4M
1C
–3M 1C
–2M 1C
–1M 1C
A (5, 15)
B (6, 11)
D (7, 8)
E (8, 6)
F (9, 5)
H (8, 15)
Figure 3.4
48 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 48Trim Size: 203.2 mm X 254 mm
Two final points related to the convexity of indifference curves should also be men-
tioned. First, we have assumed implicitly that both goods in the student’s market basket
are economic “goods” (that is, more is preferred to less), which is the general case. In other
cases, as with economic “bads,” indifference curves need not be convex. Second, a declin-
ing MRS pertains only to a movement along a given indifference curve, not to a movement from one curve to another. For example, we might be tempted to argue that if the student
has more compact discs and the same number of movie passes, the MRSCM will be lower. For Figure 3.4, this argument implies that curve U2 is flatter at point H than U1 is at point A. This is not what we are assuming; we are only assuming that the slope of each curve becomes flatter as we move down that curve.
APPLICATION 3.1
Traditionally, one of the common channels to sell cop- ies of the daily edition of a newspaper has been a kiosk. To obtain a copy, consumers provide a certain specified amount of coinage/payment after which the kiosk opens and the consumer can take a copy of the newspaper out of the stack available in the kiosk. Why has such an approach to retailing worked given that upon opening the kiosk con- sumers can take more than one copy of the newspaper? For one, the costs associated with kiosks having the means to dispense only one copy of each newspaper at a time
Diminishing MRS and Newspaper Retailing
would be significantly higher. In addition, newspaper pub- lishers rely on diminishing MRS and the fact that consum- ers’ willingness to pay for additional copies of the same edition of a newspaper that they have already bought diminishes and rapidly approaches zero with their purchase of the initial copy. The possibility that consumers, there- fore, would take more than one copy when they have paid for the initial purchase limits the probability that the kiosk will be run down of its available inventory of papers when a purchase is made.
Individuals Have Different Preferences People have different preferences, and those differences are indicated by the shapes of their
indifference curves. Consider the preferences of two consumers, Oprah Winfrey and George
W. Bush, for tacos and broiled fish. Figure 3.5a shows Winfrey’s preferences, with several of
her indifference curves relating broiled fish and taco consumption. Figure 3.5b shows Bush’s
indifference curves. Bush’s curves being steeper than Winfrey’s indicates that Bush has a
APPLICATION 3.2
In contrast to its iconic counterpart sold in the U.S. market, the Oreo in China is long, thin, four-layered, less sweet, and coated in chocolate.2 Both Oreo varieties, however, share a common trait: they are the top-selling cookies in their respective markets.
Whereas Oreos were first launched in the U.S. market in 1912, they weren’t introduced to Chinese consumers until 1996. At first, Kraft, the world’s second-largest food
Oreos in the Orient
company by revenue, sold the same Oreos in China that had done so well historically in the United States. However, in 2005, after 5 years of flat sales in China, the makeover began. As Shawn Warren, the vice president of marketing for Kraft Foods International, characterized the pre-makeover approach, the company was “practicing Albert Einstein’s definition of insanity—doing the same thing repeatedly and expecting different results.”
What Kraft researchers discovered is that Chinese con- sumers have different preferences (that is, different indif- ference maps) when it comes to cookies than do American consumers. More specifically, the traditional biscuit-like
2This application is based on “Kraft Reformulates Oreo, Scores in China,” Wall Street Journal, May 1, 2008, p. B1.
Consumer Preferences 49
C03.INDD 10:51:15:AM 08/06/2014 PAGE 49Trim Size: 203.2 mm X 254 mm
Oreo cookie was too sweet for Chinese tastes. Instead, Chinese consumers favored wafer-style cookies with less sugar content.
In 2006, Kraft introduced an Oreo in China that looked almost nothing like the original. Lower in sugar content, the new Chinese Oreo consisted of four layers of crispy wafer filled with vanilla and chocolate cream, coated in chocolate.
The reengineering efforts paid off handsomely. By the end of the year, the Oreo wafer sticks became the best- selling biscuits in China, and over the next two years Kraft doubled its Oreo revenues in China. The additional sales in the Orient propelled worldwide sales of Oreos beyond the $1 billion level for the first time by the end of 2008. Clearly, paying attention to differences in consumer preferences can pay significant dividends for companies such as Kraft.
stronger preference for tacos than Winfrey does. To understand this idea, suppose they were
both consuming the same market basket shown by point A in each graph. Because Bush’s indifference curve through this point is steeper than Winfrey’s, Bush’s MRS of fish for tacos is greater than Winfrey’s. Bush (who said on Oprah that tacos are his favorite fast food)
would be willing to trade 4 pounds of broiled fish for one more taco, but Winfrey (with her
preference for fish over tacos) would give up only 0.5 pounds of broiled fish for another taco.
Indifference curves indicate the relative desirability of different combinations of goods, so to say that Bush values tacos in terms of broiled fish more than Winfrey does is the same
as saying that Winfrey values broiled fish in terms of tacos more than Bush does. They say
nothing about how much either of them values roast turkey, for example.
Graphing Economic Bads and Economic Neuters Although our discussion of indifference curves has been restricted to the most generally
encountered case of choices among desirable goods, where more is preferred to less, we
may depict any type of preferences with a set of indifference curves. Indeed, a good test of your understanding of indifference curves is to analyze some other situations.
10 9.5
0 5 6 Tacos
A
(a) Oprah Winfrey’s indifference curves (b) George W. Bush's indifference curves
U3
U2
U1
Broiled fish (pounds)
10
6
0 5 6 Tacos
A
U3
U2
U1
Broiled fish (pounds)
Indifference Map of Two Consumers People have different preferences, and these differences show up in the shapes of their indifference maps. Oprah Winfrey’s preferences are graphed in part (a) and George W. Bush’s in part (b). The indifference maps show that Bush has a stronger preference for tacos than Winfrey does.
Figure 3.5
50 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 50Trim Size: 203.2 mm X 254 mm
For example, how would you show a person’s preferences relating weekly income and
smog with indifference curves? For a typical person, income is a desirable good, but smog
is an economic “bad.” Figure 3.6a shows income and smog on the axes. To determine the
shapes of the indifference curves, we start by picking an arbitrary market basket, point A, for example, composed of 10 units of smog and $50 in income. If we hold income constant
at $50 but increase units of smog—a move from A to B—the person will be worse off (that is, on a lower indifference curve) because smog is a “bad.” If a person inhales more smog
and is to remain on the same indifference curve, more of the “good,” income, is necessary
to compensate for the additional smog, as at point C. Thus, the indifference curve must slope upward. In addition, greater levels of well-being are shown by indifference curves
above and to the northwest: U2 is preferred to U1 if the “good” is on the vertical axis. This result can be seen by focusing on horizontal movements (more smog with the same income
makes the consumer worse off) and vertical movements (more income with the same smog
level makes the consumer better off).
Most things are either “goods” or “bads,” but an intermediate case is possible, where the
consumer doesn’t care one way or another about something. For example, we suspect most
people don’t care how many days a week the sun shines in Mongolia (unless they live in
Mongolia). Yet we can still draw indifference curves for Mongolian days of sunshine (an
economic “neuter”) and a second good such as income. Figure 3.6b shows these indiffer- ence curves as horizontal straight lines. Starting at A, we see that a horizontal move to B— more sunshine but the same income—leaves the consumer on the same indifference curve.
Thus, the indifference curves are horizontal, implying that the MRS is zero: the consumer is unwilling to give up any income for more days of Mongolian sunshine. Any vertical move-
ment—more income and the same amount of sunshine—will put the consumer on a higher
indifference curve.
By thinking of horizontal and vertical movements, you may deduce the shape of indif-
ference curves that relate various goods under different conditions. The most important
thing, however, is to understand the general case where desirable goods are on both axes,
since this case is the one most frequently encountered.
economic “neuter” a case in which the consumer doesn’t care one way or another about a particular good
$80
$50
0 10 15 Smog units inhaled (a “bad”)
A B
C
(a)
U1
U2 U3Income
(a “good”)
$70 $60 $50
0 2 3 Days of sunshine in Mongolia (a “neuter”)
A B C
(b)
U3 U2 U1
Income (a “good”)
Indifference Maps for a “Bad” and a “Neuter” (a) Indifference curves have the shapes shown here when a “good” is on the vertical axis and a “bad” on the horizontal axis. (b) They have the shapes shown here with a “good” on the vertical axis and a “neuter” on the horizontal axis.
Figure 3.6
Consumer Preferences 51
C03.INDD 10:51:15:AM 08/06/2014 PAGE 51Trim Size: 203.2 mm X 254 mm
Perfect Substitutes and Complements The shapes of indifference curves in general indicate the willingness of consumers to sub-
stitute one good for another and remain equally well off. At one extreme, certain goods are
perfect substitutes in consumption. For example, for most consumers, any dime offers the same satisfaction as any two nickels. As shown in Figure 3.7a, the typical consumer is will-
ing to trade nickels for dimes at a constant two-for-one rate while remaining equally well
off. The consumer’s indifference map thus consists of indifference curves all having a con-
stant slope of −2 nickels per dime. At the other extreme are goods that are perfect complements in consumption. To con-
sume a molecule of water, for example, we need an exact match of two atoms of hydrogen for
every atom of oxygen. Once we have one atom of oxygen and two atoms of hydrogen, as at
point A in Figure 3.7b, additional atoms of either oxygen (as at point B) or hydrogen (as at point C) keep us on the same indifference curve, U1. Only additional atoms of both hydrogen and oxygen, at a two-to-one rate, are sufficient to move us to points such as D and higher indifference curves such as U2.
Perfect complements are associated with sharply kinked, L-shaped indifference curves.
The indifference curves undergo abrupt changes in slope (from infinity to zero) at their
kinks. In our water molecule example, the kink indicates that consumers would be infinitely
willing to substitute surplus oxygen for additional hydrogen atoms when they are above
point A on indifference curve U1. Conversely, consumers would be entirely unwilling to substitute oxygen for hydrogen atoms if they are located to the right of A on indifference curve U1 and have surplus hydrogen atoms. The reason for the sharp slope change at the kink reflects the desire to consume a precise combination of the goods in question.
The typical goods on which we focus in this book fall somewhere between being perfect
substitutes and perfect complements. Namely, the typical indifference curve has a slope
that becomes gently flatter as one moves down the curve. The curve is not characterized by
a constant slope (as with perfect substitutes). Neither does the typical indifference curve’s
slope change in an abrupt manner (as with perfect complements).
perfect substitutes the case where a consumer is willing to substitute one good for another at some constant rate and remain equally well off
perfect complements goods that must be consumed in a precise combination in order to provide a consumer a given level of satisfaction
Perfect Substitutes and Perfect Complements (a) Indifference curves are straight lines when goods are perfect substitutes in consumption. (b) With perfect complements, indifference curves are L-shaped.
6
4
2
0 1 2 3 Dimes
(a) Perfect substitutes
U3
U2
U1
U3
U2
U1 A C
B D
Nickels
(b) Perfect complements
3
2
1
0 2 3 64 Hydrogen atoms
Oxygen atoms
Figure 3.7
52 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 52Trim Size: 203.2 mm X 254 mm
3.2 The Budget Constraint The preceding section examined consumer’s preferences, or tastes, for various goods. Now, we
turn to understanding budget constraints or how a consumer’s income and the prices that must be paid for various goods limit choices. Let’s begin with a simple example. Consider a student
who has a weekly discretionary income of $90 that she uses to purchase only two goods, com-
pact discs and movie passes.3 The price of each compact disc (PC) is $18, and the price of each movie pass (PM) is $9. What combinations of compact discs and movie passes may be purchased given the student’s income and the prices of the two goods? Table 3.1 provides one way of iden-
tifying the various market baskets that the consumer can purchase under these conditions.
Market basket A in Table 3.1 shows what can be purchased if all the student’s income goes to purchase movie passes. A weekly income of $90 permits the student to buy 10
movie passes at a price of $9 each, with nothing left over for compact discs. Basket Z shows the other extreme, when the student spends the entire $90 on compact discs. Because
compact discs cost $18 each, a maximum of 5 can be purchased per week, with nothing left
over for movie passes. All the intermediate baskets, B through Y, indicate the other mixes of compact discs and movie passes that cost a total of $90. In short, Table 3.1 lists all the
alternative combinations of the two goods that the student can purchase with $90.
Figure 3.8 shows a more convenient way of presenting the same information. In this dia-
gram the amount of movie passes consumed per week is measured on the vertical axis, and the
number of compact discs is measured on the horizontal axis. Both axes, therefore, measure
quantities (in contrast to a supply–demand diagram that has price on one axis). The line AZ plots the various market baskets the student may purchase from the data in Table 3.1. This line
is called the budget line, and it shows all the combinations of movie passes and compact discs the student can buy at the specified prices assuming that the student spends all of her income.
The budget line is drawn as a continuous line, not a collection of discrete points such as A and B, reflecting an assumption of continuous divisibility; that is, fractional units may be pur- chased. Although the assumption of continuous divisibility may be questioned—we can buy
zero or one haircut in a week, for example, but can we buy half a haircut?—a little reflection
shows that we can buy half a haircut per week by buying one every other week. Viewing con-
sumption as the average consumption per week, rather than the precise level of consumption in
any specific week, makes the assumption of a continuous budget line acceptable in most cases.
The consumer’s budget line identifies the options from which the consumer can choose. In our example, the student can purchase any basket on or inside line AZ. Any basket inside
budget constraint the way in which a consumer’s income and the prices that must be paid for various goods limit choices
budget line a line that shows the combinations of goods that can be purchased at the specified prices and assuming that all of the consumer’s income is expended
Table 3.1 Alternative Market Baskets the Consumer May Purchase Composition of Market Baskets
Market Basket
Movie Passes per Week
Compact Discs per Week
A 10 0
B 8 1
W 6 2
X 4 3
Y 2 4
Z 0 5
Note: Income = $90; PM = $9/movie pass; PC = $18/compact disc.
3Note that we are defining income (and, later, consumption) as a “flow” variable—the amount of income the student receives per week—as opposed to a “stock” variable—the wealth at the disposal of the student.
The Budget Constra int 53
C03.INDD 10:51:15:AM 08/06/2014 PAGE 53Trim Size: 203.2 mm X 254 mm
the line, such as G, involves a total outlay that is smaller than the student’s weekly income. Any point outside the line, such as H (7 compact discs and 6 movie passes), requires an out- lay larger than the student’s weekly income and is therefore beyond reach. Consequently,
the budget line reinforces the concept of scarcity developed in Chapter 1: the student can-
not have unlimited amounts of everything, so choices among possible options, which are
shown by the budget line, must be made.
Geometry of the Budget Line A thorough understanding of the geometry of a budget line will prove helpful later on.
Note that the intercepts with the axes show the maximum amount of one good that can
be purchased if none of the other is bought. Point A indicates that 10 movie passes can be bought if income is devoted to movie passes alone. The vertical intercept equals the stu-
dent’s weekly income (I) divided by the price of movie passes (I/PM, or [$90/($9/movie pass)] = 10 movie passes), since $90 permits the purchase of 10 movie passes costing $9 each. Similarly, point Z equals weekly income divided by the price of compact discs (I/PC, or [$90/($18/compact disc)] = 5 compact discs).
The budget line’s slope indicates how many movie passes the student must give up to
buy one more compact disc. For example, the slope at point B in Figure 3.8 is ΔM/ΔC, or −2 movie passes per 1 compact disc, indicating that a move from B to W involves sacrific- ing two movie passes to gain an additional compact disc. (Because AZ is a straight line, the slope is constant at −2 movie passes per 1 compact disc at all points along the line.) Note that the slope indicates the relative cost of each good. To get one more compact disc, the
student must give up two movie passes. A budget line’s slope is determined by the prices of
the two goods. In fact, the slope is equal to (minus) the ratio of prices:
Δ ΔM C P P/ /= − C M.
In this example we have:
Δ ΔM C/ ( )/( )= −2 movie passes 1 compact disc
because:
− = − = −
P PC compact disc movie pass movie pass
/ [($ / )/($ / )]
(
Μ 18 9
2 ees compact disc)/( ).1
The Budget Line A consumer’s budget line shows the combinations of goods that can be purchased with a given income, I, and prices of goods held constant. Line AZ shows the budget line when income is $90, PM = $9/movie pass, and PC = $18/compact disc. The line’s slope is −PC/PM.
Consumption of movie passes per week
Consumption of compact discs per week
10 2 3 4 5= I/PC
I/PM = 10
8
6
4
2
ÊM
ÊC
G
A
B
W
X
Y
Z
H (7, 6)
Figure 3.8
54 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 54Trim Size: 203.2 mm X 254 mm
To understand this important relationship, note that because movie passes cost $9 each,
the student has to purchase two fewer movie passes (∆M) to have the $18 required to buy one more compact disc (∆C). Thus, the slope of the budget line, which equals two movie passes
per compact disc, reflects the fact that compact discs are twice as expensive as movie passes.
The slope equals the price ratio: $18 per compact disc divided by $9 per movie pass equals
two movie passes per compact disc. Put somewhat differently, the slope of the budget line is a
measure of relative price—the price of one unit of a good in terms of units of the other.4
Shifts in Budget Lines We have seen how income together with the prices of goods determine a consumer’s bud-
get line. Any change in income or prices will cause a shift in the budget line. Let us explore
how the budget line shifts in response to a change in these two underlying factors.
Income Changes We’ll begin with the budget line described before, where the student’s weekly income is
$90, the price of each compact disc is $18, and the price of each movie pass is $9. We
again draw the budget line as AZ in Figure 3.9. (Note that from now on, we will use the shorthand term compact discs instead of weekly consumption of compact discs. We also will use income instead of weekly income and price instead of per-unit price.) Now suppose the student’s income rises to $180, but the prices of compact discs and movie passes do not
change. The new budget line is A′Z′. Point A′ shows the new maximum number of movie passes that can be bought if all income is allocated to movie passes. Because income is now
$180 (2I) and PM is still $9, the student can buy 20 movie passes. (Recall that the vertical intercept, A′, equals income, now 2I = $180, divided by PM.) Similarly, the student can buy a maximum of 10 compact discs if all income is spent on compact discs.
4We summarize this idea with a bit of algebra. The budget line shows market baskets where the sum of expenditures on compact discs and movie passes equals income. Thus, I = PCC + PMM, where PCC is the per-unit price times the quantity of compact discs consumed. That is, PCC is the expenditure on compact discs. Similarly, PMM is the expenditure on movie passes. Because I, PC, and PM are constants, this equa- tion defines a straight line. If we solve for M, we have M = I/PM − (PC/PM)C. The slope of this line is the coefficient of C, or −PC/PM.
Effect of an Income Change on the Budget Line A change in income when product prices remain unchanged results in a parallel shift in the budget line. When income increases from $90 to $180, the budget line shifts outward from AZ to A′Z′, but its slope does not change.
–2
+1
2I/PM = A ′
Movie passes
Z = I/PC Z ′ = 2I/PC0 Compact discs
J (2, 16)
(0, 20)
I/PM = A (0, 10)
(10, 0)(5, 0)
K (3, 14)–2 +1
Figure 3.9
The Budget Constra int 55
C03.INDD 10:51:15:AM 08/06/2014 PAGE 55Trim Size: 203.2 mm X 254 mm
Note that a change in income with constant prices produces a parallel shift in the budget line.
The slope of the budget line has not changed, because prices have remained fixed. Even with a
higher income, the student must still give up two movie passes, at $9 per pass, to consume one
more compact disc at $18 each (as shown by the move from J to K on the new budget line). The slope of any budget line—regardless of the income level—equals the price ratio; because prices
are unchanged, so is the slope. With a higher income the student can purchase more of both
goods than before, but the cost of one good in terms of the other remains the same.
Price Changes Now consider a change in the price of one good, with income and the price of the other
good held constant. Starting again with the same initial budget line AZ, reflecting a budget of $90, Figure 3.10 shows the effect of a reduction in the price of compact discs from $18 to
$9 (from PC to ′PC). The price reduction causes the budget line to rotate about point A and become flatter, producing the new budget line AZ′. The maximum number of movie passes that can be bought is unaffected, because income is still $90, and PM is still $9. However, the maximum number of compact discs that can be bought increases when the price of com-
pact discs falls. At the new price of $9 ( ′PC= 0.5 CP ), the student can buy a maximum of 10 compact discs (point Z′) if the entire $90 is spent on compact discs. A price change causes the budget line to rotate, so the slope of the line changes. When the price of compact
discs falls, the new budget line becomes flatter because its slope, − ′P PC ⁄ M is now equal to −1 movie pass per 1 compact disc. With the price of both movie passes and compact discs at $9,
the purchase of an additional compact disc now involves a sacrifice of only one movie pass:
this trade-off is illustrated by the move from S to T on the new budget line. A flatter budget line means that the real or relative price of the good on the horizontal axis is lower.
Now is an appropriate time to emphasize why the slope measures the real price. A slope of −2/1, like the slope for line AZ, means the compact disc price is double the movie pass price. Note, however, that the slope does not tell us what the nominal (absolute) prices are. If both
prices change by the same proportion—both double or are cut in half, for instance—the price
ratio does not change, so the cost of one good in terms of another is unaffected. For example,
see how pure inflation (in which all prices, including wage rates, rise proportionately) affects
the budget line. Suppose all prices and income double: the student’s income rises to $180 (or
2I), and prices increase so that ′PM = $18 (or 2PM) and ′PC = $36 (or 2PC). Despite the nominal increases in income and prices, the budget line does not change. The budget line intercepts
are now 2I/2PM and 2I/2PC, and the slope is −2PC/2PM, which all reduce to the original val- ues. The position and the slope of budget lines always reflect real, not nominal, prices.
Movie passes
Z = I/PC Z ′ = I/.5PC = I/PC ′0 Compact discs
+1
–1
+1
S (4, 6)
(0, 10)
(10, 0)(5, 0)
T (5, 5)
I/PM = A
–2
Effect of a Price Change on the Budget Line A change in the price of one good, with income and the other good’s price remaining unchanged, causes the budget line to rotate about one of the intercepts. When the price of compact discs declines from $18 to $9, the budget line rotates from AZ to AZ′.
Figure 3.10
56 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 56Trim Size: 203.2 mm X 254 mm
We have confined our attention to simple budget lines in which the consumer can buy
as much of each good as desired without affecting its price. Since individual consumers
usually buy only a tiny portion of the total quantity of any good, their purchases generally
have no perceptible effect on price. Thus, the consumer can treat price as constant, and with
prices constant the budget line is a simple straight line. As we will see later, there are cases
where budget lines are not straight lines, but they are easily dealt with after the standard
case has been thoroughly examined.
3.3 The Consumer’s Choice Indifference curves represent the consumer’s preferences toward various market baskets;
the budget line shows what market baskets the consumer can afford. Putting these two
pieces together, we can determine what market basket the consumer will actually choose.
Figure 3.11 shows the student-consumer’s budget line AZ along with several indiffer- ence curves. Remember from the foregoing section that the budget line AZ is based on income (I) of $90 and prices (PC and PM, respectively) of $18 per compact disc and $9 per movie. We assume that the student will purchase the market basket from among those that
can be afforded lying on the highest possible (most preferred) indifference curve. In other
words, the consumer will select the market basket that best satisfies preferences, given a
limited income and prevailing prices. Visual inspection of this diagram shows that the stu-
dent will choose market basket W (2 compact discs and 6 movie passes) on indifference curve U2. Indifference curve U2 is the highest level of satisfaction the student can attain, given the limitations implied by the budget line. Although the student would be better off
with any market basket on U3, none of those baskets is affordable because U3 lies entirely above the budget line. Any basket other than W on the budget line is affordable but yields less satisfaction because it lies on an indifference curve below U2. For example, basket R can be purchased, but then the student is on U1, so basket R is clearly inferior to basket W.
Note that the highest indifference curve attainable is the one that just touches, or is tan-
gent to, the budget line: U2 is tangent to AZ at W. The consumer’s optimal point is W. Since U2 and AZ are tangent at this point, the slopes of the curves are equal. Because the slopes equal the consumer’s MRS and price ratio, respectively, the consumer’s optimal choice is characterized by the following equality:
MRSCM = PC /PM.
This equality indicates that the rate at which the student is willing to substitute compact
discs for movie passes (MRSCM) is equal to the rate at which the market allows the student to make the substitution (PC/PM). To see why this equality characterizes the student’s opti- mal choice, suppose that the student buys some market basket other than point W on the budget line—for example, the basket at point R. The indifference curve through any point above W on the budget line will intersect the budget line from above, as U1 does at point R.5 Thus, at R the student’s MRSCM (three movie passes per compact disc) is greater than the price ratio (two movie passes per compact disc). At point R the student’s preferences indi- cate a willingness to exchange as many as three movie passes for another compact disc, but
at the given market prices the student needs to give up only two movie passes per compact
disc—a bargain.
In effect, the MRS measures the marginal benefit or value the student derives from con- suming one more unit of a good. At point R, for example, the marginal benefit of another compact disc is three movie passes, or the number of movie passes the student would give
marginal benefit the value the consumer derives from consuming one more unit of a good
5If we tried to draw an indifference curve through R intersecting from below, we would find that it would intersect U2, and intersecting indifference curves are impossible.
The Consumer ’s Choice 57
C03.INDD 10:51:15:AM 08/06/2014 PAGE 57Trim Size: 203.2 mm X 254 mm
up for another compact disc. On the other hand, the price ratio measures marginal cost: the
marginal cost of another compact disc is two movie passes. At R, therefore, the marginal benefit of another compact disc in terms of movie passes is greater than the marginal cost,
and the student will be better off consuming more compact discs (and fewer movie passes).
Thus, by moving along the budget line in the direction of point W, the student will reach a higher indifference curve. In these terms, the optimal point indicates that the student has
chosen a market basket so that the marginal benefit of compact discs in terms of movie
passes (MRSCM) equals the marginal cost of compact discs in terms of movie passes (PC/PM). At points below W on the budget line, similar reasoning shows that the student would be
better off consuming fewer compact discs and more movie passes, so none of these points
is optimal. At point Y, the student’s MRS is less than the price ratio. The last compact disc consumed is worth one movie pass to the student, yet its marginal cost is two movie passes,
so the student is better off by moving back up toward point W.
A Corner Solution When a consumer’s preferences are such that some of both goods will be consumed, the
optimal consumption choice is characterized by an equality between the MRS and the price ratio, as described earlier. In reality, however, there are some goods individual consumers
do not consume at all. You may wish you had a Maserati, tickets to the Super Bowl, or a
posh condominium in Palm Beach, but in all likelihood your consumption of these, and
many other, goods is zero. The reason is that the first unit of consumption of these goods,
however desirable, fails to justify the cost involved.
Figure 3.12 shows a situation in which a consumer purchases only one of the two goods
available. The optimal consumption point is A, where all the consumer’s income goes to purchase clothing. Because the slope of the indifference curve at A is less than the slope of the budget line, the first unit of Dom Perignon champagne (arguably the world’s finest) is
not worth its cost to the consumer (over $100 per bottle). In this situation, known as a
corner solution, the consumer’s optimal choice is not characterized by an equality between the MRS and the price ratio because the slopes of the indifference curve and the budget line differ at point A. The equality condition holds only between pairs of goods consumed in positive amounts.
marginal cost the cost of consuming one more unit of a good
corner solution a situation in which a particular good is not consumed at all by an individual consumer because the value of the first unit of the good is less than the cost
The Consumer’s Optimal Consumption Choice The market basket the consumer will choose is shown by point W, where the budget line is tangent to (has the same slope as) indifference curve U2. Among market baskets that can be afforded—shown by budget line AZ—basket W yields the greatest satisfaction because it is on the highest attainable indifference curve. At point W, MRSCM = PC/PM.
U3
U2
U1
R
W
D Y
6
0 2 Z = I/PC
–3M
–1M
–2M
1C
1C
1C
I/PM = A
Compact discs
Movie passesFigure 3.11
58 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 58Trim Size: 203.2 mm X 254 mm
The Composite-Good Convention We developed our analysis for a two-good world, but the general principles apply to a world
of many goods. Unfortunately, many goods cannot be shown on a two-dimensional graph.
Still, it is possible to deal with a multitude of goods in two dimensions by treating a number
of goods as a group. Suppose there are many goods: compact discs, movie passes, pagers, Big
Macs, and so on. We can continue to measure compact disc consumption, or whatever good
we wish to analyze, on one axis, but then treat all other goods as if they were one good—that
is, as a composite good. Consumption of the composite good is gauged by total outlays on it—in other words, total outlays on all goods other than compact discs.
Figure 3.13 illustrates this approach. The consumer has $180 in income, and the price of com-
pact discs is $18. The budget line’s vertical intercept occurs at $180, because total outlays on
other goods will be $180 if compact disc consumption is zero. The consumer’s income equals A. As noted earlier, the budget line’s slope equals the ratio of prices, but because a $1 outlay on
other goods has a price of one, the ratio reduces to the price of compact discs (PC /1 = PC). Thus, at any point on the budget line, such as Y, consuming one more compact disc (which costs $18) means that outlays on other goods must be reduced by $18.
Convex indifference curves can also be drawn to relate outlays on other goods and com-
pact discs, because both are presumed to be desirable goods to the consumer. We must now
state an important assumption associated with this approach: we assume that the prices of all
the other goods are constant. This assumption allows us to treat them as a single good. We
want outlays on other goods to serve as an index of the quantities of other goods consumed,
and if prices can vary, it would become a rubbery index. (A larger outlay would not nec-
essarily mean more goods were consumed unless prices were fixed.) When other prices are
held constant, the consumer’s preferences can be shown as indifference curves that identify a
unique level of well-being for each combination of compact discs and outlays on other goods.
The slope of an indifference curve, the MRS, now shows how much the consumer is willing to reduce outlays on other goods to obtain one more compact disc. With market
basket B, for example, the consumer is willing to give up $22 worth of other goods in exchange for an additional disc. Note that this MRS is still a measure of the consumer’s willingness to substitute among real goods, but now dollar outlays measure the quantity of
other goods the consumer is willing to sacrifice in return for compact discs.
Figure 3.13 shows the optimal point for the consumer is W, where the budget line is tangent to an indifference curve. At W, the consumer is just willing to give up $18 worth of other goods for another compact disc, indicating an MRS of $18 per compact disc. This
composite good a number of goods treated as a group
A Corner Solution Possibly, the consumer will not buy any of one good. In this case, the optimal choice lies at one of the intercepts of the budget line, with the consumer’s entire income spent on only one good. Here the choice is point A, with only clothing purchased.
0 Z
A
Dom Perignon
Clothing
U3 U2
U1
Figure 3.12
The Consumer ’s Choice 59
C03.INDD 10:51:15:AM 08/06/2014 PAGE 59Trim Size: 203.2 mm X 254 mm
figure equals the market price that must be paid for another compact disc (that is, the slope
of the budget line, PC/1, equals $18 per compact disc). The optimal market basket W consists of five compact discs and $90 (A′) devoted to
the purchase of other goods. The total outlay on compact discs can also be shown as the
distance AA′. Because the consumer has an income of A ($180) and, after buying compact discs, has A′ ($90) left to spend on other goods, the difference (A – A′ = AA′), or $90, is the total cost of the five compact discs. The difference between the consumer’s total income
and the amount spent on everything else except compact discs reflects the amount of the
consumer’s income spent on compact discs.
Thus, we see that using the composite-good convention does not change the substance
of our analysis. The consumer’s optimal point still involves a balancing of the relative
desirability of goods with their relative costs.
APPLICATION 3.3
While most success stories during the recent economic downturn came from the technology sector and the rapidly growing app economy, one notable exception is Chipotle Mexican Grill: a firm that has been leading the growth of the premium fast food or “fast casual” segment.6 Chipo- tle, which refers to a dried jalapeno chili pepper in Mexi- can-Spanish, became a publicly traded company in 2006 and has witnessed tremendous growth in its revenues,
Premium Fast Food: Why Chipotle Is One Hot Pepper of a Stock
employee base, profits, and number of outlets. By 2013, Chipotle had over 1,400 outlets (primarily in the United States but also some in France, Canada, and the United Kingdom), nearly $3 billion in annual revenues, and oper- ating margins of more than 25 percent. Chipotle, like other firms in the fast casual segment such as burger-oriented chains In-N-Out and Five Guys and Chop’t, an assembly- line salad chain, have been discovering that a sufficient number of today’s consumers have a more steeply-sloped indifference curve relating how much they are willing to give up in terms of other goods for an additional unit of fast food (as depicted in the composite-good approach of
6This application is based on: http://en.wikipedia.org/wiki/Chipotle_ Mexican_Grill; and http://www.slate.com/articles/business/monebox/ 2012/02/chipotle_is_apple.
Outlays on other goods
$180 = A
$90 = A ′
0 5 Z Compact discs
– $22
B
1C
W
U3
U2
U1 – $18
1C
Y
The Composite-Good Convention To deal graphically with the consumption of many goods, we group together all goods but one and measure total outlays on this composite good on the vertical axis. The slope of the budget line is then the dollar price of the good on the horizontal axis. The consumer’s optimal point, W, is once more a tangency between an indifference curve and the budget line.
Figure 3.13
60 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 60Trim Size: 203.2 mm X 254 mm
3.4 Changes in Income and Consumption Choices A change in income affects consumption choices by altering the set of market baskets
a consumer can afford—that is, by shifting the budget line. To examine the impact of a
change in income, we assume that the consumer’s underlying preferences do not change
and the prices of goods remain fixed; only income varies.
In Figure 3.14a, the student’s original optimal choice is at point W, where indifference curve U1 is tangent to budget line AZ. Consumption of compact discs is C1 and consump- tion of all other goods is A1. Now suppose that income increases such that the budget line parallel shifts out from AZ to A′Z′. The budget line’s slope (the price ratio) is unchanged since only income is assumed to vary.
The outward shift in the budget line means that the consumer is able to buy market
baskets that were previously unaffordable, but which market basket will be chosen? The
answer depends on the nature of the consumer’s preferences. For the set of indifference
curves in Figure 3.14a, the most preferred market basket along the A′Z′ budget line is at point W′, where U2 is tangent to the budget line. The consumer is better off (that is, attains a higher indifference curve) and consumes C2 compact discs and A2 other goods. For the specific indifference curves shown, an increase in income with no change in prices leads to
an increase in compact disc consumption, from C1 to C2. If income increases further such that the budget line shifts to A″Z″, the new optimal con-
sumption choice becomes point W″. Proceeding in the same way, we find that each possible income level has a unique optimal market basket associated with it. Only three optimal
consumption points are shown in the diagram, but others can be derived by considering still
different levels of income for the student. The line that joins all the optimal consumption
points generated by varying income is the income-consumption curve. It passes through points W, W′, and W″ in the diagram.
Normal Goods The relationship in Figure 3.14a is fairly typical of what happens to the consumption of
a good (compact discs, in this case) when income increases. As noted in Chapter 2, when
more of a good is purchased by an individual as income rises (prices and preferences being
unchanged) and less is purchased as income falls, the good is defined to be a normal good.
Calling such a good normal reflects the judgment that most goods are like this.
income-consumption curve the line that joins all the optimal consumption points generated by varying income
Figure 3.13) if the food is better quality and prepared more conveniently.
Chipotle is known for its large burritos, assembly-line production, and reliance on natural ingredients. For exam- ple, the majority of food is prepared on location without the use of freezers, microwaves or can openers. One of the sole exceptions is Chipotle’s barbecued meat products such as carnitas and barbacoa which are prepared through a cutting- edge approach called “sous-vide” that is normally associated with haute cuisine. At a central location in Chicago, pieces of meat are placed in an airtight bag with seasonings. The bag is then cooked slowly and at a precise temperature in a water bath until it is ready to be shipped to a particular restaurant location where it can be reheated on-site as needed. While the sous-vide technique had historically been used largely
by fancy chefs, Chipotle has figured out a way through which workers with little training can prepare premium fast food ingredients to consistently high-quality specifications.
At each Chipotle location, the production inputs of rice, beans, salsas, and meats are organized backstage and then delivered as requested by the front-line workers. The front- line workers, in turn, take and fulfill orders in parts with each worker responsible for no more than a few steps of the assembly-line production process. The production formula and realizing that many consumers will pay extra for pre- mium fast food is paying off handsomely for Chipotle. It is attracting over 750,000 customers daily, has been among the top 10 fastest-growing restaurant chains in each of the past several years, and was ranked as the best Mexican fast- food chain in 2011 by Consumer Reports.
Changes in Income and Consumption Choices 61
C03.INDD 10:51:15:AM 08/06/2014 PAGE 61Trim Size: 203.2 mm X 254 mm
Figure 3.14a indicates that compact disc consumption increases with income even
though the price of compact discs (PC) is unchanged. We can illustrate the same thing by using a different graph that shows the consumer’s demand curve, d, for compact discs. (We use a lowercase d to indicate that it is the demand curve for an individual con- sumer.) We have not derived an entire demand curve yet, but we can identify one point
on the demand curve corresponding to each income level. For example, at the original
income level, when the budget constraint is AZ and the price of a compact disc is PC, the student consumes C1 compact discs. Therefore, one point on the demand curve d in Figure 3.14b can be identified: point W indicates that compact disc consumption is C1 when the compact disc price is PC. (The other points on the curve will be taken for granted at present.)
Income Changes and Optimal Consumption Choice An increase in income parallel-shifts the budget line outward and leads the consumer to select a different market basket. Connecting the optimal consumption points (W, W′, W″) associated with different incomes yields the income- consumption curve in part (a). Part (b) shows how the consumer’s demand curve shifts when income changes.
Compact discs
Income-consumption curve
1C
$18 $18
1C1C 0
W″
A″
W ′
Z ′ Z″
A ′
W
Z
A
U2 U3U1
C1 C2 C3
(a)
(b)
A 1
A 2
A3
Other goods
$18
Compact discs0
$18 = PC
d″
W″
d ′
W ′
d
W
C1 C2 C3
Price of compact discs
Figure 3.14
62 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 62Trim Size: 203.2 mm X 254 mm
When an increase in income combined with no change in prices leads to greater con-
sumption, it is represented by a shift in the demand curve. Thus, when income rises such
that the budget line shifts from AZ to A′Z′, the entire demand curve shifts to d′, which shows an increased consumption of compact discs, C2, at an unchanged compact disc price. Recall that a demand curve shows how price affects consumption when other factors are
held constant. One of the more important factors held fixed is the consumer’s income, so a
change in income can be expected to shift the entire demand curve. Put another way, d is a demand curve that holds income constant (at the level associated with budget line AZ in Figure 3.14a) at all points along it, while d′ holds the income constant at a different level (the level associated with budget line A′Z′).
While our emphasis here is on the way budget lines and indifference curves can be used to
examine consumer choices, we should not lose sight of the fact that there are alternative ways
to approach the same problem. Both parts of Figure 3.14 show the same thing but from different
perspectives. Which approach is better depends on the problem being examined.
Inferior Goods Does an increase in income always lead to increased consumption? As discussed in Chapter 2,
not necessarily. The consumption of certain goods, termed inferior goods, is inversely
related to income. For example, consumption of Chevrolet cars may fall (and consump-
tion of Mercedes cars rise) for an individual whose income increases sharply. Rail and bus
transportation have declined over the past several decades due to rising income levels and
consumers traveling by air instead.
For the student whose preferences are shown in Figure 3.15a, hamburger is an inferior
good. At the original income level, when the budget line is AZ, the student’s optimal con- sumption point is W. Hamburger consumption is H1. When income rises such that the bud- get line parallel shifts out to A′Z′, the optimal point, W′, on the new budget line, shows that hamburger consumption drops to H2. When a good is an inferior good, the income- consumption curve connecting the optimal points, W, W′, and so on, is negatively sloped, implying lower consumption at higher income levels. In Figure 3.15b, note that the demand
curve shifts inward for an inferior good when income increases: lower consumption at an
unchanged price of hamburger (PH) implies a reduction in demand. Several other subtle points concerning normal and inferior goods should also be kept in
mind. First, a good may be inferior for some people and normal for others. Your hamburger
consumption may go up if your income rises, but someone else’s may go down if his or her
income rises. Goods themselves are not intrinsically normal or inferior: the definitions refer
to individuals’ responses to income changes, and the responses depend ultimately on the
shapes of the individuals’ indifference curves.
Second, a good may be a normal good for an individual at some income levels but an
inferior good at other income levels. In Figure 3.15a, at a low level of income, the good
is normal (as shown by the positively sloped income-consumption curve when income is
low), but it becomes inferior at higher income levels. For example, you might consume
more hamburger if your weekly income increases by $10, but you might consume less if
your weekly income increases by $1,000.
Third, an inferior good should not be confused with an economic “bad.” An inferior
good is not a “bad” (where less is preferred to more), as shown by the normally shaped
indifference curves in Figure 3.15a.
Fourth, inferior goods tend to have certain common characteristics, and understanding
these characteristics is helpful in evaluating their significance. For example, most inferior
goods are narrowly defined goods in a general category that includes several other higher-
quality (and higher-priced) goods. Take hamburger: hamburger is a narrowly defined good
belonging to the general category meat. In the meat category there are other higher-quality
and higher-priced options, such as filet mignon, prime rib, and veal. Understandably,
Changes in Income and Consumption Choices 63
C03.INDD 10:51:15:AM 08/06/2014 PAGE 63Trim Size: 203.2 mm X 254 mm
some people would consume less hamburger when their incomes go up, because they
could afford the better-quality alternatives that serve the same basic purposes but satisfy
them better.
In contrast, intuition and evidence both suggest that broadly defined goods are usually
normal. Meat is more likely to be a normal good than hamburger is, and food is more likely
to be a normal good than meat. Many applications of economic theory necessarily involve
broadly defined goods, which means that the normal-good case is likely to be the most
relevant one. For example, the food stamp program subsidizes consumption of all kinds
of food, not just hamburger; and food, considered as a composite commodity composed of
many specific items, is surely a normal good.
Income Changes and Purchases of an Inferior Good (a) An inferior good is one that the consumer will purchase less of at a higher income and unchanged price, and is characterized by a negatively sloped income-consumption curve. (b) At a higher income, the demand for an inferior good shifts inward.
Hamburger
Income-consumption curve
$2 1H1H
0
W″
A″
W ′
Z′
Z″
A′
W
Z
A
U2
U3
U1 H3 H2 H1
(a)
(b)
A1
A2
A3
Other goods
$2
Hamburger0
PH
d″
W″
d ′
W ′
d
W
H3 H2 H1
Price of hamburger
Figure 3.15
64 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 64Trim Size: 203.2 mm X 254 mm
The Food Stamp Program Under the federal food stamp program, eligible low-income families receive free food
stamps, which can be used only to buy food, F. To show how consumer theory can be used to examine this program, we’ll use a specific example in which a consumer receives $50
worth of food stamps each week. We assume that the consumer has a weekly income of
$100 and that the per-unit price of food (PF) is $5. Consider Figure 3.16a. The pre-subsidy budget line is AZ, and the consumer purchases
eight food units before receiving food stamps. The food stamp subsidy shifts the budget
line to AA′Z′. Over the AA′ range the budget line is horizontal since the $50 in free food stamps permits the purchase of up to 10 food units while leaving the consumer’s entire
income of $100 to be spent on other goods. If, however, food purchases exceed 10 units,
the consumer must buy any additional units at the full market price of $5. Thus, the A′Z′ portion of the budget line has a slope of $5 per unit of food, indicating the price of food
over this range.
In this case, the budget line is not a straight line throughout but is kinked at point A′, since the subsidy terminates at the point where the $50 in food stamps is used up. Another
way of visualizing this is by contrasting it with the budget line associated with an increase
in income of $50, resulting from the government giving the consumer $50 to spend on any
good desired. With a $50 cash grant the budget line would be A″Z′. The only way the food stamp subsidy differs from an outright cash grant is that options indicated on the upper part
of the line, A″A′, are not available to the recipient, since food stamps cannot be used to pur- chase nonfood items.
Food0
$150 = A″
W ′
Z ′
A′
W
Z
Y
U2
U3
U1
F1 (8)(10)(13) (20) (30) (5)(8) (10) (20) (30)
FG F2
(a)
$60 = A1
$85 = A2
$100 = A
Other goods
1F $5
Food0
$150 = A″
W ′
Z ′
A′
W
Z
U2
U1
F1 FG F2
(b)
$75 = A1
$100 = A
Other goods
1F $5
Effects of the Food Stamp Program on Consumption If a consumer is given AA′ in food stamps, the budget line shifts to AA′Z′. The result is identical to giving the consumer cash if preferences are like those in part (a), but the consumer would be better off if given cash for preferences like those in part (b).
Figure 3.16
Changes in Income and Consumption Choices 65
C03.INDD 10:51:15:AM 08/06/2014 PAGE 65Trim Size: 203.2 mm X 254 mm
The food stamp subsidy will affect the recipient in one of two ways. Figure 3.16a shows
one possibility. In this case, the consumer chooses market basket W′ when the budget line is AA′Z′ under the food stamp program. If the consumer receives instead a cash grant of $50, the budget line is A″Z′, and the same market basket W′ would be chosen. Food consumption increases, but only because food is a normal good, and more of a normal good is consumed
at a higher income. Note, however, that food consumption rises by less than the amount
of the subsidy. For the preferences shown in Figure 3.16a, food consumption increases by
five units, or $25. The consumer uses the remainder of the subsidy to increase purchases of
other goods, from $60 to $85. (Because other goods, taken as a group, are certainly normal
goods, too, the purchase of these goods will also increase.)
Figure 3.16b, where the indifference curves differ from those in Figure 3.16a, shows
another possible outcome of the food stamp subsidy. In this case, with a direct cash grant
of $50, the consumer prefers to consume at point W′, where U3 is tangent to the budget line A″A′Z′. The food stamp subsidy prohibits such an outcome, however; the consumer must choose among the options shown by the AA′Z′ budget line. Faced with these alternatives, the best the consumer can do is choose point A′, because the highest indifference curve attainable is U2, which passes through the market basket at point A′.
When the situation shown in Figure 3.16b occurs, the subsidy increases the consumer’s
purchases of both food and nonfood items. Indeed, regardless of whether Figure 3.16a or
Figure 3.16b is the relevant case, the food stamp subsidy cannot in reality avoid being used
in part to finance increased consumption of nonfood items. This result is particularly inter-
esting because many proponents of subsidies of this sort emphasize that the subsidy should
not be used to finance consumption of “unnecessary” goods (such as vodka or junk food).
In practice, it is difficult to design a subsidy that will increase consumption of only the sub-
sidized good and not affect consumption of other goods at the same time.
Note also that the consumer in Figure 3.16b will be better off if given $50 to spend as he
or she wishes instead of $50 in food stamps. The budget line will then be A″A′Z′, and the consumer will choose the market basket at point W′, on indifference curve U3. This obser- vation illustrates the general proposition that recipients of a subsidy will be better off if the
subsidy is given as cash. The situation in Figure 3.16a illustrates why there is a qualifica-
tion to this proposition: in some cases the consumer is equally well off under either subsidy.
There is no case, however, where the consumer is better off with a subsidy to a particular
good than with an equivalent cash subsidy.
APPLICATION 3.4
At most colleges, commencement tickets are rationed to seniors at a zero cash price. For example, each senior may be given four tickets for family and friends. We can depict the effects of such an allocation scheme in Figure 3.17a. The typical senior’s budget line is ABCE: from zero to four tickets the budget line’s slope is $0 since the first four tickets are free; the budget line becomes vertical at four tickets since no more than four tickets can be obtained; and, assuming that the college is the only source of tickets (we rule out, for now, students exchanging tickets between themselves), the price of a ticket becomes infinite beyond four tickets.
Relative to the case where the college sets a positive price for tickets such that the average senior would buy exactly four tickets, the average senior is definitely better off if four tickets
The Allocation of Commencement Tickets
are given away instead. Suppose, for example, that, as shown in Figure 3.17a, a $10-per-ticket price would result in the aver- age senior choosing market basket C along the budget line AZ; the senior would purchase four tickets at $10 each and spend A′ on other goods. Under the $10-per-ticket pricing scheme, the average senior can attain only indifference curve U1. In contrast, when the college provides four free tickets the senior can attain indifference curve U2 and thus is better off.
While the average senior depicted in Figure 3.17a may be better off, is every senior better off? The answer is no, and Figure 3.17b shows why. There are some students who may have steeply sloped indifference curves at point B—four free tickets and an income level of A. Because they come from large families and/or have many friends, they may dearly
66 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 66Trim Size: 203.2 mm X 254 mm
want more than the allocated four tickets. These students will be willing to exchange a considerable amount, in terms of dollars that could be spent on other goods, for addi- tional tickets. (The MRS at point B of the student depicted in Figure 3.17b is much greater than for the student shown in Figure 3.17a.) Under a $10-per-ticket pricing scheme, they would select market basket W on indifference curve U3 and be better off than under the existing system, where they get four free tickets and can attain only indifference curve U2.
We have so far ruled out seniors exchanging tickets among themselves. At most colleges, however, just such a resale market emerges every spring. The reason is fairly easy to see. Under the allocation scheme employed by most colleges, seniors will select a point such as B on the northeast corner of their budget line. (The northeast corner is the market basket of choice so long as their indifference curves are not either perfectly horizontal or
perfectly vertical.) Having opted to be on the same north- east corner (a point such as B) of their budget line, however, does not imply that the slopes of the indifference curves of all seniors are identical at their chosen market baskets. As can be seen by comparing Figures 3.17a and 3.17b, some students have flatter indifference curves than others at their optimal consumption points. (U2 is flatter at point B in Figure 3.17a than in Figure 3.17b.) Variations in indifference curve slopes imply differences in consumers’ willingness to exchange dollars spent on other goods for commencement tickets (that is, differences in their marginal rates of substi- tution). Such differences create an opportunity for mutu- ally beneficial exchange between the various consumers in a market. In a later chapter we will show something you may be surprised to learn: although a resale market results in what seems like high prices for otherwise “free” tickets, both individuals buying and selling the tickets gain from the development of such a market.
Commencement tickets
0
A ′ C
Z
U2
U1
(a)
A B
E
Other goods
1
4
$10
Commencement tickets
0
A ′ C
E Z
U2
U3
(b)
A B
W
Other goods
4
The Allocation of Commencement Tickets (a) If given four free tickets rather than being charged $10 per ticket, the average senior is better off. (b) However, some students may prefer the $10-per-ticket pricing scheme to receiving four free tickets.
Figure 3.17
3.5 Are People Selfish? Having set out the basic components of the economic theory of consumer choice, we now
reconsider the general nature of the analysis. In particular, we wish to evaluate a commonly
made objection to the way economists characterize individual behavior. You may already
have heard someone observe: “Economics assumes people are greedy and care only about
Are People Self ish? 67
C03.INDD 10:51:15:AM 08/06/2014 PAGE 67Trim Size: 203.2 mm X 254 mm
material possessions” or “Economics disregards the fact that individuals are benevolent and
are concerned with the welfare of other people.”
A review of our basic assumptions about preferences reveals that these criticisms are not
valid. Economic analysis does not prejudge what commodities, services or activities people
consider to be economic “goods” or “bads.” In fact, because many things are “goods” for
some people and “bads” for others (for example, liver, liquor, big-time wrestling, ballet,
chewing tobacco, video games, cigarettes, reading economic theory), any attempt to specify
in advance what all people consider desirable would frequently lead to mistakes.
If we are unable to specify which goods people find desirable, though, how can we apply
the theory to concrete situations? The answer is simple: people reveal that some commodi-
ties are desirable by the way they allocate their spending. When we observe some con-
sumers giving up money in return for Internet service, the evidence is fairly conclusive
that Internet service is a desirable good for them. We would then draw convex indifference
curves between Internet service and other goods for such consumers and investigate how
their incomes, the price of Internet service, and so on, affect consumption decisions.
People give up time and resources to pursue charitable endeavors, sacrifice material wealth
for a quiet life, or campaign for politicians. To show how economic theory can be applied to
examine the factors that influence such decisions, consider a hypothetical situation.
Samantha (Sam) and Oscar are friends. Sam earns $90,000 a year; Oscar earns only $10,000
a year. Let’s assume that Sam cares about the material well-being of Oscar—or, more precisely,
that Oscar’s income (which determines the material comforts he can enjoy) is an economic
good for Sam. Does this concern imply that Sam will give some of her income to Oscar?
Figure 3.18 illustrates how we can apply indifference curve analysis to this situation.
Figure 3.18a shows Sam’s budget line AZ, relating her income and Oscar’s. Sam’s budget line does not intersect the vertical axis because we suppose she can’t take income from
Oscar. Point A shows their initial incomes. Obviously, Sam can increase Oscar’s income by giving Oscar some of her income, but every dollar she gives to Oscar reduces her own
income by a dollar, so the slope of the budget line is −1. Figure 3.18a shows two of Sam’s
U1
0 0Z Z
U2 U1
U2
(a) (b)
A
$1
$1
Sam’s income
(in dollars)
Oscar’s income
(in dollars)
90,000
85,000
15,00010,000 10,000
W
A
Sam’s income
(in dollars)
Oscar’s income
(in dollars)
90,000
Transferring Income to Another Person Altruistic preferences can also be accommodated in the analysis. In part (a) Sam chooses to give $5,000 of her income to Oscar. When Sam’s preferences are different, as in part (b)—still altruistic but less so than in part (a)—she will not give any of her income to Oscar.
Figure 3.18
68 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 68Trim Size: 203.2 mm X 254 mm
APPLICATION 3.5
A recent study indicated that households in the United States with annual incomes of $10,000 or less gave 5.2 per- cent of their income to charity while households with incomes between $10,001, and $50,000 gave an average of 2.5 per- cent and households with more than $50,000 in income gave 2.1 percent.7 Does this evidence indicate that altruism is an inferior good? The answer is no, and Figure 3.19 indicates why in the context of the example involving Sam and Oscar.
An increase in Sam’s annual income, everything else being held constant, involves an outward parallel shift in her bud- get line, AZ. Figure 3.19 depicts two such outward shifts from an initial hypothetical budget line, AZ, that involves both Sam and Oscar having an initial annual income of $10,000. The first outward parallel shift (to A′Z′) reflects Sam’s income rising to $50,000, while the even further shift (to A″Z″) is associated with Sam’s income jumping to $100,000.
Is Altruism a Normal Good?
Based on the giving data provided, assume that Sam opts to give away $520 when her annual income is $10,000 (5.2 percent of the total); $1,250 when her annual income is $50,000 (2.5 percent of the total); and $2,100 when her annual income is $100,000 (2.1 percent of the total). By connecting the optimal consumption points (W, W′, and W″) associated with the different income levels, we see that Sam’s income-consumption curve is still posi- tively sloped even though she is giving a smaller percent- age to Oscar as her income rises. Based on the positively sloped income–consumption curve, charitable giving can be taken to be a normal good for Sam over the income range depicted in Figure 3.19.
Of course, both the percentage pattern in charitable giving and whether charitable giving is strictly a normal good over all income levels can vary with the specific income levels examined. The New Tithing Group, a char- ity that tracks giving, reported that in a recent year the top 400 American earners made an average of $174 million and gave away, on average, $25.3 million, or 14.5 percent.
7This application is based on “The Very Rich, It Now Appears, Give Their Share and Even More,” New York Times, January 1, 2004, pp. C1 and C3.
Income-consumption curve
W″ A″
A′
A
Z
U3
U2
U1
0
Sam’s income (in dollars)
Oscar’s income (in dollars)
$100,000
$50,000
$10,000
$10,000 $12,100 $10,520 $11,250
W
W ′
Z″ Z ′
Is Altruism a Normal Good? Even though Sam gives away a smaller percentage of her income as her income rises from $10,000 to $50,000 and then to $100,000, charitable giving is still a normal good for Sam over this income range. Her income-consumption curve is positively sloped.
Figure 3.19
The Uti l i ty Approach to Consumer Choice 69
C03.INDD 10:51:15:AM 08/06/2014 PAGE 69Trim Size: 203.2 mm X 254 mm
indifference curves, and because both her income and Oscar’s are economic goods, they
have the usual shapes. Given the preferences indicated by Sam’s indifference curves, at
point A she would be willing to pay more than $1 to raise Oscar’s income by $1. Thus, it is in her interest to give some of her income to Oscar, and the best-sized gift from her point of
view is $5,000.
The fact that Sam cares about Oscar’s income is not sufficient to imply that she will
always donate money to Oscar’s cause. Figure 3.18b shows an alternative set of indiffer-
ence curves that Sam might have. Because these curves slope downward, they imply that
Sam still views Oscar’s income as an economic good. At point A, however, the slope of U1 is less than the slope of the budget line. Thus, Sam might be willing to give up, for example,
$0.25 in return for Oscar having $1 more in income; unfortunately, however, it would cost
her $1 to increase Oscar’s income by $1 (the slope of the budget line), so she decides not to
contribute anything to the Oscar fund. A corner equilibrium results.
Both parts of Figure 3.18 show preferences implying that Sam views Oscar’s income as
an economic good, but the intensity of preferences differs. The differing intensity is shown
by the different slopes of the indifference curves (the MRSs) at point A in the two diagrams. Thus, the fact that Sam cares about Oscar’s income does not allow us to conclude that she
will necessarily transfer any of her income to Oscar: the intensity of her preferences and the
cost of giving play critical roles.
3.6 The Utility Approach to Consumer Choice An alternative way to understand consumer choice theory is through the concepts of total
and marginal utility. In this section we explain this alternative approach and relate it to the
indifference curve approach emphasized earlier.
Let’s assume that we can measure the amount of satisfaction a consumer gets from any
market basket by its utility. Units in which utility is measured are arbitrary, but they are
commonly referred to as utils: a util is one unit of utility. In consumers’ minds, consumption of goods provides them with utility. But we must
distinguish between the concepts of total utility and marginal utility. Table 3.2 illustrates
the difference. Suppose Marilyn, the consumer, purchases only champagne (C), so for the moment consider only the first three columns. Total utility from champagne consump- tion (TUC) is the total utils Marilyn gets from a given number of glasses of champagne. If two glasses are consumed, total utility is 38 utils. Total utility is obviously greater at
higher levels of consumption, because champagne is an economic good. The
marginal utility of champagne (MUC) refers to the amount total utility rises when con- sumption increases by one unit. When champagne consumption increases from three to
total utility assuming that it is measurable, the total satisfaction a consumer receives from a given level of consumption
marginal utility the amount by which total utility rises when consumption increases by one unit
Table 3.2 Total and Marginal Utility Champagne
(glasses) TUC MUC Perfume (ounces) TUF MUF
1 20 20 1 50 50
2 38 18 2 85 35
3 53 15 3 110 25
4 65 12 4 130 20
5 75 10 5 145 15
6 83 8 6 155 10
70 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 70Trim Size: 203.2 mm X 254 mm
four glasses, total utility rises from 53 to 65 utils, or by 12 utils. The marginal utility of
the fourth glass is thus 12 utils.8
Table 3.2 also illustrates the assumption of diminishing marginal utility. This assump- tion holds that as more of a given good is consumed, the marginal utility associated with
the consumption of additional units tends to decline, other things being equal. (In particu-
lar, the other-things-being-equal condition means that consumption of other goods is held
fixed as consumption of the good in question is varied.) In Table 3.2, the marginal utility of
the first glass of champagne is 20 utils, but it is 18 utils for the second glass (the increase in
TUC from 20 to 38 utils), and so on. Note that the MUC of each successive glass is smaller. Table 3.2 also shows the total and marginal utility associated with different levels of
consumption of a second good, perfume (F). The total utility of a market basket containing champagne and perfume is then the sum of TUC and TUF. (This statement assumes that the utility derived from perfume consumption is independent of champagne consumption, and
vice versa. While this assumption will not always be true, its use simplifies the explanation
of the theory without materially affecting the results.) With consumption of five glasses of
champagne and 3 ounces of perfume, total utility is 185 utils. Obviously, the consumer will
choose the market basket yielding the greatest total utility, subject to the limitation implied
by her income and the prices of the two goods.
The Consumer’s Optimal Choice If the consumer’s income and the prices of champagne and perfume are specified (PC and PF), we could consult Table 3.2 and by trial and error eventually find the market basket of champagne and perfume that produces the greatest total utility. With an income of $65 and
the prices of champagne and perfume at $5 and $10, we would eventually find that the mar-
ket basket composed of five glasses of champagne and 4 ounces of perfume produces more
total utility (205 utils) than any other market basket costing $65.
There is a simpler way to proceed. As it turns out, the utility-maximizing market basket
is one for which the consumer allocates income so that the marginal utility divided by the
good’s price is equal for every good purchased:
MUC/PC = MUF/PF.
A market basket of 5C and 4F satisfies this equality: MUC/PC is equal to 10 utils per glass/$5 per glass, or 2 utils per dollar, and MUF/PF is equal to 20 utils per ounce/$10 per ounce, or 2 utils per dollar. These ratios measure how much additional utility is gener-
ated by spending $1 extra on each good. With MUC/PC equal to 2 utils per dollar, $1 more spent on champagne (purchasing one-fifth of a glass) will generate 2 utils in additional
utility (one-fifth of the MU associated with the fifth champagne glass). Put slightly differ- ently, MU/P is the marginal rate of return, in terms of utility, earned by the consumer if she invests $1 extra in a good. The rule for maximizing utility thus can equivalently be stated
as allocating income among goods so that the marginal rates of return, measured in terms
of utils per dollar “invested,” are equalized across all the goods in which the consumer
invests.
If this equality is not satisfied, total utility can be increased by a rearrangement in the
consumer’s purchases. Suppose that Marilyn buys 3C and 5F. This market basket also costs $65, but total utility is now 198 utils, according to Table 3.2. With this market basket
we have MUC/PC > MUF/PF, or 15 utils/$5 > 15 utils/$10. This inequality shows that the marginal dollar devoted to champagne yields 3 utils, while the marginal dollar devoted to
diminishing marginal utility the assumption that as more of a given good is consumed, the marginal utility associated with the consumption of additional units tends to decline, other things being equal
8In Table 3.2, the marginal utility when consumption is six glasses is in the same row as six glasses of champagne. Because marginal utility refers to the change in going from five to six glasses, however, some writers prefer to place marginal utility halfway between the fifth and sixth rows. Either procedure is accept- able as long as the correct meaning is communicated.
The Uti l i ty Approach to Consumer Choice 71
C03.INDD 10:51:15:AM 08/06/2014 PAGE 71Trim Size: 203.2 mm X 254 mm
perfume yields only 1.5 utils. Since $1 spent on champagne generates a higher return in
terms of utility, shifting $1 from perfume to champagne consumption will increase total
utility. Spending $1 less on perfume reduces utility by 1.5 utils, but spending $1 more on
champagne increases it by 3 utils, a net gain of 1.5 utils. As long as an inequality persists,
the consumer should reallocate purchases from the good with a lower marginal utility per
dollar of expenditure to the good with a higher marginal utility per dollar of expenditure.
Shifting dollar outlays from perfume to champagne will eventually reestablish the equality
condition. As champagne consumption increases, its MU falls (reducing MUC/PC) because of the assumption of diminishing marginal utility; as perfume consumption falls, its MU rises (increasing MUF/PF). When the equality condition is reestablished, at 5C and 4F in this example, the consumer is maximizing utility with the given income and prices.
An analogy from the finance world may help cement your understanding of the forego-
ing rule for maximizing consumer utility. Suppose that you have a portfolio of $1 million
invested partly at Bank A and partly at Bank B. Your Bank A investment earns a 3 percent annual rate of return, while dollars saved at Bank B earn 1.5 percent per year. What would be wrong with such a financial strategy provided that investing at either bank is riskless?
The answer is probably clear. You would be better off shifting money from Bank B to Bank A as long as the rate of return at A is higher. The same principle applies if you are trying to maximize utility by purchasing various goods (investing in various “banks”) with
a limited budget.
Relationship to Indifference Curves As noted earlier in this chapter, an indifference curve shows alternative market baskets
yielding the same total utility to the consumer. Figure 3.20 shows two indifference curves.
The consumer’s MRSFC between points R and T along indifference curve U2 is ΔC/ΔF. This slope, however, can also be explained in terms of the marginal utilities of the two goods.
Along an indifference curve, the slope ΔC/ΔF equals the ratio of the marginal utilities of the two goods. Suppose ΔC/ΔF = 2C/1F. That is, 1 ounce of perfume will replace two glasses of champagne without affecting total utility. If 1F will replace 2C, then the mar- ginal utility of 1 ounce of perfume must be twice as great as the marginal utility of one
glass of champagne. Thus, the slope ΔC/ΔF equals MUF/MUC. The indifference curve slope measures the relative importance of the two goods to the consumer, which in turn is equal
to their relative marginal utilities.
We can demonstrate this conclusion more formally. In Figure 3.20, the movement from
R to S, ∆C, reduces total utility by an amount equal to ∆C × MUC. (If ∆C is two glasses and the MU per glass of C is 5 utils, then total utility falls by 10 utils.) Similarly, the movement
MRS and Marginal Utilities The slope of an indifference curve is related to the marginal utilities of the two goods. At point R the slope is ΔC/ΔF, and this ratio equals MUF/MUC.
Perfume (F)0
S
R
T
ÊF
ÊC
U2
U1
Champagne (C)Figure 3.20
72 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 72Trim Size: 203.2 mm X 254 mm
from S to T, ΔF, increases total utility by an amount equal to ∆F × MUF. Because R and T lie on the same indifference curve, the loss in utility associated with a move from R to S must be exactly offset by the gain in utility in going from S to T. Therefore, we have:
Δ ΔC MU F MU× = ×C F.
Rearranging these terms, we find:
Δ ΔC F MU MU/ / .= F C Because ΔC/ΔF equals MRSFC, we can substitute terms and obtain:
MRS MU MUFC F C= / .
Earlier in this chapter, we noted that, as long as there is not a corner solution, the consumer’s
optimal choice is where the indifference curve slope equals the slope of the budget line:
MRS P PFC F C= / .
Since MRSFC equals MUF/MUC, we can substitute terms and rewrite the optimality condi- tion as:
MU MU P PF C F C/ / .=
Then, by rearranging terms, we obtain:
MU P MU PF F C C/ / ,=
which is the condition for the optimal consumption choice when using the utility theory
approach.
An equality between the marginal utility per dollar’s worth of both goods is the same
as an equality between the MRS and the price ratio. The utility theory and the indifference curve approach are thus simply different ways of viewing the same thing.
SUMMARY
The theory of consumer choice is designed to
explain why consumers purchase the goods they do.
The theory emphasizes two factors: the consumer’s
preferences over various market baskets and the con-
sumer’s budget line, which shows the market baskets
that can be bought.
An indifference curve graphically depicts all the com-
binations of goods considered equally desirable by a
consumer.
For economic “goods,” indifference curves are assumed
to be downward sloping, convex, and nonintersecting.
The slope of an indifference curve measures the
marginal rate of substitution (MRS), which is the will- ingness of the consumer to trade one good for another.
A budget line shows the combinations of goods a
consumer can purchase with given prices for the good
and assuming all the consumer’s income is spent on the
good.
The consumer’s income and the market prices of the
goods determine the position and slope of the budget
line. The slope of the budget line is equal to the ratio of
the prices of the goods and measures the relative price of
one good compared with another.
From among the market baskets the consumer can
purchase, we assume the consumer will select the one
that results in the greatest possible level of satisfaction
or well-being. Graphically, this optimal choice is shown
by the tangency between the budget line and the indiffer-
ence curve, where the consumer’s MRS equals the price ratio.
A change in the consumer’s budget line leads to a
change in the market basket selected.
An income increase when the prices of goods are
held constant parallel shifts out the budget line. Either
an increase or a decrease in the consumption of a good
may result.
When the consumption of a good rises with an
increase in income, the good is a normal good.
An inferior good is one for which consumption falls
as income increases.
The utility approach to consumer choice does not dif-
fer significantly from the indifference curve approach.
C03.INDD 10:51:15:AM 08/06/2014 PAGE 73Trim Size: 203.2 mm X 254 mm
Review Quest ions and Problems 73
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
3.1 Imelda spends her entire income on shoes and hats. Draw the budget line for each of the following situations, identifying
the intercepts and the slope in each case.
a. Monthly income is $1,000, the price of a pair of shoes is $8, and the price of a hat is $10.
b. Same conditions as in part (a), except that income is $500. c. Same conditions as in part (a), except that income is $2,000
and the price of a pair of shoes is $16.
d. Same conditions as in part (a), except that hats cost $5 each.
*3.2 In most cases, a consumer can purchase any number of units of a good at a fixed price. Suppose, however, that a consumer
must pay $10 per visit to an amusement park for the first five visits
but only $5 per visit beyond five visits. What does the budget line
relating amusement park visits and other goods look like?
*3.3 Bill’s budget line relating hamburgers and french fries has intercepts of 20 hamburgers and 30 orders of french fries. If the
price of a hamburger is $3, what is Bill’s income? What is the per-
order price of french fries? What is the slope of the budget line?
3.4 During an economic downturn, Taco Bell increases sales per outlet and gains overall market share in the fast-food market.
By contrast, competitors such as McDonald’s and Burger King
see their sales and market share decline. Assuming that the
relative price of the items sold by the various fast-food chains
as well as other factors remain unchanged, does this evidence
indicate that the products sold by Taco Bell are normal or infe-
rior goods for the typical consumer? Explain. What about the
products sold by McDonald’s and Burger King?
3.5 Elton says, “To me, Coke and Pepsi are both the same.” Draw several of Elton’s indifference curves relating Coke and Pepsi.
3.6 People’s rankings of activities are sometimes described in terms of their “priorities.” For example, some students claim
that getting good grades takes priority over watching televi-
sion or dating. Does this ranking mean they never engage in
the latter activities and spend all their time studying? If people
do consume both high-priority and low-priority goods, what do
we mean when we say that some goods have higher priorities
than others?
3.7 Draw a set of indifference curves relating two “bads” such as smog and garbage. What characteristics do these curves have?
3.8 With the per-unit prices of broccoli (B) and pork rinds (R) equal to $2 and $1, a consumer, George, with an income of
$1,000 purchases 400R and 300B. At that point, the consumer’s MRSBR = 2R/1B. Does this mean that George would be just as well off consuming 200R and 400B? Explain with a diagram.
*3.9 Marilyn spends her entire monthly income of $600 on champagne (C ) and perfume (F ). The price of a bottle of champagne is $30 and the price of an ounce of perfume is $10.
If she consumes 12 bottles of champagne and 24 ounces of per-
fume, her MRS is 1C/1F. Is her choice optimal? Explain your answer with a diagram.
*3.10 Seat belts in cars were available as options before they were required by law. Most motorists, however, did not buy
them. Assuming that motorists were aware that seat belts
reduced injuries from accidents, were motorists irrational in
not purchasing them?
3.11 Is it inconsistent to claim that (a) people’s preferences dif- fer and (b) at their current consumption levels, their marginal
rates of substitution are equal?
3.12 Explain why the food stamp program can have the same effect on the consumption pattern and well-being of recipi-
ents as an outright cash transfer of the same cost. Why do
you think it is not converted into an explicit cash transfer
program, thereby saving the cost of printing and redeeming
food stamps?
*3.13 Is it possible for all goods a consumer buys to be nor- mal? Is it possible for all goods a consumer buys to be inferior?
*3.14 Consider two market baskets, A ($100 worth of other goods, O, and 10 video rentals, V) and B ($150 worth of other goods and 10 video rentals). If video rentals are a normal good,
will the consumer’s MRSVO be greater when basket A or bas- ket B is consumed? What if video rentals are an inferior good? Depict in a diagram.
3.15 Prior to 1979, the food stamp program required families to pay a certain amount for food stamps. Suppose a family can
receive $150 in food stamps for a payment of $50; no other
options are offered. How would this policy affect the budget
line? Compared with an outright gift of $100 in food stamps,
which is the way the program now works, would this policy
lead to more, less or the same food consumption?
3.16 Suppose that Thurston, a color-blind consumer, has $80 to spend on either pink or lime-green sweaters. Thurston does not
care what color sweater he wears but deems it very important to
buy as many sweaters as possible with the $80. Pink sweaters
cost $40 each and lime-green ones cost $20 each.
a. Draw Thurston’s budget line and indifference map. What is Thurston’s optimal consumption choice?
b. A sale on pink sweaters begins. If a consumer buys two pink sweaters at the regular price, he or she can get two additional
pink sweaters for free. Two pink sweaters must be purchased
to get the deal. Otherwise, prices are unchanged. With the
sale, depict Thurston’s new budget line and preferred con-
sumption point.
3.17 Using an indifference map, explain why U.S. parents with two children both of the same sex are are more likely to go for
a third child than are parents with one child of each gender.
3.18 An Engel curve is a relationship between the consumer’s income and the quantity of some good consumed (the price of
74 The Theory of Consumer Choice
C03.INDD 10:51:15:AM 08/06/2014 PAGE 74Trim Size: 203.2 mm X 254 mm
the good fixed). Income is measured on the vertical axis, and the
quantity of the good consumed is measured on the horizontal
axis. Draw the Engel curves for compact discs and hamburger
from Figures 3.14 and 3.15. How does the slope of an Engel
curve identify whether the good is normal or inferior?
3.19 Measure the income of Samantha on the vertical axis and the income of Oscar on the horizontal axis, as we did in
Figure 3.18. Draw several of Sam’s indifference curves under
the following circumstances.
a. Sam doesn’t care about Oscar’s income, but the higher her own income is, the better off she is.
b. Sam considers both her own income and Oscar’s income to be economic “goods,” but only as long as her income
exceeds Oscar’s. When Oscar’s income exceeds hers, Sam
considers Oscar’s income to be an economic “bad.”
3.20 If Sam’s preferences relating her own income and Oscar’s income conform to the Golden Rule (“Love thy neigh-
bor as thyself”), what would her indifference curves in a dia-
gram like Figure 3.18 look like?
3.21 Sam is subject to a 40 percent tax on income, and her after-tax income is $90,000, as in Figure 3.18. Now sup-
pose the government permits her to deduct her contributions
(to Oscar) from her income before the 40 percent tax rate is
applied. How will this deduction affect her budget line?
3.22 When we use the composite-good convention, what do we mean by a composite good and how do we measure it? What is the slope of the budget line? What is the slope of an
indifference curve? Does the consumer equilibrium involve an
equality between MRS and a price ratio?
3.23 Suppose that you have only 9 hours left to cram for final exams and you want to get as high an average numerical grade
as possible in three courses: marketing, accounting, and micro-
economics. Your grade in each course depends on the time
devoted to studying the subjects in the following manner:
Marketing Accounting Microeconomics
Hours of Study Grade
Hours of Study Grade
Hours of Study Grade
0 90 0 80 0 0
1 97 1 90 1 40
2 98 2 97 2 60
3 99 3 98 3 70
4 100 4 99 4 79
5 100 5 100 5 87
6 100 6 100 6 94
7 100 7 100 7 100
How many hours should you devote to studying each subject?
3.24 Using indifference maps, explain how Kraft’s makeover of the Oreo successfully accounted for different consumer
preferences in China versus for low-sugar, wafer-like cookies
versus higher-sugar, biscuit-like cookies in the United States.
3.25 The federal government bans the use of food stamps to buy cigarettes, alcohol or prepared foods such as deli sandwiches
and restaurant entrees. Should a similar ban be implemented
against the purchase of soda and sweetened drinks due to the
link between the consumption of sugar-sweetened beverages and
weight gain and the development of diabetes? (New York City
proposed such a ban in 2010.) Explain why or why not.
75
C04.INDD 10:52:49:AM 08/06/2014 PAGE 75Trim Size: 203.2 mm X 254 mm
CHAPTER 4 Individual and Market Demand
In Chapter 3, we developed a model using indifference curves and budget lines to explain consumer behavior and used it to examine how changes in income affect consumer choices.
In this chapter we use the consumer choice model to analyze the effects of price changes
and show how a consumer’s demand curve can be derived using the model. We also dis-
cuss how individual demand curves are aggregated to obtain the market demand curve. In
addition, we explain the concept of consumer surplus, which relates to how areas under the
demand curve can be used to measure the net benefits or costs to consumers from changes
in consumption. And finally, we cover demand estimation to show how individuals’ or
market demand curves can be estimated from real-world data.
This chapter completes our discussion of the basic elements of the theory of consumer choice. Chapter 5 will illustrate the wide range of applications of this theory to such diverse problems as pricing garbage collection, deciding how much to save and borrow, and deter-
mining how much to invest in various financial assets.
Learning Objectives
Understand how price changes affect consumption choices. Differentiate between the income and substitution effects associated with a price change on the consumption of a particular good. Explain the relation between income and substitution effects in the case of inferior goods. Explain how individual demand curves are aggregated to obtain the market demand curve. Demonstrate how consumer surplus represents the net benefit, or gain, to an individual from consuming one market basket instead of another. Investigate the relationship between own-price elasticity of demand and the price–consumption curve. Examine network effects: the extent to which an individual consumer’s demand for a good is influenced by other individuals’ purchases. Overview the basics of demand estimation. *Derive the Consumer’s Demand Curve Mathematically.
Memorable Quote “Nobody goes there no more, it’s too crowded.”
—Yogi Berra, Major League Baseball player and manager on why he no longer went to a Minneapolis restaurant, evidence of a negative network (or snob) effect at work
76 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 76Trim Size: 203.2 mm X 254 mm
4.1. Price Changes and Consumption Choices Let’s examine the way a change in a good’s price affects the market basket chosen by a
consumer. Because we wish to isolate the effect of a price change on consumption, we hold
constant other factors such as income, preferences, and the prices of other goods.
Figure 4.1a depicts a consumer deciding how to allocate a given amount of annual
income between college education (C) and all other goods. The per-credit-hour price of college education, $250, is indicated by the slope of the budget line since the per-dollar
price of outlays on all other goods can be taken to be unity. With an initial budget line of
AZ, the consumer’s optimal consumption point is W. The consumption of college education is C1, and outlays on other goods are A1.
Derivation of the Consumer’s Demand Curve A reduction in the price of college education, with income, preferences, and the prices of other goods remaining fixed, leads the consumer to purchase more units of college education. (a) The optimal market baskets associated with alternative prices for college education are connected to form the price–consumption curve. (b) The same information is plotted as the consumer’s demand curve for college education.
(a)
(b)
0 C1 Z ′ Z″
U3 U2
U1
W ′
W
W ″
A1
A2
C2
$250 $200 $150
C3
1C1C
Units of college education
Units of college education
0 C1
F
G
H
d
$250 = P1
$200 = P2
$150 = P3
Per unit price of college education
C2 C3
Z
A
Other goods
1C
P–C curve
Figure 4.1
Price Changes and Consumption Choices 77
C04.INDD 10:52:49:AM 08/06/2014 PAGE 77Trim Size: 203.2 mm X 254 mm
If the price of college education falls from $250 to $200, the budget line rotates around
point A and becomes flatter. With a price of $200, the budget line becomes AZ′, where Z′ equals the consumer’s constant income divided by the lower price of college education.
Confronted with this new budget line, the consumer selects the most preferred market basket
from among those available on AZ′. For the particular preferences shown, the preferred basket is point W′, where the slope of U2 (the marginal rate of substitution) equals the slope of the flatter budget line AZ′. Consumption of college education has increased to C2 in response to the reduction in its price. If the price of college education falls still further to
$150, then the budget line becomes AZ″, and the consumer will choose point W″, with the amount of credit hours of college education consumed equal to C3.
Proceeding in this way, we can vary the price of college education and observe the mar-
ket basket chosen by the consumer. For every possible price, a different budget line results
and the consumer selects the market basket that permits attainment of the highest possible
indifference curve. Points W, W′, and W″ represent three market baskets associated with prices of $250, $200, and $150, respectively. If we connect these optimal consumption
points, and those associated with other prices (not drawn in explicitly), we obtain the price– consumption curve, shown as the P–C curve in the diagram. The price–consumption curve identifies the optimal market basket associated with each possible price of college educa-
tion, holding constant all other determinants of demand.
The Consumer’s Demand Curve Using the procedure just described, we can determine the consumer’s demand curve for
college education. The demand curve relates consumption of college education to its price,
holding constant such factors as income, the prices of related goods, and preferences. The
price–consumption curve does the same thing, although it is not itself the demand curve. To
convert the price–consumption curve to a demand curve, we simply plot the price–quantity
relationship identified by the price–consumption curve in the appropriate graph.
Figure 4.1b shows the consumer’s demand curve d (as before, we use lowercase let- ters to indicate the individual consumer’s demand curve); it indicates the quantity of col-
lege education the consumer will buy at alternative prices, other factors held constant. The
demand curve is determined by plotting the price–quantity combinations identified by the
price–consumption curve in Figure 4.1a. For example, when the price of college education
is $250 (the slope of AZ), consumption of college education is C1 at point W in Figure 4.1a. Figure 4.1b shows the price of college education explicitly on the vertical axis. When the
price is $250, consumption is C1, so point F locates one point on the demand curve. When the price is $200 (the slope of AZ′), consumption is C2, which identifies a second point, G, on the demand curve. Other points are obtained in the same manner to plot the entire demand curve d.
Some Remarks about the Demand Curve We have just derived a consumer’s demand curve from the individual’s underlying prefer-
ences (with a given income and fixed prices of other goods). This approach clarifies several
points about the demand curve:
1. The consumer’s level of well-being varies along the demand curve. This point is clear from Figure 4.1a, where the consumer reaches a higher indifference curve when the price
of college education falls. The diagram specifies why the consumer benefits from a lower
price: the consumer can now purchase market baskets that were previously unattainable.
2. The prices of other goods are held constant along a demand curve, but the quantities purchased of these other goods can vary. For example, in Figure 4.1a, consumption of all
other goods falls from A1 to A2 when the price of college education falls from $250 to $200.
price–consumption curve a curve that identifies the optimal market basket associated with each possible price of a good, holding constant all other determinants of demand
78 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 78Trim Size: 203.2 mm X 254 mm
Because all other goods are lumped together and treated as a composite good, the way in
which consumption of any other specific good may change is not shown explicitly.
3. At each point on the demand curve, the consumer’s optimality condition MRSCO = PC/PO is satisfied. (The subscript O refers to other goods, the composite good measured on the vertical axis.) As the price of college education falls, the value of PC/PO becomes smaller, and the consumer chooses a market basket for which MRSCO, the slope of the indifference curve, is also smaller.
4. The demand curve identifies the marginal benefit associated with various levels of consumption. The height of the demand curve from the horizontal axis, at each level of
consumption, indicates the marginal benefit of the good. For example, when consumption
is C2 at point G on the demand curve (Figure 4.1b), the distance GC2 (or $200) is a measure of how much the marginal unit of college education consumed is worth to the consumer.
Why? Refer to Figure 4.1a; for a market basket selected by a consumer to be optimal, such
as W′ when the price of college education is $200, MRSCO equals PC/PO. Since PO can be taken to equal unity, this implies that MRSCO = PC. Thus, at point W′ in Figure 4.1a, MRSCO is equal to $200 per unit of college education. Because the MRS is a measure of what the consumer is willing to give up for an additional credit hour of college education, it is a
measure of the marginal benefit. Note that at every point on the demand curve the height of the demand curve equals the MRS, thereby indicating the marginal benefit of the good to the consumer. For this reason economists refer to the price at which people purchase a given good as revealing the relative importance of the good to them.
APPLICATION 4.1
Although economists ascribe an important role to price in determining the quantity demanded of a product, policy- makers often do not. A case in point is the campaign waged by policymakers since the mid-1970s to discourage alcohol abuse and thereby decrease the number of traffic-related deaths. One of the main campaign objectives has been to raise the legal age for alcohol consumption to 21 years. The reason behind this is that while people under the age of 25 represent 20 percent of all licensed drivers, they account for 35 percent of all drivers involved in fatal accidents. Alcohol is involved in more than half the driving fatalities accounted for by young drivers.
By raising the legal age for alcohol consumption to 21, policymakers hope to shift the demand curve for alcohol to the left (diminishing the portion of the population with access to alcohol) and thereby reduce both alcohol abuse and driving fatalities.
Shifting the demand curve for alcohol to the left is one way to reduce alcohol abuse and traffic fatalities. However, economic research suggests that a more effective method, even among teenagers, would be to raise the price of alcohol through higher taxes, thereby producing a movement along the demand curve for alcohol.
The federal tax on alcohol was constant in nominal dollar terms between 1951 and 1991 ($9 per barrel of beer,
Using Price to Deter Youth Alcohol Abuse, Traffic Fatalities, and Campus Violence
$10.50 per proof gallon of distilled spirits such as vodka, and so on) and, following an increase in 1991 (to $18 per barrel of beer and $13.50 per proof gallon of distilled spir- its), has remained constant since then. In real terms, con- sequently, the federal tax on alcohol has decreased since 1951. For example, the real federal tax on beer has declined by 89 percent since 1951, while the real tax on distilled spirits has decreased by 94 percent. The decline, in real terms, of the federal tax on alcohol is a major factor behind the substantial decrease in the real price of alcohol since 1951—over 90 percent in the case of beer and 60 percent for hard liquor.
A national survey of teenagers finds that, holding constant other factors such as a state’s minimum drink- ing age, religious affiliation, and proximity to bordering states with lower minimum drinking ages, the amount of alcohol consumed by the average teenager in a state is significantly influenced by the price of alcohol there.1
1According to Douglas Coate and Michael Grossman, “Effects of Alcoholic Beverage Prices and Legal Drinking Ages on Youth Alcohol Use,” Journal of Law and Economics, 31, No. 1 (April 1988), pp. 145–172, the estimated price elasticity of demand for teenage drinking ranges from 0.5 to 1.2.
Price Changes and Consumption Choices 79
C04.INDD 10:52:49:AM 08/06/2014 PAGE 79Trim Size: 203.2 mm X 254 mm
Do Demand Curves Always Slope Downward? For the specific indifference curves shown in Figure 4.1, we derived a downward-sloping
demand curve. But does the demand curve always slope downward? Is it possible for a con-
sumer to have indifference curves so that the law of demand does not hold for some goods?
Figure 4.2 suggests such a possibility. When the budget line is AZ, consumption of good X is X1 units. If the price of X falls so that the budget line becomes AZ′, consumption of X falls to X2, an apparent violation of the law of demand. Note that the indifference curves that produce this result are downward sloping, nonintersecting, and convex; that is, they do
not contradict any of our basic assumptions about preferences.
The survey findings suggest that raising taxes on alcohol offers a potent mechanism for deterring alcohol abuse and traffic fatalities among teenagers. Specifically, based on the survey’s results, had federal taxes on alcohol remained con- stant since 1951 in real purchasing power terms rather than in dollar terms, teenage drinking would have fallen more than if the minimum drinking age had been raised to 21 in all states. Raising the price of drinking and moving along the demand curve for alcohol thus promises to be more effec- tive in reducing teenage drinking than the policy pursued by most policymakers—shifting the demand curve for alcohol to the left by imposing age restrictions.
According to another study‚ the decline in the real price of alcohol also appears to have resulted in an increase in campus violence.2 Currently‚ a third of the college student population of 22 million in the United States is expected to be involved‚ in any given year‚ in some kind of campus
violence (arguments‚ fights‚ run-ins with police or college authorities‚ sexual misconduct‚ and so on). Because alco- hol consumption is positively correlated with violence‚ the study examined the relationship between prices of six-packs of beer and levels of violence at colleges around the country. The study found that a 10 percent increase in the price of beer would be sufficient to decrease cam- pus violence by 4 percent, other factors held constant. However‚ since the real price of beer has actually fallen by 10 percent since 1991—largely due to the decline‚ in real terms‚ of the federal tax on alcohol—the converse result has occurred. The results of the study indicate that campus violence has increased by 4 percent (300‚000 incidents annually) since 1991 due to the decline in the real price of beer.
2“How to Calm the Campus‚” Businessweek‚ November 1‚ 1999‚ p. 32.
A Lower Price Leading to Less Consumption A consumer purchases less of good X when its price falls, an apparent violation of the law of demand. In this case the demand curve will be upward sloping.
0
U1
U2
Other goods
Good XZ Z ′
W ′
W
X2 X1
A
Figure 4.2
80 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 80Trim Size: 203.2 mm X 254 mm
Just because we can draw a diagram that shows reduced consumption at a lower price
does not mean such an outcome will ever be observed in reality. It does suggest, how-
ever, the importance of carefully considering exactly why consumption of a good varies in
response to a change in its price. We illustrate this idea in the following sections.
4.2 Income and Substitution Effects of a Price Change When the price of a good changes, the change affects consumption in two different ways.
Normally, we cannot observe these two effects separately. Instead, when the consumer
alters consumption in response to a price change, all we see is the combined effect of both
factors. Nevertheless, it is useful to analytically break down the effects of a price change
into these two components.
The first way a price change affects consumption is the income effect. When the price of a good falls, a consumer’s real purchasing power increases, which affects consumption of
the good. A price reduction increases real income—that is, makes it possible for the con- sumer to attain a higher indifference curve.
The second way a price reduction affects consumption is the substitution effect. When the price of one good falls, the consumer has an incentive to increase consumption of that
good at the expense of other, now relatively more expensive, goods. The individual’s con- sumption pattern will change in favor of the now less costly good and away from other
goods. In short, the consumer will substitute the less expensive good for other goods—
hence the name substitution effect. To see intuitively that two different factors are at work when a price changes, compare
Figure 3.14 from Chapter 3 and Figure 4.1. In Figure 4.1, a price reduction results in the
consumer reaching a higher indifference curve. In Figure 3.14, an increase in income, with
no change in prices, also results in the consumer reaching a higher indifference curve. Appar-
ently, a common factor is at work: both a reduction in price and an increase in income raise
the consumer’s real income, in the sense of permitting attainment of greater well-being. In
both cases the budget line moves outward, allowing consumption of market baskets that
were not previously attainable. This points to one of the two ways a price reduction affects
income effect a change in a consumer’s real purchasing power brought about by a change in the price of a good
substitution effect an incentive to increase consumption of a good whose price falls, at the expense of other, now relatively more expensive, goods
APPLICATION 4.2
Since the 1960s, per capita consumption of red meat (beef, pork, veal, and lamb) in the United States has been declining, while per capita consumption of poultry (chicken and turkey) has been increasing. For example, poultry con- sumption grew from 34 to 70 pounds per capita between 1960 and 2012, while beef and pork consumption declined from 138 to 94 pounds per capita over the same time period.
Is the absolute and relative growth of poultry con- sumption due to a shift in preferences away from red meat? That is, has information that has come out about the link between red meat consumption and heart disease and other diseases led to a rightward shift in the demand for poultry? Experts who have studied the phenomenon attribute the growth in poultry consumption much more
Why the Flight to Poultry and away from Red Meat by U.S. Consumers?
to a decline in the real price of poultry.3 Between 1978 and 2000, for example, the price of a whole chicken dropped by 50 percent in constant dollars, and the real price of a skinless, boneless breast fell by 70 percent. In contrast, the price, in constant dollars, of red meat underwent no similar decline.
A decline in poultry’s relative price has induced a move- ment down a given demand curve for poultry to a higher quantity demanded of poultry. Less of a factor has been a shift in the entire demand curve due to a change in preferences.
3See, for example, B. Peter Pashigian, Price Theory and Applications, 2nd ed. (Boston: McGraw-Hill, 1998).
Income and Subst i tut ion Effects of a Pr ice Change 81
C04.INDD 10:52:49:AM 08/06/2014 PAGE 81Trim Size: 203.2 mm X 254 mm
consumption: it augments real income (by increasing the purchasing power of a given nomi-
nal income), which obviously affects consumption. This is the income effect.
Although a price reduction and an income increase both have an income effect on
consumption, there is a significant difference between them. With a price reduction the
consumer moves to a point on a higher indifference curve where the slope is lower than
it was at the original optimal consumption point (see Figure 4.1). In effect, the consumer
has moved down the indifference curve to consume more of the lower-priced good. This
result illustrates the substitution in favor of the less costly good. When income increases,
however, the consumer moves to a point on a higher indifference curve where the slope (the
MRS) is the same as it was prior to the income increase. This is so because if only income changes, the slope of the consumer’s budget line does not change. The precise distinction
between these two effects and the way they help us understand why the demand curve has
the shape it does are clarified next with a graphical treatment.
Income and Substitution Effects Illustrated: The Normal- Good Case In Figure 4.3, the consumer’s original budget line AZ relates annual credit hours of col- lege education and outlays on all other goods. At a per-credit-hour price of $250, the
optimal market basket is W, with C1 credit hours bought by the consumer. If the price of college education falls to $200, the budget line becomes AZ′, and the consumer buys C2 units. The increase in consumption of college education (from C1 to C2) in response to the lower price is the total effect of the price reduction on purchases of college education. The demand curve shows the total effect. Now we wish to show how this total effect can
be decomposed conceptually into its two component parts—the income effect and the
substitution effect.
Income and Substitution Effects of a Price Reduction The total effect (from C1 to C2) of a reduction in the price of college education can be separated into two components, the income effect and the substitution effect. A hypothetical budget line HH′ is drawn parallel to the new (after-price-change) budget line but tangent to the initial indifference curve U1 at point J. The substitution effect is then C1 to CJ, and the income effect is CJ to C2, which together give the total effect of the price decrease on the consumption of college education by the consumer.
0
U1
J
W W′
U2
Other goods
Total
S I
1C 1C $200
1C $200
A
H
Credit hours of college education per year (C)
Z Z′H′C1 C2CJ
$250
Figure 4.3
82 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 82Trim Size: 203.2 mm X 254 mm
The substitution effect illustrates how the change in relative prices alone affects consump- tion, independent of any change in real income or well-being. To isolate the substitution
effect, we must keep the consumer on the original indifference curve, U1, while at the same time confronting the individual with a lower price of college education. We do so by draw-
ing a new budget line with a smaller slope, reflecting the lower price, and then imagining
that the consumer’s income is reduced just enough (while holding the price of college educa-
tion at $200) so that the student can attain indifference curve U1. In other words, we move the AZ′ budget line toward the origin parallel to itself until it is tangent to U1. The result is the hypothetical budget line HH′, paralleling AZ′ (both reflect the $200 price) and tangent to U1 at point J. This new budget line shows that if, after the price decrease, the consumer’s income is reduced by AH, the preferred market basket will be point J on U1, the indiffer- ence curve attained by the consumer prior to the price decrease. (Remember from Chapter
3 that the height of the budget line’s intercept on the vertical axis represents the consumer’s
income when dollar outlays on all other goods are measured on the vertical axis.)
This manipulation permits us to separate the income and substitution effects so that each
can be identified independently. The substitution effect is shown by the difference between the market baskets at points W and J. The lower price of college education, looked at by itself, leads to an increased consumption of college education from C1 to CJ and reduces consumption of other goods. In effect, the substitution effect involves sliding down the
original indifference curve from point W, where its slope is $250 per credit hour, to point J, where its slope is $200 per credit hour. Consequently, the substitution effect of the lower
price increases consumption from C1 to CJ. The income effect is shown by the change in consumption when the consumer moves
from point J on U1 to point W′ on U2. This change involves a parallel movement in HH′ out to the AZ′ budget line. Recall that a parallel shift in the budget line indicates a change in income but no change in the price of college education. Thus, the income effect of the
lower price causes consumption of college education to rise from CJ to C2. The sum of the substitution effect (C1 to CJ) and the income effect (CJ to C2) measures
the total effect (C1 to C2) of the lower price on the consumption of college education. Any change in price can be separated into income and substitution effects in this manner.
Although this analysis may seem esoteric, it is highly significant. Ultimately, we are
seeking a firm basis for believing that people will consume more at lower prices—that is,
that the law of demand is valid. Separating the income and substitution effects allows us to
look at the issue more deeply.
Note that the substitution effect of any price change always implies more consumption of a good at a lower price and less consumption at a higher price. This relationship follows directly from the convexity of indifference curves: with convex indifference curves, a lower
price implies a substitution effect that involves sliding down the initial indifference curve to
a point where consumption of the good is greater. Thus, the substitution effect conforms
to the law of demand.
The income effect of a price change, however, implies greater consumption at a lower
price only if the good is a normal good. In Figure 4.3, when the budget line shifts from HH′ to AZ′ (a parallel shift), consumption of college education will rise if college education is a normal good.
The demand curve for a normal good must therefore be downward sloping. Both the substitution and income effect of a price change involve greater consumption of the good
when its price is lower.4 Because the total effect is the sum of the income and substitution
4Figure 4.3 shows the substitution and income effects for a price reduction. An increase is handled in a slightly different way. If we were considering an increase in the price of college education from $200 to $250 in the diagram, we would accomplish the separation into substitution and income effects by drawing
a hypothetical budget line with a slope of $250 (the new price) tangent to the indifference curve, U2, the consumer is on before the price increase.
Income and Subst i tut ion Effects of a Pr ice Change 83
C04.INDD 10:52:49:AM 08/06/2014 PAGE 83Trim Size: 203.2 mm X 254 mm
effects, people will consume more of a normal good when its price is lower. This conclu-
sion is a powerful one, because we know that most goods are normal goods. Some goods
are inferior goods, however. In Section 4.3 we will explore whether the law of demand
applies to them.
The Income and Substitution Effects Associated with a Gasoline Tax-Plus-Rebate Program Ever since the Arab oil embargo in 1973 and the quadrupling of oil prices that resulted from
it, there have been numerous proposals designed to encourage or force U.S. consumers to
cut back on their use of gasoline. One such proposal involves the use of a large excise tax on
gasoline (roughly 50 cents per gallon) to raise its price and thereby reduce consumption. An
excise tax is a tax on a specific good such as gasoline that allows the consumer to purchase as many units of the good at the taxed price as desired.
Realizing that a large gasoline excise tax would place a heavy burden on many families,
most proponents of the proposal recommend that the tax revenues be returned to consum-
ers in the form of unrestricted cash transfers, or tax rebates. Alternatively, the tax revenues
could be used to reduce the federal government’s outstanding debt.
Although a sizable increase in gasoline taxes has not yet been enacted into law, it poses
an interesting problem. One objection commonly raised to this plan questions whether it
would really cause gasoline consumption to fall. If the revenues from the tax are simply
distributed to the general public, why would gas consumption be curtailed? We can use
consumer choice theory to show that gasoline consumption will, in fact, be reduced by a
combination of an excise tax and a tax rebate.
The key to analyzing this policy package is realizing that the tax rebate would be a cash
transfer to each family completely unrelated to its gasoline consumption. In other words,
the proposal would not give a rebate of 50 cents for every gallon of gasoline purchased by a
family, because that policy would leave the effective price of gasoline unchanged and com-
pletely negate the effect of the tax. Instead, a family would receive a check for $500 a year,
for example, regardless of how much gasoline it purchased. On average for all families, the
rebate would equal the total tax paid, but some families would be overcompensated for
the tax while others would be undercompensated.
A Graphical Examination of a Tax-Plus-Rebate Program Now let’s examine the gasoline tax-plus-rebate plan for a representative consumer. Figure 4.4a
focuses on the effects of such a plan on the representative consumer’s budget line AZ. The excise tax by itself will increase the price from $4.00 to $4.50 per gallon and therefore rotate
the budget line inward to AZ′. This is not the end of the analysis, however, because the budget line to which the consumer will adjust must reflect both the tax and the rebate. The rebate is
shown as the outward parallel shift in AZ′ to A′Z″, similar to an increase in income, while the price of gasoline remains constant at $4.50 per gallon.
Figure 4.4b depicts what happens to the consumer’s optimal consumption point under
the tax and rebate plan. Initially, the consumer selects point E, along the original budget line AZ, with G1 gallons being purchased. With the gasoline tax and rebate the consumer selects point E′, where U1 is tangent to the new budget line A′Z″. Gasoline consumption has fallen from G1 to G2, while consumption of other goods has increased.
How far out will the tax rebate shift the after-tax budget line, AZ′? If everyone receives a rebate of the same size, and it is determined by dividing total tax revenue by the number of
consumers, then the average consumer will receive a rebate equal to the tax he or she pays.
Thus, it seems reasonable to focus the analysis on a consumer who receives a rebate equal
to the tax paid, the situation shown in Figure 4.4b.
excise tax a tax on a specific good
84 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 84Trim Size: 203.2 mm X 254 mm
A
$4.50
$4.00 1G 1G
U2
U1
E T
E ′
Other goods
A
0
(b)
G2 G1 Z″ ZZ ′ Gasoline
A ′
Other goods
A
0
(a)
Z″ ZZ ′
$4.50 $4.50 $4.00
1G 1G 1G
Gasoline
A′
$4.50
1G
Tax-Plus-Rebate Program An excise tax will reduce gasoline consumption even if the revenue is returned to taxpayers as lump-sum transfers. (a) The tax pivots the budget line to AZ′ and the tax rebate shifts it to A′Z″. (b) The combined effect reduces gasoline consumption from G1 to G2.
Figure 4.4
Income and Subst i tut ion Effects : Infer ior Goods 85
C04.INDD 10:52:49:AM 08/06/2014 PAGE 85Trim Size: 203.2 mm X 254 mm
To see that the tax and rebate are equal when the consumer’s optimal choice is point
E′, note that G2 units of gasoline will be purchased at that point. Because the AZ′ budget line shows the effect of the tax by itself, the total tax revenue is the vertical distance, E′T, between the original budget line, AZ, and the budget line, AZ′, incorporating the tax. We can see that E′T is the total tax bill by noting that if G2 gallons were purchased when the market price was $4.00, outlays on other goods would have been vertical distance, E′G2. Once the tax is levied, only TG2 in income is left (before the rebate). The vertical differ- ence, E′T, is thus the total tax. Because the rebate equals the tax, the budget line must shift up by an amount equal to E′T, and so it passes through point E′. Finally, we have already seen that point E′ represents the consumer’s optimal choice under the tax and rebate plan because the indifference curve is tangent to the final budget line at that point. By experi-
mentation, you can determine that if the rebate were any larger, it would be greater than the
amount of tax paid and, conversely, less than the tax paid if it were smaller.
The geometry of this case is slightly complicated, but the final outcome fits with
common sense. The excise tax by itself (without a rebate) has an income effect and a sub-
stitution effect. Both effects reduce gas consumption—provided, of course, that gas is a
normal good in the case of the income effect. The rebate thus offsets most of the income
effect of the tax (but not quite all of it, because the consumer does not return all the way
to the original indifference curve). Thus, the substitution effect determines the final result.
Because a higher price leads the consumer to substitute away from gasoline, the final out-
come is reduced gasoline consumption (G2 versus G1). Finally, note that this combination of tax and rebate necessarily harms the consumer.
This result is true, at least, for any consumer who receives a rebate exactly equal to the tax,
because the final outcome will be a market basket on the original budget line inferior to the
one selected in the absence of the tax and rebate. Why does anyone propose a policy that will
make the average family worse off? This is a good question. Perhaps some consequences
are not fully reflected in this analysis. For example, decreased gasoline purchases mean
decreased Middle Eastern oil imports, and possibly decreased dependence on imported oil is
beneficial in and of itself. In addition, reduced gasoline consumption means lower automo-
bile emissions and possibly improved air quality. These benefits are not incorporated into the
analysis, and if they were it is possible that consumers would be better off on balance.
4.3 Income and Substitution Effects: Inferior Goods Mechanically, the separation of income and substitution effects for a change in the price
of an inferior good is accomplished in the same way as for a normal good. The results,
however, differ in one significant respect. With a price reduction, the substitution effect still
encourages greater consumption, but the income effect works in the opposite direction. At a
lower price the consumer’s real income increases, and this, by itself, implies less consump-
tion of an inferior good. Thus, a price reduction for an inferior good involves a substitution
effect that encourages more consumption but an opposing income effect that encourages less consumption. Apparently, the total effect—the sum of the income and substitution effects—could go either way.
Figure 4.5a shows one possibility. Initially, the budget line is AZ, with the price of ham- burger at $2 per pound and H1 pounds purchased. When the price falls to $1 per pound, the budget line pivots out to AZ′, and hamburger consumption rises to H2 pounds. Once again, the hypothetical budget line HH′ that keeps the consumer on U1, the original indifference curve, is drawn in. The substitution effect is the movement from point W to point J on U1, implying an increase in consumption from H1 to HJ. Now see what happens to hamburger consumption when we move out from budget line HH′ to AZ′, a movement reflecting the income effect of the lower price of hamburger. Because hamburger is an inferior good for
86 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 86Trim Size: 203.2 mm X 254 mm
this consumer, the income effect reduces hamburger consumption, from HJ to H2. Overall, however, the total effect of the price reduction is increased consumption, because the sub-
stitution effect (greater consumption) is larger than the income effect (lower consumption).
In this situation the consumer’s demand curve for hamburger slopes downward.
For an inferior good there is another possibility, illustrated in Figure 4.5b. Good X is also an inferior good for some consumers, and a reduction in its price pivots the budget
line from AZ to AZ′. Here, however, the total effect of the price decrease is a reduction in the consumption of X, from X1 to X2. When the income and substitution effects are shown
Income and Substitution Effects for an Inferior Good (a) Hamburger is an inferior good with a normally shaped, downward- sloping demand curve, because the substitution effect is larger than the income effect. (b) Good X is an inferior good with an upward-sloping demand curve, because the income effect is larger than the substitution effect. Good X is called a Giffen good.
0
U1
J
W
W ′
U2
(a) Total
S I
HamburgerZ Z ′H ′H1 H2 HJ
A
H
0
U1
J
W
W ′
U2
Other goods
(b)
Total
S I
Other goods
A
H
Good XZ Z ′H ′X1X2 XJ
Figure 4.5
Income and Subst i tut ion Effects : Infer ior Goods 87
C04.INDD 10:52:49:AM 08/06/2014 PAGE 87Trim Size: 203.2 mm X 254 mm
separately, we see how this outcome occurs. The substitution effect (point W to point J, or increased consumption of X) still shows greater consumption at a lower price. However, the income effect for this inferior good not only works in the opposing direction (less consump-
tion, from XJ to X2), but also overwhelms the substitution effect. Because the income effect more than offsets the substitution effect, consumption falls. This consumer’s demand curve
for good X, at least for the prices shown in the diagram, will slope upward. Thus, for inferior goods, there are two possibilities. If the substitution effect is larger than
the income effect when the price of the good changes, then the demand curve will have its
usual negative slope. If the income effect is larger than the substitution effect for an inferior
good, then the demand curve will have a positive slope. This second case represents a theo-
retically possible (but rarely observed) exception to the law of demand. It can happen only
with an inferior good and, moreover, only for a subset of inferior goods in which income
effects are larger than substitution effects. We refer to a good in this class as a Giffen good, after the nineteenth-century English economist Robert Giffen, who believed that, during the
years of famine, potatoes in Ireland had an upward-sloping demand curve. Giffen observed
that as the blight diminished the supply of potatoes in Ireland and drove up their price, the
quantity demanded of potatoes appeared to increase. (The evidence in support of Giffen’s
observation is a matter of debate among economists.)
A Hypothetical Example of a Giffen Good Finding an intuitively plausible example in which the demand curve slopes upward is dif-
ficult, but consider the following hypothetical situation. The Smith family lives in Alaska
and traditionally spends the month of January in Arizona. One year the price of home heat-
ing oil increases sharply. The Smiths cut back on their use of heating oil during the other
winter months, but, nonetheless, their total heating costs rise to a point where they can no
longer afford a vacation in Arizona. Because they stay at home in January, their use of heat-
ing oil for that month increases dramatically over the amount they would have used had
they been in Arizona. On balance, annual heating oil purchases will rise if the increased
use in January is greater than the reduction achieved during the remaining winter months.
Consequently, an increase in the price of heating oil can conceivably lead to greater use of
heating oil by the Smiths. (Conversely, a decrease in the price of heating oil can result in
lower consumption of heating oil by the Smiths.)
This contrived scenario illustrates the type of situation shown in Figure 4.5b. Heating oil
is an inferior good for the Smiths; a reduction in income will lead them to spend more time
at home, which causes an increase in the use of heating oil. A price increase has an income
effect that induces them to forgo their January vacation. If the expected consumption in
January exceeds the reduced consumption of heating oil during the other winter months, a
net increase in consumption of heating oil at a higher price results.
The Giffen Good Case: How Likely? We might conceive of cases where the income effect for an inferior good exceeds the substi-
tution effect, producing an upward-sloping demand curve. However, economists believe that
most, if not all, real-world inferior goods have downward-sloping demand curves, as shown
in Figure 4.5a. This belief stems from both theoretical considerations and empirical evidence.
At a theoretical level the question is whether the income effect or the substitution effect
of a price change for an inferior good will be larger. If the substitution effect is larger, then
the demand curve will slope downward, even for an inferior good. There are good reasons
for believing that the substitution effect is larger. Consider first the income effect. Its size
relates closely to the fraction of the consumer’s budget devoted to the good. If the price of
some good falls by 10 percent, the price reduction will benefit a consumer much more (have
a larger income effect) if 25 percent of the consumer’s income is spent on the good than if
only 1 percent is spent on it. For example, a 10 percent reduction in the price of housing
Giffen good the result of an income effect being larger than the substitution effect for an inferior good, so that the demand curve will have a positive slope
88 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 88Trim Size: 203.2 mm X 254 mm
will probably influence housing consumption greatly by its income effect, but a 10 percent
reduction in the price of CDs will have a much smaller, almost imperceptible income effect.
Income effects from a change in price are quite small for most goods because they seldom
account for as much as 10 percent of a consumer’s budget. This observation is especially true
of inferior goods, which are likely to be narrowly defined goods.
In contrast, there is reason to believe that substitution effects for inferior goods will be rel-
atively large. Inferior goods usually belong to a general category that contains similar goods
of differing qualities. Take hamburger: a reduction in its price can be expected to result in a
rearrangement of a consumer’s purchases away from chicken, pork, pot roast, and so on, in
favor of hamburger, thus resulting in a large substitution effect. Consequently, price changes
for inferior goods should involve relatively large substitution effects but small income effects.
Therefore, the demand curve will slope downward, and the case shown in Figure 4.5a will be
typical. The Giffen good remains an intriguing but remote theoretical possibility.
4.4 From Individual to Market Demand We have seen how to derive an individual consumer’s demand curve and why the concepts
of income and substitution effects imply that it will typically slope downward. But most
practical applications of economic theory require the use of the market demand curve. We
begin with a discussion of individual demand because the individual demand curves of all
the consumers in the market added together constitute the market demand curve. We will
show that, if the typical consumer’s demand curve has a negative slope, then the market
demand curve must also have a negative slope.
Figure 4.6 illustrates how individual demand curves are aggregated to obtain the market
demand curve. Assume that there are only three consumers who purchase an MBA educa-
tion, although the process will obviously apply to the more important case where there are
a great many consumers. The individual demand curves are dA, dB, and dC. To derive the market demand curve, we sum the quantities each consumer will buy at alternative prices.
For example, at P2 consumer B will buy 10 credit hours, consumer C will buy 15, and consumer A will buy none. (Note that when the price is P2, consumer A will be at a corner optimum.) The combined purchases of all consumers total 25 credit hours when the price is
P2, and this combination identifies one point on the market demand curve D.
APPLICATION 4.3
The transparent chemical compound indium tin oxide (ITO) is found almost exclusively in one area of mainland China, the Dachang District.5 If one looked only at the demand for ITO in Dachang, China, it would be hard to explain why the compound is so expensive. However, once one aggregates demand curves for ITO across different geographies around the globe, the reason for the high equi- librium price becomes much more readily apparent. This is because over 90 percent of the touchscreens globally use
Aggregating Demand Curves for ITO
ITO as the conductor that detects fingertip motion. And the size of the touchscreen market has been growing rapidly— from sales of 200 million units in 2007 to 800 million units in 2012. The entire touchscreen industry is projected to double in size to $32 billion by 2018.
On account of the expensiveness of ITO, a global race is on to find lower-cost alternatives. In June 2013, for exam- ple, Kodak announced plans to invest $10 million in a man- ufacturing plant in Rochester, New York to produce a metal mesh conductor relying on nanoscale lines of copper and silver, instead of ITO, to detect fingertip motion on smart- phones, tablets, and other devices.
5This application is based on “Kodak Focusing on Touchscreen Markets,” Rochester Democrat and Chronicle, July 12, 2013.
Consumer Surplus 89
C04.INDD 10:52:49:AM 08/06/2014 PAGE 89Trim Size: 203.2 mm X 254 mm
Other points on the market demand curve are derived in the same way. If the price is
P1, A will buy 3 credit hours, B will buy 13, and C will buy 19, so total quantity demanded at a price of P1 is 35 credit hours. The process of adding up the individual demand curves to obtain the market demand curve is called horizontal summation, because the quantities (measured on the horizontal axis) bought at each price are added. Note that when the indi-
vidual demand curves slope downward, the market demand curve also slopes downward. If
all consumers buy more at a lower price, then total purchases will rise when the price falls.
A market demand curve, however, can slope downward even if some consumers have
upward-sloping individual demand curves. In a market with thousands of consumers, if a
few happened to have upward-sloping demand curves, then their contribution to the market
demand curve would be more than offset by the normal behavior of the other consumers.
So we have yet another reason not to be overly concerned about the Giffen good case. It is
possible to imagine that the Smith family in Alaska will buy more heating oil at a higher
price, but it is difficult to believe that their behavior is typical.
4.5 Consumer Surplus Consumers purchase goods because they are better off (that is, on a higher indifference
curve) after the purchase than they were before; otherwise, the purchase would not take
place. The term consumer surplus refers to the net benefit, or gain, secured by an individ- ual from consuming one market basket instead of another. For example, suppose that around
exam time you purchase six cups of espresso coffee per day at $3 per cup from the campus
coffeeshop. You have chosen to spend $18 per day on espresso and allocate the rest of your
budget to other items. Alternatively, you could choose to not buy espresso, cut your pulse
rate in half, and spend the $18 on something else; this is another possible allocation of your
budget. Because you clearly feel you are better off by consuming espresso, we say that you
consumer surplus a measure of the net gain to consumers from purchasing a good arising from its cost being below the maximum that consumers are willing to pay
0 3 10 13 15 19 25 35 Credit hours of MBA education
Price
P2
P1
dA
dB dC
D
Summing Individual Demands to Obtain Market Demand The market demand curve D is derived from the individual consumers’ demand curves by horizontally summing the individual demand curves. At each price we sum the quantities each consumer will buy to obtain the total quantity demanded at that price.
Figure 4.6
90 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 90Trim Size: 203.2 mm X 254 mm
secure a consumer surplus from being able to purchase six espressos per day at $3 per cup.
We now wish to see how this surplus, or net benefit, can be measured in dollar terms.
To obtain a measure of consumer surplus associated with espresso purchases, first ask
yourself this question: What is the maximum amount you would be willing to pay for six
cups per day from the campus coffeeshop during exam time? Your answer will be the total benefit (or total value) of the six cups per day. Your total cost is the $18 per day that you pay to the campus coffeeshop for the espressos. The difference between these two sums is
the net benefit, or consumer surplus, you receive.
The demand curve provides another, and more direct, way to measure consumer surplus.
To see how the demand curve relates to consumer surplus, consider how, in our hypothetical
example, your demand curve for espresso from the campus coffeeshop is actually generated.
To simplify the analysis, let’s initially assume that espressos are sold only in uniform unit-
cups, and start with a price so high that you wouldn’t buy any. We gradually lower the
price until you purchase one cup per day—say, when the price reaches $8. Thus, the incre-
mental value, the marginal benefit, to you of the first cup is $8; this price is the maximum amount you would pay for the first cup. Because you are willing to pay $8 for the first cup,
the $8 reflects the value you place on the first cup; that is, it is a measurement, in dollar
terms, of the benefit you derive from the espresso. Lowering the price further, suppose that
we find that at a price of $7 you will purchase a second cup; that is, the marginal benefit of
the second cup is $7. Consequently, the price at which a given unit will be purchased mea-
sures the marginal benefit of that unit to you.
Continuing this process, we can generate your entire daily demand curve for espresso
at the campus coffeeshop. In our hypothetical example (where fractions of a cup cannot be
purchased), your demand curve is the step-like curve d shown in Figure 4.7. The area of each of the tall rectangles measures the marginal benefit to you, the consumer, of a specific
cup. For instance, the tallest rectangle has an area of $8 ($8 per cup multiplied by one cup,
or $8). The marginal benefit of the first cup is $8; of the second, $7; of the third, $6; and
so on. The total benefit of consuming a given quantity is the sum of the marginal benefits. If two cups are consumed, the total benefit is $15, because you would have been willing
to pay as much as $8 for the first cup and $7 for the second. By determining the maximum
amount you will pay, we can calculate the total benefit of the espresso to you, which is
equal to the area under the demand curve up to the quantity purchased.
Now suppose, more realistically, that you can purchase each espresso cup at a price of $3.
As a rational consumer, you purchase cups up to the point where the marginal benefit of a
total benefit the total value a consumer derives from a particular amount of a good and thus the maximum amount the consumer would be willing to pay for that amount of the good
marginal benefit the incremental value a consumer derives from consuming an additional unit of a good and thus the maximum amount the consumer would pay for that additional unit
Price per cup
$8
$7
$6
$5
$4
$3 = P
0 1 2 3 4 5 Q = 6
7 8
d
E
Cups of espresso per day
Consumer Surplus The total benefit from purchasing six units at a price of $3 per unit is the sum of the six shaded rectangles, or $33. Since the six units involve a total cost of $18, the consumer surplus is $15 and is shown by the striped area.
Figure 4.7
Consumer Surplus 91
C04.INDD 10:52:49:AM 08/06/2014 PAGE 91Trim Size: 203.2 mm X 254 mm
cup is just equal to the price. Now compare the total benefit from purchasing six cups at $3
per cup with the total cost:
Total benefit Sum of marginal benefits
$8 7 6 5 4 3
= + + + + + == $ $ $ $ $ $333.
Total cost Sum of cost of each unit
Net benef
= × =$ $ .3 6 18 iit (consumer surplus) Total benefit Total cost= −
= $ .15
The total daily benefit of six cups is $33, but you have paid only $18 for the espresso, so
a consumer surplus, or a net gain of $15, accrues. Put simply, the consumer surplus is the
difference between what you would have been willing to pay for the espresso and what you
actually did pay.
Geometrically, we add the areas of the six rectangles reflecting the marginal benefits;
then we subtract the total cost (price times quantity) represented by the area of the large
rectangle, PEQ0, or $3 times six cups. The area that remains—the striped area in Figure 4.7 between the price line and the demand curve—is the geometric representation of consumer
surplus. An alternative way to see that this area measures consumer surplus is to imagine
purchasing the units of the good sequentially. The first cup is worth $8, but it costs only $3,
so there is a net gain of $5 on that unit; this gain is the first striped rectangle above the price
line. The second cup is also purchased for $3, but because you would have been willing to
pay as much as $7 for the second cup, there is a net gain of $4 on that cup. (This gain is the
second striped rectangle above the price line.) Adding up the excess of benefit over cost on
each unit purchased, we have $5 + $4 + $3 + $2 + $1 + $0 = $15, which is shown by the area between the price line and the demand curve. Note that there is no net gain on the last
unit purchased. Purchases are expanded up to the point where the marginal benefit of the
last unit is exactly equal to the price. Previous units purchased are worth more than their
price—which, of course, is why you receive a net gain.
Figure 4.8 shows the same situation, but now we assume that espresso is divisible into
small units so that a smooth demand curve D can be drawn. We also allow for more than just a single consumer of espresso (thus the uppercase D is used to express demand). Indeed, at a price of P, we assume that, across all consumers, the total amount of espresso purchased equals Q. Consumer surplus is the striped triangular area TEP between the
Price
D
Quantity0 Q
T
$3 = P E
Consumer surplus
Total cost
Consumer Surplus With a smooth demand curve, consumer surplus equals area TEP.
Figure 4.8
92 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 92Trim Size: 203.2 mm X 254 mm
demand curve and the price line. It is analogous to the areas of the rectangles above
the price line in Figure 4.7, but by letting the width of the rectangles become smaller
and smaller (fractional units may be purchased), we now have a smooth line rather than
discrete steps. In Figure 4.8, the total benefit from consuming Q units is TEQ0, the sum of the heights of the demand curve from 0 to Q. (Instead of a rectangular area, the maximum amount that consumers are willing to pay for a particular unit is represented
by the height of the demand curve at that unit when the units employed to measure purchases become very small.) The total cost is PEQ0, and the difference, TEP, is the consumer surplus garnered by all consumers, as a group, of the espresso sold by the
campus coffeeshop.
The Uses of Consumer Surplus As you might imagine, consumer surplus has many uses. To managers of business firms,
consumer surplus indicates the benefits obtained by buyers over and above the prices the
buyers are charged. As we will see in a later chapter, many product pricing strategies reflect
an effort by firms to capture more of the consumer surplus generated by their products and
to convert such surplus into profit.
The concept of consumer surplus can also be used to identify the net benefit of a change
in the price of a commodity or in its level of consumption. For example, Figure 4.9 shows
the U.S. demand curve for sugar. Suppose that at a price of 25 cents per pound, U.S. buyers
purchase Q pounds per year. The consumer surplus is given by area TAP. Now, due to trade liberalization and the possibility of imports from overseas, suppose that the price falls to 15
cents per pound. How much better off are U.S. buyers because of the decrease in the price
of sugar? There are two equivalent ways to arrive at the answer. One is to note that the
consumer surplus will be TEP′ at the lower price, which is greater than the initial consumer surplus, TAP, by the area PAEP′. Thus, the area PAEP′ is the increase in consumer surplus, and it identifies the net benefit to U.S. buyers from the lower price.
A second way to reach the same answer is to imagine U.S. buyers adjusting to the
lower price in two steps. First, total consumption is tentatively held fixed at Q. When the price falls, the same Q units can be purchased for 10 cents less per pound than before; this amount is equal to area PACP′, and it is part of the net benefit from the lower price. Second, the lower price also makes it advantageous for buyers to expand their purchases
from Q to Q′. A second net benefit is associated with this expansion because the marginal benefit of each of these pounds is greater than the per-pound price. For instance, the first
pound of sugar beyond Q pounds has a marginal benefit of just slightly under 25 cents, but it can be purchased for 15 cents thanks to trade liberalization—a net benefit of about 10
cents for that unit. The net benefit to buyers from expanding their sugar consumption from
The Increase in Consumer Surplus with a Lower Price At a price of 25 cents per pound, consumer surplus is TAP. At a price of 15 cents per pound, consumer surplus is TEP′. The increase in consumer surplus from the price reduction is thus the shaded area PAEP′: this area is a measure of the benefit to consumers of a reduction in the price from 25 to 15 cents.
Price
D
Sugar consumption
0
T
25¢ = P
15¢ = P ′
A
C E
Q Q ′
Figure 4.9
Consumer Surplus 93
C04.INDD 10:52:49:AM 08/06/2014 PAGE 93Trim Size: 203.2 mm X 254 mm
Q to Q′ pounds is the area AEC. Combining the two areas of net benefit once again yields PAEP′ as the net benefit from the lower price.
In later chapters we will see other examples of how the concept of consumer surplus can
help us to evaluate the benefits and costs of various economic phenomena.
Consumer Surplus and Indifference Curves Consumer surplus can be represented in our indifference curve and budget line diagrams.
Let’s return to our original example in which a consumer purchases six cups of espresso at
a price of $3 per cup. Figure 4.10 shows the optimal consumption at point W, the familiar tangency between an indifference curve and the budget line. Note that the consumer is on
a higher indifference curve, U2, when purchasing six cups of espresso than when buying no espresso at all. If no espresso is bought, the optimal point would be A on U1. The net benefit, or consumer surplus, from purchasing six units is clearly shown by the consumer
reaching a higher indifference curve at point W than at point A. Thus, the consumer receives a net benefit from purchasing six units instead of none. Now
let’s try to measure the net benefit in dollar terms. Starting at point A, where no espresso is purchased, let’s ask this question: What is the maximum amount of money the consumer
would give up for six cups of espresso? Paying the maximum amount means the consumer will remain on U1, the original indifference curve, and move down to point R, where six units are consumed. Distance AA2 identifies the maximum amount the consumer would be willing to pay. Note that this amount equals the sum of the amounts that would be paid for
APPLICATION 4.4
Until the advent of cable, television was not sold directly to viewers. The price of viewing broadcast programming was zero (apart from the opportunity cost of the viewer’s time, the purchase price of a television set, and the electric- ity necessary to power the set) for a household with a tele- vision and clear reception of the signal. Most of the costs of operating over-the-air networks and stations were, and still are, covered by sales of broadcast time to advertisers.
In the heyday of free television, viewing options were limited, but the consumer surplus accruing to viewers was not. In 1968, for example, the average U.S. household had access to three network stations and one independent station. The estimated annual consumer surplus garnered by viewers was $32 billion ($209 billion in 2013 dollars) due to a price of zero for broadcast television.6 The estimated consumer surplus vastly exceeded the $3.5 billion in adver- tising revenues earned by all television stations in 1968.
A prominent economic study published in 1973 indi- cated that an expansion in viewing options, in terms of the consumer surplus generated through such an expansion, would be highly valued. According to the study, a fourth network would add $4.2 billion in consumer surplus as of 1968 ($27.1 billion in 2013 dollars). Expansion, however, was precluded by regulations as well as by the fact that it was
The Consumer Surplus Associated with Free TV
not technologically feasible to charge viewers for the addi- tional programming.
The study’s results suggest why cable and satellite television have grown so rapidly over the past 45 years. Namely, by figuring out a way to exclude nonpayers and charge subscribers for their service, cable and satellite television operators have been able to capture some of the television consumer surplus from either existing or newly developed programming and convert it into profits for their companies. Television owners have found sub- scribing to cable/satellite attractive because it allows them to expand their viewing options (many cable/satellite systems now have more than 500 channels of program- ming), enhanced options that generate consumer surplus. Currently, 48 percent of U.S. television households sub- scribe to cable, and the average subscribing household spends approximately $121 per month on cable. In com- parison, the amount of advertising revenues earned by broadcast stations averages roughly $25 per month per household.
6Roger G. Noll, Merton J. Peck, and John J. McGowan, Economic Aspects of Television Regulation (Washington, D.C.: Brookings Insti- tution, 1973).
94 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 94Trim Size: 203.2 mm X 254 mm
each successive unit; that is, in moving from point A to point S, the consumer would pay $8 for the first unit, $7 for the second unit (S to T), and so on. The sum of these amounts equals AA2, or $33. The distance AA2 measures the total benefit from consuming six units, and it corresponds to the area under the demand curve in a demand curve diagram.
Total benefit is AA2. However, the consumer actually purchases six units at a cost of only AA1, or $18. The total benefit, AA2, exceeds total cost, AA1, by the distance A1A2 (also equal to the distance WR). The difference between total benefit and total cost—in this case $15—is the consumer surplus from purchasing six cups of espresso at a price of $3 per cup.
The consumer surplus can be shown either by the area between the demand curve and the
price line or as a vertical distance between indifference curves. Both this diagram and
Figure 4.7 therefore show the same thing but from different perspectives.
Note one qualification: under certain conditions, consumer surplus, as measured by
the area under a demand curve, is exactly equal to the measure obtained in Figure 4.10.
The certain conditions, however, require a special assumption: the income effect of
price changes on consumption of the good in question must be zero. This assumption is
reflected by the indifference curves being vertically parallel, having the same slope as
you move up a vertical line. In Figure 4.10, for example, the slope of U1 at point R is the same as the slope of U2 at point W. When this assumption does not hold, the area under the demand curve is only an approximation of the true measure of consumer surplus. The
approximation is still generally close enough for most applications.7
7Robert D. Willig, “Consumer Surplus without Apology,” American Economic Review, 65 (1976), pp. 589–597.
Figure 4.10 Consumer Surplus and Indiffer- ence Curves The consumer surplus associated with being able to purchase espresso at $3 per cup is shown by the consumer being on U2 rather than U1. In dollars, this net gain, or surplus, is the distance WR.
Espresso6 Z
A1
A2 U2 U1
Other goods
$33 $18
A
$8
$7 $6
$5 $4
$3
W S
T
R
$3 1E
0
8Kevin M. Murphy and Robert H. Topel, “The Value of Life: The Economic Benefits of Increased Longevity,” Capital Ideas, The University of Chicago, Graduate School of Business, February 2006, pp. 4–7.
APPLICATION 4.5
While the costs of health care have been growing mark- edly in recent decades, the benefits associated with those rising health care expenditures have also been substantial. These benefits come, among other ways, in the form of increased life expectancy. Economists Kevin Murphy and Robert Topel of the University of Chicago estimate that between 1970 and 2000, aggregate medical expenditures grew by $34 trillion, while the total benefits associated
The Benefits of Health Improvements
with those expenditures, as measured by their impact on Americans’ greater longevity (and willingness to pay for that increased life expectancy), amounted to $95 trillion.8
Price Elast ic i ty and the Pr ice–Consumption Curve 95
C04.INDD 10:52:49:AM 08/06/2014 PAGE 95Trim Size: 203.2 mm X 254 mm
4.6 Price Elasticity and the Price–Consumption Curve Price elasticity can be computed for any demand curve, whether it is the market demand
curve or an individual consumer’s curve. Admittedly, the price elasticity of market demand
is generally of greatest interest, but that elasticity depends on the underlying elasticities of
the demand curves of various consumers. In terms of our treatment of individual demand,
we can now show that the slope of the individual’s price–consumption curve provides
important information about the elasticity of demand.
Figure 4.11 shows four hypothetical price–consumption curves. In Figure 4.11a, the curve
slopes downward. This means that the price elasticity exceeds unity; that is, the consumer’s
demand curve will be elastic if plotted in a price–quantity diagram. A downward-sloping
price–consumption curve shows that the consumer’s total expenditure on college education
rises when the price of such education falls, which, by definition, is an elastic demand. (When
demand is elastic, the percentage change in quantity associated with a price change is larger,
in absolute value terms, than the percentage change in price. Total expenditure, or price times
quantity, thus moves in the same direction as quantity and in the opposite direction from price
anytime the price is altered.) Recall that the distance AA1 shows total expenditure on college education when the price of college education is given by the slope of the budget line AZ. (This is because A0 represents the consumer’s income and A10 indicates outlays on all other goods. Thus, AA1 must represent the amount that is left to spend on college education.) When price falls, the budget line rotates to AZ′, and at the new optimal point total expenditure on college education is now AA2—an increase from the original level.
Figures 4.11b and 4.11c show the cases of unit elastic and inelastic demand, respec-
tively. In Figure 4.11b, the price–consumption curve is a horizontal line, showing that total
expenditure on college education remains unchanged at AA1 when the price of college edu- cation is varied. This situation is, by definition, one of unit elastic demand. Figure 4.11c has
an upward-sloping price–consumption curve. In this case a reduction in price reduces total
expenditure on college education, by definition an inelastic demand. Expenditure is initially
AA1 but falls to AA2 when the price is reduced. Therefore, if the price–consumption curve slopes downward, then the consumer’s demand is elastic. If it is horizontal, demand is unit
elastic. If it slopes upward, it is inelastic.
Finally, Figure 4.11d shows a U-shaped price–consumption curve. The elasticity of demand varies along this curve. It is elastic along the negatively-sloped AJ portion of the curve; becomes unit elastic at point J, where the slope of the curve is zero; and is inelastic along the upward-sloping portion of the curve to the right of point J. This type of price– consumption curve is probably typical. It begins at point A because at a high enough price,
no college education would be purchased. Thus, it must be negatively sloped at relatively
The net benefits or consumer surplus generated by the increased medical expenditures thus totaled $61 trillion over the 1970–2000 time period.
The gross benefits associated with increased medical expenditures and their effect on longevity, added roughly $3.2 trillion per year to national wealth between 1970 and 2000. To put this in perspective, this amount is equal to half the size of annual gross domestic product over the period. Almost 50 percent of these gains can be attributed to progress against heart disease. The decrease in mortality from heart disease over the 1970 to 2000 period increased
the aggregate value of life for Americans by $1.5 trillion annually.
According to Murphy and Topel, further health care inno- vations promise additional significant benefits. For example, a 1 percent reduction in the mortality rate from cancer or heart disease would be worth $500 billion to current and future Americans while a cure could generate benefits on the order of $50 trillion. Improvements in the quality of life, moreover, may represent an even more valuable dimension of the bene- fits associated with health care advances—more valuable than just the benefits stemming from increased longevity.
96 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 96Trim Size: 203.2 mm X 254 mm
high prices (implying an elastic demand). On the other hand, there will generally be a finite
quantity the consumer would consume even at a zero price, so the price–consumption curve
must slope upward at relatively low prices (implying an inelastic demand). Therefore, a
consumer’s demand curve tends to be elastic at high prices and inelastic at low prices. This
knowledge does not help us determine elasticity at a specific price because we don’t know
whether that specific price is “high” or “low” in this sense.
4.7 Network Effects Until now we have assumed that a particular consumer’s demand for a good is unrelated to
other consumers’ demands for the good. However, this need not be the case. To the extent
that an individual consumer’s demand for a good is influenced by other individuals’
Price–Consumption Curves and the Elasticity of Demand The slope of a consumer’s price–consumption curve tells us whether demand is elastic, inelastic or unit elastic. (a) When the price–consumption curve is negatively sloped, demand is elastic. (b) When it has a slope of zero, demand is unit elastic. (c) When it is positively sloped, demand is inelastic. (d) When it is U-shaped, demand is elastic at high prices and inelastic at low prices.
College education
P–C curve
(a)
Z Z ′
A
A1
A2
U2 U1
0 College education
P–C curve
(b)
Z Z ′
A
A1
U2 U1
0
College education
P–C curve
(c)
Z Z ′
A
A2
A1 U2
U1
0 College education
P–C curve
(d)
U2U1
J
0 Z Z ′
Other goods
Other goods
Other goods
A
Other goods
Figure 4.11
Network Effects 97
C04.INDD 10:52:49:AM 08/06/2014 PAGE 97Trim Size: 203.2 mm X 254 mm
purchases, there is a network effect. Network effects can be positive or negative. The posi- tive case, or bandwagon effect, exists whenever the quantity of a good demanded by a particular consumer is greater the larger the number of other consumers purchasing the
same good. The negative case, or snob effect, occurs when the quantity of a good demanded by a particular consumer is smaller the larger the number of other consumers purchasing
the same good.
The Bandwagon Effect Capitalizing on bandwagon effects is critical to the marketing of some goods. For example,
clothing, toy, and food manufacturers realize that their ability to sell certain products to a
particular consumer will be enhanced the greater the number of other consumers purchas-
ing the same products. Tommy Hilfiger jeans, Barbie dolls, Pokémon cards, Nike running
shoes, and Evian water are all products characterized by such positive network effects. In
some cases, the positive network effects stem from consumers’ desires to be in fashion
and the utility derived from owning popular products. In other cases, a bandwagon effect
derives from the fact that the inherent value of a good to a consumer is enhanced by wide-
spread usage of the good among other consumers. Take the case of Facebook. The value of
Facebook to an individual consumer is increased when other consumers also use Facebook
if a larger customer base leads to more services and connections. The extent of a business
school’s alumni network similarly can increase the value of attending the school to a partic-
ular applicant. This is because a greater number of alums translates into more connections
when the applicant searches for a job after graduating.
Figure 4.12 depicts the case of a bandwagon effect. If consumers believe that only 1,000
people own a good, the demand curve is d1,000. If consumers believe that more people own the good and, consequently, the good is more desirable—either because it is in greater fash-
ion or because its inherent value is increased—the demand curve is located farther to the
right at any price: d2,000 if 2,000 people are believed to own the good; d3,000 if 3,000 people are believed to own the good; and so on.
With a bandwagon effect, the market demand curve, D, is more price elastic. To see why, suppose that consumers initially are willing to purchase 1,000 units if the price is $50. Now
consider a decrease in price from $50 to $40. Without any bandwagon effect, the quantity
network effects the extent to which an individual consumer’s demand for a good is influenced by other individuals’ purchases
bandwagon effect a positive network effect
snob effect a negative network effect
Bandwagon Effect A bandwagon effect leads to greater consumer purchases of a good the more other consumers are believed to desire the same good. The market demand curve, D, is more elastic because the bandwagon effect increases the response in quantity demanded to any change in price.
Quantity
C
B
A
E
D
$40
$50
0
Price
d1,000 d2,000 d3,000
1,0001,250 2,000 3,000
Bandwagon effect
Pure price effect
Figure 4.12
98 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 98Trim Size: 203.2 mm X 254 mm
demanded would increase to 1,250 along curve d1,000. The bandwagon effect, however, results from more people purchasing the good at the lower price, and this, in turn, increases
the willingness of consumers to purchase the good for its greater fashion or inherent value.
In the case of Figure 4.12, 2,000 units are purchased at a price of $40, and the band-
wagon effect accounts for the increase in quantity demanded from 1,250 to 2,000 in
response to the decrease in price from $50 to $40. The market demand curve is derived by
connecting the points on the curves d1,000, d2,000, and d3,000 corresponding to the quantities 1,000, 2,000, and 3,000. A, B, and C are the only points consistent with the expected quan- tities associated with each of the curves. By contrast, point E, representing a quantity of 1,250 on the curve d1,000, is inconsistent with consumers’ beliefs that overall purchases total 1,000 units and thus cannot lie on the market demand curve.
The market demand curve is more elastic than the individual curves d1,000, d2,000, and d3,000 because the bandwagon effect increases the response in quantity demanded to any change in price. In other words, the total response in quantity demanded to a change in
price is the sum of the pure price effect and the bandwagon effect. Thus, it exceeds the pure
price effect in magnitude.
The Snob Effect The snob effect is the opposite of the bandwagon effect. It occurs when a consumer is less
willing to purchase a good the more widespread its usage. A vintage Jaguar car, an original
copy of the Declaration of Independence, a custom-made Versace evening gown, a Picasso
painting, and a hand-crafted Piaget watch are all possible examples of goods associated
with snob effects. A consumer’s valuation of such goods may be greater the more exclusive
are the goods, on account of the prestige and admiration derived by the consumer from the
goods being selectively owned.
Figure 4.13 depicts a snob effect in the case of vintage Jaguar cars. The relevant
demand curves are d10, d20, and d30 if consumers believe that only 10, 20, and 30 people,
Snob Effect A snob effect leads to smaller consumer purchases of a good the more other consumers are believed to desire the same good. The market demand curve, D, is less elastic because the snob effect decreases the response to quantity demanded for any change in price.
Quantity
G
H
D
F
I
$50,000
$150,000
0
Price
d30
d10
d20
205 10 30
Snob effect
Pure price effect
Figure 4.13
Network Effects 99
C04.INDD 10:52:49:AM 08/06/2014 PAGE 99Trim Size: 203.2 mm X 254 mm
respectively, own a particular model. Note that the demand curve is farther to the right at
any given price the more exclusive the ownership of the vintage Jaguar model is believed
to be—d20 is to the right of d30, and d10 is to the right of d20. This reflects a snob effect: the quantity of a good demanded by a particular individual falls the more widely owned the
good is considered to be by other consumers.
The market demand curve is more inelastic in the case of goods characterized by a snob
effect. To see why, note that the market demand curve connects the points on curves d10, d20, and d30 associated with the quantities 10, 20, and 30. F, G, and H are the only points consistent with the expected purchases associated with curves d10, d20, and d30. Point I can- not lie on the market demand curve because a quantity of 5 is inconsistent with the expecta-
tion of 30 purchases associated with d30. An alternative, and perhaps clearer, way of understanding why the market demand
curve is more inelastic in the case of goods characterized by a snob effect is to exam-
ine the effect of a price increase, such as from $50,000 to $150,000 in Figure 4.13.
The pure price effect of this increase decreases consumption from 30 to five vintage
Jaguar cars along curve d30. However, the snob effect associated with the enhanced exclusivity resulting from the price increase works to counteract the pure price effect
and increases consumption from five to 20 units. On net, therefore, the price increase
leads to a decrease in the quantity demand from 30 to only 20 units and is less than the
decrease associated with the pure price effect. Because the snob effect runs counter to
the pure price effect, the market demand curve is less price elastic—the cumulative
impact of the pure price and snob effects on quantity demanded is less than the pure
price effect.
APPLICATION 4.6
Communications technologies are prime examples of products characterized by positive network effects.9 As the established user base of telephones‚ fax machines‚ the Internet‚ social networking sites, and e-mail grows‚ increasingly more individuals find adoption worthwhile. Consequently‚ such products tend to be characterized by relatively long developmental periods followed by rapid diffusion. Take the case of fax machines‚ a product that AT&T first introduced in 1925. The technology‚ however‚ was little used until the mid-1980s‚ when demand for and supply of fax machines exploded. While ownership of fax machines was negligible prior to 1982‚ over half of Ameri- can businesses had at least one fax machine by 1987.
Internet usage has followed a similar pattern. While the first e-mail message was sent in 1969‚ Internet traffic did
Network Effects and the Diffusion of Communications Technologies and Computer Hardware and Software
not begin to grow substantially until the late 1980s. Once it began to grow‚ however‚ it doubled annually in virtually every year after 1989.
Positive network effects are not limited to communi- cations technologies. They are also at the heart of explain- ing the diffusion of computer software and hardware‚ where popular systems enjoy a significant competitive advantage over less popular systems. Personal computers provide a telling example. A study of 110‚000 American households in 1997 suggests that the rate of adoption of personal computers may have been almost doubled on account of network effects. Moreover‚ the observed net- work effects are strongly related to usage of the Internet and e-mail.
Positive network effects also likely have played an important role in the rapid spread of popular computer software programs such as Microsoft’s Windows operat- ing system and Office (a combination of word processing‚ spreadsheet‚ database‚ and presentation programs). Both products quickly acquired shares of over 90 percent of their relevant markets.
9This application is based on Carl Shapiro and Hal R. Varian‚ Infor- mation Rules (Boston: Harvard Business School Press‚ 1999); and Austan Goolsbee and Peter J. Klenow‚ “Evidence on Learning and Network Externalities in the Diffusion of Home Computers‚” Journal of Law and Economics‚ 45, No. 2, Pt. 1 (October 2002), pp. 317–343.
100 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 100Trim Size: 203.2 mm X 254 mm
4.8 The Basics of Demand Estimation Although indifference curves and budget lines together provide an appealing theoretical
model of consumer choice, our ability to test the validity of the model and apply it to the
real world rests critically on the extent to which we are able to empirically estimate indi-
vidual consumers’ or market demand curves. Three methods are generally used to estimate
demand: experimentation, surveys, and regression analysis. We briefly outline each of
these methods as well as their accompanying pitfalls.
Experimentation McDonald’s will often run a controlled experiment to test how the demand for one of its
fast-food items responds to a change in its price or to a change in the price of a comple-
ment. For example, McDonald’s will select a number of franchises at which the price of
french fries is lowered to determine the effect of such a change on Big Mac sandwich sales.
Based on the results, the managers can estimate the sensitivity of Big Mac demand to the
price of the complement, french fries, and thereby derive a relevant cross-price elasticity of
demand.
While a valid mechanism for estimating demand, experimentation carries with it some
limitations. Among these is that to run a true, controlled experiment, only one determinant,
such as the price of french fries, should be changed at a time to determine its impact on Big
Mac sales. Suppose that the price charged by Burger King for its products rises or the local
income level falls at the same time that the McDonald’s french fries price is altered. If this
is the case, McDonald’s managers will get a contaminated and unreliable measure of the
impact of the price of their french fries on Big Mac sales.
Another limitation is one of generalizability. It may be incorrect to assume that the
experimental results obtained in one sample of McDonald’s franchises apply to all fran-
chises. The effect on Big Mac sales of a change in the price of french fries may be much
different in Ohio than in Osaka, Japan. Experimental results from a sample of Ohio fran-
chises thus may not generalize to those in Osaka.
Of course‚ the same positive network effects propelling a product’s rapid diffusion can also have adverse legal con- sequences. For example‚ the Justice Department’s antitrust case against Microsoft hinged critically on the positive net- work effects fueling the success of Windows.
The Justice Department alleged that a positive network effect allowed Microsoft to capture a dominant share of the personal computer operating system market. In turn, the built-in customer base gave Microsoft a significant edge in the browser market for its Internet Explorer product over rival Netscape Navigator, said the Justice Department, because, at no extra charge to consumers, Microsoft pack- aged Internet Explorer with its Windows operating system.
We will explore the Microsoft antitrust case more fully in Chapter 11, but one point that deserves mention here is
that positive network effects can be a two-edged sword for suppliers. Although a bandwagon effect enhances the possibility that a supplier will capture a dominant mar- ket share, it simultaneously limits the supplier’s ability to exploit that position through a price increase. As we saw in Figure 4.12, the market demand curve is more price elastic when there are positive network effects present. In the case of Windows, this implies that Microsoft’s abil- ity to exploit, through a price increase, the dominant cus- tomer base that a bandwagon effect helped to build is limited by the same bandwagon effect. Should Microsoft attempt to raise the price of Windows, customers would run toward alternative operating systems more quickly than they would without the bandwagon effect being present.
The Basics of Demand Est imat ion 101
C04.INDD 10:52:49:AM 08/06/2014 PAGE 101Trim Size: 203.2 mm X 254 mm
Surveys To anticipate the market reaction to a price increase, the Ford Motor Company regularly
conducts consumer surveys—either by mail, telephone or focus groups. A Californian will
be contacted by telephone, for instance, and asked about the extent to which her likelihood
of purchasing a Fusion automobile over the next 12 months will be diminished if the price
is raised by $400.
As with experimentation, customer surveys can generate valuable information. They are
not, however, foolproof. As with experimentation, one must be careful to choose a repre-
sentative sample. A good example of what can happen if this rule is not heeded involves
one of the first U.S. presidential polls—a survey taken prior to the 1936 election that pre-
dicted Republican Alf Landon would crush Democrat Franklin Delano Roosevelt. Exactly
the opposite occurred. The poll surveyed citizens whose names had been taken from tele-
phone directories and automobile registration rolls. Since the rich were more likely to have
telephones and cars in 1936 and were also more likely to be Republican, the poll was based
on a nonrepresentative sample of the voting population.
Furthermore, a survey’s reliability is dependent on respondents telling the truth. For exam-
ple, there is considerable evidence that polls will not accurately predict election results where
political candidates and voters are of different ethnic backgrounds. While those surveyed say
they will vote for a candidate whose ethnicity is different from their own, once in the privacy
of the voting booth they are more likely to pull the lever for a candidate similar to themselves.
Perhaps the classic case of misreporting by respondents involves the story of a
researcher who surveyed Americans regarding their sex lives. According to the research,
there was a significant difference between the number of sexual partners that males and
females in the sample reported having over their lifetimes. Men reported an average of 14
female sex partners. Women responded that they had 4 male sex partners. The samples,
moreover, appeared to be representative.
While the difference in the averages across the genders may at first blush appear provoc-
ative, some further thought should convince you that, from a statistical perspective, there
cannot be such a difference. The averages for both genders should be identical. After all,
every time heterosexual intercourse occurs with a new partner, it should raise both genders’
averages equally—regardless of how the averages are distributed within each gender.
Which gender is not telling the truth? Probably both of them. In response to this par-
ticular question, men likely have a tendency to overstate whereas women may undercount.
If this is the case, the true, common average falls in between.
Regression Analysis Private and public decisionmakers regularly rely on existing data to statistically estimate
demand. To see how this is done, suppose that you operate cable systems in 10 equal-sized
communities (10,000 homes in each community). The systems are characterized by the data
in Table 4.1. The quantity demanded of your basic tier service (Q) is represented by the number of basic tier subscribers in each community—where the basic tier features retrans-
mitted local broadcast signals and other networks such as ESPN, MTV, and CNN. The other
columns in Table 4.1 reflect the monthly basic tier price charged in each community (P), the monthly per capita income level of a community’s residents (I), and the monthly price charged per additional pay tier (PPAY) such as HBO and Cinemax.
To determine whether it would be profitable to raise or lower basic tier prices, you
would want to know how sensitive basic service demand is to its price, holding constant
other determinants such as per capita income and the pay tier price. Regression analysis, also called econometrics, is a statistical method that allows you to estimate this sensitivity
regression analysis (econometrics) a statistical method that allows one to estimate, among other things, the sensitivity of the quantity demanded of a good to determinants such as price and income
102 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 102Trim Size: 203.2 mm X 254 mm
based on existing data. It begins by assuming that we can specify an equation for the under-
lying data and that the data do not “fit” the equation perfectly.
Take the case of the information in Table 4.1. Let us start by supposing that only P influences Q and that the demand for basic cable in a community is best described by the following linear relationship:
Q a bP ei i i= + + , (1)
where the subscript i refers to the number of the system or the “observation” being consid- ered. The variable Q that the equation is seeking to explain is called the dependent variable. Any variable such as P employed to explain the dependent variable is called an explanatory variable. The error term, e, is included in the equation to account for: mistakes in data col- lection; determinants of demand other than P that are inadvertently or, due to a lack of data, intentionally omitted from the relationship; and/or the possibility that the demand for basic
cable is, to a certain extent, random and thus not predictable by economic models. The term
a is the intercept of the linear equation—the number of subscribers a system will have if the basic price and the error term equal zero. And b, the coefficient on the basic tier price, indicates by how much the number of subscribers will change per dollar change in the basic
tier price (b = ∆Q/∆P). Regression analysis usually employs the ordinary least-squares, or OLS, technique to
estimate equations such as the one we have specified for the demand for basic cable. OLS
estimates the “best fitting” intercept and coefficient for the specified relationship and the
employed data. Best fitting means that the estimated equation will be “as close as possible” to the observed data points. The technical criterion for “as close as possible” involves the
distances between the various data points and the estimated equation.
The specific manner in which OLS determines the equation that best fits the data is
beyond the scope of this book.10 Suffice it to say that the intercept and coefficient esti-
mated by OLS (where estimated is signified with a ∧, as in â and b̂) serves to minimize the distances between the data points and the estimated equation. The distances between the
data points and the estimated equation represent the errors made by the estimated equation
across the various data points.
ordinary least-squares (OLS) a technique for estimating the equation that “best fits” the data
Cable Demand Data
System Number
Basic Tier Subscribers
(Q)
Basic Tier Price (P)
Per Capita Income
(I)
Pay Tier Price
(PPAY) 1 3,300 $18 $3,900 $22
2 6,600 10 5,560 10
3 3,900 18 8,900 18
4 5,000 14 8,200 16
5 5,100 15 7,950 10
6 6,900 9 6,500 7
7 6,400 12 5,900 8
8 5,900 13 7,500 15
9 5,800 12 7,864 12
10 4,800 18 4,500 12
Average 5,370 $13.90 $6,677.40 $13
Table 4.1
10Most beginning econometrics texts explore the derivation of the intercept and coefficients through OLS in considerable detail. As long as you understand the intuition behind their derivation, that will suffice for the material in our book.
The Basics of Demand Est imat ion 103
C04.INDD 10:52:49:AM 08/06/2014 PAGE 103Trim Size: 203.2 mm X 254 mm
Figure 4.14 graphically depicts how OLS regression works. In the simple demand relation-
ship assumed by equation (1), the intercept and coefficient for the estimated equation that best
fits the data across the ten systems turn out to be ˆ , .a = 9 970 1 and ˆ .b = −330 9, respectively. The intercept estimate â implies that for the sample examined and the demand relationship assumed, if price were set equal to 0, the forecast number of basic subscribers, Q̂, would be 9,970.1. The coefficient estimate b̂ means that for every $1 increase in the monthly basic tier price, the number of basic subscribers decreases by 330.9 ( / )b Q P= Δ Δ . The slope of the estimated regression line is ΔP/ΔQ and thus equals 1 0 003/ ˆ . .b = −
The dots in Figure 4.14 represent the actual prices and quantities of the ten systems in
our sample. For the tenth system, for example, the price, P10, is $18 and the number of basic subscribers, Q10, is 4,800. The OLS-estimated number of basic subscribers for the tenth system, Q̂10, is equal to the value obtained for Q when one plugs in the tenth system’s price into the estimated regression line. Thus, ˆ , . . ( ) , .Q10 9 970 1 330 9 18 4 013 9= − = . The error made by OLS in estimating demand for the tenth system, ê10, is the difference between the actual and forecast number of basic subscribers and equals 786 1 786 110 10 10. ( . )
^ ^ Q Q e− = = . The OLS regression has calculated â and b̂ so as to minimize the sum of the squares of such errors across the ten systems in the sample. From the perspective of Figure 4.14, OLS posi-
tions the regression line, ˆ ˆa bPi+ , so as to best fit the scatterplot of data points—where the “fit” reflects the horizontal distances of the observed data points from the estimated equation.
The estimated OLS equation provides valuable demand-side information. For instance,
if we wanted to calculate the elasticity of demand for basic cable service at the average
basic tier price and number of basic subscribers in our sample, we would use the now-
familiar formula:
� = ( / )( / ),Δ ΔQ P P Q (2)
where P and Q are the average basic tier price and number of subscribers, respectively, for the sample. The first part of the right-hand side of the elasticity formula, ΔQ/ΔP, is the rate at which the number of a system’s subscribers changes per dollar change in the basic
price. This is none other than the b̂ or −330.9 estimated by OLS. Employing the average values for P and Q reported in Table 4.1, one obtains an elasticity of demand of 0.9. Since this is less than unity, demand for basic cable is inelastic when evaluated at the price and
Ordinary Least-Squares Regression Ordinary least-squares positions the regression line, ˆ ˆ ˆQ a bPi i= + , so as to “best fit” the sample data points.
Q
$18
0
P
$30.13
4,013.9
Q10
error10 = 786.1
4,800
Q10
Qi = a + bPi
a = 9,970.1
Point for system 10
Slope = 1/b = – 0.003
$1 330.9
Figure 4.14
104 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 104Trim Size: 203.2 mm X 254 mm
number of subscribers for the average system in the sample. This indicates that profit could
be increased by raising the average basic price.
Suppose that we estimated a more extensive basic cable demand relationship, such as:
Q a bP cI dPPAY ei i i i i= + + + + . (3)
This more extensive model tries to control explicitly for more of the factors that might
affect demand for basic cable. OLS regression proceeds in a fashion analogous to the one
employed in estimating the simpler equation (1). OLS calculates the intercept â and coef- ficients b̂, ĉ, and d̂ so as to best fit the observed data—now incorporating information on income, I, and the price of pay tier service, PPAY, in the estimation process. In this more extensive model, the estimated coefficient ĉ measures the independent effect of a $1 increase in per capita income on the number of basic subscribers in a community. Its value
indicates whether basic cable is a normal or an inferior good across the systems in the sam-
ple. The estimated coefficient d̂ associated with the pay tier price variable, PPAY, reflects the impact of a $1 increase in the pay tier price on the demand for basic cable, holding
constant the basic tier price and the per capita income level. Its value tells us whether pay
service is a complement or substitute for basic cable in our sample of systems.
Controlling explicitly for more variables can affect the estimate of the effect of basic
price on the demand for basic service. When equation (3) is estimated, ˆ , .a = 9 931 0, ˆ .b = −230 0, ˆ .c = −0 01, and ˆ .d = −99 5. By comparison, the estimated basic price coefficient, b̂, was larger in magnitude when equation (1) was used. Use of equation (3) rather than equation (1) causes the estimated demand elasticity to fall as well. This is probably a truer
estimate of the elasticity because it controls for additional causal factors.
We close this chapter by noting that while regression analysis is a powerful tool allowing
one to estimate the effects of individual determinants on demand while holding constant
other determinants’ effects, it also has its difficulties. For one, the intercepts and coeffi-
cients estimated by OLS regression are only as good as the data and the models to which
the analysis is applied. “Garbage in, garbage out,” as the saying goes. If the sample data are
nonrepresentative of the larger population or if the assumed demand relationship is incor-
rect (for example, linear when it should be nonlinear), unreliable estimates will result.
SUMMARY
By rotating the budget line confronting a consumer, we
can determine the market basket the consumer will select at
different prices, while factors such as income, preferences,
and the prices of other goods are held constant. The various
price–quantity combinations identified in this way can be
plotted as the consumer’s demand curve.
To determine whether a demand curve must have
a negative slope, we separate the effect of a change in
price on quantity demanded into two components, an
income effect and a substitution effect.
For a normal good, both income and substitution effects
imply greater consumption at a lower price. Thus the
demand curve for a normal good must slope downward.
For an inferior good, the income and substitution
effects of a price change operate in opposing directions.
If the income effect is larger, the demand curve will slope
upward. However, both theoretical reasoning and empiri-
cal evidence suggest this case is quite rare.
Consumer surplus is a measure of the net ben-
efit a consumer receives from consuming a good. It is
shown graphically by the area between the consumer’s
demand curve and the price line.
Consumer surplus can also show the benefit or loss a
consumer receives as a result of a change in the price of
the good.
Individual consumers’ demand curves can be aggre-
gated to obtain the market demand.
The price–consumption curve provides impor-
tant information about an individual’s elasticity of
demand.
An individual consumer’s purchases of a good may
be influenced by other individuals’ purchases through
network effects.
Three methods allow us to estimate individuals’ or
market demand curves: experimentation, surveys, and
regression analysis or econometrics.
C04.INDD 10:52:49:AM 08/06/2014 PAGE 105Trim Size: 203.2 mm X 254 mm
Review Quest ions and Problems 105
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
4.1 Explain how the indifference curve and budget line appa- ratus are used to derive a consumer’s demand curve. For a
demand curve, certain things are held constant. What are they,
and how does this approach hold them constant?
4.2 If the per-unit price of college education rises and the prices of all other items fall, is it possible for the consumer
to end up on the same indifference curve as before the price
changes? If so, will the consumer be purchasing the same mar-
ket basket? Support your answer with a diagram.
4.3 “A Giffen good must be an inferior good, but an inferior good need not be a Giffen good.” Explain this statement fully,
using the concepts of income and substitution effects.
*4.4 Assume that Joe would like to purchase 50 gallons of gasoline monthly at a price of $2.50 per gallon. However, the
$2.50 price is the result of a government price ceiling, so there
is a shortage, and Joe can only get 25 gallons. Show what this
situation looks like by using indifference curves and a budget
line. Then, show that Joe will be willing to pay a price higher
than $2.50 to get additional units of gasoline. (This result is the
demand-side reason for the emergence of a black market.)
4.5 Mary Gail has a different price–consumption curve associ- ated with each possible income level. If two of these curves
intersect, are Mary Gail’s preferences rational?
*4.6 If Edie’s income rises by 50 percent and, simultaneously, the price of automobile maintenance increases by 50 percent,
can we predict how Edie’s consumption of automobile mainte-
nance will be affected? Can we predict how, on average, Edie’s
consumption of other goods will be affected? Use the concepts
of income and substitution effects to answer this question.
4.7 Assume that Dan’s income–consumption curve for pota- toes is a vertical line when potatoes are on the horizontal axis.
Show that Dan’s demand curve for potatoes must be downward
sloping.
4.8 Given the OLS estimates for the coefficient and intercept in the basic tier cable demand equation (1), calculate the point
demand elasticity for system number 7.
4.9 Given the OLS estimates for the coefficients and inter- cept in the basic tier cable demand equation (3), calculate the
following:
a. Income elasticity of demand for basic service evaluated at the average values for monthly per capita income and the
number of basic subscribers across the sample of ten sys-
tems. Based on your calculation, is basic cable a normal or
an inferior good?
b. Cross-price elasticity of demand for basic service with respect to the pay tier price evaluated at the average values
for the pay tier price and the number of basic subscribers
across the ten systems in the sample. Based on your calcu-
lation, is pay tier service a substitute or a complement for
basic service?
c. The income elasticity of demand for basic service evaluated at the income and quantity data for system number 3.
4.10 Suppose that Lorena consumes only three different goods: steak knives, butter knives, and butcher knives. If, according
to Lorena’s preferences, butter and butcher knives are infe-
rior goods, must steak knives be a normal good? Explain your
answer.
4.11 Suppose that there are only five consumers of a software game program. The demand curve for each of the consumers
is identical. Will the market demand curve that is obtained by
horizontally summing across the five individual consumers’
demand curves be less or more price elastic at any price than
the demand curve for any of the individual consumers?
4.12 Suppose that the Downtown Athletic Club increases its monthly membership charge from $150 to $200. Among the
businesspeople belonging to the Club, would you expect lower-
level business managers to be more sensitive to the price increase
than senior managers? Explain using indifference curves.
4.13 Suppose that George is interested in only two goods, cigars and scotch. Employ the indifference curve/budget line
apparatus to show a case where a decrease in the price of cigars
leads to an increase in George’s scotch consumption. Does
this imply that cigars and scotch are complements to George?
Explain your answer.
4.14 Repeat the preceding question but assume that a decrease in the price of cigars leads to a decrease in George’s scotch
consumption. Does this imply that scotch is an inferior good in
George’s case?
4.15 In the tax-plus-rebate example discussed in the text, sup- pose that the government adjusts the size of the rebate so that
the consumer stays on her initial indifference curve (U2 in Figure 4.4). Show the results in a diagram. Can the government
achieve this result for all consumers? Why or why not?
*4.16 When the price of gasoline in Italy is $5 per gallon, Fabio consumes 1,000 gallons per year. The price rises to
$5.50 and, to offset the harm to Fabio, the Italian government
gives him a cash transfer of $500 a year. Will Fabio be better
or worse off after the price rise plus transfer? What will happen
to his gasoline consumption?
4.17 Left and right shoes are perfect complements for most people. If only the price of right shoes increased, what would
be the substitution effect of such a price change on the typi-
cal consumer’s consumption of right shoes? (Assume that the
only two goods that the consumer cares about are right and left
shoes.) What about the income effect?
4.18 If Clint’s elasticity of demand for cigars is equal to zero, are cigars a normal or an inferior good for Clint? Explain.
106 Indiv idual and Market Demand
C04.INDD 10:52:49:AM 08/06/2014 PAGE 106Trim Size: 203.2 mm X 254 mm
4.19 Define consumer surplus, and explain how you would show it in a diagram containing a demand curve for some
product. What would consumer surplus equal in Figure 4.7 if
the demand was perfectly elastic at the market price of $3 per
espresso cup?
4.20 Diamonds clearly satisfy less important needs than water, which is essential to life. Yet according to market
prices, the essential commodity, water, is worth less than the
less essential commodity, diamonds. Why would a vital com-
modity such as water sell for so much less than diamonds?
Does this imply that there is something wrong with a market
system that values diamonds more than water? Explain using
demand and supply curves for water and diamonds. In your
explanation, distinguish between the marginal and total ben-
efit of the two commodities.
4.21 Noneconomists sometimes refer to medical care as “invaluable” or “priceless.” Do you think these terms may be
simply imprecise ways of saying that the consumer surplus
associated with medical care is very large? Suppose that the
consumer surplus is immense. Explain why this is irrelevant
in deciding whether to provide more medical care. What is
relevant?
4.22 Could the snob effect ever overpower the combined sub- stitution and income effects associated with an increase in the
price of a good? Explain why or why not.
4.23 Explain how a bandwagon effect might speed up the rate at which DVD players are adopted by consumers. Do likewise
for the case of cable television subscriptions.
4.24 Suppose that the P–C curve associated with a pharmaceu- tical drug is downward sloping. If the government underwrites
a certain percentage of consumers’ drug purchases, will the
government outlays associated with such a program be greater
the larger the percentage of the purchases underwritten? Explain
why or why not. What if the P–C curve is upward sloping?
4.25 During the 1970s, 55 percent of banks with more than 15 branches installed automated teller machines (ATMs), as
opposed to 16 percent of banks with only 2 branches. Explain
why this phenomenon attests to the presence of positive net-
work effects.
C05.INDD 12:5:43:PM 08/06/2014 PAGE 107Trim Size: 203.2 mm X 254 mm
107
CHAPTER 5 Using Consumer Choice Theory
In the development of economic theory, the most important use of consumer choice analysis is to justify the negative slope of demand curves. Once consumer choice theory provides us
with a firm basis for believing the law of demand, many problems can be analyzed by using
the demand curve, without having to look more deeply into consumer behavior. Some
issues, though, are better analyzed through using the budget lines and indifference curves
developed by consumer choice theory.
In this chapter we will see how we can apply consumer choice theory to several interesting
and important questions. For example, we will see how providing “free” public schools
may lead to less consumption of education, why higher interest rates can lead people to
save less, and why lump-sum “medisave” accounts could make workers better off than the
current excise subsidy policy toward health care. Although some of the applications in this
chapter are important in themselves, they are included primarily to illustrate how the theory
of consumer choice can be applied to analyze a wide variety of problems.
Learning Objectives
Determine how an excise subsidy affects consumer welfare and why it results in a dead- weight loss. Examine how the public provision of a certain quantity of a good such as education may lead to less consumption of the good. Analyze how a voucher program would affect the quantity of educational services chosen by parents for their children. Explore the impact of per-bag charges versus a fixed annual fee on the amount of trash generated by a community, recycling, and household welfare. Develop an intertemporal model that illuminates the consumer’s choice to save or borrow and shows how changes in the consumer’s endowment and the interest rate affect that choice. Understand how the theory of consumer choice can explain what types of financial assets an individual intent on saving for the future should invest in.
Memorable Quote “If every school in the nation were to face a high level of competition both from other districts and from private schools, the productivity of America’s schools, in terms of students’ level of learning at a given level of spending, would be 28 percent higher than it is now.”
—Caroline Minter Hoxby, Stanford economist and school choice expert
108 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 108Trim Size: 203.2 mm X 254 mm
5.1 Excise Subsidies, Health Care, and Consumer Welfare An excise subsidy is a form of subsidy in which the government pays part of the per-unit price of a good and allows the consumer to purchase as many units as desired at the subsi-
dized price. For example, the government might pay half a consumer’s housing costs,
which effectively lowers the per-unit price of housing services by 50 percent. The most
common examples of excise subsidies are found in the income tax code. A tax credit for
child care expenses, for instance, reduces the taxpayer’s tax liability when these items are
purchased, and this lowers the price to the taxpayer in the same way as an outright excise
subsidy. Health care provides another example. Business has been the largest underwriter
of health care costs in the United States over the past several decades largely because the
costs of providing health care to employees have been tax exempt—in effect, subsidized by
the government. If company health benefits were not tax exempt, state and federal tax rev-
enues would be at least $100 billion higher per year.
Let’s look at an excise subsidy applied to the health care purchases of a particular con-
sumer. Suppose that the consumer has a weekly income of $100, the market price of health
care is $5 per unit, and the consumer will purchase nine units of health care without any
subsidy. The pre-subsidy optimal consumption point is shown in Figure 5.1 by the tan-
gency between indifference curve U1 and the budget line AZ. The consumer’s optimal choice is point W, consuming 9 units of health care and $55 worth of other goods. Note that total outlay on other goods, which is normally measured by A1, is also equal to the vertical distance WH1; this will be helpful to keep in mind later.
To subsidize the individual’s health care consumption, suppose that the government
pays $2 of the $5 per-unit cost of health care. The subsidy lowers the effective price of
health care from $5 to $3 and, as with any price reduction, causes the budget line to rotate.
The new budget line is AZ′. Because health care is now less costly to the consumer, more will be consumed. Given the preferences shown in the diagram, the new optimal point is
excise subsidy a form of subsidy in which the government pays part of the per-unit price of a good and allows consumers to purchase as many units as desired at the subsidized price
Excise versus Lump-Sum Subsidy A $2 per-unit excise subsidy to health care pivots out the budget line to AZ′; the optimal consumption point is W′, with a total cost of W′T. An equal-cost, lump-sum subsidy produces the budget line A′Z″, and the consumer is better off at point W″ at no additional cost to the government.
Health care
1H
0
W″
W ′
Z ′Z″
$130 = A ′
$55 = A1
$100 = A
W
Z
T
U2
U3
U1
H1 H2H3
Other goods
$3
(9) (15) (20) (26) (33.3)
$5
1H
Figure 5.1
Excise Subsidies , Health Care, and Consumer Welfare 109
C05.INDD 12:5:43:PM 08/06/2014 PAGE 109Trim Size: 203.2 mm X 254 mm
W′, where U2 is tangent to AZ′, and health care consumption is 15 units. What is the total cost to the government of supplying this subsidy? Because the subsidy is $2 per unit, and
the consumer purchases 15 units per week, the total cost is $30 per week. Figure 5.1 shows
total government outlays on the subsidy as the vertical distance W′T, the distance between the subsidized and unsubsidized budget lines at the level of consumption chosen by the
recipient. To see this, note that if the consumer had to pay the unsubsidized market price
for health care (still shown by budget line AZ) and purchased 15 units, there would only be TH2 dollars left to spend on other goods. With the subsidy, however, there are W′H2 dollars left to spend on other goods, or W′T dollars more than if the consumer had to pay the entire cost of 15 units of health care. Thus, the total cost to the subsidy recipient of consuming
15 units has fallen by W′T, or $30, because the government is absorbing that amount.
The Relative Effectiveness of a Lump-Sum Transfer So far in our analysis we haven’t done anything that couldn’t be accomplished just as easily
by using the consumer’s demand curve instead of indifference curves. We have shown that
when an excise subsidy lowers the price of a good to the consumer, consumption of the
good will increase. Now, however, consider some related issues that are better analyzed by
using consumer choice theory. Suppose that instead of an excise subsidy, the government
gives the consumer the same amount of assistance, $30 per week, in the form of a cash
grant to be spent any way the recipient wants. (Economists refer to a transfer of a fixed
amount as a lump-sum transfer.1) How will this change affect consumption, and will the consumer be better or worse off with the lump-sum grant? These questions are important
policy questions, and we can answer them by using consumer choice theory.
Substituting an unrestricted cash grant of $30 for the excise subsidy shifts the consumer’s
budget line in a different way. Since the lump-sum cash grant is the same as an increase in
the recipient’s income, it produces a parallel outward shift in the budget line, from AZ to A′Z″. Note that A′Z″ passes through point W′, the optimal market basket under the excise subsidy. The reason is that both subsidies involve the same total cost to the government. As we saw, the total cost of the excise subsidy is W′T, or $30; with a cash grant of $30 the new budget line, A′Z″, lies exactly $30 (vertically) above AZ at all points, including point T. For example, A′A equals W′T; both measure the size of the cash grant. Put differently, when the government uses a lump-sum cash grant, the consumer has the option to purchase the same
market basket that would be chosen under an excise subsidy of the same total cost.
With the new budget line A′Z″, the consumer’s preferred market basket is at point W″. The effects of substituting the cash transfer for the excise subsidy can easily be determined.
The consumer purchases less health care and more of other goods under the cash transfer and is better off (that is, reaches a higher indifference curve). These results must be true: because U2 is tangent to AZ′ at W′, it must intersect the A′Z″ budget line at point W′, indi- cating that preferred positions lie to the northwest of W′ on the A′Z″ line.
These conclusions can be explained in a different way. Although both subsidies have an
income effect that tends to expand health-care consumption, only the excise subsidy has a
substitution effect (because of the relatively lower price of health care) that further stimu-
lates health-care consumption. Confronted with a price of only $3 per unit with the excise
subsidy, the recipient purchases health care up to the point where its marginal benefit (the
MRS between health care and other goods) is only $3, at point W′. Thus, the fifteenth unit of health care purchased is worth only $3 to the recipient, but its true market cost is $5; it
was purchased only because the government absorbed $2 of its cost. The recipient would
rather have $5 worth of other goods instead of the fifteenth unit of health care. With a cash
transfer, this option is available.
lump-sum transfer a form of subsidy in which the government gives the consumer a cash grant to be spent in any way the recipient wants
1An excise subsidy is not a lump-sum transfer because it is linked to the consumption of a specific good and the amount of the subsidy varies with the amount of that good consumed.
110 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 110Trim Size: 203.2 mm X 254 mm
This analysis generally implies that a consumer will be better off if given a cash grant
instead of an excise subsidy linked to the consumption of a particular good. This result
explains why many economists favor converting welfare programs such as Medicaid, food
stamps, and housing subsidies into outright cash grants. If given cash, the recipients could
still afford to buy the same quantities of health care, food, and housing. And if they choose
to purchase something else, the presumption is that they prefer the alternative to the subsi-
dized good they were consuming.
Our analysis does not demonstrate that cash transfers are better; it only shows that the
recipients will be better off, according to their own preferences, if they receive cash. Of course, a person can take a paternalistic view, assume “subsidy recipients don’t know
what’s good for them,” and favor overriding their preferences when providing them with
a subsidy. Some people, for example, believe the poor wouldn’t spend enough of the cash
welfare payments on “necessities.” Existing data, however, indicate that low-income fami-
lies actually devote a larger proportion of their incomes to food, housing, and medical care
than do upper-income families.
Using the Consumer Surplus Approach The same analysis can be conducted from a different perspective using the concept of con-
sumer surplus (see Figure 5.2). The individual consumer’s demand curve is shown as d, and he or she is initially purchasing H1 units of health care at the market price of $5. When the government introduces the excise subsidy the price becomes $3, and the consumer now
purchases H2 units at the lower price. As we explained in Chapter 4, the benefit to the con- sumer from the lower price can be shown in the graph as area PEBP′. Now consider the cost to the government of providing this benefit. It is shown as area PCBP′, or the $2 sub- sidy per unit (PP′) times the quantity purchased with the subsidy in place (H2, or 15 units). It is clear that the cost to the government exceeds the benefit to the consumer by triangular
area ECB. This area is the deadweight loss of the excise subsidy. It is a measure of the loss in well-being that results from use of the excise subsidy.
Referring to area ECB as a deadweight loss does not mean that the consumer is worse off under the excise subsidy than with no subsidy at all. Obviously, the consumer benefits from the lower price; the benefit, in fact, is shown by area PEBP′. Instead, it means that the consumer could be better off by area ECB under an alternative subsidy of the same cost
deadweight loss also called welfare cost, a measure of the aggregate loss in well-being of participants in a market resulting from output not being at the efficient level
Excise Subsidy Using Consumer Surplus The excise subsidy reduces the price from P to P′. The benefit to the consumer (the increase in consumer surplus) equals area PEBP′, but the cost to the government is area PCBP′. The excess of cost over benefit is the deadweight loss, area ECB.
Health care0
(9) (15)
$5 = P
$3 = P ′
E
d
B
C
H1 H2
Price
Figure 5.2
Excise Subsidies , Health Care, and Consumer Welfare 111
C05.INDD 12:5:43:PM 08/06/2014 PAGE 111Trim Size: 203.2 mm X 254 mm
to the government. That is what we showed in Figure 5.1, with the consumer attaining a
higher indifference curve when an equal-cost, lump-sum cash transfer is used instead of the
excise subsidy. Indeed, the area ECB in Figure 5.2 is a measure of the monetary value to the consumer of being on U3 rather than U2 in Figure 5.1.
We can understand that area ECB constitutes a net loss in potential well-being in another way. The effect of the excise subsidy is to increase consumption from H1 to H2. How much are those additional units of health care worth to the consumer? That is shown in Figure 5.2
by the area EBH2H1. (Recall that the height of the demand curve is a measure of marginal benefit, so adding the marginal benefit of each additional unit from H1 to H2 yields area EBH2H1 as the combined value of the entire increase in consumption of health care stimu- lated by the excise subsidy.) In contrast, the cost of these additional units is equal to the
market price of health care times the additional units consumed, or area ECH2H1. (This cost is not borne fully by the consumer under the excise subsidy, but that part not paid for
by the consumer has to be paid for by someone—in this case, taxpayers in general.) We
can now see that the additional units consumed as a result of the excise subsidy have a cost
(ECH2H1) that exceeds their benefit to the consumer (EBH2H1) by area ECB. By offering the consumer an artificially low price, the excise subsidy prompts him or her to purchase
units that are worth less than their true cost.
A final connection between Figures 5.2 and 5.1 deserves mention. In Figure 5.2, the con-
sumer surplus increase produced by a $2 per-unit excise subsidy could be replicated with a
lump-sum subsidy equal to PEBP′. The excise subsidy thus costs the government more, by an amount equal to area ECB, to produce the same enhancement in consumer welfare. This offers another way to see why improving the consumer’s lot from U1 to U2 in Figure 5.1 (as measured by PEBP′ in Figure 5.2), through a lump-sum subsidy, requires less government expenditure than an excise subsidy. Whereas the excise subsidy requires $30 in government
expenditure (distance A′A on the vertical axis of Figure 5.1) to move the consumer from U1 to U2, a lump-sum subsidy would require parallel shifting out the original budget line AZ by less than $30 on the vertical axis to move the consumer from U1 to U2. Do you see why?
APPLICATION 5.1 The Price Sensitivity of Health Care Consumers
2This application is based on Charles E. Phelps, Eight Questions You Should Ask About Our Health Care System—Even if the Answers Make You Sick (Palo Alto, CA: Hoover Institute Press, 2010).
The definitive study on the price sensitivity of health care consumers was undertaken by the RAND Corporation.2 The study found that U.S. consumers are fairly sensitive to the nature of their insurance coverage and the extent to which they have “skin in the game” and pay for a portion of their health care expenses out of pocket. Individuals facing a 50 percent sharing or copayment rule on expenses up to a cer- tain “catastrophic cap” limit, beyond which they have com- plete insurance coverage, spent 33 percent less annually on face-to-face medical-visits and 28 percent less on hospital care than individuals with full insurance coverage and no required copayments. In addition, just the requirement of paying something up front and out of pocket had the big- gest observed impact on health care expenditures. That is, individuals facing, a $500 deductible whereby they are
directly responsible for the first $500 worth of any health care expenditures, spent 34 percent less on face-to-face medical visits and 10 percent less on hospital care than did individuals with no deductible.
The RAND results indicate that there is appreciable price elasticity of demand by health care consumers. Health care costs now account for one-sixth of the U.S. gross domestic product (GDP), growing from 6 percent of GDP in 1960 to 18 percent in 2014. Addressing these burgeoning costs will likely require confronting individual consumers more directly with the consequences of their decision through out-of-pocket copayments or deduct- ibles. That said, some key provisions of the ObamaCare health care reform legislation signed into law in 2010 promise to do little to “bend back” the medical cost curve. As we will see in the ensuring section, the enacted reforms are likely to end up raising total expenditures on health care through the manner in which they will further subsi- dize health insurance.
112 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 112Trim Size: 203.2 mm X 254 mm
5.2 Subsidizing Health Insurance: ObamaCare On March 23, 2010, President Barack Obama signed into law the Patient Protection
and Affordable Care Act, commonly known as ObamaCare. This sweeping legislation,
described in a bill that is over 2,500 pages long, is arguably the most significant economic
policy enacted in the past 50 years, affecting as it will almost every aspect of the health care
and health insurance markets. In this section we will use consumer choice theory to exam-
ine the effects of one part of this massive legislation: the subsidization of individual health
insurance purchases. The subsidies became effective as of 2014 while the other provisions
of the law are due to be phased in by 2020.
A major goal of ObamaCare is to reduce the number of people without health
insurance. In 2008, an estimated 46.3 million citizens in the United States lacked health
insurance, about 15 percent of the total population. The subsidies we will discuss are a
major part of the effort to reduce this number. Another part of the legislation intended
to address this issue is an expansion in existing government welfare programs (such as
Medicaid) that in effect provide health insurance to low-income people.
The Basics of ObamaCare The new ObamaCare subsidies work roughly like this: a health insurance policy offering
specific coverage (as determined by the government) is made available to those eligible for
the subsidy. It is estimated that the market (unsubsidized) cost of this type of policy will
be about $15,000 a year for a family of four people. Eligible families must purchase this
specific policy (or something similar) to receive a subsidy that will defray part of the cost
of this policy. The amount of the subsidy received depends on the family’s income. Lower-
income families receive a larger subsidy than higher-income families. In effect, the subsidy is
gradually phased out as income rises. Specifically, the lowest incomes eligible for the subsidy
are 133 percent of the federal government’s poverty line, and the highest incomes eligible are
around 400 percent of the poverty line.
At today’s poverty line (it is indexed to the consumer price index and so tends to rise
over time), this means that families (of four) with incomes between $30,000 and $90,000
are potentially eligible for the subsidies. We say “potentially” because the subsidies are
available only to families that do not receive health insurance through their employers.
Most working middle-income families receive health insurance as a fringe benefit in their
employment contracts, and they are not eligible for these subsidies.
Considering families of four persons, a family with an income of $30,000 will have to
pay about $1,000 to receive a health insurance policy that costs $15,000—a subsidy of
$14,000. (Those with lower incomes will be subsidized even more generously through
Medicaid.) A family with an income of $50,000 will have to pay about $4,000 to receive
the same policy—a subsidy of $11,000. A family with an income of $90,000 will have to
pay about $11,500 to receive the same policy—a subsidy of $3,500. It is worth noting than
median family income is about $60,000. So ObamaCare involves a major middle-class sub-
sidy; it is not just the poor who will be subsidized.
The Subsidy’s Effect on the Budget Line Let’s consider how this subsidy affects the budget line for an eligible family with an income
of $50,000. The unsubsidized budget line relating consumption of other goods and health
insurance (H) is shown as AZ in Figure 5.3, where it is assumed that H is measured in units that cost $100 each (that is, the slope of AZ is $100). Now the government introduces the subsidy: if the family pays $4,000, it will receive a health insurance policy costing $15,000
(providing 150 units of H). This does not affect the AB portion of the original budget
Subsidiz ing Health Insurance: ObamaCare 113
C05.INDD 12:5:43:PM 08/06/2014 PAGE 113Trim Size: 203.2 mm X 254 mm
line; the family can forgo the subsidy and purchase fewer than 150 units of H on its own. (A major qualification is discussed in the next subsection.) However, if the family gives
up AA1 of other goods ($4,000), it can receive 150 units of H or A1A′. Thus, the budget line becomes flat between B and A′. If the family purchases additional insurance beyond 150 units, it can do so on its own by paying $100 for each additional unit. The new (subsidized) budget line is shown as ABA′Z′.
You may find it instructive to contrast how this subsidy affects the budget line and how
the food stamps program affects it (discussed in Section 3.4). Both of these subsidies are
“in-kind” subsidies linked to consumption of particular goods. The food stamps subsidy,
however, does not require the recipient to pay anything to receive the subsidized package,
whereas the health insurance subsidy requires the recipient to pay part of the cost, with
other taxpayers covering the rest. Note also in Figure 5.3 that we can show the cost to the
government of providing the subsidy by the distance A′F (equal to A1A2, or $11,000 worth of other goods). We will use this measure when we analyze the costs and benefits of the
subsidy.
Bringing in Preferences Now, let’s bring preferences into the picture and examine some of the consequences of this
subsidy. There are a number of possible outcomes since the exact consequences depend
on the circumstances (such as income) and preferences of the family. However, we will
focus on one that may be of particular importance: how the subsidy affects families that
would choose to purchase no health insurance if not subsidized. In other words, they would
choose point A in Figure 5.3, a corner solution (see Figure 3.12). The reason that this case is of special interest is that the policy is intended in large part to see that families that would
otherwise have no insurance become insured. In fact, it turns out that nearly 40 percent of
the 46.3 million American citizens counted as uninsured prior to the passage of ObamaCare belong to families with incomes of $50,000 or more, so a corner solution is very common
among those intended to benefit from the subsidies.
Figure 5.4 shows one possible outcome for a family that would purchase no insurance
in the absence of the subsidy. This family would consume at point A on the AZ budget line in the absence of the subsidy. Confronted with the subsidy, they can attain a higher indif-
ference curve, U2, by moving to point A′ on the subsidized budget line ABA′Z′. In this case,
0
A′
Z′
B
$50,000 = A
$46,000 = A1
$35,000 = A2
H1 H2 Z
F
Health insurance
Other goods
(40) (150)
Health Insurance Subsidy’s Effect on the Budget Line With the subsidy, the budget line becomes ABA′Z′. By paying $4,000 (AA1), the family receives 150 units of health insurance (A1A′). This accounts for the flat portion of the budget line between B and A′. Beyond 150 units of health insurance, the family may obtain additional units by giving up $100 in other goods for each additional unit of health insurance. Hence, the slope of the remaining portion of the budget line, A′Z′, is the same as the slope of the initial segment AB.
Figure 5.3
114 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 114Trim Size: 203.2 mm X 254 mm
the subsidy has the intended consequence of providing health insurance coverage for a fam-
ily that would otherwise be uninsured.
The Costs and Benefits of the Subsidy Now let’s look at the costs and benefits of the subsidy. As mentioned above, the cost of
the subsidy is shown by A′F, or $11,000, in Figure 5.4. What is the benefit? Generally, the benefit is what the subsidy is worth to the recipient. We can determine that in the graph by answering this question: What unrestricted cash transfer would the family consider to be
equally beneficial as receiving the health insurance subsidy? In other words, if we get rid
of the health insurance subsidy, how much cash would we have to give the family to place
them on the same indifference curve attained by the subsidy—that is, to place them on U2? That question is answered by shifting outward the original budget line, AZ, parallel to
itself until it just touches or is tangent to U2. The required new budget line is JK, which allows the family to attain U2 by consuming at point J. Thus, a cash transfer of JA, or GF, is worth as much to the family as the health insurance subsidy that costs the government
(taxpayers) A′F. Or, for the case shown in the graph, the health insurance subsidy costs $11,000 but provides a benefit worth only $3,000 to the recipient family. The difference,
$8,000, is the deadweight loss of the subsidy.
As we saw in Section 5.1, there is an alternative way to demonstrate the existence of a
deadweight loss: we can give the recipient an unrestricted cash transfer that costs the same
as the health insurance subsidy ($11,000), and that will make the family better off. These
two ways of showing the deadweight loss—making the recipient better off at the same cost
and making the recipient equally well off at a lower cost—are simply two ways of showing
the same thing. The advantage to the approach in Figure 5.4 is that it provides a dollar mea-
sure of the size of the deadweight loss.
Other Possible Outcomes As indicated earlier, there are other possible outcomes besides the one shown in Figure 5.4.
For example, if the family’s preferences are such that they would choose a point between A′ and Z′ on the subsidized budget line, then there is no deadweight loss. In this case, the
The Optimal Consumption Choice and the Resultant Deadweight Loss The health insurance subsidy costs $11,000 (A′F) but provides a benefit worth only $3,000 (GF) to the recipient family. The difference, $8,000 (A′G), is the deadweight loss of the subsidy.
Other goods
Health insurance
U2 U1
A ′
$50,000 = A
$53,000 = J
G
H2 (150)
Z Z ′K
F
B
0
Figure 5.4
Subsidiz ing Health Insurance: ObamaCare 115
C05.INDD 12:5:43:PM 08/06/2014 PAGE 115Trim Size: 203.2 mm X 254 mm
health insurance subsidy would have the same effect as just giving the family $11,000
outright. (Compare the food stamps outcome in Figure 3.16a.) However, this outcome is
impossible for a family initially at a corner solution.3
To the extent that subsidized families would consume low or zero amounts of health
insurance in the absence of the subsidy, we would expect the deadweight losses to be sub-
stantial. Such families demonstrate by their choices that they get little benefit from health
insurance, so inducing them to consume a very large health insurance policy seems likely
to be quite inefficient—that is, to have a large deadweight loss.
But the actual results of this health insurance subsidy can be even worse than suggested
by Figure 5.4 because we have ignored two important characteristics of the subsidies.
Can a Subsidy Harm the Recipient? In general, economists argue that subsidies cannot harm the recipients of the subsidies, at
least in the case where they do not have to pay taxes to finance the subsidies. The reason
is straightforward: people are not required to accept the subsidy. If people believed they
would be worse off with the subsidy, they could simply refuse to accept it and avoid any
harm. There may be a few exceptions to this for cases where the subsidy has indirect effects
on those who do not accept the subsidies, but they are thought to be uncommon.
The health insurance subsidies in ObamaCare are different in an important way. A key
characteristic of the 2010 health care reform legislation is that all people are required by
law to have an acceptable (according to government regulations) health insurance policy.
Thus, some may be forced to accept the subsidies even if they would rather not.
This is the controversial “mandate” in the legislation. Some have argued that it is
unconstitutional for the federal government to require people to purchase a specified good
whether they want to or not. Indeed, roughly half of the state attorneys general mounted a
constitutional challenge to this mandate. We have no expertise in constitutional law, but it
is easy to use our consumer choice model to show that a mandate combined with the sub-
sidy can make some recipients worse off than they would be with no subsidy at all.
Figure 5.5 illustrates this case. Again, this is initially a corner solution, with the fam-
ily on U1 at point A on the unsubsidized budget line AZ. Note that if given the option, this family would choose to forgo the subsidy at point A rather than pay $4,000 and receive the health-insurance policy at point A′. Without a mandate, this family would be better off not accepting the subsidy. With the ObamaCare mandate, the family is required to consume at
point A′; on the lower indifference curve U0. Just as we did previously, we can use an unrestricted cash transfer, in this case negative
transfer (a tax), to show the deadweight loss. The family would be equally well off on A″Z″ (a tax of $2,000) at point A" as they are at point A′. In other words, a tax of AA" has the same effect on their well-being as the subsidy plus mandate; they are on indifference curve
U0 in both cases. The deadweight loss is A′G, larger than the cost of the subsidy. Here, the cost of the subsidy is $11,000, and the benefit to the recipient is minus $2,000, so the dead-
weight loss is $13,000.
One Other Option Families have one other option in this scenario: they can pay a fine rather than participate
in this subsidy program. The law stipulates that those who do not have “minimum essential
coverage” as defined by the government are subject to fines that can be as much as approxi-
mately $2,000 (varying with income and family size). Suppose the fine for the family in
Figure 5.5 is $1,500. Then, this family would prefer to pay the fine rather than participate
in the subsidy program, though of course they would be worse off than without the subsidy
3This is true if other goods and health insurance are both normal goods (as is almost certainly the case). Can you explain why?
116 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 116Trim Size: 203.2 mm X 254 mm
altogether. You may want to show this outcome in the graph. (Hint: Draw a budget line parallel to AZ and between AZ and A″Z″ to represent the option of the fine.)
We have no idea exactly how many people will be worse off because of the subsidy/
mandate/fine policy, but clearly many will. The Congressional Budget Office has estimated
that roughly 4 million Americans will be paying the fines in 2016; they are clearly harmed
by the policy. That estimate does not include those who participate in the program and are
harmed, such as the family shown in Figure 5.5.
We have been examining the way families are affected in their consumption choices
regarding health insurance and other goods. There is another effect on the labor market:
these health insurance subsidies may well discourage the affected families from working
and earning as much as they otherwise would.
The effect on work incentives stems from the fact that the subsidies decline as family
earnings rise. Recall that the subsidies are larger at lower income levels. A family with
an income of $66,000, for example, will have to pay $2,800 more for the same insurance
policy than if they earned $54,000. This is like a tax on their additional earnings. Typically,
this implicit marginal tax rate on earnings is around 20 to 25 percent, meaning that for each
additional dollar earned, 20 to 25 cents goes to the government. When added to the explicit
tax rates facing these families, which are in the 30 to 40 percent range, these subsidized
families will confront overall marginal tax rates on their incomes of 55 to 65 percent. That
is high enough to constitute a significant disincentive to work.
5.3 Public Schools and the Voucher Proposal Paying part of the price of some good, thereby reducing the price the consumer must pay, is
not the only way to subsidize consumption. In fact, a more common form of subsidy is one
in which the government makes a certain quantity of a good available at no cost, or perhaps
at a cost below the market price. The essential characteristic of this type of subsidy is that
Figure 5.5 A Case Where Mandated Insurance Harms Subsidy Recipients With the mandate, the family is required to consume at point A′ on the lower indifference curve U0 instead of at its optimal point A prior to the subsidy. Here, the cost of the subsidy is $11,000 (A′F), and the benefit to the recipient is minus $2,000 (a cost of FG), so the deadweight loss is $13,000 (A′G).
Health insurance
H2 (150)
Z″ Z Z ′
G
F
B
U1
U0 A ′
Other goods
$48,000 = A″
$50,000 = A
0
Publ ic Schools and the Voucher Proposal 117
C05.INDD 12:5:43:PM 08/06/2014 PAGE 117Trim Size: 203.2 mm X 254 mm
the quantity of the good being subsidized is beyond the control of the recipient. Beyond the
health insurance subsidy associated with ObamaCare and analyzed in the previous section,
the public school system provides another significant example. Public schools make a certain
quantity of educational services available to families at no direct cost;4 if a larger quantity or a
different type of schooling is desired, the only alternative is for families to use private schools
and pay the market price (tuition), because the government subsidizes only public schools.
In some cases, this type of subsidy may actually reduce consumption of the subsidized
good; that is, the provision of education through public schools may decrease the amount of
education consumed. This paradoxical outcome can occur when it is impossible, or very dif-
ficult, for the consumer to supplement the subsidized quantity of the good provided by the
government—as is true with public schooling. To clarify this point, let’s assume that we can
measure the quantity of educational services by the expenditure per pupil and that the public
school provides services at a cost of $7,000 per pupil. Now suppose that a family prefers
to spend $8,500 per child on schooling (perhaps involving smaller class sizes or more com-
puter training) and is willing to pay the $1,500 difference in cost for the greater quantity. This
option, however, is not available at the public school since the government provides only a
fixed quantity of services. Moreover, the family cannot use the services of the public school
and, by paying $1,500 extra, obtain a $8,500 education for its child. The family must either
accept the public school subsidy as is or forgo it altogether and seek out a private school
that will deliver the $8,500 in educational services it desires. In such a “take it or leave it”
situation, it would cost the family $8,500 to get $1,500 more in educational services. In this
setting the family may decide to send the child to the public school; in the absence of the public
school subsidy, however, it would have chosen to purchase a $8,500 education.
The way public schools can lead to reduced consumption of schooling is clarified further
in Figure 5.6. The pre-subsidy budget line relating educational services and all other goods
is AZ. In drawing this budget line, we assume that alternative quantities of educational services are available at private schools. In the absence of any public school subsidy the
family will choose to consume S1 of educational services (the amount referred to earlier as $8,500), and the total outlay on other goods is A1 ($41,500).
By providing a public school that children can attend at no cost, the government offers
the lower-quantity educational services of S2 (the amount referred to earlier as $7,000). As shown in Figure 5.6b, the budget line confronting the family becomes AA′RZ. The initial S2 units of educational services ($7,000, or the length of segment AA′) are available to the family at no cost. To consume more schooling than the subsidized quantity S2, the family must forgo the public school subsidy and bear the entire cost of schooling along the RZ portion of the original budget line. If the family sends its child to the public school, it can
consume S2 of schooling, leaving vertical distance A (the family’s entire disposable income of $50,000) available to spend on other goods. However, to consume more than S2 units of schooling—for example, S1S2 extra units (or $1,500)—the family will have to give up AA1 ($8,500) in other goods. This is because to consume $8,500 of educational services, the family must forgo the public school subsidy altogether and enroll its child at a private
school at a cost of $8,500. Thus, there is a very high marginal cost associated with con-
sumption beyond S2. In other words, S2 units of educational services can be consumed with no sacrifice in income, but to consume a larger quantity, the family must pay the total (not the marginal) cost of the desired education. Thus, the budget line is drawn as AA′RZ.
Confronted with the AA′RZ budget line in Figure 5.6b, the family will choose to send its child to the public school and consume at point A′; this choice places the family on a higher indifference curve, U2, which is preferable to consuming at point W on indifference curve U1. Although the family is consuming S1S2 ($1,500) less of educational services, there is a
4Public schools are usually financed by property taxes, but the taxes paid typically bear no relationship to the amount of schooling consumed. Senior citizens with no children may pay the same taxes as a family with four school-age children.
118 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 118Trim Size: 203.2 mm X 254 mm
gain of AA1 ($8,500) in other goods, and the indifference curves show the net effect to be an improvement. This analysis illustrates how the provision of public schools can lead to
a reduction in the amount of education received by some children. Keep in mind that this
analysis has been conducted for a family with a high demand for educational services, as
shown by its preference for a larger quantity of schooling than that provided by the public
schools. Other families will be affected differently. For instance, a family that may have
chosen a point along the AR portion of the AZ budget line in the absence of the public school will actually consume more education when the public school subsidy is available.
Which type of outcome is more common is not known.
Using Consumer Choice Theory to Analyze Voucher Proposals The preceding analysis also provides a framework to examine a proposal for a major
change in public school financing that has attracted a great deal of interest in recent years,
school vouchers. With a voucher program, parents receive vouchers that can be used to purchase education at any school they choose. The program operates as follows: Assume
that the public school is currently spending $7,000 per pupil. Parents receive a voucher for
each school-age child; the voucher is redeemable for $7,000 when spent on education. If a
family decides to continue to send its child to the public school, it turns in the voucher there
and nothing changes. In a voucher program, however, the family also has the option of taking
the public school subsidy to a different school (either public or private) and using the
voucher to pay for all or part of the cost of education there. For instance, to purchase a
$8,500 education at a private school, the family turns in the voucher there (and the private
school receives $7,000 from the government) and pays $1,500 of its own money.
voucher program a form of subsidy in which parents receive vouchers that can be used to purchase education at any school they choose
Fixed-Quantity Subsidy: Education (a) With no subsidy, the family confronts budget line AZ and consumes S1 units of schooling and A1 units of all other goods. (b) If the family cannot augment the quantity of the subsidized good, the result is a budget line like AA′RZ. This situation can actually lead to lower consumption of the subsidized good than would occur with no subsidy at all. The optimal consumption point with the subsidy is A′ with S2 units of schooling, less than the S1 units that would be purchased without the subsidy.
Education (dollars per pupil)
0
W ′
Z ′
$50,000 = A
$41,500 = A1
A2
A′
W
Z
(b)
R
U2 U3
U1
S1S2 S3
Other goods
($7,000) ($8,500)
U1
Education (dollars per pupil)
0
A1
Z
W
(a)
S1 ($8,500)
Other goods
$50,000 = A
Figure 5.6
Publ ic Schools and the Voucher Proposal 119
C05.INDD 12:5:43:PM 08/06/2014 PAGE 119Trim Size: 203.2 mm X 254 mm
Figure 5.6b can also be used to illustrate some of the effects of a school voucher. With
the voucher the budget line becomes AA′Z′. The family can continue to use the public school system, staying at point A′; note, however, that consuming less education is not pos- sible because the voucher can be used only to purchase educational services. Alternatively,
the family can use the voucher to help purchase a greater quantity of educational services
along the A′Z′ portion of the budget line. The exact outcome depends on the preferences of families and will vary from one family to another. The preferences shown in the diagram
depict a family who will move its child to a private school offering S3 units of educational services and pay for it with the voucher plus AA2 of its own income.
Thus, we expect the voucher program to lead some families to purchase a larger quantity
of educational services for their children—or at any rate a type of education they view as
superior to that offered in the public schools. These families will be better off, because they
are on a higher indifference curve at point W′. Other families would choose to continue receiving the same educational services at the public school.
This is a good place to emphasize that an analysis using budget lines and indifference
curves is incomplete in some cases, because it assumes that the supply side of the market
adjusts passively to consumer demands, which is not always true. Here, at least two types
of changes on the supply side of the market have been predicted to occur: one favorable to
the voucher proposal, the other not. First, with the voucher program in place, public schools
would be in direct competition among themselves, and with private schools, for students.
Under the current system public schools do not have a strong incentive to improve educa-
tional services, because they are giving away what their competitors must sell. Increased
competition could lead to a better quality of education at both public and private schools.
On the other hand, the voucher program may lead to an education system that is more seg-
regated by income, race or religion. Parents with higher incomes, for example, might use
their vouchers (as well as their financial resources) to purchase very expensive private edu-
cations for their children, reducing the resources of public schools and thereby lowering the
quality of education for children who continue to attend public schools. Of course, lower-
income families may also use the vouchers to purchase better-quality educations for their
children than their neighborhood public schools can provide, because they have the option
of using the voucher at any public school as well as at a private school. We will not try
to evaluate these consequences here because our main purpose is to show how consumer
choice theory can be used to analyze the effects of various options on consumer behavior.
APPLICATION 5.2
Starting in the 1990s with the cities of Milwaukee and Cleveland, some type of voucher program has now spread to 11 states and the District of Columbia.5 In 2011, for example, Indiana passed a state-wide voucher program offering up to $4,500 per student for families with annual household
The Demand for and Supply of School Choice
incomes up to $41,000 and lesser benefits to households with higher incomes. Based on the experience in Cleveland, the state of Ohio followed suit in 2013 with a state-wide EdChoice program that allows for up to 60,000 scholar- ships to be offered to families with incomes at or below 20 percent of the federally designated poverty line—$4,250 per student at the K-8 grade level and $5,000 for students in grades 9–12.
Voucher programs have grown notwithstanding fierce opposition from public school boards and teachers unions. Opponents claim that vouchers drain money from public schools; subsidize private, religious-based schools; and pro- mote an education system that is more segregated by race and income.
5This application is based on: http://en.wikipedia.org/wiki/ School_voucher; Milton Friedman, “The Market Can Transform Our Schools,” New York Times, July 2, 2002, p. A19; “Itching to Get Out of Public School,” Businessweek, May 10, 1999, pp. 38 and 40; Paul E. Peterson, “Charter Schools and Student Performance,” Wall Street Journal, March 16, 2010, p. A23; and Paul E. Peterson and Martin R. West, “African-Americans for Charter Schools,” Wall Street Journal, August 3, 2010, p. A15.
120 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 120Trim Size: 203.2 mm X 254 mm
Notwithstanding the opposition, there does seem to be healthy demand for greater school choice, especially on the part of low-income families. In Cleveland, Ohio, for example, more than 5,000 families are taking advantage of a voucher program that initially provided $2,250 per year per child sent by such families to a private school charging no more than $2,500. Families participated in the program even though they had to pay up to 10 percent of the private school tuition out of their own pockets to presumably attain a higher indifference curve (much as depicted in Figure 5.6b by the move from point A′ on U2 to point W′ on U3).
Moreover, while 96 percent of the children receiving vouchers attend religious schools (largely parochial, Catholic- run), the Cleveland program has survived a legal challenge that went up to the U.S. Supreme Court in 2002. By a slim 5–4 margin, the Supreme Court declared the program constitutional on the grounds that parents, and not the government, decided where the tuition dollars end up and thus that the voucher program represented true private choice as opposed to state-sponsored religion.
Furthermore, the Cleveland school choice program decreased rather than increased racial segregation. Of pub- lic school students in the Cleveland metropolitan area, 61 percent attend schools that are racially segregated (where more than 90 percent of the students are of the same racial background). By contrast, 50 percent of the low-income students attending private schools through the voucher program ended up at racially segregated schools.
On a broader basis, when philanthropists Theodore Forstmann and John Walton established the Children’s Scholarship Fund in 1999 to provide opportunities around the country for low-income families to send their children to private schools, they were overwhelmed by the demand for such scholarships. The fund initially offered $1,000 per year for four years in private school scholarship support to 40,000 kids from low-income families. The parents of more than 1.2 million low-income children applied for the scholarships, even though applying required them to commit $1,000 per year for four years of their own money toward sending their children to private schools.
The demand for school choice is driven by declining educational outcomes in many major city public school dis- tricts—outcomes that have been decreasing notwithstanding increases in real per-student public spending levels over the last several decades. In Detroit, for example, 22 percent of public school students graduated from high school in a recent year while New York City averaged 39 percent, Milwaukee 43 percent, Cleveland 44 percent, and Los Angeles 44 percent.
Beyond the demand on the part of low-income par- ents for school choice, previous experiments with voucher programs in public education programs suggest that the responsiveness of supply should not be underestimated. Post World War II‚ for example‚ the federal government enacted the G.I. bill‚ which provided vouchers for higher
education to more than 15 million military veterans for use at either private or public universities. Contrary to worries expressed at the time that colleges would not be able to expand rapidly enough to handle the flood of new students‚ enrollments nearly doubled between 1945 and 1947 (with veterans accounting for 49 percent of college students in the latter year).
Economist Milton Friedman has argued that the historical example of what happened on the supply side in the case of the G.I. bill at the college level will be mirrored through the greater use of vouchers at the elementary and secondary school levels. Moreover‚ enhancing voucher amounts may end up diminishing the role played by religious private schools—a key objection to the Cleveland experiment. As Friedman puts it:
Most schools that accept vouchers are religious for a simple reason‚ and one that is easily corrected. That reason is the low value of the voucher. It is not easy‚ perhaps not possible‚ to provide a satisfactory edu- cation for $2‚500 per student. Most private schools spend more than that. But parochial schools are able to accept the lower voucher amount because they are subsidized by their churches. Raise the voucher amount to $7‚000—the sum that Ohio state and local governments now spend [as of 2000] per child in government schools—and make it available to all students‚ not simply to students from low-income families‚ and most private schools accepting vouchers would no longer be religious. A host of new nonprofit and for-profit schools would emerge. Voucher-bearing students would then be less dependent on low- tuition parochial schools.
Beyond vouchers, public support is also growing for charter schools. Nearly 6,000 charter schools around the country now serve approximately 4 percent of the U.S. school-age population (versus 1,600 schools and less than 1 percent in 2000). Charter schools are funded by gov- ernments but typically operate independently from local school districts and receive only 60–70 percent of the per- pupil support from the government, as do traditional public schools. Charter schools have to persuade parents to select them instead of a neighborhood district school. As of 2012, 365,000 families were on waiting lists for charter schools, enough to fill an additional 1,050 charter schools.
Although a sizable portion of the U.S. public remains undecided, charter school supporters outnumber oppo- nents two to one. Among African Americans, those favor- ing charter schools outnumber opponents four to one (even though the NAACP remains opposed), and have included some leading politicians such as Senator Cory Booker, who, while as Mayor of Newark, New Jersey, spoke in favor of vouchers. Even among public school teachers, 37 percent favor charter schools, while 31 percent are opposed.
Paying for Garbage 121
C05.INDD 12:5:43:PM 08/06/2014 PAGE 121Trim Size: 203.2 mm X 254 mm
5.4 Paying for Garbage The Borough of Perkasie, a small Pennsylvania town, had a problem: throughout the 1980s
its trash collection costs rose rapidly.6 The local government devised an innovative solution
to the problem by changing the way residents paid to have their trash picked up. Historically,
Perkasie residents paid a fixed annual fee of $120 per residence for garbage collection, a system
of payment typical of many U.S. communities. Under the new plan introduced in 1988, there
is no annual fee but garbage is picked up only when it is placed in specially marked, black
plastic bags. The bags are sold by the town at a price greater than their cost; for example, each
large bag costs $1.50. The net revenue from sale of the bags (revenue less the cost of the bags
to the town) is used to finance the town’s trash collection services. This change in the system of paying for garbage collection produced some dramatic
effects in Perkasie. The amount of trash collected dropped nearly 50 percent, the average
household spent about 30 percent less on garbage collection, and the town saved 40 percent
on its garbage collection costs. Why did these changes occur? The answer lies in considering
the incentives faced by households under the two payment systems. With a fixed annual
fee, households faced an effective price of zero for the trash collection service: if a house-
hold doubled the amount of its trash, it bore no additional cost. In effect, the fixed fee gave
no incentive to cut down on the amount of trash generated. By contrast, under the bag system,
when more trash is generated, more bags have to be purchased and households bear a cost
directly associated with generating more trash. Basically, switching to the bag system
increased the price per unit households had to pay for trash collection services, and that
increase gave them an incentive to cut down on the trash they generated.
We can better understand how Perkasie’s new trash collection payment system affected households by applying consumer choice theory. In Figure 5.7, the amount of trash dis-
posal “consumed” by a typical household is measured (in pounds) on the horizontal axis, and its consumption of all other goods is measured vertically. The household’s income is
indicated by A′—the vertical height of the budget line when no trash disposal is consumed.
6Timothy Tregarthen, “Garbage by the Bag: Perkasie Acts on Solid Waste,” The Margin, September/October 1989, p. 17.
Consumer Choice: Garbage Disposal Under the fixed annual fee, the consumer’s optimal consumption point is W, where indifference curve U1 is tangent to budget line AZ. When trash disposal is sold at a price per unit, the budget line becomes A′Z′, and the optimal consumption point becomes W′; thus, less trash is generated.
Trash disposal (in pounds per household)
0
W ′
Z ′
A ′
A1
A W
Z
U2 U1
Other goods
1,500
$0.06
1
2,000
Figure 5.7
122 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 122Trim Size: 203.2 mm X 254 mm
Now consider the budget line under the original $120 annual fee system. Payment of
the fee reduces the disposable income of the household to A (A′A = $120). After paying that fee, the household can dispose of any amount of trash it wants at no additional cost. Thus,
the budget line becomes the horizontal line AZ, implying a zero price per unit (that is, the line has a zero slope) for trash disposal. Confronted with that budget line, the consumer’s optimal point is W, where the indifference curve U1 is tangent to AZ.
Before going further, we should explain the shape of U1. Note that we have drawn it with the usual shape to the left of point W, but its slope at point W is zero and then the indifference curve becomes positively sloped to the right of point W. Normally, only the downward-sloping portion of indifference curves is relevant. When consumption of a good (measured horizon-
tally) becomes large enough, however, it can become so abundant that it is a nuisance—that
is, an economic “bad.” Then the indifference curve takes on a positive slope. When positive
prices are paid for all goods, we have to be concerned with only the downward-sloping por-
tions of indifference curves, but in this case the household is confronted with a zero price.
Thus, the household will expand consumption up to the point where the marginal benefit (MRS) of another unit of trash disposal is zero, equal to the marginal cost, as shown at point W.
Now let’s consider how the situation changes when the town introduces the bag system. In effect, having to purchase bags is analogous to having to pay a price per unit of trash dis-
posal. Suppose that the price of trash disposal implicit in the price of the bags is $0.06 per
pound and that this is equal to the cost of the city’s trash disposal service. Then the household
will be confronted with the budget line A′Z′, which has a slope of $0.06. We have drawn this budget line through point W to reflect the assumption that the total cost to the household of the bag system will be the same as the annual fee if the household continues to generate the
same amount of trash. (At $0.06 per pound, 2,000 pounds of trash will cost $120, or A′A.) Basically, we want to assume that this is an average household, so that the annual fee ($120)
just covers the cost of trash disposal ($120) under the fee system. Under the bag system, of
course, the payments made will also cover the city’s cost of trash collection.
Thus, with the introduction of the bag system, the household’s budget line shifts from
AZ to A′Z′, reflecting an increase in the price per unit of trash disposal from zero to $0.06 per pound. Even though the household could continue disposing of 2,000 pounds of trash
for $120, just as it was doing under the annual fee system, the graph makes it clear that
the household will cut back on the amount of trash. The new optimal consumption point under the bag system is W′, where the amount of trash disposal has been reduced to 1,500 pounds, and the total annual cost to the household has fallen to A′A1 ($90). This analysis explains the dramatic effects observed in Perkasie in response to the bag system: when resi-
dents have to pay for each unit of trash generated, they have an incentive to cut down on the amount of trash they generate. But the graph also makes another point clear: the aver-
age household is likely to benefit from this change in the pricing arrangement. Note that the
household attains a higher indifference curve under the bag system. This will necessarily be
the case for the average household that was previously generating an amount of trash that
cost the town $120 (the amount of the annual fee) to eliminate.
Does Everyone Benefit? Why would anyone object to a system that leads to less garbage and also benefits people?
The problem is that not everyone benefits when switching from a fixed annual fee to a
per-bag pricing system. The average household—one that generates the average amount of
trash—can be expected to benefit, but not all households are average.
To see how the effects differ when households differ, consider two households with the
same income but with different amounts of trash. Suppose that under the initial fixed annual
fee system household M generates 2,500 pounds of trash, and household N generates 1,500
pounds of trash. Both households confront the AZ budget line in Figure 5.8, but M’s opti- mal consumption point is ′WMand N’s is ′WN. (Note that the two sets of indifference curves
Paying for Garbage 123
C05.INDD 12:5:43:PM 08/06/2014 PAGE 123Trim Size: 203.2 mm X 254 mm
are for two different households. We assume they have the same income, so both confront the same budget line.)
When the price-per-bag system is introduced, the budget line becomes A′Z′ for both households but affects them in different ways. Household N benefits from the change,
moving to a new optimal consumption point ′′WN on a higher indifference curve (U2N). In contrast, household M becomes worse off because of the change, moving to point
Trash Disposal: The Bag System With the same income, both households confront the same initial budget line. Household M has trash disposal of 2,500 pounds and household N, 1,500 pounds. With the bag system, the common budget line becomes A′Z′, and this change benefits household N but harms household M.
WN″
WN′
Other goods
A
U N
WM″ WM′ Z
A′
0 1,500 2,000 2,500 Z ′ Trash disposal (in pounds per household)
2
U N1
UM1
UM2
APPLICATION 5.3
There have been numerous news stories in recent years concerning the problem of garbage disposal in the United States. Despite the many admo nitions to moderate their trashy habits, Americans generate about 4 pounds of gar- bage per day, up from 2.6 pounds per day in 1960.7 In part, this behavior results from the fact that most communities use systems such as a fixed annual fee to finance trash col- lections, so residents don’t have to pay more when they discard more trash. The Perkasie experience shows that some pricing policies can give people an incentive to gener- ate less trash. Starting from close to the national average in terms of the trash produced on a per-citizen basis, Perkasie was able to cut its trash collections in half by switching to a per-bag pricing scheme.
Charging by the bag not only can reduce a municipality’s garbage collection costs and more fairly allocate expenses across households based on the amount of garbage
Trash Pricing and Recycling
generated, but it also encourages recycling, a substitute for trash generation. For example, the largest and oldest per- container billing system in the United States is in Seattle. Since the billing program’s inception in 1981, the percentage of trash recycled in Seattle has risen from 5 to 42 percent.
At the present time, about 4,000 U.S. municipalities (25 percent of the total) require households to purchase a special can, bag, tag or sticker for each unit of garbage collected. It is estimated, moreover, that increasing the curbside price of garbage from zero to 85 cents (per bag) reduces average weekly household garbage from about 30 pounds to 20 pounds—roughly one-third of the total.
The principal complaints regarding per-bag billing include that it provides an incentive for illegal dumping. Some Seattle homeowners, for example, routinely leave their garbage in apartment house dumpsters. In response, the owners of such dumpsters often padlock them. In addi- tion, there are administrative costs associated with imple- menting per-bag billing charges that need to be weighed against the benefits associated with reduced waste genera- tion and increased recycling.
7This application is based on Elbert Dijkgraaf and Raymond Gradus, “Per-Unit Garbage Charges,” Journal of Economic Perspectives, 22, No. 2 (Spring 2008), pp. 243–244.
Figure 5.8
124 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 124Trim Size: 203.2 mm X 254 mm
′′WM on a lower indifference curve (U1M) In general, households that generate the aver- age amount of trash (2,000 pounds) or less will benefit from changing to a price-per-bag
arrangement, while households that generate more than the average amount are likely to be
harmed. Under the annual fee system, households that generate little trash tend to pay more
than the cost of disposing of their trash, and that excess implicitly subsidizes heavy users of
the trash disposal service. The price-per-bag system makes each household pay the cost of
its own trash disposal, thereby removing the implicit subsidy arrangement of the annual fee
and harming those who were on balance being subsidized.
5.5 The Consumer’s Choice to Save or Borrow Saving involves consuming less than one’s current income, which makes it possible to
consume more at a later date. Borrowing makes it possible to consume more than current
income, but consumption in the future must fall below future income to repay the loan. A
decision to save (or borrow) is therefore a decision to rearrange consumption between vari-
ous time periods. By suitably adapting the theory of consumer choice, we can examine the
factors that influence decisions to save or borrow.
Let’s confront this topic in the simplest possible way. Imagine a short-lived individual, Ms.
Cher Noble, whose lifetime spans two time periods, year 1 (this year) and year 2 (next year).
Ms. Noble’s earnings in year 1 (I1) are $10,000, but they will fall to $2,200 in year 2 (I2). The interest rate, r, at which she can borrow or lend is 10 percent per year. We assume that there is no inflation in the general price level so that $1 will purchase the same quantity of goods both
years. (If there is inflation, the earnings in each year can simply be expressed in dollars of con-
stant purchasing power, and the real rate of interest can be used instead of the nominal rate.)
This information allows us to plot Ms. Noble’s budget line. In Figure 5.9, consumption in
year 1 is measured on the vertical axis and consumption in year 2 (next year’s consumption)
on the horizontal axis. Any point in the diagram therefore represents a certain level of
Consumer Choice over Two Time Periods With an interest rate, r, of 10 percent and with earnings of $10,000 in year 1 and $2,200 in year 2, the budget line relating consumption in the two years is AZ with a slope of 1/(1 + r). The optimal point is W, with saving of I1C1 in year 1. In year 2, consumption exceeds that year’s income by I2C2, which is equal to the amount saved ($3,000) plus interest on this sum ($300).
Consumption in year 2
0
($7,000) C1
($10,000) I1
($12,000) A
W
N
Z
U2
U3
U1
I2 C2
Consumption in year 1
($2,200)($5,500) ($13,200)
1.1
1.0
Saving
Figure 5.9
The Consumer ’s Choice to Save or Borrow 125
C05.INDD 12:5:43:PM 08/06/2014 PAGE 125Trim Size: 203.2 mm X 254 mm
consumption in each year. The budget line indicates what combinations are available to the
consumer. Point N, for example, identifies the consumption mix where the individual’s entire earnings are spent in each year: a market basket containing $10,000 in consumption in year 1
(equal to year 1 earnings of $10,000) and $2,200 in consumption in year 2 (equal to year 2
earnings). Point N is sometimes called the endowment point, showing the consumption mix available to the individual if no saving or borrowing takes place. Alternatively, by saving or
borrowing, the consumer can choose a different market basket.
To identify another point on the budget line, suppose that Ms. Noble’s entire year 1 income
of $10,000 is saved. In this case consumption in year 1 is zero, but in year 2 she could con-
sume $13,200, equal to the sum saved the year before ($10,000), plus interest on that sum at a
10 percent rate ($1,000 in interest), plus earnings in year 2 ($2,200). Thus, point Z shows the horizontal intercept of the budget line; if consumption in year 1 is zero, Ms. Noble can consume
$13,200 in year 2. The vertical intercept, A, shows the maximum possible consumption in year 1. This maximum is achieved by borrowing as much as possible, limited by how much can be
repaid in year 2. That is, if $2,000 is borrowed, year 1 consumption can be $12,000 ($2,000
plus year 1 earnings). Ms. Noble can borrow $2,000 at a maximum because $2,000 plus 10 per-
cent interest, or $2,200, must be repaid the next year. This amount equals total year 2 earnings
with nothing left over for consumption. So the budget line’s vertical intercept is $12,000.
Points A, N, and Z represent three points on the consumer’s budget line. Connecting these points yields AZ as the entire budget line. If the consumer, Ms. Noble, chooses a point along the NZ portion of the line, she will be consuming less than earnings in year 1, or saving, and consuming more than earnings in year 2 (by an amount equal to the previous
year’s saving plus interest). Along the AN portion of the budget line, Ms. Noble is borrow- ing in year 1 and repaying the loan in year 2.
Notice how the budget line’s slope relates to the interest rate. In fact, the slope is equal
to 1/(1 + r), where r is the interest rate. Thus, if Ms. Noble reduces consumption by $1.00 in year 1 (saves $1.00), she can increase consumption in year 2 by $1.00 plus the interest of
$0.10, or by $1.10. With an interest rate of 10 percent the slope is equal to 1/(1 + 0.10), or 0.91 (rounded). This result tells us that the present cost of $1.00 consumed in year 2 is $0.91
in year 1, since $0.91 saved today grows to $1.00 a year later at a 10 percent interest rate—or,
conversely, to have $1.00 to spend in year 2, the consumer must save $0.91 in year 1.
Now let’s bring Ms. Noble’s preferences into the picture. Because consumption in both
years is desirable (more is preferred to less), the indifference curves have the usual shape.
The slope of an indifference curve at any point is the marginal rate of substitution between
consumption in year 1 and consumption in year 2, and it shows the willingness of Ms.
Noble to reduce consumption in year 1 to have greater consumption in year 2.
For the indifference curves shown in Figure 5.9, the consumer’s optimal choice is point W. Consumption in year 1 is $7,000, and consumption in year 2 is $5,500. Note that Ms. Noble
is saving some of her year 1 earnings, as indicated by the choice to consume less than her
income of $10,000 in year 1. The amount of saving in year 1 is the difference between
income and consumption in that year, which is shown by the distance I1C1, or $3,000. In year 2, the individual’s consumption is $3,300 greater than year 2 earnings; this sum is
equal to the amount saved, $3,000, plus interest on the saving.
Thus, the consumer’s optimal consumption point is once again characterized by a tan-
gency between an indifference curve and the budget line. The only novel feature here is that
the “commodities” consumed refer to consumption in different time periods, but that does
not change the substance of the analysis.
A Change in Endowment The budget line relevant for consumption choices over time depends on current and future
income as well as the interest rate. A change in one or more of these variables will alter
the budget line and affect the market basket chosen. Let’s examine how a change in year 2
endowment point the consumption mix available to the individual if no saving or borrowing takes place
126 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 126Trim Size: 203.2 mm X 254 mm
earnings will affect consumption and saving. Continuing with the example just discussed,
suppose that Ms. Noble expects year 2 earnings to be zero rather than $2,200.
A change in earnings in either year moves the endowment point in the graph. In Figure 5.10,
budget line AZ is reproduced from Figure 5.9; it shows the opportunities available when earn- ings are $10,000 in year 1 and $2,200 in year 2. When year 2 earnings fall to zero, the endow-
ment point moves from N to I1 on the vertical axis: if Ms. Noble’s consumption equals earnings in each year, consumption will be $10,000 in year 1 and nothing in year 2. The new budget
line is I1Z′, and it is parallel to AZ because the interest rate remains unchanged [both lines have a slope of 1/(1 + r)].
A reduction in future income will not alter the relative cost of future and present
consumption, but it will influence behavior through its income effect. If consumption
in each year is a normal good—which is almost certain to be true because consumption
is a very broadly defined good—then the shift in the budget line will lead to reduced
consumption in both years. Ms. Noble spreads the loss over both years, cutting back
on consumption in year 1 and year 2. The new optimal consumption point, W′, illus- trates this situation with consumption reduced from $7,000 to $6,000 in year 1 and from
$5,500 to $4,400 in year 2.
This analysis implies that a reduction in expected future earnings causes saving in year 1
to increase. Recall that saving is the difference between consumption and income. Before the income loss, saving was I1C1, or $3,000; after the loss, saving increases to I C1 1′, or $4,000. Current saving, therefore, doesn’t depend exclusively on current income (which is unchanged in this example); it is also affected by the expected level of future income.
So far we have been looking at a person who saves in year 1. Yet some people borrow in
the present and repay the loan later. Under different circumstances the individual shown to be
a saver in Figure 5.10 could become a borrower. For example, suppose that instead of year 1
income of $10,000 and year 2 income of $2,200 (point N), earnings are $2,000 in year 1 and $11,000 in year 2 (point Y). The budget line doesn’t change: only the endowment point changes, from N to Y on AZ. The optimal consumption point remains W, but to reach that point, the individual borrows $5,000 in year 1 and repays the loan plus interest in year 2.
An Income Change and Intertemporal Choice If year 2 income is zero instead of $2,200, the budget line shifts from AZ to I1Z′, a parallel shift. The result is an increase in saving in year 1, from I1C1 toI C1 1′. Consumption in both years falls, assuming that consumption in each year is a normal good.
Consumption in year 2
0
($6,000) C ′
($7,000) C1
1
($10,000) I1
A
W
W ′
Z ′ Z
U3
U0
N ($2,200; $10,000)
Y ($11,000; $2,000)
C ′2 C2
Consumption in year 1
($4,400)($5,500) ($11,000)
Figure 5.10
The Consumer ’s Choice to Save or Borrow 127
C05.INDD 12:5:43:PM 08/06/2014 PAGE 127Trim Size: 203.2 mm X 254 mm
This analysis shows how the pattern of earnings over time is likely to affect saving and
borrowing decisions. A relatively high present income but sharply reduced future income
(such as endowment point N) is typical of middle-aged persons approaching retirement, and we expect to see them save part of their current income. A low present income but
a higher expected future income (such as endowment point Y) is typical of students and young workers, and we often see such persons acquiring debt and consuming above their
present income.
APPLICATION 5.4
Our intertemporal choice model can easily be adapted to examine the important issue of how the Social Security program in the United States affects American workers sav- ing for retirement. To do so, we make our two time periods represent the years before and after retirement. Thus, we get the lifetime budget line AZ for Justin, a typical American worker, in Figure 5.11, which shows his options regarding consumption before and after retirement. His lifetime earn- ings are 0A, and these must finance his consumption during his working years as well as his retired years. With prefer- ences shown by U1, he chooses to consume C1 before retire- ment, saving AC1 of his lifetime earnings, and this provides a retirement level of consumption of C2.
Now introduce Social Security, a program which we will discuss in greater detail in a later chapter, but here we will simplify to two key parts: a tax on earnings and a pen- sion provided by the government during retirement. The tax of AA′ by itself shifts the budget line inward to A′Z′, but that is not the whole story because in return for paying the tax Justin is promised a pension when he retires. Assume he
Social Security and Saving
expects the pension to be CG, which is the amount he would have gotten if he saved the amount of the tax for himself. Then the expected government pension shifts the after-tax budget line outward to A′RZ, with A′R equal to the govern- ment provided pension, CG. Confronted with A′RZ, Justin chooses point W, the same point as before the introduction of Social Security.
It may look like Social Security has no effect, but a closer examination reveals an important result. Justin is now saving A′C1 to supplement the government pension, and since he was previously saving AC1, we see that Social Security has caused him to reduce his saving by AA′. In fact, he has reduced his saving by exactly the amount of the Social Security tax he pays. This need not always be the outcome; it is the result of our assumption that he expects the gov- ernment pension to be the same as he would have achieved by saving the amount of the tax. (You will find it instructive to go through the analysis for the cases when he expects the government pension to be larger or smaller than shown in our diagram.)
Social Security and Saving Social Security is composed of two key parts: a tax on earnings (AA′ in this diagram); and a pension provided by the government during retirement (CG). Social Security reduces saving from S(AC1) to S′(A′C1).
A
A ′
0 C2CG
C1
ZZ ′
R
Consumption in retirement
Consumption in year before retirement
S ′ W
U1
S
Figure 5.11
128 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 128Trim Size: 203.2 mm X 254 mm
Changes in the Interest Rate Will people save more at a higher interest rate? The answer to this question may not be the
yes response commonly expected. Let’s see why.
A higher interest rate changes the relative cost of present versus future consumption,
which is reflected in a change in the budget line’s slope. If the interest rate rises from 10 to 20
percent, the budget line’s slope, 1/(1 + r), changes from 1/(1 + 0.1) to 1/(1 + 0.2), so the new budget line has a slope of 1/1.2, or 0.83. Reducing consumption by $1.00 in year 1 (saving
$1.00) permits consumption of $1.20 more in year 2. Put somewhat differently, the present
cost of consuming $1.00 in year 2 is $0.83 in year 1, because $0.83 will grow to $1.00 in
one year at a 20 percent interest rate. Or, conversely, to have $1.00 to spend in year 2, the
consumer need save only $0.83 now at the higher interest rate. Thus, a higher interest rate
reduces the cost of future consumption in terms of the present sacrifice required.
Figure 5.12a shows the way the budget line changes. The initial budget line AZ once again reflects the 10 percent interest rate, and the endowment point is N. When the interest rate rises to 20 percent, the budget line rotates about point N and becomes A′Z′. Point N is also on the new budget line because the individual can still consume I1 in year 1 and I2 in year 2 if no borrowing or saving takes place. The budget line’s slope becomes flatter through point
N because an increase in the interest rate from 10 to 20 percent increases the cost of present versus future consumption. At the higher interest rate, that is, increasing present consumption
by $1.00 (through borrowing) now requires $1.20 to be paid back in year 2. And increasing
future consumption by $1.00 (through saving) now requires only $0.83 to be banked in year 1.
Incorporating the indifference curves of the individual, we can determine the preferred
market basket associated with the new budget line. For the indifference curves shown in
Figure 5.12a, the initial optimal levels of consumption are C1 and C2, with I1C1 saving in year 1. When the interest rate rises to 20 percent, the new optimal point W′ involves con- sumption of ′C2 in year 2 and ′C1 in year 1. A lower consumption in year 1 means an increase in saving from I1C1 to I C1 1′. Consequently, this individual will save more when the interest rate rises from 10 to 20 percent.
The Case of a Higher Interest Rate Leading to Less Saving The Figure 5.12a case, however, is not the only possible outcome. For example, if the indi-
vidual’s indifference curves are as shown in Figure 5.12b, an increase in the interest rate
from 10 to 20 percent leads to less saving. With the new budget line A′Z′, the optimal point is W′, involving more consumption in both years than at W. (Note that a higher interest rate allows a saver to increase consumption in the present and still consume more in the future.)
Because consumption increases in year 1, saving falls when the interest rate rises from 10
to 20 percent for an individual with indifference curves as shown in Figure 5.12b.
What factors determine whether saving rises or falls when the interest rate increases from
10 to 20 percent? Once again, they are the familiar income and substitution effects. (To avoid complicating the diagram, we do not show these effects explicitly in Figure 5.12 but,
instead, give a verbal explanation.) The substitution effect associated with a higher inter-
est rate results from the change in the relative cost of present versus future consumption.
In interpreting this analysis, it is important to under- stand that the government is not taking Justin’s tax dollars and saving them for him (which would leave total saving unchanged). Social Security is financed on a pay-as-you- go basis, which means that the government takes Justin’s tax dollars and transfers them to people already retired.
So when Justin reduces his saving, there is no offsetting increase in government saving. This analysis thus indicates that pay-as-you-go Social Security decreases national sav- ing by reducing the incentive workers have to save for their retirement. In a later chapter we will examine how this reduction in saving affects the overall economy.
The Consumer ’s Choice to Save or Borrow 129
C05.INDD 12:5:43:PM 08/06/2014 PAGE 129Trim Size: 203.2 mm X 254 mm
A higher interest rate reduces the cost of future consumption, which implies a substitution
effect that favors future consumption at the expense of present consumption. So the substi-
tution effect encourages future consumption instead of present consumption. In contrast,
the income effect associated with a higher interest rate enriches the saver, who is able to
attain a higher indifference curve. A higher real income enables the individual to consume
more in both periods, so the income effect favors increased consumption in both periods if
consumption in each period is a normal good.
Because both the substitution and income effects favor more consumption in year 2, it
will definitely increase. However, substitution and income effects for year 1 consumption
are in opposing directions, so the outcome depends on their relative sizes. If the income
effect is greater, year 1 consumption will rise, which implies that saving will fall.
Intuitively, we can see why some people might save less at a higher interest rate.
Think of a person who is saving for a specific good, like a cruise around the world. A
higher interest rate means the consumer can purchase the cruise without committing as
many dollars to present saving. If the cruise costs $6,000, a person would have to save
$5,454 at a 10 percent interest rate but only $5,000 at a 20 percent interest rate to pur-
chase the cruise one year later. For such a focused saver the income effect of a higher
interest rate (favoring greater current consumption and thus less saving) may outweigh
A Change in Interest Rates and Intertemporal Choice A change in the interest rate rotates the budget line at endowment point N. When the interest rate rises from 10 to 20 percent, the budget line rotates from AZ to A′Z′. Whether saving rises or falls depends on the magnitude of the income effect (encourages less saving provided that year 1 consumption is a normal good) relative to the substitution effect (encourages more saving). (a) If the substitution effect outweighs the income effect, saving rises with the interest rate increase. (b) If the income effect outweighs the substitution effect, saving falls with the interest rate increase.
1 1
0.83
U3
W W ′
U2
0.91
A
A ′
I1
C1 C ′1
0 I2 C2 C′2 Z ′Z
(a)
Con- sumption in year 2
Con- sumption in year 1
SS ′
A
A ′
I1
C1 C′1
0 I2 C2 C′2 Z ′Z
(b)
Con- sumption in year 2
Con- sumption in year 1
S
U3W
N
N
W ′
U2
S ′
Figure 5.12
130 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 130Trim Size: 203.2 mm X 254 mm
the substitution effect (favoring greater future consumption through more saving). Saving
for such an individual thus may decrease as the interest rate rises.
5.6 Investor Choice The theory of consumer choice not only helps to illuminate the decision to save or borrow
but also can be applied to explain what types of financial assets an individual intent on sav-
ing for the future should purchase, or invest in. That is, should such an individual invest in
stocks, U.S. Treasury bills, gold, cattle futures or the local bank? At first glance, investing
in stocks would appear to be quite desirable given the returns generated historically by such
financial assets. For example, a portfolio composed of U.S. common stocks has averaged
an annual real rate of return of roughly 6 percent since 1802.8 (The real rate of return is the
nominal rate of return less the inflation rate.) By contrast, short-term U.S. Treasury bills
have provided an average annual real rate of return of only 0.6 percent.
The historical data on annual rates of return also indicate that a dollar invested in U.S.
common stocks in 1802 would have grown, in real terms, to nearly $800,000 today—even
when accounting for the Great Depression and the most recent significant economic down-
turn. A dollar invested in short-term U.S. Treasury bills (“T-bills” for short) in 1802 would
have grown to only about $300 after adjusting for inflation.
Given the historical data on the rates of return of various financial assets, a natural ques-
tion arises. Specifically, why would a rational investor ever choose to purchase T-bills as
opposed to stocks? It would seem logical to place all one’s financial eggs in the asset basket
offering the highest return.
The reason investors do not typically allocate all of their portfolios to the asset with
the highest rate of return involves the risk associated with such an action. If higher rates
of return are associated with greater risk and investors are averse to risk (that is, risk is a
“bad”), a trade-off exists between return and risk. While it may be true that in retrospect stocks display higher returns than T-bills, a person contemplating purchasing stocks does
not know beforehand what next year’s return will be. Stocks may average a 6 percent annual real rate of return. The average, however, may be
arrived at by a return of −6 percent in one year and 18 percent in another, with the dramatic and unpredictable gyrations averaging out to 6 percent over a two-year period. By contrast,
T-bills have a lower average return but may be a very sure thing from period to period.
There is very little chance, that is, that the U.S. government will default on its borrowing
obligations by not paying the promised amount to purchasers of T-bills that mature (must
be paid off) in a fairly short time period such as three months. An average annual real rate
of return on T-bills of 0.6 percent thus may be arrived at by a constant and more predictable
return of 0.6 percent in year 1 and 0.6 percent in year 2. Indeed, the historical volatility of
returns has been significantly lower for T-bills than for stocks.
Figure 5.13 displays what an investor’s indifference map looks like if higher expected
returns are associated with greater volatility or risk, and risk is a bad in the eyes of the investor.
As discussed in Chapter 3, the indifference curves are upward sloping if the expected return on
an asset is a good while the risk associated with the asset’s return is a bad. Holding constant the
expected return on an asset at return1, a rise in the risk associated with the return on the asset, such as a move from point A to point B, places the risk-averse investor on a lower indifference curve (U1 versus U2). Furthermore, as the risk of an asset’s return rises from risk1 to risk2, the expected return on the asset must rise from return1 to return2, as at point C, to compensate for the added risk and leave the investor’s utility level unchanged at U2. Greater levels of well- being are shown by indifference curves above and to the northwest: U2 is preferred to U1.
8Jeremy J. Siegel, Stocks for the Long Run 4th ed., (New York: McGraw-Hill, 2014).
Investor Choice 131
C05.INDD 12:5:43:PM 08/06/2014 PAGE 131Trim Size: 203.2 mm X 254 mm
The Return–Risk Trade-off Since volatility or risk is a bad and expected return is a good, the investor’s indifference curves are upward sloping.
Return
Risk0 risk1
return2
return1 A
C
U2
U1
U0
B
risk2
Figure 5.13
APPLICATION 5.5
Studies indicate that small, entrepreneurial ventures created more than 90 percent of the net new jobs in the United States in the past few decades.9 Concurrently, interest in entrepreneurship has grown. More than 1,000
Entrepreneurs and Their Risk–Return Preferences
colleges and universities now offer courses in entrepreneur- ship, and a recent Gallup poll indicates that 7 out of 10 high school students in the United States want to start and own their own businesses in their adult years.
Entrepreneurs are characterized by their willingness to take risks and their ability to pursue and seize opportuni- ties to create value, notwithstanding apparent resource constraints, through a new or existing company. Figure 5.14
9Mickie P. Slaughter, “Entrepreneurship: Economic Impact and Public Policy Implications,” Kauffman Center for Entrepreneurial Leadership, March 1996.
Return
Risk
Nonentrepreneur
0 risk 1
return2
return1 Y
Z
U 3 NE
U 3 E
U 2 E
U 1 E
U 2 NE
U 1 NE
NE NE risk 2
(a) (b)
Return
Risk
Entrepreneur
0
return2
return1 B
C
risk 1 E E
risk 2
Differences in Individuals’ Risk–Return Preferences An entrepreneur (a) has flatter indifference curves than a nonentrepreneur (b) because the entrepreneur is less risk averse.
Figure 5.14
132 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 132Trim Size: 203.2 mm X 254 mm
Investor Preferences toward Risk (a) A risk-averse investor’s total utility curve has an upward, but diminishing slope. For every return between the equally likely returns of 2 and 8 percent, the height of the total utility curve exceeds the height of the expected utility chord, AZ. (b) A risk-neutral investor has a total utility curve that is an upward-sloping straight line. The total utility curve and the expected utility chord have the same height at every return between 2 and 8 percent. (c) A risk-loving investor’s total utility curve has an upward and increasing slope. The height of the expected utility chord exceeds the height of the total utility curve over the relevant range of returns.
2
100
0
200 225
300
Total utility, U
Total utility, U
Return (%) per dollar invested in IBM stock, ri
Return (%) per dollar invested in IBM stock, ri
5
(a) (b) (c)
8 2
20
0
50
80
5 8
U
2
25
175 150
325
5 8
U
Z Z
A A
U
Z
A
Total utility, U
Return (%) per dollar invested in IBM stock, ri
0
Figure 5.15
Investor Preferences toward Risk: Risk Aversion We have so far mentioned that, other things being equal, risk-averse investors prefer a sure
thing—an asset promising the same return at lower volatility. As the volatility of an asset’s
return increases, risk-averse investors must be compensated with a greater expected
return to remain equally well off (that is, to remain on the same indifference curve, as
shown in Figure 5.13). In addition to looking at an indifference map, we can also examine
what is meant by “risk aversion” by graphing the total utility curve of an investor as a func-
tion of the expected return per dollar invested in the stock of a company such as IBM.
As depicted in Figure 5.15a, the total utility curve has a positive but diminishing slope
for risk-averse investors. To see why this is the case, let’s investigate the properties of such
a total utility curve. The height of the curve at each possible return indicates the utility of
that return to the investor. For example, as drawn, the utility of a return of 2 percent equals
100 while the utility of a return of 8 percent equals 300.
depicts the observed differences in risk-return prefer- ences for a representative entrepreneur and nonentre- preneur. The indifference curves are flatter in the case of the entrepreneur because the entrepreneur is willing to take on more risk for a given increase in the expected return (for example, from return1 to return2). That is, the
difference between risk1E and risk2E (as the entrepre- neur moves between points B and C along indifference curve U1E in Figure 5.14a) is greater than the difference between risk1NE and risk2NE (as the nonentrepreneur moves between points Y and Z along indifference curve U1NE in Figure 5.14b).
Investor Choice 133
C05.INDD 12:5:43:PM 08/06/2014 PAGE 133Trim Size: 203.2 mm X 254 mm
Suppose that, per dollar invested, IBM stock will provide either a return of 2 or 8 percent
and that the probability of either outcome is 0.5. The expected return on IBM stock can be designated as E(ri) where E is shorthand for “expected,” r represents return, and the subscript i reflects the fact that the return can take on different possible values. E(ri) is the average return that the investor can anticipate receiving from IBM stock; it is equal to the summed
value of each possible rate of return weighted by its probability: E(ri) = 2(0.5) + 8(0.5) = 5. The expected utility from holding IBM stock, E[U(ri)], is the average utility the investor
can anticipate receiving from among the different possible returns on IBM stock; it is equal to
the summed value of each possible utility weighted by its probability. The expected utility from
holding IBM stock in our example is consequently 200, E[U(ri)] = 100(0.5) + 300(0.5) = 200. Graphically, the expected utility from investing in IBM stock with an equally likely
return of 2 and 8 percent can be determined by constructing a dashed chord, AZ, that con- nects the heights of the total utility curve at returns of 2 and 8 percent. The height of chord
AZ at the expected return of IBM stock (5 percent in our example) indicates the average or expected utility from investing in IBM stock, E[U(ri)]. In Figure 5.15a, total utility rises from 100 to 300 between returns of 2 and 8 percent. Thus, to achieve a 200-unit increase
in total utility (300 − 100 = 200) as the return increases by 6 percent (8 − 2 = 6), the aver- age utility when we are at a return of 5 percent—halfway between a return of 2 and 8 per-
cent—must be half the distance between 100 and 300 units of total utility. As a result, the
expected utility of holding IBM stock is given by the height of chord AZ at the expected return of 5 percent: E[U(ri)] = 200 in the case of Figure 5.15a.
Now compare the height of the risk-averse investor’s total utility curve with the height
of chord AZ at a return of 5 percent. The height of the total utility curve at 5 percent in Figure 5.15a (225) indicates the utility the investor would derive were IBM stock to pro-
vide a sure return (with no variance or risk) of 5 percent per dollar invested. Because the
height of the total utility curve (225) exceeds the height of the straight-line segment AZ (200) at a return of 5 percent, this shows that the risk-averse investor gets more utility from
a sure return of 5 percent (and no variance) than from investing in IBM stock that generates
the same average return of 5 percent but through varying between returns of 2 and 8 per-
cent. In general, an individual is deemed to be risk averse if he or she prefers a certain return to an uncertain prospect generating the same expected return.
We can define risk aversion more formally by using some symbols for the heights of the
total utility curve and the expected utility chord AZ. While the return on IBM’s stock var- ies in our example, the average or expected return, E(ri), is 5 percent and is a fixed number. The height of the total utility curve at 5 percent thus can be written as U[E(ri)]—the utility that would be generated by a sure return of 5 percent. Since a risk-averse investor prefers a
sure return to an uncertain prospect generating the same expected return, this implies that
the height of the total utility curve exceeds the height of the expected utility chord AZ at the expected return of the risky investment:
U E r E U ri i[ ( )] > [ ( )].
The preceding inequality will be true whenever an investor’s total utility curve has a posi-
tive, but diminishing slope. Since the slope of the total utility curve is the marginal utility
associated with various returns (see Chapter 3), risk aversion is present whenever the mar-
ginal utility of payoffs to an individual declines as a function of the size of those payoffs.
With diminishing marginal utility, an investor will prefer a sure payoff of 5 percent per
dollar invested to the prospect of earning 8 percent (3 percent above the average) half of the
time and earning 2 percent (3 percent below the average) the other half. The marginal util-
ity associated with each unit increase in the return between 5 and 8 percent will not be as great as the marginal utility associated with each unit decrease in the return between 5 and
2 percent if diminishing marginal utility applies to an asset’s payoffs.
Among investors making choices involving uncertain payoffs, aversion appears to be
the most common attitude toward risk, especially when the size of the payoffs involved is
expected return the summed value of each possible rate of return weighted by its probability
expected utility the summed value of each possible utility weighted by its probability
risk averse a state of preferring a certain return to an uncertain prospect that generates the same expected return
134 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 134Trim Size: 203.2 mm X 254 mm
significant. For example, consider the risk of losing one’s home to a fire. Most families would
rather pay a $1,000 annual premium for an insurance policy covering loss due to fire than to
face the prospect that in any given year the probability of losing their home valued at $100,000
to fire is 0.01 while the probability of there being no fire is 0.99. Most families, that is, would
prefer a sure loss of $1,000 to an insurance company in the form of a premium than to take the
gamble of not having fire insurance, even though the latter gamble involves the same expected
return, E(ri), of −$1,000 per year, E(ri) = (−100,000)(0.01) + 0(0.99) = −1,000.
Investor Preferences toward Risk: Risk Neutral and Risk Loving Although risk aversion may be the most common attitude toward risk, it is not the only
possible attitude. Indeed, there are two other views people can take toward risk. They can
be neutral toward it or they can seek it out. These two alternatives are depicted by Figures
5.15b and 5.15c, respectively.
In the case of a risk-neutral investor (Figure 5.15b), the total utility curve is an upward-
sloping straight line. For the same example involving IBM stock whose return is equally
likely to be 2 or 8 percent, a risk-neutral investor will be indifferent between a sure return
of 5 percent and investing in the risky IBM stock generating an expected return of 5 per-
cent. Graphically, the height of the investor’s total utility curve (50) equals the height of the
expected utility chord, AZ, at the expected return of 5 percent. In general, an individual who is risk neutral gets the same utility from a certain return, U[E(ri)], as from an uncer- tain prospect generating the same expected return, E[U(ri)]. Formally, that is, the following equation holds for a risk-neutral investor:
U E r E U ri i[ ( )] = [ ( )].
For a risk-loving investor (Figure 5.15c), the total utility curve has an upward, but ever- increasing slope. In our IBM example, a risk-lover will prefer investing in the risky IBM
stock generating an expected return of 5 percent over a sure return of 5 percent. Graphi-
cally, the height of the investor’s total utility curve (150) is less than the height (175) of the
expected utility chord, AZ, at the expected return of 5 percent. In general, an individual who is risk loving gets less utility from a certain return, U[E(ri)], than from an uncertain prospect generating the same expected return, E[U(ri)]. Formally, the following equation holds for a risk lover:
U E r E U ri i[ ( )] < [ ( )].
An example of risk-loving behavior involves the purchase of state lottery tickets even when
one knows that the return on each dollar invested in such tickets is −50 percent. That is, the typical dollar invested in a state lottery has an expected payback of 50 cents. The return of
−50 percent is much lower than the sure return of 0 percent one could obtain by keeping the money spent on lottery tickets in one’s pockets instead.10 While such risk-loving behavior
certainly does occur, it tends to diminish as the gambling stakes increase. For example,
more individuals at the casinos in Las Vegas play the $0.25 slot machines than the $100
blackjack tables.
risk neutral a state of deriving the same utility from a certain return as from an uncertain prospect generating the same expected return
risk loving a state of deriving less utility from a certain return than from an uncertain prospect generating the same expected return
10State lotteries make most of their money, and impose such a highly negative return on investors, by stretching out payments to lottery winners. For example, you may win $10 million in a state lottery draw- ing but that winning will be paid to you in equal installments over a period of 20 years. As we saw in Section 5.5, any dollars earned in the future are equal to less than their face value in terms of today’s dollars. By contrast, the major casinos in Las Vegas pay gamblers an average return of −2 percent per dollar invested. That is, the typical dollar bet in a Las Vegas casino has an expected payoff of 98 cents. The less negative return in Las Vegas is due to the fact that casinos pay immediately when an “investor” gets lucky at a slot machine or blackjack table.
Investor Choice 135
C05.INDD 12:5:43:PM 08/06/2014 PAGE 135Trim Size: 203.2 mm X 254 mm
Minimizing Exposure to Risk Living in a world where risk abounds, how can risk-averse individuals minimize their
exposure to it? One method already mentioned is insurance. By paying a premium in return for the promise that the insurer will provide compensation for losses due to an acci-
dent, illness, fire, and so on, one effectively exchanges a gamble for a sure return. And risk-
averse individuals would prefer a sure loss of $1,000 per year in the form of a premium
paid to an insurance company over being uninsured and confronting the gamble of a 0.01
probability they will suffer a $100,000 loss versus the 0.99 probability they will not.
What is the maximum price that a risk-averse individual would be willing to pay for an
insurance premium against a 0.01 probability of suffering a $100,000 loss? We can show
that it is larger than $1,000, the expected value of the loss: $100,000(0.01) + $0(0.99) = $1,000. Figure 5.16 displays the total utility curve for an individual who owns a $100,000
home and faces the chance that the home will burn down with probability 0.01 in any given
year. The expected utility from such a prospect is given by the height (U*) of the expected utility chord, AZ, at an income of $99,000. (The expected payoff or income is $99,000 since with probability 0.99 there will be no fire and the individual’s $100,000 home will be undam-
aged while with probability 0.01 the home will be reduced to ash and a value of $0.) Note
that the same utility of U* would also be provided by a certain income of $98,300 since the height of the total utility curve equals U* at an income of $98,300. Thus, the individual whose total utility curve is depicted in Figure 5.16 would be willing to pay up to $1,700
per year to insure against a fire loss. The homeowner is indifferent between (a) paying
$1,700 (PP″) per year in exchange for a certain income of $98,300 (the insurance company will fully compensate the individual for the $100,000 loss in case of a fire), and (b) the
prospect of remaining uninsured and facing an expected, but not certain, loss of $1,000
(P′P″) and an income of $99,000. Will an insurance company be willing to supply coverage against fire loss to the home-
owner for less than the maximum the homeowner is willing to pay for such coverage? The
answer, in all likelihood, is yes. This is because insurance companies typically insure a
large pool of individuals with respect to the same risk. While the probability that any one
insurance an arrangement by which the consumer pays a premium in return for the promise that the insurer will provide compensation for losses due to an accident, illness, fire, and so on
APPLICATION 5.6
Many airline ticket counters, department of motor vehi- cles offices, federal customs checkpoints, banks, postal branches, college financial aid departments, and fast-food restaurants have a single line feeding to multiple clerks as opposed to separate lines for each clerk. This is the case even though a single-line system is unlikely to alter the average time a customer must spend waiting to see a clerk. After all, the line is not likely to affect either the total number of clerks or the number of customers waiting to see them.
A single line, however, does reduce the variance of a customer’s waiting time since a customer is less likely to get into a “slow” or “fast” line. Holding constant the expected waiting time, the reduction in the variance of the wait- ing time will be appealing to customers if they are risk averse to spending time in lines. If a sure wait of 5 minutes
Risk Aversion While Standing in Line
is preferred to the prospect of a 0.5 probability each of a 2-minute and an 8-minute wait, customers will be hap- pier under the single-line system. According to operations researchers, at least part of the reason customers prefer less variance when waiting in line is their sense of social jus- tice. Specifically, relative to an expected wait of 5 minutes, experiencing only a 2-minute wait because one is fortunate enough to get into a “fast” line appears to add less to total utility than experiencing an 8-minute wait in a “slow” line subtracts from total utility, since the longer wait brings with it the added aggravation of seeing more recent arrivals who get into a fast line receive service first.11
11Richard C. Larson, “Perspectives on Queues: Social Justice and the Psychology of Queuing,” Operations Research, 35 No. 6 (November– December 1987), pp. 895–905.
136 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 136Trim Size: 203.2 mm X 254 mm
home will burn down in a given year may be subject to significant variability and therefore
be difficult to predict, it is possible to predict much more precisely the total number of simi-
larly situated homes in a large pool that will burn down in a given year. If the probability
of any one house burning down is 0.01, for example, then an insurance company can be
reasonably confident that, out of a pool of 10 million homes it insures, 100,000 will burn
down in any given year. An apt analogy involves flipping coins. The deviation in the actual
number of heads from the theoretically predicted probability will be much greater if a coin
is flipped only once versus if the coin is flipped a large number of times.
By pooling a large enough number of similar fire risks, an insurance company will
be able to predict reasonably well (that is, with very low variance) the total expected
payments it will have to make on the fire insurance policies that it writes—$1,000 per
home insured in our example. At the bare minimum, therefore, the insurance company
should be willing to supply an insurance policy for $1,000 (plus any cost associated with
administering the policy)—distance P′P″ in Figure 5.16. And since the maximum the risk-averse homeowner is willing to pay for such insurance (PP″) exceeds the minimum the insurance company needs to be paid (P′P″), an opportunity exists for mutually benefi- cial exchange.
In addition to insurance, another method of minimizing one’s exposure to risk is through
diversifying one’s asset holdings. Diversification involves investing a given amount of resources in numerous independent projects instead of one single project. What may be an
unacceptably large risk can thus be translated into more palatable small ones. The larger the
number of independent projects, the more predictable the expected return on investment (as
in our earlier coin-flipping example) and the lower the overall risk.
Organizational structures that promote the risk-spreading advantages attendant to
diversification include partnerships and corporations whose stock is publicly traded. Sup-
pose, for example, that a promising start-up company requires $100,000 in funds to get
off the ground but there is some probability that the venture will not succeed and that any
invested funds will be lost. If you are risk averse, you are likely to find investing in the
venture more palatable if you can do it jointly with 10 other equal partners each chipping
in $10,000 than if you have to underwrite the entire venture by yourself.
diversification investing a given amount of resources in numerous independent projects instead of a single project in order to minimize exposure to risk
Pricing Insurance In our example, a risk-averse homeowner is willing to pay up to $1,700 (PP") per year to insure against a 0.01 probability of losing a $100,000 home to fire. By pooling a large enough number of equal risks, the minimum the insurance company will be willing to write an insurance policy for is $1,000 (P′P") plus any costs associated with administering the policy.
$98,300
$99,000
$100,000 Income0
Total utility, U
U
U* P″ P ′
P
Z
A
Figure 5.16
C05.INDD 12:5:43:PM 08/06/2014 PAGE 137Trim Size: 203.2 mm X 254 mm
Summary 137
As with business partnerships, public trading of corporate stocks allows investors to
diversify their asset portfolio by buying up a small number of shares of stock in a large
number of companies. By not putting their financial eggs in a single asset basket, inves-
tors are able to mitigate their exposure to risk. This is because certain risks uniquely affect
a single stock or a small group of stocks. And the investor’s overall vulnerability to such
unique risks will be lower the more diffuse a portfolio’s holdings.
For example, take the case of the unexpected death in 1983 of Henry “Scoop” Jack-
son, a senator from the state of Washington with considerable clout on defense-spending
issues. Jackson’s death had no impact on most stocks but produced a decline in the price of
Boeing Aircraft Company stock—Senator Jackson had been known as the “Senator from
Boeing” for his lobbying efforts on behalf of the Washington-state-based firm.12 Jackson’s
death also led to a rise in Lockheed Corporation’s stock price since Lockheed was the larg-
est defense manufacturer in Georgia, home to the senator expected to replace Jackson as
ranking member of the Senate Armed Services Committee, Sam Nunn. An investor holding
only Boeing or Lockheed stock at the time of Jackson’s death would have confronted much greater risk (more volatility in asset returns) than would an investor with a more diversified
portfolio. A well-diversified portfolio would include (a) many stocks that were not affected
at all by Jackson’s death, and (b) other stocks whose movement in the wake of Jackson’s
passing would work to cancel each other out (such as Boeing and Lockheed), thereby low-
ering the volatility of the overall portfolio.
SUMMARY
Consumer choice theory can be applied to a wide
range of interesting and important policy questions.
An excise subsidy is a form of subsidy in which the
government pays part of a good’s per-unit price and
allows the consumer to buy as many units as desired
at the subsidized price. Subsidies can also be made in
the form of cash, as a lump-sum transfer. Consumer
choice theory helps us discover that based on their
own preferences, consumers will be better off if they
receive cash.
In general, subsidies cannot harm the recipients of
the subsidies, at least in the case where they do not have
to pay taxes to finance the subsidies. In the case of the
ObamaCare legislation signed into law in 2010, how-
ever, subsidies can harm recipients to the extent that
there is a mandate to purchase the subsidy. The subsidies
associated with ObamaCare, moreover, provide disin-
centives to work and are associated with significant dead-
weight losses once the costs, to taxpayers, of funding
the subsidies are taken into account.
Public schools offer another common form of sub-
sidy in which the government makes a certain quantity
of a good available at no cost, or below the market price.
Voucher programs allow parents to purchase education
(with vouchers) at any school they choose. A variety of
possible consequences of a school voucher program can
be identified with the tools of consumer choice theory.
When consumers directly bear the cost of such
services as trash collection instead of paying a fixed
annual fee, they have an incentive to cut down on the
amount of the service they use, eliminating what was
in effect a subsidy for heavy users of the service.
A decision to save (or to borrow) is a decision to rear-
range consumption between various time periods. Adap-
tations of consumer choice theory allow us to examine
this decision and to see how it is affected by changes in
earnings and in the interest rate.
Consumer choice theory can help explain what types
of financial assets an individual who is saving for the
future should buy—stocks, Treasury bills, gold, and so
on. Indifference curves demonstrate the tradeoff between
expected risk and return. Total utility curves shed light
on the way in which insurance operates to reduce indi-
viduals’ risk.
12Brian E. Roberts, “A Dead Senator Tells No Lies: Seniority and the Distribution of Federal Benefits,” American Journal of Political Science, 34, No. 1 (February 1990), pp. 31–58.
138 Using Consumer Choice Theory
C05.INDD 12:5:43:PM 08/06/2014 PAGE 138Trim Size: 203.2 mm X 254 mm
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
5.1 Tony currently goes to the dentist twice a year at $150 per visit. His employer subsidizes Tony’s purchases by paying
half the cost of each dental visit in excess of one visit per year.
(No subsidy is granted on the first visit.) Show how this affects
Tony’s budget line and his purchases of dental services. Iden-
tify in a diagram how much Tony spends on dental work and
how much his employer spends in subsidizing his consump-
tion.
*5.2 The government has $100 per month with which to subsi- dize waif-like supermodel Kate Moss’s food consumption. It is
considering two alternatives: giving Kate $100 worth of food
stamps per month or paying part of the price of food, thereby
lowering the price to Kate. In the case of the excise subsidy,
the price is lowered to the point where the total cost of the
subsidy is $100 per month. Which type of subsidy will ben-
efit Kate more? Under which type of subsidy will she consume
more food?
5.3 With an excise tax of $0.20 per six-pack, Kristine pur- chases three six-packs of Diet Coke per day. If the govern-
ment eliminates the excise tax and, instead, requires Kristine
to pay $0.60 per day as a lump-sum tax (the daily tax liabil-
ity is $0.60, regardless of how much Diet Coke Kristine con-
sumes), how will Kristine’s consumption pattern and welfare
be affected?
5.4 How can the government reduce a consumer’s consump- tion of schooling by providing schooling at no cost? Explain in
words and show in a diagram.
5.5 We assumed in the text that families receive the public school subsidy at no direct cost. Yet families pay property (and
other) taxes that are used to finance schools. Does this fact
invalidate the analysis in the text?
5.6 At the “all you can eat” brunch buffet offered by Chez Paul’s Cajun Restaurant, consumers pay a price of $15 and
then can consume all they want. Show a consumer’s optimal
consumption point with a budget line and indifference curve.
Explain the shape of both curves.
5.7 Based on the assumptions of Figure 5.8, is it possible for a person to benefit from shifting from the fixed annual fee to
the price-per-unit system if that person is initially (under the
annual fee arrangement) consuming more than 2,000 units of
trash disposal? Justify your answer with a diagram.
5.8 In terms of a graph similar to Figure 5.2, explain why holi- day gift-giving may generate a deadweight loss. Specifically,
suppose that the recipients of gifts estimate that they would
have been willing to pay only 75 percent of the estimated
amount of money spent by the givers of the holiday gifts.
5.9 Draw the budget line for Beth, a consumer who must pay a higher interest rate to borrow than the interest rate she receives
on her savings.
5.10 In a two-year period, Anna has earnings of $6 million this year (year 1) and earnings of $10 million next year (year 2).
She can borrow or lend at an interest rate of 25 percent. Sup-
pose that Anna decides to borrow $1 million in year 1. Show
Anna’s optimal consumption point in a diagram. Identify the
amount borrowed in year 1, the amount repaid in year 2, con-
sumption in year 1, and consumption in year 2.
*5.11 From the position described in the preceding question, Anna’s earnings in year 1 fall to $4 million due to an unfavor-
able tax court ruling. How will this change in income affect the
amount she borrows, her consumption in year 1, and her con-
sumption in year 2? If her earnings fall by $2 million in year 1,
will her year 1 consumption fall by the same amount? Explain
your answer.
5.12 “A higher interest rate lowers the present cost of future consumption.” “A higher interest rate raises the future cost of
present consumption.” Use an example to show that both state-
ments are correct.
5.13 “For a person who saves in year 1, a higher interest rate will increase consumption in year 2 but may either increase
or reduce the amount saved in year 1.” Explain this statement,
using the concepts of income and substitution effects. Use a
diagram like Figure 5.12, and separate the income and substitu-
tion effects graphically.
5.14 What would an indifference map (with expected return on the vertical axis and risk on the horizontal axis) look like for a
risk-neutral investor? For a risk-loving investor?
5.15 Even though they enter graduate school at roughly the same age, law students typically own nicer cars than do Ph.D.
students. Assuming that there is no difference across the two
groups in terms of preferences and current incomes, explain
this phenomenon. Rely on the intertemporal choice model
developed in Section 5.5 for your explanation.
5.16 What is the maximum price that a risk-neutral individual would be willing to pay for an insurance premium against a
0.01 probability of suffering a $100,000 loss?
5.17 Explain why the delays associated with surfing the Web are greater because users are charged a flat monthly fee to
access the Web instead of per-unit use charges. Suppose that
the per-unit use charge was set at a level such that the average
Internet user ends up paying the same total amount if her con-
sumption remains unchanged subsequent to the per-unit charge
system being implemented. What will be the effect on total
units of Internet usage if the per-unit charge system is adopted?
Will the average user be better off? Explain. Will all users be
better off? Explain.
C05.INDD 12:5:43:PM 08/06/2014 PAGE 139Trim Size: 203.2 mm X 254 mm
Review Quest ions and Problems 139
5.18 Rocker Bruce Springsteen insured his vocal cords for $6 million. Dancer Fred Astaire insured his legs for $75,000 each.
And, in 1999, singer and actress Jennifer Lopez reportedly
took out a $1 billion insurance policy on her entire body. Why
might stars such as these find it advantageous to take out these
insurance policies? Why might insurance companies be willing
to offer the policies?
5.19 Prior to boarding a flight‚ a business executive opts not to buy a flight insurance policy sold from a booth she passes at
the airport terminal on the way to her departure gate. Does this
indicate that the executive is either risk neutral or risk loving?
Explain.
5.20 Americans spend almost $700 billion per year on legal- ized gambling—roughly $200 billion more than they spend on
groceries. Most Americans‚ however‚ also show evidence of
risk aversion in other aspects of their behavior such as buying
automobile‚ home‚ and life insurance. Why might Americans
spend so much on legalized gambling if they also are typically
risk averse?
5.21 If the mandate associated with purchasing health insur- ance by ObamaCare is successfully challenged in the courts by
state attorney generals, can there still be a case where the sub-
sidy would harm its recipients (provided that recipients do not
have to pay taxes to finance the subsidy)?
5.22 A key objective of health care reform was to “bend the cost curve” by finding a way to reduce demand for health care.
As analyzed in Section 5.2, to what extent will ObamaCare
succeed in achieving this objective?
140
C06.INDD 11:8:52:AM 08/06/2014 PAGE 140Trim Size: 203.2 mm X 254 mm
Exchange, Efficiency, and Prices6CHAPTER
Learning Objectives
Understand why voluntary exchange is mutually beneficial. Explain what economists mean by efficiency in exchange and the benefits associated with the promotion of such efficiency. Discuss how competitive markets promote efficient distribution of goods between consumers. Explore the extent to which price and nonprice mechanisms for rationing goods across consumers serve to promote efficiency.
Memorable Quote “By virtue of exchange, one man’s prosperity is beneficial to all others.”
—Frederic Bastiat, fifteenth-century French economist, statesman, and author
In Chapters 3 through 5 we concentrated on the way a typical consumer reacts to changes in his or her budget line. Higher or lower prices or incomes, subsidies, and taxes produce
generally predictable responses. The consequences of a consumer’s choices for partici-
pants on the selling side of markets have so far been ignored. Of course, the budget line
itself, indicating relative prices, reflects the willingness of others to trade with the consumer
on specified terms, but now we will emphasize more explicitly that a consumer’s market
choices involve exchanges between the consumer and other people.
To investigate the essential two-sidedness of market transactions, we examine pure
exchange. At the outset, our analysis will focus on two consumers who start with specified
quantities of two goods and engage in barter. This model may seem remote from real-world
behavior, and to some extent, it is. But the intention here is to focus on the nature and con-
sequences of voluntary exchanges between people, and the pure exchange model provides
the simplest means possible.
Most economic activity occurs through a series of voluntary exchanges. U.S. consum-
ers buy Samsung cell phones with some of their dollars. The dollars spent on Samsung cell
phones are eventually used by the various Samsung stakeholders (workers, management,
stockholders, and so on) to buy consumer goods of interest to them. Indirectly, therefore,
Samsung stakeholders exchange consumer goods for consumer goods. General Electric
workers exchange their labor for money and then exchange money for various consumer
goods, so indirectly they exchange labor for consumer goods. A model that permits us to
see why voluntary exchanges occur and what their consequences are thus turns out to be
quite useful.
Two-Person Exchange 141
C06.INDD 11:8:52:AM 08/06/2014 PAGE 141Trim Size: 203.2 mm X 254 mm
The model presented in this chapter is useful for two fundamental reasons. First, it
allows us to demonstrate one of the most important principles of economics: that voluntary
exchange, or trade, is mutually beneficial to the parties involved in the transaction. Although
the principle that such exchange represents a “win–win” for the parties involved is often
questioned, especially in the public policymaking arena, it is one of the most critical lessons
to be learned from the study of economics.
Second, the model developed in this chapter allows us to introduce the concept of
economic efficiency. A central tenet of economics, efficiency in exchange means that goods are distributed across consumers such that no one consumer can be made better off without
hurting another consumer. We will show why pursuing such distributions of goods is a
desirable objective and how competitive markets promote efficiency in exchange.
6.1 Two-Person Exchange People engage in exchanges, or trades, because they expect to benefit. When an exchange
is voluntary, with parties in agreement on the terms of the trade, the strong presumption is
that both benefit. Such a presumption follows from the fact that each party had the option
of refusing to trade but instead chose to engage in the exchange. If Jean-Claude van
Damme rents a tearjerker movie from Netflix, his action must mean he prefers the movie
rental to the money he exchanges for it. Also, the rental must mean Netflix prefers the
money to not renting the movie. Both parties benefit from the exchange.
The fundamental point can be stated simply: voluntary exchange is mutually beneficial. The truth of this basic economic proposition may seem obvious, but it is widely doubted.
For example, many people assume that the prosperity of successful businesspeople must
have come at the expense of their customers or workers. Economic activity, however, is
not like a sports contest in which, if there are winners, there must be losers. The voluntary
exchanges through which economic activity is organized can generate win–win outcomes.
There are, of course, some qualifications to the proposition that voluntary exchange
is mutually beneficial. First, it presupposes that fraud has not taken place. If van Damme
pays for his video rental with a bad check, Netflix will be worse off. Second, the benefit
achieved refers to the expectations of the parties at the time of the transaction. Jean-Claude
van Damme may rent a tear-jerker but after watching it decide he would have been better
off keeping the money. Nonetheless, at the time he rented the movie, he must have believed
the opportunity to view it was worth more than the money.
Setting aside these qualifications, let’s look at voluntary exchange in more detail. Sup-
pose that there are only two consumers, Mr. Edge and Ms. Worth, and only two goods,
ballet and football tickets. Assume that Edge and Worth begin with specific quantities of
each good. Edge’s initial market basket (endowment) is 35 ballet and five football tickets;
Worth’s initial market basket is five ballet and 45 football tickets. In the end, Mr. Edge
and Ms. Worth may choose not to consume their initial market baskets. Instead, they may
decide to trade with one another and end up with different combinations of football and
ballet tickets.
Under what conditions will Edge and Worth find it advantageous to trade? In this set-
ting, whether trade occurs depends crucially on the relative importance of the two goods to
each consumer. Let’s suppose that Mr. Edge, given his initial basket, would be willing to
give up five ballet tickets to obtain one more football ticket: his marginal rate of substitu-
tion is 5B/1F. Ms. Worth, on the other hand, would be willing to give up only one ballet ticket for one more football ticket: her marginal rate of substitution is 1B/1F. In this case the relative importance of the goods differs between Edge and Worth, as indicated by their
economic efficiency with regard to exchange, economic efficiency represents a distribution of goods across consumers in which no one consumer can be made better off without hurting another consumer
142 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 142Trim Size: 203.2 mm X 254 mm
different MRSs; thus, a mutually beneficial exchange can take place. Let’s see how. (The numerical data of this example are summarized in Table 6.1 for convenience.)
Given their initial holdings, Edge values football more highly than Worth does relative
to ballet. He is willing to give up as many as five ballet tickets to obtain another football
ticket. Worth, in contrast, is willing to give up one football ticket if she receives at least
one ballet ticket in return. (Note that if Worth’s MRS is 1B/1F, she will give up 1B for 1F, or, alternatively, 1F in return for 1B. For a small movement in either direction along the indifference curve, 1F trades for 1B without affecting her well-being.) Put differently, Edge would pay 5B to get 1F, while Worth would be willing to sell him 1F for 1B. There is room for a mutually beneficial trade. Suppose that Edge offers Worth three ballet tickets for one
football ticket, and she accepts. Edge will be better off after the exchange because he would
have been willing to pay as much as 5B for the football ticket, but he got it for 3B. Worth will also be better off because she would have sold the football ticket for as little as 1B but instead received 3B. Therefore, this exchange will leave both parties better off: they prefer their new market baskets to their initial holdings. (See Table 6.1.)
The Edgeworth Exchange Box Diagram The way two parties may both gain from exchange illustrates a simple idea: “You have
what I want, and I have what you want, so let’s trade!” While this statement conveys intui-
tively what is involved, a deeper understanding of voluntary exchange is important. A new
graphical device, the Edgeworth exchange box, can give us that understanding.
The Edgeworth exchange box diagram can be used to examine the allocation of fixed total quantities of two goods between two consumers.1 Figure 6.1 shows the box diagram
appropriate for the numerical example just discussed. The horizontal and vertical dimen-
sions of the box indicate the total quantities of the two goods. The length of the box
indicates the total amount of football tickets held by the two consumers, 50, and the height
of the box indicates the total amount of ballet tickets, 40.
By interpreting the box in a certain way, we can show all the possible ways 50 foot-
ball tickets and 40 ballet tickets can be divided between Mr. Edge and Ms. Worth. Let’s
measure Edge’s holdings of football tickets horizontally from point 0E and his holdings of ballet tickets vertically from the same point. In effect, 0E is the origin of the diagram for purposes of measuring the number of football and ballet tickets possessed by Edge. Point A shows the market basket for Edge that contains 35B and 5F; it is, in fact, his initial market basket from our numerical illustration.
One ingenious aspect of the Edgeworth box is that point A also indicates Worth’s market basket. Because the number of football tickets held by both parties is fixed at 50, placing
Edge at point A with five football tickets means Worth holds the remaining 45 tickets. This amount is shown in the diagram by measuring Worth’s football holdings to the left from
point 0W. Worth has 45 football tickets: point A in effect divides the horizontal dimension of the box, 50F, between Edge (5F) and Worth (45F). Point A also indicates Worth’s ballet
Edgeworth exchange box a diagram for examining the allocation of fixed total quantities of two goods between two consumers
Table 6.1 Gains from Exchange
Consumer Initial Market
Basket MRSFB Trade New Market Basket
after Trade Mr. Edge 35B + 5F 5B/1F −3B + 1F 32B + 6F Ms. Worth 5B + 45F 1B/1F +3B − 1F 8B + 44F
1The Edgeworth box is named for F. Y. Edgeworth, who hinted at such a construction in 1881 in his Mathematical Psychics: An Essay on the Application of Mathematics to the Moral Sciences (New York: August Kelly, 1953).
Two-Person Exchange 143
C06.INDD 11:8:52:AM 08/06/2014 PAGE 143Trim Size: 203.2 mm X 254 mm
holdings by measuring them down from point 0W. Since combined ballet holdings equal 40, Edge has 35, and the remaining five tickets belong to Worth.
Now consider point C, which identifies a different market basket for both Edge and Worth. Edge’s market basket at C contains 32B and 6F, while Worth’s contains 8B and 44F. (The totals still add to 50F and 40B.) In fact, the movement from point A to point C illus- trates the exchange between Edge and Worth in our numerical example. In moving from
A to C, Edge has given up 3B and gained 1F, while Worth has gained 3B and given up 1F. The movement from A to C shows graphically what happens when Edge buys one football ticket from Worth and pays for it with three ballet tickets.
The Edgeworth Exchange Box with Indifference Curves Because the points in the Edgeworth box identify alternative market baskets that each
party may consume, we can use indifference curves to represent each person’s prefer-
ences regarding the alternatives, as in Figure 6.2a. Edge’s indifference curves, UE3 and UE4, require little explanation. Because Edge’s market baskets are measured in the normal way,
from his origin at the southwest corner, these curves have their familiar shapes.
Worth’s indifference curves, UW3, UW4, and so on, may appear odd at first glance. Recall that the origin, for purposes of measuring Worth’s consumption, is the northeast corner, 0W, and her indifference curves are relative to this origin. Compared with the usual graphic rep-
resentation, all we have done is rotated Worth’s indifference map 180 degrees and placed
the origin in the northeast corner. The small insert in Figure 6.2a should make this manipu-
lation clear: it shows how the normal indifference map has been inverted to place the origin
at 0W in the box diagram. Worth’s indifference curves also have normal shapes; we are just looking at them upside down.
Edgeworth Exchange Box The horizontal dimension of the box measures the total number of football tickets, and the vertical dimension measures the total number of ballet tickets. When the origins of the consumers are 0E and 0W, each point in the box represents a specific division of the goods between the consumers.
0E
0W
Edge’s ballet
Worth’s ballet
Total ballet = 40B
35B
32B
5B
8B
5F
45F 44F
6F
Edge’s football
Worth’s football
Total football = 50F
C
A
Figure 6.1
144 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 144Trim Size: 203.2 mm X 254 mm
Now let’s use this construction for our original example once again. Point A in Figure 6.2a identifies Edge’s and Worth’s initial market baskets. The indifference curves corresponding
to the initial market baskets (passing through point A) are UE3 and UW3. Note that these curves
Gains from Trade (a) Point A shows the initial division of goods. Edge’s and Worth’s indifference curves through point A intersect, transcribing the shaded lens-shaped area. Both parties would be better off with any division of goods lying inside the lens-shaped area. It is thus in the interest of both parties to work out an exchange that will move them into this area. (b) Once the two parties reach a point such as E, where their indifference curves are tangent, no further trade is possible that will benefit both parties. Any movement from point E would harm at least one of the two parties.
0E
0W
Edge’s ballet
Worth’s ballet
Edge’s football
Worth’s football
D
A
A
G
E
H
C
0E
Edge’s ballet
Edge’s football
(b)
(a)
0W
Worth’s ballet
Worth’s football
0W Football
Ballet
UE4
UE6
UE5
UE3
UW3UW5UW6
UE3 UW3
UW4
Figure 6.2
Two-Person Exchange 145
C06.INDD 11:8:52:AM 08/06/2014 PAGE 145Trim Size: 203.2 mm X 254 mm
intersect at point A, because we assume that the marginal rates of substitution differ for the two consumers. The slope of UE3 at point A is 5B/1F, while the slope of UW3 is 1B/1F. The area that lies between these two curves—the shaded area in the graph—is highly significant.
Every point inside the shaded area represents a market basket for each consumer that is pre- ferred to basket A. In other words, within this area both consumers would be on higher indif- ference curves compared with their initial curves at point A.
The shaded lens-shaped area in Figure 6.2a illustrates the potential benefit from
exchange, and it is possible to arrange exchanges between Worth and Edge so that they
move into this area and both benefit. For example, if Edge purchases one football ticket
from Worth and pays for it with three ballet tickets (as we discussed earlier), then they
move from A to C. Note that both Edge and Worth are better off (that is, on higher indiffer- ence curves) at point C than they were at point A.
Starting at point A, any voluntary exchanges that occur will necessarily involve move- ments into the lens-shaped area of mutual gain. If they moved outside the shaded area, one
or both parties would be worse off (on a lower indifference curve), so such a trade would
never be mutually agreed to. For example, a move from point A to point D (lying outside the shaded area) would never be agreed to by Worth. It would put her on a lower indiffer-
ence curve than the one that she is on, UW3, at point A. The basic economic proposition illustrated here is that exchanges will tend to take place as long as both parties continue
to benefit. After moving from point A to point C, both parties may still benefit from fur- ther exchanges: UE4 and UW4 intersect at point C, carving out a smaller lens-shaped area of mutually beneficial potential exchanges. For example, Edge might trade his ballet tickets
for Worth’s football tickets again in such a way as to move into this smaller lens-shaped
area and make both parties better off than they are at point C. We have seen that, starting at point A, voluntary exchange can benefit both parties by
moving them into the lens-shaped area defined by indifference curves UE3 and UW3. How- ever, once they reach a point such as E in Figure 6.2b, where their indifference curves are tangent, no further trade is possible that will benefit both. Any movement from point
E would harm at least one of the two (as an inspection will show), so the injured party would never agree to such an exchange. A tangency of indifference curves implies that
the two consumers’ marginal rates of substitution are equal. The process of trading from
point A, where the MRSs differ, tends to bring the MRSs into equality. As Edge acquires more football tickets and gives up ballet tickets, his MRS becomes lower: his indifference curve is flatter at point E than it is at A. As Worth gets more ballet tickets and gives up football tickets, her MRS becomes greater: her indifference curve is steeper at point E than at A.
APPLICATION 6.1
Testament to the proposition that voluntary exchange represents a win–win for participating parties is provided by the growth of consumer-to-consumer (C2C) e-commerce. The largest supplier of C2C services‚ eBay‚ was founded in 1995 by Pierre Omidyar‚ a 31-year-old French-born Iranian software developer who wanted to help his fiancée trade Pez dispensers online. By 2013‚ eBay had over 100 million active users around the world‚ earned $16 billion in annual revenues‚ accounted for nearly 25 percent of all e-commerce in the United States, and featured more
The Benefits of Exchange and eBay
than 20‚000 categories of goods, including collectibles‚ antiques‚ sports memorabilia‚ Beanie Babies‚ jewelry‚ and books.
On a typical day of trading‚ the listed items have included a castle in Tucson‚ a designer wedding gown‚ Tickle Me Elmo toys‚ a Daewoo hatchback‚ and a photograph of young Abraham Lincoln.
While eBay offers the marketplace‚ the consumers who connect through the marketplace do all the work. When the auction is over‚ the high bidder and the seller contact
146 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 146Trim Size: 203.2 mm X 254 mm
Where the marginal rates of substitution differ, mutually beneficial trade between the parties is possible. Differing MRSs imply intersecting indifference curves and a correspond- ing lens-shaped area of potential mutual gain in the Edgeworth box diagram. Predictably,
then, voluntary exchanges will occur to realize the potential gain.
We should mention one final point. Our theory does not permit us to predict exactly
where in the lens-shaped area the consumers will end up. Although there should be a ten-
dency for trade to continue until it is no longer mutually beneficial—until a tangency is
reached—this condition does not identify a unique outcome. For example, if Edge is a very astute trader, he might persuade Worth to agree to an exchange from point A to a point near H in Figure 6.2b where Worth is scarcely any better off (she would be no better off at H); then the lion’s share of the potential mutual benefit goes to Edge. Conversely, if Worth is a
sharp bargainer, they might end up at a point near G. The reason for the indeterminacy is that we are assuming only two potential traders, one
buyer and one seller, so this setting is not a competitive one (many buyers and sellers). When
only two parties participate in the exchange process, elements of haggling and strategy
appear, because each tries to conceal from the other how much he or she wants the trade
so as to get the best terms. Thus, we are unable to predict the exact terms of the exchange,
except to note a tendency for any exchanges that occur to benefit both parties to some degree.
2Paul Resnick, Richard Zeckhauser, John Swanson, and Kate Lock- wood, “The Value of Reputation on eBay: A Controlled Experiment,” Experimental Economics, 9, No. 2 (June 2006), pp. 79–101.
each other‚ usually by e-mail‚ to finalize the purchase. The earnings to eBay come from fees paid by sellers to post items and a commission that is tied to the sale price. The fact that eBay has been consistently profitable since its founding‚ even through the economic downturns of the prior decade‚ attests both to the win–win nature of voluntary exchange and to the financial benefits that can accrue to companies that successfully provide low- cost mechanisms for parties with differing marginal rates of substitution to connect. Because their marginal rates of substitution differ‚ the connected parties’ indifference curves intersect in their relevant Edgeworth exchange box (as illustrated by a point such as A in Figure 6.2a) and there is room to move into a lens-shaped area of mutual benefit through trade.
Of course‚ there are some challenges associated with ensuring that Internet-based C2C exchange is mutually beneficial. For example‚ a seller may misrepresent the qual- ity of an item put up for auction. In an attempt to address such potential problems‚ eBay allows sellers and buyers to rate each other at the end of a transaction. Transaction insurance and escrow arrangements are also offered by eBay and allow users to validate prospective trading part- ners’ identities and past histories.
The feedback system is employed in more than half of all transactions and, perhaps surprisingly, more than 99 percent of feedback on sellers is positive. Sellers have an important means, through the rating system, to build a reputation that has value to them in future transactions. Consumers, in turn, can make better informed future purchasing decisions based on the accumulated historical performance data about sellers.
To test for the value of a positive reputation, economists Paul Resnick, Richard Zeckhauser, John Swanson, and Kate Lockwood auctioned carefully matched items from two sellers.2 One seller had a superb, long-established reputa- tion, generated through the positive feedback registered by buyers who had dealt with that seller in past transactions. The second set of items was auctioned by the same seller, operating under a newly established identity and no track record. The researchers found that the established, positive identity brought in 7.6 percent more money, on average, per transaction.
The items that are listed for sale can also be unusual and sometimes controversial. For example, the wife of Tim Shaw, a British DJ, sold Tim’s Lotus Esprit sports car with a “Buy It Now” price of 50 pence after she heard him flirting with model Jodie Marsh on air. The car was sold within five minutes, and it was requested that the buyer pick it up the same day. A seaworthy 16,000-ton aircraft carrier, formerly the British HMS Vengeance, was listed in 2004. The auction was removed when eBay determined that the vessel quali- fied as weaponry, even though all weapon systems had been removed. A man from Brisbane, Australia, attempted to sell New Zealand at a starting price of 0.01 AUD (Austra- lian dollars). The price had risen to 3,000 AUD before eBay closed the auction. In 2007, after Britney Spears shaved all of her hair off in a Los Angeles salon, it was listed on eBay for $1 million before it was taken down after considerable controversy.
Eff ic iency in the Distr ibut ion of Goods 147
C06.INDD 11:8:52:AM 08/06/2014 PAGE 147Trim Size: 203.2 mm X 254 mm
6.2 Efficiency in the Distribution of Goods Economic efficiency—or, as it is sometimes called, Pareto optimality3—is a characteris- tic, highly regarded by economists, of some resource allocations. Noneconomists do not
generally hold it in such high esteem; they frequently disparage it because they believe it
relates only to materialistic values or monetary costs and ignores human needs. This criti-
cism misconstrues the meaning of economic efficiency as economists use the term. Quite the opposite of a materialistic focus, efficiency is defined in terms of the well-being of people.
Roughly speaking, an efficient outcome is one that makes people as well off as possible. A
full treatment of the concept of efficiency is important in appreciating its use in economic
analysis.
In this chapter we consider economic efficiency as it relates to the way fixed total quantities
of goods are distributed among consumers. Two consumers and two goods are analyzed,
just as before, but the results generalize easily to larger numbers. Moreover, the concept
of economic efficiency can be applied in other settings, such as deciding what level of out-
put of a particular good is most efficient. Clearly, overall efficiency in resource allocation
involves more than just an efficient distribution of goods, but an efficient distribution is an
important part of the overall concept. We discuss other aspects of economic efficiency in
later chapters.
Suppose that we have 50 football tickets and 40 ballet tickets to divide between our
friends, Mr. Edge and Ms. Worth. There are innumerable ways to distribute 50F and 40B between two consumers. The Edgeworth box diagram not only identifies all the possi-
bilities but also shows how alternative distributions affect the well-being of both parties.
Previously, we used the Edgeworth box to show how Edge and Worth, starting with certain
market baskets, could exchange products to reach a preferred position. Now, however, the
Edgeworth box is used in a different way; we wish to consider all the points in the diagram,
not just those that Edge and Worth can reach by voluntary trade starting from some initial
endowment. In other words, let’s imagine that a philanthropist is going to give 50F and 40B to Edge and Worth and is devising ways that this might be done. Some ways are effi- cient, and some are inefficient, as we will see.
Figure 6.3 is an Edgeworth box that shows different ways of dividing the given quanti-
ties of football and ballet tickets between Edge and Worth. Several indifference curves for
both people have been drawn so that we can see their preferences among the various pos-
sibilities. We begin by defining efficiency relative to this setting. An efficient distribution of fixed total quantities of goods is one in which it is not possible, through any change in the distribution, to benefit one person without making some other person worse off.
In the diagram, the points where Edge’s and Worth’s indifference curves are tangent show
efficient distributions. Point E, for example, satisfies the definition of efficiency. If we change the distribution from point E by moving to any other point, either Edge or Worth will be worse off (that is, on a lower indifference curve). For example, a move to point L makes Ms. Worth worse off, a move to point M makes Mr. Edge worse off, and so forth. Thus, point E is an efficient distribution of football and ballet tickets. Note, though, that it is not unique; indeed, any point of tangency between indifference curves defines an efficient distribution.
At point J, for example, UW7 is tangent to UE2. Any move from point J will harm at least one of the two consumers, so point J is an efficient distribution. Point K also represents an effi- cient distribution, as do other points of tangency not drawn in. A line drawn through all the
efficient distributions is called the contract curve. It is shown as CC in Figure 6.3, and it identifies all the efficient ways of dividing the two goods between the consumers.
Pareto optimality another term for economic efficiency
contract curve in an Edgeworth exchange box, a line drawn through all the efficient distributions
3Named after the Italian economist Vilfredo Pareto (1848–1923), who first systematically formulated the concept in Cours d’Economie Politique (Lausanne: F. Rouge, 1897).
148 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 148Trim Size: 203.2 mm X 254 mm
An alternative but equivalent way of defining efficiency may be helpful. An efficient
distribution is one that makes one party as well off as possible for a given level of well-
being for the second party. For example, suppose that we consider the given level of well-being
indicated by UW2 for Worth. She is equally well off at points L and K, or any other point on UW2. Among all the possible combinations of football and ballet tickets that keep Worth on UW2, Edge is best off at point K. Point K places him on the highest indifference curve possible, assuming Worth stays on UW2. Because point K makes Edge as well off as possi- ble for a given level of well-being (UW2) for Worth, it is an efficient distribution. Similarly, if we hold Worth on UW5, point E is best for Edge. Consequently, the same set of tangency points is defined when we look at efficiency in this way.
The contract curve identifies all efficient ways to divide football and ballet tickets
between Edge and Worth. In contrast, all points off the contract curve are inefficient alloca-
tions. Inefficiency may be defined as follows: an allocation of goods in which it is possible, through a change in the distribution, to benefit one party without harming the other.
In Figure 6.3, all points off the contract curve satisfy this definition of inefficiency.
Consider point L. If we change the distribution from point L to point K, Edge will be better off without harming Worth—she remains on the same indifference curve, UW2. Alternatively, we can move from point L to point E, which benefits Worth while leaving Edge’s well-being unchanged. Note, in fact, that an inefficient allocation, such as point L, permits a change that benefits both parties. That is, a move from point L to any point on the contract curve between points E and K will leave both Edge and Worth on higher indifference curves than the ones that they are on at point L.
If we select any point off the contract curve and draw Edge’s and Worth’s indifference
curves through this point, the curves will intersect. The intersecting indifference curves will
circumscribe a lens-shaped area within which both parties will be on higher indifference
curves. Because this is true of every point off the contract curve, all these points represent
inefficient distributions of the goods.
inefficiency an allocation of goods in which it is possible, through a change in the distribution, to benefit one party without harming the other
Efficient Distributions and the Contract Curve A distribution of goods for which the consumers’ indifference curves are tangent is efficient. We cannot change the distribution without making at least one of them worse off. There are many efficient distributions, as shown by the contract curve CC, which connects the points of tangency.
J
C
C
M
L
E
K
0E
Ballet
Football
0W
Ballet
Football
UE7
UE5 UW2UW5UW7
UE2
Figure 6.3
Eff ic iency in the Distr ibut ion of Goods 149
C06.INDD 11:8:52:AM 08/06/2014 PAGE 149Trim Size: 203.2 mm X 254 mm
By this means, all the ways that the goods can be divided between Edge and Worth can
be characterized as either efficient or inefficient. Inefficient distributions are shown as points
where the indifference curves of the two parties intersect, that is, where MRS MRSFBE FBW≠ . This inequality implies, just as with our initial numerical example (Table 6.1), that the
consumers place different values on the two goods, so both will prefer a different distri-
bution. Efficient distributions are shown by the contract curve, which connects the points
of tangency between indifference curves. Thus, efficient distributions are characterized by
an equality between marginal rates of substitution, or MRS MRSFBE FBW= . At those points the consumers’ relative valuations of the two goods are equal, and no further change that will
benefit both parties is possible.
APPLICATION 6.2
While Americans spend at least $80 billion annually on gift cards, the activity results in participating individuals not ending up on a contract curve. This is due to the fact that what donors are willing to pay for the cards is greater than the value placed on the cards by recipients.4 Because physical gift cards can only be used at the store issuing the card, may be lost, and typically have an expiration date, an estimated 10 percent of the gifted cards are never redeemed. Moreover, the average resale value of a gift card is 72 percent of its face value.
Promoting Efficiency in Gift Card Giving
Entrepreneurial ventures such as GiftRocket are seeking to enhance efficiency in gift-card giving by moving par- ticipating individuals closer to the relevant contract curve. Donors can give GiftRockets, delivered instantly by email and not requiring the production and mailing of a plastic card, to recipients for use at practically any business. Recipi- ents are free to redeem the donated funds in a number of different ways (Pay Pal account, bank transfer, credit card credit or check in the mail) and can opt to apply the funds in whole, part or not at all to the business designated by the donor while using any remaining funds on a purchase of their choosing. GiftRocket charges $1 plus 5 percent of the gift amount while promoting efficiency in the gift card mar- ket and bringing the value donors are willing to pay for the cards closer to the value realized by recipients.
4This application is based on: “GiftRocket Seeks to Reinvent Gift Cards,” Forbes.com, April 14, 2011; and Wade Roush, “GiftRocket Seeks to Take the Pain (and Loss) Out of Gift Cards,” April 7, 2011, Follow@wroush.
Efficiency and Equity For any inefficient point there are many efficient points on the contract curve that both par- ties prefer. This can be seen by looking at any point such as L located off the contract curve in Figure 6.3. In this case points between E and K on the contract curve are better from both Edge’s and Worth’s viewpoints than point L. Most people would probably agree that the efficient points between E and K are preferred to the inefficient point L.
Now suppose that two efficient points are compared, points E and K, for instance. Which of these is “better”? Because both points are efficient, the concept of efficiency
provides no help in choosing between the two, yet there is a marked difference between
E and K. Edge is better off at point K than at point E, while Worth is better off at point E than at point K. Moving from K to E benefits Worth at Edge’s expense; moving from E to K benefits Edge at Worth’s expense. To judge one efficient point superior to another requires deciding whose well-being is more important, and there is no objective basis for
such a decision. In the economist’s jargon, interpersonal comparisons of utility cannot be made scientifically, so there is no objective way to demonstrate that one efficient point is
preferred to another.
If a philanthropist has to choose how to divide the football and ballet tickets between
Edge and Worth, on what basis could the choice between points E and K be made? The
150 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 150Trim Size: 203.2 mm X 254 mm
decision would have to be based on something other than efficiency because both points are
efficient. Equity, or fairness, might provide the basis for the decision. However, although we all have views of what is equitable or fair, our views differ: there is no agreed-upon def-
inition of what constitutes equity. For that reason, economics does not provide a formula
that allows us to state that one efficient point is better than any other.
Note also the critical role that the initial endowment plays in determining which of all
of the efficient points on the contract curve in Figure 6.3 are attainable through voluntary
exchange. For example, starting from endowment point L, all points between E and K on the contract curve are attainable through voluntary exchange. Any other points on the con-
tract curve, however, are not attainable. Edge would never agree voluntarily to move to
point J if the initial endowment is L. Point J lies on a lower indifference curve for Edge than does point L. Nevertheless, Edge may end up at point J if we start from an initial endowment that is less favorable to Edge than L, an endowment such as M. (Point M lies on a lower indifference curve for Edge than does point L.)
While voluntary exchange serves to promote efficiency starting from any initial endow-
ment of goods, we can see that the “fairness” of any efficient distributive outcome hinges
critically on the initial endowment. The economist’s objective criteria of Pareto optimality
does not provide us a means of judging the relative desirability of different initial endow-
ments. Much as there are no objective criteria for deciding on the relative worth of all the
different efficient outcomes lying on a contract curve, there are no objective criteria for
evaluating the desirability of different endowments such as points L and M. Once again, this judgment must be based on normative, or equity, considerations. And people disagree
on what equity implies for initial endowments, as they do over the desirability of various
possible distributive outcomes that can emerge from voluntary exchange.
In sum, economics offers an objective means to see why for any initial inefficient point
there always exist efficient points preferred by all parties. A choice among initial endow-
ment points or between different efficient distributive outcomes, however, requires that nor-
mative considerations be brought into play: some subjective judgment about which of the
parties is more deserving relative to others. The objective dictates of Pareto optimality can-
not help us make these normative judgments.
6.3 Competitive Equilibrium and Efficient Distribution In the two-person model of exchange, the exact outcome of bargaining cannot be predicted.
We can expect the parties to haggle over the terms of exchange (the price), with the result
depending on which one is the superior negotiator. In the real world, however, buyers and
sellers might not haggle over prices, especially when there are many buyers and sellers in
any given market. In such cases, each party is not limited to dealing with another specific
party—there are alternatives. It would do you little good, for example, to try to bargain over
the price of a book; if you offer a price below the going rate, the book store manager will
simply wait for another customer. Similarly, if the manager tries to extract a higher price
from you, you will probably buy the book from another source. The existence of alternative
buyers and sellers greatly limits the influence any one of them can have on the price. People
simply find alternative sources if the deal offered by one person is not as good as what can
be obtained elsewhere.
With many buyers and sellers, each individual will behave like a price taker. Consumers acting individually cannot affect the price perceptibly by haggling. They take the price as
given and buy or sell whatever quantities they wish at that price. What determines the price
in this many-person setting? It is, of course, the interaction of supply and demand, because
we are now dealing with a competitive market. Namely, the overall demand for and supply
equity the concept of fairness
price taker firms or consumers who cannot affect the prevailing price through their respective production and consumption decisions
Competit ive Equi l ibr ium and Eff ic ient Distr ibut ion 151
C06.INDD 11:8:52:AM 08/06/2014 PAGE 151Trim Size: 203.2 mm X 254 mm
of a good across all individual market participants with their respective initial holdings and
preferences determine the market price. Once this market price is determined, individual
participants must decide how much of the good they wish to buy or sell.
Let’s rejoin Edge and Worth and extend our two-person model to show the outcome
when the competitive equilibrium is attained and market participants are price takers. We
can illustrate the nature of this equilibrium for Edge and Worth by using the Edgeworth box
diagram. In Figure 6.4, Edge and Worth begin with the initial holdings shown at point A. Suppose that the market-determined price of one football ticket is three ballet tickets. Then,
both Edge and Worth confront the budget line ZZ′, which has a slope of 3B/1F. At that price, Edge prefers to move to point E, purchasing four football tickets in exchange for 12 ballet tickets. Confronted with ZZ′, Worth also prefers point E, selling four football tickets for 12 ballet tickets. If both Edge and Worth take the price (3B/1F) as given and are the only two market participants, the quantity of football tickets demanded by Edge (4) is exactly
equal to the quantity Worth wants to sell at that price. The price of three ballet tickets
per football ticket is therefore an equilibrium price.
Remember that the preceding example investigates price-taking behavior, while assum-
ing that there are only two market participants. In general, price-taking behavior results from
the presence of many market participants. And the equilibrium price will be one where the
total amount of football tickets demanded by buyers such as Edge in our preceding example
equals the quantity sellers such as Worth are willing to supply. If the price were lower, there
would be a shortage of football tickets, and the price would rise. If the price were higher,
there would be a surplus of tickets, and the price would fall. The tangency of indifference
curves at point E in Figure 6.4 simply illustrates the balance between quantities demanded and supplied at the competitive equilibrium price in a two-person setting.
Competitive Exchange In a competitive equilibrium, market actors are price takers—they all confront a uniform price for football tickets equal to three ballet tickets per football ticket. Each party faces the budget line ZZ′, and the equilibrium is at point E—an efficient outcome.
A Z
E
0E
Ballet
Football
0W
Ballet
Football
23B
3B
35B
40B
5F 9F
1F Z ′
50F
UW2 UW1
UE1
UE2
Figure 6.4
152 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 152Trim Size: 203.2 mm X 254 mm
Another implication of the competitive equilibrium should not be missed: the final equi- librium point is an efficient allocation. This can be seen in Figure 6.4 by noting that point E is a point of tangency between indifference curves and therefore lies on the contract curve.
In a pure exchange model, this conclusion illustrates Adam Smith’s famous “invisible
hand” theorem: each trader, concerned only with furthering his or her own interest, is led to
exchange to a socially efficient result. All the potential gains from voluntary exchange are
realized in a competitive market.
There is another way of showing that a competitive equilibrium produces an efficient
allocation. Specifically, recall from earlier in this chapter that an efficient distribution in our
two-person (Edge and Worth), two-good (football and ballet tickets) case requires that:
MRS MRSFBE FBW= . (1)
Because Edge and Worth make their consumption decisions independently of one another
in a competitive market setting, it might seem unlikely that this condition will be satisfied.
Consider, however, the equilibrium conditions that result when each person allocates his or
her income in the appropriate way:
MRS P PFBE F B= / ; and (2)
MRS P PFBW F B= / . (3)
Chapter 3 demonstrated that each consumer purchases a market basket such that his or her
marginal rate of substitution equals the price ratio. Because the prices are the same for both
consumers, each consumer’s MRS is equal to the same price ratio, and so the MRSs are equal to one another. Thus, condition (1) is satisfied.
Let’s look at this matter graphically. Parts (a) and (b) in Figure 6.5 show Mr. Edge and
Ms. Worth in competitive equilibrium. In part (a), on his budget line ZZ′, Edge is at point E
A Market-Determined Distribution Is Efficient (a) Edge’s optimal consumption point is E. (b) Worth’s optimal point is also E. Even though each consumer acts independently, they both seek to equate their MRS to the same market- determined price of ballet passes per football ticket, ZZ′. An efficient distribution results.
Z
E
0E
Ballet
Football
(a) (b)
Z
23B
3B
9F
1F
Z ′ Z ′0W
Ballet
Football
17B
41F
3B 1F
E
UE2
UE1 UW2 UW1
Figure 6.5
Competit ive Equi l ibr ium and Eff ic ient Distr ibut ion 153
C06.INDD 11:8:52:AM 08/06/2014 PAGE 153Trim Size: 203.2 mm X 254 mm
consuming 23 ballet tickets and 9 football tickets. Worth is consuming 17 ballet and 41
football tickets at point E in part (b). Note, however, that the slopes of their budget lines are equal. Because they face the same market prices for football and ballet, PF/PB (equal to 3B/1F in the diagram) is the same for both of them. Thus, the slope of Edge’s indifference curve at his optimal consumption point, 3B/1F, equals the slope of Worth’s indifference curve at her optimal point.
Parts (a) and (b) of Figure 6.5, of course, are nothing more than the components of
Figure 6.4, the Edgeworth box diagram for the total quantities of football and ballet,
50F and 40B, consumed by Edge and Worth. To see this, start with Edge’s budget line, ZZ′, from part (a) of Figure 6.5. Then, rotate Worth’s indifference map from part (b) 180 degrees, superimposing her optimal consumption point, E, on Edge’s optimal point and ensuring that her rotated indifference curve UW2 is tangent to Worth’s budget line ZZ′ at E in part (a). Because they are tangent to the same budget line ZZ′, Edge’s and Worth’s indif- ference curves UE2 and UW2 are also tangent to each other at point E. Thus, the distribution of goods described by points E in the two graphs is an efficient one: there is no other way to divide 50F and 40B that would not make at least one of the two consumers worse off.
Over any period of time the market distributes, or rations, goods among consum-
ers. Although each consumer acts independently in choosing a market basket, the result
is an efficient distribution. This outcome depends on two conditions. First, the prices of
goods must be the same for all consumers; this condition ensures that every consumer’s
budget line will have the same slope. Second, consumers must be able to purchase what-
ever quantity they want at those prices; this condition ensures that every consumer can
select a market basket for which the marginal rate of substitution equals the ratio of the
prices of the goods.
Reaching an efficient distribution, albeit not the only possible efficient distribution, is
no small feat. No philanthropist or government agency knows the preferences of millions
of consumers. To attain any efficient outcome in this setting is a considerable achievement.
When consumers must pay for the products they consume, self-interest leads them to utilize
their knowledge of their own preferences, and these preferences are reflected in the market
APPLICATION 6.3
The Arizona Diamondbacks baseball team won a thrilling World Series victory over the New York Yankees in 2001. While the owner of the Diamondbacks‚ Jerry Colangelo‚ had a lot to be pleased about that year‚ he groused publicly about the extent to which ticket scalping had undermined the value of his business. Ticket scalping involves the oper- ation of a secondary market and the resale of tickets by original recipients to other interested consumers.
Arizona State University economist Stephen Happel responded to the grousing by noting that the Diamond- backs owner should be thankful for the existence of ticket scalpers since they‚ if anything‚ enhanced the value of Colangelo’s organization while promoting mutually benefi- cial exchange:5
Should Ticket Scalpers Be Disparaged or Deified?
Colangelo needs to remember three points. First‚ don’t complain in the papers about scalpers’ prices when a $9 season ticket was $110 for the World Series (a 1‚122 percent markup). Second‚ don’t complain about how scalpers are making money unfairly off you. Simply wait and charge higher prices: voluntary exchange benefits both parties and your after-the-fact argument implies you want to have your cake and eat it too. Third‚ you need a secondary ticket market created by scalpers. Without vibrant resale opportunities‚ far fewer season tickets are sold‚ since fans do not want to attend 81 home baseball . . . games. They attend a select number of games and unload excess tickets in the secondary market. Scalpers bring in people who otherwise wouldn’t go‚ thereby contributing to team revenues via concession sales.
5Stephen Happel‚ “Ticket Scalpers Play Fair in Our Free-Market System‚’’ Arizona Republic‚ December 30‚ 2001‚ p. V3.
154 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 154Trim Size: 203.2 mm X 254 mm
basket they select. Because their decisions are guided by the same relative prices con-
fronting other consumers, the result is a coordination among purchase plans that would be
difficult to achieve any other way.
6.4 Price and Nonprice Rationing and Efficiency In open markets, prices serve a rationing function in determining how much of available
quantities each consumer will get. The rationing function and whether it is accomplished
efficiently or inefficiently are what this chapter is all about. We conclude our analysis with
an example that illustrates the relationship between the demand curve treatment of rationing
problems and the Edgeworth box approach emphasized in previous sections.
Figure 6.6b shows the market demand and supply curves for gasoline. Because our
emphasis is on the rationing of fixed supplies, the supply curve is drawn as vertical (perhaps
reflecting a very short-run situation). With the S supply curve, the per-gallon price is $3 and quantity is 150 gallons. Now suppose that there is a sharp reduction in supply to 100 gallons
so that the supply curve shifts from S to S′ (because of a foreign oil embargo, perhaps). The market response is an increase in price to $4 per gallon. Consumers are induced by the
higher price to restrict their use of gasoline to the available quantity.
By looking at Figure 6.6a, we see what this price increase means for the individual
consumers in the market. Once again, we consider only two consumers, Edge and
Worth, whose demand curves are dE and dW. When the price reaches $4, each consumer moves up his or her demand curve, cutting back on any gasoline use that is valued at
less than $4 per gallon. The final optimal consumption points are A and B, with Worth purchasing 70 gallons and Edge purchasing 30 gallons. The sum of their purchases, 100
gallons, is, of course, the total quantity purchased, shown in Figure 6.6b.
Their adjustment to the higher price represents an efficient rationing of the reduced
quantity available. Consider how this result would appear if the final equilibrium were
Gasoline Rationing Nonprice rationing will generally lead to an inefficient distribution. We can illustrate an inefficient distribution by differences in the demand prices of consumers. When each consumer receives 50 gallons of gasoline, Worth’s demand price is $5 and Edge’s is $3.20, implying that both would be better off with a different distribution.
(b)(a)
0 1007050300 150
$3.00
$4.00
Price
$3.00 $3.20
$4.00
$5.00
Price
Gasoline (gallons)
Gasoline (gallons)
dWdE
AB
R
A ′
B ′
S' S
D
Figure 6.6
Price and Nonpr ice Rat ioning and Eff ic iency 155
C06.INDD 11:8:52:AM 08/06/2014 PAGE 155Trim Size: 203.2 mm X 254 mm
shown in an Edgeworth box. Edge, in buying 30 gallons, consumes at a point where his
marginal rate of substitution between outlays on all other goods and gasoline is $4.00 per
gallon. Similarly, Worth, in buying 70 gallons, consumes at a point where her marginal
rate of substitution between outlays on all other goods and gasoline is $4.00 per gallon.
Their marginal rates of substitution are equal, and if this situation were depicted in
an Edgeworth box, it would look qualitatively like point E in Figure 6.4.6 Edge’s and Worth’s indifference curves would have the same slope at their optimal consumption
points and so would be tangent in the box diagram. Consequently, allowing the market
price to ration the available quantities between the consumers leads to an efficient distri-
bution of goods.
We can better appreciate the significance of the rationing problem by speculating on
how it might be resolved if price were not allowed to perform this function. For example,
suppose that when the supply of gasoline falls, the government does not allow the price to
rise but instead imposes a price ceiling at $3.00 per gallon. In that event the total quantity
demanded exceeds the total quantity available, but somehow the combined use of Edge
and Worth must be restricted to 100 gallons. Suppose that the government implements
a rationing scheme by using ration coupons, as it did in World War II. (In fact, during
the Arab oil embargo of 1973–1974, the government again proposed the use of ration
coupons and printed 3.8 billion of them, but the coupons were never used.) One hundred
6In this case, however, the vertical dimension of the Edgeworth box would measure the total amount of gasoline across the two consumers, while the horizontal dimension measures the total dollar outlay on all other goods.
APPLICATION 6.4
An equal allocation of ration coupons or a lottery are not the only nonprice ways of rationing a good that is in short supply at the existing price. Sometimes, as in the example of rent control discussed in Chapter 2, a price-controlled good is allocated to buyers on a first-come, first-served basis. This rationing-by-waiting approach has the advantage that, if there are no costs to waiting, consumers placing the highest marginal value on a good have the greatest incen- tive to get in line to purchase the good. To the extent that consumers placing the highest marginal value on a good are at the head of the line, efficiency in the distribution of the good among consumers is promoted.
Rationing by waiting, however, has its costs if consum- ers have something else that they could be doing besides waiting in line. For example, a study examined the effects of a price ceiling applied to a Chevron station in Ventura, California, in 1980 versus two competing stations not subject to the same price ceiling.7 (Stations owned and operated by integrated oil companies were subject to the price ceiling, while those operated by independent or franchised dealers
The Benefits and Costs of Rationing by Waiting
were not.) The study found that the price at the Chevron station was $0.19 per gallon lower than at the competing stations. Because of the lower price, long lines formed at the Chevron station, and the average time a consumer spent waiting in line was 15 minutes. By contrast, there was no waiting at the competing stations.
The study surveyed customers at the various stations and estimated the customers’ opportunity cost of time based on their employment and income characteristics. The study found that a significant percentage of the increase in consumer surplus generated by the price ceiling was dissi- pated through the costs of having to wait in line to buy the low-priced Chevron gasoline. Specifically, once the costs of waiting were accounted for, consumers received only 49 percent of the increase in consumer surplus generated by the price ceiling at the Chevron station. Moreover, the study pointed out that there is no guarantee that if costs are associated with waiting, the consumers placing the highest marginal value on a good will also be the ones with the lowest opportunity cost of time. To the extent that high- marginal-value customers also have a high opportunity cost of time, they will be discouraged from waiting in line and a rationing-by-waiting scheme will not allocate the good across consumers in an efficient manner.
7Robert T. Deacon and Jon Sonstelie, “Rationing by Waiting and the Value of Time: Results from a Natural Experiment,” Journal of Political Economy, 93, No. 4 (October 1985), pp. 627–647.
156 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 156Trim Size: 203.2 mm X 254 mm
ration coupons will be printed and distributed to Edge and Worth. To purchase a gallon
of gasoline, a consumer must pay $3.00 and turn in one ration coupon; because only 100
coupons are available, gasoline purchases will not exceed the available supply. Resale of
the coupons is not permitted.
The problem, then, is how to divide the ration coupons between Edge and Worth.
Suppose that each receives 50 coupons: then neither could purchase more than 50 gallons. In
Figure 6.6a, this solution puts both Edge and Worth at point R, each paying $3.00 per gallon and receiving 50 gallons. Both, however, place a value greater than $3.00 on gasoline at that
consumption level. When Worth buys 50 gallons, her marginal value of gasoline is $5.00
per gallon. This marginal value is indicated by A′, the height of her demand curve at 50 gallons. (Remember from Chapter 2 that the demand curve’s height reflects the maximum a
consumer is willing to pay for an incremental unit of a good.) A marginal value of $5.00 per
gallon also implies that Worth’s marginal rate of substitution between outlays on all other
goods and gasoline is $5.00 per gallon when only 50 gallons are available. In other words,
it is the maximum dollar outlay on all other goods that she is willing to give up to get an
additional gallon of gasoline at 50 gallons. Edge’s marginal value of gasoline is $3.20 at
50 gallons as reflected by his demand curve’s height at B′. Because Worth places a higher marginal value on gasoline than Edge does, this method of rationing gasoline is inefficient.
We could also show this coupon-rationing equilibrium in an Edgeworth box; it would be
depicted by a point where Edge’s and Worth’s indifference curves intersect. (Qualitatively,
it would look like point A in Figure 6.4.) Because Worth would be willing to pay just under $5.00 for another gallon of gas, and Edge would be willing to give up a gallon for just over
$3.20, both would be better off if Worth could buy gasoline (or coupons) from Edge. The
government will not allow her to do so in our example, however, so mutually advantageous
trades cannot occur. In this situation a black market in gasoline (or coupons) may well arise.
Thinking about how to distribute gasoline ration coupons suggests how difficult it is
to reach an efficient outcome if voluntary exchange and market-determined prices are not
allowed to perform their rationing function. The essence of efficient rationing is to dis-
tribute a good so that its marginal value is the same among consumers, but without know-
ing the preferences of all consumers, it is a virtually impossible task. For this reason any
type of nonprice rationing system is almost certain to involve some inefficiency in the way
goods are distributed among consumers.
Pointing out the inefficiency of nonprice rationing programs is not to claim that these
forms of rationing are undesirable. The purpose of price ceilings is generally to benefit
consumers at the expense of producers. Clearly, producers are harmed, but the significance
of this inefficiency is that it also diminishes the benefit to consumers. In our example,
both consumers would be better off—without further harming producers—if they could
exchange gasoline until their marginal values are brought into equality.
Of course, the long-run effect of the price ceiling on quantity supplied has been neglected
so as to focus on the rationing problem. Our purpose in this section has been to illustrate
inefficiency in the distribution of a given supply by using consumers’demand curves,
because this approach provides an alternative to the use of Edgeworth box diagrams.
SUMMARY
Voluntary exchange is mutually beneficial.
The Edgeworth exchange box shows that differing
marginal rates of substitution (MRS) imply the possi- bility of mutually beneficial exchange. The prospect of
mutual gain gives rise to voluntary exchange.
A distribution of goods between consumers is efficient
if any change in the distribution will harm at least one of
them. The many distributions that satisfy this definition are
shown as points on the contract curve in the Edgeworth
exchange box, along which consumers’ MRSs are equal.
C06.INDD 11:8:52:AM 08/06/2014 PAGE 157Trim Size: 203.2 mm X 254 mm
Review Quest ions and Problems 157
Equity is another criterion for evaluating economic
arrangements, especially when determining whether one
efficient distribution is to be preferred to another.
The distribution of goods implied by a competitive
market equilibrium is efficient. Because each con-
sumer strives to equate his or her MRS to the same
price ratio that confronts other consumers, consum-
ers’ MRSs end up being equal to one another. Some economic arrangements can lead to an inefficient
distribution of goods. Nonprice rationing schemes for gas-
oline that do not permit voluntary exchange provide such
an example.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
6.1 What do the dimensions of the Edgeworth exchange box signify? How does a point in the box identify the distribution of
goods between two consumers? What does a point on one of the
sides of the box indicate? What does it mean if we are located at
one of the corners of the box? (Examine each corner separately.)
6.2 What does a vertical movement inside the Edgeworth exchange box signify? Would a voluntary trade ever be shown
by a vertical movement? What does a horizontal movement
signify? Would a voluntary trade ever be shown by a horizon-
tal movement?
*6.3 John has 40 gallons of gasoline (G) and 20 bags of sugar (S); for that market basket, John’s MRSSG is 3G/1S. Maria has 40 gallons of gasoline and 50 bags of sugar; for that market
basket, Maria’s MRSSG is 1G/1S. Use a numerical example to explain how a trade can benefit both of them. Illustrate the
trade by using an Edgeworth exchange box, showing that both
consumers reach higher indifference curves.
6.4 Define efficiency and inefficiency in the context of the dis- tribution of goods between two consumers. If the distribution
lies inside an Edgeworth box, how does knowledge of the con-
sumers’ marginal rates of substitution permit us to tell whether
the distribution is efficient?
*6.5 John has 40 gallons of gasoline and no bags of sugar; for that market basket his MRSSG is 1G/1S. Maria has 20 gallons of gasoline and 20 bags of sugar; Maria’s MRSSG is 3G/1S. Is this arrangement an efficient distribution of goods? Explain, using
an Edgeworth box.
6.6 When John has 40 gallons of gasoline and 20 bags of sugar, his MRSSG is 5G/1S. When Maria has 40 gallons of gasoline and 50 bags of sugar, her MRSSG is 1G/1S. If John exchanges nine of his gallons of gasoline for three of Maria’s bags of sugar, both
their MRSs after the exchange are 3G/1S. Are they both better off? Show, using an Edgeworth box.
*6.7 Given his initial endowment of gasoline and sugar, John’s MRSSG is 4G/1S; given Maria’s initial endowment, her MRSSG is 2G/1S. If the government collects a tax of 3G for each unit of S traded, can John and Maria engage in mutually beneficial exchange? Compared to the absence of the tax, who is harmed
by it? Who benefits?
6.8 Bill and Hillary confront the same market prices for health care and hamburgers. Bill’s optimal consumption point is a
corner equilibrium where he consumes only hamburgers; Hill-
ary’s optimal point involves consumption of both health care
and hamburgers. Are their MRSs equal? Does the distribution lie on the contract curve? Support your answer by constructing
the relevant Edgeworth box.
*6.9 How is an equal distribution of goods shown in the Edge- worth box diagram? Is an equal distribution efficient?
6.10 Scrooge is the only moneylender in town—a monopo- list—and he charges exorbitant interest rates. If you borrow
money from Scrooge, does this practice illustrate the principle
that voluntary trade is mutually beneficial?
*6.11 An owner of an apartment building converts the units into condominiums, evicting the current tenants. Is this situa-
tion an example of voluntary trade? Is it an example of mutu-
ally beneficial trade?
6.12 “Private markets ration goods among consumers in an efficient way.” Explain what this statement means. Does it
imply that there is no basis for thinking that some other distri-
bution would be better?
6.13 Tickets to the National Football League’s (NFL’s) cham- pionship game, the Super Bowl, are sold by the League at a
below-market-clearing price. This policy produces a short-
age. To allocate the relatively scarce tickets, the NFL typically
employs a nonprice rationing scheme. For example, during a
recent season, the League allowed all interested buyers to submit
an application for up to two Super Bowl tickets. The NFL then
conducted a lottery to determine which of the submitted appli-
cations would be honored. Explain why the nonprice rationing
scheme employed by the NFL results in an inefficient distribu-
tion of goods. Can the NFL’s insistence that the state in which
the Super Bowl is played prohibit ticket scalping be justified on
efficiency or equity grounds? Explain.
6.14 Use the analytical framework of this chapter to explain why ticket scalping frequently occurs at college athletic events.
Why does it occur at some events and not at others? Who ben-
efits and who is harmed by this practice? Should steps be taken
to suppress ticket scalping?
*6.15 Denny gives each of his two daughters a glass of milk and six cookies for lunch. The two daughters want to trade so
that one will drink the two glasses of milk and the other will
158 Exchange, Eff ic iency, and Pr ices
C06.INDD 11:8:52:AM 08/06/2014 PAGE 158Trim Size: 203.2 mm X 254 mm
eat the dozen cookies. Does economic analysis imply that
Denny should allow his children to engage in this trade?
6.16 Landing fees at many airports are based on aircraft weight. However, these fees do not accurately measure the
cost associated with a landing or takeoff. This is because the
opportunity cost of a plane using a runway primarily reflects
the amount of time that other aircraft no longer have access to the
runway, and this amount of time is largely independent of
the weight of a plane. If landing fees at heavily used airports
are set below market-clearing levels and are based on aircraft
weight, explain why there may be some significant costs asso-
ciated with such an approach.
6.17 The U.S. government designates particular amounts of the electromagnetic spectrum for certain telecommunications uses:
broadcast television; wireless; radio; and so on. In the interest
of economic efficiency, should trade be allowed between holders
of government-sanctioned rights to the various portions of the
spectrum? That is, should the owner of a certain amount of spec-
trum targeted for radio be allowed to sell the asset to another
individual who believes that the value of the spectrum would be
greater if it was devoted to cellular phone service? Explain.
6.18 “An efficient distribution of goods is always to be preferred to an inefficient distribution.” Is this true‚ false or uncertain?
Explain.
159
C07.INDD 07:6:7:PM 05/30/2014 PAGE 159Trim Size: 203.2 mm X 254 mm
CHAPTER Production
In Chapters 3 through 6 we concentrated on consumer behavior, with the supply of goods taken for granted. Now we begin to analyze the factors that determine the quantities of
goods firms will produce and offer for sale.
We begin our examination of the supply side of the market by assuming that firms maxi-
mize profit. If a firm is interested in maximizing profit, two important steps must be taken.
First, for any potential output level, total cost needs to be minimized. In other words, no
more resources than necessary should be employed to produce any given level of output.
Second, having minimized the cost of producing a given output, the firm must select the
price and corresponding output level that maximize profit. As we will see, the optimal
price–output choice will depend on the market structure in which the firm operates.
This chapter and Chapter 8 focus on the first step that a firm must take to maximize profit.
Namely, we examine the firm’s technology and the input prices the firm confronts to develop
an intuitive rule for minimizing the total cost of producing a given output level. Chapters 9
through 15 address the second step toward profit maximization—choosing the right price–
output combination once the total cost of producing a given output level has been minimized.
To arrive at an intuitive rule for cost minimization, the logical starting place is to identify
the underlying technological relationships between inputs employed and output produced.
This chapter explains how economists represent the technological possibilities available to
the firm. The productivity of inputs is an important determinant of output: it specifies how
much can be produced. As we will see in later chapters, the productivity of inputs underlies
both the cost curves of the firm and the firm’s demand curves for inputs.
Learning Objectives
Establish the relationship between inputs and output. Define total, average, and marginal product, and explain the law of diminishing marginal returns in the short-run setting when at least some inputs are fixed. Investigate the ability of a firm to vary its output in the long run when all inputs are variable. Explore returns to scale: how a firm’s output response is affected by a proportionate change in all inputs. Describe how production relationships can be estimated and some different potential func- tional forms for those relationships.
Memorable Quote “Nothing is particularly hard if you divide it into small jobs.”
—Henry Ford, American inventor and industrialist
7
160 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 160Trim Size: 203.2 mm X 254 mm
7.1 Relating Output to Inputs Inputs—sometimes called factors of production—are the ingredients mixed together by a firm through its technology to produce output. For example, a motion picture studio uses
inputs such as producers, directors, actors, costume and sound designers, technicians, and
the capital invested in its lots, sound stages, and equipment to produce movies.
Inputs may be defined broadly or narrowly. A broad definition might categorize all
inputs as either labor, land, raw materials or capital. When considering some questions,
however, it may be helpful to use more narrow subdivisions within the broader categories.
For example, the labor inputs employed by a firm might include engineers, accountants,
programmers, secretaries, and managers. Raw materials may involve electricity, fuel, and
water. Capital inputs may include buildings, trucks, robots, and automated assembly lines.
The Production Function For any good, the existing technology ultimately determines the maximum amount of out-
put a firm can produce with specified quantities of inputs. By existing technology, we mean the technical or organizational “recipes” regarding the various ways a product can be pro-
duced. The production function summarizes the characteristics of existing technology. The production function is a relationship between inputs and output: it identifies the maxi-
mum output that can be produced per time period by each specific combination of inputs.
Consider the case of a firm that employs two inputs, labor (L) and capital (K), to produce output (Q). Input usage and output are measured as flows: for example, the units of capital and labor employed per day and the firm’s daily output. For simplicity, however, we gen- erally will omit the time period and refer to units of inputs or output rather than units of
inputs or output per relevant time period.
Mathematically, the firm’s production function can be written as:
Q f L K= ( ), . (1)
This function indicates what is technologically efficient—the maximum output the firm can produce from any given combination of labor and capital inputs. The production func-
tion identifies the physical constraints with which the firm must deal. We assume that the
firm knows the production function for the good it produces and always uses this knowl-
edge to achieve maximum output from whatever combination of inputs it employs. This
assumption of technological efficiency may not always be valid, but there is reason to
believe it is generally correct. Any firm operating in a technologically inefficient way is not
making as much money as possible. The firm’s cost of using a given level of inputs is the
same whether or not it uses the inputs wisely, but the revenue from the sale of the product
(and hence the profit) will be greatest when the firm produces the maximum output given
these inputs. Consequently, any profit-oriented firm has an incentive to seek out and use the
best available production technique.
7.2 Production When Only One Input Is Variable: The Short Run
Naturally, the example of a firm using the two inputs of labor and capital to produce out-
put is exceedingly simple and glosses over many of the subtleties of real-world produc-
tion technologies. Still, this simple example allows us to illustrate several key features
of the relationship between inputs and output that do characterize real-world production.
factors of production inputs or ingredients mixed together by a firm through its technology to produce output
production function a relationship between inputs and output that identifies the maximum output that can be produced per time period by each specific combination of inputs
technologically efficient a condition in which the firm produces the maximum output from any given combination of labor and capital inputs
Product ion When Only One Input Is Var iable : The Short Run 161
C07.INDD 07:6:7:PM 05/30/2014 PAGE 161Trim Size: 203.2 mm X 254 mm
One of these features is what happens to output when a firm can vary the use of only one
of its inputs over a given time period.
Resources that a firm cannot feasibly vary over the time period involved are referred to
as fixed inputs. These inputs need not be fixed in the sense that varying their use is literally impossible; rather, they are any inputs that are prohibitively costly to alter in a short time
period. For example, a commercial real estate developer in Mumbai, India, may be largely
unable to supply additional office space over the coming month in response to an increase
in market demand. This is because acquiring land and/or building permits in Mumbai over
such a short time frame is virtually impossible. Samsung’s physical plant provides another
example. In response to strong holiday consumer demand, Samsung might be able to
expand capacity for production of its tablets in a month. Doing so, however, would require
around-the-clock employment of large numbers of engineers and contractors at exorbitant
cost. In that case, practically speaking, the physical manufacturing plant associated with the
Samsung tablet is a fixed input that Samsung will not vary in the event that quick output
adjustments are required.
Suppose that in our simple scenario, the firm is stuck with a certain amount of capital for
the time being and can vary only the number of workers—the amount of labor—that it
employs. For the sake of simplicity, we assume that capital is held constant at three units and
examine how output or total product varies as the firm employs different quantities of labor. Table 7.1 shows a hypothetical relationship between output and various labor quanti-
ties. The first column is included merely to emphasize that the amount of capital input is
held constant at three units regardless of the labor used. The second and third columns
contain the important data, showing how much total product can be produced with alterna-
tive quantities of labor. With zero workers, total product is zero. As the amount of labor
increases, total product rises. One worker combined with three units of capital results in a
total product of five, using two workers raises output to 18, three workers further increases
output to 30, and so on. There is, however, a limit to the total product that the firm can
produce by increasing labor input if capital input is held constant at 3 units. In our example
the limit is reached when eight workers are employed and total product is 49. The eighth
worker adds nothing to output, and using nine workers actually causes output to fall.
Although these figures are hypothetical, the general relationship they illustrate is quite
common. To examine the relationship further, we introduce the concepts of average prod-
uct and marginal product of an input. The average product of an input is defined as the total output (or total product) divided by the amount of the input used to produce that out-
put. For example, three workers produce 30 units of total product, so the average product of
labor is 10 units of output at that employment level. The average product for each quantity
fixed inputs resources a firm cannot feasibly vary over the time period involved
total product the total output of the firm
average product the total output (or total product) divided by the amount of the input used to produce that output
Production with One Variable Input
Amount of Capital
Amount of Labor
Total Product
Average Product of Labor
Marginal Product of Labor
3 0 0 — —
3 1 5 5 5
3 2 18 9 13
3 3 30 10 12
3 4 40 10 10
3 5 45 9 5
3 6 48 8 3
3 7 49 7 1
3 8 49 6.1 0
3 9 45 5 −4
Table 7.1
162 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 162Trim Size: 203.2 mm X 254 mm
of labor is therefore derived by dividing the total product in column 3 by the corresponding
amount of labor in column 2. Note that total product, and thus the average product of labor,
depends on the amount of other inputs—in this case, capital—being used and that the
amount of nonlabor inputs is held constant throughout Table 7.1.
The marginal product of an input represents the change in total output resulting from a one-unit change in the amount of the input, holding the quantities of other inputs con-
stant. To illustrate, when labor is increased from four to five units, total output rises from
40 to 45, or by five units. So the marginal product of labor, when the fifth worker is
employed, is five units of output. What the marginal product of an input measures should
be thoroughly understood. In many applications it is the crucial economic variable,
because most production decisions relate to whether a little more or a little less of an input
should be employed. The way total output responds to this variation is what the marginal
product measures.1
Total, Average, and Marginal Product Curves The information from Table 7.1 can be conveniently graphed. Figure 7.1 shows the result.
(We have assumed that labor and output are divisible into smaller units in drawing the graphs,
so the relationships are smooth curves rather than 10 discrete points.) The total product (TP) curve in Figure 7.1a shows how the output varies with the quantity of labor employed. Just as
indicated in Table 7.1, output increases as more labor is used and reaches a maximum at 49
units, when eight workers are employed; beyond eight workers, output declines.
Figure 7.1b shows the average product (APL) and marginal product (MPL) curves for labor. Note that these curves measure the output per unit of input on the vertical axis rather
than total product, which is what the vertical axis measures in Figure 7.1a. As employment
of labor increases, MPL increases at first, reaches a maximum at two workers, and then declines. The average product of labor also increases at low levels of employment, reaches
a maximum height of 10 units per worker, and then declines.
The two panels of Figure 7.1 highlight the relationship between total and marginal
product. As long as marginal product is positive, total product rises. That is, as long as an
extra unit of labor produces some extra output (however small), the total amount produced
increases (up through seven units of labor in Figure 7.1). When marginal product is nega-
tive, total product falls (beyond eight units of labor), and when marginal product is zero,
total product is at its maximum (at eight units of labor).
A rational producer, of course, will never operate where marginal product is negative
(beyond eight workers in Figure 7.1). This is because employing a variable input at a level
where its marginal product is negative is technologically inefficient. The firm can increase
its total product and lower its production cost (provided that the price of the variable input
is positive) by using less of the variable input.
The Relationship between Average and Marginal Product Curves A definite relationship exists between the average and marginal product curves. When mar-
ginal product is greater than average product, average product must be increasing, as is
shown between one and four units of labor in Figure 7.1b. This relationship follows directly
from the meaning of the terms. If the addition to total product (marginal product) is greater
than the average, the average must rise. Think of the average height of people in a room. If
another person enters who is taller than the average (the marginal height of the extra per-
son is greater than the average), the average height will increase. Similarly, when marginal
marginal product the change in total output that results from a one- unit change in the amount of an input, holding the quantities of other inputs constant
1Note that the marginal product figures in Table 7.1 pertain to the interval between the indicated amount of labor and one unit less. Thus, the marginal product at four units of labor is 10, because total output rises from 30 to 40 when labor increases from three to four units.
Product ion When Only One Input Is Var iable : The Short Run 163
C07.INDD 07:6:7:PM 05/30/2014 PAGE 163Trim Size: 203.2 mm X 254 mm
product is less than average product, average product must decrease, as is shown for labor
beyond four units in the diagram. Because marginal product is greater than average product
when the average is rising, and less than average product when the average is falling, mar-
ginal and average products will be equal when average product is at a maximum.
The Geometry of Product Curves As we saw in discussing Table 7.1, knowing how total output varies with the quantity
of the variable input allows us to derive the average and marginal product relationships.
Similarly, we can use geometrical relationships to derive the average and marginal product
curves from the total product curve.
Total, Average, and Marginal Product Curves (a) The total product curve shows the output produced with various amounts of labor, assuming that other inputs are held constant. (b) Average and marginal product curves are derived from the total product curve.
APL
TP
MPL
(b)
(a)
Units of output
49
40
18
0
Units of output per worker
13
10
0 2 4 Units of labor
Units of labor
8
842
Figure 7.1
164 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 164Trim Size: 203.2 mm X 254 mm
Figure 7.2 illustrates how average product is derived geometrically from the total prod-
uct curve, TP. The average product of labor is total output divided by the total quantity of labor. At point B on the total product curve, average product is equal to Q2/L2, or nine units of output per worker. Note, however, that Q2/L2 equals the slope of the straight-line segment 0B drawn from the origin to point B on the total product curve. Thus, the average product at a particular point is shown geometrically by the slope of a straight line from the origin to that point on the total product curve.
Now consider points B, C, and D on the total product curve. As output expands from B to C to D, the straight-line segments 0B, 0C, and 0D become successively steeper, so that the slopes of these segments become successively greater. This shows that the average
product of labor rises over this region. At point D, in fact, average product reaches a max- imum, because the straight-line segment 0D is the steepest such segment from the origin that still touches the total product curve (0D is tangent to the total product curve at point D). Beyond point D, the slope of the straight-line segment connecting the origin to the total product curve begins to decline as the employment of labor is increased. For exam-
ple, at point H, the slope of straight-line segment 0H is less than the slope of segment 0D. This indicates that the average product of labor is smaller at point H than at point D.
Marginal product measures how much total output changes with a small change in the
use of an input, holding the use of other inputs constant. Figure 7.2 also shows how marginal
product is derived geometrically from the total product curve. The marginal product of labor at any point on the total product curve is shown by the slope of the total product curve at that point. The slope of the total product curve is, in turn, equal to the slope of a line tangent to the curve. At point A, for example, we have drawn a line tangent to the total product curve, with a slope of 5/1. Thus, the marginal product of labor at point A is five units of output.
The steeper the total product curve, the faster output rises when more input is used, which
implies a larger marginal product. In the diagram, marginal product rises as we move up the
curve from the origin to point B, but it declines (the slope becomes smaller) as we go beyond point B. At point B the total product curve is steepest, and marginal product is at a maximum. Beyond point B output rises less and less when more and more input is used. Note that at point D marginal and average product are equal since the slope of the total product curve (marginal product) equals the slope of a straight-line segment from the origin (average product). This is
the graphic representation of the proposition, noted earlier, that when average product is at a
maximum, marginal and average product are equal. When marginal product falls to zero at point
H, as implied by the zero slope of the total product curve there, total output is at a maximum.
APPLICATION 7.1
Suppose that this term in college you have been really challenging yourself. In addition to studying microeconom- ics, you enrolled in courses on theoretical physics, biomedi- cal engineering, advanced calculus, conversational Latin, and graduate-level noumenal philosophy. While you learned an incredible amount, your GPA for the term has not entirely reflected the knowledge that you have acquired, coming in at a paltry 1.5 on the traditional 4.0 scale.
Will your overall college GPA go up or down on account of your marginal term’s GPA being 1.5? As you might expect from having studied as much microeconomics as you have by now, the correct answer (as to so many other questions)
What the Marginal–Average Relationship Means for Your Grade Point Average (GPA)
is “It depends.” Namely, if your overall GPA (your average GPA across your previous terms) was greater than 1.5 (the GPA you earned in this, marginal, term), then your marginal term’s GPA performance of 1.5 will pull down your overall GPA, as averaged across all terms. If, however, your overall GPA was less than 1.5 heading into the most recent term, then your marginal term’s GPA performance of 1.5 will pull up your overall GPA. And, if your overall GPA was equal to 1.5 heading into the most recent term, then your overall GPA will remain unchanged on account of your marginal term’s GPA performance being equal to your overall GPA averaged across all terms.
Product ion When Only One Input Is Var iable : The Short Run 165
C07.INDD 07:6:7:PM 05/30/2014 PAGE 165Trim Size: 203.2 mm X 254 mm
The Law of Diminishing Marginal Returns The shapes of the product curves in Figures 7.1 and 7.2 reflect the law of diminishing marginal returns, an empirical generalization about the way output responds to increases in the employment of a variable input. The law of diminishing marginal returns holds that
as the amount of some input is increased in equal increments, while technology and other
inputs are held constant, the resulting increments in output will eventually begin to
decrease. Put more briefly, the law holds that beyond some point the marginal product of
the variable input will decline.
The law of diminishing marginal returns makes intuitive sense. If we begin with one
worker and three units of capital, that worker must be responsible for everything. A second
worker may increase total product more than the first worker does if there are advantages
to teamwork and the division of labor in producing output. For example, take the case of a
firm that delivers pianos to various buyers in a city using the two inputs of trucks (assumed
to be fixed at one unit) and labor. A piano is hard for just one worker to move. Two work-
ers working as a team, however, are likely to be more productive than one trying to do the
work alone. More teamwork and division of labor are possible as additional workers are
employed, but, eventually, the marginal product of additional units of labor falls, because
the workers’ tasks become redundant and they get in each other’s way. Imagine 20 workers
crowded around and trying to move a single piano! Ultimately, the marginal product of an
extra unit of labor becomes negative when there are so many workers relative to the other,
fixed inputs that their efforts actually lower total output. If 20 workers had to squeeze into
the firm’s one moving truck (the fixed input), there would be little room left for pianos.
law of diminishing marginal returns a relationship between output and input that holds that as the amount of some input is increased in equal increments, while technology and other inputs are held constant, the resulting increments in output will decrease in magnitude
Deriving Average and Marginal Product The average product of labor equals the slope of a straight-line segment from the origin to any point on the total product curve. Thus, at point B, the average product is shown by the slope of straight-line segment 0B, or nine units of output per unit of labor. The marginal product of labor is equal to the slope of the total product curve at each point. Thus, at point A, the marginal product is equal to five units of output per unit of labor.
C
D
F H
TP
B
A
Q2
L2
Total product
(18)
(2)
0
5
1
Labor
Figure 7.2
166 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 166Trim Size: 203.2 mm X 254 mm
In Figure 7.1b, diminishing marginal returns set in when the amount of labor increases
beyond two workers. Each additional worker beyond the second adds less to total product
than the previous one; the marginal product curve slopes downward. Note that the law of
diminishing marginal returns does not depend on workers being different in their produc-
tive abilities. We are assuming that all workers are alike.
It is entirely possible that diminishing marginal returns will occur from the very beginning,
with the second unit of labor adding less to total output than the first. More commonly, mar-
ginal returns increase at very low levels of output and then decline, as in Figure 7.1b. Note also
that the law of diminishing returns applies to labor as long as the marginal product of labor
curve is downward sloping (as it is beyond two workers). The height of the curve does not
have to be negative (as it is beyond eight workers) for the law of diminishing returns to hold.
In applying the law of diminishing marginal returns, two conditions must be kept in
mind. First, some other input (or inputs) must stay fixed as the amount of the input in ques-
tion is varied. The law does not apply, for example, to a situation where labor and capital
are the only inputs and the usage of both is increased. It does apply if the amount of capital
is held constant while workers and raw materials, for example, are varied. The key point
is that some important input is not varied. Second, technology must remain unchanged.
A change in technical know-how would cause the entire total product curve to shift.
7.3 Production When All Inputs Are Variable: The Long Run
By investigating the case where one input (capital) is fixed, the previous section was in
fact focusing on the short-run output response by a firm. The short run is defined as a period of time in which changing the employment levels of some inputs is impractical. By contrast, the long run is a period of time in which the firm can vary all its inputs. A commercial real estate developer in Mumbai can acquire the additional land and building
permits necessary to supply more office space. Samsung has sufficient time to expand
its capacity to produce tablets. There are no fixed inputs in the long run; all inputs are
variable inputs. Of course, the distinction between the short run and the long run is necessarily some-
what arbitrary. Six months may be ample time for the clothing industry to make a long-
run adjustment to a change in prevailing fashions but insufficient time for the automobile
industry to switch from production of large to small cars. Even for a given industry no
specific time period can be identified as the short run since some inputs may be variable in
short run a period of time in which changing the employment levels of some inputs is impractical
long run a period of time in which the firm can vary all its inputs
variable inputs all inputs in the long run
APPLICATION 7.2
One of the world’s most commonly used drugs‚ caf- feine‚ is a bitter‚ naturally occurring substance found in coffee and cocoa beans‚ tea leaves‚ kola nuts‚ and other plants. Caffeine is ingested when consuming coffee‚ tea‚ soft drinks, or chocolate. Through its ability to stimulate the central nervous system‚ caffeine can heighten physical per- formance‚ mental alertness‚ and wakefulness—a phenomenon well known to college students at exam time. Research indicates‚ however‚ that the law of diminishing returns
The Law of Diminishing Marginal Returns, Caffeine Intake, and Exam Performance
applies to caffeine consumption. For example‚ whereas the first cup of coffee may improve the typical student’s alert- ness and thereby her score on an upcoming test‚ excessive caffeine use results in anxiety‚ irritability‚ and trembling. For most students‚ that is, the tenth cup of coffee is likely to contribute less to performance on a test than does the ninth cup. And drinking a tenth cup may actually lead to a lower test score on account of the jitters produced by the additional caffeine.
Product ion When Al l Inputs Are Var iable : The Long Run 167
C07.INDD 07:6:7:PM 05/30/2014 PAGE 167Trim Size: 203.2 mm X 254 mm
three months, others in six months, and still others only after a year. Despite this unavoid-
able imprecision, the concepts of short run and long run do emphasize that quick output
changes are likely to be accomplished differently from output changes that can take place
over time.
Production Isoquants In the long run all inputs may be varied, so it is necessary to consider all the possibilities iden-
tified by the firm’s production function. When we consider a product produced by using two
inputs, the production options when both inputs can be varied may be shown with isoquants.
An isoquant is a curve that shows all the combinations of inputs that, when used in a techno- logically efficient way, will produce a certain level of output. Figure 7.3 shows several iso-
quants for a firm interested in maximizing output by using the two inputs of capital and labor.
Isoquant IQ18, for example, shows the combinations of inputs that will produce 18 units of output. (Note that the axes measure the quantities of the two inputs used.) Combining five
units of capital with 1 unit of labor will result in 18 units of output (point B); so will two units of capital and three units of labor (point C) or, indeed, any other combination on the IQ18 iso- quant. Isoquants farther from the origin indicate higher output levels.
The Figure 7.3 isoquants portray an important economic assumption: a firm can pro-
duce a particular level of output in various different ways—that is, by using different input
combinations, as indicated by points A, B, C, and D on IQ18. The firm can produce 18 units
isoquant a curve that shows all the combinations of inputs that, when used in a technologically efficient way, will produce a certain level of output
Production Isoquants Production isoquants show how much output a firm can produce with various combinations of inputs. A set of isoquants graphs the production function of the firm. Isoquants have geometric properties that are similar to those of indifference curves: they are downward sloping, nonintersecting, and convex. The slope of an isoquant measures the marginal rate of technical substitution between the inputs. Between points B and D the MRTSLK equals 2K/1L, implying that 1 unit of labor can replace 2 units of capital without reducing the firm’s output.
Figure 7.3
(Lf)
IQ30
IQ40
IQ48
S GF
H L
C
Ê
KÊ
A
ED
B
Units of capital
(Kf) 3
5
2
1
0 1 2
–1
+1
3 4 6 Units of labor
–2
+1
IQ18
168 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 168Trim Size: 203.2 mm X 254 mm
of output with a small amount of capital combined with a relatively large amount of labor
(point A) or with more capital coupled with less labor (point B). For example, a car can be custom-built in a local garage with very little equipment and a great deal of labor, or it can
be produced in a factory with a large quantity of specialized equipment and far less labor.
It should be emphasized that every combination of inputs shown on the isoquants in Figure 7.3 is technologically efficient: each combination shows the maximum output pos-
sible from given inputs. Since a given product can be produced in many different techno-
logically efficient ways, knowing the technological input–output relationships does not by
itself allow us to identify the best, or least costly, input combination to use. To determine
the lowest-cost way to produce a given level of output, we also need to know input costs, as
we will see in the next chapter.
Isoquants are very similar to indifference curves in their characteristics. While indiffer-
ence curves order levels of a consumer’s satisfaction from low to high, isoquants order
levels of a producer’s output. In contrast to indifference curves, however, each isoquant
reflects a measurable output level. As we discussed in Chapter 3, there is no meaning-
ful way to measure the level of satisfaction associated with each indifference curve. The
numerical labels associated with each indifference curve are useful only to the extent that
they show that higher indifference curves reflect higher levels of consumer satisfaction.
Four Characteristics of Isoquants Four characteristics of the isoquants depicted in Figure 7.3 are worth noting.
1. Downward Sloping First, the isoquants must slope downward as long as both inputs are productive—that is,
they both have positive marginal products. If we increase the amount of labor employed
(which, presumably, would by itself raise output) and wish to keep output unchanged, we
need to reduce the amount of capital. This relationship implies a negative slope.
2. Higher Output Levels for Isoquants Further to the Northeast Second, isoquants lying farther to the northeast identify greater levels of output. Assuming, again, that
inputs are productive, using more of both inputs means a higher output.
3. Non-intersecting Third, two isoquants can never intersect. Intersecting isoquants would imply, at the point
of intersection, that the same combination of inputs is capable of producing two different
maximum levels of output—a logical impossibility.
4. Typically Convex to the Origin Fourth, isoquants will generally be convex to the origin. In other words, the slope of an iso-
quant (in absolute value) becomes smaller as we move down the curve from left to right. To
see why this is likely to be true, note that the slope of an isoquant measures the ability of one
input to replace another in production. At point B in Figure 7.3, for example, five units of capital and one unit of labor result in 18 units of output. The input combination at point D, though, can also produce the same output. The slope of the isoquant between B and D is (−2 units of capital)/(+1 unit of labor), meaning that at point B, one unit of labor can replace, or substitute for, two units of capital without affecting output.
Marginal Rate of Technical Substitution (MRTS) Without the minus sign, the isoquant’s slope measures the marginal rate of technical substi-
tution between inputs. The marginal rate of technical substitution of labor for capital (MRTSLK) is defined as the amount by which capital can be reduced without changing out- put when there is a small (unit) increase in the amount of labor. Between points B and D
marginal rate of technical substitution the amount by which one input can be reduced without changing output when there is a small (unit) increase in the amount of another input
Product ion When Al l Inputs Are Var iable : The Long Run 169
C07.INDD 07:6:7:PM 05/30/2014 PAGE 169Trim Size: 203.2 mm X 254 mm
the MRTSLK is two units of capital per one unit of labor, which equals the slope when we drop the minus sign.
Convexity of isoquants means that the marginal rate of technical substitution diminishes
as we move down each isoquant. Between points C and A on IQ18 in Figure 7.3, for exam- ple, the MRTSLK is only one unit of capital per unit of labor, less than it is between points B and D. The assumption of convexity of isoquants, just as with convexity of indifference curves, is an empirical generalization that cannot be proven correct or incorrect on logical
grounds. It does, however, agree with intuition. At point B, capital is relatively abundant, and labor is relatively scarce, compared with point C. Between points B and D, one unit of the scarce input (labor) can replace two units of the abundant input (capital). Moving down
the isoquant, labor becomes more abundant and capital more scarce. It makes sense that it
becomes increasingly difficult for labor to replace capital in these circumstances, and this is
what is implied by the convexity of the curve.
We have been focusing on the long-run scenario in which a firm can vary the use of all of
its inputs, but the isoquants depicted in Figure 7.3 also show that there are diminishing returns
to both labor and capital. For example, if we hold capital constant at three units (as we did
in Section 7.2), and increase the use of labor along line KfS (Kf signifies that capital is fixed), each additional unit of labor beyond two units can be seen to add less and less to output.
Increasing labor from two to three units results in output increasing from 18 to 30 units (from
point D on IQ18 to point E on IQ30 along line KfS), thus implying that the third worker’s mar- ginal product is 12 units of output. Adding a fourth worker raises output from 30 to 40 units
(from point E on IQ30 to point F on IQ40 along line KfS), indicating that the fourth worker’s marginal product is 10 units of output. Since the fourth worker contributes less to output than
does the third worker, there are diminishing returns to labor over this range of employment.
The firm also faces diminishing returns to capital. For example, holding labor employ-
ment constant at four units (Lf where the subscript “f” indicates that labor is being held fixed) and increasing the use of capital from one to two units raises output from 18 to 30
units (from point A on IQ18 to point H on IQ30 along segment LfF). This indicates that the marginal product of the second unit of capital is 12 units of output. Further raising capi-
tal from two to three units, assuming that labor employment is still held constant at Lf, increases output from 30 to 40 units (from point H on IQ30 to point F on IQ40 along seg- ment LfF). Since the third unit of capital has a lower marginal product (10 units of output) than does the second unit of capital (12 units of output), the law of diminishing marginal
returns applies to capital over this range of capital use.
MRTS and the Marginal Products of Inputs The degree to which inputs can be substituted for one another, as measured by the marginal
rate of technical substitution, is directly linked to the marginal productivities of the inputs.
Consider again the MRTS, or slope, between points B and D in Figure 7.3. Between these two points one unit of labor can replace two units of capital, so labor’s marginal product
must be two times as large as capital’s marginal product when the slope of the isoquant
(MRTSLK) is two units of capital to one unit of labor. To check this reasoning, note that between points C and A in Figure 7.3 the slope of the isoquant is unity. Here the marginal products must be equal because the gain in output from an additional unit of labor (that is,
labor’s marginal product) must exactly offset the loss in output associated with a one-unit
reduction in capital (that is, capital’s marginal product).
Thus, the marginal rate of technical substitution, which is equal to (minus) the slope of
an isoquant, is also equal to the relative marginal productivities (MPs) of the inputs:
MRTS K L MP MPLK L K= − =( ) / /Δ Δ . (2)
Note that the isoquant’s slope does not tell us the absolute size of either marginal prod-
uct but only their ratio.
170 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 170Trim Size: 203.2 mm X 254 mm
We can also derive this relationship more formally. In Figure 7.3, consider the slope of
isoquant IQ30 between points E and H, ∆K/∆L. With a move from point E to C, the reduc- tion in capital, ∆K, by itself reduces output from 30 to 18 units. This reduction in output must equal ∆K times the marginal product of capital. For example, if ∆K = −1 unit, and the marginal product of the incremental unit of capital is 12 units of output, reducing the
amount of capital by 1 unit reduces output by 12 units. Expressing the change in output as
∆Q, we have:
Δ ΔQ K MP= × K. (3)
Similarly, when labor increases from point C to H, or by ∆L, output increases by ∆L times labor’s marginal product:
Δ ΔQ L MP= × L. (4)
For a movement along an isoquant, the output decrease from reducing capital must equal
the output increase from employing more labor, so the ∆Q terms are equal. The right-hand terms in the two expressions are thus equal, and, by substitution, we obtain:
Δ ΔK MP L MP× = ×K L. (5)
Then rearranging terms yields the suggested relationship:
Δ ΔK L MP MP/ /= L K. (6)
Using MRTS: Speed Limits and Gasoline Consumption Using isoquants can clarify a wide range of issues. Let’s say a person drives 6,000 miles
per year to and from work. The speed at which the car is driven affects both the amount of
gasoline used (driving faster reduces gas mileage) and the amount of time spent commut-
ing. We can think of gasoline and time as inputs in the production of transportation. Driv-
ing slower means using less gasoline but taking more time to get to work. This relationship
is shown by the isoquant in Figure 7.4. Suppose that the car gets 25 miles per gallon if
driven 60 miles per hour. In that case, commuting at 60 miles per hour uses 240 gallons
of gasoline and 100 hours, as shown by point A. If the car gets 26 miles per gallon when driven 55 miles per hour, commuting at 55 miles per hour uses 231 gallons and 109 hours,
as shown at point B. Driving at the slower speed saves 9 gallons but takes 9 additional hours of commuting time: the MRTS is 9 gallons/9 hours, or 1 gallon per hour.
In debates over whether gas savings justify lower speed limits, this isoquant forces us
to recognize that there is a trade-off between gasoline saved by a lower speed limit and
additional time spent in transit. The trade-off is measured by the MRTS: here 1 gallon of gasoline per hour spent commuting (between A and B). Because reducing the speed limit from 60 to 55 miles per hour means using less of one scarce resource (gasoline) but more
of another (a driver’s time), we cannot determine from the MRTS alone which speed limit is preferable. Put differently, both A and B represent technologically efficient points.
Nonetheless, the MRTS is one critical piece of information in comparing different speed limits. What else do we need to know? Basically, we need to know the relative
importance of the scarce resources, gasoline and time. If gasoline costs $4.00 per gallon,
the 55-mile-per-hour speed limit saves our commuter $36.00. But if the commuter values
time at anything more than $4.00 an hour, the lower speed limit costs the commuter more
in lost time than is saved through reduced gasoline use. Another trade-off is also rel-
evant here: lower speed limits mean greater safety. Once again, the size of the trade-off
between greater safety and time, the MRTS, is important. That trade-off, though, is much harder to measure.
Returns to Scale 171
C07.INDD 07:6:7:PM 05/30/2014 PAGE 171Trim Size: 203.2 mm X 254 mm
7.4 Returns to Scale What relationship exists between output and inputs in the long run? Because all inputs can
be varied in the long run, economists approach this problem by focusing on the overall
scale of operation. Specifically, we look at how output is affected by a proportionate change
in all inputs—for example, when the quantities of both labor and capital are doubled.
In this case three possibilities arise. First, a proportionate increase in all inputs may
increase output in the same proportion; for example, doubling all inputs exactly doubles
output. Here production is said to be subject to constant returns to scale. Second, output may increase in greater proportion than input use: output more than doubles when inputs
double. Production is then subject to increasing returns to scale. Finally, output may increase less than in proportion to input use. We then have decreasing returns to scale.
These are the possibilities, but the actual relationship is not as easy to pin down. Some
factors lead to increasing returns, and others lead to decreasing returns; which ones pre-
dominate in a particular case is an empirical question.
Factors Giving Rise to Increasing Returns To begin with, what factors may give rise to increasing returns?
1. Division and Specialization of Labor First, in a large-scale operation workers can specialize in specific tasks and carry them out more
proficiently than if they were responsible for a multitude of jobs. This factor, the specialization
and division of labor within the firm, was emphasized by the Scottish political economist Adam
constant returns to scale a situation in which a proportional increase in all inputs increases output in the same proportion
increasing returns to scale a situation in which output increases in greater proportion than input use
decreasing returns to scale a situation in which output increases less than proportionally to input use
Isoquant Relating Gasoline and Commuting Time When driving faster reduces gas mileage, there is a conventionally shaped isoquant relating gas consumption and time. The slope, or MRTS, shows the trade-off between gas and time implied by a change in speed.
240
231
0
A (60 mph)
B (55 mph)
Time spent commuting (hours per year)
Gasoline (gallons per year)
109100
6,000 miles
Figure 7.4
172 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 172Trim Size: 203.2 mm X 254 mm
Smith. In Wealth of Nations, Adam Smith noted the increasing returns that division of labor is capable of providing in a business as seemingly simple as the production of pins:2
A workman not educated to this business (which the division of labour has rendered a distinct trade), nor acquainted with the use of the machinery employed in it (to the inven- tion of which the same division of labour has probably given occasion), could scarce, perhaps, with his utmost industry, make one pin in a day, and certainly could not make twenty. But in the way in which the business is now carried on, not only the whole work is a peculiar trade, but it is divided into a number of branches, of which the greater part are likewise peculiar trades. One man draws out the wire, another straights it, a third cuts it, a fourth points it, a fifth grinds it at the top for receiving the head; to make the head requires two or three distinct operations; to put it on, is a peculiar business, to whiten the pins is another; it is even a trade by itself to put them into the paper; and the important business of making a pin is, in this manner, divided into about eighteen distinct opera- tions, which, in some manufacturies, are all performed by distinct hands, though in others the same man will sometimes perform two or three of them. I have seen a small manu- factory of this kind where ten men only were employed, and where some of them conse- quently performed two or three distinct operations. But though they were very poor, they could, when they exerted themselves, make among them about twelve pounds of pins in a day. There are in a pound upwards of four thousand pins.... Those ten persons, therefore, could make among them upwards of forty-eight thousand pins in a day. Each person, therefore, making a tenth part of the forty-eight thousand pins, might be considered as making four thousand eight hundred pins in a day. But if they had all wrought separately and independently, and without any of them having been educated to this peculiar busi- ness, they certainly could not each of them had made twenty, perhaps not one pin in a day; that is, certainly, not the two hundred and fortieth, perhaps not the four thousand eight hundredth part of what they are at present capable of performing in consequence of a proper division and combination of their distinct operations.
As is evident in this famous passage, increases in the scale of production in the pin
industry allowed firms to realize output increases that were significantly more than in pro-
portion to the increases in input use.
2. Arithmetical Relationships Second, certain arithmetical relationships underlie increasing returns to scale. For example, a
100-foot square building (with 10,000 square feet of floor space) requires 400 feet of walls,
but a 100 × 200-foot building, with twice the floor space, requires 600 feet of walls, or only 50 percent more material. For another example, a pipeline’s circumference (and hence the
amount of material that must be employed to create a unit of pipeline) equals the constant
“pi” (approximately 3.14) times twice the radius of the pipeline. In contrast, the volume of
goods such as crude oil that a pipeline is able to carry depends on the unit area of the pipeline,
which equals pi times the pipeline’s squared radius. If a pipeline’s radius is expanded from
1 to 10 feet, therefore, its circumference (and approximate construction cost) will go up by a
factor of 10 while the pipeline’s carrying capacity increases by a factor of 100.
3. Large-Scale Technologies Third, the use of some techniques may not be possible in a small-scale operation. Airline
hubs, magnetic resonance imaging (MRI) machines, an Internet backbone, assembly lines,
direct broadcast satellite television systems, and other similarly complex and expensive
techniques or equipment may be feasible only when output is sufficiently high.
The foregoing three factors (division and specialization of labor, arithmetical relation-
ships, and large-scale technologies) are generally what is meant by a phrase such as the
2Adam Smith, The Wealth of Nations (New York: Modern Library, 1937), pp. 4–5.
Returns to Scale 173
C07.INDD 07:6:7:PM 05/30/2014 PAGE 173Trim Size: 203.2 mm X 254 mm
“advantages of large-scale or mass production.” These factors, however, are inherently lim-
ited: after a certain scale of operation is reached, further expansion makes more economies
impossible. Even the arithmetical factors may be limited: as a building becomes larger, the
ceiling and walls may have to be built with stronger materials; and as a pipeline is enlarged,
stronger materials may have to be employed as well as proportionately greater amounts
spent on pumping crude oil through the pipeline.
Factors Giving Rise to Decreasing Returns Set against the factors leading to increasing returns to scale is one factor that tends to pro-
duce decreasing returns to scale: the inefficiency of managing large operations. With large
operations, coordination and control become increasingly difficult. Information may be lost
or distorted as it is transmitted from workers to supervisors to middle management and on
to senior executives, and the reverse is equally likely. Communication channels become
more complex and difficult to monitor. Decisions require more time to make and imple-
ment. Problems of this sort occur in all large organizations, and they suggest that the mana-
gerial function can be a source of decreasing returns to scale.
The relative importance of the factors leading to increasing and decreasing returns to scale is
likely to vary across industries. As a general rule, increasing returns to scale are likely to apply
when the scale of operation is small, perhaps followed by an intermediate range when constant
returns prevail, with decreasing returns to scale becoming important for large-scale operations.
In other words, a production function can embody increasing, constant, and decreasing returns
to scale at different output levels. In fact, this condition is probably the general case.
Figure 7.5 shows isoquants reflecting such a production function. Because we are talk-
ing about returns to scale, we are interested in how output varies as we move along a ray
from the origin, like 0R in the diagram. Along 0R the proportion of capital to labor is con- stant: the ratio of capital to labor is one-to-one at all points. At low output rates, increasing
returns to scale prevail: when capital and labor are both doubled, in a move between points
APPLICATION 7.3
Returns to scale play an influential role in determining cross-country trade flows. Over the last several decades, an increasing share of world trade has occurred between countries with similar characteristics in terms of the rela- tive productivity of different industries within those coun- tries. In light of this, why is the production of goods such as aircraft and cars concentrated in a few countries, and why does it occur at a relatively large scale? For example, aircraft manufacturing is dominated by the two major firms Boe- ing (United States) and Airbus (European Union) comple- mented by smaller, but still sizable, producers of regional
Returns to Scale and Cross-Country Trade Flows
jets such as Embraer (Brazil) and Bombardier (Canada). Car manufacturing is dominated by relatively large firms concentrated in a few countries such as Japan, Korea, Ger- many, the United States, England, China, Italy, France, and Sweden.
Trade theorists such as Nobel-Prize-winning econo- mist Paul Krugman argue that global trade flows can be explained both by consumers’ preferences for diversity and increasing returns to scale.3 Consumers’ preferences for diversity help explain why companies such as BMW, Volvo, Lexus, and Jaguar co-exist, and these brands get traded across countries. Increasing returns to scale explain the rela- tively large size of auto manufacturing firms.
Krugman further points out the importance of a “home- market effect:” everything else being equal, countries with a larger demand for a good shall, at equilibrium, produce more than a proportionate share of that good globally and thus be a net exporter. The home market effect explains the global preeminence of firms such as Boeing and Airbus.
3This application is based on “Paul Krugman” in Wikipedia. In the following chapter, we will introduce the broader concept of economies of scale which allows for input proportions to vary as output levels are scaled up. The concept of economies of scale is analogous to increasing returns to scale but provides a broader production-theory-based justification for the observed pattern of cross-border trade flows.
174 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 174Trim Size: 203.2 mm X 254 mm
A and D, output more than doubles (that is, it quadruples since we move from IQ10 to IQ40). Between points D and F, a range of constant returns to scale occurs: increasing both capital and labor by a half, in a move from D to F, increases output by exactly a half (that is, from IQ40 to IQ60). Finally, beyond point F, decreasing returns to scale result: a doubling of inputs increases output by only one-third (that is, from IQ60 to IQ80) in a move from F to H.
Figure 7.5 shows that as inputs are increased proportionately, the spacing of isoquants
provides a graphical method of ascertaining returns to scale. With increasing returns to
scale, that is, isoquants become closer and closer to one another as inputs are scaled up
proportionately (that is, moving from A to D). The spacing between isoquants is equidistant with constant returns to scale (moving from D to F). And the spacing between isoquants grows farther apart with decreasing returns to scale (moving from F to H).
Of course, saying that returns to scale generally will be increasing at first, then constant,
and then decreasing is not saying a great deal. The exact output range over which these
relations hold is very important; as we will see in the next chapter, it helps determine the
number of firms that can survive in an industry.
7.5 Functional Forms and Empirical Estimation of Production Functions
As with demand, production relationships can be estimated through surveys, experimen-
tation, or regression analysis. For example, a former student of ours was employed as an
intern by McKinsey & Company and assigned to a client interested in entering the pig
Returns to Scale The spacing of isoquants indicates whether returns to scale are increasing, constant, or decreasing. From A to D, there are increasing returns to scale; from D to F, constant returns to scale; and beyond F, decreasing returns to scale.
IQ80
IQ70
IQ60
IQ50
IQ40 IQ30IQ20IQ10
E
F
G
H
R
B
A
D
Capital
Labor12642
2
4
6
12
0
C
Figure 7.5
Funct ional Forms and Empir ical Est imat ion of Product ion Funct ions 175
C07.INDD 07:6:7:PM 05/30/2014 PAGE 175Trim Size: 203.2 mm X 254 mm
chow (food for pigs) business in certain Midwestern states. One of his tasks was to deter-
mine the extent of increasing returns to scale in the production of pig chow. That is, over
what range would output continue to go up more than in proportion to the increase in over-
all input use? The easiest way for the student to determine this range was through telephone
surveys of existing producers. Even though most existing producers were reluctant to talk
to him regarding the size of their operations for fear of releasing trade secrets, a surprising
number provided the relevant data.
Regression analysis offers another method for estimating production functions.
Of course, as noted in Chapter 4, such a method is not without its difficulties. Differ-
ences in technology across firms must be taken into account. Measures of the amount
of each input employed by a firm may not be easy to calculate. In the case of “labor,”
for instance, most firms employ a wide variety of different types of labor (engineers,
clerical assistants, accountants, and so on) at different wage rates. The measurement of
output may also involve some difficulties. For example, organizations may not produce
a single output, as in the case of universities, which supply both research and teach-
ing services. (We discuss multiproduct firms in Chapter 8.) Moreover, firms may have
access to the same technology but face different regulatory environments. The ability
to transform a given amount of inputs into output may be more limited in a restrictive
regulatory environment.
Linear Forms of Production Functions In employing regression analysis, care also must be exercised in selecting a functional form
for the relationship between inputs and output. Take the case of the following linear rela-
tionship between output (Q) and the two inputs of labor (L) and capital (K):
Q a bL cK.= + + (7)
Such a linear production function is straightforward to interpret and easy to estimate,
but presumes that the law of diminishing returns does not apply to either input. To see why,
suppose that the estimated intercept and coefficients are ˆ , ˆ ˆa b c= = =0 4 3and . (As noted in Chapter 4, the ^ signifies an estimated value.) If we start off employing one unit of both
inputs, the estimated output (Q̂) would be 7:
ˆ ˆ ˆ( ) ˆ( ) ( ) ( ) .Q a b= + + = + + =1 1 0 4 1 3 1 7c (8)
Fixing the level of capital at one unit, and varying the level of labor to two units would
increase output by four units, to 11:
ˆ ˆ ˆ( ) ˆ( ) ( ) ( ) .Q b c= + + = + + =a 2 1 0 4 2 3 1 11 (9)
Upon reaching this output level, varying labor further to three would increase output an
additional four units, to 15:
ˆ ˆ ˆ( ) ˆ( ) ( ) ( ) .Q a b c= + + = + + =3 1 0 4 3 3 1 15 (10)
And so on. The law of diminishing returns thus can be seen not to apply to labor because
each additional unit of labor does not add a diminishing amount but the same amount—four
units—to output. An analogous result applies to capital. Holding fixed the level of labor,
say at 1 unit, each additional unit of capital increases output by a constant (ĉ = 3), rather than a diminishing, amount.
176 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 176Trim Size: 203.2 mm X 254 mm
Multiplicative Forms of Production Functions: Cobb–Douglas as an Example Of course, there are more elaborate mathematical forms of production functions that do not
imply constant marginal products for inputs. Among the most common is the Cobb–Doug- las production function.4 In the case of our two-input example, the Cobb–Douglas pro- duction function takes this form:
Q aL Kb c= . (11)
Such a multiplicative form allows the law of diminishing returns to either apply or not
apply to individual inputs. To see why, suppose that the estimated constant a and powers associated with the inputs labor and capital (b and c, respectively) are ˆ , ˆ .a b= =2 0 5 and ĉ = 1. If we start off employing 1 unit of both inputs, the estimated output (Q̂) would be 2:
ˆ ˆ ˆ ˆ .Q L Kb c= a = 2(1 )(1 ) = 2.5 10 (12)
Fixing the level of capital at one unit and varying the level of labor to two units would
increase output by 0.83 units to 2.83:
ˆ ˆ . . ˆ ˆ .Q aL Kb c= = 2(2 )(1 ) 2(1 414)(1) 2 83.10 5 ≈ ≈ (13)
Upon reaching this output level, varying labor further to three units would increase output
by 0.63 units from 2.83 to 3.46:
ˆ ˆ ( )( ) ( . )( ) . . ˆ ˆ .Q aL Kb c= = ≈ ≈2 3 1 2 1 732 1 3 460 5 1 (14)
The law of diminishing returns can thus be seen to apply to labor for the input levels we
have considered because, holding constant employment of capital, the third unit of labor
adds less to total output (0.63 units) than does the second unit (0.83 units).
In the case of capital, however, the law of diminishing returns does not apply for the
assumed Cobb–Douglas production function and estimated constant a and powers b and c. Suppose that we start off once again by employing one unit of both inputs. As we have seen
before, the estimated output (Q̂) is two:
ˆ ˆ . ˆ ˆ .Q aL K= = =b c 2( )( )1 1 20 5 1 (15)
Now instead of holding capital constant, let’s fix labor at one unit and vary the level of
capital to two units. Total output would increase by two units to four:
ˆ ˆ ( )( ) ( )( ) . ˆ ˆ .Q aL Kb c= = = =2 1 2 2 1 2 40 5 1 (16)
Upon reaching this output level, varying capital further to three units would increase output
by two units to six:
ˆ ˆ ( )( ) ( )( ) . ˆ ˆ .Q aL Kb c= = = =2 1 3 2 1 3 60 5 1 (17)
Cobb–Douglas production function a production function that does not imply constant marginal products for inputs
4This type of production function is named after Charles W. Cobb, a mathematician, and Paul H. Douglas, an economist and U.S. senator. Cobb and Douglas did pioneering work in estimating production functions in the early part of the twentieth century.
Funct ional Forms and Empir ical Est imat ion of Product ion Funct ions 177
C07.INDD 07:6:7:PM 05/30/2014 PAGE 177Trim Size: 203.2 mm X 254 mm
The law of diminishing returns thus can be seen not to apply to capital because, holding
constant employment of labor, the third unit of capital raises output by the same amount
(two units) as does the second unit of capital.
Exponents and What They Indicate in Cobb–Douglas Production Functions In general, if the power associated with an input in a Cobb–Douglas production function is less
than unity, the law of diminishing returns applies to that input over all possible levels of input
usage. (Do you see why?) If the power associated with an input is equal to or greater than unity,
the law of diminishing returns does not apply to that input.
Furthermore, the sum of the powers associated with the inputs in a Cobb–Douglas pro-
duction function has economic significance. If the sum of the powers exceeds unity (that is,
b + c > 1), the production function is characterized by increasing returns to scale. If the sum of the powers equals unity (b + c = 1), constant returns to scale apply. Decreasing returns to scale apply when the sum of the powers is less than unity (b + c < 1).
To see why the sum of the powers associated with the inputs in a Cobb–Douglas pro-
duction function is related to returns to scale, consider what would happen to output if we
scaled up employment of all inputs by some factor, s. The scaling factor s is some number greater than unity because we are contemplating “scaling up” use of all inputs. For exam-
ple, if we considered doubling all inputs, s = 2. To check on returns to scale, we want to compare the output we get when we scale up all inputs by s:
a Ls Ksb c( ) ( ) (18)
with the original output, Q, scaled up by the same factor s:
sQ saL Kb c= . (19)
Written side by side, we are comparing whether the output we get when we scale up all
inputs is more than the scaled-up initial output:
a L K saL Kb c c( ) ( ) .s s versus b (20)
With some simple rearrangement, the comparison boils down to the following:
s aL K saL Kb c b c b c+ versus . (21)
If the sum of the powers associated with the inputs of labor and capital exceeds unity (that
is, b + c > 1), the above comparison indicates that scaling up all inputs (the left-hand side) will get us more than the scaled-up initial output (the right-hand side). This is the case
when increasing returns apply. For example, if the sum of the powers associated with the
inputs exceeds unity and s = 2, doubling all inputs will get us more than double the initial output.
If the sum of the powers associated with the inputs labor and capital equals unity
(b + c = 1), scaling up all inputs (the left-hand side of the comparison) will get us the same amount as the scaled-up initial output (the right-hand side). This holds in the case of con-
stant returns to scale. If s = 2 in such a case, doubling all inputs will produce an output that is exactly double the initial output.
Finally, if the sum of the powers associated with the inputs labor and capital is less than
unity (b + c < 1), scaling up all inputs (the left-hand side) will get us less than the scaled-up initial output (the right-hand side) and decreasing returns to scale apply. Were we to double
the use of all inputs in such a case (that is, s = 2), output would less than double.
178 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 178Trim Size: 203.2 mm X 254 mm
SUMMARY
There are two relationships between the quantities of
inputs used and the amount of output produced. In the
first, the quantities of some inputs are not changed (fixed
inputs), while the quantities of other inputs (variable
inputs) are. This is normally a short-run output response,
when varying the quantities of some inputs is not practical.
In the second relationship, the quantities of all inputs can
be varied, which is normally the case when long-run output
responses are considered.
With some inputs held fixed, the total product curve
shows the relationship between the quantity of the vari-
able input and output.
The law of diminishing marginal returns holds that
beyond some level, the marginal product of the vari-
able input will decline as more of the input is used. This
law implies that the total, average, and marginal product
curves will have the general shapes shown in Figure 7.1.
Isoquants depict all input combinations that will produce a
given output level. They show the relationship between inputs
and output when all inputs can be varied.
A set of isoquants is effectively a graphical represen-
tation of the firm’s production function.
Isoquants and indifference curves have the same geo-
metric characteristics.
The marginal rate of technical substitution shows the
technological feasibility of trading one input for another
and is equal to the slope of an isoquant.
Returns to scale refer to the relationship between
a proportionate change in all inputs and the associated
change in output. If output increases in greater propor-
tion than input use, production is said to be subject to
increasing returns to scale.
Constant and decreasing returns to scale are defined
analogously. In general, increasing returns to scale are
common at low levels of output for a firm, possibly
followed by constant returns over a certain range.
At high levels of output, decreasing returns to scale
will exist.
Although it is not without its difficulties, regression
analysis offers one means for estimating the relationship
between inputs employed and output.
Among other things, production functions can take a
linear or multiplicative form. The Cobb–Douglas pro-
duction function is an example of a multiplicative form.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
7.1 Fill in the spaces in the accompanying table associated with the firm William Perry, Inc., that delivers refrigerators in the
Chicago area, using the two inputs of labor and trucks.
Number of Trucks
Amount of Labor
Total Output
Average Product of Labor
Marginal Product of Labor
2 0 0 — —
2 1 75
2 2 100
2 3 100
2 4 380
2 5 50
2 6 75
7.2 State the law of diminishing marginal returns. How is it illustrated by the data in the table of the preceding question?
There is a proviso to this law that certain things be held con-
stant: What are these things? Give examples of situations
where the law of diminishing marginal returns is not applicable
because these “other things” are likely to vary.
*7.3 If the total product curve is a straight line through the ori- gin, what do the average product and marginal product curves
look like? What principle would lead you to expect that the
total product curve would never have this shape?
*7.4 Is it possible that diminishing marginal returns will set in after the very first unit of labor is employed? What do the total,
average, and marginal product curves look like in this case?
*7.5 Deloitte is thinking of hiring an additional employee. Should the firm be more concerned with the average or the
marginal product of the new hire?
7.6 Consider your time spent studying as an input in the pro- duction of total points on an economics test. Assume that other
inputs (what could they be?) are not varied. Draw the total,
average, and marginal product curves. Will they have the gen-
eral shapes shown in Figure 7.1? Why or why not?
7.7 Define isoquant. What is measured on the axes of a diagram with isoquants? What is the relationship between the isoquant
map and the production function?
C07.INDD 07:6:7:PM 05/30/2014 PAGE 179Trim Size: 203.2 mm X 254 mm
Review Quest ions and Problems 179
7.8 Assume that the marginal product of each input employed by Microsoft depends only on the quantity of that input
employed (and not on the quantities of other inputs), and that
diminishing marginal returns hold for each input. Explain why
Microsoft’s isoquants must be convex if these assumptions
hold.
*7.9 When United Airlines uses equal amounts of pilots and mechanics, must the isoquant drawn through this point of input
usage have a slope of −1? Could the isoquant have a slope of −1? If so, what would this characteristic tell us?
7.10 Isoquants are downward sloping, nonintersecting, and convex. Explain the basis for each of these characteristics.
*7.11 For a particular combination of capital and labor we know that the marginal product of capital is six units of out-
put and that the marginal rate of technical substitution is 3
units of capital per unit of labor. What is the marginal product
of labor?
7.12 Show how a total product curve for an input can be derived from an isoquant map. Why does the question specify
“a” total product curve rather than “the” total product curve?
7.13 If the firm’s isoquants in Figure 7.3 were straight lines, what would this imply about the two inputs?
7.14 In the commuting example in the text, we assumed that the car in question got 25 miles per gallon if driven at 60
mph and 26 miles per gallon if driven at 55 mph. If the car
gets 1 more mile per gallon for each 5-mile-per-hour reduc-
tion in speed, will the isoquant be convex? Support your
answer by identifying several more points on the isoquant
in Figure 7.4.
7.15 Does the concept of technological efficiency permit us to determine at which point on an isoquant a firm should operate?
7.16 Suppose that the number of points on an economics mid- term (P) can be characterized by the following production function:
P H B= −5 4 ,
where H is the number of hours spent studying for the exam and B is the number of beers consumed the week before the exam. Does the law of diminishing returns apply to H? To B? What does the typical isoquant look like for such a production
function? Is the production function characterized by increas-
ing, decreasing or constant returns to scale? Explain your
answers.
7.17 Among the key inputs to a houseplant’s success are light‚ temperature‚ humidity‚ soil quality‚ nutrients‚ pest control‚ and
water. Explain why increased use of any of the inputs such as
water is likely to be subject to the law of diminishing marginal
returns.
7.18 American Airlines produces round-trip transporta- tion between Dallas and San Jose using three inputs: capital
(planes), labor (pilots, flight attendants, and so on), and fuel.
Suppose that American’s production function has the following
Cobb–Douglas form:
T aK L F K L F= =b c d 0.25 0.2 0.550.02 ,
where T is the number of seat-miles produced annually, K is capital, L is labor, and F is fuel. a. If American currently employs K = 100, L = 500, and F =
20,000, calculate the marginal products associated with K, L, and F.
b. What is American’s marginal rate of technical substitution (MRTS) between K and L? How about the MRTS between K and F? Should American try to ensure that all its MRTSs are equal? Explain.
c. Does the law of diminishing returns apply to K in the pro- duction of seat-miles between Dallas and San Jose by
American? to L or F? Explain. Would the law of diminish- ing returns apply to L if c = −0.2 instead of 0.2? If c = 1.2?
d. Given that the exponent associated with F is larger than the exponent associated with L, would it be wise for American to spend all its money on either fuel or capital and none on
labor? Explain.
e. Does American’s production function exhibit constant, increasing or decreasing returns to scale? Explain. How
would your answer change if c = −0.2 instead of 0.2? If c = 1.2?
f. Does the law of diminishing returns imply decreasing returns to scale? Explain. Would decreasing returns to scale
imply the law of diminishing returns?
g. In the real world, do you think that the production of seat- miles between Dallas and San Jose is characterized by a mul-
tiplicative, Cobb–Douglas technology? If not, explain the
nature of the production function that might characterize a
typical firm producing seat-miles in this city-pair market.
7.19 Economists classify production functions as possess- ing constant, decreasing or increasing returns to scale. Yet,
from a cause-and-effect point of view, it is not readily appar-
ent why decreasing returns to scale should ever exist. That is,
if we duplicate an activity we ought to get duplicate results.
Hence, if we truly duplicate all of the inputs, we ought to get
double the output. Can you reconcile the apparent contradic-
tion between this logic and the expectation of the economist
that beyond certain output ranges firms will confront decreas-
ing returns to scale?
7.20 Suppose that you estimated a production function for various professional tennis players. The measure of output
is the percentage of matches played by a player that are won
by the player. Inputs include the average number of hours per
week spent practicing tennis. Suppose that your results indi-
cate that the marginal product associated with practice time
is 0.07 for Anna Kournikova, 0.09 for Venus Williams, and
0.16 for Maria Sharapova. If the law of diminishing marginal
returns holds, which of the three players has spent the most
time practicing?
7.21 In a year when the Los Angeles Lakers won the National Basketball Association championship, two of the leading
180 Product ion
C07.INDD 07:6:7:PM 05/30/2014 PAGE 180Trim Size: 203.2 mm X 254 mm
Lakers’ players, Shaquille O’Neal and Kobe Bryant, made 57
and 44 percent, respectively, of their field goal shots. Given
these different marginal products, wouldn’t the Lakers have
done even better in terms of overall scoring had O’Neal taken
more shots and Bryant fewer?
7.22 A fellow student states that it is best to stop studying once you reach the point of diminishing returns with regard to the
number of hours spent studying. Assess the validity of her
statement.
7.23 Nineteenth-century British economist Thomas Malthus reasoned that because the amount of land is fixed, as population
grows and more labor is applied to land, the productivity of
labor in food production would decline, leading to widespread
famine. This prediction is what led economics to be called the
“dismal science.” Malthus’s prediction failed to come to pass
as advances in technology, such as the Green Revolution,
greatly increased labor productivity in food production. Do
such technological advances contradict the law of diminishing
marginal returns?
7.24 In 1998‚ Mark McGwire hit 70 home runs while play- ing for the St. Louis Cardinals. In 1999‚ McGwire hit 65 home
runs. This decrease in marginal (home runs per season) product
led to an associated decrease in McGwire’s average (home runs
per season) product. Is this true‚ false or uncertain? Explain.
7.25 In the early days of People Express‚ the top management team at the airline was personally involved in the training and
selection of employees. This participation was key to instill-
ing spirit and dedication among the staff‚ and the organiza-
tional culture that resulted led to the airline’s successful initial
growth. (Started in 1981‚ People Express grew in a few years
from three to 80 planes‚ reached both U.S. coasts and Europe‚
and earned positive profits.) Explain why decreasing returns
might have set in with continued expansion, however‚ and thus
ultimately led to the company’s demise in 1986.
181
C08.INDD 11:11:1:AM 08/06/2014 PAGE 181Trim Size: 203.2 mm X 254 mm
The Cost of Production8
In determining how to transform inputs into output, a firm must not only determine what inputs are necessary to produce various levels of output, as we saw in Chapter 7, but it must
also consider the cost of acquiring the inputs used to produce the product. Because produc-
tion cost is important in determining a firm’s output, we should understand several aspects
about a firm’s cost: how cost is defined (what it includes), how cost varies with the output
produced, and how cost can be empirically estimated. After examining cost, we develop an
intuitive rule for minimizing the cost of producing any given output. Then, in Chapter 9, we
will discuss how revenue from a product’s sale can be incorporated into the analysis and
how production cost and sales revenue jointly determine the profit-maximizing output for
the competitive firm.
Memorable Quote “It results from our having such quantities of land to waste as we please. In Europe the object is to make the most of their land, labor being abundant; here it is to make the most of our labor, land being abundant.”
—President Thomas Jefferson explaining why Europeans tended to farm their agricultural land more intensely than Americans in the eighteenth century
Learning Objectives
Delineate the nature of a firm’s cost—explicit as well as implicit. Outline how cost is likely to vary with output in the short run and various measures of short- run cost. Detail the typical shapes of a firm’s short-run cost curves. See how a firm will choose to combine inputs in its production process in the long run when all inputs are variable. Show how input price changes affect a firm’s cost curves. Differentiate between a firm’s long-run and short-run cost curves. Explain the impact of learning by doing on production cost. Understand how the minimum efficient scale of production is related to market structure. Describe how cost curves can be applied to the problem of controlling pollution. Cover economies of scope—is it cheaper for one firm to produce products jointly than it is for separate firms to produce the same products independently? Overview how cost functions can be empirically estimated through surveys and regression analysis.
CHAPTER
182 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 182Trim Size: 203.2 mm X 254 mm
8.1 The Nature of Cost Although a firm’s cost of production is commonly thought of as its monetary outlay, this view
of cost is too narrow for our purposes. Because, as economists, we wish to study the way cost
affects output choices, employment decisions, and the like, cost should include several factors
in addition to outright monetary expenses. As discussed in Chapter 1, the relevant cost to a
firm of using its resources in a particular way is the opportunity cost of those resources—
the value the resources would generate in their best alternative use. Opportunity cost reflects both explicit and implicit costs. Recall that explicit costs arise from transactions in which the firm purchases inputs or the services of inputs from other parties; they are usually recorded
as costs in conventional accounting statements and include payroll, raw materials, insurance,
electricity, interest on debt, and so on. Implicit costs are those associated with the use of the firm’s own resources and reflect the fact that these resources could be employed elsewhere.
Although implicit costs are difficult to measure, we must take them into account in analyzing
the actions taken by a firm. Consider farmers, who not only personally supply manual labor
to their operations but also frequently own the hundreds of thousands of dollars in land and
machinery they use. Implicit costs are associated with the use of the land for farming even if
no explicit payment is being made for it. Land, for instance, can be sold or leased; the revenue
that a farmer would realize through sale or lease is sacrificed when the farmer uses the land
for farming and so reflects an implicit cost associated with its use.
For the modern large corporation, the most important implicit cost is associated with the use
of the firm’s productive assets, its capital. These resources are ultimately owned by the stock-
holders, who have provided funds to the corporation and expect to receive a return on their
investment. Let’s suppose that the stockholders could have invested their funds elsewhere and
earned an annual return of 10 percent. If the corporation does not pay at least 10 percent of the
invested capital (in dividends or higher market valuation of the stock), the stockholders have an
incentive to withdraw their investments (by selling their stock) and place the funds elsewhere
where they are guaranteed a 10 percent return. Although there is no contractual agreement, an
implicit cost is associated with the firm’s use of its own (or, more precisely, its stockholders’)
capital: the same resources could have earned 10 percent if invested elsewhere. Viewing the rate
of return that could be obtained from investing elsewhere as an implicit cost means that an aver-
age market return on investment is treated as part of the firm’s normal production cost.
That a firm’s production cost will equal the opportunity cost of the firm’s resources
emphasizes that those resources can be used to produce many things: steel can be used to
make cars, refrigerators, homes or many other goods; mechanical engineers can be used in
the production of roads, airplanes, and toys, among other things. When a resource is used
to produce one good, it can’t be used to produce something else. For example, when we
use steel to make a car, we sacrifice other products that could have been produced with
that steel. For another example, consider the firm’s employment of workers. The firm must
pay workers enough that they don’t leave and go elsewhere. If the firm’s wage is less than
a worker’s opportunity cost—that is, what the worker could earn in the best alternative
job—the worker will quit and take the better-paying alternative. Thus, to hire a worker who
could work elsewhere for $35,000, the firm must pay at least $35,000, and this wage is the
opportunity cost of the worker’s services.
8.2 Short-Run Cost of Production A firm’s production cost will vary with its rate of output. Exactly how cost is likely to vary
with output in the short run is discussed in this and the following section. First, though, we
will work through a numerical example that shows how short-run cost varies with output
Short-Run Cost of Product ion 183
C08.INDD 11:11:1:AM 08/06/2014 PAGE 183Trim Size: 203.2 mm X 254 mm
for a hypothetical firm. For the moment, understanding why cost varies with output exactly
as it does in the example is not essential; our purpose is to introduce the terminology and
explain the relationships among the various measures of cost.
Measures of Short-Run Cost: Total Fixed and Variable Costs Recall that the short run is a period of time over which the firm is unable to vary all its
inputs. Thus, some inputs are effectively fixed in the short run, whereas others are variable.
There are, however, costs associated with the use of both fixed and variable inputs. Let’s
examine Table 8.1, which shows how production cost varies at different rates of output for
the firm.
Total fixed cost (TFC) is the cost incurred by a firm that does not depend on how much output it produces. Fixed cost includes expenditures on inputs the firm cannot vary in the
short run—normally, its plant and equipment. The fixed cost will be the same, regardless of
how much output the firm produces; in particular, if the firm produces nothing, it still incurs
its total fixed cost. In Table 8.1, the firm faces a total fixed cost of $60 per day.
Total variable cost (TVC) is the cost incurred by a firm that depends on how much output it produces. This cost is associated with the variable inputs; more output requires the
use of more variable inputs, so total variable cost rises with output. To produce more in the
short run, the firm must hire more workers, use more electricity, purchase more raw materi-
als, and so on—all of which add to total variable cost as output rises. Total variable cost is
shown in column (2) of the table.
Fixed versus Sunk Costs Note that fixed cost is not necessarily identical to the concept of sunk cost introduced in
Chapter 1. A fixed cost is invariant to the output level selected by the firm; that is‚ the
firm’s expenditures on its plant and equipment are a relevant cost of production even when
output is zero. A fixed cost need not be a sunk cost‚ however‚ to the extent that the expen-
ditures made by the firm on its plant and equipment can be recouped through the sale of
such assets to another party.
Notwithstanding the subtle point that fixed costs need not be sunk, we assume for the
remainder of this book that fixed costs are indeed also sunk costs. Such an assumption is
a simplifying one made so as to allow us to more easily characterize a profit-maximizing
firm’s output and pricing decisions.
total fixed cost (TFC) the cost incurred by a firm that does not depend on how much output it produces
total variable cost (TVC) the cost incurred by a firm that depends on how much output it produces
Short-Run Costs ($) for a Hypothetical Firm
Output
Total Fixed Cost
Total Variable
Cost Total Cost
Marginal Cost
Average Fixed Cost
Average Variable
Cost
Average Total Cost
(1) (2) (3) (4) (5) (6) (7) 0 60.00 0 60.00 — — — —
1 60.00 30.00 90.00 30.00 60.00 30.00 90.00
2 60.00 49.00 109.00 19.00 30.00 24.50 54.50
3 60.00 65.00 125.00 16.00 20.00 21.67 41.67
4 60.00 80.00 140.00 15.00 15.00 20.00 35.00
5 60.00 100.00 160.00 20.00 12.00 20.00 32.00
6 60.00 124.00 184.00 24.00 10.00 20.67 30.67
7 60.00 150.00 210.00 26.00 8.57 21.43 30.00
8 60.00 180.00 240.00 30.00 7.50 22.50 30.00
9 60.00 215.00 275.00 35.00 6.67 23.89 30.56
10 60.00 255.00 315.00 40.00 6.00 25.50 31.50
Table 8.1
184 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 184Trim Size: 203.2 mm X 254 mm
Five Other Measures of Short-Run Cost Five other measures of cost are identified in the table. Before looking at them, we should
emphasize that they are all derived from the total fixed and total variable cost relationships.
No new types of costs are involved: everything that we are assuming about the firm’s costs
is shown in columns (1) and (2). The remaining columns are just different ways of present-
ing the same basic cost information in a more convenient and workable form.
Total cost (TC) is the sum of total fixed and total variable cost at each output level. For example, at five units of output, total fixed cost is $60, and total variable cost is $100, so
total cost is $160. Total cost identifies the cost of all the inputs, fixed and variable, used to
produce a certain output. Since total variable cost rises with output, so does total cost.
Marginal cost (MC) is the change in total cost resulting from a one-unit change in output. When output increases from seven to eight units, for example, total cost rises from $210 to
$240: the $30 increase in total cost is the marginal cost of producing the eighth unit. Marginal
cost can also be defined as the change in total variable cost resulting from a one-unit change
in output, because the only part of total cost that rises with output is variable cost. Basically,
then, the marginal cost relationship shows how much additional cost a firm will incur if it
increases output by one unit, or how much cost saving it will realize if it reduces output by
one unit. As we shall see, marginal cost is the most important cost concept for many purposes.
Finally, there are three measures of average cost per unit of output.
Average fixed cost (AFC) is total fixed cost divided by the amount of output. Because fixed cost is constant, the greater the output, the lower the average fixed cost. For example,
the average fixed cost associated with the first unit of output is $60, but it falls to $6.67 for
the ninth unit.
Average variable cost (AVC) is total variable cost divided by the amount of output. If eight units are produced, total variable cost is $180, so average variable cost is $180/8, or
$22.50 per unit.
Average total cost (ATC) is total cost divided by the output. We can also define it as the sum of average fixed cost and average variable cost. At eight units of output, for instance,
total cost is $240, so the average total cost is $30. Alternatively, average total cost is also
the sum of average fixed cost, $7.50, and average variable cost, $22.50.
So far we have defined the various measures of cost and explained the arithmetical rela-
tionships among them. Having covered these matters, we can now turn to the question of
what factors determine the exact way cost varies with output for the various cost measures.
Behind Cost Relationships A firm’s costs are determined by its production function, which identifies the input combi-
nations that can produce a given output, and the prices that must be paid for these inputs.
In the short run, the production function relates output to the quantity of variable inputs;
fixed inputs do not vary. Recall that in a short-run setting the law of diminishing marginal
returns is relevant. It specifies that increasing the variable input (or inputs) will, beyond
some point, result in smaller and smaller increases in total output. This assumption is the
only one we need to make about production relationships in the short run. As we will see, it
largely determines how total variable cost varies with output.
The production function indicates how much of the variable input (or inputs) the firm
needs to produce alternative levels of output. Because the total variable cost is the total
amount of money the firm must spend to acquire the necessary quantity of the variable
input, the price per unit of the variable input is crucial in determining the firm’s short-run
cost. We assume here that the firm can employ any quantity of the variable input at a given
price per unit.
Figure 8.1 illustrates how these two factors—the law of diminishing marginal returns
and a fixed price per unit of the variable input—combine to determine the way total variable
total cost (TC) the sum of total fixed and total variable cost at each output level
marginal cost (MC) the change in total cost that results from a one- unit change in output
average fixed cost (AFC) total fixed cost divided by the amount of output
average variable cost (AVC) total variable cost divided by the amount of output
average total cost (ATC) total cost divided by the output
Short-Run Cost Curves 185
C08.INDD 11:11:1:AM 08/06/2014 PAGE 185Trim Size: 203.2 mm X 254 mm
cost varies with output. Suppose that the variable input is hours of labor and the total prod-
uct (TP) curve relates the quantity of hours worked to total output. Note that the shape of the TP curve reflects diminishing marginal returns beyond point A (where the slope of the curve reaches a maximum).
The total product curve shows the amount of labor used to produce each level of out-
put. If labor is the only variable input, the cost to the firm of hiring labor is the total vari-
able cost of production. Total variable cost can also be measured on the horizontal axis by
multiplying each quantity of labor by its unit cost, here assumed to be $10 per hour. For
example, from the TP curve we know that producing 8 units of output requires 18 hours of labor, and at $10 per hour the total variable cost associated with 8 units of output is $180.
Thus, when inputs are measured in terms of their cost, the same curve relates input to
output and cost to output. The shape of the TVC curve is determined by the shape of the TP curve, which in turn reflects diminishing marginal returns. This relationship explains why understanding the law of diminishing marginal returns is important; it ultimately deter-
mines how total variable cost is related to output. In addition, the law also determines the
behavior of marginal cost and average variable cost since they are derived from the TVC relationship. We will discuss these relationships in more detail in Section 8.3.
By convention, we draw the total variable cost curve with TVC on the vertical axis and output on the horizontal axis, the reverse of the situation in Figure 8.1, which is why the
TVC curve may not look right at first glance. To see its more conventional appearance, refer to Figure 8.2a, which represents Figure 8.1 after its axes have been interchanged.
8.3 Short-Run Cost Curves Now let’s take a more careful look at the relationship between cost and output by using the
firm’s cost curves. Figure 8.2a shows the firm’s total cost curves, and Figure 8.2b shows the
per-unit cost curves. The curves have the shapes implied by the data in Table 8.1, but now we
wish to see why all short-run cost curves are expected to have these same general shapes.
A
0
4
8
Output
Hours of labor: the variable input
Cost of labor: TVC
18
$180
8
$80
TP (TVC)
From Total Product to Total Variable Cost The quantity of the variable input is related to output through the total product curve. The total product (TP) curve can be transformed into the total variable cost (TVC) curve by multiplying each quantity of the variable input by its per-unit price.
Figure 8.1
186 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 186Trim Size: 203.2 mm X 254 mm
Little more need be said about the total variable cost curve. In the previous section we
explained the way the total variable cost curve is derived from the production function.
Total fixed cost, TFC, is a horizontal line at $60, indicating a $60 fixed cost, regardless of output. The TC curve shows total cost, the sum of fixed and variable costs. Note that the TC curve is $60 higher than the TVC curve at each rate of output. The vertical distance between TVC and TC is total fixed cost: both F0 and CD equal $60.
Marginal Cost Economic analysis relies more heavily on the per-unit cost curves shown in Figure 8.2b,
so we devote more attention to them here. Examine the marginal cost curve first. It
is -shaped, with the cost of additional units of output first falling, reaching a minimum,
and then rising. Marginal cost falls at first because the fixed plant and equipment are not
designed to produce very low rates of output, and production is very expensive when out-
put is low. Consider a striking example: the marginal cost of printing a second newspaper
may be trivial compared with the marginal cost of the first (which includes composing
Short-Run Total and Per-Unit Cost Curves (a) Adding total fixed cost (TFC) of F0 to the TVC curve yields the short-run total cost curve. (b) Four per-unit cost curves—average total cost (ATC), average variable cost (AVC), average fixed cost (AFC), and marginal cost (MC)—are derived from the total cost curve.
(b)
Output
$15
$20
$40
$60 = F
Cost per unit
Total cost
10 Output
MPL = 1/4
4
MPL = 2/3
0
0 4
(a)
5
AVC
AFC
TFC
TVC
TC
ATC
MC
C
B
D
Figure 8.2
Short-Run Cost Curves 187
C08.INDD 11:11:1:AM 08/06/2014 PAGE 187Trim Size: 203.2 mm X 254 mm
the pages and setting the press). Declining marginal cost comes to an end at some point
(four units of output in Figure 8.2b), and thereafter marginal cost rises with output. Even-
tually, marginal cost must rise because the plant will ultimately be overutilized as output
expands beyond the level for which it was designed. At that point and possibly before,
marginal cost begins to rise, and each additional unit of output costs more than the previ-
ous one.
The shape of the marginal cost curve is attributable to the law of diminishing marginal
returns. To see exactly why, recall that marginal cost (MC) can be defined as the change in total variable cost (∆TVC) that is associated with a small change in output (∆q):
MC TVC q= Δ Δ/ . (1)
A one-unit increase in output, for example, requires some additional amount of the variable
input (labor), ∆L, to be employed. Using more labor increases the total variable cost by ∆L times the wage rate w. Thus, we have:
MC TVC q w L q w MP= = =Δ Δ Δ Δ/ ( )/ / L (2)
because the marginal product of labor (MPL) equals ∆q/∆L, so the reciprocal, ∆L/∆q, equals 1/MPL.
This relationship states that short-run marginal cost equals the price of the variable input
(in this case, the wage rate) divided by its marginal product. Let’s check to see why this
relationship makes sense. Suppose that at the output level of four units, w = $10, and MPL = 2/3 units of output. If an additional unit of labor increases output by 2/3 units (MPL), then one additional unit of output requires 3/2 units of labor, at a cost of $10 per unit of labor.
So the marginal cost at an output level of four units is $15, or w/MPL = $10/(2/3). Because of the law of diminishing marginal returns, the marginal product of labor
varies with the amount of output and, therefore, so must marginal cost. At low levels of
output MPL is rising, so, correspondingly, marginal cost (w/MPL) must be falling. When MPL reaches a maximum, then MC must be at a minimum. This minimum occurs at four units of output in the diagram, the rate of output at which marginal returns begin to fall.
At output levels where MPL is declining, MC must be rising. For example, at an output of 10 units in the diagram, MPL has declined to 1/4 units of output. Producing another unit of output, therefore, requires four more units of labor at a cost of $10, so the marginal
cost is $40.
If the marginal product of the variable input rises (at low rates of output) and then falls, the marginal cost curve will first fall and then rise. Thus, the law of diminishing marginal returns lies behind the MC curve. Indeed, we can restate the meaning of diminishing mar- ginal returns in a way that makes this relationship obvious. In the region of diminishing
marginal returns, each additional unit of the variable input adds less to total output. That
is, each additional unit of output requires more of the variable input than the previous unit.
More of the variable input per unit of output means higher cost, so marginal cost will rise in
the region where marginal product is falling.
Average Cost There are three average cost curves. To begin, let’s look at average variable cost in Fig-
ure 8.2b; average variable cost is equal to total variable cost divided by output. For the very
first unit of output, total variable cost, average variable cost, and marginal cost are all equal
(see Table 8.1). Then, marginal cost falls, causing the average cost to fall as well. In fact,
average cost will decline as long as marginal cost lies below it and pulls it down. Put dif-
ferently, per-unit production cost tends to fall at low rates of output (remember the second
newspaper), but beyond some point (B in Figure 8.2 or five units of output in Table 8.1), average variable cost will rise.
188 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 188Trim Size: 203.2 mm X 254 mm
The physical production relationships implied by the law of diminishing marginal
returns are also responsible for the shape of the average variable cost curve. Average vari-
able cost is total variable cost divided by output:
AVC TVC q= / . (3)
Total variable cost is simply the total amount of the variable input (L) times its unit cost (w). Thus, we have:
AVC TVC q wL q w AP= = =/ / / L (4)
because the average product of labor (APL) equals q/L (total output divided by total labor), and so the reciprocal, L/q, is equal to 1/APL.
In the previous chapter we saw that the law of diminishing marginal returns leads to an
average product curve shaped like the inverted letter ; that is, average product rises, reaches
a maximum, and then falls. As a result, the AVC curve must be -shaped. Over the region where APL is rising, w/APL (AVC) is falling. Similarly, when APL is falling, AVC must be rising. Point B, where AVC is at a minimum, therefore corresponds to the point at which APL is at a maximum. The law of diminishing marginal returns dictates the shape of both the MC and AVC curves—not a surprising conclusion since we saw that it determines TVC, and MC and AVC are derived from TVC. The shapes of the per-unit cost curves reflect the underlying physical requirements of production. As fewer units of the variable input are required per unit
of output (on average for the AVC curve and for marginal changes on the MC curve), per-unit costs fall. Conversely, they rise when input requirements per unit of output increase.
There are two other average cost curves, average fixed cost (AFC) and average total cost (ATC). Average fixed cost declines over the entire range of output as the amount of total fixed cost is spread over ever-larger rates of output. The AFC curve has an intriguing property: if its height at any output is multiplied by that output, the area of the resulting rectangle (height
times width) is the same regardless of the output level selected. Do you understand why?
The ATC curve shows average total cost. It is the sum of AFC and AVC and measures the average unit cost of all inputs, both fixed and variable. The ATC curve must also be
-shaped, although its minimum point is located at a higher output than the minimum point
of AVC. This difference occurs because ATC = AVC + AFC, and at the output where AVC is at a minimum, AFC is still falling, so the sum of AVC and AFC will continue to fall. At some point, however, the rising AVC offsets the falling AFC, and thereafter ATC rises. Finally, because AVC + AFC = ATC, the average fixed cost is the vertical distance between ATC and AVC. Note that this vertical distance becomes smaller as more output is produced, since AFC declines as output rises.
APPLICATION 8.1
According to a McKinsey Global Institute study of labor productivity growth in the United States between 1995 and 2000, the most important factor behind the observed growth in productivity per labor hour was Walmart.1 The retailing giant, which now accounts for over 12 percent of
The Effect of Walmart on Retailing Productivity, Costs, and Prices
all retail sales (excluding automobiles) in the United States, 25 percent of grocery sales, and 30 percent of toy market sales, and is the largest private-sector employer, effectively has pivoted the total product curve in Figure 8.1 counter- clockwise around the origin relative to previously prevailing production practices—thereby ensuring that the height and slope of the total product curve are greater at each level of labor input. Since the height and slope of the total product curve are greater at each level of labor input, the average and marginal products of labor are correspondingly higher
1This application is based on Pankaj Ghemawat and Ken A. Mark, “The Price is Right,” New York Times, August 3, 2005, p. A23; Thomas L. Friedman, The World Is Flat (New York: Farrar, Straus and Giroux, 2005; and “Walmart” in Wikipedia.
Short-Run Cost Curves 189
C08.INDD 11:11:1:AM 08/06/2014 PAGE 189Trim Size: 203.2 mm X 254 mm
Marginal–Average Relationships All marginal curves are related to their average (total and variable) curves in the same way.
Because we encounter average and marginal cost curves frequently in later chapters, the
relationships bear repeating here.
When marginal cost is below average (total or variable) cost, average cost will decline. Equivalently, when average cost is declining, marginal cost must be below average cost. If
average cost is currently $20, and producing one more unit costs $15, then the average cost
of all units will fall.
When marginal cost is above average cost, average cost rises. Equivalently, when average cost is rising, marginal cost must be above average cost. If average cost is currently $20, and
producing another unit costs $30, then the average cost of all units will be pulled up.
When average cost is at a minimum, marginal cost is equal to average cost. This is implicit in the other relationships, but we can explain it in a different way. At the point
where average cost is at a minimum, the curve is essentially flat over a small range of out-
put. When the curve is neither falling nor rising, a small change in output does not change
average cost. If an additional unit of output leaves the average cost unchanged, the mar-
ginal cost must equal average cost. For example, if average cost is $20, and it is still $20
after output increases by a unit, then the marginal cost of that unit must also be $20. If it
were not $20, then it would have changed average cost.
The Geometry of Cost Curves The graphical derivation of the average and marginal cost curves from a total cost
curve closely parallels the derivation of average and marginal product curves from a
total product curve explained in Chapter 7. Note that although we use the short-run
total variable cost (TVC) curve here, the same procedure applies to the derivation of the related average and marginal cost curves from any total cost curve, either short run or
long run.
Figure 8.3a shows a total variable cost curve; Figure 8.3b shows the average variable
cost and marginal cost curves derived from it. To derive AVC from the TVC curve, we draw a ray from the origin to each point on the TVC curve; the slope of the ray measures AVC at that output. At output q1, for example, ray 0A has a slope equal to Aq1/0q1 or $21.67 per unit ($65/3). The slope of the ray in Figure 8.3a is shown by the height of the AVC curve in
at each labor input level as well. And, as we have seen in the preceding discussion about cost curves, higher average and marginal products of labor imply lower average and mar- ginal cost curves at each level of firm output.
The lower costs realized by Walmart have at least partly been passed on to retail consumers—either those shop- ping at Walmart or those at the store’s competitors (since Walmart has spurred those competitors to also find ways to enhance productivity). Walmart’s entry into a market has been estimated to decrease retail prices by an average of 5 percent in urban areas and 8 percent in rural areas. It also is estimated that Walmart increases consumer surplus in the United States by roughly $200 billion annually.
Walmart has enhanced its productivity and thereby lowered its costs through a multitude of seemingly tiny
but ultimately significant operational improvements. For example, by connecting its delivery trucks through radios and satellites, Walmart ensures that its trucks don’t come back empty after dropping off their loads at designated Walmart stores. Instead, the drivers of those trucks can be directed to go a few miles down the road and pick up goods from a Walmart supplier, thereby saving the deliv- ery charges the supplier would otherwise levy. Walmart was also one of the first retailers to introduce computers to track store sales and inventory and was the first to share this information with suppliers. Through approaching its suppliers as partners, not adversaries, Walmart has devel- oped a collaborative planning, forecasting, and replen- ishment program that reduces inventory costs for both Walmart and its suppliers.
190 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 190Trim Size: 203.2 mm X 254 mm
Figure 8.3b. The flatter the ray from the origin, the lower the AVC. For example, AVC is at a minimum ($20) when output is q2, since the ray 0C is the flattest ray that touches the TVC curve. Thus, AVC falls as output increases from zero to q2 and then rises at greater rates of output.
Marginal cost is shown by the slope of the total variable cost curve at each rate of
output. At q3, for instance, producing another unit of output adds $24 to cost, as indi- cated by the slope of TVC at point D in Figure 8.3a. The height of the marginal cost curve is thus $24 in Figure 8.3b. Starting from the origin, the TVC curve becomes flat- ter as we move up to point B, implying that MC is falling; beyond point B it becomes steeper, indicating that MC is rising. At point C, where average variable cost is at a minimum, MC equals AVC.
Graphical Derivation of Average and Marginal Cost Curves (a) Average variable cost equals the slope of a ray from the origin to a point on the total variable cost curve. For example, at point A, AVC equals the slope of the ray 0A, or $21.67 per unit. Marginal cost is the slope of the TVC curve at each point. For example, at point D marginal cost is $24 per unit. (b) The entire average variable cost and marginal cost curves are shown.
$100
Cost
$65
0
0
$24
Cost per unit
$21.67
$21.67
$20
$20
Output
(a)
(b)
Output
1 1
1
A B
C
D
TVC
AVC
MC
A
B
C
D
q3 (6)
q2 (5)
q1 (3)
q3 (6)
q2 (5)
q1 (3)
$24
Figure 8.3
Long-Run Cost of Product ion 191
C08.INDD 11:11:1:AM 08/06/2014 PAGE 191Trim Size: 203.2 mm X 254 mm
8.4 Long-Run Cost of Production In the long run a firm has sufficient time to adjust its use of all inputs to produce output in
the least costly way. Because a firm can augment office space and equipment by leasing or
purchasing additional facilities, all inputs are variable. Our first task is to see how a firm
will choose to combine inputs in its production process when all factors are variable.
Isocost Lines Consider a firm that uses just two inputs, capital (its office space and equipment) and labor,
to produce its product. The firm’s costs of production can be represented by isocost lines.
An isocost line (equal-cost line) identifies all the combinations of capital and labor that can be purchased at a given total cost. Figure 8.4a shows three isocost lines corresponding to
three different levels of total cost.
Because the isocost line is a new relationship, let’s carefully examine one. Look at the
middle isocost line in Figure 8.4a. Suppose that a firm has total funds of TC2 to pay its inputs. The prices the firm must pay for inputs are w for labor (the wage rate) and r for cap- ital (the per-unit rental rate for office space and equipment, which may be an implicit cost
if the firm owns the assets). If the firm devotes all the funds to capital, it can employ TC2/r units of capital, leaving no money to hire workers. Thus, TC2/r is the vertical intercept of
isocost line a line that identifies all the combinations of capital and labor that can be purchased at a given total cost
TC3/r
TC2/r
TC1/r
TC1/w TC2/w
L2 TC3/w
IQ9 IQ6
IQ3
K2
Labor
Expansion path
(b)
Capital
A
B
C
D
E
(a)
TC1/w0 0 TC2/w
TC3/w Labor
TC3/r
TC2/r
TC1/r
Capital
Isocost Lines and the Long-Run Expansion Path (a) Three different isocost lines are shown corresponding to three different levels of total cost. (b) The point of tangency between an isoquant and an isocost line indicates the least costly combination of inputs that can produce a specified output. For example, employing L2 units of labor and K2 units of capital is the cheapest way to produce six units of output. The expansion path, which shows the least costly way of producing each output level when all inputs can be varied, is formed by connecting all points of tangency.
Figure 8.4
192 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 192Trim Size: 203.2 mm X 254 mm
the isocost line. Alternatively, if the firm devotes all the funds to hiring labor, it can hire
TC2/w units of labor, leaving no money for capital, so TC2/w is the horizontal intercept. All the intermediate positions on the line show the combinations of labor and capital the firm
can hire at a cost of exactly TC2. The slope of an isocost line is (minus) the ratio of input prices, w/r, indicating the rela-
tive prices of inputs. For example, if the wage rate is twice the rental rate of capital (w/r = 2), then hiring 1 more unit of labor, without incurring any additional cost, means the firm must
employ two units less of capital. Thus a close analogy exists between the consumer’s budget
line and a firm’s isocost line. In one respect, however, the analogy breaks down. Consumers
are usually restricted to operating within a given budget, but firms are not. Firms can expand
their use of all inputs and finance the expansion by selling the increased output. The firm’s
total cost of operation is not a constant but varies with output, which explains why there are
three isocost lines in Figure 8.4a (among the many that could be drawn) and not just one.
Least Costly Input Combinations Analyzing isocost lines and isoquants together allows us to determine what combinations of
inputs the firm will use to produce various rates of output. We have seen how a firm’s pro-
duction function can be represented by isoquants, and three (IQ3, IQ6, and IQ9) are shown in Figure 8.4b. Suppose that the firm plans to spend exactly TC2 to hire inputs. From among the input combinations it can use (shown by the middle isocost line), the firm will clearly
want to choose the labor–capital combination that yields the greatest output. The firm could
employ the input combination at point D, but only three units of output would be produced (since point D is on isoquant IQ3). By using less capital and more labor at point B, the firm can produce six units of output at the same total cost. Indeed, six units of output is the max-
imum output possible at a total cost of TC2, as shown by the fact that higher isoquants like IQ9 lie entirely above the middle isocost line. To produce the maximum output possible at a given total cost, a firm should operate at a point where an isocost line is tangent to the high-
est isoquant attainable: IQ6 is tangent to the middle isocost line at point B. In the same manner, we see that nine units is the maximum output possible at the higher
total cost of TC3, where the firm would operate at point C. Similarly, three units is the maxi- mum output possible at a total cost of TC1. We can interpret these points in a different way: they also show the least costly way of producing any given output level. Suppose, for instance, that the firm wishes to produce six units of output: it must use an input combination that lies
on IQ6. It could operate at point E and produce six units of output, but that combination of labor and capital costs TC3, because it lies on a higher isocost line and requires a larger outlay than TC2. In fact, every other way of producing six units of output involves a higher total cost than point B. Point B thus identifies the least costly combination of inputs that the firm can use to produce six units of output. Points of tangency show the maximum output attainable at a given cost as well as the minimum cost necessary to produce that output.
Interpreting the Tangency Points Let’s look at the tangency points more closely. At the tangencies, the isoquants and isocost lines
have the same slopes. Recall that the slope of an isoquant is the marginal rate of technical sub-
stitution, MRTS, and the slope of the isocost line is the ratio of the input prices. Therefore, when the firm produces an output in the least costly way, it will satisfy the following condition:
MRTS w rLK = / . (5)
This condition indicates that the firm will adjust its employment of inputs so that the rate
at which one input can be traded for another in production (MRTSLK) will equal the rate at which one input can be substituted for the other in input markets (w/r).
Long-Run Cost of Product ion 193
C08.INDD 11:11:1:AM 08/06/2014 PAGE 193Trim Size: 203.2 mm X 254 mm
Let’s explore why this equality holds. Recall that the marginal rate of technical substitu-
tion equals the ratio of the marginal products of the inputs. Thus we can write the previous
expression as:
MP MP w rL K/ / .= (6)
Rearranging terms, we obtain:
MP w MP r vL K/ / .= (7)
The last equality is sometimes termed the golden rule of cost minimization. It is equivalent to the tangency condition and indicates that to minimize cost, the firm should
employ inputs in such a way that the marginal product per dollar spent is equal across all
inputs. To illustrate, suppose that the firm is employing inputs at a point where MPL = 50 (units of output) and MPK = 60. Furthermore, suppose that the wage rate of labor is $10, and the rental cost of capital is $60. In this case, MPL/w = 50/$10 = 5/$1, implying that an additional dollar spent on labor will produce five more units of output. For capital,
MPK/r = 60/$60 = 1/$1, indicating that an additional dollar spent on capital will produce only one more unit of output. For this allocation of inputs, MPL/w > MPK/r; that is, a dol- lar’s worth of labor adds more to output than does a dollar’s worth of capital. The firm is
not producing as much output as it could given its cost (or, equivalently, it is not produc-
ing the current output at the lowest possible cost). If the firm spends $1 less on capital, it
loses one unit of output (MPK/r = 1), but spending this dollar on labor increases output by five units (MPL/w = 5). So, on balance, output will rise under such a reallocation by four units with no change in cost.
Whenever MPL/w is greater than MPK/r, a firm can increase output without increasing production cost by shifting outlays from capital, where output per dollar spent is lower, to
labor, where output per dollar spent is higher. This shifting should continue until the terms
become equal. The law of diminishing returns serves to bring the terms into equality as the
input mix is adjusted. That is, if the law of diminishing returns holds, hiring more labor
tends to reduce MPL, and hiring less capital tends to increase MPK. Note that the initial situation described could be shown in Figure 8.4b by a point like D, because at that point MRTSLK > w/r.
Perhaps the most intuitive way to think about the golden rule of cost minimization is to
remember what the ratio of an input’s marginal product to its price means. Take the case of
labor. The ratio MPL/w signifies the increase in output per dollar spent on labor. Although it’s true that each unit of labor costs $10, each of those 10 dollars increases output by five units in our example. In an important sense then, MPL/w represents the rate of return per dollar invested in labor. And the golden rule of cost minimization states that for a producer
interested in maximizing profit, the rates of return across all the inputs used should be iden-
tical at any selected level of output.
As an analogy, suppose that Larry Ellison, CEO of Oracle, has $10 billion split
equally between two banks, Bank K and Bank L. Suppose also that Bank K offers a
1 percent annual interest rate on any money invested in it while Bank L offers a 5 per-
cent annual interest rate. If there are no risk-diversification reasons for Larry to split his
$10 billion total investment equally between the two banks, the financial advice that we
would give him would be clear: namely, to maximize the total return on the investment,
Larry needs to reallocate dollars from Bank K to Bank L as long as Bank L offers the
higher interest rate.
The same principle applies to allocating dollars between inputs. Investing in different
inputs is like investing in different banks. If, at Oracle’s current level of output, the rate of
return offered by an MBA exceeds the rate of return provided by a software engineer, the
firm would be wise to reallocate dollars from the engineer “bank” to the MBA “bank” so as
to maximize total output per dollars spent on inputs.
golden rule of cost minimization a rule that says that to minimize cost, a firm should employ inputs in such a way that the marginal product per dollar spent is equal across all inputs
194 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 194Trim Size: 203.2 mm X 254 mm
The Expansion Path The points of tangency in Figure 8.4b show the least costly way of producing each indi-
cated output. Producing six units of output as cheaply as possible, for instance, involves
hiring L2 units of labor and K2 units of capital. The line formed by connecting these tan- gency points is the firm’s expansion path. It identifies the least costly input combination for each output and will generally slope upward, implying that the firm will expand the
use of both inputs (in the long-run setting) as it increases output. Note that input prices
are assumed to remain constant as the firm varies its output along the expansion path.
expansion path a line formed by connecting the points of tangency between isocost lines and the highest respective attainable isoquants
APPLICATION 8.2
After the deregulation of the domestic airline industry in the late 1970s, American Airlines became one of the largest and most profitable airlines in the United States.2 To a great extent, American’s success reflected the strategic changes implemented by CEO Robert Crandall and his manage- ment team in the wake of deregulation. The key changes included developing the best information system (SABRE) in the industry, effective marketing strategies (such as the Advantage frequent-flyer program), high-quality customer service, and a passion for cost minimization.
In terms of cost minimization, American; was among the first firms in the industry to switch to shorter-range, and more fuel-efficient aircraft–from three-engine 727s to twin-engine MD80s, 737s, and Regional Jets; developed hub-and-spoke route structures that diminished overhead costs and was compatible with shorter-range and more fuel-efficient aircraft; reduced labor costs through renego- tiated labor contracts and a two-tier wage structure offer- ing new workers employment at lower wages than existing staff; and gave thought to about every possible way to cut fuel and other nonlabor variable costs.
Most of the planes flown by American, for example, were not painted beyond their red, white, and blue logo stripes. This tactic lowered cost in terms of both paint and fuel. An unpainted 767 was approximately 400 pounds lighter and as a result cost roughly $12,000 less to fly per year.
In the mid-1980s, American reduced the internal weight of each plane by at least 1,500 pounds. The reduction was achieved by installing lighter seats, replacing metal food carts with reinforced plastic ones, relying on smaller pillows and less bulky blankets, and redesigning the service galleys. The reduction saved American at least $22,000 per year per plane in fuel costs.
American Airlines and Cost Minimization
Not even the smallest stone remained unturned by Crandall and his management crew in their pursuit of cost minimization. While on board an American flight, Crandall dumped his leftover lettuce into a plastic bag and sent it off to the director of his firm’s catering division with the message “Shrink the dinner salads!” Still not satisfied, the cost-conscious CEO then ordered the removal of one black olive from each passenger’s salad (Crandall believed customers wouldn’t notice if the olive was missing) and in the process saved the company $70,000 per year in reduced food and fuel costs. Crandall once even fired a watchdog to save a buck. As recounted by the CEO in an interview:
It’s true. We had a cargo warehouse in the Caribbean, and we had a guy there guarding it all night long. I was reviewing the budget and wanted to reduce costs. My people said we needed him to prevent thefts. So I said, ‘Put him on part time and rotate his nights so nobody knows when he will be there.’ And the next year I wanted to reduce costs, and I told them, ‘Why don’t we substitute a dog? Turn a dog loose in the warehouse.’ So we did, and it worked. Now the fol- lowing year, I needed to get the costs down some more, and my guy said, ‘Well, we’re down to a dog!’ So I said, ‘Why don’t you just record the dog snarling?’ And we did. And it worked! Nobody was really sure whether there was a dog in there or not.
Of course, the competitive advantages that Crandall and his management team were able to generate through cost minimalization in the wake of deregulation have not been sufficient to ensure the profitability of the company in more recent times. The growth of competing airlines with ever lower costs (Southwest, Jet Blue, Spirit, and so on) and the economic downturn of 2007 eroded American’s cost advantage, diminished the company’s profitability, and ulti- mately led to American filing for bancruptcy in 2012 and being acquired by rival US Airways in 2013.
2This application is based on “Airlines Resort to Penny-Pinching Ploys to Bring Their Fuel Bills Back to Earth,” Wall Street Journal, September 28, 1990, pp. B1 and B3; “The Man Who Fired a Dog to Save a Buck,” Time, October 28, 1990; and “Hold the Olives, Bring on the Lobster,” Time, May 13, 1991, p. 57.
Long-Run Cost of Product ion 195
C08.INDD 11:11:1:AM 08/06/2014 PAGE 195Trim Size: 203.2 mm X 254 mm
Is Production Cost Minimized? It is generally assumed that business firms produce output at the lowest possible cost. In
terms of Figure 8.4b, the assumption is that firms will operate somewhere on the expansion
path (exactly where depends on what output they choose to produce, as explained in the
next chapter). Because there are many ways a given output can be produced, and because
only one of these ways results in the minimum cost, what is the basis for assuming that the
firm will choose the least costly method?
The assumption that firms minimize cost is based on the belief that firms are attempting to
maximize profit. To maximize profit, a firm must produce its output at minimum cost. If a firm
could produce the same output at a lower cost, it would reduce total cost without changing total
revenue (because output is unchanged), and therefore realize a greater profit. Of course, firms
may not always be successful in minimizing cost; this is a difficult task when input prices and
technology both change over time. Nonetheless, firms do have a substantial profit incentive
to hold cost down, so the assumption that they minimize cost seems warranted. The basis for
assuming profit maximization itself will be discussed in the next chapter. It should be noted that
cost minimization, while a necessary condition for profit maximization, is not the same as profit
maximization. Cost minimization occurs at all points on the expansion path, but profit maximiza-
tion also involves selecting the most profitable output from among those on the expansion path.
That the profit motive is important in spurring organizations to minimize cost is sug-
gested by comparisons of costs of production by private firms and government agencies
that produce the same, or similar, products. A large number of studies have made such a
comparison. For example, one study concluded that the construction cost for public hous-
ing was 20 percent higher than comparable privately produced housing. Another concluded
that municipal fire departments had a 39 to 88 percent higher cost per capita than did com-
munities contracting with a private firm to provide the services. An overwhelming major-
ity of the existing studies have found government provision to be more costly than private
provision; only a few have found government provision to be less costly.3 The most likely
explanation for this cost disparity seems to be that the profit incentive is absent in govern-
ment organizations; public policymakers do not profit from reducing production cost.
APPLICATION 8.3
Since 1978, China’s gross domestic product (GDP) has grown an average of more than 9 percent annually and China has become the second largest economy in the world after the United States.4 By 2025, China is expected to surpass the United States and become the world’s largest economy. A key reason for the growth is the move by China toward greater privatization. Beginning with the de-collectivization of agriculture launched by Chinese leader Deng Xiaoping in 1978 followed by the privatization or contracting out of much of state-owned industry in the 1980s and 1990s, the private sector now accounts for over 70 percent of Chinese GDP.
Privatization and Productivity in China
Prior to pro-market reforms, the productivity of China’s agricultural sector was relatively poor and food shortages and starvation were common. After Deng and de-col- lectivized agriculture, however, agriculture output began to increase by 8.2 percent per year versus the 2.7 percent annual growth rate that prevailed pre-reform and notwith- standing less land being devoted to farming.
Of the over tenfold growth in GDP since 1978, econo- mists estimate that over 40 percent has been due to an increase in the overall productivity of the inputs of produc- tion. By contrast, from 1957 to 1978, the height of Maoist policies that focused on government control of production, total factor productivity declined by 13.2 percent in China.4This application is based on “Chinese Economic Reform,” Wikipedia.
3These studies are discussed in T. E. Borcherding, W. W. Pommerehne, and F. Schneider, “Comparing the Efficiency of Private and Public Production: The Evidence from Five Countries,” Zeitschrift fur Nation- alökonomie, 89 (1982), pp. 127–156.
196 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 196Trim Size: 203.2 mm X 254 mm
8.5 Input Price Changes and Cost Curves Remember that input prices are assumed to be constant when we construct a firm’s cost
curves. Per-unit costs vary with output along a cost curve because the productivity of the
inputs varies with the rate of output, not because the costs of inputs vary. Economists gen-
erally assume that one firm’s output will not by itself influence input prices. Rarely does
a single firm use a large portion of the total quantity of any input. Consequently, if a firm
increases its employment of land, MBAs or gasoline, this expansion will not cause a per-
ceptible increase in the market demand for these inputs, and their prices will not be affected.
Although input prices generally do not change because of any single firm’s output deci-
sion, they do occasionally change on account of underlying market forces. For example,
let’s take a look at a commercial real estate developer, CB Richard Ellis (CB). CB pro-
duces, as part of its product line, parking spaces using two inputs, concrete (C) and land (L). Importantly, these two inputs can be substituted for one another to a certain extent. Land can be conserved, for example, by using more concrete and building multistory park-
ing garages. Because the two inputs are substitutes, CB’s isoquants will have the typical
shape—that associated with a production function in which different input combinations
can produce the same total output.
Figure 8.5a illustrates the firm’s input choices. Initially, CB Richard Ellis is producing
an output of 1,000 by employing land and concrete at point E on isoquant IQ1,000 and iso- cost line MN. The firm’s original expansion path passes through point E, and from it we can derive the per-unit cost curves MC and AC, shown in Figure 8.5b. Along these curves, as along the expansion path, the prices of inputs are unchanged.
Now suppose that the price of land increases. First, let’s see how this change affects the
way inputs are combined in the least costly manner to produce the same output, 1,000, in
MC
OutputLand 1,0000
Cost per unit
Concrete
New expansion path
Expansion path
(TC2/PC) M1
(TC1/PL)′ (TC2/PL)′ (TC1/PL) N1
IQ1,000
N
(TC1/PC) M
AC
AC ′ MC ′
N ′0
E ′
E
(a) (b)
A Higher Input Price Shifts Cost Curves Upward (a) With a higher price of land the cost of producing each level of output increases. Input combination E′ becomes the least costly way to produce IQ1,000 after the increase in the land price. (b) The change in the land price shifts the AC and MC curves upward to AC′ and MC′, respectively.
Figure 8.5
Input Pr ice Changes and Cost Curves 197
C08.INDD 11:11:1:AM 08/06/2014 PAGE 197Trim Size: 203.2 mm X 254 mm
Figure 8.5a. A higher price of land means the slopes of isocost lines become steeper,
because the slope is PL/PC. If CB were to continue to incur the same total cost in hiring land and concrete, the relevant isocost line would be MN′. If, however, CB operates on MN′, it would produce less output, and we want to consider production of the original level of out-
put, 1,000. To produce 1,000 units in the least costly way at the new price of land, CB
should operate at point E′, where IQ1,000 is tangent to the isocost line M1N1. Note that M1N1 is parallel to MN′, reflecting the higher price of land. Consequently, the higher land price increases the total cost to CB Richard Ellis of producing the same output as before; MN′ shows an unchanged total cost, so M1N1, which is further from the origin, represents a higher total cost. Note also that CB will use less of the more expensive input, land, and
more of the now relatively less costly input, concrete, in its production. What the move-
ment from E to E′ really shows is the input substitution effect that results from a change in the relative cost of inputs. Because of the higher land price, a new expansion path is
defined.5
A higher price of land increases the total cost of producing a given number of parking
spaces. In turn, an increased total cost also implies that CB Richard Ellis’s AC and MC
input substitution effect the effect of a change in the price of an input on a firm’s relative use of the input to produce a given level of output
APPLICATION 8.4
To minimize the cost of producing any given level of out- put, the golden rule of cost minimization implies that, at the margin, firms should equate rates of return across inputs. To see this rule in action and to show how the input substitu- tion effect works, take the case of a commercial real estate developer producing the parking spaces associated with shopping malls. Assume that the developer uses two inputs, concrete (C) and land (L), and successfully equates the mar- ginal rates of return across the inputs, MPC/PC = MPL/PL, in the production of parking spaces in suburban locations. If the developer moves production from a suburban to an urban setting, then the price of land will likely be higher while the other factors in the equation will not change appreciably. As the price of land rises to P′L the marginal rate of return on land becomes lower than the marginal rate of return on concrete, MPC/PC > MPL/P′L. The inequality in rates of return gives the developer an incentive to shift away from land and toward concrete–hence, the phenomenon of high-rise parking structures taking up relatively little space in central business districts versus suburban parking lots that occupy a lot of land and use a minimal amount of con- crete. In Tokyo, where the price of land is very high, so are some of the parking structures: 40-story buildings that hold four cars per floor and rely on elevators to transport cars to
The Economics of Raising and Razing Buildings
upper levels. While concrete-intensive, they are parsimoni- ous in their use of expensive, horizontal real estate.
Building demolition provides another example of the golden rule of cost minimization and the input substitution effect at work. Most large, old buildings in the United States are razed through implosion or the strategic employment of dynamite. Not too long ago in Hong Kong, however, a very different method was employed. A firm charged with demolishing a building hired numerous workers who started on the building’s top story, broke down the story into small pieces with pick-axes, and then carried the pieces away in sacks on their backs. The process was repeated for each successive story from the top down to the bottom until the entire building had been manually broken apart and carted away.
Why the difference in techniques? As you probably guessed, the price of labor, PL, was lower in Hong Kong. Firms in the razing business were trying to equate marginal rates of return on dynamite (D) and labor (L), MPD/PD = MPL/PL and, due to the lower price of labor, P′L, the marginal rate of return on dynamite was lower than that mar- ginal rate of return on labor, MPD/PD < MPL/P′L. As a result, Hong Kong firms substituted toward labor and away from dynamite.
5In passing, we should warn against a common error. It is tempting to say that we can derive the firm’s demand curve for land by identifying the point at which an isoquant is tangent to MN′, by analogy to the way we derive a consumer’s demand curve. This approach is incorrect, because the firm need not continue to operate at an unchanged level of total cost; it can choose any point on the new expansion path. The point that is most profitable, which will depend on demand conditions, would occur at the same total cost only by coincidence. In a later chapter we will see how the demand curve for an input such as land is derived.
198 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 198Trim Size: 203.2 mm X 254 mm
curves shift upward to AC′ and MC′, as illustrated in Figure 8.5b. For example, a higher land price increases average cost since average cost equals total cost divided by output. The AC′ and MC′ curves are, of course, based on the new expansion path in Figure 8.5a, which shows the least costly method of producing each number of parking spaces at the higher land price.
This result is not particularly surprising, but it highlights an important distinction. Per-unit
costs can vary as a result of changing output when input prices are constant (a movement
along given cost curves), or they can vary as a result of a change in input prices (a shift in
the cost curves). Just as in the case of distinguishing between a movement along and a shift
in a demand curve, keeping this distinction in mind for cost curves is important, too.
APPLICATION 8.5
Scoring runs is the primary objective of an offense in the game of baseball.6 Under the leadership of coach Billy Beane and some sophisticated analytics, the Oak- land Athletics were able to uncover an important arbitrage opportunity in Major League Baseball from 1999 to 2003. Reaching base safely through a hit, walk or being hit by a pitch (as measured by the statistic of on-base percent- age, or the fraction of appearances at the plate in which a player reaches base safely) appeared to have, if anything, a larger impact on the runs scored by a team than did a bat- ter’s slugging prowess (as measured by slugging percentage, or the total number of bases produced by a hitter’s appear- ances at the plate divided by the number of times at bat).
In contrast to the relatively higher marginal productivity of on-base versus slugging percentage when it came to a team’s run production, the market rewarded sluggers higher salaries than it did players who focused on just reaching base when they batted, even if by means such as a walk. For example, sluggers who hit more than 25 home runs per season earned $3–$4 million more annually than the aver- age player. By comparison, there was no similar salary pre- mium for hitters with above average skill at getting on base through walks.
Realizing that the rate of return per dollar spent, at the margin, was lower for sluggers than for players who were adept at reaching base, the Oakland A’s shifted their player mix toward individuals with higher on-base percentages. The team Oakland assembled based on this information led the American League in walks in 1999 and 2001; was second or third in 2000, 2002, and 2004; and was fifth in 2003.
By relying on statistical insights and applying the golden rule of cost minimization, the Oakland A’s began to
Applying the Golden Rule of Cost Minimization to the Baseball Diamond
outperform rival teams, when it came to the efficiency of payroll spending as well as winning record and profitability. In 2001, for example, the A’s spent only $0.5 million per game won—the lowest in Major League Baseball and about one-sixth the cost of the least efficient team. (Only two teams that year spent less than $1 million to win a game.) The team went from having the fourteenth best winning record in the American League in 1997 and tenth in 1998 to fifth in 1999 and second over the entire 2000–2003 period.
Pursuing the golden rule of cost minimization allowed the A’s to win on the cheap and make the post-season play- offs every year between 2000 and 2003. Attendance at their games and revenues, as well as profits, grew accord- ingly. In 1997, the total attendance at A’s home games had been only 57 percent as high as the American League aver- age. By 2003, the A’s were drawing 101% as much as the average team in the American League due to their greater winning ways.
The arbitrage opportunity, however, did not last for the Oakland A’s beyond 2003, as other teams started apply- ing the same statistical insights to their player hiring strate- gies. The two young Ivy League graduates with quantitative backgrounds that Billy Beane had hired to evaluate person- nel ended up being recruited as general managers by the Toronto Blue Jays and the Los Angeles Dodgers in 2003. Failing to lure Billy Beane, the Boston Red Sox hired the similarly statistically inclined Theo Epstein in 2003, making him the youngest general manager in baseball history, and ended up winning the World Series in 2004.
Beginning in 2004, the salary premium for players adept at reaching base began to rise as more teams shifted their personnel decisions toward on-base and away from slugging percentage. This market adjustment in salaries awarded to players who helped teams “win ugly” by reach- ing base safely ensured that, over time, the rate of return on a dollar spent by teams on slugging percentage was equated to the rate of return per dollar spent on on-base percentage.
6 This application is based on Jahn K. Hakes and Raymond D. Sauer, “An Economic Evaluation of the Moneyball Hypothesis,” Journal of Economic Perspectives, 20, No. 3 (Summer 2006), pp. 173–185; and Michael Lewis, Moneyball: The Art of Winning an Unfair Game (New York: Norton, 2003).
Long-Run Cost Curves 199
C08.INDD 11:11:1:AM 08/06/2014 PAGE 199Trim Size: 203.2 mm X 254 mm
8.6 Long-Run Cost Curves In many respects the firm’s long-run cost curves are easier to handle than short-run cost
curves. In the short run we must distinguish among three total cost curves: total fixed cost,
total variable cost, and total (combined) cost. Because all inputs are variable in the long
run, there is only one long-run total cost curve. There is also only one long-run average cost
curve, in contrast to three in the short run.
Figure 8.6a shows a long-run total cost curve (LTC); Figure 8.6b shows the associated long- run marginal cost (LMC) and average cost curves (LAC). (To distinguish between long-run and short-run cost curves, the long-run curves are prefixed with an L and the short-run curves with an S.) Since the graphical derivation of the average and marginal cost curves from the total cost curve is the same for the long run as for the short run (see Figure 8.2), we will not repeat it here.
Our primary concern now is to explain why the curves have the shapes they do. We have
drawn the LTC curve to imply a -shaped long-run average cost curve, but why would it have this shape? With the short-run average variable cost curve, the law of diminishing
marginal returns was responsible for its shape. In the long run, there are no fixed inputs,
so the law is not directly applicable.
Nonetheless, just as the relationship between inputs and output underlies the short-run
curves, it also underlies the long-run curves. In the long run, however, returns to scale
Long-Run Cost Curves (a) The long-run total cost shows the minimum cost at which each rate of output may be produced, just as the expansion path does. (b) The long-run marginal and average cost curves are derived from the total cost curve in the same way the short-run per-unit curves are derived from the short-run total cost curve.
LTC
LMC
LAC
Output
Output
(a)
(b)
Long-run AC and MC
Long-run TC
0
0
Figure 8.6
200 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 200Trim Size: 203.2 mm X 254 mm
are the factors that determine how output varies when all inputs are varied in proportion.
As we explained in the previous chapter, increasing returns to scale are likely to be com-
mon at low output levels, while decreasing returns to scale are likely to prevail at high
output levels. Under such plausible conditions, the long-run average cost curve will have
a shape.
Let’s see why increasing returns to scale imply a declining average cost per unit of
output. Suppose that one unit of capital (1K) and six units of labor (6L) produce an out- put of three. With increasing returns to scale, a proportionate increase in inputs increases
output more than in proportion. So suppose that doubling inputs more than doubles
output. That is, suppose that employing 2K and 12L results in an output of eight. Note what this implies. When output is three, the average amount of capital per unit of out-
put is 0.33K, and the average amount of labor is 2L. At eight units of output average input requirements have fallen to 0.25K and 1.5L. Increasing returns to scale mean that at higher output levels each unit of output requires (on average) a smaller quantity of
all inputs. Because inputs are available at fixed prices, however, smaller average input
requirements imply smaller average unit costs. Therefore, increasing returns to scale imply that average unit cost is falling.
By the same reasoning, constant returns to scale imply a constant average cost, and decreasing returns to scale imply a rising average cost. Insofar as returns to scale are first increasing and thereafter decreasing, the long-run average cost curve will be -shaped.
Once again, the cost curves reflect the firm’s underlying production technology.
It is important to note that as a firm alters its level of output in the long run, it may also
want to alter its input proportions. If input proportions are changed, then the concept of
returns to scale does not apply—since not all inputs will be “scaled” up or down propor-
tionately. Instead, economists use the term economies of scale to refer to the case where a firm can increase its output more than in proportion to its total input cost. Diseconomies of scale are said to apply when a firm’s output increases less than in proportion to its total input cost. Increasing returns to scale thus imply economies of scale, but the reverse need
not be true. That is, if firms find it advantageous to alter their input proportions with their
output, then they may confront economies of scale but not increasing returns to scale. In the
more general case, therefore, a firm’s long-run average cost curve will be -shaped if econ-
omies of scale are present at low rates of output and diseconomies of scale prevail at high
output levels.7
The Long Run and Short Run Revisited It is convenient to think of the long run as a planning or investment horizon. In making
long-run decisions, a firm decides what scale of plant to build or purchase, how much and
what type of specialized equipment to install, whether to train existing workers for the new
equipment or hire new workers, and so on. The firm is planning ahead in making such deci-
sions. In effect, the firm is selecting what type of short-run situation it will be in later. Once
the plant is built, the firm must operate with that fixed input for a certain time period (the
short run) until enough time passes for a subsequent long-run adjustment to be made.
To clarify the relationship between the short run and the long run, let’s suppose that
there are only five scales of plant the firm can build, and associated with each plant size is a
short-run average total cost curve. Figure 8.7 shows the five short-run average cost curves
economies of scale a situation in which a firm can increase its output more than proportionally to its total input cost
diseconomies of scale a situation in which a firm’s output increases less than proportionally to its total input cost
7While a shape is typical, the long-run average cost curve could have a somewhat different shape. Sev- eral studies have suggested that for some firms, economies of scale occur at low rates of output, but once output reaches a certain threshold level, neither economies nor diseconomies of scale hold over an extended region before diseconomies of scale set in. When this result is true, the firm’s long-run average cost curve would have this shape: . When average cost reaches a minimum, it stays at that level over a wide range of output before it begins to rise. Do you see what the corresponding LMC curve will be shaped like in this case?
Long-Run Cost Curves 201
C08.INDD 11:11:1:AM 08/06/2014 PAGE 201Trim Size: 203.2 mm X 254 mm
as SAC1, SAC2, and so on. The firm can choose only one plant size: Which will it be? The plant size depends on what output the firm expects will be appropriate—which, of course,
depends on demand conditions not yet considered. Suppose, however, that the firm believes
q1 to be the appropriate output. The firm could build the smallest-sized plant (SAC1) and produce q1 at a unit cost of $50,000, or it could build the next larger plant (SAC2) and pro- duce at a unit cost of $55,000. Of course, the firm will build the plant size that permits it to
produce q1 at the lowest average (and hence total) cost—in this instance, the smallest-sized plant.
The long-run average cost curve is defined as the lowest average cost attainable when all inputs are variable—that is, when any plant size can be constructed. In this case, with only five options, point A on SAC1 is one point on the long-run average cost curve, because it shows the lowest unit cost for q1. Similarly, point C on SAC2 is a second point on the LAC curve. With only five options, the entire LAC curve is shown as the heavy scalloped sections of the SAC curves, because each of the segments indicates the lowest unit cost pos- sible for the corresponding level of output.
Once the firm builds the plant, its options in the immediate future will be dictated by
the SAC curve selected when the long-run decision was implemented. If the firm builds the smallest-sized plant, it is temporarily stuck with that decision even if q1 turns out to be an inappropriate level of output after all. In that event the firm must determine whether to
make another long-run decision to change its scale of operations.
In general, the firm will have many more than five scales of plant to choose from. When
a large number of options exist—for example, when there are a dozen scales of plant
between SAC1 and SAC2—the long-run average cost curve effectively becomes a smooth curve, such as LAC in Figure 8.7. Each point on this curve is associated with a different short-run scale of operation that the firm could choose.
Output
SAC2
SAC1
SAC3
SAC4
SAC5
LAC
A
B
C
q1
$55,000
$50,000
0
Average cost
Short- and Long-Run Average Cost Curves When a firm has five scales of plant from which to choose, the long-run average cost curve is the heavy scalloped portion of the five SAC curves. As the number of possible scales of plant increases, the long-run average cost curve becomes the smooth -shaped LAC curve.
Figure 8.7
202 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 202Trim Size: 203.2 mm X 254 mm
8.7 Learning by Doing Long-run cost is lower than short-run cost because of the greater flexibility in input usage a
firm possesses in the long run. Another reason long-run cost may lie below short-run cost is
learning by doing: improvements in the effectiveness with which a firm can combine inputs to produce final output that result from the experience gained over time with the pro-
duction process. Think about your own history with learning to play a new sport such as
tennis or golf. While practice may not have made you perfect‚ it likely made you a better
tennis player or golfer.
Learning by doing can result from many different factors. For example‚ workers may
become more skilled at their task the more times they repeat the task. Managers may dis-
cover superior ways‚ over time‚ to organize the production process. More effective input
combinations may be found to produce a unit of final output.
Learning by doing was first noted in the 1920s by the commander of Wright-Patterson
Air Force Base in Dayton‚ Ohio‚ who observed that the number of labor hours to manufac-
ture an airplane decreased as the number of airplanes manufactured increased. Similar stud-
ies of aircraft production during World War II indicated a drop in labor costs of 20 percent
with each doubling of cumulative output.
Learning by doing should not be confused with economies of scale. As shown in
Figure 8.8‚ suppose that a firm’s average cost curve is depicted as AC. The presence of economies of scale as output is expanded from q1 to q2 is indicated by the firm’s average cost of production falling from Aq1 to Bq2 along average cost curve AC. Learning by doing is represented by a drop in the height of the average cost curve at all output levels (from AC
learning by doing improvements in productivity resulting from a firm’s cumulative output experience
A ′
A
q2q1 Output
Learning by doing
Economies of scale
Dollars per unit
B ′
B
AC ′
AC
0
Learning by Doing Versus Economies of Scale Economies of scale are indicated by the movement from A to B along AC. Learning by doing advantages are represented by a downward shift of the average cost curve from AC to AC′.
Figure 8.8
Importance of Cost Curves to Market Structure 203
C08.INDD 11:11:1:AM 08/06/2014 PAGE 203Trim Size: 203.2 mm X 254 mm
to AC′) resulting from the experience gained from increases in cumulative output. Thanks to the benefits of learning by doing‚ for example‚ it now costs less to produce q1 in any given period (A′q1) than before (Aq1). It now also costs less to produce q2 (B′q2) than before (Bq2).
The Advantages of Learning by Doing to Pioneering Firms When there is the possibility of learning by doing‚ an industry’s pioneering firm has an
advantage over rivals in that it can‚ everything else being equal‚ attain lower production
costs through its greater cumulative production experience. Any given firm also has an
incentive to produce more in any given period so as to realize lower production costs in
future periods.
A pioneering firm’s advantage and the incentive to produce more today due to learning
by doing are mitigated‚ however‚ to the extent that the benefits associated with cumulative
experience spill over to other firms through production techniques becoming public knowl-
edge or skilled employees being hired away. Learning by doing advantages‚ moreover‚ are
attenuated in industries where general improvements in technology are rapid enough (com-
puters and semiconductors‚ to name a few) to catapult less established firms ahead of more
experienced rivals through new generations of products.
8.8 Importance of Cost Curves to Market Structure Although most firms have -shaped long-run average cost curves, the level of output at
which cost per unit reaches a minimum varies from firm to firm and from industry to indus-
try. The scale of operations at which average cost per unit reaches a minimum, called the
minimum efficient scale, may be immense for automobile producers but relatively small for apparel makers. These differences occur because the increasing returns to scale that are
responsible for the declining portion of the long-run average cost curve primarily reflect
technological factors, and the technology governing production differs significantly from
one good to another.
The minimum efficient scale for a typical firm in an industry has a major impact on
the industry’s structure—the number of firms in the industry and thus the proportion of
overall industry output accounted for by each member firm. More precisely, what mat-
ters is the output level where average cost reaches a minimum in comparison with the
total industry demand for the product. Figure 8.9 illustrates this point. When the demand
curve for legal services, for example, is D, suppose that the typical firm has a long-run average cost curve shown by LAC1. Average cost reaches a minimum at $30,000 per unit of output, and at a price of $30,000 the total quantity demanded by consumers is L1. The representative law firm can produce a total 0.05L1 at a unit cost of $30,000, or one- twentieth of the total quantity demanded by consumers. Insofar as LAC1 is typical for firms producing legal services, the industry can accommodate 20 firms, each producing 5
percent of the total output. With such a large number of firms, the industry is likely to be
highly competitive.
In contrast, suppose that production technology dictates a cost curve like LAC2. The minimum efficient scale is then half the total quantity demanded at a price of $30,000. With
such cost curves, the industry would tend to become dominated by a few (probably two
in this case) large law firms. Suppose that a small firm tried to compete in this market.
Operating on a small scale, the firm would have a unit cost of more than $30,000; it might,
for example, be operating at point A, where average cost is $50,000 per unit. Such a firm would tend to be driven out of business because it could be undersold by a larger operation
minimum efficient scale the scale of operations at which average cost per unit reaches a minimum
204 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 204Trim Size: 203.2 mm X 254 mm
Table 8.2 Minimum Efficient Plant Scales
Industry Minimum Efficient Scale
(per year) Percentage of U.S. Quantity
Demanded Beer brewing 4.5 million barrels 3.4
Cigarettes 36 billion cigarettes 6.6
Petroleum refining 73 million barrels 1.9
Glass bottles 133,000 tons 1.5
Cement 7 million barrels 1.7
Steel 4 million tons 2.6
Refrigerators 800,000 units 14.1
Car batteries 1 million units 1.9
Source: Frederic M. Scherer, Alan Beckenstein, Erich Kaufer, and R. D. Murphy, The Economics of Multi-Plant Operation: An International Comparison Study (Cambridge, MA.: Harvard University Press, 1975), pp. 80–94.
producing at a unit cost of $30,000. For this reason the industry gravitates toward a small
number of large firms. Thus, technology of production, which largely determines at what
output unit cost is at a minimum for a firm, is an important factor in determining whether
an industry is composed of a large or small number of firms, although other factors may
prove influential as well.
Economists have estimated the minimum efficient scale in different industries using
a variety of techniques. Table 8.2 reports the results of one such study. The second
A
D
Legal servicesL1
LAC1
LAC2
0.5L10.05L10
$50,000
$30,000
Dollars per unit
Cost Curves and the Structure of Industry The level of output at which long-run average cost is at a minimum relative to market demand has important implications for the structure of the industry. If all firms have LAC curves like LAC1, 20 firms can coexist in the industry; if the LAC2 curve is typical, only two firms are likely to survive.
Figure 8.9
Using Cost Curves : Control l ing Pol lut ion 205
C08.INDD 11:11:1:AM 08/06/2014 PAGE 205Trim Size: 203.2 mm X 254 mm
column shows the minimum efficient scale of operation for several major manufactur-
ing industries. The last column gives these outputs as a percentage of U.S. quantity
demanded. Although there are substantial variations, in all cases except the refrigerator
industry the minimum efficient scale tends to be small relative to the size of the market.
Other studies have also reached the conclusion that most industries can accommodate a
relatively large number of firms each operating at the minimum efficient scale. There are
exceptions, of course. For example, the production of civilian airplanes has been esti-
mated to require plant sizes that are large relative to the market to realize minimum unit
production costs.
APPLICATION 8.6
The minimum efficient scale of production varies across industries as well as over time within industries. Take the case of the beer industry in the United States where the number of breweries fell from 1,400 in 1914 to fewer than 40 in 1978 and then grew again to over 2,750 by 2014.8
While certain demand-side changes (the growth of the market) contributed to the declining number of breweries between 1914 and 1978, technological changes also served to increase the minimum efficient scale of production. Pas- teurization and the advent of at-home consumption from cans and bottles (versus a tavern’s nonpasteurized draft) favored large brewers. Moreover, the number of bottles or cans that a brewery’s filling and sealing line could pro- duce per minute increased dramatically. In addition to the increased economies to operating single plants, greater economies also developed in multi-plant operation—econ- omies in the form of lower advertising and regulatory com- pliance costs. For example, with the advent of television, it became cheaper for a big firm such as Anheuser-Busch to buy a 30-second commercial for nationwide broadcast on a major network than it was for 50 smaller breweries, each producing one-fiftieth the output of Anheuser-Busch, to
The Decline and Rise of Breweries in the United States
purchase the same advertising time on their local television stations.
What reversed the trend toward greater minimum efficient scales of production that prevailed up through the late 1970s? In October 1978, President Jimmy Carter signed a law that eliminated restrictions on home pro- duction of a small amount of beer and wine for personal consumption. Such regulatory restrictions previously had limited the ability of entrepreneurs to enter the market through experimenting with and developing craft beers that appealed to consumers seeking differentiated, niche products.
Furthermore, rising consumer income levels also have led to a shift toward premium, differentiated beers pro- duced by so-called micro-breweries and away from standard and economy brands produced by industry giants—firms such as AB InBev and MillerCoors that have continued to grow globally through mergers. As of 2014, craft beer now accounts for 7 percent of total beer sales in the United States. Furthermore, the micro-brewing seg- ment of the industry, led by firms such as Boston Beer Company (which produces Sam Adams) and Sierra Nevada Brewing Company, employs over 100,000 workers and generates over $12 billion in annual revenue. The trend toward smaller minimum efficient scales of production is expected to continue as micro-brewers acquire a larger share of the overall domestic beer market.
8This application is based on: Kenneth G. Elzinga, “The Beer Indus- try,” in The Structure of American Industry, 4th ed., edited by Walter Adams (New York: Macmillan, 1971); and “Beer in the United States,” Wikipedia.
8.9 Using Cost Curves: Controlling Pollution Many problems can be clarified if we pose them in terms of marginal cost. Let’s take
an example of some practical importance. Two oil refineries located in the Los Angeles
basin release some pollutants into the air in the process of distilling products such as
gasoline from petroleum. Because the pollutants harm people living in the Los Angeles
basin (for example, they increase the likelihood of respiratory illness), the government
206 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 206Trim Size: 203.2 mm X 254 mm
steps in and requires each refinery to curtail its pollution to 100 units (measured in some
appropriate way). This restriction limits the total pollutants discharged into the air to
200 units.
An important question here is whether the government’s program to reduce pollu-
tion to 200 units accomplishes the intended result at the lowest possible cost. To answer
it, we must think in terms of the marginal cost each refinery bears in reducing pollution.
Figure 8.10 illustrates the situation. In this diagram the amount of pollution generated by
each refinery is measured from right to left. For example, before the government restricts its activities, refinery A discharges 0P1 (300 units), and refinery B discharges 0P2 (250 units). Measuring pollution from right to left is the same as measuring pollution abatement—the
number of units by which pollution is reduced from its initial level—from left to right.
For example, if refinery B cuts back its pollution from 250 to 100 units, it has produced 150
units of pollution abatement, the distance P2X. The reason for adopting these units of measurement along the horizontal axis is that
each refinery incurs costs from reducing air pollution. The more pollution abatement, the
greater the total cost to a refinery. In fact, we can think of each refinery as having a mar-
ginal cost curve, starting from its initial position, that measures the marginal cost of pol-
lution abatement (cutting pollution by one unit) at each level of “output.” Exactly what
form this cost takes is immaterial; it may involve switching to a higher grade of petroleum,
installing scrubbers on the refinery’s smokestacks or building a more modern plant. Each
refinery would, of course, choose the least costly way of cutting back. The information
we require to find the cost of complying with the government’s regulation is each firm’s
marginal cost curve of pollution abatement. We expect these curves to rise if the law of
diminishing returns applies to pollution abatement; as more pollution is abated, eliminating
additional units becomes more expensive. In Figure 8.10 refinery A’s marginal cost curve
is MCA, and refinery B’s is MCB. (Because air pollution is a by-product of the production of other goods such as gasoline, these curves are likely to start at zero and rise throughout. Do
you see why?)
Cost of Pollution Abatement To produce a given output at the lowest possible cost, separate firms must be producing at a point where their marginal costs are equal. This condition is also true for pollution abatement. To reduce pollution to 200 units in the least costly way, refinery A should discharge 150 units and refinery B, 50 units.
MCA
MCB
X2 (50)
X1 (150)
P2 (250)
P1 (300)
X (100)
Dollars per unit
$4,000
$3,000
$2,000
PollutionPollution abatement
0
N
R
L
T
Figure 8.10
Using Cost Curves : Control l ing Pol lut ion 207
C08.INDD 11:11:1:AM 08/06/2014 PAGE 207Trim Size: 203.2 mm X 254 mm
When the government limits each refinery to 100 pollution units, refinery A will operate
at point R on MCA, and refinery B will be at point T on MCB. (We assume that this restric- tion is not so costly that either refinery is put out of business.) Now only 200 pollution units
are released into the air, a reduction from 550. Is this the least costly way to limit pollution
to 200 units? To see that it is not, consider the refineries’ marginal costs. At 100 pollution
units, the marginal cost of reducing pollution to refinery A is $4,000, but refinery B’s cost
is only $2,000. If refinery B were to cut back pollution by 1 more unit, it would add only
$2,000 to its cost. If refinery A increased pollution by 1 unit (produced 1 unit less of abate-
ment), its cost would fall by $4,000. Having refinery B pollute 1 unit less and refinery A
1 unit more leaves the total amount of pollution unchanged, but it reduces the refineries’
combined cost by $2,000 (plus $2,000 for refinery B and minus $4,000 for refinery A).
As long as the marginal costs differ, the total cost of pollution abatement can be
reduced by increasing abatement where its marginal cost is less and reducing it where
its marginal cost is greater. So shifting the production of abatement from refinery A to
refinery B should continue until B’s rising marginal cost just equals A’s falling marginal
cost. In the diagram, this equality occurs where refinery B cuts back pollution to 50 units
at point L on MCB and refinery A increases pollution by 50 units to point N on MCA. At these points the marginal cost to both refineries equals $3,000. Each step of this realloca-
tion reduces A’s cost by more than it increases B’s cost, so the two refineries still gener-
ate the same amount of total pollution (200 units) but at a lower total cost. Any further
shifting of pollution abatement from A to B would increase total cost. To limit pollution
to 200 units in the least costly way, the refineries should produce at a level where their
marginal costs are equal.
When refinery B increases its pollution abatement from P2X to P2X2, its total cost rises by the shaded area TLX2X. This area in effect sums the marginal cost of each successive unit of abatement from P2X to P2X2. For instance, the first unit (reduction in pollution from 100 units) beyond P2X adds $2,000 to total cost; the next, perhaps $2,050; the third, $2,100; and so on. So the addition to total cost from the first three units in this case is $6,150. (In
general, the area under a marginal cost curve between two levels of output measures how
much total cost rises when moving from the lower to the higher output.) Refinery A’s total
cost falls by NRXX1 when it reduces its abatement from P1X to P1X1. The cost saving to refinery A, NRXX1, is clearly larger than the additional cost to refinery B, TLX2X (recall that XX2 = XX1).
If the government simply mandates that each refinery restrict its pollution to the same
level, this policy will normally result in a higher cost for pollution control than necessary.
The only exception would be if the marginal costs of the refineries are equal when they are
polluting the same amount, and this is unlikely to happen. Our analysis, therefore, leads to
an important conclusion: to minimize the cost of pollution control, firms should operate
where their marginal costs are equal. The same rule also applies to emissions by other sta-
tionary sources as well as mobile sources of air pollutants. For example, emission standards
for automobiles should take into account the fact that the marginal cost of limiting emis-
sions varies from one make of automobile to another and between older and newer auto-
mobiles. Applying the same standards to all automobiles (as is done now) imposes a higher
total cost than is necessary.
Realistically, can the government apply this rule? In terms of Figure 8.10, to allocate
pollution abatement responsibility in the least costly way, the government needs to know
both refineries’ marginal cost curves, and such information would be virtually impossible
to obtain with any degree of accuracy. Without knowledge of the MC curves, the govern- ment cannot place a limit on each refinery’s pollution so that all refineries’ marginal costs
are equal. Does this conclusion mean that our economic analysis is of little practical value?
Not at all. As we will see in Chapter 20, there is a way for the government to limit pollution
efficiently without knowing the MC curves.
208 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 208Trim Size: 203.2 mm X 254 mm
8.10 Economies of Scope So far we have been focusing on the cost to firms of producing single products. Most firms,
however, produce more than one product. For example, airlines typically provide different
classes of service on a given city-pair route as well as serving numerous city pairs. Most
business schools produce research and teaching. And telecommunications firms are capable
of providing voice and video services through the same fiber-optic cable.
Economies of scope are present if it is cheaper for one firm to produce products jointly than it is for separate firms to produce the same products independently. Take
the case of business schools. Economies of scope are likely to be present because cer-
tain inputs such as the library, administration, office space, and faculty are capable of
contributing to both the teaching and research output of an institution. Were separate
firms to produce the two independently, they would have to build their own individual
facilities and employ separate faculty and administrators. Such duplication would in all
likelihood make the overall cost of production across the separate research and teach-
ing schools higher than the cost of joint production of research and teaching by a single
school.
Suppose that the total cost to a business school of producing research (R) and teaching (T) can be represented by TC(R,T). If only research is produced by the school, total cost equals TC(R,0). Conversely, if teaching is the only output, total cost equals TC(0,T). Econ- omies of scope can be said to be present if the following inequality holds:
TC R T TC R TC T( , ) [ ( , ) ( , )].< +0 0 (8)
In other words, when it is more expensive to produce research and teaching independently
than jointly, production is characterized by economies of scope. Diseconomies of scope would characterize production if the above inequality were reversed: if it was cheaper for
research and teaching to be produced independently by separate entities than jointly under
the same institutional roof.
Economies of scope are important since they provide an efficiency reason for the
existence of multiproduct conglomerates. For example, if there are inputs such as fiber-
optic cable and management that contribute to the provision of both voice and video
services, it will be advantageous for telecommunications firms to span both markets.
If hubs lower the total cost of providing air service between a set of different city-pair
markets, it will be cheaper to have one airline serve all city pairs through a single hub-
and-spoke system than to have separate airlines provide independent service to each par-
ticular route. The belief that the combination of AOL’s Internet customer base and Time
Warner’s ownership of entertainment programming content (such as CNN, Time maga- zine, the television series Friends, and a record label with stars such as Alanis Mori- sette) would be better able to market the services of both companies is what prompted
the $350 billion merger deal between the two companies in 2000 (the biggest corporate
merger in history as of that date)—although this belief was borne out in practice.
There is, of course, no relationship between economies of scope and economies of
scale. That is, the Wharton School of Business may face economies of scope in produc-
ing teaching and research jointly, yet face diseconomies of scale in the production of
research (holding constant the amount of teaching performed, doubling research output
may require more than a doubling of total input cost). Conversely, if Wharton faces
economies of scale in the production of both teaching and research, it does not imply
that the joint production of the two products is less expensive than if they are produced
separately.
economies of scope a case where it is cheaper for one firm to produce products jointly than it is for separate firms to produce the same products independently
diseconomies of scope a case where it is cheaper for separate products to be produced independently than for one firm to produce the same products jointly
Est imat ing Cost Funct ions 209
C08.INDD 11:11:1:AM 08/06/2014 PAGE 209Trim Size: 203.2 mm X 254 mm
8.11 Estimating Cost Functions As with demand and production functions, cost functions can be empirically estimated in
a variety of ways. Surveys are a commonly used method to determine the minimum effi-
cient scale of production in an industry. The case described in the preceding chapter of the
McKinsey & Company summer intern charged with investigating the pig chow industry
provides an example. The student used a telephone survey to determine plant sizes that the
current firms in the industry were using. This method is sometimes known as the
new entrant or survivor technique.9 Econometric estimation is another vehicle for estimating cost functions. For exam-
ple, a cubic total cost function will generate conventional -shaped MC and AC curves such as those shown earlier in this chapter. A cubic total cost function takes the follow-
ing form:
TC a bq cq dq= + + + 2 3, (9)
where q represents output, and a, b, c, and d are numerical values to be estimated. The term cubic derives from the fact that, in the assumed equation, total cost is taken to be a func- tion of output, q, up to the third power. Based on the estimated constant and coefficients for the above cubic cost function, it is possible to determine a number of things about a firm’s
production cost. For example, if the assumed cubic total cost function is valid, the inter-
cept, a, indicates the firm’s total fixed cost—it does not vary as a function of output, q; the firm’s average total cost and marginal cost curves are ATC = (a/q) + b + cq + dq2 and MC = b + 2cq + 3dq2, respectively; and the firm’s minimum efficient scale (and the extent of any economies of scale confronted by the firm) can be determined by finding the output level at
which ATC is minimized.10 Suppose that for the above cubic total cost function a = 0, b = 400, c = −50, and d = 5.
The average total cost and marginal cost curves would look as depicted in Figure 8.11a.
The minimum efficient scale of the firm would be where ATC is minimized—the output where ATC = MC. Setting ATC equal to MC and solving for q gives us an estimated mini- mum efficient scale of 5. (The corresponding height of the ATC and MC curves at this out- put is $275.)
If a quadratic total cost function was more appropriate than a cubic total cost function, we would estimate an equation of the following form:
TC a bq cq= + + 2, (10)
where the term quadratic derives from the fact that total cost is a function of output, q, up to the second power. A quadratic total cost function is also associated with a -shaped ATC curve, but not a -shaped MC curve. As depicted in Figure 8.11b, the MC curve generated by a quadratic total cost function is an upward-sloping straight line and is equal to, in our
case, MC = b + 2cq.11
new entrant/survivor technique a method for determining the minimum efficient scale of production in an industry, based on investigating the plant sizes either being built or used by firms in the industry
9The survivor technique was first proposed by economist George Stigler in his article “Economies of Scale,” Journal of Law and Economics, 1, No. 2 (October 1958), pp. 54–71. The idea helped win him the 1990 Nobel Prize in economics. 10The marginal cost curve is derived by using calculus. 11As before, this MC curve is derived by using calculus.
210 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 210Trim Size: 203.2 mm X 254 mm
SUMMARY
Cost is ultimately associated with the use of inputs
that have alternative uses.
To economists, cost means opportunity cost—sacri-
ficed alternatives when inputs are used to produce one
product rather than some other product.
In the short run, firms cannot change the quantities of
some inputs, so output can be altered only by varying the
use of variable inputs.
The law of diminishing marginal returns dictates the
shape of a short-run cost curve.
The short-run marginal cost curve rises beyond the
point at which diminishing marginal returns set in, and
it intersects the -shaped average variable and average
total cost curves at their minimum points.
In the long run, firms can vary all inputs.
The expansion path in an isocost–isoquant diagram
shows the least costly combinations of inputs required
to produce various levels of output. It identifies the
lowest total cost at which each output can be produced
when all inputs can be varied. The same information is
also conveyed by firms’ long-run cost curves.
If input proportions are held constant, returns to scale
determine the shape of a firm’s long-run cost curves.
With increasing returns to scale at low output and
decreasing returns to scale at high outputs, the long-run
average cost curve will be -shaped. It shows the lowest
per-unit cost at which each output can be produced.
The level of output at which the long-run average cost
curve reaches a minimum depends on technological con-
ditions and varies across products and firms.
In the more general case, when firms can alter input
proportions with output, the shape of their long-run
cost curves will depend on whether output changes
are more than proportionate to any change in total
input cost.
Economies of scale are present when a firm’s output
increases more than proportionally with a change in total
input cost.
All cost curves for individual firms are drawn on the
assumption that input prices are given. A change in one or
more input prices causes the cost curves to shift.
Long-run cost is lower than short-run cost because of
greater flexibility in input usage as well as any advan-
tages associated with learning by doing.
Although economies of scale determine a long-
run average cost curve’s shape as a function of a
Output (q)5
(a)
$275
Dollars per unit
ATC = 400 – 50q + 5q 2
MC = 400 – 100q + 15q 2
0
(b)
Dollars per unit
Output (q)0
ATC = + b + cq a–– q
MC = b + 2cq
b
Different Possible Cost Functions (a) A cubic total cost function such as TC = a + bq + cq2 + dq3, where q is output, is associated with -shaped average total cost and marginal cost curves. In the case depicted, the intercept and coefficients are assumed to be a = 0, b = 400, c = −50, and d = 5. (b) A quadratic total cost function such as TC = a + bq + cq2 is associated with a -shaped ATC curve but generates an upward-sloping, straight-line MC curve.
Figure 8.11
C08.INDD 11:11:1:AM 08/06/2014 PAGE 211Trim Size: 203.2 mm X 254 mm
Review Quest ions and Problems 211
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
8.1 No Pain No Gain (NPNG), Inc., is a dental practice advo- cating a “natural” approach to dentistry. Namely, NPNG
specializes in providing root canal operations without the
administration of pain-killing drugs such as Novocain. If out-
put (q) is measured as the number of root canals performed on a daily basis, define the following measures of NPNG’s short-
run cost: TFC, TVC, TC, MC, AFC, AVC, and ATC. Then fill in the spaces in the accompanying table:
q TFC TVC TC MC AFC AVC ATC
1 $100 $50
2 $30
3 $40
4 $270
5 $70
*8.2 How is the law of diminishing marginal returns related to the shape of the short-run marginal cost curve? If the marginal
product of the variable input declines from the very start, what
will the short-run marginal cost and average cost curves look
like? If the marginal product first rises and then falls, what will
the cost curves look like?
8.3 Using the data from Question 7.1, construct the following relationships: (a) the average amount of labor used per unit of
output at each level of output, and (b) the additional amount of
labor required for each additional unit of output. Show these
relationships as curves, with labor per unit of output on the
vertical axis and output on the horizontal axis. How can these
curves be converted into AVC and MC curves?
*8.4 Why do we assume that a firm will try to produce its out- put by using the lowest-cost combination of inputs possible?
Does this reason also hold for production carried out by the
U.S. Postal Service and the American Red Cross?
8.5 If the long-run marginal cost curve is -shaped, must the long-run average cost curve also be -shaped?
8.6 What is the significance of a tangency between an isoquant and an isocost line?
8.7 When a firm produces an output in the least costly way, using labor and capital, MRTSLK = w/r. Does this relationship
mean that the firm sets input prices so that the ratio equals
MRTSLK?
8.8 At point E in Figure 8.4b, is MPK/r greater or less than MPL/w? How do you know? Use this inequality to explain how the firm can increase output without increasing its total cost by
using a different combination of inputs.
*8.9 Turner Construction is employing 10 acres of land and 50 tons of cement to produce 1,000 parking spaces. Land costs
$4,000 per acre and cement costs $12 per ton. For the input
quantities employed, MPL = 50 and MPC = 4. Show this situa- tion in an isoquant–isocost diagram. Explain, and show in the
diagram, how Turner can produce the same output at a lower
total cost.
8.10 Why do we assume that the prices of inputs are constant when we draw a firm’s cost curves? How does a change in the
price of an input affect the AC and MC curves?
8.11 Suppose that the long-run average cost curve declines at first, then reaches a minimum at which level it is horizontal
over an extended region, and then rises. Draw such a curve and
the associated LMC curve.
8.12 In Figure 8.4b, assume that the firm is currently produc- ing six units of output at point B on isoquant IQ6. Suppose now that the firm plans to expand output to nine units by hiring more
workers while continuing to use K2 units of capital (capital is assumed to be a fixed input in the short run). Identify the combi-
nation of workers and capital that will be used. Is the total cost
of producing nine units of output in this way greater than, less
than, or the same as the cost at point C? Support your answer by showing what has happened using LAC and SAC curves.
8.13 “Every point on Ford’s long-run cost curve corresponds to a point on some short-run cost curve, but not every point on
one of Ford’s short-run cost curves corresponds to a point on
the long-run cost curve.” Explain.
8.14 “In the United States more than 50 firms produce tex- tiles, but only three produce automobiles. This statistic shows
that government antimonopoly policy has been applied more
harshly to the textile industry than to the automobile industry.”
Can you give an alternative explanation for the difference in
the number of firms in the two industries?
8.15 Suppose that Marriott’s production function is character- ized by constant returns to scale at all output levels. What will
particular product’s output, firms often produce an
array of products.
If it is cheaper for a single firm to produce an array of
products than it is for an array of separate firms to inde-
pendently produce distinct products, economies of scope
apply. There is no necessary relationship between econo-
mies of scale and economies of scope.
A variety of techniques may be relied upon for empir-
ical estimation of cost functions, including surveys,
experimentation, and regression analysis.
212 The Cost of Product ion
C08.INDD 11:11:1:AM 08/06/2014 PAGE 212Trim Size: 203.2 mm X 254 mm
the firm’s long-run total, average, and marginal cost curves
look like?
8.16 If the Harvard Business School faces economies of scope in producing research and teaching, what does its production
possibility frontier (PPF) look like? If the University of Phoe- nix faces diseconomies of scope in the provision of research
and teaching, what does its PPF look like?
8.17 If the cubic total cost function described in the text applies to the production of output by a firm, and a = 0, b = 400, c = −50, and d = 5, what are the equations for the firm’s TFC, TVC, MC, AFC, AVC, and ATC?
8.18 Suppose that for the cubic total cost function discussed in the text, a = 0, b = 4, c = −5/2, and d = 5/3. At what output level is average total cost minimized? At what output level is average
variable cost minimized? Marginal cost? Average fixed cost?
8.19 Suppose that a firm’s total cost is best described by the following quadratic cost function:
TC = a + bq + cq2.
What are the equations for the firm’s TFC, TVC, MC, AFC, AVC, and ATC?
8.20 Suppose that for the quadratic total cost function in the preceding problem, a = 100, b = 6, and c = 1. At what output
level is average total cost minimized? At what output level is
average variable cost minimized? Marginal cost? Average
fixed cost? Provide a graph of the firm’s TFC, TVC, MC, AFC, AVC, and ATC curves.
8.21 If learning by doing applies to race car driving and the annual Indianapolis 500 race, what would you expect the his-
torical pattern to be in the miles per hour recorded by the win-
ning driver over the 1911–2014 period, knowing that racing
was stopped during World War I and World War II?
8.22 Explain why demand-side‚ bandwagon effects (covered in Chapter 4) may be more important in explaining the mar-
ket dominance of a software firm such as Microsoft than any
production-side economies of scale.
8.23 Explain why certain college hockey teams rely on a dif- ferent relative mix of speedy versus brawny players, based on
the size of their skating rink. (There is no regulation-size rink
in college hockey.) How do differences in this mix provide
evidence of college hockey teams applying the golden rule of
cost minimization? Likewise, explain why professional base-
ball teams may rely on a different relative mix of singles versus
power hitters based on the size of their baseball stadium and
how this phenomenon is related to the golden rule of cost mini-
mization. (There is also no regulation-size stadium in profes-
sional baseball.)
C09.INDD 12:6:43:PM 08/06/2014 PAGE 213Trim Size: 203.2 mm X 254 mm
213
CHAPTER 9 Profit Maximization in Perfectly Competitive Markets
As we have seen, the basic determinants of cost are the prices and productivities of inputs. But a knowledge of cost conditions alone does not explain a firm’s output level. Cost curves
identify only the minimum cost at which the firm can produce various outputs.
For a firm interested in maximizing profit, cost and demand conditions jointly deter-
mine the optimal output level. So, to complete the model of output determination, we need
to specify the demand curve confronting the firm. The demand curve determines the sales
revenue at different volumes of output. In this chapter we concentrate on perfect competi-
tion and the demand curve facing a firm operating in such a market structure. Later chapters
focus on the demand curve confronting a firm and the firm’s optimal output when competi-
tion is imperfect.
Learning Objectives
Outline the conditions that characterize perfect competition. Explain why it is appropriate to assume profit maximization on the part of firms. Show why the fact that a competitive firm is a price taker implies that the demand curve for the firm is perfectly horizontal. Explain a competitive firm’s optimal output choice in the short run and how the firm’s short- run supply curve may be derived through this output selection. Describe the firm’s short-run supply curve. Explain how the short-run industry supply curve is derived. Define the conditions characterizing long-run competitive equilibrium. Understand how the long-run industry supply curve describes the relationship between price and industry output over the long run, taking into account how input prices may be affected by an industry’s expansion/contraction. Analyze the extent to which the competitive market model applies.
Memorable Quote “The price which society pays for the law of competition, like the price it pays for cheap comforts and luxuries, is great, but the advantages of this law are also greater still than its cost . . . for it is to this law that we owe our wonderful material development, which brings improved conditions in its train. But, whether the law be benign or not, we must say of it: It is here; we cannot evade it; no substitutes for it have been found; and while the law may be sometimes hard for the individual, it is best for the race, because it ensures the survival of the fittest in every department.”
—Andrew Carnegie, Scottish-born American industrialist and philanthropist
214 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 214Trim Size: 203.2 mm X 254 mm
We will explore how a firm’s optimal output within a perfectly competitive market
structure responds to changes in price and cost. This information can be used to derive the
firm’s supply curve and, in turn, the industry supply curve. We also address the long-run
outcome in perfect competition and contrast it with short-run responses.
9.1 The Assumptions of Perfect Competition In common usage, competition refers to intense rivalry among businesses. Microsoft and Google, Nike and Reebok, and Pepsi and Coke are all competitors in this sense. Each firm
makes a business decision—whether to introduce a new product, advertise existing prod-
ucts more forcefully, or enhance product quality—only after considering the effect on com-
petitors and their likely response.
The Four Conditions Characterizing Perfect Competition The economist’s formal model of perfect competition bears little resemblance to this pic-
ture. Perfect competition is distinguished largely by its impersonal nature. More specifi- cally, four conditions characterize a perfectly competitive industry.
1. Large numbers of buyers and sellers. The presence of a great many independent par- ticipants on each side of the market, none of whom is large in relation to total industry
sales or purchases, normally guarantees that individual participants’ actions will not sig-
nificantly affect the market price and overall industry output. In a market with many firms,
each firm recognizes that its impact on the overall market is negligible, and consequently
does not view other firms as personal rivals.
2. Free entry and exit. Industry adjustments to changing market conditions are always accompanied by resources entering or leaving the industry. As an industry expands, it uses
more labor, capital, and so on; resources enter the industry. Similarly, resources leave a
contracting industry. A perfectly competitive market requires that there be no differential
impediments across firms in the mobility of resources into, around, and out of an industry.
This condition is sometimes called free entry and exit. Examples of barriers to entry and exit include an incumbent firm with an exclusive government patent or operating license
and economies of scale that impede the entry of new firms.
3. Product homogeneity. All the firms in the industry must be producing a standardized or homogeneous product. In consumers’ eyes, the goods produced by the industry’s firms are perfect substitutes for one another. This assumption allows us to add the outputs of the
separate firms and talk meaningfully about the industry and its total output. It also con-
tributes to the establishment of a uniform price for the product. One farmer will be unable
to sell corn for a higher price than another if the products are viewed as interchangeable,
because consumers will always purchase from the lower-priced source.
4. Perfect information. Firms, consumers, and resource owners must have all the infor- mation necessary to make the correct economic decisions. For firms, for example, the
relevant information is knowledge of the production technology, input prices, and the price
at which the product can be sold. For consumers, the relevant information is a knowledge
of their own preferences and the prices of the various goods of interest to them. Moreover,
the consumers, in their role as suppliers of inputs, must know the remuneration they can
receive for supplying productive services.
Probably no industry completely satisfies all four conditions. Agricultural markets come
close, although government involvement in such markets keeps them from fully satisfying
the four conditions. Most industries satisfy some conditions well but not others. Even though
perfect competition an economic model characterized by the assumption of (1) a large number of buyers and sellers, (2) free entry and exit, (3) product homogeneity, and (4) perfect information
free entry and exit a situation in which there are no differential impediments across firms in the mobility of resources into and out of an industry
homogeneous products standardized products that, in the eyes of consumers, are perfect substitutes for one another
Prof i t Maximizat ion 215
C09.INDD 12:6:43:PM 08/06/2014 PAGE 215Trim Size: 203.2 mm X 254 mm
the number of market participants in the gasoline retailing business is large and entry into
the business is fairly easy, for example, not all gasoline brands are the same. Some brands
have higher octane and more detergents, and are better for the environment. Certain stations
are closer to particular consumers and thereby more convenient or offer better complements
such as full service, food-marts, and pumps that allow customers to pay for their purchases
by inserting a credit card. Moreover, consumers are rarely perfectly informed about the
prices all retailers are charging.
The fact that only a few industries may fully satisfy the four conditions does not mean
that the study of perfect competition is unwarranted. Many industries come close enough to
satisfying the four conditions to make the perfectly competitive model quite useful. Take
the case of gasoline retailing. Although product homogeneity and perfect information may
not fully apply, the extent to which an individual gas station has some choice over what
price to charge per gallon is probably limited to a very narrow band of just a few cents.
Such a narrow pricing power band is pretty close to having no significant impact over price,
as predicted by the competitive model.
9.2 Profit Maximization In perfectly (and imperfectly) competitive markets, is it appropriate to assume profit maximization on the part of firms? At the outset we should recognize that any profit real- ized by a business belongs to the business owner(s). For the millions of small businesses
with only one owner-manager, decisions concerning what products to carry, whom to
employ, what price to charge, and so on, will be heavily influenced by the way the owner’s
profit is affected. Owners of such businesses may well have goals such as early retirement or
expensive educations for their children. These goals, however, are not inconsistent with the
assumption of profit maximization. Since money is a means to many ends, early retirement
or college educations can more easily be afforded when the owner makes more money.
A possible problem with assuming profit maximization is that the owner-manager can-
not have detailed knowledge of the cost and revenue associated with each action that could
be taken to maximize profit. Economic theory, however, does not require that firms actually
know or think in terms of marginal cost and revenue, only that they behave as if they did.
Firms may come close enough to maximizing profit by trial and error, emulation of suc-
cessful firms, following rules of thumb or blind luck for the assumption to be a fruitful one.
When we move from the small, owner-managed firm to the large, modern corporation,
another potential criticism of the profit maximization assumption arises. A characteristic
of most large corporations is that the stockholder-owners themselves do not make the day-
to-day decisions about price, employment, advertising, and so on. Instead, salaried person-
nel of the corporation—managers—make these decisions. And so there is a separation of
ownership and control in the corporation; managers control the firm, but stockholders own
it. It is safe to assume that stockholders wish to make as much money on their investment
as possible, but it is virtually impossible for them to constantly monitor their managers’
actions. Therefore, managers will have some discretion, and some of their decisions may
conflict with the stockholder- owners’ profit-maximizing goals.
While managers may have some discretion to deviate from the profit-maximizing goals
of firms’ shareholder-owners, several factors limit the exercise of such discretion. For
example, stockholder-owners often link business managers’ compensation to profits, some-
times paying them in part with shares of stock or stock options, in order to give an incen-
tive to pursue profits more actively. In addition, the profitability managers achieve in a
given enterprise will affect their job prospects with others. And, finally, if managers do not
make as large a profit as possible, stock prices, which tend to reflect profitability (especially
projected profitability), will be lower than need be. Undervalued stock creates an incentive
profit maximization the assumption that firms select an output level so as to maximize profit
216 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 216Trim Size: 203.2 mm X 254 mm
for outsiders, or “raiders,” to buy a controlling interest in the firm and replace the inefficient
management team. A firm that neglects profit opportunities too often leaves itself open to
such a takeover bid, a fairly common occurrence in the corporate world.
Operating in a competitive market provides yet another reason it is safe to assume that a
firm will pursue profit maximization as a goal. Consider firms in the hospitality industry,
such as Hilton and Marriott. Suppose some of them virtually ignore profit, either through
ignorance, negligence, bad luck or intention. Their cost of labor is too high, they fail to
minimize waste in their food service, they neglect training for the staff who serve custom-
ers, and so forth. Other firms, whether through superior management, close attention to
costs, or the good luck of being closest to a newly enlarged convention center, produce the
right type of lodging in the appropriate quantity and in the least costly way. What will hap-
pen when these firms compete for customers? Clearly, the firms that come closer to maxi-
mizing profit will make money and prosper; the others will suffer losses. In general, in
competitive markets, firms that do not approximate profit-maximizing behavior fail; the
survivors will be the firms that, intentionally or not, make the appropriate profit-maximizing
decisions. This observation is called the survivor principle and provides a practical defense for the assumption of profit maximization.
While it goes without saying that firms must pay close attention to profit, some devia-
tion from single-minded profit maximization may still occur. Most economists, however,
believe that the profit maximization assumption provides a close enough approximation to
be useful in analyzing many problems, and it has become the standard assumption regard-
ing the firm’s behavior. While the assumption may not adequately explain why Billy, the
survivor principle the observation that in competitive markets, firms that do not approximate profit- maximizing behavior fail, and survivors are those firms that, intentionally or not, make the appropriate profit- maximizing decisions
APPLICATION 9.1
How can stockholder-owners (principals) best ensure that the managers (agents) running large, modern corporations make decisions consonant with stockholders’ profit-maximiz- ing goals? A central insight of research on this principal-agent problem is that corporate executives will be more likely to
align their actions with stock- holders’ financial objectives when their compensation is tied to their firm’s profits, thereby giving man- agers “skin in the game.”1 Doing so, however, is not as straightfor- ward as it may at first blush appear. The meltdown in finan- cial markets that precipitated the recession of 2007–2009 illus- trates why.
For example, to the extent that bank CEOs are rewarded based on the short-run earn- ings performance of their firms relative to peer firms, CEOs are incented to take on greater short-run risks so as to boost earnings relative to rival firms. Moreover, excessive risk tak- ing also is promoted when traders at financial institutions are rewarded solely on the basis of their short-run impact
principal-agent problem a situation in which a principal’s agent (such as the manager of a firm representing the firm’s stockholders) has an incentive to act in a manner inconsistent with the interests of the principal
Aligning Managerial Actions with Shareholder Interests: Lessons from the Recession of 2007–2009
on earnings and are not required to cover the costs of insur- ing against any longer-run impacts of their decisions on a bank’s performance.
The recession of 2007–2009 has led to a reevaluation of incentive-based compensation schemes and the extent to which senior bank executives and leading traders at financial institutions are rewarded through restricted stock or stock options—restricted in the sense that the manager cannot sell the shares or exercise the options for a certain period of time. Such a compensation structure diminishes the incentive managers have to engage in excessively risky short-run behavior, thereby better promoting the long-run interests of shareholders.
1This application is based on: Michael C. Jensen and William H. Meckling, “Theory of the Firm: Managerial Behavior, Agency Cost, and Ownership Structure,” Journal of Financial Economics, Vol. 3, No. 4 (October 1976), pp. 305–360; Douglas W. Diamond and Raghuram C. Rajan, “The Credit Crisis: Conjectures About Causes and Remedies,” American Economic Review, Vol. 99, No. 2 (May 2009), pp. 606–610; and Kenneth R. French et al., The Squam Lake Report: Fixing the Financial System (Princeton, NJ: Princeton University Press, 2010).
The Demand Curve for a Competit ive F i rm 217
C09.INDD 12:6:43:PM 08/06/2014 PAGE 217Trim Size: 203.2 mm X 254 mm
company president, hires his ne’er-do-well brother-in-law, Roger, or why RJR Nabisco
appears to have an excessively large fleet of corporate jets, it does not pretend to try.
Instead, it is designed to explain how a firm’s output will respond to a higher or a lower
price, a tax or a government regulation, a cost change, and so on. Recall that the ultimate
test of a theory is whether it explains and predicts well, and theories based on the assump-
tion of profit maximization have passed that test.
9.3 The Demand Curve for a Competitive Firm Assuming that firms are interested in maximizing profit, let’s examine the implications of
this assumption in a competitive market setting. Since a competitive market is character-
ized by a large number of firms selling the same product, each firm supplies only a small
fraction of the entire industry output. For example, one farmer may account for only one-
millionth of the entire corn industry’s output or one building contractor may supply 1 per-
cent of the total construction services in a particular city.
The nature of the demand curve confronting a single competitive firm follows directly
from the relatively insignificant contribution its output makes toward total supply. The
product’s price is determined by the interaction of the market supply of and demand for the
product. Because each firm produces such a small portion of the total supply, its output deci-
sions have a small effect on the market price. For simplicity the “small” effect is taken to be
a “zero” effect, and the demand curve facing a competitive firm is drawn to be perfectly
horizontal. A horizontal demand curve means the firm can sell as much output as it wants
without affecting the product’s price. Stated differently, a competitive firm is a price taker: the firm takes the price as given and does not expect its output decisions to affect price.
Figure 9.1 clarifies why the firm’s demand curve is drawn horizontally. In Figure 9.1b, the
total output of corn is measured horizontally, and the per-unit price is measured vertically.
The market demand curve is shown as D. The premise is that the total quantity offered for sale by all farms together interacts with this demand curve to determine price. If the com-
bined output of all corn farms is 15 billion bushels, the market price per bushel is $3.
Assume that the Costner farm is one of 1 million identical farms supplying corn, and it
is currently selling 15,000 bushels for $3 each, as shown in Figure 9.1a. The farm’s demand
curve is drawn horizontally as d in Figure 9.1a because its output variations will not have an appreciable effect on the market output and price. That is, if the Costner farm produced
15,000 fewer (an output of 0 bushels) or more bushels (an output of 30,000 bushels), it
would have an insignificant influence on the total industry output of 15 billion bushels.
Consequently, the market price will not be altered by the output actions of this single farm.
It is stuck with having to charge the market price of $3 no matter what output level it selects.
A horizontal demand curve has an elasticity of infinity. Since there are many homo-
geneous substitutes for any farm’s output and customers are perfectly informed in a per-
fectly competitive market, the quantity of corn demanded from the Costner farm equals
zero if the farm attempts to charge even a penny over the prevailing market price.
A firm’s average revenue (AR), or total revenue divided by output, is the same as the prevailing market price. As Figure 9.1 shows, if the market price is $3, then the Costner
farm will on average make $3 per bushel.
When a firm faces a horizontal demand curve, the market price also equals the firm’s
marginal revenue. Marginal revenue (MR) is defined as the change in total revenue when there is a one-unit change in output. A firm in a competitive market can sell one more unit
of output without reducing the price it receives for its previous units, so total revenue will
rise by an amount equal to the price. For example, if a farm is selling 15,000 bushels of
corn at a price of $3 per bushel, total revenue is $45,000. If the farm sells 15,001 bushels at
a price of $3 per bushel, as it can with a horizontal demand curve, total revenue rises from
price taker a firm or consumer who cannot affect the prevailing price through production and consumption decisions
average revenue (AR) total revenue divided by output
marginal revenue (MR) the change in total revenue when there is a one-unit change in output
218 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 218Trim Size: 203.2 mm X 254 mm
$45,000 to $45,003, or by $3. Once again the familiar average-marginal relationship can be
seen to apply. Where the average revenue is flat or constant, the marginal revenue equals
the average revenue (MR = AR). Note that the assumption of a horizontal demand curve confronting a competitive firm
does not mean that the price never changes. It just means that the firm, acting by itself, can-
not affect the going price. The market price may vary from time to time due to changes in
consumers’ incomes, technology, consumers’ preferences, and so on, but not because of
changes in the amount sold by a particular firm.
9.4 Short-Run Profit Maximization In the short run, a competitive firm operating with a fixed plant can vary its output by alter-
ing its employment of variable inputs. To see how a profit-maximizing firm decides on
what level of output to produce, let’s begin with a numerical illustration. The information
in Table 9.1 allows us to identify the output (q) that maximizes profit for a hypothetical competitive firm; it includes the short-run cost of production and revenue from the sale of
output. We know the firm is selling in a competitive market because the price (P) is con- stant at $12, regardless of the output level. Note that total revenue (TR) is equal to price times the quantity sold, and it rises in proportion to output since the price is constant. Total
cost (TC) rises with output in the familiar fashion, slowly at first and then more rapidly as the plant becomes more fully utilized and marginal cost rises. Total cost, when output is
zero, is $15, reflecting the total fixed cost.
Total profit (π) is the difference between total revenue and total cost. At low and high rates of output, profit is negative; that is, the firm would suffer losses. In particular, note
that the firm loses $15 if it produces no output at all because it still must pay its fixed
total revenue (TR) price times the quantity sold
total profit (π) the difference between total revenue and total cost
Price per bushel
Costner farm
$3.00
0 15,000
(a) (b)
d
Corn (bushels)
Price per bushel
Market
$3.00
0 15
S
D
Corn (billions of bushels)
The Competitive Farm’s Demand Curve (a) Because an individual farm supplies only a small portion of the market output, its demand curve, d, is perfectly elastic. (b) The interaction of market supply, S, and demand, D, determine the prevailing market price ($3) and output (15 billion bushels).
Figure 9.1
Short-Run Prof i t Maximizat ion 219
C09.INDD 12:6:43:PM 08/06/2014 PAGE 219Trim Size: 203.2 mm X 254 mm
cost when it shuts down. At an intermediate rate of output in this example, profit is posi-
tive. The firm, however, wishes to make as large a profit as possible, and maximum profit
occurs at an output of eight units where profit equals $11.10.
Note that maximizing total profit is generally not the same thing as maximizing average profit per unit (π/q) sold. The firm’s goal is to maximize its total profit, and that is achieved at an output of eight units. Profit per unit at that output is $1.39, but it could have
an even higher average profit, $1.57, by producing just seven units. Total profit is profit per
unit times the number of units sold, so a lower average profit can correspond to a higher
total profit if enough additional units are sold, as is true in the example of Table 9.1.
Figure 9.2 shows how we identify the most profitable level of output by using the total
revenue and total cost curves. The total revenue curve is a new relationship, but it is a
average profit per unit (π/q) total profit divided by number of units sold
Table 9.1 Short-Run Cost and Revenue of a Competitive Firm (in dollars) q P TR TC TVC ATC AVC π π/q MC MR
0 12 0 15 0 — — −15 — — — 1 12 12 25 10 25 10 −13 −13 10 12 2 12 24 33 18 16.50 9 −9 −4.50 8 12 3 12 36 40 25 13.30 8.30 −4 −1.30 7 12 4 12 48 46 31 11.50 7.80 2 0.50 6 12 MC < MR 5 12 60 54 39 10.80 7.80 6 1.20 8 12
6 12 72 63 48 10.50 6 9 1.50 9 12
7 12 84 73 58 10.40 8.30 11 1.57 10 12
8 12 96 84.90 69.90 10.61 8.70 11.10 1.39 11.90 12 MC ≈ MR 9 12 108 98 83 10.90 9.20 10 1.25 13.10 12
10 12 120 113 98 11.30 9.80 7 0.70 15 12 MC > MR 11 12 132 132 117 12 10.60 0 0 19 12
⎫
⎬
⎪ ⎪ ⎪
⎭
⎪ ⎪ ⎪
⎫ ⎬ ⎪
⎭⎪
Short-Run Profit Maximization: Total Curves Profit is maximized at the output where total revenue (TR) exceeds total cost (TC) by the largest possible amount. This occurs at output q1, where profit is equal to AB. The total profit curve (π), which plots total profit explicitly at each rate of output, also shows the point of profit maximization.
–TFC
TFC
TC TR
MR = P = $12
q0 q1 (8)
C
B
b
b
A
q3 0
Dollars
Output
1
�
Figure 9.2
220 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 220Trim Size: 203.2 mm X 254 mm
relatively simple one when we are dealing with a competitive firm. With the price per unit
constant, total revenue rises in proportion to output and is, therefore, drawn as a straight
line emanating from the origin. Its slope, showing how much total revenue rises when out-
put changes by one unit, is marginal revenue.
In terms of Figure 9.2, the firm wishes to select the output level where total revenue
exceeds total cost by the largest possible amount—that is, where profit is greatest. This
situation occurs at output q1, where total revenue, Aq1, exceeds total cost, Bq1, by AB. The vertical distance AB is total profit at q1. At lower and higher output levels, total profit is lower than AB. Note that at a lower output level, q0, for example, the TR and TC curves are diverging (becoming farther apart) as output rises, indicating that profit is greater at a higher output. This reflects the fact that marginal revenue (the slope of TR) is greater than marginal cost (the slope of TC) over this range. At q1, when the curves are farthest apart (with revenue above cost), the slopes of TR and TC (the slope of TC at B is equal to the slope of bb) are equal, reflecting an equality between marginal revenue and marginal cost.
In Figure 9.2, the level of total profit at each rate of output is also shown explicitly by
the total profit curve (π), which reaches a maximum at q1. The total profit curve is derived graphically by plotting the difference between TR and TC at each output level. For exam- ple, AB is equal to Cq1; alternatively, when output is zero, profit is negative and equal to minus total fixed cost (TFC).
Short-Run Profit Maximization Using per-Unit Curves Figure 9.3 presents the same information shown in Figure 9.2, but now we use the familiar
per-unit cost and revenue relationships. With the vertical axis measuring dollars per unit
of output, the firm’s demand curve is shown as a horizontal line, because the firm may sell
any number of units at the $12 price. The figure also shows the average total cost (ATC), average variable cost (AVC), and marginal cost (MC) curves. The most profitable output level occurs where marginal cost and marginal revenue are equal, at q1 in the figure (eight units in Table 9.1). The shaded rectangle BCDA shows total profit for that output. The height of the rectangle, CD, is average revenue ($12) minus average total cost ($10.61 ≈ $84.9/8), or the average profit per unit of output ($1.39 = $12 − $10.61). Multiplying the average profit per unit by the number of units sold (the length of the rectangle) yields total
profit ($11.10).
Understanding why the firm’s profit is at a maximum where marginal cost and marginal
revenue are equal is very important. Consider what it would mean if the firm were operating
at a lower output where MC < MR, such as q0 (seven units) in Figure 9.3. At q0, marginal revenue is still $12, but marginal cost is lower ($10). Consequently, the seventh unit of out-
put adds $12 to revenue but increases total cost by only $10; with revenue increasing more
than cost, profit (the difference) will rise from $9 (at six units) to $11 when the seventh unit
is sold. At any rate of output where marginal revenue is greater than marginal cost, the firm
can increase its profit by increasing output. As output expands, marginal cost rises, so the
addition to profit from each successive unit becomes smaller (but is still positive) until q1 is reached, where MC = MR.
At output levels beyond q1, the firm would be producing too much output. At q2 (nine units in Table 9.1), for example, marginal cost is $13.10 and is greater than marginal rev-
enue ($12). Thus, total profit could be increased by reducing output. If the firm produces
one unit less, it loses $12 in revenue. But cost falls by $13.10, so the net effect is a $1.10
increase in profit. Thus, the firm’s profit will increase by decreasing output if MC > MR. In summary, the rule for profit maximization for firms in general (whether in a competi-
tive or noncompetitive setting) is to produce where MC = MR. Because marginal revenue equals price for a competitive firm, we can also express this condition as MC = P for firms operating in the specific setting of a competitive industry. This rule does not mean that the
Short-Run Prof i t Maximizat ion 221
C09.INDD 12:6:43:PM 08/06/2014 PAGE 221Trim Size: 203.2 mm X 254 mm
firm intentionally sets price equal to marginal cost since, for a competitive firm, price is
given and beyond its control. Instead, the firm will adjust its production until the marginal
cost is brought into equality with price (= MR).2 In Chapter 11, we will see how the general rule of selecting output based on where
MR = MC applies also to firms operating in noncompetitive market settings. Maximiz- ing profit based on choosing output where P = MC is a special case of this general profit- maximizing rule. It applies in perfectly competitive settings where firms cannot affect the
prevailing industry price, and thus marginal revenue, through their output decisions.
Operating at a Loss in the Short Run A firm in a perfectly competitive market may find itself in the unenviable situation of suffer-
ing a loss no matter what output level it produces. In that event, the firm has two alternatives:
it can continue to operate at a loss or it can shut down. (Recall that we are dealing with a
short-run setting; halting production may be only a temporary move until market conditions
improve and is not necessarily the same as going out of business.) Yet shutting down will not
avoid a loss since the firm remains liable for its fixed cost whether or not it operates. The rel-
evant question is whether the firm will lose less by continuing to operate or by shutting down.
2Because the MC curve is U-shaped, MC may equal P at two different output levels. In this case the lower level is not the profit-maximizing output; in fact, it is the minimum-profit (or maximum-loss) output. If it is necessary to distinguish these two outputs, the profit-maximizing output is where MC = P and MC cuts P from below (that is, MC is rising).
q0 (7)
q1 (8)
q2 (9)
A D
B C AVC
ATC
P = MR = AR
MCDollars per unit
0
$10 $10.61
$12
$13.10
Output
Short-Run Profit Maximization: Per-Unit Curves In terms of the per-unit curves, the output that maximizes profit in the short run is where MC = MR, or q1. At lower levels of output such as q0, MC < MR, and the firm can increase profit by expanding output. At higher outputs like q2, MC > MR, and the firm can increase profit by reducing output.
Figure 9.3
222 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 222Trim Size: 203.2 mm X 254 mm
Figure 9.4 illustrates the case where the firm’s best option is to operate at a loss rather
than shut down. With the price at $8 per unit and the cost curves as shown, the firm’s “most
profitable” output (which can also mean its “least unprofitable” output, as it does here) is
the point where price equals marginal cost at q1. Because the average total cost curve lies above the price line everywhere, the firm incurs losses at all output levels. But at q1 the loss is the smallest, as indicated by the rectangle BCFE. The height of this rectangle, CF, is the difference between average cost and average revenue, or the “negative profit margin” per
unit, and the length of the rectangle, EF, is the number of units sold at a loss. How do we know that the firm loses less by producing at q1 than by shutting down? Con-
sider the larger rectangle BCDA. The height of the rectangle, CD, is the difference between ATC and AVC at q1. Thus, CD measures average fixed cost. Recall that AVC + AFC = ATC; therefore, ATC − AVC = AFC, and average fixed cost multiplied by the number of units pro- duced (the length of the rectangle) equals total fixed cost (or the area of rectangle BCDA). Even if the firm shuts down, it will still incur a loss, namely, total fixed cost. But because
that loss (BCDA) is larger than the loss (BCFE) incurred if the firm continues to operate, the firm loses less by producing q1 than by shutting down. It is better to operate and lose $100 per week than to shut down and lose $200 per week.
If the price falls sufficiently, however, the firm may lose less by shutting down, as we
will see in the following section. Moreover, even when it is in the firm’s interest to produce
at a loss, as in Figure 9.4, this equilibrium can be only temporary (short run). If the price
remains at $8, the firm will ultimately go out of business. The point here is that a firm will
not immediately liquidate its assets the moment it begins to suffer losses.
q1 Output
A D
F
B C
AVC
ATC
P = MR = AR
MC
Dollars per unit
0
$8 = E
Operating at a Loss in the Short Run In the short run, a firm may continue to produce even though its best output yields a loss. As long as average revenue covers average variable cost, it is in the firm’s interest to continue operating. Output will be q1 even though the firm’s loss is BCFE, because the loss would be even greater—BCDA—if it shut down.
Figure 9.4
The Perfect ly Competit ive F i rm’s Short-Run Supply Curve 223
C09.INDD 12:6:43:PM 08/06/2014 PAGE 223Trim Size: 203.2 mm X 254 mm
9.5 The Perfectly Competitive Firm’s Short-Run Supply Curve
The previous discussion implies a systematic relationship between a product’s price and a
firm’s most profitable output. We can analyze this relationship to derive the short-run firm supply curve in a perfectly competitive industry. Figure 9.5 depicts the firm’s average variable and marginal cost curves. Note that the average total cost curve is not shown
because it is not needed to identify the most profitable output level. The ATC curve is pri- marily useful for showing total profit or loss. If the per-unit price is $8, the firm will pro-
duce at point B, where marginal cost equals $8, so output will be q1. If the price rises from $8 to $12, then the firm’s profit will change, as will the most profitable output. Suppose that
the firm maintains an output of q1 when the price and associated marginal revenue rises to $12. Its profit will increase because revenue rises, but its cost will not change. The firm,
however, can increase profit by increasing output. At q1, marginal revenue ($12) now exceeds marginal cost ($8), signaling that an output increase will add more to revenue than
to cost. The new profit-maximizing output occurs where output has expanded until mar-
ginal cost equals the $12 marginal revenue or price, at point C on the MC curve. Thus, a higher price gives the firm an incentive to expand output from q1 to q2.
short-run firm supply curve a graph of the systematic relationship between a product’s price and a firm’s most profitable output level
q0 q1 q2
A
B
C
AVC
P2 = MR2 = AR2
P1 = MR1 = AR1
P0 = MR0 = AR0
MC
Output
Dollars per unit
0
$5
$8
$12
A Competitive Firm’s Short-Run Supply Curve The part of the MC curve lying above the minimum point on the AVC curve identifies the firm’s most profitable output at prices above $5. Below a price of $5, the firm will be better off shutting down. Above a price of $5, the competitive firm maximizes profit by producing where marginal revenue (= P) equals marginal cost. Because the MC curve slopes upward, the firm increases output at a higher marginal revenue (= P). The solid, dark red curve represents the competitive firm’s short-run supply curve.
Figure 9.5
224 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 224Trim Size: 203.2 mm X 254 mm
The MC = MR (= P) rule for maximizing profit in a competitive market structure there- fore implies that a firm will produce more at a higher price because increased production
becomes profitable at a higher price. In fact, we can think of the MC curve as the firm’s supply curve in the short run, since it identifies the most profitable output for each pos-
sible price. For example, at a price of $8 output is q1 at point B on MC, at a price of $12 output is q2 at point C on MC, and so on.
One important qualification to this proposition should be mentioned. If price is too
low, the firm will shut down. At a price of $5 the firm can just cover its variable cost by
producing at point A, where average variable cost equals the price. At this point the firm would be operating at a short-run loss; in fact, the loss would be exactly equal to
its total fixed cost, because revenue just covers variable cost, leaving nothing to set
against fixed cost. At a price below $5 the firm is unable to cover its variable cost and
would find it best to shut down. Thus, point A, the minimum level of average variable cost, is effectively the shutdown point: if price falls below that level, the firm will cease operations.
As a consequence of this qualification, only the segment of the marginal cost curve that
lies above the point of minimum average variable cost is relevant. Stated differently, the
marginal cost curve above point A identifies the firm’s output at prices above $5; at any lower price output will be zero.
Output Response to a Change in Input Prices Many factors can affect a competitive firm’s output decision, but perhaps the two most
common are variations in the product price and variations in the prices of inputs used in
production. In reality, product and input prices frequently change at the same time, but
at the outset, examining each factor in isolation is best. In the preceding paragraphs we
examined the way a product price change affects output when input prices are constant
(reflected in the fact that the cost curves did not shift). Now let’s study the impact of an
input price change while holding the product price constant.
Figure 9.6 illustrates a competitive firm initially producing the output q1 where mar- ginal cost, shown by MC, equals the $12 product price. Note that only the marginal cost curve above the shutdown point, point A, is drawn in. Now suppose that the price of one (or more) of the variable inputs falls. As we saw in Chapter 8, lower input prices cause
the cost curves to shift downward, indicating a lower cost associated with each rate of
output than before. This change is depicted in the diagram by a shift in the marginal cost
curve from MC to MC′. The rule for profit maximization in competitive markets, MC = MR (= P), determines the new output level, q2. To proceed to this conclusion somewhat more slowly, note that at the initial output level, a lower input price decreases the mar-
ginal cost from $12 to $6. After the reduction in cost, the price ($12) is higher than mar-
ginal cost ($6), indicating that an expansion in output will now (at the lower input price)
add more to revenue than to cost and thereby increase profit. Consequently, the firm will
expand output along the new MC′ curve until marginal cost once again equals marginal revenue at q2.
In this section we have reviewed a competitive firm’s reactions to changes in its eco-
nomic environment—changes communicated to the firm through variations in input
or product prices. A competitive firm by itself does not influence these prices; it is a
price taker. Nonetheless, the combined actions of all the competing firms affect product
and input prices, as we will see in the following sections. Each firm’s response to price
changes is an integral part of the adjustment process by which the overall market strives
to reach equilibrium.
shutdown point the minimum level of average variable cost below which a firm will cease operations
The Perfect ly Competit ive F i rm’s Short-Run Supply Curve 225
C09.INDD 12:6:43:PM 08/06/2014 PAGE 225Trim Size: 203.2 mm X 254 mm
MC ′
MC
A ′
A
P = MR = AR
q2q1 Output
$12
$6
0
Dollars per unit
How a Firm Responds to Input Price Changes An input price change, with an unchanged product price, alters the profit-maximizing output. If an input price falls, MC shifts to MC′, and output increases to q2, where MC′ equals the unchanged price.
APPLICATION 9.2
Feedlot operators buy one-year old cattle that weigh 750 pounds and then fatten up the animals to as much as 1,400 pounds over a 6-month period, through a corn- heavy diet, so as to prepare them for sale to slaughter- houses.3 In 2012, about 2,000 of the 77,120 feedlots in the United States shut down operations versus 20 in 2011. A similarly large number were expected to also shut down in 2013.
The reasons for the rising number of feedlot closings are several-fold. First, a multi-year drought in the early part of the 2010 decade in Kansas, Texas, and other major cattle-producing states, led to rising prices for young cattle.
Why Firms That Fatten Cattle are Seeing Their Own Numbers Thinned
Second, the drought also led to an increase in the price of corn, the main feed ingredient used by feedlots to fatten cattle. Third, Tyson Foods, a major buyer of fattened cat- tle, decided to quit purchasing animals fed with Zilmax, a popular weight-gain supplement. This decision required feedlot operators to buy more corn and take longer to fatten cattle and/or to sell animals weighing less. Finally, beef consumption in the United States was 15 percent per person lower in 2012 than a decade earlier—at least partly due to the shift in consumption toward poultry and away from red meat detailed in Application 4.2—thereby put- ting downward pressure on the prices slaughterhouses are willing to pay for fattened cattle. For the foregoing reasons, the firms which seek to fatten cattle have found their own numbers being thinned.
3This application is based on “Cheaper Feed Comes Too Late,” Wall Street Journal, August 14, 2013, p. A3.
Figure 9.6
226 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 226Trim Size: 203.2 mm X 254 mm
9.6 The Short-Run Industry Supply Curve Moving from the determination of the most profitable output of the competitive firm to the
short-run industry supply curve is a short step. In the short run, a competitive firm will produce
at a point where its marginal cost equals marginal revenue (= P), as long as the price is above the minimum point of the average variable cost curve. In other words, each firm’s marginal cost
curve indicates how much the firm will produce at alternative prices. As a first approximation, the short-run industry supply curve is derived by simply adding the quantities produced by each firm—that is, by summing the individual firms’ marginal cost curves horizontally. We assume for now that variable input prices remain constant at all levels of industry output.
Figure 9.7 illustrates how a short-run industry supply curve is derived for the three firms
(A, B, and C) composing the cement industry. Although three firms may not be enough to
constitute a competitive industry, the derivation is the same regardless of the number of firms
involved. The figure shows the portion of each firm’s marginal cost (MC) curve lying above its minimum average variable cost (AVC). Because the supply curve identifies the total quantity offered for sale at each price, we will add the outputs each firm chooses to produce individually.
At any price below P0, the minimum point on firm C’s AVC curve, industry output is zero because all three firms shut down. At P0, firm C begins to produce. Note that its output is the only output on the market until price reaches P1, because the other companies will not operate at such low prices. Thus, the lower portion of firm C’s MC curve reflects the total industry supply curve at low prices. When price reaches P1, however, firm A begins to pro- duce, and its output must be added to C’s—hence, the kink in the industry supply curve at
P1. When price reaches P2, firm B begins production, and its output is included in the supply curve at prices above P2. We refer to this derivation of the short-run industry supply curve as
short-run industry supply curve the horizontal sum of individual firms’ marginal cost curves
P4
P1
P0
P2
P3
qA0 qB qC Q Q ′
D ′ D
SS
MCA MCB MCC
Quantity of cement
Dollars per unit
The Short-Run Competitive Industry Supply Curve Each firm will produce an output where marginal cost equals marginal revenue (= P) to maximize profit. To derive the short-run industry supply curve, SS, we must horizontally sum the amounts produced by the industry’s various firms, as shown by their respective marginal cost curves.
Figure 9.7
Long-Run Competit ive Equi l ibr ium 227
C09.INDD 12:6:43:PM 08/06/2014 PAGE 227Trim Size: 203.2 mm X 254 mm
a horizontal summation of the individual firms’ short-run marginal cost curves because we
are horizontally summing quantities across firms. For example, at a price of P3, total industry supply is Q, the sum of the amounts each company produces at that price (qA + qB + qC).
Note that the short-run industry supply curve, SS, slopes upward. Remember that each firm’s MC curve slopes upward because it reflects the law of diminishing marginal returns to variable inputs. Thus, the law of diminishing marginal returns is the basic determinant of
the shape of the industry’s short-run supply curve.
Price and Output Determination in the Short Run Incorporating the market demand for the product completes the short-run competitive
model. Because the market demand relates total purchases (from all firms) to the price of
cement, it interacts with the supply curve to determine price and quantity. In Figure 9.7,
the intersection of the demand curve D with the supply curve SS identifies the price where total quantity demanded equals total quantity supplied. Thus, price P3 is the short-run equi- librium price, and total industry output is Q. Each firm, taking price as given, produces an output where marginal cost equals P3. Firm A thus produces qA; firm B, qB; firm C, qC; and the combined output, qA + qB + qC, equals Q.
Given the Figure 9.7 supply and demand relationships, if price were at a level other than
P3 for some reason, familiar market pressures would work to push price toward its equilib- rium level. For example, at a price lower than P3, total quantity demanded by consumers would exceed the total amount supplied, a temporary shortage would exist, and price would
be bid up. As price rises toward P3, quantity demanded becomes smaller while quantity sup- plied becomes larger, until the two eventually come into balance at price P3. Conversely, if price were higher than P3, quantity supplied would exceed quantity demanded, a temporary surplus would exist, and competition among firms would drive the price down.
In the short run an increase in market demand leads to a higher price and a higher out-
put. Suppose that we begin with the equilibrium just described, and then demand increases
to D′. A shortage will exist at the initial price of P3, and price will rise. The higher price will elicit a greater output response from the individual firms as they move up along their
MC curves. The new short-run equilibrium will be price P4 and quantity Q′. Note that Figure 9.7 does not explicitly identify the profits, if any, realized by the firms.
If we drew in each firm’s average total cost curve, we could show them, but we do not need
to use the average cost curves to explain the determination of price and quantity in the short
run. All the necessary information is contained in the industry supply and demand curves.
9.7 Long-Run Competitive Equilibrium At any time, the short-run scale of plant a firm has built reflects a previous long-run decision.
Therefore, we must consider how the goal of profit maximization guides firms’ long-run
decisions and what that implies, in turn, about long-run competitive equilibrium. The same
principles we used for the short-run setting apply to long-run profit maximization, but now
we employ long-run cost curves, which allow a sufficient period of time to vary all inputs.
Figure 9.8 shows the long-run cost curves LAC and LMC for a representative firm in a perfectly competitive industry. If the firm believes the marginal revenue (= P) will remain at $12, it will want to select an output level of q3, where long-run marginal cost equals price. Producing q3 involves building a scale of plant that has a related short-run average cost curve (not depicted) tangent to LAC at point A. After building the plant, the firm real- izes a total profit shown by area EFAC. No other size plant yields as much profit, since the long-run cost curves reflect all possible scales of plants the firm can select.
Figure 9.8 also illustrates why a short-run profit-maximizing outcome may be only a
temporary equilibrium. Suppose that the price is $12 but the representative firm currently
228 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 228Trim Size: 203.2 mm X 254 mm
has a scale of plant with the short-run cost curves SAC1 and SMC1. With that size plant the firm’s short-run profit-maximizing equilibrium is q1, with the firm earning profit equal to the striped rectangular area EGHI. If the firm expects the price of $12 to persist, however, it would immediately begin making plans to enlarge its scale of plant, because it could earn
significantly higher profit, area EFAC, by building the appropriate long-run scale of plant identified by point A. By its very nature, a short-run equilibrium is only temporary unless the firm’s scale of plant is consistent with the price that is expected to persist in the market.
Suppose now that the expected marginal revenue (= P) over the long run is $7 instead of $12. The horizontal demand curve at this price is just tangent to LAC at point B. For this demand curve the most profitable long-run output is q2, where price equals long-run marginal cost. Total profit, however, is zero since price just equals the average cost of production.
Zero Economic Profit The firm’s long-run equilibrium at q2, with price just covering average cost, has great sig- nificance. For one thing, the q2 equilibrium indicates zero economic profit. Since the long- run cost curves take into account the opportunity costs of inputs, this implies that the return
from employing inputs in the firm is as high as from the best alternative use of those inputs.
In the case of capital, for instance, a firm’s owners must be earning as great a return as the
one they could make if their capital was employed elsewhere. Although the firm’s books
may show a positive accounting profit, the q2 equilibrium involves a “normal” economic return being earned by inputs such as capital and thus zero economic profit.
Should a firm stay in business if it is earning zero economic profit at an output such as q2 in Figure 9.8? The answer to this question is an emphatic yes. Zero economic profit means that the various inputs, including capital provided by the owners, can earn just as much
zero economic profit the point at which total profit is zero since price equals the average cost of production
q1 q2 q3 Output
A
B
LMC
LAC
P = MR = AR
P ′ = MR ′ = AR ′
SMC1 SAC1
Dollars per unit
0
$7
$12 E
I
C
G F
H
Long-Run Profit Maximization The competitive firm maximizes its profit by producing where LMC= P. With a price of $12, output will be q3. By building a larger-scale plant and adjusting its output from q1 to q3, the firm increases its profit from EGHI to EFAC.
Figure 9.8
Long-Run Competit ive Equi l ibr ium 229
C09.INDD 12:6:43:PM 08/06/2014 PAGE 229Trim Size: 203.2 mm X 254 mm
somewhere else. Although this is true, the relevant consideration is that they can’t earn any
more elsewhere, which is why the firm has an incentive to continue production. The q2 equilibrium is also significant because it is precisely the type of equilibrium com-
petitive markets tend toward in the long run. To see why, remember that free entry and exit
of resources is one of the characteristics of a perfectly competitive market. Figure 9.8 indi-
cates how a price of $12 creates an incentive for the firm to expand output from q2 to q3 and provides the firm with a positive economic profit. Positive economic profit implies that
resources invested in the industry generate a return higher than what could be earned else-
where. Without barriers to entry, the abnormally high return results in new firms entering the
industry; investors can make more money by shifting their resources into an industry afford-
ing positive economic profits. As shown in Figure 9.9, however, new entry also results in
the industry short-run supply curve shifting to the right and a decrease in price. This process
continues until the market demand curve and the new industry supply curve (SS′) intersect at the same price (P′) where long-run marginal cost equals the minimum point on the rep- resentative firm’s long-run average total cost curve. Entry continues, in other words, until
any positive economic profit signal and hence incentive for entry are eliminated. There is no
incentive to enter the industry (that is, economic profit equals zero) when long-run average
cost equals average revenue where long-run marginal cost equals marginal revenue (= P). The long-run equilibrium shown in Figure 9.9 has three characteristics. First, the rep-
resentative competitive firm is maximizing profit and producing where LMC equals mar- ginal revenue (equals price). This condition must hold for a simple reason. If firms are not
producing the appropriate amounts, they have an incentive to alter their output levels to
increase profit. A change in firms’ outputs, however, affects the total quantity offered for
sale by the industry, which in turn changes the price at which the output is sold. Thus, the
initial price cannot be an equilibrium price if firms are not producing where LMC equals marginal revenue (equals price).
LMC
LAC
D
SS ′P = MR = AR
P ′ = MR ′ = AR ′
SS
q3q2 Output Output
$12
$7
0
(a) (b)
Dollars per unit
Q2Q10
Dollars per unit
Long-Run Competitive Equilibrium (a) The equilibrium from the perspective of a representative firm. The firm is producing the most profitable output (LMC = MR′ = P′) and is making zero economic profit (LAC = AR′). (b) An industry in long-run competitive equilibrium where SS′ and D intersect.
Figure 9.9
230 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 230Trim Size: 203.2 mm X 254 mm
Second, there must be no incentive for firms to enter or leave the industry. This condi-
tion occurs when firms are making zero economic profits. If profits were higher, other firms
seeking higher returns would enter the industry; if profits were lower, firms would leave the
industry because they could do better elsewhere. This entry or exit would affect the level of
industry output and change the price. If, however, profits in this industry are comparable to
profits in other industries, there is no reason for entry or exit to occur, and price and output
will remain stable.
Third, the combined quantity of output of all the firms at the prevailing price equals
the total quantity consumers wish to purchase at that price. If this condition were violated,
there would be either excess demand or excess supply at the prevailing price, so the price
would not be an equilibrium one.
Zero Profit When Firms’ Cost Curves Differ? When all firms in a competitive industry have identical cost curves, every firm must be mak-
ing zero economic profit in long-run equilibrium. Cost curves, however, may differ among
firms. Some businesses may have access to superior technology, patents or more skilled
workers vis-à-vis other firms. These differences will be reflected as differences in production
costs. While firms’ LAC curves may be shaped and positioned differently, their minimum points will still end up at the same cost per unit through the process described below.
If firms’ cost curves differ, we must reconsider the proposition that every firm makes zero economic profit in a long-run equilibrium. As it turns out, this proposition is still cor-
rect, but the reason it is deserves attention. To see why, consider the case of two firms that
are both in the business of raising corporate executives’ IQs, Densa and Mensa. Let’s sup-
pose that Mensa has some especially productive input that Densa does not have. For exam-
ple, suppose that Mensa and Densa both purchase office space to begin their operations.
Through foresight or chance, Mensa’s selected office site proves to be more favorable. Sup-
pose that government policymakers unexpectedly encourage businesses to set up shop in
Mensa’s vicinity while locating a nuclear waste dump adjacent to Densa, thereby scaring
off potential corporate neighbors. Because of the government actions, Mensa’s location is
more productive, even though both firms originally may have paid the same price for the
office space. The better location accounts for Mensa’s lower production cost.
Due to the unexpectedly productive location, Mensa’s income statement will indicate
a positive accounting profit. The original price Mensa paid for the space understates its
real economic value; that is, the cost of using the space is greater than its purchase price.
Remember, however, that cost reflects forgone opportunities, and once it is known that
Mensa’s location is highly productive, its market value—what other firms would pay for the
space—will rise. Provided that input markets are perfectly competitive, the opportunity cost
to Mensa of its office space will be bid up to reflect its full economic value. Once Mensa’s
office space is valued at its true opportunity cost, Mensa’s economic profit falls to zero.
In this analysis it makes no difference whether Mensa rents or owns the office space. If
Mensa rents the space, and it becomes apparent that the space is unusually productive, the
rental cost will go up because other firms will be willing to pay more for highly productive
office space. If Mensa owns the office space, the same principle applies; the original pur-
chase price is irrelevant. What counts is how much the office space is worth to other potential
users—that is, its opportunity cost. In this case the cost of using the office space is an implicit
cost, and, as explained in an earlier chapter, we include implicit costs in the cost curves.
The process by which unusually productive inputs receive higher compensation is not
restricted to office space. Suppose that you go to work for a business as a manager and are
promised a salary of $40,000. When the firm hires you, it doesn’t know whether you’ll be
any good at your job. Fortunately, you turn out to be brilliant, and even after paying your
salary, the firm’s net revenue rises by $25,000 due to your extraordinary managerial skills.
The firm’s owner is delighted, and the firm’s books show a $25,000 profit increase. Once
The Long-Run Industry Supply Curve 231
C09.INDD 12:6:43:PM 08/06/2014 PAGE 231Trim Size: 203.2 mm X 254 mm
your managerial skills are recognized, however, you are in a position to command and
receive a $25,000 raise, which will effectively eliminate the firm’s accounting profit. You
will be in such a position so long as your managerial skills are transferable across firms and
there is a competitive market for your managerial talents. Provided these conditions hold,
if you don’t get a raise, you can resign and accept a position with another firm that will pay
you a salary closer to what you’re worth—your best alternative—and your former employ-
er’s profit will decline to its original level.
Needless to say, this process doesn’t work instantaneously or with exact precision. There is
a tendency, however, for inputs to receive compensation equal to their opportunity costs, that is,
what they are worth to alternative users. This process leads to the zero-profit equilibrium. One
implication of this analysis is that the accounting measure of profit may vary widely among
firms in an industry even though economic profit is zero for each firm. This variation arises
because a firm’s assets are frequently valued on the books at their original purchase prices
instead of their opportunity costs. In our earlier example, Mensa’s accounting profit would be
higher than Densa’s if they both counted only the purchase price of the office space as a cost.
9.8 The Long-Run Industry Supply Curve To derive the competitive industry’s short-run supply curve we horizontally sum individual
firms’ supply curves. We cannot similarly derive a competitive industry’s long-run supply
curve, however, because in the long run firms enter or exit the industry in response to the
economic profits being earned. So we do not know which firms’ supply curves to sum hori-
zontally. Moreover, as an industry expands or contracts due to firm entry or exit, the prices
that firms pay for their inputs may change. For example, as the movie business expands due
to firm entry, the prices of actors and directors may be bid up. As the oil-producing industry
shrinks due to firm exit, the prices of petroleum engineers and drilling rigs may decline.
As we will see, the derivation of a long-run industry supply curve in a competitive market depends centrally on what happens to input prices as the industry expands or con-
tracts. Based on the three possible effects of industry size on input prices, competitive
industries are classified as constant-cost, increasing-cost or decreasing-cost. We address each of these cases in turn, starting with the simplest of the three: when input prices remain
constant regardless of an industry’s overall size.
Constant-Cost Industry To derive the long-run supply curve for a constant-cost industry, we start from a position of
equilibrium and trace the effects of a demand change until the industry once again returns
to a long-run equilibrium. As our example, let’s assume that the market for MBA education
is perfectly competitive, with business schools being the representative firms. We begin by
examining the industry when it is in long-run equilibrium (say in 1970).
Figure 9.10b shows the industry’s initial long-run equilibrium: when demand is D, out- put is Q1, and price is P. Assuming that we start in long-run equilibrium, point A is then one point on the long-run supply curve, LS. Figure 9.10a shows the representative firm’s posi- tion—say that of New York University (NYU). It is producing its most profitable output,
q1, and making zero economic profit at price P. To identify other points on the long-run supply curve, let’s imagine that there is an
unexpected increase in demand to D′ and work through the consequences for the industry. Demand for MBA education has grown over the past five decades for a number of reasons,
including more women joining the work force and seeking a business education, worldwide
economic growth and entrepreneurship, and the ever-advancing state of business manage-
ment knowledge.
long-run industry supply curve the long-run relationship between price and industry output, which depends on whether input prices are constant, increasing or decreasing as the industry expands or contracts
constant-cost industry an industry in which expansion of output does not bid up input prices, the long-run average production cost per unit remains unchanged, and the long-run industry supply curve is horizontal
increasing-cost industry an industry in which expansion of output leads to higher long-run average production costs and the long-run supply curve slopes upward
decreasing-cost industry a highly unusual situation in which the long-run supply curve is downward sloping
232 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 232Trim Size: 203.2 mm X 254 mm
Reflecting the increased demand, existing business schools have an immediate short-
run response—they increase output by expanding operations in their existing plants and
increasing employment of variable inputs such as faculty and staff. The response appears as
a movement along the short-run industry supply curve, SS, from point A to point B. (Note that the initial long-run equilibrium is also a short-run equilibrium. Every school is operat-
ing its existing plant at the appropriate level—where SMC equals MR equals P—so point A is also a point on the short-run supply curve.) In the short run, price rises to P′, and output increases to Q2 as the industry expands along SS.
In Figure 9.10a, we can see what this adjustment means for NYU, a representative firm
in the industry. The higher price induces NYU to increase output along its short-run mar-
ginal cost curve SMC and produce q2 where SMC = MR′ (= P′). (Recall that the summation of all the schools’ SMC curves yields the SS curve.) At this point every business school is making an economic profit.
The industry, however, has not fully adjusted to the increase in demand; this position is
only a short-run equilibrium. The key to the long-run adjustment is profit-seeking by firms. In the short-run equilibrium, existing schools are making economic profits. The return on
investment in the MBA education market is now greater than in other industries. Whenever
economic profits exist in an industry, investors will realize that they can make more money
by moving resources into the industry. If entry is costless, as it is assumed to be in the case
of perfect competition, resources will be shifted into the MBA education market, thereby
increasing its productive capacity.
Suppose that new schools enter the industry in an attempt to share in the market’s
profitability. The entry increases the industry’s total output, which, in Figure 9.10b, is
shown as a rightward shift in the short-run supply curve. Recall that a short-run supply
curve represents a fixed number of schools with given plants. When the number of firms
SMC
LAC
SS
D
A C LS
B
SS ′
D ′
MR = AR
P ′
P
P ′
P
MR ′ = AR ′
q1 q2 Output Output0
(a) (b)
Dollars per unit
Dollars per unit
Q2 Q3Q10
NYU business school MBA market
E
Long-Run Supply in a Constant-Cost Industry The long-run industry supply curve LS of a constant-cost competitive industry is horizontal. Expansion of industry output does not bid up input prices. Thus, when industry output expands from Q1 to Q3, firms’ cost curves do not shift and the long-run average production cost per unit remains unchanged.
Figure 9.10
The Long-Run Industry Supply Curve 233
C09.INDD 12:6:43:PM 08/06/2014 PAGE 233Trim Size: 203.2 mm X 254 mm
increases, the total output at each price is greater (there are more “member” firms with
their corresponding MC curves in the industry “club”)—implying a shift in the short-run supply curve. As output expands in response to entry, price falls from its short-run level
along D′ since the higher output can be sold only at a lower price. This process of firms entering, total output increasing, and price falling continues so long as the industry is
generating a positive economic profit signal. In other words, a new long-run equilibrium
emerges when schools are once again making only a “normal profit”—that is, zero eco-
nomic profit.
For a constant-cost industry the increased employment of inputs associated with expand-
ing output occurs without an increase in the price of individual inputs. This means business
school cost curves do not shift. Thus, price must fall back to its original level before profits
return to a normal level: if price is higher than P economic profits will persist, and entry of new schools will continue. Figure 9.10b shows the process of entry as a rightward shift in
SS; it continues until SS′ intersects the demand curve D′ at point C and price has returned to its original level P.
Point C is a second point on the long-run supply curve LS. With demand curve D′, industry output expands until it reaches Q3 and price is P. Each school once again makes zero economic profit. In Figure 9.10a, NYU is again confronted with price P and can do no better than cover its average cost by producing q1. The increase in industry output from Q1 to Q3 is the result of entry by new schools.
A constant-cost competitive industry is characterized by a horizontal long-run supply curve. Given time to adjust, the industry can expand with no increase in price along LS. The crucial assumption producing this result is that input prices are not affected by an industry’s
size, so new schools can enter the market and produce at the same average cost as existing
schools. The term constant-cost refers to the fact that schools’ cost curves do not shift as the industry expands or contracts; it does not mean that each school in the industry has a horizontal LAC curve.
Increasing-Cost Industry We can derive the long-run supply curve for an increasing-cost industry in the same way
we derived it for a constant-cost industry. We assume an initial long-run equilibrium; then
the demand curve shifts, and we follow the adjustment process through to its conclusion.
Figure 9.11 illustrates this case. The initial industry long-run equilibrium price and quantity
are P and Q1, respectively. The typical firm, NYU, is producing q1 and just covering its costs, shown by LAC, at the market price.
Once again, we assume an unexpected demand increase to D′. The short-run equilibrium is determined by the intersection of the short-run supply curve SS and D′. Price rises to P′, and output increases to Q2. The representative school expands output along its SMC curve to q2 and realizes economic profit. In the long run, the profit attracts resources to the industry.
Up to this point the analysis is identical to the constant-cost case. Moreover, the attain-
ment of a new long-run equilibrium involves further expansion of industry output until
economic profits return to zero, just as in a constant-cost industry. However, in an increasing-
cost industry the expansion of output leads to an increase in some input prices. To produce
more output, schools must increase their demand for inputs, and some inputs such as pro-
fessors are assumed to be available in larger quantities only at higher prices. This situation
contrasts sharply with that of the constant-cost case, where schools can hire larger quanti-
ties of all inputs without affecting their prices.
Let’s see how this difference affects the long-run adjustment process. At the short-run
equilibrium, positive economic profits lead to entry by new schools and an expansion in
industry output—a rightward shift in the short-run supply curve. Increased industry output
now tends to reduce profits in two ways. First, as we saw earlier, the higher output causes
234 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 234Trim Size: 203.2 mm X 254 mm
price to fall, which reduces profits. Second, the increased demand for inputs that accompa-
nies the expansion in industry output leads to higher input prices. Higher input prices mean
higher production costs, which also reduce profits. Profits are thus caught in a two-way
squeeze as the industry expands. The two-way squeeze continues until economic profits
equal zero and a new long-run equilibrium is attained.
Figure 9.11 shows the two-way squeeze on economic profits. Starting with the short-
run equilibrium at point B (panel b), as new schools enter the market, SS shifts to the right and price falls. Each school’s horizontal demand curve shifts downward as price
declines from P′ to P″, and this decline reduces profits. At the same time, higher input prices shift the firms’ cost curves upward, from LAC to LAC′ and LMC to LMC′. In Figure 9.11a, NYU’s profit is squeezed from above by the decline in the product price
and from below by the rising unit cost of production. A rising cost and falling price even-
tually eliminate NYU’s economic profit. This occurs at price P″, when NYU can just cover its average cost by producing at the minimum point on LAC’. Once profit is elimi- nated, there is no longer an incentive for industry output to increase further, and a new
long-run equilibrium is reached. In Figure 9.11b, SS has shifted rightward to SS′, produc- ing a price of P″ and an output of Q3. No further shift in the short-run supply curve will occur, because economic profits have fallen to zero. Thus, point C is another point on the long-run industry supply curve.
An increasing-cost competitive industry is characterized by an upward-sloping long-run supply curve. This industry can produce an increased output only if it receives a higher price, because the cost of production rises (cost curves shift upward) as the industry
expands. The term increasing cost indicates that the cost curves of all schools shift upward as the industry expands and input prices are bid up.
LAC
LAC ′
SS
D
A
C LS
B
SS ′
D ′
MR = AR
P ′
P″
P
P ′
P″
P
MR ′ = AR ′
MR″ = AR″
q1 q2 Output Output0
(a) (b)
Dollars per unit
NYU business school MBA marketDollars per unit
Q2 Q3Q10
SMCSMC ′
E
Long-Run Supply in an Increasing-Cost Industry For an increasing-cost competitive industry, the long-run supply curve LS slopes upward. Expansion of industry output bids up some input prices. Thus, when industry output expands from Q1 to Q3, firms’ cost curves shift upward from LAC to LAC′ and SMC to SMC′, and the output expansion leads to a higher long-run average production cost.
Figure 9.11
The Long-Run Industry Supply Curve 235
C09.INDD 12:6:43:PM 08/06/2014 PAGE 235Trim Size: 203.2 mm X 254 mm
Decreasing-Cost Industry A decreasing-cost competitive industry is one that has a downward-sloping long-run supply curve. You might think that such a situation is impossible, and, in fact, some economists
believe that it is. Others claim that a decreasing-cost industry is theoretically conceivable
but admit that it is a remote possibility. Because all agree that it is, at best, highly unusual,
we will deal with it only briefly here.
A decreasing-cost industry adjusts to an increase in demand by expanding output, just
as industries in the other two cases. In this instance, though, the expansion of output by
the industry in some way lowers the cost curves of the individual schools, leading to a
new long-run equilibrium with a higher output but a lower price. The tricky part is to try
to explain why the cost curves shift downward. A downward shift in cost curves usually
reflects a decrease in input prices, but for that to happen we have to explain how an increase
in demand for some input leads to an increased quantity supplied at a lower price—not an
easy task.
Although economists are in agreement about how rare decreasing-cost industries must
be, many noneconomists find the concept appealing. The reason for its appeal stems from
the observation that prices have declined sharply as output has expanded in some indus-
tries. For example, color televisions, VCR/DVD players, and pocket calculators have all
fallen in price in real terms by 80 to 90 percent since they were first introduced. There have
also been dramatic reductions over time in personal computer prices.
Before concluding that such evidence reflects decreasing-cost industries, we should
explore some other possibilities. One common feature in these examples is that the
price was high when the product was first marketed but later fell dramatically. This
suggests two possible explanations. First, the firm initially marketing the product is a
monopoly and thus has some pricing power. The price that the monopoly firm sets may
be fairly high. As other firms enter the market and begin to compete, price falls and out-
put increases. This process suggests that what we may be seeing is the rise of competi-
tion from an initial monopoly position and not a movement along an industry’s long-run
supply curve.
Second, the passage of time after a product’s introduction makes technological improve-
ments in production possible. Particularly in the case of new and complex products, tech-
nological know-how is frequently rudimentary when firms first market a product. With
experience in production over time, technological improvements may occur quickly. We
emphasize that technological know-how is assumed to be unchanged along a supply curve.
An improvement in technology shifts the entire supply curve to the right.
Consider Figure 9.12. In conjunction with demand, the long-run supply curve for
pocket calculators in 1970, LS70, determined a price of $300. This price was, in fact, the approximate price of a calculator that performed only the basic functions—when
expressed in constant 1990 dollars. After 10 years, firms developed methods that lowered
production cost, and the 1980 long-run supply curve was LS80. We assume that demand is unchanged, although it was probably increasing over the period. This shift in supply led
to a price of $100 in 1980. Further technological improvements shifted the supply curve
once again, and price fell to $10 in 1990. The combination of lower prices and higher
outputs over time resulted here from shifts in an upward-sloping long-run supply curve,
not from a slide down the negatively-sloped supply curve of a decreasing-cost industry.
This explanation of the phenomenon is especially appealing because new high-technology
items are known to show rapid improvements in technical knowledge in the first years of
their production.
These remarks do not rule out the possibility of decreasing-cost industries, but if they
exist, like the Giffen good case in demand theory, they are extremely rare. For all practical
purposes the increasing- and constant-cost cases are the relevant possibilities.
236 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 236Trim Size: 203.2 mm X 254 mm
APPLICATION 9.3
Each year, the country’s 1000-plus business schools produce a total of 155,000 MBAs. The number of accred- ited business schools has nearly tripled and the number of MBA graduates per year has increased over 20-fold since 1970. In addition, the business education industry has also expanded globally, with an estimated 14,000-plus business schools operating around the globe.4
By contrast to the demand side for faculty, only 2,000 new business Ph.D.s are produced each year, half of them by American schools. Many American business Ph.D.s end up being hired by business schools outside of the United States, decreasing the supply to and driving up the price paid for faculty by American schools. For example, Virginia Tech’s business school pays $140,000–180,000 starting salaries to business Ph.D.s for a nine-month-a-year job—a rate two to three times greater than what is paid to begin- ning faculty in most other fields at the university.
Of course, as the business school market has expanded, not all inputs have risen in price. Some have remained
The Bidding War for Business School Professors
constant or perhaps even decreased. For example, the cost of receiving/processing applications has fallen due to reliance on certain shared application platforms such as ApplyYourself.
On average, however, real input prices have increased since the 1960s. This suggests that the long-run sup- ply curve in the MBA education market is upward sloping and provides at least a partial explanation for the rise in annual tuition at private business schools.5 For exam- ple, annual MBA tuition at USC, a representative private business school, rose from $1,950 in 1970 to $55,000 in 2013. The annual USC MBA tuition in 2013 equals roughly $9,200 in constant 1970 dollars. Although other factors (such as increased reliance by universities on their business schools to generate an operating surplus) have certainly affected MBA tuition, the growth of the MBA industry and the attendant increases in demand for industry inputs and the (real) prices of those inputs surely also played a role.
4This application is based on “We’re Hiring: Many Business Schools Have Trouble Filling Faculty Positions,” Wall Street Journal, November 17, 2011.
5Tuition at public business schools is generally controlled and sub- sidized by the state.
Technological Advances Shift the Long-Run Supply Curve A price reduction and output increase over time need not imply that the long-run supply curve has a negative slope. Technological advances can occur over time, which shift the entire long-run supply curve downward, producing lower prices and higher outputs.
D Quantity of pocket calculators
0
$10
$100
$300
Dollars per unit
Q2 Q3
LS90
LS80
LS70
Q1
Figure 9.12
The Long-Run Industry Supply Curve 237
C09.INDD 12:6:43:PM 08/06/2014 PAGE 237Trim Size: 203.2 mm X 254 mm
Comments on the Long-Run Supply Curve An industry’s long-run supply curve summarizes the results of a complex and subtle pro-
cess of adjustment. Once the underlying determinants of the supply curve are understood, it
becomes a powerful tool of analysis. To use the supply curve correctly, however, we must
have a firm understanding of its underpinnings. To that end, we discuss several frequently
misunderstood points.
1. We do not derive the long-run supply curve by summing the long-run marginal cost curves of an industry’s firms. Admittedly, every firm is producing where LMC = MR (= P) at each point on LS, but as the industry adjusts along LS, firms are entering or leaving the market. Therefore, we cannot sum firms’ LMC curves as we did in the short run. In addi- tion, for an increasing-cost industry the LMC curves themselves shift because of changes in input prices.
2. Just as we did with a demand curve, we assume certain things remain unchanged at all points on the long-run supply curve. For one we assume technology is constant. An industry
expanding along its LS curve is using the same technical know-how but employing more inputs to increase output. A change in technology causes the entire supply curve to shift, as
we saw in the example about pocket calculators discussed earlier. We also assume that at
all points on the long-run supply curve the supply curves of inputs to the industry remain
unchanged. Note that we are not assuming that the prices of inputs are unchanged but rather
that the conditions of supply remain constant. In a constant-cost industry, input prices remain
constant not because we assume them to be fixed but because the input supply curves fac-
ing the industry are horizontal. In contrast, in an increasing-cost industry input prices change
because the input supply curves facing the industry are upward sloping, a condition that gives
rise to an upward-sloping, long-run industry supply curve. When relevant, other factors like
government regulations or the weather must also be assumed constant along the supply curve.
3. In reality, an industry is not likely to fully attain a position of long-run equilibrium. Real-world industries are continually buffeted by changes. For instance, input supplies,
demand, technology, and government regulations frequently change. A long-run adjust-
ment takes time, and if underlying conditions change often, an industry will find itself
moving toward a long-run equilibrium that is continually shifting. Recognizing the reality
of frequent change, however, does not undermine the usefulness of the long-run equilib-
rium concept. Although the industry may never attain a long-run equilibrium, the tendency
for the industry to move in the indicated direction is what is important, and the outcome is
correctly predicted by the theory.
4. Economic profit is zero all along a competitive industry’s long-run supply curve. This point is often misunderstood. For example, someone may say that when industry output
expands due to an increase in demand, the industry’s firms benefit. We have seen, how-
ever, that after an increase in demand, all firms will make zero economic profits in the long
run. There may be economic profits in the short run, but the benefit is not permanent.
Who does benefit as we move along the long-run supply curve? Owners of inputs
whose prices are bid up by the industry-wide expansion. With a constant-cost industry no
input prices rise, and no input owners receive a permanent gain. If firms own some of the
inputs and thus gain as input prices increase, the gain accrues to them on account of their
input ownership and not because they are firms in an increasing-cost industry. It is also
possible that firms own none of the inputs whose prices rise. It may be, for example, that
professors whose wages go up as the MBA industry expands are the sole beneficiaries on
this market’s supply side.
The tendency toward zero economic profit means that the rate of return on invested
resources will tend to equalize across industries. If invested resources yield an annual
238 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 238Trim Size: 203.2 mm X 254 mm
return of 10 percent in the restaurant industry, which is comparable to earnings elsewhere,
then the restaurant industry is earning zero economic profit. Accountants, of course,
generally call the 10 percent return a “profit,” but economists regard the 10 percent return
as a cost necessary to attract resources to the restaurant industry.
5. In deriving the long-run supply curve, we assumed that a short-run equilibrium was first established and then long-run forces came into play. Price first went up to a short-run
high and then came down to a long-run equilibrium level. The actual process of adjustment
to demand changes may not follow this pattern exactly. Identifying a short-run equilibrium
and then tracing out long-run effects is merely an expedient way of explaining the determi-
nation of the final equilibrium. In fact, following a demand increase, price may never reach
its short-run level and may never go above the ultimate long-run level. This can happen if,
for example, firms anticipate the demand increase and make adjustments before demand
actually rises.
These remarks also suggest that care must be taken in using the short-run supply curve.
The short-run supply curve can be used to identify the initial effects of a change in demand
under only two conditions: when the demand change is unexpected and when it is expected
to be temporary. If firms anticipate the demand change, they can adjust their scales of plants
or move into or out of a market before the demand change actually occurs. An unexpected
demand change catches firms unaware, and they must operate temporarily with whatever
scales of plants they have at that time. If firms expect a demand change to be temporary,
they will not expand capacity or enter a market on the basis of conditions they know will
APPLICATION 9.4
Demand for corn has grown dramatically over the last several years, partly fueled by ongoing worldwide popula- tion growth (and corn use as a food source) and an increase in the price of crude oil (which has spurred greater reliance on ethanol that can be used as a substitute for gasoline). In addition, several policymaking initiatives have contributed to heightening demand for corn: a Congressional mandate that 9 billion gallons of ethanol be blended into gasoline in the United States as of 2008; a 51-cent federal govern- ment subsidy to U.S. oil refiners for every gallon of ethanol they blend with gasoline; and the signing into law in 2007 by President George W. Bush of an energy bill to increase U.S. production of renewable motor fuels, such as ethanol, five-fold to 36 billion gallons by 2022.
The demand for corn to meet growing consumption by the food and biofuels industries pushed the price of the grain from about $2 to $3 a bushel, where it was for several years, to an 11-year high of $5 per bushel by early 2008 and $6.50 per bushel by early 2011. The price increase through 2008 came despite a record crop of over 13 billion bushels of corn being harvested in 2007, 24 percent of which went into ethanol production (up from 14 percent in 2005).
Economic models predict that the increase in demand for corn (if it is sustained over the long run) will result in zero
Cashing In on Corn
economic profits being made by the various producers in the highly competitive corn industry, as the prices of inputs, such as farmland, get bid up to reflect their opportunity costs. As output expands along the long-run supply curve in response to the increase in demand, however, the original owners of inputs whose prices are bid up do receive a permanent gain, provided that corn is an increasing-cost industry.
To date, the corn industry appears to be increasing- cost. Moreover, preliminary evidence indicates that some of the primary beneficiaries of the industry’s expansion are owners of farmland, especially farmland that is well suited to growing corn. Between 2002 and 2007, for example, the average price of an acre of farmland in Iowa (the big- gest producer of corn in the United States and home to the largest number of ethanol plants) increased by 57 percent. Between 2007 and 2008, average farm prices in Iowa increased another 20 percent from $4,200 an acre to $5,000 per acre (and to $6,500 per acre by early 2011), making Iowa one of the hottest real estate markets in the world. By contrast, home prices fell in half of the cities in the United States over the same time period, and the price of an average apartment in Manhattan grew by only 3.2 percent (and 16 percent in what was widely considered to be the red-hot real estate market in London, England).
When Does the Competit ive Model Apply? 239
C09.INDD 12:6:43:PM 08/06/2014 PAGE 239Trim Size: 203.2 mm X 254 mm
not persist. Thus, any output change will result from the firms utilizing existing plants more
or less intensively, and a short-run analysis is appropriate.
Even when appropriate, a short-run analysis identifies only an industry’s temporary resting
place, and subsequent long-run adjustments will continue moving the industry toward long-
run equilibrium. So, in most supply–demand applications we use the long-run supply curve.
9.9 When Does the Competitive Model Apply? The assumptions of perfect competition are stringent and are likely to be satisfied fully
in very few real-world markets. Still, many markets come close enough to satisfying the
assumptions of perfect competition to make the model quite useful. As we mentioned ear-
lier, product homogeneity and perfect information may not apply in the case of gasoline
retailing, yet the industry may be very close to being perfectly competitive since repre-
sentative firms have only limited pricing power.
How close is “close enough” in terms of the assumptions of perfect competition being
satisfied to make the model applicable? This question has no easy answer. For example,
consider the assumption that there must be “many” small firms, an assumption that can be
readily verified: firms can be counted. The relationship between the number of firms and an
industry’s competitiveness is not obvious. As we will see in Chapter 13, a relatively large
number of firms producing a homogeneous product may collude to fix prices above the
competitive level.
Conversely, even if there are only a few firms in an industry, each firm may face a highly
elastic demand curve if the market demand curve is very elastic or if the elasticity of sup-
ply by rival firms is high. For example, suppose that United is one of four suppliers of air
travel between Los Angeles and San Francisco, and that each of the four suppliers accounts
for one-fourth of the total air passenger travel between the two cities. Although it is by no
means a small player in this market, United may still face a highly elastic demand curve
if when it raises its price, rivals are quick to increase their output in response to United’s
price increase. A high elasticity of supply by rival airlines can thus severely limit United’s
pricing power. In Chapter 13 we will explore in greater depth how the output responses of
rivals to a firm’s price changes affect the elasticity of the firm’s demand curve.
With regard to other assumptions of perfect competition, such as free entry and exit of
resources, the competitive model frequently can be adapted to take the violations of these
assumptions into account. This adaptability is another reason why the model is important:
it can be extended to analyze the implications of specified departures from perfect com-
petition. For example, suppose that an industry is in long-run equilibrium and demand
increases, but new firms are blocked from entering the market. It is easy to show that output
will increase by less in this case, price will be higher, and economic profit will persist after
the industry’s firms have made the appropriate long-run adjustment.
Consider as well the assumption that firms produce a homogeneous product. This
assumption serves two purposes in the competitive model: it implies that a uniform price
will prevail, and it affects the elasticity of demand facing a particular firm. What if the prod-
uct homogeneity assumption is not fully satisfied? A good example of this is provided by
comparison-shopping surveys that find food prices higher in the inner city than in the sub-
urbs. For those who expect the competitive model to be an accurate description of reality,
this result violates the uniform-price implication and suggests monopoly pricing in the inner
cities. Further investigation, however, reveals profits to be no higher for inner-city food stores
than for suburban markets, a result consistent with competition but not with monopoly.
What is going on? Basically, food sold in the inner city and food sold in the suburbs are
not homogeneous products. To city residents, food sold in the inner city is worth more than
240 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 240Trim Size: 203.2 mm X 254 mm
food sold farther away. People are willing to pay something for the convenience of shop-
ping nearby, and when convenience costs the retailer something, the price of the “same”
product will differ. In comparison with costs in the suburbs, the cost of operating a food
store in the inner city is higher because rent, fire and theft insurance, and the salaries of
personnel all tend to be higher there. As a result, inner-city residents pay a higher price for
shopping near their homes.
Dropping the homogeneous product assumption means that differentials in price can
exist, but that is not often a serious problem and the competitive model can still be used.
Suppose, for example, that a city government passed a law that food prices in the inner city
cannot exceed prices in the suburbs. The competitive model applied to this setting would
predict that output (food sales) would fall in the inner city, shortages would exist, some
food stores would go out of business, and many inner-city residents would be worse off
because no food was locally available at the mandated lower price—the familiar effects of a
price ceiling. These predictions would, we suspect, be borne out in practice.
Finally, the assumption that firms and consumers have all the relevant information is also
necessary if markets are to behave exactly as the competitive model predicts. For example,
if consumers are ignorant of price, even homogeneous products can sell for different prices.
Dropping the assumption of full information does not mean, however, that we have
to replace it with one of complete ignorance. One of the results of real-world markets is
that firms and consumers have incentives to acquire the information that is important for
their economic decisions. Although they may not become fully informed, because there
are costs associated with acquiring information, we can easily suppose that they will
become well enough informed that the assumption of complete information is not too
great a distortion.
Economists have only recently begun to systematically analyze the acquisition of infor-
mation and the way “information costs” influence the workings of markets. It is too early
to make any sweeping generalizations in this area, but we do not want to suggest that the
degree of information is irrelevant to the functioning of a market. Lack of information on
the part of consumers may result in market outcomes that deviate significantly from the
competitive norm. We will discuss issues related to information in more detail in Chapter 14.
SUMMARY
A perfectly competitive industry is characterized by a
large number of buyers and sellers, free entry and exit of
resources, a homogeneous product, and perfect informa-
tion. Although few industries fully satisfy all these con-
ditions, the model of perfect competition remains quite
useful in analyzing many markets.
Since a competitive firm sells a product similar to that
sold by many other firms, the firm’s output decision has
a negligible effect on the product’s price. Thus, the com-
petitive firm’s demand curve can be approximated by a
horizontal line at the level of the prevailing price.
For any firm, the most profitable output occurs where
the marginal cost of producing another unit of output
just equals the marginal revenue from selling it.
For a competitive firm with its horizontal demand
curve, price equals marginal revenue, so it maximizes
profit by producing where marginal cost equals price.
Price, however, must be at least as high as average vari-
able cost (AVC) or the firm suffers a loss in excess of total fixed cost.
If price does not cover AVC in the short run, the firm will shut down; if price does not cover LAC in the long run, the firm will go out of business.
The competitive firm’s short-run supply curve is its
marginal cost curve, so long as marginal cost exceeds
AVC. The assumption of profit maximization allows us to pre-
dict how a competitive firm will respond to changes in the
product price or in input prices. An increase in the prod-
uct price (with unchanged cost curves) will lead the firm to
expand output as it moves up its marginal cost curve.
A reduction in one or more input prices will shift the
cost curves downward and so will lead the firm (with an
unchanged product price) to expand output.
C09.INDD 12:6:43:PM 08/06/2014 PAGE 241Trim Size: 203.2 mm X 254 mm
Review Quest ions and Problems 241
The competitive industry’s short-run supply curve is
the horizontal summation of the short-run marginal cost
curves (above AVC) of an industry’s firms. Because the law of diminishing marginal returns
implies that each firm’s marginal cost curve slopes
upward, the short-run industry supply curve also slopes
upward.
In the long run, firms have sufficient time to alter plant
capacity and to enter or leave the industry. The long-run
supply curve takes these adjustments into account, with
firms always guided by the search for profit.
Two shapes of the long-run industry supply curve
are most likely. First, an increasing-cost industry has
an upward-sloping long-run supply curve, reflecting the
increase in the prices of one or more inputs as the indus-
try expands.
Second, a constant-cost industry has a horizontal
long-run supply curve, reflecting a situation where the
industry can expand its use of inputs without affecting
their prices.
At all points along a long-run supply curve, economic
profit is zero, because only when profit is zero is there no
incentive for firms to enter or leave the industry.
Input supply curves and technology are assumed as giv-
ens in deriving a long-run supply curve; changes in these
underlying factors shift the long-run supply curve.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
9.1 What are the four assumptions of the perfectly competi- tive model? Critics are fond of pointing out that few, if any,
real-world markets satisfy all four conditions, implying that the
competitive model has little relevance for real-world markets.
How would you respond to these critics?
9.2 How can consumers of apartments in Atlanta have a downward-sloping demand curve for apartments, and yet each
Atlanta rental property owner face a horizontal demand curve?
*9.3 Assume that a competitive firm has the short-run costs given in Table 8.1. What is the firm’s most profitable output,
and how large is profit if the price per unit of output is (a) $15?
and (b) $26?
9.4 “IBM should never sell its product for less than it costs to produce.” If “costs to produce” is interpreted to mean IBM’s
average total cost, is this correct? If it is interpreted to mean
average variable cost, is the statement correct? If it is inter-
preted to mean marginal cost, is the statement correct?
9.5 In Table 9.1, suppose that variable input prices increase by 50 percent. Will the firm’s profit-maximizing output level
change? Illustrate your answer with a graph.
9.6 “If Conagra is a competitive firm, it will never operate at an output where its average total cost curve is downward slop-
ing.” True or false? Explain.
*9.7 “The difference between price and marginal cost is the amount of profit per unit of output, which the firm wishes to be
as large as possible.” True, false or uncertain? Explain.
9.8 Suppose that through laziness and/or sheer stupidity, Densa Inc. always falls 10 percent short of producing the
profit-maximizing output. Would a higher product price lead
to greater output? Would an increase in input prices lead to a
reduction in output? Does this result suggest why economists
are not overly concerned about whether the profit-maximizing
assumption is exactly correct? Explain.
9.9 Suppose that Keystone is a firm in the perfectly competitive ski resort business. If all of Keystone’s input prices unexpect-
edly double, and at the same time the product price doubles,
what will happen to Keystone’s profit-maximizing level of out-
put and its profit in the short run? In the long run? (Assume that
Keystone begins from a position of long-run equilibrium.)
9.10 Starting from a long-run equilibrium, trace the effects of an unanticipated reduction in demand for (a) a constant-cost
industry and (b) an increasing-cost industry. Note: This pro- cess is just the reverse of our derivation of the supply curves
in Section 9.8, but it is very good practice to think through the
process.
9.11 In a constant-cost industry each firm’s MC curve is upward sloping, yet all the firms together—the industry—have a hori-
zontal supply curve. Explain why there is no contradiction.
9.12 If each business school in the MBA education industry is operating where LMC = MR (= P), does it follow that the industry is in long-run equilibrium? Explain.
*9.13 “Suppose that the defense contracting business is per- fectly competitive. In long-run equilibrium the price just cov-
ers average production cost, and every contractor makes zero
profit. Thus, if the price goes down, even a little bit, due to a
decrease in government defense spending, all contractors will
go out of business.” Discuss this statement.
9.14 Assume that the MBA education industry is constant-cost and is in long-run equilibrium. Demand increases, but due to
strict accreditation standards, new firms are not permitted to
enter the market. Analyze the determination of a new long-run
equilibrium, showing the effects for a representative school as
well as for the market as a whole.
9.15 For a given increase in demand, will output increase by more if the MBA education industry is a constant-cost or
increasing-cost industry? For a given decrease in demand, will
242 Prof i t Maximizat ion in Perfect ly Competit ive Markets
C09.INDD 12:6:43:PM 08/06/2014 PAGE 242Trim Size: 203.2 mm X 254 mm
output fall by more if the industry is a constant-cost or increas-
ing-cost industry?
9.16 How would each of the following phenomena affect the long-run supply curve of apples?
a. Workers in the apple industry form a union. b. Consumers find out that apples cause cancer. c. Hard-to-control bugs that eat apples invade from Mexico. d. The government passes a law requiring apple trees to be
planted at least 60 feet apart.
e. The government sets a maximum legal price (price ceiling) at which apples can be sold.
f. Immigration laws change to permit more itinerant apple pickers to enter the country.
g. The government passes a minimum wage law for apple pickers.
9.17 “Because agricultural demand is inelastic, a technological advance that lowers production costs will reduce total revenue.
Thus, farmers have no incentive to introduce such a tech-
nique.” True, false or uncertain? Explain.
9.18 Can all schools in the MBA education market face con- stant returns to firm scale and yet the industry have an upward-
sloping long-run supply curve? If the MBA education industry
is constant-cost, what, if anything, does this imply about the
returns to scale faced by the industry’s individual schools?
9.19 As the industry moves from point B to C in Figure 9.10b, how can the industry’s output increase yet NYU’s output fall?
9.20 The American Red Cross is a supplier to the perfectly competitive domestic blood market. Unlike the other suppliers,
however, the Red Cross is strictly nonprofit—its goal is to sell
as much blood as possible without making a profit. Given this
goal, is the Red Cross supply curve equal to its average vari-
able cost curve? Explain why or why not.
9.21 Suppose that the gasoline retailing industry is perfectly competitive, constant-cost, and in long-run equilibrium. If the
government unexpectedly levies a 5-cent tax on every gal-
lon sold by gasoline retailers, depict what will happen to the
representative firm’s cost curves. What will the effects of the
tax be in the short run on industry output and price? Will the
price rise by the full five cents in the short run? In the long
run? How would your answers change if the industry was
increasing-cost?
9.22 Suppose that Mensa Inc. is a representative firm operating in a perfectly competitive industry. Mensa’s total cost of pro-
duction, TC, is given by the equation TC = 5,000 + 5q2, where q is Mensa’s output. Based on this equation, Mensa’s marginal cost is 10q. If the output price is $100, what is Mensa’s short- run profit-maximizing output? How much profit does Mensa
make at that output?
9.23 Suppose that oil tankers fall into three classes: medium (tankers capable of carrying between 10,000 and 70,000 dead-
weight tons, DWT, of crude oil—approximately 170 million
gallons); large (tonnages between 70,000 and 175,000 DWT);
and super (tonnages over 175,000 DWT). The total tonnage in
each of the three classes is 53,000,000 in medium; 76,000,000
DWT in large; and 171,000,000 in super. Finally, suppose that
the constant annual operating cost per deadweight ton is $149
for medium; $107 for large; and $84 for super. Based on the
preceding information, graph the short-run supply curves for
each of the three tanker classes as well as for the overall oil
shipping industry.
9.24 “In a competitive industry‚ high prices in response to a positive demand shock prevent higher prices.” Explain whether
this statement is true‚ false or uncertain in the context of the
competitive industry model developed in this chapter. To help
your analysis‚ think what would happen if government poli-
cymakers precluded suppliers from raising their prices in the
wake of a positive demand shock (e.g.‚ local suppliers of lum-
ber were prohibited from raising the price of their products in
the wake of a tornado ravaging a community and the resulting
need to rebuild homes in the community).
9.25 Over the past four decades, the price of personal comput- ers in real as well as nominal terms has declined markedly. Does
this mean that the personal computer industry is decreasing-cost
and that the long-run supply curve for personal computers is
downward sloping? Explain why or why not.
9.26 Suppose that Delta Airlines fills only 60 percent of the seats‚ on average, on its Boeing 737 flights—a number 20 per-
cent below the industry average. If Delta incurs incremental
costs of $2‚000 per 737 flight‚ wouldn’t it do better by schedul-
ing fewer 737 flights and thereby increasing the percentage of
seats filled‚ on average‚ per flight? Why or why not? Explain.
9.27 “In the market for cell phones‚ the amount generated in revenue per unit sold by Nokia roughly equals the associ-
ated incremental cost. Given that Nokia is losing money on its
cell phone product line‚ it would be wise for the company to
cease manufacturing such a product.” True‚ false or uncertain?
Explain.
9.28 Following the discovery of gold in 1848 at Sutter’s Mill, near the then sparsely populated Sacramento, California, there
was a tremendous rush to mine the precious metal. A sizable
influx of miners to the area dramatically increased the demand
for and price of mining implements such as shovels. A shovel,
whose nominal price had been only $1 in the Sacramento area
prior to the gold discovery, sold for as much as $50 immedi-
ately following the discovery. Using the concept of a com-
petitive firm’s short-run supply curve and a graph, explain the
reason for this huge spike in the price of shovels.
C10.INDD 12:9:21:PM 08/06/2014 PAGE 243Trim Size: 203.2 mm X 254 mm
243
CHAPTER 10 Using the Competitive Model
Learning Objectives
Show how changes in market conditions or government policies affect the welfare of consumers, producers, and market participants as a whole. Analyze the effects of an excise tax on a specific good on the welfare of consumers, producers, and market participants as a whole. Detail how regulation of the U.S. airline industry affected fares, airline company profits, and service quality. Explain how the entry restrictions imposed by most major U.S. cities on taxis affect fares and the profits earned by licensed taxi owners. Understand the effects of international trade on consumer and producer surplus and why a net gain results to a country from either imports or exports. Explore how government-specified maximum quantities, or quotas, on sugar imports affect consumers, domestic producers, and the net welfare of the United States as well as other countries that produce sugar.
This chapter builds on the theory established in the previous chapter to show how the competitive model can be used to analyze particular industries and the effects on them of
government intervention. The examples we will explore include gasoline taxes, airline reg-
ulation, taxicab licensing, and import quotas. The broad range of applications should illus-
trate the usefulness of the perfectly competitive model. The applications themselves will
also indicate that while government intervention may be justified on the grounds of helping
people, the effects of such intervention may end up being precisely counter to the objec-
tives of the proponents of the intervention.
10.1 The Evaluation of Gains and Losses Changes in market conditions or government policies always result in either gains or losses
for participants in the market. For example, rent controls (that have been used in many cities)
often benefit at least some tenants at the expense of landlords. But how large is the benefit to
tenants and how large is the cost to landlords? In total, do the benefits outweigh the losses, or
Memorable Quote “No nation was ever ruined by trade.”
—Benjamin Franklin
244 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 244Trim Size: 203.2 mm X 254 mm
is the reverse true? To answer questions like these, we need a way to measure the gains and
losses felt by market participants. In this section we will explain how we can use the con-
cepts of consumer surplus and producer surplus to measure such gains and losses.
We have already explained the concept of consumer surplus. Recall that consumer surplus is a measure of the net gain to a consumer (or group of consumers) from purchas- ing a good. (You may wish to review how it is shown by areas under the demand curve;
see Section 4.5.) The concept of producer surplus is an analogous measure of the net gain
to those involved in producing and selling a good. Because this is a new concept, we will
begin by explaining how it is related to the competitive supply curve.
Producer Surplus Producers of goods often secure gains, called producer surplus, from the sale of output to consumers. The gain results because the price often exceeds the minimum amount that
would be necessary to compensate the seller. To see how we can show the producer sur-
plus in a competitive market, consider the long-run supply of sugar curve S in Figure 10.1. Assume that the sugar industry is at point E on the supply curve, output is Q, and sugar is being sold at a price of 15 cents per pound. What we will show is that the shaded triangular
area is a measure of the net gain to sugar suppliers.
To explain this point, we begin by assuming that the industry takes a very simple form.
Assume that each firm consists of only one person (the owner-manager-worker) who can
produce only one unit of output. (We will drop this assumption later.) Now, consider
the derivation of the supply curve, starting with an output of zero. At a price lower than
5 cents, no firm produces sugar—all potential firms can do better employing their resources
elsewhere. At a price of 5 cents, firm A enters the market and produces one unit. At that
price, firm A is just barely induced into this market, and so it makes no net gain from oper-
ating in this industry. If the price is 6 cents, however, firm B enters the market and pro-
duces one unit (so total output is two units). Firm B could have earned 6 cents elsewhere
and so secures no net gain if the market price remains at 6 cents. Other firms enter the
market at higher prices; in this way we trace out the industry supply curve as a step-like
relationship.
Now consider the situation when the market price is 15 cents. At that price, firms A, B,
C, and several more are selling one unit each at a price of 15 cents. How much does each
firm benefit from selling in this market? Firm A would have been willing to sell one unit
for a minimum of 5 cents; this is shown as rectangular area A1. However, firm A sells the unit for 15 cents, shown as the rectangular area A1 plus A2, so it receives a net benefit of
consumer surplus gains to a consumer or group of consumers from purchasing a good arising from its cost being below the maximum that consumers are willing to pay
producer surplus gains to producers from the sale of output to consumers, arising from price exceeding the minimum necessary to compensate the seller
Producer Surplus Producer surplus is a measure of the net gain to producers from operating in a given market, and is shown as the area between the price line and the supply curve. It can be thought of as the sum of rectangles like A2, B2, and C2, each of which is the net gain associated with the sale of a particular unit.
Price
S
Output0 1 2 3
A1
B1
C1
A2
B2
C2
15¢
6¢ 7¢
5¢
Q
E Figure 10.1
The Evaluat ion of Gains and Losses 245
C10.INDD 12:9:21:PM 08/06/2014 PAGE 245Trim Size: 203.2 mm X 254 mm
10 cents—shown as area A2. Area A2 is the producer surplus realized by firm A. By similar reasoning, firm B receives a net benefit, or producer surplus, of area B2 (9 cents) because its revenues are 15 cents and it would have been willing to sell for 6 cents. The net gain for each
firm is therefore shown as a rectangular area like A2, B2, C2, and so on. The sum of all these areas for the firms operating in the market is the total producer surplus realized by all the
firms. Of course, if we assume more realistically that one unit of output is small relative to
total output Q, so the rectangles become much narrower, then the supply curve approaches the smooth curve S. In that case, the total producer surplus is shown as the shaded area between the supply curve and the price line.
When we drop the assumption of one-person firms and consider the general case of a
competitive industry, the triangular area shown continues to represent the total producer
surplus. Now, however, we have to reconsider exactly who gets this surplus. Recall that
economic profits are zero at every point on a long-run supply curve; producer surplus is not
the same as economic profit. Indeed, owners of the firms may not receive any producer sur-
plus at all. Producer surplus is the total net gain that goes to anyone involved in producing the good, and that includes the suppliers of inputs to the industry. To see what this implies,
suppose that the sugar industry has an upward-sloping supply curve of labor, but a horizon-
tal supply curve of every other input used to produce sugar. In this case, the sugar industry
supply curve will be upward sloping, as in Figure 10.1, because the wage rate will rise as
the industry expands and employs more workers. What is producer surplus here? It is the
net gain to the workers who would have been willing to work at the lower wage rates (when
the industry output was smaller), but who are now receiving wage rates higher than the
minimum necessary to induce them to work in the sugar industry. This total net gain will
be shown by the shaded triangular area above the product supply curve for this market. No
other input owners receive any producer surplus in this situation; only the workers.
Thus, the area between the product supply curve and the price line identifies the total
producer surplus that accrues to someone engaged in the production of the good. Even though this area identifies the total net gain to “producers,” we cannot tell from the sup-
ply curve alone exactly who receives this benefit. In general, producer surplus will accrue
to some of the owners of inputs that have upward-sloping supply curves to the industry.
Owners of inputs with horizontal supply curves to the industry receive no producer surplus.
(Note that this conclusion implies that there is no aggregate producer surplus for a constant-
cost competitive industry when it is in long-run equilibrium.)
Consumer Surplus, Producer Surplus, and Efficient Output In general, consumers and producers benefit from participation in a competitive market,
and the size of that benefit is measured by their consumer surplus and producer surplus, as
shown in Figure 10.2a. The competitive equilibrium is at an output of Q and a price of $5. The net gain to consumers is shown as the shaded triangular area A between their demand curve and the price line. Similarly, the net gain to producers is shown as the shaded triangu-
lar area B above the supply curve up to the price line. The total net gain to those who par- ticipate in this market is therefore the sum of consumer surplus and producer surplus—area
A plus area B. The sum of producer and consumer surplus is called the total surplus. The total surplus, area A plus B in Figure 10.2a, is a measure of the aggregate net gain
that is realized by participants in this market. Because it is the sum of the way the market
affects the well-being of everyone participating in it, total surplus is often used as a mea-
sure of how well the market functions relative to its potential. The amount of total surplus
will often be changed by government policies or by changes in market structure (e.g., if the
market becomes monopolized). The change in total surplus in such cases is often taken as a
measure of the gain or loss in well-being to market participants.
There is an alternative way to see that the sum of areas A and B in Figure 10.2a is equal to total surplus. This is illustrated in Figure 10.2b. The equilibrium quantity is once again Q.
total surplus the sum of producer and consumer surplus
246 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 246Trim Size: 203.2 mm X 254 mm
To identify the total surplus, we consider the net gain associated with each unit of output
from zero to Q. It is important to recall that the height of the demand curve can be inter- preted as showing the marginal benefit of the good and the height of the supply curve can be
interpreted as showing the marginal cost. Now consider the production of one unit. For the
moment, we will think of the demand and supply curves as discrete, step-like relationships.
The first unit received by consumers has a marginal benefit of $13, the sum of rectangular
areas A1 and A2. Producers must be compensated by a minimum of $2; the marginal cost of the first unit is $2, shown as area A1. The marginal benefit of the first unit is $11 greater than its marginal cost; there is a net gain associated with the first unit of $11, shown as area A2, the difference between the marginal benefit of that unit and its marginal cost. The net gain
is, of course, the total surplus associated with the first unit of output. Note, however, that
this procedure does not identify who receives this surplus—producers or consumers.
Similarly, we see that there is a net gain from the second unit of output of area B2, equal to the excess of marginal benefit over marginal cost for that unit. The sum of areas A2 and B2 is the total surplus when two units are produced. By identifying the rectangular areas of net gain of each unit from zero to the quantity actually produced and summing the net gains
over all units, we can determine the total surplus for quantity Q. When we let units of out- put become small relative to total output, the rectangles become narrower, and the demand
and supply curves become the smooth curves shown as D and S. Then the sum of the net gains, the total surplus, will be shown as the shaded area in Figure 10.2b. Note that the total
surplus arrived at by this reasoning is the same as that shown as the sum of consumer and
producer surplus in Figure 10.2a.
We have explained two different ways of identifying the total surplus associated with the
functioning of a market. The first way, illustrated in Figure 10.2a, involves determining the
surpluses realized by producers and consumers separately and then adding them together.
Dollars per unit
Output
$5 = P
$2
$13
D
SA
B
0 Q
(a)
Dollars per unit
Output
$2
$13
D
SA2
0 1 2 Q
(b)
B2
A1 B1
Competition Maximizes Total Surplus (a) Consumer surplus equals area A, and producer surplus equals area B; total surplus is the sum of consumer and producer surplus. (b) Total surplus can be thought of as the sum of the net gains (excess of marginal benefit over marginal cost) associated with the production of each unit of output from zero to Q.
Figure 10.2
The Evaluat ion of Gains and Losses 247
C10.INDD 12:9:21:PM 08/06/2014 PAGE 247Trim Size: 203.2 mm X 254 mm
This approach has the advantage of showing how the overall net gain is distributed between
consumers and producers (but recall that it doesn’t identify which producers benefit). The
second way, illustrated in Figure 10.2b, involves determining the total surplus associated
with each unit of output and summing over all units of output. It has the advantage of often
being an easier procedure to use (as we will see later), but the disadvantage of not identify-
ing the distribution of the net gain between producers and consumers.
Figure 10.2b shows why the equilibrium output of a competitive industry is also the
efficient level of output. Efficiency in output requires that output be expanded to the point where the marginal benefit equals marginal cost. To say that the output level is efficient,
moreover, is the same thing as saying that the net gain, or total surplus, from producing
the good is as large as possible. Let’s check to see if that is the case when output is Q in Figure 10.2b. If output is expanded beyond Q, then the marginal benefit of additional units will be less than their marginal cost, so the net gain of these units is negative—that is,
there is a net loss in total surplus associated with production beyond Q. Thus, the total surplus will be smaller than at output Q; it will be the shaded area minus the net loss on units in excess of Q. Any expansion of output beyond Q will therefore reduce the total surplus. Similarly, if output is less than Q, the total surplus will also be smaller than the shaded area. For example, if output is two, the total surplus is the area between the
demand and supply curves up to an output of two (roughly the sum of rectangular areas A2 and B2), which is clearly smaller than the shaded area. Thus, an output of Q will maximize the total surplus, or net gain, of participants in the market. A competitive market achieves
this result.
The Deadweight Loss of a Price Ceiling To illustrate the use of these measures of surplus, we will consider again a price ceiling
applied to rental housing, a policy we first examined in Section 2.4. In Figure 10.3a, the
initial equilibrium involves a monthly price of $800 and a quantity of Q. The government then mandates a maximum price of $600. At that price, suppliers reduce output to Q1, but consumers would like to purchase an amount of Q2; the difference is the excess demand, or shortage, caused by the price ceiling.
How does the price ceiling affect the total surplus realized in this market? As we
explained, there are different ways of evaluating total surplus and changes in it. The easiest
way, based on the Figure 10.2b approach, is to evaluate total surplus without regard to its
distribution between producers and consumers. Thus, in Figure 10.3a, total surplus before
the price ceiling was area AEF, but after the price ceiling, when only Q1 units are actually produced, total surplus is area ABCF, the sum of the net gains on each unit from zero to Q1.1 Total surplus is therefore smaller under the price ceiling; the decrease in the size of total
surplus is the area BEC. The loss in total surplus, area BEC, is the deadweight loss (also known as welfare cost) of the price ceiling. As mentioned in Chapter 5, the deadweight loss
is a measure of the aggregate loss in well-being of market participants. In this case, it is a
measure of the loss due to production of an inefficient quantity of rental housing services.
Instead of comparing the total surplus with and without the price ceiling (for outputs
Q1 and Q), we can arrive at the same conclusion by considering the change in total surplus caused by the reduction in output from Q to Q1. Consider the units not produced (because of the price ceiling) between Q and Q1. The marginal benefit of each of these units is shown by the height of the demand curve between Q and Q1; similarly, the marginal cost is shown by the height of the supply curve. As you can see, each unit not produced has a marginal
efficiency in output the condition in which output is expanded to the point where marginal benefit equals marginal cost
deadweight loss also called welfare cost, a measure of the aggregate loss in well-being of participants in a market resulting from output not being at the efficient level
1This assumes that the smaller quantity is rationed among consumers in such a way that the consumers who place the highest value on the rental housing actually get it. This may not be the case in the absence of a higher (i.e., black market) price to ration the rental housing among consumers. Then the total surplus under the price ceiling will be smaller than area ABCF.
248 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 248Trim Size: 203.2 mm X 254 mm
benefit greater than its marginal cost, so there is a net loss from not producing these units.
The net loss is the difference between marginal benefit and marginal cost, and if these net
losses are summed we arrive at area BEC as the aggregate net loss, or deadweight loss, of the price ceiling. (It may help if you think of the narrow rectangles, like area B2 in Figure 10.2b, that make up the area BEC in Figure 10.3a.)
An alternative way to proceed, based on the approach used in Figure 10.2a, considers the
effect on consumers and producers separately. This is illustrated in Figure 10.3b. Before the
price ceiling, total consumer surplus was area AEP. After the price ceiling, when consum- ers are purchasing Q1 units at a price of PC, total consumer surplus is area ABCPC, which is the sum of the net gain (marginal benefit less price) on each of the Q1 units. Compared with the situation without the price ceiling, consumers have gained area X (rectangular area PGCPC) and lost area Y (triangular area BEG). As drawn, area X is larger than area Y, so total consumer surplus has increased (by X − Y). Note that area X is the gain that consumers obtain from being able to purchase the Q1 units at a lower price; the height of the rectangle is the reduction in price. But that is not the only way consumers are affected; they also end
up consuming a lower quantity than before the price ceiling. Area Y is the consumer sur- plus they previously received on consumption from Q1 to Q; they lose this net gain due to reduced production under the price ceiling. The net impact on consumer welfare depends
on the size of these two separate effects.2
2Whether area X is larger than area Y depends on the height of the price ceiling and the elasticities of the demand and supply curves. As these curves are drawn in Figure 10.3b, area X is larger, but it is possible for area Y to be larger (e.g., imagine a price ceiling set below the level F0).
Rental housing services
0
$800 = P
Q2Q1 Q
S
B
A
D
E
C
Monthly rent per unit
$600 = Pc
(a)
F
Rental housing services
0
$800 = P
Q2Q1 Q
S
B
A
D
E Y
C
Z X
Monthly rent per unit
$600 = Pc
(b)
G
F
A Price Ceiling Reduces Total Surplus (a) The price ceiling of PC results in output of Q1; there is a deadweight loss associated with producing less than the competitive output that is shown as area BEC. (b) The price ceiling results in a gain in consumer surplus shown by area X minus area Y, and a loss in producer surplus shown by area X plus area Z. Total surplus falls by the sum of the changes in consumer and producer surplus, area Y plus area Z.
Figure 10.3
Excise Taxat ion 249
C10.INDD 12:9:21:PM 08/06/2014 PAGE 249Trim Size: 203.2 mm X 254 mm
The effect on producers is easier to understand. Producer surplus before the price ceiling
was area PEF. After the price ceiling, producer surplus is PCCF. Thus, producer surplus falls by the sum of areas X and Z (the area of PECPC). Producer surplus is unequivocally reduced by the price ceiling.
Because total surplus is the sum of consumer and producer surplus, we can determine
the change in total surplus by summing the changes in consumer and producer surplus. In
this case, the change in consumer surplus is X − Y, and the change in producer surplus is −X − Z (where a minus sign indicates a loss in surplus). Thus, the sum of the changes in consumer and producer surplus is −Y − Z. Total surplus falls by the sum of areas Y and Z, so that triangular area BEC identifies the loss in total surplus, or the deadweight loss, due to the price ceiling. Note that this is the same area identified in Figure 10.3a; these strategies
are just different approaches to measure the same thing.
One advantage to evaluating consumer and producer surplus changes separately is that it
makes clear who gains and who loses. In this case, consumers gain area X − Y and produc- ers lose area X + Z; the gain to consumers is less than the loss to producers by area Y + Z— that is, the aggregate loss in total surplus. From this we see that finding a deadweight loss is
not the same as saying that everyone is worse off. A deadweight loss measures the overall
efficiency loss when the effects on everyone are summed, but the distribution of gains and
losses may also be important to consider. If you view the well-being of consumers as suf-
ficiently more important than the well-being of producers, you might favor the price ceil-
ing in this case despite the fact that it produces a deadweight loss. (On the other hand, it is
unlikely that all consumers benefit even here; the gain in consumer surplus is an aggrega-
tion over all consumers, some of whom may be made worse off by the rent control because
they cannot find housing at the legal price.)
10.2 Excise Taxation The competitive model and the concept of efficiency introduced in the preceding sec-
tion can be employed to analyze the effects of government policies on a wide variety of
industries. Among the policies are the excise taxes levied on specific goods such as gaso-
line, cigarettes, and alcohol. For example, the 1993 Budget Reconciliation Bill increased
the federal excise tax on gasoline by 4.3 cents per gallon (from 14.1 to 18.4 cents). The
increase has raised over $85 billion in government revenue since 1993. Further increases
in the federal excise tax on gasoline, as high as 50 cents per gallon, have been advocated
as a means to reduce America’s dependence on foreign sources of oil. State and local
governments also use excise taxes. Such excise taxes applied to competitively produced
products are easily analyzed using the industry supply and demand curves. It is instruc-
tive, however, to develop the analysis step by step, showing how the tax affects indi-
vidual firms as well as the market and examining how the short-run effects differ from
long-run effects.
Let us analyze the federal government’s excise tax on gasoline. In Figure 10.4, we
assume that the industry is initially in long-run competitive equilibrium. Most govern-
ment studies in fact indicate that the industry is close to perfectly competitive. Figure 10.4a
shows a typical firm, ExxonMobil, making zero economic profit at the market-determined
price of $3.00 per gallon of gasoline. Price equals short-run and long-run marginal cost at
ExxonMobil’s initial output of q1. (We have not drawn in the long-run marginal cost curve or the short-run average cost curve. Throughout the analysis, we use only the relationships
essential to the important points to avoid cluttering the diagrams.) Figure 10.4b identi-
fies the original equilibrium price and output for the industry at $3.00 and Q1, which is determined by the intersection of the long-run supply curve LS and the demand curve D at point A. The short-run supply curve SS passes through point A because it is the sum of the
250 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 250Trim Size: 203.2 mm X 254 mm
firms’ SMC curves. Recall that SS is more inelastic than the long-run supply curve because expanding (or reducing) output is more costly in the short run (when some inputs are fixed)
than in the long run (when all inputs are variable).
The Short-Run Effects of an Excise Tax Now suppose that the government unexpectedly taxes each firm in the industry 50 cents per
gallon of gasoline sold. Such a tax is a per-unit excise tax.3 Let us first consider the imme-
diate impact of the tax on the individual firm, ExxonMobil. Before the tax, long-run aver-
age cost at output q1 is $3.00, but the tax increases ExxonMobil’s average cost of providing that output to $3.50. In fact, at every possible output, average cost is 50 cents higher than
before. Note that the tax does not affect the cost of the inputs needed to produce the product
but simply adds 50 cents per gallon to their cost. In Figure 10.4a, the 50-cent tax is shown
as a parallel upward shift in the LAC curve to LAC + T, where T equals the 50-cent per- unit tax. All other per-unit cost curves, including the marginal cost curve, shift vertically
upward also, because all per-unit production costs are increased by the amount of the tax.
If we temporarily assume that the gasoline price is unaffected, two important effects of
the tax on the individual firm can be identified. First, ExxonMobil will operate at an eco-
nomic loss at q1 because average cost ($3.50) exceeds price ($3.00). Second, at q1, marginal cost, which has increased to $3.50 at that output, exceeds the price. ExxonMobil’s immedi-
ate response will be to cut its loss by reducing output along its new short-run marginal cost
curve, SMC + T, to q*, where marginal cost equals (the assumed unchanged) price. Exxon- Mobil still incurs a loss at q*, but the loss is smaller than if it keeps output at q1.
3Another common form of excise tax is an ad valorem excise tax. An ad valorem tax is levied as a certain percentage of the market price. In contrast, a per-unit tax does not depend on the market price. The impor- tant economic effects of the two types of excise taxes are the same, however, and the per-unit tax is slightly easier to analyze.
Dollars per unit
ExxonMobil
Output
$3.50
$3.35
$3.00 A
D
c
SMC + T SMC
LAC + T
LAC ′ + T
LAC C
B
0 q* qs q1
T
Dollars per unit
Market
Output
$3.50
$3.35
$3.00
SS + T
LS + T
LS
SS
0 Q2 Q1QS
T
(a) (b)
Effects of a Per-Unit Excise Tax (a) The firm’s cost curves shift upward by the amount of the tax, and, in the short run, the firm reduces output. (b) The industry supply curves also shift upward by the amount of the tax, and industry output declines. Output decreases by more in the long run so the price rises above its short-run level.
Figure 10.4
Excise Taxat ion 251
C10.INDD 12:9:21:PM 08/06/2014 PAGE 251Trim Size: 203.2 mm X 254 mm
Because the tax applies to all firms in the industry, they all incur a loss and have an
incentive to cut output; this will affect the market price. Earlier, we assumed for the moment
that the price was unchanged to trace out the individual firm’s immediate response to the
tax. Now let’s see what happens to price when all firms in an industry reduce output. In
Figure 10.4b, the initial short-run industry supply curve SS is the sum of the individual firms’ SMC curves. Because the tax shifts the firms’ marginal cost curves vertically upward by 50 cents, the SS curve also shifts vertically upward by 50 cents, to SS + T (distance BA equals 50 cents); SS + T is the sum of the firms’ SMC + T curves. In other words, for the industry to supply any given rate of output, the price must be 50 cents higher than before
because the tax has added 50 cents to unit costs. Of course, the industry will not continue to
supply Q1 gallons. In the short run, as firms cut output along their SMC + T curves, indus- try output falls along the SS + T curve. The short-run equilibrium occurs at point C, where SS + T intersects D. Industry output is lower than the pre-tax equilibrium (QS versus Q1) and price is higher (CQS versus AQ1). In Figure 10.4a, ExxonMobil is at point c on its SMC + T curve and is still incurring a short-run loss.
The Long-Run Effects of an Excise Tax This completes the analysis of the short-run adjustment to the excise tax. Because firms are
still taking losses, however, we have not yet reached a position of long-run equilibrium.
The losses provide an incentive for firms to exit the industry altogether, allowing resources
to move to other industries where normal profits can be made. We can identify the new
long-run equilibrium from the long-run supply curve. Like the short-run supply curve, the
tax shifts the long-run supply curve vertically upward by 50 cents because it adds 50 cents
to all per-unit costs, in the short and long run. Thus, the long-run supply curve becomes
LS + T, passing through point B where distance BA equals 50 cents in Figure 10.4b. In the short run, the tax causes the industry to restrict output along SS + T; in the long run, the industry restricts production along LS + T. The final long-run equilibrium occurs at a price of $3.35 and an output of Q2, where LS + T intersects the unchanged demand curve. In the long run, the decrease in industry output is greater than in the short run (because firms have
time to exit the industry and/or adjust their scales of plant), and consequently the price to
consumers is higher.
As drawn in Figure 10.4b, the price to consumers has risen by only 35 cents (from $3.00
to $3.35) even though the per-unit tax is 50 cents. Thus, firms receive only $2.85 per unit
($3.35 minus 50 cents per unit) after paying the tax, less than they received before the
tax was levied. Does this mean that firms are operating at a loss so that Q2 is not really a long-run zero-profit equilibrium? No. We have assumed that the industry is an increasing-
cost industry (with an upward-sloping long-run supply curve), which implies that some
input prices fall as fewer of these resources are used. As a result, firms’ cost curves shift
downward as the industry contracts along its long-run supply curve. Figure 10.4a shows
the situation from ExxonMobil’s viewpoint. As input prices fall, the LAC + T curve shifts downward to LAC′ + T, so ExxonMobil makes zero economic profit at the $3.35 price.4
This step-by-step analysis shows how we can start with a firm’s cost curves and deter-
mine the effects of a tax on an industry’s supply and demand curves. It is possible, how-
ever, to analyze the effects by using just the demand and supply curves at the industry level.
For example, if we are interested in only the industry-level effects in the long run, all we
need to do is examine the effects of the shift in the long-run supply curve from LS to LS + T in Figure 10.4b. This immediately identifies the outcome: a lower output and higher price.
4Whether firms produce more or less than before the tax depends on how their LAC curves shift downward. If with lower input prices the minimum point on LAC′ + T occurs at the same output as before the tax (q1), as drawn in the graph, firm output is the same as before the tax. The reduction in industry output then results from some firms exiting the industry. The minimum point on LAC′ + T, however, can occur at a higher or lower output, depending on which input prices fall and the nature of a firm’s production function.
252 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 252Trim Size: 203.2 mm X 254 mm
However, the more systematic approach we have used helps reinforce the nature of the long-
run adjustment process and emphasizes how the process affects individual firms. This is useful
since it is not always obvious how an industry’s supply curves will be affected; in many cases,
it is important to develop the analysis by first examining how individual firms are affected.
Let’s return to the economic effects of an excise tax and review two implications of
the analysis. First, even though the tax is levied on and collected from firms, consumers
bear a cost as a result of the higher price they pay for gasoline at the pump. In our specific
example, the price to consumers rises by 35 cents in response to the 50-cent tax. As we will
explain in the following subsections, the exact amount by which the price to consumers
rises depends on the relative elasticities of the demand and (long-run) supply curves.
Second, after the long-run adjustment to the tax, firms once again make zero economic
profits and do not suffer continuing losses. (In the short run they do lose money, but the
losses are temporary.) After paying the tax, however, the firms net only $2.85 per gallon,
less than they received before the tax was levied, so someone must suffer a continuing loss
on the supply side of the market. But who? We can’t say specifically, but we know that it
will be the owners of inputs that are in less than perfectly elastic supply to the industry,
because the owners of these inputs will receive lower prices for the services they provide.
That is, as firms cut production and demand fewer inputs in response to the tax, prices for
inputs with upward-sloping supply curves will fall. Consequently, in the increasing-cost
case, the long-run burden of the tax is borne by consumers and certain input owners.
Who Bears the Burden of the Tax? Our analysis shows that excise taxes reduce the output of the taxed good and increase its
cost to consumers, but what determines how much output falls and how much the price to consumers rises? Asking how much the price to consumers rises is, of course, equivalent to
asking how much of the tax burden is borne by consumers and how much by input owners.
In our example we found that, in the long run, the price to consumers rose by 35 cents and
the price to producers fell by 15 cents (to $2.85). In such a case economists would say that
consumers bear 70 percent of the tax burden ($0.35 out of each $0.50 collected) and pro-
ducers (more precisely, owners of inputs that are in less than perfectly elastic supply) bear
the other 30 percent.
Tax Incidence: The Effect of Elasticity of Supply As we will see next, the quantitative effects on price and output depend on the elastici-
ties of demand and supply, which we can show by using the industry demand and sup-
ply curves. Let’s concentrate on the long-run effects. In Figure 10.5a, we examine how
the elasticity of supply affects the outcome. Suppose that the industry is a constant-cost industry and has the horizontal long-run supply curve shown as LS. An excise tax shifts the supply curve vertically upward to LS + T. With the demand curve D, the tax reduces output to Q2 and increases price to $3.50. In contrast to our earlier example, the price rises by the full amount of the per-unit tax, and firms continue to net $3.00 after remitting the
tax to the government. Why are the effects of the tax different for a constant-cost industry?
Recall that in the constant-cost case, when output is reduced, per-unit production costs do
not decline, so output will continue to fall until the price has risen by the amount of the tax;
only after this adjustment takes place will a new zero-profit equilibrium be achieved.
To illustrate the influence that different supply elasticities have on the distribution of the
tax burden between buyers and sellers, consider Figure 10.5a once again and suppose that at
the initial equilibrium point A the supply curve is less elastic, like LS′.5 A tax of 50 cents per
5It is important not to confuse slope and elasticity, but when we compare the elasticities of two curves at the same price–quantity combination, as we are doing here at point A, the steeper curve is the less elastic curve.
Excise Taxat ion 253
C10.INDD 12:9:21:PM 08/06/2014 PAGE 253Trim Size: 203.2 mm X 254 mm
unit shifts LS′ vertically upward, to LS′ + T. (Note that the vertical difference between each set of supply curves is the same; we are using the same tax per unit but varying the supply
elasticity to isolate the effect of different supply elasticities on price.) In this case, equilib-
rium output declines to ′Q2, the price to consumers rises to $3.15, and the net-of-tax price received by firms falls to $2.65. The consumer price does not rise by the full amount of the
tax because as output falls, per-unit production costs decline and firms can make zero eco-
nomic profits at a net-of-tax price lower than the initial $3.00 price. This also means that out-
put does not have to fall as much to restore long-run equilibrium as in the constant-cost case.
These results suggest the following generalization: for a given demand curve and tax
per unit, the more inelastic the supply curve, the smaller is the tax burden on consumers
(price rises by less), the larger is the tax burden on producers (meaning the relevant input
suppliers), and the smaller is the output reduction. The only significant exception is in the
unlikely case of a vertical demand curve. If the demand curve is vertical, the price to con-
sumers rises by the amount of the tax, regardless of the elasticity of the supply curve.
Tax Incidence: The Effect of Elasticity of Demand Now let’s see how the elasticity of demand affects the outcome. In this case, we work with a given supply curve and tax per unit and then vary the demand elasticity. Figure 10.5b illus-
trates this analysis. Initially, we have the supply curve LS and demand curve D, so the before- tax equilibrium price and quantity are $3.00 and Q1, respectively. The tax shifts the supply curve to LS + T, the price to consumers rises to $3.15, the net-of-tax price to producers falls to $2.65, and output declines to Q2. Alternatively, suppose that demand is less elastic at the initial equilibrium point A, as shown by the D′ curve. Then the price to consumers rises by more (to $3.35), the price to producers falls by less (to $2.85), and output falls by less (to ′Q2).
These results suggest the following generalization: for a given supply curve and tax per
unit, the more inelastic the demand curve, the greater the tax burden on consumers, the
Dollars per unit
Output
$3.50
$3.15
$3.00
$2.65
LS ′ + T
LS + T
0 Q2 Q2 Q1
(a)
LS ′
LS
D
A
′
Dollars per unit
Output
$3.35
$3.15
$3.00
$2.65
$2.85
LS + T
0 Q2 Q2 Q1
(b)
D ′
LS
D
A
′
How Elasticities Affect the Tax Burden (a) With the more elastic supply curve, LS, the price increase to consumers is greater. With a perfectly elastic supply curve, such as LS, the price rises by the amount of the tax. (b) With the more elastic demand curve, D, the price increase to consumers is smaller.
Figure 10.5
254 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 254Trim Size: 203.2 mm X 254 mm
smaller the tax burden on producers, and the smaller the reduction in output. The only sig-
nificant exception is in the case of a perfectly elastic supply curve—the constant-cost case.
In this case, the price to consumers rises by the amount of the tax per unit, regardless of
the elasticity of the demand curve.
In summary, the proportion of any excise tax borne by consumers depends on the rel-
ative sizes of the elasticities of demand and supply. These elasticities indicate how well
consumers and producers can “substitute away” from a tax. The higher the elasticity, the
greater the availability of substitutes in consumption or production, and the greater the abil-
ity to substitute away from the tax and impose the burden of the tax on the other party.
APPLICATION 10.1
We have seen how the extent to which the burden of an excise tax is borne by suppliers or consumers hinges on the relative sizes of the elasticities of demand and sup- ply. The more responsive consumers are to price than are suppliers, as indicated by the extent to which elasticity of demand is greater than elasticity of supply, the more easily consumers can run away from the effects of an excise tax on the price they end up paying—and consequently, the lower the percentage of the burden associated with the tax that consumers will bear relative to suppliers.
Conversely, the less responsive consumers are to price than are suppliers, as indicated by the extent to which elas- ticity of demand is smaller than elasticity of supply, the less easily consumers can run away from the effects of an excise tax on the price they end up paying. In such scenarios, con- sumers are relatively more “pinned down” to a market and will thereby end up getting stuck with more of the tax burden.
It turns out, moreover, that it doesn’t matter whether government policymakers end up levying the tax on sup- pliers (as has been assumed in the foregoing material) or on consumers. (We will explore this matter in Chapter 18, in the context of examining Social Security, a program financed through levies imposed partially on employees— suppliers of labor—and employers—consumers of labor services.) Perhaps surprisingly, only relative elasticities of demand and supply determine the extent of tax incidence rather than the relative amount of the tax that the govern- ment actually collects from suppliers or consumers.
An apocryphal story involving two hikers in Alaska can help you remember this central point. The story involves
Relative Ability to “Run” and Tax Incidence
the hikers spotting a grizzly bear on the horizon, which starts to move menacingly in their direction. The first reac- tion of both hikers is to start running away from the bear. After some running, however, one of the hikers suddenly stops, takes off her backpack, and starts removing her run- ning shoes from the backpack so that she can replace the much heavier boots that she was initially wearing. Her companion cannot believe the apparent foolishness of this move and yells back, “Why are you doing something so stupid! Don’t you know that even with running shoes, you cannot possibly outrun a grizzly bear [grizzlies can move very quickly notwithstanding their great size]?” The hiking partner takes him aback when she responds, “I don’t have to outrun the grizzly, I just have to outrun you!”
The analogy with tax incidence is straightforward. When an excise tax “grizzly” is introduced into a market, the extent to which the relevant parties to the market (suppliers and consumers) end up being harmed by the grizzly depends critically on the relative nimbleness (as measured by elas- ticities) with which they can run away from the adverse price effects imposed by the grizzly. When consumers are relatively better than are suppliers at running away from the danger, their expected harm will be relatively smaller. Con- versely, when suppliers can run away more quickly from the excise tax danger than can consumers, the expected harm to consumers will be greater. Fundamentally, when danger appears over the market’s horizon, what counts is not the nimbleness of either party (suppliers or consumers) with respect to the excise tax grizzly itself but the nimbleness of these two parties relative to each other.
The Deadweight Loss of Excise Taxation As explained in Section 10.1, any deviation in output from the competitive level is inef-
ficient and results in a deadweight loss. Because an excise tax results in an output that is
lower than in the unfettered competitive market, it produces a deadweight loss. Let’s con-
sider why in more detail. Originally, price is P and output is Q in Figure 10.6. An excise tax
Excise Taxat ion 255
C10.INDD 12:9:21:PM 08/06/2014 PAGE 255Trim Size: 203.2 mm X 254 mm
of T per unit causes the long-run supply curve to shift to LS + T, and the result is a reduction in quantity to Q1 and an increase in price to P1.
The easiest way to see that there is a deadweight loss is to consider the marginal ben-
efit and cost associated with the change in output from Q to Q1. When output declines, each unit no longer produced has a marginal benefit that varies from P to P1 to consum- ers. However, the cost saving associated with not producing each of these units is lower,
ranging from P to P2. Thus, there is a net loss associated with each of the units that is not produced (the marginal benefit sacrificed is greater than the marginal cost saved), and
the sum of these net losses is shown as the triangular area BEC. Area BEC is the dead- weight loss, and results from the excise tax because output is restricted to a level where
the product’s marginal benefit (P1) exceeds the marginal cost of production (P2); consum- ers would benefit from a higher output but the tax inhibits production of additional units
beyond Q1. In explaining the deadweight loss, note that we continue to interpret the original supply
curve as showing the marginal cost of production. From the firms’ point of view, marginal
cost is given by LS + T, but the tax is not a real cost to society. The tax simply trans- fers funds from market participants to the government; it does not reflect a use of scarce
resources in production. Consider that if output was reduced from Q to Q1 before the tax, there would have been a cost saving (shown by the LS curve) in the form of less labor, capi- tal, raw materials, and so on. When a tax leads to the same output reduction, the cost saving
in terms of the value of resources no longer employed in the market is identical. The tax is
not a real cost to society, but it does influence firms’ and consumers’ decisions and that is
why it produces a deadweight loss.
The deadweight loss of the tax can also be derived by relying on the concepts of con-
sumer and producer surplus. The tax increases the price to consumers from P to P1, thus decreasing consumer surplus by area P1BEP. In addition, the tax reduces the net price received by producers from P to P2, decreasing producer surplus by area PECP2. Combin- ing these two, we see that the reduction in consumer plus producer surplus equals area
P1BECP2. However, of this loss, area P1BCP2 is received as tax revenue by the government.
The Deadweight Loss of an Excise Tax The excise tax reduces consumer surplus by area R1 plus area X and producer surplus by area R2 plus area Y. Part of the loss is transferred to the government as tax revenue, shown as R1 plus R2. The excess of the loss over the gain to the government is the decrease in total surplus or the deadweight loss, the triangular area X plus Y.
E
D
C
X
Y
B
Dollars per unit
Output
P1
P2
R1
R2 P
LS + T
LS
0 Q1 Q
R1 + R2 = Transfer of consumer and producer surplus to government tax revenue
X + Y = Deadweight loss
Figure 10.6
256 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 256Trim Size: 203.2 mm X 254 mm
The tax revenue is also shown as the sum of rectangular areas R1 and R2, where R1 is the direct burden of the tax on consumers and R2 is the direct burden on producers.
Part of the loss in consumer plus producer surplus is a gain to the government in the
form of tax revenue. However, the tax revenue is smaller than the loss in consumer plus
producer surplus by the area BEC. Area BEC is the decrease in the total surplus (in this case, consumer plus producer plus government surplus) or the deadweight loss of the excise
tax. Consumers and producers bear a burden (the loss in consumer plus producer surplus)
that is greater than the revenue collected by the tax. That is why the deadweight loss is
sometimes referred to as an excess burden when it is produced by a tax. The excise tax produces a direct burden in the form of tax revenue collected and also
an additional, less obvious burden in the form of the deadweight loss (or excess burden).
To get an intuitive understanding of the deadweight loss as something different from the
burden of paying taxes, consider an excise tax that is so large that output falls to zero. In
that case, there is no tax revenue collected. Nonetheless, there is a deadweight loss from the
excise tax. What is the deadweight loss in this case? It is the total surplus that would have
been generated by the market absent the tax.
Our analysis should not be taken to imply that excise taxes should be avoided. All taxes
produce deadweight losses; they can’t feasibly be avoided. However, the deadweight loss
is still important because it tells us that if the public is to benefit from government expen-
ditures, more than a dollar in benefit has to be produced per dollar of government expen-
diture. For example, suppose that the gasoline excise tax revenue is $100 billion and the
deadweight loss of the tax is $40 billion. Only if the spending of the $100 billion in tax
revenue produces a benefit valued by the public at more than $140 billion does the spend-
ing compensate the public for all the costs of the excise tax.
excess burden another name for the deadweight loss produced by a tax
APPLICATION 10.2
By reducing output below the competitive equilibrium, rent control results in a deadweight loss. Employing the competitive model to compare the short- and long-run effects of rent control allows us to see how the deadweight loss associated with a government-imposed price ceiling depends on the supply elasticity.
Figure 10.7 contrasts the short- and long-run effects of rent control. In the absence of a rent ceiling, the com- petitive (short- and long-run) equilibrium is a rent of P and quantity of Q. The short-run supply curve, SS, is drawn as being relatively price inelastic. This is so because if a city unexpectedly imposes rent control today, the quantity of rental units tomorrow will be virtually unaffected by the price change. It takes time for reduced construction and maintenance to have their full effects on durable goods like dwelling units. The adverse effects on the supply side thus are relatively small in the short run. In response to a rent ceiling of Pc, for example, output decreases only a little bit in the short run to Q1 and the resulting deadweight loss is given by triangular area BEC. Area BEC is the difference between the marginal benefit (the height of the demand curve) and marginal cost (the height of the short-run supply curve) summed over each unit of output between Q1 and the competitive equilibrium of Q.
The Long and the Short (Run) of the Deadweight Loss of Rent Control
The long-run output adjustment to the rent ceiling is more substantial and begins as new construction falls and existing units are allowed to deteriorate. In the graph this is depicted by the long-run supply curve, LS, being more price elastic than the short-run supply curve, SS. In response to the rent ceiling, Pc, there is a long-run output reduction to Q2 and a deadweight loss equal to triangular area FEG. Area FEG is the difference between the marginal benefit (the height of the demand curve) and marginal cost (the height of the long-run supply curve) summed over each unit of output between Q2 and the initial competitive equilibrium of Q.
As a result, rent control’s deadweight loss is greater in the long run as suppliers have more time to reduce out- put in response to the price ceiling. All other things being equal, the greater the elasticity of supply and the larger the decline in output from the competitive equilibrium due to a price ceiling, the larger the deadweight loss. This suggests why elected local leaders may be less concerned than economists about rent control’s adverse effects. Since city government officials typically hold office for only a few years, they are not around long enough to experience the long-run output adjustment to a rent ceiling. The relevant perspective on rent control to many elected officials may
Air l ine Regulat ion and Deregulat ion 257
C10.INDD 12:9:21:PM 08/06/2014 PAGE 257Trim Size: 203.2 mm X 254 mm
10.3 Airline Regulation and Deregulation Previously we mentioned that the perfectly competitive model remains extremely useful
even if one or more of the conditions defining the model do not hold. In this section we
show how this is the case in the U.S. experience with airline regulation and deregulation.
During the period in which the domestic airline industry was regulated, entry into any given
city-pair market in the industry was restricted. Carriers operating in a particular market
were required to have an operating license from the Civil Aeronautics Board (CAB), and
the CAB issued no new licenses to airlines requesting to enter an already served market.
Even though the condition of free entry and exit was thus violated, the perfectly competi-
tive model still can be employed to analyze the effects of regulation, as well as deregula-
tion, on the airline industry.
From its formation in 1938 to the Congressional deregulation of the domestic airline
industry in 1978, the CAB controlled, among other things, fares, routes that commercial
airlines could serve, and entry of new firms into the industry. Analyzing the effects of
CAB-imposed regulations makes clear why there was widespread support for deregulation.
Our analysis will focus on the pricing policy followed by the CAB. The CAB closely
regulated the fares airlines charged, and, it turns out, kept those fares well above the level
that would have prevailed in an open market. Even when some airlines requested fare
reductions, they were regularly denied. In effect, the CAB imposed a price floor, keeping the price above the competitive level.
Throughout the regulation era, persuasive evidence existed that regulated airline fares
were artificially high. The CAB regulated only airlines engaged in interstate transporta-
tion; intrastate airlines were beyond its reach. The existence of unregulated airlines made
possible some illuminating comparisons. For instance, intrastate airlines operating in
be closer to the one represented by the short-run sup- ply curve in which the resulting deadweight loss is fairly small and the primary effect is a transfer of income from
Figure 10.7 Supply Elasticity and the Deadweight Loss of Rent Control With the rent ceiling Pc, the reduction in output in the short run is small, from Q to Q1, along the short-run supply curve SS. The associated deadweight loss, triangular area BEC, is also small. The long-run effects of the rent ceiling are more significant. Output declines to Q2 along the more elastic long- run supply curve LS. The associated deadweight loss is depicted by triangular area FEG.
E
F
D CG
B
Monthly rent per unit
Rental housing services
P
Pc
SS
LS
0 Q2 Q1 Q
landlords to tenants. If tenants have more political clout than landlords, local policymakers may find the income transfer well worth the (small) short-run deadweight loss.
258 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 258Trim Size: 203.2 mm X 254 mm
California flew the Los Angeles–San Francisco route, a distance approximately the same
as the interstate route between Washington, D.C., and Boston. The CAB-controlled fares
on the Washington–Boston route, though, were twice as high as the uncontrolled fares from
Los Angeles to San Francisco. In addition to actual price comparisons, the federal govern-
ment’s General Accounting Office estimated that, on average, fares were 22 to 52 percent
higher due to the CAB’s actions.
From this we might conclude that the CAB designed the regulations to help the airlines
at the expense of passengers. Airlines, after all, were receiving much higher fares as a result
of the CAB’s price-setting policy. Now, however, we encounter a startling fact: airlines
were not particularly profitable during the period of regulation. In fact, over the 20 years
prior to deregulation in 1978, the airline industry’s accounting profits were slightly below
the national average for all industries.
What Happened to the Profits? The apparent profits to airlines were dissipated in three ways. First, the CAB required air-
lines to operate some unprofitable routes. These routes generally provided service between
sparsely populated areas where demand was insufficient for the airline to make a profit. The
airlines had to balance the losses on these runs against the profits on other routes. Second,
airline worker unions were in a position to demand and get higher wages when the CAB
kept fares above competitive levels, and so some of the potential profits went to employees.
The third reason is perhaps the most interesting because it would, in theory, eliminate
profits even in the absence of the other two. It is nonprice competition. In any market where
prices are set and suppliers cannot compete on the basis of price, another form of competition
will emerge, as we saw with rent control in Chapter 2. What happened in the airline industry?
Airlines can make large profits at high prices only if they attract passengers. Under regu-
lation, however, they could not cut prices to attract passengers away from their competitors.
Each airline faced the problem of making itself more attractive than its competitors by some
other means than lower fares. The solution was obvious: change the nature of the product to
make it more appealing. So, airlines began scheduling more frequent flights so passengers
could fly at times convenient to them. In addition, competition evolved among airlines to
provide “frills”: gourmet meals, movies, more and better attendants, complimentary Mickey
Mouse ears for children flying to Disney World, and sometimes live entertainment. But all
these things increased the cost of providing transportation. Costs rose, prices were fixed,
and profits diminished. Indeed, economic theory would predict that in a competitive market
the process would continue until airlines no longer made a profit at all. And so we see why
the airline industry was not especially profitable despite the artificially high prices.
Let’s examine the process using graphs. Figure 10.8b shows the supply and demand
curves for airline services. For simplicity, we assume the industry is constant-cost. (Econ-
omists call this a simplifying assumption: the results are not significantly different from
the case of an increasing-cost industry, but the analysis is simpler.) In the absence of any
regulation, price and quantity are P and Q, respectively. Then, the CAB sets the price PCAB. Note that if the industry operated at point A on its supply curve, it would make a profit shown by the shaded area. Although this point is not the final outcome, it provides a con-
venient place to begin our analysis to understand why further adjustments must take place.
Corresponding to point A in Figure 10.8b, a representative firm is at point a in Figure 10.8a, operating at the minimum point on its LAC and making a profit equal to the shaded area.
With all firms making a comfortable profit, why can’t this point be an equilibrium? The
answer is that each airline can make still more money if it expands output by drawing pas-
sengers away from other airlines. This is because an individual airline’s profit-maximizing
output is where LMC equals PCAB. Because total quantity demanded is Q1 and price cannot be lowered, an airline can gain passengers only by attracting them away from other airlines
in some other way.
Air l ine Regulat ion and Deregulat ion 259
C10.INDD 12:9:21:PM 08/06/2014 PAGE 259Trim Size: 203.2 mm X 254 mm
Suppose that the airline attempts to attract passengers by scheduling more frequent
flights. Note that every airline has an incentive to initiate this practice, but more flights in
total, with an unchanged total quantity demanded, means fewer passengers on each flight.
Airlines will operate flights with empty seats, and to do so is in the interest of each airline.
Why? At a price per passenger twice as high (PCAB), it is profitable to schedule a flight even if only half the seats are filled. Flying half-filled planes, however, means a higher average
cost per passenger. Thus, in Figure 10.8a the representative airline’s LAC curve, showing the cost per unit of output, shifts upward as the number of passengers per flight declines.
This shift continues until economic profit is eliminated. The final result is an average cost
curve like LAC′, with the typical airline just covering its cost at the higher CAB price and with total output unchanged.6
After Deregulation Since the domestic airline industry was deregulated, several significant changes have taken
place. First, as our analysis would suggest, the cost of air travel to consumers has fallen. In
real terms, airline ticket prices are roughly 50 percent lower today than they were at the time of
deregulation. In part because of the fare reductions, the number of passengers flown has more
than doubled since 1977. And the annual value of consumer surplus generated by deregulation
is estimated to be roughly $29 billion (2013 dollars).7
6This analysis neglects the way the quality change affects the demand curve. Presumably, more convenient flights with “frills” are worth more to consumers, so this practice shifts the demand curve rightward to some degree. Total industry output is then somewhat greater, but airlines still end up making zero economic profit. 7Clifford Winston and Steven Morrison, The Economic Effects of Airline Deregulation (Washington, D.C.: Brookings Institution, 1997); and “Airline Deregulation: The Concise Encyclopedia of Economics,” www.econlib.org.
D
LAC ′
LAC
a
LMC
Dollars per unit
United Airlines
Output
PCAB
P
0 q1
A S
Dollars per unit
Market
Output
PCAB
P
0 Q1 Q
(b)(a)
Airline Regulation by the CAB A price floor of PCAB implies that a representative firm (a) and the industry (b) can earn profits shown by the shaded areas. However, the profits are dissipated through nonprice competition, which leads to cost curves shifting upward (LAC to LAC′).
Figure 10.8
260 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 260Trim Size: 203.2 mm X 254 mm
Second, a major restructuring of the industry has taken place since deregulation. For 40
years the CAB denied access to would-be entrants. Just within a year of deregulation the
number of airlines in interstate service rose from 36 to 98. The rapid expansion has been a
mixed blessing, however. Since 1978, for example, the overall industry has tallied losses—
including a record $40 billion between 2001 and 2005. Many established carriers, including
Braniff, Pan Am, American, United, Delta, US Airway, Northwest, and Eastern, have declared
bankruptcy (some of these have been able to reemerge from bankruptcy). A few other finan-
cially troubled airlines, such as TWA, have been absorbed by stronger competitors.
Third, after deregulation many of the industry’s new entrants such as Southwest Airlines
operated at significantly lower costs than the established carriers. One reason for the cost
differential was the union pay scales negotiated during regulation. As we mentioned ear-
lier, potential profits from higher-than-competitive fares can be dissipated by paying above-
market wages to union members. Between 1970 and 1978, for example, the consumer price
index increased by 68 percent, yet airline workers’ wages rose by over 100 percent and
fringe benefits increased by 300 percent. After deregulation the contrast in salaries paid by
established airlines and newcomers was striking. In 1983, the average worker (typically
union) at the established airlines made $39,000 a year, whereas workers (typically non-
union) at new airlines made an average of $22,000. Top (union) pilots received $150,000
per year from the established airlines, whereas the new entrants to the market paid their
pilots only $45,000.
In an attempt to reduce labor costs in the deregulated environment, the established car-
riers have cut their workforces by the thousands and implemented a two-tier wage scale
that pays new employees 30 to 50 percent less than existing ones. Already, labor costs as
a share of all expenses have fallen from 39 percent in 1979 to 33 percent today. What this
and other evidence makes clear is that unions were major beneficiaries of the CAB regula-
tions and major losers from deregulation.
Fourth, service to small communities has, on average, increased with deregulation,
but fares have gone up. Prior to deregulation, the CAB required airlines to provide ser-
vice to small cities, but because the carriers used the same jets to serve small cities, they
lost money on the small-city routes. After deregulation, the large carriers abandoned these
unprofitable routes, but new commuter airlines using smaller, more fuel-efficient planes
(e.g., turbo props and regional jets) have taken their place. In effect, the CAB regulations
required airlines to use some of their potential profits to subsidize service to smaller com-
munities, and deregulation ended the implicit subsidy.
APPLICATION 10.3
Although deregulation has lowered fares, a question that has emerged—particularly in the wake of several recent
mergers and bankruptcies—is, how many airlines are necessary in any city-pair market to ensure the competitive outcome? The number of existing suppliers may be irrelevant if airline markets are what economists term contest- able. Contestable markets are those in which competition is so perfect that the market price is
The Contestability of Airline Markets
independent of the number of firms currently serving a market, because the mere possibility of entry suffices to dis- cipline the actions of incumbent suppliers. (Remember that in perfect competition, entry and exit are assumed to be frictionless so that incumbent suppliers must be wary of potential entry if they charge a price in excess of marginal cost.) For example, even though Delta Airlines may be the sole provider on the Atlanta–Birmingham (Alabama) air route, it will charge fares equal to marginal cost if the route is contestable. If Delta charged fares above marginal cost, other airlines could costlessly (i.e., without friction) move into the market and take away Delta’s passengers.
contestable markets markets in which competition is so perfect that the market price is independent of the number of firms currently serving a market, because the mere possibility of entry suffices to discipline the actions of incumbent suppliers
Air l ine Regulat ion and Deregulat ion 261
C10.INDD 12:9:21:PM 08/06/2014 PAGE 261Trim Size: 203.2 mm X 254 mm
The Push for Reregulation Not everyone is happy with the results of airline deregulation, and in recent years propos-
als to reinstitute regulation have surfaced. Some of the support for reregulation is under-
standable since it comes from groups that have suffered financially, such as the established
airlines and their employees. Some, however, comes from groups concerned with other
questions they believe to be associated with deregulation—namely, greater congestion at
airports and issues of airline safety.
Not surprisingly, airports have been more congested following the surge in passenger
traffic after deregulation, and with the congestion have come more delayed flights, lost
luggage, and service complaints. Because deregulation has increased the number of pas-
sengers, it has contributed to these problems. But is reregulation the best way to deal with
them? Let’s consider some alternatives. Airport capacity, for example, could be expanded
to handle the increased traffic. Indeed, in a fully competitive market, this would occur
automatically, but it hasn’t because airlines don’t control airports. Airports are owned
and operated primarily by local governments, and since deregulation there has been
little expansion in airport capacity despite the fact that nearly twice as many passengers
are using them. Part of the problem with expanding capacity is that the expansion often
generates political opposition from zoning authorities, environmental groups, nearby
Although competition may be vigorous under deregula- tion, the available evidence suggests that it is not perfect and that airline markets are not contestable in the strict sense of the term. For example, holding other factors con- stant, it has been estimated that average fares on routes with two competitors are 8 percent lower than the fares on routes served by a single airline.8 A third competitor is asso- ciated with an additional decline in fares of 8 percent.
If domestic airline markets are indeed contestable, this relationship between observed average fares and the number of airlines operating on any given route will not prevail. The average fares on any given route will be insen- sitive to the number of airlines operating on that route if domestic airline markets are contestable since incumbent airlines’ pricing strategies will be affected as much by the mere possibility of entry as they are by the actual entry of rival firms.
The fact that average fares decrease as the number of air- lines flying a route increases suggests the existence of some frictions to entry and exit—some costs that impede the ability of potential competitors to constrain the pricing and profits of incumbent firms. These costs may consist of the advertising that must be done to announce one’s arrival in a market, gate space that must be acquired and modified to an entering airline’s specifications, and local staff that must be hired and trained with regard to company procedures.
Finally, the evidence suggests that there is an incentive in deregulated markets for additional entry—entry that
will benefit consumers through lower prices. Over the past two decades, companies such as Southwest Airlines, Jet Blue, Allegiant, and Spirit have made substan- tial inroads against more established firms because of their lower costs. For example, Southwest has achieved significantly lower operating costs than more established competitors by focusing on short-haul markets where the same type of fuel-efficient plane, a Boeing 737, can be used (thus lowering fuel, maintenance, and train- ing costs); meals don’t have to be served; expensive hubs are unnecessary; and planes can be turned around quickly (in 30 minutes versus an hour or more for carri- ers such as United, American, and Delta, whose planes must sit on the ground at their hubs to await connecting passengers).
Lower costs have allowed Southwest and Jet Blue to target underserved cities and offer frequent, low-fare service, often chasing out incumbents along the way. For example, in contrast to the existing one-way coach fare of $300 in the Los Angeles–San Francisco market, Southwest introduced a $20 fare between Oakland and Burbank in 1990, virtually guaranteeing the departure of American Airlines from the market. Prior to Jet Blue’s arrival in 2000 with a $120 round-trip fare in the Rochester (New York)—New York City market, the prevailing fare charged by established carriers such as US Airways was roughly $300 per round-trip. The success of carriers such as Jet Blue and Southwest even prompted both Delta and United Airlines to internally spin off low-cost versions of their own airlines—Song in the case of Delta and Ted (half of “United”) in the case of United.
8Severin Borenstein, “The Evolution of U.S. Airline Competition,” Journal of Economic Perspectives, 6, No. 2 (Spring 1992), pp. 45–73.
262 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 262Trim Size: 203.2 mm X 254 mm
residents, and perhaps the dominant airlines at the existing airport (which fear an increase
in competition if capacity is expanded).
But even if local governments are unable or unwilling to expand airport capacity,
there are other ways to deal with congestion problems. For instance, flight delays are as
much a consequence of how airports manage landings and departures as they are a result
of increased passenger service. Airlines schedule flights when consumers want them most
(early morning, midday, and late afternoon), and these are the times when congestion is the
worst. The amount of air traffic during peak hours can differ from the nonpeak hours by a
factor of 10. Currently, the landing and takeoff fees charged to airlines by airports are usu-
ally the same regardless of whether they occur during peak or nonpeak hours. Economists
would suggest increasing the peak-hour fees. Such a price differential would induce airlines
and passengers to rearrange their schedules and reduce congestion during peak travel hours.
(This technique is called peak-load pricing and is discussed more fully in Chapter 12.) Congestion is also related to airline safety, because the probability of accidents increases
with the number of airplanes in the sky at a given time. As we have seen, however, con-
gestion is not entirely the result of deregulation, but results in part from the way airports
are managed and priced by local governments. Nonetheless, the news media often suggest
a possible connection between safety and deregulation whenever reporting an accident or
near-miss. Actually, airline safety procedures are still regulated by the Federal Aviation
Administration (FAA), just as they were before deregulation. If safety problems exist, it is
not necessary to control air fares and routes to deal with them—the FAA could simply raise
safety standards. In fact, some economists argue that the FAA isn’t needed at all because
airlines have an incentive to take safety into account due to the enormous costs they bear in
the event of an accident.
The heightened concern over safety in recent years is somewhat puzzling because the
data show clearly that safety has improved in the years following deregulation—with the
notable exception of 2001, when terrorists commandeered and crashed four commercial
jetliners (acts that could have just as easily occurred in a regulated environment). Accord-
ing to the Department of Transportation, both fatality and accident rates were lower in the
years after deregulation than in the years before—even taking into account the tragic events
of September 2001. Another indication that air travel is at least as safe after deregulation is
that the companies that stand to lose the most financially from crashes—insurance compa-
nies that insure airline carriers—consistently lowered the rates they charge the major carri-
ers up until the fall of 2001, when substantial war and terrorism clauses had to be added to
most policies.
10.4 City Taxicab Markets Most major U.S. cities regulate taxis in some way. Usually, the regulations require taxis
to have city-issued licenses. The licenses look like and are often called medallions. Typi- cally, cities issue a fixed number of medallions, and new entrants must purchase one from a
current driver or cab company. The supply of cabs thus is limited by the number of issued
medallions. The effects of such an entry restriction can be analyzed through the perfectly
competitive model—much as the model allowed us to analyze the impact of domestic air-
line regulations limiting entry into individual city-pair markets.
To examine the operation of a taxicab market with a restricted number of licenses, con-
sider Figure 10.9, which shows the supply and demand curves—S and D, respectively—for taxi service absent any regulation. The supply curve is drawn as a horizontal line—a con-
stant-cost industry. Although we make this assumption partly to simplify the analysis, it is
reasonable to assume that the supply of taxi services within a city is highly, if not perfectly,
elastic. With the market conditions as illustrated, competition results in a per-mile fare
City Taxicab Markets 263
C10.INDD 12:9:21:PM 08/06/2014 PAGE 263Trim Size: 203.2 mm X 254 mm
(price) of $1.00 and an annual output of 9 million passenger-miles. Suppose that this output
is produced by 600 cabs, each of which transports passengers for 15,000 miles per year.
Next, suppose that the city requires taxis to have a medallion to operate, and it issues
600 medallions, giving one to each taxi. The medallions change nothing and have no effect
on the market as long as supply and demand conditions remain the same. Now suppose
that over time the city’s population grows and incomes rise, so demand shifts to D′. (The supply curve may also shift over time, but for simplicity we assume that it does not.)
Because the number of taxis is limited by the number of medallions, the supply curve is
vertical at 9 million passenger-miles, as shown by S′. (We are assuming that each taxi con- tinues to operate for the same number of passenger-miles as before.) Fares rise to $1.20 per
mile while output remains unchanged. In contrast, if the market had not been regulated,
output would expand along S to 12 million passenger-miles and the per-mile price would remain at $1.00.
The practice of licensing taxicabs restricts entry to the market and leads to higher con-
sumer prices. Persons on the supply side of the market tend to benefit, but exactly who they
are and how they benefit deserves some further discussion. Note that in the situation just
described, operating a taxi for a year (15,000 miles) generates a profit of $0.20 per mile, or
an annual profit of $3,000. To realize that profit, however, a person must own a medallion,
so it is the owners of medallions who tend to benefit from licensing. More precisely, those
who were originally given the medallions receive most, if not all, of the benefit. Current
owners of medallions may not receive anything but competitive returns.
To see why, let’s consider the factors determining a medallion’s price. Suppose that a
driver from another city wishes to purchase a medallion from a driver who owns one. (If
you owned a medallion, what would you sell it for?) Because ownership of the medallion
brings with it an annual gain of $3,000 (assuming market conditions remain unchanged),
its value is the same as an asset yielding $3,000 per year. If the prevailing interest rate is
10 percent, a person needs a $30,000 investment to generate an income of $3,000 per year.
Thus, the value of the medallion itself is $30,000. Buyers are willing to pay this amount
because they could get as good a return by investing $30,000 in a medallion as by investing
it where it yields an interest income of $3,000 per year.
The transferability of medallions at prices determined by the expected profitability
of operating a taxi means that those who received the medallions free when they were
first issued may get all the benefit from the licensing policy. To take an extreme example,
Licensing Taxicabs Taxicab licensing makes the supply curve vertical (under our assumptions) at an output of 9 million passenger-miles. The result is higher fares and lower output than under unregulated conditions.
D
Fare per mile
Passenger-miles driven per year
$1.20
$1.00
0 9 million 12 million
D ′
S
S ′Figure 10.9
264 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 264Trim Size: 203.2 mm X 254 mm
suppose that all the original cab operators sell their medallions to others. If this occurs,
those who benefited from the licensing policy are no longer in the market, and those now
operating taxis receive only a normal return on their investment. Nonetheless, the fare is
still above the real cost of producing taxi services; if the city allowed free entry into the
market, cab fares would fall to $1.00, as indicated by the intersection of S and D′, and the value of the medallions would fall to zero. Such a deregulation of cab markets would
impose a large loss on the current medallion owners—$30,000 per medallion in our exam-
ple—yet these cabdrivers may never have received any benefit from the restrictive policy.
A difficult ethical dilemma is implicit in this analysis: is it fair to drivers to deregulate the
market—that is, to allow unrestricted entry, lower prices, and higher output? Deregulation
would impose a large loss on medallion owners, some of whom may never have shared in
the economic profits created by the licensing policy. For example, as of 2013 the price of a
medallion in New York City was $1.1 million, so people who had invested their life savings
or borrowed money to purchase a medallion would be devastated by a return to unrestricted
entry, because that would leave them with no opportunity to earn the higher income that
made the medallion worth the $1.1 million they paid for it. If entry remains restricted, how-
ever, consumers continue to pay fares that are higher than necessary. As a further complica-
tion, the poor tend to be heavy users of taxis; they spend a larger portion of their incomes on
taxi services than do higher-income groups, so they are especially burdened by higher fares.
We have emphasized the effects of the entry restriction created by taxicab licensing, but
cities often regulate this market in other ways besides licensing. For instance, sometimes
maximum fares are specified. This price ceiling creates shortages in certain parts of the city
and at certain times of the day. For example, suppose that the maximum fare is $1.20 in
Figure 10.9. Although that price might have no effect in the middle of the day, it can create
a shortage during rush hour. At rush hour, congestion is greater and costs of operation are
higher (traffic moves more slowly, and it takes longer for a cab to travel a mile). Demand is
also likely to be higher. If the quantity supplied is lower and the quantity demanded higher,
and if a higher price is not permitted, a shortage results. Anyone who has tried to find a cab
in New York City during rush hour (or on a rainy day) knows what this means. Similarly,
fare regulations can discourage cabs from operating in more remote or dangerous areas of a
city. Because the regulated fares are not sufficiently high to cover the risk, drivers practice
a form of nonprice rationing by avoiding passengers traveling to high-risk neighborhoods.
Competition: Legal and Illegal Licensing and fare regulations have given rise to competition from both legal (e.g., Uber)
and illegal transportation services. Such competition arises because of the mutual gains
possible for the various participants. As shown in Figure 10.9, taxi services can be profit-
ably supplied at prices below $1.20 per mile (as shown by the horizontal supply curve S, which continues to show the cost of provision exclusive of the artificial medallion cost),
and consumers also gain from a lower price.
Legal forms of competition include Uber, a service linking prospective ride-sharers
through their smartphones. As of 2014, the privately-held company was valued at $18
billion and operated in 100 cities and 36 countries (although not without protests from
licensed cab drivers). The Uber app tells prospective riders how long they will wait and
monitors drivers’ safety and performance records. Uber rates average 30 percent lower than
licensed taxi fares. Payment is automated: when leaving the car, Uber charges a rider’s
credit card automatically while keeping 20 percent of the total cost as a commission—the
rest is remitted to the driver.
Uber and similar technology-based services are the modern successors to illegal forms
of competition that historically have sprung up in cities licensing taxi cabs. Sometimes
called jitneys or gypsies, these illegal, taxi service suppliers are prevalent. One study, for example, found more than twice as many illegal cabs operating in Pittsburgh as there
City Taxicab Markets 265
C10.INDD 12:9:21:PM 08/06/2014 PAGE 265Trim Size: 203.2 mm X 254 mm
are licensed cabs.9 In New York City, there are 25,000 gypsies compared with 13,000
licensed cabs (City taxi inspectors seized 6,000 gypsy cabs in 2012). The unlicensed cabs
provide service in areas, such as low-income neighborhoods, where the licensed cabs
often do not operate. In addition, fares are lower for the unlicensed cabs. For instance,
licensed cabs charged 25 percent more for a trip of a given distance than did unlicensed
cabs. Tips are uncommon in the illegal market, which makes the effective price spread
greater. Moreover, the service quality provided by the unlicensed cabs tends to be better.
One concern about the illegal market has always been that it might be unsafe. For
example, a paid advertisement in the New York Times warned riders: “Ignore unlicensed cabs. You may be putting yourself in the care of a murderer, a thief or even a rapist.”10 The
Pittsburgh study, however, finds the warning unwarranted. Much of the illegal market is
organized around station houses, which act as central clearinghouses in the neighborhoods.
The study notes, “It is apparent that station managers work hard to please customers by
having courteous, safe, and reliable drivers who will not cheat on fares.” In addition, there
was no difference in traffic accidents between the licensed and the unlicensed drivers.
9Otto A. Davis and Norman J. Johnson, “The Jitneys: A Study of Grassroots Capitalism,” Journal of Contemporary Studies, 7 (Winter 1984), pp. 81–102. 10Ibid., p. 97.
APPLICATION 10.4
In 2012, New York City’s Taxi and Limousine Commission considered raising taxi fares by 20 percent (the sixth time fares would have been increased over the last 30 years).11
At a hearing on the proposed rate increase, passengers claimed that fares were already too high while cabdrivers pled poverty. Paradoxically, both groups were right and the analysis of taxi cab markets that we have just undertaken helps us to understand why.
Fewer than 18 percent of all cabs are owner-operated in New York City at the present time. The trend toward separating medallion ownership from cab driving was set in motion in 1979 when the New York Taxi and Limou- sine Commission allowed medallions to be leased out for 12-hour shifts.
The split between medallion ownership and cab oper- ation has been the key reason why drivers’ real incomes have not increased appreciably in the wake of fare increases over the last three decades. Rather, the owners of the scarce input, the medallions, have benefited from any fare increases at the expense of passengers. With fare hikes, medallion owners simply raise the rates at which they lease out their scarce rights to operate a cab. Com- petition between a relatively large number of potential
Why New York City Cab Drivers Are Poor and Drive So Fast
drivers ensures that the lease rate is driven up sufficiently to eliminate any profits from driving a cab when fares are raised.
The rate to lease a 12-hour cab operating shift in New York City has approached $130 as of late while driv- ers’ take-home pay, above the lease cost, has slipped to less than $98 per shift as of 2012 (versus $158 in 2006). Given that the take-home pay for drivers depends on how many fares they are able to pick up during their shift, the present medallion with lease option system explains why cabbies; drive so fast; may tell rush-hour Manhattan passengers that they don’t know how to get to a location in the adjoining borough of Queens; and perhaps even why they honk so much and become otherwise upset when stuck in traffic.
In contrast to cab drivers, medallion owners have fared quite well over the past decade as the going price of a medallion has risen from $200,000 to $1.1 million. Part of the increase in medallion prices is due to medallion own- ers becoming a more sophisticated and influential lobbying group in favor of their highly concentrated interests and at the expense of the more diffuse interests of taxi passen- gers. In 2011, for example, medallion owners successfully blocked New York’s attempt to shift to hybrid taxis. In 2012, medallion owners won a legal battle halting then Mayor Bloomberg’s efforts to initiate a fleet of “green cabs” to serve New York City’s outer boroughs.
11This application is based on “The Taxi Medallion System in New York and Other Cities Raises Fares, Impoverishes Drivers, and Hurts Passengers. So Why Can’t We Get Rid of It?” Slate.com, June 6, 2012.
266 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 266Trim Size: 203.2 mm X 254 mm
10.5 Consumer and Producer Surplus, and the Net Gains from Trade
So far, our look at the way markets work has not included international trade. However,
it is a simple matter to incorporate such trade into the analysis. There are two important
reasons to examine international trade. First, the economies of nations are becoming more
interdependent. In 1960, for example, foreign trade accounted for less than 5 percent of
U.S. gross national product; today the proportion exceeds 14 percent. Second, there are
constant claims made in the political arena and the media that free trade is harmful to a
nation’s welfare. Microeconomic analysis can help us analyze the validity of these claims.
International trade arises because sellers and buyers located in different countries find it
in their interests to deal with one another. To see what effects this has, consider Figure 10.10.
In the left-hand panel we show the U.S. demand and supply curves for sugar as DUS and SUS, respectively. We assume that there is only one other country in the world, called “Rest of the World” (or RoW; this can be an aggregation of a number of other countries, of
course). The right-hand panel shows the demand and supply curves for sugar in RoW as DR and SR, respectively. Without trade, each national market would attain equilibrium where its own demand and supply curves intersect. The result would be a price of P3 in the United States and P1 in RoW.
Now think about the incentives of buyers and sellers if trade becomes possible. Consum-
ers in the United States have an incentive to buy sugar from RoW sellers, who are selling it
United States
U.S. output and consumption
0
P3
SUS
ST
DUS
E C
P2
q2q1
DR
Cents per pound
SR
A B
P1
P3
P2
P1
Rest of the World (RoW)
(a) (b)
RoW output and consumption
0 Q2Q1
Cents per pound
International Trade The total supply curve of output available for sale in the United States is derived by adding to SUS an amount equal to the difference between quantity demanded and supplied in RoW at each possible price; the result is ST. The intersection of DUS and ST determines price and total consumption in the United States; U.S. production is less than consumption by the amount of imports (q2 minus q1).
Figure 10.10
Consumer and Producer Surplus , and the Net Gains f rom Trade 267
C10.INDD 12:9:21:PM 08/06/2014 PAGE 267Trim Size: 203.2 mm X 254 mm
at a lower price than U.S. sellers. Similarly, RoW sellers have an incentive to sell sugar in
the United States, where the price is higher than they are getting in their domestic market.
Therefore, we expect RoW producers to export sugar to the United States, and our task is
to determine how much trade occurs and how it affects both markets. (Note that if the cost
of transporting sugar is higher than P3 minus P1 per unit, there would be no trade. We will assume transportation costs are negligible.)
We will proceed by deriving the total supply curve, ST, confronting U.S. consum- ers; this curve will show the total amount of sugar (from both U.S. and RoW producers
together) that will be available in the United States at each possible price. Suppose that the
price is P3; how much will RoW producers offer to sell in the United States at that price? Figure 10.10b gives us the answer. At a price of P3, RoW producers will produce at point A on their supply curve, a total amount of P3A. Of this total, they will sell P3B to consumers in RoW; the remainder, BA, is the amount they will export to the United States. The reason they sell P3B to RoW consumers is that if they sold less, the price in RoW would be higher than P3 and they would have an incentive to shift sales from the export market to their domestic market. Similarly, if they were selling more than P3B to RoW consumers, the domestic price would be lower and they would have an incentive to shift sales to the export
market. Thus, when the price is P3, RoW producers will add BA to the output U.S. produc- ers place on the market. Distance EC in Figure 10.10a is drawn equal to BA, so total output in the U.S. market is shown by the sum of domestic output, P3E, and imports, EC, or P3C. Point C is one point on the total supply curve, ST, confronting U.S. consumers.
The remainder of ST is derived in a similar fashion. At each possible price, we add to U.S. output (shown by SUS) the difference between the quantity that RoW producers would choose to sell at that price and the amount RoW buyers would purchase. The resulting
curve ST shows the total quantity available on the U.S. market at alternative prices. Note that this curve intersects the U.S. supply curve at P1, the price that would prevail in RoW in the absence of trade. This shows that if the U.S. price was the same as the RoW price in
the absence of trade, there would be no exports from RoW to the United States. At a price
lower than P1, U.S. producers would become exporters of sugar. Equilibrium can be identified by the intersection of ST and DUS in Figure 10.10a. Thus,
the U.S. price is P2, and at that price U.S. consumers purchase q2 units. However, U.S. producers are providing only q1 units to U.S. consumers; the remaining q1q2 units represent imports. Trade lowers the price of all units sold in the United States because U.S. produc-
ers cannot sell sugar at a higher price than consumers can purchase it from RoW, provided
there is no difference in quality in the sugar produced by RoW and U.S. firms.
An important implication of this analysis is that both the U.S. and RoW markets must be
in equilibrium. This is guaranteed by the way we constructed the ST curve. Observe what has happened in RoW in Figure 10.10b. The price is P2 in RoW also. At that price, total output of RoW producers is Q2, but consumption by RoW consumers is only Q1. The differ- ence, Q1Q2, is the amount exported to the United States. Q1Q2 in panel (b) is equal to q1q2 in panel (a); the amount that RoW producers want to export at P2 is equal to the amount that U.S. consumers want to purchase as imports. Both markets are in equilibrium at a
price of P2, with production, consumption, and trade as indicated in the graphs. Note also that both markets can be in equilibrium only when the price is the same in both markets,
because otherwise sellers in the lower-priced market would have an incentive to export to
the higher-priced market.
This is a versatile model, but it must be handled carefully. For example, any change in
demand or supply conditions in RoW will cause the ST curve to shift. If demand increases in RoW, for instance, ST will shift to the left and the price of sugar in the United States will increase. Similarly, changes in U.S. supply or demand conditions will result in changes
in the uniform world price, and that will cause changes in RoW output, consumption, and
exports. When nations are linked by international trade, their markets are affected by one
another in this way.
268 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 268Trim Size: 203.2 mm X 254 mm
The Gains from International Trade In the political arena and the media, we frequently hear assertions that trade is harmful to
national welfare. We can use our supply–demand model to show that these assertions are
not well founded. Figure 10.11, based on the same supply and demand curves as Figure
10.10, illustrates the argument. The United States would be at point E in Figure 10.11a in the absence of trade. With trade, the price is reduced to P2 and consumption increases, but domestic production falls. From the graph we see that consumer surplus increases by area
P3EKP2—a trapezoid defined by the difference between the no-trade and with-trade price of sugar, P3 minus P2, from the vertical axis out to the domestic demand curve, DUS. Producer surplus falls by area P3ELP2—the difference between the no-trade and the with-trade price from the vertical axis out to the domestic supply curve, SUS. The gain in consumer surplus from the lower price exceeds the loss in producer surplus by area EKL. This area measures the net gain to the United States as a whole.
To say that there is a “net gain to the United States as a whole” does not mean, of
course, that every U.S. citizen gains. Citizens involved in sugar production lose. That is
why producer groups are among the most prominent supporters of trade restrictions. The
gain to consumers is, however, larger than the loss to producers in the sense that consum-
ers could fully compensate all those who lose on the supply side and still come out ahead
(by area EKL). The fact that consumers could in principle pay compensation and still benefit is what the net gain from trade (in the case of an imported good) is all about. A
fitting analogy involves any case of technological progress. For example, would an inno-
vation such as the discovery of a low-cost cure for the common cold be a net gain for our
country? The answer is yes in the same sense that free trade is a net gain for the United
United States
U.S. output and consumption
0
P3
SUS
ST
DUS
E
(a) (b)
KL P2
DR
Cents per pound
SR
G
F
H P2
P1
Rest of the World (RoW)
RoW output and consumption
0
Cents per pound
The Gains from Free Trade (a) Trade increases consumer surplus in the United States by area P3EKP2 and decreases producer surplus by area P3ELP2. The difference, area EKL, is the net gain to the United States from trade. (b) For RoW, the net gain from trade is shown by area HGF.
Figure 10.11
Consumer and Producer Surplus , and the Net Gains f rom Trade 269
C10.INDD 12:9:21:PM 08/06/2014 PAGE 269Trim Size: 203.2 mm X 254 mm
States: the cold cure would not benefit everyone (producers of aspirin and cold “remedies”
would lose), but the gain in consumer surplus would exceed those losses.
Now let us turn to how trade affects our trading partner, RoW. This is shown in
Figure 10.11b. RoW moves from its no-trade equilibrium at point F to a higher price of P2, with production expanding and consumption falling. The gain in producer surplus is area
P2GFP1—the difference between the with-trade and the no-trade price in RoW, P2 minus P1, from the vertical axis out to RoW’s supply curve, SR. The loss in consumer surplus is area P2HFP1—the difference between the with-trade and the no-trade price in RoW from the vertical axis out to RoW’s demand curve, DR. The gain in producer surplus exceeds the loss in consumer surplus by area HGF. This area measures the net gain to RoW as a whole from trade in the same way that area EKL in Figure 10.11a measures the net gain for the United States. For RoW, however, producers gain from trade, while in the United States
consumers gain because in this case, RoW exports to the United States. In other industries,
market conditions will result in the United States being an exporter. In that case, the effect
on the U.S. market will be just like the effect on RoW in Figure 10.11b: trade will result in
a gain in producer surplus that is larger than the loss in consumer surplus.
The Link between Imports and Exports Our analysis so far shows why both imports and exports make nations better off, on net, if
supply and demand curves determine equilibrium prices and quantities. If the rest of the
world is willing to supply a good at a lower price than would prevail without trade in the
United States, our country as a whole is better off. And if the rest of the world is willing to
pay a higher price than would prevail in our domestic market without trade, our net national
well-being is also enhanced.
Imports and exports, of course, are not independent of one another: any time we import
the rest of the world’s goods, the dollars used to pay international suppliers for those goods
must come back, by necessity, to our economy in the form of international demand for U.S.
exports. To see why, suppose that the opposite was true—namely, that the dollars used to
pay for U.S. imports never came back. Suppose, for example, that foreigners either stuffed
the U.S. import dollars under their mattresses or, worse yet, burned them. If this was the
case, we could run a tremendous scam on the rest of the world by simply printing up more
pieces of paper, calling them dollars, and using them to import valuable goods such as Jap-
anese cars and oil from Saudi Arabia, for free.
The only reason the rest of the world is willing to accept dollars as payment for the goods
we import from them is that those pieces of paper are worth something in return. Specifi-
cally, the dollars foreigners earn on U.S. imports can be used to purchase U.S. goods.
Contrary to politicians’ statements about how dollars spent on foreign products are “lost”
to the United States, free trade causes no such dissipation of currency. Currency is neither
created nor destroyed in the process. The only thing that is created when other countries’
products are purchased is a net gain to the United States as a whole—a net gain not only
from the imports themselves but also from the exports that those imports serve to promote.
The difference between the political and economic views of trade is aptly summarized
by Milton Friedman:12
. . . public opinion [on foreign trade overemphasizes] the visible [versus the]... invisible effects of government policy . . . steelworkers whose jobs are threatened by imports from Japan are highly visible. They . . . can see clearly the benefit to them from restricting [steel] imports . . . . The cost is large but spread thinly. Tens of thousands of buyers of objects made with steel would pay a bit more because of the restriction. The Japanese would earn fewer dollars here and, as a result, purchase
12Milton Friedman, “In Defense of Dumping,” Newsweek, February 20, 1978.
270 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 270Trim Size: 203.2 mm X 254 mm
fewer U.S. goods. But that cost too is invisible. The man who might have had a job producing a product the Japanese would have purchased if they had been permitted to sell more steel here will have no way of knowing that he has been hurt.
Workers . . . [producing] products . . . sold to Japan to earn the yen used to buy Japanese steel are producing steel for the U.S. just as much as the men who tend the open-hearth furnaces in Gary [Indiana] . . . . We could produce bananas in hothouses, and no doubt would if the tariff on bananas was high enough. Would that make sense? Obviously not—we can produce them more efficiently indirectly by trading export goods for bananas from Central America.
As Adam Smith [noted in] The Wealth of Nations: “What is prudence in the conduct of every private family, can scarce be folly in that of a great Kingdom. If a foreign country can supply us with a commodity cheaper than we ourselves can make it, better buy it of them with some part of the produce of our own industry, employed in a way in which we have some advantage. The general industry of the country . . . will not thereby be diminished . . . but only left to find out the way in which it can be employed to the greatest advantage.”
APPLICATION 10.5
In early 2002‚ steel workers rallied outside the White House in favor of a 30-percent increase in tariffs (taxes) on steel imports.13 The workers argued that the tariffs would save U.S. jobs—an argument that both Congress and Presi- dent George W. Bush found persuasive.
In contrast to policymakers’ stated intent of saving domestic jobs‚ the increased tariffs actually ended up imposing a net job loss on the United States. They did this in two ways. First‚ the tariffs forced consumers to spend
Protecting Steel Jobs Steals Jobs
more on steel than they would otherwise have had to pay. Consumers ended up having less money to spend on other goods‚ many of which are associated with domestic jobs. Second‚ tariffs reduced imports of steel. Reduced imports decreased U.S. exports and associated domestic jobs since the dollars used to purchase imports by necessity come back to pay for U.S. exports.
Estimates indicated that the new steel tariffs cost eight American jobs for every one domestic steel job protected. Even Steel Belt states suffered a net job loss. Illinois lost five jobs for every one protected‚ Ohio three for every one‚ and Pennsylvania and Indiana two for one.13 “Man of Steel?’’ Wall Street Journal‚ March 4‚ 2002‚ p. A14.
10.6 Government Intervention in Markets: Quantity Controls Price is not the only feature of markets that policymakers may wish to control. Quantity can
also be the subject of government attention. For example, in Canada and much of Western
Europe, there are government-specified maximum quantities, or quotas, on the amount of American-produced television that can be broadcast by local stations. In Guangzhou, one
of the biggest cities in China with a population of over 16 million, annual new car sales
have been limited to 120,000 in an attempt to relieve congestion and improve air quality.
Over the years the United States has placed either voluntary or mandatory quotas on
imports of such goods as steel, sugar, cars, computer chips, dairy products, and textiles.
There are also legal limits on the number of people who can immigrate to the United States
from any other country in the world in a given year.
To illustrate how the competitive model can be employed to analyze quantity controls,
we focus on government-imposed maximum quantities—more specifically, the quota
placed on sugar imports by the United States. The effects observed in the case of the sugar
quotas government-imposed maximum quantities of goods
Government Intervent ion in Markets : Quant ity Controls 271
C10.INDD 12:9:21:PM 08/06/2014 PAGE 271Trim Size: 203.2 mm X 254 mm
import quota can be generalized to other cases of government-legislated quotas. Moreover,
the competitive model allows us to see why some of the economic effects of the quota may
be contrary to the intentions of the policymakers implementing the quota.
Sugar Policy: A Sweet Deal The U.S. climate is not well suited to the production of sugar, just as it is unsuited to the
production of coffee and bananas. In the case of sugar, however, there is a thriving domes-
tic industry; in most years we import less than 20 percent of the sugar we use. Why don’t
we import all the sugar we use, as we do with bananas and coffee? The answer is to be
found in the policies adopted by our government toward the sugar industry—policies that
include restrictions on imports from other countries.
Let’s see how the sugar market would operate in the absence of government policy. In
Figure 10.12a we show the U.S. demand and domestic supply curves for sugar as DUS and SUS. Figure 10.12b gives the supply and demand curves for the Rest of the World (RoW) as SR and DR, respectively; these are aggregated across many producing and consuming countries. From these relationships we can derive the total supply curve confronting U.S.
consumers when there is free trade. The total supply curve is shown as ST in Figure 10.12a. It is derived as the horizontal gap between DR and SR in Figure 10.12b, above P1. It con- sists entirely of imports over the relevant range, because sugar can be produced at lower
cost overseas than in the United States. With free trade, equilibrium is established at the
intersection of ST and DUS. The resulting uniform world price is P2, with the United States importing all its sugar, an amount equal to Q1Q2 per year as shown in Figure 10.12b. There is no U.S. production; P2 is not high enough to cover the cost of domestic production.
United States
U.S. sugar output and consumption
0 q2q1qq
P4
SUS S ′
ST SR
DUS DR
P5 P1
RoW sugar output and consumption
0 Q1 Q2
P1
P2
Cents per pound
Rest of the World (RoW)Cents per pound
(a) (b)
P3 qq
T
The Sugar Import Quota With free trade, the uniform world price of sugar is P2, with the United States importing Q1Q2 units of sugar and producing none. (a) With an import quota of qq, the price in the United States is P4, production is q1, consumption is q2, and imports are q1q2. (b) The price on world markets falls to P5.
Figure 10.12
272 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 272Trim Size: 203.2 mm X 254 mm
Suppose that to help domestic sugar producers the government places a quota of qq pounds per year on imports. Then the total supply curve becomes the kinked line ′ST. The new total supply curve is the same as the previous total supply curve without the quota,
ST, out to the quota amount of qq. At qq, the new total supply curve becomes vertical until a sufficiently high price is reached, P3, to encourage domestic production. Above P3, the new total supply curve parallels the domestic supply curve, SUS, and has a magnitude that exceeds SUS by the amount of the quota, qq.
The new total supply curve indicates that RoW suppliers will add an amount equal to
qq to domestic production at each possible price above P5. Although RoW suppliers would like to sell more than this in the United States, they are not permitted to do so. The result of
the import quota is a price for sugar of P4 in the United States. This price is determined by the intersection of the U.S. demand curve, DUS, and the total supply curve, ′ST. At the price of P4, U.S. sugar output is q1, domestic consumption is q2, and imports equal q1q2, or the quota amount of qq.
Domestic consumers, of course, bear the burden of the quota in the form of a higher
price (P4 versus P2 in Figure 10.12a) for sugar. The estimated annual loss in consumer surplus from the quota is $5.4 billion as of 2013, about $65 per family. Domestic sugar pro-
ducers benefit from the quota. The estimated annual increase in producer surplus is $536
million as of 2013. Thus, the quota’s cost to consumers is about $10 for each $1 of pro-
ducer surplus gained by suppliers. The reason that the producers’ benefit is so much less
than the consumers’ cost is that much of the consumer cost simply covers the significantly
higher production cost of sugar in this country. In sum, contrary to the statements made by
many political advocates of the quota, the United States is worse off on net because of the
quota: domestic consumers are harmed more than domestic producers are helped.
Given that the sugar import quota imposes a deadweight loss on the United States, why
does this policy remain in effect? The answer relates to the extent to which the benefits
and costs of the sugar import quota are distributed. Although U.S. sugar producers may
gain far less surplus than U.S. consumers lose, there are relatively few sugar producers and
many sugar consumers. The gains from the sugar quota are thus much more concentrated,
and each of the relatively small number of sugar producers has a strong incentive to lobby
policymakers to retain the quota. In contrast, practically every family in the United States
consumes sugar. The estimated annual loss per family is only $65—not much of an incen-
tive for the typical family to lobby policymakers to repeal the quota.
Because their export markets are limited, producers in other countries are also harmed
by the U.S. sugar policy. In Figure 10.12b, we see that when the United States adopts the
import quota, the price on the world market falls from P2 to P5 (P5 is the price at which the gap between the RoW supply and demand curves for sugar is exactly equal to the quota,
qq). This benefits consumers in other countries, but harms producers, and, in particular, sugar-producing countries. Ironically, many of the countries that produce sugar are less-
developed countries, such as the Philippines and Haiti, that U.S. policymakers generally try
to help. Moreover, the effects are substantial, as is suggested by the fact that in some years
the U.S. price has been more than five times the world market price!
The effects on producers in other countries are exacerbated by the fact that when cer-
tain less-developed countries are kept from more fully benefiting from trade‚ economic and
social conditions conducive to the spawning of crime and terrorism are at times created. The
sugar-cane growing area of Egypt‚ for example‚ was the breeding ground over the last two
decades for that country’s main Islamic fundamentalist terrorist group‚ Gama Islamiyya.
Terrorist group organizers found willing recruits and safe havens in the impoverished area.
The effects of the import quota also do not stop with the sugar market. For example, with
the domestic price of sugar so far above world levels, importing food products with high
sugar content has become profitable because these foods can be produced more cheaply
abroad, using sugar purchased on the world market. Because this reduces domestic sugar
purchases and puts downward pressure on the price of sugar in the United States, political
Government Intervent ion in Markets : Quant ity Controls 273
C10.INDD 12:9:21:PM 08/06/2014 PAGE 273Trim Size: 203.2 mm X 254 mm
pressure to apply import restrictions on imported goods that contain sugar has arisen. In 1985,
import quotas were placed on a number of these goods, such as cake and pancake mixes.
Beyond imports, other market forces in the form of substitutes are at work, making
maintenance of the price so far above competitive levels increasingly difficult. For exam-
ple, production of a major substitute for sugar called high-fructose corn syrup (HFCS) has
increased enormously, and this product has replaced sugar in many uses. (Today, HFCS is
the major sweetening ingredient in soft drinks; until the mid-1980s sugar played this role.)
Because HFCS is made from corn, corn producers have benefited from the import quota.
As a consequence, we now see grain growers from the cornbelt lobbying alongside sugar
growers from the South for preserving the quota.
The tremendous disparity between sugar prices in the United States and the rest of the
world also creates some unusual attempts to evade the import quotas. According to the
Department of Agriculture, some companies make money by importing products such as
iced tea mix, separating out the sugar once the mix has safely crossed the U.S. border, and
throwing away the remainder of the product. The sugar can then be sold at a cost to buyers
still lower than if it had been purchased directly from a domestic producer.
APPLICATION 10.6
In 2002‚ Kraft Foods‚ the maker of LifeSavers, announced that production facilities for its candy‚ which had been made in the United States for 90 years‚ would be moved to Canada.14 The lost domestic jobs were the direct result‚ according to Kraft‚ of sugar import quotas that made
Why Sugar Import Quotas Were Job Losers with Respect to LifeSavers
sugar‚ the main ingredient of LifeSavers‚ at least twice as expensive at that time in the United States as in Canada. During the same year‚ Brach’s‚ another candy maker‚ moved its production facilities from Chicago to Mexico‚ where sugar could be purchased‚ as in Canada‚ at the world price. Brach’s move to evade the effects of the U.S. sugar import quota resulted in 1‚100 jobs being lost at its Chicago plant.
14 George Will, “Protectionist Decisions Undercut Free Trade,” Dallas Morning News, April 14, 2002, p. J4.
Quotas and Their Foreign Producer Consequences The varied methods foreign producers use in trying to get around U.S. import quotas are
not peculiar to the case of sugar. For example, confronting limits on the number of com-
puter chips they can sell in the United States, foreign manufacturers get their chips into the
United States in an indirect fashion. They install them into personal computers and then
export the personal computers to the United States. In response to limits on U.S. textile
imports from Hong Kong, manufacturers there may move their production facilities to
Indonesia. Faced with constraints on the number of immigrants that the United States will
accept from Poland in any year, prospective Polish immigrants to the United States may
first move to Canada or Italy, countries upon which the United States places less restrictive
immigration quotas. And under the “voluntary export restraints” on automobiles negotiated
by the U.S. and Japanese governments between 1981 and 1994, Japanese car producers
responded to the limit of cars that could be shipped to the U.S. market per year in a number
of creative ways, including moving some of their production to the United States; tilting
their product mix toward more luxurious car models such as Lexus and Acura with higher
per-unit profit margins; making standard previously optional features such as air condition-
ing, and adding the fee for the features into the base sticker price; and calling certain sport-
utility vehicles such as the Suzuki Samurai and Isuzu Trooper “trucks,” thereby avoiding
the export restraint on “cars.”
274 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 274Trim Size: 203.2 mm X 254 mm
SUMMARY
A broad range of applications shows that while gov-
ernment intervention may be justified on the grounds of
helping people, its effects may be precisely counter to
the objectives of those favoring the intervention.
Producer surplus, analogous to consumer surplus, is
the gain that producers realize from the sale of output to
consumers when the price exceeds the minimum amount
necessary to compensate the seller. In general, producer
surplus will accrue to some of the owners of inputs that
have upward-sloping supply curves to the industry.
Total surplus is the total net gain to those who partici-
pate in a market—that is, the sum of consumer surplus
and producer surplus.
Price ceilings result in a deadweight loss, or welfare
cost, which is the aggregate loss in well-being of all par-
ticipants in a market.
Excise taxes produce different effects in the short run
and in the long run. In the short run, all firms incur a loss
and have an incentive to cut output, increasing price. In
the long run, some firms exit the industry, output drops
further, and the final price to consumers is higher than in
the short run. In an increasing-cost industry, consumers
and certain input owners bear the burden of the tax.
For a given supply curve and per-unit excise tax, the
more inelastic the demand curve, the greater the tax bur-
den on consumers, the smaller the tax burden on produc-
ers, and the smaller the reduction in output.
In the case of a perfectly elastic supply curve, the
constant-cost case, the price to consumers rises by the
amount of the tax per unit, regardless of the elasticity of
the demand curve.
Because an excise tax results in an output lower than
in an unfettered competitive market, it produces a dead-
weight loss by restricting output to a level where the
product’s marginal benefit exceeds the marginal cost of
production.
During the period in which the U.S. airline industry was
regulated, the Civil Aeronautics Board (CAB) imposed,
in effect, a price floor that kept price above the competi-
tive level. Since suppliers could not compete on the basis
of price, nonprice competition was waged, increasing costs
and reducing profit despite high prices.
Airline deregulation has resulted in lower ticket
prices, a major restructuring of the airline industry, and
lower costs for the firms that remain.
In most major U.S. cities, the supply of cabs is lim-
ited by an entry restriction in the form of a fixed number
of city-issued licenses, or medallions, making the supply
curve of passenger-miles fixed and vertical. As demand
increases, prices rise. In many cities, licensing and other
forms of entry restriction such as maximum fares (a form
of price ceiling) give rise to illegal markets in transporta-
tion services.
The supply–demand model shows that international
trade is beneficial to the United States as a whole,
although in some situations individual producers or con-
sumers may lose.
Imports and exports are fundamentally linked in that
the dollars that foreigners earn on U.S. imports must
ultimately be employed to purchase U.S. goods.
Governments may control quantity as well as price.
Quantity is controlled by quotas and other legisla-
tion, whose economic effects are sometimes contrary
to the intentions of policymakers. Restrictions on sugar
imports, for instance, impose a deadweight loss in that
the gains to domestic sugar producers are outweighed by
the cost to consumers caused by the significantly higher
cost of domestic sugar.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
10.1 What is producer surplus? What is consumer surplus? What is total surplus? Explain how each is shown in a supply
and demand graph.
10.2 What is the relationship between the efficient level of out- put of a good and the size of total surplus achieved? Is total
surplus greater if the output of the good is greater than the
competitive output?
10.3 Define deadweight loss. How is it related to the concept of total surplus?
10.4 Using long-run supply and demand curves, analyze the effects of an ad valorem excise tax equal to 20 percent of the
market (selling) price of gasoline. How do the effects differ
from those of the per-unit excise tax discussed in the text?
*10.5 Suppose that the gasoline market is competitive and that the government grants a subsidy to the industry’s firms of 50
cents per gallon. How will the subsidy affect price; output; and
consumer, producer, government, and total surplus? Is there a
C10.INDD 12:9:21:PM 08/06/2014 PAGE 275Trim Size: 203.2 mm X 254 mm
Review Quest ions and Problems 275
deadweight loss associated with the subsidy? Does the price
to consumers fall by more if the industry is increasing-cost or
constant-cost? (Hint: This situation is the reverse of the excise tax analysis: the supply curve shifts down in height by 50 cents
per gallon.)
10.6 Why is it not inconsistent to say that airline fares, as regu- lated by the CAB, were above the competitive level and yet
airlines did not realize economic profits?
10.7 The airlines generally favored CAB regulation and were opposed to fares being determined in open markets. Since air-
line profits apparently were not unusually high under CAB reg-
ulation, what reasons could account for this support?
*10.8 Suppose that New York City deregulates its taxicab market and, to avoid great harm to medallion owners, buys up
outstanding medallions at the prevailing market price. Is this
policy preferable to simply deregulating the market without
buying up the medallions? In your answer, identify who would
be harmed by the two alternative methods of deregulation.
10.9 How would each of the following affect the operation of a regulated taxi market and the price of a medallion?
a. A reduction in the maximum fare cabs can charge. b. An increase in the fares on subways and buses. c. An increase in parking fees in the downtown area. d. A police crackdown on the operation of illegal taxis. e. An announcement that the market will be deregulated in
five years.
10.10 Suppose that the gasoline industry is competitive and constant-cost. Suppose also that, due to an unexpected increase
in demand, the industry’s firms are making short-run eco-
nomic profits. Using graphs, depict what would happen in the
short and the long run if the government imposed a 50 percent
“windfall profits tax” on the economic profits being earned by
the industry’s firms.
10.11 In the market for organs for transplant, such as kidneys and hearts, the price is constrained to equal zero. Opposition
to any type of remuneration for donating organs has been all
but absolute from physicians and legislators. Reliance on altru-
ism, however, does not appear to be working. The number of
people waiting for organ transplants (and dying if they do not
receive them) is double the number of willing donors. Rely-
ing on graphs, explain the effect of the proscription of finan-
cial incentives for organ donations on the producer surplus and
consumer surplus in this market.
10.12 Suppose that the low-skill job market is perfectly com- petitive and that the equilibrium wage and monthly output in
the market absent government interference are $4.50 per hour
and 1,000,000 hours, respectively. Assume that the demand
and supply elasticity equal two and one, respectively. If the
federal government mandates a minimum wage of $7.25 per
hour, explain what happens to producer, consumer, and total
surplus. Is there a deadweight loss associated with the mini-
mum wage?
10.13 If all else is the same as in Question 10.12 but the demand elasticity equals zero, what is the effect of the minimum
wage on producer, consumer, and total surplus? Is there a dead-
weight loss?
10.14 “Consumers understandably like lower prices, but they should understand there is a great difference between a lower
price produced by a government price ceiling and a lower price
that comes about through normal market channels; one benefits
the consumer, the other may not.” How does our analysis of
rent control relate to this pronouncement?
10.15 Using a pair of graphs like those in Figure 10.10, illustrate a situation in which the United States would be an
exporter of the good in question, and identify the equilibrium.
10.16 In the Figure 10.10 situation, assume that P3 − P1 equals 10 cents per pound and that the cost of transporting sugar from
RoW to the United States is equal to 1 cent per pound. Explain the determination of equilibrium in this case.
10.17 If bad weather causes the supply of sugar in RoW to fall, how will this affect the U.S. market if the import quota
described by Figure 10.12 is in place? Does this explain why
the U.S. and world prices can differ greatly from year to year?
10.18 Using Figure 10.12, show the effect on consumer and producer surplus of the sugar import quota (relative to free
trade). Also show the changes in consumer and producer sur-
plus in RoW.
10.19 Suppose that sugar imports are completely prohibited by the U.S. government in Figure 10.12. What will be the new
equilibrium in the United States and RoW? Show the effect on
consumer and producer surplus of the prohibition on imports
(relative to free trade). Also show the changes in consumer and
producer surplus in RoW.
10.20 Ghana is a producer and exporter of crude oil. Since Ghana is a relatively small crude-oil-producing country, its
actions do not affect world prices; as an exporter, Ghana faces
a foreign demand curve that is perfectly elastic at a price of
$15 per barrel. The equation for the domestic demand curve
is Qd = 26 − P, where price (P) is measured in dollars per barrel and quantity demanded (Qd) is measured in billions of barrels per year. The equation for the domestic supply curve is
Qs = 10 + P, where quantity supplied (Qs) is measured in billions of barrels per year.
a. Assuming free trade, show graphically how much crude oil will be produced, consumed, and exported by Ghana.
b. Graphically, show the gains from trade. Explain who wins and who loses, and show by how much in terms of pro-
ducer and consumer surplus. Does everyone in Ghana ben-
efit from free trade? Explain why or why not. Is Ghana as a
whole better off? Explain.
c. Suppose that the government of Ghana provides a $2 per- barrel subsidy for every barrel of crude oil from Ghana
bought by foreigners. Graphically, show the effects of the
subsidy on domestic production; domestic consumption;
exports; and the welfare of producers, consumers, the gov-
ernment, and Ghana as a whole.
10.21 If output is at an inefficient level, does this imply that consumer surplus is smaller than at the competitive output?
276 Using the Competit ive Model
C10.INDD 12:9:21:PM 08/06/2014 PAGE 276Trim Size: 203.2 mm X 254 mm
Does it imply that producer surplus is smaller than at the com-
petitive output?
10.22 Explain how an excise tax levied on a constant-cost industry produces a deadweight loss. Use a graph to show the
loss in consumer and producer surplus from the excise tax. Is
the loss in total surplus the same as the deadweight loss? If not,
show the deadweight loss in the graph and explain the meaning
of the remainder of the loss in total surplus.
10.23 States do not currently collect sales taxes on consumer purchases on the Internet. Who—consumers or suppliers—
benefits from the absence of such taxes? Explain.
10.24 Most experts believe that electronic commerce cannot explain the low rate of price inflation in the United States.
These experts point to the fact that online retail sales account
for about 6 percent of total retail sales. Using the trade model
we developed in this chapter, explain the basis for the experts’
conclusion that electronic commerce cannot be exercising
a significant amount of downward pressure on the prices of
most goods.
10.25 Suppose that entry of new doctors in the U.S. medical market is limited through caps on medical school enrollments
and licensing restrictions imposed by the American Medical
Association. If the medical market is constant-cost and initially
in long-run competitive equilibrium‚ what will be the effect of
an increase in demand in the short and long run on: the num-
ber of hours worked by the average licensed doctor‚ the profits
made by licensed doctors‚ and the prevailing price charged by
doctors per unit of service?
10.26 How would your answers to the previous question differ if all the assumptions were the same but there were no entry
restrictions imposed by the American Medical Association?
C11.INDD 07:33:13:PM 08/08/2014 PAGE 277Trim Size: 203.2 mm X 254 mm
277
CHAPTER 11 Monopoly
In perfect competition, firms are price takers. In other words, firms are numerous enough to ensure that no single seller affects the market price.
Monopoly is the polar opposite of perfect competition in that it describes a market with a single seller. A monopoly firm faces the market demand curve for its product because it is
the sole seller of the product. Since it faces the market demand curve, the monopoly firm
has control over the market price: it can choose any price–quantity combination on the mar-
ket demand curve.
What price and output level should a profit-maximizing monopoly firm select? We will
see that, relative to perfect competition, monopoly results in a higher price and a lower
quantity. This has efficiency implications, and we discuss why it is illegal in the United
States to monopolize a market.
Although pure monopoly is rare, markets where a small number of firms compete with
one another are common. Chapters 13 and 14 more fully explore the strategic interactions
between firms in such markets. In general, however, the firms may have some monopoly power: some control over price, some ability to set price above marginal cost. This chapter discusses the determinants of monopoly power, how to measure it, and its implications for
product pricing.
monopoly a market with a single seller
monopoly power some ability to set price above marginal cost
Learning Objectives
Define monopoly and show what a monopolist’s demand and marginal revenue curves look like. Explain why a monopolist’s profit-maximizing output is where marginal revenue equals marginal cost. Describe why the extent to which a monopolist’s price exceeds marginal cost is larger the more inelastic the demand faced by the monopolist. Explain several important implications of monopoly analysis such as that the shutdown condition applies to monopolies as well as to firms operating in perfectly competitive environments. Outline the potential sources of monopoly power: absolute cost advantages, economies of scale, product differentiation, and regulatory barriers. Explore the efficiency effects of monopoly from static and dynamic perspectives. Provide an overview of public policy toward monopoly.
Memorable Quote “I think that it’s wrong that only one company makes the game Monopoly.”
—Steven Wright, comedian
278 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 278Trim Size: 203.2 mm X 254 mm
11.1 The Monopolist’s Demand and Marginal Revenue Curves A monopoly faces the market demand curve for its product because it is, by definition, the
only seller of the product. Thus, a monopoly’s demand curve slopes downward. This con-
trasts sharply with the horizontal demand curve faced by a competitive firm. While a com-
petitive firm is a price taker, a monopoly is a price maker. A monopoly supplies the total market and can choose any price along the market demand curve it wants. Since the
monopoly faces a downward-sloping demand curve, if it raises price, the amount it sells
will fall. Much of the analysis of monopoly and the difference in output and price between
a monopoly and a competitive industry stems from this difference in the demand curves.
Let’s consider the co-stars of Mad Men, a popular series on television in recent years. Let’s assume that the Mad Men co-stars face the demand curve depicted in Figure 11.1, are interested in maximizing profit, and must charge the same price for each new show
produced per month. According to the last assumption, while the Mad Men co-stars can operate on any price–quantity point along the demand curve they face, once they select a
price they must charge that same price for all shows sold.1
Under these assumptions, what price should the co-stars choose? Is it better to select
a very high price and produce little but make a killing from each unit sold? For instance,
if only one show is produced, the Mad Men co-stars make $1 million per show. Or is it advisable to select a lower price and sell more shows, even though the price one can
charge declines with output? The price the co-stars obtain is only $400,000 if they supply
seven shows.
price maker a monopoly that supplies the total market and can choose any price along the market demand curve that it wants
1We leave to Chapter 12 the topic of price discrimination and what happens when a monopoly firm can charge different prices for the various units of output that it sells.
The Monopolist’s (Mad Men co-stars) Demand Curve The Mad Men co-stars confront a downward-sloping demand curve. Price exceeds marginal revenue with a downward-sloping demand curve. If price falls from $800,000 to $700,000, total revenue changes by area B (the price at which the fourth unit is sold) minus area A.
E ′
E A
D = AR
B
Output (new shows per month)
$1,000,000
$800,000 = P
$700,000 = P ′
$400,000
0 Q Q ′(1) (7)
(3) (4)
Price per show
Figure 11.1
The Monopol ist ’s Demand and Marginal Revenue Curves 279
C11.INDD 07:33:13:PM 08/08/2014 PAGE 279Trim Size: 203.2 mm X 254 mm
In making its price and output decision, any profit-oriented firm will be concerned with
the relationship between output and total revenue. Will more output increase total revenue
and, if so, by how much? Recall that marginal revenue equals the change in total revenue
associated with a one-unit change in output. Marginal revenue thus indicates how an out-
put change affects total revenue. Understanding the significance of marginal revenue for a
firm’s output decision and the way marginal revenue is related to the firm’s demand curve
is central to analyzing monopoly and other noncompetitive market structures.
For a competitive firm facing a horizontal demand curve, marginal revenue is equal to
the product’s price (average revenue). With a downward-sloping demand curve, the situa-
tion is different: marginal revenue is always less than price. Figure 11.1 shows why. When
price is $800,000, the Mad Men co-stars can sell three shows, and total revenue equals rectangle PEQ0, or $2,400,000. To sell four shows, the co-stars must reduce their price to $700,000 since the demand curve slopes downward. Total revenue for 4 units sold is
P′E′Q′0, or $2,800,000. Note how total revenue changes when output increases from three to four shows. The rectangular measure of total revenue decreases by area A: this area indi- cates how much revenue is lost on the first three shows when they are sold for $700,000
instead of $800,000 (area A equals $300,000). The rectangular measure of total revenue, however, also increases by area B—the amount added to total revenue from selling the fourth show for $700,000. Area B is equal to the new price the Mad Men co-stars have chosen, $700,000. When four shows are sold instead of three, total revenue rises by area
B (the price received for the fourth show) minus area A (the reduced revenue from selling the first three shows at a lower price), or by $700,000 minus $300,000, or $400,000. The
increase in total revenue is marginal revenue, and it is less than the price (area B) because the price of the first three shows must be reduced to sell four shows. This reasoning applies
to any downward-sloping demand curve and shows why marginal revenue is always less than price when the demand curve slopes downward, except for the first unit sold.2
Another way to see the relationship between price and marginal revenue is to recall that
the demand curve is the same as the average revenue curve. If four shows are sold for
$700,000 each, the average revenue per show is the same as the price. Viewed this way, the
demand curve is a declining average revenue curve, and whenever the average is falling,
the marginal curve associated with it must lie below the average.
Marginal revenue is not a fixed amount but varies with the quantity sold. Table 11.1
illustrates a hypothetical relationship between the Mad Men co-stars’ demand schedule and total revenue (TR), marginal revenue (MR), and average revenue (AR). The first two col- umns reflect the assumption of a downward-sloping demand curve, with quantity sold (Q) rising as price (P) declines. MR = P = AR for the first show sold, but for all other outputs
Demand and Total, Marginal, and Average Revenue
P Q TR MR AR
$1,100,000 0 $0 — —
1,000,000 1 1,000,000 $1,000,000 $1,000,000
900,000 2 1,800,000 800,000 900,000
800,000 3 2,400,000 600,000 800,000
700,000 4 2,800,000 400,000 700,000
600,000 5 3,000,000 200,000 600,000
500,000 6 3,000,000 0 500,000
400,000 7 2,800,000 −200,000 400,000
Table 11.1
2In Table 11.1, where we assume that output can be produced only in whole units, marginal revenue equals price at an output of one. If we allow for output to be produced in ever-smaller and less than whole units, the marginal revenue and demand curves have the same height only at their intercepts on the vertical axis.
280 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 280Trim Size: 203.2 mm X 254 mm
price exceeds marginal revenue. When output rises from 1 to 2, for example, total revenue
rises from $1,000,000 to $1,800,000. So MR for the second show is $800,000, but P is $900,000, according to the demand curve.
11.2 Profit-Maximizing Output of a Monopoly Demand and cost conditions jointly determine the most profitable output for a monopoly,
just as they do for a competitive firm. Analytically, the only difference is that a monopoly
faces a downward-sloping demand curve while a competitive firm faces a horizontal
demand curve. Although the demand curve’s slope depends on the market setting, the
output-decision rule for maximizing profit does not. In other words, both competitive and
monopoly firms maximize profit by setting output where marginal revenue (MR) equals marginal cost (MC).
To see why the MR = MC decision rule applies to monopolies as well as to competi- tive firms, consider the demand and cost data for a monopoly firm shown in Table 11.2.
We know the firm is a monopoly from the demand data in the first two columns. These
columns show that price must be lowered to sell more output, indicating that the firm’s
demand curve slopes downward. Multiplying price times quantity for each output yields
total revenue, as shown in column (3). Column (4) identifies the long-run total cost (TC) of producing each output. Since profit (π) is the difference between total revenue and total
cost, the firm selects the output where total revenue exceeds total cost by the largest pos-
sible amount. This occurs at an output of 7 and a price of $8.80. At that output, profit is
$12.21 and MR ≈ MC. To see that profit is maximized where MR = MC, note that marginal revenue exceeds
marginal cost at output levels less than seven units, indicating that the firm can increase
profit by expanding output, but to do so, it must lower price. For example, the marginal
revenue from selling the fourth unit ($8.80) exceeds the marginal cost ($6.50). Thus, profit
will be $2.30 higher if the firm expands output from three to four units, as shown in the
fifth column. At output levels greater than seven units, marginal cost exceeds marginal rev-
enue, and the firm can increase profit by reducing output and raising its price. For example,
the marginal revenue from selling the tenth unit is $6.40, but the marginal cost of produc-
ing it is $9.00. Profit will be $2.60 higher if the firm reduces output from 10 to nine units;
that is, cost will fall by $2.60 more than revenue.
Profit Maximization by a Monopolist (in Dollars)
P Q TR TC π AR AC MR MC (1) (2) (3) (4) (5) (6) (7) (8) (9)
10.20 0 0 0 0 — — — —
10.00 1 10.00 8.00 2.00 10.00 8.00 10.00 8.00
9.80 2 19.60 15.00 4.60 9.80 7.50 9.60 7.00
9.60 3 28.80 21.00 7.80 9.60 7.00 9.20 6.00 MR > MC 9.40 4 37.60 27.50 10.10 9.40 6.88 8.80 6.50
9.20 5 46.00 34.50 11.50 9.20 6.90 8.40 7.00
9.00 6 54.00 41.80 12.20 9.00 6.97 8.00 7.30
8.80 7 61.60 49.39 12.21 8.80 7.056 7.60 7.59 MR ≈ MC 8.60 8 68.80 57.00 11.80 8.60 7.13 7.20 7.61
8.40 9 75.60 65.00 10.60 8.40 7.22 6.80 8.00 MR < MC 8.20 10 82.00 74.00 8.00 8.20 7.40 6.40 9.00
Table 11.2
⎫
⎬
⎪ ⎪⎪
⎭
⎪ ⎪ ⎪
⎫ ⎬ ⎪
⎭⎪
Prof i t-Maximizing Output of a Monopoly 281
C11.INDD 07:33:13:PM 08/08/2014 PAGE 281Trim Size: 203.2 mm X 254 mm
Graphical Analysis Figure 11.2 depicts the profit-maximizing output for a monopoly. Panel (a) shows the
monopoly’s total revenue and total cost curves. Profit is maximized at the output where
TR exceeds TC by the largest possible amount. In the figure, the profit-maximizing output is Q1 (7 units in Table 11.2), where total revenue is AQ1 ($61.60) and total cost is BQ1 ($49.39). Total profit is shown by the distance AB ($12.21). Profit is smaller at every other output. Marginal cost and marginal revenue at output Q1 are shown by the slopes of the TC and TR curves. Marginal cost is the slope of TC at point B (the slope of the line bb), and marginal revenue is the slope of TR at point A (the slope of the line aa). The slopes of the curves at these points are equal to one another since the most profitable output occurs
where MR = MC. Figure 11.2b depicts the most profitable output by using the per-unit cost and revenue
curves. Because this approach is the more useful one—and the one we will use from now
on in the text—we devote more attention to it. It is important to recognize, however, that
the total and per-unit curve approaches are equivalent ways of looking at the same problem.
Profit Maximization: Total and Per-Unit Curves (a) Profit is maximized when total revenue exceeds total cost by the largest amount possible. Maximum profit occurs at output Q1, where the slopes of TR and TC (MR and MC) are equal. (b) The per-unit revenue and cost curves illustrate the same situation shown in part (a).
A
a
a
TC
TR
Dollars
Output0
Q0 Q1
C D = AR
Output
$8.80 = P
$8.40
$7.00
MR
0
(5) (7)
B
b
b
MC
(a)
(b)
E
Dollars per unit
Q1
Figure 11.2
282 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 282Trim Size: 203.2 mm X 254 mm
Figure 11.2b shows the monopolist’s demand (average revenue) curve and the associ-
ated marginal revenue curve. As discussed in Section 11.1, for a negatively-sloped demand
curve, marginal revenue is less than price at all output levels.
The monopolist’s profit-maximizing output, Q1 (7 units) in Figure 11.2b, is identified by the intersection of the MR and MC curves, at point C. The price charged by the monopolist ($8.80, based on the Table 11.2 data) is shown by point E on the demand curve. At any other output marginal revenue is not equal to marginal cost, and profit is lower. For exam-
ple, at output Q0 (5 units) marginal revenue is $8.40 and marginal cost is $7.00. Selling an additional unit of output thus adds more to revenue ($8.40) than to cost ($7.00), and profit
will increase. At any output where marginal revenue exceeds marginal cost, the firm can
increase profit by expanding output. So, in Figure 11.2b, output should be increased up to
the point where the falling MR curve meets the rising MC curve, at point C. Figure 11.2b identifies the most profitable output, but it does not show exactly how
much profit is realized. To show the amount of profit explicitly, we must draw in the aver-
age cost (AC) curve. We do so in Figure 11.3. The most profitable output is, once more, Q1, with a price of $8.80 charged. The difference between average revenue ($8.80) and average
cost ($7.056) at Q1 is the average profit per unit—in this case $1.744. Multiplying the aver- age profit by the output, Q1 (7 units), gives total profit ($12.21), shown in the diagram by the shaded area.
We have been implicitly using long-run cost curves, as shown by the fact that there are
no fixed costs. But the same graphical analysis applies when we use short-run cost curves.
As in the competitive case, a short-run analysis is appropriate when an unexpected or tem-
porary change occurs in market conditions.
The Monopoly Price and Its Relationship to Elasticity of Demand Our analysis of monopoly has shown that to maximize profit, output should be at the level
where marginal revenue equals marginal cost, with price set above marginal cost as indicated
by the demand curve. Suppose that you are a monopolist. How would you put this analysis
to use in identifying the profit-maximizing price and output? It is plausible that you would
Profit Maximization Total profit, the shaded area, is maximized at Q1 where MC = MR.
Dollars per unit
Output
$8.80
$7.056
0 Q1
(7)
MC
D
MR
AC
Figure 11.3
Prof i t-Maximizing Output of a Monopoly 283
C11.INDD 07:33:13:PM 08/08/2014 PAGE 283Trim Size: 203.2 mm X 254 mm
know your marginal cost of production, but how do you find out what the demand curve
for your product (and, hence, the marginal revenue curve) looks like? If you were operating
in a competitive market, you would have no problem—you could simply observe the price
charged by your competitors and recognize that you could sell all you want at that price. As
a monopoly, however, you have no competitors and lack this source of information.
One way to proceed is to use your judgment and set a price, then observe the results.
You could then experiment with raising and lowering the price, and through trial and error
zero in on the profit-maximizing price. Obviously, you would make mistakes, and the mis-
takes could cost you money (in the form of sacrificed profit). Thus, you would like to find a
way to more quickly arrive at the profit-maximizing price, and economic analysis suggests
one such mechanism. Specifically, a little bit of algebra shows that if you know your mar-
ginal cost (MC) and demand elasticity (η), you should set price (P) such that:3
( )
. P MC
P − = 1
� (1)
The left-hand side is the markup of price over marginal cost expressed as a percentage
of price. This expression shows that to maximize profit, the price markup should equal the
inverse of the demand elasticity. The smaller the demand elasticity, the greater the price
markup. The formula can be rewritten to give price directly as a function of marginal cost
and the demand elasticity:
P MC= −/ [ ( / )].1 1 � (2)
If you know your demand elasticity and marginal cost, this expression can be used to
calculate the profit-maximizing price.4 For example, take the case of the only seller of
gasoline on a particular corner of a major intersection; the seller is a monopolist due to the
station’s location. Suppose also that the station is located far from the airport (the impor-
tance of this assumption will be apparent shortly), marginal cost is $3 per gallon, and
the station’s demand elasticity is 20 (a fairly high number due to the nearby presence of
other stations) and is constant over all ranges of the demand curve.5 Based on the inverse
elasticity pricing formula, the station should charge a price equal to $3/[1 − (1/20)] = $3/(19/20) = $(60/19) ≈ $3.16. With a demand elasticity of 20, in other words, the profit-
maximizing price-marginal cost markup is 5 33. percent.
3Refer to Figure 11.1 and note that the change in total revenue (∆TR) associated with a change in quantity sold (∆Q) is equal to area B minus area A. Area B equals P (∆Q) and minus area A equals Q(∆P)—note that ∆P is a negative number in Figure 11.1. Thus: ∆TR = P(∆Q) + Q(∆P). (1n) Since ∆TR/∆Q is marginal revenue, dividing (1n) by ∆Q yields: MR = P + (∆P/∆Q)Q. (2n) Since the elasticity of demand η equals (when it is expressed as a positive number) −(∆Q/Q)/(∆P/P), ∆P/∆Q equals (−1/η)(P/Q). Substituting (−1/η)(P/Q) for ∆P/∆Q in equation (2n) produces: MR = P + Q[(−1/η)(P/Q)] = P − (P/η) = P[1 − (1/η)]. (3n) At the profit-maximizing output, MC = MR, so: MC = P[1 − (1/η)]. (4n) Subtracting P from both sides of equation (4n) and then multiplying through by −(1/P) yields: (P − MC)/P = 1/η. (5n) 4The formula has one difficulty: it holds exactly only at the point of profit maximization, and because mar- ginal cost and elasticity may vary with output, you may need to use this expression repeatedly to locate the profit-maximizing price. However, if marginal cost and elasticity vary only a little over the range of output you are considering, this formula can approximate the profit-maximizing price quite closely. 5Demand curves with a constant elasticity have the nonlinear, convex shape depicted in Figure 11.4. As explained in Section 11.3, the elasticity varies along a linear demand curve.
284 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 284Trim Size: 203.2 mm X 254 mm
Why do gas stations located near airports often charge more for gasoline than others
that are not? Our inverse elasticity pricing rule suggests an answer. To avoid the hefty refu-
eling charges levied by rental car companies on vehicles returned with a near-empty gas
tank (almost double the going price) and because they may have little time to shop around
before catching their flight, renters are willing to pay more per gallon if they haven’t filled
up prior to reaching the airport. These stations thus hold more monopoly power than do
non-airport stations. The average consumer at an airport gas station is less price sensitive
and the demand elasticity facing the typical airport gas station is smaller.
Say that because car renters are less price sensitive and account for a significant portion
of airport gas station business, the typical airport gas station has a demand elasticity of 3.
According to our inverse elasticity pricing rule, and with a marginal cost of $3 per gallon,
the airport station’s profit-maximizing price is $3/[1 − (1/3)] = $3/(2/3) = $(9/2) = $4.50. The airport station’s price–marginal cost markup is thus 50 percent, nearly 10 times greater
than for the non-airport gas station examined earlier (5 33. percent) facing the same mar-
ginal cost but having a higher demand elasticity of 20.
Figure 11.4 illustrates MC, D, and MR curves for the airport and non-airport gas stations that we have just described. Since the non-airport station faces a more elastic demand, its
price–marginal cost markup is lower than that of the airport station. In the limiting case,
if the demand for non-airport stations was infinitely elastic (instead of equal to 20, as we
have assumed), the inverse elasticity pricing formula shows that price equals marginal cost,
a conclusion familiar from our analysis of perfectly competitive markets. This is shown
in Figure 11.4 through the Dpc = MRpc curves. If the elasticity of a firm’s demand curve is infinity, the price–marginal cost markup equals zero.
In sum, if you know your marginal cost, the only other thing you need to know is the
demand elasticity to determine what price to charge. How can you determine the demand
elasticity? One way is to estimate it statistically, as outlined in Chapter 4. Data from sur-
veys or market experiments offer alternative methods. The important point is that you don’t
need to know the entire demand curve for your product; you need to know just how quan-
tity demanded varies relative to price as summarized by the demand elasticity.
The Inverse Elasticity Pricing Rule The more elastic demand is at the profit-maximizing output, the smaller the markup of price over marginal cost.
Dollars per gallon
Output
($4.50) Pairport
($3.16) Pnon-airport
Dpc = MRpc
Dnon-airport
Dairport
MRairport
MRnon-airport
($3.00) Ppc
0 Q1
MC
Figure 11.4
Further Impl icat ions of Monopoly Analys is 285
C11.INDD 07:33:13:PM 08/08/2014 PAGE 285Trim Size: 203.2 mm X 254 mm
11.3 Further Implications of Monopoly Analysis In this section we extend our discussion of monopoly to clarify several less obvious points:
1. We are so accustomed to analyzing markets in supply and demand terms that it is tempting to apply the same reasoning to a monopoly, but doing so can lead to mistakes.
For example, if demand for a monopolist’s product rises and the monopolist has an
upward-sloping marginal cost curve, we might anticipate that both output and price
will rise. Take a look again, however, at Figure 11.4. With demand Dnon-airport, price is Pnon-airport, and output is Q1. When demand increases to Dairport, the new marginal revenue curve MRairport intersects the MC curve at the original output. Output remains at Q1, but price rises to Pairport.
To guard against thinking of supply and demand (appropriate for a competitive model
but not for a monopoly), we note that a monopoly has no supply curve. A supply curve
delineates the unique relationship between price and quantity supplied when firms have no
control over price. In perfect competition, where firms are price takers, demand shifts trace
out the unique price–quantity combinations (that is, the supply curve). There is no such
unique relationship between price and output in monopoly because the output and price
selected by a monopolist depend on both marginal cost and demand (the monopolist’s
marginal revenue curve is determined by the demand curve). A rise in demand can
consequently lead to an increase in both price and quantity, an increase in quantity but no
increase in price, or an increase in price but no increase in quantity (as in Figure 11.4).
The peculiar outcome shown in Figure 11.4 is not the typical response of a monopoly
to increased demand. Instead, it occurs because the higher demand curve is much less
elastic at the initial quantity. As a general proposition, we suspect that monopolies find
it profitable to expand output when demand increases. For example, if the demand curve
shifts outward parallel to the original curve, or if it rotates about the price axis, output will
rise, as will price, so long as the marginal cost curve slopes upward.
2. Monopolies are usually thought of as making huge profits, but in fact, they may not make a profit at all. A monopoly can always charge a price above cost, but it cannot
force consumers to purchase at that price. The position of the demand curve ultimately
limits its money-making ability. If the long-run average cost curve lies entirely above the
demand curve, as depicted in Figure 11.5, any output the firm produces will have to be
APPLICATION 11.1
Why would a parking spot at a Jack in the Box restaurant cost $990 for a Sunday night? Normally, such a spot would be costless at most times and at most Jack in the Box res- taurant locations. However, on February 6, 2011, Super Bowl Sunday, the Jack in the Box located right across the street from Cowboys Stadium in Dallas, Texas (the site of the Super Bowl), charged precisely such a price for each of its park- ing spots. The $990 amount was equivalent to the restau- rant selling 248 Jumbo Jack hamburgers, 496 tacos, 4,960 ounces of soda, and 8,680 French fries. It also far surpassed
Demand Elasticity and Parking at Jack in the Box
the $50 charged by the Jack in the Box for the same parking spaces for regular Dallas Cowboys home games.
Why the price difference? You guessed it: price elastic- ity of demand. Due to its unique location, the Jack in the Box across the street from Cowboys Stadium has some monopoly or price-making power on the days of home football games. And, because the demand elasticity for the parking spots at the Jack in the Box was much lower on the day of the Super Bowl in 2011, the price that the restaurant charged for the spots was so much higher.
286 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 286Trim Size: 203.2 mm X 254 mm
sold at a loss. Since average total cost lies above average revenue (that is, LAC1 > AR1) at the output (Q1) where MR equals MC, the monopoly depicted in Figure 11.5 will do better to produce nothing in the long run. Just as in perfect competition, shutting down may be
the best option.
Each year thousands of monopolists find out that monopoly power does not guarantee
profits. This group includes those who receive patents on their inventions. Many items
granted patents—which give the inventor the exclusive right to sell the product—are never
marketed at all because businesses believe that potential customers will not pay enough to
cover the production cost. For example, the following items have been given patents and
not proven marketable: a chewing gum preserver, a safety coffin (with an escape tunnel and
alarm so that people mistakenly buried alive can “on recovery of consciousness, ascend. . .
or ring the bell [thus averting] premature death”), and goggles for chickens (to keep them
from pecking one another in order to establish flock hierarchy, a pecking order).6
3. A monopoly’s demand curve is elastic where marginal revenue is positive. An elastic demand curve means that a decrease in price and the associated increase in output will
increase total revenue (total revenue moves in the same direction as output and in the
opposite direction as price when demand is elastic) and when marginal revenue is greater
than zero, total revenue, by definition increases as output rises. In Figure 11.6, the
demand elasticity (η) exceeds unity along the upper portion of the straight-line demand curve (between the outputs of zero and Qtrmax) because marginal revenue is positive over this range. When marginal revenue is zero (at Qtrmax), total revenue remains constant when an additional unit is sold so demand is unit elastic (the effects of the decrease in
price and the associated increase in output on total revenue exactly offset one another in
the case where demand is unit elastic and total revenue thus remains unchanged as output
increases). When marginal revenue is negative (at quantities beyond an output of Qtrmax in the graph), a decrease in price and the associated increase in output reduce total revenue
so the demand curve is inelastic (the effect of the output increase on total revenue is less
than the effect of the price decrease).
Dollars per unit
Output0 Q1
LAC1
AR1
MC
MR
D = AR
LAC
6A. E. Brown and H. A. Jeffcott, Jr., Absolutely Mad Inventions (New York: Dover, 1960).
Monopoly and the Shutdown Condition The shutdown condition applies to monopolies, just as it does to competitive firms. If LAC is greater than AR at the output, Q1, where MR equals MC, zero is the most profitable output.
Figure 11.5
Further Impl icat ions of Monopoly Analys is 287
C11.INDD 07:33:13:PM 08/08/2014 PAGE 287Trim Size: 203.2 mm X 254 mm
As shown in Figure 11.6b, in the case of a straight-line demand curve, a monopolist’s
total revenue curve has the shape of an upside-down bowl. Total revenue peaks at the
output, Qtrmax, where marginal revenue is zero and demand is unit elastic. Total revenue equals zero in two cases: where at least Pmax is charged and zero units are sold and where a price of zero is charged and Qmax units are sold.
One bit of geometry may be useful to keep in mind when drawing marginal revenue
curves for straight-line demand curves: the slope of the MR curve is, in absolute value, exactly twice the slope of the demand curve. The MR curve falls twice as fast and becomes zero at an output exactly halfway between the origin and the level of output where the
demand curve intersects the quantity axis. In Figure 11.6a, marginal revenue becomes zero
at Qtrmax (13 units), while the demand curve reaches zero at Qmax (26 units).
4. Monopolists are frequently thought to make more money if demand for their products is inelastic. Yet we can easily see that a profit-maximizing monopolist will always be
selling at a price where demand is elastic. If, for some reason, a monopoly is producing
an output where demand is inelastic, it can increase its profit by cutting back output and
raising price. Lower output means higher total revenue (when demand is inelastic) and
Monopoly Demand, Marginal Revenue, and Total Revenue At each output, the MR curve’s height shows how much total revenue changes when one unit more or less is sold. The height of the MR curve at any output thus equals the slope of the TR curve at that output. The demand elasticity equals unity, and total revenue is maximized where marginal revenue is zero. The total revenue maximizing price, Ptrmax, is less than the profit-maximizing price, Pπ.
Output
P�
Ptrmax
Output
TR
0
(26)
MC
= 1
> 1
< 1
D = AR
Dollars
0 Q� (7)
QmaxQtrmax (13)
(b)
(a)
MR
(26) Qmax
(13) Qtrmax
Dollars per unit
Pmax
�
�
�
Figure 11.6
288 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 288Trim Size: 203.2 mm X 254 mm
lower total cost, so profit will necessarily increase. The monopoly should reduce output
until it is operating somewhere along the elastic portion of its demand curve. Another way
to see this is by recalling that profit is maximized when marginal revenue equals marginal
cost. Since marginal cost is always greater than zero, marginal revenue must be positive
when profit is maximized. But a positive marginal revenue implies an elastic demand curve
since it means that greater output (lower price) will increase total revenue.
Simple as this point is, notice how it allows us to see the inconsistency in the following
statements: (a) “the oil companies collude with one another, charging a monopoly price for
gasoline”; and (b) “gasoline is a virtual necessity that is in highly inelastic demand.” These
statements cannot both be correct. If gasoline is in inelastic demand at the current price,
that price is not a monopoly price. If the price is a monopoly price, the demand must be
elastic. Yet many people believe that both statements are correct.
11.4 The Measurement and Sources of Monopoly Power As you might suspect, pure monopoly, in which there is only one supplier, is rare. More
common are markets populated by at least several firms selling products that are reason-
ably close substitutes for one another. Even when there are several firms operating in the
same market, however, each firm is likely to face a downward-sloping demand curve and
thus have some monopoly power: some control over price, some ability to charge a price above marginal cost. In this section we explain why this is the case, as well as how the
extent of any individual firm’s monopoly power may be measured and the general sources
of monopoly power.
Consider the aspirin market and suppose that Bayer is one of five (equal-sized) sellers
in it. Suppose also that Bayer assumes that rival suppliers behave as competitive firms in
determining their output. The latter assumption is a simplifying one, not meant to down-
play other types of strategic behavior in which suppliers may engage when making price
and output decisions. Chapters 13 and 14 more fully explore the strategic interactions
between firms when the number of firms operating in a market is small. For now, however,
we ignore alternative forms of strategic behavior to show that even when rival suppliers
behave as competitive firms, an individual firm may have some monopoly power if the
number of rivals is not too great.
Under these assumptions, Figure 11.7 shows how Bayer’s demand curve, d, can be derived from the market demand curve, D, and the supply curve, SO, of all other firms in the market. In Figure 11.7a, if Bayer produces nothing, the market price will be $10 per
bottle, and Bayer’s demand curve will begin at $10 on the vertical axis. How many bottles
can Bayer sell at a price of $9? At $9, other firms will supply 12 million bottles along their
supply curve, but consumers are willing to purchase 15 million bottles, so Bayer can sell
the difference, 3 million bottles. Now we have a second point on Bayer’s demand curve, d. It is obviously highly elastic, with a point elasticity of 9 at an output of 3 million bottles.
Although Bayer’s demand curve is much more elastic than the market demand curve (the
latter has a point elasticity of 0.6 at an output of 15 million bottles), the important point is
that it is not perfectly elastic. And because its demand elasticity is less than infinity, Bayer
has some monopoly power, some ability to set price above its marginal cost.
Given its demand curve, how much should Bayer produce to maximize profit? Once
again, output should be set where mr = MC (3 million in Figure 11.7a). Bayer’s profit- maximizing price is the height of its demand curve ($9) at the output where mr = MC. Note that the price exceeds Bayer’s marginal cost ($8) and that Bayer thus has some monopoly
power even though it is not a pure monopoly. The presence of four other suppliers limits
Bayer’s monopoly power but does not eliminate it.
The Measurement and Sources of Monopoly Power 289
C11.INDD 07:33:13:PM 08/08/2014 PAGE 289Trim Size: 203.2 mm X 254 mm
Measuring Monopoly Power To measure a firm’s monopoly power, economists often rely on the Lerner index (named after economist Abba Lerner). The Lerner index is nothing more than the markup of price
over marginal cost, expressed as a percentage of a product’s price:
Lerner index of monopoly power = −( ) / .P MC P (3)
We noted before that, at the profit-maximizing output, the price-marginal cost markup
equals the inverse of the firm’s demand elasticity, (P − MC)/P = 1/η. Thus, the smaller the firm’s demand elasticity at the profit-maximizing output, the greater the price-marginal cost
markup, and the larger the firm’s degree of monopoly power as measured by the Lerner index.
The Lerner index varies between zero and one. In perfect competition, the elasticity of
the firm’s demand curve is infinite, and price equals marginal cost, so the Lerner index
equals zero. The larger the Lerner index value, the greater a firm’s monopoly power. In the
aspirin example just discussed, Bayer’s demand curve has an elasticity of 9 at the profit-
maximizing output. The degree of Bayer’s monopoly power, as measured by the Lerner
index, thus equals the inverse of the elasticity or 1/9 (roughly 0.11).
The Sources of Monopoly Power What factors determine the extent to which a firm has monopoly power? Our Bayer exam-
ple suggests two: the elasticity of the market demand curve and the elasticity of supply by
other firms. Note how, in Figure 11.7, as the price is dropped from $10 to $9 per bottle, the
expansion of total purchases by consumers (by 1 million bottles) and the reduction in out-
put by other firms (2 million bottles) determine how much Bayer can sell at the lower price.
The size of the expansion in total purchases is determined by the elasticity of the market
Lerner index a means of measuring a firm’s monopoly power that takes the markup of price over marginal cost expressed as a percentage of a product’s price
Output (millions of bottles)
$10 $9 $8
MC
mr
d
Dollars per bottle
Bayer Aspirin market
0 3
(a)
Output (millions of bottles)
So
D
Dollars per bottle
0 15
1412
(b)
$9 $10
Monopoly Power When There Are Several Suppliers Under the assumptions made in the text, Bayer’s demand curve is d. It is derived by subtracting the quantity supplied by other firms (indicated by SO) from the total amount consumers wish to purchase (shown by D) at each price.
Figure 11.7
290 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 290Trim Size: 203.2 mm X 254 mm
demand curve; the size of the reduction in output by other firms is determined by the elas-
ticity of the other firms’ supply curve. Thus, the more elastic D and SO are, the greater the elasticity of Bayer’s demand curve.
If the market demand curve is perfectly elastic, any individual supplier such as Bayer
has no monopoly power. This would be the case even if Bayer were the only aspirin sup-
plier and thus a pure monopolist. Even a pure monopolist, that is, would be unable to set
price above marginal cost. Any attempt to do so would lead to total purchases equaling zero
since consumers are hypersensitive to the price charged by the (albeit pure) monopolist if
the market demand curve is perfectly elastic.
The monopoly power possessed by any one firm is also more limited the greater the num-
ber of rival firms. This is because as rivals become more numerous, the elasticity of supply
by rival firms, as a group, tends to increase and the ability of any one firm to set price above
marginal cost is impeded.7 In Figure 11.7, for example, if SO were more elastic such that production by other firms, as a group, falls to 1 million bottles as the price declines from
$10 to $9, Bayer’s demand curve would be much more price sensitive: the price decline
would result in Bayer’s sales rising from zero (at $10) to 14 million bottles (at $9).
Barriers to Entry A barrier to entry is any factor that limits the number of firms operating in a market and thereby serves to promote monopoly power on the part of incumbent suppliers. Such fac-
tors fall into four general categories: absolute cost advantages, economies of scale, product
differentiation, and regulatory barriers.
1. Absolute cost advantages. An absolute cost advantage occurs where an incumbent firm’s production cost (its long-run average total cost) is lower than potential rivals’
production costs at all relevant output levels. This cost disparity may be due to unique access
to a production technique or an essential input. For example, KFC has a proprietary recipe
for “‘finger-lickin’ good” chicken.8 The Aluminum Company of America (Alcoa) was the
sole producer of aluminum in the United States from the late nineteenth century until the
1940s, because it controlled all domestic sources of bauxite—the ore from which aluminum
is made. In the field of music, Kenny Chesney and Taylor Swift have unique access to
their personal singing abilities. Cisco is the leading maker of networks that use the Internet
because of the superior design technology for routers and servers that the company has
proprietary access to.
2. Economies of scale. All firms (incumbents as well as potential entrants) may have the same cost curves, but the production technology may be such that one large firm
can supply an entire market at a lower per-unit cost than several smaller firms that share
the market. In other words, the long-run average total cost curve for all firms may slope
downward over the relevant range of market output. Consequently, to have more than one
firm operating is wasteful since production cost is minimized if one firm supplies the entire
output. The industry is thus characterized by economies of scale and is termed a natural monopoly.
Natural monopoly is common in the local distribution of power, water, and telephone
services. It is cheaper, that is, to have one electric company serve an entire neighborhood
than to have each home in the neighborhood rely on a separate company with its own
barrier to entry any factor that limits the number of firms operating in a market and thereby serves to promote monopoly power
absolute cost advantage a situation in which an incumbent firm’s production cost (its long- run average total cost) is lower than potential rivals’ production costs at all relevant output levels
economies of scale a situation in which a firm can increases its output more than proportionally to its total input cost
natural monopoly an industry in which production cost is minimized if one firm supplies the entire output
7We continue to ignore, for now, the strategic behavior in which firms may engage when selecting their price and output. As we will see in Chapters 13 and 14, it is possible for a large number of firms to collude in setting a monopoly price and a small number of firms to interact in such a manner as to ensure the com- petitive outcome. We will also see, however, why collusion tends to be less likely as the number of firms in a market grows. 8Reportedly, only two company officials know the recipe.
The Measurement and Sources of Monopoly Power 291
C11.INDD 07:33:13:PM 08/08/2014 PAGE 291Trim Size: 203.2 mm X 254 mm
distinct transmission lines. The single electric company dictated by economies of scale,
however, has the potential to exercise monopoly pricing power.
3. Product differentiation. Product differentiation is a third type of barrier to entry. Consumers may perceive the product sold by an incumbent firm to be superior to that
offered by prospective rivals. Based on this perception, consumers are willing to pay
more for the incumbent firm’s product. For example, Ray-Ban sunglasses may be
sufficiently differentiated in consumers’ eyes to give the company some pricing latitude
over potential competitors—even though the competitors have access to the same
production technology.
4. Regulatory barriers. Finally, a firm may have a limited number of rivals due to regulatory barriers such as government-granted patents, copyrights, franchises, and licenses. A patent, for example, grants the exclusive right to use some productive
technique or to produce a certain product for a period of 17 years. Patents thus amount
to the legal right to a temporary monopoly. Although patents are an instance of
government-created monopoly power, there is an economic rationale for their use—
namely, that firms and individuals will be less inclined to invest in the research and
development of new products if others can immediately copy the results. As we will see
in Chapter 20, this rationale is generally regarded as a valid argument for granting some
protection to inventors. But some economists believe that 17 years is too long; others
believe that given the length of time needed to develop and market a product, 17 years is
not long enough.
Governments also occasionally block entry by requiring firms to have a public
operating license or franchise. Licensing is sometimes defended as a method of ensuring
minimum standards of competency, but it can be (and many feel has been) used as a
barrier to entry that insulates existing holders of licenses from new competition. For
example, one cannot enter the mail delivery, broadcasting, public utility, and trucking
markets without a public license. Similarly, hundreds of occupations require licenses,
among them hair stylists, funeral directors, taxi drivers, contractors, bartenders, and
tailors. Often these licenses are granted by state government boards composed largely of
existing license holders.
In the case of cable television, the ability to provide service to any given community
requires a franchise from the local city government. Until recently, these franchises
were typically exclusive. Under exclusive franchising, no more than one operator is
allowed to serve a community. Exclusive franchising is often predicated on the belief
that economies of scale exist in local cable television distribution. However, studies have
found that any such economies of scale are relatively minor, while the pricing power
conferred by exclusive franchises appears to be substantial.9 The average monthly basic
service rate charged in “overbuild” communities (communities served by more than one
operator) is 20 to 35 percent lower than in comparable communities served by only one
cable operator.
Regulatory barriers can also take the form of the government making its purchases
from particular firms or limiting nonprice forms of competition such as advertising.
Restrictions on advertising exist in many states for products such as legal services,
prescription drugs, health care, and eyeglasses. In general, prices are higher where there
are limits on advertising. For example, researchers find that eyeglass prices are 25 to
30 percent higher in states with total advertising bans than in states with weak or no
restrictions on advertising.10
product differentiation a means by which consumers perceive the product sold by an incumbent firm to be superior to that offered by prospective rivals
regulatory barriers barriers to entry created by the government through vehicles such as patents, copyrights, franchises, and licenses
9Thomas Hazlett, “Duopolistic Competition in Cable Television: Implications for Public Policy,” Yale Journal on Regulation, 7, No. 1 (Winter 1990), pp. 65–119. 10Lee Benham, “The Effect of Advertising on the Price of Eyeglasses,” Journal of Law and Economics, 15, No. 2 (October 1972), pp. 337–352.
292 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 292Trim Size: 203.2 mm X 254 mm
APPLICATION 11.2
Massage is one of the world’s oldest professions.11 The ancient physician Hippocrates touted its benefits and almost all cultures have relied on it including the early Egyptians, Chinese, Greeks, and Romans. While massage fell out of favor in the early 20th century with the advent of biological-science-based medical care, it has become pop- ular since then due to its successful use by athletes and a growing number of insurance plans underwriting alternative medical treatments.
The Effect of State Licensing on One of the World’s Oldest Professions
As massage has grown in popularity, so have efforts by states to regulate it, often with the intention of eliminating prostitution to which it has had some historical links. Ohio was the first state to require massage therapists to obtain a license to practice. By 1975, 10 states had adopted a similar regulatory approach and, as of 2010, 36 states and the District of Columbia license massage therapists.
The effects of the regulatory barriers on the supply of massage therapy and the prevailing price have been predictable. Licensing reduces the number of massage therapists working in a state and drives up the earnings of massage therapists by 12 to 16 percent. Moreover, there does not appear to be any compensating increase in the quality of services rendered in states that regulate massage.
11This application is based on Robert J. Thornton and Edward J. Timmons, “Licensing One of the World’s Oldest Professions: Massage,” Journal of Law and Economics, 56, No. 2 (May 2013), pp. 371–388.
APPLICATION 11.3
The National Collegiate Athletic Association (NCAA) conducts its annual basketball tournament every March.12
Sixty-eight teams are invited to participate in the tourna- ment to determine a national champion during a period called “March Madness.” The tournament is a tremendous revenue generator for the NCAA. This is because, among other things, the exclusive right to broadcast tournament games for an 11-year period has been awarded by the NCAA to CBS Television for $6 billion.
The NCAA used to have more competition in March from competing events. As noted by Jeffrey Kessler, an attorney with Dewey Ballantine:
Once upon a time, the National Invitation Tourna- ment [NIT], which is older than the NCAA tourna- ment and culminates every year in Madison Square Garden, provided strong competition to the NCAA tournament and attracted many of the top teams in the country. In 1950, for example, City College of New York played in and won both tournaments. In 1962, Loyola, Mississippi State, Dayton, the Univer- sity of Houston, and St. John’s all chose to partici- pate in the NIT rather than accept invitations to the NCAA tournament. In 1970, Marquette, one of the best teams in the nation that year, chose to go to the
March Monopoly Madness
NIT over the NCAA tournament, which provoked an outcry by the powers that ran the NCAA tournament. During this time, college basketball fans benefited from the excitement, and the lower ticket prices, cre- ated by having a healthy competition between two premier postseason tournaments.
In the early 1980s, the NCAA sought to limit competi- tion from the NIT. The NCAA tournament was expanded from 32 to 64 teams. In addition, any school invited by the NCAA to its basketball tournament was required to forgo playing in the NIT under the pain of significant penalties.
As a consequence of its actions, according to Kessler, the NCAA was able to acquire an absolute cost advantage over potential rival tournaments such as the NIT through its con- trol of an essential input (top college basketball teams). This control over an essential input permitted the NCAA to alleg- edly solidify its monopoly power and earn significantly higher profits at the expense of rival tournaments and basketball fans. Kessler provides an analogy from the board game Monopoly:
Imagine that a new rule in that board game required that all of the best properties—Boardwalk, Park Place, the railroads, and the rest—be assigned to only one player (the NCAA), while the remaining contestants could try to acquire only Baltic Avenue, Mediterra- nean, and the other streets with much lower values. Could these other players effectively compete? We all know the answer. The NCAA would always win the
12This application is based on Jeffrey L. Kessler, “Tournament Has Become March Monopoly Madness,” New York Times, March 28, 2004, p. 10.
The Measurement and Sources of Monopoly Power 293
C11.INDD 07:33:13:PM 08/08/2014 PAGE 293Trim Size: 203.2 mm X 254 mm
Strategic Behavior by Firms: Incumbents and Potential Entrants A common belief is that the degree of monopoly power exercised by firms in any market
is related to the number of firms: the more firms there are, the less monopoly power each
has. From our earlier discussion of the determination of Bayer’s demand curve, it is easy to
see why some such relationship might be expected. For instance, if there were four (equal-
sized) firms instead of five in our example, then each firm would have a less elastic demand
curve and, therefore, more monopoly power. However, the relationship is not exact, and
sometimes focusing on the number of firms can be misleading. The elasticity of each firm’s
demand curve depends not only on the number of competing firms, but also on the elasticity
of the market demand, the elasticity of the supply curve for other firms, the extent to which
the products produced by the various firms in the industry are homogeneous, and the nature
of the competition among the firms. As we will see in Chapter 13, firms might choose not
to compete perfectly in terms of the prices they charge. At the extreme, an industry’s firms
might even opt to form a cartel and behave in a collusive manner.
Another factor is likely to be of even greater importance: the possibility of entry by new
firms into the market. After all, it is not only the number of firms already operating in a
market that matters. The potential for entry and the elasticity of supply of such potential
entrants can also play an influential role.
The possibility of entry by new firms can greatly constrain the exercise of monopoly
power. To see how, suppose that you possess some monopoly power by virtue of owner-
ship of a patent that enables you to produce CD-ROMs at a lower cost than other firms.
Your marginal cost curve is shown as MC in Figure 11.8, and the market demand curve is shown as TD. If no other firms could sell CD-ROMs, you would maximize your profit by charging $20 per CD-ROM, and producing QM. However, suppose that other firms could sell them at a price of $16. At that price, they will sell whatever quantity consumers wish to
purchase. How will this affect your price and output?
Obviously, if you try to charge a price higher than $16, other firms will enter the market
and you will be undercut; you would not be able to sell any at a price above $16. The demand
curve you confront in this situation is basically horizontal at $16 out to the market demand
curve. Any output between zero and Q units can be sold for a price of $16, but no higher, so the relevant demand curve is the horizontal line PC over that range of output. Higher levels of output beyond Q can still be sold for prices lower than $16, so the CD segment of the market demand curve is still relevant. As a result, your monopoly demand curve effectively becomes
PCD on account of the threat of entry.13 When the demand curve changes, so does your marginal revenue curve. Over the range
where the demand curve is horizontal, P = MR since you can sell additional CD-ROMs without lowering price. Thus, PC is also your marginal revenue curve up to an output of Q. At greater output levels the original demand curve is unchanged, so the FM segment of the original marginal revenue curve associated with the CD portion of the demand curve is still
13We are assuming that other firms effectively have a supply curve that is horizontal at $16. If their supply curve is upward sloping, the PC portion of your demand curve will be negatively sloped but more elastic than the market demand curve.
game and generate huge profits from its properties, and there would be no fair competition at all. That is how the NCAA tournament dominates today.
Keller and his firm represented the NIT in pursuing some legal remedies (that will be discussed later in this chapter) against the NCAA.
294 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 294Trim Size: 203.2 mm X 254 mm
relevant. The entire marginal revenue curve is therefore PCFM. The curve is discontinuous at an output of Q. Think about what the discontinuous (CF) segment of the new marginal revenue curve means. Suppose that Q is 150 units. If output increases one unit, from 149 to 150, then the marginal revenue of the 150th unit is $16 (equal to CQ) since both the 149th and 150th units can be sold for $16. To sell 151 units, however, your firm must reduce
price, say to $15.90. Thus, the marginal revenue of the 151st unit is only about $1 (equal to
FQ).14 Marginal revenue drops abruptly from $16 to $1 at an output of Q. Once we recognize how the threat of entry affects the demand and marginal revenue
curves, the rest of the analysis is straightforward. (However, note that we have not drawn in
the average cost curve. For the analysis to be correct, we must assume that average cost is
low enough for it to be profitable for the monopolist to continue to operate.) With the threat
of entry, QM is no longer the profit-maximizing output; at this output, marginal revenue (now $16) exceeds the unchanged marginal cost. This means that profit can be increased by
expanding production. Note that marginal revenue exceeds marginal cost until output has
increased to 150 units, implying that profit rises as output is expanded up to Q. But the mar- ginal revenue associated with the sale of the 151st unit, FQ, is less than the marginal cost, so it does not pay to produce that unit. The new profit-maximizing output is Q.
Reflect carefully on the implications of the foregoing analysis. You are the only seller in
this market, but your pricing power is rather limited. The threat of entry leads you to charge
a lower price than you would if you could be assured entry would not occur. In general,
depending on the conditions that would attract entry, you may have very little monopoly
power—as suggested in the graph where your price is only slightly above marginal cost.
The example illustrates the important point that the threat of entry, as well as actual
entry, can have a significant impact on the pricing behavior of firms. In addition, it shows
why the number of firms operating in a market does not always have a direct relationship
with the amount of monopoly power exercised.
14Total revenue from selling 150 units at $16 each is $2,400. Total revenue from selling 151 units at $15.90 each is $2,400.90. Thus, marginal revenue from selling the 151st unit is $0.90.
Potential Entry and Monopoly The possibility of entry can affect a monopoly’s price and output. If other firms are willing to sell the product at a price of P, then the monopoly’s demand curve is PCD, and the monopoly will sell Q units at a price of P rather than QM units at a price of PM.
Dollars per unit
Output
$20 = PM
$16 = P
T
0 QM Q
(150)
MC
F
C
D
M
Figure 11.8
The Eff ic iency Effects of Monopoly 295
C11.INDD 07:33:13:PM 08/08/2014 PAGE 295Trim Size: 203.2 mm X 254 mm
11.5 The Efficiency Effects of Monopoly The way a market structure affects the functioning of a market has always been a major concern
in economics. Having examined competitive and monopoly markets separately, we should now
turn to a careful comparison of the two market forms. To do so, we need to determine how
a change in market structure—from competition, for example, to pure monopoly—will affect
price and industry output.
To make the comparison, let’s reexamine the aspirin industry and assume that it is ini-
tially perfectly competitive and constant cost. The constant-cost assumption means that
input prices are the same under competition and monopoly and allows us to isolate more
easily the impact of monopoly in the output market. In Figure 11.9, the market demand and
supply curves are D and LS, so the competitive outcome is a price of P ($11) and output of Q (10 million bottles). The marginal revenue curve associated with the market demand curve is MR, but it plays no role in determining the competitive output since each firm adjusts to its own marginal revenue curve. With perfect competition, each of the numerous
firms faces a horizontal marginal revenue curve at the market-determined price.
Now suppose that the aspirin industry becomes a pure monopoly. The monopoly faces
the industry’s demand and marginal revenue curves, but what about the monopoly’s cost curves? If we assume that the monopoly can operate the separate plants at the same costs as
those of the individual competitive firms, the competitive supply curve is the monopoly’s
average cost curve. Because this curve is horizontal, implying constant average cost regard-
less of output, marginal cost equals average cost. Thus, the horizontal competitive supply
curve is the same as the monopoly’s average and marginal cost curves.
At the initial competitive output Q, the monopoly’s marginal cost (CQ) is greater than marginal revenue (EQ), so the monopolist is in a position to increase profit (from the zero- profit level of the competitive equilibrium) by reducing output. By restricting output, the
monopolist is able to charge a higher price. The profit-maximizing output occurs where
MR = MC at an output QM (5 million bottles). The monopoly will produce QM, charge a price of PM ($15), and realize an economic profit of PMBAP. For the same demand and cost conditions, price will be higher and output lower under monopoly than under competition. This is one of the most important and best-known conclusions of microeconomic theory.
Because a monopoly reduces the output of aspirin, from Q to QM in Figure 11.9, a net loss in total surplus results. The net loss is, of course, the deadweight loss of monopoly. To see why there is a net loss, note that at the monopoly output of QM, price ($15 per bottle) is above marginal cost ($11). Thus, consumers value additional aspirin bottles more than
they cost the monopolist to produce. (Remember that the height of the demand curve at any
quantity reflects the marginal value of a good.) If output is 5 million bottles, the incremental
bottle is worth $15 to consumers, but it uses resources that can produce other goods worth
only $11 (marginal cost). Consequently, a gain of $4, or BA, results if an additional bottle is produced. Each successive unit of output yields a smaller net benefit than the previous
one until output reaches 10 million, where price equals marginal cost. For example, when
the monopoly chooses not to produce the 7.5 millionth bottle, consumers lose a product
worth FQ1 ($13) to them; not producing that bottle permits the production of other goods to increase, but these goods are worth only GQ1 to consumers ($11, equal to marginal cost), so a net loss of FG, or $2, on that bottle results.
The excess of value over cost associated with increasing output from 5 to 10 million
is triangular area BCA. Area BCA is the sum of the loss in net benefits for all the aspirin bottles from 5 to 10 million. This area is a measure of the deadweight loss due to the monopoly restriction of output. The aspirin monopoly chooses not to produce these bottles, so consumers are unable to realize the potential net gain. Under competition, output expands
to 10 million, where price is equal to marginal cost.
296 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 296Trim Size: 203.2 mm X 254 mm
Another way to see that area BCA is a net loss is through the use of consumer and pro- ducer surplus. When the price rises from $11 to $15, consumer surplus falls by area PMBCP. This area measures the loss to consumers from the monopoly price; it is not the deadweight loss because there is a corresponding gain in producer surplus accruing to the monopoly
firm. The gain in producer surplus equals area PMBAP, the difference between the monopoly price and marginal cost over the range of output (5 million) produced by the monopoly.
However, the loss to consumers from the monopoly price, PMBCP, is larger than the pro- ducer surplus gain to the monopoly, PMBAP, by the area BCA. Consumers lose more than the monopoly gains, and the difference—area BCA—is the deadweight loss of monopoly.
The deadweight loss of monopoly, then, is due to an inappropriate level of production.
Monopolies produce an inefficient (too low) level of output, and the triangular area BCA is a dollar measure of the net loss involved. Consumers bear this cost in addition to the cost
they bear from paying the higher monopoly price for the product.
A Dynamic View of Monopoly and Its Efficiency Implications The preceding comparison of monopoly and perfect competition employs what economists
term a static analysis. Basically we took a snapshot at one point in time. We started from a perfectly competitive outcome and assumed no changes in market demand and production
cost. We then investigated what would happen if the industry moved from being perfectly
competitive to monopolized. As we saw, price ended up being higher, output fell, and a
deadweight loss was generated in the process.
While our static analysis indicates that, relative to perfect competition, monopoly
imposes a deadweight loss on society, there is another important way of evaluating the effi-
ciency effects of monopoly. This other way relies on dynamic analysis: looking over time at why monopolies are created in the first place. In contrast to the static analysis, dynamic
analysis suggests an important reason why the existence of certain monopolies should be
viewed more favorably from a social welfare perspective. Specifically, monopoly power
may stem from firms generating better products through either devising ways to lower pro-
duction cost (creating absolute cost advantages) or differentiating the product in consum-
ers’ eyes (product differentiation). If this is in fact the case, monopoly serves to enhance
social welfare from a dynamic perspective since it reflects the creation of better products.
static analysis a form of economic analysis that looks at the efficiency of a market at any one point in time
dynamic analysis a form of economic analysis that looks, over time, at the efficiency of a market
The Deadweight Loss of Monopoly The competitive long-run supply curve is LS; if the industry is competitively organized, output is Q and price is P. With monopoly, LS is assumed to be the same as the monopolist’s long-run AC and MC curves, and the profit-maximizing output is QM at a price of PM. Price is higher and output lower under monopoly. The shaded rectangular area shows monopoly profit. Triangular area BCA is the deadweight loss associated with the reduced output under monopoly.
Dollars per bottle
Output (millions of bottles)
$15 =
$11 = P
$13
0 Q1 Q
(5) (7.5) (10)
F
B
C
E
GA LS = AC = MC
Deadweight loss
D
MR
Monopoly profit
QM
PM
Figure 11.9
The Eff ic iency Effects of Monopoly 297
C11.INDD 07:33:13:PM 08/08/2014 PAGE 297Trim Size: 203.2 mm X 254 mm
Figure 11.10 shows the market for personal computers and contrasts the dynamic view
with the static perspective on monopolies that we have previously outlined. In the early
1970s, the market for personal computers was virtually nonexistent. The absence of a
market at that time reflected the cost of producing such a good (ACbefore = MCbefore) being greater than the amount consumers were willing to pay for it. In other words, even though it
was technologically possible to manufacture computers for personal use in the early 1970s,
the minimum cost producers would have had to be paid to produce computers (as measured
by the height of the ACbefore = MCbefore curve) exceeded the maximum price consumers were willing to pay (as measured by the height of the demand curve) for them.
According to the dynamic perspective, a monopoly is created in this market when a
company figures out a way to lower the production cost of personal computers—say, to a
level of AC = MC. With this innovation, total surplus increases relative to the early 1970s outcome of no personal computers being marketed. There are two reasons for this. First, the
company developing the lower-cost production method will be rewarded for its innovation
by being able to exercise monopoly pricing power, charging a price of PM and earning pro- ducer surplus equal to area PMBEP. Second, consumers also benefit from the innovation. Relative to a world without personal computers, the monopoly outcome of QM increases consumer surplus from zero to an area equal to ABPM. The increase in total surplus (pro- ducer plus consumer surplus) is ABEP.
Of course, once the lower-cost production technique is developed, the static perspec-
tive on monopolies still applies since competition would serve to further increase total
surplus. That is, suppose that there are 20 firms with access to the same cost-reducing
technology (AC = MC) as the innovating company, and 20 firms is sufficient to ensure the perfectly competitive price of P and output of Q. In this situation, relative to the monopoly
Figure 11.10 A Dynamic View of Monopoly Relative to a world where production cost (ACbefore = MCbefore) exceeds the value consumers place on a good (the height of the D curve) and output is zero, a firm possessing monopoly pricing power because it has figured out a way to lower production cost to AC = MC serves to increase total surplus by its actions. The cost-reducing innovation increases consumer surplus by ABPM and producer surplus by PMBEP. P
A
Dollars per unit
0 Q Output
B
CE
ACbefore = MCbefore
D
MR
AC = MC
PM
QM
298 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 298Trim Size: 203.2 mm X 254 mm
price of PM and output of QM, consumer surplus increases by PMBCP, producer surplus decreases by PMBEP, and total surplus increases by triangular area BCE. Just as from a dynamic perspective innovation and pure monopoly are better than no firms possessing
the lower-cost production technology, 20 firms and a perfectly competitive outcome are
preferable, from the static perspective, to pure monopoly, once the lower-cost production
method exists.
Which approach, the static or the dynamic, is the most appropriate to employ in
analyzing monopoly? It turns out that both approaches have merit. As we will see in
Chapter 20 regarding the debate on patents and the length of time for which they should
be granted, monopolies should be encouraged to the extent that they result from the
development of innovative products. Ex post monopoly pricing power provides an ex ante incentive to innovate. Exactly how much incentive should be provided to induce innovation, however, is an open question. The static perspective informs us that, relative
to monopoly, competition increases total surplus once an innovation has been made. And
the longer monopolies retain their pricing power, the more the total-surplus-enhancing
benefits of competition are forestalled—even though such delays serve to induce innova-
tion from a dynamic perspective.
The decision about which approach to employ in analyzing monopoly not only is of
academic interest, but also has considerable policy relevance. Section 11.6 offers a brief
overview of public policy toward monopoly in the United States.
APPLICATION 11.4
Some of the insights from comparing static versus dynamic views of monopoly can also be applied to how we analyze the “market” for government power. In a recent article, Clemson University economists Robert Fleck and Andrew Hanssen show that tyrannical (i.e., monopoly) con- trol of a city-state in ancient Greece was more likely to result in a transition to democracy by promoting (albeit unin- tentionally) the development of a broad base of citizens (wealthy elites, traders, craftsmen, and small land-holders) interested in more competitive, growth-promoting policies.15
Of course, how long it takes to transition from tyranny to a more competitive form of government also matters
Static Versus Dynamic Perspectives on Monopoly Control of Government
as does the damage done to a society by a monopoly- tyrant along the way. For example, in 1485 the sultan of the Ottoman Empire banned printing presses for fear that they would promote attempts to overthrow his rule. The prohi- bition was kept in place by successive Ottoman rulers until 1727 when a single printing press was allowed, although whatever was printed in Arabic still had to be approved by a panel of three religious and legal scholars.
The opposition to printing by Ottoman rulers had a dramatic and lasting impact on literacy, education, and economic success. In 1800, the literacy rate of adult citi- zens in the Ottoman Empire was 2–3 percent versus 60 percent for adult males and 40 percent for adult females in England. The dramatically lower literacy rates and education levels were a key reason why the Ottoman Empire began to lag its European counterparts economi- cally and, by the mid-1800s, became known as the “Sick Man of Europe.”
15This application is based on: Robert K. Fleck and F. Andrew Hanssen, “How Tyranny Paved the Way to Democracy: The Demo- cratic Transition in Ancient Greece,” Journal of Law and Economics, 56, No. 2 (May 2013), pp. 389–412; and Daron Acemoglu and James A. Robinson, Why Nations Fail (New York: Crown Business, 2012).
11.6 Public Policy toward Monopoly U.S. policy toward monopoly has been largely guided by the static view of monopoly.
According to the static view, a monopoly results in an inefficient resource allocation by
producing too low an output level. In comparison with a competitive market structure, it
Publ ic Pol icy toward Monopoly 299
C11.INDD 07:33:13:PM 08/08/2014 PAGE 299Trim Size: 203.2 mm X 254 mm
also transfers income from consumers to the monopoly owners. For both these reasons, it
has been deemed desirable to use public policy to limit the acquisition and exercise of
monopoly power. In the United States, the primary means to achieve these goals have been
antitrust laws, a series of codes and amendments intended to promote a competitive mar- ket environment.
There are three major statutes governing antitrust policy. The first is the Sherman Act, passed in 1890. The Sherman Act makes illegal any activities “in restraint of trade or commerce among the several States.” An example of forbidden activities is price fix-
ing, whereby firms attempt to secure prices above the competitive level. The Sherman Act
also states that “every person who shall monopolize, or attempt to monopolize. . . shall be
deemed guilty of a felony.” Although this appears to make being a monopolist illegal, this
is not how the courts have actually interpreted the provision. Instead, being a monopolist is
a crime only when certain practices are employed.
Partly because of vagueness in the Sherman Act’s wording, Congress passed two
more important pieces of antitrust legislation in 1914. The Clayton Act explicitly out- laws specific business practices believed to be monopolistic, such as price discrimination
(examined in Chapter 12) and predatory pricing (pricing designed to drive competing firms out of business and/or deter prospective entrants so that the incumbent firm engag-
ing in such behavior can eventually charge higher prices). However, these actions are
illegal only if they “substantially lessen competition, or tend to create a monopoly.” The
Federal Trade Commission Act was also passed in 1914, creating a new federal agency charged with enforcing the antitrust laws (a duty it shares with the Justice Department)
and having the authority to prohibit “unfair” methods of competition, such as deceptive
advertising.
These laws form the cornerstone of antitrust policy. How well they have worked is a
matter of some dispute, and assessing the evidence is beyond the scope of this book. More-
over, over the past three decades, the extent to which the antitrust laws have been applied to
deter monopolies has diminished, with the notable exception of some recent cases against
Microsoft and Intel.
Part of the decline in the use of antitrust statutes is accounted for by the growing influ-
ence of the dynamic view of monopoly in the policymaking area. With international
competition growing and the pace of technological change accelerating, any control by a
supplier of a market at a given point in time is rendered more vulnerable, from a dynamic
perspective. For example, it is much harder for policymakers to attempt to prosecute Gen-
eral Motors or Ford today for having too large a share of domestic output, given the stiff
competition these companies now face from international rivals. An antitrust case against
semiconductor chip manufacturer Intel is today all the more difficult to prosecute, given
both the vigorousness of the competition Intel faces from overseas firms and the rapidity of
technological innovation in the market for semiconductor chips.
Of course, the election and tenure of some conservative and/or more middle-of-the-road
presidents since 1980 also partially explain the decline in the use of antitrust statutes. Well-
publicized cases against companies such as IBM, AT&T, and the top four ready-to-eat
cereal manufacturers were dropped during the 1980s, following the election of Republican
Ronald Reagan in 1980.
Regulation of Price Besides antitrust statutes, policymakers also rely on price regulation to deal with monop-
oly. In the case of local cable television distribution, for example, policymakers have relied
on rate controls to limit the prices that can be charged by firms, of which there is typically
only one per community. For example, whereas the profit-maximizing price might be $40,
policymakers can impose a ceiling, say $30, on the rate a local cable monopoly charges its
customers for monthly basic service.
antitrust laws a series of codes and amendments intended to promote a competitive market environment
300 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 300Trim Size: 203.2 mm X 254 mm
APPLICATION 11.5
One of the practices explicitly forbidden by the antitrust statutes is talking with one’s rivals in a market about fixing prices. The following conversation in 1982 between the CEOs of American (Robert Crandall) and Braniff (Howard Putnam) Airlines, who were engaged at the time in a fierce competition for passengers into and out of Dallas, provides an example of what one should not say to one’s rival about prices:16
Crandall: I think it’s dumb as hell . . . to sit here and pound the (deleted) out of each other and neither one of us making a (deleted) dime. . . . We can both live here and there ain’t no room for Delta. But there’s, ah, no reason to put both companies out of business. Putnam: Do you have a suggestion for me? Crandall: Yes, I have a suggestion for you. Raise your (deleted) fares 20 percent. I’ll raise mine the next morning. . . . You’ll make more money and I will, too. Putnam: We can’t talk about pricing! Crandall: Oh (deleted), Howard. We can talk about any (deleted) thing we want to talk about.
This conversation, secretly taped and turned over to the Justice Department by Putnam, led to price-fixing charges against Crandall and American Airlines, charges Crandall vigorously and successfully fought (on the basis of no price having actually been agreed to) in federal court.
What Not to Say to a Rival on the Telephone
Whereas talking directly to one’s rivals about fixing prices is explicitly forbidden by the Sherman Act, an airline industry tradition whose competitive implications are more difficult for antitrust authorities to assess is the practice of publishing fares with the Airline Tariff Publishing Company, a collectively owned computer network.17 Centralizing the price data made it easier for airlines to convey information to travel agents about the over 100,000 domestic fare changes occurring daily in the industry. Critics of the system contended, how- ever, that it enabled airlines to communicate pricing inten- tions to one another. According to these critics, the most questionable practice involved one airline that was typically not the dominant provider of service in a particular city try- ing to increase its passenger traffic by lowering its fares. The lower fares were entered in the computer system. The domi- nant carrier at the affected airport not only matched the new fares but also lowered its fares in other markets served by the carrier initiating the fare decrease. The dominant carrier sometimes even attached special codes to the new fares to emphasize its message. For example, certain carriers prefixed their new fares with the impolite code letters “FU” to con- vey their displeasure. In the end, the carrier initiating the fare reduction often canceled the change, and consumers ended up the losers. While airline officials denied any wrongdoing, the Justice Department scrutinized such fare games for their anticompetitive consequences and eventually found them to be in violation of antitrust statutes.
16“American Air Accused of Bid to Fix Prices,” Wall Street Journal, February 24, 1983, pp. 3 and 22.
17“Airlines May Be Using a Price-Data Network to Lessen Competition,” Wall Street Journal, June 28, 1990, pp. A1 and A6.
APPLICATION 11.6
The static versus dynamic views of monopoly were at the heart of the Microsoft antitrust case that was resolved through a consent decree that took effect in 2001 and a final judgment entered in 2002.18 The Justice Department,
Static versus Dynamic Views of Monopoly and the Microsoft Antitrust Case
which brought the antitrust case on behalf of the U.S. gov- ernment, alleged that Microsoft monopolized the market for personal computer (PC) operating systems. As of the late 1990s, Microsoft accounted for more than 90 percent of the U.S. market for PC operating systems through the dominance of its Windows product. The Justice Department further alleged that Microsoft had attempted to extend its monopoly power from the PC operating system market to the market for Internet browsers by tying its Internet Explorer to Windows, at the expense of the rival Navigator product produced by Netscape.
In its defense, Microsoft pointed to the dynamic nature of competition in the market for computer hardware and
18This application is based on David S. Evans, Franklin M. Fisher, Dan- iel L. Rubinfeld, and Richard L. Schmalensee, Did Microsoft Harm Con- sumers? Two Opposing Views (Washington, D.C.: AEI Press, 2000); Thomas L. Friedman, The Lexus and the Olive Tree (New York: Farrar, Straus and Giroux, 1999); Bill Gates, “We’re Defending Our Right to Innovate,” Wall Street Journal, May 20, 1998, p. A14; and Thomas W. Hazlett and George Bittlingmayer, “As Goes Microsoft, So Goes the Computer Industry,” Wall Street Journal, May 26, 1998, p. A18.
Publ ic Pol icy toward Monopoly 301
C11.INDD 07:33:13:PM 08/08/2014 PAGE 301Trim Size: 203.2 mm X 254 mm
We have already analyzed the effects of a price ceiling in a competitive market and
have seen that the results include reduced output, a shortage at the controlled price, and
nonprice rationing. A monopoly, however, may not respond to a price ceiling in the same
way. Indeed, under certain conditions a mandatory price reduction for a monopoly may
lead to increased output.
How can a lower price lead to greater output? Recall that a monopoly restricts output
in order to charge a higher price. A price ceiling means that a restriction in output can-
not result in a higher price, so the price ceiling eliminates the monopolist’s reason for
restraining output.
We can better understand the problem by focusing on how a price ceiling affects
the profit-maximizing output of a local cable monopoly. In Figure 11.11, the demand
curve is TD and the marginal revenue curve is TM. In the absence of any regulation, the most profitable output is QM, since marginal revenue and marginal cost are equal at that output, and the firm charges a $40 price. Now the government imposes a maximum
price of $30. As a result, the monopoly demand and marginal revenue curves effectively become PCD and PCFM—as in the case of a monopoly confronting a threat of entry that we examined in Section 11.4. Once we recognize the way price regulation affects the
demand and marginal revenue curves, the remainder of the analysis is straightforward.19
With the price regulation, QM is no longer the profit-maximizing output; at this output, marginal revenue (now $30) is greater than the unchanged marginal cost. What this
means is that the monopoly can recoup some of the lost profit by expanding production.
Note that marginal revenue exceeds marginal cost until output has been increased from
QM to Q, implying that profit rises as output expands over that range. The new profit- maximizing output is Q.
software products. For example, manufacturers boast that they will have newer, faster models of their computers out every five months, and none nowadays would dare think of guaranteeing, as IBM did with its AT (Advanced Tech- nology) desktop model in 1985, that a product will remain state-of-the-art for five years. Software that may domi- nate a market at any given moment is constantly vulner- able to being overthrown by superior versions produced by rival suppliers. Take the case of WordStar, the leading word-processing software program in the early 1980s. WordStar lost its position of preeminence to WordPerfect in the late 1980s. WordPerfect, in turn, lost its market- leading position to Word by the mid-1990s. Likewise, Lotus 1-2-3, the leading electronic spreadsheet software program throughout most of the 1980s, lost its position to Excel in the 1990s.
Due to the dynamic nature of competition in PC hard- ware and software markets, Microsoft argued that, over time, consumers stand to gain a great deal (in terms of consumer surplus) from the innovations that result from intense competition between rival suppliers. Tying Inter- net Explorer to Windows, according to Microsoft, is but one example of such innovation that makes consumers better
off. For example, numerous reviews from the trade press praised Microsoft when it integrated Internet Explorer into Windows because of the benefits that were likely to accrue to consumers.
Moreover, Microsoft argued that government interven- tion against the exercise of monopoly power at any given point in time by a supplier who has brought a superior prod- uct to market hurts the producer of the product, as well as discourages other firms from innovation, and thereby harms consumers. Indeed, in a study by George Bittlingmayer and Tom Hazlett, the stock market valuation of a broad portfolio of computer companies is, in general, adversely affected by judicial and regulatory decisions against Microsoft.
The Justice Department’s reasoning in the Microsoft antitrust case would predict the opposite effect. Microsoft’s actions, according to the Justice Department, diminished competition in the software market and thereby harmed consumers. The stock prices of computer companies thus should rise when Microsoft is restrained by government intervention. The Bittlingmayer and Hazlett results, how- ever, suggest that antitrust restraints against Microsoft deter competition and innovation in computer software markets and thereby diminish consumer welfare.
19We again assume that average cost is low enough for it to be profitable for the monopolist to continue to operate.
302 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 302Trim Size: 203.2 mm X 254 mm
In this case, the mandatory lower price leads to greater output and reduced profit for the
monopoly. Although the firm recoups some of the initial loss in profit by expanding output
from QM to Q, the net result is still a loss in profit. We can see this even without using the average cost curve, by noting that the monopoly could have chosen to produce Q at a price of $30 before the price control, but did not because profit was higher at QM and a price of $40. (In fact, profit is higher by the area ACF. Can you see why?) The regulation essen- tially confronts the monopoly with a horizontal demand curve over the zero-to-Q range of output, just like the demand curve facing a competitive firm, and therefore eliminates the
reason for restricting output.
Thus, this price regulation reduces monopoly profit and benefits consumers by lowering
price. But in this case it also does more: it increases output to the efficient level, eliminating
the deadweight loss (from a static perspective) of monopoly! At the initial monopoly out-
put, the deadweight loss is area BCA, the sum of the excess of the marginal value of units from QM to Q over their marginal cost. (This assumes that the monopoly marginal cost curve shows the relevant opportunity cost of the resources, which may not always be the
case—but we will assume to be true here.20) By inducing an expansion in output to Q, the price regulation results in the monopoly reaching a level of production where the marginal
value of the good is equal to marginal cost, as required for efficiency.
Of course, this beneficial outcome is not as easy to achieve as our analysis suggests.
First, the outcome depends on where the price ceiling is set. In the diagram, if the price is
set at either a higher or lower level than $30, the monopolist will choose to produce less.
(You may want to confirm this.) From the point of view of promoting efficiency, the price
should be set where the marginal cost curve intersects the demand curve, but because the
government doesn’t know the monopolist’s marginal cost curve that outcome may not be
achieved. Second, the price must not be lowered to the point where the monopolist suffers
20For a discussion of the case when the monopoly’s marginal cost curve is upward sloping, see Edgar K. Browning, “Comparing Monopoly and Competition: The Increasing-Cost Case,” Economic Inquiry, 25, No. 3 (July 1987), pp. 535–542.
Figure 11.11 Price Ceiling Applied to Monopoly When a price ceiling of P is applied to the monopoly, the demand curve becomes PCD, and the marginal revenue curve becomes PCFM. The most profitable output is Q, the efficient level of output.
$40 = P
$30 = P
T
Dollars per unit
0 OutputQ Q
B
C
A
D
F
M
MCM
M
Publ ic Pol icy toward Monopoly 303
C11.INDD 07:33:13:PM 08/08/2014 PAGE 303Trim Size: 203.2 mm X 254 mm
losses and goes out of business. Third, the monopoly has an incentive to skirt the price ceil-
ing by reducing the quality of its product. Producing a lower-quality, lower-cost product is
one way the monopolist can avoid the drop in profit that the price control otherwise causes.
If the monopolist can pursue this strategy, a price ceiling will not achieve efficiency since
the wrong-quality product will be produced.
APPLICATION 11.7
Over the past three decades, most European Union (EU) countries have sought to keep pharmaceutical prices from rising in real terms.21 This is in contrast to the United States, which does not directly regulate pharmaceutical prices.
The static efficiency effects of the EU policy have been predictable. Between 1986 and 2004, pharmaceutical prices rarely rose more than a few percentage points faster than general consumer prices in any given year. By com- parison, real pharmaceutical prices in the United States increased by 45 percent over the 19-year time period and only moderated in years where there appeared to be a political threat of price regulation. For example, in 1993, the Clinton administration proposed instituting regulations that would limit pharmaceutical price increases at or below the average consumer price inflation. In response to this pro- posed legislation, 21 major pharmaceutical firms pledged to limit their price increases to remain, at least temporarily, below the average rate of consumer price inflation.
The dynamic effects of the different policy approaches in the EU versus the United States on consumers and efficiency appear to be different from the static effects. Whereas EU-based pharmaceutical companies were out- spending their American counterparts on research and development by about 24 percent in 1986, by 2004 U.S.
Efficiency and the Regulation of Pharmaceutical Drug Prices in the European Union and the United States
pharmaceutical firms were outspending their European counterparts by 15 percent. For the 1986 to 2004 time period, economists Joseph Golec and John Vernon of the University of Connecticut estimate that price regulation by EU countries resulted in $5 billion in forgone R&D spend- ing, 1,680 fewer research jobs, and 46 new medicines never being introduced. Furthermore, Golec and Vernon project that the long-run costs associated with the forgone R&D spending—in terms of jobs lost and new medicines never introduced—may be 10 to 20 times larger for the EU once appropriate account is made for the gestation period for developing new medicines.
Golec and Vernon predict that were the United States to impose pharmaceutical price regulation—either through direct means or through promotion of greater reimporta- tion of drugs sold by U.S. companies in other countries— the dynamic losses would be even greater.22 For example, they estimate that U.S. companies would have forgone $427 billion (2004 dollars) in R&D spending and devel- oped 974 fewer medicines over the long run had the United States also limited pharmaceutical price increases to the general rate of inflation.
21This application is based on Joseph H. Golec and John A. Vernon, “Financial Effects of Pharmaceutical Price Regulation on R&D Spending by EU Versus U.S. Firms,” Pharmacoeconomic 28, No. 8 (2010), pp. 615–628.
22In the next chapter we will see the reasons pharmaceutical com- panies may find it profitable to sell their products for a lower price in countries such as Canada and Australia than they do in the United States. At the present time, for the purported reason of quality assurance, the Food and Drug Administration prohibits the reimportation of pharmaceuticals sold by U.S. companies in other countries back to the United States market.
SUMMARY
A monopoly is the sole seller of some product with-
out close substitutes.
A monopoly confronts the market demand curve for
the product it sells, and the demand curve will generally
be downward sloping.
With a downward-sloping demand curve the monop-
oly’s marginal revenue is less than price because price
must be reduced to sell a larger output.
If a monopoly can select any price–quantity combi-
nation on its downward-sloping demand curve but must
304 Monopoly
C11.INDD 07:33:13:PM 08/08/2014 PAGE 304Trim Size: 203.2 mm X 254 mm
charge the same price to all customers, profit is maxi-
mized by producing the output for which marginal cost
equals marginal revenue. The price of the product will
be higher and its output lower under monopoly than
under competitive conditions.
From a static perspective, the output restriction
characteristic of monopoly represents a misallocation
of resources and involves a deadweight loss. Because
price is above marginal cost, greater output would be
worth more to consumers than it would cost to produce.
The size of the deadweight loss due to restricted out-
put is shown by the triangular area between the demand
and marginal cost curves from the monopoly output to the
competitive output (where price equals marginal cost).
Antitrust laws and price regulation are two policies that
can, in principle, reduce monopoly’s static deadweight loss.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solutions given in Answers to Selected Problems at the back of the book (pages 528–535).
11.1 Because they result in higher prices than perfect competi- tion, policymakers often blame monopolies for causing infla-
tion, where inflation is a persistent increase in the general price
level. Is it appropriate to assign such blame to monopolies?
Explain.
11.2 Suppose that we, as consumers, have the option of having an AIDS vaccine produced by a monopoly or of not having the
vaccine produced at all. Under which option would we be bet-
ter off? Why?
11.3 “Because a monopoly is the only source of supply, con- sumers are entirely at its mercy. There is no limit to the price
the monopoly can charge.” Evaluate this statement.
11.4 Why will Disneyland never set its admission price at a level where its demand curve is inelastic? Use the total revenue
and total cost curves to illustrate your answer.
11.5 At the profit-maximizing output the price of Tommy jeans is twice as high as marginal cost. What is the elasticity of
demand? (Hint: Solve MR = P[1−(1/η)] for η and remember that MC = MR.)
11.6 When a ski resort with some monopoly power is maximiz- ing profit, price is greater than marginal cost. Thus, consum-
ers are willing to pay more for additional lift tickets than the
tickets cost to produce. So why does the ski resort not charge a
lower price per lift ticket and increase output?
11.7 “A competitive firm will never operate where marginal cost is declining, but a monopoly may.” True or false? Explain.
*11.8 Show how the most profitable output and price are determined for a monopoly that can produce its product at
zero cost (MC = AC = 0). Explain the deadweight loss that exists in this case.
11.9 Draw a diagram to show the deadweight loss of a monop- oly with a marginal cost curve that is vertical at the profit-
maximizing output level.
*11.10 “The concept of opportunity cost teaches us that pro- ducing more of any good, including a good produced by a
monopoly, means that we must produce less of other goods.
Thus, there is no objective basis for saying that an increase in
a monopolist’s output is worthwhile.” Evaluate this statement.
11.11 Suppose that there is a single seller of gasoline in a partic- ular town. Suppose that policymakers, outraged by the prices
charged by this monopoly seller, impose a price ceiling. Will
the seller’s output increase? Explain your answer.
11.12 “Since the shutdown condition applies only to competi- tive firms, it is not a relevant factor when considering what
profit-maximizing output level a monopolist such as Amazon
should produce.” Explain why you agree or disagree with this
statement.
11.13 Suppose that the MC faced by Skechers is a constant $10 per shoe. If the demand elasticity for Skechers shoes is also
constant and is equal to 5, what price should Skechers charge
for its shoes?
11.14 Suppose that the (inverse) market demand curve for a new drug, Adipose-Off, designed to painlessly reduce body fat, is
represented by the equation P = 100 − 2Q, where P is the price in dollars per dose and Q is the annual output. (The marginal revenue curve is thus given by the equation MR = 100 − 4Q.) Suppose also that there is a single supplier of the drug who
faces a marginal cost, as well as average cost, of producing the
drug, equal to a constant $20 per dose. What are the monopo-
list’s profit-maximizing output and price? What is the resulting
deadweight loss relative to the competitive outcome?
11.15 Suppose that in the preceding problem, the government levies an excise tax of $5 per dose on the monopolist. What
would happen to the monopolist’s profit-maximizing output
and price? What would happen to consumer and producer sur-
plus? How much money would the government collect due to
the tax? What would be the size of the resulting deadweight
loss relative to the competitive outcome?
11.16 Address all the questions in the preceding problem but assume that instead of a tax of $5 per dose the government
offers a subsidy of $5 per unit.
11.17 “A monopolist like Spago (a famous Hollywood restau- rant frequented by movie stars) can fully pass on all marginal
cost increases to its diners through higher prices since it is a
price maker and can charge any price it wishes.” Do you agree
or disagree with this statement? Explain your answer.
Publ ic Pol icy toward Monopoly 305
C11.INDD 07:33:13:PM 08/08/2014 PAGE 305Trim Size: 203.2 mm X 254 mm
11.18 Calculate the Lerner index for the monopoly described in Question 11.14. How would the value of this index change when
the tax described in Question 11.15 is imposed on the monopo-
list? If the subsidy in Question 11.16 is imposed instead?
11.19 “A monopoly’s marginal cost curve is the monopoly’s supply curve.” True or false? Explain your answer.
11.20 Explain the determinants of a firm’s monopoly power. How can a firm have monopoly power if it is not the sole sup-
plier of a product?
11.21 Suppose that a monopoly is producing at an output where its average total cost of production is minimized and equals
$50 per unit. If marginal revenue equals $60, is the monopoly
producing at the profit-maximizing output level? Explain why
or why not.
11.22 Provide an example of a firm with a Lerner index value of (a) zero and (b) unity. Why does the Lerner index take on
a value between these two extremes? Explain why the Lerner
index measures a firm’s monopoly power.
11.23 Marin County Enterprises has a monopoly on the pro- duction of lunar-powered homes and has the normal U-shaped
average cost curve. At its present profit-maximizing output and
price, it is able to earn a positive economic profit. Graphically
show the effects in the product market (output, price, profit,
and so on) of each of the following changes:
a. Lunar-powered homes become a nationwide fad. b. The cost of labor (a variable factor of production) rises. c. The rent for the firm’s office space (a fixed factor of produc-
tion) rises.
If the Federal Alternative Power Commission can regulate
the prices of lunar-powered homes and the promotion of effi-
ciency is the commission’s goal, what price should it set? What
will happen to the output and profit of Marin County Enter-
prises as a result?
11.24 The City of Berkeley is currently considering alterna- tive ways of providing cable service to its citizens. Based on
an econometric analysis of several recently awarded cable
franchises in other cities, economists have determined that the
total cost, TC, and inverse demand curves for a cable company in Berkeley would be:
TC Q Q Q P Q
= − + = −
2 0 1 0 005
20 0 5
2 3. .
. ,
and
where output, Q, is measured in thousands and P is the monthly basic tier price.
a. Given this information, what are the equations for the total and marginal revenue curves, and what do these curves look
like on a graph?
b. City Councilor A believes the city should own and operate a cable system for the purpose of making as much profit as
possible. The profit would be used to lower the city govern-
ment’s deficit. If Councilor A gets her way, what will be the
price and output of cable and by how much will the city-
owned system be able to reduce the city’s deficit?
c. Councilor B believes the city should produce as much cable service as possible without losing money (i.e., the
city should provide cable to its citizens on a nonprofit
basis). If Councilor B gets his way, what output and price
will result?
d. Councilor C believes that the private sector should provide cable to the city but that the single, private firm that gets
the city’s franchise should pay 10 percent of its total rev-
enue back to the city in the form of an annual franchise fee.
If Councilor C gets her way and the franchise is awarded
to the firm promising to pay the largest franchise fee, what
price and output will result? What will be the size of the
annual franchise fee?
11.25 Suppose that the Berkeley City Council takes 10 years to award its first cable television franchise for the sake of ensur-
ing that the price the franchised operator charges is as close to
average cost as possible. Explain why such a strategy may do
less to promote consumer surplus than the alternative strategy
of awarding the franchise right away to an operator who will
charge a monopoly price.
11.26 The prices of seats on major financial exchanges have plummeted dramatically in recent years. For example, at the
Chicago Board of Trade, the world’s biggest futures exchange,
a membership seat sold for $300,000 in 2012, down 65 per-
cent from a record of $858,000 in 1997. Explain, using a graph,
why the decline in the value of such seats may be related to the
growth of electronic trading.
11.27 Movie actors and directors are often paid a percentage of a film’s total revenue since revenue is easier to maintain than
film studio profit. In light of this, explain why actors and direc-
tors are incented to push for a lower price/higher output com-
bination than the one preferred by the studios associated with
their films.
306
C12.INDD 12:44:56:PM 08/06/2014 PAGE 306Trim Size: 203.2 mm X 254 mm
Our analysis of monopoly has thus far been based on the assumption that the monopo- list charges a single price to all customers, and we have identified the profit-maximizing
price and output based on this assumption. In many cases, however, firms with monopoly
power charge different prices to different customers or even to the same customers depend-
ing on the quantity purchased. The practice of charging different prices for the same prod-
uct when there is no cost difference to the producer in supplying the product is called price discrimination.
Examples of price discrimination include amusement parks, such as Disneyland, that
use season passes to offer the first day of admission at a relatively high price and any addi-
tional day of admission for the rest of the year at a price of zero; the sale of discounted air-
line seats, hotel rooms, and rental cars through online merchants such as Travelocity; hotel
chains and metropolitan bus services that feature discount rates for senior citizens; laundry
services that charge more to dry clean women’s blouses than men’s shirts; book publishers
price discrimination the practice of charging different prices for the same product when there is no cost difference to the producer in supplying the product
Learning Objectives
Explain price discrimination, the various degrees of price discrimination, and how price discrimination can increase a firm’s profit. Spell out the three necessary conditions for a firm to be able to engage in price discrimina- tion. Demonstrate how, under third-degree price discrimination, market segments that have less elastic demand end up being charged a higher price, all else being equal. Show how intertemporal price discrimination, a type of third-degree price discrimination, can increase a firm’s profit. Explore how two-part tariffs, a form of second-degree price discrimination, can increase a firm’s profit.
Memorable Quote “I have received hundreds of emails from iPhone customers who are upset about Apple dropping the price of iPhone by $200 two months after it went on sale. After reading every one of these emails, I have some observations . . . even though we are making the right decision to lower the price of iPhone, and even though the technology road is bumpy, we need to do a better job tak- ing care of our early customers as we aggressively go after the new ones with a lower price. Our early customers trusted us, and we must live up to that trust with our actions in moments like these. Therefore, we have decided to offer every [prior] iPhone customer . . . a $100 store credit towards the purchase of any [Apple] product.”
—Steve Jobs, Former Apple CEO
Product Pricing with Monopoly Power12CHAPTER
Price Discr iminat ion 307
C12.INDD 12:44:56:PM 08/06/2014 PAGE 307Trim Size: 203.2 mm X 254 mm
that charge more for the early printings of a title (the hardback edition) than later printings
(the paperback edition); the frequent-flyer programs offered by airlines; and restaurants in
Paris that list a higher price for a dish on the English version of their menu than they list for
the same dish on the French version.
Why these more complicated pricing practices arise and what consequences they have
are this chapter’s subjects. As we will see, price discrimination can increase a firm’s profit
as well as the total surplus (consumer surplus plus producer surplus) generated by a market.
12.1 Price Discrimination Let’s begin our discussion of price discrimination with a simple example. Consider a
monopoly that produces a product of which each consumer will purchase no more than one
unit. An example might be a monthly fee for Internet access—few people would purchase
two subscriptions for the same Internet access, even at a very low price. The demand curve
confronting the monopoly, shown as D in Figure 12.1, slopes downward because consum- ers are willing to pay different prices for Internet access. At a lower price, a larger quantity
of subscriptions can be sold as more consumers sign up for Internet access, but each con-
sumer purchases only one subscription. Marginal cost is constant at $10 per monthly sub-
scription. If the monopoly must set one price for all consumers—the assumption we made
in Chapter 11—price is $15, output is 100 units, and profit is given by the rectangular area
A. (We have not drawn in the marginal revenue curve to keep the diagram simple.)
First-Degree Price Discrimination Now let’s see how the monopolist might be able to increase its profit by charging nonuni-
form prices. Note that there is a person willing to pay just below the $15 price, say $14.95,
for Internet access. If the monopoly could charge her $14.95, she would purchase the
Price Discrimination Can Increase Profit When a uniform price is charged, the maximum profit is shown by area A, assuming MR = MC at Q1. However, if a different price can be charged for each unit sold, it may be possible to realize areas A plus B plus C as monopoly profit.
$15 = P1
$20
$10
Dollars per subscription
0 Output (subscriptions)Q1 Q2
B
C
A
D
MC
(100) (200)
Figure 12.1
308 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 308Trim Size: 203.2 mm X 254 mm
product. Moreover, the monopoly’s profit would rise by $4.95 from the transaction ($14.95
minus the $10 marginal cost of the 101st unit) as long as the monopoly does not have to
reduce the $15 price to the first 100 customers. (Recall that it is the lost revenue from low-
ering the price on initial sales that makes marginal revenue lower than price. If you don’t
have to lower the price on the first 100 units, the price received for the 101st unit is its mar-
ginal revenue.) But the monopolist doesn’t have to stop there. If the 102nd unit can be sold
for $14.90, then profit will rise by another $4.90, as long as the prices charged for the first
101 units are not lowered. Indeed, additional profit can be realized in this way all the way
out to an output of 200: each unit can be sold for more than it costs to produce. In this way,
the monopolist can increase profit by area B (which is the sum of the excess of price paid over marginal cost of each unit beyond 100).
Note that this pricing procedure doesn’t have to be restricted to units from 101 to 200;
all units can be priced at the maximum price each consumer will pay. The very first unit
can be sold for, say, $20, the second for $19.95, and so on. Under these conditions, the
marginal revenue curve relevant for the monopoly’s output decision coincides with the
demand curve, so the MR = MC rule for profit maximization leads to an output of Q2, as already explained. Then the monopoly’s profit is given by the sum of areas A, B, and C, substantially higher than when the single price of $15 is charged for all units.
This pricing policy, in which each unit of output is sold for the maximum price a con-
sumer will pay, is called first-degree or perfect price discrimination. It is perfect from the monopolist’s point of view because the monopolist makes the maximum profit given
the demand curve. If any higher price is charged on any unit, then that unit would not be
sold and profit would be smaller. In effect, the monopoly has been able to capture all of the
consumer surplus (areas A + B + C ) as its profit. A monopolist can do no better than that. Perfect price discrimination, if a monopoly can practice it, has some notable conse-
quences. In addition to increasing the monopolist’s profit, the resulting output is efficient.
This is in sharp contrast to the market outcome when a monopolist can charge only a single
price and (at least from a static perspective) output ends up being less than the efficient
output. With perfect price discrimination, even though consumers receive no net benefit (or
just the tiniest amount necessary to induce them to buy), every unit with a marginal value
to consumers greater than the marginal cost of producing it is, in fact, produced. All the
potential net benefit from producing the good goes to the monopolist as profit, but it is just
a transfer of income from consumers to the monopolist and not a net loss to society. All the
potential benefit is realized by some member of society, which is what efficiency entails.
Implementing First-Degree Price Discrimination Implementing first-degree price discrimination is, as you might expect, not easy. It requires
some mechanism by which a monopolist can determine the maximum amount each poten-
tial customer is willing to pay for the product. Asking potential customers is no good
because customers have an incentive to understate what they are willing to pay if they will
be charged accordingly. Furthermore, there tends to be no indirect means of securing such
information from potential customers. For example, car dealers do not (yet!) have a device
that automatically registers the maximum amount a prospective customer is willing to pay
for a car when the customer walks into the showroom.
Although perfect price discrimination may be rare (if not nonexistent), there are cases in
which different consumers are charged different prices, and the preceding analysis explains
why they arise: producers are trying to increase their profits by approximating this type
of pricing. For example, lawyers and doctors often charge wealthy customers more than
poor customers. Many car dealers also strive to approximate first-degree price discrimina-
tion. Although they are unable to perfectly estimate what each potential customer is will-
ing to pay for a car, most dealers employ certain tactics to elicit at least a rough guess. For
instance, the salesperson may claim to share the customer’s goal of getting the best possible
first-degree (perfect) price discrimination a policy in which each unit of output is sold for the maximum price a consumer will pay
Price Discr iminat ion 309
C12.INDD 12:44:56:PM 08/06/2014 PAGE 309Trim Size: 203.2 mm X 254 mm
deal and promise to bargain very hard with the dealership owner on the customer’s behalf.
However, the sales representative will first attempt to get the customer to make an initial
bid without committing the dealership to any selling price—all the while sizing up how
badly the customer would like to buy a new car, how stuck the customer is on a particular
model and color, the customer’s financial resources, and so on. Once the customer is com-
mitted to a bid, the representative disappears (supposedly to meet with the owner of the
dealership on the customer’s behalf) and then reappears, typically to report that the custom-
er’s bid isn’t quite good enough. The owner wants at least X dollars more, at a bare mini- mum. Another round of negotiation is initiated, through which the salesperson attempts to
get a higher bid from the customer. Because of this tactic, various buyers may pay a wide
variety of prices for the same make and model of car, and dealers’ profits are higher than
they otherwise would be.
Second-Degree Price Discrimination Price discrimination occurs when different prices are charged for the same good, with the
per-unit price varying either across consumers or across separate units purchased by the
same consumer, or both. In the case of first-degree price discrimination, each consumer is
charged a different price equal to the maximum amount he or she is willing to pay. In the
case where consumers are willing to buy more than one unit, they are charged a different
price for each successive unit, with the schedule of prices set to extract their entire con-
sumer surplus.
As discussed in the preceding section, first-degree price discrimination is rare to nonex-
istent, if for no other reason than the difficulty of knowing each consumer’s demand curve.
Some pricing practices, however, represent rough attempts to approximate perfect price
discrimination.
Second-degree price discrimination, or block pricing, is the name given to a schedule of prices such that the price per unit declines with the quantity purchased by a particular
customer. It is distinguished from first-degree price discrimination in that the same price
schedule confronts all consumers; the price schedule is not perfectly individually tailored
as in the first-degree case.
An example of second-degree price discrimination is depicted in Figure 12.2. Suppose
that an electric utility prices the first 100 kilowatt hours per month at $0.12 per kilowatt
hour, the second 100 at $0.10, the third at $0.08, and so on. Just as in the first-degree case,
block pricing can increase a firm’s profit by extracting additional consumer surplus on ini-
tial units consumed. It also tends to result in greater (more efficient) output because heavy
users pay prices closer or equal to marginal cost. It does not, however, convert all potential
consumer surplus into monopoly profit—as perfect price discrimination does. To see this,
note that based on the electricity consumer’s demand curve depicted in Figure 12.2, the
consumer is willing to pay more than $0.12 per unit for each kilowatt hour less than 100
and thus realizes some surplus when the utility charges only $0.12 per output unit. Like-
wise, the consumer is willing to pay more than $0.10 per hour for each kilowatt hour in
excess of 100 but less than 200 and so realizes some added consumer surplus when the util-
ity charges $0.10 per unit over this output range.
Third-Degree Price Discrimination Third-degree price discrimination, or market segmentation, occurs when each con- sumer faces a single price and can purchase as much as desired at that price, but the price
differs among categories of consumers. Because this is probably the most common type of
price discrimination, we will examine it more fully. There are many examples of this pric-
ing practice. Your college bookstore quite possibly sells books to faculty members at a dis-
count, charging them a lower price than it charges students for the same books. Telephone
second-degree price discrimination (block pricing) the use of a schedule of prices such that the price per unit declines with the quantity purchased by a particular consumer
third-degree price discrimination (market segmentation) a situation in which each consumer faces a single price and can purchase as much as desired at that price, but the price differs among categories of consumers
310 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 310Trim Size: 203.2 mm X 254 mm
APPLICATION 12.1
Second-degree price discrimination is actively practiced by most major airlines. Frequent flyers pay less as they fly more. Frequent-buyer programs are also employed by hotels, fast-food chains, airport parking lots, supermar- kets, and financial services firms. For example, Marriott’s Honored Guest program gives participating consumers a free weekend night at any domestic Marriott hotel after they have accumulated a certain number of points in the program. (Each dollar the customer spends on a Marriott guest room earns 10 points.) Park and Fly, a chain that oper- ates off-terminal parking lots near major airports in sev- eral major U.S. cities, has a “frequent-parker” program that offers customers a week’s free parking after they have paid for 35 days.
Most major fast-food chains offer value packages whereby consumers are offered a discount the more food they order at any one point in time. McDonald’s, for example, offers a “Value Meal” whereby the price of a Big Mac, french fries, and soft drink is lower if the items are purchased as a package than if each item is purchased separately.
Supermarkets have also been moving toward a tiered- pricing system by offering lower prices to shoppers with greater loyalty. As of 2013, more than 75 percent of U.S. households belonged to at least one supermarket fre- quent shopper club. Supermarkets relying on such pro- grams employ computerized scanning systems that sort
Giving Frequent Shoppers the Second Degree
customers based on the volume of their purchases and also can dispense coupons at the checkout line accordingly.
In the financial services area, Merrill Lynch offers a client-reward program that bases the annual fee on the amount of money investors maintain in Merrill accounts. “Bronze” investors who maintain at least $100,000 in accounts are charged an annual fee of 1.5 percent and are given 12 commission-free stock or bond trades per year. “Platinum” investors who maintain at least $5 million in Merrill accounts are charged an annual fee of 0.84 per- cent and receive 75 commission-free stock or bond trades per year.
One of the hottest recent Internet startups, Groupon, also relies on second-degree price discrimination by orga- nizing prospective customers for a particular product or service into groups. The site offers subscribers at least one deal a day in their city—say half off a meal at a local Ital- ian restaurant or a theatrical performance at a nearby stage. Provided that a certain number of people commit to a particular promotion, the deal kicks in, or “tips,” in Grou- pon parlance. Once the deal tips—for example 50 people commit to buying a $40 coupon toward the $80 theatrical show—the seller and Groupon split the associated revenues equally, and a group of customers get a bargain because of the collective volume of their purchases. By late 2010, Groupon had grown so rapidly that it was able to turn down a $6 billion buyout offer from Google.
Second-Degree Price Discrimination: Block Pricing An electric utility charges $0.12 per kilowatt hour per month for the first 100 units, $0.10 for the second 100 units, $0.08 for the third 100 units, and so on. The unit price (depicted by the step-like red curve) thus depends on the quantity purchased by a consumer.
$0.12
$0.10
$0.08
Dollars per kilowatt
0 Consumption (kilowatt hours)
D
100 200 300
Figure 12.2
Three Necessary Condit ions for Pr ice Discr iminat ion 311
C12.INDD 12:44:56:PM 08/06/2014 PAGE 311Trim Size: 203.2 mm X 254 mm
companies charge higher monthly rates for business phones than for home phones. Many
drugstores offer senior citizens discounts on drug purchases. Movie theaters typically
charge lower prices to children, senior citizens, and students. Grocery stores offer certain
items at lower prices to customers with coupons. In all these cases the same product is sold
to different groups for different prices.
12.2 Three Necessary Conditions for Price Discrimination Although there are many examples of price discrimination, especially third-degree price
discrimination, all of them are predicated on the satisfaction of certain conditions. First, the
product seller must possess some degree of monopoly power, in the sense of confronting
a downward-sloping demand curve. (It isn’t necessary that the firm be a pure monopoly—
that is, the only seller—just that the firm have some monopoly power.) In the absence of
monopoly power, a seller is not able to charge some customers higher prices than others.
Second, the seller must have some means of at least roughly approximating the maxi-
mum amount buyers are willing to pay for each unit of output. To practice third-degree
price discrimination, for example, the seller must be able to separate customers into two or
more identifiable market segments and the price elasticity of demand must differ among the
segments. As we explain next in more detail, this condition makes it profitable for the seller
to charge a higher price to the market segment with the more inelastic demand.
Third, the seller must be able to prevent resale, or arbitrage of a product among the mar- ket segments. If this condition is violated, the likelihood that a seller will be able to engage
in price discrimination is significantly undermined. Suppose, for example, that General
Motors tries to price discriminate by selling automobiles to senior citizens at a 20 percent
discount. How many automobiles would it sell at the higher, normal price? Very few, we
would predict. Senior citizens would simply buy cars at a discount and then resell them at a
higher price (still below GM’s normal price). A similar result would occur if people got
their parents or grandparents to purchase cars for them. Resale of a product undermines the
seller’s ability to sell at the higher price.
resale arbitrage of a product among market segments
APPLICATION 12.2
Professor Ian Ayres of Yale University employed a team of research assistants to explore whether car retailers priced their products differently on the basis of race or gender.1
Apart from race and gender, the research assistants were selected for uniformity of age, education, economic class, occupation, address, and attractiveness. The research assistants were also all trained to use the same bargaining tactics.
Based on the study, Professor Ayres found that after multiple bargaining rounds, the lowest price offered by
The Third Degree by Car Dealers
a dealer for a new car results in an average profit to the dealer of $362 per white male customer, $504 per white female customer, $783 per black male customer, and $1,237 per black female customer. The most likely expla- nation for this pattern is third-degree price discrimination, according to Ayres. To the extent that white men, on aver- age, are believed by car dealers to have superior access to information about the car market and less aversion to bar- gaining, profit-maximizing behavior by car dealers would encourage such a pricing pattern. Naturally, this conclu- sion presumes that information and competition in the market for new cars are sufficiently imperfect to allow an individual dealership some pricing power over individual customers.
1Ian Ayres, “Fair Driving: Gender and Race Discrimination in Retail Car Negotiations,” Harvard Law Review, 104, No. 4 (February 1991), pp. 817–872.
312 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 312Trim Size: 203.2 mm X 254 mm
If resale of a product is relatively easy, price discrimination can’t be very effective.
How, then, can resale be prevented? Sometimes, the nature of the product itself prevents
resale. Electricity provided to a local business can’t be resold to a nearby homeowner. If
you receive a medical checkup, there is no way you can transfer it to a friend. Children who
attend movies cannot reproduce the entertainment for their parents. In general, goods that
are immediately consumed—a common characteristic of services—are not as susceptible
to resale. In contrast, manufactured items, like automobiles, appliances, and clothes, can be
purchased by one person and later turned over to someone else. As a result, price discrimi-
nation is less common in the sale of manufactured goods.
In addition to the foregoing three necessary conditions for practicing price discrimina-
tion, it is essential to understand that price discrimination is not costless and may often
require establishing mechanisms to be able to determine how much different consumers are
willing to pay for the same product as well as to prevent resale across consumers. These
costs may be nontrivial, and as we will see in a later application (Application 12.7), may be
sufficient in magnitude to preclude the effective practice of price discrimination.
APPLICATION 12.3
Over 10 million Americans turn to Canada for their pre- scription drugs.2 America’s senior citizens are increasingly being joined by union members and even beneficiaries of government plans who are ignoring the warnings of the U.S. Food and Drug Administration (FDA) that such purchases may be unsafe, threats to choke off supplies by pharma- ceutical companies who manufacture the drugs, and even a U.S. law that bans certain cross-border purchases (those that are not for personal use, in small amounts, and for pre- scriptions exceeding 90 days).
The basic reason for the cross-border purchases is the arbitrage opportunity that exists because of the third-degree price discrimination engaged in by pharmaceutical compa- nies across countries. Prices for the same prescription drug tend to be at least 40 to 50 percent lower in Canada than in the United States (and the same is true in other countries, such as Mexico, vis-à-vis the United States). At least part of the reason drug prices are lower in other countries such as Canada is government regulation. Canada, for example, has a universal health-care system that applies price controls to pharmaceuticals and therefore ends up requiring drug com- panies to charge a lower price in the Canadian market than they are able to charge in the U.S. market.
There are three basic ways that American consumers are arbitraging the price differences in prescription drugs across countries. First, some Americans physically travel to other
Gray Hairs and Gray Markets
countries such as Canada and Mexico to have their prescrip- tions filled at brick-and-mortar pharmacies. For instance, day-long bus tours to Canada and Mexico for $99 and catering largely to senior citizens have sprung up through- out bordering states such as Minnesota, Washington, Michi- gan, Arizona, Texas, and New York.
Second, many American drug consumers opt to make purchases from retailers located in other countries through mail order or the Internet. It is estimated that more than 1 million Americans currently use mail order or Internet chan- nels to access drugs from Canada. And representatives from the FDA have testified before Congress that they do not intend to interfere with such Internet sales.
Third, U.S. organizations that purchase larger quantities of prescription drugs (such as pharmacy chains and health maintenance organizations) can buy the drugs from foreign wholesalers and then turn around and sell them in the U.S. market, marked up to just below current domestic prices. In 2003, for example, Springfield, Massachusetts (total popu- lation of 152,000), became the first U.S. city with a Cana- dian drug plan. Springfield supplied 1,600 of its insured employees, retirees, and dependents with drugs purchased from a company headquartered in Ontario, Canada, at an average savings of 40 percent per prescription and a total annual savings of $9 million to the city. Even U.S. Veterans Administration (VA) clinics around the country have begun purchasing prescription drugs from wholesalers in countries such as Canada, Colombia, Pakistan, India, Singapore, and Israel. The VA purchases are made at substantial discounts from the prices prevailing for drugs in the United States, and at the same time that another agency of the federal government, the FDA, is warning the American public that it is dangerous to make such cross-country purchases.
2This application is based on “Pfizer Moves to Try to Stop Drugs from Canada,” New York Times, January 14, 2004, pp. 1 and 7; “‘Buy Canada’ Drug Plan Sweeping U.S.,” Toronto Star, October 26, 2003, p. A6; “Bitter Pill,” Tulsa World, November 2, 2003, p. G6; and “Cheaper Drugs May Carry a Higher Price,” Insight, December 8, 2003, p. 29.
Price and Output Determinat ion with Pr ice Discr iminat ion 313
C12.INDD 12:44:56:PM 08/06/2014 PAGE 313Trim Size: 203.2 mm X 254 mm
12.3 Price and Output Determination with Price Discrimination
Imagine a monopoly that is initially selling 1,500 units of output at the uniform price of $10.
Suppose that it can practice third-degree price discrimination by separating customers into
two identifiable market segments, segment A and segment B, and preventing resale of the
product between the segments. Therefore, the monopolist may charge a different price to each
segment. However, for price discrimination to be worthwhile, the monopolist must be able to
sell the 1,500 units for a higher total revenue by charging each segment a different price.
Price Determination When the demand elasticities differ for the two market segments, the monopolist can increase
total revenue from selling a given quantity by charging different prices. Suppose that when
both segments are charged the $10 price, the elasticity of demand for segment A is 1.25, and
for segment B it is 5.0. Recall from Chapter 11 that the formula for marginal revenue is:
MR P= −[ ( / )].1 1 3η (1)
A difference in elasticities means that the marginal revenue from selling in the two market
segments differs. For segment A:
MRA = − =$ [ ( / . )] $ .10 1 1 1 25 2 (2)
For segment B:
MRB = − =$ [ ( / )] $ .10 1 1 5 8 (3)
3See footnote 3 in Chapter 11. The formula implies that the more elastic the demand, the closer marginal revenue is to the price of the product. At the extreme, when the elasticity of demand is infinity (a horizontal demand curve), marginal revenue equals price: MR = P[1 − (1/∞)] = P(1 − 0) = P. When demand is elastic (η > 1), marginal revenue is less than price but greater than zero. For example, when η = 3, marginal revenue is two-thirds the price: MR = P[1 − (1/3)] = P(2/3) = (2/3)P. When demand is unit-elastic, marginal revenue equals zero: MR = P[1 − (1/1)] = P(0) = 0. And when demand is inelastic (η < 1), marginal revenue is nega- tive. For example, when η = 1/2, marginal revenue is equal to the negative of the price: MR = P[1 − (1/0.5)] = P(1 − 2) = −P.
These “gray market” purchases, through which consum- ers arbitrage price differences across countries, reduce the profitability of price discrimination and have been exten- sively fought by pharmaceutical companies. In recent years, for instance, Pfizer instituted a new sales policy through which distributors in Canada have been authorized to deal with only approved buyers of Pfizer products. Exports of Pfizer products are forbidden, and distributors are required to get approval to sell to new pharmacies or to any phar- macy whose purchases exceed a certain limit. The let- ter outlining the new policy goes on to prohibit “selling, transferring or distributing products to any person that you know, or have reasonable grounds for believing, will or may export Pfizer products out of Canada....Any breach of the
terms of this letter will result in Pfizer refusing all further sales to you.”
Glaxo, another pharmaceutical company, has adopted the strategy of only selling directly to retail pharmacies. Through such a strategy, Glaxo hopes to curtail supplies to online and mail-order drugstores and any wholesaler that might sell to gray-market vendors.
Notwithstanding the efforts of drug companies to cur- tail gray-market purchases, arbitrage across different coun- tries by consumers is likely to continue, given the significant price differences that persist for various drugs between the United States and other countries. The Congressional Bud- get Office has estimated that Americans could save $40 billion annually by purchasing their drugs from Canada.
314 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 314Trim Size: 203.2 mm X 254 mm
Thus, if one unit less is sold to segment A, the monopolist loses $2 in total revenue, but if
that unit is sold to segment B, revenue from that segment will rise by $8. Consequently,
transferring a unit of output from segment A to segment B increases total revenue by $6.
Reducing sales to segment A raises the price to segment A, while segment B’s price falls
as sales increase there. This policy means that the segment with the more inelastic demand,
segment A, pays a higher price.
Figure 12.3 illustrates the way to divide 1,500 units of output between the two segments
to maximize total revenue. Segment A’s demand curve is to the left of the origin, and seg-
ment B’s is to the right. Initially, the monopolist charges a flat $10 price. At that price seg-
ment A purchases 500 units and segment B, 1,000 units. Total revenue is $15,000. Because
segment A’s demand curve is less elastic than segment B’s, however, marginal revenue is
lower for segment A ($2) than for segment B ($8). Shifting sales from the market segment
where the marginal revenue is low to where it is high increases total revenue. As long as
marginal revenue is higher for segment B, such reallocation will increase total revenue, so
it should continue until the marginal revenues in the two market segments are equal.
When the monopolist transfers 200 units from segment A to segment B, marginal reve-
nue in both market segments is equal at $7.50. The restriction of sales in segment A, where
demand is less elastic, raises price sharply for this segment, to $12.75. But the increase in
sales in market segment B, where demand is highly elastic, reduces price only slightly, to
$9.75. The relative differences in the price changes explain why total revenue increases.
Price rises sharply for the less elastic demand segment but falls only slightly for the highly
elastic demand segment. With sales allocated so that the marginal revenues are equal, total
revenue is now $15,525 [(300 × $12.75) + (1,200 × $9.75)], higher than the $15,000 in total revenue the monopolist earns when both market segments are charged the same price.
Output Determination When the monopoly can charge different prices to the two segments, total revenue from
the sale of any given output is highest when the marginal revenues are equal. This result
always means a higher price for the segment with the less elastic demand. Note, however,
that the rule of equating marginal revenues holds for any output, but it does not tell us what
Gains from Price Discrimination If demand elasticities differ and if the seller can segment a market, then it pays the seller to charge a higher price in the market segment with the less elastic demand. To maximize total revenue from the sale of 1,500 units, the seller divides output between the market segments so that the marginal revenues are equal: 1,200 units in segment B and 300 units in segment A. The seller charges a higher price in segment A ($12.75) than in segment B ($9.75).
Dollars per unit
Output in segment B
Output in segment A
0500 300
DA
PA = $12.75
PB = $9.75 P = $10.00
P = $10.00
DB
MRA
MRB
(1,500)
$8.00
$7.50
$2.00
(1,500)
1,000 1,200
Figure 12.3
Price and Output Determinat ion with Pr ice Discr iminat ion 315
C12.INDD 12:44:56:PM 08/06/2014 PAGE 315Trim Size: 203.2 mm X 254 mm
level of output is most profitable. Should the monopolist produce more than 1,500 units?
The marginal revenue from an additional unit of output is now $7.50 in whichever mar-
ket segment it is sold, so if marginal cost is less than $7.50, profit will be higher if output
increases. When sales are divided between market segments in this way, the decision of
how much output to produce is made by comparing the common value of marginal revenue
(because it is equal in both market segments) with marginal cost.
Figure 12.4 shows how a monopolist determines the most profitable level of total out-
put. As we just explained, the monopolist should compare the marginal cost with the com-
mon value of marginal revenue for the two separate market segments. The common value
of marginal revenue is derived by horizontally summing the separate marginal revenue
curves, and the result is the darker curve ΣMR. This curve shows one value of marginal revenue (since it is the same in both separate segments) for each level of total output. The
output level where MC equals ΣMR is consequently the most profitable output. In Figure 12.4, the most profitable output is 1,500 units where marginal cost is $7.50 and
equals the marginal revenue in both market segments. To determine how this total output
is divided between segments A and B, we identify the output at which marginal revenue is
equal to $7.50 in each segment. To do so, we move horizontally to the left from the inter-
section of MC and ΣMR until we reach each segment’s separate marginal revenue curve. This occurs at points F and G, so sales to segment A are 300 units and sales to segment B are 1,200 units. Price is higher for A than for B.
Whether a monopolist who price discriminates is in any sense worse than one who
charges a uniform price is not clear. Compared with a single-price monopoly, price dis-
crimination benefits one group of consumers, those with the more elastic demand who are
charged a lower price, and harms the other group of consumers. Frequently, those who ben-
efit have lower incomes than those harmed because in some markets low-income persons
are more sensitive to price (i.e., have higher demand elasticities), so perhaps this outcome
is favorable. The monopoly also benefits from price discrimination by obtaining a higher
profit (or else it wouldn’t price discriminate), and that outcome is often regarded as unde-
sirable in the public policy arena. Matters become more complicated when we recognize
that total output may be greater (i.e., more efficient) under price discrimination than under
Price and Output Determination under Price Discrimination The most profitable output occurs where the sum of the separate MR curves, ΣMR, intersects MC at QT. Thus, the monopoly sells QB in market segment B at a price of PB and QA in market segment A at a price of PA.
Dollars per unit
Output
$12.75 = PA
$9.75 = PB
$7.50
0 QA QTQB
(300) (1,200) (1,500)
DA
F
G
DB
MRA
MRB
�MR
MC
Figure 12.4
316 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 316Trim Size: 203.2 mm X 254 mm
single-price monopoly. For these reasons no blanket condemnation of price discrimination
seems appropriate, and each case should be judged separately.
The identities of the monopoly and its customers may also play a role in evaluating price
discrimination, as a further example will suggest. Price discrimination is sometimes found
in international markets when a firm charges a higher price in its domestic market than it
charges abroad. This procedure is sometimes called dumping, and it occurs when the inter- national demand for a product is more elastic than the demand in the domestic market. The
difference in elasticities occurs because there is more competition in world markets. For
instance, Japanese firms have been alleged to dump products in the United States by selling
them at lower prices here than in Japan. In this case of price discrimination, U.S. consum-
ers might applaud the practice because they are the ones who benefit. If we can get TVs,
stereos, radios, steel, and cars from Japan more cheaply than we can produce them here, the
average real income of U.S. consumers rises.
12.4 Intertemporal Price Discrimination and Peak-Load Pricing
Intertemporal price discrimination is a form of third-degree price discrimination. When different market segments are willing to pay different prices depending on the time at which
they purchase the good, a firm can increase its profit by tailoring its prices to the demands
of the various market segments.
Take the case of video programming. Distributors of television programs and motion
pictures discriminate among audiences by releasing their products at different times (known
as windows) and through different channels. Historically, movies were released through a series of “runs,” beginning with first-run theaters in big cities and working down to small
community theaters. Over the past three decades, the typical domestic release sequence for
a successful U.S. feature film has changed to cinema, home video/DVD, first cable run,
intertemporal price discrimination a form of third-degree price discrimination in which different market segments are willing to pay different prices, depending on the time at which they purchase the good
APPLICATION 12.4
A large number of the 1,500 private four-year colleges in the United States use statistical analysis to determine how much financial aid to offer prospective students and thereby increase the schools’ tuition revenue.4 By offering less financial aid, a college in effect charges a higher tuition price to a prospective student. If the prospective student opts to attend, the college earns more revenue than it would have made by offering the student a more generous financial aid package.
Statistical models attempt to take into account the price sensitivity of various applicant groups. For example, eager applicants who apply for early admission tend to be less
The Cost of Being Earnest When It Comes to Applying to Colleges
price sensitive and thus constitute a market segment that can be offered less financial aid (that is, charged a higher tuition price). Students who come for on-campus inter- views are statistically more likely to enroll and so need less aid to entice them. Expressed pre-med majors also tend to be less price sensitive and thus can be offered less financial aid as well.
In contrast to basing financial aid offers overwhelmingly on a student’s demonstrated financial need (as was histori- cally the norm in college admissions), taking into account a student’s price sensitivity to college costs pays off for the colleges that take this route. The National Center for Enroll- ment Management, one of the consulting groups offer- ing “financial-aid-leveraging” statistical models, states that its average client college increases its annual tuition rev- enue by nearly $500,000 by factoring in applicants’ price sensitivity.
4This application is based on “Colleges Manipulate Financial-Aid Offers, Shortchanging Many,” Wall Street Journal, April 1, 1996, pp. A1 and A4.
Intertemporal Pr ice Discr iminat ion and Peak-Load Pr ic ing 317
C12.INDD 12:44:56:PM 08/06/2014 PAGE 317Trim Size: 203.2 mm X 254 mm
broadcast network, second cable run, and syndication to local television stations. Even
more recently, the ability to download films through the Internet has created an additional
window for video producers. Through this variety of windows and type of release sequence,
distributors allow buyers to sort themselves according to how much they are willing to pay
to view the product at different points in time after the initial release of a program.
Figure 12.5 illustrates how a motion picture distributor can increase profit by engag-
ing in intertemporal price discrimination versus charging the same markup of price over
marginal cost irrespective of the viewing window. Suppose that consumers eager to view
a new film as soon as it is released to theaters are represented by the demand and marginal
revenue curves DE and MRE. In contrast, consumers willing to wait to view the film, on broadcast television perhaps, are represented by the demand and marginal revenue curves
DW and MRW. Assume, for simplicity’s sake, that the cost of serving all consumers is con- stant and equal to MC.
If resale can be prevented, the film distributor can increase profit by charging different
prices to the two market segments: PE to customers eager to see the film at a cinema as soon as it is released and PW to customers opting to wait to view the film at home. Such
PE
PW
MRE
MRW
0 QE
MC = ATC
DW
QW Output
Dollars per unit of output
DE
Intertemporal Price Discrimination When two different market segments are willing to pay different prices based on the time that a good is purchased, a supplier can increase profit by employing a pricing strategy that takes this into account. Compared to charging a common price to the two segments, profit is increased by charging PE to the market segment more eager to purchase the good and PW to the segment willing to wait to make the purchase.
Figure 12.5
318 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 318Trim Size: 203.2 mm X 254 mm
5Increased competition in the personal computer market as well as falling production cost, however, might also explain the historical decline in prices.
third-degree price discrimination results in profit being greater than if a common price is
charged to both market segments. As can be easily verified, charging a common price such
as PE or PW, or a price in between PE and PW, yields lower total profit than if the two seg- ments are charged different prices based on the outputs where their respective marginal
revenue curves intersect the marginal cost curve.
Another example of intertemporal price discrimination involves different fares for seats
on the same flight depending on how far in advance an airline ticket is purchased. Some
economists believe that computer hardware manufacturers engage in intertemporal price
discrimination when they introduce a new product. The price charged by IBM for a per-
sonal computer fell steadily over time after the product was first introduced.5 Book publish-
ing provides another example. Relative to the hardcover edition, the paperback version of a
book typically comes out six months later and is priced lower, not solely because of lower
production cost but also in recognition of the fact that consumers willing to wait that long
to read the book are more price sensitive than hardcover buyers.
APPLICATION 12.5
To increase their profits, major airlines rely on “yield man- agement:” sophisticated computer programs to determine how many seats on a given flight they should make available at a particular fare. Through a fare structure based on past demand for a flight (e.g., 30 seats might be earmarked for sale at the lowest discount fare) and restrictions to prevent resale (advanced purchase requirements, Saturday-night stay conditions, cancellation penalties, and so on), airlines strive to take advantage of the fact that air travelers differ in their sensitivity to prices and restrictions. For example, the lowest-price fare typically requires an advance purchase of 21 days, necessitates a Saturday-night stay, and is non- refundable. Such deep-discount fares generally are unat- tractive to business travelers, who are willing to pay more for a ticket (because the tickets are charged to an expense account) and cannot finalize their travel plans that far in advance or stay over a Saturday night. Prior to spinning it off as a separate unit in 1996, American Airlines estimated that SABRE, its computer-reservation system, accounted for 75 percent of the company’s net worth, thanks to its effectiveness at promoting price discrimination.
To exploit the fact that airlines release more lower- price fares when demand for seats on a flight is lower than expected, travel agents rely on software programs that con- tinuously scan a reservation system, snagging low fares as they become available. Airlines have retaliated by impos- ing hefty fees, based on the number of computer key- strokes made by the agencies, to discourage extensive fare searches. Airlines are also increasingly seeking to bypass travel agents by directly selling to customers through their own Web sites.
Yield Management by Airlines
Airlines’ efforts to limit resale through their computer- reservation systems and travel restrictions have not proved entirely successful. For example, suppose that a business customer needs to travel twice from Atlanta to Chicago for meetings on two successive Wednesdays, on April 7 and 14. Rather than buying a ticket for each round-trip that originates in Atlanta on Tuesday afternoon and returns from Chicago on Wednesday night (fares that retail for $500 to $800), the traveler can buy two back-to-back Super-Saver tickets that require a Saturday night stay (each costing only $200 to $400): The first originates out of Atlanta on Tues- day, April 6, and returns from Chicago on Wednesday, April 14; and the second originates out of Chicago on Wednes- day, April 7, and returns from Atlanta on Tuesday, April 13. In this manner, business travelers can circumvent the restric- tions imposed by airlines while accessing the deepest-dis- count fares.
Individual airlines occasionally crack down on back- to-back ticketing ploys by refusing to let travelers board a flight if they do not plan to stay over on Saturday night at the destination city on a Super-Saver fare6—that is, when the carrier knows that the traveler has an overlapping ticket issued on the same airline. The ability of airlines to prevent such arbitrage by business travelers in any sustained fashion is questionable, however, given that a business traveler can buy one back-to-back ticket from one airline and the other ticket from a competing airline.
6“Airlines Crack Down on Agents over Fare Plays,” Wall Street Journal, September 12, 1997, pp. B1 and B2.
Intertemporal Pr ice Discr iminat ion and Peak-Load Pr ic ing 319
C12.INDD 12:44:56:PM 08/06/2014 PAGE 319Trim Size: 203.2 mm X 254 mm
Peak-Load Pricing In Figure 12.5 we assumed that the marginal cost associated with selling output at various
points in time is constant. Such an assumption is not always valid. Sometimes producers
charge different prices at different points in time because, in addition to demand, the cost
of producing the “same” product varies with the time it is produced. In such instances dif-
ferent prices reflect not only different demands (i.e., price discrimination), but also different
costs. An important case of this type involves the provision of telephone service.
Telephone usage varies greatly over a 24-hour period. Typically, total use is greatest
in the daytime, during normal business hours. Residential use tends to be greater in the
evening than during the day, but total use is lower in the evening than in the daytime. Late-
evening use is lowest of all. (There are also often systematic variations in use over the
year; telephone usage is lowest during vacation months such as August and skyrockets on
specific days such as Mother’s Day.) Thus, a different demand curve exists for telephone
service at different times of the day. Economists refer to the period when demand is highest
as the “peak” period and when it is the lowest as the “off-peak” period.
Just as the demand for telephone service differs, so does the cost of producing it. If tele-
phone switching capacity (the ability to connect one caller to another) could be stored at
negligible cost, the marginal cost of providing telephone calls during peak and off-peak
periods would not differ; that is, the telephone company could operate at a constant rate of
production over the day, store the surplus switching capacity during the off-peak period,
and sell it during the peak period. Unfortunately, switching capacity can’t be stored; it has
to be used when it is produced. Because production must be greater during the peak period
than during the off-peak period, telephone companies must have the switching capacity to
meet the peak demand. As a consequence, much of the switching capacity needed during
periods of peak demand sits idle in off-peak periods. Moreover, the marginal cost of pro-
viding telephone service is higher during peak periods when capacity is strained and lower
during off-peak periods when only the most efficient switching capacity is employed.
As we will see, charging a higher price for telephone service during the peak period than
during the off-peak period serves to promote efficiency. We can compare the consequences
of a uniform price for telephone service with prices set to reflect different demands and
marginal costs of service over the day, with the aid of Figure 12.6. We assume a short-run
Peak-Load Pricing With demand of D1 in the peak period and D2 in the off-peak period, peak-load pricing involves charging a price of P1 in the peak period and P2 in the off-peak period.
Dollars per unit
0 Output
P1
SMC
A
D1
D2
Q2 Q1Q′1Q2′
P2
P
Figure 12.6
320 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 320Trim Size: 203.2 mm X 254 mm
setting in which the scale of operation has been selected; the switching capacity, buildings,
and telephone lines are already built. The short-run marginal cost of telephone service is
shown as SMC, and it slopes upward for reasons already explained in Chapter 8. We further assume that demand varies between two periods, with the demand curve for the peak period
shown as D1 and the demand curve for the off-peak period shown as D2. Now suppose that a public utility commission requires the provider of telephone service,
a regulated monopoly, to sell output at the price P, which just covers the average cost of producing telephone service in peak and off-peak periods. At P, the monopoly will provide an output of Q2 during the off-peak period, but what will be the most profitable output for the peak period? During the peak period the telephone company would like to produce an
amount equal to PA, where MC = MR (= P). Note, however, that if the telephone company produced PA, a shortage would result during the peak period. To avoid such a shortage, the public utility commission may require that the telephone company be able to meet the
demand at the regulated rate. Thus, we assume that the monopoly will produce Q1 during the peak period. However, because marginal revenue (= P) is less than marginal cost at Q1, the interests of the telephone company and the public utility commission, as well as the
public, would be at odds. (In the case of electric utilities, where uniform pricing is much
more common than for telephone service, some have argued that the occasional blackouts
and brownouts during periods of heavy use are manifestations of suppliers’ reluctance to
provide adequate capacity to meet peak demand at regulated rates.)
In contrast to the uniform-price policy, peak-load pricing calls for a different price in peak and off-peak periods. If regulators wish to promote efficiency, the price in each period
would be set where SMC intersects the relevant demand curve. Thus, price would be P2 in the off-peak period and P1 in the peak period. Faced with these prices, consumers would purchase Q2′ in the off-peak period and Q1′ in the peak period. With this price structure, con- sumers have an incentive to be more economical in their use of telephone service at the
time when the cost of providing it is highest. For example, they may shift their telephone
usage from the peak period to the off-peak period.7
The Advantages of Peak-Load Pricing Peak-load pricing has two advantages relative to uniform pricing. First, a more efficient
distribution of telephone usage between the peak and off-peak periods results. Note that
people curtail their telephone usage when it is more costly and increase their usage when
it is less expensive, so the total cost of producing a given amount of telephone service is
reduced. More formally, when use in the peak period falls from Q1 to Q1′ total cost falls by the area under the SMC curve over this range, and total benefit falls by the area under the demand curve. Total cost falls by more than total benefit, however, so there is a net gain,
as shown by the shaded area. Similarly, total benefit rises by more than total cost when
consumption is increased in the off-peak period; the net gain is shown by the shaded area
between D2 and SMC. If the monopoly is regulated so that it makes zero economic profit, this efficiency gain will be realized by the consumers of telephone service. The gain to
consumers is easiest to see if we assume that the total output of telephone service remains
unchanged (Q2′ − Q2 = Q1 − Q1′ ). Then, the total cost of producing this telephone service is reduced. If total revenue just covers total cost, total revenue from consumers also falls: the
average price of telephone service is reduced by using peak-load pricing.
peak-load pricing a pricing policy in which different prices are charged for peak and off- peak periods
7The extent to which people shift telephone usage from the peak to the off-peak period will be greater than shown in the diagram because the demand curves themselves will shift. Demand curve D1 is drawn for a given price of telephone service in the off-peak period—in this case, a price of P. When the price is P2 in the off-peak period, the demand curve in the peak period will shift to the left since consumption in the two periods are substitutes. Similarly, demand in the off-peak period will rise. The interdependence between the demand curves is ignored in the text; taking it into account strengthens the case for peak-load pricing.
Intertemporal Pr ice Discr iminat ion and Peak-Load Pr ic ing 321
C12.INDD 12:44:56:PM 08/06/2014 PAGE 321Trim Size: 203.2 mm X 254 mm
A second advantage of peak-load pricing becomes apparent when we turn from a
short-run to a long-run setting. In choosing a scale of operation, the telephone company
must have the capacity to meet the peak-period demand. Under uniform pricing, handling
the peak demand means being able to produce Q1. With peak-load pricing the quantity demanded in the peak period is less, so a smaller scale of operation is feasible. In terms
of building adequate capacity, peak-load pricing means that the telephone company has to
build and maintain less switching capacity. This cost saving also represents an efficiency
gain from peak-load pricing.
To a significant extent, the efficiency gains from peak-load pricing depend on the ability
of users to curtail their consumption when confronted with a higher price during the peak
period. The options here are greater than might be imagined. Some adjustments are quite
simple. In the case of electricity production in Vermont, for example, a system of peak-
load pricing has been used since 1974. Vermont families commonly fill dishwashers after
dinner but do not turn them on until late at night, when rates fall.
Businesses are also capable of adjusting their demand in response to a system of
peak-load pricing. A case in point is provided by the Kohler Corporation, in Kohler,
Wisconsin. When the day-time electricity price was raised to 2.03 cents per kilowatt
hour and the night-time price was lowered to 1.01 cents, Kohler responded by shifting
250 of its workers to the night-time shift. To compensate its workers for a less desirable
work schedule, Kohler paid an extra $50,000 in wages, but it cut its annual electric bill
by about $464,000.
APPLICATION 12.6
In recent decades, the price of highway construction in the United States has skyrocketed, while the state and federal gasoline tax revenues used to finance such construction have diminished (at least partly due to the improvements in the fuel efficiency of automobiles). In light of these trends, questions have begun to emerge about whether U.S. driv- ers will be able to continue relying on the government to provide free use of a sufficiently uncongested highway sys- tem that keeps pace with a growing population.8
To address the dilemma, Orange County in Southern California began building a series of toll roads in the mid- 1990s that could be a model for future highways. The plan is for the toll roads to be paid for by the drivers who use them. That is, individual drivers who opt to travel on the toll roads confront a system of “congestion prices” that vary from 25 cents to $5 per trip, depending on the length
Using Peak-Load Pricing to Combat Traffic Congestion
of the particular toll road and the amount of other traffic on the road. The peak-load prices are adjusted continu- ously, and ease of use is promoted through automation; cars can be equipped with transponders that automatically bill drivers’ accounts for usage and thus do not require stops at toll plazas. In the case of Route 91, for example, a highly congested freeway artery, 10 miles of the median was licensed to the California Private Transportation Company so that it could develop two toll lanes (out of a total of six) to implement a peak-load pricing scheme. The prices for this 10-mile stretch range from $1.40 to $9.55, depending on the time of day and day of the week.
Critics liken the experimental toll roads to a polite form of “highway robbery.” One policymaker has bemoaned the attempt to chip away at the concept of free roads, one of the last things shared equally by the rich and poor. He goes on to argue, “I don’t like special roads being developed for richer people, while ordinary people end up with potholes and congestion.”
Most drivers, however, seem to accept the concept as a means to ensure continuation of the California lifestyle they have come to accept. As one legal secretary who uses a toll road almost daily puts it, “I love it. . . . it’s expensive, but when you measure that against the frustration of being stuck on the freeway, it’s worth it. It’s cheaper than therapy!”
8This application is based on “Tolls Seen as Road to Expansion,” Los Angeles Times, March 23, 1997, pp. A3 and A29; and Jeff McCles- key, “Peak-Load Pricing and Traffic Congestion,” February 23, 2010, www.sustainbusper.com/peak-load-pricing-and-traffic-congestion- california-usa. Chapter 20 more fully explores another fundamen- tal reason for traffic congestion: drivers in free freeways not having to pay for the full costs, including increased congestion, that they impose on other drivers.
322 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 322Trim Size: 203.2 mm X 254 mm
Although we have examined peak-load pricing in the context of regulated monopolies,
it is also relevant for other forms of market organization. When the conditions are appro-
priate, it tends to arise naturally in unregulated markets. For example, hotels and motels in
resort areas charge more during vacation periods when demand is high, restaurants charge
more at dinner than at lunch, and movie theaters charge more in the evening than in the
afternoon. In these and similar cases, the different prices charged result from the fact that
demand and cost vary systematically over time and that cost varies over time because the
product cannot be stored.
12.5 Two-Part Tariffs A two-part tariff is a form of second-degree price discrimination. Under a two-part tariff a firm charges consumers a fixed fee (per time period) for the right to purchase the product at
a uniform per-unit price. For example, consumers might have to pay $50 per month (regard-
less of how much of the product they purchase); having paid this entry fee, they can then purchase the product at $10 per unit. In this manner, consumers pay a lower average price
per unit with the more units they purchase.
An example of a two-part tariff is a tennis club for which you must pay an annual mem-
bership fee plus a charge each time you use the tennis courts. Another example is telephone
service for which you pay a monthly fee plus a charge for calls placed. Mail-order book
retailers and member-based discount warehouses employ two-part tariffs when they charge
a customer a fee to join their shoppers’ clubs and then offer discounts (20 to 50 percent off
the list price) on any purchase made.
To employ two-part tariffs, a firm must have a degree of monopoly power and must be
able to prevent resale of the product; in these respects the situation is analogous to price
discrimination. Resale must be prevented because consumers have incentives to avoid the
entry fee by having one consumer pay the entry fee and then resell the product to other con-
sumers who have not.
How does a firm that uses a two-part tariff decide how to set the entry fee and the per-
unit price? Of course, the firm is guided by a desire to maximize its profit, but determining
what combination of price and entry fee will maximize profit is often not an easy matter.
In one case, however, it is simple: when all consumers have the same demand curve for
the product, and the firm knows this demand curve. Figure 12.7 illustrates this case. In
Figure 12.7a, a single consumer’s demand curve for minutes of local telephone service (per
month) is shown as D, and the marginal and average total costs to the firm providing local telephone service are assumed to be constant at MC = ATC. For the purpose of illustration, we assume that the provider of local telephone service is not subject to any price regulation.
(As Chapter 15 will explain, this assumption is not valid in reality; public utility commis-
sions limit the rates that local telephone suppliers can charge.) To maximize profit, the firm
charges an entry fee shown by the triangular shaded area T and a per-unit price of P. The consumer pays the entry fee and consumes Q minutes of local telephone service. The firm makes a profit (from this consumer) equal to the shaded area (the entry fee) because the
revenue from selling at price P just covers the production cost. How do we know that this combination of entry fee and price will maximize profit? In
general, a monopolist cannot make a profit greater than the maximum consumer surplus
that a consumer would attain if the product is priced at marginal cost. The maximum con-
sumer surplus is the shaded area, and in our example the firm realizes this amount as profit.
In fact, the consumer receives no net gain at all from purchasing the product; all of the
potential gain goes to the firm as profit. (Practically speaking, the firm might have to use a
slightly lower entry fee to ensure that the consumer participates.) If the firm tried to raise
the entry fee (with price fixed), the consumer would be better off not participating in this
two-part tariff a form of second-degree price discrimination in which a firm charges consumers a fixed fee per time period for the right to purchase a product at a uniform per-unit price
entry fee the fixed fee charged per time period in the case of a two-part tariff
Two-Part Tar i f fs 323
C12.INDD 12:44:56:PM 08/06/2014 PAGE 323Trim Size: 203.2 mm X 254 mm
market at all. Similarly, if the firm tried to raise the price (with the entry fee fixed), the con-
sumer would be better off exiting the market altogether.
The situation from the consumer’s point of view can be clarified with the aid of Figure
12.7b. The consumer’s income is given by A; line AZ has a slope equal to the price of the local telephone monopolist’s product. AZ would be the consumer’s budget line if the firm charged price P and no entry fee. Obviously, if the consumer could choose a point on AZ, he or she would be better off than if consuming none of the good at point A; there would be a net gain, or consumer surplus, in this case. The local telephone supplier, however, sets
the entry fee to extract all this potential gain. In this case, the entry fee is given by AA′. After paying that entry fee, the consumer can purchase the product along the A′Z′ budget line (which has a slope of P). Note that the entry fee is set so that the consumer’s preferred point on A′Z′, point E, is on the same indifference curve the consumer realizes at point A: the consumer is indifferent between purchasing the product (and paying the entry fee) and
not participating in this market at all (i.e., staying at point A). With many consumers who have identical demands, all would choose to participate and
the local telephone provider would realize all the potential consumer surplus as its profit.
In terms of the outcome, note that it is the same as when the firm can practice first-degree,
or perfect, price discrimination. In both cases, the firm is able to capture all the consumer
surplus as profit. In addition, the firm is producing the efficient rate of output, because it is
producing where marginal cost equals the price (marginal benefit). All the potential gain
from producing this product has been realized, but it has been realized by the firm as profit
rather than by the consumers as consumer surplus.
Many Consumers, Different Demands A firm would like to charge each consumer an entry fee that extracts the entire potential
consumer surplus. When consumers have different demand curves, however, a different
A Two-Part Tariff (a) Using a two-part tariff pricing strategy, the firm charges an entry fee shown by area T and a per-unit price equal to P, extracting all potential consumer surplus as profit. (b) The entry fee is shown as AA′. The consumer selects point E but is no better off than at point A.
Minutes of local telephone service
P
T
MC = ATC
D
Dollars per minute
0
(a)
Minutes of local telephone service
A
A ′
E
U1
Other goods
0 Q Z ′ ZQ
(b)
Figure 12.7
324 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 324Trim Size: 203.2 mm X 254 mm
entry fee must be charged to each consumer. If that can be done, the outcome is the same
as we have just explained. Typically, however, a firm may find that it must charge the same
entry fee to all consumers, perhaps because it does not have enough knowledge of each
consumer’s demand curve, and acquiring such knowledge would be prohibitively expen-
sive. In this case, the joint determination of the entry fee and a price that maximizes profit
is more difficult. In fact, there is no general rule that determines the most profitable policy.
Instead, firms have to proceed on the basis of trial and error, first setting an entry fee and
then varying price, and vice versa, until they find the combination that maximizes profit.
Let’s consider what this combination is likely to look like.
Assume there are two consumers of local telephone service, Martha and Donald, with
demand curves DD and DM, respectively, in Figure 12.8. (The analysis also applies, of course, if there are a large number of consumers, with equal numbers having each demand
curve.) To simplify the analysis, we have drawn Donald’s demand curve such that he would
consume exactly twice as much as Martha at each possible price. The total demand curve
facing the local telephone supplier is then DT, and the supplier’s marginal cost of produc- tion is assumed to be constant, as before. (The supplier is assumed to be free of any regu-
latory rate controls.) Now suppose that the local telephone supplier is initially charging a
price equal to marginal cost and sets the entry fee equal to the shaded area (thereby extract-
ing all of Martha’s consumer surplus). Profit is equal to twice the shaded area because the
entry fee is collected from both consumers. Note that if the entry fee is increased a small
amount, Martha would drop out of the market and so profit would fall; only Donald would
pay the entry fee.9
This combination of price and entry fee is not, however, the one that maximizes profit.
To see why, suppose that the firm raises price to P′ and simultaneously reduces the entry fee to ensure that both consumers remain in the market. (If the price is increased and the
entry fee remains unchanged, Martha would exit the market. In Figure 12.7b, this would
have the effect of making the budget line steeper at point A′, and Martha would be better off at point A than at any point on the new budget line.) The maximum entry fee that can be charged and still keep Martha in the market is now the area TSP′, so the entry fee has been reduced by area P′SRP. Our problem is to see whether this combination of a higher
9If Donald’s demand is more than twice as large as Martha’s, it would pay the firm to raise the entry fee to extract all of Donald’s consumer surplus and let Martha exit the market.
A Two-Part Tariff with Different Demands When consumers have different demand curves, the entry fee is set lower and the price of the product is set above marginal cost. Here, we see that the firm will make a larger profit by charging price P′ with entry fee TSP′ instead of price P with entry fee TRP.
Minutes of local telephone service
P
J
K
LS
R M P ′
T
MC = ATC
Dollars per minute
0
DT
DDDM
Q1 Q
Figure 12.8
Two-Part Tar i f fs 325
C12.INDD 12:44:56:PM 08/06/2014 PAGE 325Trim Size: 203.2 mm X 254 mm
price and lower entry fee produces a larger profit for the firm. To see that it does, note that
profit is now equal to area P′JKP (from sales at a price above cost) plus twice the area TSP′ (from the entry fee charged to each consumer). Compared with the initial situation, profit
has increased by area P′JKP minus twice the area P′SRP. Because area P′JKP is larger than twice area P′SRP, profit has increased. (Because Donald’s demand is exactly twice Martha’s, twice the area P′SRP exactly equals area P′LMP, so profit has increased by area P′JKP minus P′LMP, or by area LJKM.)
Thus, the telephone supplier can increase its profit by reducing the entry fee and rais-
ing price above marginal cost. Note that this contrasts with our earlier analysis of the case
where all consumers have the same demand curves. In that case, price was set equal to mar-
ginal cost. When demands differ, however, the firm has an incentive to alter both the entry
fee and price. In Figure 12.8, for example, the firm has an incentive to continue reducing
the entry fee and raising price as long as profit can be further increased. Where this process
ends depends on the specific pattern of consumer demand curves confronting the firm, but
we can show that the result will usually be a price lower than the simple monopoly price.
Why the Price Will Usually Be Lower Than the Monopoly Price Consider Figure 12.9, in which the consumer demands and marginal cost are the same as
in the previous graph. If the firm were a simple monopoly charging a uniform price and no
entry fee, the price would be P and output, Q, would be half the competitive output (for linear demand curves and constant marginal cost). Profit would be shown by the rectangle
PJHN. Now we can see that the firm can do better by using a two-part tariff, if that is fea- sible. If it charges an entry fee equal to the area TSP and continues to charge price P, profit will increase by twice the entry fee. Thus, a two-part tariff will result in larger profit than
would a uniform price. However, it does not mean that this specific entry fee and price will
maximize profit. In fact, we can show that an increase in the entry fee coupled with a price
lower than P will increase profit. Let us evaluate how an increase in the entry fee to TRP′, coupled with a reduction in price
to P′, will affect the firm’s profit. Ignoring the entry fee for the moment, we see that the price
Effect of Two-Part Tariff on Price Profit can be increased by charging a price lower than the simple monopoly price when a two-part tariff is used. Here, we see that the firm will make a greater profit by charging price P′ with entry fee TRP′ rather than price P with entry fee TSP.
MC = ATC
Dollars per minute
T
P
N
P ′
0 Q Q1 Minutes of local telephone service
J
F L
M K
R S
H G
DM
DD DT
Figure 12.9
326 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 326Trim Size: 203.2 mm X 254 mm
reduction will affect profit in two opposing ways. Profit will be increased by area KFGH (additional output at a price above cost) and reduced by area PJKP′ (reduced profit on initial output). If the initial price P is the simple monopoly profit-maximizing price, these two areas will be approximately equal. (In other words, a small change in price in the neighborhood
of the profit-maximizing price will have a negligible effect on total profit.) Thus, the profit
from sales at a price above marginal cost is approximately unchanged by the price reduction.
Recalling now that the firm also collects higher entry fees, equal to twice area PSRP′, we can see that the combination of a lower price and higher entry fee will increase total profit.
(Proceeding more slowly, profit increases by area KFGH, minus area PJKP′, and plus twice area PSRP′ = area PLMP′. Thus, profit rises by area KFGH minus area LJKM.)
This analysis is obviously somewhat complicated and would be even more so if we con-
sidered a case with more than two consumers. Nevertheless, we have been able to reach
some interesting conclusions about two-part tariffs. First, a firm can realize more profit by
using this pricing strategy than by simply using a uniform price. Second, the price charged
will be lower than the simple monopoly profit-maximizing price but higher than marginal
cost. Third, and an implication of the second point, output will be higher than under simple
monopoly and therefore the deadweight loss will be smaller.
APPLICATION 12.7
Employing a pricing strategy such as a two-part tariff is not without costs. The Disney theme parks provide a case in point. Prior to 1980, Disney required customers to purchase a “passport” that granted admission to its theme park and included a set number of tickets to each of the park’s various rides. Additional ride tickets could be purchased, with the price varying depending on the ride. Disney set the highest prices for “E” rides such as Space Mountain, since custom- ers favoring such rides tended to be more fanatical (that is, less price sensitive) in their preferences. While fanatical riders could not be identified at the admission gate, they could be sorted out through the two-part pricing scheme, and so Disney could extract more of their consumer surplus.
The Costs of Engaging in Price Discrimination
Despite the effectiveness of the two-part pricing scheme for extracting consumer surplus, Disney adopted a simpler, single-price admission policy in 1980; the com- pany began charging a higher entry fee but eliminated the additional per-ride charges. The reason was that Disney found the cost of administering the more complicated two-part pricing scheme in terms of labor and paperwork (additional staff were needed to sell and collect tickets for the various rides) outweighed the benefit. Although Disney’s theme parks still employ other forms of price discrimina- tion (multi-day and season passes, senior-citizen discounts, passes ensuring shorter wait times for certain popular rides, and so on), the two-part passport pricing strategy did not enhance company profit because of its administrative costs.
SUMMARY
Monopolies can engage in pricing tactics not avail-
able to competitive firms. One example is price discrimi-
nation, the charging of nonuniform prices.
A firm with monopoly power has an incentive to
engage in price discrimination because it can increase
profit, provided it is not too costly to identify what dif-
ferent potential customers are willing to pay and that
resale can be prevented.
Perfect, or first-degree, price discrimination means
selling each unit of output for the maximum price a con-
sumer will pay. It is perfect from the monopolist’s point
of view because it produces the maximum amount of
profit and reduces consumer surplus to zero.
Second-degree price discrimination, or block pricing,
occurs when the per-unit price declines as a function of the
quantity purchased. Provided that such a declining per-unit
Review Quest ions and Problems 327
C12.INDD 12:44:56:PM 08/06/2014 PAGE 327Trim Size: 203.2 mm X 254 mm
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
*12.1 Apply the theory of price discrimination to a monopoly that faces a downward-sloping demand curve for its domestic
sales but a horizontal demand curve for sales in international
markets. (Do you see how tariffs and trade restrictions could
produce this situation?)
12.2 Assume that all consumers have identical demand curves for local telephone service, and the producer of such service is
a monopoly. Compare price, output, profit, and consumer sur-
plus when (a) the monopoly sets a uniform price for the prod-
uct; and (b) the monopoly uses a two-part tariff.
12.3 How can the supplier of local telephone service determine the optimal two-part tariff if its customers have different (but
known to the supplier) demand curves?
12.4 In Figure 12.8, how will the profit realized by raising the price and reducing the entry fee be affected if Donald’s
demand curve is only slightly greater than Martha’s (instead
of twice as large, as shown in the graph)? In Figure 12.9, how
will the profit realized by reducing the price and increasing the
entry fee be affected if Donald’s demand curve is only slightly
greater than Martha’s? What do these results suggest about
how the profit-maximizing price and entry fee will vary in the
two cases?
*12.5 Car rental firms often charge a daily rental fee for cars plus an additional cost per mile driven. Is this an example of a
two-part tariff?
12.6 What is peak-load pricing? How is it similar to price dis- crimination? How is it distinguished from price discrimina-
tion?
*12.7 Food consumption peaks at dinnertime and is very small between midnight and 6:00 A.M. In view of this systematic
variation in consumption over the day, why is peak-load pric-
ing not used more extensively for food?
12.8 The text states that if conditions are appropriate, peak- load pricing arises naturally under competitive conditions.
Explain why peak-load pricing will emerge, starting from a
point where all firms are charging a uniform price.
12.9 “Suppose that Cornell University faces a downward- sloping linear demand curve for the undergraduate education
that it provides. If Cornell is able to engage in perfect, first-
degree price discrimination (through obtaining detailed finan-
cial information from each prospective student and offering
different levels of financial aid), then Cornell’s marginal and
average revenue curves will be identical.” Explain why this
statement is true, false or uncertain.
12.10 The year is 2020, and the U.S. airline industry has been radically transformed through a recent wave of mergers. Only
one company, MONO Airlines, has managed to survive the
succession of price wars, labor–management disputes, and
government policy reversals that plagued the industry in previ-
ous years. MONO now seeks to make the most of its exclusive
hold on the market. To that end, it adopts a new slogan, “Fly
MONO—or Walk,” and then hires you as a consultant to offer
advice on its pricing policy. Specifically, MONO asks you for
advice on how much to charge for its one-way flight from Bos-
ton to New York City. You are informed that the one-way mar-
ginal cost for each passenger is $40. You are also told that there
are two types of customers: well-paid business executives, and
less-advantaged students and tourists. The demand by each of
these types of customers is shown in the following table:
One-Way Trips Demanded per Year (in Thousands)
Price of One- Way Ticket
Executives Students/ Tourists
$140 0 0
$130 8 0
$120 9 1
$110 10 2
$100 11 3
$90 12 4
$80 13 5
$70 14 6
You recommend that MONO charge different prices to the
two different customer groups. If MONO charges different
fares, what fare would maximize the profit earned from each
customer group?
pricing schedule does not reflect only cost considerations
(e.g., economies of scale), it can increase a monopoly’s
profit by allowing the firm to take advantage of the fact
that it faces a downward-sloping demand curve.
Third-degree pricing, or market segmentation, occurs
when the price differs among categories of consumers.
The same item may be sold to different market segments
at different prices, depending on such factors as a seg-
ment’s demographic features and sensitivity to the time
of purchase, as well as the extent to which the market
segment is informed about the prices charged by com-
peting firms.
328 Product Pr ic ing with Monopoly Power
C12.INDD 12:44:56:PM 08/06/2014 PAGE 328Trim Size: 203.2 mm X 254 mm
12.11 Consider your answer to the preceding problem. Rela- tive to the case where MONO charges a single price to all its
passengers, would the price discrimination scheme you recom-
mended raise or lower MONO’s total profit? By how much?
12.12 Hard Bodies is a new entrant to the local health club scene. The owners of Hard Bodies realize that profit can be
increased through price discrimination. Accordingly, the firm
employs several different pricing schemes. For each of the fol-
lowing schemes explain whether it is price discrimination and,
if so, what degree of price discrimination it is.
a. An annual membership to the club sells at a 50 percent dis- count of the total rate charged customers who choose to pay
on a month-by-month basis (e.g., the annual fee is $300
while the regular monthly rate is $50).
b. Obese customers weighing at least 300 pounds get a 20 per- cent discount on all regular rates.
c. Spouses of members belonging to the club qualify for a 30 percent discount on all regular rates.
d. Hard Bodies offers to beat (through a 20 percent discount) any rate that a customer is offered by a rival health club.
12.13 A private golf club has two types of members. Serious golfers each have the demand curve Q = 350 − 10P, where Q represents the number of rounds played per year and P is the per-round price. Casual golfers have the demand curve Q = 100 − 10P. The club has 10 serious and 100 casual golfing members and faces a constant marginal cost of $5 per round
played by either type of member. If the club can engage in
third-degree price discrimination, what prices should it charge
to the two types of members?
12.14 In the preceding problem, suppose that the club can employ a two-part pricing scheme but must charge all mem-
bers the same annual membership (entry) fee. What entry fee
and per-round price should the club charge?
12.15 Suppose that the golf club described in Question 12.13 can employ a two-part pricing scheme and can charge different
entry fees to different members. What entry fee and per-round
price should the club charge to each member type?
12.16 Assume that the marginal cost to a grocery of selling a bottle of salad dressing to customers who use coupons versus
those who don’t is identical and equal to $1.50. If the elastic-
ity of demand of coupon users is 5 versus 1.25 for noncou-
pon users, how much of a per-unit discount should the store
make available through coupons? What if coupon users have a
demand elasticity equal to 2 versus 1.25 for noncoupon users?
12.17 A video game producer has costs of $25,000 per month that are fixed with regard to output. The firm’s marginal cost
is $5 per unit of output for output between 1 and 15,000 units.
Information available from the market research group indi-
cates that 15,000 units could be sold each month in the firm’s
primary market if the price was set at $6.80 per unit and that
14,000 units could be sold at $7 per unit. The market research
group also suggests that it is reasonable to assume that price
and quantity demanded have a linear relationship in this market
not only between those two points, but also well beyond them.
a. One officer of the firm feels that price should be set at the level that would maximize revenue. At what price would
this objective be accomplished? What would price elasticity
and marginal revenue be at this price? Is this the price the
firm should establish? Why or why not?
b. Other officers are concerned with profit. What price should be set to maximize profit? What output will prevail in the
market at this price? What would price elasticity and mar-
ginal revenue be at this price? What is the profit?
The firm has the opportunity to sell in a second market that is
separated from the first in such a way that buyers in one mar-
ket cannot resell to buyers in the other market. For the second
market, the market research group has estimated the demand
relationship to be:
P Q2 27 0 00001= − . ,
where P2 is the price in the second market and Q2 is the quan- tity of the firm’s product sold in that market each month.
c. Some officers of the firm believe this second market offers an opportunity for additional profit. They argue that if
production is constrained to 15,000 units, the limit within
which marginal cost is $5, it is worthwhile to sell some of
these units in the second market. Should the firm sell any
units in this market? Should it sell only units that would not
be absorbed in the primary market at the profit-maximizing
price? Should it divert some units from the primary to the
secondary market? What price would you set in each mar-
ket? What are the elasticity and marginal revenue in each
market? What is the profit if your policy suggestion is fol-
lowed? How much profit do you attribute to each market?
Explain why your suggestion is the best policy.
d. One of the firm’s production managers has pointed out that 15,000 units of output per month is not the absolute limit
on production. The physical limit, she points out, may be
closer to 30,000 units. The problem is that for each unit of
output above the 15,000-unit level, marginal cost will rise
by $0.001, so that unit 15,001 will increase total cost by
$5.001, unit 15,002 will increase it by $5.002, and so on.
She wonders if the two markets together could not advan-
tageously absorb more than 15,000 units considering this
production situation. What total output do you recommend?
How much should go into each market? Is it worthwhile
to push beyond 15,000 units of output per month? Why or
why not?
12.18 You run a rather plush ride concession at an amusement park. It costs you $500 per day to have the ride available to
patrons of the park. For each rider you have, the incremental
cost is $1.
The patrons of the park appear to fall into two groups.
Members of the first group are not concerned with taking a
variety of rides but are quite responsive to ride price and will
take the same ride many times. Members of the second group
Review Quest ions and Problems 329
C12.INDD 12:44:56:PM 08/06/2014 PAGE 329Trim Size: 203.2 mm X 254 mm
like variety in their rides and will pay a good deal to have at
least one turn on a particular ride.
The daily demand for rides on your concession by a patron
of the park in each of the two groups is shown in the table:
Price Patron in Group 1 Patron in Group 2 $5.00 0 1
4.00 0 1
3.00 0 1
2.75 1 1
2.50 2 1
2.25 3 1
2.00 4 2
1.75 5 2
1.50 6 2
1.25 7 2
1.00 8 2
0.75 9 2
0.50 10 2
Each day each group includes about 100 patrons.
a. If the amusement park limits you to a one-part pricing structure consisting of a price per ride that is the same for
every ride taken, what price will you charge?
b. If you could charge a two-part tariff consisting of a fee for access to your concession plus a charge for each ride taken,
then what access fee and ride charge would you set? How
much would you be willing to pay to the amusement park’s
owners to permit you to use this pricing structure?
c. If you could charge each patron a declining amount for each ride the patron took—that is, $3 for the first ride, $2.50 for
the second, and so on—could you do better for yourself and
for the amusement park than you could with either a single
price or a two-part tariff? Explain why or why not.
330
C13.INDD 10:49:55:AM 08/06/2014 PAGE 330Trim Size: 203.2 mm X 254 mm
13CHAPTER
Learning Objectives
Explain how price and output are determined under monopolistic competition. Describe the characteristics of Oligopoly and the Cournot Model. Compare several key noncooperative oligopoly models, including Stackelberg and the dominant firm. Show how price and output are determined under the cooperative oligopoly model of cartels.
Competition and pure monopoly lie at opposite ends of the market spectrum. Competi- tion is characterized by many firms, unrestricted entry, and a homogeneous product, while
a pure monopoly is the sole producer of a product. Yet many real-world market structures
seem to be incompatible with either the competitive or the pure monopoly model. How do
we analyze a market situation, then, where there are a dozen similar but slightly different
brands of aspirin or only three companies supplying breakfast cereals?
Falling between competition and pure monopoly are two other types of market structure:
monopolistic competition and oligopoly. Monopolistic competition is closer to competi-
tion; it has many firms and unrestricted entry, like the competitive model, but the firms’
products are differentiated. Fast-food chains, for example, may be viewed as monopolistic
competitors. They supply the same general product, fast food, but one chain’s specialty
burger, say the Big Mac, is “different” from another’s, such as the Whopper. Oligopoly is
more like pure monopoly; it is characterized by a small number of large firms producing
either a homogeneous product like steel or a differentiated product like cars.
This chapter examines monopolistic competition and oligopoly market structure models,
noting the similarities with, as well as the differences from, perfect competition and pure
monopoly. We also explore the case of cartels, whereby firms in an industry attempt to
coordinate price and output decisions so as to act, in concert, as a pure monopoly and maxi-
mize their joint profit.
Memorable Quote “People of the same trade seldom meet together, even for merriment or diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices. It is impossible indeed to prevent such meetings, by any law which either could be executed, or would be consistent with liberty and justice. But though the law cannot hinder people of the same trade from sometimes assembling together, it ought to do nothing to facilitate such assemblies; much less to render them necessary.”
—Adam Smith, Wealth of Nations, Book I, Chapter X
Monopolistic Competition and Oligopoly
Price and Output under Monopol ist ic Competit ion 331
C13.INDD 10:49:55:AM 08/06/2014 PAGE 331Trim Size: 203.2 mm X 254 mm
13.1 Price and Output under Monopolistic Competition Monopolistic competition resembles both competition and monopoly. As with competition, entry into and exit from the industry are unrestricted, often resulting in a large number of inde-
pendent sellers. However, in contrast to competition, the firms do not produce a homogeneous
product. Instead, their products are heterogeneous, or differentiated. A differentiated product is one that consumers view as different from other similar products. For example, Wheaties
and Cheerios are differentiated products in the general category of breakfast cereals. Consum-
ers are not indifferent among brands of cereals; they perceive differences in taste, crunchiness,
caloric content, and nutritional value. In a competitive market, by contrast, consumers view
the product of one firm as identical to (a perfect substitute for) any rival firm’s product.
Product differentiation may reflect real differences among products (e.g., Special K cereal is lower in fiber but higher in protein than Post Raisin Bran), or it may be based only
on the belief that there are differences (e.g., a three-year-old may perceive Fruit Loops to be sweeter than Frosted Flakes, but their sugar content is the same). The content of most
aspirins is virtually identical, but many consumers believe Bayer to be superior to other
brands. In blind taste tests, many consumers claiming to have strong preferences for Coca-
Cola over Pepsi cannot select their preferred brand. This outcome doesn’t affect the theory,
however; the important point is that consumers, or at least a substantial number of them,
believe the products to be different. There are many aspects to product differentiation. For example, products may be dif-
ferentiated by physical features such as function, design or quality, or by advertising, brand
names, logos (such as the apple on Apple products) or packaging (such as Oscar Mayer
Lunchables). They may also be differentiated by conditions related to the sale, such as
credit terms, availability or congeniality of sales help, location or service. (Have you ever
shopped at a nearby 7-Eleven because of its convenience? Do you get your hair cut at a
particular beauty salon because of the conversational rapport you have with a specific styl-
ist and/or a fondness for how the person cuts your hair?) As this list suggests, many of the
goods you purchase are differentiated products. Clothing, medication, cosmetics, restaurant
meals, and many types of food products are prominent examples.
Determination of Market Equilibrium The first step in analyzing monopolistic competition is understanding the demand curve that
confronts a single firm. When a firm sells a differentiated product with close substitutes, it has
some degree of monopoly power—hence, the “monopolistic” in monopolistic competition.
In other words, the demand curve confronting the firm is downward sloping. However, the
degree of monopoly power will typically be slight because of the availability of close sub-
stitutes. For instance, because McDonald’s is the only firm selling Big Macs, the quantity of
Big Macs sold is unlikely to fall to zero if McDonald’s charges a slightly higher price than its competitors. But at a higher price for Big Macs, many consumers might switch to a Burger
King Whopper or a Wendy’s Double Bacon Cheeseburger. Thus, the demand curve facing each firm in a monopolistically competitive market is downward sloping but fairly elastic.
Assume that the market for jeans is monopolistically competitive. In Figure 13.1a, we
show the demand curve, D, for one firm in this market, Tight Jeans. The demand curve’s position depends strongly on the prices of other jeans, as well as the variety available.
Thus, in drawing the demand curve for Tight Jeans, we assume that the number of other
firms in the industry is fixed. Furthermore, we assume that the prices charged by other
firms do not change when Tight Jeans varies its price. (Changes in the prices charged by
other firms would cause the demand curve for Tight Jeans to shift.) The basis for assuming
monopolistic competition a market characterized by unrestricted entry and exit and a large number of independent sellers producing differentiated products
differentiated product a product that consumers view as different from other similar products
332 Monopol ist ic Competit ion and Ol igopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 332Trim Size: 203.2 mm X 254 mm
other firms’ prices as given is that Tight Jeans represents only a small part of the total jeans
market. While a lower price for its jeans will cause some customers to shift from other
brands, the loss for each brand will be small enough to be unnoticeable, or at least not to
provoke a reaction.
Given the behavior of other firms in the market, let’s consider how Tight Jeans deter-
mines price and output. Because its demand curve is downward sloping, marginal revenue
is less than price, and profit maximization calls for operating where marginal revenue equals
marginal cost. If the firm has the cost curves shown in Figure 13.1a, it produces an output of
Q1 and charges a price of P1. The resulting economic profit equals the shaded area. In terms of the diagram, the position of the monopolistically competitive firm resembles
that of a monopoly. However, there are two important differences. First, Tight Jeans is only
one among a number of firms producing a similar product, and so the demand curve is not
the market demand curve for jeans; it is only the demand curve for jeans produced by one
firm. Second, under monopolistic competition, as distinct from pure monopoly, entry into
the market is unrestricted. When existing firms are making profits, other firms are attracted
to the market. Thus, the equilibrium in Figure 13.1a cannot be a long-run equilibrium
because profits are being realized. It could represent a short-run equilibrium, but with the
entry of other firms, the demand curves facing each existing firm will shift.
Under monopolistic competition, long-run equilibrium is attained as a result of firms entering (or leaving) the industry in response to profit incentives. In the present example, entry continues to occur until firms in the market are no longer making economic profits.
How will the entry of other firms affect existing firms like Tight Jeans in Figure 13.1a? New
firms will increase the industry’s total output, as well as provide for a wider variety of dif-
ferentiated products. Both of these effects shift existing firms’ demand curves downward,
leading to a general reduction in the industry’s level of prices and, from that, lower profits.
(It is also possible that entry will lead to higher prices for some inputs, causing cost curves
to shift upward as in an increasing-cost competitive industry, but we will ignore this pos-
sibility.) Entry and output adjustments by existing firms will continue until economic profits
are zero; only then will there be no further incentive for other firms to enter the market.
Dollars per unit
Dollars per unit
(a)
0 Q1
P1
MC
AC
D
MR
Output Output
(b)
0 Q2
P2
MC
T S
AC
D ′
MR ′
Monopolistic Competition (a) In the short run, a firm in a monopolistically competitive market may make a profit. (b) Attracted by the prospect of profits, new firms enter the market. As entry continues, the demand curve for existing firms shifts downward until a zero-profit, long-run equilibrium is attained.
Figure 13.1
Price and Output under Monopol ist ic Competit ion 333
C13.INDD 10:49:55:AM 08/06/2014 PAGE 333Trim Size: 203.2 mm X 254 mm
Figure 13.1b shows a position of long-run equilibrium for Tight Jeans. The firm’s
demand curve has shifted down to D′, a position where it is just tangent to the average cost curve at point T. (If the demand curve intersected the average cost curve, then there would be a range of output over which profit would be positive; the final equilibrium must be a
tangency.) The profit-maximizing output is now Q2 with a price of P2; at this price and output, Tight Jeans makes zero economic profit. All rival firms will be in a similar situa-
tion, making zero economic profit in long-run equilibrium. Their cost and demand curves,
however, need not be identical because they are not producing exactly the same products.
For this reason, there may be a range of prices prevailing in equilibrium. Given the similar-
ity among the differentiated products within a monopolistically competitive market, prices
are likely to vary over a small range. A Big Mac and a Whopper need not be the exact same
price, for example, but it would be surprising if the prices differed substantially.
Firms in a monopolistically competitive industry compete not only on price, but also
by variations in their products intended to attract customers. The range of differentiated
products in a market is not fixed, and firms often introduce new variations they believe
will be profitable. For instance, when Coca-Cola introduced its caffeine-free Coke, it was
betting that enough consumers wanted to limit caffeine intake for the line to be profitable.
The company was right, and for a time it found itself in the position shown in Figure 13.1a,
making a profit. But once it was recognized that this was a profitable way to differenti-
ate cola drinks, other firms followed suit. Coca-Cola’s profit eroded as the market moved
toward a long-run equilibrium.1
Note that the long-run equilibrium position is similar to both the competitive and the
monopolistic equilibria. As with perfect competition, each firm’s demand curve is tan-
gent to its long-run average cost curve, so economic profit is zero. As with monopoly, the
demand curve is downward sloping, so price exceeds marginal cost at the equilibrium.
However, because the firm’s demand curve is relatively elastic, price will normally not
exceed marginal cost by very much. For instance, demand elasticities for monopolistically
competitive firms can easily exceed 10. If the firm’s demand elasticity is 15, we can use
the markup formula explained in Chapter 11 [(P − MC)/P = 1/η] to see that when profit is being maximized, the price/marginal-cost markup would be only about 7 percent of price.
Monopolistic Competition and Efficiency In Chapter 10 we saw that a competitive industry tends to be efficient, while in Chapter 11
we saw that a monopoly is inefficient (produces a deadweight loss). Because monopolistic
competition combines elements of both monopoly and competition, it is natural to consider
whether it is an efficient market structure, like competition, or inefficient, like monopoly.
Monopolistic competition has been charged with inefficiency in two respects. We can
examine both with the aid of Figure 13.2, which shows a monopolistically competitive firm
in long-run equilibrium. (Ignore the D* demand curve for now.) The first aspect of the alleged inefficiency involves the fact that the firm does not operate at the minimum point on
its long-run average cost curve. In the diagram, the firm operates at point A, where average cost per unit is greater than at point S. Every firm in the monopolistically competitive industry is in a similar position. By contrast, firms in a competitive industry operate at the
minimum points on their long-run AC curves. When firms fail to produce at lowest possible average cost, they are sometimes said to have excess capacity.
A failure to operate at minimum average cost is potentially inefficient because it is pos-
sible to produce the same industry output at a lower cost. To verify this notion, suppose that there are currently 10 firms like the one in Figure 13.2, each producing 100 units of output
excess capacity the result of firms failing to produce at the lowest possible average cost
1Not all new product variations, of course, are successful. For example, McDonald’s introduced the McLean burger during the 1980s, hoping that it would satisfy the tastes of health-conscious fast-food consumers. The McLean burger never proved profitable and came to be known as the “McFlopper” by industry analysts.
334 Monopol ist ic Competit ion and Ol igopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 334Trim Size: 203.2 mm X 254 mm
at an average cost of $15. The total cost of producing the 1,000 units would therefore be
$15,000 (10 × 100 × $15). If the average total cost is at a minimum of, say, $11 per unit at an output of 125 units per firm, then eight firms could produce the same 1,000 total output for less total cost. The total cost would now be $11,000 (8 × 125 × $11).
A monopolistically competitive market has also been alleged to be inefficient because it
produces the wrong total output from a social perspective. (Note that in discussing excess
capacity, we were concerned with an unchanged total output.) Each firm is producing an
output where price is greater than marginal cost. This condition suggests, by analogy to the
case of pure monopoly, that additional output is worth more to consumers than the cost of
producing it. There is a deadweight loss from producing too little.
Figure 13.2 shows a monopolistically competitive firm producing an output of Q where price, AQ, is greater than marginal cost, BQ. It is tempting to apply the same reasoning we did in the case of pure monopoly and argue that the magnitude of the deadweight loss
equals triangular area ACB. By performing the same calculation for each firm in the indus- try and adding up the results, we could arrive at the deadweight loss for the entire monopo-
listically competitive industry. Tempting as it is, this procedure is incorrect and overstates
the industry’s total potential deadweight loss.
To understand this, consider the firm in Figure 13.2 expanding output to the apparently
efficient point C on its demand curve. Recall that the firm’s demand curve is drawn on the assumption that rival firms keep their prices unchanged. This is the appropriate assump-
tion if we are examining a price change by one firm alone. However, the prospective inef-
ficiency here is that all firms in this industry are producing too little. If all expand their output, the demand curve confronting each firm must shift downward. Consequently, it is
not desirable for every firm to expand output to the point where marginal cost intersects its
initial demand curve, since that curve shifts in reaction to output and price changes by the other firms.
There is a complicated interdependence between individual firms’ demand curves in
an industry composed of several firms, and that interdependence must be accounted for in
evaluating efficiency. (Note that this problem did not arise with pure monopoly because
there was only one firm in the industry, or with a competitive market where we worked
with industry, and not firm, demand curves.) One way to account for the interdependence
is to draw the demand curve confronting the firm when it and all other firms in the industry
simultaneously expand output. This demand curve, shown as D* in Figure 13.2, captures the interdependence among the firms and shows that the marginal value, or price, of the
firm’s product falls more rapidly when rival firms are also producing more units. Point R
Figure 13.2 Alleged Deadweight Loss of Monopolistic Competition The monopolistic competitor’s demand curve is D when it alone varies price; the demand curve D* is relevant when all firms simultaneously change output. The deadweight loss is shown by area ARB; similar areas for the other monopolistically competitive firms can be added to this area to obtain the total deadweight loss due to restricted output in this market.
D*
Dollars per unit
P
MC AC
D
MR
B
R
A S
C
Q Output0
Price and Output under Monopol ist ic Competit ion 335
C13.INDD 10:49:55:AM 08/06/2014 PAGE 335Trim Size: 203.2 mm X 254 mm
now represents the efficient output of this firm, and this is consistent with every other firm
also having expanded output to the point where their price is equal to marginal cost. We
can thus sum the areas like ARB to arrive at the total deadweight loss resulting from each firm producing at a point where price exceeds marginal cost. Of course, the important point
is that the industry’s total deadweight loss is smaller than if we erroneously sum up the
areas like ACB.
Is Government Intervention Warranted? While monopolistic competition has been charged with being inefficient, there are three rea-
sons government intervention probably is not warranted. First, any deadweight loss associ-
ated with monopolistic competition is likely to be small, due to the presence of competing
firms and free entry. Put differently, each firm’s demand curve is relatively elastic, and so
the excess of price over marginal cost is typically small. In the case of pure monopoly, this
is not necessarily true. (Note that this excess, P − MC, is the height of the deadweight loss triangle, AB, in Figure 13.2.) For the same reason, the cost associated with excess capacity will also tend to be small.
APPLICATION 13.1
Refractive eye surgery has become very popular in recent years. Roughly 1.3 million Americans undergo the procedure each year in order to correct their vision.2 As the refrac- tive eye surgery industry has grown, it has evidenced all the characteristics of monopolistic competition. Entry and exit into the industry are relatively unrestricted. There are a large number of independent sellers who do not produce a homogeneous product. For example, under the Lasik pro- cedure, a surgeon first creates a flap in the eye and then uses a laser on the area underneath to correct a patient’s vision. PRK, another form of laser eye surgery, consists of a surgeon using a laser on the eye’s surface to correct vision. Some patients opt for corneal rings, prescription inserts intended to correct mild nearsightedness.
Monopolistic Competition: The Eyes Have It (When It Comes to Refractive Surgery)
The various sellers of refractive eye surgery services tout the advantages of their differentiated products. It has been estimated that surgery centers spend as much as $200 per eye corrected on advertising. Sports personalities such as Tiger Woods (golf), LeBron James (basketball), and Brandon Knight (baseball) have appeared in testimonials as have actors/singers such as Brad Pitt, Reese Witherspoon, Brooke Shields, and Alan Thicke.
Finally, sellers of refractive eye surgery appear to also compete on price and respond to profit-based incentives for entry and exit. While the cost of laser eye surgery was as high as $3,000 per eye in the late 1990s, it has fallen to as low as $299 today (depending on a prospective patient’s existing vision quality). Lasik Vision Institute has opened centers across the United States over the past decade. When Lasik Vision Institute entered the market in Cincin- nati in 2002 with an introductory offer of $500 per eye, its chief existing rival in town, LCA-Vision, immediately matched the price.
2This application is based on: “Cost Cuts Debated by Doctors: Sur- gery Often ‘On Sale’,” Cincinnati Enquirer, November 15, 2002, P. B10; and “Turning Surgery Into a Commodity,” New York Times, December 9, 2000, pp. B1 and B4.
Second, and perhaps most importantly, any possible inefficiency cost must be weighed
against the product variety produced by monopolistic competition and the benefits of such
variety to consumers. Similarly, it is probably desirable for firms to continue to have a
dynamic incentive to introduce new differentiated products that better satisfy consumer
tastes, and that incentive could be undermined by regulation.
Third, any sort of intervention has its own costs, which must also be balanced against
the potential gain from expanding output. The costs of operating a regulatory agency may
exceed the noted deadweight loss associated with monopolistic competition. Moreover,
336 Monopol ist ic Competit ion and Oligopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 336Trim Size: 203.2 mm X 254 mm
regulators can find it difficult to obtain the information necessary to achieve a more efficient
output, and mistakes may be made.
13.2 Oligopoly and the Cournot Model Oligopoly is an industry structure characterized by a few firms producing all, or most, of the output of some good that may or may not be differentiated. The number of competitors
is the distinguishing feature of this market structure. With competition (and monopolistic
competition) there are “many” sellers, whereas with pure monopoly there is only one seller.
Oligopoly falls between these extremes. In the United States, there are a number of exam-
ples of oligopolistic industries, including aluminum, cellular phone service, network televi-
sion, and military aircraft.
When there are a small number of competitors, their market decisions will exhibit
strong mutual interdependence, and this characteristic of oligopoly is what makes analy- sis of it difficult. By mutual interdependence, we mean that a firm’s actions (setting price,
for example) have a noticeable effect on its rivals, and so they are likely to react in some
way. In this way, the firms are interdependent. As an example, suppose that General Motors
(GM) is considering a 10 percent cut in the price of its Chevrolet line. This action will have
a significant effect on Ford and Toyota. If they maintain their prices, they will lose sales to
GM. If they cut prices, they can avoid losing sales, but they will make a smaller profit per
car. To complicate matters further, Ford and Toyota have the options of cutting prices by
more or less than the 10 percent cut by GM.
Now consider what this situation means for GM: the results of its own 10 percent price
cut cannot be predicted without knowing how its rivals will respond. For instance, GM’s
sales will rise more if Ford and Toyota maintain their prices than they would if those com-
panies also reduce their cars’ prices. GM must base its decisions on some guess, or conjec-
ture, about its competitors’ responses. What guess should it make? The problem for GM,
and also for us as we try to understand the market, is that it is far from clear which predic-
tion is appropriate. The market functions differently depending on which predictions about
responses each firm makes and acts on. Furthermore, over time the firm may learn that
some of its predictions were wrong and alter its behavior accordingly. But its competi-
tors will also be learning and trying to outguess it. Complex questions of strategy arise in
this setting.
In view of this complicated interdependence, it is perhaps not surprising that we do not
have one agreed-upon theory of oligopoly. In fact, dozens of models have been suggested.
Some of them appear to successfully explain the behavior in some industries over some periods of time, but none appears to explain all oligopolistic behavior. We will discuss a
few of the more important models that have been developed, but be forewarned that deter-
mining when each model applies (if at all) is often difficult.
In addition to smallness in the numbers of competitors, there are two other features of
oligopoly that are likely to have a bearing on how the market performs. First is whether the
product is homogeneous or differentiated. Some oligopolies produce a homogeneous prod-
uct, like aluminum or steel, while others produce differentiated products, like diapers or
airline service in smaller city-pair markets. When the product is differentiated, advertising
(which we will discuss in more detail in the following chapter) is likely to become a more
important influence in the market.
A second important oligopoly feature is the nature of barriers to entry, if any. Oligopo-
listic firms are often thought to realize economic profits, and whenever there are profits
there is incentive for entry. Something must impede entry for profits to persist. Moreover,
just as in the case of monopoly (see Chapter 11), potential entry can influence oligopolists’
pricing behavior.
oligopoly an industry structure characterized by a few firms producing all or most of the output of some good that may or may not be differentiated
Oligopoly and the Cournot Model 337
C13.INDD 10:49:55:AM 08/06/2014 PAGE 337Trim Size: 203.2 mm X 254 mm
The Cournot Model We begin our discussion of oligopoly by considering one of the earliest models, introduced
by French economist Augustin Cournot in 1838.3 Cournot considered a duopoly, an indus- try with just two firms. To illustrate his analysis, Cournot assumed that the two firms sold
water from the only two mineral springs in the area. To follow tradition, we will consider
two firms that sell bottled water, Artesia and Utopia. No entry of new firms is possible. The
bottled water is a homogeneous product, so that only one price can prevail in the market;
that price is determined by the combined output of the two firms in conjunction with the market (industry) demand curve for bottled water. To further simplify the analysis, we
assume that both firms have constant and equal long-run marginal cost curves, and that the
market demand curve is linear.
The key element in the Cournot model is that each firm determines its output based on the assumption that any other firms will not change their outputs. This assumption (which may be an unreasonable one, as we will see) allows us to determine the market price and out-
put. To see how we can do this, consider Figure 13.3, where the market demand and mar- ginal revenue curves are shown as D and MR, and each firm’s marginal cost (MC) and average total cost (ATC) curves are assumed to be constant. Now, let’s examine Artesia’s output decision. Artesia’s most profitable output will depend on how much Utopia is produc-
ing, so first we consider how much Artesia will produce for each possible output of Utopia.
Suppose that Utopia produces nothing. In the Cournot model, Artesia assumes that Uto-
pia will continue to produce nothing whatever output Artesia chooses. In this situation,
Artesia confronts the entire market demand curve and behaves as a monopolist, producing
QM (48), where Artesia’s marginal revenue curve (the same here as the market marginal revenue curve) intersects marginal cost. In the analysis that follows, it will be helpful to
remember that with linear demand and constant marginal cost, the marginal revenue curve
intersects marginal cost at an output half as large as that at which marginal cost intersects
duopoly an industry with just two firms
Cournot model a model of oligopoly which assumes that each firm determines its output based on the assumption that any other firms will not change their outputs
3Augustin Cournot, Réchêrches sur les Principes Mathématiques de la Théorie des Richesses (Paris, 1838), trans. Nathaniel Bacon (New York: Macmillan, 1897).
Figure 13.3 The Cournot Model When Utopia’s output is 32, the vertical axis relevant for Artesia’s output decision is BQU, and Artesia’s demand curve is the BD portion of the market demand curve. Artesia’s marginal revenue curve is then MR(32), and its most profitable output is 32, so combined output is 64.
Dollars per unit
0 QU (32)
QM (48)
QC (96)
OutputQU + QA (64)
B Monopoly price and quantity
Cournot price and quantity
Competitive price and quantity
MR
D
MC = ATC
MR(32)
338 Monopol ist ic Competit ion and Oligopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 338Trim Size: 203.2 mm X 254 mm
the demand curve. In this case, Artesia’s output of 48 is half as large as the output that
would be produced under competition, 96, as shown by the intersection of demand and
marginal cost.
Artesia’s output depends on how much Utopia produces. We have just seen that Artesia
will produce 48 units when Utopia produces nothing. Alternatively, suppose that Utopia
produces 32 units. Then how much will Artesia produce? Artesia believes Utopia will con-
tinue to produce 32 units, regardless of how much Artesia produces (and thus regardless of
what happens to the market price, which will be determined by the two firms’ combined
output). At each price, Artesia can sell 32 fewer units than total quantity demanded, as
shown by the market demand curve. So Artesia’s demand curve is the market demand curve
shifted leftward by 32 units. This idea can be shown in a simpler, yet equivalent, fashion
by moving Artesia’s vertical axis rightward by 32 units without repositioning the demand
curve. Taking the origin for Artesia now to be QU, the demand curve confronting Artesia is the BD portion of the original demand curve. This makes sense. If Artesia produces noth- ing, total output will be 32 (Utopia’s output), price will be BQU, and as Artesia produces and adds to Utopia’s output, price will fall along the BD portion of the demand curve.
Confronted with the demand curve BD, Artesia’s marginal revenue curve is MR(32), the marginal revenue curve when Utopia’s output is fixed at 32. In this situation, Artesia pro-
duces where its marginal revenue curve, MR(32), intersects marginal cost; thus Artesia’s output is 32 units, while the total output of the two firms is 64 units. Note that Artesia’s
output is half the difference between the competitive output (96) and Utopia’s output (32);
this is because MR(32) intersects MC halfway between the new vertical axis for Artesia and the output at which MC intersects the demand curve.
We can now see how Artesia’s output depends on how much Utopia produces. For each
possible output by Utopia, Artesia will produce half the difference between Utopia’s output
and the output at which MC intersects D (96 units). If Utopia produces nothing, Artesia will produce 48; if Utopia produces 10, Artesia will produce 43; if Utopia produces 20, Artesia
will produce 38; and so on. Now that we know what Artesia will do, what about Utopia?
Because the firms have the same costs and because we also make the Cournot assumption
for Utopia (that is, it will take Artesia’s output as a given in determining its output), the
same relationship holds for Utopia. In other words, Utopia will produce 48 units if Artesia
produces nothing, 43 if Artesia produces 10, and so on.
So where will the market equilibrium be? Equilibrium is reached when neither firm has
any incentive to change its output. This occurs when each firm is producing the output it
prefers, given the other firm’s output. In this example, that occurs only when both firms
produce 32 units. To check this, we note in Figure 13.3 that Artesia’s most profitable output
when Utopia produces 32 units is also 32 units. Because Utopia has the same marginal cost
curve, it will also maximize profit by producing 32 units when Artesia produces 32. Neither
firm has any incentive to change its output of 32 when the other firm is producing 32. (The
implication of equal output here arises because the firms have the same costs; if costs differ,
outputs will differ, but the reasoning remains the same.)
Reaction Curves There is another way to arrive at the foregoing conclusion about the Cournot model—by
using reaction curves. Each firm’s reaction curve shows its profit-maximizing output for each possible output by the other firm. In fact, we have already explained the relationships
above. In Figure 13.4, RA is Artesia’s reaction curve. It shows that Artesia will produce 48 units when Utopia’s output (measured on the vertical axis) is zero, 36 units when Utopia’s
output is 24, and so on. Utopia’s reaction curve is RU; it is the same relationship as for Arte- sia but looks different in the graph because the firms’ outputs are on different axes. We can
see how equilibrium can be attained in a step-by-step process, although this should not be
thought of as the actual adjustment process, because if both firms started producing 32,
reaction curve a relationship showing one firm’s most profitable output as a function of the output chosen by the other firm(s)
Oligopoly and the Cournot Model 339
C13.INDD 10:49:55:AM 08/06/2014 PAGE 339Trim Size: 203.2 mm X 254 mm
there would be no reason for either to change. To begin, if Utopia produces nothing, then
Artesia produces 48. When Artesia produces 48, however, we can see by looking directly
above 48 to Utopia’s reaction curve that it will produce 24. With Utopia producing 24,
Artesia would prefer to change its output to 36. And with Artesia producing 36, Utopia will
produce 30. The adjustments follow the arrows, and the firms are both not satisfied with their outputs until they reach the point where each is producing 32 units. Put differently, the
Cournot equilibrium occurs at the intersection of the two reaction curves. As depicted in Figure 13.3, the Cournot equilibrium involves a combined output of 64
units by the two firms in the industry. It is important to note that this amount is greater than
the pure monopoly output (48 units) and less than the competitive output (96 units). A total
output lying between that of pure monopoly and competition is characteristic of most oli-
gopoly models. Some other things are clear from inspection of Figure 13.3. Price exceeds
marginal cost, and, because average cost equals marginal cost, price is above average cost
and both firms realize economic profit. However, their combined profit is less than the
maximum combined profit possible if the firms together produce the monopoly output. This
fact is significant—it means that if the firms colluded instead of behaving independently as
Cournot duopolists, they could increase their combined profit.
Evaluation of the Cournot Model Is it reasonable for a firm to assume, in choosing its output, that the output of a rival
remains constant? Not in the duopoly setting we have just studied, if the market is still
adjusting toward the Cournot equilibrium. Away from the Cournot equilibrium—that is,
when Artesia changes its output on the assumption that Utopia will keep its output fixed—
it will observe that the assumption is wrong: Utopia does change its output in response to
Artesia’s actions. Yet at each step in the adjustment process, the firms continue to behave
based on an assumption they can see is wrong. Thus, the key assumption of the Cournot
model, that each firm takes the other firms’ outputs as given, appears to be suspect if the
market is still adjusting toward equilibrium.
While this criticism is significant, there are some things that can be said in defense of the
Cournot model. First, note that if the equilibrium is somehow established, firms will not see
Figure 13.4 The Cournot Model with Reaction Curves Each reaction curve shows one firm’s most profitable output as a function of the other firm’s output. For example, when Artesia’s output is 48, RU shows 24 to be Utopia’s most profitable output. The Cournot equilibrium is shown by the intersection of the reaction curves, with each firm producing its most profitable output given the output of the other firm.
Output of Utopia
96
48
32 30 24
0 32 36
48 Output of Artesia
RA
RU
96
340 Monopol ist ic Competit ion and Oligopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 340Trim Size: 203.2 mm X 254 mm
the assumption invalidated. When Artesia sees Utopia producing 32 units, and decides on
32 for itself, based on the assumption that Utopia will not change its output, it will be right.
The assumption becomes implausible only for adjustments to the equilibrium.
Second, the assumption is more plausible the larger the number of firms in the market.
(The Cournot model can readily accommodate any number of firms greater than two, and,
in general, the greater the number of firms, the larger the total industry output as a percent-
age of the competitive output.) With 10 equal-sized firms, if one changes its output by, say,
10 percent, it will represent only a 1 percent change in industry output, which will have a
small effect on price. The other firms may not associate such a small price change with the
actions of one firm because other things, like shifting market demand, can also affect price.
13.3 Other Oligopoly Models The Cournot model serves as a good introduction to oligopoly models by highlighting the
importance of how firms handle the mutual interdependence in such markets. In this section
we explore two other models of oligopoly. Although they by no means represent all the
models that have been suggested, they do indicate some different assumptions a firm in an
oligopoly market might make about rival firms’ actions.
The Stackelberg Model Recall that in the Cournot model, each firm takes other firms’ outputs as constant in deter-
mining its own output. We saw, however, that this assumption may not be valid. So now
suppose that in the same two-firm example, we have one firm that continues to behave in
the naive Cournot fashion, while the other firm wises up and realizes that it should not
assume its rival’s output doesn’t change. In fact, let’s assume that Artesia realizes how Uto-
pia chooses its output (from its reaction curve) and see whether Artesia can use that infor-
mation to realize greater profit. Artesia is the “leader” firm in this case; it chooses its best
output taking Utopia’s reaction into account. Utopia is the “follower” firm; it selects output
in exactly the same way as in the Cournot model, by taking the output of the other firm as
given. This is the essence of the Stackelberg model: a leader firm selects its output first, taking the reactions of naive Cournot follower firms into account.4
Figure 13.5 illustrates the Stackelberg model as it operates for Artesia and Utopia. The
marginal cost, average total cost, and market demand curves are shown in Figure 13.5a;
they are the same as in Figure 13.3. Figure 13.5a shows how Artesia selects its output.
Given Artesia’s output, Utopia’s output can be read off its reaction curve, RU, reproduced directly below in Figure 13.5b. Remember that we are assuming that Artesia knows Uto-
pia’s reaction curve, so that it knows how much output Utopia will produce for each output
Artesia may choose.
Our first task is to determine Artesia’s demand curve under these conditions. This will
not be the market demand curve, but what is referred to as a residual demand curve, which shows how much Artesia can sell at each price. The amount that Artesia can sell at
each price is less than total quantity demanded (as shown by the market demand curve) by
the amount that Utopia produces. For example, suppose that Artesia produces 0. From Uto-
pia’s reaction curve, Artesia knows that Utopia will then produce 48 units. Thus, total
(combined) output is 48 units when Artesia produces nothing, and the market price in Fig-
ure 13.5a will be S. At the other extreme, if Artesia produces 96 units, Utopia will produce nothing, and Artesia will be at point C in Figure 13.5a. That gives two points on Artesia’s
Stackelberg model a model of oligopoly in which a leader firm selects its output first, taking the reactions of follower firms into account
residual demand curve a firm’s demand curve, based on the assumption that the firm knows how much output rivals will produce for each output the firm may choose
4Heinrich von Stackelberg, Marktform und Gleichewicht (Vienna: Julius Springer, 1934). As with the Cournot model, the Stackelberg model can readily be adapted to account for a larger number of firms.
Other Ol igopoly Models 341
C13.INDD 10:49:55:AM 08/06/2014 PAGE 341Trim Size: 203.2 mm X 254 mm
residual demand curve, S and C. Artesia’s residual demand curve (with the assumed linear demand and cost conditions) is just the straight line connecting these points between out-
puts of zero and 96 units. Beyond 96 units of output, Artesia’s residual demand curve coin-
cides with the market demand curve (along CD), since Utopia will produce zero if Artesia produces in excess of 96 units.
To see that straight-line segment SC represents Artesia’s residual demand curve for out- puts between zero and 96 units, suppose that Artesia produces 48 units. From Utopia’s
reaction curve, Artesia knows that Utopia will produce 24, so total output will be 72.
When total output is 72, the price is given by point B on the market demand curve. Thus, Artesia can sell 48 units when price is at the height of point E (the same height as point B), which gives us a third point on Artesia’s demand curve. Note that the horizontal distance
between Artesia’s demand curve and the market demand curve is Utopia’s output. As you
can see, Utopia’s output becomes smaller as Artesia increases output along its demand
curve. In fact, for each one-unit increase in output by Artesia, Utopia reduces output by
one-half unit (as can be seen from Utopia’s reaction curve), so the two firms’ total out-
put increases by only half as much as the increase by Artesia. That is why price declines
less rapidly along Artesia’s residual demand curve than along the market demand curve.
(Artesia’s demand curve is flatter; in fact, the slope is exactly half the slope of the market
demand curve.)
With knowledge of its demand curve, profit maximization by Artesia is straightforward.
With demand curve SCD, the marginal revenue curve is MR, intersecting the marginal cost curve halfway between zero output and the output where marginal cost intersects demand.
Therefore, Artesia’s profit-maximizing output is 48 units, with price shown by the height of
point E. Utopia is producing 24 units, so total industry output is 72 units.
Figure 13.5 The Stackelberg Model When Artesia knows that Utopia will choose output as in the Cournot model, Artesia confronts the demand curve SCD, and its most profitable output is 48. Utopia will produce 24, so total output is 72, higher than when both firms behave as Cournot duopolists.
B
C E
MR
Output
MC
D
(24)
(48)
(a)
A
48 96
48 96 Output of Artesia
RU
(b)
Output of Utopia
S
0
48
0
24
Dollars per unit
342 Monopol ist ic Competit ion and Oligopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 342Trim Size: 203.2 mm X 254 mm
Because we are using the same demand and cost conditions as we did with the Cournot
model, it is instructive to compare the outcomes. Note that total output is higher with the
Stackelberg model (72 versus 64), so price to consumers is lower. Output is closer to the
competitive result than in the Cournot model, but still lies between the competitive and
monopoly outputs. In addition, Artesia is making a larger profit and Utopia a smaller profit
than in the case of a Cournot equilibrium. (This is not shown in the graphs but is easily
verified.) This outcome is to be expected: Artesia is exploiting its superior knowledge of
how Utopia will respond to make a larger profit at Utopia’s expense.
Our discussion highlights a key point: the conjectures a firm makes in an oligopoly mar-
ket about how its rivals will respond can affect firms’ outputs and profits as well as total
industry output. For example, total industry output is higher in a Stackelberg model than in
a Cournot model. And the firm that is a Stackelberg leader can take advantage of its lead-
ership position to set a larger (firm) output, thereby enhancing its profit at the expense of
firms that follow its lead in naive Cournot fashion.
Whether the Stackelberg or Cournot model better describes an oligopoly depends on
the particular market being examined. Where an oligopoly is composed of roughly equal-
sized firms, none with superior knowledge or exercising a leadership position, the Cournot
model is likely to be more apt. However, when one firm is more sophisticated about how
rival firms will react and uses this information to operate as a leader in terms of output,
pricing, and/or the introduction of new products, the Stackelberg model is more appropri-
ate. The leadership role played by Intel in terms of setting price and introducing new prod-
ucts in the computer chip market over the past three decades provides a possible example
of the latter case.
The Dominant Firm Model In the Stackelberg model, the leader firm assumes that rivals display Cournot behavior and
plans its output and price accordingly. We now will examine an alternative model in which
the leader firm makes a different conjecture about the behavior of rival follower firms. In
this model, known as the dominant firm model, the leader or dominant firm assumes that its rivals behave as competitive firms in determining their output. (Sometimes this model is
referred to as the dominant firm with a competitive fringe model because the competitive firms are on the fringe.) The dominant firm model has been used by economists to analyze
the performance of many industries.
Figure 13.6 shows how this market structure operates. To determine what price will
maximize its profit, the dominant firm must know its demand curve. As with the Stack-
elberg model, the dominant firm’s demand curve is a residual demand curve that shows
what it can sell after accounting for other firms’ output. In this case, the other firms in
the market are assumed to behave as competitive firms: they will accept whatever price
is set by the dominant firm and produce an output where their marginal cost equals that
price. The output of the competitive fringe firms can therefore be determined from their
supply curve because they collectively behave as a competitive industry. This supply
curve is shown as SF in the diagram. The market demand curve is DD′. At any price, the dominant firm can sell an amount equal to the total quantity demanded
at that price (as shown by DD′) minus the quantity the fringe firms produce (as shown by SF). For example, the dominant firm’s demand curve begins at P1 because at that price the fringe firms will supply as much as consumers wish to purchase, and the dominant firm
could sell nothing. At the other extreme, if the dominant firm charges a price less than P2, it faces the entire market demand curve because the fringe firms will produce nothing at such
a low price. Between P1 and P2, the dominant firm’s residual demand curve is P1A, where the horizontal distance between this demand curve and the market demand curve at each
possible price shows the output of the fringe firms.
dominant firm model a model of oligopoly in which the leader or dominant firm assumes that its rivals behave like competitive firms in determining their output
Other Ol igopoly Models 343
C13.INDD 10:49:55:AM 08/06/2014 PAGE 343Trim Size: 203.2 mm X 254 mm
Armed with a knowledge of its demand curve, P1AD′, the dominant firm also knows its marginal revenue curve, MRD, and maximizes profit by producing where marginal rev- enue equals marginal cost. With a marginal cost curve shown as MCD, the dominant firm’s profit-maximizing output equals QD. The price is P, the height of the dominant firm’s resid- ual demand curve (not the market demand curve) at output QD. At price P, other firms pro- duce QF as shown by their supply curve, and total output, QT, is the sum of their output and the dominant firm’s output. At price P, consumers wish to purchase an output of QT, and so the market is in equilibrium. At the equilibrium, note that price is above marginal cost
for the dominant firm, but it is equal to marginal cost for the fringe firms, SF. This implies that total output is less than if the industry were competitive. The competitive output for
the dominant firm is where MCD intersects the residual demand curve; at that point both it and the other firms are producing where marginal cost equals price. Total output and price
under competitive conditions are indicated by point C on the market demand curve. One interesting implication of this model is that the share of total industry output produced
by the dominant firm may not indicate how close output comes to the competitive result.
For example, suppose that the supply curve of the fringe firms is perfectly elastic (as with a
constant-cost competitive industry) at price P: the supply curve coincides with the horizontal dotted line in the graph. Then the dominant firm’s residual demand curve is also given by this
horizontal dotted line, and marginal revenue will equal P out to output QT. The dominant firm will produce where its marginal cost curve intersects this horizontal line. Industry output will
be the same, QT, but now the dominant firm is producing about 90 percent of it. Furthermore, price is equal to marginal cost, as under competition, even though one firm is contributing 90
percent of total output. This example illustrates the critical importance of the elasticity of the
competitive firms’ supply curve for the functioning of this sort of market structure.
Recall that this model differs from the Stackelberg model only in what the leader, or
dominant, firm assumes about rival firms’ output. In the Stackelberg model, the leader
firm assumes Cournot behavior on the part of rivals; in this model, it assumes competitive
behavior. The dominant firm model is more appropriate when there are a sufficiently large
number of fringe firms for the assumption of competitive behavior to be plausible.
Figure 13.6 The Dominant Firm Model With the supply curve of fringe firms shown as SF, the residual demand curve of the dominant firm is derived by subtracting the quantity supplied by fringe firms at each price from total quantity demanded at that price; the result is curve P1AD′. The dominant firm maximizes profit by producing QD and charging price P; fringe firms produce QF at that price, so total output is QT.
Dollars per unit
P1
P2
P
0 QD QF QT Output
(= QD + QF)
D
C
A
D′MRD
SF
MCD
344 Monopol ist ic Competit ion and Ol igopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 344Trim Size: 203.2 mm X 254 mm
The Elasticity of a Dominant Firm’s Demand Curve Based on the fact that a dominant firm’s output is equal to the total market output minus the
quantity the fringe firms supply, we can derive the dominant firm’s elasticity of demand as
follows:
� � "D M SF= ⎛ ⎝⎜
⎞ ⎠⎟
+ − ⎛ ⎝⎜
⎞ ⎠⎟
1 1 1
MS MS , (1)
where ηD is the elasticity of the dominant firm’s demand; ηM is the elasticity of the market demand; MS is the dominant firm’s market share; and εSF is the elasticity of supply of the fringe firms.5 To see how to apply the formula, consider the case of the pharmaceutical firm
Pfizer, whose brand-name product Viagra is the market-leading erectile dysfunction (ED)
drug. Suppose, for illustrative purposes, that Pfizer can be taken to be the dominant firm in
the ED market, it has a 50 percent market share, it faces a competitive fringe of firms that
produce the generic equivalent of Viagra, the elasticity of supply by the competitive fringe
is equal to 2, and the elasticity of market demand for ED drugs is equal to 1. Using these
assumptions, we can calculate the elasticity of demand for the dominant firm’s product,
Viagra, as follows:
�D = ⎛ ⎝⎜
⎞ ⎠⎟
+ − ⎛ ⎝⎜
⎞ ⎠⎟
=1 1 0 5
2 1
0 5 1 4
. . . (2)
Even though its output is equal to one-half the entire market output, Pfizer faces a residual
demand with an elasticity of 4. Thus, if the company raises Viagra’s price by just 5 percent,
it will lose one-fifth (20 percent) of its sales.
The formula for the dominant firm’s demand elasticity shows that the demand elasticity
becomes larger when (a) the dominant firm’s market share becomes smaller, (b) the elastic-
ity of the market demand becomes greater, and (c) the elasticity of supply by the competi-
tive fringe becomes greater. For example, if Pfizer’s market share was 10 percent instead of
50 percent, the elasticity of demand for Viagra would be greater:
�D = ⎛ ⎝⎜
⎞ ⎠⎟
+ − ⎛ ⎝⎜
⎞ ⎠⎟
⎡
⎣ ⎢
⎤
⎦ ⎥ =1
1
0 1 2
1
0 1 1 28
. . (3)
versus the 4 already calculated. If the elasticity of the demand for ED drugs was 5 instead
of 1, the elasticity of demand for Viagra would be:
�D = ⎛ ⎝⎜
⎞ ⎠⎟
+ − ⎛ ⎝⎜
⎞ ⎠⎟
⎡
⎣ ⎢
⎤
⎦ ⎥ =5
1
0 5 2
1
0 5 1 12
. . (4)
5To derive this formula, we start with the fact that the dominant firm’s output (QD) equals the market output (QM) minus the output of the competitive fringe (QSF):
QD = QM − QSF. (1n) This relationship also holds for a given change in output that results from a price change:
ΔQD = ΔQM − ΔQSF. (2n) Now divide by QD and multiply the two terms on the right by QM/QM and QSF/QSF, respectively:
Δ = Δ
⎛ ⎝⎜
⎞ ⎠⎟
⎛ ⎝⎜
⎞ ⎠⎟
− Δ ⎛ ⎝⎜
⎞ ⎠⎟
⎛ ⎝⎜
⎞ ⎠⎟
Q Q
Q Q
Q Q
Q Q
Q Q
D
D
M
M
M
D
SF
SF
SF
D
. (3n)
Dividing this expression by ΔP/P yields the formula in the text. Note that the minus sign on the right-hand side became a plus sign because we are treating the elasticity of demand as a positive number; QM/QD equals 1/MS; and QSF/QD equals (QM − QD)/QD or (1/MS) − 1.
Cartels and Col lus ion 345
C13.INDD 10:49:55:AM 08/06/2014 PAGE 345Trim Size: 203.2 mm X 254 mm
instead of 4. And if the elasticity of the fringe supply was 5 instead of 2, the elasticity of
demand for Viagra would be:
�D = ⎛ ⎝⎜
⎞ ⎠⎟
+ − ⎛ ⎝⎜
⎞ ⎠⎟
⎡
⎣ ⎢
⎤
⎦ ⎥ =1
1
0 5 5
1
0 5 1 7
. . (5)
instead of 4.
13.4 Cartels and Collusion In the oligopoly models we have examined so far, individual firms were assumed to behave
independently. Each firm makes a specific conjecture regarding how other firms will
respond to its actions and then maximizes its own profit accordingly, without any concern
for how it affects other firms’ profits. An alternative class of oligopoly models is based on
various types of cooperation among firms. The firms coordinate their pricing and output
decisions in an attempt to increase their combined profit, thereby increasing their individual
profits as well.
The most important cooperative model of oligopoly is the cartel model. A cartel is an agreement among independent producers to coordinate their decisions so each of them will
earn monopoly profit. Because cartels are illegal under the antitrust laws in the United
States (though, surprisingly, not in many other countries), they are not common here. There
have been a number of international cartels, however, and we examine one of the most
famous, OPEC, later in this section. Familiarity with the cartel model is useful, because
collusive practices that fall short of outright cartel agreements can be investigated with it.
We begin by considering what happens if firms in a competitive industry form a cartel, and
then extend the results to oligopolistic markets.
cartel an agreement among independent producers to coordinate their decisions so each of them will earn monopoly profit
APPLICATION 13.2
As the patent on a brand-name pharmaceutical expires, the producer of the drug typically confronts competition from generic manufacturers. Generic manufacturers do little research of their own; rather, they specialize in copy- ing brand-name products after their patents expire. Generic manufacturers tend to become both more numerous and more capable of expanding their output capacity the lon- ger that a brand-name drug is “off patent.” In such a setting, therefore, the brand-name drug producer can be taken to be the dominant firm, with its market share decreasing and the competitive fringe’s elasticity increasing the more years the brand-name drug is off-patent.
What does the dominant firm model predict about the price charged by a brand-name drug maker and the sensitivity of consumers to the brand-name drug’s price as the number of years that the drug has been off-patent
The Dynamics of the Dominant Firm Model in Pharmaceutical Markets
increases? As the fringe supply curve tends to shift right- ward (see Figure 13.6) as the time since the brand-name drug maker’s patent expired increases, it works to shift the dominant firm’s residual demand curve leftward and put downward pressure on the price charged by the brand- name drug maker. Moreover, as both the brand-name mak- er’s market share decreases and the elasticity of the fringe supply increases, equation (1) indicates that the demand elasticity facing the brand-name maker tends to increase with the time since patent expiration. The available empiri- cal evidence bears out these theoretical predictions gener- ated by the dominant firm model.6
6Richard Caves, Michael Whinston, and Mark Horwitz, “Patent Expi- ration, Entry, and Competition in the U.S. Pharmaceutical Industry,” Brookings Papers on Economic Activity: Microeconomics (1991), pp. 1–48.
346 Monopol ist ic Competit ion and Ol igopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 346Trim Size: 203.2 mm X 254 mm
Cartelization of a Competitive Industry Let’s see how a group of firms in a competitive market can earn monopoly profits by coor-
dinating their activities. We assume that the industry is initially in long-run equilibrium,
and then we identify the short-run adjustments (with existing plants) that the industry’s
firms can make to reap monopoly profits for themselves. Figure 13.7b shows the industry
equilibrium with a price of P and an output of 1,000 units. Figure 13.7a shows the com- petitive equilibrium for one of the firms in the industry. Note that initially, the firm faces
the horizontal demand curve d at the market-determined price and produces an output of 50 units. To simplify matters, suppose that there are 20 identical firms in the industry, each
producing 50 units of output.
Next, the firms form a cartel and agree to restrict output to attain a higher price. Each firm
agrees to produce an identical level of output, equal to one-twentieth of total industry out-
put because there are 20 firms. The cartel agreement has the effect of changing the demand
curve facing each firm. Before the agreement, if one firm alone reduced output, its action
would not appreciably affect price, as shown by the firm’s horizontal demand curve d. Now, however, other firms match a restriction in output by one firm, so when one firm cuts output
by 15 units, all firms match the reduction, industry output falls by 300 units, and price rises
significantly. The demand curve showing how price varies with output when firms’ output
decisions are coordinated in this way is the downward-sloping curve d* in Figure 13.7a. At any price, the quantity on the d* curve is 1/20 that on the industry demand curve.
Faced with this downward-sloping demand curve, the firm’s profit-maximizing output
occurs where its short-run marginal cost curve SMC intersects the new marginal revenue curve mr*. Output is 35 units, and because all 20 firms reduce production to the same out- put level, total output falls to 700 units and the price rises to P1. Each firm is now mak- ing an economic profit. Indeed, the idealized cartel result is just the same as if the industry
Figure 13.7
(a)
Dollars per unit
P1
P
0 q1 (35)
q (50)
q2 Output
SMC
SAC
d
d* mr*
Firm
(b)
Dollars per unit
P1
P
0 Q1 (700)
Q (1,000)
Output
SS
DMR
Market
A Cartel Under competitive conditions, industry output is Q and price is P. If the firms in the industry form a cartel, output is restricted to Q1 in order to charge price P1, the monopoly outcome. Each firm produces q1 and makes a profit at price P1.
Cartels and Col lus ion 347
C13.INDD 10:49:55:AM 08/06/2014 PAGE 347Trim Size: 203.2 mm X 254 mm
were supplied by a monopoly that controlled the 20 firms. Figure 13.7b illustrates the result, with the short-run supply curve SS (the sum of the SMC curves of the firms) intersecting the industry marginal revenue curve MR at an output of 700 units and a monopoly price of P1. By forming a cartel and restricting output to achieve the monopoly equilibrium, the firms
maximize their combined profit. Figure 13.7b shows the total market effect of the coordinated
output reduction by the 20 firms; Figure 13.7a shows the effects on each firm individually.
Firms can always make a larger profit by colluding rather than by competing. Acting alone, com-
petitive firms are unable to raise price by restricting output, but when they act jointly to limit the
amount supplied, price will increase. As we will see in the next subsection, however, achieving a
successful cartel in practice is not as simple as it may seem.
APPLICATION 13.3
The common wisdom is that the Internet serves to promote competition among suppliers, thereby creat- ing bargains for surfing shoppers.7 Indeed, a survey by Erik Brynjolfsson and Michael Smith of MIT finds that prices on the Internet are 9 to 16 percent lower than in retail outlets.
Although the Internet lowers the cost of searching and thus makes it easier for buyers to shop around for a lower price, Hal Varian, chief economist for Google and the fund- ing dean of the School of Information at the University of California, Berkeley, cautions that there is a good reason the Web might actually result in higher prices for consumers. This is because if there are only a few sellers the availabil- ity of low-cost information about the prices they are charg- ing could make it easier for them to coordinate their pricing through the Web. As a historical example, Varian points to the Joint Executive Committee set up by the major U.S. rail- roads in the 1880s prior to the enactment of antitrust laws. The committee served as a cartel by collecting and publish- ing information about the prices individual railroads charged and their weekly shipments. The railroads often cheated on the published prices by offering lower rates to shippers in secret in exchange for more business. Such cheating would have been mitigated by a public Internet exchange,
Does the Internet Promote Competition or Cartelization?
according to Varian, since each railroad could readily moni- tor the prices charged by others. Any attempt to offer a lower price could quickly be countered, thereby making the cartel less vulnerable to cheating.
A modern parallel to the activities of the Joint Executive Committee is the manner in which airlines post fares online: an airline will announce its rates and associated terms and then watch to see how competitors respond. In the late 1980s airlines began to use online reservation systems to signal their pricing intentions to each other. For example, United would post its intended fare changes at 2 on Thursday. If rival airlines followed suit by 6 , United’s price remained in effect. If not, United would restore its price to the pre-2 level. In 1992, the Department of Justice brought an antitrust suit against several large airlines in an effort to limit the extent to which online reservations systems serve as a signaling device and thereby promote cartelization.
In the long run, the key will be the number of sellers in any online category of goods or services. If there are many sellers, the extra flow of information through the Internet is likely to work to the benefit of buyers, pushing prices down. But in Web-based exchanges where there are only a few sellers and many buyers, the availability of more timely price information may serve to promote cartelization, thereby increasing prices to consumers.
7This application is based on Hal R. Varian, “Online Commerce Cre- ates Strange Competition,” New York Times, August 24, 2000, p. C2.
Why Cartels Fail If cartels are profitable for the members, why aren’t there many more? One reason is that in
the United States, they are illegal. But even before there were laws against collusive agree-
ments, cartels were rare except when actually supported by government; when they did
exist, they were short-lived. Three important factors appear to contribute to cartel instability:
1. Incentive to cheat. A cartel achieves monopoly profit through its members restricting output below the levels that each would individually choose, and this reduction results in
a higher price. Once a higher price is achieved, individual cartel members could earn even
348 Monopol ist ic Competit ion and Oligopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 348Trim Size: 203.2 mm X 254 mm
more profit by expanding output. Each firm would like to enjoy the cartel’s benefit—a
higher price—without incurring the cost—lower output. If only one firm expands output,
price will not fall appreciably, but the additional sales at the monopoly price will add
significantly to that firm’s profit. It is thus in each firm’s self-interest to violate the cartel
agreement to restrict output.
Figure 13.7 illustrates the incentive to cheat on the cartel agreement. If the firm in Fig-
ure 13.7a adheres to the cartel agreement, it will produce q1 and sell at price P1. However, note what happens if the firm expands output beyond q1 while the other firms continue to abide by the cartel agreement’s restrictions. In this event the firm faces a horizontal
demand curve at price P1; that is, one firm expanding output alone will not affect price. Remember that the downward-sloping demand curve d* is relevant only for simultaneous expansion and contraction of output by all firms. The firm acting alone can increase its
profit significantly by expanding output, since marginal revenue (equal to price with the
horizontal demand curve) is above marginal cost at q1. Profit will be maximized if the firm increases sales to q2 at the price of P1.
Every firm has the same incentive to expand output and cheat on the cartel agreement.
Yet if many firms do so, industry output increases significantly, and price falls below the
monopoly level. It is in each firm’s interest to have other firms restrict their output while it
increases its own. Every firm’s self-interest is therefore a threat to the cartel’s survival. To
be successful, the cartel must have some means of monitoring and enforcing its agreement.
The foregoing suggests why government backing generally is so essential to ensuring a
cartel’s stability. Government provides the means of monitoring and enforcing a cartel agree-
ment. The caviar cartel provides a good example of this.8 Prior to the collapse of the Soviet
Union, the Ministry of Fisheries in Moscow set stringent quotas for the annual sturgeon
catch, the source of caviar, one of the world’s most expensive delicacies. Close government
monitoring limited poaching and illegal dealing in caviar. However, as the Soviet Union
disintegrated, four new independent states and two autonomous regions appeared around the
Caspian Sea—the location of over 90 percent of the world’s sturgeon stocks. Central author-
ity evaporated, and the independent actions of numerous caviar poachers and illegal traders
ripped the formerly tightly regulated cartel wide open. Caviar prices plummeted.
2. Difficulty of Reaching Agreement. In Figure 13.7, we assumed that the firms in a cartel have identical cost curves, making agreement on the profit-maximizing cartel output and
price relatively easy. But when firms differ in size, cost conditions, and other respects,
agreement will not come as easily since the firms will have different goals. If, for example,
the cartel members’ costs differ, they will disagree on what price the cartel should set. The
problems become even more acute when the firms must make long-run investment deci-
sions. Every cartel member will want to expand its capacity and share of total output and
profit, but not all can be allowed to do so.
These problems are basically political, and no matter what policy the cartel follows, it
will reflect a compromise among divergent views. As happens with any compromise, some
firms will be unhappy with the outcome, and those firms are all the more likely to refuse to
join the cartel or join but violate any cartel agreement on output.
Agreement will also be more difficult the less homogeneous the product. For example,
in the United States there are two primary areas in which oranges are grown: Florida and
California–Arizona. Through regulations instituted in 1937, the U.S. government (as an
exception to antitrust laws) has allowed growers’ cartels to control prices and supplies
in the two areas. The organization of a growers’ cartel has been much more problem-
atic in Florida than in California–Arizona.9 This reflects the longer growing season and
greater varieties of oranges that can be produced in the climate and soil conditions there.
8“Bootleggers Thrive, Sturgeons Flounder, as Caviar Cartel Splits,” Washington Post, June 1, 1992, pp. 1 and 8. 9Gary D. Libecap and Elizabeth Hoffman, “The Failure of Government-Sponsored Cartelization and the Development of Federal Farm Policy,” Economic Inquiry, 33, No. 3 (July 1995), pp. 365–382.
Cartels and Col lus ion 349
C13.INDD 10:49:55:AM 08/06/2014 PAGE 349Trim Size: 203.2 mm X 254 mm
Because there are more product dimensions that must be taken into account, Florida orange
growers have been less successful at reaching an effective cartel agreement—despite the
U.S. government’s official approval of such an agreement.
3. Profits attracting entry. If a cartel achieves economic profits by raising the price, new firms have an incentive to enter the market. If the cartel cannot block entry of new firms,
price will be driven back down to the competitive level as production from the “outsiders”
reaches the market. Indeed, if an increase in the number of firms in the market causes the
cartel to break down, then price will temporarily fall below the cost of production, forcing
losses on the cartel members. The prospect of entry by new competitors eager to share in
the profits is probably the most serious threat to cartel stability.
To be successful, therefore, a cartel must be able to get its members to comply with
cartel policy (limiting output) and to restrict entry into the market. These tasks are not eas-
ily accomplished, and history is strewn with examples of cartels that flourished for a short
time only to disintegrate because of internal and external pressures.
APPLICATION 13.4
The Organization of Petroleum Exporting Countries (OPEC) has relied on such measures as accounting firms to monitor member nations’ outputs to eliminate cheating on production quotas.10 These measures, however, have not been entirely successful. This is perhaps not surprising, given that member states possess differing production costs, petroleum reserves, time horizons, and goals. Enforcing agreements is also difficult when member states are at war with one another, as in the case of Iran and Iraq in the 1980s.
To demonstrate the difficulties facing OPEC—or any other cartel—in fixing prices, Raymond Battalio, a Texas A&M professor, conducted an experiment with a class of 27 intro- ductory economics students. In the experiment, each stu- dent was asked to write either a 0 or a 1 on a slip of paper. A 1 indicated a willingness to adhere to a collusive agree- ment, whereas a 0 signified a desire to cheat. There were real money payoffs associated with the game, and they were
The Difficulty of Controlling Cheating
structured so that if everyone chose 1, the total payoff to all the students would be maximized. The payoff to any indi- vidual student, however, was maximized if all other students opted for 1 and the individual cheated and selected 0. If more than one student cheated, the payoff for cheating decreased as the number of cheaters increased, but it was still greater than the payoff to noncheaters. Significantly, the payoffs associated with all possible outcomes were disseminated to students at the start of the experiment.
On the first round of the experiment, there were six non- cheaters and 21 cheaters. Allowed to discuss the outcome, the students quickly realized that they could all be better off if they all voted to honor the agreement and place a 1 on their ballots. A student “leader” proposed, “Let’s all put down 1, and nobody cheat.” After the students reached an explicit oral agreement to all vote 1 another vote was taken. There were four 1 votes and 23 0 votes. The ringleader muttered, “I’ll never trust anyone again as long as I live.” Asked how he voted, the ringleader replied, “Oh, I voted 0.” The results suggest that in a cartel setting, self-interest leads to efforts to maximize returns individually, even when the risk is great of lowering overall returns both to oneself and to one’s “partners.”
10This application is based on “OPEC May Monitor Members’ Out- put on Site to Reduce Cheating on Quotas,” Wall Street Journal, November 13, 1987, p. 44; and “All for One . . . One for All? Don’t Bet on It,” Wall Street Journal, December 4, 1986, p. 1.
APPLICATION 13.5
While collusive agreements between different firms typically come to mind when cartels are mentioned, the model has wider applicability. Essentially, any firm with multiple production facilities or distribution channels and some market power faces a cartel management problem.
The Rolex “Cartel”
The firm must coordinate the output and pricing decisions of its various plants and distribution channels to ensure that its total economic profit is maximized. To the extent, for example, that Rolex has some market power, it needs to ensure that one of its licensed dealers does not attempt to
350 Monopol ist ic Competit ion and Oligopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 350Trim Size: 203.2 mm X 254 mm
Oligopolies and Collusion Much of what has been said about the consequences of cartel formation in a competitive
industry also applies to an oligopolistic industry. Firms in an oligopolistic industry can
increase their profits (they may already be making some pure economic profits to begin
with) by colluding. The key to understanding this is our earlier finding that oligopoly out-
put is usually greater than that of pure monopoly. This means that the oligopolistic firms’
combined profit can be increased if total output is restricted to the monopoly level, and,
with an appropriate sharing arrangement, each firm can realize greater profit. The incentive
to collude—the prospect of higher profit—also exists in an oligopoly.
We generally expect collusion to be more common in oligopolies than in competitive
markets because there are only a few firms in an oligopolistic industry. The limited number
of firms means that fewer parties must participate in the collusive agreement, making reach-
ing agreement easier. Additionally, monitoring is simpler when few firms are involved; it is
easier to detect cheaters.
Nonetheless, factors that inhibit the formation and maintenance of cartels in competitive
markets are also present when oligopolistic firms collude. Each party to the cartel agree-
ment still stands to make more profit by cheating than by abiding by the cartel agreement.
Even though one firm does not face a horizontal demand curve if it cheats—as does the
competitive firm in Figure 13.7a—it will face a more elastic demand curve when cheating
than when complying, so its profit will be greater if it cheats and the other firms abide by
the agreement. Furthermore, the higher price achieved by collusion can still prompt entry
by new firms, and that also undermines the success of the cartel.
A final point is important. It is not necessary for all of a market’s firms to participate
in a cartel for it to be worthwhile. Several firms producing most of an industry’s output
can increase their profit by colluding, although not by as much as when the entire industry
is cartelized. In this case, the colluding firms will behave as if they were one giant firm
and collectively exploit whatever monopoly power they have. The remaining firms then
benefit from the higher price set by the colluding firms. In this case of partial market car-
telization, the results can be analyzed using the dominant firm model illustrated earlier in
Figure 13.6. In that graph, the marginal cost curve of the dominant firm is interpreted now
as the summed marginal costs of the colluding firms, and they collectively maximize profit
subject to the residual demand curve. Therefore, partial cartelization of an industry pro-
duces price and output lying between those found with pure monopoly and competition.
We will see an example of this model’s application in the next subsection.
cheat on the Rolex “cartel” by selling more than a certain number of watches at the agreed-upon price. While such cheating may benefit the individual dealer, it undermines the overall profitability of the cartel. To limit the undermin- ing, Rolex may set up exclusive territories for its various dealers, allot a set number of watches to each dealer, and specify a manufacturer’s suggested retail price.
In the mid-1980s, Rolex became disgruntled with Carl Marcus, a Beverly Hills retailer of its watches.11 Marcus had been buying Rolex watches from other dealers around the country who were unable to sell their allocated number.
He then sold those watches in the Beverly Hills area at a discount of 5 to 25 percent below the price most jewelers were charging for the same items.
Rolex launched an advertising campaign against Mar- cus, alleging that he had sold a used watch to a customer while claiming it was new. The advertisements asked Mar- cus’s former customers to have their watches inspected by Rolex to verify that they had not been similarly duped. The underlying motive of the advertising campaign was not so much to protect unsuspecting customers as it was to pre- serve Rolex’s market power and profit. By offering to inspect the watches Marcus sold, Rolex could obtain the registration numbers and identify the dealers who had undermined the cartel by selling some of their allotments to Marcus.
11“The Rolex War Rages on for Beverly Hills Jeweler,” Los Angeles Times, November 13, 1987, pp. 1 and 17.
Cartels and Col lus ion 351
C13.INDD 10:49:55:AM 08/06/2014 PAGE 351Trim Size: 203.2 mm X 254 mm
The Case of OPEC Few Americans took notice when OPEC was formed in 1960. Originally containing just
five member nations, the cartel grew to an ominous 13 countries by 1973. During these
years, however, OPEC could not be judged successful. World oil prices actually declined
slightly in the 1960s. OPEC had not yet learned to use its potential market power.
All that changed with the 1973 Arab–Israeli War. During the war, OPEC’s Arab mem-
bers temporarily cut off oil exports. The result was dramatic: oil prices nearly quadrupled
in a matter of months. The price of a barrel of oil on the world market was $3 in 1973 but
jumped to $11 in 1974, providing a graphic demonstration of what an output restriction
could accomplish. OPEC continued to hold down output after resuming exports; by acci-
dent, OPEC had learned how to run a cartel! After the war’s end, oil prices remained rela-
tively stable (in real terms) until they received another jolt in 1978, when revolution swept
Iran. Iranian exports, which accounted for 20 percent of all OPEC exports, fell to almost
zero. The world oil price rose to an average of $20 per barrel in 1979. Soon, thereafter, the
Iran–Iraq War, which began in 1980, resulted in the widespread destruction of oil-produc-
ing facilities in both countries and reduced oil exports further. The world oil price jumped
again to over $30 per barrel in 1980 and averaged $35 in 1981.
Before continuing the historical saga of developments in the world oil market, let’s
examine the market as an example of a cartel. Specifically, because OPEC does not con-
trol all supply sources (in 1972, OPEC production was 64 percent of noncommunist world
output), we will use the dominant firm model. In view of the history of OPEC, this is a bit
of an oversimplification—OPEC has not always been monolithic enough to act as a single
profit-maximizing firm. Nonetheless, we can gain insight into the way this market has func-
tioned by treating OPEC as a dominant firm.
In Figure 13.8 we show world demand for oil as DD′ and non-OPEC supply as SF. These relationships should be interpreted as short-run, where the elasticities are relatively low.
As we will see, low elasticities of demand and supply from fringe firms have a significant
effect on the outcome. The marginal cost curve for OPEC is shown as MCO, exhibiting
Figure 13.8 OPEC Cartel as a Dominant Firm World oil demand is shown by DD′ and non-OPEC supply as SF. The residual demand curve confronting OPEC is then P1AD′. With its marginal cost curve MCO, OPEC produces QO at price P. Non-OPEC output at that price is QF, so total output is QT.
T
D SF
MCO
MRO
C
A
D ′ QF QO QT Output
Dollars per unit
P1
PC
0
P
352 Monopol ist ic Competit ion and Oligopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 352Trim Size: 203.2 mm X 254 mm
substantially lower costs than other sources of supply, another characteristic of this mar-
ket. OPEC’s residual demand and marginal revenue curves are derived as before. Under
competitive conditions (which prevailed before 1973), when OPEC and non-OPEC produc-
ers are producing where marginal cost equals price, price is PC and total output is shown at point C on the demand curve. When OPEC behaves as a dominant firm, however, it produces where marginal cost equals its residual marginal revenue, an output of QO with a price of P. Note that price has risen sharply compared with the competitive price. This result should be contrasted with that shown earlier in Figure 13.6. In that representation of
the dominant firm model, we assumed that demand and fringe supply were more elastic.
A comparison of these graphs indicates why low elasticities of demand and fringe supply
imply a higher cartel price. Equation (1) on page 344 also suggests this: decreases in the
elasticity of the market demand and the fringe supply both work to lower the elasticity of
demand faced by the dominant firm.
The Reasons for OPEC’s Early Success Because much of OPEC’s early success was dependent on low elasticities, we need to be
sure we understand why this was a characteristic of the world oil market.
The price elasticity of demand for oil is quite low in the short run. When oil prices rose, consumers were caught with a stock of energy-consuming capital designed for cheap oil.
Houses, office buildings, and gas-guzzling automobiles could not be replaced overnight;
it took time for higher energy prices to have a substantial effect on energy consumption. A
low price elasticity means, of course, that moderate output restrictions will produce a large
price increase, just what we see in Figure 13.8.
The price elasticity of supply of oil from non-OPEC suppliers is also quite low in the short run. The biggest threat to a cartel is increased production by noncartel members or entry of new firms. OPEC’s ability to raise oil prices successfully depended on total (OPEC
and non-OPEC) output; if OPEC output restrictions were matched by substantial increases
by others (as would happen with highly elastic non-OPEC supply), price would be largely
unaffected. How much non-OPEC suppliers could increase output depended on their price
elasticity of supply. Because non-OPEC producers were already producing at near-capacity
levels, their immediate ability to increase output was limited. Furthermore, new oil fields
could be brought into production only after a lengthy process of seismic exploration, drill-
ing, and installation of pipelines requiring several years. Thus, non-OPEC producers were
unable to respond quickly to the higher prices produced by the cartel.
These characteristics of the world oil market help explain OPEC’s early success. In addi-
tion, OPEC also enjoyed another advantage: oil-importing nations frequently adopted poli-
cies that strengthened OPEC’s position. In the United States, for example, price controls
on oil kept the price received by domestic oil producers artificially low and discouraged
production and exploration. Price controls on natural gas discouraged production of this
alternative energy source. Similarly, tough environmental restrictions on the mining and
use of coal slowed the transition to coal as another energy alternative. Finally, price con-
trols on oil products such as gasoline and heating oil kept the prices of such products below
world market levels and encouraged consumption. Encouraging domestic consumption and
discouraging domestic production implies an increase in demand for oil from OPEC, and
the United States inevitably became more dependent on imported oil during the 1970s.
The Rest of the OPEC Story Our story, however, does not end with the $35-per-barrel price of 1981. As economic the-
ory tells us, long-run elasticities are greater than short-run elasticities. As time passed con-
sumers responded to higher energy prices by switching to more energy-efficient homes,
appliances, and cars, while non-OPEC suppliers increased output. Both responses put
increased pressure on OPEC in the early 1980s. More specifically, world oil consumption
Cartels and Col lus ion 353
C13.INDD 10:49:55:AM 08/06/2014 PAGE 353Trim Size: 203.2 mm X 254 mm
fell from a 1979 high of 51 million barrels per day to less than 45 million in 1985 (particu-
larly impressive since the prior long-run trend had been a 3 percent annual rate of increase
in consumption). At the same time, non-OPEC production increased sharply, from 15 mil-
lion barrels per day in 1977 to 24 million barrels in 1985. With total consumption down
and non-OPEC production up, OPEC had to restrict its output sharply to try to hold the
price up. Even though OPEC cut production from 31 million barrels per day in 1979 to
barely 16 million barrels in 1985, the price of oil still fell below its previous level. The per-
barrel price dropped to $32 in 1982 and to $27 in 1984, and it continued falling gradually
until it stabilized at around $18 in 1987.
The experience of OPEC after 1981 was almost a textbook case of the operation of eco-
nomic forces that undermine cartels. According to Businessweek:12 “The OPEC cartel is facing strains that it has never experienced before . . . . OPEC is so divided by cutthroat
competition and internal political bickering that the organization is unlikely to find a way to
agree anytime soon . . . Under-the-table discounts . . . are common.”
External pressures that diminished OPEC’s power were also at work:13 “Demand for
OPEC oil . . . collapsed under the combined impact of . . . surprisingly high conservation,
expanded use of alternative energy sources such as coal and gas, and rising non-OPEC
production of oil from Mexico, the North Sea, and the North Slope of Alaska.” (Long-run
price elasticities of demand are higher than short-run price elasticities and were being felt
as consumers switched to substitutes. Entry at the high cartel price was taking place.)
To illustrate in analytic terms how the passage of time affected the market, long-run
demand and non-OPEC supply responses can be incorporated into Figure 13.8. To avoid
cluttering the diagram, we have not done so, but we recommend that you work this out.
Specifically, assume that the original competitive equilibrium was a long-run equilibrium.
Then a more elastic long-run demand curve will pass through point C on the short-run demand curve, and a more elastic supply curve will pass through point T on the non-OPEC supply curve. With these curves, where will the long-run cartel equilibrium lie? You will
find that in the long run, price will be lower, non-OPEC output will be greater, and OPEC
output will be smaller (all compared to the short-run cartel equilibrium shown in our
graph)—just the events we saw emerge in the 1980s.
Most observers concluded that OPEC’s power had largely eroded by the late 1980s. In
August 1990, however, Iraq invaded Kuwait and supply was disrupted again. The price
of oil quickly rose to $40 per barrel, before falling back to about $20 per barrel after the
United States defeated Iraq in the Gulf War in early 1991. These events do not mean that
OPEC was exercising control, of course. Small changes in supply can have pronounced
effects on price in a competitive market in the short run when short-run demand is very
inelastic. As long as political turmoil pervades the Mideast, a major source of world oil,
volatility is likely to characterize this market.
By 1998, the price of oil had fallen below $12 a barrel, the lowest level in a decade,
because of increased global production and a decrease in OPEC’s share of total output to
40 percent (from more than 50 percent in the early 1970s). The decline in the price of oil
was arrested in 1999, due to a pact negotiated between the three largest exporters of oil to
the United States: Saudi Arabia, Venezuela, and Mexico. The first two countries belong
to OPEC, but Mexico does not. The three countries respectively account for 15, 18, and
14 percent of U.S. imports. The agreement, known as the Riyadh Pact, consisted of pledges to reduce output by 1.5 to 2.0 million barrels a day, about 2 to 3 percent of world produc-
tion of 73 million barrels a day. The price of crude oil rose above $30 in the pact’s wake.
By 2008, the price of crude oil rose further to over $120 per barrel due to, among other
things, fears of a disruption of supplies from the Mideast related to Iran’s nuclear weapons
12“The Leverage of Lower Oil Prices: On World Economies, on OPEC Countries, on the Oil Industry,” Businessweek, March 22, 1982, p. 69. 13Ibid., p. 69.
354 Monopol ist ic Competit ion and Oligopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 354Trim Size: 203.2 mm X 254 mm
program, a worldwide economic recovery, and increased overall demand—especially from
countries such as China and India (over the past three decades China has gone from being
a net exporter of crude oil to the world’s second-largest net importing nation).14 While the
macroeconomic downturn put downward pressure on oil prices between 2008 and 2010
(to $40 per barrel in early 2009), a global economic recovery starting in the second half of
2010 saw prices once again rise to over $100 per barrel by 2011. Tensions in the Mid-East
in 2012 to 2013 in places such as Libya, Egypt, and Syria drove prices to a further high of
over $110 per barrel by mid-2013.
The recent higher price of crude oil appears to largely reflect changes in worldwide demand
rather than any increase in OPEC’s pricing power since by 2011 OPEC’s share of total world
output had fallen to under 33 percent. And any successful effort by OPEC to exercise pricing
power (such as the Riyadh Pact) is bound to be short-lived on account of the reasons given ear-
lier for cartel failure. For example, the Mexican officials who were key to brokering the Riyadh
Pact, didn’t hold any illusions about its long-run prospects. Then-president of Mexico, Ernesto
Zedillo, recalled that when he was an economics student at Yale University, one of his pro-
fessors engaged his class in a game to test the limits of oligopolistic behavior (see Applica-
tion 13.4).15 “Within a few minutes, somebody would start cheating; it never failed,” Zedillo
observed. “No market with more than two participants can sustain a cartel.”
While the long-run prospects for any cartel in the oil market may not be so sanguine, the
last four decades also make clear that this market will continue to display fairly dramatic
swings in price in the short run (due to the inability of consumers and producers to respond
in a significant fashion to unexpected changes in demand and/or supply). The magnitude
of any response in consumption and production, however, will continue to be greater the
more time consumers and producers are given to respond. For example, in the wake of
largely low oil prices in the 1990s, Americans switched back to using less fuel-efficient
motor vehicles. (SUVs accounted for 54 percent of all motor vehicles in the United States
by 2008 versus 5 percent in 1990.) This switch is likely to be reversed if the higher oil
prices witnessed in the most recent few years persist.
14A devaluation of the dollar relative to other currencies also played a role. This devaluation, reflecting looser monetary policy by the Federal Reserve, led to a rise in the dollar prices of goods, such as crude oil, imported by the United States. 15“ ‘Big 3’ Exporters’ Pact to Cut Oil Output Signals Seismic Shifts,” Wall Street Journal, June 23, 1998, pp. A1 and A10.
SUMMARY
A monopolistically competitive market is one in
which there are a large number of competing firms, but
each firm produces a differentiated product.
Each firm in a monopolistically competitive market con-
fronts a demand curve that is fairly, but not perfectly, elastic.
In long-run equilibrium, firms in a monopolistically
competitive market make zero economic profits. They
are not, however, producing at the minimum point on
their long-run average cost curves since the LAC curve
is tangent to a downward-sloping demand curve.
Oligopoly is characterized by a few firms that
together produce all or most of the total output of some
product. A pronounced mutual interdependence among
the decisions of firms in the industry results.
What one firm does has a decided impact on other
firms in an oligopoly, but the way that other firms will
react is uncertain.
There are three common oligopoly models: Cournot,
Stackelberg, and dominant firm.
Oligopoly firms can collude and operate as a cartel.
The variety of models for studying oligopoly and the
differences in their implications make generalizing about
the effects of oligopoly difficult. In some cases, prices
are predicted to be near the monopoly level. In other
cases, they hover near the competitive level.
Oligopolistic outcomes lie somewhere between the
monopoly and competitive results and differ from one
industry to another.
Review Quest ions and Problems 355
C13.INDD 10:49:55:AM 08/06/2014 PAGE 355Trim Size: 203.2 mm X 254 mm
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
13.1 What are the assumptions of the theory of monopolistic competition? In what way do these assumptions differ from
those of the perfectly competitive model?
13.2 Explain the relationship between the demand elasticity and the excess capacity that occurs for a monopolistic competi-
tor.
13.3 Explain how, in the Cournot model, the output of one firm depends on the output of other firms. Specifically, in
Figure 13.3, what will be the output of Artesia if Utopia pro-
duces 32 units? If Utopia produces 48? If Utopia produces 64?
13.4 Starting from the Cournot equilibrium in Figure 13.3, sup- pose that the marginal and average total cost curves (which
are the same for both firms) shift downward. Explain how the
firms adjust to a new Cournot equilibrium.
13.5 How does the Stackelberg model differ from the Cournot model? Which model predicts that output will be higher?
13.6 Explain how the residual demand curve confronting the dominant firm in the dominant firm model is derived. In
this derivation, what is assumed regarding how the output of
other firms is determined? How does it differ from the Cournot
assumption?
13.7 Explain how equilibrium is determined in the dominant firm model. If market demand increases, how will a new equi-
librium be determined?
*13.8 Suppose that the supply curve of the “follower” firms in the dominant firm model is perfectly horizontal. Does the dom-
inant firm still have the power to set the industry price?
13.9 “Because firms in any industry can always make greater profits by colluding, there is an inevitable tendency for com-
petitive industries to become cartels over time.” Is the first part
of this statement correct? Is the second part? Explain.
13.10 What problems usually make cartels collapse? How was OPEC able to avoid this fate, at least through the mid-1980s?
13.11 Consider the dominant firm model and treat OPEC as the dominant firm. Explain how OPEC would determine the
price of oil and the level of output produced by the cartel. How
would OPEC’s price and output be affected by new discoveries
of oil that shift the supply curve of oil for non-OPEC members
to the right?
13.12 “There is no general theory of oligopoly.” Explain this statement.
13.13 Under which oligopoly model does the outcome most nearly resemble that obtained with pure monopoly? Under
which oligopoly model does the outcome most nearly resemble
that obtained with perfect competition?
13.14 In an analysis of the automobile industry, what factors would you consider in determining whether to use the competi-
tive model, the monopoly model or one of the oligopoly models?
13.15 Suppose that there were three identical firms instead of only two under the cost and market demand conditions out-
lined in Section 13.2. What would be the Cournot equilibrium
in terms of each firm’s output as well as the total market out-
put? If there were four identical firms sharing the market?
13.16 In a two-firm Stackelberg model of oligopoly, can both firms be “leaders”? Explain why or why not.
13.17 Suppose that Iran and Iraq are Cournot duopolists in the crude oil market and face the following market demand func-
tion:
P q q= − +100 1 2( ),
where qi represents the output levels of the two countries with Iran being 1 and Iraq being 2, and P is the per-barrel price. The marginal revenue schedules facing the two countries are:
MR q q MR q q
1 1 2
2 2 1
100 2
100 2
= − − = − −
and
.
Each country has a marginal cost curve of the form:
MC qi i= , where i = 1, 2. a. Determine each country’s reaction function. b. Does a Cournot equilibrium exist? If so, find the outputs
and prices of crude oil in the two countries.
c. Suppose that the two countries collude and become a cartel. What will be the resulting price and outputs for crude oil for
the two countries? [Note that the market marginal revenue
is 100 − 2(q1 + q2).] d. Can it be said that because collusive profits are strictly
greater, it is true that these countries should necessarily
collude? Are there any potential pitfalls in such a collusive
arrangement?
13.18 Suppose that Iraq is the Stackelberg leader in the preced- ing problem. What will be each country’s reaction function?
How much will each country produce, and what will its prof-
its be?
13.19 Repeat the preceding problem but assume that Iran is the Stackelberg leader.
13.20 It is the not-too-distant future, and all important human needs have been eliminated. Thanks to the firm Bioeconotek,
a miracle drug, Needless, has been invented that genetically
suppresses a patient’s desire to think in terms of having needs.
The drug stimulates the patient’s cognitive ability to perceive
substitute products as being ever more attractive the greater the
price of any item.
Needless does not come cheap. Given its curative powers,
Bioeconotek charges $1,633 per dose and sells 62,020,000
356 Monopol ist ic Competit ion and Ol igopoly
C13.INDD 10:49:55:AM 08/06/2014 PAGE 356Trim Size: 203.2 mm X 254 mm
doses annually despite the presence of a competitive fringe
of generic suppliers, which collectively sell 51,650,000 doses
annually at the same $1,633 price.
Bioeconotek managers estimate that the fringe supply curve
is described by the equation:
P Q= +600 50 000
FS
, ,
and that the market demand is given by:
P Q= −13 000 10 000
, ,
.M
In the preceding equations P is the price per dose, QFS is the quantity supplied by the competitive fringe, and QM is the quantity consumed in the total market.
The marginal cost associated with producing and distribut-
ing Needless is a constant $600 per dose. Bioeconotek initially
invested $400 million in developing the drug but has faced no
additional fixed costs since the initial investment.
a. At the current price of $1,633, a Bioeconotek manager esti- mates the demand elasticity for Needless as:
Elasticity = = − ⎛ ⎝⎜
⎞ ⎠⎟
=
Δ Δ
Q P
P Q
( , ) ,
, ,
. .
10 000 1 633
113 670 000
0 14
Since this value represents an inelastic demand, he argues that Needless’s price should be raised. Is there something
wrong with this reasoning? Briefly explain.
b. Write an equation that gives the demand for Bioeconotek’s own output—that is, demand as seen by Bioeconotek given
market demand and the fringe supply.
c. What is the profit-maximizing price for Bioeconotek to charge, given the current market demand and the current
fringe supply? (Recall that if a demand curve can be written
as P = a − bQ, then MR = a − 2bQ.) d. If Bioeconotek wants to implement a “limit-pricing” strat-
egy that successfully eliminates fringe suppliers from the
market, what annual output level does it need to select for
its own product, Needless?
e. Another Bioeconotek manager worries about future growth of the fringe supply and argues that the company should
set the price determined in part d so as to drive the current
fringe out of business and eliminate any incentives for fur-
ther entry. As briefly as possible, explain why this strategy
cannot be more profitable than the pricing strategy you gave
in part c.
13.21 Suppose that the market for Web search engines can best be characterized as monopolistically competitive. If this is the
case, should firms that operate in the market such as Yahoo and
Google be prosecuted by antitrust authorities on account of the
potential inefficiency of the market outcome? Explain why or
why not.
13.22 College textbooks have been increasing in price at a rate greater than the general price level over the past several
decades. The reason for this phenomenon appears to be that
publishers have realized that by having authors revise their
texts more frequently, competition from used texts is reduced.
Using the dominant firm model, explain how such a more-
frequent-revision strategy leads to a higher price for a given
textbook.
13.23 Historically, officials from 23 elite northeastern colleges with selective admissions policies and high tuition met each
spring to compare financial aid packages for more than 10,000
common applicants. The meetings, known as “Overlap,” were
designed to eliminate any differences in the financial aid pack-
ages offered by the various colleges. What would you predict
would be the effect on the net price (tuition less financial aid)
paid by applicants to the colleges participating in the Overlap
meetings? Are there any factors that would work to undermine
the ongoing viability of the Overlap practice? If so, what might
these factors be?
13.24 In 2002, drug manufacturer Schering-Plough’s monop- oly (through patent protection) on the huge-selling allergy
medication Claritin was set to expire. To protect its $3 billion
in annual sales from Claritin, Schering-Plough sued 10 pro-
spective manufacturers of generic substitutes for Claritin. The
suits alleged that anyone who swallowed a generic version
would produce in their livers a separate chemical compound
or metabolite on which Schering-Plough had patent protection
through April 2004. Using the dominant firm model, explain
what Schering-Plough’s actions were intended to accomplish
in the market for allergy medications.
13.25 Assume that Intel can be treated as the dominant firm in the market for computer chips. What three basic factors deter-
mine the elasticity of demand confronted by Intel? Explain
whether increases or decreases in these three factors will
increase Intel’s elasticity of demand, and provide an intuitive
explanation why.
357
C14.INDD 10:46:44:AM 08/06/2014 PAGE 357Trim Size: 203.2 mm X 254 mm
CHAPTER 14 Game Theory and the Economics of Information
This chapter covers two topics that have been intensively studied by economists in recent years: game theory and the economics of information. Game theory is a mathematical tech-
nique developed to study choice under conditions of strategic interaction. It has become
the major approach to the study of oligopoly, and our discussion looks at how it has been
adapted to give insight into problems of oligopolistic interdependence. In particular, we
will see that it helps us understand the difficulties firms face in colluding to raise prices.
In most economic analysis, whether of competition or oligopoly, it is assumed that con-
sumers have complete knowledge of prices and product characteristics. It is clear that this
assumption is often violated in the real world, and recent research has begun to analyze
what effect this has on the way markets function. When information is costly, consumers
are not fully informed and lack either knowledge of the prices different firms charge or
knowledge of the qualities of the products they sell, or both. Under these circumstances,
the prices, quantities, and qualities of goods traded can be quite different than when con-
sumers have “perfect information.” Two especially interesting applications of the effects
of costly information involve insurance markets and advertising. We discuss them later in
the chapter.
Learning Objectives
Understand the basics of game theory: a mathematical technique to study choice under conditions of strategic interaction. Describe the prisoner’s dilemma and its applicability to oligopoly theory as well as many other situations. Explore how the outcome in the case of a prisoner’s dilemma differs in a repeated-game versus a single-period setting. Analyze asymmetric information and market outcomes in the case where consumers have less information than sellers. Explain how insurance markets may function when information is imperfect and there is the possibility of either adverse selection or moral hazard. Show how limited price information affects price dispersion for a product. Investigate advertising and the extent to which it serves to artificially differentiate products versus provide information to consumers about the availability of products and their prices and qualities.
Memorable Quote “Tis better to have loved and lost than never to have loved at all.”
—Alfred Lord Tennyson
358 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 358Trim Size: 203.2 mm X 254 mm
14.1 Game Theory Game theory is a method of analyzing situations in which the outcomes of your choices depend on others’ choices, and vice versa. In these strategic situations, there is mutual inter-
dependence among the choices made by the decision makers: each decisionmaker needs to
account for how the others are affected by the choices made and how the other decision mak-
ers are likely to respond since their responses may affect what is the best choice to make.
Among market structures, oligopoly is the setting in which such interactions are likely to be
most important, and game theory has become widely used to analyze oligopolies.
Game theory is a general approach to analyzing strategic interactions, however, and it
can be applied to issues other than oligopoly. For example, in determining the defense bud-
get, the U.S. government must not only consider the impact of its budgetary decisions on
the military decisions of potential enemies, but also recognize that these potential enemies
are trying to predict how the United States will respond to their budgetary decisions. Simi-
larly, a politician thinking about conducting negative campaign tactics must recognize that
his or her opponent might respond in a variety of possible ways. Whether it is wise to pur-
sue a negative campaign strategy depends on how the opponent is likely to respond. Game
theory can be applied to a wide range of phenomena like these, but our main concern is
with how it helps us understand the functioning of oligopolistic markets.
All game theory models have at least three elements in common: players, strategies, and
payoffs. The players are the decisionmakers whose behavior we are trying to predict. In the case of oligopoly, the players are the firms. The strategies are the possible choices of the players. Outputs produced and prices charged are strategies in this sense, but so too are
advertising budgets, new product introductions, and product differentiation. For oligopolis-
tic firms, all of these actions can affect rivals. The payoffs are the outcomes, or conse- quences, of the strategies chosen. For firms, it is natural to express the outcomes as the
profits or losses realized. It is important to remember, however, that a specific strategy (say,
producing 300,000 microwave ovens) does not uniquely determine the profit (payoff)
because that will depend on the strategies followed by the other players.
Determination of Equilibrium In addition to the players, strategies, and payoffs, other elements may play a role in deter-
mining the outcome in a game theory model, but for the moment, we’ll use some simple
examples to illustrate how the model links the various elements. Consider a market with
just two firms, A and B. Each firm must choose either a low or a high output; these are the
only possible “strategies.” However, each firm’s profit depends not only on its own output
but also on the output of the other firm.
To go further, we have to be precise about exactly how profits are affected by both firms’
choices. To show the possibilities with some illustrative numbers, we will use a payoff
matrix. A payoff matrix is a simple way of representing how each combination of choices affects firms’ profits. Table 14.1 is the payoff matrix for our example. There are four cells
in this case. The upper-left cell shows that profit is 10 for firm A and 20 for firm B when
both firms choose the low-output strategy. The upper-right cell shows that profit is 9 for
firm A and 30 for firm B when A chooses a low output and B chooses a high output. (Thus,
when B increases its output from low to high and A holds its output constant at low, B’s
profit increases from 20 to 30, but A’s falls from 10 to 9—perhaps because B’s higher out-
put reduces the demand confronting firm A.) Similarly, the lower-left cell shows what hap-
pens to profits when A’s output is high and B’s output is low. Finally, the lower-right cell
shows the outcome when both firms choose a high output.
In interpreting this payoff matrix, we assume that firms act independently. Note that this
means that firm A has only two possible strategies: it can choose one of the two horizontal
rows in the payoff matrix. If it chooses the top row (low output), its profit will be either 10 or 9,
game theory a method of analyzing situations in which the outcomes of your choices depend on others’ choices, and vice versa
players in game theory, decision- makers whose behavior we are trying to predict and/or explain
strategies in game theory, the possible choices of the players
payoffs in game theory, the outcomes or consequences of the strategies chosen
payoff matrix a simple way of representing how each combination of choices affects players’ payoffs in a game theory setting
Game Theory 359
C14.INDD 10:46:44:AM 08/06/2014 PAGE 359Trim Size: 203.2 mm X 254 mm
depending on what B chooses. Firm A cannot choose which cell it will end up in because it
doesn’t control B’s choice. In the same way, firm B can choose only between the two vertical
columns. If it chooses the right column, for example, its profit will be either 30 or 25, but it
doesn’t know which unless it can predict A’s output choice. (We assume that the firms select
their outputs simultaneously so neither knows with certainty what the other’s choice will be.)
The purpose of this exercise is, of course, to predict what the market equilibrium will be.
So let’s see if we can figure out what each firm will do. Consider firm A. If firm B chooses
a low output, firm A is better off selecting a high output because its profit would then be 20
rather than 10. On the other hand, if B chooses a high output, A is also better off selecting a
high output because its profit would then be 18 rather than 9. Consider carefully what this
means: A is better off with a high output regardless of what B does. In this case A has a dominant strategy because the high-output choice is best no matter what B does. In this situation it is easy to predict that firm A will produce a high output.
Now apply the same reasoning to firm B. If firm A chooses a low output, B is better off
selecting a high output because its profit would then be 30 instead of 20. Alternatively, if
A chooses a high output, B is still better off selecting a high output because its profit would
then be 25 rather than 17. Thus, B also has a dominant strategy: its high-output strategy is
best regardless of what A does. We would predict that firm B will produce a high output.
dominant strategy a case where a player is better off adopting a particular strategy, regardless of the strategy adopted by the other player
APPLICATION 14.1
In baseball, it is advantageous for any forced base run- ner (i.e., runner who will be forced to advance to the next base if the batter gets on base) to run on the pitch when there are two outs and the count on the batter stands at three balls and two strikes.1 This is so because, if the batter doesn’t swing at the pitch, either the pitch is a third strike and the inning ends or it is a fourth ball, and the runner and
Dominant Strategies in Baseball
batter advance. If the batter swings and fouls off the pitch, the runner simply returns to the initial base (if the foul isn’t caught) or the inning ends (if the foul is caught). And if the batter swings and hits a fair ball, a runner who left base on the pitch has a better chance of advancing or scoring in the case when the defensive team doesn’t make an out on the play and is no worse off if an out is made. Thus, running on the pitch is a dominant strategy for a forced runner when there are two outs and the count is three balls and two strikes on the batter.
1This application is based on Avinash K. Dixit and Barry J. Nalebuff, Thinking Strategically (New York: W. W. Norton, 1991).
Dominant-Strategy Equilibrium: A Simple Oligopoly Game
Low Output
Low Output
Firm A
High Output
High Output
10
20
17 25
30
20 18
9
Firm B
Table 14.1
360 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 360Trim Size: 203.2 mm X 254 mm
In this example both firms have dominant strategies, and so we expect the equilibrium
outcome to include high output from both firms. This is called a dominant-strategy equi- librium and is perhaps the simplest game that can be imagined. As you might expect, how- ever, not all situations are so easily analyzed.
Nash Equilibrium A slightly more involved situation is shown in Table 14.2. Here, we have changed only one
number from the previous table. When both firms choose a low output (upper-left cell), A’s
profit is now 22 (rather than 10 as before). Now reconsider A’s output choice. If B chooses
a low output, A’s best choice is a low output (profit of 22 rather than 20), but if B chooses
a high output, A’s best choice is a high output (profit of 18 rather than 9). Firm A now does
not have a dominant strategy: its best choice depends on what firm B does. What output should firm A choose?
If firm A can predict what B will do, it will then know which choice is best. So A has
to try to figure out what B’s choice will be before making its own decision. This is charac-
teristic of most game-theory applications: players have to evaluate and predict what their
rivals will do since their own decisions depend on the rivals’ choices. Note that this was not
necessary in our first example, when both firms had dominant strategies. Each firm’s best
choice was a high output regardless of the actions of the other. So let’s consider how firm
A might try to figure out what B will do. In this case it is fairly simple. Firm B still has a
dominant strategy. For B, a high output is best, regardless of what A does. Thus, A might
reasonably predict that B will choose a high output. And if B selects a high output, A’s
profit is higher when it also selects a high output. Consequently, our analysis suggests that
two rational, profit-maximizing firms would both, in the Table 14.2 setting, select a high
output, and that that would be the game’s equilibrium.
The equilibrium we have identified in Table 14.2 is not a dominant-strategy equilibrium,
however, because that term refers to a case where both firms have dominant strategies, and
here only one does. To analyze the outcomes in many games, we need a more general concept
of equilibrium than a dominant-strategy equilibrium. The concept most widely used is called
a Nash equilibrium, after mathematician John Nash, who formalized the notion and won the Nobel Prize in economics for his contribution. (Nash’s genius and lengthy struggle with men-
tal illness were the subjects of the award-winning book and movie A Beautiful Mind.) A Nash equilibrium is a set of strategies such that each player’s choice is the best one
possible given the strategy chosen by the other player(s). To see that the high output–high
dominant-strategy equilibrium the simplest game theory outcome, resulting from both players having dominant strategies
Nash equilibrium a set of strategies such that each player’s choice is the best one possible, given the strategy chosen by the other player(s)
Another Oligopoly Game
Low Output
Low Output
Firm A
High Output
High Output
22
20
17 25
30
20 18
9
Firm B
Table 14.2
The Pr isoner ’s Di lemma Game 361
C14.INDD 10:46:44:AM 08/06/2014 PAGE 361Trim Size: 203.2 mm X 254 mm
output set of strategies in Table 14.2 is a Nash equilibrium, we evaluate it as follows. What
we want to see is whether firm A’s best choice is high output when firm B chooses high out- put, and simultaneously whether firm B’s best choice is high output when firm A chooses high output. If firm B chooses high output, then firm A is best off by choosing high out- put (profit of 18 rather than 9), so the lower-right cell passes the first part of the test. If A
chooses high output, B is best off by choosing high output (profit of 25 rather than 17), so
the lower-right cell passes the second part of the test also. The two strategies shown in the
lower-right cell, then, represent a Nash equilibrium.
Applying this same reasoning to the equilibrium we identified in Table 14.1, you will
see that the high-output combination of strategies there is also a Nash equilibrium. A domi-
nant-strategy equilibrium (as in Table 14.1) is a special case of a Nash equilibrium. That is,
all dominant-strategy equilibria are also Nash equilibria, though not all Nash equilibria are
dominant-strategy equilibria (as the Table 14.2 example illustrates). Unfortunately, not all
games have a Nash equilibrium. Thus, while the concept works in many cases, it does not
let us determine the outcome in all strategic situations.
The Nash equilibrium is closely related to the analysis of the Cournot oligopoly model
discussed in Chapter 13. Recall that the Cournot equilibrium is one in which each firm is
producing its best output given the outputs of other firms. This is exactly the description of a Nash equilibrium, and sometimes the equilibrium in such a market situation is called a
Cournot–Nash equilibrium. More importantly, the application of game theory to the stra-
tegic interaction in that sort of model has provided a reason why firms might rationally
choose to behave in the way Cournot described.
14.2 The Prisoner’s Dilemma Game The most famous game theory model is the prisoner’s dilemma. We will first describe the game in the form that made it famous, thereby illustrating how it got its name. Although it
may not be immediately apparent, the prisoner’s dilemma is widely applicable—to oligop-
oly theory as well as many other situations.
Two individuals, Nancy and Sid, are picked up on a public nuisance charge. While ques-
tioning the suspects, the district attorney begins to suspect that they may be two key players
in an international drug ring. The district attorney, however, does not have enough evi-
dence to convict them of the more serious charge, so she comes up with the following ploy
in an attempt to extract a confession. She separates the two prisoners so they cannot com-
municate with one another and tells each the following:
1. If both confess to drug trafficking, they will both go to jail for 10 years. 2. If neither confesses, they will be charged and convicted of the nuisance offense, and each will receive a 2-year sentence.
3. If one confesses (turns state’s evidence) and the other does not, the one who confesses will get a reduced sentence of 1 year; the one who doesn’t will be convicted and go to jail
for 15 years.
Table 14.3 shows the relevant payoff matrix for this game. Now we can apply our knowl-
edge of game theory to determine what each suspect will do. Let’s look at the situation con-
fronting Sid first. If Nancy confesses, Sid’s best strategy is to confess also because he then gets
10 years instead of 15 years. If Nancy doesn’t confess, Sid’s best strategy is still to confess
because he then gets only 1 year instead of the 2 years he gets if he doesn’t confess. To confess is a dominant strategy for Sid: Regardless of what Nancy does, Sid does better by confessing.
The situation is exactly symmetrical for Nancy. If Sid confesses, Nancy’s best strategy
is to confess (10 versus 15 years), and if Sid doesn’t confess, Nancy still does better to con-
fess (1 versus 2 years). Confession is a dominant strategy for Nancy also.
prisoner’s dilemma the most famous game theory model, in which self-interest on the part of each player leads to a result in which all players are worse off than they could be if different choices were made
362 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 362Trim Size: 203.2 mm X 254 mm
Thus, the dominant-strategy equilibrium (and a Nash equilibrium) is for both parties to
confess, and that is the expected outcome. This may not seem surprising until you realize
fully what it means. The predicted outcome is one where both suspects are worse off than they would be if neither confessed (in which case they would get only 2 years each, the
lower-right cell). Since they both know this, why do they confess? Because it is in each one’s
self-interest to confess, even though the collective outcome of each pursuing their self-interest is inferior for both. By the way, do not mistakenly think that the reason each confesses is that
they believe the other will also. The reason for the predicted outcome is stronger than that: it
is in each suspect’s interest to confess, regardless of the other’s actions.
APPLICATION 14.2
Economist Russell Roberts of George Mason University likens the difficulty a representative democracy encounters in restraining spending to the case in which one’s bill at a res- taurant is spread evenly across 100 other diners. If you were responsible for your entire bill, you would usually spend only $6 on a meal and wouldn’t splurge on a second drink and des- sert that would add $4 to the tab. On the other hand, add- ing the $4 drink and dessert costs only 4 cents when the bill is spread out evenly over all 100 patrons of the restaurant:
Splurging is easy to justify now. In fact, you won’t just add a drink and dessert, you’ll upgrade to the steak and add a bottle of wine. Suppose you and everybody else orders $40 worth of food. The tab for the entire restaurant will be $4,000. Divided by the 100 diners, your bill will be $40. While you’ll get your “fair share” this outcome is a disaster. When you dined alone, you spent $6. The extra $34 of steak and other treats were not worth it. But in competition with the others,
The Congressional Prisoner’s Dilemma
you chose a meal far out of your price range whose enjoyment fell far short of its cost—self-restraint goes unrewarded. If you go back to ordering your $6 meal in hopes of saving money, your tab will be close to $40 anyway, unless the other 99 diners cut back also. The good citizen starts to feel like a chump. And so we read of the freshman Congressman eager to cut pork out of the budget but in trouble back home because local projects will also come under the knife. Instead of being proud to lead the way, he is forced to fight for the projects, to make sure his district gets its “fair share.”2
In Roberts’ judgment, therefore, the average represen- tative confronts a prisoner’s dilemma wherein promoting spending on projects in one’s district is a dominant strategy (akin to “confessing” in Table 14.3). The resulting equilib- rium—an overall high level of government spending (and the taxes that must support such spending) across districts—is inferior to the outcome that would emerge if representatives exercised more restraint in their pursuit of spending on gov- ernment projects in their individual districts.
2Russell Roberts, “If You’re Paying, I’ll Have Top Sirloin,” Wall Street Journal, May 18, 1995, p. A18.
The Prisoner’s Dilemma
Confess
Confess
Sid
Don’t Confess
Don’t Confess
10 years in prison
10 years in prison
1 year in prison
2 years in prison
15 years in prison
15 years in prison
2 years in prison
1 year in prison
Nancy
Table 14.3
The Pr isoner ’s Di lemma Game 363
C14.INDD 10:46:44:AM 08/06/2014 PAGE 363Trim Size: 203.2 mm X 254 mm
When they first encounter the prisoner’s dilemma, many people try to figure out some
way in which the prisoners could realize the best all-around outcome, the 2-year sentence
drawn if neither one confesses. By adding some additional elements to the scenario, it is
indeed possible to spin a game-theoretic tale where both refuse to confess. For example,
if Nancy and Sid are lovers, such that each feels as much pain if the other goes to jail as if
they go to jail themselves, they would not confess. Another possibility might be that each
suspect believes that if he or she is the only one to confess, the suspect who subsequently
does 15 years would be willing to commit murder on release in revenge. In that case, there
would probably be no confessions. (Note that the payoffs would be different than just the
jail sentences shown in the table.) Taken on its own terms, however, the prisoner’s dilemma
does show how the individual pursuit of self- interest can, in certain situations, produce
results that are inferior for all players.
The prisoner’s dilemma has wide-ranging applicability. It was used to model the interac-
tions between the United States and the Soviet Union in the days of the Cold War. It has
also been employed to explain political ticket-splitting (when voters cast ballots for candi-
dates from different parties in various races in the same election) and why the U.S. health
care system is so often characterized as being both overly expensive and bureaucratic. In
this chapter we explore how the prisoner’s dilemma can be applied to explain cheating by
members of a cartel, the effects of a “curve”-based grading system on student study effort,
and certain aspects of World War I trench warfare.
The Prisoner’s Dilemma and Cheating by Cartel Members Firms in oligopolistic industries often find themselves in a prisoner’s dilemma where each
firm acting in its self-interest produces an outcome in which all firms are worse off. Con-
sider a very simple setting, with just two firms in an industry. We will use the example
developed in Chapter 13 to illustrate the Cournot model. Utopia and Artesia are the firms,
and they sell a homogeneous product. We will consider the possibility that the two firms
will coordinate their activities to increase their profits. In other words, the firms try to form
a cartel. The cartel agreement would have each firm restrict its output so that the market
price will be high. However, each firm might consider raising its output beyond the cartel
quota (cheating on the agreement). Thus, for each firm the two possible strategies are to
comply and to cheat. How will the firms decide which strategy to choose?
The firms’ various alternatives are shown in Table 14.4. The numbers in the cells of the
payoff matrix represent the firms’ profits in each of the possible outcomes. Importantly, the
numbers are not just arbitrarily chosen, but reflect the actual economic environment being
Cheating in a Cartel
Cheat
Cheat
Utopia
Comply
Comply
10
10
25 20
5
5 20
25
Artesia
Table 14.4
364 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 364Trim Size: 203.2 mm X 254 mm
investigated. (Or at least the relationship between the numbers, their rank ordering, reflects
the problem being studied.) Let’s see how we can apply economics to explain each firm’s
profit for each outcome.
Let’s assume that Artesia and Utopia have the same cost curves, so that when they col-
lude, they agree to produce the same output. Together they will produce the monopoly out-
put for the industry because that maximizes their combined profit. With both producing the
same amount, each receives half the monopoly profit. This situation is shown in the lower-
right cell, where both firms restrict their output, thereby complying with the cartel agree-
ment, and each realizes a profit of 20. Now consider the upper-left cell, where both firms
cheat. Because cheating means producing a higher output, price will be lower and, more
significantly, profits will be lower. Because the firms are identical, let’s suppose that both
firms have the same “cheating output” (which we might assume to be the Cournot output),
and so will realize the same profit, which we take to be 10.
Next consider the lower-left cell, where Artesia cheats and Utopia complies. The easiest
way to see how this affects the profit of each is to imagine starting with the perfect cartel
outcome at the lower right. Then Artesia cheats. This increases Artesia’s profit from 20
to 25 for the reason we discussed in Chapter 13: the demand curve confronting a firm is
more elastic when it changes output than when all cartel members simultaneously change
output. This action, however, reduces Utopia’s profit because its output is unchanged, but
the price of the product is reduced by Artesia’s output expansion. Moreover, the reduction
in profit for Utopia is greater than the increase for Artesia because we know their combined
profit has to be less than 40, the maximum (monopoly) profit shown in the lower-right cell.
Here, we assume profit is 5 for Utopia when it complies and Artesia cheats, although the
important point is that Utopia’s profit falls by more than Artesia’s profit increases. Finally,
by analogous reasoning, we obtain the figures in the upper-right cell, which is the mirror
image of the lower-left cell.
Now that we understand why the payoffs are as shown, we can analyze the behavior of the
two firms. For Artesia, cheating is better if Utopia either complies (profit of 25 versus 20) or
cheats (profit of 10 versus 5). Thus, cheating is a dominant strategy for Artesia. For Utopia,
cheating is better if Artesia either complies (profit of 25 versus 20) or cheats (profit of 10
versus 5), so cheating is a dominant strategy for Utopia also. Thus, we expect both firms to
cheat, and they end up with a profit of 10 each, even though if they both complied they would
realize a profit of 20 each. As you might suspect from the outcome, this situation is a version
of the prisoner’s dilemma game, so the outcome is not surprising. The model shows, in very
clear fashion, why firms have an incentive to cheat on a cartel agreement.3
It would be a mistake, however, to conclude from this analysis that firms will never suc-
cessfully form a cartel. Other factors not incorporated here can increase the likelihood of
collusion. For example, suppose that cartel agreements are legal and will be enforced. If
Artesia and Utopia enter into a contract stipulating that both will comply, each firm will
have an incentive to sign since the cartel outcome is better than the outcome realized when
both cheat. What this model makes clear, however, is that the agreement must be enforce-
able because each firm individually still has an incentive to cheat on the agreement. Of
course, in the United States cartels are illegal, so if firms illegally enter into a collusive
agreement they must have some way of enforcing the agreement on their own to prevent
cheating. In other words, for a cartel to be stable, there must be some way to punish firms
that cheat. We will consider how this goal might be accomplished in the next section.
3Note that this example appears very similar to Table 14.1: in both cases there are two firms, each choosing between a “low” and a “high” output. Table 14.1 is not, however, a prisoner’s dilemma, while Table 14.4 is. What accounts for the difference is that the interdependence of outputs is not as pronounced between the firms in Table 14.1. For example, it may be that they are producing goods that are not very good substitutes for one another, while in Table 14.4 the goods are perfect substitutes. Can you see why that would account for the difference in the payoffs shown in the two tables?
The Pr isoner ’s Di lemma Game 365
C14.INDD 10:46:44:AM 08/06/2014 PAGE 365Trim Size: 203.2 mm X 254 mm
This game theory model of cartel behavior can be extended to explain why cheating is
more likely the higher the number of firms in the cartel. Suppose there are 10 firms initially
complying, with each firm making a profit of 20. Then one firm cheats. In our two-firm
model, this action increases the cheater’s profit from 20 to 25, but with 10 firms involved,
the increase in profit will be greater. This is simply because the cheater will confront a more
elastic demand curve when it is one among 10 firms than when it is one of two firms. (If
one firm among 10 increases its output by, say, 20 percent, price will fall by less than after
an increase of 20 percent by one of two firms.) In addition, the damage the cheater does to
the complying firms will be shared among the remaining nine, and so less noticeable for
each individual, noncheating firm. For this reason, the more firms in the cartel, the more
common we expect cheating to be. With more firms, cheating produces a greater increase in
profit for the cheater while the loss for each complying firm is smaller and harder to detect.
This game theory illustration supports our earlier explanation as to why collusion is likely
to be less common among a larger number of competitors. Each member has a greater
incentive to cheat and so undermine the cartel the greater the number of firms involved.
A Prisoner’s Dilemma Game You May Play To see an application of the prisoner’s dilemma that may be relevant in your academic life,
consider students competing for grades in a class. The professor grades “on the curve” and
assigns a certain distribution of grades in the class regardless of what the absolute grades
are. For example, the professor may assign 20 percent of the class a grade of “A,” regard-
less of whether 20 percent or only 5 percent score above 90. The professor may have found
that in the past about 20 percent of her students do “A” quality work. If on a particular test
only 5 percent get grades above 90, she may reason that the test was unfairly difficult or
that she graded more harshly than usual. Thus, she sets the “A” range to include the top 20
percent, even though that means giving an “A” to students with absolute scores below 90.
If students’ motivation for studying is to get a good grade, they will find themselves in a
prisoner’s dilemma in this class.
Consider two students, Kaitlyn and Scott—taken as representative of the entire class. In
this class, grades are curved so that the average is a “B.” If they each study four hours per
week, they will both get an absolute score of 85 and a grade of “B,” as shown in the lower-
right cell of the Table 14.5 payoff matrix. However, if they each study only one hour per
week, they both get an absolute score of 60, but still receive a grade of “B” because 60 is
now the average score for the class and the average receives a “B.” This outcome is shown
in the upper-left cell. Although their letter grades are the same in both these cells, the stu-
dents are better off in the upper-left cell: they have the same grades at a lower cost in study
time. However, if Kaitlyn studies one hour and Scott studies four hours, he will receive an
“A” and she will receive a “C”—the lower-left cell. (Kaitlyn gets the same absolute score
for the course but ranks lower in the class now.) We will assume that a student prefers to
earn an “A” with four hours of studying over a “B” with one hour. The same reasoning
establishes the grades in the upper-right cell.
The way the students rank the Table 14.5 payoffs gives rise to the prisoner’s dilemma.
Each student acting independently has a dominant strategy: studying four hours per week.
Thus, the outcome is the lower-right cell, a dominant-strategy equilibrium. However, the
two students would be better off if they both studied one hour per week and attained the
same “B” at a lower cost in time. Students might try to collude and attain the upper-left
cell, but they would find enforcing the collusive agreement difficult because each student
has an incentive to cheat (that is, studying four hours and “spoiling the curve”), and it is
difficult to detect cheating and enforce sanctions. As a result, the students are caught in a
prisoner’s dilemma.
Does this model explain why students study as much as they do? Actually, it is clear that
other factors play a role in studying decisions. For example, if a professor has strict absolute
366 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 366Trim Size: 203.2 mm X 254 mm
standards (even if you have the highest score in the class, you don’t get an “A” unless it is
above 90), there will be no prisoner’s dilemma. A more serious objection to the analysis con-
cerns the assumption that students care only about the grades they receive. If they care more
about learning than about the grades, there is no prisoner’s dilemma. Because this is undoubt-
edly true in your economics classes, the model may not be applicable there. However, it may
give you some insight into student behavior in some of your other classes.
14.3 Repeated Games Oligopolists often find themselves in a prisoner’s dilemma if they attempt to collude so as
to increase their profits. Because the prisoner’s dilemma game has a dominant strategy
equilibrium where all firms cheat (that is, don’t collude), it appears that successful collu-
sion never occurs unless binding contracts are permitted and enforced by an external
authority. Such a conclusion, however, is overreaching. As explained so far, the implicit
assumption is that the prisoner’s dilemma game is played only once: the decision to cheat
or comply is made just once. This is an appropriate assumption in some settings (as in the
case of Sid and Nancy), but it is often inappropriate when applied to firms’ pricing and out-
put decisions. Firms in an oligopoly play against one another repeatedly as they make deci-
sions week after week. In the jargon of game theorists, the appropriate model for these
market conditions is a repeated-game model rather than a single-period model. Let us now reconsider the prospects for effective collusion in a duopoly. We will use the
data in Table 14.4 explained in the last section. Now, however, we want to imagine that the
payoffs refer to the profits for Artesia and Utopia each week and that each week the firms
make a choice of either complying (low output) or cheating (high output). In the repeated-
game setting, the range of strategies available to the firm increases enormously. It would be
impossible to consider all the permutations, but one point should be clear: each firm now
has a way to punish its rival for any past transgressions. Because the game is repeated, if
Artesia cheats in the fourth week, then it is possible for Utopia to cheat in the fifth week, a
tactic that imposes a cost on Artesia. In other words, the firms have a way of enforcing the
cartel agreement by punishing one another for cheating.
What strategy will firms adopt in this setting, and what is the equilibrium likely to be?
These are difficult questions without definitive answers, but we can easily see that collusion
is more likely in the repeated-game setting. Suppose that Utopia adopts a tit-for-tat
repeated-game model a game theory model in which the “game” is played more than once
Beating the System
1 Hour
1 Hour
Scott
4 Hours
4 Hours
B (60)
B (60)
C (60) B (85)
A (85)
A (85) B (85)
C (60)
Kaitlyn
Table 14.5
Repeated Games 367
C14.INDD 10:46:44:AM 08/06/2014 PAGE 367Trim Size: 203.2 mm X 254 mm
strategy. Under a tit-for-tat strategy, Utopia will comply in a given week as long as Artesia complied in the previous week. However, if Artesia cheats in one week, the following week
Utopia will cheat (tit-for-tat), and will continue cheating in each successive week until
Artesia complies, after which Utopia will revert to complying. There is nothing sacrosanct
about the tit-for-tat strategy, but it is simple and in a computer simulation turns out to be
quite effective.4 We can see why by considering the consequences of adopting this strategy.
Let’s examine some of the options open to Artesia when Utopia plays the tit-for-tat
strategy and Artesia is aware of it. Table 14.6 shows two possible scenarios that might
unfold over a succession of weeks. In the first scenario, both firms comply in the first week
and realize a profit of 20 (see Table 14.4 and the discussion there for the source of the profit
numbers). In the second week, Utopia complies but Artesia cheats and obtains a higher
profit of 25. In this scenario, we investigate the effects if Artesia continues cheating in the
following weeks. In the third week, Utopia plays tit-for-tat and cheats. Artesia continues to
cheat so both firms realize a profit of 10. As long as Artesia continues to cheat, so will Uto-
pia, and they will continue to obtain a profit of 10 each.
Now imagine you are Artesia, and consider whether you would be better off following
the strategy of Scenario 1 (comply in the first week and cheat thereafter) or complying every
week. In Scenario 1, you get profits of 20, 25, 10, 10, and so on. By complying, however,
you get profits of 20, 20, 20, 20, and so on, because Utopia will comply as long as you do.
Clearly, you are better off by complying because the short-term gain you get by cheating in
the second week is more than offset by the lower profits you suffer every week thereafter.
As an alternative, however, you might consider cheating one week and then complying
the next. Scenario 2 shows the evolution of the game when Artesia cheats in the second
week and then complies in the third week, inducing Utopia to revert to complying in the
fourth week. Once again, the sum of profits over time is higher for Artesia if it complies in
every period than if it follows this strategy.5
This example shows why it is rational for Artesia to comply in every period when it
knows that Utopia is playing a tit-for-tat strategy. This does not prove that the collusive
outcome results because we have just assumed that Utopia adopted this particular strategy
and that Artesia is aware of it. It does, however, suggest why in a repeated-game setting the
collusive outcome is more likely: firms realize that if they cheat in the current period they
are likely to suffer losses in subsequent periods, and that realization diminishes the incen-
tive to cheat.
tit-for-tat a strategy in which each player mimics the action (e.g., cheat, comply) taken by the other player in the preceding period
Table 14.6 Cheating in a Repeated Cartel Game Scenario 1 Scenario 2
Period Artesia Utopia Artesia Utopia Week 1 Comply (20) Comply (20) Comply (20) Comply (20)
Week 2 Cheat (25) Comply (5) Cheat (25) Comply (5)
Week 3 Cheat (10) Cheat (10) Comply (5) Cheat (25)
Week 4 Cheat (10) Cheat (10) Comply (20) Comply (20)
· · · · ·
· · · · ·
· · · · ·
4Of course, the effectiveness of any strategy depends on what strategy opposing players are using. Robert Axelrod, in The Evolution of Cooperation (New York: Basic Books, 1984), finds that, on average, tit-for-tat works better than any of the other strategies investigated. 5This conclusion depends on the sum of the profits in weeks 2 and 3 (cheat-comply and comply-cheat), here 30, being less than the sum of the profits when Artesia complies in both periods, here 40. Refer back to the discussion of the payoffs in Table 14.4 to see why the relationship between the payoffs must be this way.
368 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 368Trim Size: 203.2 mm X 254 mm
Do Oligopolistic Firms Always Collude? Our analysis of the repeated prisoner’s dilemma game seems to turn our earlier conclusion
on its head. The initial analysis suggested that firms would inevitably cheat, and now we
have an analysis suggesting that over time collusion is likely. But the real world is, alas,
even more complicated than the repeated-game model may suggest. Consider, for example,
the restrictive assumptions (not all explicit) underlying our repeated prisoner’s dilemma
game: there are only two firms, no entry into the market occurs, the firms have identical
costs and produce the same product, each firm has complete knowledge of both firms’ pay-
offs for all strategy combinations, demand and cost conditions do not vary over time, and
the game is repeated indefinitely. Relaxing almost any of these assumptions makes it less
likely that collusion will be a stable outcome in an oligopolistic industry.
To see how changing one assumption can affect the outcome, consider the number of
time periods the game is played. Our previous analysis implicitly assumed that the game
was repeated forever. To contrast that, suppose that the two firms know that if they collude
for 10 weeks, new firms will have time to enter the market, and with new entry in the elev-
enth week, collusion will become ineffective. Artesia and Utopia then know that the game
will last for 10 weeks; there are ten time periods in which the data in Table 14.4 are rel-
evant. The easiest way to see how this affects the analysis is to work backward by consid-
ering first the decisions of the firms in the last (tenth) week. In that week, it is rational for
both firms to cheat because neither can be punished by the other after that, when the market
will effectively be competitive. In the last period, the one-period prisoner’s dilemma model
that we initially discussed is relevant.
Both firms realize that it is in the other’s interest to cheat in week 10. Now consider the
choice in the ninth week. Because Artesia knows that Utopia will cheat in the tenth week,
APPLICATION 14.3
Cooperation between rivals can emerge in some unlikely places. A poignant example involves the “live-and-let-live” system that surfaced in World War I trench warfare.6 The system emerged despite the passions of battle and the mil- itary logic of “kill or be killed.”
The situation along the Western Front of World War I can be represented as a repeated-game prisoner’s dilemma. In any given locality, opposing units could either “cheat” (shoot to kill) or “cooperate” (withhold fire or shoot in such a way as to miss). Cheating was the dominant one-period-game strat- egy for both sides. This is so because weakening the enemy through cheating increased the cheating side’s chances of survival. Cheating by both sides, however, resulted in an outcome—heavy losses inflicted on both sides for little or no gain—that was inferior to the one produced by coopera- tion. And opposing units interacted with each other for what appeared, at least to them, indefinite periods of time.
The diaries, letters, and reminiscences of the trench fighters testify to the “live-and-let-live” (i.e., cooperative)
Cooperation in the Trenches of World War I
equilibrium that eventually emerged. One British staff officer touring the trenches was “astonished to observe German soldiers walking about within rifle range behind their own lines. Our men appear to take no notice.” A soldier commented: “It would be child’s play to shell the road behind the enemy’s trenches, crowded as it must be with ration wagons and water carts, into a bloodstained wilder- ness . . . but on the whole there is silence. After all, if you prevent your enemy from drawing his rations, his remedy is simple: he will prevent you from drawing yours.” Another British officer recounted: “I was having tea with A Com- pany when . . . suddenly a [German] salvo arrived but did no damage. Naturally, both sides got down and our men started swearing at the Germans, when all at once a brave German got on to his parapet and shouted out ‘We are very sorry about that; we hope no one was hurt. It is not our fault, it is that damned Prussian artillery [behind the front lines].’”
Believing that tacit truces would undermine troop morale, the high commands of both sides began rotating troops and ordering raids (whose success or failure could be monitored by headquarters staff) in an effort to destroy the “live-and-let-live” system.
6Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984). The quotes that follow come from Chapter 4.
Asymmetr ic Information 369
C14.INDD 10:46:44:AM 08/06/2014 PAGE 369Trim Size: 203.2 mm X 254 mm
Artesia has no reason to comply in the ninth week. (The only reason for complying is to
avoid its rival cheating in the next period, and it is going to do that anyway.) The same
holds true for Utopia. Both firms will cheat in the ninth week, and both know this. Working
backward in this way, we find that the firms’ incentive is now to cheat in every period! This
outcome occurs because of the common knowledge that the repeated game comes to an end
at some point in the future.
Modifying the other assumptions in the repeated-game analysis also makes the collu-
sive outcome more difficult to achieve. When demand conditions are variable, for example,
the price each firm receives can change either because of cheating by its rival or because
of falling market demand. How can the firm pinpoint the cause? If it assumes cheating
whenever the price falls, the collusive arrangement can fall apart as the firms retaliate even
though cheating may not have occurred.
Reaching a general conclusion about the repeated-game model of cartels is difficult
because many factors can clearly influence the outcome. The analysis does, however, sug-
gest how collusion might emerge and be enforced for at least some period of time. Still, the
factors we emphasized in our discussion of cartels in Chapter 13—the possibility of entry,
cheating, and the difficulty of reaching agreement, especially with heterogeneous firms and
products—are forces that tend to undermine collusive arrangements.
Game Theory and Oligopoly: A Summary We have given only a brief introduction to game theory as it relates to the study of oli-
gopoly. Our purpose has been to convey the nature of this approach to studying markets in
which strategic interactions are important. As is apparent from our discussion of collusion
as a possible outcome in a repeated game, the strategic interactions become very compli-
cated. Game theory provides a technique that is suited to investigate such interactions, but
as applied in the research literature it is a far more mathematical treatment than might be
suggested by our attempt to explain its nature in a simple way.
Because game theory has become the dominant approach to the analysis of oligopoly over
the last few decades, let’s examine its lessons. Unfortunately, the use of game theory has
not produced a general theory of oligopoly—a theory that would tell us, for example, that a
market characterized by factors A and B will operate as in the Cournot model, while a market characterized by factors C and D will operate as a dominant firm model. A large number of possible outcomes have been enumerated in game theory models, but we do not know when
or whether these outcomes will occur in real-world markets. In this sense, the state of oligop-
oly theory is much as it was before the application of game-theoretic techniques.
Progress has been made, however, in illuminating more clearly the complexity of the
whole topic of oligopoly analysis, and many theorists expect further research to provide
the basis for a better understanding of the way oligopolistic markets function. For now, we
have to be content with the recognition that the outcome in such a market may fall any-
where between the monopoly (the perfect cartel case) and the competitive outcome. And
whether the outcome is closer to monopoly or to competition depends on the interplay of a
wide variety of factors in ways that are not yet fully understood.
14.4 Asymmetric Information All of the models presented so far—the competitive model as well as the imperfectly com-
petitive models—have been based on the assumption that market participants have all the
information needed to make informed choices. For firms, this means knowing technology,
input costs, and the prices consumers will pay for different products. For consumers, this
means knowing product characteristics and prices. Although this assumption regarding
370 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 370Trim Size: 203.2 mm X 254 mm
knowledge is sometimes referred to as the perfect information assumption, the term is an exaggeration. Consumers and firms do not have to know everything for the analyses to be
valid. Nonetheless, the assumption places meaningful restrictions on the models, and it is
important to consider how markets function when there is imperfect information and mar- ket participants lack some information relevant to their decisions. We will begin by consid-
ering a common feature in many markets: when consumers have difficulty determining the
quality of products prior to purchase.
The “Lemons” Model7
We are all familiar with “lemons”: products that repeatedly break down or perform unsatisfac-
torily relative to what we expected. The model we will examine suggests that a high propor-
tion of goods may be lemons in a market where buyers are less well informed about product
quality than sellers. The basic assumption is one of asymmetric information: participants on one side of the market (in our case, sellers) know more about a good’s quality than do partici-
pants on the other side (buyers). One market where this characteristic seems prevalent is the
one for used cars. A seller of a used car normally has extensive experience using the car and
can be expected to know its defects. It is often difficult for a prospective buyer to determine
how good the car is until after having purchased and driven it for a while. At the time of the
transaction, buyers are likely to have less information than sellers about product quality.
Before examining how asymmetric information affects the functioning of used car
markets, let’s first consider, for purposes of contrast, a market where all parties are fully
informed. Suppose that there are only two types of used cars, “good” (high-quality) cars
and “bad” (low-quality) cars. Consumers are willing to pay $12,000 for a good car and
$6,000 for a bad car. To simplify, assume that the market demand curves are perfectly elas-
tic at these prices. In this situation, there will be two markets, one for good cars and one for
bad cars, and the prices will be $12,000 and $6,000, respectively. Assume also that sales of
each type of car are 50,000. These markets will function as the competitive model predicts.
Now assume that buyers cannot distinguish between good and bad cars, but the sellers
know the difference. How will this affect the market for used cars? Consider the buyers.
Only one price will now prevail in the used car market because buyers cannot distinguish
between good and bad cars at the time of purchase. Buyers do know, however, that the car
may turn out to be either good or bad. How much the buyer is willing to pay for a used car
will then depend on how likely it is that the car will be good or bad. To contrast this market
with the full-information outcome given previously, let’s initially suppose that consumers
believe there is a 50 percent chance that the car they purchase will be a good one and a 50
percent chance it will be a bad one (because 50 percent of the cars sold in the full-informa-
tion model were of each type). Then we expect the typical buyer to be willing to pay about
$9,000 for a used car—because this is the average value of a used car when half turn out to
be good (worth $12,000) and half turn out to be bad (worth $6,000).
From this discussion, we might expect that the used car market would now be in equilib-
rium at a single price of $9,000. But that is premature. We have not considered the response
of the used car sellers. The sellers know the qualities of their cars. When sellers of high-
quality used cars confronted a price of $12,000 (in the full-information model), we assumed
they would choose to sell 50,000 units. Now these sellers will get only $9,000, and so we
expect them to offer fewer units for sale. With an upward-sloping supply curve, fewer good
cars will be sold when the price is lower. Similarly, if 50,000 bad cars would be sold at a
price of $6,000, when the price is $9,000 we expect more owners of lemons to unload their
cars. Suppose that at a price of $9,000, sales of good cars would be 25,000 and sales of bad
cars would be 75,000.
imperfect information the case when market participants lack some information relevant to their decisions
asymmetric information a case in which participants on one side of the market know more about a good’s quality than do participants on the other side
7The model explained in this section is based on George A. Akerlof, “The Market for ‘Lemons’: Quality Uncer- tainty and the Market Mechanism,” Quarterly Journal of Economics, 84, No. 3 (August 1970), pp. 488–500.
Asymmetr ic Information 371
C14.INDD 10:46:44:AM 08/06/2014 PAGE 371Trim Size: 203.2 mm X 254 mm
If the price were $9,000, we would expect consumers to become aware that it is more
likely they will get a low-quality than a high-quality car in this market. That will affect
their willingness to pay. If consumers perceive correctly that three-fourths of the time they
get a bad car, they will be willing to pay only $7,500 for a used car (a weighted average of
the $12,000 and $6,000 values of good and bad cars, with a weight of three-fourths on the
$6,000 figure). But if the price is $7,500, this will decrease the quantities of both types of
used cars. However, it is likely to increase the proportion of bad used cars in the market, leading to a further revaluation downward in the price consumers will pay.
Exactly where this process ends—where equilibrium will be reached—depends on the
supply elasticities of good and bad used cars. It is possible that the process continues until
only low-quality used cars are sold at a price of $6,000. It is also possible that there will be
an equilibrium in which both good and bad cars are sold. Which outcome results is not as
important as recognizing that the proportion of used cars sold that are high quality will be lower than when consumers know the qualities before making the purchases. Low-quality products tend to drive out high-quality products when there is asymmetric information.
Our use of used cars as an example should not be interpreted to mean that the analysis
applies only to used products or that it applies to all used products. This analysis may apply
in any market in which consumers have difficulty determining product quality. This is often
true of the markets for technologically sophisticated products like personal computers, cel-
lular phones or DVD players. It is also true of some services: Did you know the quality of
instruction at your college when you enrolled? The problem also often arises in purchasing
the services of plumbers, carpenters, doctors, and dentists, to give only a few examples.
Market Responses to Asymmetric Information Do low-quality products really drive out high-quality products? Certainly this is not always
the case and may never be the case. The preceding analysis is intended to draw out the impli-
cations of the assumption that consumers have no way of judging quality. Consumers often
do have ways of distinguishing low- from high-quality products. Before discussing a few of
them, let’s consider why the preceding analysis is incomplete. It is incomplete because it sug- gests there are substantial mutual gains to be realized if consumers seeking high-quality prod-
ucts can be paired with sellers of high-quality products. (In our example, sellers of good used
cars would be better off selling them at, say, $10,000 than not selling any, and buyers would
also be better off.) This means that people have the incentive to acquire (or disseminate)
information that allows consumers to know when they are getting a high-quality product. Of
course, information itself is a scarce good, and there are costs to acquiring and disseminating
information. That is one reason why people are not fully informed: the benefits from acquir-
ing information about product quality will not always be worth its costs.
The way in which information is acquired and used by consumers depends on many fac-
tors, including the nature of the product and its price, and therefore will differ from market to
market. For low-priced products that are frequently purchased (e.g., ballpoint pens), personal
experience may be the most economical source of information. When goods are higher priced
and are purchased infrequently (automobiles, stereos, and so on), it becomes more important
to not be stuck with a lemon, and consumers take some care before making a purchase. In
the case of a used car, for example, they are likely to want a test drive and may take it to a
mechanic whose quality they already know. For many products, consumers can consult publi-
cations such as Consumer Reports for disinterested evaluations of product quality. Similarly, prospective college students may consult several of the available college guides before choos-
ing a college. The opinions of people you know and trust can also provide useful information.
In many cases, it is more efficient for sellers to take the initiative in providing informa-
tion about product quality. There are several ways to accomplish this goal. For example,
sellers of high-quality products may offer guarantees or warranties. This step communi-
cates to consumers that the products are high quality, and firms are willing to incur the
372 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 372Trim Size: 203.2 mm X 254 mm
costs of the guarantees because they can charge more when consumers believe they are
getting a high-quality product. A key issue here is whether the information provided by the
seller is believable. A guarantee is more believable if offered by an established firm than if
offered by a stranger peddling “gold” watches on a street corner. In a similar fashion, many
firms take actions to develop a reputation, or a brand name, for selling high-quality prod-
ucts. The brand name can provide reliable information about the quality of a firm’s prod-
ucts. Finally, liability laws give firms an incentive to avoid at least the most serious quality
defects because firms can be bankrupted by suits from consumers.
The Relevance of the Lemons Model Do these various ways of acquiring and disseminating information imply that the lem-
ons model tells us nothing about real-world markets? It would be a mistake, we think, to
reach that conclusion. The provision and acquisition of information has its costs, and for
products where these costs are high we would expect the lemons model to provide some
insight. There are, for example, still many purveyors of “effortless weight loss” programs
and wrinkle-removal creams, even though their “products” have been discredited (to the
satisfaction of most people who have investigated them—but investigation has costs).
Even where the cost of product quality information is relatively low and markets func-
tion well, it is worthwhile to understand that the markets do function differently than they
would if consumers were fully informed without cost.
A final point is important: when a real-world market functions differently than it
would if consumers were fully informed, it does not mean that the market is necessarily
inefficient. Informing consumers is costly, and that usually means that it is efficient for
consumers to be something less than fully informed. That is, the costs of informing con-
sumers may be greater than the benefits produced.
Is There a Lemons Problem in Used Car Markets? For reasons we have explained, it is doubtful that bad cars will totally drive out good cars
in used car markets. However, it is possible that asymmetric information has an effect on
the way markets for used cars operate. One study, by economist James Lacko of the Federal
Trade Commission, examined this issue through the use of survey data.8 Lacko reasoned
that to the extent the lemons problem was relevant, the quality of used cars should vary by
type of seller. Cars purchased from used car dealers (who may provide warranties and have
reputations at stake) and from friends or relatives should be of higher quality than those
purchased from unknown individuals through a newspaper ad.
Quality is difficult to measure, of course; three different measures were designed for this
study. One was based simply on a buyer’s own evaluations of a car’s mechanical condition
using a 10-point scale, with 1 being a lemon and 10 being a “gem.” On this scale, the average
used car’s condition was rated at 6.65. After controlling for various factors (such as age and
mileage), the researcher found that for cars between 1 and 7 years old, there were few differ-
ences among the various types of sellers. This runs contrary to the lemons model. On the other
hand, for older cars (8 to 15 years), cars purchased from dealers and friends were rated higher
than those purchased through a newspaper ad from a stranger. Cars purchased from a used car
dealer, for example, were rated 0.91 points higher than those purchased through an ad.
The Lacko study suggests that asymmetric information about the quality of used cars
has no effect for cars less than 8 years old and a limited effect for older cars. Apparently,
consumers do obtain enough information through the sorts of channels we discussed to
avoid the extreme outcome predicted by the lemons model.
8James M. Lacko, Product Quality and Information in the Used Car Market, Bureau of Economics Staff Report, Federal Trade Commission (Washington, D.C.: U.S. Government Printing Office, 1986).
Adverse Select ion and Moral Hazard 373
C14.INDD 10:46:44:AM 08/06/2014 PAGE 373Trim Size: 203.2 mm X 254 mm
It must also be kept in mind, moreover, that markets are dynamic, and where there are
any further opportunities for mutual gain, entrepreneurs have an incentive to exploit such
prospective gains. In the case of the used car market, for example, recent years have seen
the advent and rapid growth of firms such as Auction Direct that specialize solely in being
high-volume purveyors of used cars. By better tapping into the reputational mechanisms
that facilitate matching consumers seeking high-quality used cars with sellers of such high-
quality used cars and also by capitalizing on any learning-by-doing and scale advantages,
firms such as Auction Direct have been acquiring increasing market share.
signal In cases of asymmetric information, a method through which high-quality sellers can distinguish
themselves from low-quality sellers to prospective buyers
APPLICATION 14.4
The job market is characterized by asymmetric informa- tion. That is, employers are unsure about the productive capabilities of different prospective employees and thus worry about hiring a “lemon” as opposed to someone with your extensive abilities and favorable values.
Employers, of course, have numerous ways to distin- guish between low- and high-quality employees. Inter- viewing and reference checking are among these, albeit imperfect, screening techniques. In turn, prospective employees have some mechanisms by which to signal that they are high- as opposed to low-quality material. Invest- ing in education is one of these mechanisms. Indeed, in a paper that helped him win a Nobel Prize, Stanford Gradu- ate School of Business professor and former dean Michael Spence shows how you may want to invest in education, such as obtaining an MBA, so as to signal to an employer that you have desirable abilities and values.9
Job Market Signaling
Signals can serve to distinguish high- from low-quality employees if they are costly. An MBA education certainly meets this requirement with its associated tuition and opportunity costs. Moreover, the costs of signaling need to be inversely correlated with productive capability. To under- stand this point, imagine what would happen in a world in which lower-quality prospective employees could more easily get admitted into MBA programs.
Spence shows that, under the foregoing conditions, even if education such as an MBA does not contribute anything to an employee’s productivity, high-quality employees will invest in buying more education so as to signal their higher produc- tivity to employers. In turn, employers will use education as a way to estimate the quality of potential employees and pay higher wages to individuals with higher education levels.
In a world in which education may not contribute any- thing to an employee’s productivity but serves a signal- ing purpose, Jeff Pfeffer, a faculty colleague of Spence’s at Stanford, has quipped that individuals who get admitted to a top-tier business school shouldn’t even bother to attend. Pfeffer’s advice to these individuals is to instead take their offer of admission to the employer for whom they ulti- mately want to work and thereby bypass the need to invest further in costly signaling.
9Michael Spence, “Job Market Signaling,” Quarterly Journal of Economics, 87, No. 3 (August 1973), pp. 355–374.
14.5 Adverse Selection and Moral Hazard Our analysis of asymmetric information has emphasized markets in which consumers have
less information than sellers. In some important instances, for example, insurance markets,
it is the firms that are less well informed.10 Our discussion in Chapter 5 of insurance mar-
kets was based on the assumption of full information (that is, both firms and consumers
know the risks). Now let’s see how insurance markets may function when this assumption
is modified.
10Another example, as discussed in greater detail in Application 14.4, is labor markets: when a firm hires workers, it is less well informed about the quality of the workers’ labor services than the workers.
374 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 374Trim Size: 203.2 mm X 254 mm
Adverse Selection The most important information affecting the operation of insurance markets is the prob-
ability that the insured-against event will occur. Insurance companies have amassed sta-
tistical data that enable them to estimate these probabilities. They may find, for example,
that one out of a thousand houses they insure burns down each year. What they often don’t
know is how the probability varies from one homeowner to another. Thus, it is quite pos-
sible that some homeowners have better information than the insurance companies. In
the case of a potential arsonist, this is certain to be the case, but even in less extreme cir-
cumstances (e.g., people who store flammables alongside electrical wiring in their attics)
consumers may know whether their risks are much higher than average. Thus, it is quite
possible that at least some consumers have better information than the firms do. This asym-
metric information can have profound effects on the operation of the market.
Assume that some homeowners are much more at risk of suffering fire damage to their
homes, and that these homeowners know it. Insurance companies know the average risks
based on their experience, and they have to charge premiums based on that. What consum-
ers will find this an attractive deal? Clearly, the high-risk homeowners will find the price
attractive; think of how much insurance an arsonist would buy when $100 in coverage can
be purchased for $1, where this fee reflects the average risk. On the other hand, the insur-
ance is much less attractive to low-risk homeowners. So the insurance companies will find
most of their customers coming from the high-risk group; they get an adverse selection from the pool of potential customers. The “undesirable” customers, the high-risk home-
owners, are more likely to appear in the market (this is the adverse selection), and the insur-
ance companies cannot distinguish high-risk from low-risk homeowners when they sell
policies.
Imagine where this process can lead. As mostly high-risk persons buy insurance, the
insurance company finds that it has to pay off on a larger share of policies than initially
predicted. The average riskiness of its customers is thus higher than for the population as
a whole, and this causes the price of insurance to rise. The higher cost of insurance drives
away more low-risk customers and further raises the cost of insurance. In the end, it is
possible that only high-risk customers are served, and low-risk customers must go without
insurance that could potentially benefit them if it reflected their true risk status. (In a full-
information world, high- and low-risk customers would simply be charged different prices,
reflecting the difference in risk.) This analysis should sound familiar, for it is essentially
the lemons problem in a new setting. Adverse selection was the driving force in our used
car example, also. There, sellers of low-quality used cars were adversely selected because
buyers could not distinguish between good and bad cars. Here, high-risk customers are
adversely selected because firms cannot distinguish between high- and low-risk customers.
The adverse selection problem may be important in many insurance markets. In life and
medical insurance markets, customers often have a better idea of their risk status (from
their own medical or family history) than do insurance companies. Similarly, automobile
drivers who speed or drink and drive are more likely to have accidents, and they know more
about their driving behavior than an insurance company does when providing them cover-
age. Doctors who purchase medical malpractice insurance may also be better informed of
their lawsuit risks than their insurance companies. In all these markets, the adverse selec-
tion problem possibly leads to a situation where mostly high-risk customers are insured and
many low-risk customers choose to remain uninsured.
Market Responses to Adverse Selection Of course, the outcome is not likely to be as dire as the analysis so far suggests, and the
reason is the same as in the lemons model: there are potential gains to market participants
from adjusting their behavior to account for the adverse selection problem. For example,
adverse selection a situation in which asymmetric information causes higher-risk customers to be more likely to purchase or sellers to be more likely to supply low-quality goods
Adverse Select ion and Moral Hazard 375
C14.INDD 10:46:44:AM 08/06/2014 PAGE 375Trim Size: 203.2 mm X 254 mm
homeowners’ insurance covers only the market value of structures and contents. By placing
an upper limit on the potential losses, insurance firms reduce the costs imposed by high-risk
customers, and this lowers the cost of insurance. (Imagine if potential arsonists could insure
the family photo album for its “sentimental value” of $100,000.)
Other insurance company practices also make more sense when the adverse selection
problem is understood. For health and life insurance policies, companies often require
physical exams (to help distinguish high- from low-risk people) and a waiting period before
a policy is in force (some maladies may not be apparent in a physical, even though the
consumer is aware of them). In some cases, insurance companies use indirect measures to
help identify the riskiness of customers. For instance, men aged 15 to 24 have car accidents
with about twice the frequency of women the same age, so gender can be used as an indicator
of riskiness. Similarly, women live longer than men and so are at lower risk of dying at any
age; hence, life insurance is less costly for them.
Finally, group health plans have been developed partly in response to the adverse selec-
tion problem. These plans offer policies covering all of a firm’s employees. Because all
employees must be enrolled, the likelihood that high-risk people will be overrepresented is
smaller; the insured workers are more likely to be representative of the average population.
Adverse selection can still occur (high-risk workers choose to work for firms with exten-
sive coverage), but that is less likely to be a problem than in the sale of individual policies,
so group plans can provide insurance coverage at lower costs. (There are also some other
reasons for the lower cost of group plans, such as lower administrative costs.)
APPLICATION 14.5
Fewer than 4 percent of all blood donors are paid for their donation. A key reason why so-called commercial blood is not more common is adverse selection.11 Critics of for-profit giving argue that when people are motivated to
Adverse Selection and the American Red Cross
donate blood for the sake of a financial reward, blood banks are more likely to attract donors who are desperate for the money because they are addicted to drugs or alcohol or have a serious infectious disease. Indeed, numerous studies have found that hepatitis, a disease that can be transmitted through blood transfusion and that inflames the liver and can occasionally be fatal, is much more likely to be pres- ent in commercially collected blood than in blood that is donated on a nonprofit basis.
11This application is based on Michael L. Katz and Harvey S. Rosen, Microeconomics (Boston: Irwin/McGraw-Hill, 1998); and Alvin W. Drake, Stan N. Finkelstein, and Harvey M. Sapolsky, The American Blood Supply (Cambridge, MA: MIT Press, 1982).
Moral Hazard Moral hazard is another problem endemic to insurance markets. It occurs when, as a result of having insurance, an individual’s behavior changes in such a way that the probability of
the unfavorable outcome increases or its cost is greater when it does occur. An uninsured
homeowner, for example, would take great precautions against fire. The homeowner might,
for example, avoid the use of kerosene heaters, install smoke detectors, prohibit smoking,
and have an electrician inspect the wiring frequently. The incentives are altered when the
costs of replacing the house are covered by insurance. The benefits from avoiding a fire are
now smaller, and so the homeowner is likely to devote fewer resources to that use. What
this behavior means is that fires become more likely when the parties are insured, and this
leads to higher prices of insurance coverage.
The extent of the moral hazard problem is likely to vary across individuals and types of
insurable events. People with car insurance may be more careless about locking their cars
moral hazard a situation that occurs when, as a result of having insurance, an individual becomes more likely to engage in risky behavior
376 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 376Trim Size: 203.2 mm X 254 mm
or may park them in riskier neighborhoods. But does health insurance coverage lead people
to exercise less, eat poorly, and smoke more? One reason the problem may be less severe in
this instance is that health insurance does not cover all the costs of illness. Although it may
cover all the medical costs, for many illnesses the costs to the patient of the pain, suffering,
and disfigurement can be as much or greater. This gives people an incentive to take actions
to avoid the illness even when the medical costs are covered. But the moral hazard problem
is an important reason why you can’t get insurance to compensate you fully for all (medical and other) costs you bear from illness: that sort of insurance would give you no reason for
taking care of yourself.
The moral hazard problem arises when insurance companies lack knowledge of the
actions people take that may affect the occurrence of unfavorable events. If the actions
are observable, the policies can be made contingent on performance of those actions. This
can work to the advantage of insurance companies as well as insured parties who receive
more favorable rates. For instance, homes with security systems and smoke detectors may
receive coverage at lower rates, giving homeowners incentives to take actions that lower
the probability of theft and fire. Similarly, smokers are generally charged higher health
insurance premiums than nonsmokers, making smokers bear the higher cost of medical
care that results from their behavior, and at the same time giving them an incentive to quit
smoking. “Good driver” policies reward automobile owners with unblemished records
through lower rates.
Another type of moral hazard problem is particularly significant to health insurance.
Consider a policy that covers all hospital costs. When a person is hospitalized, he or she
will, in effect, face a zero price for treatment. This is likely to lead the patient to consume
hospital services beyond the point where they are worth what they cost. Any form of medi-
cal care that has any benefit, no matter how small, will seem worthwhile if the insurance
company is absorbing the expense. In this setting, doctors are likely to prescribe expensive
tests and sophisticated treatments, knowing that financial responsibility falls to a faceless
third party. But when patients and doctors behave this way, the cost of hospitalization goes
up and insurance premiums rise to cover the cost.
APPLICATION 14.6
In the second half of 2007, the U.S. economy began to contract. The contraction worsened as subprime home mortgage loan sellers, such as Countrywide, began to see a decreased probability that their loans would be repaid by borrowers on time and in full. (In the home mortgage loan industry, it was sometimes, only half-facetiously, noted that subprime mortgage loans were available even for those individuals with “no income, no job, and no assets.”) In turn, major financial institutions such as Bear Stearns, Citigroup, Lehman Brothers, and UBS—which had invested heavily in tradable subprime-mortgage- based assets—began to see their firms’ worth erode dra- matically. For example, the price of Bear Stearns stock fell from a high of $172 per share in January 2007 to $93 by February 2008 and further to only $2 in March 2008 (a point in time when the Federal Reserve stepped in and helped orchestrate the sale of Bear Stearns to the
Moral Hazard and Subprime Home Mortgages
rival Wall Street firm JPMorgan Chase through a loan of $30 billion).
Financial market analysts noted how moral hazard may have led major financial institutions to make too many loans to individuals with questionable repayment abilities and/or invest in assets backed by such subprime mortgage loans. To the extent that major financial institutions believed that the low interest rate environment fostered by monetary policymakers for a lengthy period of time prior to the sub- prime crisis would continue and that government authori- ties also would do everything in their power to prevent a recession from developing and to keep asset prices up, the inhibition against making risky investments was diminished. The macroeconomic consequences of such moral hazard have been substantial—as have been the costs to the aver- age taxpayer, who has ended up footing the bill for the gov- ernment-sponsored bailout of those financial institutions.
Limited Pr ice Information 377
C14.INDD 10:46:44:AM 08/06/2014 PAGE 377Trim Size: 203.2 mm X 254 mm
Market Responses to Moral Hazard A number of practices have evolved in the medical insurance market to deal with moral haz-
ard. One is limitations on the services covered by insurance (no more than three days in the
hospital for an appendectomy, for example). Although patients may believe this reflects stin-
giness on the insurer’s part, it actually leads to lower insurance premiums and can be in the
long-run interest of the insured. A second approach is to require the insured person to pay
part of the costs. Patients might be required to pay 20 percent of the hospital bill; the share of
the cost borne by the patient is called the coinsurance rate. Use of coinsurance reduces the cost of insurance directly, but also indirectly lowers it by giving patients incentive to be more
economical in their use of hospital services. A third approach is the use of deductibles. Using a deductible means that a patient must pay, for example, the first $500 of hospital costs before
insurance coverage is effective. This gives patients incentive to take account of all costs in the
case of minor medical treatments, and also leads to lower insurance premiums.
Insurance markets are profoundly affected by asymmetric information, as this discussion
indicates. There is no doubt that these markets function very differently from the way they
would if all parties had perfect information. Understanding the problems created by asym-
metric information helps us see why certain practices have emerged to mitigate them.
APPLICATION 14.7
Since 1979, holders of New York City taxi medallions have been legally permitted to lease out the rights associ- ated with the medallions to operate a taxi for 12-hour shifts (see Application 10.4). As a result of this policy change, 18 percent of cabs in New York City are now owner-operated while the rest are driven by lessees.
Given that lessees pay only part or none of the vari- able costs associated with their driving—including main- tenance, repair, replacement, and insurance—a moral hazard problem exists. At the core of the problem is that it is hard to contractually specify the level of vehi- cle care for lessees and to monitor their driving behav- ior at reasonable cost. The problem is compounded by the fact that leased taxis are often operated by multiple drivers. Thus, the level of care that lessees can be expected to exercise when driving a cab will be less than owner-operators. As goes the old saying: “No one ever washes a rental car…only owners wash their automobiles.”
Moral Hazard on the Streets of New York City
How significant is the moral hazard problem in the case of New York City cab drivers? Professor Henry Schneider of Cornell finds that long-term lessees are involved in 64 per- cent more driving violations and 62 percent more accidents per mile than are cabs driven exclusively by owner-drivers.12
Moreover, lessees failed vehicle safety inspections at a 67 per- cent higher rate than cabs operated solely by owner-drivers. After controlling for vehicle usage and driver characteristics, Professor Schneider estimates that moral hazard explains 21 percent of lessees’ driving violations, 16 percent of their acci- dents, and 30 percent of leased taxis’ higher vehicle inspec- tion failure rates.
The findings suggest that passengers looking to hail a cab in the Big Apple would be advised to find a way to travel with owner-operators instead of with lessees.
12Henry Schneider, “Moral Hazard in Leasing Contracts: Evidence from the New York City Taxi Industry,” Journal of Law and Economics, 53, No. 2 (November 2010), pp. 783–805.
14.6 Limited Price Information Consumers may lack information about product prices as well as product qualities. In the
competitive model, where consumers are fully informed, if one firm raises its price above
the competitive level (the price charged by the other firms), it will lose all its sales because
its customers know that the other firms charge a lower price and so will purchase from
378 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 378Trim Size: 203.2 mm X 254 mm
them. When consumers have perfect information about prices, all firms have to charge the
same price and thus each firm faces a horizontal demand curve.
By contrast, if consumers do not know the prices at stores other than the one where they
are currently shopping, retailers can raise prices without losing all customers. Consumers’
lack of information about competitors’ prices means that each store confronts a downward-
sloping demand curve, which, in turn, gives it some market power. Exactly how such a
market functions depends on a number of factors, including the extent of consumer igno-
rance and the cost of acquiring price information. In general, however, we expect that the
firms in markets where many or all consumers are uninformed will be charging different
prices. There is likely to be a range of prices or price dispersion for the same product. Now consider consumer behavior in a market where there is price dispersion. Consum-
ers wish to purchase from the firm offering the product for the lowest price, but they don’t
know which firm it is. They can find out, but there is a cost of acquiring the information.
Search costs are the costs that consumers incur in acquiring information; they include such things as time (making telephone calls, buying newspapers, reading the ads) and transpor-
tation between stores. There is also a benefit from acquiring price information, of course,
because consumers can buy the product for a lower price. However, consumers are unlikely
to search until they are fully informed about all the prices being charged by various stores.
The reason for this is that the expected marginal benefit from additional search declines the
longer the search goes on.
A simple example illustrates this point. Suppose 20 stores sell a compact disc you want
to purchase. You call one store and find its price is $12. Should you call a second store? If
all stores charge different prices, the probability that the second store will charge a lower
price is about one-half. So you call a second store and find its price is $13; you lost the coin
flip. Now the probability that a third telephone call will generate a price lower than either of
these is only about one-third. Even after calling 19 stores, there is a slight probability that
the lowest-price store is the one you haven’t called, but you would be unlikely to take the
time to call all 20 stores just to guarantee you find the lowest price. In general, because the
expected marginal benefit of additional search declines with the amount of search, you will
not keep placing calls until you become fully informed. You may stop after three or four
telephone calls and purchase from the lowest-priced source you have located, although that
will often not be the lowest price available in the market.
The Effect of Search Intensity on Price Dispersion For a given consumer, the amount of search undertaken will have no effect on the actual
price dispersion in the market. But if many or all consumers increase search intensity,
becoming better informed about price, the price dispersion will be reduced. As consumers
become better informed, high-priced firms lose business relative to low-priced firms, and
high-priced firms are forced to reduce their prices. Taken to the limit, if consumers become
fully informed, only one price can prevail.
How does this help us understand the amount of price dispersion for a given product?
The theory predicts that the dispersion falls when consumers search more (e.g., become bet-
ter informed). They will search more when the benefit from search is higher than the cost. A
little thought will convince you that the benefit will be higher the greater the product’s price.
Finding a store that sells a compact disc for 1 percent less than another may not be worth an
extra phone call, but saving 1 percent on the price of a car is likely to be worth it. Thus, it is
not surprising that empirical studies have found less relative price dispersion (price variation
compared with the average price) for higher-priced products. Prices are less widely dispersed
for cars of the same make and model, for example, than for washing machines.13
price dispersion a range of prices for the same product, usually as a result of customers’ lacking price information
search costs the costs that customers incur in acquiring information
13George J. Stigler, “The Economics of Information,” Journal of Political Economy, 69, No. 2 (June 1961), pp. 213–225.
Advert is ing 379
C14.INDD 10:46:44:AM 08/06/2014 PAGE 379Trim Size: 203.2 mm X 254 mm
Finally, we should not ignore the possibility that when consumers are not fully informed
and there is price dispersion in a market, it may be in the interest of the low-price firms to
inform consumers about price. This is one reason why firms advertise, which leads us to the
next section.
14.7 Advertising Firms advertise in an attempt to increase the demand for their products. On that point econ-
omists agree. But there is some disagreement over whether advertising expenditures serve
the useful function of providing information to consumers or the baleful function of wast-
ing resources and distorting consumption choices.
Economists have a long tradition of skepticism regarding the benefits of advertising. In
this view, competitive firms have no need to advertise because they can sell as much as
they want at the market price (although this point is disputed), so the very existence of
advertising implies that firms have some monopoly power. But advertising itself may not
only be a symptom of market power—it can also help firms achieve and maintain market
power. Firms may use Madison Avenue’s tools of persuasion to convince consumers that
their products are different from and better than those of competitors. This is sometimes
referred to as artificial product differentiation and, if successful, increases the demand for the product and also makes demand more inelastic, conferring additional market power
on the firm. For example, the chemical composition of Bayer aspirin is nearly indistin-
guishable from that of generic aspirin, but many consumers think Bayer is better and pay a
substantially higher price for that brand.
It is also charged that advertising can operate as a barrier to entry. If a new firm attempts
to enter a profitable industry, it may find that advertising by the established firms has cre-
ated a captive audience of consumers who are reluctant to try a new brand. It may be nec-
essary for an entrant to wage a massive advertising campaign to get consumers to give its
product a try, and the prospect of that cost could be an effective deterrent. On the other
hand, advertising can also be a means of breaking into an entrenched market, giving an
entrant a way to increase sales quickly so that economies of scale can be realized.
In its most extreme form, criticism of advertising holds that it manipulates consum-
ers and leads them to choose products they don’t want or need. In this view, consumers’
tastes are not formed independently, but are actually created by advertisers. Thus, instead
of consumers’ underlying wants being the factor behind the position of demand curves (and
thereby the pattern of production), producers are held to play the central role of determin-
ing consumers’ wants through their promotional activities.
Although few economists hold this extreme view on advertising today, there continues
to be a belief that advertising may enhance market power, deter entry, and lead to more
concentrated industries. Numerous studies have tried to find a connection between advertis-
ing and industrial concentration or profitability. By and large, their results have been incon-
clusive: about as many find advertising to be unassociated with industrial concentration
or profitability as the reverse. This evidence has dispelled the worst fears of the critics of
advertising, but there remains the possibility that advertising has harmful effects that are
simply not large enough to be empirically identified.
Advertising as Information A view that advertising is benign in its effects has more recently emerged as an outgrowth of
research on the economics of information. Advertising is held to be a low-cost way of pro-
viding information to consumers about the availability of products and their prices and qual-
ities. It may make markets more competitive and even lead to lower prices for consumers.
artificial product differentiation the use of advertising to differentiate products that are essentially the same
380 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 380Trim Size: 203.2 mm X 254 mm
To see why, imagine an industry where there is no advertising. As we explained in the
previous section, firms will then face downward-sloping demand curves because it is costly
for consumers to find out about alternatives. In Figure 14.1, D1 is a typical firm’s demand curve when no firm in the industry advertises. Now if this firm alone advertises, its demand curve shifts outward to D2 and may become more inelastic, as drawn. This appears to sup- port the negative view of advertising—that market power is enhanced as demand becomes
more inelastic. However, we must recognize that other firms are also free to advertise,
and the effects of all firms advertising may be quite different from the effects when only
one does. So, suppose that our typical firm continues to advertise, but now its competi-
tors also try to persuade consumers of the virtues of their products. This will cause this
firm’s demand curve to shift leftward from D2 and become more elastic, as shown by D3, the demand curve when all firms are advertising. It is more elastic than demand when no
firms advertise (D1) because consumers will be more aware of the alternatives available on the market, and an increase in this firm’s price will cause more consumers to shift to other
products than when the consumers are not aware of the alternatives. Thus, it is possible that
advertising, when undertaken by many or all competing firms, actually confronts firms with
more elastic demand curves and reduces their market power.
Advertising and Its Effects on Products’ Prices and Qualities In the previous section, we explained how price dispersion arises in a market when con-
sumers are uninformed. The existence of price dispersion gives consumers an incentive
to incur search costs to obtain information. It also, however, gives firms an incentive to
advertise. For example, if some firms charge a higher price to uninformed consumers and
if firms charging lower prices could inform consumers of this fact, the low-price firms
can enhance the demand for their products and potentially their profits. Moreover, if con-
sumers switch from high-price to low-price firms, it will reduce the average price in the
market. It could also lead the high-price firms to reduce their prices, which reduces the
price dispersion and further lowers the average price. In this way, advertising can lead to
lower prices.
It is also possible for advertising to solve the lemons problem discussed in Section 14.4.
Recall that in this model, consumers cannot determine the quality of the product before
Advertising and a Firm’s Demand Curve When there is no advertising in an oligopoly, a firm’s demand curve is D1. If only this firm advertises, its demand curve shifts out to D2. If all firms advertise, the firm’s demand curve becomes D3.
Output
Dollars per unit
No firms advertise
All firms advertise
One firm advertises
D2D3 D1
0
Figure 14.1
Advert is ing 381
C14.INDD 10:46:44:AM 08/06/2014 PAGE 381Trim Size: 203.2 mm X 254 mm
purchase, and as a result low-quality products drive out high-quality ones. Through adver-
tising, high-quality sellers can inform consumers that their products are of high quality.
The problem arises in deciding whether such claims are true. Won’t the low-quality sell-
ers make the same claims? Possibly so, but surprisingly, high-quality sellers will normally
have a greater incentive to advertise. Such is the importance of repeat purchases and word-
of-mouth endorsements to a firm’s future sales. A firm selling a lemon can at best convince
a consumer to purchase one time, but it can be sure there will be no future sales to that
customer, and it may lose future sales from other consumers that hear of this buyer’s bad
experience. By contrast, a firm selling a high-quality product can gain future sales in addi-
tion to the current sale, so it has a greater incentive to promote its product.
One interesting aspect of this analysis is that it suggests consumers may gain useful
information from advertising even when that advertising does nothing more than assert a
product is “great tasting” (that is, of high quality). Higher-quality (or higher-value) prod-
ucts will tend to be more heavily advertised, and if consumers are influenced by the amount
of advertising they may be led to try the better products.
There is some evidence that advertising does work to the benefit of consumers, as this
analysis suggests. For example, according to one study, the average price of eyeglasses in
states permitting advertising is about 25 percent lower than the prices in states where such
advertising is prohibited.14 Eyeglass prices are also lower in states prohibiting advertisers
from mentioning price but permitting claims of high quality. (Prices are lowest in states
where prices can be advertised.) Thus, even when advertising does not mention prices, it
can apparently contribute to lower prices for consumers. Other studies have also produced
evidence that advertising about price lowers the average price consumers pay for products
such as liquor, drugs, contact lenses, toys, and gasoline.
Advertising, the Full Price of a Product, and Market Efficiency Even when advertising does not lead to lower prices, it may be advantageous for consumers.
Without advertising, consumers have to incur search costs to find out about products. The
true prices they pay are then the sum of the money price and the search costs they bear; this
is sometimes referred to as the full price of the product. One effect of advertising is that it is a substitute for the consumer’s own search efforts. Thus, advertising can reduce consum-
ers’ search costs. Even if the money price is unchanged or rises somewhat, the full price
consumers pay may fall as a result of advertising. For example, even if they do not lead to a
decrease in the money price of products, Internet advertisements promise to decrease the
full price consumers pay for products by lowering consumers’ search costs.
The advertising-as-information view suggests that advertising is a low-cost way of con-
veying useful information to consumers about alternative products and their prices, and
thus makes markets work more efficiently. Not all economists accept this view, and televi-
sion advertising has been singled out for special criticism. (Significantly, however, less than
20 percent of all advertising expenditures are devoted to television; newspapers account for
a larger share and the Internet is a rapidly growing medium.) This is not surprising, because
television advertising is more intrusive than most other forms and is more difficult to target
to interested consumers than, say, advertising in specialized magazines. Moreover, there
is the lingering suspicion that some consumers (of course, not us) may be swayed by the
emphasis on visual images and emotional appeals.
In sum, probably each view of advertising contains an element of truth. It may be that
advertising works well in certain types of markets and for certain types of products but has
deleterious effects in other cases.
full price the sum of the money price and the search costs that consumers incur
14Lee Benham, “The Effect of Advertising on the Price of Eyeglasses,” Journal of Law and Economics, 15, No. 2 (October 1972), pp. 45–74.
382 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 382Trim Size: 203.2 mm X 254 mm
SUMMARY
Game theory analyzes strategic interactions among
decisionmakers and has been applied extensively to
understanding oligopolistic markets.
A payoff matrix clarifies the nature of the problem
confronting decisionmakers and helps identify the
equilibrium.
Two equilibrium types are dominant strategy and
Nash.
A particularly important type of game is the prison-
er’s dilemma, in which self-interest on the part of each
player results in all players being worse off then they
could be if different choices were made.
The prisoner’s dilemma game helps us see why firms
have an incentive to cheat on a cartel agreement. When
the game is repeated, however, the analysis becomes more
complicated, and there is a greater possibility of collusion.
APPLICATION 14.8
Online dating is the most popular subscription service sold on the Internet.15 Each year, more than 40 million Americans participate in online dating services such as Match.com, eHarmony.com, Craigslist, Gay.com, and Yahoo Singles. As part of such a service, a participant provides a brief advertisement of his or her personal characteristics and dating objectives.
How accurate are the individual dating participants’ per- sonal advertisements? According to one economic study of daters in San Diego and Boston, there does appear to be some misrepresentation—particularly on dimensions that each gender perceives to be of interest to the opposite gender. For example, of the male online daters, 4 percent claimed to make over $200,000 annually (versus 1 per- cent for the national average) and to be 1 inch taller than the average American adult male. Of the female online daters, 70 percent claimed to have “above average” looks, and 24 percent asserted that they had “very good looks.” The average listed weight was 20 pounds less than the average for adult American females, and 28 percent noted that they were blonde (a number far above the national average).
In addition to finding apparent misrepresentation, the study also uncovered some brutal honesty. For instance, 8 percent of the male online daters said that they were mar- ried. (Only 9 percent of these married online male daters, however, posted their photos online!)
Of course, there are mechanisms to promote veracity in online personals. The surest way to fail at online dating, for example, is to not post a photo with one’s personal adver- tisement. As the authors of Freakonomics note:
The Effectiveness of Internet Advertising: The Case of Online Dating
A low-income, poorly educated, unhappily employed, not-very-attractive, slightly overweight, and bald- ing man who posts his photo stands a better chance of gleaning some emails [from females] than a man who says he makes $200,000 and is deadly hand- some but doesn’t post a photo. There are plenty of reasons someone might not post a photo . . . but as in the case of a brand-new car with a for-sale sign, pro- spective customers will assume he’s got something seriously wrong under the hood.
There are also a variety of services that, for a small fee, will provide a check of someone’s criminal record, address, and so on. In addition, the Internet and social networking sites can also serve as a means to do background checks on a particular online dater. Witness what happened to upstate New York Congressman Chris Lee, when he posted a shirt- less photo of himself in early 2011 on Craigslist, claiming to be a divorced lobbyist. One of the women viewing Lee’s personal advertisement decided to Google him. To her dis- may, she quickly found out that he was still married and a member of Congress, discrepancies she reported to a national media outlet. Within 3 hours of the discrepancies being noted in the national media, Lee resigned.
Ultimately, as noted by the authors of the study, the effectiveness of online dating needs to be measured against other forms of seeking romantic hookups—forms such as hanging out in bars, being set up on blind dates by family members and friends, and so on. Problems such as imper- fect information, misrepresentation, and adverse selection doubtlessly affect these alternative forms of searching for matches. The fact that so many Americans subscribe to online dating services suggests that, in relative terms, the benefits of such services related to broadening one’s dat- ing market, enhancing the quality of one’s prospective matches, and lowering the cost of search are significant— notwithstanding the associated imperfections.
15This application is based on Steven D. Levitt and Stephen J. Dubner, Freakonomics: A Rogue Economist Explores the Hidden Side of Everything (New York: HarperCollins, 2005); and Gunter J. Hitesh, Ali Hortacsu, and Dan Ariely, “Matching and Sorting in Online Dating,” American Economic Review, 100, No. 1 (March 2010), pp. 130–163.
Review Quest ions and Problems 383
C14.INDD 10:46:44:AM 08/06/2014 PAGE 383Trim Size: 203.2 mm X 254 mm
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
14.1 What is a dominant-strategy equilibrium? What is a Nash equilibrium? Is it possible for a Nash equilibrium to exist
where neither player has a dominant strategy?
*14.2 Construct and explain the payoff matrix for two firms that operate in a competitive market. How does it differ from
the situation illustrated in Table 14.1?
14.3 What is a prisoner’s dilemma game? Why is it relevant in evaluating the likelihood of cheating in a cartel?
14.4 Tables 14.1 and 14.4 both involve two firms each choos- ing between low and high outputs, but only one of the tables
illustrates the prisoner’s dilemma. Explain why the nature of
the market in which firms interact may sometimes produce a
prisoner’s dilemma and sometimes not.
14.5 Is a repeated- or single-period game more appropriate for the study of oligopolies? In which setting is collusion more
likely to be a stable outcome? Explain your answer.
14.6 Construct a payoff matrix to examine the determination of outputs in the Cournot duopoly model. What type of equilib-
rium exists for this model? Does the game-theoretic approach
make this model any more plausible?
14.7 Why do you think that game theory has become the preferred method of analyzing oligopolistic markets? What
advantages does it have over simply assuming, say, Cournot
behavior?
14.8 What is the basic assumption about information in the lemons model?
14.9 In the lemons model, there is only one price even though the products differ in quality. Why is that? What factors deter-
mine that price? How does the price affect the quantities traded
of the different quality goods?
14.10 College instructors know more about the quality of their courses than prospective students. Does this mean that the
lemons model is appropriate? How does this market differ from
the one assumed in the pure lemons model?
14.11 What is adverse selection in insurance markets and how does it relate to the lemons model?
*14.12 Consider insurance covering the costs of cancer when there is no way to determine how likely it is that any given
individual will contract the disease. How will the price of the
policy be determined? Now suppose that it is determined that
smokers have 10 times the risk of nonsmokers. How will the
price be affected if insurance companies cannot determine who
smokes and who doesn’t?
14.13 How does the moral hazard problem differ from the adverse selection problem in markets for medical insurance?
14.14 Why don’t consumers become fully informed about the prices different firms charge? If consumers are not fully
informed, why is a firm likely to possess some degree of mar-
ket power?
14.15 Suppose that a college town has a large number of firms selling a homogeneous product—pizza—and that there are two
types of consumers in the town. The town’s permanent resi-
dents are fully informed about the prices charged by all firms
and always shop at the firm or firms with the lowest price. On
the other hand, the students attending college in the town (tem-
porary residents) are completely uninformed; they do not know
anything about prices and simply choose among firms on a ran-
dom basis. Explain why, in such a setting, a single price may
prevail in the market for pizza.
14.16 How does advertising affect the demand curve confront- ing a single firm? How does the outcome depend on whether
other firms also advertise? If all firms in an industry advertise,
how will this shift the industry demand curve for the product?
*14.17 “Because advertising adds to firms’ costs of production, it cannot lead to lower product prices.” True or false? Explain.
14.18 A considerable number of initial public offerings (IPOs) of stock in a company evidence a substantial run-up in price
during early trading. Explain why asymmetric information
Markets can work very differently when consumers or
firms are not fully informed about prices and/or product
characteristics. When consumers cannot determine the
quality of a product before purchase, the lemons model
suggests that low-quality products will predominate.
Market forces limit the extent of the lemons problem.
In the case of insurance, firms may have less informa-
tion than do consumers. Adverse selection can then lead
to only high-risk customers being insured.
When consumers are not informed about the prices
charged by all firms, it is possible for more than one
price to prevail in the market. It is not necessary, how-
ever, for all customers to be fully informed for a single
price to emerge. In general, the larger the proportion of
informed consumers, the less the price dispersion in the
market.
Advertising is a particularly important way in which
information is provided to consumers by firms. It can
lead markets to operate more efficiently, but there is
also the possibility that it can distort consumer choices
and make it difficult for new firms to enter profitable
industries.
384 Game Theory and the Economics of Information
C14.INDD 10:46:44:AM 08/06/2014 PAGE 384Trim Size: 203.2 mm X 254 mm
between the investment banks organizing the offerings and pro-
spective investors may be the reason for such a phenomenon.
14.19 In multidivision corporations where division heads are allocated an annual budget, explain why the “use-it-or-lose-it”
phenomenon occurs and is a reflection of a prisoner’s dilemma.
14.20 If there is asymmetric information between the owners of a baseball team for which a given player plays and other teams’
owners, would you predict that players who opt to become free
agents and end up getting traded to another team will spend
more days on the disabled list, after being traded, than players
who remain with their existing team? Explain why or why not.
14.21 Imagine that several players are each asked to pick a number from 0 to 100. The prize associated with winning the
game is awarded to the player who comes closest to the num-
ber that is half the average of what all the other players select.
Viewed from a game-theoretic setting, what number will be
selected by the players in a Nash equilibrium? (Economist Hal
Varian of Berkeley notes that real people playing such a game
do not typically make such a selection and that this calls into
question the assumption that players are always fully rational
in game-theoretic settings.)
14.22 Economists have proposed retaining the tax-exempt sta- tus of health care while lifting the requirement that employees
spend all of the tax-free fringe benefits solely on health care. If
retirees had a similar option to select such “medisave’’ accounts
versus retain their traditional Medicare coverage, explain why
average government payments to retirees opting to retain their
traditional Medicare coverage could end up rising.
14.23 In 2002, Allergan, the company making wrinkle-fighter Botox, put together a $50 million advertising campaign to
hook the masses on the anti-aging drug. Botox is derived from
the neurotoxin that causes botulism, a deadly food poisoning.
Injected under the skin, it relaxes facial muscles causing lines.
Botox was approved by the Food and Drug Administration in
1989 to treat crossed eyes and uncontrolled blinking. Word
spread, however, in the late 1990s regarding its ability to also
control wrinkles. A 15-minute doctor’s office treatment costs
around $500. Customers have to get repeated treatments since
the effects wear off in a few months. In light of this, evaluate
the statement made by a political activist stating that Botox
advertising should be banned by the federal government since
it promotes a product that has “no socially redeeming value
and [that] merely manipulates vulnerable consumers into buy-
ing a product with short-lived positive effects that they don’t
really need.”
14.24 In 1994, Congress passed legislation exempting aircraft from product liability claims in court if they were older than
18 years and had fewer than 20 seats. The probability that such
aircraft would be involved in an accident declined in the wake
of the passage of the legislation. Given what you have learned
in this chapter, explain why.
385
C15.INDD 10:48:36:AM 08/06/2014 PAGE 385Trim Size: 203.2 mm X 254 mm
CHAPTER 15 Using Noncompetitive Market Models
In Chapters 11 through 14, we examined a number of models in which firms have varying degrees of market power. We found that prices tend to be higher and output lower than
under competitive conditions. These models also provide analytical frameworks that can
be used to examine other issues relevant to the functioning of noncompetitive markets and
to understand certain pressing social problems. In this chapter we look at several examples,
including the magnitude of the deadweight loss associated with monopoly and the effect
of monopoly on innovation. We also examine how best to regulate monopolies that arise
because economies of scale characterize the production of a good over the relevant range of
market output. In such natural monopoly cases, average production cost is minimized if a single firm supplies the entire market.
15.1 The Size of the Deadweight Loss of Monopoly In Chapter 11, we explained how monopoly, at least from a static perspective, results in a
deadweight loss. Figure 15.1 illustrates this analysis for a market that would be a constant-
cost industry under competitive conditions. The competitive outcome is an output of Q (1,000) and a price of P ($1.00). If the industry becomes a pure monopoly and the monop- oly can produce under the same cost conditions (so the competitive supply curve becomes
the monopolist’s marginal cost and average cost curves), the monopoly outcome is an out-
put of QM (500) and a price of PM ($2.00). The deadweight loss of monopoly is shown as the triangular area BCA.
Learning objectives
Determine the relative magnitude of the deadweight loss of monopoly. Ascertain the extent to which, if any, monopolies suppress innovations. Explore whether government intervention can promote efficiency in the case of natural monopoly. Explore the concepts of iterated dominance and commitment in the context of game theory models.
Memorable Quote “I’m sorry that we have to have a Washington presence. We thrived during our first 16 years with- out any of this. I never made a political visit to Washington and we had no people there. It wasn’t on our radar screen. We were just making great software.”
—Bill Gates, co-founder of Microsoft
386 Using Noncompetit ive Market Models
C15.INDD 10:48:36:AM 08/06/2014 PAGE 386Trim Size: 203.2 mm X 254 mm
Determining the magnitude of this deadweight loss is important for public policy rea-
sons. This is so because, as we saw in Chapter 11, the deadweight loss indicates how much
lower total (consumer plus producer) surplus is in a market on account of monopoly. Pub-
lic policies promoting greater competition in such a market hence offer the potential for
increasing total surplus and attaining efficiency in output.
Note that the deadweight loss is given by the area of the triangle BCA (at least when we assume linear demand and supply curves), and the area of a triangle equals one-half its base
times its height. Thus, if we can determine the base and the height of the deadweight loss tri-
angle, we can calculate the magnitude. The height of the triangle, distance BA in Figure 15.1, is the excess of the monopoly price over marginal cost of production (PM − MC). The base of the triangle, distance AC, is the restriction in output due to the monopoly (Q − QM). This restriction in output can be calculated if we know price and marginal cost at the monopoly out-
come and the price elasticity of demand (assuming marginal cost is constant). For example,
the demand curve in Figure 15.1 has an arc elasticity of one between QM and Q, and because the monopoly price is 100 percent above the competitive price, we know the competitive out-
put is 100 percent above the monopoly output. (The ratio of the percentage change in output
and percentage change in price equals the elasticity.) Therefore, for the figures given in the
graph, we can calculate the deadweight loss as (½) (BA)(AC), or (½) ($1)(500), or $250. This is equal to one-fourth of the total consumer outlay on the product in our example.
Several economists have estimated the magnitude of the deadweight loss due to monopo-
listic restrictions in output in the real world. The first was Arnold Harberger, whose 1954
study concluded that the deadweight loss of monopoly in U.S. manufacturing corporations
equaled a scant 0.1 percent (or one-thousandth) of gross national product (GNP).1 (Manufac-
turing, however, accounted for only about one-fourth of GNP, so if monopoly is as important
in other sectors as in manufacturing, the economy-wide deadweight loss would have been
0.4 percent.) A number of more recent studies, using different data and methodologies, have
tended to support Harberger’s conclusion that the deadweight loss is not large relative to GNP.
The Deadweight Loss of Noncompetitive Output When price is PM and output QM, the deadweight loss due to the monopolistic restriction of output is triangular area BCA. If the product is produced by an oligopoly, output is likely to be higher (say, Q2) and the deadweight loss smaller (area HCV).
Dollars per unit
$2.00 = PM
$1.20 = P2
$1.50 = P1
$1.00 = P
0
E
B
H
D
A G V
MR
C
Output
(500)
Q
(1,000)
Q1
LS = MC = AC
(750)
Q2
(900)
QM
Figure 15.1
1Arnold Harberger, “Monopoly and Resource Allocation,” American Economic Review, 44, No. 2 (May 1954), pp. 77–87.
The Size of the Deadweight Loss of Monopoly 387
C15.INDD 10:48:36:AM 08/06/2014 PAGE 387Trim Size: 203.2 mm X 254 mm
A comprehensive survey of the research concludes that the deadweight loss of monopoly in
the United States lies somewhere between 0.5 and 4 percent of GNP.2
Why Are the Estimates of the Deadweight Loss Not Large? There are several reasons why estimates of the deadweight loss of monopoly in relation
to GNP are not large. One is that we are comparing the deadweight loss not to the size of
the monopolized sector, but to the size of the whole economy (GNP). In our example in
Figure 15.1, the deadweight loss is 25 percent of the total outlay in this market, but if only
20 percent of the economy is monopolized (to this degree), then the total deadweight loss
relative to GNP would be one-fifth of 25 percent, or 5 percent.
An even more important factor is that there are few, if any, instances of pure monopoly
in the United States. Most examples of noncompetitive markets involve markets dominated
by several large firms (that is, oligopolies) rather than pure monopolies. Studies of the
deadweight loss of monopoly are really examining monopoly power in industries such as
soft drink manufacturing, as opposed to sole-producer industries.
Although we do not have a single satisfactory theory of oligopoly, recall that in most
oligopoly models, the output is greater than the pure monopoly output. To see the impor-
tance of this point, suppose that output is actually three-fourths the competitive output in
Figure 15.1—because the industry is an oligopoly. Then output is Q1 (750) and the price is P1, or $1.50, because the demand curve is linear. The deadweight loss is then equal to area ECG, or (½)($0.50)(250), which equals $62.50. Note that although the output restriction in this case (250) is half as large as the output restriction under pure monopoly (500), the
deadweight loss is only one-fourth as large.
Of course, we cannot directly measure the restriction in output in any industry; we can
observe only actual output. What we can try to measure directly is the excess of price over
marginal cost—that is, the height of the deadweight loss triangle. This is what is done in
practice, and the P − MC estimates, along with estimated or assumed demand elastici- ties, allow us to infer the restriction in output. Unfortunately, it is far from straightforward
to estimate how much the actual price exceeds marginal cost. If we assume that average
cost and marginal cost are equal, as in our diagram, then the excess of price over marginal
cost is also the economic profit per unit of output. Thus, data on profits may tell us how
much higher than cost price is. But profit data always report accounting profits, and these
typically exceed the pure economic profits we wish to measure. Nonetheless, looking at
accounting profits data and making adjustments to try to estimate economic profits provide
interesting clues to the excess of price over marginal cost in different industries.
Considering all U.S. corporations, how much accounting profit do you think businesses
on average make per dollar of sales? The typical response from college students in Gal-
lup opinion surveys is 45 cents. The correct answer is actually 5 cents of after-tax profit
per dollar of sales. (In the same survey, students also indicated that they thought a “fair”
profit for corporations would be 25 cents per dollar of sales.) The 5 percent profit margin is
based on the accounting definition of profit; the economic profit margin tends to be smaller.
Of course, we are considering the average for all corporations, and for those with market
power the profit margin can be much greater. But it turns out to be very rare for any indus-
try to have a profit margin, even based on accounting data, as large as 20 percent. That is
not really so surprising given that a 20 percent profit margin would be four times the aver-
age and would constitute a powerful incentive for entry to occur in the market.
So, instead of basing our deadweight loss estimate on the output restriction, as we did in
assuming output was three-fourths the competitive level, let’s assume that in the industries
2Frederic M. Scherer and David Ross, Industrial Market Structure and Economic Performance, 3rd ed. (Boston: Houghton Mifflin, 1990), p. 667. The estimates rise to 3–4 percent if one allows for nonunitary demand elasticity. See Jonathan B. Baker, “The Case for Antitrust Enforcement,” Journal of Economic Perspectives, 17, No. 4 (Fall 2003), pp. 27–50.
388 Using Noncompetit ive Market Models
C15.INDD 10:48:36:AM 08/06/2014 PAGE 388Trim Size: 203.2 mm X 254 mm
with market power, price is 20 percent above marginal cost. This situation is shown in
Figure 15.1 by an output of Q2 and a price of P2, or an output of 900 and price of $1.20. The deadweight loss is area HCV, which equals (½)($0.20)(100), or $10. This is about 1 percent of the industry’s total revenue. However, the arc elasticity of demand between points H and C on the demand curve is only 0.58, and if we used the more commonly assumed value of 1 for our exercise, the deadweight loss would be nearly $20, or about 2 percent of the
industry’s total revenue. If one in five industries exercise this degree of monopoly power,
the economy-wide deadweight loss from monopoly would be 0.4 percent of GNP.
These considerations have convinced many economists that the deadweight loss of
monopolistic restrictions in output in the U.S. economy is almost certain to be less than
1 percent of GNP. Presumably, this is because most of the economy is reasonably com-
petitive, with a relatively small number of exceptions. But there may be other types of
deadweight loss associated with noncompetitive markets, as we will see next.
Other Possible Deadweight Losses of Monopoly We have emphasized two consequences of monopoly widely viewed as undesirable: the
restriction of output and the redistribution of income in favor of the owners of the monop-
oly. Only the restriction of output involves a net loss in total surplus, and both theory and
the available evidence suggest it is not very large compared with GNP. Other potential
effects of monopoly, however, should also be mentioned.
When comparing monopoly and competition, we assumed that the costs of production
under monopoly would be the same as under competition. If the monopoly maximizes
profit, this assumption is correct. For example, the monopoly would produce the 500 units
of output in Figure 15.1 at a cost of $500. If profit is to be maximized, whatever output the
firm produces must be produced at the lowest possible cost. The pressure on a monopoly to
minimize production cost, however, is not as strong as the pressure on a competitive firm.
If cost should increase because of slack cost controls in a competitive firm, the firm will
start losing money and eventually close operations. For the monopolist, though, higher-
than-necessary cost may just mean a smaller profit, not a loss. In the absence of competi-
tion with other firms, the monopolist is under less pressure to minimize cost. As Adam
Smith observed, “monopoly... is a great enemy to good management.”3
If production cost is unnecessarily high under monopoly, this is another deadweight loss.
For instance, if in Figure 15.1 the monopoly produces 500 units at a cost of $1.50 per unit,
the total cost of producing 500 units is therefore $250 higher under monopoly than it would
be under competition. Part of the potential monopoly profit of $500 would be dissipated
through the cost increase. Some of the $500 profit rectangle, PMBAP, no longer would be a transfer of income from consumers to the monopolist; instead, the money is absorbed by
higher-than-necessary cost and is a net loss.
A similar outcome results if the monopoly incurs costs in acquiring or maintaining its
market position. Our analysis implicitly assumed that production cost is the only cost of
the monopoly. But a monopoly may have to expend resources to ensure continuation of
its monopoly power. A lobbying effort may be necessary to secure favorable government
policies to block competition by other firms. Management may spend more time worry-
ing about protecting its market from encroachers than making business decisions regarding
output and cost. Legal and accounting staffs may be required to fend off antitrust suits by
the Justice Department. Because an average of seven years is needed to see an antitrust suit
through to conclusion, litigation can be quite costly for both the government and the firm.
For instance, in 1974 the Justice Department brought an antitrust suit against American
Telephone and Telegraph (AT&T). By 1981, AT&T estimated that it had spent $250 mil-
lion on the case—$25 million in direct legal costs, such as lawyers’ fees and briefs, and
3Adam Smith, The Wealth of Nations (New York: Modern Library, 1937), p. 147.
Do Monopol ies Suppress Invent ions? 389
C15.INDD 10:48:36:AM 08/06/2014 PAGE 389Trim Size: 203.2 mm X 254 mm
another $225 million on supporting paperwork—and that pretrial proceedings had involved
more than 40 million pages.4 In January 1982, as the trial was nearing its conclusion,
AT&T and the Justice Department settled the case out of court.
Due either to the absence of competitive pressures or to the expenses associated with
securing monopoly power, cost may be higher than necessary under monopoly. Conse-
quently, measures of the deadweight loss of monopoly based on the welfare triangle, which
considers only the output restriction, may underestimate monopoly’s true deadweight
loss. We must emphasize that the analysis does not imply that cost will be higher under monopoly, only that it may be. There is little current evidence to suggest how quantitatively important this other deadweight loss really is. Moreover, as discussed in Chapter 11, it is
important not to overlook the dynamic perspective on monopoly. While a monopolist may
face less pressure to minimize cost than a competitive firm, monopoly power may have
been acquired in the first place through discovering a way to build a better mousetrap or an
existing mousetrap at lower cost. From a dynamic perspective, therefore, monopolies may
be associated with a lower, rather than higher, production cost.
15.2 Do Monopolies Suppress Inventions? A bit of folklore is that firms sometimes suppress inventions that would benefit consumers.
One version of this idea is the belief that manufacturers design products to wear out quickly
(planned obsolescence) so that consumers periodically will have to replace them.
An economist would assess these beliefs by looking first at the internal consistency of
the argument. The basic premise is that a firm will make a larger profit by suppressing a
worthwhile invention than by marketing it. Under what conditions will this be true?
To avoid ambiguity, we’ll define a “worthwhile” invention as one that allows a firm to
produce a higher-quality product at an unchanged cost or to produce the same-quality product
at a lower cost. Suppression of such an invention would be unambiguously harmful.
Under competitive conditions a firm would never suppress a worthwhile invention. Sup-
pose that the invention permits the production of the same-quality product at a lower cost.
The first firm to introduce the process will have a lower production cost than its rivals, and
this guarantees a profit. Even if the invention cannot be patented, the firm can earn a profit
until other firms have had time to copy it.
What about the monopoly case? Let’s look at an example and see whether it’s likely a
monopoly will suppress a worthwhile invention. Suppose that the market for light bulbs
is monopolized and that the monopoly sells light bulbs that last for 1,000 hours. Then the
monopolist acquires an invention that permits production of bulbs that last for 10,000 hours
at the same unit cost as the 1,000-hour bulbs. Obviously, consumers would purchase many
fewer light bulbs per year if each one lasted 10 times as long. Does this mean that the
monopoly will make more money if it continues to sell the 1,000-hour light bulb and with-
holds the superior product?
The answer is no. To see why, suppose that consumers want 10,000 hours of light per
year. Initially, they purchase 10 1,000-hour bulbs at $1.00 each, involving a total outlay of
$10.00. If it costs the monopolist $0.50 to make each bulb, the firm makes a profit of $5.00
per consumer. Each consumer will be willing to pay at least $10.00 (more if convenience
counts) for one 10,000-hour bulb, because a 10,000-hour bulb yields the same light as 10
1,000-hour bulbs that together cost $10.00. The monopolist, however, can produce each
10,000-hour bulb for $0.50, so profit is $9.50 on the sale of one 10,000-hour bulb but only
$5.00 on the sale of 10 1,000-hour bulbs.
4“Out of the Quagmire,” Wall Street Journal, January 30, 1981, p. 1.
390 Using Noncompetit ive Market Models
C15.INDD 10:48:36:AM 08/06/2014 PAGE 390Trim Size: 203.2 mm X 254 mm
The Effect of Inventions on Output The foregoing example assumes that customers continue to purchase just enough light
bulbs for 10,000 hours of light in both cases. While such an assumption may not be valid,
the result that the monopolist will make more money by selling the superior light bulb con-
tinues to hold even when the assumption does not.5 A graphical analysis shows why. There
are two ways to proceed. One is to consider the demand curve for light bulbs but to recog-
nize that the demand curve for 10,000-hour light bulbs differs from the curve for 1,000-
hour bulbs. A simpler approach is to recognize that what consumers are really purchasing is
the services of light bulbs—that is, hours of lighting—and the demand curve defined in this
way does not shift. What changes when we switch from 1,000- to 10,000-hour bulbs is the
cost and price per hour of lighting, not the demand curve itself.
Figure 15.2 illustrates this latter approach. On the horizontal axis we measure kilohours
of lighting; each kilohour equals 1,000 hours, the service provided by each of the first type
of bulbs. For simplicity, average and marginal cost are assumed to be constant at $0.50 per
1,000-hour bulb (per kilohour).
The initial pre-invention equilibrium at Q1 involves 100 kilohours (100 1,000-hour bulbs) sold for $1.00 each. Each 1,000-hour bulb costs $0.50 to produce, so total profit is
$50. The invention of the 10,000-hour bulb, which the firm can produce at the same unit
cost ($0.50), means that the cost per kilohour falls to $0.05. Thus, the average cost curve if
the new light bulb is produced is AC′. Operating with this lower-cost curve, the monopolist can make more profit, and the new profit-maximizing output of kilohours (not bulbs) is
Q2. Price falls to $0.75 per kilohour. Profit rises from $50 to $105: the cost per kilohour is $0.05, and the price is $0.75, so the profit per kilohour is $0.70; $0.70 times 150 kilohours
Monopoly and Inventions If a monopoly can produce a 10,000-hour light bulb at the same cost as a 1,000-hour bulb, the invention effectively reduces the cost per kilohour of light from $0.50 to $0.05. The monopoly will make a larger profit by producing and selling the superior bulb.
Dollars per kilohour
$1.00 = P1
$0.75 = P2
$0.50
$0.05
0
D
MR
Kilohours of lighting
Q1
AC = MC
AC ′ = MC ′
(100)
Q2
(150)
Figure 15.2
5Our analysis, of course, ignores the subtle complication that a consumer may want some amount of hours of light per year (such as 15,000) that is not a whole-number multiple of 10,000, while light from the supe- rior bulbs must be purchased in 10,000-hour increments. Such a complication can still be addressed, how- ever, to the extent that a consumer wants light in more than one year. That is, any excess light from a superior bulb not used in one year can be employed in the next year. A consumer who wants 5,000 hours of light in each of two years thus could use one of the 10,000-hour bulbs over the course of those two years rather than five of the 1,000-hour bulbs in each of the two years.
Do Monopol ies Suppress Invent ions? 391
C15.INDD 10:48:36:AM 08/06/2014 PAGE 391Trim Size: 203.2 mm X 254 mm
yields a total profit of $105. Note that the new equilibrium corresponds to the sale of 15
10,000-hour bulbs at $7.50 each; fewer bulbs are sold.
This analysis suggests that a monopolist has no reason to suppress a worthwhile invention.
In fact, the reverse is true. We have examined the more difficult case of a higher-quality product
to show how we can analyze quality changes by focusing on product services (hours of lighting)
rather than the product itself. We reach the same conclusion for an invention that lowers the
cost of producing a product of unchanged quality. In that case the cost curves for an unchanged-
quality product shift downward due to the invention, implying more profit for a monopolist.
Because a monopolist can increase profit by marketing a worthwhile invention, economists
tend to be skeptical of allegations that businesses suppress them. As with many other gener-
alizations in economics, though, we can conceive of an exception, a case where a monopoly
would find it profitable to suppress an invention. For instance, suppose the firm would lose its
monopoly position by introducing the invention. Once the invention of the 10,000-hour light
bulb becomes public knowledge, other firms could produce and sell it for $0.50. The monop-
oly would then find itself in a competitive market. If the firm can patent the invention, how-
ever, the monopoly may be able to retain its monopoly position and market the invention.6
There are many instances of worthwhile inventions being marketed, both by competi-
tive firms and by firms with various degrees of monopoly power. Thus, the generalization
that profit incentives will lead to the introduction of worthwhile inventions seems reliable.
While there may be other deadweight losses associated with monopoly, widespread sup-
pression of inventions does not appear to be one of them.
APPLICATION 15.1
Kodak provides perhaps the archetypical case of the costs associated with being slow to develop innovations even though those innovations threaten to cannibalize existing sales. By being slow to move from film to a digital format, Kodak saw its overall imaging market share, profit- ability, and employment decline after the early 1980s. In 1982, for example, Kodak was the dominant worldwide producer of film, operated the largest industrial complex in the northeastern United States, and employed more than 62,000 workers at its headquarters site in Roch- ester, New York. In 2012 Kodak filed for bankruptcy, and its Rochester-area employment level had fallen to 7,100 workers and was predicted to fall even further over the coming years. It wasn’t until the late 1990s that Kodak entered the digital camera market in a significant way, despite the fact that the first digital camera was actually invented by a Kodak engineer, Steve Sasson, in 1975. The market value of San Jose, California–based Adobe has sig- nificantly exceeded that of Kodak over the past decade, even though most of Adobe’s digital-photography-based products arguably could have been part of Kodak’s portfo- lio had Kodak moved more quickly into the digital sphere.
The Cost of Not Cannibalizing
A similar phenomenon has been at work in the book market, as digital readership grows. Amazon, which intro- duced the Kindle digital book reader in 2007, saw its share of the overall book market (print plus digital) increase from 15 percent in 2007 to nearly 30 percent by 2013. By con- trast, Borders, which had led Amazon in overall book sales through 2006 but failed to introduce a digital reader and stuck instead to its traditional print and bookstore model, saw its market share fall steadily over the 2007–2011 period and was forced to file for bankruptcy in early 2011. Even though Amazon’s entry into the digital book market doubt- lessly cannibalized some of its existing print sales, had Amazon not made such a move, competitors such as Apple (through the iPad), Sony, and Barnes & Noble (through the Nook) would have cannibalized Amazon’s print sales instead.
As the Kodak and Borders cases attest, firms with seem- ing monopoly power in their respective markets suppress innovations at their own long-run risk when technology is prone to change. Not cultivating the mind-set of cannibal- izing one’s own existing sales through technologically supe- rior products carries with it the prospect that rival firms more open to newer technologies will do the cannibalizing instead.
6This is a long-run analysis. In the short run, firms, whether they are monopolistic or competitive, may not introduce an invention immediately. When the time comes to replace worn-out equipment, however (and that time will come more quickly when a lower-cost process is available), the firm will introduce the invention.
392 Using Noncompetit ive Market Models
C15.INDD 10:48:36:AM 08/06/2014 PAGE 392Trim Size: 203.2 mm X 254 mm
15.3 Natural Monopoly In some cases monopoly results because one large firm can produce at a lower per-unit cost
than several smaller firms together accounting for the same total output. If the production
technology is such that economies of scale (declining average cost per unit of output)
extend to very high output levels, a large firm can undersell small firms, and one or a few
large firms will eventually dominate the industry. The extreme case is one in which the
average cost of a single enterprise declines over the entire range of market demand. As
mentioned earlier, this is called a natural monopoly. Natural monopoly presents a challenging public policy dilemma. On the one hand,
it implies that efficiency in production will be better served if a single firm supplies the
entire market. On the other hand, natural monopoly results in the absence of any firms that
actively compete with the monopolist. The monopolist thus will be tempted to exploit its
natural monopoly power and to restrict output and raise price. And inefficiency in output (a
deadweight loss) will occur if the monopolist takes these actions to increase its profit.
Figure 15.3 illustrates the natural monopoly case. Graphically, a natural monopoly exists
when the long-run average cost curve of a single firm is still declining at the point where it
intersects the total market demand curve for the product—at point A in the diagram. One firm can produce an output of Q2 at an average cost of AQ2. In this situation, the market, if unregulated, will be dominated by a single firm. If, instead, there are several small firms,
each producing Q1, for example, price will have to be at least P1. Yet any one firm could expand output, sell at a lower price, and ultimately drive the smaller firms out. Monopoly is
the “natural” result. Moreover, forcing a competitive structure on this market is undesirable
in terms of attaining efficiency in production. The real cost of serving the market will be
higher than necessary if there are several small, high-cost firms.
Drawing cost curves that imply a natural monopoly is easy, but the key question is
whether natural monopoly is prevalent in the real world. In fact, natural monopoly con-
ditions are not common, but they do exist for several products. Economists believe, for
example, that natural monopoly characterizes the provision of electricity, water, natural
gas, telephone services, and possibly cable television to specific geographic localities.
Consider electricity, providing it requires that homes be physically connected to the gen-
erating facility through underground or overhead lines. If several separate firms served
homes in a given community, each firm would have to run its own connecting power
lines. The cost of duplicating connecting lines (implying higher average costs) could be
avoided by using just one set of lines. This situation is depicted in Figure 15.3. Unit costs
are higher when several firms supply a few homes (Q1) than when one operation provides electricity to all homes (Q2).
When natural monopoly conditions exist, there are four ways public policymakers
can deal with the situation. One is to leave the market alone. In this case a monopoly
will result, and the monopoly will not choose to supply Q2 at a cost of AQ2 per unit (in Figure 15.3). Instead, it will choose the profit-maximizing output of QM with a price of PM. The second option is to permit a monopoly to operate but to regulate its activities. The third option is to have government ownership and operation of the facility (the U.S.
Postal Service, for example). A fourth option involves a government-sponsored competi-
tion for the right to operate a natural monopoly. Ideally, the operating right is awarded to
the bidder promising to charge the lowest price. Competition for the right to be the sole
supplier can serve to promote efficiency in output even though once the award is made,
there is only one supplier.
In the United States, the regulatory option has generally been pursued. A privately
owned firm is given the legal right to be the monopoly provider, but a public agency is cre-
ated to regulate the firm’s behavior.
natural monopoly an industry in which production cost is minimized if one firm supplies the entire output
Natural Monopoly 393
C15.INDD 10:48:36:AM 08/06/2014 PAGE 393Trim Size: 203.2 mm X 254 mm
Regulation of Natural Monopoly: Theory The public agencies charged with regulating natural monopolies, usually called public utili-
ties, generally set the prices that may be charged. Before investigating how this is accom-
plished, let’s examine the economic principles behind the price-setting approach.
In Figure 15.4, the natural monopoly’s average and marginal cost curves are AC and MC (ignore AC′ for the moment). If we have complete knowledge of cost and demand condi- tions, two logical prices can be set. One is the price at the level where the average cost
curve intersects the demand curve, a price of P1. This solution is called average-cost pric- ing. If the monopoly produces Q1, the price of P1 just covers its average cost, implying zero
Natural Monopoly When the average cost of producing a good declines over the entire range of market demand, a natural monopoly exists. It is less expensive for one firm to produce the entire market output than for several small firms to share the market. One firm can produce Q2 at a unit cost of AQ2, which is less than the cost when several firms each produce Q1 at a unit cost of P1. However, if the firm is allowed to produce monopolistically, output will be QM and price PM.
Dollars per unit
P1 PM
0
D
A
MR
OutputQ1
AC MC
QM Q2
Figure 15.3
Regulation of Natural Monopoly If price is set equal to MC at P2, the monopoly cannot cover its production cost. If price is set equal to AC at P1, output will be inefficient, the monopoly may have little incentive to minimize cost, and the average cost curve could rise to AC′.
Dollars per unit
P1 P2
0
D
A
B
C
MR
OutputQ1
ACAC′
Q M Q2
MC
Figure 15.4
394 Using Noncompetit ive Market Models
C15.INDD 10:48:36:AM 08/06/2014 PAGE 394Trim Size: 203.2 mm X 254 mm
economic profit. Moreover, the monopoly has an incentive to produce Q1 if a maximum price of P1 can be charged. As explained in Chapter 11, in the case of a price ceiling, the demand curve facing the monopoly becomes P1AD, so marginal revenue (equal to P1 up to an output of Q1) exceeds marginal cost as output expands to Q1, but MR drops below MC at higher output levels. Indeed, when a price of P1 is set, any output other than Q1 yields a loss, because average cost is above P1 at lower rates of output.
At an output of Q1, price exceeds marginal cost. (Because average cost is falling at point A, marginal cost must be below it.) Thus, consumers value additional units of output at more than they cost to produce, which suggests a second option—to set price at the level
where the marginal cost curve intersects the demand curve (point B). This option is called marginal-cost pricing. There is, however, a major obstacle to marginal-cost pricing: if we set price at P2, the monopoly incurs a loss. Because marginal cost is below average cost at Q2, setting price equal to marginal cost will put the firm out of business. A subsidy can be used to enable the firm to produce Q2 at a price of P2, but the cost of implementing and financing the subsidy generally makes this solution impractical.
Thus, the most practical alternative seems to be average-cost pricing. Output is greater
than the unregulated monopoly output of QM, and because expansion of output from QM to Q1 provides more benefits to consumers than the additional production costs, there is an efficiency gain. (Said another way, part of the deadweight loss arising from restricted out-
put by an unregulated monopoly is eliminated.) The price to consumers is lower than under
unregulated monopoly, and the monopoly’s owners receive no profit.
Regulation of Natural Monopoly: Practice In practice, regulators do not have complete knowledge of cost and demand conditions. They
generally attempt to attain the average-cost-pricing outcome by focusing on the rate of return
on invested capital (accounting profit) earned by a monopoly. It works this way: if the real-
ized rate of return is higher than what is thought to be a normal return (suggesting economic
profit), then the current price must be above average cost, and the result signals regulators to
reduce the price. Conversely, if the realized rate of return is lower than normal (suggesting
economic losses), regulators raise the allowed price. Proceeding in this trial-and-error fashion,
regulators locate the price at which profit is normal—that is, where price equals average cost.
There are several problems with this approach, but perhaps the most serious is that it dimin-
ishes the monopolist’s incentive to minimize cost. If cost rises, regulators permit a higher price
so that the monopoly still earns a normal rate of return. Thus, managers have an incentive to pad
expense accounts, pay themselves and their colleagues higher-than-necessary wages, and incur
numerous other costs that would normally be avoided because they cut into profit. Unnecessary
costs will not reduce profit if the regulatory agency permits a price increase to cover the costs.
Figure 15.4 also illustrates the consequences of this behavior. The AC curve continues to show minimum unit production cost, but cost padding shifts the actual cost curve to AC′. Because losses would occur at a price of P1, regulators grant a price increase to cover the higher costs (point C). Most regulatory agencies recognize the perverse incentive of the regulation, and to overcome it, they frequently become involved in monitoring the costs of
the monopoly. However, to determine the need for a particular cost is not easy, so average
cost probably drifts upward to some degree.
This form of regulation may also lead the monopoly to suppress or slow down the intro-
duction of inventions, which would not occur in an unregulated environment. The slow-
ness with which AT&T introduced automated switching equipment is a good example.
Automatic panel switches to replace operators were invented in the 1920s, but not until
50 years later, in the mid-1970s, did AT&T replace the old switches—even though the
automatic switches permitted a greater number of connections and more rapid switching
between them at a much lower cost and with much simpler maintenance. Recent advances
in switching equipment, primarily digital technology, have produced further speed and cost
Natural Monopoly 395
C15.INDD 10:48:36:AM 08/06/2014 PAGE 395Trim Size: 203.2 mm X 254 mm
economies, permitting additional services such as call waiting, call forwarding, and interna-
tional direct dialing. AT&T did not convert all exchanges to digital switching until the turn
of the century, however, so even though the relevant technology existed, some customers
were unable to purchase these services for many years.
The slowness with which regulated monopolies introduce new products and technology
may be a natural response to a price ceiling. If a monopoly discovers a cost-saving tech-
nology, it is unable to keep the increased profit because regulators will in turn reduce its
rates. Similarly, the monopoly has reduced incentive to engage in research and develop-
ment activities designed to decrease costs. Further, it is under no competitive pressure to
offer new services quickly because its customers are unlikely to have a better alternative.
For these reasons, economists have become increasingly critical of the regulation of nat-
ural monopolies. One famous study compared electric rates in regulated and unregulated
states between 1912 and 1937—before all states regulated rates—and found no difference
in the rates charged.7 However, the alternatives to regulation when natural monopoly con-
ditions prevail—alternatives such as unregulated monopoly and government ownership—
may not be particularly attractive either. Thus, there may be no completely satisfactory
solution to the natural monopoly problem.
APPLICATION 15.2
An alternative to rate regulation in the case of natu- ral monopoly is public ownership. This is the approach adopted in the case of mail collection and delivery in the United States. It is assumed that the average cost of service will be lower if there is a single designated producer—the U.S. Postal Service (USPS)—and that public ownership of production can be relied upon to ensure that prices equal average cost.8
Public ownership, however, is associated with some notable drawbacks. For example, because the objective of a public enterprise is to ensure that price equals average cost instead of profit maximization, the incentive to innovate and/or to encourage cost minimization is attenuated. The managers of a public enterprise generally cannot benefit from introducing an innovative product and/or cost-saving technology since the profit of the enterprise is constrained to equal zero. This situation is in marked contrast to an unregulated, for-profit setting, where such improvements can translate into a healthier bottom line.
The absence of a profit motive likely explains why the USPS was slow to permit credit card payments, increase its hours of operation, close smaller offices around the country that are holdovers from the time when postal jobs were the principal form of patronage that elected federal policymakers could offer to their supporters, and offer ancillary products (e.g., the packaging services sup- plied by for-profit competitors such as Mail Boxes Etc.).
Regulating Natural Monopoly through Public Ownership: The Case of the USPS
The nonprofit constraint also suggests why USPS was not the originator of overnight package delivery service (an innovation introduced by for-profit Federal Express) and has been slow to respond to competition from other forms of communication, such as e-mail. Finally, while economies of scale may be present in the collection and delivery of mail service, enshrining one publicly owned firm to provide the service at cost may result in costs not being minimized. Indeed, the USPS’s own internal Postal Inspection Service indicates that on the 10 percent of all rural routes that are contracted out to private companies, the quality of service is greater and the cost is significantly lower (by over 50 percent) than it would be if the USPS provided the service. The difference in the cost of service between contracted- out and noncontracted-out routes is by and large due to lower wages and fringe benefits.
The USPS reported a record loss of $15.9 billion for its fiscal year 2012 and warned that it would go broke without further help from Congress. The loss reflected ongoing declines in revenue from first-class mail due to ever-growing competition from other forms of communi- cation, such as e-mail, cell phones, and texting. The loss also stemmed from $11.1 billion in prepayment of USPS retiree health care benefits.
8This application is based on Mark A. Zupan, “Let the Market Deliver the Mail,” New York Times, August 7, 1993, p. 11.
7George J. Stigler and Claire Friedland, “What Can Regulators Regulate? The Case of Electricity,” Journal of Law and Economics, 5, No. 2 (October 1962), pp. 1–16.
396 Using Noncompetit ive Market Models
C15.INDD 10:48:36:AM 08/06/2014 PAGE 396Trim Size: 203.2 mm X 254 mm
15.4 More on Game Theory: Iterated Dominance and Commitment
In Chapter 14, we introduced the subject of game theory and showed how it related to
the study of noncompetitive markets. In all of the applications we covered in Chapter 14,
at least one of the two players had a dominant strategy. We were therefore able to easily
determine the equilibrium—either a dominant-strategy equilibrium (as in the case of the
prisoner’s dilemma game where both players had a dominant strategy of cheating/confess-
ing) or the Nash equilibrium of Table 14.2, where only one of the two players had a domi-
nant strategy.
What if neither player has a dominant strategy? Can a Nash equilibrium still emerge?
The answer is “yes.” In this section, we will show one way in which this can happen
through the concept of iterated dominance. We will also examine the possibility that a player can make a commitment that alters the relevant payoff matrix in such a manner that a different equilibrium will emerge—an equilibrium that is more favorable to the player mak-
ing the commitment. We thus can see that the usefulness of game theory is not limited to
situations where at least one player has a dominant strategy and that a player may have the
ability to take an action that affects the strategies selected by other players.
Iterated Dominance Suppose that only two companies, Walmart and Best Buy, compete in the consumer elec-
tronics market. The two firms each need to select one of three strategies regarding their
products’ prices: high, medium or low. The relevant payoff matrix, representing each firm’s
profit based on the two firms’ selected strategies, is depicted in Table 15.1.
What equilibrium will emerge given the Table 15.1 payoff matrix? Clearly, neither com-
pany has a dominant strategy. If Walmart chooses a high price, Best Buy’s best strategy is a
medium price (with a profit of 105). If Walmart chooses a low price, Best Buy’s best strategy
is a high price (a profit of 11 versus a profit of 7 associated with choosing a medium price).
Thus, Best Buy does not have a best strategy irrespective of Walmart’s strategy.
Likewise, Walmart does not have a best strategy irrespective of Best Buy’s strategy. For
example, if Best Buy chooses a high price, Walmart’s best strategy is a medium price (with
a profit of 105). If Best Buy chooses a low price, Walmart’s best strategy is a high price (a
profit of 11 versus a profit of 7 associated with choosing a medium price).
A More Complex Game
90
High Price
High Price
Medium Price
Best Buy
Walmart
90
44
105
Low Price
11
97
105
Medium Price
44
50
50
7
40
97
Low Price
11
40
7
–10
–10
Table 15.1
More on Game Theory: I terated Dominance and Commitment 397
C15.INDD 10:48:36:AM 08/06/2014 PAGE 397Trim Size: 203.2 mm X 254 mm
With neither player having a dominant strategy in Table 15.1, we cannot as readily derive
the equilibrium as we did in the games examined in Chapter 14. It turns out, however, that
there is a Nash equilibrium associated with the Table 15.1 payoff matrix. To determine this
equilibrium we need to rely on the concept of iterated dominance: ruling out any strategy that is inferior to, or dominated by, another strategy. That is, if a certain strategy yields lower payoffs for Walmart than another strategy irrespective of the strategy selected by Best Buy,
Walmart would never select such a strategy. The strategy that is dominated by another strat-
egy thus can be effectively eliminated from Walmart’s menu of strategic possibilities and the
dimensions of the relevant payoff matrix thereby reduced. Whenever Walmart has a strategy
that is dominated by another strategy, eliminating the dominated strategy effectively reduces
the number of rows in the Table 15.1 payoff matrix. Analogously, when Best Buy faces a
strategy that is dominated by another strategy, eliminating Best Buy’s dominated strategy
effectively reduces the number of columns in the Table 15.1 payoff matrix.
Consider the low-price strategy for Walmart. Irrespective of Best Buy’s pricing strategy,
a medium price consistently yields a higher payoff for Walmart than does the low-price
strategy. If Best Buy selects the high column, a medium price for Walmart yields a payoff
of 105 versus a payoff of 97 associated with a low price. If Best Buy opts for the medium
column, a medium price for Walmart yields a payoff of 50 versus a payoff of 40 associ-
ated with a low price. And, if Best Buy chooses the low price, a medium price remains a
better choice for Walmart than a low price (a payoff of 7 versus −10). Since Walmart is never better off choosing the low price, the low row can effectively be eliminated from the
Table 15.1 payoff matrix: it would never be selected by Walmart in favor of the medium
strategy.
Reasoning in a similar manner allows us to eliminate the low column for Best Buy.
This strategy yields consistently lower payoffs than the medium option for Best Buy,
regardless of Walmart’s pricing strategy. If Walmart selects the high row, a medium price
for Best Buy yields a payoff of 105 versus a payoff of 97 associated with a low price. If
Walmart opts for the medium row, the medium option for Best Buy yields a payoff of 50
versus a payoff of 40 associated with a low price. And, if Walmart chooses a low price,
the medium option remains a better choice for Best Buy than the low price (a payoff of 7
versus −10). Once both the low row for Walmart and the low column for Best Buy are eliminated
from consideration, the three-by-three dimensional Table 15.1 matrix is reduced to the two-
by-two Table 15.2 matrix: both players have only two strategies from which to choose.
In the two-by-two payoff matrix of Table 15.2, both players have a dominant strategy.
Walmart’s payoffs associated with the medium price are always greater than those associated
with the high price (105 versus 90 if Best Buy opts for the high-price strategy, 50 versus 44 if
Best Buy selects the medium price). Best Buy’s payoffs associated with a medium price are
also always greater than those associated with a high price (105 versus 90 if Walmart opts
for the high row, 50 versus 44 if Walmart selects the medium row). Consequently, once we
take the steps—or iterations—of eliminating players’ dominated strategies, we end up with
the prediction that the equilibrium associated with the Table 15.1 game will be both players
selecting the medium-price strategy and, as a result, receiving a payoff of 50.
The predicted equilibrium for the Table 15.1 game is a Nash equilibrium. That is, each
player’s choice is the best one given the strategy chosen by the other player. Specifically,
if, through the concept of iterated dominance, Best Buy finds that a medium price is the
best strategy, Walmart’s best choice is the medium option: it generates the highest payoff
for Walmart when Best Buy chooses the medium column. Likewise, if Walmart opts for a
medium price, the medium option is Best Buy’s best choice: it generates the highest payoff
for Best Buy when Walmart chooses the medium row.
By eliminating dominated strategies, therefore, the concept of iterated dominance
allows us to predict what equilibrium will emerge in more complex games, such as the one
iterated dominance the concept of eliminating any strategy that is inferior to or dominated by another strategy
398 Using Noncompetit ive Market Models
C15.INDD 10:48:36:AM 08/06/2014 PAGE 398Trim Size: 203.2 mm X 254 mm
depicted in Table 15.1. Even though neither of the players in Table 15.1 has a dominant
strategy, a Nash equilibrium can be shown to exist once clearly inferior strategies are elimi-
nated from consideration.
Commitment The payoff matrix in Table 15.2 is another example of a prisoner’s dilemma. Once the domi-
nated low-price strategy of Table 15.1 is eliminated for both players, each player’s dominant
strategy is a medium price and results in a payoff of 50. The predicted outcome is one where
both players are worse off than they would be if they both chose a high price and earned a
payoff of 90. While choosing the medium option is in the self-interest of both players, the col-
lective outcome of each player pursuing their self-interest is inferior for both.
Apart from the ways discussed in Chapter 14 by which players confronting a prisoner’s
dilemma game might overcome the problem and realize the best all-around outcome
(through repeated games, altruism, and so on), there is another possible mechanism through
which a player can ensure that acting on the basis of self-interest results in the maximum
feasible payoff. Specifically, a player might find it desirable to make a commitment to a particular course of action and, by constraining one’s choice of strategies, increase the
player’s equilibrium payoff.
It may seem paradoxical that constraining the set of strategic choices can generate a
higher payoff. To see why this may be so, consider the case where both Walmart and Best
Buy vow that “they will not be undersold” in the context of Table 15.2. That is, if Walmart
chooses its medium row, Best Buy will not opt for a high-price strategy. Likewise, if Best
Buy chooses a medium price, Walmart will not select a high-price strategy. The commit-
ment to not be underpriced on the part of both players reduces the number of possible
outcomes in this game from four to two: it effectively eliminates the northeast cell where
Walmart is underpriced by Best Buy and the southwest cell where Best Buy is underpriced
by Walmart. As depicted in Table 15.3, the only two possible outcomes that remain are
if both players simultaneously opt for a high or low price. Thus, we can predict that both
players will choose the high-price strategy and earn a profit of 90.
Note what the commitment to “not be undersold” on the part of Walmart and Best Buy
has accomplished. While appearing to promote competition in pricing, it allows the two
players to overcome the prisoner’s dilemma they previously confronted and results in con-
sumers being charged the highest possible price to the benefit of the two sellers.
Of course, to be effective, a commitment must be credible. For example, Best Buy’s vow
not to be undersold implies that if Walmart chooses a medium price in Table 15.3, Best Buy
will not choose a high price, and vice versa for Walmart if Best Buy chooses the medium
option. To make their commitments credible, each player may promise to match or even
beat (by, say, $25) the best price a customer can obtain from the rival seller. In this manner,
Best Buy and Walmart effectively use customers to bind them to their commitments to not
commitment the strategy of adopting a particular course of action, constraining one’s choice of strategies, in order to increase your equilibrium payoff
Eliminating Dominated Strategies
90
High Price
High Price
Medium Price
Medium Price
Best Buy
Walmart 90
44
105
105
44
50
50
Table 15.2
More on Game Theory: I terated Dominance and Commitment 399
C15.INDD 10:48:36:AM 08/06/2014 PAGE 399Trim Size: 203.2 mm X 254 mm
be undersold—customers who, in the course of enforcing Best Buy’s and Walmart’s com-
mitments, end up promoting a high-price equilibrium.
Commitments need not be limited to vows to not be undersold. In many game-theoretic
situations, a player can take other actions to alter the relevant payoff matrix so that it will
be in the player’s interest to follow through on a particular strategy. For example, to deter
a political rival from entering a race, a politician may build up a substantial war chest—a
money reserve that effectively binds the politician to competing vigorously to retain his or
her elected post. On overnight cross-country flights that are half-empty, after the airplane
door is closed passengers usually scramble to sit in the middle seat of any empty row of
three seats. Committing oneself to the middle seat diminishes the chances that a fellow pas-
senger will choose either of the two empty seats next to you and increases the likelihood
that you will have an empty row of seats in which to stretch out and sleep once the plane
reaches a comfortable cruising altitude. As another example, Delco’s shutting down of a
production line that produces spark plugs specifically tailored for the Ford Motor Company
may convince General Motors to divert more of its spark plug purchases to Delco because
Delco can devote its attention to better satisfying General Motor’s needs.
The Role of Commitment
90
High Price
High Price
Medium Price
Medium Price
Best Buy
Walmart 90
50
50
Table 15.3
APPLICATION 15.3
Invading Mexico in the sixteenth century, Spanish explorer Cortés ordered the fleet of ships that had carried his army to the New World burned. At first blush, his order may appear to have been an act of madness, given that the Spanish army was vastly outnumbered. However, restricting the
Why It May Be Wise to Burn Your Ships
menu of strategic choices available to them committed Cortés’s men to fighting. Not fighting and returning to Spain was no longer an option once the ships were burned. And upon being committed to fighting, Cortés’s army success- fully accomplished its objective of conquest.
SUMMARY
Several examples illustrate the functioning of non-
competitive markets, the deadweight loss associated
with monopoly, and the effect of monopoly on inno-
vation.
Public policies promoting greater competition in
a monopoly market offer the potential for increasing
total surplus and attaining efficiency in output. Thus, it
is important to determine the relative magnitude of the
deadweight loss of monopoly.
If the product is produced by an oligopoly, output is
likely to be higher and the deadweight loss smaller than
in the case of monopoly.
In addition to restriction of output, other deadweight
losses may occur in cases of monopoly. For instance,
400 Using Noncompetit ive Market Models
C15.INDD 10:48:36:AM 08/06/2014 PAGE 400Trim Size: 203.2 mm X 254 mm
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
15.1 Explain why a certain triangular area is a measure of the deadweight loss of monopoly. What information do you
require in order to calculate the size of this triangle?
*15.2 In an oligopolistic industry with constant marginal cost, output is 20 percent lower and price is 20 percent higher than
competitive levels. How large is the deadweight loss as a per-
centage of the total consumer outlay on the product?
15.3 Studies have concluded that the deadweight loss of monopoly power in the United States is less than 0.5 percent of
GNP. From your knowledge of the determinants of the dead-
weight loss, explain why such a small figure is plausible.
15.4 Suppose that the government levied a lump-sum tax on a monopolist. How would such a tax affect the monopolist’s
pricing and output decisions and profit?
15.5 Compare the effects of a $1-per-unit excise subsidy when applied to a monopoly and to a competitive industry with the
same cost and demand conditions. In which case will price fall
more? In which case will output increase more?
15.6 “If a business sells a product that wears out in a month, you will have to buy 12 a year, and the business will make 12
times as much money as it would selling a product that lasts a
year.” Evaluate this statement. Why don’t businesses sell prod-
ucts that wear out in a day? In an hour?
15.7 Businesses frequently own patents on a number of prod- ucts they do not produce and sell. This is sometimes cited as
evidence that businesses suppress inventions. Is it?
15.8 Explain what natural monopoly is in terms of the relation- ship between cost curves and the demand curve. If the market
is left to itself, what price and output will result?
15.9 Use a diagram to illustrate the “hoped for” result of nat- ural monopoly regulation that attempts to set a price equal to
average cost. What are the difficulties in achieving this out-
come? Would an unregulated natural monopoly be preferable
to a regulated natural monopoly?
15.10 In Table 15.2, if only Best Buy commits to not being undersold, what will be the outcome?
15.11 The manufacturer of a drug that has had a monopoly, due to patent protection, commits to pricing at cost and ensur-
ing that no firm in the market will make a profit should a rival
manufacturer enter the market once the drug’s patent wears off.
Is such a commitment credible? Explain.
15.12 Two companies each own property (and mineral rights) in an oil field. Each firm therefore has the legal right to drill for
oil on its land and take out as much oil as it can. The problem,
of course, is that one company’s actions affect how much oil
the other can produce.
The following matrix represents how each of these compa-
nies views the situation. The terms outside the matrix repre-
sent oil output by each firm (low, medium or high), while the
numbers in each cell show the present value of all oil to be
extracted by each company, given the two extraction policies.
The first number represents the value to Company A, and the
second number represents the value to Company B.
As an example, if Company A pumps at a “low” rate and
Company B pumps at a “low” rate, then the value to Company
A of all the oil it expects to take over the life of the field is
$100 while the value to Company B of its oil is $8.
with the absence of competition with other firms, the
monopolist may be under less pressure to minimize costs
of production or may incur costs in acquiring or main-
taining its market position.
Suppression of inventions does not appear to be a
result of monopoly because a monopolist can increase
profit by marketing a worthwhile invention.
Natural monopoly exists when the average cost of a
single enterprise declines over the entire range of market
demand. This situation implies that the monopolist can serve
the entire market more efficiently than many small firms
but leads to the possibility that the firm will be tempted to
exploit its power and restrict output and raise price.
There are four policy options in dealing with natural
monopolies: leave the market alone; regulate the monop-
oly’s activities; allow the government to own and oper-
ate the facility; or sponsor a competition for the right
to operate a natural monopoly, which goes to the bid-
der promising to charge the lowest price. In the United
States, the second option has generally been chosen.
In principle, average-cost pricing offers the most
practical alternative for public agencies that regulate the
behavior of natural monopolies.
In practice, regulators lack complete knowledge of cost
and demand conditions and seek to promote average-cost
pricing by focusing on the rate of return on invested capital.
The problem with this approach is that it diminishes the
firm’s incentive to minimize cost and to innovate.
In a game-theory setting, iterated dominance allows
us to assume that a firm will rule out any strategy that is
inferior to, or dominated by, another strategy.
Commitment to a particular course of action is
another strategy that can increase one’s equilibrium pay-
off in a prisoner’s dilemma situation.
Review Quest ions and Problems 401
C15.INDD 10:48:36:AM 08/06/2014 PAGE 401Trim Size: 203.2 mm X 254 mm
(B and C) Total Number of Competitors’ Flights
5 6 7 8 9 10 11 12 13 14 15 16 17 18
A’s Own Number of Flights
2 88.6 72.5 60.0 50.0 41.8 35.0 29.2 24.3 20.0 16.3 12.9 10.0 7.4 5.0
3 108.8 90.0 75.0 62.7 52.5 43.8 36.4 30.0 24.4 19.4 15.0 11.1 7.5 4.3
4 120.0 100.0 83.6 70.0 58.5 48.6 40.0 32.5 25.9 20.0 14.7 10.0 5.7 1.8
5 125.0 104.5 87.5 73.1 60.7 50.0 40.6 32.4 25.0 18.4 12.5 7.1 2.3 −2.2 6 125.5 105.0 87.7 72.9 60.0 48.8 38.8 30.0 22.1 15.0 8.6 2.7 −2.6 −7.5 7 122.6 102.3 85.0 70.0 56.9 45.3 35.0 25.8 17.5 10.0 3.2 −3.0 −8.8 −14.0 8 116.9 97.1 80.0 65.0 51.8 40.0 29.5 20.0 11.4 3.6 –3.5 −10.0 −16.0 −21.5
a. What extraction rates maximize the total value of the oil field? b. Do the extraction rates maximizing the value of the field
represent a stable situation? Explain.
c. Is there a dominant strategy (extraction rate) for either or both players? Explain.
d. Is there a Nash equilibrium set of extraction rates? If so, does it maximize the total value of the oil field?
e. Is there a mutually beneficial exchange inherent in this matrix—one that could solve the problem these two com-
panies face? If Company A were to purchase Company B’s
oil rights, how much would it have to pay? Is this a feasible
transaction?
15.13 If the latest computer chip produced by Intel has twice the storage capacity as the previous-generation chip, Intel
would find it advantageous to market the new chip even though
its sales of the old chip would plummet. True or false? Explain
why. Would your answer change if Intel operated in a fully
competitive market versus having monopoly power in the sup-
ply of computer chips?
15.14 Some have argued that the distribution of cable televi- sion service in a community is subject to economies of scale.
Namely, it is cheaper to have just one company supply every
household in the community with the service than to have sev-
eral providers, each having to string separate cables throughout
the community and each having to have its own satellite down-
load facilities. On account of this apparent natural monopoly,
communities employ franchise bidding to regulate local cable
companies. Companies interested in supplying service to a
community are required to bid ex ante for the right to be the sole supplier ex post. Explain why such franchise bidding com- petitions can serve to promote efficiency in markets character-
ized by natural monopoly.
15.15 Microsoft spends over $2 million per year supporting a wide variety of political candidates. To what extent does such
spending reflect a deadweight loss? Explain.
15.16 Economist Bill Samuelson suggests a problem center- ing around three air carriers competing for passengers on a
given city-pair route. Namely, the fare that can be charged
on the route is fixed at $225, while the size of the market
is fixed at 2,000 passengers per day. There are three com-
peting airlines: A, B, and C. Each airline gets passengers in
proportion to its share of total flights. For example, if all
three airlines offered the same number of flights, then they
would each get one-third of the passengers. If Airline A
offered six flights and B and C each offered three, then A
would get 50 percent of the market, while B and C would
get 25 percent each. Each plane holds a maximum of 300
passengers. Each plane trip costs $20,000, whether the plane
is full or not.
a. Confirm firm A’s profit equals $450,000[a/(a + b + c)] – $20,000a, where a, b, and c represent the number of flights by firms A, B, and C, respectively.
b. Confirm to yourself that the table below gives the profits to A as a function of its flights and its competitors’ flights per
day.
c. Consider a strategy for any one of the firms to be a policy of flying a certain number of flights per day. Is there a domi-
nant strategy for A—that is, a number of flights that gives
higher profits no matter what the competitors do?
d. Is there a Nash equilibrium in this game? That is, is there a set of strategies (numbers of flights a, b, and c) such that each airline’s strategy is optimal given what the others are
doing? Or, said another way, is there a set of strategies in
which “unilateral defection” does not pay?
e. Are there any strategies of A’s that are dominated by other strategies? That is, can you rule out one or more of A’s
strategies because they are always worth less than some-
thing else?
f. Follow the foregoing logic to its end: if you rule out some of A’s strategies, can you also rule out some for B and C?
And if you can do that, can you then go back and rule out
more of A’s strategies? Can you continue this process of
“iterated dominance” to convince yourself of how many
flights A should fly?
402
C16.INDD 10:42:13:AM 08/06/2014 PAGE 402Trim Size: 203.2 mm X 254 mm
Employment and Pricing of Inputs16CHAPTER
In previous chapters we emphasized factors determining the output and price of final products. In this and the next two chapters we turn our attention to factors determining
the employment levels and prices of inputs used to produce final products. There
are many similarities in the processes of analyzing product markets and input mar-
kets, since both involve the interaction of buyers and sellers. The leading actors’ roles,
however, are reversed. Firms are suppliers in product markets but demanders in input
markets. Households and individuals are demanders in product markets and suppliers in
input markets.
An examination of input markets helps us answer many interesting and important ques-
tions, such as “What determines a country’s income level and distribution?” “Why do
women earn less than men and African Americans less than whites?” “Why have the sala-
ries of professional baseball players risen so rapidly during the past four decades?” “Does
immigration make the United States better off?” and “What are the effects of unions, and
why have unions in the United States been declining in the private sector but growing in the
public sector?”
This chapter discusses the basic principles common to all input market analysis, whether
the input is labor, capital, land or raw materials. We first examine the demand for inputs by
competitive firms, then turn to the supply of inputs, and finally bring the two together to
complete the general model. The chapter’s final two sections analyze input markets under
noncompetitive conditions.
Learning Objectives
Explore the factors influencing the demand for an input by an individual competitive firm. Derive the market demand curve for an input by aggregating the demand curves of the various firms interested in hiring the input. Investigate the general shape of an input supply curve. Explain how the price and employment of inputs are determined in an industry. Show how an input’s price and employment level is determined in a multi-industry market. Examine input demand and employment by an output market monopoly. Define what is meant by monopsony in input markets.
Memorable Quote “It is not the employer who pays the wages. Employers only handle the money. It is the customer who pays the wages.”
—Henry Ford
The Input Demand Curve of a Competit ive F i rm 403
C16.INDD 10:42:13:AM 08/06/2014 PAGE 403Trim Size: 203.2 mm X 254 mm
16.1 The Input Demand Curve of a Competitive Firm The market demand curve for an input shows the total quantity of the input that will
be purchased at various prices by all demanders as a group. To determine the market
demand, we begin with the factors influencing the employment decision of the individual
demander, usually a business firm. Then, we aggregate the demands of individual firms to
obtain the total or market demand for the input. This treatment is similar to the derivation
of the market demand curve for a consumer good. In that case, we first derived the indi-
vidual consumer’s demand curve; then, we combined those curves to obtain the market
demand curve.
In this section we learn how a change in an input’s price affects its use by a competi-
tive firm. At the outset, note that there are no new assumptions in the analysis. We use
the same competitive model developed in Chapters 8 and 9, but now our focus is on input
rather than output markets. For example, we still assume the firm’s goal is profit maxi-
mization, but now we want to see why profit maximization, along with the competitive
model’s other assumptions, implies that a firm will employ more of an input when its price
is lower.
How does a firm determine how many workers to hire when its goal is to maximize
profit? For the moment, we assume that workers are homogeneous—that is, interchange-
able as far as the firm is concerned—so the only question is how many the firm should
employ. Each additional worker hired adds to the firm’s cost, since the firm must pay the
going wage rate. At the same time, each additional worker also adds to the firm’s reve-
nue, since a larger work force produces more output. Thus, benefits (greater revenue) and
costs (wages) are associated with the firm’s employment decision. The firm will increase
profit by hiring additional workers as long as the additional revenue generated by the output
expansion exceeds the wages paid. A comparison of the marginal benefit of hiring workers,
in the form of added revenue, to the marginal cost of hiring workers, in the form of added
wage costs, guides the firm’s decision of how many workers to employ.
The Firm’s Demand Curve: One Variable Input Imagine a short-run setting where the quantities of nonlabor inputs (e.g., raw materials
and machines) are fixed and only the number of workers can be varied. In this setting the
law of diminishing marginal returns applies to labor: beyond some point, each additional
worker results in a smaller addition to output. The contribution to output made by increas-
ing the number of workers is an important determinant of the firm’s demand for labor, and
the marginal product curve (as described in Chapter 7) contains the relevant information.
In Figure 16.1, the downward-sloping portion of the marginal product curve for labor is
MPL. The marginal product curve indicates that if the firm employs 20 workers per day, the output produced by an additional worker (MPL) is three units; if employment increases to 25 workers, the marginal product of labor is lower (in this case two units) due to the law of
diminishing marginal returns.
Starting from any given employment level, let’s consider how hiring an additional
worker affects the firm’s total revenue. If 20 workers are employed, one more worker
increases final output by three units. If the final output sells for a price, P, of $100 per unit, the additional three units of output generated by hiring another worker add $300 to reve-
nue. Multiplying the marginal product by the price per unit of output (MPL × P) gives the marginal value product of labor (MVPL). In general, the marginal value product measures the extra revenue a competitive firm receives by selling the additional output generated when employment of an input is increased by one unit.
marginal value product (MVP) the extra revenue a competitive firm receives by selling the additional output generated when employment of an input is increased by one unit
404 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 404Trim Size: 203.2 mm X 254 mm
A downward-sloping marginal value product curve is derived by multiplying the
constant price of output (recall that we are dealing with a competitive firm; the price is
unchanged when more output is sold) by the declining marginal product of labor. The mar-
ginal value product curve coincides with the marginal product curve. The only difference is
that now we are measuring the marginal product of labor in terms of what it sells for on the
vertical axis. For example, when 20 workers are employed and the price of output is $100,
MPL is three units and MVPL is $300; when 25 workers are employed, MPL falls to two units and MVPL to $200.
The marginal value product curve is the firm’s demand curve for a given input when all other inputs are fixed. To see this relationship, suppose that the daily wage rate is $300 per worker. The firm can hire as many workers as it wants at this wage rate, so each addi-
tional employee adds $300 to the firm’s total cost. Every extra worker, however, also adds
an amount equal to the marginal value product to the firm’s total revenue. Comparing the
cost and revenue effects tells the firm how many workers to hire. For instance, if the firm
is currently employing 15 workers, the marginal value product of an additional worker is
$400. Hiring another worker thus adds more to revenue ($400) than to cost ($300), so profit
increases by employing more workers. The firm should expand employment up to the point
where the marginal value product has fallen to $300, the wage rate. In Figure 16.1, the most
profitable employment level is 20 workers when the wage rate is $300. If the firm hires
more than 20 workers, total cost goes up by more than total revenue (the wage rate, $300, is
greater than MVPL beyond 20 workers), so profit declines. Note that at a lower wage rate, hiring more workers is profitable. For instance, if the
wage rate drops to $200 per day, the firm maximizes profit by expanding employment to
25 workers. At the initial employment level of 20 workers and the lower wage rate, the
marginal value product of hiring another worker ($300) is now greater than the wage cost
($200), so the firm adds more to revenue than to cost by employing 5 more workers.
Two important conclusions emerge from this analysis. First, the marginal value product
curve identifies the most profitable employment level for the input at each alternative cost.
The firm will hire up to the point where the input’s marginal value product equals its cost.
Second, the marginal value product curve—the firm’s demand curve when other inputs are
not varied—slopes downward. This follows directly from the law of diminishing marginal
returns: If an input’s marginal product declines as more is employed, so must the marginal
value product.
The preceding analysis assumes that the firm is a profit maximizer in a competitive mar-
ket. It may be helpful to relate this analysis to the discussion in Chapter 9 that emphasized
A Competitive Firm’s Demand for Labor: One Variable Input With labor as the only variable input, we can convert labor’s marginal product curve, MPL, into the marginal value product curve, MVPL, by multiplying the marginal product of labor by the price of the commodity produced. The MVPL curve is the competitive firm’s demand curve for labor if other inputs cannot be varied.
4
3
2
1
0
MPL and MVPL(MPL × P)
Quantity of labor15 2520
$400
$300
$200
$100
Marginal product (MPL)
Marginal value product (MPL × P)
Figure 16.1
The Input Demand Curve of a Competit ive F i rm 405
C16.INDD 10:42:13:AM 08/06/2014 PAGE 405Trim Size: 203.2 mm X 254 mm
the most profitable output for the competitive firm. We have just seen that the firm maxi-
mizes profit by employing an input—in this case, labor—up to the point where its MVP equals the cost of the input—in this case, the wage rate w. When profit is at a maximum, therefore, the following condition holds:
w MVPL= . (1)
Because MVPL equals MPL × P, if we divide both sides of equation (1) by MPL, we obtain:
w MP PL/ = . (2)
Recall from Chapter 8 that the ratio w/MPL is equal to the marginal cost (MC) of pro- ducing one more unit of output by using additional amounts of labor. Therefore, equation
(2) is equivalent to the price-equals-marginal-cost condition for profit maximization in a
competitive output market. When the competitive firm is hiring workers so that w = MVPL, then MC = P, and vice versa. We have been looking at the same process of profit maximi- zation that we examined in earlier chapters, but now from the perspective of its implica-
tions for the employment decisions of the firm.
The Firm’s Demand Curve: All Inputs Variable In identifying a firm’s MVP curve as its demand curve for an input, we assumed that the quantities of other inputs are fixed. In general, however, a change in an input’s price leads
a firm to alter its employment of not only that input but also other inputs. For example, a
reduction in the cost of computers may lead to the employment of more computer program-
mers as well as more computers. Consequently, an input demand curve should allow a firm
to adjust its use of other inputs as well as the one whose price has changed.
We can easily extend the analysis to allow for variation in the quantities of all inputs.
Suppose a competitive firm is initially in equilibrium, employing the appropriate quantities
of all inputs. In Figure 16.2, the firm is operating at point A on MVPL, employing 20 work- ers when the daily wage rate is $300. Note that the quantity of capital, assumed to be the
A Competitive Firm’s Demand for Labor: All Inputs Variable When all inputs are variable, an input’s MVP curve shifts with changes in the employment of other inputs. The firm’s labor demand curve is then the d curve, which takes into account the way changes in the amount of capital employed affect the MVP of labor.
Wage
$300 = w
$200 = w′
0
B
d MVPL(K = 12)
MVPL(K = 10)
A
C
Quantity of labor
20 3025
′
Figure 16.2
406 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 406Trim Size: 203.2 mm X 254 mm
only other input, is constant at 10 units at all points along MVPL. Now suppose the wage rate falls to $200. If the quantity of capital is kept constant at 10 units, the firm increases its
employment of workers to 25 units, at point B on MVPL. This increased employment does not represent a complete adjustment to the lower wage rate, since it will normally be in the
interest of the firm to expand its employment of capital, too.
An increase in the quantity of capital, though, shifts the MVPL curve upward. If the quantity of capital increases to 12 units, the MVPL curve shifts to MVPĹ. With 12 units of capital, each worker has more “tools” to work with than before, so the marginal productiv-
ity of workers is greater. The greater marginal productivity, coupled with an unchanged
product price, implies that the new marginal value product curve has a greater height at
each possible employment level than before. The adjustment leads to a further increase in
the employment of workers, to point C on the marginal value product curve for a constant 12 units of capital. Thus, the firm’s full response to the lower wage rate is an employment
increase from 20 to 30 workers.1
A and C are two points on the firm’s labor demand curve, d, when the employment of all inputs can be varied in response to a change in the wage rate. The firm still employs work-
ers up to the point where their marginal value product equals the wage rate, but we have
now allowed for the effect of variations in the employment of other inputs on the marginal
productivity of labor. We can think of the demand curve, d, therefore, as a generalized marginal value product curve. Since this curve allows the firm to vary all inputs, it is the
competitive firm’s long-run demand curve for an input. Note that in deriving this long-
run demand curve, we assume that other inputs’ prices are unchanged (only the quantities
employed are variable) and that the final product’s price is also constant.
The Firm’s Demand Curve: An Alternative Approach We can gain further insight into a firm’s adjustment to an input price change with an alter-
native approach. Our first method has the advantage of linking the demand for an input
to its marginal productivity but the disadvantage of obscuring what happens in the output
market and the market for other inputs. The approach we develop next also shows that the
demand curve for an input slopes downward, but pays more explicit attention to the output
market and the demand for other inputs.
Figure 16.3b shows the firm’s position, q1, in its output market where its marginal cost curve MC crosses the horizontal demand curve at point E. Figure 16.3a shows the same initial situation from the perspective of the firm’s employment of inputs. The least-cost
method of producing output q1 occurs at the tangency between the IQ1 isoquant and the isocost line AZ at point E. The firm employs 20 workers and 10 units of capital at point E.
Now let’s work through the effects of a reduction in the wage rate. First, how will a
lower wage rate affect the firm if we tentatively assume it continues to produce the same output? Recall from Chapter 8 that the isocost line’s slope equals the ratio of the wage rate
to the price of capital. A lower wage rate and an unchanged price of capital imply that iso-
cost lines will be flatter because labor becomes cheaper relative to capital when the wage
rate falls. Isocost line A′Z′ in Figure 16.3a reflects the lower wage rate, and the least costly way of producing q1 units of output occurs at E1, where A′Z′ is tangent to isoquant IQ1.
1Our analysis deals with the situation, thought to be typical, where labor and capital are complements. (Two inputs are complements if an increase in the quantity of one leads to an increase in the marginal product of the other.) When the firm uses more labor and the same amount of capital at point B, the marginal product curve of capital increases so the firm also expands its use of capital. The demand curve for labor, however, still slopes downward if the two inputs are substitutes. (They are substitutes if an increase in the quantity of one decreases the marginal product of the other.) In that case, when the firm increases labor (from point A to point B), the marginal product of capital declines and the firm employs less capital. A reduction in capital increases the marginal product of labor when the inputs are substitutes, so the MVPL curve shifts upward in this case as well, just as it does in Figure 16.2.
The Input Demand Curve of a Competit ive F i rm 407
C16.INDD 10:42:13:AM 08/06/2014 PAGE 407Trim Size: 203.2 mm X 254 mm
To produce an unchanged output, the firm uses more labor and less capital when the rela- tive cost of labor falls; that is, the firm substitutes labor for capital.
In Figure 16.3b, point E1 on MC′ shows the same adjustment. A wage rate reduction lowers the entire marginal cost curve to MC′ (see Chapter 8 for a more detailed discussion), so when output is q1, the product’s price is greater than its marginal cost. The firm, there- fore, has an incentive to expand output as a result of the lower wage rate. Now consider the
subsequent effects as the firm expands output from q1 to the new profit-maximizing level q2. Figure 16.3a shows this effect as the movement along the new expansion path (based on
the lower wage rate and unchanged price of capital) from point E1 to point E2. As the firm produces more output, it moves to a higher isocost line, A1Z1, and it employs more of both inputs than it did at point E1. At E2 the firm employs 30 workers and 12 units of capital. The wage rate reduction thus increases labor employment from 20 to 30 workers.
This approach to an input’s demand curve involves separating the total effect of a price
change into two components, similar to what we did with a consumer’s demand curve. The
change in input employment when output is held constant and one input is substituted for
another is called the substitution effect of an input price change. The movement along the IQ1 isoquant from point E to point E1 (from 20 to 24 workers) shows the substitution effect in our case when the wage rate falls. The change in input employment when output is altered
is called the output effect of an input price change and is shown by the movement along the new expansion path from point E1 to point E2 (from 24 to 30 workers).2 When summed, these two effects identify a firm’s full input employment response to an input price change.
substitution effect of an input price change the change in input employment when output is held constant and one input is substituted for another in response to an input price change
output effect of an input price change the change in input employment when output is altered in response to a change in the price of an input
Figure 16.3
Capital
A
12
New expansion path
10
A1
A′
0 20 24
E
LaborZ1
E2
E1
IQ1
IQ2
Z′
Z
(a)
Dollars per unit
P3
0 Outputq1 q2
E2E
MC MC′
d
E1
(b)
30
A Competitive Firm’s Demand for Labor: All Inputs Variable (a) A lower wage rate causes the firm to substitute toward labor and away from capital– the move from E to E1. (b) At a lower wage rate, output expands from q1 to q2, as the lower wage rate shifts the marginal cost curve downward. This output effect further increases labor employment–the move from E1 to E2 in part (a).
2In contrast to a consumer’s demand curve, the firm’s labor demand curve is not derived by pivoting the isocost line at point A in Figure 16.3a. That approach would be valid only if the firm continues operating at the same total cost, which is generally untrue. The most profitable total cost for the firm to incur depends on the demand and cost conditions shown in Figure 16.3b.
408 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 408Trim Size: 203.2 mm X 254 mm
Since the substitution and output effects both imply greater employment at a lower input
price, the firm’s input demand curve slopes downward. The firm employs more workers at
a lower wage rate because it uses more labor per unit of output (the substitution effect) and
because it is profitable to increase output when production cost falls (the output effect).
16.2 Industry and Market Demand Curves for an Input To derive the market demand curve for an input, we aggregate the demand curves of the
various firms interested in hiring the input. Since an input such as labor is likely to be
demanded by firms in many different industries, we proceed in two steps. First, we need
to determine each industry’s demand curve for labor. Second, we (horizontally) aggregate
each industry’s demand curve for labor to obtain the total or market demand curve for it.
A Competitive Industry’s Demand Curve for an Input In determining a particular industry’s demand curve for an input such as labor, we must
recognize one problem. When deriving the competitive firm’s input demand curve, we
assumed that the product’s price was fixed. Recall, though, that when the wage rate fell, the
firm expanded output and sold the larger output at an unchanged price. The assumption of a
fixed product price is appropriate when we are dealing with just one firm. But now we are
interested in the response of all firms in an industry to a lower wage rate. When all firms simultaneously increase output, they can sell more total industry output only at a lower price.
Figure 16.4 illustrates how this factor affects the derivation of an industry’s demand
curve for an input. In Figure 16.4a, d is the labor demand curve for a single firm, assuming
Figure 16.4
The firm
0 20 3027
C
A
B
Labor
d(P = $100)
(a)
Wage
$300 = w
$200 = w′
(b)
The market
0 2,000 3,0002,700
C′
A′
B′
D
Labor
�d (P = $100)
�d ′(P = $80)d ′(P = $80)
Wage
$300 = w
$200 = w′
The Competitive Industry’s Demand for Labor (a) The competitive firm’s demand curve for labor, d, assumes a given product price. (b) In deriving the industry demand curve for labor, we must take into account that as industry output changes, so will the product price. The industry demand curve for labor is D.
Industry and Market Demand Curves for an Input 409
C16.INDD 10:42:13:AM 08/06/2014 PAGE 409Trim Size: 203.2 mm X 254 mm
that the price, P, of the final product is $100 at all points along d. If the initial wage rate is $300, the firm in Figure 16.4a hires out to point A on its demand curve d and employs 20 workers. With 100 identical firms in the industry interested in hiring labor, total employ-
ment is 2,000 workers at the $300 wage rate, or point A′ in Figure 16.4b. Point A′ lies on the ∑d curve, which is the horizontal summation of the d curves of the firms in the industry.
By assumption, A′ in Figure 16.4b is a point on the industry demand curve for labor. The simple summation of the firms’ demand curves, ∑d, does not, however, show the amount of labor demanded by the industry at other wage rates. To see this, suppose that the wage rate
falls to $200. In Figure 16.4a, individual firms begin expanding employment from point A to point B, and in the process output rises. As industry output rises, the product price falls since consumers will buy the larger output only at a lower price. But a lower product price
shifts each firm’s labor demand curve downward, since labor’s marginal value product
curve is lower when the value (price) of the product is reduced. (Recall that the marginal
value product of labor is equal to the marginal product of labor multiplied by the price of
the output. Since the output price has fallen, the marginal value product of labor is lower
for each level of input use.) If the product price falls to $80 per unit, for example, the firm’s
demand curve becomes d′ in Figure 16.4a, and the firm employs 27 workers. In Figure 16.4b, all the firms together begin by increasing employment from point A′ to point B′, but the increase is cut short by the falling product price, and they end up at point C′. Point C′ is a second point on the industry demand curve for labor, D.
This derivation of an industry’s labor demand curve accounts for the effect of increased
employment, and hence output, on the product price. The industry demand curve is less
elastic than the ∑d curve, which is based on a fixed product price but still slopes down- ward. This relationship becomes easier to see when we recognize that there are still substi-
tution and output effects, implying greater employment at the lower wage rate even when
the declining product price is accounted for. The product price falls only because more out-
put is produced, and greater production involves the use of more labor (the output effect).
In addition, firms use more labor per unit of output when the wage rate is lower because
they substitute labor for other inputs (the substitution effect) at each given output level.
Note that we assume the demand curve for the final product is fixed when deriving an
industry’s demand curve for an input. In fact, economists often refer to an industry’s input
demand curve as a derived demand: the textile industry’s demand for workers, for exam- ple, is derived from the demand of consumers for textiles. Firms will pay workers to pro-
duce textiles only because consumers are willing to pay for textiles. If the demand curve for
textiles shifts, the textile industry’s demand curve for workers also shifts. The consumer
demand curve for a product is thus an important determinant of the industry demand curve
for an input used in the production of the product.
The Elasticity of an Industry’s Demand Curve for an Input The price elasticity of an industry’s demand for an input is defined and measured in the
same way as a consumer demand curve. The magnitude of the price elasticity of demand for
an input can be critical. For example, in evaluating the minimum wage law, whether a 10
percent increase in the legal minimum wage reduces employment of low-wage workers in
the fast-food industry by 20 percent (an elasticity of 2.0) or 5 percent (an elasticity of 0.5)
makes a big difference.
There are four major determinants of the elasticity of an industry’s demand for an input.
First, the greater the elasticity of demand for the product produced by the industry, the
more elastic the input demand. Recall that an industry’s input demand curve is a derived
demand curve. If consumers will purchase a great deal more of the good at a slightly lower
price (highly elastic product demand), firms in the industry will produce much more when
an input price falls, and employment will increase sharply. An elastic product demand gives
derived demand another name for an industry’s input demand curve, reflecting the fact that the industry’s demand for an input ultimately derives from consumers’ demand for the final product produced by that input
410 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 410Trim Size: 203.2 mm X 254 mm
rise to a large output effect, which, in turn, contributes to the elasticity of the industry’s
demand for inputs. For instance, consider Figure 16.4b, and suppose the wage rate falls
from $300 to $200. If the consumers’ demand curve was perfectly elastic, the greater indus-
try output could be sold at an unchanged $100 price. The firms would expand employment
to point B′, and, in fact, ∑d would be the industry’s demand curve for labor in this case. Second, an industry’s input demand is more elastic when it is easier to substitute one
input for another in production. This condition refers to the technology of production
reflected in the curvature of the production isoquants. When it is technically easy to sub-
stitute among inputs, the substitution effect of an input price change is large, implying a
large (elastic) employment change. For example, if machines can adequately do the work
performed by workers and at only a slightly higher cost, a wage increase can lead firms in
an industry to switch entirely to machines and reduce employment of workers to zero—
implying a highly elastic industry demand for workers.
Third, an industry’s demand for an input is more elastic when the supply of other inputs
is more elastic. If machine prices rise sharply when firms switch from workers to machines
(implying an inelastic supply), only a limited amount of profitable switching can occur,
resulting in a small substitution effect and a low elasticity of demand for workers. The
output effect reinforces this impact. If machines rise in price as more are used when output
increases, the additional output firms can profitably produce as wage rates fall is limited.
Fourth, the longer the time allowed for adjustment, the more elastic an industry’s
demand for an input becomes. This is true because substitution possibilities among inputs
become greater as firms in an industry have more time to alter their usage of various inputs.
For example, a rise in the wage rate may mean that replacing workers with machines is
profitable. It takes time, however, for machines to be built and installed, so in the short run
few workers will be discharged.
APPLICATION 16.1
Employee wages in the domestic airline industry were significantly higher when the industry was subject to regula- tion. For example, veteran union pilots at the major carriers made $150,000 per year in 1978, just prior to airline dereg- ulation. By contrast, firms entering the industry following deregulation paid their top pilots only $45,000 per year.
The higher pilot wages during regulation stemmed from the actions of the Civil Aeronautics Board (CAB). Specifi- cally, the CAB limited competition between airlines on any given route during regulation. Only a few carriers were ini- tially issued licenses to operate on each route and the CAB refused to grant additional licenses to other firms wanting to enter a market.
With limited competition, airlines with operating licenses for particular routes faced more inelastic demand curves for their output. More inelastic demands for final output implied that the derived demands for inputs, such as pilots, also were less price sensitive. In other words, because the cost of pilot salaries could more easily be passed on to an airline’s customers in the form of higher air fares, domes- tic airlines were less price sensitive in their demand for
Explaining Sky-High Pilot Salaries under Airline Regulation
pilots during regulation. Consequently, pilot’s wages in the domestic airline industry were higher under regulation than after deregulation.
Among categories of airline employees, the relative dif- ference between salaries at established firms prior to dereg- ulation and new firms subsequent to deregulation also was greatest for pilots. For example, the average unionized non- pilot worker at established airlines prior to deregulation made slightly less than double the amount at new firms post-deregulation ($39,000 versus $22,000) compared to the nearly 3.33 times higher salaries of pilots before versus after deregulaion. This reflects the second determinant of the elasticity of an industry’s demand for an input identified above: pilots are arguably the most difficult to substitute for when it comes to producing output in the airline business. Since airlines have less relative ability to substitute away from pilots to other inputs, any wage increase demanded by pilots during regulation was more likely to be agreed to by airlines relative to a wage increase proposed by unions representing other employees such as flight attendants, gate agents, baggage handlers, and so on.
The Supply of Inputs 411
C16.INDD 10:42:13:AM 08/06/2014 PAGE 411Trim Size: 203.2 mm X 254 mm
The Market Demand Curve for an Input Once we have derived an industry’s demand curve for an input, determining the market
demand curve for the input is straightforward. We need to recognize that firms in more than
one industry may be vying for the services of a particular input. For example, the automo-
bile industry is not the only buyer of steel or the only employer of engineers. The aerospace
and construction industries also may compete for the services of the same inputs.
The market demand curve for an input is determined by (horizontally) aggregating the
various industry demand curves for the input. The aggregation is analogous to the manner
in which the demand curves of individual consumers are aggregated to obtain the market
demand curve for a product. As we will see in a later section, it is the market demand for an
input (the total demand of all industries using the input) that interacts with the total supply
to all industries to determine input prices.
16.3 The Supply of Inputs The supply side of input markets deals with the quantities of inputs available at alterna-
tive prices. This subject is somewhat complicated because the shape of the supply curve is
likely to differ according to the type of input. In this section we make some general obser-
vations that are applicable to all inputs; we defer a discussion of specific inputs until the
next chapter.
A broad definition of inputs might classify them as either labor, land or capital. A nar-
row definition might distinguish between skilled and unskilled workers, land in New York
City and Iowa, and buildings and trucks. The appropriate definition depends on the prob-
lem. The broad classification serves to make the general points here, but in the next two
chapters we will see examples of cases for which it is fruitful to be more specific.
People own the inputs used by firms to produce goods. Our problem is to understand the
conditions under which the owners of inputs will offer them for sale or rent. At the outset
we should distinguish between the amount of inputs in existence at any given time—the
stock of resources—and the amount offered for sale or rent. At any time, a fixed number of
people are capable of working. There is a fixed area of land, and a fixed number of build-
ings, machinery, and other capital equipment. The amount in existence can differ signifi-
cantly from the amount owners offer for use. Since the amount that owners offer depends on
the price they are paid, we must be concerned with the supply curves of inputs, not just the
stock of inputs in existence.
The general shape of an input supply curve depends critically on the market for which the supply curve is drawn. Consider the supply curve of labor to all industries in the econ- omy. To simplify, suppose that all workers are identical, so there is only one wage rate.
If the wage rate goes up, will the total amount of labor offered increase? We will give a
fuller analysis in the next chapter, but here we simply note that the total amount of labor
can increase only if workers decide to work longer hours or if more people enter the labor
force. Such responses to a higher wage rate may be so small, though, that the supply curve
of labor to all industries together will be approximately vertical, as in Figure 16.5a. A verti-
cal supply curve indicates that an increase in the wage rate from w1 to w2 leaves the number of workers unchanged, at 100 million. We are not asserting that the supply curve will nec-
essarily be vertical, but it could be; so for the moment let’s suppose it is.
Although the supply of labor to all industries taken together may be vertical, this does
not imply that the supply curve of labor confronting any particular industry is vertical. While the total number of workers employed in the economy may not change, the num-
ber employed by a particular industry is subject to great variation. If the wage rate paid
412 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 412Trim Size: 203.2 mm X 254 mm
to software programmers increases, workers in other industries will leave their jobs to go
to work as software programmers. This adjustment doesn’t change the total number of
workers employed, but it does change employment in the software programming industry.
The supply curve of workers to the software programming industry thus slopes upward, as
illustrated in Figure 16.5b. An increase in the software programming wage from w1 to w2 induces 10,000 workers to move from other jobs into the software programming industry.
Both labor supply curves in Figure 16.5 are correct in the sense that they can both
exist simultaneously. Figure 16.5a shows the supply curve of labor to the entire economy; Figure 16.5b shows the supply curve of labor to the software programming industry. Because the software programming industry is only a small part of the entire (economy-wide) labor
market, its labor supply curve is more price sensitive than the supply curve of labor for the
economy. Indeed, if its share were as small as the numbers used in Figure 16.5, the software
programming industry would likely face a virtually horizontal labor supply curve.
This discussion explains why referring to the supply curve of an input is ambiguous. We must always specify that it is the input supply curve to a particular set of demanders.
Otherwise, we can fall into the trap of thinking that the supply curve of engineers to the
defense contracting industry, for example, is vertical, because in the short run there are
only a limited number of trained engineers. In fact, the supply curves of most inputs to most industries are likely to be upward sloping, as in Figure 16.5b, regardless of the shape of the
supply curve of the input to the economy as a whole, because most industries employ only
a small portion of the total amount of any input.
This concept applies to other inputs besides labor. Although the total supply of land to
the economy may plausibly be fixed (a vertical supply curve, as in Figure 16.5a), the supply
available to any given industry is not. The corn industry can bid land away from other uses
if it expands, just as homeowners can bid land away from farmers for building homes.
Consequently, for individual industries, input supply curves will generally be quite elastic.
However, supply curves of inputs to more broadly defined markets will be less elastic, and
Figure 16.5
0
S
Labor (in all industries)
(a)
Wage
w2
L1
w1
(100,000,000)
0
S
Labor (in software programming industry)(b)
Wage
w2
L1
w1
(5,000)
L2
(15,000)
The Supply Curve of Labor to the Economy and to a Particular Industry Distinguishing (a) the supply curve of an input to all industries together from (b) the supply curve of the input to one industry is important. The supply curve to one industry will always be more elastic than the supply curve to the economy as a whole.
Industry Determinat ion of Pr ice and Employment of Inputs 413
C16.INDD 10:42:13:AM 08/06/2014 PAGE 413Trim Size: 203.2 mm X 254 mm
Figure 16.6
0
s
d D
Labor Labor
(a)
Wage
$300 = w
l
(500)
0
S
(b)
Wage
$300 = w
L0 L
$250 = w1
(6,000)
L1
(10,000) (12,000)
The firm The industry
The Equilibrium Wage and Employment Level for a Competitive Industry (a) The position of the firm in equilibrium is shown. Each firm faces a horizontal supply curve at the industry-determined wage rate of $300. (b) Supply and demand in an industry determine the equilibrium employment level and wage rate: 10,000 workers at $300 per day.
there are some types of situations where we should use this kind of supply curve. The proper
pairing of supply and demand concepts is important to the analysis of input markets, as we
will show in detailed examples ahead.
16.4 Industry Determination of Price and Employment of Inputs
Firms in an industry compete with one another to acquire inputs, and the industry demand
curve for an input summarizes the way an input’s price influences the firms’ hiring deci-
sions. Input owners provide resources to firms in the industry, and the supply curve of an
input to the industry reflects the way the input owners’ decisions depend on the price they
receive. As with other competitive markets, the interaction of supply and demand deter-
mines the equilibrium price and quantity. Figure 16.6 illustrates this process, once again
using labor as an example. As before, we assume that all workers are identical.
Figure 16.6b shows a particular industry’s demand and supply curves for labor. Perfect
competition results in an equilibrium where the industry’s firms employ 10,000 workers
at a (daily) wage rate of $300 (w). To understand why the behavior of firms and work- ers results in this outcome, consider what would happen if some other rate prevailed.
Suppose, for example, that the wage rate is $250 (w1) instead. At that wage, only 6,000 (L0) workers agree to work. With such low labor costs, firms as a group find it profitable to employ 12,000 (L1) workers, but only 6,000 are available. A shortage of labor will exist at the $250 wage rate, resulting in a tight labor market. Firms will advertise for workers but
414 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 414Trim Size: 203.2 mm X 254 mm
have too few applicants at the current wage. Workers currently employed by one firm will
receive job offers from others at higher wages. Firms unable to recruit workers from within
the industry will try to hire workers from other industries, but to do so, they will have to
offer higher wages.
As a result, the wage rate will not stay at $250; it will rise. As firms bid up the wage,
workers will quit their jobs in other industries to seize the better opportunity in this market,
resulting in an increase in the quantity supplied—a movement up the supply curve. As the
wage increases, firms will find that it is no longer profitable to try to fill 12,000 jobs, and
the quantity demanded will decrease—a movement up the demand curve. The process will
continue until the wage rate reaches the point where the number of workers willing and
able to work for firms in the industry equals the number of workers firms are willing to
employ. Graphically, the equilibrium is shown by the intersection of the industry demand
and supply curves.
When the labor market for a particular industry is in competitive equilibrium, the
situation from an individual firm’s perspective is depicted in Figure 16.6a. The equi-
librium wage is $300, and each firm faces a horizontal supply curve at that wage. The
individual firm is a small part of the total market, so it has no option but to pay the going
wage determined in the broader market (illustrated in Figure 16.6b). And at the equilib-
rium wage, it can hire more or fewer workers without appreciably affecting the wage
rate. Faced with the $300 wage, the firm in Figure 16.6a maximizes its profit by hiring
500 workers.
One implication of this analysis is that the market-determined input price equals the marginal value product of the input. Each firm is in the position shown in Figure 16.6a, employing an input quantity for which the marginal value product equals the input price.
Recall that an input’s marginal value product is the value consumers place on the addition
to output made by the input. Thus, in a competitive market, input owners are compensated
according to how much value the inputs they supply add to output. This happens because
input demands are derived demands. Consumers, in their product purchases, are indirectly
expressing how valuable the services of the inputs are to them.
Process of Input Price Equalization across Industries Several different industries often employ the same inputs. Most industries, for example,
use land, electricity, unskilled labor, and buildings. To understand how the prices of these
widely used inputs are determined, we must look beyond the boundaries of a single indus-
try and recognize that there is competition among many industries for these inputs. We
emphasize one characteristic of this situation here: the tendency for identical inputs to
receive the same price, regardless of the industry employing them.
Suppose that the aerospace and telecommunications industries both employ computer
programmers. Suppose also that for some reason, wages are higher in the aerospace indus-
try. The wage difference won’t persist for long because programmers can move from one
industry to another. Programmers in the low-paid telecommunications industry will leave
their jobs and seek work in the aerospace industry, where pay is higher. This movement
will simultaneously reduce the supply of programmers to the telecommunications industry
and increase the supply of programmers to the aerospace industry. With more programmers
seeking employment in the aerospace industry, the wage rate there will decline, and the
wage rate will rise in the telecommunications industry, where supply has decreased. This
process will continue until wages in the two industries are equal, since only then do pro-
grammers have no further incentive to change jobs.
Figure 16.7 illustrates the process of input price equalization. Let’s assume that the aero-
space and telecommunications industries together employ 1,500 programmers: aerospace
employs 500, while telecommunications hires 1,000. At their respective levels of employ-
ment, the daily wage rate is $400 in the aerospace industry (Figure 16.7a) and $300 in the
Industry Determinat ion of Pr ice and Employment of Inputs 415
C16.INDD 10:42:13:AM 08/06/2014 PAGE 415Trim Size: 203.2 mm X 254 mm
telecommunications industry (Figure 16.7b). To show why these markets are not in equi-
librium and how they will ultimately adjust to an equilibrium, we use a specially defined
supply curve. Consider the momentary, or very-short-run, supply curve of programmers
that identifies the number of programmers in each industry at a specific time. These supply
curves will be vertical; that is, at any given time each industry employs a certain number of
programmers. These curves are not the supply curves we use in most applications. We use
them here for the specific purpose of illustrating explicitly how the movement of program-
mers from one industry to another affects both markets.
Initially, the momentary supply curve of programmers in the aerospace industry is SSA, and in the telecommunications industry it is SST. Given the initial allocation of program- mers between the industries, the wage rate of the 500 aerospace programmers is $400, and
the wage rate of the 1,000 telecommunications workers is $300. Clearly, this result is not
an equilibrium since telecommunications programmers have an incentive to quit their jobs
and seek employment in the aerospace industry. Shifts in the SS curves show the movement of programmers between industries. The momentary supply curve in the telecommunica-
tions industry shifts to the left as programmers leave, and when these programmers seek
jobs in the aerospace industry, the momentary supply curve there shifts to the right. The
movement of workers decreases the difference in wages between the two industries: wages
fall in aerospace and rise in telecommunications. Moreover, the movement of programmers
will persist as long as the wage rate is higher in the aerospace industry, meaning it will
continue until wages in the two industries are equal. In the diagram, equilibrium occurs
when 200 programmers have moved from telecommunications to aerospace, and the com-
mon wage rate of $350 is established.
Labor
(a)
Aerospace industry
Wage
$400 = wA
$350 = w′ A
0 LA (500)
L′ (700)
SSA
DA
SS′A
Labor
(b)
Telecommunications industry
Wage
$300 = wT
′
0 LT (1,000)
$350 = wT
′LT (800)
D T
SS T′SS T
A
Figure 16.7 Input Price Equalization Across Industries When several industries employ the same input, the input tends to be allocated so that its price is equalized across industries. If this were not true–if, say, programmers were receiving $400 (wA) in the aerospace industry and $300 (wT) in the telecommunications industry–input owners would have an incentive to shift inputs to industries where pay is higher. This process tends to equalize input prices.
416 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 416Trim Size: 203.2 mm X 254 mm
Note that this process works without requiring all programmers to change jobs in search
of higher salaries. Relatively few programmers need to relocate to bring wages in the two
industries into equilibrium. In our example, only 200 of the 1,500 programmers have to
change jobs to produce a uniform wage rate. In fact, in some cases it is unnecessary for the
workers to move, geographically speaking, at all. Suppose that the aerospace industry is in
California and the telecommunications industry is in Colorado. Suppose also that program-
mers in Colorado are paid less but are unwilling to relocate to California. Labor immobil-
ity of this sort does not forestall the adjustment process. The wage differential creates an
incentive for aerospace firms to relocate in Colorado and take advantage of the lower wage
rate there, so the impact on wages will be the same as if workers had moved from Colorado
to California—a uniform wage will be established.3
As this discussion suggests, competitive markets establish uniform input prices across
firms, industries, and regions when identically productive inputs are compared. Competi-
tive markets thus promote “equal pay for equal work.” This conclusion is true if equal work is interpreted to mean equally productive work from the viewpoint of consumers—that is,
equal in terms of marginal productivity. The possibility that discrimination by employers
leads to differences in wage rates among equally productive workers, an exception to this
conclusion, is discussed in Chapter 18.
16.5 Input Price Determination in a Multi-Industry Market When several industries compete for the available supply of a particular input, the impact
of any one industry on the input’s price is likely to be slight, since it usually composes only
a small part of the total demand for the input. The broader multi-industry conditions of
demand and supply determine the input’s price. Now we wish to see how a single industry
fits into this broader input market and, in particular, to identify the factors that determine
the shape and position of the input supply curve confronting each industry.
Let’s consider the hypothetical market for engineers, workers we assume to be identi-
cal. Several industries employ engineers. The demand curve for engineers by the building
construction industry (Industry B) is DB in Figure 16.8b, and DA in Figure 16.8a reflects the demand for engineers by all other industries, excluding the building construction industry. Think of industry A as a group of industries; each has a demand curve for engi- neers, and their demands are aggregated as DA. Therefore, A and B together constitute the total market demand for engineers. The total market demand curve for engineers, the sum
of DA and DB, is DT in Figure 16.8c. The market supply curve of engineers to all indus- tries together is ST, and we have drawn it as upward sloping on the assumption that higher wage rates will be needed to encourage more people to enter the engineering profession.
(Note that we are looking at a time period long enough for people to complete their train-
ing. In the short run the market supply curve for engineers will be more inelastic.) The interaction between the number of people willing and able to work as engineers and the total demand for engineers by all firms and industries determines the wage rate for engineers. This interaction is shown in Figure 16.8c, with an equilibrium involving employment of 6,000 engineers at a daily wage rate of w, or $350. Each individual indus- try will then employ the number of engineers it wants at that wage. The building construc-
tion industry (Industry B) will hire 1,000 engineers, and all other industries (Industry A)
will hire 5,000, for a total of 6,000.
Our primary purpose in this section is to explain what determines the shape of the sup-
ply curve of engineers to a particular industry—in this example, the building construction
3In some cases, wages can differ between locations because of workers’ geographic preferences, but we defer that topic to Chapter 17.
Input Pr ice Determinat ion in a Mult i- Industry Market 417
C16.INDD 10:42:13:AM 08/06/2014 PAGE 417Trim Size: 203.2 mm X 254 mm
industry (Industry B). We have already identified one point on this supply curve, point F in Figure 16.8b. At F, 1,000 engineers are willing to work in the building construction indus- try at a wage of $350. Recall that we derive the supply curve of output for an industry by assuming a shift in the demand curve and tracing the consequences. We can use the same
approach to derive the supply curve of an input to an industry—that is, the supply curve of engineers to the building construction industry.
Let’s assume that Industry B’s demand for engineers increases to ′DB, perhaps because of an increase in consumer demand for the building construction industry’s output. As a
result, the total market demand for engineers increases, but the effect on the market demand
is proportionately less because demand has increased in only one market segment. With
the market demand rising to ′DT (equal to DA plus the new demand by B, ′DB), the wage of engineers is bid up to w′, and total employment increases to 6,500. At the new wage of $400, the building construction industry hires 2,000 engineers, at point G. Point G is a second point on the supply curve of engineers to the building construction industry; that is,
at a wage of $400, 2,000 engineers are willing to work in the building construction indus-
try (Industry B). In all industries (Industry A) where the demand curve has not shifted,
employment falls to 4,500 when the wage rises to $400.
Note that the additional 1,000 engineers employed in Industry B come partly from
Industry A, where 500 fewer are employed, and partly from an expansion in the total num-
ber of engineers from 6,000 to 6,500. In effect, as the building construction industry bids
for more engineers by offering higher wages, it attracts some from other industries and
also induces some new entrants into the engineering profession. Having derived the supply
curve of engineers to the building construction industry, we can understand more easily
why the supply curve of an input to a particular industry will normally be highly elastic.
In the example given here, when the wage rate in the building construction industry rose
from $350 to $400, the number of engineers willing to work there increased from 1,000 to
Figure 16.8
Industry A (all other industries)
(a)
Wage
$350 = w A ′$400 = w A
0
SA
′SA
DA
LA (5,000)
′LA (4,500)
Labor
Industry B (building construction
industry)
(b)
Wage
0 LB (1,000)
F
G
DB
DB′
SB
LB (2,000)
′ Labor
$400 = w′ $350 = w
(c)
Total labor market
Wage
0 LT (6,000)
ST
DT = DA + DB′ ′
LT (6,500)
′ Labor
DT = DA + DB
$350 = wB ′$400 = wB
Input Price Determination in a Multi-Industry Setting Total labor demand DT is the sum of the demands of industries B (building construction) and A (all other industries), and it intersects with total supply ST in part (c) to determine the uniform wage rate. In part (b) the supply curve confronting Industry B (the building construction industry) alone is derived by assuming that DB increases. Supply curve SB is highly elastic because Industry B is a small part of the total labor market.
418 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 418Trim Size: 203.2 mm X 254 mm
2,000, implying an elasticity of supply of about 5 (using the formula for arc elasticity). The
elasticity of the supply curve of engineers to the total market, however, is only about 0.6.
The reason for this difference is straightforward: the building construction industry is only
a part of the market for engineers, and it can bid some away from other industries—perhaps
only a few from each of dozens of industries—without greatly affecting the general wage
level for them all. The smaller the share of the total market accounted for by an industry, the more elastic its input supply curve. In our example, the building construction industry initially employed one-sixth of the total number of engineers, but this proportion rose to
nearly one-third after the demand increase. In many real-world cases, a single industry will
compose a much smaller part of the total market, so its input supply curve can easily be
perfectly elastic (horizontal). Recall the significance of high-input supply elasticities for the
elasticity of the output supply curve, discussed in Chapter 9.
We should also note that the increase in demand by the building construction industry
causes the supply curve of engineers to all other industries to shift. An input supply curve to
a given industry is based on given demand conditions in other industries (in drawing SB, we assumed that DA was fixed). When other industries compete more aggressively for inputs, Industry A will find its workers being bid away, causing a reduction in input supply to
Industry A. The result is a higher wage in Industry A as well. Remember that input prices
will be equalized across industries, so Industry A will be unable to retain engineers if it
pays less than the building construction industry (Industry B).
If we were concerned solely with the building construction industry, we would simply
need to consider Figure 16.8b. We should, however, integrate the supply of an input to
a particular industry into the broader market for the input, a market that usually contains
several industries. Indeed, the concept of an industry—a group of firms producing the
same product—was designed primarily to study how output markets work. For that pur-
pose, grouping the firms producing the same product makes sense. Relying on the same
classification scheme when analyzing input markets is much less helpful since many differ-
ent industries compete for the same supply of inputs. The notion of a multi-industry input
market (Figure 16.8c) is more appropriate.
16.6 Input Demand and Employment by an Output Market Monopoly
A monopoly is defined as a firm that is the sole seller of some product, but a firm that has
monopoly power in its output market does not necessarily have market power in its input
markets. A firm can be the sole seller of a product and still compete with a large number of
firms in hiring inputs. In that case, the firm is a monopoly in its output market and a com-
petitor in its input markets; this situation is the subject of this section’s discussion.
Like a competitive firm, a monopoly bases its decisions about input use on the way
profit is affected. It expands input employment as long as hiring one more unit adds more
to revenue than to cost. The price that must be paid for an input measures the added cost
of employing it—just as it did for a competitive firm. The difference in the two market set-
tings rests on the way hiring one more input unit affects the firm’s revenue.
For a competitive firm, employing one more input unit adds to revenue an amount equal
to the marginal value product. The marginal value product is the additional output pro-
duced multiplied by the price at which it can be sold. For a monopoly, one more input unit
also adds to revenue by expanding output, but revenue does not increase by the price at
which the additional output is sold. Recall that to sell more, a monopoly must reduce the
price for all its output; that is, the price received for an incremental unit of output is greater
than its contribution to revenue. Marginal revenue, which is always lower than price,
Input Demand and Employment by an Output Market Monopoly 419
C16.INDD 10:42:13:AM 08/06/2014 PAGE 419Trim Size: 203.2 mm X 254 mm
measures the effect on revenue of selling one more unit of output. Consequently, for a
monopoly, the contribution to revenue from employing one more input unit is the addi-
tional output (the input’s marginal product) multiplied by the marginal revenue associated
with the additional output. The product of marginal product and marginal revenue is called
the input’s marginal revenue product. Consider a situation in which all inputs but labor are fixed in quantity for the firm. In
Figure 16.9, the marginal value product curve, MVPL, would be the demand curve for the input under competitive conditions, as we explained in Section 16.1. If the firm is a
monopoly, the marginal revenue product curve, MRPL, is the demand curve for the input. For a monopoly, marginal revenue is below price at each level of output and at each level
of labor employment, so the MRPL curve lies below the MVPL curve. The monopoly’s demand curve for labor, the MRPL curve, slopes downward for two reasons. First, the mar- ginal product of labor declines as more labor is employed. (This relationship also holds true
in the competitive case.) Second, the marginal revenue associated with selling more output
also declines as more labor is employed, since the additional output can be sold only at a
lower price.
In Figure 16.9, at a wage rate of $400, the monopoly employs L1 workers. This amount is the profit-maximizing level of employment. At any lower level of employment the rev-
enue generated by hiring another worker exceeds the cost (MRPL > w), so profit increases if employment increases to L1, where MRPL = w. If the monopoly employed more workers, they would add more to cost (w) than to revenue (MRPL), and profit would be lower.
Employment of labor, or any other input, is lower under monopoly than under competi-
tion. Under competition, L2 workers are employed at the point where w = MVPL; only L1 workers are employed under monopoly. This result should come as no surprise, since it was
already implicit in our conclusion in Chapter 11 that a monopoly produces less output than
a competitive industry. To produce less, it uses fewer inputs. Figure 16.9 therefore depicts
the monopolistic reduction in output from the perspective of the input market.
Figure 16.9 also shows an output market monopoly’s deadweight loss. The marginal
value product of labor measures how much one more worker’s output is worth to consum-
ers. At the monopoly outcome, L1, MVPL equals $700, indicating that consumers are will- ing to pay more for another worker than it costs the monopolist to hire the worker (a wage
of $400). The benefit to consumers of more output is greater than the cost of producing the
output, but the monopoly does not hire more workers to expand output. This discussion
describes the same deadweight loss due to monopoly explained in Chapter 11, but focuses
on the input side of the picture.
marginal revenue product the product of an input’s marginal product and the marginal revenue that can be derived from selling that marginal product
An Output Monopolist’s Demand for an Input When only one input is variable, we derive a monopoly’s input demand curve by multiplying the input’s marginal product by the marginal revenue, MR, from selling the commodity produced: MRPL = MPL × MR. The marginal revenue product curve MRPL is the monopoly’s demand curve, and it lies below the competitive demand curve, MVPL.
Wage
$700
$400 = w
0 L2L1 Labor
S
MRPL = MPL × MR
MVPL = MPL × P
Figure 16.9
420 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 420Trim Size: 203.2 mm X 254 mm
We can extend the analysis of the monopoly demand for an input to a case where all
inputs are variable in the same manner as we did for the competitive firm. No significant
new conclusions emerge, and two important points remain. First, the input demand curves
of an output monopoly slope downward, both because the input’s marginal productivity
declines and because marginal revenue from selling output declines as more of any input is
consumed. Second, the input demand curves of a monopoly are lower than they would be if
the output market were competitive.
Finally, it is important to note that our focus in this section has been on the input demand
curve of a firm that is a monopoly in its output market. An output market monopoly is not the same as an input market monopoly. The latter involves a single seller of an input con- fronting the entire market demand curve for the input (where the market demand curve for
an input is derived by aggregating the demands for the input across various firms operating
in either competitive or monopoly output markets). We address the topic of input market
monopoly in the next chapter in the context of labor unions.
16.7 Monopsony in Input Markets Monopsony means “single buyer.” Pure monopsony in input markets occurs when a firm is the sole purchaser of an input. An example of pure monopsony is purchases by General
Motors of an automobile part that has been uniquely tailored by suppliers for GM cars. As a
sole buyer, a monopsony faces the market supply curve of the input, a curve that is often
upward sloping. An upward-sloping supply curve means that the monopsonist has market
power in the input market and can reduce the price paid without losing all the input.
An input market monopsony is analogous to an output market monopoly. An output
market monopoly has some discretion over its product’s price (as determined by the
downward-sloping demand curve), while an input market monopsony has some discretion
over the input’s price (as determined by the input’s upward-sloping supply curve).
Graphically, an upward-sloping supply curve for an input confronting the firm indicates
the presence of monopsony. We’ll use labor once more as an example. An upward-sloping
supply curve means that the firm must pay a higher wage rate to increase the number of
workers it employs. (Up until now, we have assumed that every firm faces a horizontal
supply curve for each input.) When the firm faces an upward-sloping input supply curve,
marginal input cost is not the same as average input cost. For example, suppose that a firm employs 10 workers at a wage of $300, but to employ 11 workers, the firm must pay a
wage rate of $310 to them all. The marginal cost of hiring the eleventh worker is therefore
$410, because total labor cost rises from $3,000 to $3,410 when employment increases by
one worker. Put differently, we can think of the wage rate as the average cost of labor
(ACL). When the average cost rises as more workers are employed (as is the case with an upward-sloping labor supply curve), the marginal cost of labor (MCL) must be greater than average cost. In this case the $310 wage rate is equal to the average cost of labor, or the
total wage bill divided by the number of workers, and the marginal cost is $410, or the
additional cost associated with hiring an extra worker.
We can identify the profit-maximizing employment level of a monopsony by comparing
the impact on total revenue from a change in employment with the effect on total cost, just
as we did before. Now, however, the effect on cost is determined by the marginal cost of
labor, which is not equal to the wage rate. Figure 16.10 illustrates how we determine the
profit-maximizing level of employment. The demand curve of the firm indicates the amount
that hiring an additional worker adds to revenue. (The demand curve will be the MVPL curve if the firm is a competitor in its output market; it will be the MRPL curve if the firm is a monopoly in its output market. Conceivably, a firm could be competitive in its output
monopsony an input market in which a firm is the sole purchaser of an input
marginal input cost the cost of using an additional unit of an input
average input cost the total cost of an input divided by the units of that input used by a firm
Monopsony in Input Markets 421
C16.INDD 10:42:13:AM 08/06/2014 PAGE 421Trim Size: 203.2 mm X 254 mm
An Input Market Monopsony An input market monopsony faces an upward-sloping input supply curve, so the marginal cost of employing the input is greater than the input price (the average cost of the input): MCL lies above S = ACL. The intersection of MCL and D determines employment, but the wage rate is determined by the height of the supply curve.
Wage
$400
MCL
S = ACL
D
$300 = w
0 L1 L2 Labor
Figure 16.10
APPLICATION 16.2
Until arbitration led to its dismantling in 1975, a provision in Major League Baseball known as the reserve clause lim- ited players to negotiating salary with the first team to sign them.4 That is, unless the first team to offer a player a con- tract terminated or traded him, he was at the mercy of the team’s owners. The reserve clause thus effectively tied each player to one team and eliminated competition between teams for individual players.
As the monopsony model predicts, when the reserve clause was in effect, players’ salaries were lower than their contributions to total revenues (as measured by gate receipts, broadcast fees, concession sales, and so on). In the late 1960s, for example, superstar hitters received an
Major League Monopsony
average salary of $68,000 per season but contributed $384,000 to their teams’ total revenues (net of any train- ing and transportation costs). Superstar pitchers averaged annual salaries of $86,000 but added $480,000 to their teams’ total revenues.
The dismantling of the reserve clause in 1975 and resulting competition between teams over individual players has caused a fly-up in players’ salaries. As of 1976, players who have served a team for five years are eligible to become free agents and sell their skills to the high- est bidder. Average annual salaries shot up by 175 percent between 1975 and 1980, from $53,000 to $146,000. By 2012, the average salary of a Major League Baseball player had risen to $3.2 million, as competition between teams over individual players further intensified and the value of players’ contributions to the output market also increased due to the growth in revenues earned by teams from tele- vision, cable, ticket sales, sports paraphernalia, and so on.
4Susan Lee, “The Baseball Strike Is a Monopolists’ Slugfest,” Wall Street Journal, June 30, 1981, p. 15; and Gerald W. Scully, “Pay and Performance in Major League Baseball,” American Economic Review, 64, No. 6 (December 1974), pp. 915–930.
market and be a monopsony in its labor market, although this situation is unlikely.) The
supply curve to the firm slopes upward—the graphical characteristic of monopsony—so the
marginal cost of labor curve, MCL, lies above the supply curve. Note that the average cost of labor curve, ACL, is identical to the supply curve if the firm can select any employment level along the labor supply curve. But once the level is selected, each worker hired is paid
the same wage. For example, if each worker is paid $300 per day, the average daily wage
per worker is also $300.
The firm maximizes profit by employing L1 workers. At that point the marginal cost of labor equals the addition to total revenue from hiring one more worker. The firm, however,
422 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 422Trim Size: 203.2 mm X 254 mm
does not pay an amount equal to the worker’s marginal contribution to revenue ($400).
Instead, the firm pays a wage of $300. The intersection of MCL and the demand curve determines the most profitable employment level, while the wage rate is determined by the
height of the supply curve at the corresponding level of employment.
In comparison with competitive input market conditions, employment is lower under
monopsony and so is the wage rate. If there were competition in the input market in Figure
16.10, employment would be L2, and the wage would be higher, since the supply curve slopes upward. A similarity between monopsony and monopoly becomes apparent. An out-
put market monopoly restricts output to obtain a higher price; an input market monopsony
restricts employment to pay a lower wage. An output market monopoly is able to charge a
higher price because it faces a downward-sloping demand curve for its product; an input
market monopsony is able to pay a lower wage because it faces an upward-sloping input
supply curve.
SUMMARY
A competitive firm’s demand for any input slopes
downward. This can be shown either by focusing on
the marginal value product curve or by considering the
substitution and output effects of a change in the input’s
price.
A competitive industry’s demand curve for an input
is derived by aggregating firms’ demand curves for the
input, but this summation is not a simple one, since a
product’s price changes as total industry employment
and thus output vary.
In terms of input supply, the demander or set of
demanders needs to be specified. Firms that are perfect
competitors in input markets face horizontal input sup-
ply curves.
An input’s supply curve to an industry may be either
horizontal or upward sloping.
The supply curve of an input to the economy as a
whole may be very inelastic.
In competitive markets the interaction of supply
and demand determines input prices and employment,
although the relevant market is frequently broader than
a single industry.
When many industries employ the same input, the
input price tends to be equalized across them, and the
supply curve of the input to any single industry will be
very elastic; no single industry will have much effect on
the input price.
Firms that are monopolies in output markets also
have downward-sloping input demand curves, termed
marginal revenue product curves.
Compared with competitive firms, output market
monopolies have lower input demand curves. This out-
come does not necessarily imply, however, that an out-
put market monopoly will pay lower prices for inputs,
since it may face a horizontal input supply curve and
thus has to pay the market price for an input.
An input market monopsony is the sole employer of
the input, and so it faces the market supply curve. If that
supply curve slopes upward, the monopsony’s employ-
ment decision affects the price of the input: The monop-
sonist has market power in the input market.
Compared to the competitive outcome, the price and
employment of an input will be lower with input market
monopsony.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
16.1 The data in Table 7.1 relate total output to the amount of labor employed when the amounts of other inputs such as
capital are held constant. Use it to answer the following: If
the price of final output is $10 and the wage rate is $120, how
many workers will be hired? (Assume that the output market
is competitive.) Illustrate your answer with a graph similar to
Figure 16.1.
16.2 What factors (other than the wage rate) affect the amount of labor a firm that operates in perfectly competitive output
markets will hire? How will a change in each of these factors
affect the firm’s demand curve for labor?
Review Quest ions and Problems 423
C16.INDD 10:42:13:AM 08/06/2014 PAGE 423Trim Size: 203.2 mm X 254 mm
16.3 “The law of demand does not apply to professional base- ball players. Since each team already has the maximum num-
ber of players allowed on its squad, a reduction in the wage
rate that must be paid for baseball players would not lead to
any more being hired.” Evaluate this statement.
*16.4 If the demand for personal computers rises, the produc- tivity of workers engaged in making personal computers does
not necessarily increase. Why, then, does the demand for such
workers increase?
16.5 Distinguish between the short-run and the long-run sup- ply curves of geologists to the domestic economy. Which curve
will be more inelastic? Why?
16.6 Rank the following labor supply curves in terms of their elasticities. How does your answer depend on whether you con-
sider short-run or long-run supply curves? Explain your answer.
a. The supply of economists to the federal government. b. The supply of taxi drivers to Chicago. c. The supply of college professors to Ohio State University. 16.7 Discuss the determination of equilibrium input price and employment by a competitive industry. Concerning the equi-
librium, firms would prefer to pay less for an input; why don’t
they? Input owners would prefer to receive a higher price; why
don’t they refuse to supply the input unless the price is higher?
16.8 “Employers set wage rates equal to marginal value prod- ucts.” True or false? Explain.
16.9 “If the supply of labor increases and depresses wages in a competitive industry, this outcome will benefit firms at the
expense of workers.” True or false? Explain.
16.10 Writing about the nineteenth century, C. Vann Wood- ward observes: “There was nothing but the urging of con-
science and the weak protest of labor to keep employers from
cutting costs at the expense of their workers.” Analyze this
statement. Was it conscience that kept wages from being zero?
16.11 “College teachers are no more productive today than they were 50 years ago, yet they are paid three times as much today.
They are obviously not being paid according to their marginal
productivity.” Discuss.
16.12 If the demand for automobiles rises sharply, how will the price of refrigerators be affected? (Steel is an important input
in the production of both products.)
16.13 “Recently, the demand for DVD players has increased rapidly, while the demand for radios has hardly budged. There-
fore, the fact that workers are better paid in the DVD player
industry is not surprising.” Would you be surprised if the asser-
tion about wages turned out to be correct? Support your posi-
tion with a graphical analysis.
16.14 “If among many good-hearted employers there is one determined to exploit workers, the actions of the single
employer may suffice to neutralize the good-heartedness of the
rest. This is because if a single employer succeeds in paying
lower wages, his fellow employers may have no alternative but
to follow suit, or to see themselves undersold in the product
market.” Why is this statement wrong?
16.15 “If Mexicans are allowed to immigrate to the United States, they will take jobs away from U.S. citizens.” Evaluate
this statement.
16.16 Explain why an output market monopoly will employ more of an input when its price is lower in terms of the substi-
tution and output effects of the lower input price. (Hint: How would Figures 16.3a and 16.3b be different for a monopoly?)
16.17 Do output market monopolies cause unemployment?
16.18 What is a monopsony? Graphically, what distinguishes a monopsony from a competitive employer of inputs? What does
this difference imply for the relative levels of employment and
input prices under monopsony versus competition?
16.19 “Along a downward-sloping competitive industry’s demand curve for labor, such as the one depicted in Figure 16.4, firm
profits will be greater the lower the wage rate.” Explain why this
statement is true, false or uncertain.
16.20 Workers in to the Malevolent Association of Microeco- nomics Teaching Assistants Union at your school are disgrun-
tled over their salaries and would like to request a $100 increase
in their weekly wage. Is their request more or less likely to be
honored if alternative means of assisting faculty in the instruc-
tion of economics, such as study guides, YouTube videos, and
robots, become more plentiful? If the demand for the under-
graduate education provided by your school becomes more
price elastic due to greater competition from other schools?
16.21 Bad Breath, Inc., sells its output at $1 per unit into com- petitive markets. Bad Breath’s factory is the only employer of
labor in Gilroy, California (garlic capital of the world). It faces
a supply from competitive workers of QL = w, where QL is the number of workers hired per year and w is the annual wage. Each additional worker hired adds 1 less unit of output than was
added by the previous worker. The 30,000th worker adds noth-
ing to total output. Bad Breath must pay all workers the same
wage and, because it has to raise wages to get more labor, each
additional worker costs the company 2QL dollars per year. To maximize profit, how much labor should Bad Breath hire and
what wage should it pay? Does efficiency prevail in the Gilroy
labor market? If not, what is the size of the deadweight loss?
16.22 If nominal wages in the South are less than in the North, does this imply that the economic theory of labor mobility is
invalid? Explain why or why not.
16.23 Per capita income is 600 percent higher in the United States than in Mexico. No other two countries sharing a border
have a wider disparity in income levels. Explain why this leads
to immigration of workers from Mexico to the United States. Is
such immigration beneficial, on net, to the United States?
16.24 Due to OPEC, the price of jet fuel increased by over 700 percent between 1970 and 1980. Explain why airlines
responded hardly at all to the increase in the price of jet fuel
in the short run (an estimated demand elasticity of between 0
424 Employment and Pr ic ing of Inputs
C16.INDD 10:42:13:AM 08/06/2014 PAGE 424Trim Size: 203.2 mm X 254 mm
and 0.15). Also explain why the long-run response was more
substantial and involved substituting Boeing 737s and 767s
for 727s and 747s and investing in superior tracking sys-
tems whereby a take-off can be delayed until a landing slot is
assured at the destination city.
16.25 Studies find that controlling for other factors, university professors’ earnings tend to decline with experience. In other
words, the more seniority a faculty member has with a particu-
lar institution, the lower his or her salary after accounting for
other factors such as productivity, degree, and field of study.
Explain why this phenomenon may reflect monopsony power
being exercised by universities because of the greater moving
costs faced by more senior faculty members.
16.26 Suppose that firms in an industry use two inputs, labor and capital. If the price of labor increases, then the firms will
demand less labor and more capital. True, false or uncertain?
Explain your answer.
16.27 Economists Ross Eckert and Richard Leftwich have noted that in the early 1950s, over 60 percent of MBA gradu-
ates from leading business schools took their first jobs in
manufacturing and 10 percent in investment banking and
consulting. Nowadays, no more than 20 percent of MBA
graduates from leading business schools take their first jobs
in manufacturing and over 50 percent in investment banking
and consulting. Between the 1950s and today, furthermore,
MBA starting salaries in investment banking and consulting
have risen dramatically relative to starting salaries in manu-
facturing. Using a graphical analysis, explain why this phe-
nomenon is related to a decline in demand for manufactured
products and an increase in demand for services over the last
half century.
425
C17.INDD 10:38:4:AM 08/06/2014 PAGE 425Trim Size: 203.2 mm X 254 mm
CHAPTER 17 Wages, Rent, Interest, and Profit
The general principles discussed in Chapter 16 apply to the analysis of the market for any type of input. However, some special issues arise in conjunction with particular input
markets, and these deserve further attention. In this chapter, we extend the general analy-
sis to specific input markets to see how wages, rent, interest, and profit are determined.
Because labor earnings account for about 75 percent of total U.S. national income, we con-
tinue to emphasize labor markets.
17.1 The Income–Leisure Choice of the Worker In our discussion of consumer demand in Chapters 3 and 4, we assumed the consumer’s
income to be fixed. For most people, however, income is not fixed; among other things, it
depends on the decision about how much time to spend working. To investigate the worker’s
Learning Objectives
Investigate a worker’s decision concerning how many work hours to supply. Examine the income and substitution effects of a higher wage rate and whether the net result of a wage increase involves a worker supplying more work hours. Analyze the general level of wage rates. Explain why wage rates differ among jobs. Define what economists mean by the term rent. Explore selling or monopoly power in input markets and show how unions attempt to exercise such power in labor markets. Explain how the interest rate is determined through the interplay of the supply of and demand for capital. Investigate investment and the marginal productivity of capital. Describe the relation between saving, investment, and the interest rate. Overview why interest rates differ across specific credit markets.
Memorable Quote “What the hell has Hoover got to do with it? Besides, I had a better year than he did.”
—George Herman “Babe” Ruth, Jr., responding to a reporter in 1930 who questioned him on the appropriateness of his salary ($80,000) being greater than President Herbert Hoover’s salary ($75,000)
426 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 426Trim Size: 203.2 mm X 254 mm
decision concerning how many work hours to supply, we will assume that the individual
worker is paid a fixed hourly wage and can work any number of hours desired at that wage.
This analysis utilizes the indifference curve–budget line technique developed in
Chapters 3 and 4 for the analysis of consumer choice. In Figure 17.1, the vertical axis
measures the individual worker’s total weekly income, and the horizontal axis, from left
to right, measures the worker’s leisure time. The term leisure refers to the portion of the worker’s time when he or she is not receiving compensation by an employer. Any worker
has 168 hours a week available, 24 hours a day, 7 days a week. We divide this time into
two mutually exclusive categories, work and leisure. Working time plus leisure time per
week must equal 168 hours. In Figure 17.1, Z is the total time available, and the point L1 indicates that the individual consumes L1 hours of leisure and supplies the remaining time, ZL1 hours, to an employer as work effort. Measuring leisure in this way, from left to right, is therefore equivalent to measuring work effort (hours of labor supplied) to the left from
point Z. With income and leisure measured on the axes, a worker’s budget line reflects the
wage rate received per hour of work provided. In the diagram, the budget line is AZ, and it shows the combinations of income and leisure available to the worker. Note that the more
hours the individual works (i.e., the less leisure time consumed), the higher the worker’s
income. For example, if ZL1 hours are worked (leisure of L1), income is Y1, but if work effort increases to ZL2 hours (leisure decreases to L2), income rises to Y2. If no work is done (point Z), income is zero since we assume that the worker has no nonlabor sources of income. Also observe that the slope of this budget line equals the worker’s wage rate. For example, a movement from point F to point G indicates that the worker is providing one more hour of labor (giving up one more hour of leisure) and in return receives an additional
$20 in wage income. Thus, the hourly wage rate is $20.
Leisure the portion of a worker’s time when he or she is not receiving compensation from an employer
Income–Leisure Choice of the Worker Measuring leisure from left to right is the same as measuring hours worked from right to left from point Z. The budget line AZ has a slope equal to the hourly wage rate. The optimal point is E, with the individual working ZL1 hours and earning $800 per week.
Weekly income
A
B
E
G F
U3
U2
U1
0
$20 1 hr.
L2 L1 (128)
Z (168 hours per week)
Leisure
Leisure
Work
Y2
$800 = Y1
Figure 17.1
The Income–Leisure Choice of the Worker 427
C17.INDD 10:38:4:AM 08/06/2014 PAGE 427Trim Size: 203.2 mm X 254 mm
To the worker, both income and leisure are desirable economic goods. For a given
amount of leisure, more income is preferred to less, and for a given amount of income,
more leisure is preferred to less. As we pointed out in Chapter 3, whenever economic goods
are measured on the axes, indifference curves have their normal shapes (downward sloping,
convex, and nonintersecting). The slope of an indifference curve relating income and lei-
sure measures the willingness of the worker to give up leisure for more money income and
therefore indicates the relative importance of these two goods to the individual. Figure 17.1
shows three of the worker’s income–leisure indifference curves.
The optimal point for the worker is E since this point represents the most preferred com- bination of income and leisure from among those on the budget line. Work effort is ZL1 (40 hours per week) and weekly income is Y1 ($800). As usual, the optimal point involves a tangency between the budget line and an indifference curve. In this case the tangency indi-
cates that the subjective marginal valuation of the worker’s own leisure time is equal to the
market valuation of the individual’s work time, the wage rate. At point E the worker must be paid at least $20 to give up one more hour of leisure, since the marginal rate of substitu-
tion (MRS) between income and leisure is $20 per hour at that point. Note that the optimal point in Figure 17.1 does not result in the worker’s income being
maximized. For example, the worker could earn a higher income by working longer hours,
such as at point B. But the extra Y1Y2 income is worth less than the L1L2 hours of leisure time that must be sacrificed to earn the additional income, as shown by the fact that point
B is on a lower indifference curve than point E. The worker thus is better off by forgoing the higher income (point B) for the sake of more leisure (point E). At point B, the marginal value of leisure exceeds the opportunity cost of leisure (forgone income); thus the worker
will increase his or her utility by pursuing additional leisure until the MRS between income and leisure just equals the wage rate.
Is This Model Plausible? Our analysis rests on the assumption that workers can choose how many hours to work. A
common objection to the model is that workers don’t really have the ability to vary their
work hours. Most employment contracts specify that employees will work a certain length
of time—perhaps 35 hours per week, for example. Even if an individual employee prefers
a 30-hour week combined with a one-seventh reduction in pay, the employer may not pro-
vide that option. In short, the argument holds that the workweek is fixed by the employer
and is beyond the control of the individual worker. There is an element of truth to this
point, but it is a less serious criticism than it might appear at first glance.
One way to justify the assumption of variable work effort is to recognize that most
workers can, in fact, exercise some degree of control over how much they work, although
perhaps not on a daily or weekly basis. Overtime, vacation leave, leaves without pay,
moonlighting, sick leave, and early retirement are options available to many workers. At
a more basic level, each person has a range of options in selecting a job in the first place.
Some jobs entail long hours, some short, and some permit considerable variation in work
effort. For example, it is not unusual for entrepreneurs to average 80-plus hours of work per
week in start-up firms striving to bring new products to market. Many of the same entrepre-
neurs could opt for less demanding jobs in the nonprofit, public or private sectors (e.g., at
established firms) that, while requiring fewer hours of work per week, are associated with
less promising financial rewards.
Another justification for the variable work effort assumption is that at a more fun-
damental level the analysis may still be valid for many purposes, even if it is impos-
sible for a worker to vary his or her workweek even slightly. Although the employer
fixes the workweek, consider the economic factors that determine the level at which it
is set. Employers are profit maximizers, and it is therefore in their interest to cater to
428 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 428Trim Size: 203.2 mm X 254 mm
the preferences of workers, just as they are led by profit motives to cater to the prefer-
ences of consumers. If a firm’s employees prefer a 30-hour workweek, but the firm
requires them to work 35 hours, it will lose workers to other firms that do a better job of
satisfying the employees’ preferences. Competition for workers thus leads firms to set
workweeks that correspond to worker preferences. So, the assumption that workers can
choose how much they will work should yield a reasonably correct analysis, although it
does not precisely describe reality.
One qualification should be mentioned. Employers have an incentive to cater to work-
ers’ preferences on average, but not necessarily to each employee’s preferences. The tech-
nology of production requires most firms to have a common workweek for all employees
(though it can, and does, differ among firms). A fixed workweek means that workers with
preferences different from the group average will not like the workweek schedule. Thus,
our model may not be strictly accurate for any specific worker, but it does provide a basis
for analyzing work effort decisions involving groups of workers.
17.2 The Supply of Hours of Work Will workers work longer hours at a higher wage rate? The work–leisure choice model
can help answer this question. Figure 17.2 examines the effect of a higher wage rate for a
particular worker. When the hourly wage rate is $20, the budget line is AZ, and the work- er’s preferred point is E with ZL1 work hours supplied. Remember that we measure work effort from right to left in the diagram. If the hourly wage rate rises to $25, the budget
Worker’s Response to a Change in the Wage Rate A higher wage rate pivots the budget line from AZ to A′Z and work effort increases from ZL1 to ZL2. We show the income and substitution effects of the change in the wage rate by using the hypothetical budget line HH′ that is parallel to the A′Z budget line but just tangent to the initial indifference curve U1. The substitution effect, L1L3, involves more work, but the income effect, L3L2, involves less. In this case the combined effect implies greater work effort at the higher wage rate.
E1
Weekly income
$25
A′
E′
E
U2
U1
H
A
0
1 hr.
$25
1 hr.
$20 1 hr.
I
S
TE
H′ ZL3 L2 L1 Leisure
Figure 17.2
The Supply of Hours of Work 429
C17.INDD 10:38:4:AM 08/06/2014 PAGE 429Trim Size: 203.2 mm X 254 mm
line rotates about point Z and becomes steeper. The new budget line is A′Z with a slope of $25 per hour. Note that at the higher wage rate income is greater for any level of work
effort. Given the specific preferences of this worker, when confronted with the higher wage
rate, the new optimal point E′ involves an increase in hours of work, from ZL1 to ZL2. Does a higher wage always lead a worker to work more? The answer is no, and we can
see why by considering the income and substitution effects associated with a change in
the wage.
The substitution effect of a higher wage rate encourages a worker to supply more hours of labor. When the hourly wage rate rises from $20 to $25, the sacrifice for leisure con-
sumption is greater, since each hour of leisure now means giving up $25 in income instead
of $20. Since leisure has become more costly in terms of income lost, the worker is encour-
aged to substitute away from leisure toward income—that is, to work more.
An income effect is also associated with a higher wage rate but has an opposite result on work effort from the substitution effect. A wage increase makes the individual better
off, permitting the worker to reach a higher indifference curve. A higher real income tends
to increase the consumption of all normal goods, and leisure for most people is a normal
good. The income effect of a wage rate increase thus encourages the consumption of leisure
and leads the worker to work less. Because of the higher wage, the worker can afford to
work less; it is possible to work fewer hours and still achieve a higher money income than
before the wage increase.
Figure 17.2 shows that the substitution effect of a higher wage rate encourages more work, and the income effect encourages less work. The hypothetical budget line HH′ is drawn tangent to the worker’s original indifference curve U1. Its slope reflects the higher wage rate of $25. The substitution effect is shown as the movement along U1 from E to E1. Because leisure has become more costly, the worker consumes less, and work effort increases from ZL1 to ZL3. The income effect is shown as the movement from E1 to E′ when we allow the individual to move from the HH′ budget line to the parallel A′Z budget line, reflecting the increase in real income associated with the rise in the wage rate. Since leisure
is a normal good, the income effect involves more leisure, from L3 to L2, which is the same as saying it encourages less work—that is, ZL2 instead of ZL3. The total effect of the higher wage rate is the sum of the income and substitution effects. Although these effects operate
in opposite directions, in this case the substitution effect is larger, so the total effect is an
increase in work hours from ZL1 to ZL2.
Is a Backward-Bending Labor Supply Curve Possible? For the worker whose preferences are depicted in Figure 17.2, the supply curve of work
hours slopes upward, at least between wage rates of $20 and $25, since a higher wage
leads to a greater quantity of labor supplied. This outcome need not always be the case,
however. The income effect of a higher wage may be larger than the substitution effect,
resulting in a reduction in work hours at higher pay. The intuition behind such an alterna-
tive outcome is straightforward. Beyond some point individuals may prefer to work a little
less, take some time off, and enjoy the higher income made possible by a higher salary.
For example, 30 hours a week at a wage of $25 per hour means more income and more
leisure time with which to enjoy the income when compared to 40 hours at a wage of
$5 per hour.
Figure 17.3a shows an individual who will choose to work longer hours when the
wage increases from $20 to $25 but will then decide to work somewhat less when the
wage rises again. When the wage increases from $20 to $25, work hours increase from
ZL1 to ZL2, but at a wage of $30, hours worked falls to ZL3. (We have not separated the income and substitution effects in the diagram, but you may wish to do so.) Figure
17.3b shows the same information plotted as a labor supply curve of weekly work hours.
430 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 430Trim Size: 203.2 mm X 254 mm
The supply curve slopes upward between wage rates of $20 and $25 but bends backward
as the wage rate rises further.
A supply curve of work hours can thus be backward bending beyond some wage rate.
Note that a backward-bending labor supply curve does not depend on an unusual set of
circumstances, as does an upward-sloping demand curve; rather, it just requires that the normal income effect of a higher wage rate exceed the substitution effect. Leisure is a nor- mal good, so the income and substitution effects always work in opposing directions in the
case of a labor supply curve because the worker is a seller of labor services. A rise in the price of something an individual sells (such as labor services) has a positive effect on the
individual’s income and thus leads to greater consumption of all normal goods, including
leisure. (In contrast, when the price of something you purchase as a buyer increases, the effect on your income is negative. The negative income effect reinforces, rather than coun-
teracts, the substitution effect for a normal good.) Moreover, the income effect of a change
in pay is likely to be large relative to the substitution effect since most income derives from
providing labor.
The Market Supply Curve To go from an individual’s supply curve of work hours to the market supply curve, we
need only horizontally sum the responses of all workers competing in a given labor market.
Thus, the market supply curve can also slope upward, bend backward or show a combina-
tion of the two, as shown in Figure 17.3b. Theoretical considerations alone do not permit
us to predict the exact shape of the market supply curve. The historical empirical evidence
Weekly income
A″
E″
E
E′
A′
A
L2 L3
U3 U2
U1
L1 Z
(a)
Leisure0
1 hr.
$30
1 hr.
$25
1 hr.
$20
Hourly wage
E″
S
E
E′
$30
$25
$20
ZL1
ZL3 ZL2
(b)
Labor0
Figure 17.3 An Individual Worker’s Weekly Supply of Work (a) An individual’s choices of how much to work at three different wage rates are represented by E, E′, and E″. (b) These labor supply choices are plotted as the supply curve of weekly work hours. When the income effect exceeds the substitution effect, the supply curve becomes backward bending.
The Supply of Hours of Work 431
C17.INDD 10:38:4:AM 08/06/2014 PAGE 431Trim Size: 203.2 mm X 254 mm
based on the U.S. labor market, however, suggests that it slopes upward (at least for wage
rates near present levels), with an elasticity somewhere between 0.1 and 0.3.2 An elasticity
of 0.1 means that a 10 percent increase in the wage rate increases the quantity of labor sup-
plied by 1 percent. Such an inelastic supply curve slopes upward and is almost vertical.
A highly inelastic aggregate supply curve, as suggested by the historical empirical evi-
dence from the U.S. labor market, appears plausible. Casual observation suggests that the
amount of time most people work stays the same over moderate periods of time despite
changes in wage rates. If individual supply curves were sharply upward sloping or back-
ward bending, we would see substantial changes in individuals’ work hours in jobs where
market forces have produced large wage rate changes. Work hours per worker in most jobs
seem to be quite stable over time, suggesting that the effect of wage rate changes on work
hours is not pronounced. This result does not mean that people are completely unrespon-
sive to wage rate changes, only that the responses that occur are modest.
When should we use the aggregate labor supply curve? In the preceding chapter we
emphasized that the labor supply curve to a particular industry or occupation is likely to
APPLICATION 17.1
As recounted in a story in the Wall Street Journal:1 Randy Bryson and his brother-in-law Larry Fazioli are both medical professionals in their 40s who practice in Pennsylvania. The similarity ends there. . . . Dr. Bryson works four days a week, drives a Mercedes, and lives in a 4,000-square-foot house with a pool. He and his wife, who works part-time in the same practice, together take home more than $500,000 a year. . . . Dr. Fazioli . . . works between 55 and 80 hours a week, and his annual income of less than $180,000 has been stagnant or down the past few years. He drives a Chevrolet. The key to their different lives: Dr. Bryson is a dentist, and Dr. Fazioli is a family-practice physician.
In general, whereas dentists used to make less than every type of physician as of 1988 (e.g., general dentists made an average of $78,000 that year, two-thirds the level of the average internal medicine doctor), they now make more. In 2010, the average general dentist earned $220,000 versus $205,441 for general internal medicine doctors. The discrepancy in hourly average earnings is even greater when one factors in the average number of hours
An Example of a Backward-Bending Labor Supply Curve: The Work Effort Choices of Dentists Versus Physicians
worked: dentists tend to put in 40-hour weeks, and physi- cians average 50–55 hours worked per week, according to the American Medical Association (AMA).
The difference in hours worked between dentists and physicians appears to provide an example from the medi- cal field of a backward-bending labor supply curve. The difference in work effort has fundamentally resulted from the fact that dentists’ wages have been less impacted by managed care and third-party pressure to restrict pay- ments received. (About 44 percent of all dental care is paid for by patients out of their own pockets versus just 10 percent of all physician and clinical costs.) As the effec- tive price ceilings associated with managed care have put downward pressure on physician versus dentist wages, dentists have ended up working fewer hours than their physician counterparts at the higher hourly wages they are earning vis-à-vis physicians. The income effect associated with the higher wage rate that dentists can earn appears to exceed the substitution effect that would otherwise encourage dentists to work more hours than physicians. And, as recounted by the illustrative tale of two doctors by the Wall Street Journal, the average general dentist in the United States now appears to be better off than the aver- age physician since he or she both enjoys more leisure (i.e., works fewer hours per week) and more income than the average physician.
1“Tale of Two Docs: Why Dentists Are Earning More,” Wall Street Journal, January 10, 2005, pp. A1 and A9.
2For a representative survey of the empirical evidence, see U.S. Congressional Budget Office, An Analysis of the Roth-Kemp Tax Cut Proposal (Washington, D.C.: U.S. Government Printing Office, 1978).
432 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 432Trim Size: 203.2 mm X 254 mm
APPLICATION 17.2
While the previously cited historical evidence from the United States indicates that the aggregate labor supply curve is highly inelastic, recent cross-country studies sug- gest that the elasticity value may be appreciably greater than zero. Specifically, Nobel Prize-winning economist Edward Prescott of the Federal Reserve Bank of Minnesota and Arizona State University has analyzed labor market statistics from the Organization for Economic Co-operation and Development (OECD) and found that Americans aged 15–64 work 50 percent more than do the French, Italians, and Germans.3 The differences in labor supply are not related to cultural differences or to institutional factors such as unemployment benefits. Rather, differences in marginal tax rates explain virtually all of the cross-country differences in labor supply. If someone in France, Italy, and Germany works more and produces 100 additional Euros of output, that individual pays 60 Euros in taxes and gets to keep only 40 Euros for his or her own consumption. By contrast, the marginal tax rate on labor supply in the United States is only 40 percent—that is, an individual worker gets to keep 60 cents of each additional dollar produced.
The Prescott evidence runs counter to the conventional wisdom that the French, for example, prefer leisure more than do Americans or, conversely, that Americans like to work more. Indeed, the evidence, indicates that the labor supply in European industrialized countries such as France
Why Do Americans Work More Than Europeans?
either exceeded or was comparable to that of the United States from 1970 to 1974, when marginal tax rates across the different countries were equivalent.
A further study also points out the role of the untaxed “underground economy” across various countries and how it responds to the marginal tax rate. Germans, for example, spend the same amount of time working as Americans, but the proportion of nontaxable home work time to taxable market time is greater in Germany, where the marginal tax rate is higher. The Italians, moreover, are not necessarily working less than Americans—they are just not being taxed for some of their “underground economy” labor. The Ital- ian government, in fact, increased its measured output by nearly 25 percent recently in an effort to reflect the role of the underground sector.
The bottom line is that workers the world over appear to make similar choices about labor and leisure and that policymakers need to recognize the impacts of their actions on aggregate labor supply. As Prescott notes:4
In 1998, Spain flattened its tax rates in similar fashion to the U.S. rate cuts of 1986, and the Spanish labor supply increased by 12%. In addition, Spanish tax revenues also increased by a few percent. . . . Free European workers from their tax bondage and you will see an increase in gross domestic product. The same holds true for Americans.
3Edward C. Prescott, “Why Do Americans Work So Much More Than Europeans?” Federal Reserve Bank of Minneapolis Quarterly Review, 28, No. 1 (July 2004), pp. 2–13.
4Edward C. Prescott, “Why Do Americans Work More Than Euro- peans?” Wall Street Journal, October 21, 2004.
slope upward and be quite elastic. We must distinguish that type of labor supply curve,
however, from the one discussed here. In looking at the supply to an industry or occupa-
tion, we see that the quantity of labor services can increase sharply with a rise in the wage rate in that job, but the increase results mainly from an influx of workers from other jobs
or industries and not from a change in work hours of current workers. The supply curve of
labor to a specific job or industry therefore depends mainly on how the number of workers
varies with wage rates in those specific occupations.
The aggregate labor supply curve is used in cases when the movement of workers
between jobs is not likely to be significant but the possible change in the hours supplied
by workers in their current jobs is. For example, how would a 10 percent increase in all
wage rates affect the total quantity of labor supplied? Since all jobs will pay proportionately
more, people will have little incentive to change jobs. The only way the total labor supply
will increase, therefore, is if people work longer hours (which includes the possibility that
some people will enter the labor force for the first time at the higher wage rate). For this
type of aggregate analysis, which involves the total quantity of labor supplied across all
industries, the market supply curve of work hours is appropriate. We discuss an example
using this supply curve in the next section and offer more examples in Chapter 18.
The General Level of Wage Rates 433
C17.INDD 10:38:4:AM 08/06/2014 PAGE 433Trim Size: 203.2 mm X 254 mm
17.3 The General Level of Wage Rates In analyzing the determination of wage rates, it is convenient to divide the subject into two
parts: determination of the general level of wage rates and consideration of why they differ
among jobs. In this section we focus on the factors that influence the level of real wages, or
the average wage rate; we defer until the next section a discussion of the factors that cause
a variation in wage rates around the average.
Supply and demand are still the applicable concepts for investigating the level of wage
rates. The supply curve of labor indicates the total quantity of labor supplied by all persons
at various wage levels. The appropriate supply concept is therefore the aggregate supply
curve of work hours discussed in the previous section. This supply curve is probably quite
inelastic, like curve S1 in Figure 17.4. The aggregate demand curve for labor reflects the marginal productivity of labor to
the economy as a whole. Indeed, it is convenient to think of the wage rate as being paid in units of national output (each unit composed of the combination of goods consumed by
the average person) to emphasize that we are dealing with the level of real wage rates. In constructing the demand curve relevant for a particular time period, the following factors
are held constant: capital (including land, buildings, and equipment), technology, and the
skills, knowledge, and health of the labor force. If these factors are fixed, an increase in the
total quantity of labor is subject to the law of diminishing marginal returns. Consequently,
the aggregate marginal product curve slopes downward and is the aggregate demand curve
for labor.
At any particular time, if the supply curve is S1 and the demand curve is D1 as in Figure 17.4, then the (average) real wage rate is w1 and employment is L1. At the high degree of aggregation used in this analysis, the model is necessarily abstract and ignores a multi-
tude of factors that could influence the positions of the demand or supply curves. Yet it
highlights the importance of the productivity of labor in determining the level of wages.
National output and national income are two sides of the same coin, and with labor receiv-
ing about 75 percent of national income (output), it is clear that the factors that determine
the output level produced from a given quantity of labor play a central role in the analysis.
These factors are primarily (but not exclusively) technology, the skill level of the labor
force, and the amounts of other inputs, which in this example we refer to as capital.
Determination of the General Wage Level The aggregate demand curve for labor interacts with the aggregate supply curve to determine the general wage level. Over time, normally both supply and demand increase. If demand increases faster than supply, wage rates tend to rise over time.
Wage
w2 w1
0 L1
D1 D2
S1 S2 S3
L2 Labor
Figure 17.4
434 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 434Trim Size: 203.2 mm X 254 mm
This concept explains why real wages are so much higher in the United States than
in less developed countries: the (marginal) productivity of labor is greater. Marginal pro-
ductivity is higher because of the factors determining the position of the demand curve:
capital, technology, and skills. In the United States, the amount of capital per worker is
about $250,000, contributing to high average and marginal productivity of labor. Techno-
logical knowledge is superior in this country, and the U.S. labor force is well educated and
highly skilled by international standards. Consequently, output per worker in the United
States is much higher. Other factors, such as climate, efficiency of the economic system
in allocating the available resources, the degree of political stability, and attitudes toward
work, play a role in determining national real wage levels. We can analyze their influence
on the productivity of the labor force by examining the way they affect the positions of the
demand or supply curves.
Over time, both the demand and supply curves shift. Because of saving and investment,
the amount of capital tends to grow over time, and this growth shifts the demand curve
outward: more capital per worker at each employment level increases marginal productiv-
ity. Similarly, technological progress and improvements in the labor force’s skill level also
increase marginal productivity. On the supply side, the supply curve shifts to the right as
the population grows over time. Whether these changes lead to a rising or falling real wage
depends on how much demand shifts relative to supply. In Figure 17.4, if demand increases
relatively more, to D2, while supply increases to S2, the real wage goes up to w2. Such an outcome has been the experience in most industrial countries over the past two centuries.
Because of capital accumulation, technological progress, and the availability of workers with
greater skills, the demand for labor has increased faster than the supply, pulling real wages
and living standards up in the process. Unfortunately, this happy outcome is not inevitable.
APPLICATION 17.3
To most U.S. consumers, who had become accustomed to uninterrupted economic progress since the 1930s, the decade of the 1970s was a shock. Foremost among the decade’s economic woes was a failure of many families’ incomes to keep pace with inflation. Median family income in constant dollars increased by 33 percent in the 1950s and again in the 1960s. In the 1970s, however, real median family income increased less than 1 percent before taxes. After taxes were paid, many families lost ground.
What went wrong with the U.S. economy in the 1970s? Observers point to a variety of possible culprits: inflation, higher energy prices, reduced national saving and invest- ment, heightened international competition from countries such as Japan, increased government taxes and regulation, and others. The basic problem was a failure of real wage rates to rise as much as they had in previous decades. Thus, an explanation of the situation should focus on what hap- pened to the supply of and demand for labor. Focusing on supply and demand conditions doesn’t rule out the pos- sibility that some of the factors mentioned above could have caused wages to grow less rapidly, but it suggests that a major role may have been played by a factor that most commentators fail to recognize.
The Malaise of the 1970s
During the 1970s, the labor force in the United States grew dramatically. Following increases of 12 percent in the 1950s and 19 percent in the 1960s, the labor force grew by 29 percent in the 1970s—the largest increase on record in a single decade. Referring to Figure 17.4, we can see the consequences of such a large increase in the labor force. If the normal increase in supply due to labor force growth over a decade is from S1 to S2 (by about 15 percent, for example), then the increase during the 1970s would be shown as the much greater shift in labor supply from S1 to S3. With the same increase in demand from D1 to D2, the abnormally large increase in labor supply means that over that period wage rates will not rise as much as usual, or might not rise at all.
Table 17.1 provides data on the growth in the labor force and real hourly earnings for the 1950 to 1980 period. A large increase in the labor force in the 1970s was associated with a 2.3 percent decline in real wages over the entire decade. Simple supply and demand forces appear to have been at work.
The increase in the labor force in the 1970s was primar- ily the result of two factors. First, the number of young persons reaching working age during the decade was unusually large. The “baby boom” generation of the 1950s
Why Wages Differ 435
C17.INDD 10:38:4:AM 08/06/2014 PAGE 435Trim Size: 203.2 mm X 254 mm
17.4 Why Wages Differ From the market forces that determine the general level of wage rates, we turn now
to the question of why there is such wide variation in the wages received by different
individuals. In Chapter 16, we explained why wage rates across firms or industries tend
to equalize. That analysis depended on the assumptions that workers were identical and
that they evaluated the desirability of the jobs only in terms of the money wage rates.
Dropping these assumptions, as we must for a fuller understanding of labor markets,
suggests that wage rates can differ among jobs and among people employed in the same
line of work. People are different in the type of work they are both able and willing to
perform, and these differences on the supply side of labor markets produce differences
in wage rates.
Perhaps the best way to see what is involved is to take a hypothetical, but plausible,
example. Figure 17.5a shows the labor market for clerks, and Figure 17.5b shows the labor
market for engineers. Under competitive conditions, the intersection of supply and demand
curves in each market yields a wage rate for engineers that is twice that for clerks. We sug-
gest that these markets are in full equilibrium with no tendency for the wage rates to equal-
ize. Is this result possible, and if so, why? Why don’t some clerks leave their low-paying
jobs and become engineers, a movement that would tend to equalize wage rates in the two
Table 17.1 Size of Labor Force and Wage Rates, 1950–1980 Labor Force
(Millions) Increase over
Previous Period (Percentage)
Index of Real Hourly Earnings
(1971 = 100)
Increase over Previous Period
(Percentage) 1950 62.2 — 64.0 —
1960 69.6 11.9 81.4 27.2
1970 82.8 19.0 95.7 17.6
1980 106.9 29.1 93.5 –2.3
Source: Economic Report of the President, 1982.
had grown up. Second, the share of females in the labor force increased from 43 percent in 1970 to 52 percent in 1980. (The labor force participation rate of females had never exceeded 40 percent before 1966.) Each of these factors was significant in itself, but together they spelled an unusually large increase in labor supply. Labor markets, however, adjusted to accommodate this influx of workers, and employment increased by millions more than it had in any previous decade. The relatively large shift in supply meant that wage rates rose less than in previous decades.
What roles did inflation, energy prices, decreased sav- ing, and the other factors so frequently mentioned play in affecting real wage rates? Several caused the demand for labor to grow less rapidly than it had in previous decades and thus reinforced the tendency for wage rates to rise more slowly. For example, a reduced rate of investment and
a higher price of oil—an input purchased in large quantities from other countries—both depressed the rate of increase in labor demand. Theory suggests, however, that some of the other factors frequently mentioned, like inflation and higher imports of consumer goods from Japan, would have little effect, since they do not significantly affect produc- tivity. The exact quantitative contribution of each of these factors to the slowdown in the growth of real wages is an unresolved issue and has been the subject of some debate among economists.
The role of other factors notwithstanding, the massive increase in the number of workers in the labor force appears to have had a significant impact on real wages during the 1970s in the United States. Moreover, the supply–demand model appears to offer the correct approach to the question of what caused the decline in real wages during this period.
436 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 436Trim Size: 203.2 mm X 254 mm
occupations? Why don’t employers of engineers seek out workers presently employed as
clerks and offer them engineering jobs at better pay (but at wage rates slightly below wE) since their demand curve indicates that they would be willing to hire more engineers at a
lower wage?
Thinking about these questions suggests several possible answers. First, workers cur-
rently employed as clerks may prefer their jobs despite the financial difference; that is, they
don’t want to work as engineers. Second, acquiring the skills to become an engineer may
have a significant cost. The wage for engineers may not be sufficiently high to compensate
clerks for the training costs they would have to bear to become engineers. Third, even if
there were no training costs, clerks may not have the aptitude for science and mathematics
necessary to work as engineers.
Training costs as well as differences in workers’ abilities and/or preferences for particu-
lar jobs can lead to equilibrium differences in wage rates among persons and jobs; there is no tendency toward adjustments that would wipe out wage differentials due to such factors.
Now let’s take a somewhat more detailed look at these factors.
Compensating Wage Differentials Monetary compensation is not the only factor, and sometimes not even the most important
factor, influencing individuals’ job choices. People routinely make decisions to take jobs
paying less in monetary terms than they might earn elsewhere. Many academic economists,
for example, could earn 50 percent more working for government or industry but choose to
remain in a scholarly environment. Similarly, some people might agree to work at the same
job for less money if they could live in the Sun Belt instead of the Northeast.
When workers view some jobs as intrinsically more attractive than others, the forces of
supply and demand produce differences in the wages paid. These differences are called
Figure 17.5
Wage
wC
0 LC
D
S
Labor services of clerks
(a)
Wage
wE
0 LE
D
S
Labor services of engineers
(b)
Equilibrium Wage Differences (a) The labor market for clerks results in wage rate wC. (b) The labor market for engineers results in wage rate wE. Although the wage rate for engineers is higher, there is no tendency for these wage rates to equalize: they are equilibrium wage differences.
Why Wages Differ 437
C17.INDD 10:38:4:AM 08/06/2014 PAGE 437Trim Size: 203.2 mm X 254 mm
compensating wage differentials because the less attractive jobs must pay more to equal- ize the real (monetary and nonmonetary) advantages of employment across jobs.
We can illustrate the implications of differences in job attractiveness. Suppose a certain
number of potential workers are identical in their abilities to work as police officers or fire
fighters. At equal wages, they would all prefer to be fire fighters, perhaps because they
think police work is a thankless task. Only if the police wage is at least 25 percent higher
than the fire fighting wage will they choose to enter the police force. If market conditions
determine wages, then wages for police officers will be 25 percent higher than those for
fire fighters. Only then would the real wage, in the eyes of the workers, be the same for the two jobs.
Differences in money wages are necessary to equate the quantity of labor supplied
and demanded in different occupations when the nonmonetary attractiveness of jobs
differs. Some cities have ignored this fact and set the same wages for police officers and
fire fighters. The outcome? A surplus of fire fighter recruits and a shortage of police
force applicants.
compensating wage differentials differences in wages paid that are created by the forces of supply and demand when workers view some jobs as intrinsically more attractive than others
APPLICATION 17.4
Each year hundreds of “jumpers,” or “glow boys,” are hired by public utilities to fix steel pipes in aging power plants.5
The jumpers work for only 10 minutes at a time and are paid the equivalent of 12 hours of average maintenance wages in addition to free travel and per diems. The only catch is the job site: the pipes are located inside nuclear power plants,
Twelve Hours’ Pay for Ten Minutes’ Work
where radiation is so intense that a jumper can stay for only a limited period of time before, in industry parlance, “burn- ing out.” The maximum amount of radioactivity to which a jumper may be exposed is 5,000 millirems per year (the equivalent of 250 chest X-rays). The Nuclear Regulatory Commission, which sets the limit, estimates that if workers are exposed to even that level for 30 years, 5 percent will die of cancer as a result. The high pay earned by jumpers thus illustrates a compensating wage differential: riskier jobs pay more.
5“Ten Minutes’ Work for 12 Hours’ Pay? What’s the Catch?” Wall Street Journal, October 12, 1983, pp. 1 and 21.
Differences in Human Capital Investment Our ability to perform useful services can be augmented by training, education, and expe-
rience. People can become more productive workers, and more productive workers
receive higher wage rates. The process by which people augment their earning capacity is
sometimes called human capital investment. In these terms, human beings are viewed as capable of generating a flow of productive services over time, much as capital assets can.
When they bear the costs of training or education themselves, they are investing in their
own earning capacity by attempting to increase the productive services they can provide
as workers.
Education and training are much like other investments. Initially, people incur costs.
For college students, for example, the explicit and implicit costs include tuition, fees, and
forgone income. The payoff to the investment comes several years later; then, you find out
how profitable the investment was. If earning capacity has increased sufficiently, the higher
earnings in later years cover the initial investment cost. No student needs to be told that this
form of investment, like many others, is risky.
Jobs requiring large human capital investments tend to pay higher wages. The reason is
simple: if the wages weren’t higher, few people would be willing to incur the training costs.
human capital investment the process by which people augment their earning capacity
438 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 438Trim Size: 203.2 mm X 254 mm
The higher wages associated with highly skilled work are, in part, the returns on past invest-
ments in human capital. According to this view, it is no accident that college graduates on
average earn more than high school graduates or that neurosurgeons earn more than typists.
In terms of our labor supply–demand model, the supply curve tells us that the amount
of labor supplied to jobs requiring large investments in human capital will be forthcoming
only at higher wage rates. Thus, in Figure 17.5, the supply curve of engineers is positioned
higher (vertically) than the supply curve for clerks.
Differences in Ability Workers’ productive capacities depend not only on their training and experience (human
capital investment) but also on certain inherited traits. The relative importance of these
two factors is greatly disputed. For years people have debated whether genetic or envi-
ronmental factors are more important in explaining IQs. There is no doubt, however, that
inherited traits play a significant role in determining what we are capable of doing. No
amount of training could turn all of us into nuclear physicists, basketball stars, entertain-
ers, models, politicians or business executives. People differ in strength, stamina, height,
mental ability, physical attractiveness, motivation, creativity, and numerous other traits.
These traits, or the lack of them, influence the work we are capable of accomplishing.
Similarly, possessing abilities that are scarce is no guarantee of a high wage. For
example, the ability to wiggle your ears may be uncommon but is unlikely to generate a
high income. What matters is the supply of persons with abilities required to perform cer-
tain jobs relative to the demand for their services. If consumers were unwilling to pay to
watch athletes throw balls through a hoop, being able to slam-dunk in 12 different ways
would not command a million-dollar salary. Obviously, then, some human abilities are in
limited supply relative to demand, and, as a result, those endowed with such abilities can
command higher wages.
APPLICATION 17.5
Although investing in a college education involves some risk, the average return, net of any costs, appears to be both substantial and increasing.6 As of 2011, according to the U.S. Census Bureau, workers whose highest educational attainment is a college degree made 97 percent more than workers with only a high school diploma (an average of $58,904 per year for college degree holders versus $29,869 for workers with only a high school diploma). As recently as 1988, college degree holders made just 48 percent more than workers with only a high school diploma. Researchers estimate that as much as 30 to 50 percent of the relative wage gains made by college graduates in recent years reflect the spread of computer technology and the fact that
The Returns to Investing in a BA and an MBA
college-educated workers tend to be more proficient at using computers than their non-college-educated counterparts.
The Census Bureau does not track annual income for holders of an MBA degree. According to the Graduate Management Admission Council, however, the median full-time MBA student earned $85,000 (excluding sign- ing bonus) after completing MBA studies in 2012 versus $50,000 before obtaining the MBA. In addition, the annual wage growth for workers with professional degrees, includ- ing MBAs, has exceeded inflation by 2 percent over recent decades. By comparison, workers with only undergraduate degrees have seen their wages annually increase by 1 per- cent over inflation over the same time period. This differ- ence means that the expected annual earnings gap between workers with MBAs and those with only undergraduate degrees should more than double—in real terms—over the course of their respective professional careers.
6“The Payoff from Computer Skills,” Businessweek, November 3, 1997, p. 30; and “ROI MBA,” BizEd, January/February 2005.
Economic Rent 439
C17.INDD 10:38:4:AM 08/06/2014 PAGE 439Trim Size: 203.2 mm X 254 mm
17.5 Economic Rent In ordinary usage, rent refers to payments made to lease the services of land, apartments, equipment or other durable assets. Economists use the term differently. Economic rent is defined as that portion of the payment to an input supplier in excess of the mini- mum amount necessary to retain the input in its present use. The economic rent accruing
to suppliers in input markets is thus analogous to the concept of producer surplus in
output markets.
In the history of economics, the term rent was originally associated with the payments to landowners for the services of their land. To see why, let’s suppose that the supply of
land is fixed, so its supply curve is vertical. (This assumption is a slight exaggeration since
the supply of usable land is not absolutely fixed; without care, land can erode, become
overgrown or lose its fertility.) A vertical supply curve means that landowners will place
the same quantity on the market regardless of its price. Even at a zero price, the same
amount of land would be available. Thus, the minimum payment necessary to retain land
in use is zero, and any actual payment above zero exceeds the minimum amount necessary
to call forth the supply. All the payments to landowners therefore satisfy the definition of economic rent.
Figure 17.6 illustrates this point. The vertical supply curve of land interacts with
the demand curve to determine the equilibrium price. The price and quantity are speci-
fied per month to indicate that we are concerned not with the sale price of the land
but rather with the price for the services yielded by the use of land. The shaded area
indicates the monthly income received by land suppliers.7 The shaded area also equals
the economic rent received by landowners, because all payments for the services of
land exceed the (zero) amount necessary to have a supply of Q. The position of the demand curve determines the amount of economic rent. If demand increases, the price
of land services goes up and the shaded area becomes larger, but the larger shaded area
would still be all rent.
economic rent that portion of the payment to an input supplier in excess of the minimum amount necessary to retain the input in its present use
Economic Rent with a Vertical Supply Curve When the input supply curve is vertical, the entire remuneration of the input (the shaded area) represents economic rent, since the same quantity would be available even at a zero price.
Price
P
S
0 Q Land
D
Figure 17.6
7In practice, owners often do not receive the income from the use of their land in monetary form. Home- owners, for example, secure the services of the land on which their houses rest. Since they own the land, the income from its use is in the form of these services directly received, but these services have a monetary value determined by supply and demand.
440 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 440Trim Size: 203.2 mm X 254 mm
Because nineteenth-century economists regarded the supply of land as fixed, they
viewed payments to landowners as economic rent. Today, economists recognize that sup-
pliers of other inputs may receive economic rent as well. The prices received by the owners
of any input in fixed supply are entirely rent, and part of the prices received by inputs with
upward-sloping supply curves are also rent. Consider the supply of university professors.
For clarity, let’s examine the discrete case involving a small number of persons. The sup-
ply curve is then the step-like relationship in Figure 17.7. In equilibrium, five professors
are working: individuals A, B, C, D, and E. We assume that they are identically productive as professors but not identical in their abilities to perform other jobs.
Individual A will be willing to work as a professor if paid area A1. Even at a very low wage, he will opt for the academic life because his next best employment opportunity is
as a dishwasher, where he would earn very little. Individual B has a better alternative: she could sell used cars. She will work in academia if she receives at least area B1. And so we progress up the supply curve. The supply curve slopes upward because to attract
more people, colleges must bid them away from increasingly more attractive alternative
employments.
When this labor market is in competitive equilibrium, individuals A to E are paid a wage of w. A wage of w means that individual A receives an amount greater than the minimum necessary to induce him to work as a professor. Recall that the minimum amount is area A1, but he is being paid A1 plus A2, so A2 represents an economic rent. Individuals B, C, D, and E receive rents equal to B2, C2, D2, and zero, respectively. Individual E is paid just enough to induce her to supply her services as a professor. If the wage were any lower, she would
work elsewhere, so none of her earnings represents rent.
In this example, part of the earnings of professors (except for E) are economic rent, and the remainder is the payment required to keep the individuals from leaving their aca-
demic jobs and working elsewhere. The total rent received is the sum of the colored areas.
Whenever the supply curve of any input slopes upward, part of the payment to inputs will
be rent, as in this example. The more inelastic the supply curve, the larger the rent as a
fraction of the total payment to an input. In the extreme case of perfectly inelastic supply
(Figure 17.6), all the payment represents rent.
Economic Rent with an Upward-Sloping Supply Curve With an upward-sloping input supply curve, part of the payment to input owners represents rent. In this case individuals A, B, C, and D receive rent equal to areas A2, B2, C2, and D2, respectively.
Wage
w
0 2 3 4 5 6 College professors1
S
D
A2
B2
C2
D2
A1
B1
C1
D1
E1
Figure 17.7
Monopoly Power in Input Markets : The Case of Unions 441
C17.INDD 10:38:4:AM 08/06/2014 PAGE 441Trim Size: 203.2 mm X 254 mm
Figure 17.8
Wage
wU
wC
F
0 LU LC Workers
D
S
A
E
B
MR
The Effect of an Input Market Monopoly An effectively organized union represents a monopoly in the market for labor services. To maximize economic rent, the union selects the employment level where the marginal revenue, in terms of additions to the union’s wage bill, from having an additional worker hired (the height of the MR curve) equals the opportunity cost needed to induce the worker to work (the height of the S curve). The union employment level is LU at a wage rate of wU. The economic rent accruing to the union equals area wUABF.
17.6 Monopoly Power in Input Markets: The Case of Unions In Chapter 16 we examined the hiring of inputs by firms operating in either perfectly
competitive or monopoly output markets. We also explored the effects of buying or mon-
opsony power in input markets. We did not, however, address selling or monopoly power
in input markets. We do so here by focusing on labor unions. If effectively organized,
labor unions are an input’s sole supplier of the labor services of the union’s members and
thus have some influence over the wages union members are paid. Much as a monopolist
in a product market seeks a price and an output level so as to maximize profit, a labor
union seeks a wage and employment level so as to maximize the economic rent accruing
to its members.
Consider Figure 17.8, where the demand and supply of labor to a market are given
by the curves D and S. These curves depict the quantity of labor services demanded and
442 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 442Trim Size: 203.2 mm X 254 mm
supplied, respectively, at various possible wage rates absent any monopsony or monopoly
power. Under competitive conditions, LC workers are hired at a wage rate of wC. Now suppose that workers unionize and effectively cartelize the supply side of the mar-
ket depicted in Figure 17.8. By organizing, the union’s members acquire monopoly power
and can exercise that power by selecting any wage–employment level combination on the
demand curve for labor services confronting the union.8
What point on the demand curve for their services should a union’s members select if
they are interested in maximizing their economic rent? Analogous to the case of monop-
oly power in output markets, the union needs to take into account the marginal revenue
curve, MR, associated with the demand curve for its services. The marginal revenue curve represents the additional wages earned by the union membership as a whole when
an incremental worker is hired. The marginal revenue curve lies below the demand curve
at all employment levels since the hiring of an additional worker requires the union to
lower the wage level paid to all previously employed workers. The reduced payments
to previously employed workers must be subtracted from the wage earned by the incre-
mental worker (the height of the demand curve) to derive the additional wages earned
by the union as a whole (the height of the marginal revenue curve) when the incremental
worker is hired.
To maximize economic rent, the union compares the additional wages resulting from the
hiring of another worker (MR) with the bare minimum the worker needs to be paid to work in this market—the worker’s opportunity cost. The height of the supply curve represents
the bare minimum that needs to be paid to induce an incremental worker to work at various
possible total employment levels and thus is effectively the union’s marginal cost curve.
Much as profit in an output market is maximized by setting marginal revenue equal to mar-
ginal cost, the economic rent earned by a labor union’s members is maximized by selecting
an employment level, LU, where MR intersects S in Figure 17.8. Up to LU, every worker contributes more to the union’s total wage payments than it costs to induce him or her to
work. The wage rate set by the union-monopoly is wU, and the economic rent accruing to the union membership as a whole is depicted by area wUABF—the excess of the wage rate over workers’ opportunity costs from zero to the quantity of labor services sold, LU, by the union-monopoly.
Note that in Figure 17.8, fewer workers end up being employed as a result of the union
than under competitive conditions. This is because the union is presumed to be interested
in maximizing the economic rent accruing to its membership rather than the number of
workers having jobs. The economic rent from the monopoly outcome of LU and wU (area wUABF) exceeds the economic rent associated with the competitive outcome of LC and wC (area wCEF).
By restricting output in the market for labor and raising the prevailing wage rate, the
union illustrated in Figure 17.8 produces a deadweight loss similar to the one resulting
from monopoly in output markets. Triangular area AEB is the deadweight loss associated with the labor union. This can be seen by noting that in order to maximize economic rent,
the union restricts labor employment to LU from LC. Each worker between LU and LC can generate more in terms of the value of output produced (the height of the D curve) than it costs to induce the worker to work (the height of the S curve). These potential net gains are not realized, however, because a labor union interested in maximizing economic rent does
not care about the total value of the output produced by an incremental worker (the height
of the demand curve). Rather, such a union cares about the incremental worker’s contribu-
tion to the union’s total wage bill (the height of the MR curve).
8For ease of discussion we assume that the union cannot engage in price discrimination, securing different wages for its various members. When an input supplier can price discriminate, an analysis analogous to the one presented in Chapter 12 on price discrimination in output markets is applicable.
Monopoly Power in Input Markets : The Case of Unions 443
C17.INDD 10:38:4:AM 08/06/2014 PAGE 443Trim Size: 203.2 mm X 254 mm
APPLICATION 17.6
Labor unions in the United States have undergone dra- matic membership changes over the past half century.9 In the private sector, the percentage of nonagricultural wage and salary workers who are union members plummeted from 38 percent in the early 1950s to 7 percent by 2012. By contrast, over the same time period, public sector union membership skyrocketed from 10 to 35 percent.
What accounts for the different fortunes of private-and public-sector unions in the United States over the past few decades? Public-sector unions have grown in rela- tive importance due to changes in labor laws favorable to the unionization of government workers at all levels of government throughout the 1960s and 1970s. Prior to the 1960s, collective bargaining by most government workers was illegal.
In the private sector, the growth of the high-technology sector explains some of the overall decline in union mem- bership in the United States. Only a small fraction of the 5 million employees who work for computer semiconduc- tor and software companies are unionized. This is due to demand for such workers growing much more quickly, over the past 30 years, than the supply and companies promising generous salaries and stock options to recruit and retain workers—rewards that diminish the relative benefits workers can obtain through unionization.
Furthermore, growing international competition in cer- tain industries, such as automobile production, and the enhanced competitiveness of other deregulated indus- tries, such as trucking and airlines, have made the output demand curves for firms operating in these industries more price elastic. As we saw in the previous chapter, one of the factors that determines the elasticity of demand for an input (and hence the monopoly power of sellers of that input) is the elasticity of the demand for the final product produced by the input. For example, the monopoly power possessed by the United Auto Workers has been lessened as U.S. car manufacturers have faced growing foreign competition. In the 1950s, when foreign competition was more limited, the demand for General Motors (GM) cars was more price inelastic and any wage increase demanded by the United Auto Workers from GM
The Decline and Rise of Unions and Their Impact on State and Local Government Budgets
could more readily be passed on to final consumers and so agreed to by GM.
Rising unionization rates in the public sector have created some budgetary challenges and led to political wrangling over the past few years. Roughly 80 cents of every dollar in California, for example, goes to cover employee wages and benefits, and these costs have been rising appreciably. Between 2000 and 2010, spending on California’s state employees increased three times as quickly as revenues, crowding out spending on programs such as higher education, environmental protection, parks and recreation, and so on. Due to unfunded pension and retirement health care promises made to state employees, coupled with inappropriate pension-fund accounting that overestimated future investment returns, California now must deal with $550 billion of retirement debt. The cost of servicing this debt has been growing at over 15 percent annually since 2000. In 2010, California spent over $6 billion on retirement benefits for employees, an amount exceeding that spent on higher education.
Between 2007 and 2010, the private sector in Cali- fornia shed 1 million jobs, and the retirement accounts of private-sector employees declined in value. Over the same time period, however, public-sector employment remained virtually flat, and the defined retirement plans of state gov- ernment employees—which private-sector workers have to underwrite—rose in value. Whereas few Californians have $1 million in savings, that was the value, as of 2010, of the retirement package promised to public employees who opt to retire at 55, at a guaranteed inflation-protected $3,000 per month.
It is no wonder, then, that the growing wage and retire- ment packages promised to public-sector employees have brought so much greater scrutiny to public-sector union- ization and its effects. The state of Wisconsin effectively experienced a shutdown in early 2011, when protests by public employees erupted in the wake of efforts by the newly elected governor to pass legislation designed to strip the state workers’ union of its ability to collectively bargain for pension benefits.
Economists Robert Novy-Marx of the Simon Business School at the University of Rochester and Joshua Rauh of the Stanford Graduate School of Business estimate the size of state and municipal pension liabilities to be $5 trillion greater than what state and local governments have pres- ently set aside for the next 30 years. Novy-Marx and Rauh argue that unfunded state and local unfunded pension liabilities are the second most significant fiscal problem in the United States today.
9This application is based on “Public and Private Unionization,” Journal of Economic Perspectives, 2, No. 2 (Spring 1988), pp. 59–110; “Union Members Summary,” U.S. Department of Labor, Bureau of Labor Statistics, January 23, 2013; Arnold Schwarzenegger, “Public Pensions and Our Fiscal Future,” Wall Street Journal, August 27, 2010; and “$5 Billion Price Tag for Public Pensions,” USNews .com, November 8, 2012.
444 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 444Trim Size: 203.2 mm X 254 mm
17.7 Borrowing, Lending, and the Interest Rate This chapter has so far focused on labor markets. We now turn to the market for another
critical input, capital. We begin our discussion with a term, interest rate, that has two dif- ferent meanings in economics. Interest rate sometimes refers to the price paid by borrowers for the use of funds. When a person borrows $100 this year and must return $110 to the
lender a year later, the additional $10 is interest, and the ratio of the interest to the principal
amount, 10 percent, is the annual interest rate. Interest rate can also refer to the rate of return earned by capital as an input in the production process. When a person purchases a
machine for $5,000, and the use of that machine generates $500 in income each year there-
after, the rate of return is 10 percent. Economists designate both the return on loaned funds
and the return on invested capital as interest rates because there is a tendency for these returns to become equal.
To explore the determinants of the interest rate, let’s start with a simple example.
Suppose that no capital investment is taking place; that is, no investors are borrowing
to finance building construction or equipment purchases. Although there is no borrowing
for investment purposes, there are still people who wish to lend money and others who
interest rate (1) the price paid by borrowers for the use of funds, and (2) the rate of return earned by capital as an input in the production process
Some Alternative Views of Unions and an Assessment of the Impact of Unions on Worker Productivity While acknowledging the potential deadweight loss associated with unions, some econo-
mists have argued for a more benign evaluation of their role since they may be organized
to counteract monopsony power possessed by input-buying firms.10 For example, in this
view, the United Auto Workers union arose out of the need to protect workers’ rights and
wages from the actions of the “Big Three” automobile companies. Furthermore, to the
extent that unions set up effective grievance procedures and give workers a “voice” with
their employers, they may actually make workers more satisfied and productive on the
job (the height of the labor demand curve shifts upward if workers become more produc-
tive, all else being equal).
There is some empirical evidence that, controlling for other factors, union workers not
only receive higher wages but also are more productive than nonunion workers. Most econ-
omists, however, are skeptical that unions actually improve workers’ productivity. One
would expect the value of the final output produced by an additional worker to increase the
fewer the number of workers hired. Fewer workers have more of the other, nonlabor inputs
to share between themselves and thus, if the law of diminishing returns holds, have higher
marginal products.
In Figure 17.8, for example, the height of the labor demand curve reflects the value of the
final output produced by an incremental worker and could be used as a measure of worker
productivity. According to such a measure, workers are more productive at the union
employment level of LU than at the competitive level of LC. This does not imply, however, that the union has made each worker more productive. The height of the demand curve at an
employment level LU is just as high under a union-monopoly as it is with perfect competition. Once again it is important to distinguish between a movement along a given demand curve
(in this case a labor demand curve) and a potential shift of the entire demand curve.
10See, for example, “Unions Need Not Apply,” New York Times, July 26, 1999, pp. C1 and C14; and Richard B. Freeman and James L. Medoff, What Do Unions Do? (New York: Basic Books, 1984).
Borrowing, Lending, and the Interest Rate 445
C17.INDD 10:38:4:AM 08/06/2014 PAGE 445Trim Size: 203.2 mm X 254 mm
wish to borrow. Households whose current incomes are high in comparison with their
expected future incomes may be willing to lend to others and use the repayment of the
loan to augment their otherwise lower ability to consume in the future. Other households
may wish to borrow in order to consume more than their current income in the present,
repaying the loan in the future. These households are the potential suppliers and demand-
ers of consumption loans, and their interaction determines a price, or interest rate, for
borrowed funds.
The level of the interest rate affects both the quantity supplied and quantity demanded
of loanable funds. The suppliers of funds are saving—that is, consuming less than their
current income allows. As noted in Chapter 5, a higher interest rate has opposing income
and substitution effects on the amount saved. The substitution effect encourages less pres-
ent consumption (more saving) because each dollar saved returns more in the future. The
income effect, however, increases the real incomes of savers, encouraging more present
consumption and less saving. On balance, a higher interest rate may lead to more or less
saving. The supply of loanable funds from savers is likely to slope upward at low interest
rates but can become backward bending at sufficiently high interest rates. This curve is
analogous to the supply curve of hours of work.
On the other side of the market, the demand curve for funds from borrowers must slope
downward. To a borrower, a higher interest rate has income and substitution effects, both
of which reduce the level of desired borrowing. The substitution effect reflects the fact that
a higher interest rate makes it more expensive to finance increased consumption from bor-
rowed funds, inhibiting borrowing. The income effect reflects the fact that a higher interest
rate reduces the real income of borrowers, so they cannot afford to borrow as much.
Figure 17.9 illustrates the interaction of borrowers and lenders in the market for con-
sumption loans. The supply of saving (funds to lend) is SS, and the demand for these funds by borrowers for consumption purposes is DC. Equilibrium occurs when lenders provide $10 million to borrowers in the present time period, with borrowers agreeing to repay the
loans at an interest rate of 5 percent.
In this simple example, the funds saved are not used in a way that increases the out-
put of goods through capital accumulation. Yet saving is productive. It provides the
means for borrowers to have a consumption pattern over time that is more to their liking
than having to live within their incomes each year, and these preferences explain why
A Borrowing–Lending Equilibrium If there is no investment demand for funds, the demand for consumption loans and the supply of saving determine the interest rate. Here people borrow $10 million to finance present consumption, with the commitment to repay it plus 5 percent interest.
Interest rate (i)
SS
DC
5% = i
0 $10 million Consumption loans
Figure 17.9
446 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 446Trim Size: 203.2 mm X 254 mm
borrowers are willing to pay for the use of borrowed funds. A positive interest rate
emerges because lenders must receive compensation for sacrificing their use of the funds
for present consumption.
17.8 Investment and the Marginal Productivity of Capital In the consumption loan example, the only outlet for saver-supplied funds is borrowing by
other households to finance consumption. Now let’s expand the analysis to account for the
fact that saving also provides funds used to finance investments. Firms may borrow money
for the purpose of enlarging their stock of capital equipment, and they must compete for
the limited supply of saving with households that are borrowing to finance consumption.
Why are firms willing to incur interest costs to finance investments in capital? Basically,
the reason is that capital contributes sufficiently to production to repay the interest costs.
For example, if Robinson Crusoe fishes by hand, he may be able to catch 20 fish a week.
If he takes 1 week off to weave a net, he can then catch 25 fish a week with the net until it
wears out in 10 weeks. Unfortunately, Crusoe is going to get very hungry if he doesn’t eat
for a week. However, if he borrows 10 fish from his Man Friday, on the condition that he
pays back the 10 fish plus an extra five fish, he can catch an additional five fish a week.
The five additional fish caught per week for 10 weeks (50 fish) are a measure of the
gross marginal productivity of the net. Whether the net (the capital) is productive depends on whether the gain in output outweighs the cost of not fishing for a week so he could
weave a net. Since Crusoe could have caught 20 fish that week, the cost of the net is a sacri-
fice of 20 fish plus the five fish paid as interest on the 10 borrowed fish. (The 10 borrowed
fish are not a cost; he borrowed 10 and returned 10, so the cost is zero.)
By sacrificing 25 fish, Crusoe gains 50 fish over the life of the net, representing a gain
of 25 fish. In this example, capital—the net—is productive in the sense that its net (no pun intended) marginal productivity is 25 fish. When the capital’s net marginal productivity is positive, investment—the act of adding to the amount of capital—allows Crusoe to produce
a larger output even after “netting” out the cost of the capital. (Note as a brief aside that
Man Friday also benefits, as he picks up an extra five fish for his trouble.)
Next, we need a measure for the net productivity of investment in capital that allows us
to compare the productivities of various projects. The annual percentage rate of return is
convenient for this purpose. Suppose a machine costs $100 to construct this year and its use
next year will add $120 to output. For simplicity we will let the machine wear out after one
year’s use. Then the net gain is $20 in one year, an annual rate of return on the initial $100
investment of 20 percent. The 20 percent figure is the net marginal productivity of invest-
ment, and it measures how much the capital investment will add to output one year hence
per unit of present cost. It is essentially the rate of return on the investment. Denoting this
rate of return measure of productivity by g, we can calculate it from the formula:
C R g g
= + = +
/ ( )
$ $ / ( ),
1
100 120 1
or
(1)
where C is the initial cost and R is the resulting addition to output (capital’s gross marginal value product) the next year. When capital equipment yields services over more than one
year, the normal case, the principle is the same but the formula is slightly more complicated.
Suppose the equipment lasts for two years before wearing out and adds $60 to output after
the first year (R1) and $81 in the second year (R2). Then, we calculate the rate of return from:
C R g R g g g
= + + + = + + + [ / ( )] [ / ( ) ]
[$ / ( )] [$ / ( ) ].
1 2 2
2
1 1
100 60 1 81 1
or
(2)
gross marginal productivity the total addition to productivity that capital investment contributes
net marginal productivity the total addition to productivity that capital investment contributes, less the cost of capital
Investment and the Marginal Product iv i ty of Capita l 447
C17.INDD 10:38:4:AM 08/06/2014 PAGE 447Trim Size: 203.2 mm X 254 mm
Given the initial cost of the equipment and the contribution to output in each year, we can
solve this expression for g. In this example the rate of return is 0.25, or 25 percent per year.11
The Investment Demand Curve The net productivity of a capital investment, g, depends on the value of additional output generated (the R terms) and on the initial cost (the C term). In turn, the C and R terms depend on many factors, among them the size and skills of the labor force, the amount
and costs of other inputs such as natural resources, technology, and the degree of politi-
cal stability. When we hold these other factors constant, the net productivity of investment
also depends on the rate of investment undertaken per time period. The greater the rate of
investment, the lower the net marginal productivity of still more investment. Consider the rate of investment for the economy as a whole. As investment increases,
each additional dollar’s worth of capital adds less to output than the previous dollar’s worth
because of the law of diminishing marginal returns; more capital applied to a given labor
force and quantity of land causes its marginal product to decline. This factor reduces the
R terms in the formula and contributes to a lower rate of return. In addition, an increase in investment also means more demand for machinery, vehicles, buildings, and other types of
capital, which tends to raise their prices. This factor increases the initial cost of investment,
the C term, and also contributes to a lower rate of return. Taken together, these factors imply that a lower rate of return will be associated with a
greater amount of investment, other things being equal. The downward-sloping DI curve in Figure 17.10 illustrates this point. When investment is at a level of I0, g is equal to 15 percent. If investment increases to I1, the rate of return on invested capital falls to 10 percent.
The DI curve, indicating the rate of return generated by different investment levels, is the investment demand curve. This idea is easiest to understand if we suppose that firms and individuals finance their investments by borrowing. If the interest rate is 10 percent, invest-
ment expands to I1, where the return on the investment just covers the cost of the borrowed funds. At any lower investment level, the rate of return on investment will be higher (15
percent at I0, for example), yielding a pure economic profit to firms and individuals who are only paying 10 percent for the funds they are investing. An expansion in investment will
occur as long as the rate of return is greater than the cost of borrowed funds, and such an
expansion causes the rate of return to fall. Equilibrium results when investments yield a
investment demand curve the relationship between the rate of return generated and various levels of investment
11In the general case where equipment lasts n years, the formula is C = [R1/(1 + g)] + [R2/(1 + g)2] + [R3/(1 + g)3] +. . .+ [Rn/(1 + g)n].
Investment Demand Curve The DI curve shows the rate of return generated by alternative investment levels. It is the investment demand curve, and it shows the amount invested at various interest rates.
i,g
15%
10%
D I
0 I0 I1 Investment
Figure 17.10
448 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 448Trim Size: 203.2 mm X 254 mm
return just sufficient to cover the interest rate on borrowed funds, at I1 when the interest rate is 10 percent. Thus, the rate of return on capital investment (g) tends to equal the interest rate for borrowed funds (i).
Even if investment is not financed by borrowing, I1 still is the equilibrium investment level if the interest rate is 10 percent. For example, a firm with $1 million in retained earn-
ings can use this sum to finance an equipment acquisition, but it will not do so unless the
investment yields at least 10 percent. Why? Because the firm can lend the $1 million at the
10 percent interest rate; if the investment project yields less than 10 percent, the firm can do
better by lending funds rather than investing them. The opportunity cost of investing is the
same whether the funds are borrowed or acquired in any other way, so the 10 percent inter-
est rate guides investment decisions in either case.
Our discussion so far assumes that firms and individuals know the rate of return gener-
ated by investments. In most cases, however, investors do not know, since rates of return
depend on future outcomes. The investment demand curve reflects what investors expect the outcome of investment projects to be. After the fact, some investments expected to
yield 10 percent fail to do so, and others do better. For given expectations on the part of
actual and potential investors, the demand curve still slopes downward. A change in expec-
tations causes the entire curve to shift. If investors expect nationalization of industry or
violent revolution to occur in a country, the demand shifts far to the left, one reason firms
are reluctant to invest in countries with unstable political regimes.
17.9 Saving, Investment, and the Interest Rate Households with a demand for consumer loans, and firms and persons with investment
projects, compete for saver-supplied funds. So far we have discussed these elements sepa-
rately; Figure 17.11 brings them together. The consumer loan demand curve is DC, and DI is the investment demand curve. The horizontal summation of these curves yields DT, the total demand for funds supplied by savers. The total demand in conjunction with the supply
of savings determines the interest rate. At the market-determined interest rate of 10 percent,
consumer loans are L1 and investment is I1, with the sum equal to T. As drawn, the investment demand is much greater than consumer loan demand. This tends to be the real-world case.
The Equilibrium Levels of Saving, Investment, Consumer Loans, and the Interest Rate The total demand for funds supplied by savers, DT, is the sum of the investment demand curve, DI, and the demand for consumer loans, DC. The intersection of DT and SS determines the interest rate. Investment is I1, saving is T, and consumer loans are L1.
Interest rate (i )
10%
0 L1
SS
D I
D T = DC + D I
DC
I1 T Investment, saving, and loans
Figure 17.11
Saving, Investment, and the Interest Rate 449
C17.INDD 10:38:4:AM 08/06/2014 PAGE 449Trim Size: 203.2 mm X 254 mm
In this model, the gross saving of households and firms, T, is not equal to investment, I1; consumer loans account for the difference. Economists sometimes define saving as net of
consumer loans (T − L1), and in that definition net saving equals investment. (We have also ignored the government’s demand for funds to help finance budget deficits. Including that
factor would create another demand for funds supplied by savers.)
The Figure 17.11 analysis corresponds to the aggregate labor demand and supply model
since we are talking about the supply and demand for funds aggregated across all saving–
lending and borrowing–investment markets. It is important to recognize that there is not
just one market that determines the interest rate but many closely interrelated markets that
we are summarizing. Bond markets, stock markets, mortgage borrowing, credit card loans,
bank deposits and loans, investment of retained earnings by firms, and other markets are
all involved in this process of allocating funds, provided by a multitude of sources, among
different competing uses. We lose some detail by using an aggregated model (as we did
in the labor market model), but we gain the important advantage of emphasizing the com-
mon underlying factors affecting all the interrelated markets: namely, the willingness to
consume less than current income and the existence of profitable investment opportunities.
Investment in excess of the amount required to replace worn-out capital adds to the
stock of productive capital, which, in turn, increases the productive capacity of the
economy in subsequent periods. Figure 17.12 illustrates this effect. In the present year,
a society’s production possibility frontier (PPF) relating the attainable output of con- sumer goods and capital goods is FF′. The production of capital goods requires the use of resources (inputs) that can otherwise be used to produce consumer goods, so if more
capital goods are produced, fewer consumer goods are available, and vice versa. The mar-
ket forces summarized in Figure 17.11 determine where we are located on this frontier.
Investment of I1 in Figure 17.11 means an addition to the stock of capital of that amount and corresponds to a point such as K1 in Figure 17.12. With more capital available the following year, the productive capacity of the economy increases. When K1 is invested in year 1, the PPF becomes F1F1′ in year 2. If investment was greater in year 1—for instance, at K2 (point E′)—then the PPF would move out even farther, to F2F2′. The effect of capital investment on the position of the PPF in subsequent periods is due to the productivity of capital discussed in Section 17.8. (There are also other reasons why the PPF shifts over time—most importantly, growth in the labor force, investment in human capital, and
technological progress.)
The Level of Investment and Productive Capacity Greater investment in the present—E′ rather than E—means less current consumption—C2 versus C1—but greater capacity to produce goods in the future. When K1 is invested in year 1, the production frontier in year 2 is F1F1′; if K2 is invested, it is F2F2′.
Consumer goods output
F2 F1
F C1 C2
0 K1 K2
E′
E
F1′F′ F2′ Capital goods output
Figure 17.12
450 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 450Trim Size: 203.2 mm X 254 mm
Investing more in the present time period thus means that incomes and consumption are
greater in the future. Still, we should avoid drawing the conclusion that an ever-increasing
expansion in investment is desirable. To invest more, less must be consumed; that is, for
an increase in investment from K1 to K2, consumption must fall from C1 to C2. Thus, the cost of greater investment (yielding higher future income) is increased saving (reduced
consumption) in the present. The saving supply curve shows how much households
must be paid in return for providing funds for investment purposes. At the Figure 17.11
optimal point, households will supply more saving than T only if they receive an interest rate higher than 10 percent. Investors, however, will be unwilling to pay savers more than
10 percent for additional funds since the increased investment yields less than a 10 per-
cent return. As a result, the market is in equilibrium with investment of I1.
Equalization of Rates of Return There is a tendency for capital to be allocated across firms and industries so that the rate
of return is equal everywhere. This parallels the tendency of labor to be allocated so that
wages are equal. In the case of labor, however, we pointed out a number of reasons why
wages differ because of differences in the productivities and preferences of people. With
capital, fewer qualifications are necessary. A firm purchasing a machine is indifferent to
the source of funds for financing; one person’s money can purchase as much capital as
another’s. Moreover, a person supplying funds usually doesn’t care whether the funds are
used to finance a computer in the aerospace or construction industry; lenders care only
about the rate of return earned on the saving. There are, of course, some reasons rates of
return won’t be exactly equal, such as differing degrees of risk, but they are generally not
quantitatively as important as the factors that produce differences in wages.
The process by which rates of return tend to equalize is much like the process for labor.
Let’s say that two industries, E (for entertainment) and S (for steel), are initially in equilib- rium, earning the same rates of return on invested capital. Next, let an unexpected shift in
demand occur. Suppose consumers’ demand for E increases while demand for S falls. As explained in Chapter 9, the short-run effects are economic profits in industry E and losses in industry S. In terms of the return to capital, capital now earns an above-normal rate in industry E but a below-normal rate in industry S. (A normal rate of return signifies the rate of return investors can expect to earn on capital invested elsewhere and thus represents the
opportunity cost of capital.) The owners of capital in industry S have an incentive to shift their investments to industry E, where the return is higher. As capital (and other resources) moves from S to E, the output of industry E expands, and price falls until the industry earns a normal return (zero profit). The opposite happens in industry S: the industry con- tracts, investors withdraw capital, price rises, and once again the industry earns a normal
return. This process describes how industries adjust in response to a change in consumer
demand from the perspective of the market for the input, capital.
17.10 Why Interest Rates Differ Although we have been discussing the interest rate and its relationship to the rate of return on capital, you should know that this description is a simplification, just as it was
when we discussed the wage rate in the aggregate labor supply–demand model. There is, in fact, a range of interest rates, and the interest rate in an aggregate model is best thought of as shorthand for “the general level of interest rates.” In that sense the model
developed in the preceding sections helps to pinpoint important determinants of the level
of interest rates.
Why Interest Rates Differ 451
C17.INDD 10:38:4:AM 08/06/2014 PAGE 451Trim Size: 203.2 mm X 254 mm
Differences can exist in specific interest rates in equilibrium, although the differences
are less pronounced than differences in wage rates. The four most important reasons for
differences in interest rates are as follows:
1. Differences in risk. There is a possibility that a loan will not be repaid. The greater the risk that a borrower will default, the higher the interest rate that a lender will charge.
If there is a one-in-five chance of a default, for example, the lender will have to charge an
interest rate about 25 percent higher than for no-risk loans to receive the same expected
return. For this reason corporate bonds pay higher interest rates than do government bonds,
and loans secured by collateral (like home mortgages) involve lower interest rates than
installment credit.
2. Differences in the duration of the loan. Borrowers will generally pay more for a loan that does not have to be repaid for a long time since it gives them greater flexibility.
Usually, lenders must also receive a higher interest rate to part with funds for extended
periods. This is one reason why savings accounts, where the funds can be withdrawn on
short notice, pay lower interest rates than six-month certificates of deposit.
3. Cost of administering loans. A small loan usually involves greater bookkeeping and servicing costs per dollar of the loan than a large loan. Loans repaid in frequent
installments, such as automobile loans, also involve higher administrative costs. When
greater costs are associated with administering loans, the borrower must cover these costs
and thus pay a higher interest rate.
4. Differences in tax treatment. The way the tax system treats interest income and investment income is very complex, and it sometimes leads to divergences in interest rates
that would otherwise be more nearly equal. For example, the interest paid by state and
local governments on municipal bonds is not taxable under the federal income tax, but the
interest on otherwise comparable corporate bonds is. The after-tax returns are what guide
lenders’ decisions, and they will tend to be brought into equality, implying a difference in
the before-tax (market) rates of interest. The result is that state and local governments can
borrow at lower rates of interest than corporations.
Although these factors lead to divergences in interest rates in specific credit markets,
they should not obscure the common factors that affect all interest rates. For example, if the
public decides to save less (a leftward shift in the SS curve), all these rates would go up, as would the rate of return on invested capital.
SUMMARY
The aggregate labor supply curve identifies the total
number of hours of work that will be supplied to the econ-
omy as a whole at different wage rate levels. It is derived
from the income–leisure choices of individual workers.
Since the income and substitution effects operate in
opposite directions, the aggregate labor supply curve is
likely to be quite inelastic and probably bends backward
above some wage rate.
A major application of the aggregate labor supply
curve, in connection with the aggregate labor demand
curve, is in analyzing the determination of the gen-
eral level of real wages. The aggregate demand curve
is essentially the marginal product of labor curve. Its
position depends on the amount of other inputs available
(such as capital), the level of technology, and the pro-
ductive characteristics of the labor force.
Wage rates can differ among jobs and workers for
several reasons, with no tendency for the rates to become
equal. Equilibrium wage differences can be due to com-
pensating wage differentials, differences in human capi-
tal investments, and/or differences in ability.
Economic rent arises when an input supplier is paid
more than the minimum amount necessary to retain the
input in its present use. Supply and demand analysis
shows that part of an input’s price represents rent when-
ever the input is in less than perfectly elastic supply.
452 Wages, Rent , Interest , and Prof i t
C17.INDD 10:38:4:AM 08/06/2014 PAGE 452Trim Size: 203.2 mm X 254 mm
Borrowing–lending and saving–investment markets
interact to determine both the interest rate on loaned
funds and the rate of return on invested capital. In equi-
librium, the interest rate on loaned funds and the rate of
return on invested capital tend to be equal, ignoring any
differences in risk, tax treatment, and the like. The inter-
est rate thus acts to equalize the willingness of people
to give up present consumption for future consumption
(marginal rates of substitution) and the real net marginal
productivity of investment.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
17.1 Edie chooses to work 90 hours per week when the wage rate is $16 per hour. If she is offered time-and-a-half ($24
per hour) for “overtime work” (i.e., hours in excess of 90 per
week), will she choose to work longer hours? Support your
results with a diagram.
17.2 If workers differ widely in their preferred hours of work at a given wage rate would you expect to see all firms set the
workweek at the average preferred level? Why or why not?
*17.3 Jeeves works 40 hours a week at a $7.50 wage rate. He unexpectedly inherits a trust fund that pays him $200 per
week. How does inheritance of the trust fund affect his budget
line? Will he continue to work 40 hours a week? Will his total
income rise by $200 a week?
*17.4 Can a person choose to work so much less at a higher wage that her total earnings decline? Show the income and
substitution effects implied by that outcome.
17.5 “Kevin (a highly skilled businessman) earns $150,000 a year. If he were to return to his native New Guinea (where
there are more limited business opportunities), Kevin would be
able to earn only $5,000 a year. This proves both that Kevin
doesn’t deserve his high salary and that wages are arbitrarily
set in the United States.” Evaluate this statement.
17.6 In general, the aggregate demand for and the supply of labor increases over time. Can we predict what will happen to real
wage rates and employment over time? What factors are respon-
sible for the shift in the demand curve? In the supply curve?
*17.7 “Since only a few people have the ability to become nuclear physicists, the long-run supply curve of nuclear physi-
cists is vertical.” True or false? Explain.
17.8 Discuss the three reasons for equilibrium wage rate differ- ences given in the text. Which one, or more, accounts for differ-
ences in wage rates between engineers and elementary school
teachers? College professors and high school teachers? Basket-
ball superstars and basketball coaches? Doctors and lawyers?
17.9 If workers had identical preferences and were equally able to perform any job, would all wage rates be equal?
17.10 If Miley Cyrus’ (Hannah Montana’s) supply curve of hours of work to the entertainment industry is vertical, does this
imply that all of her labor earnings represent economic rent?
17.11 What do we mean when we say that capital is productive? How do we measure this productivity? What does productivity
have to do with the investment demand curve?
17.12 “In equilibrium, the interest rate equates the willingness of people to give up present goods in return for future goods
and the ability of the economy to transform present goods into
future goods.” Explain.
17.13 How does the interest rate serve to equalize the desired rate of saving and the desired rate of investment? Would this
function be served if the government placed an upper limit on
the interest rate lenders could charge? How would such a law
affect the amount of investment undertaken?
*17.14 In Figure 17.12, will the aggregate demand for labor in year 2 be higher if investment is K1 or K2 in year 1? Why?
17.15 If the union shown in Figure 17.8 wishes to maximize the total wage payments made to its members, what wage and
employment levels should be selected? Explain your answer.
17.16 Why might a medical doctor who has established a lucra- tive practice reduce her patient load and spend more time on
the golf course or tennis court? Explain using a graph of a
labor supply curve.
17.17 High-tech industries are not extensively unionized in the United States. Explain why this is the case, knowing that soft-
ware programming, semiconductor manufacturing, and even
customer service representative positions can be outsourced by
companies such as Intel, Microsoft, and Amazon to locations
in India, Singapore, and Malaysia. Rely on the model of input
market monopoly developed in Section 17.6 in your answer.
17.18 On average, television news reporters are better looking than newspaper reporters. Explain why this might be the case, relying
on the discussion in Section 17.4 regarding wage differences.
17.19 Workers in countries such as Sweden and Germany work fewer hours per week, on average, than do workers in the
United States. Does this mean that Swedes and Germans are
lazier than Americans? Explain how laziness could be inter-
preted in the context of the work–leisure choice model devel-
oped in this chapter. What other information might you want to
gather to determine whether Swedes or Germans were indeed
less work oriented than Americans?
17.20 A biotech start-up has to pay a higher interest rate on any funds it borrows relative to established pharmaceutical giant
Eli Lilly. Why would this be the case if both firms have access
to the same capital markets and lending institutions?
453
C18.INDD 10:36:38:AM 08/06/2014 PAGE 453Trim Size: 203.2 mm X 254 mm
CHAPTER18 Using Input Market Analysis
Some of the most interesting and important issues in economics center on the way input markets function. This is not surprising. People understandably have great interest in a mar-
ket to which they supply inputs, since the functioning of that market largely determines
their standard of living. As consumers, we have only a small stake in each of many product
markets, but as producers, we have a great interest in one or a few. For example, few con-
sumers are sufficiently motivated to stage demonstrations over milk price supports raising
the price of the milk they buy by 20 percent, but the prospect of a 20 percent gain or loss in
your salary will not leave most people so unmoved.
In addition to our interest in matters affecting our personal incomes, we are affected
by the operation of input markets generally. The way input markets work strongly influ-
ences an economy’s income distribution. Some people are poor because the wage rates they
receive are low; others are wealthy because their wage rates are high. Many public policies
have been designed to redistribute income toward low-income households, and input mar-
ket analysis is vital to understand the consequences of such policies. This chapter uses input
market analysis to examine institutions, policies, and practices affecting workers’ incomes.
These include the minimum wage, Social Security, discrimination, the NCAA cartel, and
immigration.
Learning Objectives
Analyze the effects of the minimum wage on the employment of unskilled workers. Determine the extent to which employers versus employees bear the burden of the Social Security program. Explore an important hidden cost of Social Security that results from the program’s long-run impact on saving and capital accumulation. Explain the benefits to firms from colluding in hiring an input through examining the NCAA cartel. Show how employment discrimination can affect wage rates and employment. Outline the benefits and costs of immigration.
Memorable Quote “This is no simple reform. It really is a revolution. Sex and race because they are easy and visible differences have been the primary ways of organizing human beings into superior and inferior groups and into the cheap labor in which this system still depends. We are talking about a society in which there will be no roles other than those chosen or those earned. We are really talking about humanism.”
—Gloria Steinem, U.S. journalist and feminist
454 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 454Trim Size: 203.2 mm X 254 mm
18.1 The Minimum Wage In 1938, Congress passed the Fair Labor Standards Act, which established a nationwide
minimum wage of $0.25 per hour. The federal minimum wage has been raised periodically
since its establishment, most recently to $7.25 in 2009. Furthermore, as of 2013, proposals
have been made by both President Obama and certain members of Congress to raise the
minimum wage to $9–$15 per hour. Of course, the consequences of the policy depend criti-
cally on how high the legal minimum is relative to prevailing wage rates in the economy.
In 1938, the minimum wage of $0.25 represented 40 percent of the average manufacturing
wage, while the current $7.25 minimum is 45 percent of the average wage in manufacturing.
As these comparisons may suggest, most workers are unaffected by the minimum wage
because their wages are well above the legal minimum. Only the most unskilled and low-
paid workers are directly affected by the minimum wage.
Most people think of the minimum wage as a policy designed to help the poor, specifi-
cally the “working poor.” If some people are poor because of low wages, requiring employ-
ers to pay a “living wage” seems a straightforward remedy. Some simple economic analysis
along with a bit of empirical evidence, however, suggests that a minimum wage may not be
the best way to help the poor. In fact, it is quite possible that the most recent increase in the
minimum to $7.25 actually reduced the total income of the poor!
To begin our examination of this policy, let’s assume that unskilled workers are identi-
cal in all relevant respects so that a single wage rate prevails; we will relax this assump-
tion later. Figure 18.1 depicts the labor market for unskilled workers. Because of the law’s
broad coverage (most, but not all, jobs are covered), an aggregate analysis is appropriate.
As a result, we can assume that the supply of labor for unskilled lobs is fairly inelastic, as
shown in the diagram. In the absence of the minimum wage, let’s assume that the wage
rate would be $5 per hour and total employment would be L1. Next, let the government impose a minimum wage of $7.25. In accordance with the law of demand, when forced to
pay a higher wage, employers will hire fewer workers. With the labor demand curve D, employment falls from L1 to L2. Discussions of the minimum wage frequently overlook this predictable response of employers to a higher wage rate. Economists refer to the
reduction in employment of L1L2 as the disemployment effect of the minimum wage.
disemployment effect the tendency of employers to respond to a higher wage rate by hiring fewer workers
The Minimum Wage A minimum wage of $7.25 per hour reduces employment of unskilled workers from L1 to L2 and increases the quantity of labor supplied from L1 to L3. The difference between quantity demanded and quantity supplied, L2L3, is the unemployment created by the minimum wage.
Wage
$7.25 = w1
$5 = w
0 L2
D
L1 L3 Unskilled labor
SFigure 18.1
The Minimum Wage 455
C18.INDD 10:36:38:AM 08/06/2014 PAGE 455Trim Size: 203.2 mm X 254 mm
Note also that at the $7.25 wage, the number of workers looking for work has increased to
L3, which exceeds the number employed, L2, by L2L3. This is the unemployment created by the minimum wage, which differs slightly (with an inelastic supply) from the disem-
ployment effect. This difference reflects the fact that at a higher wage more people may
enter the job market.
Does the minimum wage benefit unskilled workers? On the basis of this analysis, an
unqualified answer is not possible. Workers able to get jobs at the $7.25 wage do indeed
benefit, but other workers—those who lose their jobs or are unable to find jobs—are left
worse off. Some people argue that if the total earnings of unskilled workers as a group rise,
we should say they gain as a group. Note that the effect on total earnings depends on the
elasticity of demand, which is beyond the control of policymakers. If demand is inelastic,
total earnings will indeed rise. Whether or not the total earnings of unskilled workers rise,
however, some unskilled workers suffer a major loss due to the law.
We gain more insight into who is likely to lose jobs when we drop the assumption that
all unskilled workers are identical. For example, suppose that in the absence of the mini-
mum wage, some workers would be earning just $4 an hour, while others would be earn-
ing $7. With a minimum wage of $7.25, which group would suffer a greater reduction in
employment? Assuming that the demand elasticity is the same for both groups, clearly
those with the lowest wages would he hardest hit. This minimum wage represents an 80
percent wage increase for those initially earning $4, but only a 4 percent increase for those
initially earning $7. Employment therefore will decline much more for workers whose
wages are initially the lowest. In short, the most disadvantaged, unproductive workers are
the ones most likely to lose their jobs and be priced entirely out of the market by the mini-
mum wage. This finding is notable because presumably this is the group most in need of
assistance. Even if the total earnings of affected workers rise, the most vulnerable among
affected workers will be harmed.
Apart from those who lose their jobs, who else bears a cost from the minimum wage?
Most people assume that employers bear the cost of paying the mandated higher wage.
But this conclusion is unlikely to be true. The costs of the minimum wage are certain to
be spread widely through society in the form of higher prices for products produced by
unskilled workers and possibly lower prices for complementary factors of production. No
one knows exactly who bears these costs. But as we will see later, some of these costs will
fall on the poor themselves.
Further Considerations We ignored several complications in the preceding analysis. Among them are the following:
1. The reduction in employment can take the form of a reduction in hours each worker is employed rather than a reduction in the number of workers employed. In other words,
instead of one out of ten workers losing a job, each worker might be able to work only 90
percent as many hours as desired. Whatever is more profitable to employers determines
the outcome. Since overhead costs are associated with each worker hired, independent of
hours worked, employers will probably cut back on workers rather than on work hours.
Nonetheless, there is some evidence of reductions in hours per worker. For example, about
half of all low-wage workers work only part time—although this practice is surely, at least
in part, a matter of choice.
2. When the government requires firms to pay a higher money wage, employers will respond, if possible, by reducing fringe benefits of employment. The fringe benefits that
may be reduced include pensions, health insurance, and on-the-job training. Reducing
fringe benefits means that the real wage employers pay rises by less than the money wage.
Although employment will not fall as much, the intended impact of the minimum wage is
456 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 456Trim Size: 203.2 mm X 254 mm
mitigated: if employers reduce fringe benefits, the real wage (minimum wage plus fringe
benefits) may not change at all. The importance of this reaction is unclear, but it could
partly explain why low-wage workers tend to have such poor private pension and health
insurance coverage and the on-the-job training provided by employers to their workers has
fallen over the past several decades.
3. Our analysis assumed that the minimum wage law covers all unskilled jobs. Actually, not all employers must pay the minimum wage; 75 percent of all unskilled workers have
uncovered jobs. The Fair Labor Standards Act contains a long list of exemptions, including
tipped workers, seasonal recreation employees, charitable organizations, mom-and-pop
businesses, farm workers, and Samoan laborers. For the 25 percent of unskilled workers
in covered jobs, the analysis of Figure 18.1 holds: employment falls. With an uncovered
sector, workers who are unable to find jobs at the minimum wage in the covered sector
may seek employment in uncovered jobs. As the supply of workers to the uncovered sector
increases, the wages prevailing in that sector go down, thereby harming workers already
employed in the uncovered sector.
4. With a surplus of workers created by the minimum wage (L2L3 in Figure 18.1), employers can be more selective about whom they hire. If employers have prejudices
related to the gender, race, age, weight or religion of their workers, they are in a
better position to indulge their “tastes.” When there is a glut of workers from which to
choose, employers can more easily hire someone with characteristics they prefer. That
is, if there are more applicants than jobs, the cost of discriminating falls. Insofar as
employers have prejudices, the harmful effects of the minimum wage are more likely
to be borne by workers with characteristics considered undesirable in the eyes of
employers.
Does the Minimum Wage Harm the Poor? The minimum wage is often portrayed as a policy designed to help the poor, but as we
have already suggested, there is a more than negligible possibility that it harms the poor. In explaining this observation, it must be kept in mind that the federal government’s definition
of poverty is based on total family income, not on the individual workers’ wage rates. For
example, the poverty line for a family of four in 2013 in the contiguous United States was
$23,550 ($19,530 for a family of three); families with lower incomes were deemed poor.
You might think that a strong correlation exists between low wages and low incomes (and
hence poverty), but that turns out to be false. Most people who are paid low wage rates are
in families that are not poor, and most people in poor families who work are paid more than
the minimum wage.
Actually, most families are poor not because of low wage rates but because they have
no earnings at all. Among poor working-age adults, roughly 60 percent do not work at all.
Only one in seven poor working-age adults works full time. Considering the minority who
do work at least part time, approximately 80 percent earn more than the minimum wage of
$7.25. These facts tell us that a large majority of the poor are not directly affected by the
minimum wage since they either don’t work or already have a wage rate well above the
minimum.
In contrast, most of those directly affected by the minimum wage live in households
with incomes well above the poverty line: over 80 percent of low-wage workers are in fam-
ilies with incomes above their respective poverty lines. Many of them, such as teenagers
in families where both parents work, are in families with incomes in the upper part of the
income distribution.
Therefore, if the goal is to help the poor, the minimum wage is not well designed. As
economists Richard Burkhauser and Joseph Sabia put it: “The majority of the working poor
The Minimum Wage 457
C18.INDD 10:36:38:AM 08/06/2014 PAGE 457Trim Size: 203.2 mm X 254 mm
are not helped by a minimum wage hike and the vast majority of those who are helped do
not live in poor families.”1 But it gets even worse. Using evidence provided by Burkhauser
and Sabia, we can conclude that it is likely that the increase in the minimum wage to $7.25
in 2009 actually reduced the total income of the poor. Let’s see how that could happen.
Of course, we have already noted that if the labor demand curve is elastic, the mini-
mum wage will reduce total labor earnings, but we do not have to assume a large reduc-
tion in employment to conclude that the total real incomes of the poor will be reduced.
To show how this effect can occur, we will use some interesting evidence provided by
Burkhauser and Sabia in a recent study.2 They estimated how much the earnings of work-
ers would rise from an increase in the minimum wage to $7.25 (from $5.15) if there were
no negative employment effects. Low-wage workers would gain $18.3 billion in that case.
However, most of this gain would go to the nonpoor, for reasons we have just discussed.
Earnings of poor households rise by just $2.3 billion; only 12.7 percent of the benefits of
the minimum wage go to poor households. Remarkably, 42 percent of this $18.3 billion
benefit goes to families with incomes more than three times their poverty lines.
Figure 18.2 illustrates how Burkhauser and Sabia made their estimate. Assuming the
wage in the absence of the minimum wage is w and the minimum wage is w1, their estimate of $18.3 billion in extra earnings is shown by the area w1ABw. It is simply how much earn- ings would rise from the higher wage rate if employment remained at L1.
Now let’s see how the incomes of the poor would be affected if there was a small reduc-
tion in employment. We will now interpret Figure 18.2 as applying only to workers in poor
families, not all low-wage workers. Thus, area w1ABw is equal to the $2.3 billion gain in earnings for poor workers. However, let’s now assume that the minimum wage results in
a 5 percent decline in employment among poor workers, from L1 to L2 in the graph. This corresponds to a highly inelastic demand curve since the average percentage increase in the
wage rate is about 20 percent. (The average wage of workers earning between $5.15—the
old minimum wage—and $7.25 is about $6, so the $7.25 wage represents a 21 percent
1Richard V. Burkhauser and Joseph J. Sabia, “Why Raising the Minimum Wage Is a Poor Way to Help the Working Poor.” Employment Policies Institute, July 2004, p. 3. 2Richard V. Burkhauser and Joseph J Sabia, “Raising the Minimum Wage: Another Empty Promise to the Working Poor,” Employment Policies Institute, August 2005.
The Effects of an Increase in the Minimum Wage An increase in the wage from w to w1 raises workers’ wage income by area w1GFw (the extra amount unskilled workers that continue to be employed at the higher wage earn) less area FBL1L2 (the lost wage earnings of unskilled workers who lose their jobs due to the wage increase).
Wage
w1
w
0
D
B
F
G A
L2 L1 Unskilled labor
SFigure 18.2
458 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 458Trim Size: 203.2 mm X 254 mm
increase in the average wage of affected workers.) Thus, we are assuming that a 20 percent
wage hike reduces the quantity demanded by 5 percent, an elasticity of only 0.25.
Taking into account the reduction in employment, how much will the total earnings of
the poor rise? The increase was $2.3 billion (w1ABw) if there was no reduction in employ- ment, but GABF is (by our assumption) 5 percent of w1ABw, so this reduces the gain by $0.115 billion (0.05 × $2.3 billion). However, the unemployed workers not only lose the
wage increase (GABF) by becoming unemployed, but they also lose their previous earnings of FBL1L2. At an average wage of $6, this area is 4.8 times as large as GABF, which puts it at $0.552 billion. In total, then, the earnings of the poor rise by $2.3 billion minus $0.667 billion
(area GAL1L2), or by about $1.6 billion. Because we are assuming an inelastic demand, total earnings still rise when we take account of the reduction in employment, but by less than
Burkhauser and Sabia’s estimate, which was based on no change in employment.
If this were the only way the poor were affected by the increase in the minimum wage,
we would still conclude that the total income of the poor as a group would be increased by
$1.6 billion. But this gain is diminished by two other factors. First is a result of the way
taxes and welfare programs affect the poor. The poor won’t see their net incomes rise by
$1.6 billion because they will pay taxes (especially payroll taxes) on these extra earnings
and, more importantly, lose welfare benefits as a result of having higher earnings. Most
welfare programs reduce benefits provided to the poor sharply as the poor gain higher mar-
ket incomes. This, in combination with outright taxes on the earnings, means that at least
half of this $1.6 billion in extra earnings will be lost to the government, leaving the poor
with at most an increase in net income of $0.8 billion.
Second, the poor (as well as others) will pay higher prices for goods and services as
a result of the increase in production costs of firms from the higher wage rates they must
pay. The total $18.3 billion gain in wages for all low-wage workers will raise prices so that
consumers as a group end up losing about $18.3 billion in purchasing power. Only a small
part of this $18.3 billion economy-wide cost falls on the poor, of course. The poor, who
make up 12 percent of the total population, consume about 6 percent of all goods and ser-
vices, so we will assume they bear 6 percent of this $18.3 billion cost to the economy. This
means the poor lose about $1.1 billion in real income from the higher prices produced by
the minimum wage. Combined with the increase in real income of $0.8 billion calculated
in the previous paragraph, we conclude that when all these effects of the $7.25 minimum
wage are evaluated, the poor as a group ended up with lower total real income, a reduction
of about $0.3 billion.
These calculations are obviously not intended to be precise, but they do strongly suggest
that the poor as a group probably gain little or nothing from the minimum wage when all of
its effects are evaluated. Even a small reduction in employment is enough to tip the balance
in the direction of overall net losses for the poor. It is not surprising that most economists
are very critical of the minimum wage as an antipoverty policy.
The Minimum Wage: An Example of an Efficiency Wage? Some analysts have argued that the minimum wage represents an efficiency wage: an above-market wage serving to increase firms’ profits.3 According to this theory, paying an
above-market wage may be profitable to a firm if there are costs associated with searching
for, selecting, and training new workers. Facing these types of costs, a firm may want to
pay higher wages to reduce labor force turnover and the costs associated with it. Addition-
ally, if there is imperfect information in labor markets and if a higher wage is likely to
attract a larger pool of more-qualified job applicants, a firm may find it profitable to raise
efficiency wage a wage higher than the prevailing market- determined level that serves to increase firms’ profits by lowering the costs of searching for, selecting, and training new workers
3George Akerlof and Janet Yellen, eds., Efficiency Wage Models of the Labor Market (Cambridge, England: Cambridge University Press, 1986).
The Minimum Wage 459
C18.INDD 10:36:38:AM 08/06/2014 PAGE 459Trim Size: 203.2 mm X 254 mm
wages above the level needed to attract qualified applicants. The increased quality in job
applicants may more than offset the costs. Moreover, if hired workers are prone to moral
hazard (see Chapter 14) and working below their fullest capacity upon being given an
employment contract, a wage exceeding the prevailing market level may induce a greater
increase in productive effort than the wage premium—provided that workers are motivated
by the higher pay to put forth the additional effort.
Henry Ford’s 1914 decision to institute a $5-per-day wage for workers in his automobile
factory, twice the prevailing market rate, is often cited as an example of an efficiency wage.
The benefits of the higher wage included 75 percent declines in worker turnover and absen-
teeism, a dramatic increase in the number and quality of applicants for positions in the fac-
tory, and a productivity improvement of 50 percent in the year following the wage increase.
The profitability of Ford’s company doubled in the wake of the wage hike.
Although Ford’s $5-per-day pay contract may be a striking example, the evidence that
a government-legislated minimum wage represents a similar efficiency wage is more tenu-
ous. Moreover, if the minimum wage is indeed an example of an efficiency wage, the ques-
tion arises of why the government has to legislate a minimum wage for unskilled workers
in the first place. An efficiency wage implies that firms have an incentive to pay an above-
market wage without any government intervention, since such a wage increases profits.
APPLICATION 18.1
Most of the economic research on the minimum wage has concentrated on its effect on employment. Does the minimum wage reduce employment and, if so, by how much? Researchers have conducted numerous stud- ies, and the evidence generally supports the proposition that the minimum wage reduces employment, espe- cially among the least skilled workers. But no consensus has been reached concerning the size of the disemploy- ment effect.
The debate over certain studies of the effects of an hourly minimum wage increase from $3.35 to $4.25 between 1990 and 1991 suggests why there is no general consensus on the extent of the disemployment effect of the minimum wage.4 The studies are controversial because of both their unorthodox approach and their findings. For example, economists Lawrence Katz of Harvard and Alan Kreuger of Princeton conducted a telephone survey of fast-food chain managers in Texas and found that only 11 percent of the surveyed chains reduced their employ- ment of nonmanagement workers following the minimum wage hike in the spring of 1991. The authors argue that the
The Disemployment Effect of the 1990–1991 Minimum Wage Hike
results indicate a fairly minor disemployment effect asso- ciated with the minimum wage hike.
Critics of the Katz–Kreuger study contend that it under- estimates the disemployment effect because not all other factors have been held constant. For example, the demand for labor appears to have been unusually strong in the period examined; as a result of rising exports to Mexico, the overall unemployment level in Texas declined from 6.9 to 5.8 percent. Moreover, three to four months may not be sufficient time to gauge the full effects of the mini- mum wage hike. While they may not fire existing workers, employers may reduce new hiring. Furthermore, the study does not account for the fact that the reduction in employ- ment may occur in the form of fewer hours per worker, reduced fringe benefits or an increase in employee pro- ductivity per hour worked. As one manager of several fast- food restaurants put it: “Our head count didn’t increase or decrease. . . . [We’re] just going to manage a lot more tightly to get more out of the people we’ve got.”
In contrast to the Katz–Kreuger results, a statistical study by Donald Deere and Finis Welch of Texas A&M and Kevin Murphy of the University of Chicago finds a more sizable disemployment effect associated with the 1990–1991 mini- mum wage hike. The Deere, Murphy, and Welch study finds that, holding constant other factors, the 27 percent increase in the minimum wage (from $3.35 to $4.25) decreased the employment of male and female teenagers by 12 and 18 percent, respectively.
4This application is based on “Forging New Insight on Minimum Wages and Jobs,” New York Times, June 29, 1992, pp. C1 and C7; Alan Reynolds, “Cruel Costs of the 1991 Minimum Wage,” Wall Street Journal, July 7, 1992, p. A14; and Gary S. Becker, “It’s Simple: Hike the Minimum Wage and You Put People Out of Work,” Businessweek, March 6, 1995, p. 22.
460 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 460Trim Size: 203.2 mm X 254 mm
18.2 Who Really Pays for Social Security? The U.S. Social Security system was begun during the Great Depression and is intended
to provide older citizens with a secure source of income after retirement. The payments
made to retired workers under the system are financed by a payroll tax composed of two
equal-rate levies, one collected from current employers and the other from employees.
In 2014, each rate was 6.2 percent. Workers earning $20,000 per year, then, have $1,240
deducted from their paychecks to cover the employee portion of the tax, and this amount
is matched by a $1,240 payment collected from their employers. The total tax is thus
$2,480.
By splitting the tax between current employers and employees, Congress apparently
intended to divide the burden of the Social Security program between them. Whether this
has actually been accomplished is far from clear. Economists believe that the way the
tax is divided into employer and employee portions has no effect on who actually bears
its cost.
Before discussing who actually pays the Social Security tax, we need to see whether
the division of the tax into employer and employee portions makes a difference to wage
rates and employment levels. We will compare two extreme cases: one where employees
pay the entire tax and the second where employers pay it. In Figure 18.3, before any tax is
levied, the supply and demand curves for labor are S and D, the wage rate is $10 per hour, and employment is L1. Now suppose the government levies a payroll tax on employers, which requires them to pay $2 to the government for each hour of labor they employ.
To understand how we incorporate the tax into the analysis, recall that the demand
curve for labor shows the maximum amount per hour that employers will pay for each
alternative quantity of labor. For example, the demand curve in Figure 18.3 means that
employers will pay a maximum of $10 per hour to hire L1 units of labor. With the tax in place, employers will still pay no more than $10 per hour for the quantity L1, but since they must pay $2 to the government, the amount that employers will be willing to pay for
L1 units of labor falls to $8 per hour. In the diagram the effect of the tax is thus shown as a vertical shift downward by $2 in the demand curve to D’. The downward shift in the demand curve means that with a $2-per-hour tax, employers pay $2 less to workers at each
level of employment. With the supply curve S, the tax reduces employment to L2, and the
Tax on Employers versus Tax on Employees A tax of $2 per hour of employment has the same effect regardless of whether it is collected from employees or employers. When collected from employers, it is analyzed with a $2 downward shift in demand to D′; when collected from employees, it is analyzed with a $2 upward shift in supply to S′. In both cases workers receive $8.50 per hour, and firms pay $10.50 per hour.
S
Wage
$10.50 = wB
$2
$2
$10 = w
$8.50 = wA
0 L2
S′ = S + T
D
D′ = D – T
B
A
L1 Labor
Figure 18.3
Who Real ly Pays for Socia l Secur i ty? 461
C18.INDD 10:36:38:AM 08/06/2014 PAGE 461Trim Size: 203.2 mm X 254 mm
wage rate paid to workers falls to wA, or $8.50. To employers, the cost of labor including the tax is now $10.50 per hour.
Alternatively, if employees pay the $2-per-hour tax entirely, the supply curve shifts ver-
tically upward by $2, or to S′, without affecting the demand curve. The shift in supply reflects the fact that workers must receive $2 more per hour to yield the necessary after-tax
wage to compensate them for supplying each alternative quantity of labor. For example, if
workers must pay the $2-per-hour tax, they will continue to supply L1 hours of labor only if they receive a net (after-tax) payment of $10 per hour. When the workers pay the tax, the
intersection of S′ and D determines the new equilibrium, involving employment of L2 and a wage rate of $10.50. Since workers must remit $2 to the government, their take-home
pay is $8.50.
Note that the real effects of the tax are exactly the same whether the tax is collected from employers or employees. When collected from employers, employment is L2, and firms incur a labor cost of $10.50 per hour, with $2 going to the government and the remain-
ing $8.50 to workers. When employees pay the tax, employment is again L2, and firms pay $10.50 as before; although the workers receive $10.50, they keep only $8.50 and must remit
$2 to the government. In both cases the $2-per-hour tax, distance AB, reflects the difference between the gross-of-tax cost of labor to employers and the net-of-tax payment to workers.
The way the tax is collected gives the government no control over who ultimately bears
its cost. The results are the same whether employers or employees pay the tax. Although
we have shown that the effects are identical for the extreme cases (when the employer or
the employee pays the entire tax), this conclusion holds for the intermediate cases, too. For
instance, if $1 of the tax is collected from employers and $1 from employees, firms would
pay $9.50 to workers (plus $1 to the government for a total unit cost of labor of $10.50, as
before), and workers would receive a gross wage of $9.50 and get to keep $8.50, since they
would turn over $1 to the government.
But Do Workers Bear All the Burden? We have just shown that the real effects of the payroll tax are the same regardless of its
division, for collection purposes, between employers and employees. This is not the same
as saying that workers bear all the burden of the tax. Note that in our example the $2-per-
hour tax led to a $1.50 reduction in the net wage rate ($10 to $8.50), so in that case workers
did not bear the full burden of the tax in the form of a lower wage rate. Most economists
specializing in tax analysis, however, believe that workers bear most, if not all, of the cost
of the tax in the form of reduced wages.
Exactly how far (net-of-tax) wages fall when a payroll tax is collected depends on the
elasticities of supply and demand. Of particular importance in the case of the payroll tax is
the elasticity of supply. In Figure 18.3, the labor supply curve is drawn as moderately elas-
tic, with lower wage rates leading to significant reductions in the quantity of labor supplied.
This is done to simplify the graph for the point being made. Now let’s look more closely at
the labor supply curve relevant for analyzing the payroll tax.
The payroll tax applies to virtually all jobs, industries, and occupations. Since it applies
across the economy, the relevant supply curve for use in analyzing the payroll tax is the
aggregate supply curve of hours of work. As explained in Chapter 17, the aggregate labor
supply curve is likely to be quite inelastic and, in fact, is vertical if income and substitution
effects exactly offset one another. Consequently, let’s examine the impact of the payroll tax
when the supply curve is vertical.
Figure 18.4 illustrates this case. In the absence of the tax, employment is L1 and the wage is $10. Next, the government levies a tax of $2 per hour. Since it makes no difference
if it is collected from employers or workers, let’s assume employers pay the entire tax. As
a result, the demand curve shifts vertically downward by $2 to D′. With a vertical supply
462 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 462Trim Size: 203.2 mm X 254 mm
curve the wage received by workers falls by $2—the amount of the tax—to $8. Employ-
ment, however, does not fall, because workers choose to supply the same number of hours
at the lower wage. When the supply curve is vertical, the net wage received by workers falls by the full amount of the tax, so workers bear the entire burden of the tax.5
Note that the cost of labor, including the tax to employers, has not risen; it is still $10.
Now, however, $2 goes to the government and $8 to workers, rather than $10 to workers.
Since labor costs have not risen, there is no effect on product prices. Much popular discus-
sion in the media of payroll taxes, especially the employer portion, holds that higher payroll
taxes add to labor costs and thus contribute to higher prices. The analysis here shows that
total labor costs do not rise: when taxes go up, wages go down.
This analysis suggests two conclusions, both important and routinely misunderstood
in discussions of the Social Security payroll tax. First, whether the tax is collected from
employers or employees makes no difference. The employer portion is no more borne
by business than is the employee portion. Second, if the aggregate supply of labor is
highly inelastic, workers bear all, or virtually all, of the tax burden in the form of lower
after-tax wages. In particular, workers bear the cost of the employer portion to the same
degree that they bear the cost of the employee portion. Many people are unaware of the
existence of the employer portion of the tax, but it depresses their take-home pay just
as much as the employee portion of the tax. These points are important to keep in mind.
Recent advocates of national health insurance, for example, have proposed financing the
program through increased payroll taxes, with employers paying a disproportionately
larger share. As our analysis shows, the cost of any health insurance scheme financed
through payroll taxes will be borne by workers in the form of lower wages, regardless of
how the tax is assigned.
5If the tax is collected from the workers, the supply curve shifts vertically upward by $2, which means that it does not visibly shift at all. Workers still continue supplying L1, and employers still pay $10 per hour. But after sending $2 per hour to the government, workers keep $8 per hour, just as when the tax is collected from employers.
The Burden of the Social Security Tax When the supply curve of labor is vertical, a tax on wage income reduces the hourly wage rate workers receive by the amount of the tax, from $10 to $8.
Wage
Total revenue
$10 = w
$8 = w′
0 L1
S
D
D′ = D – T
Labor
$2
Figure 18.4
The Hidden Cost of Socia l Secur i ty 463
C18.INDD 10:36:38:AM 08/06/2014 PAGE 463Trim Size: 203.2 mm X 254 mm
18.3 The Hidden Cost of Social Security Our Social Security system provides retirement benefits to the elderly and finances these
benefits on a pay-as-you-go (PAYGO) basis from payroll taxes collected from workers. In
other words, taxes paid by workers are not invested in a fund on their behalf, but are instead
transferred directly to retired persons—that is what PAYGO means. As we saw in Chapter 5,
this policy is likely to lead workers to save less for their own retirement since they antici-
pate receiving a Social Security pension. This reduction in saving has a series of further
consequences that we will now examine.
In the last chapter we explained how saving provides funding for investment projects,
leading to an increase in the stock of real capital—buildings, vehicles, computers, equip-
ment, and so on. Figure 18.5 reproduces in an abbreviated form our earlier model of how
the investment demand curve and saving supply curve interact to determine the level of
investment. (Compared to Figure 17.11, we have netted out the portion of the saving supply
curve that finances consumption loans, so the S curve in Figure 18.5 shows saving available to finance investment.) In the absence of PAYGO Social Security, the level of investment
is I1 and the interest rate (rate of return on investment) is i1. Now suppose PAYGO Social Security is introduced. That results in workers reducing
their saving for retirement, shown as a leftward shift in the saving supply curve to S′. (It is assumed, plausibly, that the retirement benefits provided to the elderly are consumed,
not saved, by them, so the reduction in workers’ saving implies a decline in total saving.)
The result is that investment declines from I1 to I2, and there is a small increase in the inter- est rate. (The small effect on the interest rate follows from having an elastic investment
demand curve, which is suggested by the empirical evidence.)
What the graph does not show explicitly are the long-run consequences of a reduction in
investment in real capital. Lower investment means that the capital stock of the economy
grows more slowly over time, and that in turn implies that the rate of economic growth will
decline. Future gross domestic product (GDP) will be lower than it would have been had
saving not fallen in response to Social Security.
The Effect of Social Security on Investment and the Interest Rate By reducing saving for retirement, the Social Security program results in a reduction in investment from I1 to I2 and an increase in the interest rate from i1 to i2.
i2
i1
Investment0
Interest rate S
DI
I1I2
S′ Figure 18.5
464 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 464Trim Size: 203.2 mm X 254 mm
It takes a long time for an economy to fully adjust to a PAYGO Social Security pro-
gram because workers’ saving for retirement is lower not just this year but each year in the
future. Ultimately, the future reduced saving by workers may be offset by a reduction in the
amount of negative saving (dissaving) by retirees, but until that happens national saving
and capital accumulation are lower as a result of Social Security.6 Let’s try to put a rough
quantitative perspective on how the policy would affect the economy over time.
The total capital stock in the United States today is roughly $48 trillion (i.e., $48,000
billion), and annual GDP is about $17 trillion. Suppose PAYGO Social Security is intro-
duced full blown this year (it didn’t exist previously), transferring $900 billion from
workers to retirees (which is approximately the current level when Medicare and the cash
pensions of Social Security are combined). As we have explained, saving and capital
accumulation will fall. We will assume that saving (equal to investment) would have been
$1,000 billion without Social Security, but when that policy is introduced saving (and
investment) falls to $600 billion, or declines by $400 billion. This decline in saving is less
than the amount of taxes collected ($900 billion), in contrast to our discussion in Chapter 5,
where we explained that a worker might reduce saving by the amount of the tax. However,
there are some factors ignored in that analysis, and the empirical evidence suggests that
saving will fall by somewhat less than the amount of taxes collected.
Given these data, how will the economy be affected next year? Without Social Security,
the capital stock would have grown from $48 to $49 trillion (due to the $1,000 billion
in saving), but with Social Security, the capital stock instead grows from $48 to $48.6 trillion.
Thus, next year the capital stock is about 0.8 percent lower than it would have been
had Social Security not been introduced. That means output will also be lower, but since
capital is only one of the factors of production (and labor is quantitatively more impor-
tant), the reduction in output will be smaller than 0.8 percent. A reasonable estimate is that
next year’s output will be 0.2 percent less with a capital stock of $48.6 trillion rather than
$49 trillion.7
This does not mean that next year’s output will be lower than this year’s, of course, only
that it will be lower than it would have been without Social Security. Since the economy
normally grows around 3 percent each year, the effect of Social Security over the next year
might show up as a growth rate of 2.8 percent instead of 3.0 percent—too small a differ-
ence to be easily detected. That is why we refer to this as a hidden cost of Social Security.
The Rest of the Story If this were the end of the story, it wouldn’t be worth telling. But saving and capital accu-
mulation will be lower in many future years as long as PAYGO Social Security continues,
and the loss in GDP grows over time. If the capital stock grows by $400 billion less each
year for 30 years, for example, the capital stock will be $12 trillion less at the end of that
time. That is large enough to imply that GDP will be about 6 percent less than it would
have been had Social Security never existed.
PAYGO Social Security has been around a long time, and the discussion above explains
how our current GDP may have been significantly reduced as a result of many years’
reduced saving and capital accumulation. Figure 18.6 illustrates how economic analysis
6Total saving equals saving by workers plus saving by retirees. In the absence of Social Security, saving by retirees is negative, as the elderly sell off their accumulated assets to provide retirement income. At first, Social Security reduces saving by workers but doesn’t have an effect on dissaving by retirees. But in later years, as people retiring have not accumulated as many real assets in anticipation of relying on the Social Security pension, the amount of dissaving also declines, which tends to offset the decline in saving by workers. 7The $400 billion reduction in capital would have added about $36 billion to next year’s output (at a 9 per- cent rate of return, roughly equal to the annual real return to capital), which is about 0.2 percent of the $17 trillion GDP.
The Hidden Cost of Socia l Secur i ty 465
C18.INDD 10:36:38:AM 08/06/2014 PAGE 465Trim Size: 203.2 mm X 254 mm
suggests the growth path of GDP over time would be affected by PAYGO Social Security.
(GDP is measured using logarithms so that a straight line implies a constant rate of eco-
nomic growth.) The line ABC shows how GDP would grow over time without any PAYGO policy. Instead, Social Security is introduced in, say, 1940 (that was the first year benefits
were paid), and as a result the economy grows as shown by the BDE path. The rate of growth is lower over the BD range, but ultimately returns to its former level after 60 years (point D: this is just assumed as we don’t know exactly how long this adjustment would take8). From the year 2000 onward, the economy is growing as fast as it would have had
Social Security not existed, but the level of GDP is lower in every subsequent year than it would have been had Social Security never been implemented.
This analysis thus implies that GDP is today lower than it would be had Social Security
never been introduced. How much lower is an empirical issue that is much disputed.
Several studies suggest that today’s GDP is perhaps 5 to 15 percent lower as a result of
Social Security.9
The Effect on Labor Markets One implication of the foregoing analysis of PAYGO Social Security may not be obvious:
how it affects labor markets. Consider our aggregate labor market analysis, as developed in
Section 17.3. Figure 18.7 is drawn to reflect current conditions, with a labor demand curve
shown by D and supply curve shown by S. Employment is at L1 and the wage is w1. Social Security affects the position of the demand curve for labor because that curve reflects the
marginal productivity of labor, and labor’s marginal productivity depends in part on the
capital stock. As mentioned above, the capital stock of the U.S. economy is about $48 tril-
lion, and that underlies the position of the D curve: the D curve is drawn on the assumption that workers have $48 trillion in capital to augment their productivity.
The Long-Run Effect of Social Security on GDP While the rate of growth of GDP ultimately returns to its former level, the actual level of GDP is lower in every year subsequent to the implementation of a Social Security system.
Year1940 20000
A
B
D
E
C Log GDPFigure 18.6
8The fact that the growth rate isn’t expected to be permanently lower is the result of the process described in footnote 6. 9See, for example, Martin Feldstein and Andrew Samwick, “The Economics of Prefunding Social Security and Medicare Benefits,” National Bureau of Economic Research, Working Paper No. 6055, June 1997; and Laurence J. Kotlikoff, “Privatization of Social Security: How It Works and Why It Matters,” in Tax Policy and the Economy, 10 (Cambridge, MA: MIT Press, 1996), pp. 1–32.
466 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 466Trim Size: 203.2 mm X 254 mm
Figure 18.7 is intended to represent the situation today, after we have had PAYGO
Social Security in place for more than half a century. But we have seen that one result of
Social Security is that the capital stock is, after a lengthy period of adjustment, substan-
tially lower than it would have been without that policy’s effect on saving. Some estimates,
for example, suggest that the capital stock would be 40 percent higher today (that’s $19.2
trillion more capital!) if we never had Social Security.10 If we had a larger capital stock,
there would be more capital per worker, worker productivity would be higher, and the
demand curve would shift outward. Thus, if we never had Social Security, the demand
curve for labor today would be D′ rather than D and wage rates of workers would be higher. In summary, workers today bear two significant costs from Social Security. One they
are well aware of: payroll taxes on their earnings.11 The other they probably don’t even
know exists: their before-tax wage rates are lower as a result of Social Security’s effect on
saving and capital accumulation in the past. Note that even if Social Security were elimi-
nated immediately, wage rates would not quickly rise from w1 to w2. They would rise, but the adjustment process would take many years to complete as increased saving would take
many years to add substantially to the capital stock.
18.4 The NCAA Cartel In Chapter 13, we explained why a group of firms has an incentive to reach a collusive
agreement, or a cartel, to limit production and raise the price of their product. In that chapter,
we examined a cartel in the output market; it is also possible to have a cartel among buyers
in input markets. The motivation is the same: by colluding, firms can make higher profits.
When firms collude in hiring some input, the goal is to reduce the price paid for the input,
which in turn lowers production costs. In effect, firms that collude are attempting to exercise
their collective monopsony power in input markets. If effective, it has a detrimental effect on
the price and economic rent received by suppliers of the input facing the buyers’ cartel.
The Effect of Social Security on Before-Tax Wage Rates Without Social Security, capital accumulation would be greater, the demand curve for labor would increase in magnitude (from D to D′), and the before-tax wage rate would be higher (w2 versus w1).
S
D
D′
Labor0 L1
w1
w2
L2
Wage
Figure 18.7
10Ibid. 11It is possible that the taxation of earnings affects labor supply in Figure 18.7, but we are ignoring that possibility. (If the labor supply curve is vertical, or very inelastic, the tax would have only a small effect on labor supply.)
The NCAA Cartel 467
C18.INDD 10:36:38:AM 08/06/2014 PAGE 467Trim Size: 203.2 mm X 254 mm
This section examines an effective input buyers’ cartel: the National Collegiate Athletic
Association (NCAA) regulation of student-athlete compensation. To fully understand the
effects of this regulation, we need to first discuss the theory relating to the exercise of mon-
opsony power in input markets by cartels.
An Input Buyers’ Cartel We’ll begin with a competitive equilibrium in an input market and then explain how collu-
sion among the firms would increase their profits. Figure 18.8b shows the market demand
and supply curves for a certain type of labor, with an equilibrium wage rate of w and employment level of L1. Figure 18.8a shows the equilibrium from the viewpoint of one of the competitive firms in the market. As noted in Chapter 16, in a competitive environment
the firm faces a horizontal supply curve, s, for the input at the market wage rate w, so with a demand curve of d, the firm employs l1 workers.
Each firm acting alone has no incentive to reduce the number of workers it employs,
since a single firm’s decision to hire fewer workers will not affect the wage—all that will
happen is that the firm produces less output and sees its profits fall. If all firms simultane-
ously reduce hiring, however, the combined effect of their actions reduces the wage if the
APPLICATION 18.2
Beyond the impact of Social Security on workers’ wages through payroll taxes and reducing saving and capital accumulation, there are two other important neg- ative political–economic consequences associated with the PAYGO manner in which Social Security is financed in the United States as well as in many other countries.12
For one, there is only a modest link between what an individual has paid into the system and the retirement benefits “guaranteed” by the government to him or her. The growing number of Social Security recipients have become an important political force lobbying to ensure their “fair share” of benefits. Indeed, the first several gen- erations of retirees received benefits that, in real terms, appreciably exceeded their accumulated Social Security tax payments.
For example, despite improvements in the physical and mental well-being of older persons, political pres- sures in many nations with PAYGO systems have led to restructuring of Social Security payouts to promote ear- lier retirements and thereby access to benefits before the originally established age of 65. In the United States, many retirements occur at age 62 or sooner. In Italy, retirements frequently occur in the mid-50s, and early retirement is common in other European countries with PAYGO financ- ing systems, such as Germany and Belgium.
Other Hidden Costs of PAYGO Social Security
Second, due to demographic factors, the Social Secu- rity Trust Fund in the United States (as well as in many other PAYGO nations) has been collecting more revenues in recent decades than it has been paying out. This is due to the fact that members of the baby boom generation are in their prime earning years, thereby leading to a higher ratio of paying workers to retirees. The “surplus” in the Social Security Trust Fund is counted toward the consolidated federal budget account and thus makes the reported defi- cit an underestimate of how much government spending in any given year actually exceeds tax revenues. During fiscal year 2010, for example, the U.S. government’s budget defi- cit would have been $120 billion larger than the reported amount of $1,300 billion were it not for the surplus of Social Security tax receipts over benefit payments.
By providing a near-term cushion, Social Security sur- pluses create a temptation for federal policymakers to increase government spending. This is especially true because many policymakers will no longer be in office and therefore will not have to face the consequences of more constricted federal government budgets when the Social Security annual surpluses turn into deficits as the baby boomers enter their retirement years. In addition, the future workers who will be supporting the baby boomers in their retirement years are not all of voting age at the present time, and thus their voices are not sufficiently factored into political decisions about how much of the Social Security surplus should be used to help underwrite current spending on other government programs.
12This application is based on Gary S. Becker, “Why I Support a Privatized Individual Account Social Security System,” Capital Ideas, Chicago Graduate School of Business, May 2005, pp. 4–6.
468 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 468Trim Size: 203.2 mm X 254 mm
market supply curve is upward sloping, as it is in this example. Lower wages (i.e., lower
production costs) are what an input buyers’ cartel tries to achieve. To show the conse-
quences somewhat more formally, assume for simplicity that the firms are identical and
that they agree to coordinate hiring decisions so they simultaneously increase or reduce
employment. The cartel agreement has the effect of changing the labor supply curve facing
each firm.
If one firm reduces employment, all other firms will as well, and the combined effect
depresses the wage. With coordinated hiring decisions, the individual firm faces the
upward-sloping supply curve s* in Figure 18.8a. This curve is each firm’s pro rata share of the total labor market supply curve S in Figure 18.8b. It reflects the firm’s average cost of labor at alternative hiring levels, provided that all firms coordinate their hiring decisions.
Because each firm’s labor supply curve is upward sloping with collusion, the marginal
cost of employing labor is greater than the average cost, or wage rate, as shown by mcL* in Figure 18.8a. Firms find that at their initial level of employment, the marginal cost of
hiring labor, Al1, is greater than the marginal value product of labor, El1. Thus, profits can be increased by reducing employment, since labor cost will fall by more than revenue. The
profit-maximizing employment level is where the marginal cost of labor equals the mar-
ginal value product—that is, where mcL* intersects the demand curve d. Firms therefore reduce employment to l2 and pay a wage rate of w′. This outcome is identical to the monop- sony equilibrium explained in Chapter 16.
A buyers’ cartel acts just as if a single buyer controlled hiring by all firms and acted
to maximize combined profit. This result is illustrated in Figure 18.8b, which shows how
the cartel, acting as a single buyer, employs L2 workers where MCL intersects the market demand curve. MCL and S are the sum of the mcL* and s* curves facing the firms when they coordinate their hiring decisions, reflecting the fact that the buyer side of the market has
Figure 18.8
d
Wage Firm
w s
B
A s**
E
w′
0 l2 l1 l3 Labor
D
Wage Market
w
S
MCL mcL
w′
0
(b)(a)
L2 L1 Labor
An Input Buyers’ Cartel Under competitive labor market conditions, the wage rate is w, with each firm facing a horizontal supply curve at that wage rate and employing I1 workers. If the firms form a cartel, total employment is restricted to L2, permitting firms to pay a lower wage, w′.
The NCAA Cartel 469
C18.INDD 10:36:38:AM 08/06/2014 PAGE 469Trim Size: 203.2 mm X 254 mm
restricted employment to pay a lower wage rate. That is, the firms are collectively behaving
as if they were a pure monopsony.
In this way, buyers can profit by colluding to force down prices of goods they purchase.
If this type of behavior is so advantageous to buyers, why aren’t buyer cartels more com-
mon? Just like seller cartels, buyer cartels are difficult to organize and maintain. In Chapter
13, we identified three reasons why seller cartels in output markets tend to fail; they are
relevant in this case as well.
First, firms have an incentive to cheat on the cartel agreement. Each firm, acting indi-
vidually, can raise its profit still further by increasing employment to l3 at wage rate w′ in Figure 18.8a. If many firms cheat, the cartel will fail. Second, participating firms will find
it difficult to reach agreement on the levels of permitted employment (l2) and the wage rate, w′. This is especially true when firms differ in size, structure, and aspiration. Third, the lower wage rate invites entry into the market by other firms that are not parties to the
cartel, making it more difficult for the cartel to restrict competition for labor and hold the
wage rate down.
Two additional factors, unique to input markets, hinder input buyers’ cartels. First, an
input market is often composed of many firms in several different industries. Coordinat-
ing hiring decisions among a large number of firms within and across industries is difficult.
Second, a firm usually hires many different inputs, and the potential profit from reducing the
price of only one input may be small. Typically, a firm stands to gain more by increasing the
price of what it sells (usually one or a few items) than by reducing the price of one input.
For these reasons the effective exercise of monopsony power by an input buyers’ cartel
is rare. There are, however, a few interesting examples, and we discuss one next.
The NCAA As a Cartel of Buyers The NCAA is a private organization empowered to regulate various aspects of college ath-
letics. Currently, roughly 800 colleges and universities belong to the NCAA. When founded
in 1906, the NCAA’s goal was to control violence in college football, but its powers gradu-
ally expanded over time. As athletics became a major source of revenue for many colleges,
the NCAA began to promulgate first voluntary guidelines and later rules governing recruit-
ment and financial aid. Today the NCAA limits the number of athletic scholarships schools
may award as well as the amount of financial assistance they may give. Basically, student-
athletes can receive assistance only for room, board, and tuition (although this oversimpli-
fies the NCAA’s 411-page rulebook).
In essence, the NCAA determines the maximum financial reward a student-athlete can
receive and the number of student-athletes who may be recruited with scholarships at each
school. If the rules are effectively enforced, the NCAA thus has the power to operate a
buyers’ cartel on behalf of member schools. We can see this by reinterpreting Figure 18.8
as it applies to the NCAA. The NCAA sets total “employment” by restricting the number
of student-athletes receiving scholarships, and it controls the maximum “wage” by limiting
assistance to room, board, and tuition. Each school is given a quota (l2) smaller than the number of student-athletes it would like to hire (l3) at the lower wage.
Applying a model developed for profit-maximizing firms to the nonprofit world of higher
education may seem strange, but there is no doubt that some colleges and universities try
to maximize the “profit” from their athletic programs. College athletic programs generate
annual revenues equaling over $2 billion in ticket sales and television rights. A winning
basketball or football team (the only true money-making sports) can provide enough rev-
enue to finance the entire athletic program, so it is not surprising that some schools treat
athletics as a business.
The NCAA rules clearly have the potential to be used to establish a monopsony-like
result for college athletics, but are the rules really applied strictly enough to produce
this outcome? We have several pieces of evidence that they are. First, let’s consider the
470 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 470Trim Size: 203.2 mm X 254 mm
marginal value products (or marginal revenue products) of college athletes. As shown in
Figure 18.8a, the wage paid to an athlete is below the athlete’s marginal value product (the
height to the demand curve, or Bl2) in a monopsony equilibrium. A number of estimates for selected college athletes have shown how much revenue they personally generated for their
colleges. Doug Flutie, for example, brought $3 million in television revenue during his last
two years on the football team at Boston College, and there are many other cases of college
“superstars” generating revenues in excess of $1 million a year for their schools.13 Clearly,
these athletes are paid far less than they are worth to their colleges.
A second piece of evidence is the significant number of NCAA rules violations by col-
leges and universities. Recall that any individual school has an incentive to cheat on the
cartel agreement, since superior athletes are worth more to the school than the NCAA
permits them to be paid. If the sports press can be believed, rules violations are common.
Reports of athletes receiving cash payments, cars, apartments, free tickets, jobs without
work requirements, and so on are widespread. Such instances of cheating all indicate that
the official wage permitted by the NCAA is below the market-clearing level.
Why doesn’t cheating cause the NCAA cartel’s demise? If cheating could not be con-
trolled, disintegration would be the likely result, but the NCAA can apply sanctions to
punish cheaters who are caught, and these sanctions have become increasingly severe. For
instance, the NCAA can reduce the number of athletic scholarships a school can grant. The
NCAA can also limit television appearances and ban a school from appearing in lucra-
tive post-season tournaments. An appearance in the national championship football bowl
game can be worth $20 million to a school (although these revenues must be shared with
many other schools in one’s conference). In an extreme example of the penalty the NCAA
can impose on a school, Southern Methodist University’s entire 1987 football schedule
was canceled. The NCAA clearly has powers enabling it to enforce, at least partially, the
employment restrictions necessary to achieve the monopsony outcome.
Eliminate the Cartel Restrictions on Pay? Should colleges be permitted to pay student-athletes? That question cannot be answered
using economics alone since, as the “should” in the question indicates, value judgments
as well as positive analysis are involved. The topic is too provocative and interesting to
ignore, however, so let’s consider some of the relevant issues. Positive economic analysis
indicates that NCAA policies result in a monopsony-like market for student-athletes. Thus,
we know that student-athletes are paid less than they are worth as revenue generators. So if
student-athletes are harmed by the NCAA restrictions on pay, who benefits? At a general
level, colleges and universities with big-time athletic programs gain: they receive sizable
revenues without having to pay the people largely responsible for generating them. But
since colleges are nonprofit institutions, tracking down the ultimate beneficiaries is difficult.
Perhaps nonrevenue-generating sports are subsidized from the profits, perhaps tuition is a
bit lower, or perhaps the salaries of coaches and economics professors are higher.
What is clear, however, is that those who benefit from the low pay of college athletes
are wealthier than those who are harmed by it. Student-athletes typically come from
poorer households than the average college student; often they come from quite disad-
vantaged backgrounds. One justification for the imbalance between revenue generators
and revenue receivers is that student-athletes will soon be making megabucks as profes-
sional athletes. This is incorrect; fewer than 1 percent of college athletes make it to the
pros. Thus, not only does the NCAA impede the smooth functioning of this labor market,
but those who are most obviously harmed are also often relatively poor. These income
13John Fizel and Rodney Fort, eds., The Economics of College Sports (Westport, CT: Greenwood Press, 2004). Fizel and Fort provide systematic evidence that the marginal revenue product is around $400,000 for a premium college football player and $1,200,000 for a premium college basketball player.
The NCAA Cartel 471
C18.INDD 10:36:38:AM 08/06/2014 PAGE 471Trim Size: 203.2 mm X 254 mm
distribution implications are the reason why certain analysts support breaking up the
NCAA cartel. Others remain unconvinced, so let’s briefly examine three of the most com-
monly heard arguments in favor of the current system:
1. Some schools would have to drop their athletic programs if they had to pay their athletes a competitive “wage.” Although this might be true for some schools, since more student-athletes would participate in college sports at a higher wage, many schools
could run the same size or expanded athletic programs. Nonetheless, someone must bear
the cost when schools lose “profits” made from not paying their student-athletes what
they are worth. If a school currently uses its basketball and football programs’ profit to
subsidize other sports, for example, these programs will suffer. The basic issue is whether
the student-athletes in profitable sports should bear the cost of supporting other, less
profitable, programs.
2. Paying college athletes would destroy their amateur status and turn college athletics into a business. Giving schools the option of paying student-athletes does not imply that every institution would have to do so. If a school wanted to maintain an amateur athletic
program, it could, and probably many would choose this option. Some student-athletes
might actually prefer this alternative, although schools choosing not to pay athletes a
competitive wage would have difficulty attracting the best athletes. Under the current
system, however, student-athletes are forced to be amateurs rather than having a choice.
3. Paying athletes might adversely affect the education they receive. Currently, fewer than one-third of athletes in the revenue-producing sports graduate from college, and
of those who do, many take unchallenging courses, so the situation is already cause for
APPLICATION 18.3
Whereas college athletes are restricted from earn- ing anything above a scholarship that can be revoked if the player does not perform well on the field, there are no limits imposed by the NCAA on what coaches can earn. The earnings of coaches at major colleges has grown sig- nificantly over the years as sports programs, notably foot- ball and basketball, have generated increasing amounts of television, ticket, and ancillary (paraphernalia, parking, food, and so on) revenues.14 The athletes whose talent dispro- portionately generates these revenues have remained out of the money, because the NCAA, in conjunction with col- lege presidents and athletic directors as well as conference commissioners, constrains their wages to zero (beyond any scholarship support). Economist Roger Noll of Stanford University notes that “the rising dollar value of the exploita- tion of athletes is obscene…out of control.”
Between 2006 and 2013, the average salaries of major college football coaches rose over 70 percent to $1.64 million. Major conference men’s basketball coaches whose
The Differing Fortunes of College Athletes and Coaches
teams made the post-season March Madness tourna- ment saw their salaries rise by 20 percent to $2.25 mil- lion between 2010 and 2012. By contrast, Texas A&M star quarterback Johnny Manziel, who was the first fresh- man ever to win the coveted Heisman Trophy in 2012, was forced to sit out the home opener of the 2013 season because he had been observed signing autographs earlier in the year. The NCAA and Texas A&M fined Manziel for failing to realize that brokers would end up profiting from his signature.
The hypocrisy of the Manziel case was exposed by ESPN analyst Jay Bilas who noted that while the NCAA prohibits schools from selling players’ jerseys with their names on the back, ShopNCAASports.com was listing a No. 2 Texas A&M jersey (Manziel’s number) for $64.95. A legal case also was also brought in 2009 by former UCLA basketball star Ed O’Bannon against colleges, conferences, and the NCAA for profiting from using players’ names on video games while players are prohibited from benefiting from such lucrative promotions.
Duke University economist Charles Clotfelter observed that “universities are quick to lecture society…but here is a situation where we’re not living up to our best selves.”
14This application is based on “Athletes at Big-Time Football Schools are Cheap Labor Who Bring in Millions. It’s Time They Got Paid for Their Work,” Time, September 16, 2013, pp. 36–42.
472 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 472Trim Size: 203.2 mm X 254 mm
concern. The joke about the university basketball coach telling a player who received four
Fs and a D, “Looks like you’re spending too much time on one subject,” may be a more
accurate reflection of the priorities of some college athletic departments than they care
to admit. Perhaps the goals of having exciting sports programs with superb athletes and
providing a quality college education to those athletes are incompatible. Whether or not
this is true, it is not clear that permitting colleges to pay athletes makes it harder for them
to require athletes to perform well academically and complete their education.
Although these remarks do not dispose of all the objections to permitting the wage rate
of college athletes to be determined in an open market, they suggest how an economic
approach to the question might help in understanding some of the issues.
18.5 Discrimination in Employment Discrimination can take many forms, but in this section we are concerned with only one
aspect of this complex issue: how employment discrimination can affect wage rates and
employment. Our main concern is whether, and under what circumstances, wage rates
of equally productive workers will differ because of discriminatory hiring practices on
grounds unrelated to a person’s productivity, such as race, religion or gender. If wage dif-
ferentials unrelated to productivity can result from discrimination, we need to qualify the
analysis in Chapter 16, where we concluded that wages have a tendency to equalize.
We will examine this issue by considering a labor market in which some employers dis-
criminate and others do not. To be specific, let’s assume that a fixed total of 2,000 equally
productive workers wish to work in this market. Of these workers, 1,200 are men and 800
are women. Half of the employers (group A) are prejudiced against women workers and will
not hire them under any circumstances; the other employers (group B) don’t care whether
their workers are male or female. What will be the wages paid to men and women in this
market, and how many will be hired by group A and group B employers, respectively?
To answer this question, it will simplify matters to assume that no employers dis-
criminate, identify a nondiscriminatory labor market equilibrium, and then work out the
consequences when group A employers begin to discriminate against women. In Figure
18.9, the demands for labor by groups A and B are shown in separate graphs as DA and DB, respectively. In the initial nondiscriminating equilibrium, workers will be divided between the two markets so that the wage rates are equal, as shown by the supply curves
SA and SB. The wage rate is w, or $10, in both markets, and 1,000 workers are employed by each group. Since this market does not include any discrimination yet, we assume
that men and women are employed by groups A and B in proportion to their total labor
market representation—that is, each group of employers hires 600 males and 400 females
(600M + 400F). When employers do not discriminate, equally productive men and women receive the
same wage rates. Now let us suppose that group A employers begin to discriminate against
women. They fire their 400 female employees. In the graph, the immediate, or very short-
run, effect of this action would be shown as a shift in the supply curve of labor to the A
market from SA to SA′ : the total supply is now 600 men. The wage rate in the A market paid to men will be bid up to wA, or $12. What happens to the 400 discharged women employees? Of course, they will look for work elsewhere—in this case, in the B market.
Their availability increases the total supply of labor to the B market, shown as a shift in
supply from SB to SB′ . The result is a lower wage rate of wB, or $8, and an increase in total employment in market B to 1,400. All of the women are now employed by nondiscriminat-
ing employers and none by discriminating employers.
Discr iminat ion in Employment 473
C18.INDD 10:36:38:AM 08/06/2014 PAGE 473Trim Size: 203.2 mm X 254 mm
This position is not, however, the final labor market equilibrium. Indeed, it may not materialize at all, as incentives for other changes simultaneously affect the behavior of mar-
ket participants. To see that the position described is not an equilibrium, consider the 600
men still employed in market B. They are now receiving an $8 wage rate—while the wage
rate in market A is $12 and market A is not discriminating against men. Thus, men in mar- ket B would look for jobs in market A. (Alternatively, employers in market A would try
to lure male employees away from market B because they wouldn’t have to pay the men
as much as their current $12 wage.) The shift of male workers from B to A is shown as a
rightward shift in SA′ and a simultaneous leftward shift in SB′ . This trend will tend to reduce the wage rate in A and increase it in B. How far will this process continue? As long as the
wage rate remains higher in market A and any men continue to work in market B, male
workers in market B have incentive to move to market A. The process continues until the
wage rates are the same in the two markets.
The final discriminating equilibrium occurs when SA′ has shifted to SA″ (coinciding with SA) and SB′ has shifted to SB″ (coinciding with SB). Wage rates are the same in both markets, and each market employs 1,000 workers—just as in the case when no employers discrimi-
nate. The difference appears in the composition of employment in the two markets. The
discriminating employers now employ 1,000 workers—all men; the nondiscriminating
employers now employ 1,000 workers—200 men and 800 women.
This effect is a remarkable, and little understood, implication of basic economic analysis.
It demonstrates that widespread discrimination can exist, and yet have no effect on the
Figure 18.9 Discrimination in Labor Markets When employers in market A begin to discriminate against women, the initial effect is a leftward shift in the total labor supply curve in market A, SA, and a simultaneous rightward shift in SB, as the discharged women move to the nondiscriminating market. This trend creates incentives for men in market B to move to market A, reversing the initial supply curve shifts. The final result is that wage rates of men and women are equal despite the presence of discrimination.
DA
Wage
$10 = w
0 Labor (total) [composition]
$12 = wA
SA′ SA = SA
600 [600M ]
1,000 600M 400F
[1,000M ] [ [
Market A (Discriminators)
DB
Wage
$10 = w
0 Labor (total) [composition]
$8 = wB
′SB = SB ″″
1,400 600M 800F[ [
1,000 600M 400F[ [
[ [
Market B (Nondiscriminators)
200M 800F
SB
474 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 474Trim Size: 203.2 mm X 254 mm
wage rates of the discriminated-against group. Discrimination will tend to produce segre-
gated employment patterns, but not wage differentials.
Can discrimination by employers ever lead to lower wages for discriminated-against
groups? In our model, this result is possible if the proportion of employers who discrimi-
nate is sufficiently large relative to the size of the discriminated-against group. Specifically,
if women represent 40 percent of the relevant labor force and discriminating employers
account for more than 60 percent of total employment, then discrimination could reduce
women’s wages (and raise those of men). But as this analysis makes clear, discrimina-
tion must be very widespread to affect wages. In the case of a small minority like African
Americans, who account for about 12 percent of the total labor force, discriminating employ-
ers would have to control an even larger portion of total employment (more than 88 percent
in that case) for discrimination to produce lower wage rates. The key issue is whether enough
nondiscriminating employers exist to absorb the discriminated-against group.
Even if discrimination is prevalent enough to produce lower wage rates for the
discriminated-against group, a second market force tends to limit its consequences—the
profit motive. Firms that engage in discrimination bear a cost in the form of sacrificed
profits. Suppose, for example, that widespread discrimination against women produces
lower wage rates for female workers. To be specific, assume that firms can hire male
workers for $35,000 per year and can hire equally productive female workers for $25,000
per year. For every male worker hired in this situation, a firm loses $10,000 in profit
because it incurs a higher cost than necessary. If the firm is a profit maximizer, it would
hire only female workers, which would give it a cost advantage over (discriminating)
firms that hire only males. This incentive for firms—to employ the lowest-cost available
inputs (of comparable productivity)—works to promote equality in input prices.
The stronger the profit motive, the less likely that discrimination will exist or influence
wage rates. Even a moderate difference in wage rates creates the opportunity for a firm to
greatly increase its profit by not discriminating. On average, labor costs account for approx-
imately 70 percent of all production costs, while before-tax profits average about 9 percent.
If the average firm could reduce its labor cost by just 15 percent, it could nearly double its
profit. How many businesses would be willing to forgo doubling their profits just to dis-
criminate against certain groups of workers?
These arguments do not mean that discrimination in employment is nonexistent or that it
cannot affect wages, particularly when we recognize that discrimination in hiring can reflect
preferences of consumers or co-workers rather than employers’ prejudices. (Employers
may discriminate against certain groups of workers if customers are likely to refuse to pur-
chase goods provided by that group of workers or if other workers are unwilling to work
with them.) The analysis does suggest, however, that discriminatory attitudes must be
widespread and that discriminators must be willing to bear nontrivial reductions in profit to
indulge their prejudices, if significant wage differentials are to be maintained.
What Causes Average Wage Rates to Differ? It is well known that incomes and earnings differ among groups. In recent years, for exam-
ple, the median earnings of African American males who work full-time have been about
79 percent as much as those of white males. Among full-time workers, the median earn-
ings of women have been about 76 percent as much as men. To many people, these and
similar numbers provide evidence of discrimination and its consequences. Our theoretical
analysis does not prove this view to be wrong, but it does suggest that we should consider
other possible explanations for these differences. Average wage rates of groups can differ
for reasons other than discrimination. Differences may exist in the average current labor
market productivity among groups, for example. Although productivity is notoriously dif-
ficult to measure directly, we can measure a number of characteristics thought to be related
to productivity.
Discr iminat ion in Employment 475
C18.INDD 10:36:38:AM 08/06/2014 PAGE 475Trim Size: 203.2 mm X 254 mm
The median age of African Americans, for example, is 30; for whites, the median age
is 38. Productivity and earnings tend to rise with age and experience, so whites would be
expected to have higher average earnings when the entire groups are compared. More than
half of all African Americans, but only one-third of white Americans, live in the South,
where wage rates (and living costs) are lower. The careers of African Americans, espe-
cially if they are older, may also bear the lingering effects of past discrimination by schools
and labor unions. On average, whites have slightly more schooling than African Americans
(though the difference has been narrowing in recent decades); 26 percent of whites but only
17 percent of African Americans have completed college, and only 15 percent of whites
but 22 percent of African Americans have failed to complete high school. Since education
is related to earnings, these discrepancies account for some of the difference in average (or
median) earnings. African Americans also score lower on standardized tests on average
than whites, and some believe that these test scores reflect productive potential better than
years spent in school. Even marriage patterns can play a role in determining earnings. The
percentage of African American men who are single is nearly double the percentage for
white men, and married men work longer hours and earn more than single men. These sorts
of differences between the groups would produce sizeable differences in earnings even if
there were no discrimination.
What about women? Perhaps the most important factor here is the different ways that
marriage and child-rearing affect men and women. Statistically, married men make more
money than single men, but married women make less than single women. Perhaps this
occurs because, historically, marriage has freed men of household duties and permitted
them to give more single-minded attention to their jobs, whereas it has had the opposite
effect for women. For whatever reason, among full-time workers, men work 10 percent
longer hours than women. When we compare annual earnings, finding that women earn 76
percent as much as men, this difference in hours means that women are earning 84 percent
as much per hour of work.
To test whether marriage and choices made within marriages cause much of the differ-
ence in average wages between men and women, we may consider only men and women
young enough that marriage and procreation are not so widespread. It turns out that young
women earn substantially more relative to young men than older women do relative to
older men. For example, among college graduates working full-time, and ages 18–24,
women earn 94 percent as much as men, and that is without any adjustment for differences
in college majors or hours of work. Another way to try to isolate the effect of marriage on
earnings is to examine the earnings of men and women who have never married. Among
never-married men and women who work full-time, are college educated, and ages 40–64,
women earn 17 percent more than men. In trying to determine whether wage differences are due to discrimination, we should
compare groups of workers who are equally capable, motivated, and experienced. Such a
comparison is not easy, given the numerous factors that affect the current earnings of any
worker. Using statistical techniques, economists and other social scientists have tried to take
into account easily measured factors like age, years of education, standardized test scores,
region, and hours of work. They have generally found that from one-half to 100 percent
of the overall differences in earnings between African Americans and whites, and between
men and women, can be explained by these factors.15 Whether any remaining differences
found in some studies reflect discrimination or something else remains a controversial issue.
15Ronald Ehrenberg and Robert Smith, Modern Labor Economics, 3rd ed. (Glenview, IL: Scott, Foresman, 1988). All of the difference in gross earnings between young African American and white males has been explained in a study that takes account of the results of achievement tests administered by the military since World War I—achievement tests used by the armed services to test individuals for fitness for military ser- vice. The tests measure verbal and mathematical skills and reflect the quality of schooling received as well as the effects of parental background. See June O’Neill, “The Role of Human Capital in Earnings Differ- ences Between Black and White Men,” Journal of Economic Perspectives, 4, No. 4 (Fall 1990), pp. 25–45.
476 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 476Trim Size: 203.2 mm X 254 mm
18.6 The Benefits and Costs of Immigration One in eight U.S. residents today is an immigrant; in 1970, only one in 21 was an immigrant.
This dramatic change is a result of U.S. immigration policy, especially the 1965 immigra-
tion law, which led to large increases in annual legal immigration, but also lax enforcement
of immigration laws, leading to unprecedented illegal immigration. Over the past 15 years,
the United States has absorbed about 1.3 million immigrants per year, composed of roughly
equal proportions of legal and illegal immigrants. In the 1950s, the level of immigration was
about a fifth as large.
In this section, we examine some of the economic costs and benefits associated with
immigration. For most of the issues we examine, it is the total number of immigrants that
matters and not how many are legal and illegal, so we will ignore that distinction.
The major economic impact of immigration flows from its effect on labor supply—it
increases the supply of labor. To examine the consequences of this effect, it is important
to use an aggregate model of the labor market since virtually all individual labor markets
are affected. In Figure 18.10 we use the aggregate model discussed in Section 17.3, with
one major simplification: here we assume that the labor supply curve is perfectly inelastic
(vertical).
Before proceeding, we should explain one implication of this model that was not noted
earlier. Recall that the quantity of capital (interpreted as all inputs except for labor) is held
constant in deriving the demand curve for labor, and the height to the demand curve at each
level of potential labor supply is equal to the marginal product of labor. With supply curve S and demand curve D, the wage rate will be w and total labor earnings are shown by the rect- angle, wBL0. The important new point here is that this graph also shows total capital earnings.
Because the height to the demand curve is the additional output generated by each extra
unit of labor, the total area under the demand curve out to a specific quantity of labor is the
economy’s total output (or gross domestic product). In the graph, if S is the supply curve, then the total output is equal to the area ABL0. Of this total, wBL0 goes to workers, and the remainder of total output must go to capital, the only other input in this model. Thus, capi-
tal owners receive the area ABw as their payment for supplying capital. This simple model explains how a society’s output (income) is distributed between workers and capitalists.
The Effects of Immigration on the U.S. Labor Market Immigration shifts the labor supply curve from S to S’, increasing employment to L1 and lowering the wage to w1. The income of U.S. capital owners increases by wBGw1 while native laborers lose wBFw1 in earnings. The United States, as a whole, is better off by area BGF.
Wage
w1
A
w
0
D
G
F
B
L L1 Labor
S S′
Figure 18.10
The Benef i ts and Costs of Immigrat ion 477
C18.INDD 10:36:38:AM 08/06/2014 PAGE 477Trim Size: 203.2 mm X 254 mm
With this background, it is straightforward to analyze the consequences of immigration.
Suppose that immigration has shifted the supply curve of labor from S to S′ (L1 − L is the number of immigrants added to the labor force.) The new equilibrium at point G involves a lower wage rate and greater employment. Notably, the analysis implies that U.S. native
workers (i.e., those in the United States prior to immigration) receive lower wage rates, so
the increased competition from immigrant workers tends to harm U.S. workers. The aggre-
gate earnings of U.S. workers fall by the amount wBFw1 in Figure 18.10. But that is not the whole story. While U.S. workers lose from immigration, U.S. capi-
tal owners gain. Indeed, they gain more than workers lose, so on balance the nation as a
whole experiences a net gain. Note in Figure 18.10 what happens to capital income as a
result of immigration: it increases from ABw to AGw1, or by wBGw1. The increase in the income of capital owners, wBGw1, is larger than the loss to U.S. workers, wBFw1, by the triangular area BGF. The total income of U.S. native residents is thus increased by an amount measured by the triangle BGF. When economists say that the nation as a whole gains from immigration, they are referring to this triangle. We will have more to say about
this gain later.
Some people, including a few economists, have argued that immigration does not lead to
lower earnings for U.S. residents. How could this be? As shown in Figure 18.10, if immi-
gration causes the demand for labor to increase at the same time it increases the supply of
labor, it is possible that employment increases but the wage rate does not change. Does
immigration lead to an increase in labor demand?
Recall that we held the quantity of capital fixed in deriving the demand curve. If the
capital stock increases, the labor demand curve shifts upward because a larger capital stock
means more capital per worker, and therefore a higher marginal productivity for labor.
Currently, the amount of capital per worker in the United States is about $250,000. Thus,
if immigrant workers were to bring with them $250,000 each and invest this money in
the U.S. economy, labor demand would increase, the amount of capital per worker would
remain unchanged, and wage rates could stay the same.
This scenario seems unlikely. But here we should distinguish the short-run effect of
immigration from its long-run effect. Most immigrants arrive in the United States with
little financial capital, so in the short run our analysis implying a reduction in wage rates
seems valid. Over time, however, as immigrants assimilate economically (in the sense of
becoming like U.S. residents and owning capital valued at $250,000 per worker), the effect
on wage rates can reverse. This process probably takes a long time, at least a generation
or two, and will be offset if new immigration continues to occur. Therefore, for the fore-
seeable future, our conclusion that immigration lowers the wage rates of U.S. residents is
probably correct.
One other matter deserves attention. In Figure 18.10, we have implicitly assumed that
immigrants are just like U.S. workers in terms of skills. When this is true, wage rates in all
labor markets are expected to fall in roughly the same proportion. However, recent (since
1965) immigrants have tended to be less skilled than U.S. workers, at least as indicated by
their level of education. According to a study by the Congressional Budget Office, immi-
grants now account for 45 percent of workers with less than a high school diploma, whereas
they make up only about 15 percent of the total labor force.16
This means that immigration has increased the supply of unskilled labor by much more
than it has increased the supply of skilled labor. As a result, we expect wage rates to be
reduced by more (proportionately) in unskilled labor markets than in skilled labor mar-
kets. One implication of this finding is that immigration has contributed to the increase
in income inequality in the United States. It actually does so in two ways. First, immi-
grants have lower incomes than U.S. native residents, and so this fact directly increases
16Congress of the United States, Congressional Budget Office, “The Role of Immigrants in the U.S. Labor Market,” November 2005. Calculated from Table 4.
478 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 478Trim Size: 203.2 mm X 254 mm
inequality since immigrants’ (both legal and illegal) incomes are included in measures of
U.S. inequality. Second, immigration indirectly increases inequality by lowering the wage
earnings of U.S. residents (especially unskilled workers) and increasing the incomes of
capital owners.
Despite this effect, our analysis does imply that immigration brings a net gain for U.S.
residents, workers, and capital owners. Let’s take a closer look at this gain, and some other
consequences of immigration.
More on Gains and Losses As we have seen, U.S. (native) workers are harmed by immigration but capital owners
benefit, and the benefits to capital owners exceed the costs to workers by the triangular
area BGF shown in Figure 18.10. The size of this net gain “to the nation as a whole” is of obvious importance in an evaluation of immigration policy. The President’s Council of
Economic Advisers (CEA) reported its estimate of the size of this gain: an amount equal
to 0.28 percent of GDP.17 This estimate seems surprisingly small in view of the sizable
volume of immigration in recent decades. Let’s examine why economists believe this gain
is relatively small.
The CEA does not explain how it constructed its estimate, but we can infer roughly how
it was done. Consider Figure 18.11, where we want to know the size of BGF relative to the size of OAGL1(GDP). Since we are concerned with relative magnitudes, we can arbi- trarily specify that the wage in the absence of immigration, w, is equal to one and the size of the U.S. native labor force, L, is 100. We know that immigrants account for about 15 percent of the total labor force, so immigration has added 18 to the labor force (18/118 is
15 percent). (Note that we are examining the cumulative effect of past immigration, not one
year’s immigration.) This gives us the base, FG, of the triangle we want to estimate. The height of the triangle, BF, is equal to the reduction in the average wage of U.S.
native workers, which we cannot observe directly. Its magnitude depends on the elasticity
of the labor demand curve, which is not known with certainty. Suppose the demand curve
is D in Figure 18.11, and so the increase in labor supply reduces the wage rate to 0.95, or by
A Further Look at the Effects of Immigration The net gains from immigration to the United States, as a whole, appear to be small for the reasons spelled out in this graphical analysis.
Wage
(1.0)w
(.95)w1
0
D
D′
G
F
B
A
L (100)
L1 (118)
Labor
S S′
Figure 18.11
17Council of Economic Advisors, Economic Report of the President. Washington, D.C.: U.S. Government Printing Office, 2007.
The Benef i ts and Costs of Immigrat ion 479
C18.INDD 10:36:38:AM 08/06/2014 PAGE 479Trim Size: 203.2 mm X 254 mm
5 percent. (This is roughly in line with some empirical estimates.) Now we have the base
and height of the triangular gain, and we can calculate its size as 1/2 × 0.05 × 18, or 0.45. Now let’s compare this gain to total GDP. Total labor earnings in Figure 18.11 are 112
(0.95 × 118), and we also know that total labor income is approximately 70 percent of GDP. Thus, GDP is 160 (112/0.70). Therefore, our calculation concludes that the gain from
immigration is equal to 0.28 percent of GDP (0.45/160), which is exactly the percentage
reported by the CEA.
What this exercise explains is why the net gain from immigration is likely to be quite
small relative to GDP. With immigrants constituting 15 percent of the labor force, this gain
would be larger only if the wage rates of U.S. workers fall by more than 5 percent when
immigrants increase labor supply by 18 percent. This would occur if the labor demand
curve is less elastic. For example, suppose the demand curve is D′, and this implies that wage rates fall by 15 percent. Then the triangular area of gain (not shown in the graph)
would be three times as large, or 1.35—still quite small relative to the total economy. Few
if any economists believe that immigration has depressed wages to this degree.
The CEA’s estimate thus appears to be based on the assumption that immigration has
depressed wages by about 5 percent. If this is the case, then it represents a loss to U.S.
native workers of about $400 billion a year. But recall that this is not a net loss, because
capital owners gain, and gain more than the loss to workers (and many people are simulta-
neously workers and capital owners). We could sum up these economic effects by saying
that workers lose $400 and capital owners gain $442 billion from immigration.
There are other economic consequences of immigration that do not result from its
impact on the aggregate labor market. Perhaps the most important of these consequences
is the way immigrants affect government budgets. Immigrants pay taxes and also receive
services and benefits from government expenditure policies. The question is: Do immi-
grants receive more in government benefits than they contribute in taxes? If they do, then
U.S. native taxpayers must pay for the difference. This would be a net loss to U.S. citizens
that could offset the net gain produced by immigrants in labor markets (the triangular area
just discussed).
Because immigrants have lower incomes than U.S. natives, it would be expected
that they pay less in taxes and receive more in benefits than the average U.S. taxpayer.
Several studies have investigated this issue, and they generally conclude that this statement
is true. For example, according to studies prepared for the National Research Council of the
National Academy of Sciences, the net cost on native taxpayers from immigration’s effect
on government budgets was in the range of $11–20 billion per year in the mid-1990s.18
Current costs would probably be at least double this amount, which would put this cost at
about $22–40 billion. Interestingly, this is essentially the same order of magnitude as the
gain from the labor market effect, which we put at $42 billion.
Let us sum up the discussion of the costs and benefits of immigration: on average,
native-born households in the United States receive a net gain from the increase in labor
supply resulting from immigration, but that net gain is small, about 0.28 percent of GDP.
Offsetting this gain is a net loss to native-born households resulting from immigrants’
participation in government tax and expenditure policies, and that net loss is approximately
the same magnitude as the net gain in labor markets. Combining these, and recognizing
the uncertainties surrounding all such estimates, it is probable that the overall average eco-
nomic effect on native-born households is close to zero.
In addition, immigration tends to increase inequality by depressing wages (especially
those of less skilled workers) and increasing capital incomes. This effect is a transfer, or
redistribution, of income, not a net loss, and requires value judgments to decide whether it
is good or bad.
18See James P. Smith and Barry Edmonston, eds., The New Americans: Economic, Demographic, and Fiscal Effects of Immigration (Washington, D.C.: National Academy Press, 1997).
480 Using Input Market Analys is
C18.INDD 10:36:38:AM 08/06/2014 PAGE 480Trim Size: 203.2 mm X 254 mm
Supporters of current immigration policy, and those advocating increased immigration,
often say that it is “good for the economy.” They are presumably referring to the triangular
net gain in labor markets, BGF, discussed earlier. But if we base immigration policy on this effect, we would like that gain to be as large as possible. It is possible to make this gain
larger without increasing the number of immigrants; this could be done by admitting more
skilled immigrants and fewer unskilled immigrants. For a given number of immigrants,
skilled immigrants produce a larger net gain because the height of the net gain triangle (BF) is the absolute reduction in the relevant wage rate produced by immigration, and it will be
larger for skilled than unskilled immigration.
Greater emphasis on skills in determining immigration would have two other economic
effects generally regarded as beneficial. First, skilled workers usually pay more in taxes
than they receive in government benefits, so skilled immigrants contribute a net gain to
U.S. taxpayers, in contrast to the net loss from current immigration policy. Second, increas-
ing the supply of skilled workers would, of course, decrease wages at the top of the distri-
bution, thereby reducing inequality rather than increasing it, as current policy does.
SUMMARY
The way input markets work strongly influences an
economy’s income distribution.
The 1938 Fair Labor Standards Act established
a nationwide minimum wage, which, by 2009, was
increased to $7.25 per hour. Although the minimum
wage was designed to help low-wage workers, economic
analysis suggests that it may not be the best way to do
so. Employers’ response to the minimum wage is often to
hire fewer workers, according to the law of demand. The
resulting disemployment effect of the minimum wage,
and the fact that the cost of paying the higher wage is
spread widely through society in the form of higher goods
prices, makes it difficult to assess the minimum wage as
an unqualified success in assisting unskilled workers.
Social Security is intended to provide older citizens
with a secure source of income after retirement by tax-
ing both those currently employed and their employers.
Economists believe that the way the tax is divided has no
effect on who actually bears the cost of Social Security
and that employees in fact bear most (but not all) of the
burden in the form of lower after-tax wage rates, particu-
larly if the aggregate supply of labor is highly inelastic.
The Social Security program has some important hid-
den costs on both GDP and before-tax wages through its
long-run impact on saving and capital accumulation.
The National Collegiate Athletic Association is an
input buyers’ cartel that, in essence, determines the
maximum financial reward a student-athlete can receive
and the number of student-athletes who can be recruited
with scholarships at each school. At the same time, its
policies result in a monopsony-like market for student-
athletes, who are paid less than they are worth as rev-
enue generators for their schools.
Economic analysis shows that widespread employ-
ment discrimination can exist, yet have no effect on the
wage rates of the group that is discriminated against.
Even if discrimination is prevalent enough to produce
lower wage rates for a particular group, firms that fail
to hire workers in that group pay higher labor costs than
necessary and realize correspondingly lower profits.
While immigration makes the United States, as a
whole, better off, economic analysis reveals that the net
gains associated with immigration at the present time
appear to be small.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
18.1 “Proponents of minimum wage laws stress society’s obli gation to act through its elected representatives to ensure an
adequate standard of living for all working citizens.” Evaluate
the extent to which minimum wage laws achieve this goal.
18.2 Why does economic theory imply that the most harmful effects of the minimum wage law will fall on the most disad-
vantaged and least productive workers?
*18.3 “The employer Social Security tax is just like any other labor cost to firms. A higher employer tax will thus increase
labor costs, reduce employment, and increase prices.” Explain
why this reasoning is incorrect.
Review Quest ions and Problems 481
C18.INDD 10:36:38:AM 08/06/2014 PAGE 481Trim Size: 203.2 mm X 254 mm
18.4 When the Social Security Administration attempts to compare the retirement benefits a worker receives with
the taxes paid, it usually bases the comparison on only the
employee portion of the tax. Do you think this comparison is
appropriate?
18.5 As the owner of a retail store, you would like to be able to pay your sales people lower wages. What problems would you
confront if you attempt to establish a cartel among employers
to force down wage rates?
18.6 Articulate and defend your position regarding paying col- lege athletes. What is the role played by positive analysis ver-
sus value judgments in your argument?
18.7 You are a business manager, and you believe that the differences in wages between men and women and between
African Americans and whites reflect discrimination and not
productivity differences. What type of employment policy
should you adopt?
*18.8 Suppose that some consumers of personal training ser- vices refuse to purchase such services from health clubs that
employ any members of a minority group, the Amazons. How
will this refusal affect the employment and wage rates of Ama-
zons and other (assumed equally productive) workers?
18.9 Suppose that all employers in a perfectly competitive industry do not have a preference for discriminating against
African Americans but that all the employees at one firm do.
Describe what will happen to the profit-maximizing output
level of this firm in the long run if the employees’ racial prefer-
ences influence the firm’s hiring decisions.
18.10 If the prevailing market wage for low-skilled workers is $7 per hour and a minimum wage law is passed, dictating $6
per hour, what will be the effect of the law on employment and
the prevailing wage rate?
18.11 Suppose that Korean Americans on average earn more than Polish Americans. Does this imply that Polish Ameri-
cans are discriminated against by employers relative to Korean
Americans?
18.12 Economists Daniel Hamermesh (University of Texas) and Jeff Biddle (Michigan State University) examined the dif-
ference in attractiveness of private-sector lawyers, who have to
woo clients, and public-sector lawyers, who do not. The results
of their study indicate that private-sector firms drew more
attractive lawyers to begin with and that the gap in attractive-
ness grew over time. What do their results say about the role
customers, fellow employees, and managers play in promoting
discrimination based on attractiveness in hiring decisions at
private-sector law firms?
18.13 Explain why it is more difficult to organize and maintain an input buyers’ cartel than an output sellers’ cartel.
18.14 In terms of the cost to employees, it would make no dif- ference regarding the extent to which the financing of national
health insurance was underwritten through various possible
shares of the required tax being formally assigned to employ-
ers. True, false or uncertain? Explain.
18.15 Explain why the Social Security program reduces before- tax wage rates and why this effect is greater over the long run
than in the short run.
482
C19.INDD 10:35:43:AM 08/06/2014 PAGE 482Trim Size: 203.2 mm X 254 mm
Learning Objectives
Delineate the difference between partial and general equilibrium analysis. Explain the concept of economic efficiency. Outline the three conditions necessary for the attainment of economic efficiency. Examine efficiency in production and what this implies about input usage across different industries. Show how efficiency in output is related to the production possibility frontier. Demonstrate how perfect competition satisfies all three conditions for economic efficiency. Spell out the reasons why economic efficiency may not be achieved.
General Equilibrium Analysis and Economic Efficiency19CHAPTER
The analysis developed in previous chapters focused on individual markets in isolation. Price and quantity in each market, whether it was a product or an input market, were deter-
mined by supply and demand conditions in that specific market. The analysis largely
ignored events in other markets.
We know, of course, that markets are interrelated. Changes in the market for gaso-
line, for example, affect the automobile market, and changes in the automobile market
in turn affect the gasoline market. Consequently, an analysis that focuses on one mar-
ket in isolation is incomplete. To see how the interdependence of individual markets
can be taken into account, this chapter provides a brief introduction to general equilibrium analysis, the study of how equilibrium is determined in all markets simul- taneously.
In addition to exploring their interdependence, this chapter also evaluates the extent to
which markets promote the well-being of the members of a society as a whole. We revisit
the concept of economic efficiency first introduced in Chapter 6 and discuss the conditions
that must be met for an economy to ensure efficiency in allocation of inputs across firms,
distribution of products among consumers, and output mix. We show how perfect competi-
tion satisfies these conditions as well as consider the reasons why markets may fail to pro-
mote efficiency.
general equilibrium analysis the study of how equilibrium is determined in all markets simultaneously
Memorable Quote “Obviously, the highest level of efficiency is that which can utilize existing material to the best advantage.”
—Jawharlal Nehru, twentieth-century Indian prime minister
Part ia l and General Equi l ibr ium Analys is Compared 483
C19.INDD 10:35:43:AM 08/06/2014 PAGE 483Trim Size: 203.2 mm X 254 mm
19.1 Partial and General Equilibrium Analysis Compared In previous chapters we have employed partial equilibrium analysis almost exclusively.
Partial equilibrium analysis focuses on the determination of an equilibrium price and quantity in a given product or input market, where the market is viewed as largely self-
contained and independent of other markets. An analysis of the gasoline market using sup-
ply and demand curves, for instance, is a partial equilibrium analysis. The supply and
demand curves for gasoline are drawn on the assumption of given and unchanging prices in
other product and input markets. In effect, these assumptions allow us to focus on the gaso-
line market and ignore others.
Characteristic of a partial equilibrium approach is the assumption that some things—like other prices—that conceivably could change do not. In many situations this assumption may be reasonable. For example, a tax on gasoline that raises its price is unlikely to have a
measurable effect on the price of wristwatches. A change in the price of gasoline could con-
ceivably cause a change in the price of wristwatches by raising or lowering their demand,
but in a partial equilibrium analysis of the gasoline market, we assume it does not. In con-
trast, a higher gasoline price would probably have a significant effect on the market for
automobiles. In that case the partial equilibrium assumption that the price of automobiles
does not change could be seriously flawed.
Partial equilibrium analysis therefore tends to ignore some of the interrelationships
among prices and markets. Formally, this is accomplished through the “other things being
equal” assumption. By contrast, in a general equilibrium analysis all prices are considered
variable, and the analysis focuses on the simultaneous determination of equilibrium in all
markets.
The Mutual Interdependence of Markets Illustrated Before turning to the discussion of a model of general equilibrium, let’s examine what is
meant by the interrelationship, or mutual interdependence, among markets. Consider two markets where the interdependence on the demand side is likely to be fairly pronounced—
the markets for margarine and butter. Margarine and butter are close substitutes, so a higher
price for margarine shifts the demand curve for butter upward; similarly, a higher price for
butter causes the demand for margarine to increase.
Figure 19.1a shows the margarine market, and Figure 19.1b shows the butter market.
Initially, assume both markets are in equilibrium, with the price of margarine at $2 per
pound and the price of butter at $3 per pound. From our earlier analysis of demand curve
relationships, recall that the prices of other goods are assumed to be fixed at all points along
a given demand curve, as are consumers’ incomes and tastes. Our emphasis here will be on
prices in other markets. Thus, at all points along DM the price of butter is $3 per pound, and at all points along DB the price of margarine is $2 per pound.
To illustrate the significance of mutual interdependence, let’s examine the effects
of an excise tax of $0.75 per pound on margarine. Using the familiar partial equilibrium
approach, we could analyze an excise tax by shifting the supply curve in Figure 19.1a
upward to ′SM, which causes the price to rise to $2.50 and the quantity to fall to ′QM. Now, however, let’s see how the partial equilibrium approach ignores the mutual interdepen-
dence between the margarine and butter markets and what implications this procedure has
for the analysis.
The foregoing partial equilibrium analysis neglects two types of consequences. First, the change in the margarine market has a spillover effect on other markets, which disrupts the equilibria there. In our example, the higher margarine price causes the demand for butter to
partial equilibrium analysis the study of the determination of an equilibrium price and quantity in a given product or input market viewed as self-contained and independent of other markets
spillover effect a change in equilibrium in one market that affects other markets
484 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 484Trim Size: 203.2 mm X 254 mm
rise because butter and margarine are substitutes. Don’t forget that in drawing DB we held the price of margarine fixed at $2 per pound; when the tax raises the price of margarine to
$2.50 per pound, we must redraw the demand curve for butter on the basis of the higher
margarine price. Thus, ′DB is the demand for butter when the margarine price is $2.50. In short, the tax on margarine leads to an increase in the demand for butter, which in turn
increases the price of butter to $3.50 per pound.
If this type of spillover effect from the margarine market to the butter market were the
only effect neglected by partial equilibrium analysis, there would be little cause for con-
cern. The analysis of the margarine market would remain exactly correct. But a second effect is neglected: the induced change in the butter market has a feedback effect on the margarine market. So far, the tax on margarine has led to a higher price for butter. Now consider the demand curve for margarine once again. We constructed the original demand
curve, DM, on the assumption that the price of butter was $3. Since the price of butter has risen, the demand curve for margarine will shift upward. When the price of butter is $3.50,
for example, the demand curve for margarine is ′DM. So the partial equilibrium analysis of the margarine market, which identified ′PM and ′QM as the equilibrium price and quantity, does not correctly identify the final result. Partial equilibrium analysis, by assuming that prices
in other markets remain unchanged, rules out the possibility of such a feedback effect.
This example illustrates what economists mean by mutual interdependence among mar-
kets: what happens in one market affects others (spillover effects) and is affected by other
markets (feedback effects). The butter–margarine example is a simple case of mutual inter-
dependence since just two markets, related only on the demand side, are involved. In the
real world a change in one market may affect the operation of hundreds of other markets
and, in turn, be affected by conditions in those markets. In addition, the interdependence
need not be restricted to the demand side of the markets. The employment and pricing of
inputs in one market will affect the supply curves in others that employ the same or closely
related inputs. For example, the increase in defense spending promoted by the Reagan
administration in the 1980s led to an expansion in the demand for inputs such as land in
Boston and Southern California, where a large number of defense contractors are based.
feedback effect a change in equilibrium in a market that is caused by events in other markets that, in turn, are the result of an initial change in equilibrium in the market under consideration
Figure 19.1 Interdependence between Markets: Butter and Margarine By raising the price of margarine, a tax on margarine increases the demand for butter. The higher butter price causes the margarine demand curve to shift to ′DM. This result illustrates economic interdependence between markets.
SM
Price
$2.00 = PM
0 QM
DM (PB = $3.00; . . . )
Margarine
$0.75
Price
$3.00 = PB
0 QB
DB (PM = $2.00; . . . )
SB
Butter
(b)(a)
$2.50 = PM′
QM′
′′
SM
DM (PB = $3.50; . . . )
′
′′DB (PM = $2.50; . . . )
′$3.50 = PB
QB′
Part ia l and General Equi l ibr ium Analys is Compared 485
C19.INDD 10:35:43:AM 08/06/2014 PAGE 485Trim Size: 203.2 mm X 254 mm
This increase in demand for land made it more expensive for other industries such as bank-
ing, entertainment, and education to do business in the same cities and shifted their respec-
tive output supply curves leftward.
When Should General Equilibrium Analysis Be Used? The first 18 chapters of this book concentrated on partial equilibrium analysis, and this topic
would not have received such emphasis if economists believed that it was an unreliable
framework for analysis. Yet we have seen that partial equilibrium analysis neglects some mar-
ket interdependencies that can affect the way a given market functions. General equilibrium
analysis, in contrast, accounts for the interrelationships among markets. On these grounds,
the general equilibrium approach would appear superior, so it is worthwhile to explain why
economists continue to rely on partial equilibrium analysis to study many issues.
Partial equilibrium analysis explicitly ignores some factors that could have a bearing on
the analysis, but in many cases these neglected factors may be quantitatively unimportant in the sense that if they were taken into account, the conclusions would be affected only to
a trivial degree. For instance, in our butter–margarine example, the excise tax on margarine
may affect the price of butter only slightly, and this in turn will have an even smaller effect
on the demand curve for margarine. In that case, ignoring the market interdependencies and
assuming that the margarine demand curve “stays put” yields a result that is a sufficiently
close approximation to the true outcome.
This does not imply that partial equilibrium analysis can always be used. There are cases
where the implications of a partial and general equilibrium analysis differ significantly.
A reasonable guideline is that partial analysis is usually accurate in cases involving a
change in conditions primarily affecting one market among many, with repercussions on
other markets dissipated throughout the economy. When a change in conditions affects
many, or all, markets at the same time and to the same degree, however, general equilib-
rium analysis tends to be more appropriate.
An example will clarify this distinction. Suppose that a price control is applied to one
product—say, rental housing. Rent control is sure to have a major impact on the rental
housing market, but the impact on other markets is likely to be slight and uncertain. Most
economists would agree that a partial equilibrium analysis focusing on the rental housing
market is adequate to investigate this issue.
By contrast, imagine that the government applies price controls to all goods simultane-
ously. With all markets affected at the same time, and to a large degree, a general equi-
librium analysis is required. In fact, a partial equilibrium analysis may give misleading
results. Suppose, for example, that the government mandates a 50 percent reduction in the
prices of all goods except rental housing and a 5 percent reduction for that. If we looked at
the rental housing market using partial equilibrium analysis, we would be tempted to say
the output of rental housing would fall and a shortage would result. The opposite is more
likely to be the case because this set of price controls increases the relative price of rental housing compared with all other goods. Resources would shift from industries where prices
are most depressed to those where they are least depressed, increasing output in the latter.
Only a general equilibrium analysis is capable of accurately evaluating this situation.
Thus, both general and partial equilibrium approaches are quite valuable, with their rela-
tive usefulness depending on the issue under investigation. Earlier chapters give numerous
examples of topics that can be fruitfully studied by using the partial equilibrium approach.
This section has shown how the separate markets form an interconnected system and attain
a general equilibrium. In the remainder of this chapter we will see how we can use the gen-
eral equilibrium model to evaluate the efficiency with which an economy allocates
resources. Chapter 6 first introduced the concept of economic efficiency in the context of
the distribution of fixed quantities of goods among consumers. Now we are concerned with
efficiency in a more general sense. In our discussion the terms efficient and Pareto optimal
Pareto optimal the condition in which it is not possible, through any feasible change in resource allocation, to benefit one person without making some other person or persons worse off
486 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 486Trim Size: 203.2 mm X 254 mm
are used interchangeably; the latter term is named after the Italian economist Vilfredo
Pareto, who first gave careful attention to the concept.1
19.2 Economic Efficiency Let’s begin with a formal definition of economic efficiency and its corollary, economic
inefficiency. An allocation of resources is efficient when it is not possible, through any fea- sible change in resource allocation, to benefit one person without making any other person worse off. In other words, when the economy is operating efficiently, there is no scope for further improvement in anyone’s well-being unless they are benefited at the expense of
other people.
An allocation of resources is inefficient when it is possible, through some feasible change in the allocation of resources, to benefit at least one person without making any other person worse off. Inefficiency implies waste, in the sense that the economy is not sat- isfying the wants of people as well as it could.
These abstract definitions become clearer when we employ a diagram. To simplify mat-
ters, let’s assume that society consists of only two people, Scrooge and Tiny Tim, although
we can easily extend the analysis to larger numbers. In Figure 19.2, Scrooge’s welfare is
measured horizontally and Tiny Tim’s welfare vertically. Since no objective way exists to
attach units of measurement to a person’s utility or welfare, the welfare measure is entirely
ordinal. In other words, a rightward movement in the diagram implies that the resource
allocation has changed in a way beneficial to Scrooge, but it does not tell us how much
better off Scrooge is. All we know from the diagram is that the farther to the right we are,
the higher is the indifference curve Scrooge attains. Upward movements similarly imply a
change beneficial to Tiny Tim.
The levels of well-being attained by Scrooge and Tiny Tim depend on their consump-
tion of goods. There are limits, though, to how much they can consume, because limited
efficient a condition that is the same as Pareto optimality
inefficient the condition in which it is possible, through some feasible reallocation of resources, to benefit at least one person without making any other person worse off
1Vilfredo Pareto, Manuel d’Economie (Paris: V. Giard and E. Briere, 1903).
Figure 19.2 The Welfare Frontier The welfare frontier shows how any resource allocation affects the well-being of both consumers. Any point on the frontier is an efficient point. Points inside the frontier, like point D, are inefficient since both parties can be made better off.
W
A L
B
C D
Welfare of Tiny Tim
0 W′ Welfare of Scrooge
F
E
Economic Eff ic iency 487
C19.INDD 10:35:43:AM 08/06/2014 PAGE 487Trim Size: 203.2 mm X 254 mm
quantities of resources are available to produce those goods. Scarcity places upper limits on
the well-being of Scrooge and Tiny Tim, and these limits are shown in Figure 19.2 by the
welfare frontier WW′. The welfare frontier separates welfare levels that are attainable from those that cannot be reached, given the available resources. Any point on or inside the fron-
tier is attainable. For example, different allocations of resources would place Scrooge and
Tiny Tim at points A, B, C, D, E or F. Any point beyond the frontier, like L, is unattainable. The economy cannot produce enough goods and services to make Scrooge and Tiny Tim as
well off as the point indicated by L. A welfare frontier illustrates how the allocation of resources affects the well-being of
members of society. To use it correctly, we must understand how it is derived from the
underlying characteristics of an allocation of resources, and we will do this in the remain-
der of the chapter. For now, let’s take the existence of the frontier for granted and use it to
illustrate several points about the nature of economic efficiency.
Any resource allocation resulting in a point on the WW′ frontier is efficient, or opti- mal; that is, it satisfies the definition of efficiency given earlier. Consider point A, for example. Since it is impossible to move beyond the frontier, there is no move from
point A that can benefit one person without making the other worse off. The same is also true of point B; any move from point B harms at least one of the two persons. Thus, point B also represents an efficient allocation of resources. Indeed, every point lying on the welfare frontier satisfies the definition of economic efficiency. In fact, all points on the frontier are equally efficient, and no point on the frontier is more efficient than any
other.
Any point inside the welfare frontier represents an inefficient resource allocation.
Point D, for instance, is inefficient because resources can be reallocated so as to benefit one person without harming the other. A vertical move from D to B makes Tiny Tim better off and leaves Scrooge’s welfare unchanged. Alternatively, a horizontal move
from D to C benefits Scrooge without harming Tiny Tim. Every point lying inside the welfare frontier represents an inefficient allocation of resources. Note, also, that an inefficient point means it is possible to reallocate resources in a way that makes all par- ties better off, such as a move from point D to a point between B and C on the welfare frontier.
Efficiency As a Goal for Economic Performance The notions of efficient and inefficient resource allocations, as summarized by the points
on and inside the welfare frontier, naturally lead to an emphasis on the factors that affect
the level and distribution of well-being. But this focus does not allow us to identify one
resource allocation as being better than any other. To see why, consider a choice among
the points on the welfare frontier, all of which are efficient. Is one better than another?
Note that the points differ in terms of the distribution of well-being; a movement
from one point to another—for example, from E to B—benefits one person and harms another. Since there is no objective way to compare one person’s gain with another per-
son’s loss—interpersonal utility comparisons cannot be made objectively—economics
must remain silent on this issue. As individuals we might believe for normative rea-
sons that B is superior to E, but we can’t rest our judgment on positive, efficiency-based considerations.
To see that economic efficiency is a reasonable goal, notice what an inefficient alloca-
tion of resources, like point D, implies. Inefficiency indicates that it is possible to reallocate resources in a way that benefits some, perhaps all, people without harming anyone, a noble
goal. Since a move from inside the frontier at D to a point on the frontier between B and C benefits both parties, would anyone oppose such a change? In this context we must realize
that when we talk about people being better off, we mean better off according to their own
welfare frontier a curve that separates welfare levels that are attainable from those that cannot be reached given the available resources
488 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 488Trim Size: 203.2 mm X 254 mm
preferences: Tiny Tim views himself as better off at B than at D. Accepting efficiency as a goal means accepting the premise that each person is the best judge of their own welfare.
One could quarrel with this view, but it appears a reasonable assumption in most situations.
If granted, we could conclude that changes benefiting some and harming no one—that is,
movements from inefficient to efficient points—are desirable, which is why we use effi-
ciency as a goal.
We cannot conclude, however, that any efficient position is better than any inefficient position. For example, although a move between points D and E is a change from an inef- ficient to an efficient allocation, the change in this case greatly benefits Scrooge while
impoverishing Tiny Tim. In comparing these points, we cannot simply note that one is effi-
cient and the other inefficient; we must also take into account the change in the distribution
of well-being. Taking equity considerations into account, we might judge point D to be superior to point E. By making such a judgment, however, we recognize that efficiency is not the only goal: the distribution of well-being counts, too. Even so, efficiency goals are
not irrelevant since there are still efficient points between B and C making both Scrooge and Tiny Tim better off than point D.
Almost all real-world resource allocation changes involve both a move to a more (or
less) efficient position and a change in the distribution of well-being, like the move from
D to E in the diagram. In these cases, demonstrating that there is a gain in efficiency does not prove that the change is desirable, since distributional effects are important as well.
Consequently, economists are generally reluctant to claim that one resource allocation is
superior to any other. Economics can sometimes prove that one situation is more efficient
than another, but there are other goals besides efficiency.
19.3 Conditions for Economic Efficiency Notwithstanding the importance of other goals, such as distributional equity, let us turn
now to deriving the conditions necessary for achieving economic efficiency. In general, any
economy must solve three fundamental economic problems:
1. How much of each good to produce; 2. How much of each input to use in the production of each good; and 3. How to distribute goods among consumers.
Each of these problems can be solved in different ways, but not all solutions are equally
efficient. For example, consider the distribution of goods among consumers, an issue we
discussed in Chapter 6. Recall that in a simple two-person, two-good setting, an Edgeworth
exchange box shows all possible distributions of goods between consumers. Only some of
these distributions, however, are efficient. The contract curve identifies which distributions
of goods across consumers are efficient; distributions located off the contract curve are
inefficient. At all points along the contract curve, the marginal rate of substitution (MRS) between any two goods is the same across consumers. Thus, we can concisely express the
condition for efficiency in the distribution of goods as:
MRS MRS MRSi1 2= = =. . . , (1)
where the superscripts 1, 2, . . . i represent the i consumers in an economy. If this condition does not hold, at least some consumers can be made better off without harming other con-
sumers by a change in the distribution of goods.
We now turn to developing similar conditions for efficiency in both production and out-
put. We do so by employing the now-familiar constructs of an Edgeworth box and a pro-
duction possibility frontier (PPF).
Eff ic iency in Product ion 489
C19.INDD 10:35:43:AM 08/06/2014 PAGE 489Trim Size: 203.2 mm X 254 mm
19.4 Efficiency in Production To address the issue of efficiency in production, think about a simplified world in which
there are only two inputs (labor and land) and two possible consumer goods produced with
the inputs (food and clothing). Assume that consumers’ incomes are earned through the
sale of the services of their labor and land and are spent on food and clothing. All inputs are
homogeneous; each worker, for instance, is interchangeable with any other. Most impor-
tantly, the overall quantities of land and labor are taken to be in fixed supply. In other
words, the aggregate labor and land supply curves are vertical. Finally, although the total
supply of each input is fixed, the amount employed in each industry is not. The food indus-
try can employ more labor, for instance, but only by bidding workers away from the cloth-
ing industry, since the total employment by the two industries together is fixed.
Although these assumptions describe about the simplest economy imaginable, under-
standing how the pieces fit together is still complex. Note that there are six identifiable
markets: labor employed in producing food, labor employed in producing clothing, land
employed in producing food, land employed in producing clothing, and the two prod-
uct markets for food and clothing. Moreover, these markets are interrelated. For the food
industry to expand, for example, it will have to bid away inputs from the clothing industry.
Consequently, we must determine a pattern of prices and quantities in which the quanti-
ties demanded and supplied in each of the six markets are brought into equality simultane-
ously—that is, a general equilibrium.
The Edgeworth Production Box An Edgeworth production box (analogous to the Edgeworth exchange box introduced in Chapter 6) identifies all the ways labor and land can be allocated between the food and
clothing industries in our simplified economy. As shown in Figure 19.3, the length of the
Edgeworth production box a diagram that identifies all the ways two inputs, such as labor and land, can be allocated between industries in a simplified economy
Figure 19.3 Edgeworth Production Box With fixed total input supplies, we can show all the possible ways of allocating inputs between food and clothing production with an Edgeworth production box. Point B, for example, indicates employment of 10 units of land (A) and 40 units of labor (L) in clothing production; the remaining inputs, 40 units of land and 40 units of labor, are employed in food production.
Land in clothing production
Total land = 50A
B
20A
10A
0C
C
Labor in clothing production
40L 60L
Total labor = 80L
Land in food production
30A
0F
40A
Labor in food production
40L 20L
490 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 490Trim Size: 203.2 mm X 254 mm
box indicates the amount of labor employed by the two industries, in this case 80 units; the
height of the box shows the amount of land employed by the two industries, 50 units.
Through the Edgeworth production box, we can determine the quantities of labor and
land employed by both industries. Let’s measure the employment of labor in clothing pro-
duction horizontally from the southwest corner, point 0C, and the employment of land in
clothing production vertically from the same point. Point B, for example, indicates that 40 units of labor (L) and 10 units of land (A) are used in clothing production; point C implies 60 units of labor and 20 units of land. Because the total available input quantities are shown
by the dimensions of the box, a given point also identifies employment levels in the food
industry. Since total employment of labor in both industries together is 80 units, if the
clothing industry employs 40 units, the food industry must be employing the remaining 40
units. In the diagram we measure the employment of labor and land in the food industry
to the left and down, respectively, from the northeast corner, point 0F. Thus, at point B the clothing industry employs 10A and 40L, and the food industry employs the remaining 40A and 40L. A move from B to C would indicate an expansion in employment of both inputs in clothing production to 20A and 60L, coupled with a decline in employment of both inputs in food production to 30A and 20L. Put differently, a move from B to C shows that the clothing industry has bid 20 labor units and 10 land units away from the food industry,
although the total employment in both industries together remains unchanged.
Having shown how the allocation of inputs is indicated by a point in the Edgeworth pro-
duction box, we next want to identify the levels of food and clothing output corresponding
to each possible input allocation. This is accomplished by incorporating clothing and food
isoquants into the diagram, since these curves identify the output level associated with each
combination of inputs. In Chapter 7 we explained how a firm’s production function could be graphed as a set of isoquants. Here, though, isoquants are used to represent the produc-
tion function of an entire industry composed of all the separate firms producing each good. The industry isoquants have the same characteristics as those of individual firms.
Figure 19.4 incorporates the clothing and food isoquants into the Edgeworth production
box. For clothing, several isoquants are drawn with their origin at point 0C—100C, 150C,
Figure 19.4 General Equilibrium in Input Markets A competitive equilibrium involves an input allocation lying somewhere on the contract curve connecting points of tangency between food and clothing isoquants. Input prices depend on exactly where on the contract curve the equilibrium lies.
G
300F B
k
10A
100C
220F 260F
160C 150C
0C 40L
20A
Land in clothing production
Labor in clothing production
Land in food production
0F
Labor in food production
10L
1L
1A
D
k
200C 100F
g
H
2L 1A
g
Eff ic iency in Product ion 491
C19.INDD 10:35:43:AM 08/06/2014 PAGE 491Trim Size: 203.2 mm X 254 mm
160C, and so on—each labeled to indicate the amount of clothing produced with each combina- tion of land and labor. Note that the isoquants have the familiar shapes discussed in Chapter 7.
The food isoquants—100F, 220F, 260F, and so on—are drawn relative to the origin at point 0F, with the employment of labor measured to the left and the employment of land measured
down from 0F. In effect, the food isoquants are turned upside down, which accounts for their
unconventional appearance. Nonetheless, the food isoquants embody the familiar properties,
with isoquants lying closer to the origin at point 0F representing lower food output. With the isoquants drawn in, each point in the Edgeworth production box indicates the
employment of labor and land in both industries as well as the output of food and clothing.
For example, point B implies employment of 40L and 10A by the clothing industry, with an output of 100 clothing units; point B also indicates employment of 40A and 40L by the food industry (see point B in Figure 19.3 to understand this explicitly; to avoid cluttering the diagram, we have not shown these input use levels on the Figure 19.4 axes), with an output
of 300 food units. In the same way, point D shows employment of 20A and 10L producing 100 food units, with the remaining 30A and 70L used to produce 200 clothing units.
The Production Contract Curve and Efficiency in Production The Figure 19.4 box shows every conceivable way of allocating labor and land between
the two industries. Still, only some of these resource allocations are efficient in the sense that the output of one good cannot be increased without decreasing the output of the other.
Indeed, only input allocations where the isoquants are tangent to one another represent efficient resource allocations. The production contract curve running from one origin to the other and passing through points B, H, and D connects all the points where food and cloth- ing isoquants are tangent.
To see why only points on the production contract curve represent efficient input alloca-
tions, consider point G, where a food (220F) and a clothing (150C) isoquant intersect. At point G, a lens-shaped area lies between the intersecting food and clothing isoquants. The significance of the lens-shaped area is that every allocation of inputs identified by a point
inside the area involves larger outputs of both goods than at point G. For instance, point H implies greater production of both clothing (160) and food (260) than point G (150 clothing units and 220 food units). Thus, a move from point G to H, which involves shifting some labor from the food to the clothing industry and some land from the clothing to the food
industry, will increase the output of both goods at no additional cost. This result is true at
every point in the box where isoquants intersect. Thus, any point where food and clothing
isoquants intersect, which includes all points not on the production contract curve, cannot
represent efficiency in production.
General Equilibrium in Competitive Input Markets If perfect competition prevails in input markets, a point on the production contract curve
(i.e., efficiency in production) will be attained. To see why, recall from Chapter 16 that in
competitive markets, the price of an input tends to be equalized across firms and industries.
Here, this tendency means that the wage rate earned by laborers will be the same in the
clothing and food industries, and similarly for the rental price of land. Furthermore, every
firm will minimize cost by employing inputs in quantities so that the ratio of marginal prod-
ucts (MP) equals the ratio of input prices (Chapter 8). For a wage rate of w and a rental price of land of v, the condition for cost minimization is:
w v MP MP MRTS/ /L A LA= = . (2)
Geometrically, this equality is shown by the tangency between an isocost line, with a slope
of w/v, and an isoquant where the isoquant’s slope, the marginal rate of technical substitu- tion (MRTSLA), is equal to the ratio of marginal products (MPL/MPA).
492 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 492Trim Size: 203.2 mm X 254 mm
In a competitive equilibrium, each food producer operates at a point where the slope
of its food isoquant equals the ratio of input prices, w/v. In addition, each clothing pro- ducer operates at a point where the slope of its clothing isoquant also equals the same input price ratio. Therefore, the slopes of the clothing and food isoquants must equal one another since both are equal to the same input price ratio. Consequently, the equi- librium must lie on the contract curve, which identifies resource allocations where the
slopes of clothing and food isoquants are equal. For example, if the wage rate is half
the rental price of land, isocost lines have a slope of 1/2, as illustrated by line kk in Figure 19.4. To minimize the cost of producing 100 clothing units, clothing producers
would operate at point B where the 100C isoquant is tangent to kk. Similarly, food pro- ducers also minimize the cost of producing 300 food units when they use the remaining
inputs, since the 300F isoquant is also tangent to kk at point B. Only when the isoquants are tangent can both industries be minimizing cost when confronted with the same input
prices.
For these reasons the competitive equilibrium can exist only at a point on the production
contract curve in Figure 19.4. Exactly where on the contract curve the equilibrium will lie
depends on the consumers’ demands for clothing and food. If the demand for clothing is
relatively high, for example, equilibrium will occur at a point like D, where a large quantity of clothing and very little food is produced. On the other hand, if the demand for clothing is
relatively small (which is equivalent to saying that the demand for food is relatively great),
not much clothing and a large quantity of food will be produced, as at point B.
19.5 The Production Possibility Frontier and Efficiency in Output
To bring the output markets for food and clothing clearly into focus, we use a concept
introduced in Chapter 1, the production possibility frontier (PPF). The PPF shows the alternative combinations of food and clothing that can be produced with fixed supplies of
labor and land. The same information is already contained in the Edgeworth production
box, but the PPF presents it more clearly. Glance once again at the contract curve in Figure 19.4. Each point identifies a certain combination of food and clothing output that can be
produced with available inputs. A movement between points on the contract curve, as from
B to H, shows that as more clothing is produced, food output must fall. The PPF is derived from the contract curve in Figure 19.4 by plotting the various pos-
sible output combinations directly. In Figure 19.5, the frontier is the bowed-out curve ZZ′. Points B, H, D, and G in Figure 19.5 correspond to these same points in Figure 19.4. For example, point B indicates an output combination of 100C and 300F in both diagrams. The frontier slopes downward, indicating that more clothing can be produced only by giving up
food output, since land and labor must be transferred from the food to the clothing industry
to produce more clothing.
With the available quantities of labor and land, firms can produce any combination of
output lying on or inside the PPF. Points lying inside the frontier represent allocations that are production inefficient. That is, in Figure 19.5 a point like G inside the frontier corre- sponds to a point where isoquants intersect in the box of Figure 19.4; more of both goods
can be obtained from the available resources, such as at point H. As discussed in Chapter 1, the PPF is typically bowed out, or concave to the origin. The
PPF slope thus becomes more steep in absolute value as we move down the curve from point Z to point Z′. Like many other slopes in economics, the PPF slope has a special name, the marginal rate of transformation (MRT). At any point on the frontier, the MRT indi- cates the rate at which one product can be “transformed” into the other. Of course, once
marginal rate of transformation (MRT) the rate at which one product can be “transformed” into another
The Product ion Poss ibi l i ty Front ier and Eff ic iency in Output 493
C19.INDD 10:35:43:AM 08/06/2014 PAGE 493Trim Size: 203.2 mm X 254 mm
food is produced, it cannot usually be changed into clothing, but clothing output can be
increased by transferring land and labor from food to clothing production, thereby gaining
more clothing at a cost of reduced food output. At point B the marginal rate of transforma- tion is 0.5F/1.0C, indicating that production of one more clothing unit requires removing resources from food production by an amount that will reduce food output by 1/2 unit. Far-
ther down the frontier, at point D, the MRT is 3F/1C, implying that increasing clothing out- put by one unit necessitates a sacrifice of three food units.
As Chapter 1 shows, the marginal rate of transformation, or the PPF slope, reflects the opportunity cost of one good in terms of the other. The marginal rate of transfor-
mation also equals the ratio of the monetary marginal cost (MC) of clothing production to the monetary marginal cost of food production. At any point on the frontier, the slope, or MRT, equals MCC /MCF. To see why, suppose that at current output levels the marginal costs of clothing and food are $100 and $200, respectively. How much food
would we have to give up to produce one more clothing unit? (That is, what is the MRT between clothing and food?) Producing one more clothing unit utilizes $100 worth of
resources (labor and land), so MCC = $100. If we remove $100 worth of resources from food production, food output falls by half a unit since the marginal cost of a food unit is
$200. Thus, we must give up half a food unit to produce one more clothing unit—that is,
MRT = 0.5F/1.0C = 1F/2C. This ratio is also equal to the marginal cost ratio: MCC/MCF = $100/$200 = 1F/2C.
When the marginal cost of food is twice that of clothing, we know that one more
clothing unit requires a sacrifice of half a unit of food. Point B illustrates this situation. As we move down the frontier, the marginal cost of clothing increases as more is pro-
duced, and the marginal cost of food declines as less is produced. (This means that the
ratio MCC/MCF rises because the numerator increases and the denominator decreases.) At point D, for example, MCC/MCF = 3F/1C, showing that the marginal cost of clothing is three times the marginal cost of food, so three food units must be given up to produce
one more clothing unit.2
Figure 19.5 The Production Possibility Frontier Revisited The PPF plots the output combinations from the contract curve in Figure 19.4. It is normally bowed out from the origin. The slope of the frontier, called the marginal rate of transformation, shows how much of one good must be given up to produce more of the other.
Food (F)
Z
300 0.5F 1C
100
0 100 200 Z′
B
H
G
Clothing (C)
D 3F
1C
2As mentioned in Chapter 1, it is possible for the PPF to be a straight line with a constant slope. This situa- tion occurs when both industries are constant cost so that each good’s marginal cost is constant over all out- put levels. The ratio MCC/MCF is constant at all output combinations, implying a linear production frontier.
494 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 494Trim Size: 203.2 mm X 254 mm
Efficiency in Output Now we come to the question of where on the PPF we should operate—that is, what combination of food and clothing should be produced? Producing an efficient output mix
requires balancing the subjective wants, or preferences, of consumers with the objective
conditions of production. More specifically, efficiency in output is attained when the rate at
which consumers are willing to exchange one good for another (the marginal rate of sub-
stitution, or MRS, of those consumers) equals the rate at which, on the production side, one good can be transformed into another (the marginal rate of transformation, or MRT):
MRS MRS MRS MRTi1 2= = = =. . . , (3)
where the superscripts refer to an economy’s i consumers. To see why the foregoing equality needs to be satisfied to ensure efficiency in output,
recall from Chapter 4 that, in the context of our example, MRSCF reflects a consumer’s will- ingness to pay for an additional unit of clothing by consuming less food. As we saw in the
preceding section, MRTCF represents the ratio of the marginal cost of clothing relative to the marginal cost of food.
Now, suppose that the MRSCF is equal across all consumers (i.e., we have achieved effi- ciency in distribution) and equal to three units of food per clothing unit, while the MRTCF equals one food unit per clothing unit. Such an outcome would not represent an efficient
output mix. When consumers’ MRSCF is 3F/1C, the marginal benefit to a consumer of one more clothing unit is equal to three food units—this is the maximum amount of food
that a consumer is willing to give up for one more clothing unit. In contrast, when MRTCF is 1F/1C, the marginal cost of one more clothing unit is only one food unit. Thus, when MRSCF is greater than MRTCF, the marginal benefit to consumers of more clothing output exceeds the marginal cost of producing it—both expressed in food units. Additional cloth-
ing is worth more to consumers than it costs to produce, and consumers can be made better
off by moving along the PPF to a point where more clothing and less food is produced. As more clothing and less food is produced, MRTCF rises, and as consumers consume
more clothing and less food, their marginal rates of substitution tend to decline. The pro-
cess of producing more clothing and less food tends to bring MRTCF and MRSCF closer together. This movement along the production frontier can continue to benefit consumers
until the two terms are exactly equal. In sum, it is always possible to change the output mix and leave consumers better off whenever their common marginal rates of substitution are not equal to the marginal rate of transformation.
We can illustrate the preceding analysis with a diagram. In Figure 19.6, the points on the
production possibility frontier ZZ′ all represent efficient input allocations in our two-good economy. Of these points, an efficient output mix is represented by a point such as P on the frontier. At point P, the marginal rate of transformation between food and clothing (MRTCF) equals the slope of a representative consumer’s indifference curve, the marginal rate of sub-
stitution between food and clothing (MRSCF), at the optimal consumption point, E, chosen by the consumer. The slope of the production frontier at P is reflected by the slope of line gg and is equal to two units of food per one clothing unit. The slope of the representative consumer’s indifference curve Ui at consumption point E is represented by the slope of the line hh and also equals two units of food per clothing unit.3
By calling for the marginal rates of substitution and transformation to be equal, the con-
dition for efficiency in output is simply restating what we discussed earlier in Chapter 10:
namely, efficiency in output requires the output of any good to be expanded to the point
3Figure 19.6 does not rule out the possibility that consumers differ in their incomes and, from that, in how far removed their budget lines may be from the origin. For efficiency in output mix to result, all that is nec- essary is that, at the consumption points selected by individual consumers, the slopes of various consumers’ indifference curves be identical and equal to the slope of the PPF.
The Product ion Poss ibi l i ty Front ier and Eff ic iency in Output 495
C19.INDD 10:35:43:AM 08/06/2014 PAGE 495Trim Size: 203.2 mm X 254 mm
where marginal benefit (MRS) equals marginal cost (MRT). The only difference of note is that in Chapter 10 we employed a partial equilibrium perspective and showed how, at any
given output level, the marginal benefit to consumers from a good is represented by the
height of the demand curve, while the height of the supply curve represents the marginal
cost. Relying on a partial equilibrium approach, we saw how efficiency in output thus is
realized where the demand and supply curves intersect.
In this chapter we have adopted a general equilibrium approach and avoided the assump-
tion that we can treat the market for any good in isolation. The heights of demand and
supply curves for a good at any given output level can be influenced by the operation of
markets for other goods.
An Economy’s PPF and the Gains from International Trade Our general equilibrium analysis has been developed for a society that does not trade
with other countries. It can, however, be easily extended to show the consequences of
international trade. In fact, a general equilibrium approach provides a particularly vivid
demonstration of the sense in which a country can be said to gain from participation in
international trade.
In Figure 19.7, suppose that we are a nation initially isolated from world trade and in
equilibrium at point B on our production possibility frontier, ZZ′. The assumed pretrade price ratio between the two goods is shown by the slope of gg, or 2F/1C. Now assume that we begin engaging in international trade. The terms on which other countries are willing to
trade are measured by the world price ratio. If that ratio is, for instance, 1F/1C, then other countries are willing to supply us with one food unit for each clothing unit we supply to
them, or to sell us one clothing unit for one food unit. In this model we view our exports of
one product as paying for our imports, with the world price ratio indicating how much of
one product can be exchanged for the other. (In reality, money is used to pay for imports,
but since the money received by international sellers is used to purchase our exports,
exports are really being exchanged for imports.)
Figure 19.6 Efficiency in Output The output mix at point P, together with the consumption point E selected by a representative consumer, satisfies the condition for efficiency in output. For efficiency in output to be realized, the slope of the PPF (the marginal rate of transformation) must equal the slope of all consumers’ indifference curves (the marginal rates of substitution) at their selected consumption points.
Food
Z
P
E
Ui 2F
1C2F
1C
g
h
h
g
0 Z′ Clothing
496 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 496Trim Size: 203.2 mm X 254 mm
If the opportunity for international trade at a price ratio of 1F/1C exists and the domestic price ratio is initially 2F/1C, then the domestic economy will undergo a number of changes to adjust to a new equilibrium position. With the world price of clothing lower than the domestic
price, clothing consumers will switch from domestic to imported apparel and expand total
clothing purchases. As a consequence, the domestic clothing industry will lose customers and
have to contract its output. (In partial equilibrium terms, the domestic industry moves down
its supply curve until it operates where the domestic price equals the world price.)
At the same time that the clothing industry contracts in the face of foreign competition,
the food industry expands. Since the world price of food (one C for one F) is higher than the initial domestic price (one-half C for one F), food producers can make higher profits by exporting food. Food production will expand until the domestic food price rises to equal the
world price of food.
So, domestic clothing production falls and domestic food production rises until the
domestic price ratio equals the world price ratio. These adjustments are illustrated in
Figure 19.7. The world price ratio is equal to the slope of line ww. Since domestic pro- duction adjusts until the domestic price ratio equals the world price ratio, the output mix
produced by the economy shifts to point J on the production frontier; at point J, the slope of ZZ′ equals the slope of ww, the world price ratio. Point J thus identifies the output mix produced by the economy. It does not, however, identify the domestic consumption of food
and clothing since the country can, by engaging in trade, consume a different bundle of
products than it produces. In fact, through trade the country may consume any output com-
bination along line ww. Exactly where on ww consumption will take place depends on domestic demands for
the two products, which are not shown in the diagram. Suppose, however, that consumers
choose to consume at point K. To arrive at K, an amount of food equal to F1 minus F2 is exported and used to pay for clothing imports of C2 minus C1. Note that consumption of both products is greater at point K than at the no-trade equilibrium of point B. This differ- ence shows clearly the nature of the gain made possible when a nation engages in interna-
tional trade: it is possible to consume more of all goods. In effect, trade makes possible the
consumption of goods beyond the domestic production possibility frontier.
Figure 19.7 The PPF and the Gains from International Trade Without trade, equilibrium is at point B. When trade becomes possible at terms of 1F/1C, production takes place at point J, and exports of F1 minus F2 food units are exchanged for imports of C2 minus C1 clothing units. The country can attain more of both goods through trade (at point K) than it did without trade (at point B).
Food
Z
F1
F2
0 C1 C2 Z′ Clothing
Imports
Exports
w
g
K
B
J
w (slope = 1F/1C)
g (slope = 2F/1C)
Competit ive Markets and Economic Eff ic iency 497
C19.INDD 10:35:43:AM 08/06/2014 PAGE 497Trim Size: 203.2 mm X 254 mm
Does this analysis demonstrate that trade is beneficial for the country as a whole? As
discussed in Chapter 10, the answer to this question depends on what is meant by “country
as a whole.” Not everyone is likely to benefit. Since the adjustment to trade implies that
the domestic clothing industry will contract, the owners of resources specialized in cloth-
ing production are likely to find their real incomes reduced. Alternatively, in our general
equilibrium model, as the output of food expands at the expense of clothing (see Figure
19.4), the wage rate falls relative to the rental price of land, so workers may lose. Our
analysis shows that trade makes it possible for everyone to consume more of all goods, not
that everyone actually will. This idea has significant policy implications. The analysis indi-
cates, for example, that the trade liberalization brought about by the North American Free
Trade Agreement (NAFTA) and the World Trade Organization (WTO) may not result in all
Americans consuming more of all goods, though it does make it possible for them to do so.
Trade liberalization raises the average standard of living, but not everyone is average.
19.6 Competitive Markets and Economic Efficiency Economists tend to be advocates of perfect competition because it satisfies all three condi-
tions for economic efficiency. Let’s consider each of the three conditions in turn and show
why this is the case:
1. As shown in Chapter 6, a perfectly competitive economy results in an efficient distribution of products among consumers. To see why this is the case, recall that, in maximizing utility,
each consumer will select a basket of goods where the consumer’s MRS between any two goods (say, food and clothing) equals the ratio of the prices of the two goods, or:
P P MRSC F CF/ = . (4)
Since competitive markets establish a uniform price for each good, the ratio PC/PF confronting all consumers is the same. Because all consumers equate their individual
marginal rates of substitution to the same price ratio, the marginal rates of substitution
between goods will end up being the same across consumers:
MRS MRS MRSiCF CF CF1 2= == . . . , (5)
where the superscripts signify the i consumers in an economy. This is the condition for an efficient distribution of goods across consumers. A competitive equilibrium therefore
implies an efficient distribution of goods.
2. A perfectly competitive economy results in efficiency in production. This point was covered in Section 19.4. Recall that each firm producing a good minimizes cost by
employing inputs in quantities such that the marginal rate of technical substitution between
the inputs equals the input price ratio. In the context of a two-input economy (labor and
land), each firm thus equates:
w v MRTS/ LA= . (6)
Since competitive markets equalize input prices across firms and industries, the ratio w/v is the same for all firms. Therefore, the marginal rates of technical substitution end up being
the same across all firms and industries:
MRTS MRTS MRTSiLA LA LA1 2= == . . . , (7)
where the superscripts indicate the j producers in an economy. This is the condition for efficiency in production. It requires that we end up on the contract curve of the economy’s
Edgeworth production box and that the slopes of all producers’ isoquants are identical.
498 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 498Trim Size: 203.2 mm X 254 mm
3. A perfectly competitive economy results in efficiency in output. This can be shown by considering the equilibrium conditions of firms in output markets and the equilibrium
conditions of consumers. When clothing producers produce the profit-maximizing outputs,
they operate where marginal cost equals price, or:
P MCC C= . (8)
For food producers, the profit-maximizing condition is:
P MCF F= . (9)
Dividing equation (8) by equation (9) yields:
P P MC MCC F C F/ /= . (10)
We know that MCC/MCF equals the marginal rate of transformation between food and clothing (MRTCF) and that utility-maximizing consumers will equate their marginal rates of substitution between food and clothing to the ratio of prices of these two goods (that is,
MRSCF = PC/PF). By substitution, therefore, we obtain:
MRS MRS MRS MRTiCF CF CF CF1 2= = == . . . , (11)
where the superscripts refer to an economy’s i consumers. This is the condition for effi- ciency in output.
The preceding formal manipulations show that if perfect competition prevails, then all three conditions for economic efficiency are satisfied. Perhaps the most intuitive way of understanding why perfect competition efficiently solves the three basic economic prob-
lems of distribution, production, and output is to note that a competitive economy relies on
voluntary exchanges. Whenever any possible change in the allocation of either inputs or
goods promises mutual benefits to market participants, people have an incentive to work
out exchanges to realize these gains. If all mutually beneficial exchanges are consummated,
as they are in competitive markets, then no further change will benefit some without harm-
ing others. The outcome is efficient.
This discussion is proof, at an abstract level, of Adam Smith’s famous “invisible hand”
theorem: namely, that people pursuing their own ends in competitive markets promote an
important social goal—economic efficiency—that is not actually their intention and that
they may not even understand. In terms of our welfare frontier construct, it means that
competitive markets attain a point on the frontier. Although the one point on the welfare
frontier that represents a competitive equilibrium is not the only efficient resource alloca-
tion, we should recognize that competitive markets get us to a point on the welfare frontier,
which is no easy task.
The Role of Information Before closing our discussion of economic efficiency and why perfect competition attains it,
we should emphasize the important role of information in the process. When showing what
an efficient resource allocation looks like, we assumed that all the relevant information was
known: consumer preferences, production functions, and the quantities and productive capa-
bilities of inputs. Clearly, in the real world, with millions of consumers, firms, and products,
no one person knows or could possibly ever know all the relevant information needed to
attain economic efficiency. Take the case of even a simple lead pencil. Producers of pencils
purchase wood, graphite, steel, paint, and rubber from other people. Pencil producers cannot
produce these inputs themselves. Despite the fact that no one individual knows how to make
a pencil, much less an automobile or a personal computer, these items are produced. How?
The answer lies in the nature of a market system: partial bits of information possessed by
many different people are coordinated to produce a result that no one fully comprehends. The
Competit ive Markets and Economic Eff ic iency 499
C19.INDD 10:35:43:AM 08/06/2014 PAGE 499Trim Size: 203.2 mm X 254 mm
only information individual consumers or producers need to know about the rest of the econ-
omy to adjust their behavior is conveyed through prices. For example, if the supply of cotton
expands while the supply of wool declines, the price of cotton clothing will fall relative to the
price of wool clothing. Buyers will substitute cotton for wool in their apparel, using more of
the plentiful fabric and economizing on the scarce. This efficient response can, and probably
will, occur without anyone knowing why prices changed the way they did.
A market system can function efficiently without any single individual understanding how.
In this sense markets economize on the information people individually require to coordinate
their economic activities. An immense amount of information must be utilized to achieve an
efficient resource allocation. Perhaps the most significant implication of our analysis is that, in
principle, an efficient outcome can be accomplished by decentralized, voluntary transactions
among people, each of whom has only a tiny portion of the requisite information.
APPLICATION 19.1
During the early part of the twentieth century, econo- mists Oscar Lange and Abba Lerner argued that it would be possible to attain Pareto optimality through central planning. All that was necessary, according to them, was individual consumers and producers reporting to a central planning board information about underlying consumer preferences, production technology, input availability, and so on. The central planning board could then specify the amount of each commodity to be produced, the input usage levels to be used in production, and the distribution of goods among consumers. The Lange and Lerner scheme gave credence to the fascist and communist governments coming into power at that time and raised the possibility that such forms of government might be more effective in promoting efficiency than governments relying on decen- tralized, market-based economies.
Among the earliest critics of the hypothesis advanced by Lange and Lerner was Austrian economist Friedrich Hayek.4 Hayek argued that central planning inevitably must fail because it can never fully accommodate the particular and changing information about costs and/or demand pos- sessed by individual consumers and producers. Moreover, central planning also undermines the incentive consumers and producers have to acquire information and to act on the information they have acquired. As Hayek stated:
[Knowledge of this kind] cannot be conveyed to any central authority in statistical form. The statistics . . . would have to be arrived at precisely by abstracting from minor differences between the things, by lump- ing together, as resources of one kind, items which dif- fer in regards to location, quality, and other particulars,
Can Centralized Planning Promote Efficiency?
in a way which may be very significant for the specific decision. It follows from this that central planning based on statistical information by its nature cannot take direct account of these circumstances of time and place and that the central planner will have to find some way or other in which the decisions depending on them can be left to the “man on the spot.” . . .
[T]he ultimate decisions must be left to the people who are familiar with these circumstances, who know directly of the relevant changes and of the resources immediately available to meet them. We cannot expect that this problem will be solved by first communicating all this knowledge to a cen- tral board which, after integrating all knowledge, issues its orders. We must solve it by some form of decentralization.
[Where] knowledge of the relevant facts is dispersed among many people, prices can act to coordinate the actions of separate people. . . . The marvel is that in a case like the scarcity of one raw material, with- out an order being issued, without more than per- haps a handful of people knowing the cause, tens of thousands of people whose identity could not be ascertained by months of investigation, are made to use the material or its products more sparingly. . . . I have deliberately used the word “marvel” to shock the reader out of the complacency with which we often take the working of this [price] mechanism for granted. I am convinced that if it were the result of deliberate human design, and if the people guided by the price changes understood that their deci- sions have significance far beyond their immediate aim, this mechanism would have been acclaimed as one of the greatest triumphs of the human mind. Its
4This application is based on Friedrich A. Hayek, “The Use of Knowl- edge in Society,” American Economic Review, 35, No. 4 (September 1945), pp. 519–530.
500 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 500Trim Size: 203.2 mm X 254 mm
misfortune is the double one that it is not the prod- uct of human design and that the people guided by it usually do not know why they are made to do what they do. But those who clamor for “conscious direction” (i.e., central planning)—and who can- not believe that anything which has evolved with- out design (and even without our understanding it) should solve problems which we should not be able
to solve consciously—should remember this: The problem is precisely how to expand the span of our utilization of resources beyond the span of control of any one mind; and, therefore, how to dispense with the needs of conscious control and how to pro- vide inducements which will make the individuals do the desirable things without having to tell them what to do.
19.7 The Causes of Economic Inefficiency A market may fail to satisfy the conditions for Pareto optimality for several reasons, includ-
ing market power, imperfect information, and externalities/public goods. We briefly dis-
cuss each of these reasons.
Market Power The preceding section showed that competitive markets, without government intervention,
will result in the attainment of Pareto optimality. This will not be the outcome, however, if
producers or consumers have some market power and perfect competition does not prevail.
To see why monopoly or monopsony power results in economic inefficiency, consider the
case of a monopoly in an output market. Suppose, for example, that in our simplified two-
good economy, the food industry is competitive while the clothing industry is controlled by
a monopoly seller.
The clothing monopoly maximizes its profit by reducing output below the competitive
level and setting the per-unit clothing price above the marginal cost. Since the price of food
equals its marginal cost (we assumed that the food market is competitive), the relative price
of clothing exceeds its relative marginal cost:
P P MC MCC F C F/ /> . (12)
The preceding inequality indicates that an output market monopoly violates the output
efficiency condition. With a monopoly seller in the clothing market, the rate at which con-
sumers are willing to trade food for clothing (MRSCF, which equals PC/PF provided con- sumers are utility maximizers) exceeds the marginal rate of transformation between food
and clothing on the production side (MRTCF, or MCC/MCF).5 More clothing and less food should be produced since the marginal benefit of more
clothing, in terms of food, exceeds its marginal cost at the monopoly equilibrium. Such a
change in output mix, however, will not occur, because it would be contrary to the profit-
maximizing interest of the monopoly clothing supplier.
A similar inefficiency results in the case of monopoly in input markets. For example,
suppose that the market for labor in the clothing industry is monopolized by a union.
5In the context of Figure 19.6, the clothing monopoly results in a point on the PPF being realized some- where between Z and P (less than the optimal amount of clothing produced). The slope of the PPF at the realized point is less than the slope of all consumers’ indifference curves (as indicated by line hh in Figure 19.6). That is, the rate at which consumers are willing to trade food for clothing (given by the slope at line hh) is greater than the marginal rate of transformation between food and clothing (as indicated by the slope of the PPF at the realized point between Z and P with a clothing monopoly).
The Causes of Economic Ineff ic iency 501
C19.INDD 10:35:43:AM 08/06/2014 PAGE 501Trim Size: 203.2 mm X 254 mm
This implies that in our simple, two-good economy, the ratio of the labor wage rate (w) to the rental price of land (v) will be higher in the clothing industry than in the food industry (provided that competition characterizes all other input markets):
w v w vC F/ /> . (13)
Since profit-maximizing firms equate their MRTS between labor and land to the relative input costs (that is, MRTSLA = w/v), the rate at which producers are willing to exchange land for labor units (MRTSLA) will be greater in the clothing than in the food industry. In equilibrium, that is, labor is relatively more productive if employed in clothing manufactur-
ing than in food production. The input-pricing actions of the union, however, prevent the
movement of labor from the food to the clothing industry.
In the context of a PPF such as that in Figure 19.5, the union’s action results in an out- come such as point G inside the frontier. Note that this outcome occurs not because resources are unemployed (since we assume that all inputs are employed), but because of an inefficient
allocation of inputs between the food and the clothing industries—inefficiency in production.
Total output is lower than it could be, as is total consumption. At least in terms of total eco-
nomic surplus, society is not as well off at point G as it could be if it were operating on the production frontier.
APPLICATION 19.2
While the goal of the New Deal was to get Americans back to work, a policy at the heart of the recovery plan, the National Industrial Recovery Act (NIRA), actually served to lower employment.6 NIRA covered more than 500 industries, ranging from autos, steel, ladies’ hosiery, and poultry production, and it removed the applicability of antitrust laws. The relevant industries were permitted to collusively raise prices, provided that they shared their monopoly rents with workers by increasing wages beyond those justified by productivity growth. NIRA distorted the economy by promoting output restrictions, artificially raising prices and wages, and diminishing productive capacity through quotas on industry investment in new plants and equipment.
The negative impact of NIRA on the efficiency of the U.S. economy was significant. Economists Harold Cole of the University of Pennsylvania and Lee Ohanian of UCLA have estimated that manufacturing wages were as much as 25 percent higher in industries than they would have been without NIRA. Prices and wages in other sectors such as agriculture not covered by NIRA witnessed no similar increase.
While the artificially higher wages benefited workers able to secure jobs in industries covered by NIRA’s codes, they appreciably depressed overall production and employment.
How Government Prolonged the Great Depression
Total private-sector hours worked per adult were 27 percent lower during the New Deal (1933–1939) than they were in 1929, before the start of the Great Depression, and com- pared to 18 percent lower between 1930 and 1932, prior to the launch of the New Deal.
Even though NIRA was declared unconstitutional in 1935, there were virtually no antitrust actions pursued by the Justice Department despite overwhelming Federal Trade Commission evidence of price fixing and production limits in many industries. Moreover, the National Labor Relations Act of 1935 substantially enhanced unions’ collective-bargaining power, legislation that resulted in large wage hikes above NIRA levels and helped precipitate the economic downturn of 1937–1938.
In a 1938 speech, President Franklin Delano Roosevelt acknowledged that the U.S. economy had become a “con- cealed cartel system like Europe,” resulting in the Justice Department re-initiating antitrust activity. In addition, union bargaining power was curtailed when the courts ruled sit-down strikes to be illegal toward the end of the 1930s and the National War Labor Board, after the onset of World War II, limited union wage settlements to cost-of- living increases.
Cole and Ohanian estimate that New Deal labor and industrial policies prolonged the Great Depression by seven years. The economic boom of World War II, moreover, reflected not only dramatically increased government spend- ing on military goods and services but also the erosion of New Deal labor and industrial policies.
6This application is based on Harold L. Cole and Lee Ohanian, “How Government Prolonged the Depression,” Wall Street Journal, February 2, 2009, p. A17.
502 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 502Trim Size: 203.2 mm X 254 mm
APPLICATION 19.3
It is commonly believed that government intervention is warranted when information is imperfect. An example that such a belief may not be well founded, however, is provided by efforts to determine the health consequences of ciga- rette smoking. Economist John Calfee, formerly of the Fed- eral Trade Commission, has examined the historical role of government in improving information about the health risks associated with cigarettes and found that, if anything, the pri- vate sector did more than the government in the early half of the twentieth century to point out the potential risks.7
Prior to stronger scientific evidence in the 1950s and early 1960s showing that smoking cigarettes had adverse health consequences, competition between different sup- pliers served to highlight the dangers. That is, while smok- ing was widely considered glamorous from the 1920s to the 1950s, it was also popularly described in such unglamorous terms as “coffin nails,” “smoker’s cough,” “gasper,” and “lung duster”—despite a lack of scientific evidence about smok- ing’s long-run effects on mortality rates.
Rather than suppressing smokers’ fears or arguing that they were unfounded, cigarette manufacturers relied on these fears as an advertising tool. Among the slogans employed by particular brands were “Not a cough in a car- load” (Chesterfield), “Not a single case of throat irritation due to smoking Camels,” and “Why risk sore throats?” (Old Gold).
In the early 1950s, two well-designed scientific studies sponsored by the American Cancer Society linked smoking with lung cancer. Newspapers and magazines such as Reader’s Digest provided extensive coverage of the academic stud- ies, and Consumer Reports began publishing tar and nicotine ratings for all cigarette brands. Cigarette producers, notably the smaller firms, began introducing brands with filters that greatly reduced tar and nicotine. Filtered brands grew from
Deterring Cigarette Smoking
1 to 10 percent of the total cigarette market between 1950 and 1954. The companies selling them sought to spur their sales by scaring smokers about rival brands: “Filtered smoke is better for your health” (Viceroy), “just what the doctor ordered” (L&M), and “[Kent] takes out more nicotine and tars than any other leading cigarette—the difference in protec- tion is priceless.” Television advertisements showed the dark smoke left by competing unfiltered brands on Kent’s filter.
Producers’ competitive actions thus informed smokers about the health risks associated with cigarettes. Accord- ing to Calfee, these actions, coupled with the reports in the popular press, did more to deter smoking through the early 1960s than government actions. If anything, Calfee argues, government actions actually hindered the dissemination of information about smoking’s risks after a 1955 Federal Trade Commission regulation forbidding producers from making any tar and nicotine claims until “it has been established by competent scientific proof that the claim is true, and if true, that such difference or differences are significant.” Although the guidelines explicitly permitted the advertising of taste and pleasure, any references to the presence or absence of physical effects of smoking were banned.
Predictably, cigarette advertisements changed to stress- ing taste and pleasure rather than the fear associated with smoking. Kent advertisements changed from “signifi- cantly less tars and nicotine” to “satisfies your appetite for a real good smoke.” Duke, one of the new low-tar brands, switched from “lowest in tars” to “designed with your taste in mind.” Cigarette sales ended their several-year decline in 1955 and rose significantly in the ensuing decade. In 1966, acceding to appeals from the American Cancer Soci- ety, the Federal Trade Commission reversed its policy and authorized tar and nicotine advertising. It was not until 1965, moreover, that the Federal Trade Commission began requiring all cigarette packages to carry a warning from the Surgeon General. The same warning was mandated on all cigarette advertisements effective 1972.
7John E. Calfee, “The Ghost of Cigarette Advertising Past,” Regulation, 10, No. 2 (November/December 1986), pp. 35–45.
Imperfect Information If consumers or producers are not accurately informed, they may take actions that run coun-
ter to the dictates of Pareto optimality. For example, consumers may mistakenly think that a
certain “miracle” gel can help reverse hair loss. Upon applying the gel to their scalps, the con-
sumers may be disappointed to find that their hair continues to thin. Worse yet, some consum-
ers may end up bald. Likewise, if a computer manufacturer is unaware that a particular chip
has a computational glitch, it may use the chip extensively in the production of computers.
Of course, the existence of imperfect information does not in and of itself imply that
government intervention can best remedy the problem. After all, information may be con-
sidered to be just another good for which private markets provide better production and
consumption incentives than do government edicts.
Review Quest ions and Problems 503
C19.INDD 10:35:43:AM 08/06/2014 PAGE 503Trim Size: 203.2 mm X 254 mm
Externalities/Public Goods Sometimes, in the process of producing or consuming certain goods, harmful or benefi-
cial side effects called externalities are borne by people not directly involved in the market activities. Take the case of a motorist choosing to travel an urban freeway during rush hour.
The motorist may impose congestion costs on other drivers, costs for which the motorist
is not directly accountable. Other goods, known as public goods, simultaneously provide benefits to multiple consumers. For example, the same parade, park, or B-1 bomber may
enhance the well-being of more than one consumer.
Because the benefits or costs of a good may not be fully accounted for by market
actors, externalities and public goods can result in inefficiency—even if competitive
markets prevail. In the following chapter, we discuss the reasons why as well as the best
mechanisms for promoting efficiency in the case of externalities and public goods.
SUMMARY
Partial equilibrium analysis concentrates on one mar-
ket at a time, viewing that market as independent from
other markets.
General equilibrium analysis views the economy
as a network of interconnected markets, with events in
one market affecting others and, in turn, being affected
by others. Mutual interdependence among markets is
emphasized.
The concept of economic efficiency, or Pareto optimal-
ity, defines a situation in which no person’s well-being
can be further improved unless someone else is harmed.
Three conditions determine whether an economy is
operating efficiently. First, the goods produced must be
efficiently distributed among consumers. Efficient distri-
butions occur at points on the contract curve in an Edge-
worth exchange box.
Second, inputs must be allocated efficiently in the
production of goods. Efficiency in production is shown
by points on the contract curve in an Edgeworth produc-
tion box or, equivalently, by the economy’s operation
on, rather than inside, its production possibility frontier.
Third, the output mix produced must be efficient. Effi-
ciency in output is identified by a point on the production
frontier where, when the outputs are distributed among
consumers, each consumer’s marginal rate of substitu-
tion between any two goods equals the marginal rate of
transformation between those two goods.
If perfect competition prevails, all three efficiency
conditions are satisfied and an economy will end up at a
point on its welfare frontier. The reasons why economic
efficiency may not be realized include market power,
imperfect information, externalities, and public goods.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
19.1 What do economists mean when they say markets are mutually interdependent? Give an example to support your explanation.
*19.2 In the Figure 19.1 butter–margarine example, would there be any spillover effect on the butter market from the margarine
market if the supply curve of margarine were horizontal?
19.3 What does the contract curve in an Edgeworth production box signify? Why do competitive markets generate equilibri-
ums that lie on the contract curve?
19.4 What is the relationship between the PPF and the contract curve in an Edgeworth production box?
*19.5 If all industries are in competitive equilibrium, and the price of personal computers is 10 times the price of cellular
telephones, what is the MRT between the two goods?
19.6 The domestic computer chip manufacturing industry argues that permitting free trade will cost the jobs of thousands
of computer chip workers. How does general equilibrium anal-
ysis help in responding to this argument?
19.7 What factors determine whether a particular economic issue can be adequately analyzed by using a partial rather than
a general equilibrium approach?
504 General Equi l ibr ium Analys is and Economic Eff ic iency
C19.INDD 10:35:43:AM 08/06/2014 PAGE 504Trim Size: 203.2 mm X 254 mm
19.8 Is every efficient allocation of resources preferred to every inefficient allocation of resources?
19.9 Explain why, when all markets are competitive and in equilibrium, all three conditions for efficiency are satisfied.
Does this result indicate that society’s welfare is maximized?
*19.10 According to Albert Einstein, “The economic anarchy of capitalist society as it exists today is, in my view, the main
cause of our evils. Production is carried on for profit, not for
use.” Is there a conflict between “production for profit” and
“production for use”?
19.11 If Cisco has a monopoly in the server market, what effi- ciency condition is violated? Would the regulation of Cisco
and the elimination of Cisco’s profit lead to a more efficient
allocation of resources? Will all members of society benefit?
*19.12 Ignoring rationing problems and black markets, under rent control (or any price ceiling that produces a shortage) the
price paid by consumers equals the marginal cost of producing
the good. Does this mean the output level is efficient? Explain.
19.13 “Using efficiency as a criterion biases the analysis in favor of the status quo, since any change is certain to harm
someone.” Discuss.
19.14 In each of the cases below, state whether one of the con- ditions for economic efficiency is violated. “Uncertain” is an
acceptable response. If one of the efficiency conditions is vio-
lated, indicate which one and whether the resources in question
are overused or underused.
a. The Rapid Transit District charges reduced bus fares to its senior citizens.
b. There is a limit to the number of people who can legally immigrate to the United States from India per year.
c. Some neighborhood families do not regularly mow their lawns.
d. The market for hot dogs is perfectly competitive. Michael Jordan’s consumption of a hot dog leaves fewer hot dogs
available for the rest of the world.
e. The Federal Trade Commission provides free pamphlets helping potential used car buyers identify whether a car’s
odometer has been rolled back.
f. A per-unit tax is applied to clothing in the context of a two-good (food-and-clothing), two-input (labor-and-capital)
economy.
g. A selective minimum wage is imposed by the government on labor employed in clothing production in the context of a
two-good (food-and-clothing), two-input (labor-and-capital)
economy.
19.15 Why might a resource allocation that achieves efficiency in production not satisfy the condition for efficiency in output?
Provide a real-world example.
19.16 Suppose that in the production of computer software, the marginal rate of technical substitution between engineers
and marketers is 5 for IBM and 3 for Microsoft. Explain why
this outcome violates the condition for efficiency in produc-
tion and how a voluntary exchange could make both compa-
nies better off.
19.17 Most former communist governments of Eastern Europe subsidized food production (both in absolute terms and rela-
tive to any subsidies provided other goods). Explain the effect
of this policy on the relationship between the typical Eastern
European consumer’s marginal rate of substitution between
food and all other goods (treated as a composite good) and
the marginal rate of transformation between food and all other
goods.
*19.18 In the international trade example, we implicitly assumed that the world price ratio was unaffected when the
domestic country engaged in trade. Under what conditions is
this assumption reasonable? If the world price ratio is affected,
how will it change? How will this change affect the analysis?
19.19 Under marketing orders instituted during the 1930s and administered by the U.S. Department of Agriculture, orange
growers in California and Arizona have been successful in
behaving as a cartel in the fresh orange market. Despite the
ability of California and Arizona growers to rely on market-
ing orders to cartelize the fresh fruit market, explain why, from
a general equilibrium perspective, marketing orders have had
only a limited effect on grower profits because of the fact that
fruit can be diverted to secondary, processed food markets such
as orange juice concentrate.
19.20 All points on the welfare frontier depicted in Figure 19.2 are efficient. There is no reason, therefore, why one point on the
frontier should be preferred to another. True, false or uncertain?
Explain.
19.21 Suppose that the United States limits the amount of steel that can be imported from other countries. Using a
PPF that puts units of steel on the horizontal axis and units of another good, such as food, on the vertical axis, explain
how such a steel import quota will affect production of food
and steel in the United States and alter our consumption pos-
sibilities. Will the quota make the United States better off as a
whole? If not, will it make anyone in the United States better
off? Explain.
505
C20.INDD 10:34:15:AM 08/06/2014 PAGE 505Trim Size: 203.2 mm X 254 mm
As we have seen in several preceding chapters, government intervention in markets may fail to promote economic efficiency. For example, Chapter 10 showed how rent control and
a quota on sugar imports can diminish the total surplus realized by market participants as a
group. Without discounting the impediments to efficiency that may be associated with gov-
ernment intervention, this chapter looks at two important reasons markets left to themselves
may also not function efficiently: public goods and externalities. Public goods are those that benefit all consumers, such as national defense. A public good will be undersupplied
by a market when consumers cannot be excluded from sharing in its benefits and thus have
no incentive to pay for its production.
public goods goods that benefit all consumers, such as national defense
Memorable Quote “The buffalo nobody’s property Went o’er the plains clippity cloppity In thunderous herds, where now only birds Fly and rabbits go hippity, hoppity. The cow, now, is kept on the farm And flourished and came to no harm For its owners to thrive Had to keep it alive So property worked like a charm.”
—Ode to Property Rights by economist Kenneth E. Boulding, contrasting the different fates of buffalo and cow herds in the late 1880s.
Learning Objectives
Explain what economists mean by the term public goods and the free-rider problem. Describe efficiency in the provision and distribution of a public good. Define external benefits and external costs and show how their presence results in nonopti- mal output levels for goods characterized by such aspects. Show how clearly defined and enforced property rights can resolve externality problems and thereby ensure an efficient outcome. Demonstrate how air pollution can more efficiently be controlled through the establishment of an overall industry pollution target and the assignment of tradable emissions permits to the industry’s firms.
Public Goods and Externalities20CHAPTER
506 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 506Trim Size: 203.2 mm X 254 mm
Externalities are present when all of the costs or benefits of a good are not fully borne by market participants. For example, an oil refinery may not have to pay for some of the air
pollution generated by its production process and may consequently produce more oil than
is economically efficient. Individuals may not obtain a flu shot if some of the benefits of the
vaccination against such a communicable disease accrue to society at large rather than fully
to them.
When public goods or externalities lead markets to generate an inefficient allocation of
resources, government can intervene, at least in theory, with an appropriate policy that will
improve things. This chapter analyzes how public goods and externalities may adversely
affect the way resources are allocated by markets as well as the remedies, government regu-
lation among others, to such impediments to economic efficiency.
20.1 What Are Public Goods? The term public good, as used by economists, does not necessarily refer to a good provided by the government. Instead, economists define a public good by the characteristics of the
good itself. Two are important: nonrival in consumption and nonexclusion.
A good is nonrival in consumption if, with a given level of production, consumption by one person need not diminish the quantity consumed by others. Although this defini-
tion may sound peculiar, such goods do exist. Consider a nuclear submarine that reduces
the likelihood of enemy attack. Your property and person are protected, and so are those
of others. The protection you receive in no way diminishes the extent to which others are
protected. Another example is a flood control project that reduces the probability of flood
damage. Less flood damage to one home does not mean more flood damage to another;
all persons in a given area simultaneously benefit in the form of a reduced likelihood of
flooding.
In effect, nonrival consumption means potential simultaneous consumption of a good
by many persons. By contrast, most goods are rival in consumption. For a given level of
production of shoes (or soft drinks, T-shirts, cars or hamburgers), the more you consume,
the less is available for others. In these cases consumption is rival because the economic
system must ration output among competing (rival) consumers. When a good is nonrival
in consumption, the good need not be rationed. Once it is produced, the good can be made
available to all consumers without affecting any individual’s consumption level.
The second characteristic of a public good is nonexclusion. Nonexclusion means that confining a good’s benefits (once produced) to selected persons is impossible or prohibi-
tively costly. Thus, a person can benefit from a good’s production regardless of whether he
or she pays for it. Although the concepts of nonrivalry and nonexclusion often go together,
they are distinct. Nonrivalry means that consumption by one person need not interfere with consumption of others; although a good may be nonrival in consumption, restricting con-
sumption to selected persons may still be possible.
For example, when a Web site is posted, anyone with Internet access can go to the Web
site and view its contents without interfering with another person’s ability to view the same
site. (An exception would be if the site suddenly got a huge number of hits, overloading
the server.) It is possible to deliberately exclude access to a Web site, however, and in
fact, it is often done. For instance, the Web site for the Wall Street Journal, www.wsj.com, includes free content that anyone can access. However, only subscribers can access more
detailed information, such as front-page stories from the Journal. Clicking on those areas brings up the message, “The page you requested is available only to subscribers.” Subscrib-
ers must supply a user number and a password, and nonsubscribers are denied access. The
Web illustrates how some things can be nonrival and yet have the possibility of exclusion.
Thus, it does not meet the criteria to be a public good.
externalities the harmful or beneficial side effects of market activities that are not fully borne or realized by market participants
nonrival in consumption a condition in which a good with a given level of production, if consumed by one person, can also be consumed by others
nonexclusion a condition in which confining a good’s benefits, once produced, to selected persons is impossible or prohibitively costly
What Are Publ ic Goods? 507
C20.INDD 10:34:15:AM 08/06/2014 PAGE 507Trim Size: 203.2 mm X 254 mm
In contrast, national defense is an example of a good with both characteristics. A given
defense effort protects (or endangers) everyone simultaneously, and to limit the protection
to certain people is impossible. The benefits of defense are thus nonrival to the population,
and exclusion of selected persons is infeasible.
A good that is nonrival in consumption and has high exclusion costs creates problems
for a market system. Once such a good is produced, many people will automatically benefit
regardless of whether they pay for it, because they cannot be excluded. As we will see, this
feature makes it unlikely that private producers will provide the good efficiently.
The Free-Rider Problem Even when a public good is worth more to people than it costs to produce, private markets
may fail to provide it. To see why, consider the construction of a dam that will lessen the
probability of flooding for a community’s residents; the dam is a public good for the residents.
It may have a total cost of $1,000,000, and business firms will be willing to build it if someone
will provide the funds. If 10 persons live in the community and the benefit of the dam to each
person is $200,000, then the total benefit of the dam to all 10 residents is $2,000,000—twice
as much as it costs. All 10 people would be better off if each contributed $100,000 to finance
construction costs, since each would then receive a benefit valued at $200,000 from the dam.
Even though it is in each resident’s interest to have the dam built, there is a good chance
that it won’t be built if private markets are relied upon to organize the construction. To
finance the dam, residents must jointly agree to contribute, but many will realize that they
get the benefit of the dam once built, regardless of whether they contribute toward its con-
struction. Each resident, therefore, has an incentive to understate what the dam is worth in an effort to secure the benefit at a lower, or zero, cost. If enough people behave this way— as a free rider—voluntary contributions will be insufficient to finance the dam, and it won’t be built. Viewing the provision of public goods as a prisoner’s dilemma, free riding
is the equivalent of “cheating” in the prisoner’s dilemma game discussed in Chapter 14.
free rider a consumer who has an incentive to underestimate the value of a good in order to secure its benefits at a lower, or zero, cost
APPLICATION 20.1
In 2000, horror writer Stephen King became the first major author to self-publish online.1 King asked readers to pay him $1 for each chapter of a serial e-novel, titled The Plant, they downloaded and warned that he would not post
An Online Horror Tale
new installments unless he received payments for at least 75 percent of the downloads. Voluntary contributions for King’s e-novel appeared to be plagued by the free-rider problem, as only 46 percent of the downloads were paid for. King promptly called it quits on publishing The Plant online in order to work on other, more conventional books from which it is easier to exclude nonpayers.
1This application is based on “A Stephen King Online Horror Tale Turns into a Mini-Disaster,” www.nytimes.com, November 29, 2000.
Free Riding and Group Size When public goods are involved, free riding is rational, but it hinders the ability of private
markets to cater efficiently to the demand for a public good. In the example just discussed,
enough people could conceivably contribute so that the dam would be financed by volun-
tary agreements. With just 10 people involved, only a small number need to agree to con-
tribute. The severity of the free-rider problem, however, varies with the number of people
involved. The larger the number of people receiving benefits from a public good, the less
likely that voluntary cooperation will ensure its provision.
508 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 508Trim Size: 203.2 mm X 254 mm
As the group size increases, it is more likely that everyone will behave like a free rider, and the public good will not be provided. To illustrate, let’s change our example slightly and assume that a dam now benefits 1,000 people, each by $2,000. (Note that the total
benefit is still $2,000,000, just as before.) In this case, faced with deciding whether or how
much to contribute voluntarily, each person will realize that one single contribution has
virtually no effect on whether the dam is built. Put differently, the outcome depends mainly
on what the other 999 people do, and whether any one person contributes will not affect the
others’ decisions. In this case each person gets the same benefit whether or not any contri-
bution is made, and choosing not to contribute is the most rational behavior. Because this
is true for everyone, few people will contribute, and the good likely will not be provided.
Many real-world examples provide evidence of free-rider behavior with public goods.
A particularly clear-cut example occurred in 1970 (before mandatory pollution controls)
when General Motors tried to market pollution control devices for automobiles at a price of
$20. The emission controls would have reduced the pollution emitted by 30 to 50 percent.
Pollution abatement is, of course, a public good, at least over a certain geographic area. It is
reasonable to suppose that the benefits of a 30 to 50 percent reduction in automobile pollu-
tion far outweighed the cost of $20 per car. Yet GM withdrew the device from the market
because of poor sales. This example illustrates the large-group free-rider problem at work.
Everyone might have been better off if all drivers used the device, but it was not in the
interest of any single person to purchase it because the overall level of air quality would not
be noticeably improved as a result of any one individual’s action.
When the benefits of a public good are nonrival over a large group, private markets
probably will not provide it. Even if some amount of the public good is provided through
the contributions of a few people, it will be at a suboptimal quantity. This result is true
even when it is in the interest of people to have the good provided—that is, even when
the benefits exceed the costs. Competitive markets cannot in general supply public goods
efficiently. This fact provides a major justification for considering governmental alterna-
tives. In the dam example of 1,000 persons, for instance, the government could levy a tax
of $1,000 on each person and use the $1,000,000 in tax revenue to finance the dam. Each
person would be made better off by this policy, receiving $2,000 of benefit from the dam
at a cost of $1,000 in taxes. The government expenditure of $1,000,000 on the dam would
lead to a more efficient allocation of resources than reliance on private markets.
20.2 Efficiency in the Provision of a Public Good What is the efficient output of a public good? As usual, we must compare the marginal
benefit and marginal cost associated with different levels of output. The marginal cost of a
public good is the opportunity cost of using resources to produce that good rather than others,
just as it is in the case of the nonpublic, or private, goods discussed in previous chapters.
Because of the nonrival nature of the benefits of a public good, though, its marginal benefit
differs from that of a private good. With a good like a cheeseburger, the marginal benefit
of producing an additional unit is the value of the cheeseburger to the single person who
consumes it. With a public good like defense, the marginal benefit is not the marginal value
to any one person alone because many people benefit simultaneously from the same unit.
Instead, we must add the marginal benefits of every person who values the additional unit
of defense, and the resulting sum indicates the combined willingness of the public to pay
for more defense—that is, its marginal benefit.
Figure 20.1 shows how we derive the demand, or social marginal benefit curve for a public good like submarines. For simplicity, assume that only two people, Ted and Jane,
benefit from the defense services of submarines. Each person has a demand curve for sub-
marines, shown as dT and dJ. These demand curves are derived from each person’s
social marginal benefit curve the demand curve for a public good
Eff ic iency in the Provis ion of a Publ ic Good 509
C20.INDD 10:34:15:AM 08/06/2014 PAGE 509Trim Size: 203.2 mm X 254 mm
indifference curves, just as would their demand curves for a private good. Recall that the
height of a consumer’s demand curve indicates the marginal benefit of another unit of the
good. To derive the social, or combined, demand curve, we must add the marginal benefits of the two consumers. Geometrically, the combined demand curve involves a vertical sum- mation of the consumers’ demand curves. For example, the marginal benefit to Ted of the first submarine ($400) is added to the marginal benefit Jane receives from the first subma-
rine ($250) to determine the social marginal benefit ($650) for the first unit. This vertical
addition of marginal benefits identifies one point on the social marginal benefit curve, indi-
cating that the combined marginal benefit of Ted and Jane for the first submarine is $650.
At alternative quantities of submarines, we continue to add the heights of each consumer’s
demand curve to trace out the entire social marginal benefit curve, MBS. We can now determine the efficient output of submarines. At any output where MBS lies
above the marginal cost curve MC—drawn here for simplicity as horizontal at $500—Ted and Jane are willing to pay more for the unit than its marginal cost, so efficiency requires
a higher output. Thus, an additional unit could be financed in a way that makes both of
them better off—with each contributing somewhat less than the maximum amount they are
willing to pay. When MC lies above MBS, however, too much of the public good is being produced—the combined marginal benefit as shown by MBS is less than marginal cost over this output range. Thus, the efficient output is 10 submarines, where Ted’s marginal benefit
of $325 plus Jane’s marginal benefit of $175 just equals the marginal cost of $500.
In general, the efficient output of a public good occurs where MBS, obtained by ver- tically summing the demand curves of all consumers, intersects the marginal cost curve. There is no presumption, however, that this output will be the actual output. In fact, we
have already seen that the free-rider problem generally means that private markets will not
produce the efficient output. The government has the power to finance the efficient output
from tax revenues, but whether it actually does so depends on how political forces deter-
mine public policy.
Government financing of a public good overcomes one aspect of the free-rider problem—
the tendency of people to withhold payment. There is another aspect of free riding that gov-
ernment financing does not overcome: people have no incentive to accurately reveal their
demands for the public good—especially if they will be taxed commensurate to the ben-
efits they report receiving. To determine the efficient output, we must know every person’s
vertical summation (of demand curves) the derivation of a social marginal benefit curve through the summing of consumers’ marginal benefit curves
The Efficient Output of a Public Good Because the benefits of a public good are nonrival, the social marginal benefit is the sum of the marginal benefits of the separate consumers. Graphically, the social marginal benefit curve is constructed by vertically summing the consumers’ demand curves. The efficient output is identified by the intersection of MBS and MC.
Figure 20.1 Dollars per submarine
$650
$500
$400
$325
$175
0 1 10
MBS
MC
dT
dJ
Submarines
$250
510 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 510Trim Size: 203.2 mm X 254 mm
demand curve so that we can vertically add them to obtain MBS. How can we determine the true worth of a public good like defense to millions of people? Needless to say, obtaining
this information is problematic.
For example, to obtain the efficient output level of a public good such as the one depicted
in Figure 20.1, the government can tax people according to the heights of their reported mar-
ginal benefit curves. Thus, for 10 submarines, Ted and Jane are taxed $325 and $175 per unit,
respectively. The total amount paid in taxes ($500 per submarine) is just sufficient to cover
the marginal cost of producing an additional submarine at the efficient output of 10.
Where citizens are taxed according to their reported marginal benefit curves, however,
they have an incentive to understate their benefits. For example, Jane may be tempted to
say that she gets no benefit from submarines (and thus pay no taxes) and free ride on any
payments made by Ted—since any submarines Ted pays for through taxes also benefit
Jane. Ted has the same incentive to understate the benefit he gets from submarines. Under-
statement of demand implies suboptimal provision of the public good.
Efficiency in Production and Distribution In Chapter 19, we pointed out that there are three conditions for efficiency. These condi-
tions also apply to public goods. So far, we have emphasized only the condition for an
efficient level of output. A second condition is that the output be produced by using the
least costly combination of inputs. In Figure 20.1, that condition is implicit in the assump-
tion that a marginal cost of $500 is the minimum cost necessary to produce a submarine.
The third condition relates to the efficient distribution of the good among consumers. For a
private good this condition requires an equality of marginal rates of substitution. But how is
a public good rationed efficiently?
With a public good, there is no rationing problem. If 10 submarines are produced, both Ted and Jane simultaneously benefit, and the benefit to one in no way diminishes the benefit
to the other. For example, suppose it were possible in some hard-to-imagine way to have Ted
protected by 10 submarines but Jane protected by only five. In other words, if exclusion were
APPLICATION 20.2
Lojack is a hidden radio-transmitter device used for retriev- ing stolen vehicles. The annual cost of installing Lojack is roughly $100 per car owner, while the estimated marginal social benefit per unit of Lojack installed is roughly $1,500 per year in terms of reduced auto thefts.2 As noted by econ- omists Ian Ayres and Steve Levitt, there are various reasons why the presence of Lojack is associated with significant external benefits that are not fully captured by the individual consumer opting to install the device in his or her car:
First and foremost, Lojack disrupts the operation of “chop-shops” where stolen vehicles are disas- sembled for resale of parts. In the absence of Lojack,
The Lowdown on Why Lojack Installations Are Lower Than the Efficient Output
identifying chop-shops requires time-consuming, resource-intensive sting operations. With Lojack, police following the radio signal are led directly to the chop-shop. In Los Angeles alone Lojack has resulted in the breakup of 53 chop-shops. Second, data col- lected in California suggest that the arrest rate for stolen vehicles equipped with Lojack is three times greater than for cars without Lojack (30 percent versus 10 percent). Since most thieves are repeat offenders, arrests that lead to incarceration . . . also provide social benefits via reductions in victimiza- tions while the criminal is behind bars.
Since drivers who install Lojack in their vehicles obtain only a portion of the total social benefits, we can predict that private markets will produce less than the efficient amount of Lojack installations.
2This application is based on Ian Ayres and Steven D. Levitt, “Mea- suring Positive Exernalities from Unobservable Victim Precaution: An Empirical Analysis of Lojack,” Quarterly Journal of Economics, 113, No. 1 (February 1998), pp. 43–77.
Eff ic iency in the Provis ion of a Publ ic Good 511
C20.INDD 10:34:15:AM 08/06/2014 PAGE 511Trim Size: 203.2 mm X 254 mm
possible, would there be any advantage in excluding Jane from the services of all 10 subma-
rines? The answer is no, because when 10 submarines are available, Jane’s receiving the ser-
vices of only five does not make any more submarines available for Ted. Consequently, Jane
is harmed, and no one benefits. By definition, this outcome is clearly inefficient.
Recall our definition of a public good as one characterized by nonrival consump-
tion and nonexclusion. When a good has both characteristics, it would be impossible to
exclude anyone from its benefits, even if we wanted to. What about a good with nonrival
benefits where exclusion is possible? The analysis above suggests that it is inefficient to exclude anyone even if we could. Before accepting that as a general rule, let’s examine an important public policy issue dealing with a good where benefits are nonrival but exclu-
sion is possible.
Patents As explained in Chapter 11, a patent grants temporary legal monopoly power to an inven-
tor. A patent gives the inventor the right to make and sell some new product or to use some
new production process for a period of 17 years. But what do patents have to do with the
exotic world of nonrival benefits and nonexclusion? Surely, you say, a vibrating toilet seat
(patent number 3,244,168, granted in 1966) is not a public good.
Admittedly, most of the products granted patents are not themselves public goods. But
what about the knowledge required to make, for example, a vaccine to prevent AIDS? This
knowledge of “how to do it” has nonrival benefits. Once the knowledge exists, any number
of people can use it without interfering with each other’s use. One person’s use of this special
knowledge does not leave less for someone else. Simultaneous consumption of knowledge is
therefore possible, but could people be excluded from its use? Whether exclusion is possible
depends on the type of knowledge involved, but in some cases the use of knowledge can be
prohibited if producing or selling its tangible embodiment is made illegal. For example, if it is
illegal for you to manufacture and sell the AIDS vaccine, you would be effectively excluded
from using the knowledge of how to make it. This is exactly what patents do. They exclude
all but the inventor from making use of the knowledge he or she produced.
Thus, at least some types of knowledge have nonrival benefits, but exclusion is pos-
sible. Now let’s consider efficiency in resource allocation in connection with knowledge.
Although new knowledge is sometimes produced by accident, much of it results from
expenditures on research and development. In making such expenditures, the efficient out-
put of new knowledge requires that resources be devoted to producing it and is accom-
plished by equating the vertically summed marginal benefits with marginal cost, just as
in Figure 20.1. Yet once the knowledge exists, using it efficiently requires that no one be
excluded. Both aspects of efficiency are important.
To see how this discussion relates to patents, suppose that the inventor of an AIDS vac-
cine could not exclude others from copying and selling the product. Would the inventor
devote a million dollars to develop such a vaccine? If this investment were successful, oth-
ers would immediately copy and sell it, driving the price down to a level that just covered
production cost and leaving no way for the inventor to recoup research costs. For this rea-
son inventors would have little financial incentive to produce the knowledge in the first
place, even though that knowledge might be highly beneficial. Too few resources would be
devoted to research and development, because those who bear the costs could not charge
others who use the knowledge for the benefit they receive. In other words, private markets
would not produce the efficient quantity of the public good, new knowledge.
Patents can encourage a greater, more efficient output of new knowledge. Because
inventors receive a temporary monopoly right, they get a return above the cost of produc-
ing new products to compensate for the research costs. The prospect of this gain stimulates
inventors to devote resources to the production of new knowledge. This example illustrates
how private markets can produce a good with nonrival benefits when exclusion is possible.
512 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 512Trim Size: 203.2 mm X 254 mm
Encouraging a greater, more efficient output is the beneficial result of using patents,
but there is a cost. Once the new knowledge is produced, it is inefficiently employed,
since some people are legally excluded from using it. That is, the AIDS-preventing vac-
cine will be monopolistically produced for 17 years, which inefficiently restricts the use,
or consumption, of the vaccine. This cost must be weighed against the gain—namely,
that the vaccine might never have been developed without the incentive created by patent
protection.
Private markets can produce goods with nonrival benefits when exclusion is possible,
as the patent example shows. Private markets, however, do not function with perfect effi-
ciency because of the exclusion of some people who could potentially benefit. Whether it
is possible to devise a better arrangement is uncertain and requires a more detailed case-by-
case evaluation. In any event, the degree of inefficiency in market provision for a nonrival
good will be far less when exclusion is possible than when it is not. The combination of the
nonrival and nonexclusion characteristics creates more severe problems for market provi-
sion, and in this case a more active role for government may be required. No one has deter-
mined, for example, how national defense could be provided by private markets.
20.3 Externalities Sometimes, in the process of producing or consuming certain goods, harmful or beneficial
side effects called externalities are borne by people not directly involved in the market
exchanges. These side effects are called external benefits when the effects are positive and external costs when they are negative. The term externality is used because these effects are felt beyond, or are external to, the parties directly involved in generating the effects.
Immunization against a contagious disease is a good example of a consumption activ-
ity that involves external benefits. For instance, if Barney decides to get an inoculation,
he benefits directly because his chance of contracting the disease is reduced. This benefit
is not the external benefit, since Barney himself receives it. The external benefit is the one
other people receive in the form of a reduced likelihood they will catch the disease because
an inoculated Barney is less likely to transmit it. The central point is that Barney’s deci-
sion about whether to be inoculated is unlikely to be swayed by how his inoculation affects
other people: he is concerned mainly with the effect on his own health. Thus, the benefit his
inoculation creates for others is external to, and doesn’t influence, his decision.
Pollution provides a classic example of an external cost. Driving a car or operating
a factory with a smoking chimney pollutes the atmosphere that others breathe; thus, the
operation of a car or factory imposes costs on people not directly involved in the activity.
Similarly, operating a boom box or motorcycle produces a level of noise that is often irritat-
ing to those nearby, just as the noise level of an airplane may be annoying to people living
near airports. Congestion is also an external cost: when a person drives during rush hour,
the road becomes more congested, not only for this person but for other drivers as well.
Externalities and Efficiency Spam e-mail provides a current example of an external cost. When such e-mail is sent,
costs are borne by individuals who receive the unwanted communication and must figure
out a way to screen out and remove the messages. These costs appear to be significant. For
example, a study by Nucleus Research estimates that over $71 billion in worker productiv-
ity is lost annually in the United States due to spam.
At a formal level, externalities and public goods are very similar, and recognizing the
similarity makes understanding externalities easier. If Barney is inoculated against a con-
tagious disease, there are nonrival benefits: both he and others simultaneously benefit from
external benefits positive side effects of ordinary economic activities
external costs negative side effects of ordinary economic activities
External i t ies 513
C20.INDD 10:34:15:AM 08/06/2014 PAGE 513Trim Size: 203.2 mm X 254 mm
his inoculation. In addition, to exclude other people from the benefits would be very dif-
ficult. When a person produces new knowledge, this action confers an external benefit on
others who can use the knowledge profitably. Pollution is also like a public good (except
here it should perhaps be called a public bad) since there are nonrival costs. A large number
of people are simultaneously harmed if the atmosphere is polluted, and, obviously, to have
the atmosphere in a particular area polluted for some and not for others would be difficult.
The sole distinction between externalities and public goods is that external effects are unin- tended side effects of activities undertaken for other purposes. People don’t pollute because they enjoy breathing polluted air; they simply want to transport themselves conveniently in
a car from one place to another.
Externalities are likely to lead to an inefficient allocation of resources, just as public goods do. Market demands and supplies will reflect the benefits and costs of market par- ticipants only; the benefits and costs that fall on others will not be taken into account in
determining resource allocations. For example, Barney may decide against being inocu-
lated because the improvement in his health is not worth the extra cost. If the benefits of
improved health for others are added to his benefit, the combined benefit might exceed the
cost. In this case Barney’s decision to not be inoculated would represent an inefficient use
of resources.
External Costs A closer look at a case involving external costs will help clarify the issues involved. Sup-
pose that firms in a constant-cost competitive industry produce some type of waste materi-
als as a byproduct of their activities. They dispose of these wastes by dumping them in a
nearby river. From the firms’ point of view, this method of disposal is the least costly. Peo-
ple living downstream, however, suffer, because the river no longer serves recreational pur-
poses. The firms impose external costs on those living downstream. Because these external
costs are not taken into account by the firms, the allocation of resources is inefficient.
Let’s see how this situation appears diagrammatically. In Figure 20.2, the competitive
demand and supply curves, where of course the supply curve is the private marginal cost
curve, are D and MCP. The equilibrium output is Q1 with a price of $20 per unit. Each unit of output generates a specific quantity of wastes, so the greater the industry output,
the greater the amount of water pollution. The harm done by the pollution is shown by
External Costs and Taxes The marginal external cost curve, MCE, shows the external cost associated with production of the good. Vertically adding this curve to the private marginal cost curve, MCP, yields the social marginal cost curve, MCS. Its intersection with the demand curve identifies the efficient output, QE, which is less than the market equilibrium output, Q1.
Figure 20.2 Dollars per unit
QE
MCS
MCP
MCE
D
MC′ T
Q1 Quantity
$27 $25
$20
$7 $5
0
514 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 514Trim Size: 203.2 mm X 254 mm
the marginal external cost, or MCE, curve. It slopes upward, reflecting the assumption that additional amounts of pollution inflict increasing costs on people living downstream. (The
marginal external cost curve results from vertically summing the marginal cost of each per-
son harmed, since the harmful effects are nonrival over many persons.) At the market out-
put of Q1, the marginal external cost is $7, implying that people downstream would be $7 better off with one unit less of the product and the waste associated with it.
With external costs, the competitive output is too large from a social perspective. Firms
expand output to where the price consumers pay just covers their private production costs (as reflected by the private marginal cost curve, MCP), but the resulting price does not cover all costs, since pollution is also a cost of producing the product. At Q1, firms incur a cost of $20 per unit, which is just covered by the price paid by consumers, but there is also a
$7 per-unit pollution cost borne by people downstream. At the competitive output, Q1, the product is not worth what it costs to produce. The social, or combined, marginal cost of
production is $27, while the marginal benefit to consumers is only $20. The social marginal
cost is shown by the curve MCS, obtained by vertically summing the MCE curve and the private marginal cost curve, MCP. It identifies all the costs associated with producing the
product, not just the costs borne by producers. Efficiency requires that output be expanded
to the point where the marginal benefit to consumers equals the social marginal cost of pro-
duction. This point is shown by the intersection of D and MCS at an output of QE. Competitive market pressures, however, lead to an output of Q1, larger than the effi-
cient output. The government can do several things to improve the situation. One approach
would be to levy a tax on the product to induce firms to produce at the efficient level. A
tax of $5 per unit would shift the private marginal cost curve up by $5 to MC′, and firms would curtail production to QE, with consumers paying a price of $25. The result is the efficient output level, where the marginal benefit to consumers equals the social marginal
cost of production. Note that pollution is not completely eliminated; it is simply reduced to
the point where a further reduction in production and pollution would cost more than it is
worth. In general, external costs should not be totally eliminated even though those who are
harmed might like to see them reduced to zero. Instead, the gain from reduced pollution to
people downstream must be weighed against the cost to consumers of reduced output.
In this example we assumed that each unit of output is invariably associated with a
certain amount of pollution. In the more general case the amount of pollution per unit of
output is variable. Automobiles, for example, can produce various amounts of emissions.
When this situation is the relevant case, as it usually is, the tax should be levied on pollu-
tion itself, not on the product. Then, as discussed in Chapter 8, firms have an incentive to
curtail pollution—the external cost—in the least costly manner.
APPLICATION 20.3
While an individual motorist’s decision to drive at rush hour may cost only a few extra minutes of commuting time, the external congestion costs imposed by such motorists as a group can add up to millions of dollars per year in a major
Traffic Externalities: Their Causes and Some Potential Cures3
urban area. For example, an average driver in the Wash- ington, D.C. area spends 67 hours a year stuck in traffic, at a cost of over $1,400 per person in wasted time and gas. Across 498 urban areas in the United States, the average driver spends 38 hours stuck in traffic per year, at a cost of $121 billion in gasoline and reduced productivity.
Beyond congestion costs, rush-hour commuters pay for only a fraction of what they impose on the community at large in terms of road construction and pollution costs. For example, significant road construction subsidies exist to better accommodate the needs of rush-hour drivers.
3This application is based on “In Singapore Driving a Car Is Easy but Owning a Car Isn’t,” Los Angeles Times, August 17, 1991, pp. A1 and A14; Texas Transportation Institute, What Does Congestion Cost Us? College Station, TX: Texas A&M, 2007; and “The 2012 Urban Mobility Report,” Texas Transportation Institute, The Texas A&M University System, http://mobility.tamu.edu.
External i t ies 515
C20.INDD 10:34:15:AM 08/06/2014 PAGE 515Trim Size: 203.2 mm X 254 mm
APPLICATION 20.4
The British Petroleum (BP) oil spill in 2010 is the largest accidental marine oil spill in the history of the petroleum industry. Between the explosion of the Deepwater Hori- zon drilling platform on April 20, 2010, and the time when the blown-out well was finally capped 3 months later, 206 million gallons of oil poured into the Gulf of Mexico. The U.S. Travel Association estimated that the economic impact on tourism across the Gulf Coast over the 3-year period after the spill exceeded $23 billion. The insurance firm Willis Group Holdings, furthermore, calculated that the economic losses associated with the lost well, redrill- ing, third-party liability, seepage, and pollution costs to the Gulf fishing and tourism industries totaled over $30 billion.
One of the legal issues coming to the fore in the wake of the BP spill is the statutory cap of $75 million on tort damages for any offshore incident. University of Chicago law professor Richard Epstein argues that “$75 million
Liability Caps and the British Petroleum Gulf Oil Disaster4
is chickenfeed. Fortunately, the law removes that cap if the incident was caused by ‘the gross negligence or will- ful misconduct’ of any party, or its failure to comply with any ‘applicable Federal safety, construction or operating regulation.’”
To ensure that oil drilling companies appropriately take into account the full costs of their actions and thereby are appropriately incented to prevent future oil spills, Epstein argues for removal of the liability cap and requiring operators “to purchase insurance—amount- ing to tens of billions if necessary—when they operate in dangerous waters or terrains.” Epstein notes that the proposed tougher liability system “sorts out the wheat from the chaff—so that in this case companies with weak safety profiles don’t get within a mile of an oil derrick. Solid insurance underwriting is likely to do a better job in pricing risk than any program of direct government over- sight. Only strong players, highly incentivized and fully bonded, need apply for a permit to operate.” In addi- tion to removing the liability cap for offshore oil drilling, Epstein also asserts that the Price-Anderson Act’s $375 million cap on damages associated with nuclear power accidents deserves reconsideration.
4This application is based on Richard A. Epstein, “BP Doesn’t Deserve a Liability Cap,” Wall Street Journal, June 16, 2010, p. A21.
According to one study, the subsidy totals $500 annually per rush-hour commuter in the Los Angeles area. Overall, California state gasoline taxes amount to only one-sixtieth of the estimated cost that rush-hour drivers impose on the community at large in terms of congestion, road construc- tion, and pollution costs.
Some examples from overseas suggest mechanisms by which motorists could be held more fully accountable for the burden they impose on a community. Singapore’s efforts to control traffic are legendary for promoting free- flowing roads and cleaner air and also for their draconian methods of enforcement. Whereas the neighboring capitals of Southeast Asia such as Bangkok, Thailand, and Jakarta, Indonesia, are notorious for their smog and day-long grid- lock, Singapore’s policies keep the skies clean and rush- hour traffic delays to a minimum. Among the policies is a requirement that cars entering the city center pay an Elec- tronic Road Price (ERP) that varies in amount depending on the time of day. For example, the ERP is $2.50 for 8–9
; $2 for 9–9:30 and 6–6:30 ; $1.50 for 5:30–6 ; $1 for 9:30–10 , 12:30–5:30 , and 6:30–7 ;
$0.50 for 7:30–8 ; and free at all other times.
The government of Singapore also adds more than 200 percent in duties to every auto’s purchase price. Moreover, every car owner must pay a sizable annual road use tax (analogous to a registration fee in the United States) to operate a vehicle. The road tax on a small Toy- ota is $800. When even the sizable costs of purchasing an automobile failed to curb the growth in car ownership, Singapore began limiting the absolute number of cars that can be sold in any year to 50,000. The quota is designed to keep new purchases to 4 percent of the total cars on the road.
The top speed limit on the island of Singapore is set at 45 miles per hour, to promote safe driving and thereby minimize the chance of accident-related traffic jams. Taxis are required to have a chime built into their dashboards that sounds continuously and annoyingly when the speed limit is violated. Trucks are mandated to have a yellow light on their roofs that is activated when the speed limit is exceeded. Police, as well as hidden cameras hooked up to remotely operated radars, are employed to catch traffic vio- lators. If speeding is caught on camera, a ticket is sent to the violator by mail within days of the infraction.
516 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 516Trim Size: 203.2 mm X 254 mm
External Benefits External benefits can be analyzed in a similar fashion to external costs. Let’s suppose that
the consumption of some product generates external benefits—that is, people other than
the direct consumers of the product benefit from its consumption. Some economists have
argued that education may be such a product if a better-educated citizen not only makes
himself or herself better off in the process but also improves the society of which he or
she is a part. In Figure 20.3, the competitive supply and demand curves where demand
identifies the private marginal benefit are S and MBP. The private marginal benefit curve, MBP, reflects the marginal benefits of the good only to the consumers of the product, and its intersection with the supply curve determines the market equilibrium with an output
of Q1 and a price of $10. The marginal external benefit curve, MBE, shows the external benefits per unit of consumption. This curve is derived by vertically summing the demands
of people other than the immediate consumers of the product. Vertical sum mation is used
because people other than direct consumers simultaneously receive benefits—that is, the
benefits are nonrival.
When external benefits exist, the competitive output is inefficient. At Q1, the marginal benefit to consumers of another unit of the product is $10, as given by the height of the
MBP curve. If another unit is consumed, people other than the direct consumers also receive a marginal benefit valued at $3, as shown by the height of MBE at Q1. Thus, the com- bined marginal benefit for all those affected by consumption of another unit is $13, and
this amount exceeds the $10 marginal cost of producing an additional unit. The combined,
or social, marginal benefit is shown by MBS, which is derived by vertically adding MBP and MBE (again because the benefits are nonrival).5 The competitive output is too small
5At an output in excess of the level at which the marginal external benefit becomes zero, the MBS and MBP curves coincide. When consumption is so great that additional consumption yields zero marginal benefits to other people, the only ones who receive any benefits from further increasing consumption are the direct consumers themselves, and their marginal benefits are shown by the MBP curve.
External Benefits and Subsidies The marginal external benefit curve, MBE, shows the external benefits generated by consumption of the good. Vertically adding this curve to the private marginal benefit curve, MBP, yields the social marginal benefit curve, MBS. Its intersection with the supply curve identifies the efficient output, QE, which is greater than the market equilibrium output, Q1.
Figure 20.3 Dollars per unit
MBS
Q1 QE
MBE
Quantity
D = MBP
S
S′
$13
$10
$8
$3 $2
0
External i t ies 517
C20.INDD 10:34:15:AM 08/06/2014 PAGE 517Trim Size: 203.2 mm X 254 mm
because the marginal benefit of additional units of output exceeds the marginal cost of pro-
ducing them. Yet there is no tendency for competitive pressures to produce a larger output,
because the additional benefit to the direct consumers is less than the $10 price they must
pay per unit. Thus, it is not in the consumers’ interest to purchase more than Q1 units; only when the combined benefits to consumers and other people are considered is it apparent why greater production is worthwhile.
Figure 20.3 illustrates the general tendency of an activity to be underproduced when
external benefits are involved and when production is determined in competitive mar-
kets. The competitive output is Q1, whereas the efficient output is QE, where MBS inter- sects S. To achieve the efficient output, the government could step in with a policy designed to increase output beyond the market-determined level. In this case, a subsidy
would be appropriate. If the government pays firms $2 for every unit of output sold, the
supply curve confronting consumers would shift to S′. Although the marginal cost of production is still $10, the government in effect bears $2 of the cost through the sub-
sidy, so consumers pay only $8. At the lower price, consumers purchase QE units, the efficient output.
APPLICATION 20.5
The use of cell phones by drivers can impose some exter- nal costs on society.6 Among these costs is the heightened risk of accidents, some of which involve serious injury or death. Indeed, it is estimated that several hundred people die annually in the United States due to motor vehicle accidents precipitated by cell phone use. Although this is still a small number relative to the 41,000 people who die annually in the United States in all car accidents, a grow- ing number of states and municipalities have begun to consider either banning the use of cell phones by drivers altogether or restricting usage in certain ways such as requir- ing a hands-free device to be used in conjunction with a cell phone.
While there are negative externalities, there are also some positive externalities associated with cell phone usage by drivers. For example, because conversation is known to help individuals from becoming drowsy, the use of cell phones may thus keep drivers from falling asleep and thereby prevent accidents from happening. Being able to use a cell phone to call and notify relevant parties when one is running late for a meeting can also keep drivers from feeling they need to speed and thereby promotes greater
Should Cell Phone Use While Driving Be Banned?
traffic safety. Moreover, motorists either involved in or wit- nessing an accident or a breakdown can summon help to the scene more quickly through the use of a cell phone. Similarly, motorists with cell phones have the capability to report drunk or otherwise unsafe drivers to authorities more rapidly.
Beyond the positive externalities, the ability to use a cell phone while driving provides significant benefits to consumers. There are now more than 160 million cellu- lar phone subscribers in the United States, and the esti- mated total annual consumer surplus associated with being able to make phone calls while driving is well over $50 billion.
Because of the sizable benefits to consumers from being able to use cell phones while driving, as well as the external benefits associated with cell phone usage, leading studies indicate that banning cellular phone usage by motorists is a bad idea. The estimated deadweight loss from imposing a nationwide ban is over $40 billion per year.
Moreover, the analysis also suggests that even a less draconian measure such as mandating a hands-free device fails to pass the cost–benefit test. Evidence from accident data in Japan and North Carolina indicates that only 15 per- cent of the collisions caused by cellular phones might have been avoided had drivers been using a hands-free device. The costs from instituting a hands-free mandate in terms of inconvenience and equipment purchase generally appear to be larger than the benefits from the projected reduction in cell-phone-caused collisions.
6This application is based on Robert W. Hahn, Paul C. Tetlock, and Jason Burnett, “Should You Be Allowed to Use Your Cellular Phone While Driving?” Regulation, 23, No. 3 (2000), pp. 46–55; “Right Now,” Harvard Magazine, November/December 2000, pp. 18 and 20; and CTIA: The Wireless Association, “Safe Driving Overview” www.wow-com.com.
518 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 518Trim Size: 203.2 mm X 254 mm
20.4 Externalities and Property Rights External effects may appear intrinsically different from normal costs and benefits, but that
appearance is partly misleading. When a firm uses your labor services, it imposes a cost on
you, since you sacrifice the option to use your time in other ways. When a firm pollutes the
river passing by your home, it imposes a cost on you, since you sacrifice the option to use
the river for recreation. These costs are not fundamentally different: they both imply that
you are unable to use economic resources in other, valuable, ways. Why, then, do we call
pollution, but not the firm’s employment of your labor services, an external cost?
One glaring difference in these two cases is that the firm must pay you for your labor
services, but you are not compensated when the river is polluted. Since the firm must pay
you at least enough to persuade you to give up alternative uses of your time, it will have
an incentive to take this cost into account in deciding whether to employ you; that is, when
the firm bears a direct cost associated with the use of a resource, that cost enters into its
production decision. But if the firm can use the river in a way that harms you without com-
pensating you for the damage, it has no reason to consider this cost in making its output
decision—the firm treats the river as a zero-priced input.
Why must the firm pay to use your labor services but not to use the river? Fundamen-
tally, the answer to this question involves property rights to the use of economic resources.
You have well-defined and legally enforceable rights to your own labor services, meaning
that no one can use them without securing your permission, which is normally acquired by
paying you. There are, however, no such clearly defined property rights to the water that
flows past your home. In fact, ownership of the river and who has the right to decide how it
will be used are uncertain. Consequently, the firm can use it as a convenient garbage dump.
If you had property rights to pure water flowing past your home, the firm would have to
buy your permission to dump waste in the river. The firm might still pollute, but would do
so only if the gain from polluting was greater than the compensation required to be paid.
The situation would then be just analogous to the case of labor services. Pollution would
no longer be a cost external to the firm’s calculations; the cost would be taken into account,
and the allocation of resources would be efficient.
Reasoning along these lines suggests that externalities are intimately connected with the
way property rights are defined. Indeed, in most cases dealing with externalities, we can
usually trace the source of the problem to an absence or inappropriate assignment of prop-
erty rights. Accordingly, the government may not need to use taxes, subsidies or regula-
tion at all; it may only have to define and enforce property rights, and the resulting market
exchanges will produce an efficient resource allocation.
As an example, imagine a beautiful beach on the California coast and suppose that no
one owns it, just as no one owns the river. How will this scarce economic resource be used?
It is not beyond fancy to conceive of masses of people crowding the beach trying to enjoy
the sand, sun, and surf. Radios could blare, dune buggies roar up and down the beach, dirt
bikes spray sand, litter lie strewn across the beach, and surfboards crash into swimmers.
Externalities would be rampant.
Most would agree that this is not an efficient use of scarce oceanfront property, but because
no one owns it, no one has an incentive to see the beach used in the most valuable way. The
situation is far different when someone has property rights to the beach. In that case, use of the
property will be guided by whoever pays the most for its use—that is, by who benefits most.
The owner may still use the property as a beach, but now it will be operated differently. Admis-
sion might be charged, which will diminish the overcrowding that reduces the attractiveness of
the nonowned beach. The owner might enforce rules regarding radios, litter, surfboards, and so
on, further enhancing the benefits to consumers. In short, the external cost is no longer exter-
nal when someone owns the beach. The owner has an incentive to see that the beach yields as
much benefit to consumers as possible, since they will then pay more for its use.
External i t ies and Property Rights 519
C20.INDD 10:34:15:AM 08/06/2014 PAGE 519Trim Size: 203.2 mm X 254 mm
The beach example is hypothetical, but it helps explain why some highways, parks, and
beaches are overcrowded and inefficiently utilized. “Publicly owned” property is, in effect,
sometimes owned by no one, in the sense that no one has the incentive and the right to see
that it is used in the most valuable way.
The Coase Theorem These examples suggest that the assignment of property rights can make an important con-
tribution to resolving issues involving externalities—but who is to have exactly what right
to use the resource in question? Should a factory have the right to discharge smoke into the
atmosphere, or should a nearby resident have the right to pure air? A case can certainly be
made that both of these parties have a reasonable claim to use the atmosphere for their own
purposes, yet giving the resident the right to clean air denies the factory the right to use a
smokestack, and vice versa.
Ronald Coase addressed this issue in one of the most widely read papers in the history
of economics.7 Coase developed his analysis by considering a rancher and a farmer with
adjoining properties. The rancher’s cattle occasionally stray onto the farmer’s property and
destroy some of the crops—an external cost associated with cattle raising if this cost is
not properly taken into account. Now suppose the farmer has the right to grow crops in a
trample-free environment. The rancher would then be legally liable for the damage caused
by the straying cattle. Since the rancher will have to compensate the farmer for the crop
damage, the cost of straying cattle will become a direct cost to the rancher and will be taken
into account in the rancher’s production decision. An efficient outcome will result, prob-
ably one involving fewer straying cattle.
This conclusion is familiar, but Coase went further and argued that even if the rancher
were not liable for damages, an efficient outcome would still result! This situation corre-
sponds to giving the rancher the right to allow his or her cattle to stray. Coase explained
that the farmer then has an incentive to offer to pay the rancher to reduce the number of
straying cattle because a reduction in crop damage increases the farmer’s profit. The harm
done by straying cattle necessarily implies that the farmer will be willing to pay something
to avoid that harm. An agreement would therefore be struck that would reduce cattle stray-
ing to the efficient level.
As far as efficiency is concerned, the Coase theorem states that whether the farmer or the rancher is initially assigned the property rights doesn’t matter. As long as the rights are clearly defined and enforced, bargaining between the parties can ensure an efficient out- come. The distributional effects, though, depend on the exact definition of property rights. If the rancher is liable, the rancher will compensate the farmer; alternatively, if the rancher
is not liable, the farmer will pay the rancher to reduce the cattle straying. In both cases cat-
tle straying and crop damage are reduced to the efficient level, but different people bear the
cost and secure the benefit.
Simply assigning property rights, however, will not resolve all externality problems. In the
case discussed earlier, a firm pollutes a river and many people living downstream are harmed. If downstream residents are given the right to have clean river water, would bargaining between
parties lead to an efficient level of water pollution? Most likely not. This is because thousands
of people are affected by the pollution, and a firm would have to negotiate an agreement with
all of them simultaneously to be allowed to pollute. Whenever the effects are nonrival over a large group and exclusion is not feasible, the free-rider problem hinders the process of achiev- ing agreement among all concerned. The negotiation process likely would be so costly and time consuming as to become a practical impossibility. Consequently, with such an assignment
of property rights, there would be no pollution in the river—but that may be as inefficient as (or
perhaps even more inefficient than) allowing the firm to pollute freely.
Coase theorem the idea that as long as property rights are clearly defined and enforced, bargaining between two parties can ensure an efficient outcome
7Ronald H. Coase, “The Problem of Social Cost,” Journal of Law and Economics, 3, No. 2 (October 1960), pp. 1–45.
520 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 520Trim Size: 203.2 mm X 254 mm
Our earlier conclusion that markets would be inefficient is correct, therefore, in the case
where the external effects simultaneously fall on many people. Assigning property rights
can solve externality problems when there are small numbers of parties involved but not
as readily when there are large numbers, because of the free-rider problem. Many issues of
great importance, such as defense, pollution, and police protection, are large-group exter-
nalities or public goods, and private markets are thus unlikely to function effectively in
these areas without some form of government intervention.
APPLICATION 20.6
More than 50 million Americans have registered for the Federal Trade Commission’s do-not-call registry list that allows households to block commercial telemarketing calls but not calls from nonprofit organizations such as religious groups, charities or political parties. The do-not-call list has grown as a result of the negative externalities imposed by telemarketers, who disrupt the privacy of families without having to pay for it. (Telemarketers only have to pay for the labor and communication charges associated with placing calls but not for the disruption costs that they impose on families receiving their unwanted calls.)
Rather than banning telemarketing entirely, Professor Ian Ayres of the Yale Law School has proposed an innovative solution for internalizing the externalities imposed by com- mercial telemarketers.8 Ayres argues that the property rights to telemarketing should be assigned to households and not to commercial telemarketing firms (as was historically the case) and that bargaining between prospective telemarket- ers and households could then ensure a much more efficient outcome than an outright ban on such calls:
Local phone companies could set up a kind of reverse “900” number system where customers would get paid for each minute they listen to a sales pitch—in fact, companies have already offered long-distance
Making Telemarketers Pay
service based on this model (customers can earn free minutes by, for example, listening to phone sales pitches). By encouraging compensated calling, Congress could save the jobs of tens of thousands of telemarketers. It would even give telemarketers the opportunity to make new types of calls . . . so long as the sender meets the household’s prerequisites, the intermediary should be authorized to connect, say, prerecorded calls or faxes. Finally, allowing families to choose their telemarketers could also serve the public interest. The current hang-up mentality has wreaked havoc on polling organizations, which now average response rates as low as 15 . . . or 20 percent. Under this system, perhaps the response ratio will improve, and we will have more reliable information about public opinion.
By making commercial telemarketers pay for the full costs of their actions, the solution proposed by Professor Ayres promises to result in a more efficient outcome than a ban on telemarketers. The current policy approach of ban- ning telemarketing calls leaves no opportunity to realize the benefits of mutually beneficial exchange between callers who are willing to pay households at least as much as they need to be compensated to have their privacy interrupted. Such benefits can still be realized in a politically more pal- atable manner by giving households the right to accept telemarketing calls and then allowing bargaining between households and telemarketing firms.
8This application is based on Ian Ayres, “Dialing for Dollars,” New York Times, September 30, 2003, p. A29.
20.5 Controlling Pollution, Revisited Perfect competition ensures efficiency in industry output if demand and supply curve
heights reflect the full marginal benefits and costs associated with a particular product.
There are cases, however, in which assuming that demand and supply curve heights reflect
a product’s full marginal benefits and costs is not valid. Oil refineries, for example, may
not be fully accountable for the costs associated with their productive activities. That is, the
Control l ing Pol lut ion, Revis i ted 521
C20.INDD 10:34:15:AM 08/06/2014 PAGE 521Trim Size: 203.2 mm X 254 mm
refineries may not have to pay for the air pollution caused by their operations and imposed
on surrounding communities.
As we saw earlier in this chapter, when demand and supply curve heights do not reflect
a product’s full marginal benefits and costs, the industry output attained by perfect compe-
tition is generally not efficient. Still, even if an industry does not attain efficiency in out-
put because certain benefits and costs are external to the decisionmaking of consumers and
firms, perfect competition results in efficiency in production. The industry output that is
produced, in other words, is produced at the lowest possible cost even though it may not be
the efficient output.
To see how competition ensures production efficiency even if output efficiency is not
attained, reconsider the case of two oil refineries first introduced in Chapter 8. Refiner-
ies A and B are located in the Los Angeles basin. Suppose they impose air pollution
costs that they do not have to pay for on their surrounding communities. As a result, the
firms’ individual output decisions are based on only part of the costs associated with their
productive activities and (as we saw in Section 20.3) the realized industry output is not
efficient.
Policymakers may seek to ensure that the refineries account for the costs of their air
pollution by levying a tax per unit of air pollution emitted. The tax creates an incentive for
each refinery to curtail pollution, because the refinery saves the amount of the tax per unit
of pollution not emitted. If reducing pollution by one unit costs the refinery less than the
tax, the refinery has an incentive to engage in pollution abatement.
The appropriate tax policymakers should levy per pollution unit to ensure output effi-
ciency may, of course, be difficult to determine. Regardless of the amount of the tax, how-
ever, the total amount of pollution abatement across all refineries in response to the tax will
be produced at the lowest possible cost (and will achieve efficiency in production) if perfect
competition prevails.
Say that a tax of $3,000 per unit of air pollution emitted is imposed on refinery A.
Figure 20.4 shows how such a tax would affect the level of air pollution. As in Chapter 8,
pollution is measured from right to left, and the refineries’ marginal cost curves for pollu-
tion abatement are shown as MCA and MCB. (Ignore refinery B for the moment.) A tax of $3,000 per unit of pollution can be shown as a horizontal line at $3,000. A refinery’s total
tax liability is $3,000 times the number of units of air pollution. If refinery A continues to
pollute at its initial level, 0P1 (when there was no pollution tax), it will have to pay a tax equal to 0T ($3,000) times 0P1, which is the area 0TT1P1.
The tax gives refinery A an incentive to curtail pollution, because each abatement unit
saves $3,000 in taxes. Looked at from left to right, the horizontal line T1T is like a demand curve for pollution abatement: It shows a gain in net revenue of $3,000 per unit of abate-
ment produced. Thus, refinery A faces a horizontal demand curve for abatement, and its
marginal cost curve indicates the cost of abatement. To maximize profit, the refinery has
a strong incentive to curtail pollution. Specifically, if the refinery can eliminate a unit
of pollution for less than $3,000, doing so adds more to net revenue (reducing taxes by
$3,000) than it adds to cost. In the interest of profit, refinery A will curtail pollution to
0X1 (producing P1X1 in abatement), where the marginal cost of abatement exactly equals the $3,000 per-unit tax. Cutting back further is not worthwhile because the cost of more
abatement exceeds the tax saving. With pollution of 0X1, refinery A still pays total taxes of 0TNX1, but this sum is significantly less than the tax cost associated with the initial pol- lution level 0P1.9
9This analysis does not show how the pollution tax affects the refinery in its product market. Of course, the tax increases production cost and shifts the cost curves upward. To be accurate here, we should explicitly assume that the tax is not so large that it becomes unprofitable for the refinery to stay in business.
522 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 522Trim Size: 203.2 mm X 254 mm
Now let’s turn to refinery B. The same analysis is relevant for refinery B, which has
an incentive to cut back pollution to 0X2, where its marginal cost of abatement equals the $3,000 per-unit tax. Note what this means: refinery A and refinery B are both operating at
an abatement level where marginal cost equals $3,000. Their marginal costs are the same,
which implies that the total amount of abatement of 350 units (P1X1 = 300 − 150 = 150 units by refinery A plus P2X2 = 250 − 50 = 200 units by refinery B) is achieved in the least costly way. And this outcome happens without the government’s knowing either refinery’s
marginal cost curve. By applying the same tax to both refineries, the government gives each
refinery the same incentive to curtail pollution. The result is efficient coordination of their
independent production decisions.
To better understand why the total cost of achieving 350 abatement units is minimized
by relying on incentives and competitive market forces, consider a reallocation of abate-
ment units among the two refineries. For example, consider what happens if refinery A
produces 50 additional units and refinery B produces 50 fewer units (so combined output
remains unchanged). Refinery A’s production cost goes up by the sum of the marginal cost
of each unit from 151 to 200, shown by area X1NRX. Refinery B’s production cost falls by the sum of the marginal cost of each unit it ceases to produce, shown by area XULX2. Because refinery A’s cost increase exceeds refinery B’s cost saving (as can be seen in
Figure 20.4, by recalling that the widths of the shaded areas are equal), the total cost of
producing 350 abatement units is higher under such a reallocation. Similar reasoning shows
Figure 20.4
Dollars per unit
Dollars per unit
MCA
R
N
U
LS
MCB
$3,000 = T1 T = $3,000
TA
TB
P1 (300)
P2 (250)
X1 (150)
0X
(100)
X2 (50)
Pollution abatement Pollution
A Tax on Pollution A per-unit tax on pollution makes refineries competitive producers of pollution abatement. With a tax of $3,000 per unit of pollution, refinery A produces P1X1 abatement units (0X1 pollution units), and refinery B produces P2X2 abatement units (0X2 pollution units).
Control l ing Pol lut ion, Revis i ted 523
C20.INDD 10:34:15:AM 08/06/2014 PAGE 523Trim Size: 203.2 mm X 254 mm
that any other way of having the refineries produce a total of 350 abatement units results in
a higher total production cost than that achieved by the competitive refineries individually
choosing their profit-maximizing outputs.
The foregoing discussion indicates why many economists favor taxation as a pollution
control strategy. Notwithstanding the difficulty of determining the “appropriate” level of
the tax that will result in efficiency in output (a tax that equals the marginal external costs
associated with pollution), a pollution tax effectively creates market incentives for firms to
reduce pollution in the least costly manner and thereby ensures efficiency in production.
Moreover, the size of the tax can be changed to regulate the amount of pollution: a larger
tax per unit will reduce pollution further.
In the United States, most environmental policies rely on a command-and-control
approach: regulations and quantity limitations rather than taxes. Many economists have
been critical of these non-market-oriented policies, in part because they believe the taxa-
tion approach can achieve the same results at lower cost. And the taxation approach is more
than a theorist’s pipe dream. For example, Germany has successfully used pollution taxes
to regulate waste discharge into the Ruhr River for over 50 years.
The Market for Los Angeles Smog An alternative market-oriented approach to controlling pollution involves the setting of an
overall industry pollution level, with each firm receiving permits to emit a certain amount
of pollution units and allowing firms to exchange their permits. This Coasean approach was
adopted by policymakers in 1994 in an attempt to control smog in the Los Angeles basin.
Tradable permits to pollute were issued to each of the 390 companies producing four or
more tons of emissions annually. The overall emission level allowed through the permits
was set below the existing level and further reduced each year by 5 to 8 percent so that by
the year 2004, nitrogen oxides were cut by 75 percent and sulfur oxides by 60 percent. Pol-
lution permits were allocated across firms more or less according to their formerly prevail-
ing emissions.
The L.A. smog market allows an overall emission target to be achieved in the least
costly manner. To see why, suppose that in our simple example the goal is to reduce air
pollution by 350 units (the same reduction achieved by a tax of $3,000 per emission unit)
and that the two refineries are issued tradable permits to emit 100 pollution units each (200
total units across the two refineries). As shown in Figure 20.4, with no pollution control
whatsoever, 550 pollution units would be produced—300 by refinery A (0P1) and 250 by refinery B (0P2).
Under the tradable-permit scheme, the potential exists for mutual gains from trade
between the two refineries. This is because at its allotted 100 emission units, refinery A’s
marginal abatement cost (TA) exceeds refinery B’s abatement cost (TB). Since, at the mar- gin, refinery A is willing to pay more to increase its emission (cutting its abatement costs
TA) than refinery B needs to be paid to decrease its emission (incurring abatement costs TB), there is room for the two refineries to exchange pollution permits for a price somewhere
between TA and TB, making both sides better off. Of course, the exact price at which the permits will be exchanged will depend on, among
other things, the bargaining abilities of the two refineries. Moreover, the bounds around
the exchange price will narrow as more permits are sold by refinery B to refinery A. The
bounds around the exchange price will narrow because as refinery A buys more permits,
the maximum amount it is willing to pay per permit declines from TA; the cost of abatement to refinery A decreases as it pollutes more and moves to the left along its marginal cost of abatement curve. In addition, as refinery B sells additional permits, the minimum amount it needs to be paid for each rises from TB; the cost of abatement to refinery B rises as it pol- lutes less and moves right along its marginal cost curve.
524 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 524Trim Size: 203.2 mm X 254 mm
How many permits will be exchanged? The total will be 50 in the case depicted in
Figure 20.4. When refinery B sells this many permits to refinery A, their marginal costs of
abatement are identical. Refinery B, because it has the lower marginal cost of abatement
curve, ends up emitting 0X2 (50) pollution units, abating P2X2 (250 − 50 = 200) units, and having a marginal cost of abatement of T1. Refinery A ends up emitting 0X1 (150) pollution units, abating P1X1 (300 − 150 = 150) units, and having a marginal cost of abatement of T1. The price for the fiftieth exchanged pollution permit equals $3,000 since it is perfectly con-
stricted by the two refineries’ marginal costs of abatement (i.e., T1). By generating the “proper” marginal permit price (“proper” in terms of achieving an
overall level of 200 pollution units and 350 units of pollution abatement) and confront-
ing both refineries with that price, the L.A. smog market ensures attainment of the overall
emission target in the least costly way. If, instead of allowing permit trading, regulators
limited each refinery to 100 pollution units, the total abatement cost would be higher. That
efficiency in production would not be served through such a command-and-control device
is evidenced by the fact that in Figure 20.4 the cost to refinery A of reducing its pollution
from 150 to 100 units (area X1NRX) exceeds the cost to refinery B of reducing its pollution from 100 to 50 units (area XULX2). The same overall emission target of 200 units can thus be achieved at lower total cost if refinery A is allowed to emit 150 units and refinery B is
permitted 50 units.
In sum, market-based pollution control mechanisms such as tradable emission permits
or per-emission-unit taxes promote efficiency in production. Although such mechanisms do
not necessarily guarantee the attainment of output efficiency, they do ensure that any abate-
ment amount produced by an industry is produced at lowest possible cost.
The cost savings associated with market-based pollution control mechanisms can be
substantial in the real world. For example, economists have estimated that the L.A. smog
market saves $1,000 per year in abatement costs per resident household relative to a pol-
icy of mandating proportional, across-the-board reductions in emissions and not allow-
ing pollution permit trading.10 Significant abatement cost savings could also be realized if
emission trading programs were more broadly implemented on a national as well as inter-
national basis. For example, according to the consulting firm Charles River Associates, the
cost for reducing greenhouse-gas emissions according to the Kyoto Protocol (a greenhouse-
gas reduction treaty signed in 1997 as part of the United Nations Framework Convention
on Climate Change) is estimated to be $280 per ton if no trading in emissions permits is
allowed.11 By contrast, the estimated cost drops to $60 per ton if a completely open market
in emissions permits is authorized.
Among some of the other market-oriented pollution control mechanisms with which
policymakers have experimented over the past four decades are “bubbles,” through
which a firm can treat an existing plant with multiple emission sources as if it were
a single source—a bubble allows a firm to adjust its various emission sources to meet
an overall emission target for the plant in the least costly manner; banking of pollu-
tion abatement credits, whereby a firm can hold onto emission reduction credits for
future use or sale; and offsetting—a major new emission source in regions failing to
meet national air quality standards can compensate for its added pollution with emis-
sion reductions of an equal or greater amount achieved through internal or external
trades. All of these market-based approaches promise significant efficiencies in produc-
tion over the more commonly employed command-and-control mechanisms for dealing
with pollution.
10David Harrison, Jr., and Albert L. Nichols, Market-Based Approaches to Reduce the Cost of Clean Air in California’s South Coast Basin (Cambridge, MA: National Economic Research Associates, November 1990). 11“Letting the Free Market Clear the Air,” Businessweek, November 6, 2000, pp. 200–204.
Review Quest ions and Problems 525
C20.INDD 10:34:15:AM 08/06/2014 PAGE 525Trim Size: 203.2 mm X 254 mm
SUMMARY
Public goods are characterized by nonrival consump-
tion and nonexclusion. When a good has these two char-
acteristics, the free-rider problem arises and makes it
difficult to ensure that the efficient quantity will be pro-
duced through voluntary arrangements.
An efficient output of a public good is that at which
the vertically summed demand curves of individuals
intersect the marginal cost, or supply, curve.
Externalities are the harmful or beneficial side effects
of market activities that are borne or realized by people
not directly involved in the market exchanges. They rep-
resent costs or benefits that are not incorporated in the
private supply and demand curves that guide economic
activity. Once again, the result is an inefficient resource
allocation.
In some cases it is only necessary to define prop-
erty rights appropriately for externalities to be taken
into account. In other cases, principally those involving
large numbers of people, this solution will likely not
work and other types of government policies should be
considered.
With regard to accounting for the externalities associ-
ated with air pollution, government policies that can pro-
mote efficiency in production include per-emission-unit
taxes and tradable permits to emit a certain amount of
pollution.
REVIEW QUESTIONS AND PROBLEMS
Questions and problems marked with an asterisk have solu- tions given in Answers to Selected Problems at the back of the book (pages 528–535).
20.1 What two characteristics define a public good? Which of the following are public goods: parks, police services, welfare
payments to the poor, production of energy, space exploration?
20.2 Why will private markets produce an inefficient output of a public good? Explain how the efficient level of a public good
is determined.
20.3 What is meant by the free-rider problem? How does it relate to the provision of public goods? How can it be over-
come?
*20.4 Suppose there are three consumers—two “hawks” and one “dove.” The dove receives negative benefits from (i.e.,
is harmed by) national defense, but the hawks value defense.
Show graphically how an efficient output of defense is deter-
mined in this case.
20.5 From a public good perspective, critique the use of pat- ents.
20.6 “External costs are bad, and government intervention to reduce them is justified. External benefits, however, are good,
and there is no reason for government intervention in this
case.” Evaluate these statements.
20.7 Education is sometimes cited as a source of external ben- efits. In what way, if at all, does your receiving a college edu-
cation benefit other people?
*20.8 Suppose that property rights change so that students no longer have exclusive rights to the use of lecture notes they
take in classes. (All notes are collected after class, and any-
one can borrow notes for 24 hours on a first-come, first-served
basis.) How would this policy affect note taking, class atten-
dance, and studying? Would students learn more or less? What
does this example illustrate about the relationship between
externalities and property rights?
20.9 A piece of state legislation proposes banning smoking in nearly all public facilities and private businesses. The major
argument for the bill is that it “is needed to protect nonsmokers
from the health hazards of cigarettes.” Prepare an evaluation of
the economic case for this legislation. (Assume that smoking
adversely affects the health of nearby nonsmokers.)
*20.10 “When public goods or externalities lead to inefficient resource allocation, government intervention is justified.” Is it?
Why?
20.11 In an otherwise competitive economy there is an exter- nality in the form of pollution. Show what the private market
equilibrium implies in terms of where we are on (or inside of)
the welfare frontier.
20.12 In Figure 20.2, suppose the government placed a quota instead of a tax on the output of the product that limited output
to a maximum of QE. Would this policy achieve an efficient allocation of resources?
20.13 In discussing Figure 20.3, the text states that the private equilibrium output, Q1, is inefficient. By definition, inefficiency is supposed to mean that everyone could be better off with a
different allocation of resources. Does the subsidy shown in the
diagram benefit everyone, including the taxpayer who pays to
finance it? If not, what type of policy could be used that would
benefit everyone?
20.14 A miracle drug that cures obesity is developed and produced competitively at a constant cost to producers of 10
cents per dose. The companies that produce the drug, however,
526 Publ ic Goods and External i t ies
C20.INDD 10:34:15:AM 08/06/2014 PAGE 526Trim Size: 203.2 mm X 254 mm
discharge chemical wastes resulting from the production pro-
cess into the nation’s rivers. The damage to the economy in
terms of reduced commercial and recreational use of the water-
ways is estimated to be 1 cent per dose. The social demand for
the drug is Q = 1,000,000 − 10,000P, where Q indicates the number of doses produced and P is the price per dose measured
in cents.
a. How much of the miracle drug is produced? At what price? Give both numerical and graphical answers.
b. An economist testifies before Congress that obesity cures are being overproduced and sold too cheaply. What price
and quantity do you think the economist would advocate?
What is she likely to estimate as the cost to the economy of
the supposed overproduction? Give numerical and graphi-
cal answers.
c. The economist argues that a tax should be placed on every unit of output that is accompanied by a waste discharge.
What size tax per unit of output would an efficiency-seek-
ing economist advocate? With this tax, what would the
price and output be for the miracle drug? Give both numeri-
cal and graphical answers.
20.15 Suppose that in the preceding problem drug producers invent another production process that discharges no waste and
can be used at a constant cost of 10.6 cents per dose.
a. What will the price and output of the miracle drug be if the tax advocated in part (c) of the preceding problem is
imposed?
b. What will the price and output be if the no-discharge pro- cess had a cost of 11.4 cents per dose?
c. Suppose that the production cost under the new process rises to 13 cents per unit and producers are forbidden to use
the original waste-discharging process. What price and out-
put will result? By the economist’s criteria, are there appro-
priate amounts of obesity cures and pollution? What is the
deadweight loss, if any, that results? Give both numerical
and graphical answers.
20.16 Comment on the following hypothetical remarks by the Congressional representatives who questioned the economist in
Problem 20.14. Do you agree or disagree? (You don’t have to
play the role of an economist in your answer.)
Senator Anthony: “No value can be placed on the benefit this drug brings to humanity.”
Senator Loeb: “The cure for obesity should be provided freely to all who need it.”
20.17 Distinguish between efficiency in production and effi- ciency in output. Can an industry achieve efficiency in produc-
tion and still produce at an inefficient output level? Explain.
20.18 Explain why efficiency in production is not realized if both refineries in Figure 20.4 are limited to emitting 100 pollu-
tion units each.
20.19 A clothing factory is located downwind from a copper smelting plant. The copper smelting plant emits particulates
into the air that cause $100,000 in damage per year to the cloth-
ing factory in terms of fabric discoloration. The plant could
eliminate its emissions of particulates by installing a superior
scrubbing technology at a cost of $50,000 per year. Existing
regulations do not prohibit the emission of these particulates.
The only way to remedy the inefficiency associated with the
emissions is through government intervention. True or false?
Explain your answer.
20.20 The defense services provided by submarines are a public good. Suppose that the equation relating the marginal
benefit Ted derives from the quantity of submarines produced
(Q) is MB = 600 − 10Q. The equation relating the marginal benefit Jane derives from submarines is MB = 400 − 10Q. The marginal cost of producing each submarine is a constant
$400.
a. If Ted and Jane are the only two individuals who benefit from the defense services provided by submarines, what is
the efficient output of submarines?
b. Without any coordination between Ted and Jane and/or government intervention, what will be the output of subma-
rines? Explain your answer. What will be the size of any
resulting inefficiency?
c. Suggest a taxing scheme to ensure the attainment of effi- ciency in the provision of submarines.
d. Suppose that the defense services afforded by submarines are still nonrival in consumption but that the cost of exclud-
ing a demander is zero. If submarines are produced by a
monopolist practicing third-degree price discrimination
(that is, a producer that charges different prices to different
demanders), what will be the profit-maximizing output and
price that the monopolist will charge Ted and Jane? How
does your answer compare to that in part a?
20.21 Should entrance fees be charged at our national parks? Explain why or why not.
20.22 The Copenhagen Accord of 2009 required most industrial nations to reduce their carbon dioxide and other
greenhouse-gas emissions during the ensuing decade by
5 to 40 percent, depending on the country. The United States
and Canada, for example, committed to reducing emission by
17 percent while India commited to a 20–25 percent reduction,
Mexico 30 percent, Israel 20 percent, and so on. Is such an
agreement efficient? Explain why or why not.
20.23 As of the late 1990s, there had been instances of Rus- sians shooting at Japanese, Tunisians shooting at Italians, and
Portuguese shooting at Spaniards. This is just a partial list of
the heated conflicts occurring on the high seas between aggres-
sive fishing fleets and well-armed navy and coast guard vessels
jealously protecting a lucrative and declining resource. At the
source of the conflict appears to be the inability to define prop-
erty rights to fish. Explain why.
20.24 The military alliance NATO (North Atlantic Treaty Organization) was formed during the Cold War to pre-
vent military action against its member states by the
Review Quest ions and Problems 527
C20.INDD 10:34:15:AM 08/06/2014 PAGE 527Trim Size: 203.2 mm X 254 mm
former Soviet Union. During its heyday, the seven countries
belonging to NATO with the largest gross national prod-
ucts (GNPs) spent 7.8 percent of their combined GNP on
defense. The seven nations with the smallest GNPs spent
only 4.3 percent. Provide an economic explanation for this
phenomenon.
20.25 When film crews rent a house to shoot a scene for a movie, external costs may be imposed on neighboring house-
holds in the forms of occupied/blocked parking spaces, bright
lights, and explosions. Relying on the Coase theorem, explain
why neighbors have been known to go to some unusual lengths
(revving up lawn mowers or stereos, turning on leaf blowers,
blowing foghorns, and banging garbage pail lids against iron
gates) in cases such as this, especially when filmmakers are
about to start shooting a scene.
20.26 Because the federal government covered the lion’s share of the costs, Los Angeles built the 17.4-mile-long Red Line sub-
way system in the 1990s, at a cost of $4.7 billion ($270 million
per mile). The fully allocated costs are estimated to be $1.15
per passenger mile, whereas the initial fare was $0.07 per mile
for the journey. Using a graph, explain why output efficiency
was or was not served by the subway’s construction. (Note: To cover the city’s portion of the construction costs, local bus fares
were raised, and daily bus ridership declined by more in the
wake of the fare increase than the number of residents taking
the Red Line on an average day.)
528
BANSWERS.INDD 12:15:56:PM 08/06/2014 PAGE 528Trim Size: 203.2 mm X 254 mm
2.14. The price of crack cocaine would rise and so would the total consumer outlay on crack cocaine since its demand is
inelastic. Since the funds to purchase crack cocaine are obtained
by stealing, the amount of crime would rise.
Chapter 3 3.2. The budget constraint is AMZ, kinked at point M. Along the AM portion the slope is $10/1V, while the slope is $5/1V along the MZ portion.
Other goods
A
0 5 Z Amusement park visits (V)
$10
1V
M
$5 1V
3.3. Income is $60, the price of hamburgers times the ham- burger intercept. The price of french fries is $2, income of $60
divided by the french fries intercept, 30. The slope is 20H/30F = PF/PH = $2/$3.
3.9. Marilyn would be at point N on the budget line (see next page for a graph). At this point, her MRS < PC/PF: the slope of the indifference curve at N is 1F/1C while the budget line’s slope is 3F/1C. Thus, N is not an optimal consumption point for Marilyn. The optimal point is M.
3.10. Not necessarily. Seat belts have costs as well as the benefit of reduced risk of injuries, and if motorists considered
the marginal cost of purchasing and using seat belts greater
than the marginal benefit, they would be rational not to pur-
chase them. This would be shown as a corner equilibrium, as
in Figure 3.12.
Chapter 1 1.4. False. The opportunity cost of the land to RAND equals the amount RAND could sell the land for. This amount is likely
to be substantial, given the prime location of the land.
Chapter 2 2.1. First, because the increase in the overall price level over the year is not given, it is not clear how the relative price of
textbooks has changed. Second, it is possible that the demand
curve has shifted over the year, perhaps due to such things as
changes in college enrollments or greater interest in economics
and business courses.
2.5. (a) $0.70 and 120,000. (b) The demand curve confronting producers becomes DAD1. Output will be 140,000 with consum- ers purchasing 110,000 and the government 30,000 (equal to
the output consumers do not purchase—140,000 less 110,000).
(c) Consumer demand becomes D″D″′, and the demand curve confronting producers becomes D″A′D1. Total output and price do not change, but consumer purchases rise to 121,000 and
government purchases fall to 19,000 (140,000 less 121,000).
Price
$0.80
$0.70
0 110 120
121 140 Apples (1,000's)
D‴
D1
S
A A ′
D″
D ′
D
2.8. (a) Will increase demand. (b) Will increase demand. (c) Will reduce supply. (d) Will increase demand and reduce
supply. (e) Will increase demand. (f) Will increase demand.
2.10. False. The demand elasticity for a more narrowly defined good (such as a particular cigarette brand) is always greater than
for a less narrowly defined good (cigarettes in general).
Answers to Selected Problems
Answers to Selected Problems 529
BANSWERS.INDD 12:15:56:PM 08/06/2014 PAGE 529Trim Size: 203.2 mm X 254 mm
and Joe’s optimal consumption point is M. Since the MRS at M is greater than $2.50 per gallon, Joe will pay more than $2.50 to
get additional gasoline beyond 25 gallons.
Other goods
$2.50
M
U2 U1
1G
A
0 G′ (25)
50 Z Gasoline
4.6. The budget line pivots from AZ to A′Z. The income effect implies greater car maintenance consumption (with mainte-
nance a normal good), while the substitution effect implies less
consumption of car maintenance, so we cannot predict the over-
all effect. (That illustrates why income must be held constant in
drawing a demand curve; in the present case, the demand curve
has shifted outward.) However, both the income and substitution
effects imply greater consumption of other goods.
Other goods
A′
A
Z Car maintenance
0
4.16. Fabio will be better off and consume less gasoline. The price increase changes the budget line from AZ to AZ′ (see next page)
Perfume (F)
60 A
24 1F
M
N U2 U1
3F
1C
1C
0 12 Z (20)
Champagne (C)
3.13. Yes; at a higher income, it is possible for a consumer to purchase more of all goods. No; since all income is spent on
something (remember, saving is considered one of the goods), at
a higher income more of at least one good must be purchased.
3.14. Market baskets are shown as points A and B. If good V is a normal good, MRSVO—the slope of the indifference curve—must be greater at B, as shown in the diagram. Otherwise the consumer would not purchase more V at a higher income. If V is an inferior good, MRSVO at point B would be less than at point A.
Other goods (O)
$150
$100
0 10 Video rentals (V)
B
A
U2 U1
Chapter 4 4.4. If Joe could purchase as much gas as desired at a price of $2.50, the budget line would be AZ. However, since only 25 gallons can be acquired at that price, the budget line is AMG′,
530 Answers to Selected Problems
BANSWERS.INDD 12:15:56:PM 08/06/2014 PAGE 530Trim Size: 203.2 mm X 254 mm
$4,000,000, the endowment point shifts to N′ and the budget line becomes A′Z′. With consumption in both years being a normal good, consumption in both years will fall due to the
income effect (there is no substitution effect), as shown by the
new optimal point E′. Borrowing has increased, so year 1 con- sumption has fallen by less than the reduction in earnings.
Consumption in year 1
7,000,000
4,000,000
0 8,750,000 I2
U2
U1
Consumption in year 2
I1
Z′ Z
A
E
N
A′
E′
N′
Chapter 6 6.3. At point A, John would be willing to give up three gallons of gasoline for one bag of sugar and remain equally well off.
Thus giving up two gallons of gasoline for one bag of sugar as
in a move from point A to point B would make John better off (John would end up on a higher indifference curve, U2J versus U1J). Maria would be willing to participate in the trade repre- sented by the move from point A to point B since she is willing at point A to give up one gallon of gasoline in exchange for one bag of sugar. A move from point A to point B results in Maria getting two gallons of gasoline for one bag of sugar and makes
Maria better off (she ends up on indifference curve U2M versus U1M at point A).
0M
0J
40 A B
38 80G
70S
2021
U J 2
U J 1
UM 2 U
M 1
and distance ET = $500 (= 1,000 gallons times $0.50). Thus, a cash transfer of $500 shifts the budget line to A1Z1, which passes through point E since A1A = ET = $500.
E ′
E
T
U2
U1
0 1,000 Z′ Z1 Z Gasoline
Other goods
A1
A
Chapter 5 5.2. Kate will be better off with food stamps but will consume more food with the excise subsidy. Her optimal point with the
food stamps is E on budget line AA′Z′, with the $100 cost shown by the distance ET. Her optimal point with the excise subsidy is E′ on budget line AZ1 with the $100 cost equal to E′T′. The trick is to locate the size of the excise subsidy per unit that costs
exactly $100, and this implies that the optimal point with the
excise subsidy is where Kate’s price-consumption curve inter-
sects A′Z′. Note that with a larger excise subsidy per unit (lower price), total cost would exceed $100, and vice versa.
Other goods
0 F1 F2 Z FoodZ′
A A′
E′
T′
E
U2
U1
Z1
T
5.11. The initial endowment point is N, AZ is the budget line, and Anna’s optimal point is E. When year 1 earnings fall to
Answers to Selected Problems 531
BANSWERS.INDD 12:15:56:PM 08/06/2014 PAGE 531Trim Size: 203.2 mm X 254 mm
Total product (TP)
0 Input
TP
7.4. The total product curve has a slope that becomes flatter right from the start, as shown in the above diagram. The mar-
ginal and average product curves are depicted below. Note that
AP = MP for the first unit.
Output per unit of input
0 Input
AP
MP
7.5. Since the marginal product indicates how much output will rise if the worker is hired, the firm should be concerned with the
marginal product.
7.9. The isoquant might have a slope of (minus) one at this point, but it does not have to. If it does, this implies that the
marginal products of both inputs are equal when equal quanti-
ties are employed. Remember that the slope of the isoquant
equals the ratio of the marginal products.
7.11. Eighteen units of output.
Chapter 8 8.2. With MP falling from the start, MC and AC are rising from the start (as shown on the next page). If MP first rises and then falls, the cost curves have their conventional U shapes.
6.5. The question describes the situation shown at point A on one of the sides of the box.
0M
0J
(40)A 20
60G
20S
U J 1
UM 1
Although the MRSs differ, the distribution is efficient since there is no way to move without harming one of the two persons.
The equality-of-MRSs’ condition is necessary for efficiency only when both parties consume some positive amount of both
commodities.
6.7. No. John will pay a maximum of 4G for one S, so after the tax is collected, that leaves only one G for one S, which is less than Maria requires to part with one S. Both John and Maria are harmed by the tax since they don’t get the potential gain from
exchange; no one is benefited since the government collects no
tax revenue.
6.9. An equal distribution is shown at the midpoint of the diag- onal line connecting the origins. This distribution would be effi-
cient only if the consumers’ MRSs happen to be equal at that point. If the consumers have different preferences, this will not
be so.
6.11. Voluntarism in trade includes the right to stop trading with someone if you can do better disposing of the good in
another way. Presumably, the apartment owner and the future
purchasers of condominiums will both benefit from the conver-
sion if the owner’s judgment is correct. Note that this situation
is no different in principle from the situation when you stop
eating at McDonald’s and patronize Burger King instead.
6.15. No. It only implies that both children would consider themselves better off after the trade. The parent may wish
to override their wishes on the ground that the children don’t
appreciate the benefits of a more balanced diet.
Chapter 7 7.3. The average and marginal product curves would coin- cide and be a horizontal line: AP and MP are constant regard- less of the amount of input used. If the law of diminishing
marginal returns holds, the total product curve would not have
this shape.
532 Answers to Selected Problems
BANSWERS.INDD 12:15:56:PM 08/06/2014 PAGE 532Trim Size: 203.2 mm X 254 mm
9.7. False. The difference between price and marginal cost is the change in profit (positive or negative) that would be real-
ized by a unit change in output. The firm wishes to maximize
total profit, which implies that price minus marginal cost
equals zero.
9.13. False. Enough defense contractors need to exit such that the short-run industry supply curve shifts sufficiently to the left
to ensure that the price can once again cover average cost for
the contractors remaining in the industry.
Chapter 10 10.5. With a constant-cost industry, with LS1 and D, P and Q are the equilibrium price and quantity. The subsidy shifts LS1 to LS1 − S and price falls to P1, or by the amount of the subsidy per unit S (20 cents). The deadweight loss is equal to triangular area ABC. With an increasing-cost industry, the supply curve is LS2 and shifts to LS2 − S when the subsidy is applied. Price falls by less in this case (to P2).
Dollars per unit
P P2 P1
0 Q Q2 Q1 Output
LS1
LS1 – S
LS2
LS2 – S
S
S
A
C
B
D
10.8. Either method of deregulation would lead to a competi- tive market in taxi services, but the cost would be imposed on
different people. With the government purchasing the medal-
lions, the cost is borne by taxpayers; without purchases, the cost
would be borne by the current owners of the medallions. To
decide which is better, a value judgment must be made concern-
ing who should bear the cost.
Chapter 11 11.8. In this case, total revenue is equal to profit, so the monopolist produces where total revenue is at a maximum,
or where marginal revenue equals zero. Thus, output is Q and price is P in the diagram (see next page). The deadweight loss is shown by the area ADQ.
Cost per unit
0 Output
MC
AC
8.4. Basically, the reason is that we assume firms attempt to maximize profit. If the same output can be produced at a lower
total cost, a firm would make a larger profit by doing so. This
reasoning does not provide a basis for assuming that the U.S.
Postal Service and the American Red Cross minimize cost since
they are not profit-making operations.
8.9. Coldwell Banker is at point N on isocost line AZ. At point N, the isoquant has a slope of 50C/4L = 12.5, which is flatter than the isocost line with a slope of 4,000C/12L = 333.33. By operating at point M on the lower isocost line A′Z′, Coldwell Banker can produce the same output at a lower cost.
Cement
50
0 Z′ 10 Z (10.15)
Land
333.33
1
IQ1,000
M
N
12.5 1
(3,383.33) A
A′
(C)
(L)
Chapter 9 9.3. (a) Output is zero and losses equal the fixed cost of $60,000 (profit equals minus $60,000) since the firm cannot
cover AVC. (b) Output is seven units and the firm loses $28,000 (profit equals minus $28,000), but this is less than it would lose
if it shut down.
Answers to Selected Problems 533
BANSWERS.INDD 12:15:56:PM 08/06/2014 PAGE 533Trim Size: 203.2 mm X 254 mm
may simply cover these costs while the mileage fee covers the
additional costs associated with driving the automobiles.
12.7. Food need not be consumed at the same time it is pur- chased; it can be stored. The peak-load phenomenon occurs
only when production and consumption must occur simultane-
ously, and there is a systematic variation in demand over time.
Chapter 13 13.8. With the supply curve of the follower firms shown as S0, the dominant firm faces a demand curve PAD and marginal rev- enue curve PALM. The dominant firm can only affect the price if it sets a price below P. If the dominant firm’s marginal cost curve is MC1, price will be P with both the dominant firm and other firms supplying a total output of Q. If MC2 is the cost curve, the dominant firm supplies the entire market at P and produces Q. Only if the MC curve is lower than MC3 will price and total output differ from P and Q. In most cases, a horizontal supply curve of other firms means the dominant firm will have
no effect on price and output.
Dollars per unit
P
D
M
L
A S0
MC1
MC2
MC3
Q Quantity0
Chapter 14 14.2. Assuming “high output” is the long-run competitive equi- librium output for each firm, the matrix would look something
like the one depicted below.
–10 –10 –10
–10
0
00 0
Firm B
Firm A
Low output High output
High output
Low output
Dollars per unit
P
Q Output
MR
D
A
0
11.10. The deadweight loss due to monopoly means that con- sumers would be willing to give up the necessary quantities
of other goods to have more of the monopolized product. This
is shown by the fact that at the monopolist’s output, price
(the marginal value to consumers) is greater than marginal
cost (the value of other goods that could be produced by the
resources required to produce another unit of the monopo-
list’s product).
Chapter 12 12.1. The ΣMR curve is ABD2. Total output is Q2. Q1 is sold in the home market at P2, while Q2 − Q1, or Q1Q2, is sold in the international market at the lower price, P1.
Dollars per unit
P2
P1
0 Q1 Q2 Output
B
MC
D2 (= MR2)
D1
MR1
A
12.5. Probably not. There are a large number of car rental firms, and the presumption is that the market is competitive;
competitive firms cannot use monopolistic pricing tactics. Note
that there are costs to the firms from renting cars independent
of how much they are driven (depreciation, some maintenance,
insurance, administrative overhead, and so on), and the daily fee
534 Answers to Selected Problems
BANSWERS.INDD 12:15:56:PM 08/06/2014 PAGE 534Trim Size: 203.2 mm X 254 mm
16.8. False. Each employer must take the market wage rate as given. Employers expand employment at the market-determined
wage rate up to the point where MVP equals that wage rate.
16.9. In the short run, an unanticipated increase in labor sup- ply will lead to economic profits for firms and somewhat lower
prices for consumers. In the long run, economic profits of the
firms will be zero. Consumers, who can get the product at a
lower price, are the primary beneficiaries of an increase in labor
supply in the long run.
16.10. No single firm except a monopsony has the ability to cut wage rates without losing employees. This quotation illus-
trates the common error of treating thousands of separate firms
as if they could collude to form one gigantic monopsony. One
could just as well argue that consumers force down the prices
of products they buy at the expense of enormous losses to
firms.
16.12. If the supply curve of steel confronting the combined automobile and refrigerator industries is upward sloping, the
price of refrigerators will rise. The analysis would be the same
as in Figure 16.8. However, steel is used by a number of other
industries, so the share of total steel output used by these
two may be fairly small, in which case the steel supply curve
will be horizontal and the price of refrigerators would not
be affected.
Chapter 17 17.3. Initially, Jeeves is at point E on AZ. The inheritance produces a parallel shift in the budget line to A′Z′Z, where Z′Z = $200. There is only an income effect here since the wage
rate has not changed. If leisure is a normal good, Jeeves will
consume more of it—that is, work less. For example, the new
optimal point might be at E′, and his total weekly income would be $400, up only $100 because he is earning $100 less
than before.
Income
$400
$300
$200
0 L1 L2 Z Leisure
E'
Z′
E
U2
U1
A′
A
Each firm’s profit depends only on its own output; changes in
one firm’s output have no effect on the other firm’s profit. There
is no interdependence between the firms to take account of, so
game theory is not necessary or useful.
14.12. If 1 in 50 people gets cancer, the cost of a policy would be about one-fiftieth of the cost of medical care in a competitive
market. If the probability is 1 in 100 for nonsmokers and 1 in
10 for smokers, and smokers know this, then insurance would
become much more attractive for smokers; there would be an
adverse selection of smokers. Firms would find that they have
to pay out on more than 1 in 50 policies, and the price would
rise. In the limit, the price would rise to one-tenth the cost of
medical care, and only smokers would purchase insurance.
14.17. There are at least three reasons the statement is false. First, price may be above average cost in the absence of advertis-
ing, and equal to average cost when advertising increases com-
petition. Second, the firm may operate at a lower point on the
higher average cost curve as a result of advertising. Even though
advertising results in the firm operating on AC′ rather than AC, the firm may be at point B with advertising and point A without advertising (because advertising leads to more elastic demand
curves). Third, even if the money price of the product does rise
as a result of advertising, the full price may still be lower.
Dollars per unit
Output
A B
AC′ AC
0
Chapter 15 15.2. Let P and Q be the competitive price and quantity, respectively. The oligopoly price and output are therefore 1.20P and 0.8Q, respectively. Thus, the deadweight loss is 0.5(0.2P) (0.2Q), or 0.02PQ. Because the total outlay on the product equals (1.2P)(0.8Q), or 0.96PQ, the deadweight loss is 2.1 per- cent as large as the total consumer outlay.
Chapter 16 16.4. The marginal value product of any given number of workers will be higher when product demand increases since
consumers pay a higher price for the product.
Answers to Selected Problems 535
BANSWERS.INDD 12:15:56:PM 08/06/2014 PAGE 535Trim Size: 203.2 mm X 254 mm
the marginal value (or MRS) of rental housing to consumers. The marginal value of rental housing would be greater than
marginal cost, so output is too low.
19.18. If the domestic country’s exports and imports are a small part of total world trade in these products, then the partici-
pation of the domestic country in world trade would have a neg-
ligible effect on world prices. If the world price ratio is affected,
the relative price of the good imported by the domestic coun-
try will rise. In Figure 19.12, for example, if the world price
prior to trade is 1F/1C, the final price might be 1F/1.5C after the domestic country begins to trade. The gain to the domestic
country from trade is less in this case.
Chapter 20 20.4. The dove, individual C, has a negative demand for defense. This must be added vertically to the demands of the
hawks A and B. The result, Σd, is lower than it would be if we summed the demands of A and B only. The intersection of Σd and MC identifies the efficient output.
Dollars per unit
$12
$9
$6
0 QE Defense
dC
dB
dA
MC
�d
–$3
20.8. We would predict that class attendance would rise, poorer notes would be taken and students would study the text more.
Students would probably learn less.
20.10. No. Government intervention could make a bad situ- ation worse. Only an explicit analysis of the effects of gov-
ernment intervention could show whether it would improve
matters, in the sense of leading to a more efficient resource
allocation.
17.4. If money income is a normal good, a worker will always choose to earn more of it when the wage rate rises, so there is
a limit on how backward bending a labor supply curve can be.
A higher wage rate has an income effect implying greater con-
sumption of money income (if it is a normal good) and a substi-
tution effect in favor of money income also, so the net effect is
unambiguous.
17.7. False. Not everyone who has the ability to be a nuclear physicist is one. At a higher wage for physicists, others with the
requisite ability would be attracted to the profession. At a lower
wage, fewer would.
17.14. Aggregate labor demand will be higher if investment is K2 in year 1. The more investment today, the greater the amount of capital per worker in the future. More capital per worker
means the marginal productivity of workers will be greater, and
hence demand for workers will be greater too.
Chapter 18 18.3. Labor costs will not increase if the money wages that must be paid go down by the amount of the increase in the
employer tax. As explained in the text, if the supply curve of
labor is perfectly vertical, this will happen.
18.8. No Amazons would be employed by health clubs catering to discriminating customers (Type A clubs); all the Amazons
would be employed by Type B clubs serving nondiscriminat-
ing customers. Wage rates for non-Amazons and Amazons will
remain equal, however, unless there are fewer non-Amazons
initially employed by Type B clubs than there are Amazons
employed by Type A clubs. To see this, imagine that wage rates
did go down in Type B clubs and up in Type A clubs. Then
the non-Amazons initially in Type B clubs have an incentive to
shift to Type A clubs, and this process would continue until the
wage rates at the two types of clubs were again equal.
Chapter 19 19.2. No, at least not from the demand side of the market since the price of margarine would not change in this case, and thus
the demand curve for butter would not shift.
19.5. Ten cellular telephones per personal computer. The MRT is equal to the ratio of marginal costs, and since marginal cost
equals price in equilibrium, the MRT equals the ratio of the price of personal computers to the price of cellular telephones.
19.10. No, if markets are competitive.
19.12. No. Consumers cannot purchase as much as they want at the controlled rent, and so the controlled rent does not measure
536
BGLOSS.INDD 12:14:36:PM 08/06/2014 PAGE 536Trim Size: 203.2 mm X 254 mm
Absolute cost advantage A situation in which an incumbent firm’s production cost (its long-run average total cost) is lower
than potential rivals’ production costs at all relevant output
levels.
Accounting costs Costs reported in companies’ net income statements generated by accountants.
Adverse selection A situation in which asymmetric informa- tion causes higher-risk customers to be more likely to purchase
or sellers to be more likely to supply low- quality goods.
Antitrust laws A series of codes and amendments intended to promote a competitive market environment.
Arc price elasticity of demand formula
Artificial product differentiation The use of advertising to dif- ferentiate products that are essentially the same.
Asymmetric information A case in which participants on one side of the market know more about a good’s quality than do
participants on the other side.
Average fixed cost (AFC) Total fixed cost divided by the amount of output.
Average input cost The total cost of an input divided by the units of that input used by a firm.
Average product The total output (or total product) divided by the amount of the input used to produce that output.
Average profit per unit (π/q) Total profit divided by number of units sold.
Average revenue (AR) Total revenue divided by output.
Average total cost (ATC) Total cost divided by the output.
Average variable cost (AVC) Total variable cost divided by the amount of output.
Bandwagon effect A positive network effect.
Barrier to entry Any factor that limits the number of firms oper- ating in a market and thereby serves to promote monopoly power.
Black market An illegal market for a good.
Block pricing or second-degree price discrimination The use of a schedule of prices such that the price per unit declines with
the quantity purchased by a particular consumer.
=
⎡
⎣ ⎢
⎤
⎦ ⎥
⎡ ⎣⎢
⎤ ⎦⎥
Δ ⎛ ⎝⎜
⎞ ⎠⎟
+( ) Δ
⎛ ⎝⎜
⎞ ⎠⎟
+( )
Q
Q Q
P
P P
d
d d 1
2
1
2
1 2
1 2
Budget constraint The way in which a consumer’s income and the prices that must be paid for various goods limit choices.
Budget line A line that shows the combinations of goods that can be purchased at the specified prices and assuming that all of
the consumer’s income is expended.
Cartel An agreement among independent producers to coor- dinate their decisions so each of them will earn monopoly
profit.
Coase theorem The idea that as long as property rights are clearly defined and enforced, bargaining between two parties
can ensure an efficient outcome.
Cobb–Douglas function A commonly used functional form in economics.
Cobb–Douglas production function A production function that does not imply constant marginal products for inputs.
Commitment The strategy of adopting a particular course of action, constraining one’s choice of strategies, in order to
increase your equilibrium payoff.
Compensating wage differentials Differences in wages paid that are created by the forces of supply and demand when
workers view some jobs as intrinsically more attractive than
others.
Complements Two goods that tend to be consumed together, so consumption of both tends to rise or fall simultaneously.
Composite good A number of goods treated as a group.
Constant-cost industry An industry in which expansion of out- put does not bid up input prices, long-run average production cost
per unit remains unchanged, and the long-run industry supply
curve is horizontal.
Constant returns to scale A situation in which a proportional increase in all inputs increases output in the same proportion.
Consumer surplus A measure of the net gain to a consumer or group of consumers from purchasing a good arising from
its cost being below the maximum that consumers are willing
to pay.
Contestable markets Markets in which competition is so per- fect that the market price is independent of the number of firms
currently serving a market, because the mere possibility of entry
suffices to discipline the actions of incumbent suppliers.
Contract curve In an Edgeworth exchange box, a line drawn through all the efficient distributions.
Glossary
Glossary 537
BGLOSS.INDD 12:14:36:PM 08/06/2014 PAGE 537Trim Size: 203.2 mm X 254 mm
Corner solution A situation in which a particular good is not consumed at all by an individual consumer because the value of
the first unit of the good is less than the cost.
Cournot model A model of oligopoly that assumes each firm determines its output based on the assumption that any other
firms will not change their outputs.
Cross-price elasticity of demand A measure of how respon- sive consumption of one good is to a change in the price of a
related good.
Deadweight loss or welfare cost A measure of the aggregate loss in well-being of participants in a market resulting from out-
put not being at the effficient level.
Decreasing-cost industry A highly unusual situation in which the long-run supply curve is downward sloping.
Decreasing returns to scale A situation in which output increases less than proportionally to input use.
Derived demand Another name for an industry’s input demand curve, reflecting the fact that the industry’s demand for an input
ultimately derives from consumers’ demand for the final prod-
uct produced by that input.
Differentiated product A product that consumers view as dif- ferent from other similar products.
Diminishing marginal utility The assumption that as more of a given good is consumed, the marginal utility associated with the
consumption of additional units tends to decline, other things equal.
Diminishing MRS A consumer’s willingness to give up less and less of some other good to obtain still more of the first good.
Diseconomies of scale A situation in which a firm’s output increases less than proportionally to its total input cost.
Diseconomies of scope A case in which it is cheaper for sepa- rate products to be produced independently than for one firm to
produce the same products jointly.
Disemployment effect The tendency of employers to respond to a higher wage rate by hiring fewer workers.
Disequilibrium A situation in which the quantity demanded and the quantity supplied are not in balance.
Diversification Investing a given amount of resources in numer- ous independent projects instead of a single project in order to
minimize exposure to risk.
Dominant firm model A model of oligopoly in which the leader or dominant firm assumes its rivals behave like competi-
tive firms in determining their output.
Dominant strategy A case where a player is better off adopting a particular strategy regardless of the strategy adopted by the
other player.
Dominant-strategy equilibrium The simplest game theory out- come, resulting from both players having dominant strategies.
Duopoly An industry with just two firms.
Dynamic analysis A form of economic analysis that looks, over time, at the efficiency of a market.
Econometrics (regression analysis) A statistical method that allows one to estimate, among other things, the sensitivity of the
quantity demanded of a good to determinants such as price and
income.
Economic “bads” Commodities of which less is preferred to more over all possible ranges of consumption.
Economic cost or opportunity cost The sum of explicit and implicit costs.
Economic efficiency With regard to exchange, economic effi- ciency represents a distribution of goods across consumers in
which no one consumer can be made better off without hurting
another consumer.
Economic “goods” Commodities of which more is better than less.
Economic “neuter” A case in which the consumer doesn’t care one way or another about a particular good.
Economic rent That portion of the payment to an input supplier in excess of the minimum amount necessary to retain the input
in its present use.
Economies of scale A situation in which a firm can increase its output more than proportionally to its total input cost.
Economies of scope A case where it is cheaper for one firm to produce products jointly than it is for separate firms to produce
the same products independently.
Edgeworth exchange box A diagram for examining the allo- cation of fixed total quantities of two goods between two
consumers.
Edgeworth production box A diagram that identifies all the ways two inputs such as labor and land can be allocated
between industries in a simplified economy.
Efficiency in output The condition in which output is expanded to the point where marginal benefit equals marginal cost.
Efficiency wage A wage higher than the prevailing market- determined level that serves to increase firms’ profits by low-
ering the costs of searching for, selecting, and training new
workers.
Efficient An outcome that is Pareto optimal.
Elastic The situation in which price elasticity of demand exceeds 1.0 or unity.
Elasticities Measures of the magnitude of the responsiveness of any variable (such as quantity demanded or supplied) to a
change in particular determinants.
Endowment point The consumption mix available to the indi- vidual if no saving or borrowing takes place.
538 Glossary
BGLOSS.INDD 12:14:36:PM 08/06/2014 PAGE 538Trim Size: 203.2 mm X 254 mm
Entry fee The fixed fee charged per time period in the case of a two-part tariff.
Equilibrium A situation in which quantity demanded equals quantity supplied at the prevailing price.
Equity The concept of fairness.
Excess burden Another name for the deadweight loss produced by a tax.
Excess capacity The result of firms failing to produce at lowest possible average cost.
Excise subsidy A form of subsidy in which the government pays part of the per-unit price of a good and allows consumers
to purchase as many units as desired at the subsidized price.
Excise tax A tax on a specific good.
Expansion path A line formed by connecting the points of tan- gency between isocost lines and the highest respective attain-
able isoquants.
Expected return The summed value of each possible rate of return weighted by its probability.
Expected utility The summed value of each possible utility weighted by its probability.
Explicit costs Money used in the pursuit of a goal that could otherwise have been spent on an alternative objective.
External benefits Positive side effects of ordinary economic activities.
External costs Negative side effects of ordinary economic activities.
Externalities The harmful or beneficial side effects of mar- ket activities that are not fully borne or realized by market
participants.
Factors of production Inputs or ingredients mixed together by a firm through its technology to produce output.
Feedback effect A change in equilibrium in a market that is caused by events in other markets that, in turn, are the
result of an initial change in equilibrium in the market under
consideration.
First-degree (perfect) price discrimination A policy in which each unit of output is sold for the maximum price a consumer
will pay.
Fixed inputs Resources a firm cannot feasibly vary over the time period involved.
Free entry and exit A situation in which there are no differen- tial impediments across firms in the mobility of resources into
and out of an industry.
Free rider A consumer who has an incentive to underestimate the value of a good in order to secure its benefits at a lower, or
zero, cost.
Full price The sum of the money price and the search costs that consumers incur.
Game theory A method of analyzing situations in which the outcomes of your choices depend on others’ choices, and vice
versa.
General equilibrium analysis The study of how equilibrium is determined in all markets simultaneously.
Giffen good The result of an income effect being larger than the substitution effect for an inferior good, so that the demand curve
will have a positive slope.
Goal-oriented behavior The behavior of market participants interested in fulfilling their own, personal goals.
Golden rule of cost minimization A rule that says that to minimize cost, the firm should employ inputs in such a way
that the marginal product per dollar spent is equal across all
inputs.
Gross marginal productivity The total addition to productivity that capital investment contributes.
Homogeneous products Standardized products that, in the eyes of consumers, are perfect substitutes for one another.
Human capital investment The process by which people aug- ment their earning capacity.
Imperfect information The case when market participants lack some information relevant to their decisions.
Implicit costs Costs associated with the individual’s use of his or her own time.
Income-consumption curve The line that joins all the optimal consumption points generated by varying income.
Income effect A change in a consumer’s real purchasing power brought about by a change in the price of a good.
Income elasticity of demand A measure of how responsive consumption of some item is to a change in income, assuming
the price of the good itself remains unchanged.
Increasing-cost industry An industry in which expansion of output leads to higher long-run average production costs and the
long-run supply curve slopes upward.
Increasing returns to scale A situation in which output increases in greater proportion than input use.
Indifference curve A plot of all the market baskets the con- sumer views as being equally satisfactory.
Indifference map A set of indifference curves that shows the consumer’s entire preference ranking.
Inefficiency An allocation of goods in which it is possible, through a change in the distribution, to benefit one party with-
out harming the other.
Inefficient The condition in which it is possible, through some feasible reallocation of resources, to benefit at least one person
without making any other person worse off.
Inelastic The situation in which price elasticity of demand is less than 1.0 or unity.
Glossary 539
BGLOSS.INDD 12:14:36:PM 08/06/2014 PAGE 539Trim Size: 203.2 mm X 254 mm
Inferior goods Those goods whose consumption falls when income rises.
Input substitution effect The effect of a change in the price of an input on a firm’s relative use of the input to produce a given
level of output.
Insurance An arrangement by which the consumer pays a pre- mium in return for the promise that the insurer will provide com-
pensation for losses due to an accident, illness, fire, and so on.
Interest rate (1) The price paid by borrowers for the use of funds, and (2) the rate of return earned by capital as an input in
the production process.
Intertemporal price discrimination A form of third-degree price discrimination in which different market segments are
willing to pay different prices depending on the time at which
they purchase the good.
Investment demand curve The relationship between the rate of return generated and various levels of investment.
Isocost line A line that identifies all the combinations of capital and labor that can be purchased at a given total cost.
Isoquant A curve that shows all the combinations of inputs that, when used in a technologically efficient way, will produce
a certain level of output.
Iterated dominance The concept of eliminating any strategy that is inferior to or dominated by another strategy.
Law of demand The economic principle that says the lower the price of a good the larger the quantity consumers wish to
purchase.
Law of diminishing marginal returns A relationship between output and input that holds that as the amount of some input
is increased in equal increments, while technology and other
inputs are held constant, the resulting increments in output will
decrease in magnitude.
Law of supply The economic principle that says the higher the price of a good, the larger the quantity firms want to produce.
Learning by doing Improvements in productivity resulting from a firm’s cumulative output experience.
Leisure The portion of a worker’s time when he or she is not receiving compensation from an employer.
Lerner index A means of measuring a firm’s monopoly power that takes the markup of price over marginal cost expressed as a
percentage of a product’s price.
Long run A period of time in which the firm can vary all its inputs.
Long-run industry supply curve The long-run relationship between price and industry output, which depends on whether
input prices are constant, increasing or decreasing as the indus-
try expands or contracts.
Lump-sum transfer A form of subsidy in which the govern- ment gives the consumer a cash grant to be spent in any way the
recipient wants.
Macroeconomics The study of aggregate economic factors.
Marginal benefit The incremental value a consumer derives from consuming an additional unit of a good and thus the maxi-
mum amount the consumer would pay for that additional unit.
Marginal cost (MC) The change in total cost that results from a one-unit change in output.
Marginal input cost The cost of using an additional unit of an input.
Marginal product The change in total output that results from a one-unit change in the amount of an input, holding the quanti-
ties of other inputs constant.
Marginal rate of substitution (MRS) A measure of the consum- er’s willingness to trade one good for another.
Marginal rate of technical substitution The amount by which one input can be reduced without changing output when there is
a small (unit) increase in the amount of another input.
Marginal rate of transformation The rate at which one product can be “transformed” into another.
Marginal revenue (MR) The change in total revenue when there is a one-unit change in output.
Marginal revenue product The product of an input’s marginal product and the marginal revenue that can be derived from sell-
ing that marginal product.
Marginal utility The amount by which total utility rises when consumption increases by one unit.
Marginal value product (MVP) The extra revenue a competi- tive firm receives by selling the additional output generated
when employment of an input is increased by one unit.
Market baskets Combinations of goods.
Markets The interplay of all potential buyers and sellers of a particular commodity or service.
Market segmentation or third-degree price discrimination A situation in which each consumer faces a single price and can
purchase as much as desired at that price, but the price differs
among categories of consumers.
Microeconomics The study of the behavior of small economic units, such as consumers and firms.
Minimum efficient scale The scale of operations at which aver- age cost per unit reaches a minimum.
Monopolistic competition A market characterized by unre- stricted entry and exit and a large number of independent sellers
producing differentiated products.
Monopoly A market with a single seller.
Monopoly power Some ability to set price above marginal cost.
Monopsony An input market in which a firm is the sole pur- chaser of an input.
Moral hazard A situation that occurs when, as a result of having insurance, an individual becomes more likely to engage in risky
behavior.
540 Glossary
BGLOSS.INDD 12:14:36:PM 08/06/2014 PAGE 540Trim Size: 203.2 mm X 254 mm
Movement along a given demand curve A change in quan- tity demanded that occurs in response to a change in price, other
factors holding constant.
Movement along a given supply curve A change in quantity supplied that occurs in response to a change in the good’s sell-
ing price, other factors holding constant.
Nash equilibrium A set of strategies such that each player’s choice is the best one possible given the strategy chosen by the
other player(s).
Natural monopoly An industry in which production cost is minimized if one firm supplies the entire output.
Net marginal productivity The total addition to productivity that capital investment contributes, less the cost of capital.
Network effects The extent to which an individual consumer’s demand for a good is influenced by other individuals’ purchases.
New entrant/survivor technique Method for determining the minimum efficient scale of production in an industry based on
investigating the plant sizes either being built or used by firms
in the industry.
Nominal price The absolute price, not adjusted for the chang- ing value of money.
Nonexclusion A condition in which confining a good’s ben- efits, once produced, to selected persons is impossible or pro-
hibitively costly.
Nonrival in consumption A condition in which a good with a given level of production, if consumed by one person, can also
be consumed by others.
Normal goods Those goods for which an increase in income leads to greater consumption.
Normative analysis A nonscientific value judgment.
Oligopoly An industry structure characterized by a few firms producing all or most of the output of some good that may or
may not be differentiated.
Opportunity cost or economic cost The sum of explicit and implicit costs.
Ordinary least-squares (OLS) A technique for estimating the equation that “best fits” the data.
Output effect of an input price change The change in input employment when output is altered in response to a change in
the price of an input.
Pareto optimal The condition in which it is not possible, through any feasible change in resource allocation to benefit one person
without making some other person or persons worse off.
Pareto optimality Another term for economic efficiency.
Partial equilibrium analysis The study of the determination of an equilibrium price and quantity in a given product or input mar-
ket viewed as self-contained and independent of other markets.
Payoff matrix A simple way of representing how each combina- tion of choices affects players’ payoffs in a game theory setting.
Payoffs In game theory, the outcomes or consequences of the strategies chosen.
Peak-load pricing A pricing policy in which different prices are charged for peak and off-peak periods.
Perfect competition An economic model characterized by the assumption of (1) a large number of buyers and sellers, (2)
free entry and exit, (3) product homogeneity, and (4) perfect
information.
Perfect complements Goods that must be consumed in a pre- cise combination in order to provide a consumer a given level of
satisfaction.
Perfect (first-degree) price discrimination A policy in which each unit of output is sold for the maximum price a consumer
will pay.
Perfect substitutes The case where a consumer is willing to substitute one good for another at some constant rate and
remain equally well off.
Players In game theory, decision makers whose behavior we are trying to predict and/or explain.
Point price elasticity of demand formula = Δ Δ
( )
( )
Q Q P P d d/
/
Positive analysis Assessment of expected, objective outcomes.
Preferences or tastes The feelings of consumers about the desirability of different goods.
Price ceiling A legislated maximum price for a good.
Price-consumption curve A curve that identifies the optimal market basket associated with each possible price of a good,
holding constant all other determinants of demand.
Price discrimination The practice of charging different prices for the same product when there is no cost difference to the pro-
ducer in supplying the product.
Price dispersion A range of prices for the same product, usu- ally as a result of customers’ lacking price information.
Price elasticity of demand A measure of how sensitive quan- tity demanded is to a change in a product’s price.
Price elasticity of supply A measure of the responsiveness of the quantity supplied of a commodity to a change in the com-
modity’s own price.
Price floor A legislated minimum price for a good.
Price maker A monopoly that supplies the total market and can choose any price along the market demand curve that it wants.
Price taker Firms or consumers who cannot affect the prevail- ing price through their respective production and consumption
decisions.
Price theory Another term for microeconomics.
Principal–agent problem A situation in which a principal’s agent (such as the manager of a firm representing the firm’s
Glossary 541
BGLOSS.INDD 12:14:36:PM 08/06/2014 PAGE 541Trim Size: 203.2 mm X 254 mm
stockholders) may have an incentive that is not entirely conso-
nant with the interests of the principal.
Prisoner’s dilemma The most famous game theory model, in which self-interest on the part of each player leads to a result
in which all players are worse off than they could be if different
choices were made.
Producer surplus Gains to producers from the sale of output to consumers, arising from price exceeding the minimum neces-
sary to compensate the seller.
Product differentiation A means by which consumers may perceive the product sold by an incumbent firm to be superior to
that offered by prospective rivals.
Production function A relationship between inputs and output that identifies the maximum output that can be produced per
time period by each specific combination of inputs.
Production possibility frontier (PPF) A depiction of all the different combinations of goods that a rational actor with cer-
tain personal goals can attain with a fixed amount of resources.
Profit maximization The assumption that firms select an output level so as to maximize profit.
Public goods Goods that benefit all consumers, such as national defense.
Quotas Government-imposed maximum quantities of goods.
Rational behavior The behavior of market participants based on a careful, deliberative process that weighs expected benefits
and costs.
Reaction curve A relationship showing one firm’s most profit- able output as a function of the output chosen by other firms.
Real price The nominal price adjusted for the changing value of money.
Regression analysis (econometrics) A statistical method that allows one to estimate, among other things, the sensitivity of the
quantity demanded of a good to determinants such as price and
income.
Regulatory barriers Barriers to entry created by the govern- ment through vehicles such as patents, copyrights, franchises,
and licenses.
Rent control Price ceilings applied to rental housing units.
Repeated-game model A game theory model in which the “game” is played more than once.
Resale Arbitrage of the product among market segments.
Residual demand curve A firm’s demand curve based on the assumption that the firm knows how much output rivals will
produce for each output the firm may choose.
Risk averse A state of preferring a certain return to an uncertain prospect that generates the same expected return.
Risk loving A state of deriving less utility from a certain return than from an uncertain prospect generating the same expected return.
Risk neutral A state of deriving the same utility from a cer- tain return as from an uncertain prospect generating the same
expected return.
Scarce resources Insufficient time, money or other resources for individuals to satisfy all their desires.
Search costs The costs that customers incur in acquiring information.
Second-degree price discrimination or block pricing The use of a schedule of prices such that the price per unit declines with
the quantity purchased by a particular consumer.
Shift of a demand curve A change in the demand curve itself that occurs with a change in factors besides price such as
income, the price of related goods, and preferences and affects
the quantity demanded at each possible price.
Shift of a supply curve A change in the supply curve itself that occurs when the other factors, besides price, that affect output
change.
Shortage Excess demand for a good.
Short run A period of time in which changing the employment levels of some inputs is impractical.
Short-run firm supply curve A graph of the systematic rela- tionship between a product’s price and a firm’s most profitable
output level.
Short-run industry supply curve The horizontal sum of the individual firms’ marginal cost curves.
Shutdown point The minimum level of average variable cost below which the firm will cease operations.
Snob effect A negative network effect.
Social marginal benefit curve The demand curve for a public good.
Spillover effect A change in equilibrium in one market that affects other markets.
Stackelberg model A model of oligopoly in which a leader firm selects its output first, taking the reactions of follower firms
into account.
Static analysis A form of economic analysis that looks at the efficiency of a market at any one point in time.
Strategies In game theory, the possible choices of the players.
Substitutes Goods that can replace one another in consumption.
Substitution effect An incentive to increase consumption of a good whose price falls, at the expense of other, now relatively
more expensive, goods.
Substitution effect of an input price change The change in input employment when output is held constant and one
input is substituted for another in response to an input price
change.
Sunk costs Costs that have already been incurred and are beyond recovery.
542 Glossary
BGLOSS.INDD 12:14:36:PM 08/06/2014 PAGE 542Trim Size: 203.2 mm X 254 mm
Surplus Excess supply of a good.
Survivor principle The observation that in competitive mar- kets, firms that do not approximate profit-maximizing behavior
fail, and that survivors are those firms that, intentionally or not,
make the appropriate profit-maximizing decisions.
Tastes or preferences The feelings of consumers about the desirability of different goods.
Technologically efficient A condition in which the firm pro- duces the maximum output from any given combination of
labor and capital inputs.
Third-degree price discrimination or market segmenta- tion A situation in which each consumer faces a single price and can purchase as much as desired at that price, but the price
differs among categories of consumers.
Tit-for-tat A strategy in which each player mimics the action (e.g., cheat, comply) taken by the other player in the preceding
period.
Total benefit The total value a consumer derives from a par- ticular amount of a good and thus the maximum amount the
consumer would be willing to pay for that amount of the good.
Total cost (TC) The sum of total fixed and total variable cost at each output level.
Total fixed cost (TFC) The cost incurred by the firm that does not depend on how much output it produces.
Total product The total output of the firm.
Total profit (�) The difference between total revenue and total cost. Total revenue (TR) Price times the quantity sold.
Total surplus The sum of producer and consumer surplus.
Total utility Assuming that it is measurable, the total satisfac- tion a consumer receives from a given level of consumption.
Total variable cost (TVC) The cost incurred by the firm that depends on how much output it produces.
Two-part tariff A form of second-degree price discrimination in which a firm charges consumers a fixed fee per time period
for the right to purchase the product at a uniform per-unit price.
Unit elastic The situation in which price elasticity of demand equals 1.0 or unity.
Variable inputs All inputs in the long run.
Vertical summation (of demand curves) The derivation of a social marginal benefit curve through the summing of consum-
ers’ marginal benefit curves.
Voucher program A form of subsidy in which parents receive vouchers that can be used to purchase education at any school
they choose.
Welfare cost or deadweight loss A measure of the aggregate loss in well-being of participants in a market resulting from an
inefficient output level.
Welfare frontier A curve that separates welfare levels that are attainable from those that cannot be reached given the available
resources.
Zero economic profit The point at which total profit is zero since price equals the average cost of production.
543
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 543Trim Size: 203.2 mm X 254 mm
Ability differences, wage differentiation
and, 438
Absolute cost advantage, monopoly
and, 290
Accounting costs, 7–8
Adjustment to changes in demand or
supply, 22–25, 22f
Adverse selection, 373–375
Advertising, 379–382
artificial product differentiation and,
379
cigarettes and, 502
firm’s demand curve and, 380, 380f
full price of product and, 381
as information, 379–380
Internet advertising, 382
market efficiency and, 381
monopoly and, 291
product prices and qualities and,
380–381
African Americans, wage discrimination
and, 474–475
Aggregate demand, 88
Aggregate labor supply curve, 431–432
Airline regulation and deregulation,
257–262
after deregulation, 298–299
contestability of airline markets,
299–300
pilot salaries and, 410
profits and, 258–259
push for reregulation and, 261–262
Airlines, yield management by, 318
Altruistic preferences, 66–69, 67f
Amazon, 391
American Airlines, 194, 300
American Cancer Society, 502
American Red Cross, 375
American vs. European work hours, 432 Anheuser-Busch, 205
Antitrust laws, monopoly and, 299–301
AOL, 208
Apple, 391
Arc elasticity formula, 33
Arithmetical relationships, returns to
scale and, 172
Artificial product differentiation, 379
Asymmetric information, 370–373
imperfect information and, 370
jobs market and, 373
lemon’s model, 370–371
market responses to, 371–372
relevance of lemon’s model, 372
signal and, 373
used car markets and, 372–373
AT&T, 299, 394
Automobile imports, 273
Average cost, short-run cost curves and,
187–188
Average-cost pricing, 394
Average fixed cost, 184
Average input cost, 420
Average product, 161–162
Average product curves, relationship
with marginal curves, 162–163
Average profit per unit, short-run profit
maximization and, 218
Average revenue, demand curves for
competitive firms and, 217
Average total cost, 184
Average variable cost, 184
Ayres, Ian, 311, 520
Baby boomers, 467
Backward-bending supply curves,
429–430
Bandwagon effect, 97–98, 97f
Barnes & Noble, 391
Barriers to entry, monopoly and,
290–291
Battalio, Raymond, 349
Beane, Billy, 198
Becker, Gary, 30
Beer industry, minimum efficiency
scales and, 205
Bittlingmayer, George, 301
Black markets, 28–29
Bliss, Michael, 29
Block pricing, 309, 310f
Boeing Aircraft, 137
Borders, 391
Borrowing. See Saving and borrowing Borrowing-lending equilibrium,
444–446, 445f
Braniff Airlines, 300
British Petroleum (BP), 583
Brynjolfsson, Erik, 347
Budget constraints, 52–56
Budget line, 52–56, 55f
geometry of, 53–54
income changes, 54–55, 54f
price changes, 55–56, 55f
shifts in, 54–56
Building demolition, 197
Bush, George W., 238, 270
Cable television
consumer surplus and, 93
demand elasticity and, 33
monopoly and, 291
regression analysis and, 102–103,
102t
Caffeine intake, 166
Calfee, John, 502
Campus violence, 78–79
Canadian prescription drugs, 312–313
Car dealerships, price discrimination
and, 311
Cartels, 345–354
cartelization of competitive industry,
346–347
cheating by, 363–365, 363f
difficulty of reaching agreements and,
348–349
incentive to cheat and, 347–348
National Collegiate Athletic Associa-
tion (NCAA) example, 467–472
OPEC example, 351–354
profits attracting entry and, 349
Index
544 Index
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 544Trim Size: 203.2 mm X 254 mm
Cell phone use by drivers, 517
Centralized planning, economic
efficiency and, 499–500
Children’s Scholarship Fund, 120
Cigarette producers, 502
Cigarette smoking deterrence, 502
City taxicab markets, 262–265
Clayton Act, 299
Coase, Ronald, 519
Coase theorem, 519–520
Cobb-Douglas production function,
176–177
Cole, Harold, 501
College education, investing in, 438
Collusion, oligopolies and, 350
Commencement ticket allocation, 65–66
Commitment, 398–399
Communication technologies, network
effects and, 99–100
Compensating wage differentials,
436–437
Competitive equilibrium and efficient
distribution, 150–154
Competitive exchange, 151–152, 151f
Competitive firms, input demand curves
and, 403–409
Competitive industry, cartelization of,
346–347
Competitive markets, 497–500
Competitive model. See also Perfectly competitive markets
airline regulation and deregulation,
257–262
city taxicab markets, 262–265
evaluation of gains and losses,
243–249
excise taxation, 249–257
net gains from trade, 266–270
quantity controls, 270–273
Complements, 16
Composite-good convention, 58–59, 59f
Computer chips, 273
Conditions of supply of inputs, 20
Congestion and airline safety, 262
Congestion costs, 514
Congressional prisoner’s dilemma, 362
Constant-cost industry, long-run
industry supply curve and,
231–233, 232f
Constant per-unit opportunity costs,
10–11
Constant purchasing power, 5
Constant returns to scale, 171
Consumer choice
budget constraints, 52–56
composite-good convention and,
58–59, 59f
consumer preferences, 43–51
consumer selfishness, 66–69
corner solution and, 57, 58f
income and consumption choice
changes, 60–66
marginal benefit and, 56
marginal cost and, 57
optimal consumption choice, 56, 57f
utility approach to consumer choice,
69–72
Consumer choice theory
excise subsidies, health care, and
consumer welfare, 108–111
investor choice, 130–137
ObamaCare, 112–116
paying for garbage, 121–124
public schools and voucher proposals,
116–120
saving and borrowing and, 124–130
Consumer preferences, 43–51
curvature of indifference curves,
45–48
economic “bads,” 44
economic “goods,” 44
graphed as indifference curves,
44–45, 44f
graphing economic bads and
economic neuters, 49–50
indifference maps, 45, 45f, 49f, 50f
individuals and different preferences,
48–49
perfect substitutes and complements,
51
Consumer price index (CPI), 4
Consumer surplus, 89–95, 110–111,
110f
free TV and, 93
gains and losses evaluation and, 244
increase in with lower price, 92, 92f
indifference curves and, 93–94
marginal benefit and, 90
total benefit and, 90
uses of, 92–93
Consumer-to-consumer (C2C) e-com-
merce, 145–146
Consumer’s demand curve, 77–80
Contestable markets, 260–261
Contract curve, 147, 148f
Convex indifference curves, 58
Cooperation evolution, 368
Corn demand, 238
Corner solution, 57, 58f
Cost curves
controlling pollution and, 205–207,
206f
input price changes and, 196–198,
196f
long-run, 199–201, 199f
market structure and, 203–205, 204f
minimum efficient scale and, 203–205
short run, 185–190, 186f
Cost function estimation, 209, 210f
Cost minimization, 193–195, 198
Cost relationships, 184–185, 185f
Cournot model, 336–340, 337f
Cournot-Nash equilibrium, 360
Craigslist, 382
Crandall, Robert, 194, 300
Crooker, John, 7
Cross-country labor supply
differences, 432
Cross-price elasticity of demand, 37
Cubic total cost function, 209
Deadweight loss
consumer surplus and, 110
excise taxation and, 254–257, 255f
monopolistic competition and,
333–335, 334f
monopoly and, 295–296, 296f,
385–389
of a price ceiling, 247–249, 248f
of rent control, 256–257
unions and, 444
Decreasing-cost industry, long-run
industry supply curve and, 231,
235–236
Decreasing returns to scale, 171
Deere, Donald, 459
Demand, monopolist’s, 278–280
Demand and supply
demand or supply adjustments, 22–25
elasticities, 30–38
equilibrium price and quantity
determination, 21–22
price controls, 25–30
supply of inputs, 413–415, 413f
Demand and supply curves, 14–20
average revenue and, 217
Index 545
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 545Trim Size: 203.2 mm X 254 mm
backward-bending, 429–431
competitive firms and, 217–218,
218f
competitive industry’s for an input,
408–409
complements and, 16
demand curve, 14–16, 14f
determinants of demand other than
price, 16–17
inferior goods and, 16
law of demand, 14
law of supply, 19
marginal revenue and, 217
market demand curve for an input,
411
normal goods and, 16
price elasticity of industry’s for input,
409–410
price takers and, 217
shifts in vs. movements along demand curve, 17–19, 17f
shifts in vs. movements along supply curve, 20
substitutes and, 16
supply curve, 19–20, 19f
tastes or preferences and, 16
Demand elasticities
estimation of, 35–36, 35t
variation among goods, 34–36
Demand estimation, 100–104
experimentation and, 100
regression analysis and, 101–104
surveys and, 101
Dentists, work effort choices of, 431
Deregulation, airline, 257–262
Derived demand, 409
Digital readers, 391
Diminishing marginal rate of
substitution, 47
Diminishing marginal returns, 165–166
Diminishing marginal utility, 70
Discrimination in employment,
472–475, 473f
Diseconomies of scale, 200
Diseconomies of scope, 208
Disemployment effect, 454, 459
Disequilibrium, 21
Disney theme parks, 326
Diversification, minimizing risk expo-
sure and, 136
Do-not-call registry, 520
Dominant firm model, 342–345, 343f
Dominant-strategy equilibrium, 359,
359f
Duopoly, 337
Dynamic analysis, monopoly and,
296–298, 297f
E-commerce, growth of consumer-to-
consumer, 145–146
eBay, 145–146
Econometrics, 101–104, 208
Economic “bads,” 44, 49–50
Economic cost, 6–8
Economic efficiency. See also Exchange, efficiency, and prices
centralized planning and, 499–500
competitive markets and, 497–500
concept of, 141
conditions for, 488
definition of, 486
efficiency in output, 492–497
efficiency in production, 489–492
externalities/public goods and, 503
as goal for economic performance,
487–488
imperfect information and, 502
inefficiency causes, 500–501
market power and, 500–501
production possibility frontier and,
492–497
role of information and, 498–499
welfare frontier and, 486–487, 486f
Economic “goods,” 44
Economic neuters, 49–50
Economic rent, 439–440, 439f, 440f,
442
Economies of scale, 200, 202, 202f, 290
Economies of scope, 208
Edgeworth exchange box, 142–143,
143f
Edgeworth exchange box with indiffer-
ence curves, 143–146
Edgeworth production box, 489–491,
489f
Education and training, 437–438
Efficiency, externalities and, 512–513
Efficiency in distribution of goods,
147–150
competitive equilibrium and, 150–154
contract curve and, 147
efficiency and equity, 149–150
inefficiency and, 148
Pareto optimality and, 147
price and nonprice rationing and,
154–156
Efficiency in output, 247
Efficiency wage, 458–459
Efficient allocation, 152
Efficient distribution, competitive
equilibrium and, 150–154
eHarmony.com, 382
Elasticities, 30–38
cross-price elasticity of demand, 37
demand elasticities vary among
goods, 34–36
excise taxation and, 252–253, 253f
income elasticity of demand, 36
monopoly and, 282–288
price elasticity of demand, 30–34, 32f
price elasticity of supply, 37–38
Ellison, Larry, 193
Empirical estimation, production
functions and, 174–177
Employment and pricing of inputs
industry and market demand curves
for an input, 408–411
industry determination of price and
employment of inputs, 413–416
input demand and employment by an
output market monopoly, 418–420
input demand curve of competitive
firm, 403–408
input price determination in a multi-
industry market, 416–418
monopsony in input markets,
420–422
supply of inputs, 411–413
Employment discrimination, 472–475,
473f
Endowment changes, saving and
borrowing and, 125–127
Endowment point, 125
Entry fees, two-part tariffs and, 322
Entry possibilities, monopoly and,
293–294, 294f
Epstein, Richard, 30
Equalization of rates of return, 450
Equilibrium determination, game theory
and, 358–360, 359f
Equilibrium price and quantity
determination, 21–22, 21f
disequilibrium and, 21
shortage and, 21
surplus and, 22
Equity concept, 150
546 Index
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 546Trim Size: 203.2 mm X 254 mm
Evolution of cooperation, 368
Exam performance, 166
Excess burden, excise taxation and, 256
Exchange, efficiency, and prices
competitive equilibrium and efficiency
distribution, 150–154
efficiency in distribution of goods,
147–150
price and nonprice rationing and ef-
ficiency, 154–156
two-person exchange, 141–146
Excise subsidies, 108–109, 108f
Excise taxation, 249–257
burden of, 252
consumer responsiveness and, 254
deadweight loss and, 254–257, 255f
elasticity of demand and, 253–254
elasticity of supply and, 252–253, 253f
excess burden and, 256
gasoline tax-plus-rebate program, 83
long-run effects of, 251–252
short-run effects of, 250–251
Exclusive franchising, 291
Expansion path, 194
Expected return, investor choice and,
133
Expected utility, 133
Experimentation, demand estimation
and, 100
Explicit costs, 6, 182
Exponents, Cobb-Douglas production
functions and, 177
External benefits, 512, 516–517
External costs, 512–515
Externalities
efficiency and, 512–513
external benefits, 512, 516–517
external costs, 512–515
property rights and, 518–520
ExxonMobil, 249
Fairness concept, 150
Federal Trade Commission, 502
Federal Trade Commission Act, 299
Feedback effect, 484
Fenn, Aju, 7
Finance professors, bidding war for, 236
Financial aid offers, college applications
and, 316
Financial services
managerial incentives and shareholder
interests and, 216
First-degree price discrimination,
307–309
Fixed inputs, production and, 161
Fixed vs. sunk costs, 183 Food stamp program, 64–65, 64f
Ford, 299
Ford, Henry, 459
Foreign producer consequences, quotas
and, 273
Forstmann, Theodore, 120
Freakonomics, 382 Free entry and exit, perfect competition
and, 214
Free television, 93
Frequent-buyer programs, 310
Friedman, Milton, 119, 269
Full price of product, advertising and,
381
Functional forms, production,
174–177
Gains and losses evaluation
competitive model and, 243–249
consumer surplus and, 244
deadweight loss of a price ceiling
and, 247–249, 248f
efficiency in output and, 247
producer surplus and, 244–245, 244f
total surplus and, 245–247, 246f
Game theory
commitment and, 398–399
determination of equilibrium,
358–360
dominant-strategy equilibrium and,
359, 359f
iterated dominance and, 396–398
Nash equilibrium, 360–361
payoff matrix and, 358–359
payoffs and, 358
players and, 358
prisoner’s dilemma game, 361–366
repeated games, 366–369
strategies and, 358
Garbage disposal, 121–124, 121f
bag system and, 122, 123f
trash pricing and recycling, 123
Gasoline consumption, speed limits and,
170–171, 171f
Gasoline rationing, 154–156, 154f
Gasoline tax-plus-rebate program,
83–85, 84f
Gay.com, 382
General equilibrium analysis
competitive markets and economic
efficiency, 497–500
economic efficiency and, 486–488,
500–503
efficiency in output, 492–497
efficiency in production, 489–492
feedback effect and, 484
market interdependence and, 483–485
Pareto optimal and, 485–486
partial and general analysis compared,
546–549
production possibility frontier and,
492–497
spillover effect and, 484
uses of, 485–486
General Motors, 299
Geometry of cost curves, 189–190, 190f
G.I. bill, 120
Giffen good, 87–88
Glaxo, 313
Goal-oriented behavior, 5–6
Golden rule of cost minimization,
193–194, 198
Golec, Joseph, 303
Government interventions
monopolistic competition and, 335
price controls and, 25–30
quantity controls and, 270–273
Grade point average (GPA), 164
Gray market purchases, 312–313
Great Depression, 501
Gross marginal productivity, 446
Groupon, 310
Gulf oil disaster, 515
Hayek, Friedrich, 499
Hazlett, Tom, 301
Health care. See also ObamaCare benefits of improved, 94–95
moral hazard problem and, 376
price controls and, 29
price sensitivity and, 111
Homogenous products, perfect competi-
tion and, 214–215
Hong Kong Hilton, 9
Human capital investment, 437–438
IBM, 299
Illegal transportation services, 264
Immigration
benefits and costs of, 476–480, 476f
Index 547
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 547Trim Size: 203.2 mm X 254 mm
government budgets and, 479
skill levels and, 479–480
unskilled labor supply and, 477
wage effects and, 478–479
Imperfect information
asymmetric information and, 370
economic inefficiency and, 502
Implicit costs, 6–7, 182
Income and substitution effects of price
change, 80–85, 81f
excise tax and, 83
gasoline tax-plus-rebate program,
83–85, 84f
inferior goods and, 85–88, 86f
normal-good case, 81–83
Income changes, budget line and,
54–55, 54f
Income changes and consumption
choices, 60–69, 61f
food stamp program and, 64–65, 64f
income-consumption curve, 60
inferior goods, 62–63, 63f
normal goods and, 60–62
Income elasticity of demand, 36
Income-leisure choice of the worker,
425–428, 426f
Increasing-cost industry, long-run
industry supply curve and, 231,
233–234, 234f
Increasing per-unit opportunity costs,
10–11
Increasing returns to scale, 171
Indifference curves
consumer choice and, 71–72
consumer preferences graphed as,
44–45, 44f
consumer surplus and, 93–94, 94f
curvature of, 45–48
diminishing MRS and, 47–48
Edgeworth exchange box with,
143–146
marginal rate of substitution and, 46,
47f
Indifference maps, 45, 45f, 49f, 50f
Individual and market demand
consumer surplus, 89–95
demand estimation basics, 100–104
income and substitution effects of
price change, 80–85
from individual to market demand,
88–89
inferior goods and, 85–88
network effects, 96–100
price changes and consumption
choices, 76–80
price elasticity and price-consumption
curve, 95–96
Industry and market demand curves for
input, 408–411
competitive industry’s demand curve,
408–409
derived demand and, 409
elasticity of industry’s demand curve,
409–410
market demand curve for an input,
411
Industry determination of price and
employment of inputs, 413–416, 413f
Inefficiency and goods allocation, 148
Inelastic, defined, 31
Inelastic demand, monopoly and,
287–288
Inferior goods
defined, 16
income and substitution effects and,
85–88, 86f
income changes and consumption of,
62–63, 63f
Information, advertising as, 379–380
Information role in economic efficiency,
498–499, 502
Input combinations, long-run cost of
production and, 192
Input demand and employment by an
output market monopoly, 418–420
Input demand curve for competitive
firms, 403–408
all inputs variable, 405–406, 405f
alternative approach, 406–408, 407f
marginal value product and, 403–405
one variable input, 403–405, 404f
output effect of an input price change,
407
substitution effect of input price
change, 407
Input market analysis
employment discrimination and,
472–475
immigration benefits and costs and,
476–480
minimum wage and, 454–459
NCAA cartel and, 466–472
social security and, 460–462
Input markets
borrowing, lending, and interest rates
and, 444–446
economic rent and, 439–440
general equilibrium in, 491–492
general level of wage rates and,
433–435
income-leisure choice of the worker,
425–428
industry and market demand curves
for, 408–411
industry determination of price and
employment of inputs, 413–416
input demand and employment by an
output market monopoly, 418–420
input demand curve of competitive
firm, 403–408
input price determination in multi-
industry market, 416–418
interest rate differentiation and,
450–451
investment and marginal productivity
of capital and, 446–448
monopoly power in, 441–444
monopsony in, 420–422
process of input price equalization
across industries, 414–416
savings, investment, and interest rate
and, 448–450
supply of hours of work and,
428–432
supply of inputs, 411–413
unions and, 441–444
wage differentiation and, 435–438
Input price changes
cost curves and, 196–198, 196f
output effect of, 407
output response to change in,
224–225, 225f
substitution effect of, 407
Input price determination in multi-
industry market, 416–418, 417f
Input price equalization across
industries, 414–416, 415f
Input substitution effect, 197
Insurance, minimizing risk exposure
and, 135–136
Intel, 299
Interest rates
borrowing and lending and, 444–446,
445f
differentiation in, 450–451
investment and, 448–450
548 Index
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 548Trim Size: 203.2 mm X 254 mm
Interest rates (continued) loan administration costs and, 451
loan duration and, 451
risk differences and, 451
saving and borrowing and, 128–130,
129f
tax treatment differences and, 451
International trade, 266–270, 266f
gains from, 268–269, 268f
link between imports and exports and,
269–270
political and economic views on,
269–270
production possibility frontier and,
495–497
steel tariffs and, 270
Internet, cartelization and, 347
Internet advertising, 382
Intertemporal price discrimination,
316–318, 317f
Inventions
monopolies and, 389–391, 390f
output effects of, 390–391
Investment. See also Savings and bor- rowing
demand curve, 447–448, 447f
interest rates and, 448–450
marginal productivity of capital and,
446–448
Investor choice, 130–137
diversification and, 136–137
entrepreneurs and risk-return,
131–132
expected return and, 133
expected utility and, 133
insurance and, 135–136
minimizing risk exposure, 135–137
return-risk tradeoff, 130–131, 131f
risk aversion, 132–134
risk loving behavior, 134
risk neutral behavior, 134
Isocost lines, 191–192, 191f. See also Long-run cost of production
Isoquants, production, 167–168,
167f
Iterated dominance, 396–398
Jack in the Box, 285
Jackson, Henry “Scoop,” 137
James, LeBron, 7
Jet Blue, 261
Job market signaling, 373
Katz, Lawrence, 459
Kessler, Jeffrey, 292
Kindle digital reader, 391
Kodak, 391
Kraft Foods, 48–49, 273
Kreuger, Alan, 459
Labor division, returns to scale and,
171–172
Labor force size, wage rates and,
434–435
Labor markets
average wage differences in, 474–475
discrimination in, 472–475, 473f
Social Security effects on, 465–466,
466f
Labor specialization, returns to scale
and, 171–173
Labor unions. See Unions Lange, Oscar, 499
Large-scale technologies, returns to
scale and, 172–173
Law of diminishing marginal returns,
165–166
Learning by doing, 202–203
pioneering firms and, 203
vs. economies of scale, 202–203, 202f Least costly input combinations, long-
run cost of production and, 192
Leisure, 425–428, 426f
Lemons model, 370–373
Lerner, Abba, 499
Lerner index, 289
Liability caps, 515
Licensing, monopoly and, 291
Limited price information, 377–379
Linear forms of production functions,
175–176
Loan administration costs, interest rates
and, 451
Loan duration, interest rates and, 451
Local government budgets, unions
effect on, 443
Lockwood, Kate, 146
Long-run competitive equilibrium
perfectly competitive markets and,
227–231
profit maximization and, 227–228,
228f
zero economic profit and, 228–230
zero profit when firms’ cost curves
differ, 230–231
Long-run cost curves, 199–201, 199f
diseconomies of scale and, 200
economies of scale and, 200
learning by doing and, 202–203
long and short run revisited, 200–201,
201f
Long-run cost of production, 191–195
cost minimization and, 195
expansion path and, 194
golden rule of cost minimization and,
193, 198
interpreting the tangency points,
192–193
isocost lines, 191–192, 191f
least costly input combinations, 192
production when all inputs are
variable, 166–171
Long-run effects, excise taxation and,
251–252
Long-run industry supply curve
constant-cost industry and, 231–233
decreasing-cost industry and, 231,
235
increasing-cost industry and, 231,
233–234
perfectly competitive markets and,
231–239
technological advances and, 235, 236f
Los Angeles smog, 523–524
Loss in the short-run, 221–222, 222f
Lotteries, 155
Lump-sum transfers, 109–110, 108f
Macroeconomics, 2
Major League Baseball
cost minimization and, 198
dominant strategies in, 359
monopsony and, 421
March Madness, 292–293
Marcus, Carl, 350
Marginal-average relationships, short-
run cost curves and, 189
Marginal benefit, 56, 90
Marginal cost
consumer choice and, 57
natural monopoly and, 393, 394
short-run cost curves and, 186–187
short-run cost of production and, 184
Marginal input cost, 420
Marginal product, 162
Marginal product curves, relationship
with average curves, 162–163
Index 549
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 549Trim Size: 203.2 mm X 254 mm
Marginal productivity of capital, invest-
ment and, 446–448
Marginal products of inputs, marginal
rate of technical substitution and,
169–170
Marginal rate of substitution, 46, 47f
Marginal rate of technical substitution,
168–171
marginal products of inputs and,
169–170
speed limits and gasoline consump-
tion and, 170, 171f
Marginal rate of transformation,
492–493
Marginal returns, diminishing, 165–167
Marginal revenue curves, monopoly
and, 278–280
Marginal revenue demand curves for
competitive firms, 217–218
Marginal revenue product, 419
Marginal tax rates, 432
Marginal utility, 69, 69f
Marginal value product, 403–404
Market analysis, 4–5
Market baskets, 44
Market demand. See Individual and market demand
Market demand curve for an input,
411
Market-determined distribution and
efficiency, 152–153, 152f
Market efficiency, advertising and, 381
Market equilibrium, monopolistic
competition and, 331–333
Market interdependence, 483–485,
484f
Market outcomes, supply-demand
model and, 24–25, 24f
Market participants, basic assumptions
about, 5–6
Market power, economic inefficiency
and, 500–501
Market responses
adverse selection and, 374–375
moral hazard and, 377
Market segmentation, 309
Market structure, cost curves and,
203–205, 202f
Marriott hotel, 308
Match.com, 382
MBA education, 373
McDonald’s, 310
Medical insurance market. See Health care Medicare, 29
Merrill Lynch, 310
Microeconomic theory
basic assumptions about market
participants, 5–6
economic vs. accounting costs, 8 good theories, 2–3
market analysis, 4–5
nature and role of, 2–3
opportunity cost, 6–9
positive vs. normative analysis, 3–4 production possibility frontier, 9–11
real vs. nominal prices, 4–5 scope of, 2
sunk costs, 8–9
Microsoft, 100, 299, 300
Minimum efficient scale, 203–205, 204f
Minimum wage, 454–459, 457f
discrimination and, 455–456
disemployment effect and, 455, 459
effects on poor and, 456–458
efficiency wage and, 458–459
fringe benefits and, 456
uncovered sector and, 456
unskilled workers and, 456
work hours and, 455
Monopolistic competition
deadweight loss and, 333–334, 334f
efficiency and, 333–335
excess capacity and, 334–335
government intervention and, 335–336
market equilibrium and, 331–333
price and output under, 331–336
product differentiation and, 331
Monopoly. See also Noncompetitive market models; Price discrimina-
tion; Product pricing with monopoly
power
absolute cost advantage and, 290
antitrust laws and, 299
barriers to entry and, 290–293
deadweight loss and, 295–298,
385–389
demand and marginal revenue curves,
278–280
dynamic analysis of, 296–298, 297f
economies of scale and, 290
efficiency effects of, 295–298
elasticity of demand and, 282–285
incumbents and potential entrants
and, 293–294
inelastic demand and, 288
input markets and, 441–444
inventions and, 390–391
inverse elasticity pricing rule, 284,
284f
Lerner index and, 289
maintaining market position and, 388
marginal revenue curves and,
278–280
measuring monopoly power, 289
monopoly analysis implications,
285–288
natural monopoly, 290, 392–395
output market monopoly, 418–420
power when there are several
suppliers, 288–289, 289f
price ceiling and, 301–302
price maker and, 278
price regulation and, 299, 301–303
product differentiation and, 290
production costs and, 388
profit-maximizing output and, 280–285
public policy toward, 298–303
regulatory barriers and, 290–291
shutdown condition and, 286, 286f
sources of monopoly power, 289–290
strategic behavior by firms and,
293–294
Monopsony in input markets, 420–422,
421f
Moral hazard, 375–376
Movements along demand curve, 17–19
Movements along supply curve, 19–20
Multi-industry market, input price
determination in, 416–418, 417f
Multiplicative forms of production,
176–177
Murphy, Kevin, 94, 459
Nash equilibrium, 360–361, 360f, 397
National Collegiate Athletic
Association (NCAA), 292, 466–472
National Industrial Recovery Act
(NIRA), 501
National Labor Relations Act of, 1935,
501
Natural monopoly, 392–395
economies of scale and, 290
new product introductions and,
394–395
public ownership and, 395
regulation of, 394–395
550 Index
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 550Trim Size: 203.2 mm X 254 mm
Nature of cost, 182
Net marginal productivity, 446
Network effects, 88–98
bandwagon effect, 97–98, 98f
snob effect, 98–99, 98f
New Deal, 501
New entrant/survivor technique, 209
Nominal price, 4–5, 5t
Noncompetitive market models
iterated dominance and commitment,
396–399
monopolies and invention
suppression, 389–391
natural monopoly, 392–395
size of deadweight loss of monopoly,
385–388
Nonprice rationing and efficiency,
154–156
Nonsatiation, 43
Normal goods
definition of, 16
income changes and consumption
of, 60
Normative analysis, 3–4
Obama, Barack, 29
ObamaCare, 112–116. See also Health care
basics of, 112
budget line effects, 112–113, 113f
costs and benefits and, 114
other options and, 115–116
possible harms and, 115, 116f
possible outcomes and, 114–115
Ohanian, Lee, 501
Oil drilling, 515
Oil prices, short-run price gyrations, 38
Oligopoly, 336–340. See also Cartels cartel model and, 345–354
collusion and, 350, 368–369
Cournot model and, 336–340
dominant firm model and, 342–343
reaction curves and, 338–339
residual demand curve and, 340–341
Stackelberg model and, 340
Omidyar, Pierre, 145
Online dating, 382
Operating at loss in short run, 224, 223f
Opportunity cost, 6–9, 182
Optimal choice, consumer’s, 70–71
Oracle, 193
Ordinary least-squares, 102, 103f
Oreos in the Orient, 48–49
Organ transplants, 30
Organization of Petroleum Countries
(OPEC), 351–352
difficulty of controlling cheating and,
349
early success of, 352
historical saga of, 351
passage of time and, 352–354
Output determination in the short run,
227
Output determination with price
discrimination, 314–316, 315f
Output effect of an input price change,
407
Output efficiency, 492–497
international trade and, 495–497
marginal rate of transformation and,
492–493
Output market monopoly, 418–420
Output response to change in input
prices, 224–225, 225f
Outputs to inputs, 160
factors of production and, 160
production function and, 160
technological efficiency and, 160
Pareto optimality, 147, 485
Partial equilibrium analysis, 483–486
Patient Protection and Affordable Care
Act. See ObamaCare PAYGO Social Security. See Social
Security
Payoffs, game theory and, 358
Peak-load pricing, 319–320, 319f
Per-unit curves, short-run profit maxi-
mization and, 218–222, 221f
Perfect complements, 51, 51f
Perfect information, perfect competition
and, 214
Perfect price discrimination, 307–310
Perfect substitutes, 51, 51f
Perfectly competitive markets. See also Competitive model; Profit maxi-
mization in perfectly competitive
markets
applications of, 239–240
assumptions of, 214–215
demand curves for, 217–218, 218f
four conditions characterizing,
214–215
free entry and exit and, 214
large numbers of buyers and sellers
and, 214
long-run competitive equilibrium and,
227–231
long-run industry supply curve and,
231–239
perfect information and, 214
product homogeneity and, 214
short-run firm supply curve and,
223–225
short-run industry supply curve and,
226–227
short-run profit maximization and,
218–222
Pfizer, 312–313
Pharmaceutical markets
cross-border purchases and, 312–313
dominant firm model in, 345
Physicians, work effort choices of, 431
Pilot salaries, 410
Pioneering firms, learning by doing and,
203
Players, game theory and, 358
Point elasticity formula, 32
Pollution control, 520–524
using cost curves and, 205–207, 206f
Positive analysis, 3–4
Postal Service, 395
Poultry consumption, 80
Poverty, minimum wage effects on,
456–458
Predatory pricing, 299
Prescott, Edward, 432
Prescription drugs, cross-border
purchases and, 312–313
Price and nonprice rationing and
efficiency, 154–156
Price ceilings, 25, 30, 301–302, 302f
Price changes
budget line and, 55–56, 55f
consumer demand curve and, 77–80
consumption choices and, 76–80
demand curve slopes and, 79–80
income and substitution effects of,
80–85
price-consumption curve and, 77
Price-consumption curve, 77, 95–96,
96f
Price controls, 25–30
black markets, 28–29
health care reforms and, 29
price ceilings, 25, 30
Index 551
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 551Trim Size: 203.2 mm X 254 mm
price floors, 25
rent control, 25–27, 26f
Price determination in short run, 227
Price determination with price discrimi-
nation, 313, 314f
Price discrimination
costs of engaging in, 326
defined, 306
first-degree price discrimination,
307–308
intertemporal price discrimination,
316–318
monopoly power and, 311
peak-load pricing and, 319–322
price and output determination with,
313–316
profit and, 307, 307f
resale prevention and, 311
second-degree price discrimination,
309
third-degree price discrimination,
309–311
three necessary conditions for,
309–310
two-part tariffs and, 322–326
Price dispersion, 378–379
Price elasticity, price-consumption
curve and, 95–96, 96f
Price elasticity of demand, 30–32, 32f
arc elasticity formula, 33
calculating, 32–34
inelastic, 31
point elasticity formula, 32–33
unit elastic, 31
Price elasticity of industry’s demand
curve for input, 409–410
Price elasticity of supply, 37
Price equalization across industries,
414–416, 415f
Price floors, 25
Price maker, 278
Price regulation, monopoly and, 299,
301–303
Price takers, 150–151, 217
Price theory, 2
Prisoner’s dilemma game, 361–366,
362f
Private sector unions, 443
Privatization, productivity and, 195,
196f
Process of input price equalization
across industries, 414–416, 415f
Producer surplus, gains and losses
evaluation and, 244–245, 244f
Product curves
geometry of, 163–165, 165f
total, average, and marginal, 162, 163f
Product differentiation, monopolistic
competition and, 290, 296–297
Product homogeneity, perfect competi-
tion and, 214
Product pricing with monopoly power.
See also Monopoly; Price discrimi- nation
intertemporal price discrimination
and, 316–322
necessary conditions for price
discrimination, 311–313
peak-load pricing and, 319–320
price and output determination with
price discrimination, 313–316
price discrimination and, 307–311
two-part tariffs and, 322–326
Production
factors of, 160
functional forms and empirical
estimation of production functions,
174–177
production function, 160
production when all inputs are
variable: long run, 166–171
production when only one input is
variable: short run, 160–166
relating output to inputs, 160
returns to scale, 171–174
Production contract curve, 491
Production costs
controlling pollution and, 205–207
cost curves importance to market
structure, 203–205
economies of scope, 208
estimating cost functions, 209–210
input price changes and cost curves,
196–198
learning by doing, 202–203
long-run cost curves, 199–201
long-run cost of production, 191–195
monopoly and, 388
nature of cost, 182
short-run cost curves, 185–190
short-run cost of production, 182–185
Production efficiency, 489–492
Edgeworth production box and,
489–491
general equilibrium in input markets,
491–492
production contract curve and, 491
Production functions
Cobb-Douglas production function,
176–177
functional forms and empirical
estimation of, 174–177
linear forms of, 175
multiplicative forms of, 176–177
Production isoquants, 158–178, 167f.
See also Marginal rate of technical substitution
downward sloping, 168
higher output levels for isoquants
further to the northeast, 168
non-intersecting, 168
typically converse to the origin, 168
Production possibility frontier (PPF),
9–11, 9f, 12f, 492–497, 493f
Production when all inputs are variable,
166–171
marginal rate of technical
substitution, 168–169
production isoquants, 167–168, 167f
Production when only one input is
variable, 160–166, 161f
average product and, 161–162
fixed inputs and, 161
geometry of product curves, 163–165,
165f
law of diminishing marginal returns
and, 165–166
marginal product and, 163
relationship between average and
marginal product curves, 162–163
total, average, and marginal product
curves, 162, 163f
total product and, 161
Productivity, privatization and, 195, 196f
Profit maximization in perfectly
competitive markets
assumptions of perfect competition,
214–215
competitive model applications,
237–240
demand curve for a competitive firm,
217–218
long-run competitive equilibrium,
227–231
long-run industry supply curve,
231–239
552 Index
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 552Trim Size: 203.2 mm X 254 mm
Profit maximization in perfectly
competitive markets (continued) profit maximization, 215–217
short-run industry supply curve,
226–227
short-run profit maximization,
218–222
short-run supply curve, 223–225
survivor principle and, 216
Profit-maximizing output of monopoly,
280–285, 281f
elasticity of demand and, 282–284
graphical analysis and, 281–282
inverse elasticity pricing rule, 284,
284f
total and per-unit curves, 281–282,
281f
Property rights, externalities and,
518–520
Public goods and externalities. See also Externalities
controlling pollution and, 520–524
economic inefficiency and, 503
efficiency in provision of public
goods, 508–512
externalities, 512–520
free-rider problem and, 507
free riding and group size,
507–508
nonexclusion and, 506
nonrival in consumption and, 506
patents and, 510–511
production and distribution efficiency
and, 510–511
social marginal benefit curve and,
508–509
vertical summation of demand curves
and, 509
Public ownership, regulating natural
monopoly and, 395
Public schools, voucher proposal and,
116–120
Public sector unions, 443
Putnam, Howard, 300
Quadratic total cost function, 209
Quantity controls, 270–273
quotas and their foreign producer
consequences, 273
sugar policy and, 270–273
Quantity determination, 21–22
Queuing psychology, 135
Quotas, 270–273
Rates of return equalization, 450
Ration coupons, 155, 156
Rational behavior, 6
Rationing, 154–156
Rationing by waiting, 155–156
Reaction curves, 338–339, 339f
Reagan, Ronald, 299
Real price, 4–5, 5t
Red meat consumption, 80
Regression analysis, demand estimation
and, 102–104
Regulatory barriers, monopoly and, 291
Rent control, 25–28, 26f, 256–257
Repeated games, 366–369
Reregulation, airline, 261–262
Resale prevention, price discrimination
and, 311–312
Reserve clause in major league
baseball, 421
Residual demand curve, 340–342
Resnick, Paul, 146
Retail productivity, 188–189
Returns to scale, 171–174
arithmetical relationships and, 172
constant returns to scale, 171
decreasing returns, 173–174
decreasing returns to scale, 171
division and specialization of labor
and, 171–172
increasing returns, 171–173
increasing returns to scale, 171
large-scale technologies and,
172–173
Risk aversion, 132–134
Risk differences, interest rates
and, 451
Risk loving behavior, 134
Risk neutral behavior, 134
Riyadh Pact, 38
Roberts, Russell, 362
Rolex “cartel,” 349–350
Roosevelt, Franklin, 501
Sarbanes-Oxley Act of 2002, 8
Saving and borrowing
consumer’s choice and, 124–130,
124f
endowment changes and, 125–127
endowment point and, 125
interest rate changes and, 128–130,
129f
social security and, 127–128, 127f
Savings, interest rates and, 448–450
Scale, returns to, 171–174
Scarce resources, 6
Search costs, 378
Search intensity, price dispersion and,
378–379
Second-degree price discrimination,
309, 310f
Selfishness, consumer, 66–69
Sherman Act, 299
Shifts of demand curve, 16, 14f
Shifts of supply curve, 16–17
Short-run cost curves, 185–190,
186f
average cost, 187–188
geometry of cost curves, 189–190,
190f
marginal-average relationships,
189
marginal cost, 186–187
Short-run cost of production, 182–185
average total cost and, 184
average variable cost and, 184
averaged fixed cost and, 184
behind cost relationships, 184–185
fixed vs. sunk costs, 183 marginal cost and, 184
total cost and, 184
total fixed cost and, 183
total variable cost and, 183
Short-run effects, excise tax and,
250–251
Short-run firm supply curve
output response to a change in input
prices, 224–225, 225f
perfectly competitive firms and,
223–225, 223f
shutdown point and, 224
Short-run industry supply curve
perfectly competitive markets and,
226–227, 226f
price and output determination and,
227
Short-run production when only one
input is variable, 160–166
Short-run profit maximization
average profit per unit and, 219
Index 553
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 553Trim Size: 203.2 mm X 254 mm
operating at a loss in the short run,
221–222, 222f
perfectly competitive markets and,
218–221
total profit and, 218–219
total revenue and, 218
using per-unit curves, 220–221, 221f
Shortage, 21
Shutdown condition, monopoly and,
285–286, 286f
Shutdown point, 224
Signaling, 373
Singapore traffic control, 515
Smith, Adam, 172
Smith, Michael, 347
Snob effect, 98–99, 98f
Social Security, 460–466
baby boomers and, 467
employer and employee portions and,
460–461, 460f
fair share issue and, 467
government spending and, 467
hidden cost of, 463–464
labor market effects of, 465–466,
465f
long-run effects of, 465, 465f
savings and, 127–128, 127f
tax burden of, 461–462
Soft-drink producers, minimum efficient
scales and, 205
Sony, 391
Southwest Airlines, 261
Specialization, labor, returns to scale
and, 171–173
Speed limits, gasoline consumption and,
170–171, 171f
Spence, Michael, 373
Spillover effect, 483
Stackelberg model, 340–342, 341f
State budgets, unions effect on, 443
State of technology knowledge, 17
Static analysis, 296
Steel tariffs, 270
Strategies, game theory and, 358
Subletting, 28
Subprime home mortgages, 376
Subsidies, external benefits and,
516–517, 516f
Substitutes, 16
Substitution effect of input price
change, 407
Substitution effects of price change,
80–85, 81f
Sugar policy, 271–273, 271f
Sunk costs, 8–9
Supply curves. See Demand and supply curves
Surplus, 22
Surveys, demand estimation and, 101
Survivor principle, 216
Swensen, John, 146
Tangency points interpretation,
192–194
Tastes or preferences, 16
Taxation
external costs and, 513–514, 513f
interest rates and, 451
pollution and, 520–523, 522f
Taxicab licensing, 262–263, 263f
Taxicab markets, 262–264
Technological efficiency, production
function and, 160
Telemarketing, 520
Telephone services, peak-load pricing
and, 319–320
Television regulation, 93
Thatcher, Margaret, 195
Third-degree price discrimination, 309,
311
Ticket scalping, 153
Time Warner, 208
Tit-for-tat strategy, 366–367
Toll roads, 321
Topel, Robert, 94
Total, average, and marginal product
curves, 162, 163f
Total benefit, 90
Total cost, 184
Total fixed cost, 183
Total output, 161
Total profit, short-run profit
maximization and, 218–219
Total revenue, short-run profit
maximization and, 218
Total surplus, gains and losses
evaluation and, 245–247, 248f
Total utility, 69
Total variable cost, 183
Traffic congestion, peak-load pricing
and, 321
Traffic externalities, 514–515
Traffic fatalities, 78
Training, 437
Transitivity, 43
Trash pricing and recycling, 123
Two-part tariffs, 322–326, 323f
entry fees and, 322
many consumers, different demands,
323–325, 324f
price of compared to monopoly price,
325–326, 325f
Two-person exchange, 141–146
Edgeworth exchange box, |
142–143
Edgeworth exchange box with
indifference curves, 143–146
Ueberroth, Peter, 471
Underground economy, 432
Unemployment rate, law of demand
and, 15
Unions, 441–444
deadweight loss and, 442
economic rent maximization
and, 442
state and local government budgets
and, 442–443
unionization rates, 442–443
worker productivity and, 444
U.S. Postal Service, 395
Used car markets, lemons problem in,
372–373
Utility approach to consumer choice,
69–72
consumer’s optimal choice, 70–71
diminishing marginal utility, 70
marginal utility, 69–70, 71f
relationship to indifference curves,
71–72
total utility, 69
Variable work effort, 425–428
Varian, Hal, 347
Vernon, John, 303
Voluntary export restraints, 273,
498
Voucher proposals
consumer choice theory analysis of,
119–120
demand and supply of school choice,
119–120
public schools and, 116–118
554 Index
BINDEX.INDD 09:37:56:AM 08/06/2014 PAGE 554Trim Size: 203.2 mm X 254 mm
Wage differentiation, 435–438, 436f
ability differences and, 438
compensating wage differentials and,
436–437
discrimination and, 474–475
human capital investment and,
437–438
Wage rates
backward-bending supply curves and,
429–431
general levels of, 433–435, 433f
labor force size and, 434–435
substitution and income effect and,
428–429
Walmart, 188
Walton, John, 120
Welch, Finis, 459
Welfare frontier, 486–487. 486f
Women, wage discrimination and, 475
Work-leisure choice, 428–431
Worker productivity, unions and, 444
Yahoo Singles, 382
Yield management by airlines, 318
Youth alcohol abuse, 78–79
Zeckhauser, Richard, 146
Zero economic profit, long-run competi-
tive equilibrium and, 228–231
Zhang, Ivy, 8
WILEY END USER LICENSE AGREEMENT Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.
- Cover������������
- Title Page�����������������
- Copyright
- Contents���������������
- Chapter 1: An Introduction to Microeconomics���������������������������������������������������
- 1.1 The Scope of Microeconomic Theory��������������������������������������������
- 1.2 The Nature and Role of Theory����������������������������������������
- What Is a Good Theory?�����������������������������
- 1.3 Positive versus Normative Analysis���������������������������������������������
- 1.4 Market Analysis and Real versus Nominal Prices���������������������������������������������������������
- Application 1.1 Real Versus Nominal Presidential Salaries����������������������������������������������������������������
- 1.5 Basic Assumptions about Market Participants������������������������������������������������������
- 1.6 Opportunity Cost���������������������������
- Application 1.2 Why the King Left Cleveland in 2010, and Can Benefits in Sports Be Measured?���������������������������������������������������������������������������������������������������
- Economic versus Accounting Costs���������������������������������������
- Application 1.3 The Accounting and Economic Costs of SOX���������������������������������������������������������������
- Sunk Costs�����������������
- Application 1.4 Why It Was Profitable to Demolish a Profitable Hong Kong Hotel�������������������������������������������������������������������������������������
- 1.7 Production Possibility Frontier������������������������������������������
- Constant versus Increasing per-unit Opportunity Costs������������������������������������������������������������
- Chapter 2: Supply and Demand�����������������������������������
- 2.1 Demand and Supply Curves�����������������������������������
- The Demand Curve�����������������������
- Application 2.1 The Law of Demand at Work for Non-work�������������������������������������������������������������
- Determinants of Demand Other Than Price����������������������������������������������
- Shifts in versus Movements along a Demand Curve������������������������������������������������������
- Application 2.2 The Rise and Fall of Cigarette Consumption in the United States��������������������������������������������������������������������������������������
- Application 2.3 The Occasional Interplay Between Price and Non-Price Factors in Determining Quantity Demanded��������������������������������������������������������������������������������������������������������������������
- The Supply Curve�����������������������
- Shifts in versus Movements along a Supply Curve������������������������������������������������������
- 2.2 Determination of Equilibrium Price and Quantity����������������������������������������������������������
- 2.3 Adjustment to Changes in Demand or Supply����������������������������������������������������
- Application 2.4 Why Holiday Home Prices in Switzerland are Soaring�������������������������������������������������������������������������
- Using the Supply–Demand Model to Explain Market Outcomes���������������������������������������������������������������
- 2.4 Government Intervention in Markets: Price Controls�������������������������������������������������������������
- Rent Control�������������������
- Who Loses, Who Benefits?�������������������������������
- Black Markets��������������������
- Application 2.5 Health Care Reform and Price Controls������������������������������������������������������������
- Application 2.6 Price Ceilings Can Be Deadly for Buyers��������������������������������������������������������������
- 2.5 Elasticities�����������������������
- Price Elasticity of Demand���������������������������������
- Calculating Price Elasticity of Demand���������������������������������������������
- Application 2.7 Demand Elasticity and Cable Television Pricing���������������������������������������������������������������������
- Demand Elasticities Vary among Goods�������������������������������������������
- The Estimation of Demand Elasticities��������������������������������������������
- Application 2.8 Why Canadians Are Flying South of the Border�������������������������������������������������������������������
- Three Other Elasticities�������������������������������
- Application 2.9 Price Elasticities of Supply and Demand and Short-Run Oil Price Gyrations������������������������������������������������������������������������������������������������
- Chapter 3: The Theory of Consumer Choice�����������������������������������������������
- 3.1 Consumer Preferences�������������������������������
- Consumer Preferences Graphed as Indifference Curves����������������������������������������������������������
- Curvature of Indifference Curves���������������������������������������
- Application 3.1 Diminishing MRS and Newspaper Retailing��������������������������������������������������������������
- Individuals Have Different Preferences���������������������������������������������
- Application 3.2 Oreos in the Orient������������������������������������������
- Graphing Economic Bads and Economic Neuters��������������������������������������������������
- Perfect Substitutes and Complements������������������������������������������
- 3.2 The Budget Constraint��������������������������������
- Geometry of the Budget Line����������������������������������
- Shifts in Budget Lines�����������������������������
- 3.3 The Consumer’s Choice��������������������������������
- A Corner Solution������������������������
- The Composite-Good Convention������������������������������������
- Application 3.3 Premium Fast Food: Why Chipotle Is One Hot Pepper of a Stock�����������������������������������������������������������������������������������
- 3.4 Changes in Income and Consumption Choices����������������������������������������������������
- Normal Goods�������������������
- Inferior Goods���������������������
- The Food Stamp Program�����������������������������
- Application 3.4 The Allocation of Commencement Tickets�������������������������������������������������������������
- 3.5 Are People Selfish?������������������������������
- Application 3.5 Is Altruism a Normal Good?�������������������������������������������������
- 3.6 The Utility Approach to Consumer Choice��������������������������������������������������
- The Consumer’s Optimal Choice������������������������������������
- Relationship to Indifference Curves������������������������������������������
- Chapter 4: Individual and Market Demand����������������������������������������������
- 4.1 Price Changes and Consumption Choices������������������������������������������������
- The Consumer’s Demand Curve����������������������������������
- Some Remarks about the Demand Curve������������������������������������������
- Application 4.1 Using Price to Deter Youth Alcohol Abuse, Traffic Fatalities, and Campus Violence��������������������������������������������������������������������������������������������������������
- Do Demand Curves Always Slope Downward?����������������������������������������������
- Application 4.2 Why the Flight to Poultry and away from Red Meat by U.S. Consumers?������������������������������������������������������������������������������������������
- 4.2 Income and Substitution Effects of a Price Change������������������������������������������������������������
- Income and Substitution Effects Illustrated: The Normal-Good Case������������������������������������������������������������������������
- The Income and Substitution Effects Associated with a Gasoline Tax-Plus-Rebate Program���������������������������������������������������������������������������������������������
- A Graphical Examination of a Tax-Plus-Rebate Program�����������������������������������������������������������
- 4.3 Income and Substitution Effects: Inferior Goods����������������������������������������������������������
- A Hypothetical Example of a Giffen Good����������������������������������������������
- The Giffen Good Case: How Likely?����������������������������������������
- 4.4 From Individual to Market Demand�������������������������������������������
- Application 4.3 Aggregating Demand Curves for ITO��������������������������������������������������������
- 4.5 Consumer Surplus���������������������������
- The Uses of Consumer Surplus�����������������������������������
- Application 4.4 The Consumer Surplus Associated with Free TV�������������������������������������������������������������������
- Consumer Surplus and Indifference Curves�����������������������������������������������
- Application 4.5 The Benefits of Health Improvements����������������������������������������������������������
- 4.6 Price Elasticity and the Price–Consumption Curve�����������������������������������������������������������
- 4.7 Network Effects��������������������������
- The Bandwagon Effect���������������������������
- The Snob Effect����������������������
- Application 4.6 Network Effects and the Diffusion of Communications Technologies and Computer Hardware and Software��������������������������������������������������������������������������������������������������������������������������
- 4.8 The Basics of Demand Estimation������������������������������������������
- Experimentation����������������������
- Surveys��������������
- Regression Analysis��������������������������
- Chapter 5: Using Consumer Choice Theory����������������������������������������������
- 5.1 Excise Subsidies, Health Care, and Consumer Welfare��������������������������������������������������������������
- The Relative Effectiveness of a Lump-Sum Transfer��������������������������������������������������������
- Using the Consumer Surplus Approach������������������������������������������
- Application 5.1 The Price Sensitivity of Health Care Consumers���������������������������������������������������������������������
- 5.2 Subsidizing Health Insurance: ObamaCare��������������������������������������������������
- The Basics of ObamaCare������������������������������
- The Subsidy’s Effect on the Budget Line����������������������������������������������
- Bringing in Preferences������������������������������
- The Costs and Benefits of the Subsidy��������������������������������������������
- Other Possible Outcomes������������������������������
- Can a Subsidy Harm the Recipient?����������������������������������������
- One Other Option�����������������������
- 5.3 Public Schools and the Voucher Proposal��������������������������������������������������
- Using Consumer Choice Theory to Analyze Voucher Proposals����������������������������������������������������������������
- Application 5.2 The Demand for and Supply of School Choice�����������������������������������������������������������������
- 5.4 Paying for Garbage�����������������������������
- Does Everyone Benefit?�����������������������������
- Application 5.3 Trash Pricing and Recycling��������������������������������������������������
- 5.5 The Consumer’s Choice to Save or Borrow��������������������������������������������������
- A Change in Endowment����������������������������
- Application 5.4 Social Security and Saving�������������������������������������������������
- Changes in the Interest Rate�����������������������������������
- The Case of a Higher Interest Rate Leading to Less Saving����������������������������������������������������������������
- 5.6 Investor Choice��������������������������
- Application 5.5 Entrepreneurs and Their Risk–Return Preferences����������������������������������������������������������������������
- Investor Preferences toward Risk: Risk Aversion������������������������������������������������������
- Investor Preferences toward Risk: Risk Neutral and Risk Loving���������������������������������������������������������������������
- Application 5.6 Risk Aversion While Standing in Line�����������������������������������������������������������
- Minimizing Exposure to Risk����������������������������������
- Chapter 6: Exchange, Efficiency, and Prices��������������������������������������������������
- 6.1 Two-Person Exchange������������������������������
- The Edgeworth Exchange Box Diagram�����������������������������������������
- The Edgeworth Exchange Box with Indifference Curves����������������������������������������������������������
- Application 6.1 The Benefits of Exchange and eBay��������������������������������������������������������
- 6.2 Efficiency in the Distribution of Goods��������������������������������������������������
- Application 6.2 Promoting Efficiency in Gift Card Giving���������������������������������������������������������������
- Efficiency and Equity����������������������������
- 6.3 Competitive Equilibrium and Efficient Distribution�������������������������������������������������������������
- Application 6.3 Should Ticket Scalpers Be Disparaged or Deified?�����������������������������������������������������������������������
- 6.4 Price and Nonprice Rationing and Efficiency������������������������������������������������������
- Application 6.4 The Benefits and Costs of Rationing by Waiting���������������������������������������������������������������������
- Chapter 7: Production����������������������������
- 7.1 Relating Output to Inputs������������������������������������
- The Production Function������������������������������
- 7.2 Production When Only One Input Is Variable: The Short Run��������������������������������������������������������������������
- Total, Average, and Marginal Product Curves��������������������������������������������������
- The Relationship between Average and Marginal Product Curves�������������������������������������������������������������������
- The Geometry of Product Curves�������������������������������������
- Application 7.1 What the Marginal–Average Relationship Means for Your Grade Point Average (GPA)������������������������������������������������������������������������������������������������������
- The Law of Diminishing Marginal Returns����������������������������������������������
- Application 7.2 The Law of Diminishing Marginal Returns, Caffeine Intake, and Exam Performance�����������������������������������������������������������������������������������������������������
- 7.3 Production When All Inputs Are Variable: The Long Run����������������������������������������������������������������
- Production Isoquants���������������������������
- Four Characteristics of Isoquants����������������������������������������
- Marginal Rate of Technical Substitution (MRTS)�����������������������������������������������������
- MRTS and the Marginal Products of Inputs�����������������������������������������������
- Using MRTS: Speed Limits and Gasoline Consumption��������������������������������������������������������
- 7.4 Returns to Scale���������������������������
- Factors Giving Rise to Increasing Returns������������������������������������������������
- Factors Giving Rise to Decreasing Returns������������������������������������������������
- Application 7.3 Returns to Scale and Cross-Country Trade Flows���������������������������������������������������������������������
- 7.5 Functional Forms and Empirical Estimation of Production Functions����������������������������������������������������������������������������
- Linear Forms of Production Functions�������������������������������������������
- Multiplicative Forms of Production Functions: Cobb–Douglas as an Example�������������������������������������������������������������������������������
- Exponents and What They Indicate in Cobb–Douglas Production Functions����������������������������������������������������������������������������
- Chapter 8: The Cost of Production����������������������������������������
- 8.1 The Nature of Cost�����������������������������
- 8.2 Short-Run Cost of Production���������������������������������������
- Measures of Short-Run Cost: Total Fixed and Variable Costs�����������������������������������������������������������������
- Fixed versus Sunk Costs������������������������������
- Five Other Measures of Short-Run Cost��������������������������������������������
- Behind Cost Relationships��������������������������������
- 8.3 Short-Run Cost Curves��������������������������������
- Marginal Cost��������������������
- Average Cost�������������������
- Application 8.1 The Effect of Walmart on Retailing Productivity, Costs, and Prices�����������������������������������������������������������������������������������������
- Marginal–Average Relationships�������������������������������������
- The Geometry of Cost Curves����������������������������������
- 8.4 Long-Run Cost of Production��������������������������������������
- Isocost Lines��������������������
- Least Costly Input Combinations��������������������������������������
- Interpreting the Tangency Points���������������������������������������
- Application 8.2 American Airlines and Cost Minimization��������������������������������������������������������������
- The Expansion Path�������������������������
- Is Production Cost Minimized?������������������������������������
- Application 8.3 Privatization and Productivity in China��������������������������������������������������������������
- 8.5 Input Price Changes and Cost Curves����������������������������������������������
- Application 8.4 The Economics of Raising and Razing Buildings��������������������������������������������������������������������
- Application 8.5 Applying the Golden Rule of Cost Minimization to the Baseball Diamond��������������������������������������������������������������������������������������������
- 8.6 Long-Run Cost Curves�������������������������������
- The Long Run and Short Run Revisited�������������������������������������������
- 8.7 Learning by Doing����������������������������
- The Advantages of Learning by Doing to Pioneering Firms��������������������������������������������������������������
- 8.8 Importance of Cost Curves to Market Structure��������������������������������������������������������
- Application 8.6 The Decline and Rise of Breweries in the United States�����������������������������������������������������������������������������
- 8.9 Using Cost Curves: Controlling Pollution���������������������������������������������������
- 8.10 Economies of Scope������������������������������
- 8.11 Estimating Cost Functions�������������������������������������
- Chapter 9: Profit Maximization in Perfectly Competitive Markets����������������������������������������������������������������������
- 9.1 The Assumptions of Perfect Competition�������������������������������������������������
- The Four Conditions Characterizing Perfect Competition�������������������������������������������������������������
- 9.2 Profit Maximization������������������������������
- Application 9.1 Aligning Managerial Actions with Shareholder Interests: Lessons from the Recession of 2007–2009����������������������������������������������������������������������������������������������������������������������
- 9.3 The Demand Curve for a Competitive Firm��������������������������������������������������
- 9.4 Short-Run Profit Maximization����������������������������������������
- Short-Run Profit Maximization Using per-Unit Curves����������������������������������������������������������
- Operating at a Loss in the Short Run�������������������������������������������
- 9.5 The Perfectly Competitive Firm’s Short-Run Supply Curve������������������������������������������������������������������
- Output Response to a Change in Input Prices��������������������������������������������������
- Application 9.2 Why Firms That Fatten Cattle are Seeing Their Own Numbers Thinned����������������������������������������������������������������������������������������
- 9.6 The Short-Run Industry Supply Curve����������������������������������������������
- Price and Output Determination in the Short Run������������������������������������������������������
- 9.7 Long-Run Competitive Equilibrium�������������������������������������������
- Zero Economic Profit���������������������������
- Zero Profit When Firms’ Cost Curves Differ?��������������������������������������������������
- 9.8 The Long-Run Industry Supply Curve���������������������������������������������
- Constant-Cost Industry�����������������������������
- Increasing-Cost Industry�������������������������������
- Decreasing-Cost Industry�������������������������������
- Application 9.3 The Bidding War for Business School Professors���������������������������������������������������������������������
- Comments on the Long-Run Supply Curve��������������������������������������������
- Application 9.4 Cashing In on Corn�����������������������������������������
- 9.9 When Does the Competitive Model Apply?�������������������������������������������������
- Chapter 10: Using the Competitive Model����������������������������������������������
- 10.1 The Evaluation of Gains and Losses����������������������������������������������
- Producer Surplus�����������������������
- Consumer Surplus, Producer Surplus, and Efficient Output���������������������������������������������������������������
- The Deadweight Loss of a Price Ceiling���������������������������������������������
- 10.2 Excise Taxation���������������������������
- The Short-Run Effects of an Excise Tax���������������������������������������������
- The Long-Run Effects of an Excise Tax��������������������������������������������
- Who Bears the Burden of the Tax?���������������������������������������
- Tax Incidence: The Effect of Elasticity of Supply��������������������������������������������������������
- Tax Incidence: The Effect of Elasticity of Demand��������������������������������������������������������
- Application 10.1 Relative Ability to “Run” and Tax Incidence�������������������������������������������������������������������
- The Deadweight Loss of Excise Taxation���������������������������������������������
- Application 10.2 The Long and the Short (Run) of the Deadweight Loss of Rent Control�������������������������������������������������������������������������������������������
- 10.3 Airline Regulation and Deregulation�����������������������������������������������
- What Happened to the Profits?������������������������������������
- After Deregulation�������������������������
- Application 10.3 The Contestability of Airline Markets�������������������������������������������������������������
- The Push for Reregulation��������������������������������
- 10.4 City Taxicab Markets��������������������������������
- The Illegal Market�������������������������
- Application 10.4 Why New York City Cab Drivers Are Poor and Drive So Fast��������������������������������������������������������������������������������
- 10.5 Consumer and Producer Surplus, and the Net Gains from Trade�����������������������������������������������������������������������
- The Gains from International Trade�����������������������������������������
- The Link between Imports and Exports�������������������������������������������
- Application 10.5 Protecting Steel Jobs StealsJobs��������������������������������������������������������
- 10.6 Government Intervention in Markets: Quantity Controls�����������������������������������������������������������������
- Sugar Policy: A Sweet Deal���������������������������������
- Application 10.6 Why Sugar Import Quotas Were Job Losers with Respect to LifeSavers������������������������������������������������������������������������������������������
- Quotas and Their Foreign Producer Consequences�����������������������������������������������������
- Chapter 11: Monopoly���������������������������
- 11.1 The Monopolist’s Demand and Marginal Revenue Curves���������������������������������������������������������������
- 11.2 Profit-Maximizing Output of a Monopoly��������������������������������������������������
- Graphical Analysis�������������������������
- The Monopoly Price and Its Relationship to Elasticity of Demand����������������������������������������������������������������������
- Application 11.1 Demand Elasticity and Parking at Jack in the Box������������������������������������������������������������������������
- 11.3 Further Implications of Monopoly Analysis�����������������������������������������������������
- 11.4 The Measurement and Sources of Monopoly Power���������������������������������������������������������
- Measuring Monopoly Power�������������������������������
- The Sources of Monopoly Power������������������������������������
- Barriers to Entry������������������������
- Application 11.2 The Effect of State Licensing on One of the World’s Oldest Professions����������������������������������������������������������������������������������������������
- Strategic Behavior by Firms: Incumbents and Potential Entrants���������������������������������������������������������������������
- Application 11.3 March Monopoly Madness����������������������������������������������
- 11.5 The Efficiency Effects of Monopoly����������������������������������������������
- A Dynamic View of Monopoly and Its Efficiency Implications�����������������������������������������������������������������
- Application 11.4 Static Versus Dynamic Perspectives on Monopoly Control of Government��������������������������������������������������������������������������������������������
- 11.6 Public Policy toward Monopoly�����������������������������������������
- Regulation of Price��������������������������
- Application 11.5 What Not to Say to a Rival on the Telephone�������������������������������������������������������������������
- Application 11.6 Static versus Dynamic Views of Monopoly and the Microsoft Antitrust Case������������������������������������������������������������������������������������������������
- Application 11.7 Efficiency and the Regulation of Pharmaceutical Drug Prices in the European Union and the United States�������������������������������������������������������������������������������������������������������������������������������
- Chapter 12: Product Pricing with Monopoly Power������������������������������������������������������
- 12.1 Price Discrimination��������������������������������
- First-Degree Price Discrimination����������������������������������������
- Implementing First-Degree Price Discrimination�����������������������������������������������������
- Second-Degree Price Discrimination�����������������������������������������
- Third-Degree Price Discrimination����������������������������������������
- Application 12.1 Giving Frequent Shoppers the Second Degree������������������������������������������������������������������
- Application 12.2 The Third Degree by Car Dealers�������������������������������������������������������
- 12.2 Three Necessary Conditions for Price Discrimination���������������������������������������������������������������
- Application 12.3 Gray Hairs and Gray Markets���������������������������������������������������
- 12.3 Price and Output Determination with Price Discrimination��������������������������������������������������������������������
- Price Determination��������������������������
- Output Determination���������������������������
- Application 12.4 The Cost of Being Earnest When It Comes to Applying to Colleges���������������������������������������������������������������������������������������
- 12.4 Intertemporal Price Discrimination and Peak-Load Pricing��������������������������������������������������������������������
- Application 12.5 Yield Management by Airlines����������������������������������������������������
- Peak-Load Pricing������������������������
- The Advantages of Peak-Load Pricing������������������������������������������
- Application 12.6 Using Peak-Load Pricing to Combat Traffic Congestion����������������������������������������������������������������������������
- 12.5 Two-Part Tariffs����������������������������
- Many Consumers, Different Demands����������������������������������������
- Why the Price Will Usually Be Lower Than the Monopoly Price������������������������������������������������������������������
- Application 12.7 The Costs of Engaging in Price Discrimination���������������������������������������������������������������������
- Chapter 13: Monopolistic Competition and Oligopoly���������������������������������������������������������
- 13.1 Price and Output under Monopolistic Competition�����������������������������������������������������������
- Determination of Market Equilibrium������������������������������������������
- Monopolistic Competition and Efficiency����������������������������������������������
- Is Government Intervention Warranted?��������������������������������������������
- Application 13.1 Monopolistic Competition: The Eyes Have It (When It Comes to Refractive Surgery)��������������������������������������������������������������������������������������������������������
- 13.2 Oligopoly and the Cournot Model�������������������������������������������
- The Cournot Model������������������������
- Reaction Curves����������������������
- Evaluation of the Cournot Model��������������������������������������
- 13.3 Other Oligopoly Models����������������������������������
- The Stackelberg Model����������������������������
- The Dominant Firm Model������������������������������
- The Elasticity of a Dominant Firm’s Demand Curve�������������������������������������������������������
- Application 13.2 The Dynamics of the Dominant Firm Model in Pharmaceutical Markets�����������������������������������������������������������������������������������������
- 13.4 Cartels and Collusion���������������������������������
- Cartelization of a Competitive Industry����������������������������������������������
- Application 13.3 Does the Internet Promote Competition or Cartelization?�������������������������������������������������������������������������������
- Why Cartels Fail�����������������������
- Application 13.4 The Difficulty of Controlling Cheating��������������������������������������������������������������
- Application 13.5 The Rolex “Cartel”������������������������������������������
- Oligopolies and Collusion��������������������������������
- The Case of OPEC�����������������������
- The Reasons for OPEC’s Early Success�������������������������������������������
- The Rest of the OPEC Story���������������������������������
- Chapter 14: Game Theory and the Economics of Information���������������������������������������������������������������
- 14.1 Game Theory�����������������������
- Determination of Equilibrium�����������������������������������
- Application 14.1 Dominant Strategies in Baseball�������������������������������������������������������
- Nash Equilibrium�����������������������
- 14.2 The Prisoner’s Dilemma Game���������������������������������������
- Application 14.2 The Congressional Prisoner’s Dilemma������������������������������������������������������������
- The Prisoner’s Dilemma and Cheating by Cartel Members������������������������������������������������������������
- A Prisoner’s Dilemma Game You May Play���������������������������������������������
- 14.3 Repeated Games��������������������������
- Application 14.3 Cooperation in the Trenches of World War I������������������������������������������������������������������
- Do Oligopolistic Firms Always Collude?���������������������������������������������
- Game Theory and Oligopoly: A Summary�������������������������������������������
- 14.4 Asymmetric Information����������������������������������
- The “Lemons” Model�������������������������
- Market Responses to Asymmetric Information�������������������������������������������������
- The Relevance of the Lemons Model����������������������������������������
- Is There a Lemons Problem in Used Car Markets?�����������������������������������������������������
- Application 14.4 Job Market Signaling��������������������������������������������
- 14.5 Adverse Selection and Moral Hazard����������������������������������������������
- Adverse Selection������������������������
- Market Responses to Adverse Selection��������������������������������������������
- Application 14.5 Adverse Selection and the American Red Cross��������������������������������������������������������������������
- Moral Hazard�������������������
- Application 14.6 Moral Hazard and Subprime Home Mortgages����������������������������������������������������������������
- Market Responses to Moral Hazard���������������������������������������
- 14.6 Limited Price Information�������������������������������������
- Application 14.7 Moral Hazard on the Streets of New York City��������������������������������������������������������������������
- The Effect of Search Intensity on Price Dispersion���������������������������������������������������������
- 14.7 Advertising�����������������������
- Advertising as Information���������������������������������
- Advertising and Its Effects on Products’ Prices and Qualities��������������������������������������������������������������������
- Advertising, the Full Price of a Product, and Market Efficiency����������������������������������������������������������������������
- Application 14.8 The Effectiveness of Internet Advertising: The Case of Online Dating��������������������������������������������������������������������������������������������
- Chapter 15: Using Noncompetitive Market Models�����������������������������������������������������
- 15.1 The Size of the Deadweight Loss of Monopoly�������������������������������������������������������
- Why Are the Estimates of the Deadweight Loss Not Large?��������������������������������������������������������������
- Other Possible Deadweight Losses of Monopoly���������������������������������������������������
- 15.2 Do Monopolies Suppress Inventions?����������������������������������������������
- The Effect of Inventions on Output�����������������������������������������
- Application 15.1 The Cost of Not Cannibalizing�����������������������������������������������������
- 15.3 Natural Monopoly����������������������������
- Regulation of Natural Monopoly: Theory���������������������������������������������
- Regulation of Natural Monopoly: Practice�����������������������������������������������
- Application 15.2 Regulating Natural Monopoly through Public Ownership: The Case of USPS����������������������������������������������������������������������������������������������
- 15.4 More on Game Theory: Iterated Dominance and Commitment������������������������������������������������������������������
- Iterated Dominance�������������������������
- Commitment�����������������
- Application 15.3 Why It May Be Wise to Burn Your Ships�������������������������������������������������������������
- Chapter 16: Employment and Pricing of Inputs���������������������������������������������������
- 16.1 The Input Demand Curve of a Competitive Firm��������������������������������������������������������
- The Firm’s Demand Curve: One Variable Input��������������������������������������������������
- The Firm’s Demand Curve: All Inputs Variable���������������������������������������������������
- The Firm’s Demand Curve: An Alternative Approach�������������������������������������������������������
- 16.2 Industry and Market Demand Curves for an Input����������������������������������������������������������
- A Competitive Industry’s Demand Curve for an Input���������������������������������������������������������
- The Elasticity of an Industry’s Demand Curve for an Input����������������������������������������������������������������
- Application 16.1 Explaining Sky-High Pilot Salaries under Airline Regulation�����������������������������������������������������������������������������������
- The Market Demand Curve for an Input�������������������������������������������
- 16.3 The Supply of Inputs��������������������������������
- 16.4 Industry Determination of Price and Employment of Inputs��������������������������������������������������������������������
- Process of Input Price Equalization across Industries������������������������������������������������������������
- 16.5 Input Price Determination in a Multi-Industry Market����������������������������������������������������������������
- 16.6 Input Demand and Employment by an Output Market Monopoly��������������������������������������������������������������������
- 16.7 Monopsony in Input Markets��������������������������������������
- Application 16.2 Major League Monopsony����������������������������������������������
- Chapter 17: Wages, Rent, Interest, and Profit����������������������������������������������������
- 17.1 The Income–Leisure Choice of the Worker���������������������������������������������������
- Is This Model Plausible?�������������������������������
- 17.2 The Supply of Hours of Work���������������������������������������
- Is a Backward-Bending Labor Supply Curve Possible?���������������������������������������������������������
- The Market Supply Curve������������������������������
- Application 17.1 An Example of a Backward-Bending Labor Supply Curve: The Work Effort Choices of Dentists Versus Physicians����������������������������������������������������������������������������������������������������������������������������������
- Application 17.2 Why Do Americans Work More Than Europeans?������������������������������������������������������������������
- 17.3 The General Level of Wage Rates�������������������������������������������
- Application 17.3 The Malaise of the 1970s������������������������������������������������
- 17.4 Why Wages Differ����������������������������
- Compensating Wage Differentials��������������������������������������
- Differences in Human Capital Investment����������������������������������������������
- Application 17.4 Twelve Hours’ Pay for Ten Minutes’ Work���������������������������������������������������������������
- Application 17.5 The Returns to Investing in a BA and an MBA�������������������������������������������������������������������
- Differences in Ability�����������������������������
- 17.5 Economic Rent�������������������������
- 17.6 Monopoly Power in Input Markets: The Case of Unions���������������������������������������������������������������
- Application 17.6 The Decline and Rise of Unions and Their Impact on State and Local Government Budgets�������������������������������������������������������������������������������������������������������������
- Some Alternative Views of Unions and an Assessment of the Impact of Unions on Worker Productivity��������������������������������������������������������������������������������������������������������
- 17.7 Borrowing, Lending, and the Interest Rate�����������������������������������������������������
- 17.8 Investment and the Marginal Productivity of Capital���������������������������������������������������������������
- The Investment Demand Curve����������������������������������
- 17.9 Saving, Investment, and the Interest Rate�����������������������������������������������������
- Equalization of Rates of Return��������������������������������������
- 17.10 Why Interest Rates Differ��������������������������������������
- Chapter 18: Using Input Market Analysis����������������������������������������������
- 18.1 The Minimum Wage����������������������������
- Further Considerations�����������������������������
- Does the Minimum Wage Harm the Poor?�������������������������������������������
- The Minimum Wage: An Example of an Efficiency Wage?����������������������������������������������������������
- Application 18.1 The Disemployment Effect of the 1990–1991 Minimum Wage Hike�����������������������������������������������������������������������������������
- 18.2 Who Really Pays for Social Security?������������������������������������������������
- But Do Workers Bear All the Burden?������������������������������������������
- 18.3 The Hidden Cost of Social Security����������������������������������������������
- The Rest of the Story����������������������������
- Application 18.2 Other Hidden Costs of PAYGO Social Security�������������������������������������������������������������������
- The Effect on Labor Markets����������������������������������
- 18.4 The NCAA Cartel���������������������������
- An Input Buyers’ Cartel������������������������������
- The NCAA as a Cartel of Buyers�������������������������������������
- Eliminate the Cartel Restrictions on Pay?������������������������������������������������
- Application 18.3 The Differing Fortunes of College Athletes and Coaches������������������������������������������������������������������������������
- 18.5 Discrimination in Employment����������������������������������������
- What Causes Average Wage Rates to Differ?������������������������������������������������
- 18.6 The Benefits and Costs of Immigration�������������������������������������������������
- More on Gains and Losses�������������������������������
- Chapter 19: General Equilibrium Analysis and Economic Efficiency�����������������������������������������������������������������������
- 19.1 Partial and General Equilibrium Analysis Compared�������������������������������������������������������������
- The Mutual Interdependence of Markets Illustrated��������������������������������������������������������
- When Should General Equilibrium Analysis Be Used?��������������������������������������������������������
- 19.2 Economic Efficiency�������������������������������
- Efficiency as a Goal for Economic Performance����������������������������������������������������
- 19.3 Conditions for Economic Efficiency����������������������������������������������
- 19.4 Efficiency in Production������������������������������������
- The Edgeworth Production Box�����������������������������������
- The Production Contract Curve and Efficiency in Production�����������������������������������������������������������������
- General Equilibrium in Competitive Input Markets�������������������������������������������������������
- 19.5 The Production Possibility Frontier and Efficiency in Output������������������������������������������������������������������������
- Efficiency in Output���������������������������
- An Economy’s PPF and the Gains from International Trade��������������������������������������������������������������
- 19.6 Competitive Markets and Economic Efficiency�������������������������������������������������������
- The Role of Information������������������������������
- Application 19.1 Can Centralized Planning Promote Efficiency?��������������������������������������������������������������������
- 19.7 The Causes of Economic Inefficiency�����������������������������������������������
- Market Power�������������������
- Application 19.2 How Government Prolonged the Great Depression���������������������������������������������������������������������
- Imperfect Information����������������������������
- Application 19.3 Deterring Cigarette Smoking���������������������������������������������������
- Externalities/Public Goods���������������������������������
- Chapter 20: Public Goods and Externalities�������������������������������������������������
- 20.1 What Are Public Goods?����������������������������������
- The Free-Rider Problem�����������������������������
- Application 20.1 An Online Horror Tale���������������������������������������������
- Free Riding and Group Size���������������������������������
- 20.2 Efficiency in the Provision of a Public Good��������������������������������������������������������
- Efficiency in Production and Distribution������������������������������������������������
- Application 20.2 The Lowdown on Why Lojack Installations Are Lower Than the Efficient Output���������������������������������������������������������������������������������������������������
- Patents��������������
- 20.3 Externalities�������������������������
- Externalities and Efficiency�����������������������������������
- External Costs���������������������
- Application 20.3 Traffic Externalities: Their Causes and Some Potential Cures������������������������������������������������������������������������������������
- Application 20.4 Liability Caps and the British Petroleum Gulf Oil Disaster����������������������������������������������������������������������������������
- External Benefits������������������������
- Application 20.5 Should Cell Phone Use While Driving Be Banned?����������������������������������������������������������������������
- 20.4 Externalities and Property Rights���������������������������������������������
- The Coase Theorem������������������������
- Application 20.6 Making Telemarketers Pay������������������������������������������������
- 20.5 Controlling Pollution, Revisited��������������������������������������������
- The Market for Los Angeles Smog��������������������������������������
- Answers to Selected Problems
- Glossary
- Index
- EULA
