Review Questions

ECONOMICS OF THE PUBLIC SECTOR

FOURTH EDITION

ECONOMICS OF THE PUBLIC SECTOR

FOURTH EDITION

JOSEPH E. STIGLITZ | JAY K. ROSENGARD

n W. W. NORTON & COMPANY, INC. NEW YORK • LONDON

W. W. Norton & Company has been independent since its founding in 1923, when William Warder Norton and Mary D. Herter Norton f irst published lectures delivered at the People’s Institute, the adult education division of New York City’s Cooper Union. The f irm soon expanded its program beyond the Institute, publishing books by celebrated academics from America and abroad. By midcentury, the two major pillars of Norton’s publishing program—trade books and college texts— were f irmly established. In the 1950s, the Norton family transferred control of the company to its employees, and today—with a staf f of four hundred and a comparable number of trade, college, and professional titles published each year—W. W. Norton & Company stands as the largest and oldest publishing house owned wholly by its employees.

Copyright © 2015 by W. W. Norton Company, Inc.

Copyright © 2000, 1988, 1986 by Joseph E. Stiglitz, the Trustee of Edward Hannaway Stiglitz Trust, the Trustee of Julia Hannaway Stiglitz Trust, and the Trustee of the Trust for the Benef it of Joseph E. Stiglitz’s Children

All r ights reserved Printed in the United States of America

Editor: Jack Repcheck Editorial Assistant: Theresia Kowara Project Editor: Sujin Hong Managing Editor, College Digital Media: Kim Yi Production Manager: Vanessa Nuttry Marketing Manager, Economics: Janise Turso Design Director: Jil l ian Burr Permissions Manager: Megan Jackson Composition: Cenveo® Publisher Services Manufacturing: Quad/Graphics—Taunton

Library of Congress Cataloging-in-Publication Data.

Stiglitz, Joseph E. Economics of the public sector / Joseph E. Stiglitz, Jay K. Rosengard.—Fourth edition. pages cm Includes bibliographical references and index. ISBN 978-0-393-92522-7 (pbk.) 1. Finance, Public—United States. 2. Fiscal policy—United States. I. Rosengard, Jay K. II. Title. HJ257.2.S84 2015 336.73—dc23 2014048383

W. W. Norton & Company, Inc., 500 Fif th Avenue, New York, NY 10110-0017 ww nor ton.com

W. W. Norton & Company Ltd., Castle House, 75/76 Wells Street, London W1T 3QT

1 2 3 4 5 6 7 8 9 0

To our f irst teachers, Nat and Charlotte Jordan and Betty

vii

BRIEF CONTENTS

Preface xxvii

PART 1 ROLE AND SIZE OF THE PUBLIC SECTOR 1

1 DEFINING PUBLIC SECTOR RESPONSIBILITIES 3

2 MEASURING PUBLIC SECTOR SIZE 26

PART 2 FUNDAMENTALS OF WELFARE ECONOMICS 59

3 MARKET EFFICIENCY 61

4 MARKET FAILURE 81

5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS 101

6 EXTERNALITIES AND THE ENVIRONMENT 129

7 EFFICIENCY AND EQUITY 163

PART 3 PUBLIC EXPENDITURE THEORY 197

8 PUBLIC PRODUCTION OF GOODS AND SERVICES 199

9 PUBLIC CHOICE 230

viii BRIEF CONTENTS

PART 4 PUBLIC EXPENDITURE IN PRACTICE 267

10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY 269

11 EVALUATING PUBLIC EXPENDITURE 296

12 DEFENSE, RESEARCH, AND TECHNOLOGY 329

13 HEALTH CARE 357

14 EDUCATION 394

15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME 428

16 SOCIAL INSURANCE 470

PART 5 TAXATION IN THEORY 503

17 INTRODUCTION TO TAXATION 505

18 TAX INCIDENCE 538

19 TAXATION AND ECONOMIC EFFICIENCY 574

20 OPTIMAL TAXATION 606

21 TAXATION OF CAPITAL 636

PART 6 TAXATION IN PRACTICE 665

22 THE PERSONAL INCOME TAX 667

23 THE CORPORATION INCOME TAX 709

24 A STUDENT’S GUIDE TO TAX AVOIDANCE 746

25 REFORM OF THE TAX SYSTEM 762

PART 7 FURTHER ISSUES 799

26 INTER GOVERNMENTAL FISCAL RELATIONS 801

27 SUBNATIONAL TAXES AND EXPENDITURES 832

28 FISCAL DEFICITS AND GOVERNMENT DEBT 851

References 880 Index 893

ix

CONTENTS

Preface xxvii

PART 1 ROLE AND SIZE OF THE PUBLIC SECTOR 1

1 DEFINING PUBLIC SECTOR RESPONSIBILITIES 3

The Economic Role of Government 4

The Mixed Economy of the United States 4 Different Perspectives on the Role of Government 6 An Impetus for Government Action: Market Failures 7 Achieving Balance between the Public and Private Sectors 10 The Emerging Consensus 11

Thinking Like a Public Sector Economist 13

Analyzing the Public Sector 15 Economic Models 17 Case Study Musgrave’s Three Branches 18 Normative versus Positive Economics 19

Disagreements among Economists 21

Differences in Views on How the Economy Behaves 21 Disagreement over Values 23 Case Study Public Sector Economics and

the Global Economic Crisis 23

Review and Practice 24

Summary 24 Key Concepts 24 Questions and Problems 25

x

2 MEASURING PUBLIC SECTOR SIZE 26

What or Who Is the Government? 27

Types of Government Activity 29

Providing a Legal System 30 Government Production 30 Government’s Infl uence on Private Production 33 Government Purchases of Goods and Services 36 Government Redistribution of Income 36 Overview of Government Expenditures 40

Gauging the Size of the Public Sector 42

Growth in Expenditures and Their Changing Composition 42 Case Study Estimating the Full Budgetary and Economic Costs of War 43 Comparison of Expenditures across Countries 45

Government Revenues 47

Taxes and the Constitution 47 Federal Taxation Today 48 State and Local Government Revenues 49 Comparison of Taxation across Countries 50

Defi cit Financing 51

Playing Tricks with the Data on Government Activities 55 Review and Practice 56

Summary 56 Key Concepts 57 Questions and Problems 57

PART 2 FUNDAMENTALS OF WELFARE ECONOMICS 59

3 MARKET EFFICIENCY 61

The Invisible Hand of Competitive Markets 61

Welfare Economics and Pareto Effi ciency 63

Case Study On the Prowl for Pareto Improvements 64 Pareto Effi ciency and Individualism 65 The Fundamental Theorems of Welfare Economics 66 Effi ciency from the Perspective of a Single Market 68

Analyzing Economic Effi ciency 69

The Utility Possibilities Curve 69 Exchange Effi ciency 70 Production Effi ciency 74 Product Mix Effi ciency 78

CONTENTS

xi

Review and Practice 79

Summary 79 Key Concepts 80 Questions and Problems 80

4 MARKET FAILURE 81

Property Rights and Contract Enforcement 82 Case Study Property Rights and Market Failures:

The Tragedy of the Commons Revisited 82

Market Failures and the Role of Government 83

1. Failure of Competition 83 2. Public Goods 86 3. Externalities 86 4. Incomplete Markets 87 Case Study Student Loans: Incomplete Reform of an

Incomplete Market 89 5. Information Failures 91 6. Unemployment, Infl ation, and Disequilibrium 93 Interrelationships of Market Failures 93

Case Study Market Failures: Explanations or Excuses? 94

Redistribution and Merit Goods 95

Two Perspectives on the Role of Government 97

Normative Analysis 97 Positive Analysis 98

Review and Practice 99

Summary 99 Key Concepts 99 Questions and Problems 100

5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS 101

Public Goods 102

Public Goods and Market Failures 103 Paying for Public Goods 103 The Free Rider Problem 105 Case Study Economists and the Free Rider Problem 106 Pure and Impure Public Goods 107 Case Study Property Rights, Excludability, and

Externalities 110

Publicly Provided Private Goods 111

Rationing Devices for Publicly Provided Private Goods 113

Effi ciency Conditions for Public Goods 116

Demand Curves for Public Goods 117 Pareto Effi ciency and Income Distribution 122

CONTENTS

xii

Limitations on Income Redistribution and the Effi cient Supply of Public Goods 122 Distortionary Taxation and the Effi cient Supply of Public Goods 123

Effi cient Government as a Public Good 124

Review and Practice 125

Summary 125 Key Concepts 125 Questions and Problems 126

APPENDIX: The Leftover Curve 127

6 EXTERNALITIES AND THE ENVIRONMENT 129

The Problem of Externalities 130

Private Solutions to Externalities 132

Internalizing Externalities 132 The Coase Theorem 133 Using the Legal System 134 Case Study The Exxon Valdez Oil Spill 135 Failures of Private Solutions 136

Public Sector Solutions to Externalities 138

Case Study Double Dividend 138 Market-Based Solutions 139 Regulation 145 Innovation 146 Information Disclosure 148 Compensation and Distribution 149

Protecting the Environment: The Role of Government in Practice 150

Air 151 Water 155 Land 156

Concluding Remarks 159

Review and Practice 159

Summary 159 Key Concepts 160 Questions and Problems 160

7 EFFICIENCY AND EQUITY 163

Effi ciency and Distribution Trade-Offs 164

Analyzing Social Choices 164

Determining the Trade-Offs 166 Evaluating the Trade-Offs 169 Two Caveats 173

CONTENTS

xiii

Social Choices in Practice 174

Measuring Benefi ts 175 Ordinary and Compensated Demand Curves 178 Consumer Surplus 179 Measuring Aggregate Social Benefi ts 181 Measuring Ineffi ciency 181 Case Study Drawing a Poverty Line 182 Quantifying Distributional Effects 184 Case Study The Great Gatsby Curve 185

Three Approaches to Social Choices 186

The Compensation Principle 186 Trade-Offs across Measures 186 Weighted Net Benefi ts 187

The Trade-Off between Effi ciency and Fairness Revisited 188

Review and Practice 189

Summary 189 Key Concepts 190 Questions and Problems 190

APPENDIX: Alternative Measures of Inequality 192

The Lorenz Curve 192 The Dalton–Atkinson Measure 194

PART 3 PUBLIC EXPENDITURE THEORY 197

8 PUBLIC PRODUCTION OF GOODS AND SERVICES 199

Natural Monopoly: Public Production of Private Goods 201

The Basic Economics of Natural Monopoly 202 Regulation and Taxation (Subsidies) 206 No Government Intervention 209 Government Failures 210 Case Study Rent Control and Agricultural Price Supports:

Case Studies in Government Failure 211

Comparison of Effi ciency in the Public and Private Sectors 213

Case Study National Performance Review 214

Sources of Ineffi ciency in the Public Sector 216

Organizational Differences 216 Individual Differences 217 Bureaucratic Procedures and Risk Aversion 220

Corporatization 221

Case Study Privatizing Prisons 224

CONTENTS

xiv

A Growing Consensus on Government’s Role in Production 225

Review and Practice 227

Summary 227 Key Concepts 228 Questions and Problems 228

9 PUBLIC CHOICE 230

Public Mechanisms for Allocating Resources 230

The Problem of Preference Revelation 231 Individual Preferences for Public Goods 232 The Problem of Aggregating Preferences 236 Majority Voting and the Voting Paradox 237 Arrow’s Impossibility Theorem 238 Single-Peaked Preferences and the Existence of a

Majority Voting Equilibrium 240 The Median Voter 243 The Ineffi ciency of the Majority Voting Equilibrium 243 The Two-Party System and the Median Voter 246 Case Study Social Choice Theory 248

Alternatives for Determining Public Goods Expenditures 249

Lindahl Equilibrium 249

Politics and Economics 252

Why Do Individuals Vote? 252 Elections and Special Interest Groups 253 The Power of Special Interest Groups 254 Other Aspects of the Political Process 255 Case Study Campaign Finance Reform 256 The Altruistic Politician? 257 The Persistence of Ineffi cient Equilibrium 258

Review and Practice 259

Summary 259 Key Concepts 260 Questions and Problems 260

APPENDIX: New Preference-Revelation Mechanisms 262

PART 4 PUBLIC EXPENDITURE IN PRACTICE 267 10 FRAMEWORK FOR ANALYSIS OF

EXPENDITURE POLICY 269

Need for a Program 270

Market Failures 271

Case Study Higher Education in the United States 272

Alternative Forms of Government Intervention 272

CONTENTS

xv

The Importance of Particular Design Features 274

Private Sector Responses to Government Programs 275

Effi ciency Consequences 277

Income and Substitution Effects and Induced Ineffi ciency 277

Distributional Consequences 281

Evaluating the Distributional Consequences 284 Case Study Incidence of Education Tax Credits 284 Fairness and Distribution 286

Equity–Efficiency Trade-Offs 287

Public Policy Objectives 290

Political Process 291

Review and Practice 294

Summary 294 Key Concepts 294 Questions and Problems 294

11 EVALUATING PUBLIC EXPENDITURE 296

Private Cost–Benefi t Analysis 297

Present Discounted Value 297

Social Cost–Benefi t Analysis 299

Consumer Surplus and the Decision to Undertake a Project 300

Measuring Nonmonetized Costs and Benefi ts 303

Valuing Time 304 Valuing Life 304 Case Study Children, Car Safety, and the Value of Life 305

Valuing Natural Resources 307

Shadow Prices and Market Prices 308

Discount Rate for Social Cost–Benefi t Analysis 309

Case Study Climate Change and Discount Rates 312

The Evaluation of Risk 314

Risk Assessment 317

Distributional Considerations 318

Cost Effectiveness 319

Post-Expenditure Evaluation: Assessing and Improving Government Performance 323

Case Study Taking a Bite Out of Crime in the Big Apple 324

Review and Practice 326

Summary 326 Key Concepts 327 Questions and Problems 327

CONTENTS

xvi

12 DEFENSE, RESEARCH, AND TECHNOLOGY 329

Defense Expenditures 330

The Value of Marginal Analysis 333 Defense Strategy 334 Case Study Game Theory, the Arms Race, and

the Theory of Deterrence 336 Case Study Converting Swords into Plowshares 338

Increasing the Effi ciency of the Defense Department 339

Defense Procurement 339

Defense Conversion 343

Accounting and the Defense Department 344

Research and Technology 345

Market Failures 348 Case Study The Scope of the Patent: Can the

Human Body Be Patented? 350 Government Direct Support 353

Review and Practice 354

Summary 354 Key Concepts 355 Questions and Problems 355

13 HEALTH CARE 357

The Health Care System in the United States 360

The Private Sector 364 The Role of Government 364 Other Expenditure Programs 365 Tax Expenditures 366

Rationale for a Role of Government in the Health Care Sector 367

Imperfect Information 368 Limited Competition 369 Absence of Profi t Motive 370 Special Characteristics of the U.S. Market 371 The Role of the Health Insurance Industry 372 Case Study Medical Malpractice 372 Insurance and Excessive Expenditures on Health Care 374 Consequences of Ineffi ciencies in Health Care Markets 379

Poverty, Incomplete Coverage, and the Role of Government 381

Reforming Health Care 382

Cost Containment 383 Case Study Comprehensive Health Care Reform 384 Extending Insurance Coverage 385

CONTENTS

xvii

Medicare Reform: Easing Long-Term Fiscal Strains 387 Reforming Medicaid 390

Review and Practice 391

Summary 391 Key Concepts 392 Questions and Problems 392

14 EDUCATION 394

The Structure of Education in the United States 397

Federal Tax Subsidies to Private and Public Schools 399

Why Is Education Publicly Provided and Publicly Financed? 401

Is There a Market Failure? 401 The Federal Role 403

Issues and Controversies in Educational Policy 403

Education Outcomes 404 Do Expenditures Matter? 405 School Vouchers: Choice and Competition 407 Case Study Vouchers: The San Jose and Milwaukee

Experiments 412 School Decentralization 413 Performance Standards: No Child Left Behind and Race to the Top 414 Inequality 416

Aid to Higher Education 418

Review and Practice 422

Summary 422 Key Concepts 423 Questions and Problems 423

APPENDIX: How Should Public Educational Funds Be Allocated? 425

15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME 428

A Brief Description of Major U.S. Welfare Programs 430

AFDC and TANF 430 Earned Income Tax Credit 431 Food Stamps/SNAP 432 Medicaid 434 Housing 435 Other Programs 436

Rationale for Government Welfare Programs 437

Dimensions of the Problem 438

CONTENTS

xviii

Analytic Issues 440

Labor Supply 440 Cash versus In-Kind Redistribution 444 Ineffi ciencies from In-Kind Benefi ts 445 Are In-Kind Benefi ts Paternalistic? 450 Categorical versus Broad-Based Aid 451 Is Means Testing Objectionable in Its Own Right? 453 Other Distortions 453 Case Study Conditional Cash Transfer Programs 454

Welfare Reform: Integration of Programs 456

The Welfare Reform Bill of 1996 458

Block Granting 458 Analytics of State Responses to Block Grants 459 Time Limits 461 Mandatory Work 461 The Welfare Reform Debate of 1996 462 Case Study The Person or the Place? 464

Concluding Remarks 466

Review and Practice 466

Summary 466 Key Concepts 467 Questions and Problems 468

16 SOCIAL INSURANCE 470

The Social Security System 472

Social Security, Private Insurance, and Market Failures 475

High Transactions Costs 476 Risk Mitigation 477 Lack of Indexing: The Inability of Private Markets to

Insure Social Risks 477 Adverse Selection, Differential Risks, and the Cost of Insurance 478 Moral Hazard and Social Security 480 Retirement Insurance as a Merit Good 481 Social Security, Forced Savings, and Individual Choice 481

Is There a Need to Reform Social Security? 482

The Nature of the Fiscal Crisis 484 Savings 487 Labor Supply 488 The Rate of Return 490 Inequities 491

Reforming Social Security 492

Reducing Expenditures 492 Increasing Revenues 494

CONTENTS

xix

Structural Reforms 495 Case Study Social Security Abroad 496

Review and Practice 500

Summary 500 Key Concepts 501 Questions and Problems 501

PART 5 TAXATION IN THEORY 503

17 INTRODUCTION TO TAXATION 505

Background 506

Forms of Taxation 507 Changing Patterns of Taxation in the United States 508 Comparisons with Other Countries 509

The Five Desirable Characteristics of Any Tax System 511

Economic Effi ciency 512 Administrative Costs 517 Case Study Corrective Taxes and the Double Dividend 518 Flexibility 520 Transparent Political Responsibility 521 Fairness 523 Case Study Corruption-Resistant Tax Systems 531

General Framework for Choosing among Tax Systems 532

Utilitarianism 533 Rawlsian Social Welfare Function 534

Review and Practice 536

Summary 536 Key Concepts 537 Questions and Problems 537

18 TAX INCIDENCE 538

Tax Incidence in Competitive Markets 540

Effect of Tax at the Level of a Firm 540 Impact on Market Equilibrium 542 Does It Matter Whether the Tax Is Levied on Consumers

or on Producers? 543 Case Study The Incidence of Government Benefi ts 544 Ad Valorem versus Specifi c Taxes 545 The Effect of Elasticity 546 Taxation of Factors 548 Case Study The Philadelphia Wage Tax 549

CONTENTS

xx

Tax Incidence in Environments without Perfect Competition 552

Relationship between the Change in the Price and the Tax 554 Ad Valorem versus Specifi c Taxes 556 Tax Incidence in Oligopolies 556 Equivalent Taxes 557 Income Tax and Value-Added Tax 557 Equivalence of Consumption and Wage Taxes 558 Equivalence of Lifetime Consumption and Lifetime Income Taxes 559 A Caveat on Equivalence 560

Other Factors Affecting Tax Incidence 560

Tax Incidence under Partial and General Equilibrium 560 Case Study Behavioral Economics, Managerial Capitalism,

and Tax Incidence 561 Short-Run versus Long-Run Effects 564 Open versus Closed Economy 564 Associated Policy Changes 565 Case Study Tax Incidence of Specifi c Tax Provisions 566

Incidence of Taxes in the United States 566

Review and Practice 570

Summary 570 Key Concepts 570 Questions and Problems 571

APPENDIX: Comparison of the Effects of an Ad Valorem and Specifi c Commodity Tax on a Monopolist 572

19 TAXATION AND ECONOMIC EFFICIENCY 574

Effect of Taxes Borne by Consumers 575

Substitution and Income Effects 576

Quantifying the Distortions 577

Measuring Deadweight Loss Using Indifference Curves 578 Measuring Deadweight Loss Using Compensated

Demand Curves 580 Calculating the Deadweight Loss 582

Effect of Taxes Borne by Producers 584

Effects of Taxes Borne Partly by Consumers, Partly by Producers 587

Taxation of Savings 588

Quantifying the Effects of an Interest Income Tax 591

Taxation of Labor Income 591

Effects of Progressive Taxation 593 Case Study The 1993, 2001, and 2003 Tax Reforms 596 Secondary Labor Force Participants 597

CONTENTS

xxi

Measuring the Effects of Taxes on Labor Supplied 597

Statistical Techniques Using Market Data 598 Experiments 600 Review and Practice 603

Summary 603 Key Concepts 604 Questions and Problems 604

APPENDIX: Measuring the Welfare Cost of User Fees 605

20 OPTIMAL TAXATION 606

Two Fallacies of Optimal Taxation 607

The Fallacy of Counting Distortions 607 Misinterpretations of the Theory of the Second Best 607

Optimal and Pareto Effi cient Taxation 608

Lump-Sum Taxes 609 Why Impose Distortionary Taxes? 609 Case Study Estimating the Optimal Tax Rate 610 Case Study Rent Seeking, Inequality, and

Optimal Taxation 611 Designing an Income Tax System 611 Why Does More Progressivity Imply More Deadweight Loss? 612 A Diagrammatic Analysis of the Deadweight Loss of Progressive Taxation 614 Choosing among Flat-Rate Tax Schedules 615 Case Study The 1993 Tax Increase on Upper-Income Individuals: A Pareto Ineffi cient Tax? 616 General Equilibrium Effects 617 Case Study Flat-Rate Taxes Arrive on the Political Scene 618

Differential Taxation 621

Ramsey Taxes 621 Differential Commodity Taxes in Advanced Countries

with Progressive Income Taxes 625 Interest Income Taxation and Commodity Taxation 626

Taxes on Producers 627

The Dependence of Optimal Tax Structure on the Set of Available Taxes 629

Review and Practice 630

Summary 630 Key Concepts 631 Questions and Problems 631

APPENDIX A: Deriving Ramsey Taxes on Commodities 632 APPENDIX B: Derivation of Ramsey Formula for Linear Demand Schedule 634

CONTENTS

xxii

21 TAXATION OF CAPITAL 636

Should Capital Be Taxed? 638

Relationship among Consumption Taxes, a Wage Tax, and Exempting Capital Income from Taxation 638 Equity Issues 638 Effi ciency Arguments 639 Administrative Problems 640

Effects on Savings and Investment 641

Effects of Reduced Savings in a Closed Economy 641 The Distinction between Savings and Investment 642 National Savings and Budget Neutrality 644 Effects of Reduced Savings in an Open Economy 646

Impact on Risk Taking 648

Why Capital Taxation with Full Loss Deductibility May Increase Risk Taking 649 Case Study Tax Incentives for Risk Taking 650 Why Capital Taxation May Reduce Risk Taking 651

Measuring Changes in Asset Values 652

Capital Gains 653 Case Study Equity and the Reduction in

Capital Gains Taxes 654 Depreciation 657 Case Study Distortions from Depreciation 657 Neutral Taxation 659 Infl ation 659

Review and Practice 662

Summary 662 Key Concepts 663 Questions and Problems 663

PART 6 TAXATION IN PRACTICE 665

22 THE PERSONAL INCOME TAX 667

Outline of the U.S. Income Tax 667

Legislated versus Actual Tax Rates 672 Case Study A Loophole in the Earned Income

Tax Credit? 673 Other Taxes 673

Principles Behind the U.S. Income Tax 677

The Income-Based Principle and the Haig–Simons Defi nition 677 The Progressivity Principle 679 The Family-Based Principle 680 The Annual Measure of Income Principle 683

CONTENTS

xxiii

Practical Problems in Implementing an Income Tax System 684

Determining Income 684 Timing 690 Personal Deductions 690 Deductions versus Credits 698 Case Study Temporary Tax Changes 699

Special Treatment of Capital Income 699

Housing 700 Savings for Retirement 701 Interest on State and Municipal Bonds 703 Capital Gains 704

Concluding Remarks 706

Review and Practice 707

Summary 707 Key Concepts 707 Questions and Problems 708

23 THE CORPORATION INCOME TAX 709 The Basic Features of the Corporation Income Tax 711

The Incidence of the Corporation Income Tax and Its Effect on Effi ciency 713

The Corporation Income Tax as a Tax on Income from Capital in the Corporate Sector 713 Shifting of the Corporate Tax in the Long Run 715 The Corporation Tax for a Firm without Borrowing Constraints 717 Incidence of the Corporation Income Tax with Credit-Constrained Firms 718 The Corporation Tax as a Tax on Monopoly Profi ts 720 Managerial Firms: An Alternative Perspective 721

Depreciation 726

Combined Effects of Individual and Corporate Income Tax 728

Distributing Funds: The Basic Principles 728 The Dividend Paradox 730 Mergers, Acquisitions, and Share Repurchases 731 Does the Corporate Tax Bias Firms toward Debt Finance? 732

Distortions in Organizational Form Arising because Some Firms Do Not Have Taxable Income 734 Are Corporations Tax Preferred? 735 Calculating Effective Tax Rates 736

The Corporation Tax as Economic Policy 737

Case Study The Proposed Incremental Investment Tax Credit of 1993: An Idea before Its Time? 738

CONTENTS

xxiv

Taxation of Multinationals 739

Case Study Foreign Income and the Corporation Income Tax 741

Should There Be a Corporation Income Tax? 742

Why Is There a Corporate Income Tax at All? 743

Review and Practice 744

Summary 744 Key Concepts 744 Questions and Problems 745

24 A STUDENT’S GUIDE TO TAX AVOIDANCE 746

Principles of Tax Avoidance 747

Postponement of Taxes 747 Shifting and Tax Arbitrage 749 Case Study Shorting against the Box 751

Tax Shelters 752

Case Study The Economics of Tax Avoidance 753 Who Gains from Tax Shelters 753 Middle-Class Tax Shelters 755

Tax Reform and Tax Avoidance 756

The 1986 Tax Reform 756 Minimum Tax on Individuals 757 Subsequent Tax Acts 758

Equity, Effi ciency, and Tax Reform 758

Review and Practice 760

Summary 760 Key Concepts 760 Questions and Problems 760

25 REFORM OF THE TAX SYSTEM 762

Fairness 764

Horizontal Equity Issues 764 Vertical Equity 766

Effi ciency 769

Case Study Marginal Tax Rates and the 1986 Tax Reform 772 Base Broadening 773 Interaction of Fairness and Effi ciency Concerns 777

Simplifying the Tax Code 778

Assessing Complexity 778 Increasing Compliance 779 Reducing Tax Avoidance 781 Reducing Administrative and Compliance Costs 781 Sources of Complexity 782 The 1986 Tax Reform 784

CONTENTS

xxv

Transition Issues and the Politics of Tax Reform 785

Tax Reforms for the Twenty-First Century 787

Reforms within the Current Framework 787 Major New Reforms 788 Case Study Ordinary Income versus Capital Gains 794 Case Study IRAs and National Savings 796

Review and Practice 797

Summary 797 Key Concepts 797 Questions and Problems 797

PART 7 FURTHER ISSUES 799

26 INTERGOVERNMENTAL FISCAL RELATIONS 801

The Division of Responsibilities 802

Other Interaction between the Federal Government and the State and Local Governments 805 The Size of Financial Transfers 806 Case Study Unfunded Mandates 807

Principles of Fiscal Federalism 808

National Public Goods versus Local Public Goods 808 Case Study International Public Goods 809 Do Local Communities Provide Local Public Goods

Effi ciently? 810 Tiebout Hypothesis 810 Market Failures 812 Redistribution 814 Other Arguments for Local Provision 818

Production versus Finance 819

Effectiveness of Federal Categorical Aid to Local Communities 821 The Federal Tax System and Local Expenditures 825

Concluding Remarks 827

Review and Practice 830

Summary 830 Key Concepts 830 Questions and Problems 830

27 SUBNATIONAL TAXES AND EXPENDITURES 832

Tax Incidence Applied to Local Public Finance 832

Local Capital Taxes 833 Property Tax 834 Case Study The U.S. Property Tax Revolt 834 Income, Wage, and Sales Taxes 836 Distortions 837

CONTENTS

xxvi

Limitations on the Ability to Redistribute Income 837 Rent Control 838

Capitalization 839

Incentives for Pension Schemes 840 Choice of Debt versus Tax Financing 840 Short-Run versus Long-Run Capitalization 841 Who Benefi ts From Local Public Goods? The Capitalization

Hypothesis 841 Absolute versus Relative Capitalization 843 The Use of Changes in Land Rents to Measure Benefi ts 843 Testing the Capitalization Hypothesis 844

Public Choice at the Local Level 845

Problems of Multi jurisdictional Taxation 847

Review and Practice 849

Summary 849 Key Concepts 849 Questions and Problems 849

28 FISCAL DEFICITS AND GOVERNMENT DEBT 851

The U.S. Defi cit Problem since the 1980s 853

Sources of the Defi cit Problem 853 Factors Not Contributing to the Defi cit Problem 857 Success in Taming the Defi cit: The Experience of the 1990s 858

Case Study Measuring Budget Defi cits: What’s Large, What’s Real, and What’s Right? 858

Consequences of Government Defi cits 866

How Defi cits Affect Future Generations When the Economy Is at Full Employment 869 Alternative Perspectives on the Burden of the Debt 870 Case Study Austerity in a Recession: Expansionary or Contractionary? 871

Improving the Budgetary Process 873

Budget Enforcement Act and Scoring 873 Capital Budgets 874 Other Strategies 874

The Long-Term Problem: Entitlements and the Aged 875

Review and Practice 877

Summary 877 Key Concepts 878 Questions and Problems 878

References 880

Index 893

CONTENTS

xxvii

PREFACE

It has been more than a decade since the last revision of this textbook, and much has happened in the intervening years—two recessions, one from which we have yet to fully recover, two wars, a major health care reform, numerous tax reforms, and budget battles, one of which actually led to a government shutdown. Much has happened also in the development of the economics of the public sector. This edition incorporates this chang- ing economic and intellectual landscape. I have been fortunate enough to be joined by Jay Rosengard, a long-term practitioner of the subject and teacher at Harvard’s Kennedy School of Government, in writing this edition.

When the last edition published, I commented that I had been lucky to have been an active participant in many of these changes, as member and chairman of the President’s Council of Economic Advisers. That edition was written from the unique perspective of a public sector economist who had the chance to be involved in the decision-making process not only in the United States, but also in many other countries. Following my stint at the White House, I worked as chief economist and senior vice president of the World Bank, which is involved in advising developing countries concerning their public sector policies. Since then, I have continued with my passionate involvement in debates about public sector policies, as an adviser to many governments and as member or chair of several interna- tional commissions. I have been able to participate in debates around the world on the central questions with which this book is concerned: What should be the role of government? How should it design its programs in areas ranging from health and education, to Social Security and welfare? How should tax systems be designed to promote economic effi ciency and be consistent with basic views of fairness?

xxviii

In one sense, my experiences confi rmed many of the views and approaches I had developed in previous editions. Indeed, it gave me great pleasure to see the extent to which the ideas and perspectives, many of which seemed so new when they were presented in the fi rst edition of this book, were being integrated into thinking about policy, not only in the United States, but throughout the world. I have become increasingly convinced that the kind of analysis presented in this book can—and has— signifi cantly improve the formation of policies in the public sector.

The economics of the public sector is a subject that is always in fl ux. While there are some general principles that are as applicable today as they were two decades ago, new issues have risen to the top of the pol- icy agenda and old issues have waned in importance. Debates today often hinge on diff erent questions than they did even a decade ago. Even the language in which some of the debates are couched has changed. We have tried in this edition not only to incorporate the many changes in expen- diture policies and tax laws, but also to refl ect some of these changing approaches and themes. There is, for instance, an increased emphasis on understanding why government is often ineffi cient, and on improving the effi ciency of government—to use the phrase popularized by the Clinton administration—to “reinvent government.” Growing inequality, espe- cially in the United States, but also in most other countries around the world, has rightly become a subject of concern, and this book refl ects these concerns, with special attention given to the distributive consequences of diff erent policies.

Our major aim in writing this edition remains the same as when I wrote the fi rst: the belief that an understanding of the issues addressed in this book is central to any democratic society. Among the most import- ant of these are the appropriate balance between the public and private sectors, the ways in which the public and private sectors can comple- ment each other, and how governments can more eff ectively meet their objectives. Issues in public sector economics often become highly charged politically, but we tried to present the analysis in an impartial manner, with a clear delineation between the analysis of the consequences of any policy and the value judgments associated with assessing the desirabil- ity of the policy. We have tried to be clear about what economic theory and empirical research had to say on all sides of the debate, identifying where—and why—there is frequent uncertainty about the outcomes of certain policies, and clarifying why disagreements about the desirability of diff erent policies persist. In this edition, we continue with the commit- ment to present to the student a balanced account of these often heated debates. The favorable reception of the previous editions by instructors of a wide variety of political persuasions suggests that we have succeeded

PREFACE

xxix

in doing so. And the publication of many foreign editions of the book, in countries ranging from Russia, China, Japan, Germany, Italy, and Spain, to Latvia, Turkey, and the Czech Republic has shown that the approach has met with favor not just in the United States, but in countries facing quite diff erent circumstances and problems.

My experiences, at the White House, in the World Bank, and in dialogues around the world have made me even more convinced of the importance of this endeavor. Democracies can only succeed if there are meaningful public debates on the central public policy issues of the day. Too often, in too many countries, good policies fl ounder because of a lack of widespread understanding of basic economic issues. Writing an under- graduate textbook such as this thus present both a great challenge and a great opportunity: the challenge to present complex and complicated ideas in simple enough terms that they can be understood by someone with a relatively limited background in economics (at the most, a single year of a principles course); and the opportunity, if one succeeds in doing so, to infl uence the ways in which public policy debates are approached.

Public sector issues are some of the most exciting in all of economics. Health, defense, education, Social Security, welfare programs, and tax reform all receive steady attention in the news media, and economic anal- ysis brings special insights to the debates. Should education be publicly provided? What is the long-term outlook for our Social Security program? How do current proposals for tax reform match our knowledge of who really bears the tax burden? What determines the effi ciency and equity consequences of various taxes? These kinds of questions breathe life into the course, which is why we give them careful attention.

Examining specifi c tax and expenditure programs off ers an additional benefi t: it underscores the importance of design features. One of the les- sons we learned in the past decade is that good intentions are not enough. There are numerous examples where legislation has not been successful in achieving its objectives, and in which there are often unintended con- sequences. For instance, the 2001 and 2003 tax reforms, counter to their intentions, may have actually led to less investment. We use examples like these not only to enliven the course, but also to instill in students the important habit of testing theories against the complex environment in which public sector decisions are enacted and implemented.

The organization of this book is based on the principle of fl exibility. The sequence we follow is to introduce in Parts 1 and 2, the fundamental questions, institutional details, and a review of the microeconomic the- ory underlying the role of the public sector. Part 3 develops the theory of public expenditures, including public goods, public choice, and bureau- cracy, while Part 4 applies the theory of the fi ve largest areas of public

PREFACE

xxx

expenditure in the United States: health, defense, education, Social Secu- rity, and welfare programs. Parts 5 and 6 repeat this pattern, presenting the theory of taxation and its analysis, respectively. Part 7 takes up two additional topics: issues concerning state and local taxation and expen- diture and fi scal federalism; and issues concerning fi scal policy, with particular emphasis on the relation between microeconomic analysis and macroeconomic performance. The ups and downs of the defi cit are among the major changes we confront in this book. In the fi rst two edi- tions, defi cits were at the center of attention, but then, in the third edition, the defi cits changed to surpluses. As I noted then, there was a risk that this change would be temporary, and so it was, and defi cits are once again part of the economic debate. In this edition, we try to come to an under- standing of these marked fl uctuations in the U.S. defi cits.

A perfectly workable alternative to this sequence would be to cover taxation before expenditures. Parts 5 and 6 have been carefully developed so that instructors wishing to go straight to taxation after Part 1 can do so without losing continuity. Further tips on how courses can be orga- nized, as well as lecture notes, test questions, and coverage of advanced topics that instructors may wish to include in their lectures are in the Instructor’s Manual.

The list of those to whom I am indebted is a long one. My teachers at Amherst College, James Nelson and Arnold Collery, not only stimulated my interest in economics and in the particular subject of this course, but also laid the foundations for my later studies. They also showed me, by exam- ple, what good teaching meant. I hope that some of what I learned from them is refl ected in this book. At M.I.T., Dan Holland and E. Cary Brown introduced me to the formal study of public economics. Again, I hope some of the blend of policy, theory, and institutional detail that marked their work is refl ected here. The insights of my colleagues and collaborators at the institutions at which I have worked (M.I.T., Yale University, Stanford University, Princeton University, Oxford University, Cambridge Univer- sity, and the National Bureau of Economic Research) and the government agencies (Council of Economic Advisers, Treasury, Labor, Interior, Energy, Agency for International Development, State of Louisiana, State of Texas) and international organizations (World Bank, Interamerican Develop- ment Bank, Organization of Economic Cooperation and Development) for which I have worked and consulted have also proved invaluable. I should mention Henry Aaron (Brookings Institution), Alan J. Auerbach (Berkeley), Greg Ballantine (former Assistant Secretary of the Treasury for Tax Policy), William Baumol (Princeton University), Charles T. Clotfelter (Duke University), Partha Dasgupta (Cambridge University), Peter A. Diamond (M.I.T.), Avinash Dixit (Princeton University), Martin Feldstein

PREFACE

xxxi

(Harvard University), Harvey Galper (Brookings Institution), Robert E. Hall (Stanford University), Jon Hamilton (University of Florida), Arnold G. Harberger (University of Chicago and University of California, Los Angeles), Charles E. McClure (Hoover Institution; former Deputy Assis- tant Secretary of the Treasury), James A. Mirrlees (Cambridge University), Alvin Rabushka (Stanford University), Michael Rothschild (Princeton), Agnar Sandmo (Norges Handelshøgskole, Norway), Eytan Sheshinski (Hebrew University), Nick Stern (London School of Economics), Lawrence Summers (Harvard University), and in particular Anthony B. Atkinson (Oxford University), Peter Mieskowski (Rice University), Kumar Sah (Uni- versity of Chicago), and Steven L. Slutsky (University of Florida).

Comments and suggestions I received from those who taught from the book or read various stages of the manuscript have been enormously help- ful in shaping this text. Here I particularly want to thank Justin Barnette (Kent State University), Donald N. Baum (University of Nebraska), Jim Bergin (Queens University, Canada), Michael Boskin (Stanford University), Lawrence Blume (Cornell University), the late David Bradford (Princeton University), Bradley Braun (University of Central Florida), Douglas M. Brown (Georgetown University), Donald Bruce (University of Tennessee), Neil Bruce (University of Washington), John Burbidge (McMaster Uni- versity), Paul M. Carrick (Naval Postgraduate School), Donald Cole (Drew University), Paul N. Courant (University of Michigan), Lieutenant Colonel Floyd Duncan (Virginia Military Institute), Stephen Erfl e (Dickinson College), J. Eric Fredland (US. Naval Academy), Victor R. Fuchs (Stanford University), Don Fullerton (University of Illinois), Ted Gayer (Brook- ings Institution), Malcolm Getz (Vanderbilt University), Roger Gordon (University of California, San Diego), Timothy Gronberg (Texas A&M University), William F. Hellmuth (Virginia Commonwealth University), Mervyn King (New York University), Laurence J. Kotlikoff (Boston Uni- versity), Sally Kwak (Johns Hopkins University), the late Robert Lampman (University of Wisconsin), Jerry Miner (Syracuse University), Yasuhide Okuyama (The University of Kitakyushu, Japan), Umut Ozek (American Institute for Research), the late Joseph A. Pechman (Brookings Institu- tion), Harold Pollack (University of Chicago), Jim Poterba (M.I.T.), the late Anora Robbins (UNC Greensboro), Balbir S. Sahni (Concordia Uni- versity, Montreal), Catherine Schneider (Boston College), Robert Sherry (Keene State College), John Shoven (Stanford University), Joel Slemrod (University of Michigan), Anne Winkler (University of Missouri at St.  Louis), Sun-Tien Wu (Chung Hsing University, Taipei), and Qiang Zeng (Tsiing Hua University, Beijing).

In its second edition, this book benefi ted tremendously from the insights of Karla Hoff , who served as both research assistant and critic.

PREFACE

xxxii

In that eff ort, I was aided by a group of outstanding graduate research assistants from University of Maryland: Amy Harris, Kosali Ilayper- uma, Steven Karon, and Diana Stech. I also owe a special thanks to Janet McCubbin for her contributions on the fi rst draft of the Third Edition. Janet read each word of the manuscript, and her invaluable feedback put the revision on a fi rm footing from the get-go.

For this fourth edition, I also wish to thank Eamon Kircher-Allen for his assistance in shepherding this edition to completion.

Finally, we are deeply indebted to the fi ne people at W. W. Norton and Company, a truly outstanding publishing fi rm, who brought this project to fruition. For this edition, I owe special thanks to Jack Repcheck for his excellent guidance, as well as the rest of the superb team at Norton: Sujin Hong, Theresia Kowara, Vanessa Nuttry, Marina Rozova, Carson Russell, and Stefani Wallace. I also remain eternally grateful to Ed Parsons, Claire Acher, Kate Barry, Joan Benham, Margaret Farley, Roseanne Fox, and Mark Henderson for their excellent work on earlier editions.

J.E.S. Fall 2014

PREFACE

ROLE AND SIZE OF THE PUBLIC SECTOR

At the center of any country’s political life are some basic economic questions: How does the government affect the economy? What is the appropriate role and size of government? Why are some economic activities undertaken in the public sector and others in the private sector? Should government do more than it is currently doing, or less? Should government change what it is doing, and how it is doing it?

To answer these questions, we must begin by understanding what the government does today. How have governments changed over time? How do the size and scope of government in one country compare with those of other countries? What might explain these dif ferences?

Part One provides this context for public sector economics. Chapter 1 gives an overall perspective on the economic role of government. It sets forth the basic questions that are addressed by public sector econo- mists, and explains some of the reasons why there are disagreements among them about appropriate policies. Chapter 2 addresses challenges in measuring the size of the public sector and provides comparative data on the magnitude of the public sector around the world today.

PART ONE

3

DEFINING PUBLIC SECTOR RESPONSIBILITIES

From birth to death, our lives are aff ected in countless ways by the activ- ities of government. For example, in the United States:

• We are born in hospitals that are publicly subsidized, if not publicly owned. Our arrival is then publicly recorded (on our birth certifi cates), entitling us to a set of privileges and obligations as American citizens.

• Most of us (almost 90 percent) attend public schools. • Virtually all of us, at some time in our lives, receive money from the

government, through programs such as student loans, unemployment or disability payments, antipoverty programs, Social Security, and Medicare.

• All of us pay taxes to the government—sales taxes; taxes on such com- modities as gasoline, liquor, telephones, air travel, perfumes, and tires; property taxes; income taxes; and Social Security (payroll) taxes.

• More than a sixth of the work force is employed by the government, and for the rest, the government has a signifi cant impact on employment conditions.

• In many areas of production—be it cars, sneakers, or computers— profi ts and employment opportunities are greatly aff ected by whether

1 1. What are the central questions concerning the economics of the public sector? 2. What are the diff ering

views concerning the economic role of govern- ment? How have they changed over the years and what has given rise to those changes?

3. How do economists go about studying the economics of the public sector?

4. What are the principal sources of disagreement among economists about appropriate policies that government should pursue?

FOCUS QUESTIONS

4 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

the government allows foreign competitors to sell goods in America without a tariff or quota.

• What we eat and drink, where we can live, and what kinds of houses we can live in are all regulated by government agencies.

• We travel on public roads and publicly subsidized railroads. In most communities our garbage is collected and our sewage is disposed of by a public agency; in some communities the water we drink is provided by public water companies.

• Our legal structure provides a framework within which individuals and fi rms can sign contracts with one another. When there is a dispute between two individuals, the two may turn to the courts to adjudicate the dispute.

• Without environmental regulations, many of our major cities would be choked with pollution, the water of our lakes and rivers would be undrinkable, and we could neither swim nor fi sh in them.

• Without safety regulations, such as those requiring seat belts, highway fatalities would be even higher than they are.

THE ECONOMIC ROLE OF GOVERNMENT

Why does government engage in some economic activities and not others? Why has the scope of its activities changed over the past hundred years, and why does it have diff erent roles in diff erent countries? Does the gov- ernment do too much? Does it do well what it attempts to do? Could it per- form its economic role more effi ciently? These are the central questions with which the economics of the public sector is concerned. To address them, we will fi rst consider the economic role of government in modern economies, how ideas about the role of government have emerged, and the changing role of government in the twenty-fi rst century.

THE MIXED ECONOMY OF THE UNITED STATES

The United States has what is called a mixed economy, where many eco- nomic activities are undertaken by private fi rms, while others are under- taken by the government. In addition, the government alters the behavior of the private sector through a variety of regulations, taxes, and subsidies.

5The Economic Role of Government

By contrast, in the former Soviet Union, most economic activities were undertaken by the central government. Today, only North Korea and Cuba give the government such primacy. In many western European econo- mies, national governments have had a larger role in economic activity than in the United States. For instance, the government of France once participated in a range of economic activities, including the production of cars, electricity, and airplanes. Since the 1980s, however, privatization— converting government enterprises into private fi rms—has been the trend in Europe, although the economic role of government generally remains larger there than in the United States.1 Some of these privatizations have been far less successful than hoped. In the United States, President Bush proposed partially privatizing Social Security, but in the aftermath of the 2008 fi nancial crisis, support for such eff orts vanished, as Americans real- ized the magnitude of the losses—and the insecurity that they would have faced—had that initiative proceeded. Indeed, around the world, the cri- sis brought on the most signifi cant increase in the role of the government in more than half a century, with governments taking over or providing massive subsidies to the fi nancial sector and a host of other industries.

The origins of the mixed economy of the United States lie in the ori- gins of the country itself. In formulating the United States Constitution, the founders of the republic had to address explicitly key issues concern- ing the economic role of the new government. The Constitution assigned the federal government certain responsibilities, such as running the post offi ce and printing money. It provided the foundations for what we now call “intellectual property rights” by giving the government the right to grant patents and issue copyrights to encourage innovation and creativity. It gave the federal government certain rights to levy taxes, although those did not include taxes on exports, income, or net wealth. Most importantly, for the future evolution of the country, Article 1, Section 8, Clause 3 gave the federal government the right to regulate interstate commerce. Because so much of economic activity involves goods produced in one state and sold in another, this clause, interpreted broadly, has been used to justify much of the federal government’s regulatory activities.

Throughout the history of the United States, the economic role of the government has undergone important changes. For instance, one hun- dred years ago some highways and all railroads were private; today, there are no major private roads and most interstate railroad passenger travel is by Amtrak, a publicly owned and subsidized enterprise. It is because mixed economies constantly face the problem of defi ning the appropriate

1 For more on the case of France, see H. Dumez and A. Jeunemaitre, “Privatization in France: 1983– 1993,” in Industrial Privatization in Western Europe: Pressures, Problems, and Paradoxes, ed. Vincent Wright (London and New York: Pinter Publishers, 1994), pp. 83–105, 194.

6 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

boundaries between government and private activities that the study of the economics of the public sector in these countries is both so important and so interesting.

DIFFERENT PERSPECTIVES ON THE ROLE OF GOVERNMENT

To better understand contemporary perspectives on the economic role of government, it can be helpful to consider the diff erent perspectives that have evolved in the past.2 Some of the central ideas of the eighteenth and nineteenth centuries have been critical to economic history in the twentieth century, and continue to be important today.

One dominant view in the eighteenth century, which was particularly persuasive among French economists, was that the government should actively promote trade and industry. Advocates of this view were called mercantilists. It was partly in response to the mercantilists that Adam Smith (who is often viewed as the founder of modern economics) wrote The Wealth of Nations (1776), in which he argued for a limited role for gov- ernment. Smith attempted to show how competition and the profi t motive would lead individuals—in pursuing their own private interests—to serve the public interest. The profi t motive would lead individuals, competing against one another, to supply the goods other individuals wanted. Only fi rms that produced what was wanted and at as low a price as possible would survive. Smith argued that the economy was led, as if by an invisi- ble hand, to produce what was desired—and in the best possible way.

Adam Smith’s ideas had a powerful infl uence both on governments and on economists. Many of the most important nineteenth-century economists, such as the Englishmen John Stuart Mill and Nassau Senior, promulgated the doctrine known as laissez faire. In their view, the gov- ernment should leave the private sector alone; it should not attempt to regulate or control private enterprise. Unfettered competition would serve the best interests of society.

Not all nineteenth-century social thinkers were persuaded by Smith’s reasoning. The grave inequalities in income that they saw around them, the squalor in which much of the working classes lived, and the unemployment that workers frequently faced concerned them. While nineteenth-century writers like Charles Dickens attempted to portray the plight of the working classes in novels, social theorists, such as Karl Marx, Jean Charles Léonard de Sismondi, and Robert Owen, developed theories that not only attempted

2 See A. O. Hirschman, Shifting Involvements: Private Interest and Public Action (Princeton, NJ: Princeton University Press, 1982). Hirschman has put forth an interesting theory attempting to explain the constant changes in views on the appropriate role of the government.

7The Economic Role of Government

to explain what they saw but also suggested ways in which society might be reorganized. Many attributed the evils in society to the private ownership of capital; what Adam Smith saw as a virtue, they saw as a vice. Marx, if not the deepest of the social thinkers, was certainly the most infl uential among those who advocated a greater role for the state in controlling the means of production. Still others, such as Owen, saw the solution neither in the state nor in private enterprise, but in smaller groups of individuals getting together and acting cooperatively for their mutual interest.

On one hand, private ownership of capital and unfettered free enter- prise; on the other, government control of the means of production—these contrary principles were to become a driving force for international pol- itics and economics in the twentieth century, embodied in the Cold War. Today, the countries of the former Soviet Union and the Eastern bloc are in the midst of a monumental transition to market systems—a fundamen- tal transformation of government’s role in those economies. In the United States, the economic role of government has also changed, but the changes have arisen more gradually, in response to economic events throughout the century. There is now widespread agreement that markets and private enterprises are at the heart of a successful economy, but that government plays an important role as a complement to the market.

The precise nature of that role, however, remains a source of contention. It diff ers both between countries and within nations over time, depending largely on society’s expectations for government and what the members of a society are willing to pay to meet these expectations, sometimes referred to as a “social compact” or “social contract.” For example, the citizens of countries in northern Europe generally expect their governments to provide health, education, and social services that are provided largely by the private sector in the United States, and are willing to pay relatively higher taxes to fi nance these public services. Several of these countries have succeeded in creating public health care systems that deliver better health outcomes at much smaller costs than the largely private American system. The debate over the appropriate role of the government took a sharp turn in 2008, when it became evident that only the government could save the economy from an economic crisis that had largely been created by private markets.

AN IMPETUS FOR GOVERNMENT ACTION: MARKET FAILURES

Prior to the 2008 crisis, the Great Depression—in which the unemploy- ment rate reached 25 percent and national output fell by about a third from its peak in 1929—was the event that most fundamentally changed

8 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

attitudes toward government. There was a justifi ed widespread view that markets had failed in an important way, and there were enormous pressures for government to do something about this market failure. The great English economist John Maynard Keynes, writing in the midst of the Great Depression, argued forcefully that the government not only should do something about economic slumps, but also that it could. The belief that governments should and could stabilize the level of economic activity was eventually embedded in legislation in the United States, in the Full Employment Act of 1946, which at the same time established the Council of Economic Advisers, to counsel the President on how best to accomplish these objectives.

The economy’s seeming inability to provide jobs was not the only problem that drew attention. The depression brought to the fore problems that, in less severe form, had been there for a long time. Many individu- als lost virtually all their money when banks failed and the stock mar- ket crashed. Many elderly people were pushed into dire poverty. Many farmers found that the prices they received for their products were so low that they could not make their mortgage payments, and defaults became commonplace.

In response to the depression, the federal government not only took a more active role in attempting to stabilize the level of economic activ- ity, but also passed legislation designed to alleviate many of the specifi c problems: unemployment insurance, Social Security, federal insurance for depositors, federal programs aimed at supporting agricultural prices, and a host of other programs aimed at a variety of social and economic objectives. Together, these programs are referred to as the New Deal.

After World War II, the country experienced an unprecedented level of prosperity. However, it became clear that not everyone was enjoying the fruits of that prosperity. Many individuals, by the condition of their birth, seemed to be condemned to a life of squalor and poverty; they received inadequate education, and their prospects for obtaining good jobs were bleak.

These inequities provided the impetus for many of the govern- ment programs that were enacted in the 1960s, when President Lyndon B. Johnson declared his “War on Poverty.” Whereas some programs were aimed at providing a “safety net” for the needy—for instance, programs to provide food and medical care to the poor—others, such as job retraining programs and Head Start, which off ers preschool education for under- privileged children, were directed at improving the economic opportuni- ties of the disadvantaged.

Could government actions alleviate these problems? How was suc- cess to be gauged? The fact that a particular program did not live up to

9The Economic Role of Government

the hopes of its most enthusiastic supporters did not, of course, mean that it was a failure. Medicaid, which provides medical assistance to the indigent, was successful in narrowing the diff erences in access to med- ical care between the poor and the rich, but the gap in life expectancy between these two groups was not eliminated. Medicare, which provides medical care for the elderly, relieved the elderly and their families of much of the anxiety concerning the fi nancing of their medical expenses, but it left in its wake a national problem of how to fi nance rapidly increasing medical expenditures. Even though the Social Security program provided the aged with an unprecedented level of economic security, it too has run into fi nancial problems that some critics say have cast doubt on whether future generations will be able to enjoy the same benefi ts.

Fifty years after the War on Poverty began, poverty has not been erad- icated from America. Government programs have signifi cantly reduced poverty from what it otherwise would have been, but both critics and supporters of the government’s programs agree that we still have con- siderable challenges to overcome if we are to eliminate poverty, and that good intentions often have unintended negative consequences. Many programs designed to alleviate the perceived inadequacies of the market economy have had eff ects markedly diff erent from those their proponents anticipated. Urban renewal programs designed to improve the quality of life in inner cities have, in some instances, resulted in the replacement of low-quality housing with high-quality housing that poor people cannot aff ord, thus forcing them to live in even worse conditions. Homelessness has become an increasing concern. Although many programs designed to promote integration of public schools have succeeded, because of res- idential segregation, public schools in some districts are no better inte- grated than private schools. A disproportionate share of the benefi ts of farm programs has accrued to large farms; government programs have not enabled many of the small farms to survive.

Supporters of continued government eff orts claim that critics exag- gerate the failures of government programs. They argue that the lesson to be learned is not that the government should abandon its eff orts to solve the major social and economic problems facing the nation, but that greater care must be taken in the appropriate design of government pro- grams. In other words, the limitations of government, or “government failures,” should not prevent the government from trying to mitigate mar- ket failures.

More recently, attention has shifted to two other market failures: excessive volatility, evidenced by the crisis of 2008 and by more than a hundred other crises around the world since 1980 when the era of deregulation began, and growing inequality, accompanied by a decline

10 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

in economic opportunity. These problems are prevalent in many other countries, but they have become a special focus of attention in the United States, which now has the highest level of inequality among the advanced industrial countries and one of the lowest levels of “opportunity”; that is, an American child’s life prospects are more dependent on the income and education of his or her parents than in other advanced countries. Even worse, some of the wealth at the top (in particular, that associated with fi nancial markets) has come as a result of exploitation of those at the bot- tom, through predatory and discriminatory lending, abusive credit card practices, and exploitation of market power. These pervasive concerns about the market suggest that there is an important role for government; but in some quarters, there is the worry that government has not only failed to “correct” the market failures, but it may also have actually con- tributed to the problems.

ACHIEVING BALANCE BETWEEN THE PUBLIC AND PRIVATE SECTORS

Markets often fail, but governments often do not succeed in correcting the failures of the market. Today, economists, in ascertaining the appro- priate role of government, attempt to incorporate an understanding of the limitations of both government and markets. There is agreement that there are many problems that the market does not adequately address; more generally, the market is fully effi cient only under fairly restrictive assumptions (see Chapters 3 and 4).

The recognition of the limitations of gov- ernment, however, implies that government should direct its energies only at areas in which market failures are most signifi cant and where there is evidence that government intervention can make a signifi cant diff erence. Among American economists today, the domi- nant view is that limited government interven- tion could alleviate (but not solve) the worst problems; thus, the government should take an active role in maintaining full employment and alleviating the worst aspects of poverty, but private enterprise should play the cen- tral role in the economy. The prevalent view attempts to fi nd ways for government and markets to work together, each strengthening

THE MIXED ECONOMY • The United States is a mixed economy, in which

both the public and private sectors play an important role.

• The roles played by government—and views concerning what they should be—have changed markedly over time.

• An important motivation for government’s undertaking certain activities is actual or perceived failures of the market.

• There has been increasing recognition of the limitations of government—of “government failures” as well as market failures.

11The Economic Role of Government

the other. For instance, ensuring that governments rely more heavily on markets and market-like mechanisms.

Controversy remains, though, over how limited or how active the government should be, with views diff ering according to how serious one considers the failures of the market to be and how eff ective one believes government is in remedying them. Typically, economists who have served Democratic administrations believe that there is a larger role for govern- ment to play; those who have served Republican administrations are more doubtful. Part of the disagreement arises from the importance they attach to market failures; part of it arises from the importance they attach to inequality—even when markets are effi cient, even pro-market economists agree that markets may lead to an unacceptably high level of poverty. However, much of the diff erence arises from politics: their assessment of the ability or likelihood that government will eff ectively deal with the market failure, without creating problems of its own. Some of those who agree that the market produces “too much” inequality believe that the costs of even the most effi cient eff orts to reduce poverty are too great, whereas others believe that, in practice, government eff orts will prove ineff ective.

THE EMERGING CONSENSUS

As important as they are, the diff erences in views of government’s eco- nomic role are far smaller than the diff erences a hundred years ago, when socialists advocated a dominant role for government and laissez-faire economists advocated no role for government at all. Contemporary rethinking of the role of government has been refl ected in two concurrent initiatives, deregulation and privatization.

The fi rst, begun in the United States under President Carter, reduced the role of government in regulating the economy. For instance, the gov- ernment stopped regulating prices for airlines and long-distance trucking. While there is recognition that regulations have a cost, there is increasing awareness that not regulating may have even greater costs. Regulations have continued to grow, partly in response to the growing recognition of market failures, such as those associated with environmental degrada- tion and the near collapse of the banking system twice in the past three decades. The Clinton administration, in its “Reinventing Government” initiative, sought a balance: while recognizing the need for regulation, it also recognized that some regulations were overly burdensome, their benefi ts were less than their costs, and there might be more eff ective ways of obtaining the desired objectives. Major reforms were instituted in such areas as banking, telecommunications, and electricity. In some of these

12 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

areas, such as telecommunications, it was hoped that, with new technolo- gies, the scope for competition would be larger than had previously been thought. Parallel reforms occurred throughout the world.

In some cases, including the United States, the enthusiasm for deregu- lation seems to have been carried too far. The economic crisis in East Asia in 1997—as the savings and loan debacle in the United States had done a decade earlier and the global economic crisis did in 2008—brought home the importance of fi nancial market regulation. These crises ultimately end up costing workers, businesses, and even taxpayers dearly. So, too, the hoped-for competition in telecommunications did not emerge.

Like the Clinton administration, the Obama administration claims to be striving to fi nd an appropriate balance. Critics worry whether, like the Clinton administration, it has come too much under the infl uence of the special interests that benefi ted from deregulation or of the ideologies that place excessive confi dence in the marketplace.

The second initiative, privatization, sought to turn over to the private sector activities previously undertaken by government. The privatization movement was much stronger in Europe, where telephones, railroads, airlines, and public utilities were all privatized. In the United States, because government ran few enterprises, there was much less scope for privatization. Perhaps the most important, and controversial, privatiza- tion was that of the United States Enrichment Corporation, the govern- ment agency responsible for enriching uranium. (Low-enriched uranium is used in nuclear power plants; highly enriched uranium is used to make atomic bombs. The same process and plants are used to make both.) The privatization, which was approved in 1997 and completed in 1998, raised profound implications for U.S. national security. For instance, it compli- cated subsequent nuclear disarmament discussions because of confl icts of interest between the privatized fi rm and national security. To many, this privatization appeared to be a case of the ideology of privatization gone amok—government had lost the sense of balance between the private and public sector required to make a mixed economy work.

As recently as 2005, however, there was a major privatization eff ort in the United States: to privatize a substantial part of the public old age retirement program (Social Security), following similar eff orts in Chile, the United Kingdom, and other countries. The global fi nancial crisis exposed major problems with these initiatives, including increased insecurity for the elderly. Other privatization eff orts, such as roads in Mexico and rail- roads in the United Kingdom, similarly encountered major problems. With the 2008 crisis, many governments were forced to take a much more active role in the economy, in some cases nationalizing, or renationalizing, private enterprises (especially banks). In many resource-rich countries, private oil

13Thinking Like a Public Sector Economist

and mining companies had driven such unfair bargains that governments demanded new contracts, or again, in some cases, the nationalization, or even renationalization, of the oil or gas fi elds or mines.

THINKING LIKE A PUBLIC SECTOR ECONOMIST

Economists study scarcity—how societies make choices concerning the use of limited resources. They inquire into four central economic questions:

1. What is to be produced?

2. How is it to be produced?

3. For whom is it to be produced?

4. How are these decisions made?

Like all economists, public sector economists are concerned with these fundamental questions of choice, but their focus is the choices made within the public sector, the role of the government, and the ways gov- ernment aff ects the decisions made in the private sector.

1. What is to be produced? How much of our resources should be devoted to the production of public goods, such as defense and highways, and how much of our resources should we devote to the production of pri- vate goods, such as cars, TV sets, and video games? We often depict this choice in terms of the production possibilities schedule, which traces the various amounts of two goods that can be produced effi ciently with a given technology and resources. In our case, the two goods are public goods and private goods. Figure 1.1 gives the various possible combina- tions of public goods and private goods that society can produce.

Society can spend more on public goods, such as national defense, but only by reducing what is available for private consumption. Thus, in moving from G to E along the production possibilities schedule, pub- lic goods are increased, but private goods are decreased. A point such as I, which is below the production possibilities schedule, is said to be ineffi cient: society could get more public goods and more private goods. A point such as N, which is above the production possibilities schedule, is said to be infeasible: it is not possible, given current resources and technology, to have that quantity of public goods and that quantity of private goods at the same time.

14 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

2. How should it be produced? Under this question are subsumed such deci- sions as whether to produce privately or publicly, to use more capital and less labor or vice versa, or to employ energy-effi cient technologies.

Other issues are also subsumed under this second question. Gov- ernment policy aff ects how fi rms produce the goods they produce: environmental protection legislation restricts pollution by fi rms; pay- roll taxes that fi rms must pay on the workers they employ may make labor more expensive and thus discourage fi rms from using production techniques that require much labor.

3. For whom is it to be produced (the question of distribution)? Government decisions about taxation or welfare programs aff ect how much income diff erent individuals have to spend. Similarly, the government must decide what public goods to produce. Some groups will benefi t from the production of one public good, others from another.

4. How are choices made? In the public sector, choices are made collec- tively. Collective choices are the choices that a society must make together—for instance, choices concerning its legal structure, the size of its military establishment, its expenditures on other public goods, and so on. Texts in other fi elds of economics focus on how individuals make their decisions concerning consumption, how fi rms make their decisions concerning production, and how the price system works to ensure that the goods demanded by consumers are produced by fi rms.

SOCIETY’S PRODUCTION POSSIBILITIES SCHEDULE

This depicts the maximum level of private goods that society

can enjoy for each level of public goods. If society wishes to enjoy more public goods, it must give

up some private goods.

FIGURE 1.1

Public goods

Production possibilities

schedule

G N

E

I

Private goods

15Thinking Like a Public Sector Economist

Collective decision making is far more com- plicated, for individuals often disagree about what is desirable. After all, just as some indi- viduals like chocolate ice cream and some like vanilla ice cream, some individuals get greater enjoyment out of public parks than do others. But whereas with private goods the individual who likes chocolate ice cream can simply buy chocolate ice cream and the individual who likes vanilla ice cream can buy vanilla ice cream, with public goods we must make decisions together. Anyone who has lived in a family knows something about the diffi culties of collective decision making (should we go to the movies or go bowling?).

Public decision making is far more com- plex. Increasingly, though, we have come to understand that decision making in the private sector, especially in large corporations, is far more complex than depicted in simplistic models of fi rms that have a single owner. Within the corporation, there are large diff erences in views about what should be done, partially motivated by judgments about the consequences of diff erent actions (how well will a product sell?), but also by diff erences in “values”—the extent, for instance, to which the fi rm should focus on the short term or the long. One of the objectives of public sector economics is to study how collective choices (or, as they are sometimes called, social choices) are made in democratic societies.

The recognition of this divergence of views is important in itself. It should make us wary of expressions such as “It is in the public interest” or “We are concerned with the good of society.” Diff erent policies may be good for diff erent individuals. One should carefully specify who will benefi t from a given policy and who will be harmed by it.

ANALYZING THE PUBLIC SECTOR

In addressing each of the fundamental economic questions, there are four general stages of analysis: describing what the government does, analyzing the consequences of government action, evaluating alternative policies, and interpreting the political forces that underlie the decisions government makes.

KEY ECONOMIC QUESTIONS • What is produced?

Public or private goods? • How is it produced?

Within the public sector or the private? • For whom should it be produced?

Taxes affect amount different individuals have to spend.

Different government programs benefi t different groups.

• How are these decisions made?

How are collective decisions, such as those concerning the supply of publicly provided goods and taxes, made?

16 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

1. Knowing what activities the public sector engages in and how these are organized. The complexity of the government’s operations is so great that it is diffi cult to assess what its total expenditures are and what they go for. The budget of the U.S. federal government alone is a docu- ment that is more than 1,000 pages long, and within the budget, activ- ities are not easily compartmentalized. Some activities are undertaken in several diff erent departments or agencies. Research, for instance, is funded through the Department of Defense, the National Science Foun- dation, the National Institutes of Health, and the National Aeronautics and Space Administration, among others. Furthermore, a department like the Department of Health and Human Services undertakes a myr- iad of activities, some of which are only vaguely related to others.

Further, taxes and expenditures occur at several diff erent levels: in some places, individuals pay not only federal and state taxes but also sep- arate taxes to their school district, their township, their county, the juris- dictions that provide their water and sewage, and their public library.

2. Understanding and, insofar as possible, anticipating the full conse- quences of these governmental activities. When a tax is imposed on a corporation, who bears the tax? At least part of the tax will be passed on to consumers through higher prices, or on to employees as wages fall. What are the consequences of the government’s changing the age of retirement for Social Security? Of a tax credit or deduction for col- lege tuition? Will universities respond by raising tuition so a college education will be hardly more aff ordable than before? Will a tax on the global income of American corporations (eliminating the provi- sion that allows them to postpone taxes until they bring their income back to the United States) reduce their incentive to outsource jobs and increase government revenue? Or will it simply encourage corpora- tions to move their headquarters abroad?

The consequences of government policies are often too compli- cated to predict accurately, and even after a policy has been intro- duced, there is often controversy about what its effects are. This book attempts not only to present all sides of some of the major controversies, but also to explain why such disagreements have persisted, and why they are difficult to resolve.

3. Evaluating alternative policies. To do this, we need not only to know the consequences of alternative policies, but also to develop criteria for eval- uation. First we must understand the objectives of government policy, and then we must ascertain the extent to which a particular proposal meets (or is likely to meet) those criteria.

17Thinking Like a Public Sector Economist

Many government programs have mul- tiple objectives. For example, the United States has a program to clean up hazardous waste sites, not only to protect health, but also because such sites may be an impedi- ment to economic development. Some pol- icies are better at achieving one objective; others may be better at achieving others. We need a framework for decision making in such situations: How do we think systemati- cally about the trade-off s in evaluating alter- native policies?

4.  Interpreting the political process. Collective decisions such as whether to subsidize farmers or to build a supercollider, or how much to spend on education, get made through political processes. How can we explain which alternatives are chosen? Economists identify the various groups that benefi t or lose from a government program and analyze the incentives facing these groups to attempt to mobilize the political process to promote outcomes favorable to them. They also ask how the structure of government—the “rules of the game” (the rules by which Congress works, whether the President can veto specifi c items within a bill or only the bill as a whole, and so on)—aff ects the out- comes. In many quarters, there is a concern that the Supreme Court decision in Citizens United v. Federal Election Commission, 558 U.S. 310 (2010), which seemingly gave corporations unbridled opportunities to make campaign contributions, may have tilted the political playing fi eld toward corporations. Then they try to push the question further: What determines how the rules of the game are chosen? In addressing these questions, economics and political science merge. Economists, however, bring a distinct perspective to the analysis: they emphasize the importance of economic incentives in the behavior of participants in the political process, and therefore of economic self-interest in determining outcomes.

ECONOMIC MODELS

A central part of the analysis of the economics of the public sector is understanding the consequences of diff erent policies. Economists, how- ever, sometimes disagree over what those consequences will be. The standard way that science has found to test competing theories is to

ANALYZING THE PUBLIC SECTOR • Knowing what activities the public sector engages

in and how they are organized

• Understanding and anticipating the full conse- quences of these government activities

• Evaluating alternative policies

• Interpreting the political process

18 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

carry out an experiment. With luck, the results of the experiment will bear out the predictions of only one theory while discrediting others. However, economists ordinarily do not have the possibility of doing controlled experiments. Instead, what economists can observe are the uncontrolled experiments that are being done for us in diff erent markets and in diff erent time periods; the historical evidence, unfortunately, often does not permit us to resolve disagreements about how the econ- omy behaves.

To analyze the consequences of various policies, economists make use of what are called economic models. Just as a model airplane attempts to replicate the basic features of an airplane, so too a model of the economy attempts to depict the basic features of the economy. The actual economy is obviously extremely complex; to see what is going on, and to make predictions about what the consequences of a particular change in policy will be, one needs to separate out the essential from the inessential features. The features on which one decides to focus in constructing a model depend on the questions one wishes to address. The fact that models make simplifying assumptions—that they leave out many details—is a virtue, not a vice. An analogy may be useful. In going on a long road trip, you may use several maps. One map, depicting the interstate highway system, provides an overview, enabling you to

MUSGRAVE’S THREE BRANCHES

R ichard Musgrave, one of the great public fi nance economists of the twentieth century, thought of the government as having three economic branches. The fi rst was the stabilization branch; its responsibility was to ensure that the economy remained at full employment with stable prices. How this was to be done was the principal subject of courses in macroeconomics. The second branch was the allocation branch. Here, the govern- ment intervened in how the economy allocated its resources. It did this directly, by buying goods such as defense and education, and indirectly, through taxes and subsidies, which encouraged some activ- ities and discouraged others. The third branch, the

distribution branch, was concerned with how the goods that were produced by society were dis- tributed among its members. This branch was con- cerned with issues such as equity and the trade-offs between equity and effi ciency. The economics of the public sector focuses on the latter two branches, though the issues arise in other economic courses as well, such as those that deal with regulation.

Today, we recognize that government activities in all three branches are intertwined and cannot be neatly compartmentalized in the way that Musgrave envisaged. Still, his “three branches” provide a con- venient way of looking at the myriad of activities in which the government is engaged.

19Thinking Like a Public Sector Economist

see how to get from the general area where you are to the general area where you wish to go. You then use detailed maps to see how to get from your point of origin to the expressway, and from the expressway to your fi nal destination. If the interstate highway map showed every street and road in the country, it would be so large that its usefulness would be limited; the extra detail, though important for some purposes, would simply get in the way.

All analysis involves the use of models, of simple hypotheses concern- ing how individuals and fi rms will respond to various changes in govern- ment policy, and how these responses will interact to determine the total impact on the economy. Everyone—politicians as well as economists—uses models in discussing the eff ects of alternative policies. The diff erence is that economists attempt to be explicit about their assumptions, and to be sure that their assumptions are consistent with one another and with the available evidence.

NORMATIVE VERSUS POSITIVE ECONOMICS

In their analysis, economists also try carefully to identify the points in their analysis at which values enter. When they describe the economy and construct models that predict either how the economy will change or the eff ects of diff erent policies, they are engaged in what is called positive economics. When they attempt to evaluate alternative policies, weighing the various benefi ts and costs, they are engaged in what is called normative economics. Positive economics is concerned with what “is,” with describing how the economy functions; normative economics deals with what “should be,” with making judgments about the desirability of various courses of action. Normative economics makes use of positive eco- nomics. We cannot make judgments about whether a policy is desirable unless we have a clear picture of its consequences. Good normative eco- nomics also tries to be explicit about precisely what values or objectives it is incorporating. It tries to couch its statements in the form “If  these are your objectives . . . , then this is the best possible policy.”

Consider the positive and normative aspects of a proposal to levy a $1-per-case tax on beer. Positive economics would describe the eff ect the tax would have on the price of beer—would the price rise by the full $1, or would producers absorb some of the tax? On the basis of that analysis, economists would go on to predict how much beer consumption would be reduced, and who would be aff ected by the tax. They might fi nd, for instance, that because lower-income individuals spend a larger fraction of their income on beer, these people would be aff ected proportionately more.

20 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

Studies may have indicated that there is a systematic relationship between the quantity of beer consumed and road accidents. Using  this  informa- tion, economists might attempt to estimate how the beer tax would aff ect the number of accidents. These steps are all part of describing the full consequences of the tax, without making judgments. In the end, however, the question is, should the tax be adopted? This is a normative question; in responding to it, economists will weigh the benefi ts of the tax revenue, the distortions it induces in consumption, the inequities caused by the fact that proportionately more of the tax is borne by lower-income indi- viduals, and the lives saved in road accidents. Furthermore, in evaluating the tax, economists will also want to compare it with other ways of rais- ing similar amounts of revenue.

This example is typical of many such situations that we face in eco- nomic policy analysis. Through positive economic analysis, we identify some gainers (the roads are safer) and some losers (consumers who pay higher prices, producers who have lower profi ts, workers who lose their jobs). Normative economics is concerned with developing systematic pro- cedures by which we can compare the gains of those who are better off with the losses of those who are worse off , to arrive at some overall judg- ment concerning the desirability of the proposal.

The distinction between normative statements and positive state- ments arises not only in discussions of particular policy changes but also in discussions of political processes. For instance, economists are concerned with describing the consequences of the majority vot- ing system in the United States, in which the proposal that gets the majority of votes wins. A major group of economists, led by Nobel Prize winner James Buchanan of George Mason University, has focused on describing the impact of political processes on social choices (hence, these economists are often referred to as the social choice school). What will be the consequences—in terms of patterns or levels of taxa- tion or expenditure, or the speed with which these change in response to changed circumstances—of requiring a two-thirds majority for increments in public expenditures exceeding a certain amount? What will be the consequences of increasing politicians’ pay? Of restricting private contributions to political campaigns? Of imposing campaign spending limits, or a variety of other proposals for reforming the financing and conduct of political campaigns? Of public support for political campaigns?

But economists are also concerned with evaluating alternative political processes. Are some political processes better, in some senses, than others? Are they more likely to produce consistent choices? Are some political pro- cesses more likely than others to yield equitable or effi cient outcomes?

21Disagreements Among Economists

DISAGREEMENTS AMONG ECONOMISTS

Unanimity is rare in the central questions of policy debate. Some indi- viduals think affi rmative action or bilingual education is desirable; some do not. Some think that the income tax should be more progressive (i.e., that wealthy individuals should pay a higher percentage of their income in taxes, while poor individuals should pay a lower percentage); some believe it should be less progressive. Some agree with the decision to pro- vide a tax credit for college tuition; some believe the money could have been spent in better ways, including ways that are more eff ective in pro- viding education for the poor. Some believe that capital gains should be taxed like any other form of income; others think capital gains should receive preferential treatment. One of the central concerns of policy anal- ysis is to identify these sources of disagreement.

Disagreements arise in two broad areas. Economists disagree about the consequences of policies (about the positive analysis) and about values (about the normative analysis).

DIFFERENCES IN VIEWS ON HOW THE ECONOMY BEHAVES

As we have seen, the fi rst question economists ask in analyzing any pol- icy is, what are its full consequences? In answering this question, they have to predict how households and fi rms will react. In 1696, for exam- ple, England imposed a tax on windows, under the Act of Making Good the Defi ciency of the Clipped Money. At the time windows were a lux- ury, and the houses of the wealthy had more windows than those of the poor. The window tax could be thought of as a rough substitute for an income tax, which the government did not have the authority to impose. The government should have asked, how much do people value light in their houses? One could imagine a policy debate among the king’s advisers about what fraction of the population would value light so little that, rather than pay the tax, they simply would survive with window- less houses. At the time, there were no statistical studies on which the king could rely. (In fact, many people did not value light highly, so the government raised less revenue than anticipated, and more homes were darker than anticipated.)

22 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

Today, economists often disagree about the best model for describing the economy, and even after agreeing about the nature of the economy, they may disagree about quantitative magnitudes. For instance, they may agree that increased taxes discourage work, but disagree about the size of the eff ect.

A standard model that many economists employ assumes that there is perfect information and perfect competition—every fi rm or individual is so small that the prices it pays for what it buys and receives for what it sells do not depend at all on what it does. Although most economists rec- ognize that information and competition are both imperfect, some believe that the model of perfect information and perfect competition provides a close enough approximation to reality to be useful; others believe that—at least for some purposes, such as the health care market—the deviations are large, and that policy must be based on models that explicitly incorpo- rate imperfect information and competition.

We cannot resolve these disagreements, but what we can do is to show how and when diff erent views lead to diff erent conclusions.

Even when economists agree about the kind of response a particular policy will elicit, they may disagree about the magnitude of the response. This is one of the sources of dispute about the consequences of President Obama’s Aff ordable Care Act of 2010. Most economists believe that pro- viding health insurance to more people will lead individuals who previ- ously did not have insurance to consume more health care—one of the motivations of the program is that many of those without health insur- ance are getting inadequate care. However, there is disagreement about how much more they will consume. There is even some disagreement over whether total consumption of health care services by the uncovered will decrease: many of these people previously got medical care, but only in an ineffi cient way (e.g., in high-cost emergency rooms) and only after their illnesses had been allowed to fester, increasing the overall cost of treatment. There is also disagreement over how much of the expense of increased coverage will be off set by initiatives to improve effi ciency in the delivery of health services. The answers to these questions dramatically aff ect the projected cost of this new law.

Although a central concern of modern economics is ascertaining the magnitude of the response of, say, investment, to an investment tax credit, of consumption to a change in the income tax rate, of savings to an increase in the interest rate, and so on, it is an unfortunate fact that var- ious studies, using diff erent bodies of data and diff erent statistical tech- niques, come up with diff erent conclusions. As economists obtain more data and develop better techniques for analyzing the limited available data, some of these disagreements may be resolved.

23Disagreements Among Economists

PUBLIC SECTOR ECONOMICS AND THE GLOBAL ECONOMIC CRISIS

A s the world strives to recover from the recent global economic crisis, disagree-ments about how economies function and how we want them to function (positive and nor- mative economics) have intensifi ed. The impact of the crisis has been deep and sustained: the United States experienced its worst recession since the Great Depression of the 1930s; the U.S. unemploy- ment rate remains persistently high, well in excess of levels that would represent full employment; the global economy contracted in 2009 for the fi rst time since the International Monetary Fund (IMF) began collecting data in 1970. The public sector response to this crisis has also been unprecedented: in addition to injections of substantial liquidity by the world’s central banks and widespread publicly fi nanced recapitalization of many of the world’s largest fi nancial institutions, countries have also enacted large fi scal stimulus packages, totaling 13.5 percent of 2007 GDP in China (4 trillion yuan), 6.8 percent in the United States (the Bush Eco- nomic Stimulus Act of 2008 of $168 billion and the Obama American Recovery and Reinvestment Act

of 2009 of $787 billion), and 5.5 percent in Japan (¥27 trillion). Although the stimulus measures did not restore the economy to full employment, U.S. unemployment peaked at a far lower level than it otherwise would have, and without government help, the banking system may well have collapsed, bringing on an even deeper downturn. In Europe, concerns over looming budget defi cits resulted in widespread cutbacks, and after a shallow recovery, Europe sank back into recession, with unemploy- ment reaching record levels.

The global crisis has resulted in a major re-examination of the role of government. There is a broad consensus that fi nancial markets took on too much risk on their own and imposed huge costs on the economy, and that markets on their own recover too slowly. Strong government fi nancial regulation had succeeded in preventing fi nancial crisis for decades, after they were enacted in the 1930s, in response to the Great Depression, but these regu- lations were stripped away beginning in the 1980s. And government intervention, even if imperfectly designed, has played a role in the recovery.

DISAGREEMENT OVER VALUES

Whereas the two previous sources of disagreement—concerning the best model for describing the economy and about quantitative magnitudes, such as the size of the response of savings to interest rates—arise within positive economics, the fi nal source of disagreement lies within norma- tive economics. Even if there is agreement about the full consequences of some policy, there may be disagreement about whether the policy is desirable. As has already been noted, there are frequently trade-off s: a policy may increase national output but also increase inequality; it may increase employment but also increase infl ation; it may benefi t one group

24 CHAPTER 1 DEFINING PUBLIC SECTOR RESPONSIBILITIES

but make another group worse off . Any policy, in other words, may have some desirable consequences and some undesirable consequences. Indi- viduals may weigh these consequences in diff erent ways, some attaching more importance to price stability than to unemployment, others attach- ing more importance to growth than to inequality.

On questions of values, there is no more unanimity among economists than there is among philosophers. This book presents the major views and assesses some of the criticisms that have been leveled against each.

SUMMARY

1. In mixed economies, such as the United States, economic activity is carried on by both private enterprise and the government.

2. Since the time of Adam Smith, economic theory has emphasized the role of private markets in the effi cient supply of goods. However, economists and others have come to recognize important limitations in the ability of the private sector to produce effi cient outcomes and meet certain basic social needs. The attempt to correct these market failures has led to the growth of government’s role in the market economy.

3. The government, however, also has its limitations when intervening to mitigate market failures. These government failures sometimes result in government programs with unintended adverse consequences.

4. The United States has a federal government struc- ture, with certain activities primarily the respon- sibility of states and localities (e.g., education) and other activities primarily the responsibility of the federal government (e.g., defense).

5. Economics is the study of scarcity—how resources are allocated among competing uses. Public sec- tor economics focuses on choices between the public and private sectors and choices within the public sector. It is concerned with four basic

issues: what gets produced, how it gets produced, for whom it gets produced, and the processes by which these decisions are made.

6. In studying the public sector, positive econom- ics looks at the scope of government activity and the consequences of various government policies. Normative economics attempts to evaluate alter- native policies that might be pursued.

7. Disagreements about the desirability of policies are based on disagreements about the appropriate assumptions for describing the economy, such as how competitive the economy actually is, disagree- ments about how strongly the economy will respond to policy initiatives, and disagreements about values.

KEY CONCEPTS

Deregulation

Economic models

Full Employment Act of 1946

Laissez faire

Mercantilists

Mixed economy

Normative economics

Positive economics

Privatization

Production possibilities schedule

REVIEW AND PRACTICE

25Review and Practice

QUESTIONS AND PROBLEMS

1. Consider the following discussion of a program of price supports for farmers:

a. The objective of our farm program is to ensure that all farmers have a reasonable standard of living. The way it does this is to ensure that farmers receive fair prices for their commod- ities. It is no more right that farmers should produce for substandard prices than that workers should work for substandard wages.

b. Our farm program has been a failure. The ben- efi ts of the price subsidies accrue largely to large farmers (because they produce more). Many farmers still have incomes below the poverty line. The high prices have induced increased production, which has meant high costs for the government. Acreage restrictions have had only limited eff ect because farmers have kept their best land in production. Direct grants to farmers would be preferable to our price support program.

Which of the statements in this discussion are normative, and which are positive? (The fact that you disagree with a normative statement or that you think a particular “positive” statement is inaccurate does not change the nature of the statement.)

Identify the sources of disagreement: Are they due to diff erences in values and objectives? To diff erences in perceptions about the nature of the economy? Or to a failure on one (or the other) side of the debate to take into account the full consequences of the government’s action?

2. For each of the following programs, identify one or more “unintended” consequences:

a. Rent control

b. Minimum wages

c. Medicare (free hospital care to the aged)

d. Improved highways making suburbs more accessible to the city

e. Forced integration of central-city schools

f. Agricultural price supports

g. Lowering the speed limit to 55 miles an hour to save gasoline

h. Providing health insurance to children who currently are underinsured

i. Banning advertising of cigarettes (Hint: Con- sider the consequences of increased life spans for the Social Security system.)

j. National testing standards for schools

3. There has been considerable concern that our Social Security (old age and survivors insur- ance) program is not adequately fi nanced: with expected birth rates, death rates, and increases in payroll tax collections, the current level of ben- efi ts can be sustained only with increases in tax rates. Some believe that the appropriate response is to reduce the current level of benefi ts, others that the appropriate response is to increase taxes in the future. Still others, worried about the eff ects of even higher tax rates but believing that lowering the benefi ts of those currently receiving Social Security would be unfair, argue that bene- fi ts should be cut in the future.

In this discussion, separate the positive state- ments from the normative statements. To what extent are the disagreements attributable to dif- ferences in views of the economy?

26

MEASURING PUBLIC SECTOR SIZE

2

A central topic of debate in the United States, and in other mixed econ- omies, is the appropriate size of the public sector. Some believe that the public sector is too large. They are skeptical of government’s ability to solve social and economic problems because of the kinds of government failures we discussed in Chapter 1—for example, government’s limited control over private market responses. Or they may believe in limited gov- ernment on philosophical grounds, because of a fear that big government undermines economic and political freedom.1 Others believe that the pub- lic sector is too small. In their view, greater government spending could solve the problems of blighted inner cities and inadequate schools. Whatever view you take, there is no doubt that the U.S. government today is far larger than it was before World War I. In 2010, tax revenues (and other nontax receipts)2 collected at all levels of government were $3.6  trillion, or 25 percent of total U.S. production, and government

1�A leading proponent of this view, a form of libertarianism, is Robert Nozick. His ideas are summarized in the preface of his Anarchy, State, and Utopia (Oxford, UK: Basil Blackwell, 1974). See also Milton Friedman, Capitalism and Freedom (Chicago: University of Chicago Press, 1962). 2�Nontax receipts include, for instance, fees the government receives for various services.

27What or Who Is the Government?

expenditures were 34 percent.3 By contrast, in 1913, prior to World War I, taxes and government expenditures were less than 10 percent of total production. How do we account for this dramatic change in the size of government? What does the government spend all this money on? This chapter gives an overview of the scope of the U.S. public sector and how it has broadened over time. It also shows the ways in which gov- ernment actions aff ect private markets. Chapter 4 takes up the economic rationale for government intervention in markets. These chapters will not resolve the debate over whether the U.S. public sector is too big or too small, but they provide a basis for formulating a reasonable position on this issue and provide a context for comparison with other countries.

WHAT OR WHO IS THE GOVERNMENT?

Throughout this chapter we have referred to “the government.” But what precisely is the government? We all have some idea about what institu- tions are included: Congress and state and local legislatures, the president and state governors and mayors, the courts, and a host of alphabet-soup agencies, such as the FTC (Federal Trade Commission) and the IRS (Internal Revenue Service). The United States has a federal governmental structure—that is, governmental activities take place at several levels: federal, state, and local. The federal government is responsible for national defense, the post offi ce, the printing of money, and the regulation of inter- state and international commerce, whereas the states and localities have traditionally been responsible for education, police and fi re protection, and the provision of other local services, such as libraries, sewage, and garbage collection. Even though the Constitution asserts that all rights not explicitly delegated to the federal government reside with the states and the people, commonly referred to as subsidiarity, the Constitution has proven to be a suffi ciently fl exible document that the exact boundaries are ambiguous. Although education is primarily a local responsibility, the federal government has become increasingly involved in its support. The constitutional provision giving the federal government the right to con- trol interstate business has provided the basis for federal regulation of almost all businesses, as almost all businesses are involved in one way or another, in interstate commerce.

3�Offi ce of Management and Budget, Budget of the U.S. Government, Fiscal Year 2012, Historical Tables, Tables 15.1 and 1.2.

1. What are the principal activities of government?

2. What does government spend its money on? How have these expenditure patterns changed over time, and how do they compare across countries? Which expenditures occur at the national level, and which at the state and local level?

3. How does the government fi nance its expenditures? How do the sources of tax revenues diff er between the national government and state and local gov- ernments? How have they changed over time?

FOCUS QUESTIONS

28 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

At the local level, there are frequently several separate governmental structures, each having the power to levy taxes and the responsibility for administering certain programs. In addition to townships and counties, there are school districts, sewage districts, and library districts. In 2007, there were 90,000 such governmental entities in the United States, down from 155,000 in 1942.4

The boundaries between what are public institutions and what are not are often unclear. When the government sets up a corporation, a pub- lic enterprise, is that enterprise part of the “government”? For instance, Amtrak, which was set up by the federal government to run the nation’s interstate passenger railway services, receives subsidies from the fed- eral government but otherwise is run like a private enterprise. Matters become even more complicated when the government is a major stock- holder in a company, but not the only stockholder. For instance, prior to 1987 the British government owned up to 50 percent of the shares of British Petroleum.

What distinguishes the institutions that we have labeled as “govern- ment” from private institutions? There are two important diff erences. First, in a democracy the individuals who are responsible for running pub- lic institutions are elected, or are appointed by someone who is elected (or appointed by someone who is appointed by someone who is elected . . .). The “legitimacy” of the person holding the position is derived directly or indirectly from the electoral process. In contrast, those who are respon- sible for administering General Motors are chosen by the shareholders of General Motors, and those who are responsible for administering private foundations (such as the Rockefeller and Ford foundations) are chosen by a self-perpetuating board of trustees.

Second, the government is endowed with certain rights of compulsion that private institutions do not have. The government has the right to force you to pay taxes (and if you fail, it can confi scate your property and/ or imprison you). The government has the right to seize your property for public use provided it pays you just compensation (this is called the right of eminent domain).

Not only do private institutions and individuals lack these rights, but the government actually restricts the rights of individuals to give to oth- ers similar powers of compulsion. For instance, the government does not allow you to sell yourself into slavery.

In contrast, all private exchanges are voluntary. I may need your prop- erty to construct an offi ce building, but I cannot force you to sell it. I may think that some deal is advantageous to both of us, but I cannot force you to engage in the deal.

4�U.S. Census Bureau, Statistical Abstract of the United States, 2011, Table 426.

29Types of Government Activity

Government is thus fundamentally diff erent from other institutions in our society. It has strengths—its ability to use compulsion means that it may be able to do some things that private institutions cannot do. But it also has weaknesses, as we discuss in greater detail in later chapters. Understanding these strengths and weaknesses is an essential part of assessing what should be the role of the government in our mixed economy, and of deter- mining how government can most eff ectively fulfi ll that role.

TYPES OF GOVERNMENT ACTIVITY

A primary role of government is to provide the legal framework within which all economic transactions occur. Beyond that, the activities of gov- ernment fall into four categories: (1) the production of goods and services; (2) the regulation and subsidization or taxation of private production; (3)  the purchase of goods and services, from missiles to the services of street cleaners; and (4) the redistribution of income—that is, payments like unemployment benefi ts to particular groups of individuals that enable them to spend more than they could otherwise. Payments that transfer money from one individual to another—but not in return for the provision of goods or services—are called transfer payments.

These four categories—production, regulation, purchase, and redistribution—are simply a convenient way of grouping the vast array of government activities. However, they do not correspond to the way the federal government organizes its budget or divides responsibilities among its various departments—Commerce, Health and Human Services, Interior, and so on. Moreover, government activities are undertaken at the state and local levels as well as at the federal level, with the relative importance of state, local, and federal expenditures of various types hav- ing changed over time.

A fi nal complication is that the nature of some government expendi- tures is ambiguous. For example, government subsidies to small farmers could be considered a production subsidy or a redistributive (transfer) payment. Pension payments to military retirees are often counted as transfer payments, but they are more appropriately treated as part of the cost of national defense, just as the pension costs of a private fi rm are counted among its labor costs.

Thus, the task of constructing a quantitative description of the gov- ernment’s activities is a formidable one.

30 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

PROVIDING A LEGAL SYSTEM

An important activity of the government, but one that accounts for very little expenditure, is the establishment of the legal framework within which fi rms and individuals can engage in economic interactions. Econo- mists and philosophers often try to imagine what life would be like in the complete absence of government. Without laws defi ning property rights, for instance, only the exercise of force would stop one individual from stealing from another. Without the ability to protect property, individuals would have little incentive to accumulate assets. Needless to say, eco- nomic activities would be severely restricted.

The U.S. legal system does much more than just protect property rights. It enforces contracts between individuals. It also imposes restric- tions on the kinds of contracts that are legally enforceable. Our bank- ruptcy laws limit the liability of investors. Product liability laws have an important eff ect on the quality of goods produced. Antitrust laws attempt to encourage competition among fi rms: they restrict mergers, acquisi- tions, and unfair business practices.

The eff ects of our criminal justice system are pervasive, but expendi- tures for running it are relatively small: about 5 percent of total govern- ment expenditures.5

GOVERNMENT PRODUCTION

The U.S. government undertakes certain types of production directly. Much of this is similar to corresponding activities carried out by private fi rms. For instance, both private and government enterprises produce and sell electricity (the most famous of the latter is, perhaps, the Tennessee Valley Authority). In addition, under the Constitution, the federal govern- ment takes responsibility for running the postal service and for printing money.6

At the local level, many communities provide water and collect gar- bage, services that in other communities are provided by private fi rms. Most elementary and secondary school students go to public schools— schools run by the government—although others go to private schools,

5�Based on 2007 data, the most recent available. See Offi ce of Management and Budget, Budget of the U.S. Government, Fiscal Year 2012, Historical Tables, Table 3.1; and Bureau of Justice Statistics, “Direct Expenditure by Level of Government, 1982–2007,” in Justice Expenditure and Employment Extracts, 2011. 6�Although the U.S. Postal Service has a monopoly on the delivery of fi rst-class mail, private carriers, such as United Parcel Service, Federal Express, and others, play a major role in the delivery of parcels and express mail.

31Types of Government Activity

some of which are run by nonprofi t organizations like churches and a few run on a for-profi t basis.

Comparing the public and private sectors in various countries, we see that some industries frequently fall within the public sector, whereas other industries seldom do. Agriculture and retail trade are seldom in the public sector. On the other hand, in most countries, at least part of the radio and TV broadcasting industry lies in the public sector. In many countries, the banking system is at least partially owned and operated by the government; in the United States it is closely regulated but pri- vately owned.7

The line between public and private production shifts over time. During the past two decades in Europe, many countries have converted public enterprises into private enterprises, a process called privatization. (The process of converting private enterprises to government enterprises is called nationalization.) For instance, the British government has pri- vatized enterprises in industries ranging from telecommunications to energy, automobiles, aerospace, and steel. In France, a wave of privatiza- tion, which began in 1986, included the privatization of enterprises that had been nationalized earlier in the decade when the socialist party was in power.

The distinction between public and private production has also become more blurred. Public enterprises are sometimes corporatized or commer- cialized, so they are managed like private enterprises although they are still government owned, as are many of Singapore’s government-linked companies (GLCs). This can also be a transitional step in the privatization process. There are also many hybrid models characterized by coopera- tion between the public and private sectors, such as public–private part- nerships (PPP) and private sector participation (PSP) in the provision of public goods and services.8

For technical reasons, the best way to measure the size of govern- ment production is to look at employment, as in Figure 2.1. In 2009, public employees, including public education and the armed forces, represented 16.9 percent of total employment. This was almost double the percent- age in 1929 when it was 8.9 percent of the labor force. The fi gure shows

7�The Federal Reserve Banks, which are responsible for the management of the banking system, are pub- licly owned. Most of their profi ts are turned over to the U.S. Treasury. In 2009, their comprehensive income prior to distribution was $53.4 billion. (The Federal Reserve Banks Combined Financial Statements as of and for the Years Ended December 31, 2009 and 2008 and Independent Auditors’ Report.) 8�For more on privatization, see William L. Megginson et al., “The Financial and Operating Performance of Newly Privatized Firms: An International Empirical Analysis,” Journal of Finance 49, no. 2 (1994), and Pierre Guislain, The Privatization Challenge: A Strategic, Legal, and Institutional Analysis of International Experience (Washington, DC: World Bank, 1997). For more on alternatives to privatization, see José A. Gómez-Ibáñez, “Alternatives to Infrastructure Privatization Revisited: Public Enterprise Reform from the 1960s to the 1980s,” Policy Research Working Paper 4391 (Washington, DC: World Bank, 2007), and Darrin Grimsey and Mervyn K. Lewis, Public Private Partnerships: The Worldwide Revolution in Infra- structure Provision and Project Finance (Cheltenham, UK: Edward Elgar, 2004).

32 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

a marked increase in the ratio of public employment from 1929 through 1936 (both in the Hoover and Roosevelt administrations), a burst of pub- lic employment during World War II, and a return to prewar levels by 1947. Although there was a slight decrease in the pace of growth during the Eisenhower years, employment in the public sector did not begin to decline until the Nixon and Ford administrations. This decline continued until the current global economic crisis. In fact, by 2007 the federal gov- ernment’s share of total employment had fallen to 3.2 percent, the same as it was in 1932, before the New Deal.

It is also important to note the variations in the relative roles played by the federal, state, and local governments, as suggested by the bottom line in Figure 2.1. Comparing it to the top line, we see that total govern- ment employment and federal government employment do not always move together. Federal employment as a percentage of total employ- ment declined in the early 1970s, but this decline was off set by the rise in employment at the state and local level. It is important to bear this in mind: reductions in federal expenditures or employment do not of

GOVERNMENT EMPLOYMENT AS A

PERCENTAGE OF TOTAL EMPLOYMENT, 1929–2009

Government employment as a percentage of all employ- ment provides a view of the

government’s role as producer. Employment is defi ned here

as the number of full-time equivalent employees.

FIGURE 2.1

SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National

Economic Accounts, Tables 6.5A–6.5D. 1925 0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

22%

24%

26%

28%

30%

32%

34%

1935 1945 1955 1965 1975 1985 1995 2005

Federal government employment as a percentage of total employment

State and local government employment as a percentage of total employment

Total government employment as a percentage of total employment

33Types of Government Activity

themselves necessarily imply a reduction in government expenditures or employment. More of a burden may simply be placed on states and localities.

GOVERNMENT’S INFLUENCE ON PRIVATE PRODUCTION

In industries in which the government is neither a producer nor a con- sumer, it may nevertheless have a pervasive eff ect on the decisions of private producers. This infl uence is exercised through subsidies and taxes—both direct and indirect—and through regulations. There are many motives for such government infl uence. There may be dissatisfaction with particular actions of fi rms, such as pollution. There may be concern about the monopoly power of some fi rms. Special interest groups may convince Congress that they are particularly deserving of help. Private markets may fail to provide certain goods and services that are felt to be important.

SUBSIDIES AND TAXES�Government subsidizes private production in three broad ways: direct payments to producers, indirect payments through the tax system, and other hidden expenditures. The most exten- sive of the U.S. government’s subsidy programs is for agriculture. Direct payments to farmers rose precipitously during the 1980s, from $1.3 billion in 1980 to a peak of $16.7 billion in 1987, when they began a steady decline until they reached $7.3 billion in 1995. In 1987, direct payments amounted to 37 percent of income from wheat, 40 percent of income from rice, and 20 percent of income from all crops. At least one of every fi ve dollars in farm income was a gift from the government. Despite passage of the “Freedom to Farm” bill of 1996, which was designed to further reduce reliance of U.S. farmers on government payments, subsequent legislation and volatile markets have resulted in marked fl uctuation of agricultural subsidies over the past decade, ranging from a low of $11.9 billion in 2007 to a high of $24.4 billion just two years earlier.9

The tax system also sometimes serves to subsidize production. If the government gives a grant to a producer to assist in buying a machine, it appears as an expenditure. But suppose the government allows the pro- ducer to take a tax credit on the expenditures on machines—that is, if the

9�United States Department of Agriculture, Economic Research Service, Economic Indicators of the Farm Sector: National Financial Summary, 1992, Tables 14 and 22, January 1994; Survey of Current Business, Table B-10, June 1997, p. D-24; Direct Government Payments by Program, United States, 1933–95; and Direct Government Payments by Program, United States, 1996–2009.

34 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

producer buys a $100 machine with a 7 percent tax credit, it will get a $7 tax credit, which reduces by $7 the taxes it otherwise would have paid. Although it is not accounted as such, for all intents and purposes the tax credit is equivalent to government expenditure, and is thus referred to as tax expenditure. The value of federal tax expenditures has become very large in recent years, amounting to 30 percent of direct expenditures for fi scal year 2010.10

Finally, many government subsidies show up neither in the statis- tics on government expenditures nor in those on tax expenditures. For instance, when the government restricts the importation of some foreign good or imposes a tariff on its importation, this raises the prices of that good in the United States. American producers of competing goods are helped. In eff ect, there is a subsidy to American producers, paid not by the government but directly by consumers.

GOVERNMENT CREDIT�A special type of subsidy is government pro- vision of credit below market interest rates, in the form of low-interest loans and loan guarantees. Government subsidies tend to lead to the expansion of the subsidized industry, by lowering its cost of doing busi- ness. This is as true for subsidies to credit as it is for other forms of sub- sidies. Although such subsidies were once hidden, the Credit Reform Act of 1990 required the government to treat as expenditures any diff erence between the interest rates it paid and the interest rates it charged (taking into account the probability that the borrower might not repay).

In addition to loan subsidies, other government programs aff ect the allocation of credit, and thus of productive resources. In the United States, the subsidies are often used to buy particular goods and services. For instance, government-sponsored enterprises (GSEs) encourage lend- ing to enable people to buy homes and go to school. In 1993, the Clinton administration began lending funds directly to college students.

REGULATING BUSINESS Government regulates business activity in an attempt to protect workers, consumers, and the environment; to pre- vent anticompetitive practices; and to prevent discrimination.

The Occupational Safety and Health Administration attempts to ensure that workers’ places of employment meet certain minimal stan- dards. The National Labor Relations Board attempts to ensure that management and unions deal fairly with each other. The Federal Trade Commission attempts, among other things, to protect consumers from misleading advertising. The Environmental Protection Agency attempts

10�Budget of the United States Government, Fiscal Year 2012, Analytical Perspectives, Table 17-1, and Historical Tables, Table 1.1.

35Types of Government Activity

to protect certain vital parts of our environment by regulating, for instance, emissions from automobiles and toxic waste disposal.

In addition to these broad categories, some regulations apply to specifi c industries. The banking industry is regulated both by the Federal Reserve Board and the Comptroller of the Currency. Trucking is regulated by the Federal Highway Administration. The airlines are regulated by the Federal Aviation Administration. The telecommunications industry is regulated by the Federal Communications Commission. The securities industry is reg- ulated by the Securities and Exchange Commission. Beginning in the late 1970s, there was a concerted eff ort to reduce the extent of federal regula- tion. As noted earlier, the process of reducing or eliminating regulations is referred to as deregulation. There has been signifi cant deregulation in the airline, natural gas, electricity, trucking, and telecommunications industries.

Although the overall trend has entailed reduced regulation, there have been some instances of tightened regulation: the massive failure of the sav- ings and loan associations in the 1980s was attributed in part to lax banking regulation, and legislation enacted in 1989 provided for heightened scru- tiny. Likewise, repeal in 1999 of the Glass–Steagall Act of 1933, which had separated commercial and investment banking, contributed to the recent global economic crisis, so the Dodd–Frank Wall Street Reform and Con- sumer Protection Act was passed in 2010 to strengthen fi nancial sector reg- ulation. In other cases, the focus has been on changing regulation to refl ect changing circumstances. For example, several cases of deaths from food poisoning reinforced the importance of food safety regulations, but there was increasing recognition that the visual inspection system (did the meat smell and look rotten?) that had been employed since the beginning of the twentieth century needed to be replaced with a more scientifi c process.

Today, there is a move toward “smart regulation” that tries to balance the need for government oversight with its potential burden. It began with President Clinton’s issuance of an executive order in 1993 calling for a cost–benefi t review of government regulations. The initiative was called “Reinventing Government” or “National Performance Review,” and focused on making government agencies more client oriented and employing more market-like regulatory mechanisms. This approach was reaffi rmed and extended by a subsequent executive order issued by Presi- dent Obama in 2011. Both directives stress the principle that a regulation’s benefi ts must justify its costs and “impose the least burden on society, consistent with obtaining regulatory objectives.”11

11�See President Clinton’s Excutive Order 12866 of September 30, 1993, and President Obama’s Executive Order 13563 of January 18, 2011 (the source of the quotation in the text), for specifi c instructions on improving regulation and regulatory review, including acknowledgement of the diffi culty in quantifying many costs and benefi ts, as well as the importance of a participatory process when reviewing government regulations.

36 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

Federal outlays for the regulatory agencies are minuscule relative to the rest of the budget. However, these expenditures do not give an accu- rate view of the impact of the federal regulatory agencies. The extent to which these agencies infl uence virtually every aspect of business practices goes well beyond the simple measure of government expenditures. Many regulations have eff ects that are similar to those of taxes and subsidies. For example, regulations on utility prices may reduce prices for certain users below the free-market level, while raising the price to other users.

GOVERNMENT PURCHASES OF GOODS AND SERVICES

Every year the government buys billions of dollars’ worth of goods and ser- vices to support our national defense, maintain a network of highways, and provide education, police protection, fi re protection, and parks. These pur- chases of goods and services amount to nearly one-fi fth of the total produc- tion in the United States. In 2010, total government purchases were $3 trillion. Of these purchases, 17 percent was for investments in infrastructure such as roads and bridges that increase the economy’s future productivity.12

What we characterize as government purchases are amounts spent for goods and services made available to the public, such as national defense, public schools, and highways. Government payments to the aged through the Medicare program to fi nance their hospital expenses or to the poor through the food stamp program are categorized as transfer payments, not as direct government purchases. They are discussed in the next sec- tion, on government redistribution of income.

GOVERNMENT REDISTRIBUTION OF INCOME

The government takes an active role in redistributing income; that is, in taking money away from some individuals and giving it to others. There are two major categories of explicit redistribution programs: public assistance programs, which provide benefi ts to those poor enough to qualify; and social insurance, which provides benefi ts to the retired, disabled, unemployed, and sick.

As we saw earlier, outlays for explicit redistribution programs are called transfer payments. These expenditures are qualitatively diff erent from gov- ernment spending on, say, roads or bombers. Transfer payments are simply

12�The investment numbers do not include investments in people—human capital, for either education or health—but they do include investments in military aircraft and other hardware that enhance the country’s future defense capabilities, in transportation infrastructure such as highways and airports, and in natural resource ventures such as pollution control facilities.

37Types of Government Activity

changes in who has the right to consume goods. In contrast, a government out- lay for a road or a bomber reduces the amount of other goods such as private consumption goods that society can enjoy. Transfer payments aff ect the way in which society’s total income is divided among its members, but transfers do not aff ect the total amount of private goods that can be enjoyed (neglecting here losses of output due to distorted incentives associated with transfers).

PUBLIC ASSISTANCE PROGRAMS�Public assistance programs take two forms. Some provide cash, whereas others provide payment only for specifi c services or commodities. The latter are referred to as in-kind benefi ts. Of the cash programs, the largest are Temporary Assistance to Needy Families (TANF), the program that replaced the long-standing Aid to Families with Dependent Children (AFDC) in 1997, and Supple- mental Security Income (SSI), which provides cash to the poor who are aged, blind, or disabled. The largest in-kind public assistance program is Medicaid, which covers the medical costs of the poor, and accounts for about one half of total public assistance.

Table 2.1 lists the main public assistance programs with their date of enactment and their benefi ts. (In-kind benefi ts are valued at government cost in the table; we will see later that this may be diff erent from their value to the recipients.) The table shows that most benefi ts are in-kind, not cash,

TABLE 2.1 SELECTED GOVERNMENT PUBLIC AS SIS TANCE PROGR AMS (IN MILLIONS OF 2008 DOLL ARS) . PROGRAM DATE ENACTED 1990 OUTLAY 2000 OUTLAY 2008 OUTLAY

CASH BENEFITS

AFDC/TANF from 1998 1935 32,109 22,950 18,874

SSI 1972 27,897 39,423 44,062

Earned income tax credit* 1975 9,877 40,196 50,669

Total 69,884 102,569 113,605

IN-KIND BENEFITS

Medicaid 1965 130,828 255,169 352,170

Food stamps/Supplemental Nutrition Assistance Program (SNAP) from 1998 1964 24,217 18,128 36,442

National School Lunch Program 1946 5,379 6,837 8,265

Total 160,423 280,133 396,877

TOTAL 230,307 382,702 510,482

In-kind benefi ts share of total 69.7% 73.2% 77.7%

*Includes only the refunded portion.

SOURCES: Census Bureau, Statistical Abstract of the United States, 2011, Tables 538 and 568, and Statistical Abstract of the United States, 1992, Table 565; Internal Revenue Service, Statistics of Income Bulletin, Historical Table, Table 4.

38 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

and that this share has risen from roughly two- thirds to more than three-quarters since 1990.

SOCIAL INSURANCE PROGRAMS�Social insurance diff ers from public assistance in that an individual’s entitlements are partly depen- dent on his or her contributions, which can be viewed as insurance premiums. To the extent that what individuals receive is commensurate with their contributions, social insurance can be viewed as a government “production activity,” not a redistribution activity. Because what some receive is far in excess of what they contribute (on an actuarial basis), however, a large element

of redistribution is involved in government social insurance programs. The largest of these programs is the Old-Age, Survivors, and Disabil-

ity Insurance Program (OASDI, the proper name for Social Security). It provides income not only for the retired, but also for their survivors (in particular, widows and widowers) and the disabled. The other major social insurance programs are unemployment insurance, which provides temporary benefi ts after an individual loses a job; workers’ compensation, which provides compensation for workers injured at work; and Medicare. The Medicare program, providing medical services to the aged, has (like Medicaid) grown rapidly since it was fi rst introduced in 1965, and now is the second largest social insurance program. Figure 2.2 gives the relative size of the various social insurance and public assistance programs.

The Social Security and Medicare programs are sometimes referred to as middle-class entitlement programs, because the main benefi ciaries are the middle class, and benefi ts are provided not on the basis of need but because the benefi ciaries satisfy certain other eligibility standards (e.g., age). As soon as they satisfy these criteria, they become entitled to receive the benefi ts.

HIDDEN REDISTRIBUTION PROGRAMS�The government affects the distribution of income not only through direct transfers but also through the indirect effects of the tax system and other government programs. One could imagine the government taxing everyone at the same rate but then giving grants to those whose income fell below a cer- tain level. This would have the same effect as taxing the lower-income individuals at a lower rate. Thus, there is a certain arbitrariness in dis- tinguishing between transfer payments through spending programs and the implicit transfers through the tax system.13

13�Some of the tax expenditures can be viewed explicitly as forms of social insurance. The fact that unemploy- ment insurance and Social Security are only partially taxed and disability benefi ts not taxed at all, means that a dollar of direct expenditures for those purposes goes further than it would if subjected to taxation.

GOVERNMENT ACTIVITIES • Providing a legal system—required if a market

economy is to function

• Producing goods—defense, education, mail

• Affecting what the private sector produces, through subsidies, taxes, credit, and regulation

• Purchasing goods and services from the private sector, which are then supplied by the government to fi rms and households

• Redistributing income

39Types of Government Activity

RELATIVE IMPORTANCE OF SOCIAL INSURANCE AND PUBLIC ASSISTANCE PROGRAMS, 2009

Social Security (OASDI) and Medicare are by far the largest social insurance expenditures, as well as the largest overall transfer programs. Medicaid is the largest public assistance expenditure and the third larg- est overall transfer program.

FIGURE 2.2

The major example of a transfer program run through the tax system is the earned income tax credit (EITC), which actually provides supple- mental income to low-income earners. Under the Clinton administration, expenditures for the EITC were expanded in an eff ort to enhance the incentives for low-skilled workers to stay off welfare.

The government also redistributes income in the guise of subsidy programs and quotas. Our agricultural programs in eff ect redistrib- ute income to farmers. The oil import quotas of the 1950s redistributed income to owners of oil reserves. The alleged reason for the quotas was to ensure the energy independence of the United States; nonetheless, the

Social insurance total: $1,333.5 billion

Other (1.3%)

Workers’ compensation and

temporary disability (1.7%)

Unemployment insurance

(9.7%)

Medicare (37.5%)

OASDI (49.8%)

A

Public assistance total: $732.4 billion

Other (15.0%)Veterans’ benefits

(6.8%)

EITC (9.1%)

SNAP (7.5%)

Family assistance (2.7%) SSI

(6.5%)

B

Medicaid and other

medical care (52.4%)

SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Table 3.12.

40 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

redistributive eff ects were among the primary consequences, and they may indeed provide the true motivation for the legislation.

Spending for goods and services also has its redistributive conse- quences: subsidies to urban bus transport may help the poor, whereas subsidies to suburban rail lines may help the middle class.

OVERVIEW OF GOVERNMENT EXPENDITURES

We can now put together the discussion of purchases and transfers to get an overview of government expenditures. Figure 2.3 shows the distribution of government expenditures in 2009. In panel A, which combines outlays at all levels of government, we can see that purchases of goods and services

GOVERNMENT EXPENDITURES BY TYPE, 2009

(A) Today, almost half of all government expenditures at the

federal, state, and local levels are transfers, (B) excluding an

additional one out of every seven dollars of federal

expenditures for grants-in-aid to state and local government.

SOURCE: Economic Report of the President, 2011, Tables B-82,

B-84, and B-85.

FIGURE 2.3

Total government expenditures, 2009:

$4,998.7 billion

Subsidies (1.2%)

Interest (7.2%)

Transfers (43.3%)

Purchases (48.3%)

A

Federal government expenditures, 2009:

$3,457.4 billion

Subsidies (1.7%)

Purchases (28.6%)

Transfers (48.4%)

Interest (7.3%)

Grants-in-aid to state and local

government (14.0%)

B

41Types of Government Activity

primarily for defense and education constitute almost half of expenditures, and transfer payments comprise the bulk of the rest. However, in panel B, we can see that at the federal level, almost two-thirds of all expenditures are for transfer payments to individual, states, and local governments.

Figure 2.4 shows the purpose of the expenditures by broad categories, both for the federal government’s expenditures and for all government expenditures. Note that at the federal level, Social Security (OASDI) and defense purchases play major roles. For total government expenditures, education appears as a major category because it is the largest type of expenditure at the state and local level.14

14�There is some inherent imprecision in any classifi cation. For instance, veterans’ benefi ts, which are typically not included in defense expenditures, can be thought of as expenditures for previously deliv- ered defense services, and some of the expenditures on space are motivated by defense concerns.

SOURCES: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Tables 3.1, 3.12, 3.9.5, and 3.15.5; Economic Report of the President, 2011, Tables B-82, B-84, and B-85.

GOVERNMENT EXPENDITURES BY PURPOSE, 2009

At the federal level, the most important expenditures, other than transfers, are for defense; at the state and local level, the most important expenditures are for education.

FIGURE 2.4

Total government expenditures, 2009:

$4,998.7 billion

Subsidies (1.2%)

Interest (7.2%)

Other transfers (30.0%)

OASDI (13.3%)

Other purchases (19.9%)

Defense purchases (13.3%)

Education purchases

(15.1%)

A

Other purchases

(9.3%)

Federal government expenditures, 2009:

$3,457.4 billion

Subsidies (1.7%)

OASDI (19.2%)

Defense purchases (19.2%)

Other transfers (29.2%)

Interest (7.4%)

Grants (14.0%)

B

42 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

There have been marked changes in the relative importance of expenditures at diff erent levels of government over the past century. For instance, the federal share of all nondefense government spending grew from slightly less than one-fi fth in 1902 to almost two-thirds in 2009.

GAUGING THE SIZE OF THE PUBLIC SECTOR

Because the government’s impact on the private economy depends on its regulatory and tax policies as well as on its outlays, no single number can provide an accurate indicator of the government’s eff ect on the econ- omy. Nonetheless, one indicator that economists have found particularly convenient to use is the size of public expenditures relative to the size of the total economy. A standard measure of the size of the total economy is gross domestic product (GDP), which is a measure of the value of all the goods and services produced in the economy during a given year.

GROWTH IN EXPENDITURES AND THEIR CHANGING COMPOSITION

During the past eighty years, public expenditures as a share of GDP have quadrupled. In 1929, they were 10 percent of GDP, whereas in 2010, they represented 40 percent of GDP, as we see in Figure 2.5.15

DEFENSE EXPENDITURES Figure 2.5 also shows that between 1967 and 1979 defense expenditures as a percentage of GDP declined from 10.0 percent to 5.7 percent. They then increased during the Reagan years until 1987, peaking at 7.4 percent, but after the end of the Cold War they once again declined to 3.7 percent of GDP in 2000. They were back up to 5.6 percent in 2010, primarily due to the confl icts in Iraq and Afghanistan.

To avoid the misleading impressions that can be caused by failing to take account of infl ation, economists like to express expenditures in dollars of con- stant value. Thus, if last year the government spent $1 billion on some pro- gram, and this year it spends $1.1 billion but overall prices have increased by 10 percent, we say that the current expenditures (measured in last year’s prices) are $1 billion—in constant dollars, expenditures have not increased  at  all.

15�Recall from the earlier discussion the arbitrariness of this measure. For instance, if the government switches from providing aid to education through direct grants to providing it through tax expendi- tures, these statistics would show a fall in the share of public expenditures.

43Gauging the Size of the Public Sector

SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Tables 1.1.5, 3.1, and 3.9.5.

FEDERAL DEFENSE OUTLAYS AND TOTAL GOVERNMENT EXPENDITURES AS A PERCENTAGE OF GROSS DOMESTIC PRODUCT, 1929–2010

Total government expenditures as a percentage of GDP have increased markedly since 1929.

FIGURE 2.5

ESTIMATING THE FULL BUDGETARY AND ECONOMIC COSTS OF WAR

For a more detailed discussion on calculating the full costs of war, see Joseph E. Stiglitz and Linda J. Bilmes, The Three Trillion Dollar War: The True Cost of the Iraq Confl ict (New York: W. W. Norton & Company, 2008), and http://threetrilliondollarwar.org.

A nnual defense allocations rarely include the full budgetary costs of war, and certainly exclude the economic costs beyond direct fi nancial outlays. For example, in 2008, fi ve years after the United States invaded Iraq, the Congres- sional Budget Offi ce estimated the cost of the Iraq war at between $1 trillion and $2 trillion, depend- ing on troop levels and duration of the confl ict. This estimate consists of outlays beyond normal defense and veterans affairs expenditures. It includes war zone operations and related logistical support,

repair and replacement of some equipment, and a portion of medical expenses for wounded com- batants. However, if other costs are included, such as the long-term costs of providing medical care and disability compensation for war veterans, the economic costs of lives lost and disrupted, and the impact of rising oil prices, the 2008 estimate of other analysts rises to $4 trillion. Relatively con- servative full-cost accounting estimates in 2011 are between $4 trillion and $6 trillion.

0% 1925 1935 1945 1955 1965

Total government expenditures as a percentage of GDP

Federal defense outlays as a percentage of GDP

1975 1985 1995 2005

10%

20%

30%

40%

50%

60%

44 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

Thus, in constant 2005 dollars, defense expenditures shrank from $518 billion in 1968 to $311 billion in 1977. They then increased over the next decade to $482 billion in 1989 before declining again to $346 billion in 1998. Since then they have risen to a post-Vietnam War high of $612 billion in 2010.

TRANSFER PAYMENTS AND INTEREST Growth in expenditures for Social Security and Medicare account for much of the rise in public expendi- tures since 1950—social insurance’s share of expenditures increased almost fi vefold during this period (see Figure 2.6). Interest payments also increased signifi cantly, owing to the huge defi cits that began under President Reagan, as the government spent more than it received. It was not until President Clinton succeeded in passing a defi cit reduction act in 1993 that the defi cit was brought under control; interest payments actu- ally came down, as interest rates fell by more than the defi cit increased.

SOURCES: U.S. Department of Commerce, Bureau of Economic Analysis, National

Income and Product Accounts, Tables 3.1 and 3.12; and Offi ce of

Management and Budget, Historical Tables, Table 15.4.

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Public assistance

Social insurance

Net interest

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

22%

24%

26%

28%

30%

GOVERNMENT TRANSFER AND INTEREST PAYMENTS

AS A PERCENTAGE OF TOTAL EXPENDITURES,

1950–2009

Social insurance payments (largely OASDI and Medicare)

have grown from 5.9 percent of total public expenditures in 1952

to 26.7 percent of expenditures in 2009. Public assistance

share of total expenditures has risen from 6.1 percent in

1963 to 14.6 percent in 2009. Net interest payments grew from a low of 4.7 percent of

expenditures in 1975 to a high of 9.7 percent in 1995. Although

the national debt has grown tremendously over the past

decade, interest payments fell to 3.9 percent of expenditures in 2009 as a result of near-zero

interest rates on government borrowing.

FIGURE 2.6

45Gauging the Size ofthe Public Sector

Although public assistance is often blamed for the growth in public expenditures, its share of total government expenditures increased only from 6.1 percent of expenditures in 1963, before President Johnson’s War on Poverty, to 14.6 percent in 2009.

A major source of increased expenditure during the past two decades has been health care—Medicaid and Medicare—and there is real concern that these expenditures will con- tinue to soar in coming decades as well.

Figures 2.5 and 2.6 tell one other interest- ing story. Total nondefense expenditures in general, as well as social insurance expendi- tures in particular, increased rapidly from the mid-1960s through the mid-1970s, under both Democratic and Republican administrations. The most signifi cant break in the rate of increase of expenditures occurred under President Carter (1976–1980). Further- more, although Presidents Reagan and both Bushes preached that they would cut back the size of government, they failed to do so.

One reason for this is the tremendous inertia in the fi scal system. The full economic consequences of the Medicare program, enacted in 1965, were not felt until many years later. The scope for discretion—for chang- ing directions—within any administration is accordingly limited.

COMPARISON OF EXPENDITURES ACROSS COUNTRIES

The share of government, measured by total government expenditures as a percentage of GDP, is smaller in the United States than in most other high-income countries (see Figure 2.7), and its relative growth has also been slower than in most other advanced economies. The share of govern- ment in most middle- and low-income countries is considerably smaller than it is in high-income countries.

Because defense expenditures play a larger role in the United States, the relative size of nondefense expenditures is particularly low, viewed from this international perspective (see Figure 2.8).16

16�Comparisons across countries always need to be treated with caution. Particular problems are raised by the treatment of public enterprises. The fact that tax expenditures are relatively more important here than abroad may result in an understatement of the “eff ective” relative size of the public sector in the United States. On the other hand, regulations are perhaps less important in the United States than in most other developed countries.

GAUGING THE SIZE OF GOVERNMENT • The size of the U.S. government today is much

larger than it was a hundred years ago.

• During the past eighty years, public expendi- tures as a share of GDP have grown rapidly. In 1930, they were 9 percent of GDP. In 2010, they represented 36 percent of GDP.

• Growth in expenditures for Social Security, Medi- care, Medicaid, and interest account for much of the increase in public expenditures since 1950.

• The size of government relative to the economy is much smaller in the United States than in most European countries.

46 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

GOVERNMENT EXPENDITURES AS A PERCENTAGE OF

GDP IN 2009

Despite the growth of govern- ment in the United States, it has

one of the smallest public sec- tors of the eleven high-income

countries shown here—only those of Australia and South

Korea are smaller.

FIGURE 2.7

COMPOSITION OF GOVERNMENT

EXPENDITURES IN 2009

Spending priorities differ sig- nifi cantly between countries, as

seen, for example, by health and defense spending in the United States, social protection spend-

ing in Germany, and economic affairs spending in South Korea.

FIGURE 2.8

0%

High income Upper middle income

Low income

Sw ed

en US

A

Au str

ali a

Ge rm

an y

Ca na

da UK Ja

pa n

Ita ly

Sp ain

Gr ee

ce

S. Ko

re a

Ch ile

Ru ss

ia

Tu rke

y

S. Af

ric a

Eg yp

t

10%

20%

Total 54.1

Total 40.9

Total 35.1

Total 47.6

Total 41.8

Total 49.9

Total 41.9

Total 51.4

Total 43.1

Total 52.3

Total 24.2

Total 24.1

Total 42.7

Total 38.3

Total 38.4

Total 30.2

30%

40%

50%

60%

Employee compensation Goods and services Social benefits Interest Other

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0% USA Australia Germany Japan Spain S. Korea Egypt

Defense

Economic Affairs General Services

Education OtherHealth

Social Protection

SOURCES: IMF, Government Finance Statistics Yearbook 2010, Tables W3

and W5; and World Bank, World Development Indicators.

SOURCES: IMF, Government Finance Statistics Yearbook 2010, Tables W3

and W6; and World Bank, World Development Indicators.

47Government Revenues

GOVERNMENT REVENUES

Now that we have examined what the government spends its money on, we will briefl y survey the methods by which government raises rev- enue to pay for these expenditures. The government levies a variety of taxes. When the revenues that it receives from taxes are less than its planned expenditures, it must either cut back expenditures or borrow the diff erence.17

TAXES AND THE CONSTITUTION

The issue of taxation was very much in the thoughts of the founders of the republic. Indeed, the American Revolution began as a tax revolt with the  Boston Tea Party, which was a protest against the tax on tea, and with  the slogan “Taxation without representation is tyranny.” The fi rst article of the Constitution provides, “The Congress shall have power to levy and collect Taxes, Duties, Imposts, and Excises, to pay the Debts and pro- vide for the Common Defense and General Welfare of the United States.”

Three restrictions were imposed: the government could not levy taxes on exports; “all Duties, Imposts, and Excises” had to be “uniform throughout the United States” (referred to as the uniformity clause); and “no capitation or other direct tax shall be laid, unless in proportion to the Census or Enumeration herein before directed to be taken” (referred to as the apportionment clause). (A capitation tax is a tax levied on each person. These taxes are also called head taxes or poll taxes. They are no longer levied by any state.)

The constitutional provision restricting direct taxes proved to be a problem. Congress levied an income tax during the Civil War and reen- acted it in 1894 as a tax on very high incomes. However, it was declared unconstitutional by the Supreme Court in 1895. The Court held that the individual income tax was, in part, a direct tax, which the Constitution stipulates must be apportioned among the states according to their pop- ulation. Widespread criticism of this rule led to a constitutional amend- ment. The Sixteenth Amendment, ratifi ed in 1913, declares that “Congress shall have the power to levy and collect taxes on incomes, from whatever sources derived, without apportionment among the several states, and without regard to census or enumeration.”

17�In many countries when there is a gap between expenditures and revenues, the diff erence is fi nanced by printing money. This is how the Continental Congress fi nanced the Revolutionary War. (The expression “not worth a continental” arose from the fact that the currency was not highly valued.)

48 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

The apportionment provision, however, still may restrict Congress’s ability to impose some taxes. Several countries impose national property taxes or wealth taxes, but these are likely to be considered direct taxes and thus precluded in the United States by the apportionment provision.

FEDERAL TAXATION TODAY

The federal government currently relies on fi ve major forms of taxation: (1) the individual income tax, (2) payroll tax (to fi nance Social Security and Medicare benefi ts), (3) corporate income tax, (4) excise tax (taxes on specifi c commodities, such as gasoline, cigarettes, airline tickets, and alcohol), and (5) customs tax (taxes levied on selected imported goods). Although the individual income tax was the single largest source of tax revenue for the federal government after World War II and accounted for half of government revenues in 2001, its share has dropped in recent years and is now surpassed by social insurance payroll taxes. In 2010, the cor- poration income tax accounted for another 12.8 percent and customs and excise taxes 4.5 percent of federal government revenue.18

Just as there has been a marked shift in the composition of expendi- tures over the past fi fty years, so too has there been a marked change in the source of government revenues. With the two exceptions mentioned previously, the federal government did not impose any income tax on indi- viduals before 1913. The individual income tax accounted for 20 percent or less of government tax revenues in the years preceding World War II, when rates were quadrupled to pay for the war. Instead, the federal gov- ernment relied heavily on excise, customs, and corporate income taxes.19 Since then, the individual income tax had been the largest single source of federal revenues until it was recently overtaken by social insurance and retirement receipts, as shown in Figure 2.9. The corporation income tax has played a decreasing role, falling from 36 percent of federal revenues in 1927 to 23 percent in 1960 and 13 percent in 2010.

Between 1789 and 1909, the federal government received almost all its revenues from excise taxes and customs. Today, those taxes are relatively unimportant. On the other hand, the payroll tax, which was introduced by the Social Security Act of 1935, increased from 10 percent of federal revenues in 1953 to 41 percent in 2010.

18�U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Table 3.2. 19�For a historical summary of the major federal taxes, see Joseph Pechman, Federal Tax Policy, 5th ed. (Washington DC: Brookings Institution, 1987), Appendix A.

49Government Revenues

STATE AND LOCAL GOVERNMENT REVENUES

Unlike the federal tax system, state and local tax systems rely heavily on sales and property taxes. As shown in Figure 2.10, until the 1970s property taxes were their major source of revenue. Today, property and sales taxes each contribute about 20 percent of their total revenue. State and local individual income taxes amount to only 15 percent of the total, whereas corporate income taxes are between 2 and 3 percent.

Competition among states for industry discourages the use of some state and local taxes, especially corporate income taxes. The federal government provides substantial aid to state and local governments, much of it directed at specifi c programs, such as road construction, mass transit, bilingual education, vocational education, and libraries. Over the past decade, fed- eral grants to state and local governments have provided one-fi fth of their revenue; this share rose to one-fourth of their revenue in 2009 and 2010 as part of the federal stimulus to combat the nation’s deep economic recession.

SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Table 3.2.

DISTRIBUTION OF FEDERAL RECEIPTS BY SOURCE, AS PERCENTAGES OF TOTAL FEDERAL RECEIPTS, 1929–2010

The individual income tax and contributions to social insurance (primarily Social Security payroll taxes) are now by far the most important source of federal revenue. The shares of revenue provided by the corporate income tax and by customs and excise taxes have fallen sharply over the past forty years.

FIGURE 2.9

1925 1935 1945 1955 1965 1975 1985 1995

Corporate income tax

Social insurance contributions (payroll tax)

Personal income tax

Customs and excise tax

2005 0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

55%

60%

65%

50 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

COMPARISON OF TAXATION ACROSS COUNTRIES

As indicated in Figure 2.11, patterns of taxation diff er from country to country. Although in most European countries the income tax is less import- ant than in the United States—in 2006, it was almost half of tax revenue in the United States but averaged only 36 percent in Organisation for Economic Cooperation and Development (OECD) countries—taxes on goods and ser- vices are more important, averaging almost a third of tax revenue in OECD countries but only 17 percent in the United States. Outside the United States, the value-added tax (a tax imposed on the value of the output of a fi rm less the value of goods and services purchased from other fi rms) is a major source of revenue. Social Security taxes comprise about the same share of govern- ment revenues in the United States and other OECD countries. In contrast to

SOURCES: U.S. Department of Commerce, Bureau of Economic

Analysis, National Income and Product Accounts, Table 3.3; and OECD, Government at a

Glance 2009, Table 2.4.

DISTRIBUTION OF STATE AND LOCAL GOVERNMENT

RECEIPTS BY SOURCE, AS PERCENTAGES OF TOTAL

RECEIPTS, 1929–2010

Sales taxes and federal grants have increased in importance

while property taxes have decreased in importance

as a source of state and local revenues.

FIGURE 2.10

1925 1935 1945 1955 1965

Federal grants

Sales tax

Property tax

Personal income tax

Corporate income tax

1975 1985 1995 2005 0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

55%

60%

65%

70%

Note: Percentages refl ect tax revenues in 2006.

51Deficit Financing

high-income countries, lower-income countries tend to rely more on indirect than direct taxes, as the small size of their formal economies make it very diffi cult to collect payroll-based income and social security taxes.20

DEFICIT FINANCING

The major source of fi nancing of government expenditures is taxes. But many governments, especially in recent years, have found tax revenues insuffi cient to pay for their expenditures. A defi cit in any period is the excess of spend- ing over revenues. A defi cit is fi nanced by borrowing. The cumulative value of borrowing by a fi rm, household, or government is its debt.

20�IMF, Government Finance Statistics Yearbook 2010, Tables W3 and W4.

SOURCE: IMF, Government Finance Statistics Yearbook 2010, Tables W3 and W4.

GOVERNMENT REVENUE AS A PERCENTAGE OF GDP, 2009

The United States has a comparatively diverse set of revenue sources when all levels of government are aggregated, as do the United Kingdom, Italy, and Tunisia.

FIGURE 2.11

High income

Upper middle income

Lower middle income

US A

Au str

ali a

Ge rm

an y

UK Ita ly

Gr ee

ce Ch

ile

Tu rke

y

S. Af

ric a

Th ail

an d

Tu nis

ia

Total 30.8

Total 33.5

Total 44.5

Total 40.3

Total 46.6

Total 38.1

Total 21.7

Total 33.9

Total 34.4

Total 20.8

Total 32.4

0%

5%

10%

15%

25%

20%

35%

30%

45%

40%

50%

Income, profits, and capital gains taxes Property taxes

Social contributionsGoods and services taxes

Other revenue

52 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

A fi rm or household that runs a defi cit cannot continue to borrow indefi nitely, but will be forced into bankruptcy once its debt gets too large. Because of the federal government’s ability to tax, and the huge poten- tial revenue sources it can tap, its defi cits do not cause the same kinds of problems that large debts incurred by private fi rms or individuals would. Lenders will continue to willingly fi nance the federal government’s debt, provided the interest rate is high enough.

In the early 1980s, the size of the federal defi cit, both in dollar terms and, more importantly, as a fraction of GDP and of the budget, reached all-time highs (for peacetime); see Figure 2.12. The size of the defi cits in the 1980s caused great consternation both in and outside Washington. To fi nance the defi cit, the role of the federal government as a borrower in U.S. credit markets soared.

SOURCES: Offi ce of Management and Budget, Historical Tables, Tables 1.1

and 1.2; and U.S. Department of Commerce, Bureau of Economic

Analysis, National Income and Product Accounts, Table 1.1.5.

THE FEDERAL BUDGET DEFICIT AS

A PERCENTAGE OF EXPENDITURES AND

OF GDP, 1929–2010

The defi cit increased markedly during the early 1980s, fell, and then increased again during the

early 1990s. A federal budget surplus was achieved from 1998

to 2001, but the United States resumed running defi cits in

2002, culminating in 10 percent of GDP in 2009.

FIGURE 2.12 Budget deficit as a percentage of government expenditures

Budget deficit as a percentage of GDP

Great depression

World War II

Iraq/Afghanistan & great recessionVietnam

-40%

-30%

-20%

-10%

0%

10%

20%

30%

40%

50%

60%

70%

1925 1935 1945 1955 1965 1975 1985 1995 2005

53Deficit Financing

The dollar value of the debt goes up each year by the amount of that year’s federal defi cit (a federal surplus reduces the federal debt). However, the real value of the debt also depends very much on infl ation. To see what this means, assume you promise to pay someone $100 next year. If the prices of all goods and services rise by 10 percent, next year that person will be able to purchase with $100 the same goods that he or she could have purchased with $91 this year. The “real value” of what you have to pay has declined by $9.

Figure 2.13 traces the changes since 1940 in the real value of the total gross federal debt as well as the portion of this debt owed to U.S. citizens and foreigners, commonly refered to as the publicly held federal debt. In real terms, the increase in the debt after 1980 is dramatic. As a result of the high defi cits and the fall in the infl ation rate, the period 1981–1989 saw a doubling of the publicly held real debt, from $1.65 trillion in 1981 to $3.30 trillion in 1989 (both amounts measured in 2005 prices). To put it another way, during the Reagan administration, the total increase in real debt of the federal

SOURCE: Offi ce of Management and Budget, Historical Tables, Tables 1.3 and 7.1.

GROSS FEDERAL DEBT (2005 PRICES), 1940–2010

Federal debt held by the public doubled from 1981 to 1989, and then doubled again over the next two decades.

FIGURE 2.13

Total gross federal debt

Gross federal debt held by the public

1940 $0

$1,000

$2,000

$3,000

$4,000

$5,000

$6,000

$7,000

$8,000

$9,000

$10,000

$11,000

$12,000

1950 1960

Billions of 2005 USD

1970 1980 1990 2000 2010

54 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

government was equal to the total real debt accumulated over the fi rst two hundred years of existence of the United States, including the entire debt required to fi nance U.S. participation in World War II.

In 1998, the expanding economy, which had grown strongly since 1993, the tax increases enacted in 1993, and the limitations on expen- ditures that had been imposed for almost a decade, beginning in 1990, fi nally achieved their long-sought goal: there was a $70 billion surplus. Further surpluses were achieved through the end of the Clinton admin- istration, but subsequent substantial tax cuts and the cost of the confl icts in Iraq and Afghanistan resulted in renewed annual defi cits and a rapidly growing federal debt. These trends were exacerbated by the recent reces- sion; in 2009, the defi cit was 10 percent of GDP and the publicly held real debt had more than doubled again, totaling $6.80 trillion.

To put the United States in a comparative international perspective (see Figure 2.14), the United Kingdom ran a central government defi - cit of 10.0  percent of GDP and Japan 9.0 percent in 2009, close to the

CENTRAL GOVERNMENT DEFICITS AND GENERAL

GOVERNMENT DEBT, 2009 (% OF GDP)

The United States ran defi cits comparable to those of the

United Kingdom and Japan in 2009, and only Australia had a signifi cantly lower debt in 2009. Japan’s chronic large

annual defi cits have resulted in a debt at more than double its GDP, while Chile stands out

with remarkably low defi cit and debt fi gures.

SOURCES: IMF, Government Finance Statistics Yearbook 2010, Table W3; IMF,

World Economic Outlook Database, April 2011; and World Bank, World

Development Indicators.

FIGURE 2.14

8 4

.6 %

US A

Au str

ali a

Ge rm

an y

UK

Ja pa

n* Ita

ly

Gr ee

ce Ch

ile

Tu rke

y

S. Af

ric a

Th ail

an d

Tu nis

ia

10 .4

%

-10%

10%

30%

50%

70%

90%

110%

130%

150%

170%

190%

210%

230%

17 .6

% 1.9

%

73 .5

%

68 .3

% 10

.0 %

9. 0

%

4 .9

% 11

6. 1% 12

6. 8

%

21 6.

3%

14 .5

%

6. 2%

2. 5%

4 5.

5% 5.

0 %

30 .9

% 4

.6 %

4 5.

2% 1.0

%

4 2.

9% –1

.8 %

2. 1%

High income Upper middle income

Lower middle income

General government gross debt Central government deficit

* Japan deficit is for general government.

55Playing Tricks with the Data on Government Activities

10.4 percent federal defi cit, while both Australia and Germany’s defi cits were much lower, at 1.9 and 2.1 percent of GDP, respectively. However, at 17.6 percent of GDP, only Australia’s general government debt was sig- nifi cantly lower than that of the United States (84.6 percent), the United Kingdom (68.3  percent), and Germany (73.5 percent), whereas Japan’s stood at an alarming 216.3 percent of GDP in 2009. In contrast to these high-income countries, Chile, an upper-middle-income country, ran a defi cit of just 2.5 percent of GDP and had a debt of only 6.2 percent of GDP in 2009.

PLAYING TRICKS WITH THE DATA ON GOVERNMENT ACTIVITIES

The budgets of the federal, state, and local governments set out their expenditures and receipts. As we have seen, however, budgets provide only a partial view of the size of government and the eff ect of govern- ment on economic activity. As a result, one must treat with caution any comparisons of the size of the public sector either over time or across countries.

Earlier in this chapter, the sections on government subsidies and cred- its discussed how tax expenditures may result in misleading conclusions concerning not only the size of the public sector but also the composition of its expenditures. If the federal government wishes to hide the size of its subsidies to business, it provides tax credits to businesses. It hides the extent of its subsidies to states and localities by providing “tax expendi- tures” in the form of tax deductions on the federal individual income tax for most state and local taxes and tax exemption for interest on state and local bonds.

There is a second method by which the budget may be manipulated: by recording the revenues obtained when assets are sold, but not the cost— the reduction in the assets of the government. Such tricks were important in President Reagan’s attempt to reduce the defi cit. For instance, he accel- erated the sale of off shore oil and gas leases.

Speeding tax collections by increasing withholding or by increasing penalties for failing to pay taxes in a timely fashion is another one-time way of reducing a current defi cit.

The overall size of the public sector (but not the defi cit) can be decreased by setting up independent agencies and enterprises. It makes no real diff er- ence whether the post offi ce is a department of the U.S. government or, as is

56 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

the case today, a separate “corporation” receiving a subsidy from the federal treasury. If it is a department, however, all its income and all its expendi- tures will be included in the government budget; if it is a separate enter- prise, only the defi cit (the diff erence between its expenditures and income) is recorded. Similarly, placing special government funds off budget, such as those established to pay government pension obligations, also decreases the reported overall size of the public sector. Excluding quasi-fi scal activ- ities of the central bank, such as subsidies for special credit programs or preferred exchange rates, also understates the size of government. The U.S. federal government switched to a unifi ed (consolidated) budget in 1969 to address many such presentational issues.21

Although these problems provide considerable room for politicians to select statistics to support their views, the pattern of changes in the level and structure of expenditures and taxation in the United States since World War II has been signifi cant enough that there can be little question about three major observations that have been made in this chapter:

1. The public sector exerts a major infl uence on the production of goods and the distribution of income in the United States.

2. Social insurance has been the fastest-growing category of government expenditures in the past thirty years. Since 1960, the rapid growth in non- defense expenditure by government was largely accounted for by Social Security, government retirement programs, Medicare, and interest.

3. The individual income tax and social insurance payroll taxes have become the principal sources of federal revenue, whereas the role of the corporation income tax as a revenue source has dwindled.

21�For a detailed discussion of what the United States includes and excludes in its budget presentations, see Offi ce of Management and Budget, “Coverage of the Budget,” Analytical Prespectives, Budget of the United States Government, Fiscal Year 2012.

SUMMARY

1. The government performs many roles:

a. It provides the basic legal framework within which we live.

b. It regulates economic activities. It encourages some activities by subsidizing them and dis- courages others by taxing them.

c. It produces goods and provides credit, loan guarantees, and insurance.

REVIEW AND PRACTICE

57Review and Practice

d. It purchases goods and services, including many that are produced by private fi rms, such as weapons manufacturers.

e. It redistributes income, transferring income from some individuals to others.

f. It provides social insurance, for retirement, unemployment, disability, and medical care for the aged.

2. The size of the government relative to GDP is much larger now than it was forty years ago. Much of this increase is accounted for by increased payments for social insurance.

3. The relative size of the public sector in the United States is smaller than in most western European countries.

4. The three major areas of government expen- ditures are defense, education, and transfers. Together, these accounted for 72 percent of gov- ernmental expenditures in 2009.

5. The major source of revenue for the federal gov- ernment is the payroll tax, followed by the indi- vidual income tax, corporation tax, and customs and excise taxes.

6. The major sources of revenue for state and local government are the sales tax, the property tax, and the income tax.

7. The Constitution provides the basic framework for the government of the United States. It pro- vides some restrictions on the taxes that can be imposed, but no eff ective restrictions on what the government can spend its money on.

8. The deficit—the difference between the gov- ernment’s expenditures and revenues—grew enormously, beginning in 1981, with the total real debt accumulated from 1981 to 1988 alone equaling the total real debt accumulated over the first two hundred years of the country’s existence. Deficit reduction measures, begun in 1990 and extended in 1993, combined with a growing economy, enabled a surplus to be achieved in 1998. Subsequent tax cuts and the cost of the Iraq and Afghanistan con- flicts, together with the recent recession, have resulted in a doubling of the publicly held real

debt since then, and a 2009 deficit equal to 10 percent of GDP.

KEY CONCEPTS

Corporate income tax

Customs tax

Defi cit

Excise tax

Income tax

In-kind benefi ts

Middle-class entitlement programs

Nationalization

Payroll tax

Social insurance

Transfer payments

Value-added tax

QUESTIONS AND PROBLEMS

1. To see what is going on, economists often “adjust” the data to refl ect changes in the economy. For instance, in the text, we discussed the adjust- ments in dollar amounts made to correct for infl a- tion. Another adjustment that is frequently made is to take into account the increase in population. What adjustments might you make in looking at education expenditures? At Social Security expenditures?

2. In each of the following areas, give one or more examples (where possible) in which the gov- ernment is involved as a producer, a regulator, or a purchaser of fi nal goods and services dis- tributed directly to individuals or used within government:

a. Education

b. Utilities

c. Transportation

d. Credit markets

e. Insurance markets

f. Food

g. Housing

58 CHAPTER 2 MEASURING PUBLIC SECTOR SIZE

3. In each of the following areas, give an example of a tax expenditure and a conventional expen- diture. Explain how the same results could be obtained by converting the tax expenditure into a conventional expenditure.

a. Medicine

b. Housing

c. Education

4. Assume you were President and your planned expenditures exceeded your receipts. Describe some of the tricks you might use to reduce the apparent budget defi cit while maintaining current levels of services and transfers (subsidies).

Assume, on the other hand, that you had run on a platform of keeping the growth in total gov- ernmental expenditures down to 3 percent. Once in offi ce, you see, however, that you would like expenditures to rise by 5 percent. How might you do this while appearing to keep your election promises?

FUNDAMENTALS OF WELFARE ECONOMICS

Most economies today are mixed economies, in which there is both a private and a public sector. At the core of the economy are profit-maximizing firms interacting with households in competitive markets. Under certain idealized conditions, a competitive economy is efficient. If those condi- tions were satisfied, there would be a very limited role for government. To understand the role of the public sector then, we have to understand when markets work well, and when—and in what ways—they do not. That is the objective of this part of the book.

Chapter 3 explains what is entailed by efficiency, and why, under ide- alized conditions, competitive economies are efficient. Chapter 4 then explains the variety of reasons why and circumstances in which markets may fail to produce efficient outcomes, and why, even if the economy were efficient, there might be a role for government in redistributing income. Chapter 5 builds on this overview of market failure to produce efficient and equitable outcomes by exploring the unique nature of pub- lic goods, why these are undersupplied by private markets, and the ratio- nale for publicly provided private goods. Chapter 6 examines another type of market failure in greater depth, namely, externalities.

PART TWO

The hardest choices facing the public sector involve trade-offs, in particular, trade-offs between increased efficiency and a more equi- table distribution of income. Chapter 7 provides a conceptual frame- work for thinking about these trade-offs, some tools that are used by governments in attempting to quantify them, and an analysis of the circumstances in which one can have both more efficiency and more equity.

61

MARKET EFFICIENCY

In most modern industrial economies, primary reliance for the production and distribution of goods lies in the private rather than the public sector. One of the most enduring tenets of economics holds that this form of eco- nomic organization leads to an effi cient allocation of resources—but if private markets are effi cient, why should there be an economic role for gov- ernment? To answer this question, a precise understanding of the meaning of economic effi ciency is needed. That is the aim of this chapter. The next chapter will consider why private markets may fail to achieve effi cient out- comes and how government may respond to these market failures.

THE INVISIBLE HAND OF COMPETITIVE MARKETS

In 1776, Adam Smith, in The Wealth of Nations—the fi rst major work of modern economics—argued that competition would lead the individual

3 1. What do economists mean when they say the economy is effi cient?2. What conditions have to be satisfi ed if markets are to be effi cient?

3. What role does compe- tition play in ensuring effi ciency?

FOCUS QUESTIONS

62 CHAPTER 3 MARKET EFFICIENCY

in the pursuit of his or her private interests (profi ts) to pursue the public interest, as if by an invisible hand:

[H]e intends only his own gain, and he is in this, as in many other cases, led by an invisible hand to promote an end which was no part of his intention. Nor is it always the worse for the society that it was no part of it. By pursuing his own interest he frequently promotes that of the society more eff ectually than when he really intends to promote it.1

The signifi cance of Smith’s insight is clarifi ed by a look at the views about the role of government commonly held prior to Smith. There was widespread belief that achieving the best interests of the public (however that might be defi ned) required an active government. This view was particularly associated with the mercantilist school of the seventeenth and eighteenth centuries, which argued that government should promote industry and trade. Indeed, many European governments had actively promoted the establishment of colonies, and the mercantilists provided a rationale for this.

Some countries (or some citizens within them) had benefi ted greatly from the active role taken by their government, but other countries, whose governments had been much more passive, had also prospered. And some countries with strong, active governments had not prospered, as their resources were squandered on wars or on a variety of unsuccess- ful public ventures.

In the face of these seemingly contradictory experiences, Smith addressed himself to the question: Can society ensure that those entrusted with governing actually pursue the public interest? Experience had shown that although at times the policies governments pursued seemed consistent with the public good, at other times, the policies pur- sued could not by any reasonable stretch of the imagination be recon- ciled with the public good. Rather, those in the position of governing sometimes seemed to pursue their private interests at the expense of the public interest. Moreover, even well-intentioned leaders often led their countries astray. Smith argued that it was not necessary to rely on government or on any moral sentiments to do good. The public interest, he maintained, is served when each individual simply does what is in his or her own self-interest. Self-interest, Smith argued, is a much more persistent characteristic of human nature than a concern to do good, and therefore provides a more reliable basis for the organization of society. Moreover, individuals are more likely to ascertain with some accuracy what is in their own self-interest than they are to determine what is in the public interest.

1�Adam Smith, The Wealth of Nations (New York: Modern Library, 1937); originally published in 1776.

63Welfare Economics and Pareto Efficiency

The intuition behind Smith’s insight is simple: if there is some commodity or service that individuals value but that is not currently being produced, then they will be willing to pay something for it. Entrepreneurs, in their search for profi ts, are always looking for such opportunities. If the value of a certain commodity to a consumer exceeds the cost of production, there is a poten- tial for profi t, and an entrepreneur will produce the commodity. Similarly, if there is a cheaper way of producing a commodity than that which is pres- ently employed, an entrepreneur who discovers this cheaper method will be able to undercut competing fi rms and make a profi t. The search for profi ts on the part of enterprises is thus a search for more effi cient ways of production, and for new commodities that better serve the needs of consumers.

In this view, no government committee needs to decide whether a com- modity should or should not be produced. It will be produced if it meets the market test—that is, if what individuals are willing to pay exceeds the costs of production. Nor does any government oversight committee need to check whether a particular fi rm is producing effi ciently: competition will drive out ineffi cient producers.

There is widespread consensus among economists that competitive forces do lead to a high degree of effi ciency, and that competition does provide an important spur to innovation. However, over the past two hun- dred years, economists have come to recognize that in some important instances the market does not work as perfectly as the more ardent sup- porters of the free market suggest. Economies have gone through peri- ods of massive unemployment and idle resources; the Great Depression of the 1930s left many who wanted work unemployed; pollution has choked many of our larger cities; and urban decay has set in on others.

WELFARE ECONOMICS AND PARETO EFFICIENCY

Welfare economics is the branch of economics that focuses on what were termed normative issues in Chapter 1. The most fundamental normative issue for welfare economics is the economy’s organization—what should be produced, how it should be produced, for whom, and who should make these decisions. In Chapter 1, we noted that the United States and most other economies today are mixed, with some decisions made by the government but most left up to the myriad fi rms and households. But there are many “mixes.” How are we to evaluate the alternatives? Most economists embrace a criterion called Pareto effi ciency, named after the great Italian economist

64 CHAPTER 3 MARKET EFFICIENCY

ON THE PROWL FOR PARETO IMPROVEMENTS

A lthough fi nding Pareto improvements is diffi cult, economists are constantly on the lookout for such opportunities. Two recent proposals illustrate some of the problems that may be encountered.

One proposal concerned offshore oil wells. The federal government leases the land to oil compa- nies in return for a royalty, usually around 16 per- cent. The oil companies compete for these leases in competitive auctions; the lease goes to the fi rm offering the highest bid. As oil wells get old, the cost of extraction increases, often to the point at which, with the royalty taken into account, it pays to shut down the well. If the price of oil is $20 a barrel, for instance, and there is a 16 percent roy- alty, it pays to shut down the well when the cost of extraction exceeds $16.80 ($16.80 plus the $3.20 royalty equals the $20 received). This seems ineffi - cient, as the value of the oil ($20) exceeds the cost of production. Hence, there have been propos- als to eliminate royalties on old wells and to allow the oil companies to pay a fi xed up-front fee. The government is no worse off (because if the well is shut down it receives no revenue), and, provided the fee is set low enough, the oil company is bet- ter off (because if the well is shut down it receives nothing). The oil companies have resisted the pro- posal: they prefer that the government simply elim- inate royalties. Although the proposal is a Pareto improvement over the status quo, they would prefer

to garner for themselves more of the potential gains from the increased economic effi ciency.

A second proposal involved allowing private companies to construct improved turbines at hydro- electric sites, increasing the energy output. They would be allowed to sell the electricity at market prices. Hydroelectric energy is particularly attrac- tive, since it generates no pollution. There would be no adverse environmental impacts, as the develop- ments would occur only at sites already being used. This too appeared to be a Pareto improvement: economic effi ciency would be increased as cheaper hydroelectric power replaced power relying on fossil fuels; the benefi ts of the improved effi ciency would be shared among consumers, investors, and the government; and future generations would be better off as a result of the more favorable envi- ronmental impacts. This proposal was opposed by utility companies that currently got electricity from these dams at below-market prices. Although the proposal did not alter the current level of prefer- ential treatment, they were worried that once the principle that electricity from hydroelectric sites could be sold at market prices was established, their preferential treatment would be threatened. Even though the proposal as framed was a Pareto improvement, the utilities saw the long-run conse- quences of the proposal as a gain in effi ciency at the expense of their future welfare.

and sociologist Vilfredo Pareto (1848–1923). Resource allocations that have the property that no one can be made better off without someone being made worse off are said to be Pareto effi cient, or Pareto optimal. Pareto effi ciency is what economists normally mean when they talk about effi ciency.

Assume, for instance, that the government is contemplating building a bridge. Those who wish to use the bridge are willing to pay more than enough

65Welfare Economics and Pareto Efficiency

in tolls to cover the costs of construction and maintenance. The construction of this bridge is likely to be a Pareto improvement; that is, a change that makes some individuals better off without making anyone worse off . We say “likely” because there are always others who might be adversely aff ected by the construction of the bridge. For example, if the bridge changes the traffi c fl ow, some stores might fi nd that their business is decreased, and they are worse off ; or an entire neighborhood may be aff ected by the noise of bridge traffi c and the shadows cast by the bridge superstructure.

Frequently, on summer days or at rush hour, large backups develop at tollbooths on toll roads and bridges. If tolls were raised at those times and the proceeds used to fi nance additional tollbooths or more peak-time toll collectors, everyone might be better off . People would prefer to pay a slightly higher price in return for less waiting. Even this change might not be a Pareto improvement, though: among those waiting in line may be some unemployed individuals who are relatively not concerned about the waste of time but who are concerned about spending more money on tolls.

Economists are always on the lookout for Pareto improvements. The belief that any such improvements should be instituted is referred to as the Pareto principle.

“Packages” of changes together may constitute a Pareto improvement, when each change alone might not. Thus, although reducing the tariff on steel would not be a Pareto improvement (because steel producers would be worse off ��), it might be possible to reduce the tariff on steel, increase income taxes slightly, and use the proceeds to fi nance a subsidy to the steel industry. Such a combination of changes might make everyone in the country better off , and make those abroad—the foreign exporters of steel—also better off .

PARETO EFFICIENCY AND INDIVIDUALISM

The criterion of Pareto effi ciency has an important property that requires comment. It is individualistic, in two senses. First, it is concerned only with each individual’s welfare, not with the relative well-being of diff erent indi- viduals. It is not concerned explicitly with inequality. Thus, a change that made the rich much better off but left the poor unaff ected would still be a Pareto improvement. Some people, however, think that increasing the gap between the rich and the poor is undesirable. They believe that it gives rise, for instance, to undesirable social tensions. Less developed countries often go through periods of rapid growth during which all major segments of society become better off but the income of the rich grows more rapidly than that of the poor. To assess these changes, is it enough simply to say that everyone is better off ? There is no agreement on the answer to this question.

66 CHAPTER 3 MARKET EFFICIENCY

Second, it is each individual’s perception of his or her own welfare that counts. This is consistent with the general principle of consumer sovereignty, which holds that individuals are the best judges of their own needs and wants; that is, of what is in their own best interests.

THE FUNDAMENTAL THEOREMS OF WELFARE ECONOMICS

Two of the most important results of welfare economics describe the relationship between competitive markets and Pareto effi ciency. These results are called the fundamental theorems of welfare economics. The fi rst theorem tells us that if the economy is competitive (and satisfi es certain other conditions), it is Pareto effi cient.

The second theorem asks the reverse question. There are many Pareto effi cient distributions. By transferring wealth from one individual to another, we make the second individual better off and the fi rst worse off . After we make the redistribution of wealth, if we let the forces of competi- tion freely play themselves out, we will obtain a Pareto effi cient allocation of resources. This new allocation will be diff erent in many ways from the old. If we take wealth away from those who like chocolate ice cream and give it to those who like vanilla, in the new equilibrium, more vanilla ice cream will be produced and less chocolate, but no one can be made better off in the new equilibrium without making someone else worse off .

Let’s say there is a particular distribution that we would like to obtain. Assume, for instance, that we care particularly about the aged. The sec- ond fundamental theorem of welfare economics says that the only thing the government needs to do is redistribute initial wealth. Every Pareto effi cient resource allocation can be obtained through a competitive market process with an initial redistribution of wealth. Thus, if we do not like the income distribution generated by the competitive market, we need not abandon the use of the competitive market mechanism. All we need do is redistribute the initial wealth, and then leave the rest to the competitive market.

The second fundamental theorem of welfare economics has the remarkable implication that every Pareto effi cient allocation can be attained by means of a decentralized market mechanism. In a decentral- ized system, decisions about production and consumption (what goods get produced, how they get produced, and who gets what goods) are car- ried out by the myriad fi rms and individuals that make up the economy. In contrast, in a centralized allocation mechanism, all such decisions are concentrated in the hands of a single agency—the central planning

67Welfare Economics and Pareto Efficiency

agency—or a single individual, who is referred to as the central planner. Of course, no economy has even come close to being fully centralized, although under communism in the former Soviet Union and some of the other Eastern bloc countries, economic decision making was much more concentrated than in the United States and other Western economies. Today, only Cuba and North Korea place heavy reliance on central planning.

The second fundamental theorem of welfare economics says that to attain an effi cient allocation of resources, with the desired distribution of income, it is not necessary to have a central planner, with all the wis- dom an economic theorist or a utopian socialist might attribute to him or her; competitive enterprises, attempting to maximize their profi ts, can do as well as the best of all possible central planners. This theorem thus provides a major justifi cation for reliance on the market mechanism. Put another way, if the conditions assumed in the second welfare theorem were valid, the study of public fi nance could be limited to an analysis of the appropriate governmental redistributions of resources.

Why the competitive market, under ideal conditions, leads to a Pareto optimal allocation of resources is one of the primary subjects of study in standard courses in microeconomics. Because we will be concerned with understanding why, under some circumstances, competitive markets do not lead to effi ciency, we fi rst need to understand why competition, under ideal conditions, leads to effi ciency. Before turning to this, though, it is important to emphasize that these results are theorems; that is, log- ical propositions in which the conclusion (the Pareto effi ciency of the economy) follows from the assumptions. The assumptions refl ect an ideal competitive model, in which, for instance, there are many small fi rms and millions of households, each so small that it has no eff ect on prices; in which all fi rms and households have perfect information, say, concerning the goods that are available in the market and the prices which are being charged; and in which there is no air or water pollution.2 The accuracy of these assumptions in portrayal of our economy and the robustness of the results—the extent to which the conclusions change when the assump- tions change—are two of the main subjects of debate among economists. In the next chapter, we look at some of the important ways in which mar- kets fail to deliver effi cient outcomes; that is, we identify important cir- cumstances in which the ideal conditions underlying the fundamental theorems of welfare economics are not satisfi ed.

2�There are also a number of technical assumptions.

FUNDAMENTAL THEOREMS OF WELFARE ECONOMICS • Every competitive economy is Pareto effi cient.

• Every Pareto effi cient resource allocation can be attained through a competitive market mecha- nism, with the appropriate initial redistributions.

68 CHAPTER 3 MARKET EFFICIENCY

EFFICIENCY FROM THE PERSPECTIVE OF A SINGLE MARKET3

We can see why competition results in economic effi ciency using tradi- tional demand and supply curves. The demand curve of an individual gives the amount of the good the individual is willing to demand at each price. The market demand curve simply adds up the demand curves of all individuals: it gives the total quantity of the good that individuals in the economy are willing to purchase, at each price. As Figure 3.1 illustrates, the demand curve is normally downward sloping: as prices increase, individuals demand less of the good. In deciding how much to demand, individuals equate the marginal (additional) benefi t they receive from consuming an extra unit with the marginal (additional) cost of pur- chasing an extra unit. The marginal cost is just the price they have to pay.

The supply curve of a fi rm gives the amount of the good the fi rm is willing to supply at each price. The market supply curve simply adds up the supply curves of all fi rms: it gives the total quantity of the good that fi rms in the economy are willing to supply, at each price. As Figure 3.1 illustrates, the supply curve is normally upward sloping: as prices increase, fi rms are willing to supply more of the good. In deciding how much of a good to produce, competitive fi rms equate the marginal (addi- tional) benefi t they receive from producing an extra unit—which is just 3�This is often called the partial equilibrium approach, in contrast to the general equilibrium approach that looks at all markets simultaneously. We take the latter approach in the next section.

Supply curve

Demand curve

Quantity

Price

E

FIGURE 3.1

EFFICIENCY FROM THE PERSPECTIVE OF A

SINGLE MARKET

In deciding how much to demand, individuals equate the

marginal benefi t they receive from consuming an extra unit

with the marginal cost, the price they have to pay. In deciding

how much to supply, fi rms equate the marginal benefi t they

receive, which is just the price, with the marginal cost. At the

market equilibrium, where sup- ply equals demand, the marginal

benefi t (to consumers) is equal to the marginal cost (to fi rms)—

and each equals the price.

69Analyzing Economic Efficiency

the price they receive—with the marginal (additional) cost of producing an extra unit.

Effi ciency requires that the marginal benefi t associated with produc- ing one more unit of any good equal its marginal cost—for if the marginal benefi t exceeds the marginal cost, society would gain from producing more of the good; if the marginal benefi t was less than the marginal cost, society would gain from reducing production of the good.

Market equilibrium occurs at the point at which market demand equals supply, point E in Figure 3.1. At this point, the marginal benefi t and the mar- ginal cost each equal the price; thus, the marginal benefi t equals the mar- ginal cost, which is precisely the condition required for economic effi ciency.

ANALYZING ECONOMIC EFFICIENCY

To develop a deeper analysis that goes beyond the basic supply and demand framework just presented, economists consider three aspects of effi ciency, all of which are required for Pareto effi ciency. First, the econ- omy must achieve exchange effi ciency; that is, whatever goods are pro- duced have to go to the individuals who value them most. If I like chocolate ice cream and you like vanilla ice cream, I should get the chocolate cone and you the vanilla. Second, there must be production effi ciency. Given the society’s resources, the production of one good cannot be increased without decreasing the production of another. Third, the economy must achieve product mix effi ciency so that the goods produced correspond to those desired by individuals. If individuals value ice cream a lot relative to apples, and if the cost of producing ice cream is low relative to apples, then more ice cream should be produced. The following sections examine each of these types of effi ciency in turn.

THE UTILITY POSSIBILITIES CURVE

In preparation for learning what is entailed by each of the three aspects of Pareto effi ciency, the concept of the utility possibilities curve is use- ful. Economists sometimes refer to the benefi ts that an individual gets from consumption as the utility that the individual gets from the com- bination of goods he or she consumes.4 If the person gets more goods,

4�The concept of utility is only a useful way of thinking about the benefi ts that an individual gets from consumption. There is no way of measuring utility other than indirectly by looking at what individuals are willing to pay; no machine can ascertain the number of “utiles,” or whatever the unit of measure- ment of utility might be called, derived from eating a pizza or listening to a CD.

70 CHAPTER 3 MARKET EFFICIENCY

his or her utility has increased. The utility possibilities curve traces out the maximum level of utility that may be achieved by two consumers. Figure 3.2 shows a utility possibilities frontier for Crusoe and Friday, showing Friday’s maximum level of utility, given Crusoe’s level of utility (and vice versa). Recall the defi nition of Pareto effi ciency: an economy is Pareto effi cient if no one can be made better off without making someone else worse off ; that is, we cannot increase the utility of Friday without decreasing the utility of Crusoe. Thus, if an economy is Pareto effi cient, it must be operating along the utility possibilities frontier. If the economy were operating at a point below the utility possibilities frontier, such as at point A in Figure 3.2, it would be possible to increase the utility of Friday or Crusoe without decreasing the utility of the other, or to increase the utility of both.

The fi rst fundamental theorem of welfare economics says that a com- petitive economy operates along the utility possibilities frontier; the sec- ond fundamental theorem of welfare economics says that we can attain any point along the utility possibilities frontier using competitive mar- kets, provided we redistribute initial endowments appropriately.

EXCHANGE EFFICIENCY

Exchange efficiency concerns the distribution of goods. Given a particular set of available goods, exchange efficiency provides that those goods are distributed so no one can be made better off without

Friday’s utility

Crusoe’s utility

A

FIGURE 3.2

THE UTILITY POSSIBILITIES CURVE

The utility possibilities curve gives the maximum level of utility that one individual (Friday) can achieve,

given the level of utility of the other individual (Crusoe). Along the frontier, it is not possible for Crusoe to consume more unless Friday consumes less. Therefore,

the utility possibilities curve is downward sloping: the higher Crusoe’s utility, the lower the

maximum level of Friday’s utility.

71Analyzing Economic Efficiency

someone else being made worse off. Exchange efficiency thus requires that there is no scope for trades, or exchanges that would make both parties better off.

Assume that Crusoe is willing to give up one apple in exchange for one orange, or to get one apple in exchange for giving up one orange. Assume that Friday, on the other hand, is willing to give up three apples if he can get one more orange. At the margin, Friday values oranges more highly than does Crusoe. Clearly, there is room for a deal: if Crusoe gives Friday one of his oranges, and Friday gives Crusoe two of his apples, both are better off . Crusoe would have required only one apple to make him just as well off , but he gets two in exchange for his orange. Friday would have been willing to give up three apples, but he gave up only two, so he is clearly better off .

The amount of one commodity that an individual is willing to give up in exchange for a unit of another commodity is called the marginal rate of substitution. As long as Crusoe and Friday’s marginal rates of sub- stitution diff er, there will be room for a deal. Thus, exchange effi ciency requires that all individuals have the same marginal rate of substitution.

We now will see why competitive economies satisfy this condition for exchange effi ciency. To do so, we need to review how consumers make their decisions. We begin with the budget constraint—the amount of income a consumer can spend on various goods. Crusoe has $100, which he can divide between apples and oranges. If an apple costs $1 and an orange $2, Crusoe can buy 100 apples or 50 oranges, or combinations in between, as illustrated in Figure 3.3. If Crusoe buys one more orange, he has to give up two apples. Thus, the slope of the budget constraint is equal to the ratio of the prices.

Budget constraint

50 Oranges

100

Apples

CRUSOE’S BUDGET CONSTRAINT

Given income of $100, the price of oranges of $2, and the price of apples of $1, an individual can purchase any combination of apples and oranges along or to the left of the budget constraint. Any combination to the right of the budget constraint is unaffordable. The slope of the budget constraint is based on the relative price of oranges and apples.

FIGURE 3.3

72 CHAPTER 3 MARKET EFFICIENCY

50 60

Oranges 5917 18

100

89

80

18

17

Apples

B

A

E

F

D C

l1 l0

Crusoe chooses the point along the budget constraint that he most pre- fers. To see what this entails, we introduce a new concept: Indiff erence curves give the combinations of goods among which an individual is indif- ferent or which yield the same level of utility. Figure 3.4 shows indiff erence curves for apples and oranges. For example, the indiff erence curve I0 gives all the combinations of apples and oranges that the consumer fi nds just as attractive as 80 apples and 18 oranges (point A on the indiff erence curve). If points A and B are on the same indiff erence curve, the consumer is indif- ferent between the two combinations of apples and oranges represented by the two points. The indiff erence curve also shows how much of one good (apples) the consumer is willing to give up in return for one more unit of another good (oranges). The amount of one good the individual is willing to give up in return for one more unit of another good is just the marginal rate of substitution, which we defi ned earlier. Thus, the slope of the indiff erence curve equals the marginal rate of substitution. In Figure 3.4, in moving from point A to point B, Crusoe gives up one orange, but he is just as well off if he is compensated with nine extra apples. Note that the number of apples that he needs to compensate him for having one less orange is much higher

FIGURE 3.4

THE CONSUMER’S CHOICE PROBLEM

The budget constraint gives the combinations of apples and

oranges that Crusoe can buy, given his income and given the

price of apples and oranges. The indifference curve gives

the combinations of apples and oranges among which Crusoe is

indifferent. A and B are on the same indifference curve; Crusoe

is indifferent between them. Other individuals prefer combi-

nations of apples and oranges that are on a higher indifference curve. Thus, point F is preferred to either A or B. Crusoe chooses the point along the budget con- straint that he most prefers; that

is, the point at which the indif- ference curve I0 is tangent to the

budget constraint (point E ).

73Analyzing Economic Efficiency

when he moves from A to B than when he moves from C to D. When he has 60 oranges, he is much more willing to give up one of his oranges: he needs only one more apple to compensate him. Thus, the marginal rate of substitu- tion diminishes as the number of oranges that Crusoe consumes increases. This explains why the indiff erence curves have the shape depicted.

Clearly, individuals are better off if they have more apples and oranges; that is why combinations of goods along a higher indiff erence curve give a higher level of utility. Thus, any of the points on I1 in Figure 3.4 are more attractive than the points on I0. By defi nition, a consumer does not care at which point along an indiff erence curve he or she is placed; but the consumer wants to be along the highest indiff erence curve possible. Cru- soe would like to get to any point along the indiff erence curve I1, but he cannot: all these points lie above the budget constraint, and so are not feasible. The best that Crusoe can do is to choose point E, at which the indiff erence curve is tangent to the budget constraint.

At the point of tangency, the slope of the indiff erence curve is identical to the slope of the budget constraint, but the slope of the indiff erence curve is the marginal rate of substitution, and the slope of the budget constraint is the price ratio. Thus, individuals choose a combination of apples and oranges at which the marginal rate of substitution is equal to the price ratio.

Because all consumers face the same prices in a competitive economy, and each sets his or her marginal rate of substitution equal to the price ratio, they all have the same marginal rate of substitution. Earlier, we showed that the condition for exchange effi ciency was that all individuals have the same marginal rate of substitution. Thus, competitive markets have exchange effi ciency.

Another way to represent exchange effi ciency is illustrated in Figure 3.5. For simplicity, we continue the example of Crusoe and Friday: whatever Cru- soe does not get, Friday gets. Thus, we can represent all possible allocations in a box (called an Edgeworth–Bowley box after two early–twentieth- century English mathematical economists) in which the horizontal axis represents the total supply of oranges and the vertical axis represents the total supply of apples. In Figure 3.5, what Crusoe gets to consume is mea- sured from the bottom left corner (O), and what Friday gets is measured from the top right corner (O9). At the allocation denoted by the point E, Crusoe gets OA oranges and OB apples, while Friday gets the remainder (O9A9 oranges and O9B9 apples). We then draw Crusoe’s indiff erence curves, such as U�c. We have also drawn Friday’s indiff erence curves. His indiff er- ence curves look perfectly normal if you turn the book upside down.

Let us now fi x Crusoe’s utility. Pareto effi ciency requires us to maximize Friday’s utility, given the level of utility attained by Crusoe. Thus, we ask, given that Crusoe is on the indiff erence curve U�c, what is

74 CHAPTER 3 MARKET EFFICIENCY

the highest indiff erence curve that Friday can get to? Remember that Friday’s utility increases as we move down and to the left (Friday is get- ting more goods, Crusoe fewer goods). Friday attains his highest utility when his indiff erence curve is tangent to Crusoe’s, at E. At this point, the slopes of the indiff erence curves are the same; that is, their marginal rates of substitution of apples for oranges are the same.

PRODUCTION EFFICIENCY

If an economy is not productively effi cient, it can produce more of one good without reducing production of other goods. Along the production possibil- ities frontier in Figure 3.6, the economy cannot produce more of one good without giving up some of another good, given a fi xed set of resources.5

The analysis used to determine whether an economy is productively effi - cient is similar to the one we used above for exchange effi ciency. Consider Figure 3.7. In place of the budget constraint we have an isocost line, giving the diff erent combinations of inputs that cost the fi rm the same amount. The slope of the isocost line is the relative price of the two factors. The fi g- ure also shows two isoquants. These trace out the diff erent combinations

5�The production possibilities schedule has the shape it does because of the law of diminishing returns. As we try to produce more and more oranges, it becomes harder and harder to produce an additional orange. Thus, as we give up apples, we get more oranges, but for each additional apple we give up, we get fewer and fewer extra oranges.

EXCHANGE EFFICIENCY

The sides of this Edgeworth– Bowley box give the available

supplies of apples and oranges. OA and OB give Crusoe’s con- sumption of the two commod- ities. Friday gets what Crusoe

does not consume; that is, O’A’ and O’B’. Pareto effi ciency

requires the tangency of the two indifference curves (one such point is at E), where the

marginal rates of substitution of apples for oranges are equal.

FIGURE 3.5

Friday’s consumption of apples

Friday’s consumption of oranges

Friday’s indifference

curves

Crusoe’s indifference curve (U c)

Crusoe’s consumption of oranges

Crusoe’s consumption

of apples

Apples

Oranges

Friday’s indifference

curves

B¿

O¿A¿

EB

AO

75Analyzing Economic Efficiency

PRODUCTION EFFICIENCY AND THE PRODUCTION POSSIBILITIES FRONTIER

Points inside the frontier are attainable but ineffi cient. Points along the frontier are feasible and effi cient. Points outside the frontier are unattainable, given the resources of the economy.

FIGURE 3.6Apples

Oranges

Labor

Isocost line

Isoquants

Land

Q1

Q0

ISOQUANTS AND ISOCOST LINES

An isoquant gives combinations of inputs (land and labor) that yield the same output. The isoquant labeled Q1 represents a higher level of output than the isoquant labeled Q0. The slope of the isoquant is the marginal rate of technical substitution. The isocost line gives the combinations of inputs that cost the same amount. The slope of the isocost line is given by the relative prices of the two inputs. The fi rm maximizes its output, given a particular level of expenditures on inputs, at the point where the isoquant is tangent to the isocost line. At that point, the marginal rate of technical substitution equals the relative price.

FIGURE 3.7

76 CHAPTER 3 MARKET EFFICIENCY

of inputs—in this case, land and labor—that produce the same quantities of outputs. Thus, isoquants are to the analysis of production what indiff er- ence curves are to the analysis of consumption. Economists call the slope of an isoquant the marginal rate of technical substitution. In Figure 3.7, the marginal rate of technical substitution is the amount of land required to compensate for a decrease in the input of labor by one unit. When rela- tively little labor is being used, it is hard to economize further in its use, so if one less worker is used, there must be a large increase in land if output is to remain unchanged. That is why the isoquants have the shape they do. There is a diminishing marginal rate of technical substitution.

Just as exchange effi ciency requires that the marginal rate of substi- tution between any pair of commodities be the same for all individuals, production effi ciency requires that the marginal rate of technical substi- tution be the same for all fi rms. Assume the marginal rate of substitu- tion between land and labor is 2 in producing apples and 1 in producing oranges. This means that if we reduce labor by one in oranges, we need one more unit of land. If we reduce labor by one in apples, we need two more units of land. Conversely, if we increase labor by one in apples, we need two fewer units of land. Thus, if we take one worker from producing oranges and put him or her to work in apples, and we take one unit of land, and switch it from producing apples to producing oranges, production of oranges is unchanged but production of apples is increased. Whenever the marginal rates of substitution diff er, we can switch resources around in a similar way, to increase production.

A fi rm maximizes the amount of output that it produces at a given level of expenditures on inputs by fi nding the point at which the isoquant is tangent to the isocost line. At the point of tangency, the slopes of the two curves are the same—the marginal rate of technical substitution is equal to the ratio of the prices of the two inputs. In a competitive economy all fi rms face the same prices, so all fi rms using labor and land will set their marginal rate of technical substitution equal to the same price ratio. Hence, all will have the same marginal rate of technical substitution—the condition that is required for production effi ciency.

In Figure 3.8, we see the same principle diagrammatically, using another Edgeworth–Bowley box. We wish to know how to allocate a fi xed supply of inputs to ensure productive effi ciency. We represent the fi xed supply of the two inputs by a box, with the total available supply of land measured along the vertical axis and the total supply of labor along the horizontal axis. We measure inputs into orange production from the bot- tom left-hand corner. Thus, point E means that the amount OB of land is used in orange production, and OA of labor. That, in turn, means that the remaining inputs are used in the production of apples. Thus, we measure

77Analyzing Economic Efficiency

inputs into apples from the upper right-hand corner. At E, the amounts OˇBˇ of land and OˇǍ of labor go into apple production.

The isoquants also appear in the fi gure. Q0 gives a typical orange iso- quant. Remember that the quantities of inputs going into apple produc- tion are measured from Oˇ. That is why the isoquants for apples have the shape they do; they look perfectly normal if you turn the book upside down. Clearly, production effi ciency requires that for any level of production of oranges the output of apples is maximized. As we move down and to the left in the box, more resources are being allocated to apple production; hence, isoquants through those points represent higher levels of apple output. If we fi x the output of oranges at the level corresponding to isoquant Q0, it is clear that the output of apples is maximized by fi nding the apple isoquant that is tangent to isoquant Q0. Given that we produce Q0 of oranges, producing Q1 of apples (at, say, point C) means that some resources are unused. Producing along Q0, but not at E (at, say, point D), means that all resources are used, but not effi ciently; we can produce the same number of oranges and more apples at E. The economy cannot produce more than Q1 of apples and still produce Q0 of oranges; producing Q2 of apples would require producing less than Q0 of oranges. Only at point E are all resources used effi ciently and Q0 of oranges produced. At the point of tangency, the slopes of the isoquants are the same; that is, the marginal rate of substitution of land for labor is the same in the production of apples as it is in the production of oranges.

PRODUCTION EFFICIENCY

The sides of this Edgeworth– Bowley box give the available supply of resources—land and labor. Resources used in the production of oranges are given by OA and OB; resources not used in the production of oranges are used in the production of apples, O’A’ and O’B’. Production effi ciency requires the tangency of the isoquants. At tangency points, such as E, the marginal rate of substitution of land for labor is the same in the production of apples and oranges.

FIGURE 3.8

Land input into apples

Labor input into apples

Orange isoquant

Apple isoquants

Apple isoquants

Labor input into oranges

Land input into oranges

Land

Labor

B¿

O¿A¿

E

D

C

B

AO

Q1 Q0

Q2

E1

78 CHAPTER 3 MARKET EFFICIENCY

PRODUCT MIX EFFICIENCY

To choose the best mix of apples or oranges to produce, we need to con- sider both technical feasibility and individuals’ preferences. For each level of output of apples, we can determine from the technology the maximum feasible level of output of oranges. This generates the production possibili- ties schedule. Given the production possibilities schedule, we wish to get to the highest possible level of utility. For simplicity, we assume all individuals have identical tastes. In Figure 3.9, we depict both the production possi- bilities schedule and the indiff erence curves between apples and oranges. Utility is maximized at the point of tangency of the indiff erence curve to the production possibilities schedule. The slope of the production possibil- ities schedule is called the marginal rate of transformation; this tells us how many extra apples we can have if we reduce production of oranges by one. At the point of tangency, E, the slopes of the indiff erence curve and the production possibilities schedule are the same, that is, the marginal rate of substitution of apples for oranges is equal to the marginal rate of transformation.

Under competition, the marginal rate of trans- formation will be equal to the relative price of apples to oranges. If, by reducing production of apples by one, fi rms can increase the production of oranges by, say, one, and sell the oranges for more than the price of apples, profi t-maximizing fi rms will clearly expand production of oranges. We have shown why, under competition, consum- ers’ marginal rates of substitution will equal the price ratio. Because both the marginal rates of substitution and the marginal rate of transforma- tion will equal the price ratio, the marginal rate of transformation must equal consumers’ marginal rates of substitution. Hence, under ideal compet- itive markets, all three conditions required for Pareto effi ciency are satisfi ed.

BASIC CONDITIONS FOR PARETO EFFICIENCY 1. Exchange effi ciency: The marginal rate of substi-

tution between any two goods must be the same for all individuals.

2. Production effi ciency: The marginal rate of tech- nical substitution between any two inputs must be the same for all fi rms.

3. Product mix effi ciency: The marginal rate of transformation must equal marginal rate of substitution.

Competitive economies satisfy all three conditions.

79Review and Practice

SUMMARY

1. Resource allocations that have the property that no one can be made better off without someone else being made worse off are called Pareto effi - cient allocations.

2. The Pareto principle is based on individualistic values. Whenever a change can make some indi- viduals better off without making others worse off , it should be adopted. Most public policy choices, however, involve trade-off s, under which some individuals are better off and others are worse off .

3. The principle of consumer sovereignty holds that individuals are the best judges of their own needs and pleasures.

4. Pareto effi ciency requires exchange effi ciency, production effi ciency, and product mix effi ciency.

5. The fundamental theorems of welfare econom- ics provide conditions under which a competitive economy is Pareto effi cient, and under which every Pareto effi cient allocation can be obtained through markets, provided that the appropriate redistribu- tion of initial endowments (incomes) occurs.

6. Exchange effi ciency means that, given the set of goods available in the economy, no one can be made better off without someone else being made worse off ; it requires that all individuals have the same marginal rate of substitution between any pair of commodities. Competitive markets in which individuals face the same prices always have exchange effi ciency.

7. Production effi ciency requires that, given the set of resources, the economy not be able to produce more of one commodity without reducing the output of some other commodity; the economy must be operating along its production possibili- ties curve. Production effi ciency requires that all fi rms have the same marginal rate of technical substitution between any pair of inputs; compet- itive markets in which fi rms face the same prices always have production effi ciency.

8. Product mix effi ciency requires that the marginal rate of transformation—the slope of the produc- tion possibilities curve—equal individuals’ mar- ginal rate of substitution. Competitive markets have product mix effi ciency.

REVIEW AND PRACTICE

PRODUCT MIX EFFICIENCY REQUIRES THAT THE MARGINAL RATE OF TRANSFORMATION EQUAL CONSUMERS’ MARGINAL RATE OF SUBSTITUTION

To reach the highest level of consumers’ utility, the indiffer- ence curve and the production possibilities schedule must be tangent (point E). At any other point, such as E’, consumer utility is lower than at E.

FIGURE 3.9

Oranges

Apples Indifference

curves

Production possibilities

curve

E¿

E

80 CHAPTER 3 MARKET EFFICIENCY

KEY CONCEPTS

Centralized allocation mechanism

Consumer sovereignty

Exchange effi ciency

Fundamental theorems of welfare economics

Indifference curves

Invisible hand

Isocost line

Isoquants

Marginal (additional) benefi t

Marginal (additional) cost

Marginal rate of substitution

Marginal rate of technical substitution

Marginal rate of transformation

Pareto effi ciency

Pareto improvement

Pareto principle

Production effi ciency

Product mix effi ciency

Trades

Utility possibilities curve

QUESTIONS AND PROBLEMS

1. Explain why an economy in which airlines charge diff erent passengers diff erent prices for the same fl ight will not have exchange effi ciency.

2. Doctors often charge patients diff erent amounts depending on their judgment concerning the patients’ ability to pay. What implications does this have for exchange effi ciency?

3. Can you think of other common practices and poli- cies that might interfere with exchange effi ciency?

4. Explain why a tax that is levied only on the use of capital by corporations will interfere with the production effi ciency of the economy. (Com- pare the marginal rates of technical substitu- tion between corporations and unincorporated enterprises.)

5. Advocates of small businesses often argue that they should receive special tax treatment. Assume that small businesses had to pay only half the Social Security tax that is imposed on large corporations. What eff ect would that have on production effi ciency?

6. Consider an economy that produces two goods, cars and shirts. Explain why if a tax is imposed on the consumption of cars but not on shirts, the economy will not exhibit product mix effi ciency.

7. An individual is indiff erent among the combina- tions of public and private goods shown in the following table.

COMBINATION PUBLIC GOODS PRIVATE GOODS

A 1 16

B 2 11

C 3 7

D 4 4

E 5 3

F 6 2

Draw the individual’s indiff erence curve. Assum- ing that the economy can produce one unit of pub- lic goods and ten units of private goods, but that it can produce one more unit of public goods by reducing its production of private goods by two units, draw the production possibilities sched- ule. What is the maximum production of pri- vate goods? The maximum production of public goods? Can it produce fi ve units of public goods and one unit of private goods? Which of the feasi- ble combinations maximizes utility?

81

MARKET FAILURE

The last chapter explained why markets play such a central role in our economy: under ideal conditions, they ensure that the economy is Pareto effi cient. However, there is often dissatisfaction with markets. Some of the dissatisfaction is of the “grass is always greener on the other side” variety: people like to think that an alternative way of organizing the economy might make them better off . Some of the dissatisfaction is real, though: markets often seem to produce too much of some things, such as air and water pollution, and too little of others, such as support for the arts, research into the nature of matter, or the causes of cancer. Furthermore, markets can lead to situations in which some people have too little income to live on. Over the past fi fty years, economists have devoted enormous eff orts to understanding the circumstances under which markets yield effi cient outcomes, and the circumstances in which they fail to do so.

This chapter looks both at these market failures and at the reasons why governments intervene in markets even when they are effi cient.

4 1. What are the principal reasons why markets fail to produce effi cient outcomes? 2. What role does govern-

ment play in making it possible for markets to work at all?

3. Why might the govern- ment intervene in the market’s allocation of resources, even when it is Pareto effi cient? What are merit goods? What is government’s role in redistribution?

4. What is the “market fail- ures” approach to the role of government? What are alternative perspectives in thinking about the role of government?

FOCUS QUESTIONS

82 CHAPTER 4 MARKET FAILURE

PROPERTY RIGHTS AND CONTRACT ENFORCEMENT

Chapter 3 explained why markets result in Pareto effi cient outcomes. For markets to work, however, there needs to be a government to defi ne property rights and enforce contracts. In some societies, land is held in common: anyone can graze cattle and sheep on it. Because no one has the property right to the land, though, no one has an incentive to ensure

PROPERTY RIGHTS AND MARKET FAILURES: THE TRAGEDY OF THE COMMONS REVISITED

T he tragedy of the commons is a term that encapsulates how diffi cult it is to achieve the appropriate level of property rights. Oft- cited examples of this are the common pasture that is ruined by overgrazing or the common lake that is depleted by overfi shing. Although the action of each individual is a rational attempt to promote one’s own short-term self-interest, the group’s col- lective actions are not in the community’s long-term best interests. This dilemma raises the perplexing and hotly debated policy question: How does one manage a resource that formally does not belong to anyone?

For a long time, the common response was either conversion of common resources to private property or the external regulation of these common resources. If a resource is converted to private prop- erty, the owner should have both the profi t incentives and the protection of property rights to manage the entire resource responsibly. If the resource is regu- lated by the government, rules could be imposed on individual users for the common good.

However, in 1999, Nobel Prize-winning econ- omist Elinor Ostrom revisited these solutions

because they sometimes left the commons worse off. Private control over previously public resources could create the usual problems resulting from monopoly business practices, whereas exter- nal oversight could generate inappropriate and poorly implemented regulations. She noted a third response to the tragedy of the commons, one that she documented empirically by drawing on many cases from around the world: utilization of commu- nity social capital to devise creative and effective local solutions.

Another twist on this debate is referred to as the tragedy of the anticommons, a term coined by Michael Heller of Columbia Law School. In con- trast to the problems created by ambiguous own- ership of a community resource, this refers to the opposite: excessive private ownership of a com- munity resource that prevents achievement of a desirable outcome for society. The tragedy of the anticommons is now commonly applied to exces- sive property rights in areas such as biomedical research, for instance, in the patenting of genes (see Chapter 12).

83Market Failures and the Role of Government

that overgrazing does not occur. In the former communist countries, property rights were not well defi ned, so people had insuffi cient incentive to maintain or improve their apartments. In market economies, though, the benefi ts of such improvements are refl ected in the market price of the property.

Similarly, if individuals are to engage in transactions with each other, the contracts they sign must be enforced. Consider a typical loan, in which one person borrows money from another and signs a contract to repay it. Unless such contracts are enforced, no one would be willing to make a loan.

At an even more primitive level, unless there is protection of private property, people will have insuffi cient incentive to save and invest, as their savings might be taken away.

Government activities aimed at protecting citizens and property, enforcing contracts, and defi ning property rights can be thought of as providing the foundations on which all market economies rest.

MARKET FAILURES AND THE ROLE OF GOVERNMENT

The fi rst fundamental theorem of welfare economics asserts that the economy is Pareto effi cient only under certain circumstances or condi- tions. Markets are not Pareto effi cient under six important conditions, referred to as market failures, which provide a rationale for government activity.

1. FAILURE OF COMPETITION

For markets to result in Pareto effi ciency, there must be perfect com- petition; that is, there must be a suffi ciently large number of fi rms that each believes it has no eff ect on prices. However, in some industries— supercomputers, operating systems and chips for PCs, aluminum, ciga- rettes, and greeting cards, for instance—there are relatively few fi rms, or one or two fi rms have a large share of the market. When a single fi rm supplies the market, economists refer to it as a monopoly; when a few fi rms supply the market, economists refer to them as an oligopoly. Even when there are many fi rms, each may produce a slightly diff erent good, and may thus perceive itself facing a downward-sloping demand curve.

84 CHAPTER 4 MARKET FAILURE

Economists refer to such situations as monopolistic competition. In  all these situations, competition deviates from the ideal of perfect competition, in which each fi rm is so small that it believes there is noth- ing it can do to aff ect prices.

It is important to recognize that under these circumstances, fi rms may still seem to be competing actively against each other, and that the market economy may seem to “work” in the sense that goods are being produced that consumers seem to like. The fi rst fundamental theo- rem of welfare economics—the result that market economies are Pareto effi cient—requires more than just that there be some competition. As we saw in the last chapter, Pareto effi ciency entails stringent conditions, such as exchange, production, and product mix effi ciency; these conditions typically are satisfi ed only if each fi rm and household believes that it has no eff ect on prices.

There are a variety of reasons why competition may be limited. When average costs of production decline as a fi rm produces more,1 a larger fi rm will have a competitive advantage over a smaller fi rm. There may even be a natural monopoly, a situation in which it is cheaper for a single fi rm to produce the entire output than for each of several fi rms to produce part of it. Even when there is not a natural monopoly, it may be effi cient for only a few fi rms to operate. High transportation costs mean that goods sold by a fi rm at one location are not perfect substitutes for goods sold at another location. Imperfect information may also mean that if a fi rm raises its price it will not lose all its customers; it only faces a downward-sloping demand curve.

Firms may also engage in strategic behavior to discourage competi- tion. They may threaten to cut prices if potential rivals enter; such threats may both be credible and serve to discourage entry.

Finally, some imperfections of competition arise out of government actions. Governments grant patents—exclusive rights to an invention—to innovators. Although patents are important in providing incentives to innovate, they make competition in the product market less than perfect. Of course, even without patents, the fact that an innovator has some infor- mation (knowledge) that is not freely available to others may enable it to establish a dominant market position.

It is easy to see why imperfect competition leads to economic inef- fi ciency. We saw earlier that under competition, fi rms set output at the Pareto effi cient level. They set price equal to the marginal cost of pro- duction. Price can be thought of as measuring the marginal benefi t of consuming an extra unit of the good. Thus, with competition, marginal benefi ts equal marginal costs. Under imperfect competition, fi rms set

1�Declining average costs correspond to increasing returns: doubling inputs more than doubles output.

85Market Failures and the Role of Government

the extra revenue they obtain from selling one unit more—the marginal revenue—equal to the marginal cost. With a downward-sloping demand curve, the marginal revenue has two components. When a fi rm sells an extra unit, it receives the price of the unit, but to sell the extra unit, it must lower the price it charges on that and all previous units—the demand curve is downward sloping. The revenue gained from selling the extra unit is its price minus the revenue forgone because the expansion in sales lowers the price on all units. Thus, marginal revenue is less than price. Figure 4.1 shows the demand curve facing a fi rm and the marginal reve- nue, which lies below the demand curve. Competitive equilibrium occurs at Qc, whereas the imperfect competition equilibrium occurs at Qi, a much lower level of output. This reduction in output is the ineffi ciency associ- ated with imperfect competition.

Of course, if there is a natural monopoly, with declining average costs, and with marginal costs below average costs,2 competition is not viable; if a fi rm charged a price equal to the marginal cost (as would be the case under competition), it would operate at a loss, as the mar- ginal cost is lower than the average cost. Even then, however, a private monopoly would typically charge more than a government-run monop- oly; the private monopoly would seek to maximize profi ts, whereas the government-run monopoly, which did not receive any subsidy, would only seek to break even.

2�When average costs are declining, marginal costs always lie below average costs; it is the low value of the marginal cost—the cost of producing the last unit—that brings down the average costs.

MONOPOLY PRICING

Monopoly output is lower than competitive output, or the output at which profi ts are zero. There is a resulting welfare loss.

FIGURE 4.1

Marginal revenue

Marginal cost (= average cost)

Demand

Price

Output

B

Qi Qc

C

A

86 CHAPTER 4 MARKET FAILURE

2. PUBLIC GOODS

Some goods either will not be supplied by the market or, if supplied, will be supplied in insuffi cient quantity. An example on a large scale is national defense; on a small scale, navigational aids (such as buoys). These are called pure public goods. They have two critical properties. First, it costs nothing for an additional individual to enjoy their benefi ts: for- mally, there is zero marginal cost for the additional individual enjoying the good. It costs no more to defend a country of one million and one indi- viduals than to defend a country of one million. The costs of a lighthouse do not depend at all on the number of ships that sail past it. Second, it is, in general, diffi cult or impossible to exclude individuals from the enjoy- ment of a pure public good. If I put a lighthouse in a rocky channel to enable my ships to navigate safely, it is diffi cult or impossible for me to exclude other ships entering the channel from its navigational benefi ts. If our national defense policy is successful in diverting an attack from abroad, everyone benefi ts; there is no way to exclude any single individual from these benefi ts.

The market either will not supply, or will not supply enough of, a pure public good. Consider the case of the lighthouse. A large shipowner with many ships might decide that the benefi ts it receives from a lighthouse exceed the costs; but in calculating how many lighthouses to put in place, it will look only at the benefi ts it receives, not at the benefi ts received by others. Thus, there will be some lighthouses for which the total bene- fi ts (taking into account all of the ships that make use of the lighthouse) exceed the costs but for which the benefi ts of any single shipowner are less than the costs. Such lighthouses will not be put into place, and that is ineffi cient. The fact that private markets will not supply, or will supply too little of, public goods provides a rationale for many government activities. Public goods are discussed in detail in the next chapter.

3. EXTERNALITIES

In many cases, the actions of one individual or one fi rm aff ect other indi- viduals or fi rms, such as when one fi rm imposes a cost on other fi rms but does not compensate them, or, alternatively, when one fi rm confers a benefi t on other fi rms but does not reap a reward for providing it. Air and water pollution are examples. When I drive a car that is not equipped with a pollution control device, I lower the quality of the air, and thus impose a cost on others. Similarly, a chemical plant that discharges its chemicals into a nearby stream imposes costs on downstream users of the

87Market Failures and the Role of Government

water, who may have to spend a considerable amount of money to clean up the water to make it usable.

Instances in which one individual’s actions impose a cost on others are referred to as negative externalities. Not all externalities are nega- tive, though. There are some important instances of positive externali- ties, in which one individual’s actions confer a benefi t on others. If I plant a beautiful fl ower garden in front of my house, my neighbors may benefi t from being able to look at it. An apple orchard may confer a positive exter- nality on a neighboring beekeeper. An individual who rehabilitates his or her house in a neighborhood that is in decline may confer a positive exter- nality on the neighbors.

There are a large number of other examples of externalities. An addi- tional car on a crowded highway will add to road congestion, both reduc- ing the speed at which other drivers can travel safely and increasing the probability of an accident. An additional fi sherman fi shing in a given pond may reduce the amount of fi sh that others will be able to catch. If several oil wells are drilled in the same oil pool, taking more oil from one of the wells may reduce the amount of oil extracted by the other wells.

The crisis of 2008 made it clear that the fi nancial sector could (and did) impose large externalities on the rest of the economy. In fact, many described what they did using language similar to that associated with environomental externalities: the fi nancial sector was accused of pollut- ing the global economy with toxic mortgages.

Whenever such externalities exist, the resource allocation provided by the market will not be effi cient. Because individuals do not bear the full cost of the negative externalities they generate, they will engage in an excessive amount of such activities; conversely, because individuals do not enjoy the full benefi ts of activities generating positive externalities, they will engage in too little of these. Thus, for example, without govern- ment intervention of some kind, the level of pollution would be too high.

Externalities and environmental policy are discussed in detail in Chapter 6.

4. INCOMPLETE MARKETS

Pure public goods and services are not the only goods and services that private markets fail to provide adequately. Whenever private markets fail to provide a good or service even though the cost of providing it is less than what individuals are willing to pay, there is a market failure that we refer to as incomplete markets (because a complete market would provide all goods and services for which the cost of provision is less than

88 CHAPTER 4 MARKET FAILURE

what individuals are willing to pay). Some economists believe that pri- vate markets have done a particularly poor job of providing insurance and loans, and that this provides a rationale for government activities in these areas.

INSURANCE AND CAPITAL MARKETS The private market does not provide insurance for many important risks that individuals face, although insurance markets are much better today than they were ninety years ago. The government has undertaken a number of insurance pro- grams, motivated at least in part by this market failure. In 1933, following the bank failures of the Great Depression, the government set up the Fed- eral Deposit Insurance Corporation. Banks pay the corporation annual premiums, which provide insurance for depositors against a loss of sav- ings arising from the insolvency of banks. The government has also been active in providing fl ood insurance. Following urban riots in the summer of 1967, most private insurance companies refused to write fi re insurance in certain inner-city areas, and again the government stepped in.

Similarly, government has provided farmers with crop insurance, partly because of the failure of markets to do so; it provides unemploy- ment insurance; and until Medicare, the government health insurance program for the aged, was introduced in the 1960s, many of the elderly found it diffi cult to procure health insurance in the market. To protect investors against the eff ects of infl ation, the government has been issuing Treasury infl ation-protected securities (TIPS) since 1997.

In recent decades, the government has taken an active role not only in remedying defi ciencies in risk markets but also in ameliorating the eff ects of imperfect capital markets. In 1965 the government passed legislation providing for government guarantees on student loans, making it less dif- fi cult for individuals to obtain loans to fi nance their college education. Today, after it was discovered that private lenders were making massive profi ts out of student loans and charging excessive interest rates, the gov- ernment has become a major provider of student loans. This is only one of several government loan programs: the government provides loans to businesses engaged in international trade through the Export–Import Bank, it provides loans for small business through the Small Business Administration, and so forth. In each of these credit markets, there were allegations that access to credit was restricted prior to the introduction of the government program.

The question of why capital and insurance markets are imperfect has been the subject of extensive research during the past four decades. At least three diff erent answers have been put forward; each may have some validity. One focuses on innovation: we are used to new products

89Market Failures and the Role of Government

constantly coming onto the market; but there are also innovations in how the economy functions—innovations in creating new markets, including inventing new securities and new insurance policies. Indeed, those work- ing in the insurance and securities industries refer to these advances as new products. However, there is often an undersupply of innovations, which is why there is an important role for government in research.

The introduction of many of these new products is related to the sec- ond explanation: transactions costs. It is costly to run markets, to enforce contracts, and to introduce new insurance policies. An insurance fi rm may be reluctant to go to the trouble of designing a new insurance policy if it is unsure whether anyone will buy the policy, or if even if the product is successful, whether it will be able to reap the rewards as other compet- itors come into the market.

The third set of explanations centers around asymmetries of informa- tion and enforcement costs. The insurance company is often less informed

STUDENT LOANS: INCOMPLETE REFORM OF AN INCOMPLETE MARKET

Government guarantees opened the door to student loans, but the student loan market has not always worked well. Many of the problems arise out of the undue infl uence of the fi nancial sector and the private education sector in shaping the student loan program. For a long while, although the government guaranteed the loans, the private sector charged interest rates as if they were risky, imposing enormous costs on students. By replacing government-guaranteed loans by govern- ment loans, students and the government were able to save an estimated $80 billion over ten years. Many for-profi t schools entice poorly informed students; they fail to provide them with the skills they need to get a job. However, these schools have successfully repelled attempts to regulate them effectively, or even to deny them access to government loan pro- grams, without which they would not survive. Mak- ing matters worse, several pieces of legislation (most

recently, in the so-called reform of bankruptcy) made it extraordinarly diffi cult to discharge student debt, to get a fresh start, even if the school did not deliver the education and jobs promised.

There are increasing worries too that the pri- vate sector will attempt to “skim the cream,” offer- ing loans to the best risks, leaving the government with the highest risks.

With student debt now exceeding a trillion dollars—exceeding even credit card debt—and many young Americans saddled with seemingly crushing debts, there is a growing sense that there must be better ways of fi nancing these important investments.

Australia has explored one such approach, in which the amount students repay depends on their income. Income-contingent loan programs are now being experimented with by several countries around the world (see Chapter 14).

90 CHAPTER 4 MARKET FAILURE

about the nature of some risks than the person purchasing insurance. When the two parties to a transaction have diff erent information of this kind, we say that there is an information asymmetry. Thus, a fi rm might well wish to buy insurance against the risk that the demand for its prod- uct will decline. The insurer, on the other hand, may well reason: I want to estimate the risk, and charge a premium based on that estimate. If I overestimate the risk, the premium will be too high, and the fi rm will refuse to buy my policy; whereas if I underestimate the risk, the premium will be too low, in which case, the fi rm will buy my policy, but on average, I will lose money. I am in a heads-you-win-tails-I-lose situation. When information asymmetries like this are large, markets will not exist.

Similarly, in capital markets, lenders worry about getting repaid. They may not be able to tell which borrowers are likely to repay. This is partic- ularly a problem with loans, such as student loans, for which there is no collateral. (In the case of a loan on a house, if the borrower defaults, at least the lender can sell the house and recoup most or all of what it has put out.) The bank fi nds itself in a dilemma: if it increases the interest rate to refl ect the fact that many loans are not repaid, it may fi nd that the default rate actually increases; those who know that they are going to repay refuse to borrow, while those who are not planning to repay care very lit- tle about the amount the lender is nominally charging, as in all likelihood they will not pay that amount anyway. The phenomenon is called adverse selection. It may turn out that there is no interest rate that the bank can charge for, say, student loans (without a government subsidy) at which it can reap an expected return commensurate with what it can obtain on other investments.

This basic principle—that when there are asymmetries of information and enforcement problems markets may not exist—has been shown to provide part of the explanation of many missing markets.3 We shall exam- ine these problems in greater depth in the context of health insurance, in Chapter 13.

The reasons why markets do not exist may have implications for how governments might go about remedying the market failure. Government, too, faces transactions costs, enforcement problems, and asymmetries of information, although in many instances they are diff erent from those faced by the private sector. Thus, in designing loan programs or interven- tions in capital markets, governments need to bear in mind that they too are often less informed than the borrower.

3�The literature in this area is extensive. The basic articles are George Akerlof, “The Market for Lemons: Qualitative Uncertainty and the Market Mechanism,” Quarterly Journal of Economics 84 (1970): 488–500; and Michael Rothschild and Joseph Stiglitz, “Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Information,” Quarterly Journal of Economics 90 (1976): 629–650.

91Market Failures and the Role of Government

COMPLEMENTARY MARKETS Finally, we turn to the prob- lems associated with the absence of certain complementary markets. Suppose all individuals enjoy only coff ee with sugar. Assume, moreover, that without coff ee there is no market for sugar. Given that sugar was not produced, an entrepreneur considering whether to produce coff ee would not do so, because he would realize that he would have no sales. Likewise, given that coff ee was not produced, an entrepreneur consider- ing whether to produce sugar also would not do so, since she too would realize that she would have no sales. If, however, the two entrepreneurs could get together, there would be a good market for coff ee and sugar. Each acting alone would not be able to pursue the public interest, but acting together they could.

This particular example is deliberately quite simple, and in this case coordination (between the potential sugar producer and the potential coff ee producer) might easily be provided by the individuals themselves without government intervention. In many cases, however, particularly in less developed countries, large-scale coordination is required; this may require government planning. Similar arguments have been put forward as justifi cation for public urban renewal programs. Redeveloping a large section of a city requires extensive coordination among factories, retail- ers, landlords, and other businesses. One of the objectives of government development agencies is to provide that coordination (if markets were complete, the prices provided by the market would perform this “coordi- nation” function).

5. INFORMATION FAILURES

A number of government activities are motivated by imperfect informa- tion on the part of consumers, and by the belief that the market, by itself, will supply too little information. For instance, the Truth in Lending Act requires lenders to inform borrowers of the true rate of interest on their loans. The Federal Trade Commission and the Food and Drug Adminis- tration have both adopted a number of regulations concerning labeling, disclosure of contents, and the like. At one time, the Federal Trade Com- mission proposed that used-car dealers be required to disclose whether they had tested various parts of the car, and, if so, what the outcomes of the tests were. These regulations generated a considerable amount of con- troversy, and under pressure from Congress, the FTC was forced to back down.

Opponents of regulations on information disclosure contend that they are unnecessary (the competitive market provides incentives for fi rms to

92 CHAPTER 4 MARKET FAILURE

disclose relevant information), irrelevant (consumers pay little attention to the information the law requires fi rms to disclose), and costly, both to government that must administer them and to the fi rms that must comply with the regulations. Proponents of these regulations claim that, though they are sometimes diffi cult to administer eff ectively, they are still critical to the aff ected markets.

The government’s role in remedying information failures goes beyond these simple consumer and investor protections, however. Information is, in many respects, a public good. Giving information to one more individual does not reduce the amount others have. Effi ciency requires that information be freely disseminated or, more accurately, that the only charge be for the actual cost of transmitting the infor- mation. The private market will often provide an inadequate supply of information, just as it supplies an inadequate amount of other public goods. The most notable example of government activity in this area is the National Weather Service. Another example is the information pro- vided to ships by the U.S. Coast Guard.

Various other market failures are associated with imperfect information. One of the assumptions that went into the proof of the fundamental theorems of welfare economics was that there was per- fect information, or, more precisely, that nothing fi rms or households did had any eff ect on beliefs or information. In fact, much economic activity is directed at obtaining information, from employers trying to fi nd out who are good employees, to lenders trying to fi nd out who are good borrowers, investors trying to fi nd out what are good investments, and insurers trying to fi nd out who are good risks. Later, we shall see that information problems lie behind several government programs. For instance, many of the problems in the health sector in general and health insurance markets in particular can be traced to problems of information.

Resources devoted to producing new knowledge—research and development (R&D) expenditures—can be thought of as a particularly important category of expenditures on information. Again, the funda- mental theorems of welfare economics, which form the basis of our belief in the effi ciency of market economies, simply assume that there is a given state of information about technology, begging the question of how the economy allocates resources to research and development. Chapter 12 will explain why the market, on its own, may engage in an insuffi cient amount of at least certain types of R&D.4

4�For an extended discussion of the market failures associated with incomplete markets and imperfect information, see B. Greenwald and J. E. Stiglitz, “Externalities in Economies with Imperfect Informa- tion and Incomplete Markets,” Quarterly Journal of Economics 105 (May 1986): 229–264.

93Market Failures and the Role of Government

6. UNEMPLOYMENT, INFLATION, AND DISEQUILIBRIUM

Perhaps the most widely recognized symptoms of market failure are the periodic episodes of high unemployment, both of workers and machines, that have plagued capitalist economies during the past two centuries. Even though these recessions and depressions were greatly moderated in the period between World War II and 2008, perhaps partly because of government policies, the unemployment rate still climbed over 10 percent in 1982 and reached that level again in 2009. Although this is low compared to the rate during the Great Depres- sion, when unemployment reached 24 percent in the United States, high unemployment rates during the current Great Recession per- sisted despite government interventions to kickstart the economy. The national unemployment rate also masks significantly higher rates in especially hard-hit parts of the country and among highly vulnerable populations. At one time, more than one of six Americans who wanted a full-time job could not get one. The global economic crisis has hit several countries in Europe particularly hard; by 2013, average unem- ployment had reached record levels, and several countries were in a depression, with average unemployment in excess of 25 percent and youth unemployment in excess of 50 percent.

Most economists take the high levels of unemployment as prima facie evidence that something is not working well in the market. To some econ- omists, high unemployment is the most dramatic and most convincing evidence of market failure.

The issues raised by unemployment and infl ation are suffi ciently important, and suffi ciently complicated, that they warrant a separate course in macroeconomics. Some aspects of these issues are touched on in Chapter 28, which is concerned with the consequences of government defi cits and attempts to survey some of the important ways that these macroeconomic con- siderations aff ect the design of tax policy.

INTERRELATIONSHIPS OF MARKET FAILURES

The market failures we have discussed are not mutually exclusive. Information prob- lems often provide part of the explanation of

SIX BASIC MARKET FAILURES 1. Imperfect competition

2. Public goods

3. Externalities

4. Incomplete markets

5. Imperfect information

6. Unemployment and other macroeconomic disturbances

94 CHAPTER 4 MARKET FAILURE

missing  markets. In turn, externalities are often thought to arise from missing markets: if fi shermen could be charged for using fi shing grounds— if there were a market for fi shing rights—there would not be overfi shing. Public goods are sometimes viewed as an extreme case of externalities, where others benefi t from my production of the good as much as I do. Much of the recent research on unemployment has attempted to relate it to one of the other market failures.

MARKET FAILURES: EXPLANATIONS OR EXCUSES?

T he agricultural price support program pro-vides an illustration of an instance in which the appeal to market failures is more of an excuse for a program than a rationale. There are important market failures in agriculture. Prices and output are highly variable. Farmers typically cannot buy insurance to protect them against either price or output fl uctuations. Even though they could reduce their exposure to price risk somewhat by trading in futures and forward markets, these mar- kets are highly speculative, and farmers worry that they are at a marked disadvantage in trading in them. For example, there are fi ve very large trad- ers in wheat that have access to more information; as a result, farmers view trading on futures markets with these informed traders as playing on an unlevel playing fi eld.

What farmers really care about, of course, is not price variability, but income variability. Programs to stabilize prices do not fully stabilize income, as income depends both on the price received and the quantity produced. Indeed, in some cases, sta- bilizing prices may actually increase the variability of income. Normally, prices rise when, on average, quantities fall. If prices rise proportionately, then income may vary very little, with price increases

just offsetting quantity decreases. In such a sit- uation, stabilizing prices will increase income variability.

Price support programs are also justifi ed as helping poor farmers—refl ecting the failure of markets to provide an appropriate distribution of income. Critics ask, though, why are poor farmers particularly deserving of aid, rather than poor peo- ple in general? Moreover, the price support pro- grams give aid on the basis of how much a farmer produces. Thus, large farmers gain far more than small farmers do.

If the objective of the farm programs were to address these market failures, then the farm pro- gram would be designed in a markedly different manner. In fact, a major objective of the farm pro- gram is to transfer resources—to subsidize farm- ers (and not just poor farmers)—not to correct a market failure. The program is designed to keep a large part of its cost hidden: only a part of the cost is refl ected in the federal budget; the rest is paid for by consumers in the form of higher prices. The market failure approach has provided some of the rhetoric for the program, but not the rationale. For that, we have to look into politics and the role of special interest groups.

95Redistribution and Merit Goods

REDISTRIBUTION AND MERIT GOODS

The sources of market failure discussed thus far result in economic inef- fi ciency in the absence of government intervention. Even if the economy were Pareto effi cient, though, there are two further arguments for gov- ernment intervention. The fi rst is income distribution: the fact that the economy is Pareto effi cient says nothing about the distribution of income; competitive markets may give rise to a very unequal distribution, which may leave some individuals with insuffi cient resources on which to live. One of the most important activities of the government is to redistrib- ute income. This is the express purpose of welfare activities, such as food stamps and Medicaid. How we think systematically about issues of distri- bution is the subject of Chapter 7.

The second argument for government intervention in a Pareto effi cient economy arises from concern that individuals may not act in their own best interests. It is often argued that an individual’s perception of his or her own welfare may be an unreliable criterion for making welfare judg- ments. Even fully informed consumers may make “bad” decisions. Indi- viduals continue to smoke, for instance, even though it is bad for them, and even though they know it is bad for them. Individuals fail to wear seat belts, even though wearing seat belts increases the chances of survival in an accident, and even though individuals know the benefi ts of seat belts. The same holds true for motorcycle helmets. There are those who believe that the government should intervene in such cases, in which individuals seemingly do not do what is in their own best interest; the kind of inter- vention that is required must be stronger than simply providing informa- tion. Goods that the government compels individuals to consume, such as seat belts and elementary education, are called merit goods.

The view that the government should intervene because it knows what is in the best interest of individuals better than they do themselves is referred to as paternalism. The paternalistic argument for government activities is quite distinct from the externalities argument discussed ear- lier. One might argue that smoking causes cancer, and that because indi- viduals who get cancer may be treated in public hospitals or fi nanced by public funds, smokers impose a cost on nonsmokers. This, however, can be dealt with by making smokers pay their full costs—for instance, by impos- ing a tax on cigarettes. Alternatively, smoking in a crowded room does indeed impose a cost on nonsmokers in that room. But this, too, can be dealt with directly. Those who take a paternalistic view might argue that individuals should not be allowed to smoke, even in the privacy of their

96 CHAPTER 4 MARKET FAILURE

own homes, and even if a tax that makes the smokers take account of the external costs imposed on others is levied. Although few have taken such an extreme paternalistic position with respect to smoking, this paternal- istic role undoubtedly has been important in a number of areas, such as government policies toward drugs (illegalization of marijuana) and liquor (prohibition in the 1930s).

In contrast to the paternalistic view, many economists and social phi- losophers believe that the government should respect consumers’ prefer- ences. Though there may occasionally be cases that merit a paternalistic role for the government, these economists argue that it is virtually impos- sible to distinguish such cases from those that do not. And they worry that once the government assumes a paternalistic role, special interest groups will attempt to use government to further their own views about how individuals should act or what they should consume. The view that government should not interfere with the choices of individuals is some- times referred to as libertarianism.

There are two important caveats to economists’ general presumption against government paternalism. The fi rst concerns children. Someone— either the parents or the state—must make paternalistic decisions on behalf of children, and there is an ongoing debate concerning the proper division of responsibility between the two. Some treat children as if they were the property of their parents, arguing that parents alone should have responsibility for taking care of their children. Most argue, however, that the state has certain basic responsibilities, such as, for instance, ensur- ing that every child gets an education and that parents do not deprive their children of needed medical care or endanger them physically or emotionally.

The second caveat concerns situations in which the government can- not, at least without diffi culty, commit itself to refrain from helping indi- viduals who make poor decisions. For instance, individuals who do not save for their retirement become a burden on the government; this pro- vides part of the rationale for Social Security. In other instances, individ- uals who fail to take appropriate precautions become a burden to society, and a sense of compassion makes it diffi cult, in the face of a crisis, to sim- ply say, “You should have taken appropriate precautions.” Government accordingly responds by forcing, or at least encouraging, precautionary behavior. Individuals, for example, who neither buy earthquake insur- ance nor build homes that can withstand the eff ects of an earthquake become a burden on the government when an earthquake strikes. The government fi nds itself compelled to act compassionately, even if the vic- tims’ dire situation is partly of their own making. Recognizing this, the government may compel individuals to take adequate precautions against

97Two Perspectives on the Role of Government

the event of an earthquake by, for instance, enforcing high standards for earthquake-resistant construction and making the purchase of earth- quake insurance mandatory.

TWO PERSPECTIVES ON THE ROLE OF GOVERNMENT

We saw in Chapter 1 that there are two aspects of the analysis of public sector activities: the normative approach, which focuses on what the gov- ernment should do, and the positive approach, which focuses on describ- ing and explaining both what the government actually does and what its consequences are. We can now relate our discussion of market failures, redistribution, and merit goods to these two alternative approaches.

NORMATIVE ANALYSIS

The fundamental theorems of welfare economics are useful because they clearly delineate a role for the government. In the absence of market fail- ures and merit goods, all the government needs to do is worry about the distribution of income (resources). The private enterprise system ensures that resources will be used effi ciently.

If there are important market failures—imperfect competition, imperfect information, incomplete markets, externalities, public goods, and unemployment—there is a presumption that the market will not be Pareto effi cient. This suggests a role for the government, but there are two important qualifi cations.

First, it must be shown that there is, at least in principle, some way of intervening in the market to make someone better off without making anyone worse off ; that is, of making a Pareto improvement. Second, it must be shown that the actual political processes and bureaucratic structures of a democratic society are capable of correcting the market failure and achieving a Pareto improvement.

When information is imperfect and costly, the analysis of whether the market is Pareto effi cient must take into account these information costs; information is costly to the government, just as it is to private fi rms. Markets may be incomplete because of transactions costs; the government, too, would face costs in establishing and running a public

98 CHAPTER 4 MARKET FAILURE

insurance program. These costs must be considered in the decision to set up such a program.

Recent research has established a variety of circumstances under which, although the government has no advantage in information or trans- actions costs over the private market, the government could, in principle, bring about a Pareto improvement. The fact that there may exist govern- ment policies that would be Pareto improvements does not, however, nec- essarily create a presumption that government intervention is desirable. We also have to consider the consequences of government intervention, in the form it is likely to take, given the nature of our political process. We have to understand how real governments function if we are to assess whether government action is likely to remedy market failures.

In the 1960s, it was common to take a market failure, show that a gov- ernment program could lead to a Pareto improvement (someone could be made better off without making anyone worse off ), and conclude that gov- ernment intervention was called for. When programs were enacted and failed to achieve what they were supposed to, the blame was placed on petty bureaucrats or political tampering. But, as we shall see in Chapters 8 and 9, even if bureaucrats and politicians behave honorably, the nature of government itself still may help explain government’s failures.

Public programs—even those allegedly directed at alleviating some market failure—are instituted in democracies not by ideal governments or benevolent despots, but by complicated political processes.

POSITIVE ANALYSIS

The market failure approach to understanding the role of the government is largely a normative approach. The market failure approach provides a basis for identifying situations in which the government ought to do something, tempered by considerations of government failure.

The popularity of the market failure approach has caused many programs to be justifi ed in terms of market failures. However, this may simply be rhetoric. There is often a signifi cant diff erence between a pro- gram’s stated objective (to remedy some market failure) and its design. Political rhetoric may focus on the failure of markets to provide insur- ance against volatile prices and the consequences that this has for small farmers, but government agricultural programs may, in practice, transfer income to large farmers. Insight into the political forces at work and the true objectives of the programs may be gained more easily by looking at how the programs are designed and implemented than by looking at the stated objectives of the legislation.

99Review and Practice

Some economists believe that economists should focus their attention on positive analysis, on describing the consequences of government pro- grams and the nature of the political processes, rather than on norma- tive analysis, what the government should do. However, discussions by economists (and others) of the role that government should play consti- tute an important part of the political process in modern democracies. Beyond that, an analysis of institutional arrangements by which public decisions get made may lead to designs that enhance the likelihood that the public decisions will refl ect a broader set of public interests, not just special interests. These matters will be taken up in further detail in later chapters.

SUMMARY

1. Under certain conditions, the competitive market results in a Pareto effi cient resource allocation. When the conditions required for this are not sat- isfi ed, a rationale for government intervention in the market is provided.

2. Government is required to establish and enforce property rights and enforce contracts. Without this, markets by themselves cannot function.

3. There are six reasons why the market mecha- nism may not result in a Pareto effi cient resource allocation: failure of competition, public goods, externalities, incomplete markets, information failures, and unemployment. These are known as market failures.

4. Even if the market is Pareto effi cient, there may be two further grounds for government action. First, the competitive market may give rise to a socially undesirable distribution of income. And second, some believe that individuals, even when well informed, do not make good judgments con- cerning the goods they consume, thus providing a rationale for regulations restricting the consump- tion of some goods, and for the public provision of other goods, called merit goods.

5. Even though the presence of market failures implies that there may be scope for government activity, it does not imply that a particular gov- ernment program aimed at correcting the market failure is necessarily desirable. To evaluate gov- ernment programs, one must take into account not only their objectives, but also how they are implemented.

6. The normative approach to the role of govern- ment asks: How can government address mar- ket failures and other perceived inadequacies in the market’s resource allocation? The positive approach asks: What is it that the government does, what are its eff ects, and how does the nature of the political process, including the incentives it provides bureaucrats and politicians, help explain what the government does and how it does it?

KEY CONCEPTS

Adverse selection

Incomplete markets

Libertarianism

Marginal revenue

Market failures

REVIEW AND PRACTICE

100 CHAPTER 4 MARKET FAILURE

Merit goods

Monopolistic competition

Monopoly

Natural monopoly

Negative externalities

Oligopoly

Paternalism

Perfect competition

Positive externalities

Pure public goods

Research and development (R&D)

Tragedy of the anticommons

Tragedy of the commons

QUESTIONS AND PROBLEMS

1. For each program listed below, discuss what market failures might be (or are) used as a partial rationale:

a. Automobile safety belt requirements

b. Regulations on automobile pollution

c. National defense

d. Unemployment compensation

e. Medicare (medical care for the aged)

f. Medicaid (medical care for the indigent)

g. Federal Deposit Insurance Corporation

h. Federally insured mortgages

i. Law requiring lenders to disclose the true rate of interest they are charging on loans (truth- in-lending law)

j. National Weather Service

k. Urban renewal

l. Post offi ce

m. Government prohibition of the use of narcotics

n. Rent control

2. If the primary objective of government pro- grams in each of these areas is the alleviation of some market failure, how might they be better designed?

a. Farm price supports

b. Oil import quotas (in the 1950s)

c. Special tax provisions for energy industries

3. Many government programs both redistribute income and correct a market failure. What are the market failures associated with each of these programs, and how else might they be addressed if there were no distributional objectives?

a. Student loan programs

b. Public elementary education

c. Public support for universities

d. Social Security

4. Draw the average and marginal cost curves for a natural monopoly. Draw the demand and mar- ginal revenue curves.

a. Show the effi cient level of output, at which price equals marginal cost. Explain why if the fi rm charged a price equal to marginal cost, it would operate at a loss. Show diagrammati- cally the necessary subsidy.

b. Show the monopoly level of output, at which marginal revenue equals marginal cost. Explain why the monopoly level of output is smaller than the effi cient level of output.

c. Show the level of output of a government monopoly that was instructed to just break even. How does this level of output com- pare with the effi cient level of output and the private monopoly level of output? (Show diagrammatically.)

101

PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

The government supplies a wide variety of goods, from national defense to education to police and fi re protection. Some of these goods, like edu- cation, are also provided privately; others, like national defense, are the exclusive province of government. What are the economic properties of such goods? How do they diff er from goods such as ice cream, automo- biles, and the myriad of other goods that are provided principally through private markets?

Earlier chapters noted the central role played by prices in market economies. Because of the price system, markets result in an effi cient allo- cation of resources. Prices ration private goods. Those consumers who are willing and able to pay the requisite price obtain the good. This chapter asks: What is distinctive about the goods typically provided by govern- ment? What prevents them in many cases from being provided privately? And if they are provided privately, why is the private supply likely to be inadequate?

5 1. What distinguishes public goods—those that are typically provided by governments—from privately provided goods? What do economists mean by pure public goods?

2. Why will private markets undersupply pure public goods? What is the free rider problem?

3. What are the special char- acteristics and challenges of global public goods?

4. Why do governments provide goods that are not pure public goods?

5. What determines an effi cient supply of pure public goods? How is the effi cient supply aff ected by concerns about income distribution? How is it aff ected by the fact that taxes required to pay for the public good typically introduce distortions in the economy?

6. In what sense is effi cient government a public good?

FOCUS QUESTIONS

102 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

PUBLIC GOODS

To distinguish between private and public goods, economists ask two basic questions. First, does the good have the property of rival con- sumption? Rival consumption means that if a good is used by one person, it cannot be used by another. For instance, if Lynn drinks a bottle of apple juice, Fran cannot drink that same bottle of apple juice. By contrast, non-rival consumption refers to cases in which one person’s consumption does not detract from or prevent another person’s consumption.

The classic example of non-rival consumption is national defense. If  the government creates a military establishment that protects the country from attack, all citizens are protected. National defense costs are essentially unaff ected when an additional baby is born or an additional individual immigrates to the United States. This stands in sharp contrast to private goods. It costs additional resources to provide another bottle of apple juice so Lynn and Fran can each have one. This is the only way for Lynn and Fran each to enjoy a bottle of apple juice. For a non-rival good, such as a lighthouse, although it would indeed cost more to build more lighthouses, there is essentially no additional cost for an additional ship to make use of an existing lighthouse.

The second question we ask to distinguish between private and public goods relates to the property of exclusion. Is it possible to exclude any individual from the benefi ts of the public good (without incurring great costs)? A ship going past a lighthouse, for instance, cannot be excluded from the benefi ts the lighthouse provides. Likewise, if the country is defended against attack by foreigners, then all citizens are protected; it is diffi cult to exclude anyone from the benefi ts. Clearly, if exclusion is impossible, then use of the price system is impossible, because consumers have no incentive to pay. By contrast, private goods always have the prop- erty of excludability: individuals can be excluded from enjoying a good unless they pay for it.

Generally, private goods have the properties of rival consumption and excludability; public goods are characterized by non-rival con- sumption and non-excludability. Goods for which there is no rivalry in consumption and for which exclusion is impossible are pure public goods. To develop a more complete picture of public goods (and pure public goods), we now examine the properties of non-rivalry and non-exclusion in greater detail. We will see how these properties may lead to market failures, creating a rationale for the public provision of public goods.

103Public Goods

PUBLIC GOODS AND MARKET FAILURES

To isolate the role of excludability and rivalrousness in consumption, we consider instances in which a good has one property but not the other. For some goods, consumption is non-rival but exclusion is possible. For instance, the marginal cost of an additional individual turning on a tele- vision and watching a show is zero; the number of times I watch American Idol does not detract from the number of times you can watch it. Exclu- sion is possible, however, as illustrated by the proliferation of channels available only through subscription cable and satellite TV.

Even if exclusion were possible, when a good is non-rival, there is no impetus for exclusion from the standpoint of economic effi ciency. Charging a price for a non-rival good prevents some people from enjoy- ing the good, even though their consumption of the good would have no marginal cost. Thus, charging for a non-rival good is ineffi cient because it results in underconsumption. The marginal benefi t is positive; the mar- ginal cost (of the extra person watching the show) is zero. The undercon- sumption is a form of ineffi ciency.

But if there is no charge for a non-rival good, there will be no incen- tive for supplying the good. In this case, ineffi ciency takes the form of under-supply.

Thus, there are two basic forms of market failure associated with pub- lic goods: underconsumption and undersupply. In the case of non-rival goods, exclusion is undesirable because it results in underconsumption. Without exclusion, however, there is the problem of undersupply.

PAYING FOR PUBLIC GOODS

If exclusion is possible, even if consumption is non-rival, governments often charge user fees to those who benefi t from a publicly provided good or service. Toll roads are fi nanced by user fees. The airline ticket tax can be thought of as a user fee; revenues from the ticket tax go to fi nance airports and the air traffi c control system. User fees are often thought of as an equitable way of raising revenues, as those who use the public facility the most (and therefore presumably benefi t from it the most) pay the most. However, when consumption is non-rival, user fees introduce an ineffi ciency. We can use the sort of analysis that will be introduced in Chapter 7 to measure the ineffi ciency.

This is illustrated in Figure 5.1 for the case of a bridge. We have drawn the demand curve for the bridge, describing the number of trips taken

104 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

as a function of the toll charged. Lowering the toll results in increased demand for the bridge. The capacity of the bridge is Qc�; for any demand below Qc, there is no congestion and no marginal cost associated with use of the bridge. As long as the bridge is operating below capacity, con- sumption is non-rival; additional consumption by one individual does not detract from what others can enjoy. Because the marginal cost of usage is zero, effi ciency requires that the price for usage be zero. Clearly, though, the revenue raised by the bridge will then be zero.

This is where the diff erence between public provision and private is clearest: with a single bridge, the monopoly owner would choose a toll to maximize its revenue, and would build the bridge only if those revenues equaled or exceeded the cost of the bridge. The government would face a more complicated set of calculations. It might charge the toll required to just cover the costs of construction, to break even. In doing so, it would recognize that with any toll, the usage of the bridge would be reduced, and some trips whose benefi ts exceed the social cost (here, zero) would not be undertaken. Thus, it might charge a toll less than that required to break even, raising the revenue required to fi nance the bridge in some other way. It might not even charge any toll. In mak- ing these decisions, it would weigh equity considerations—the principle that those who benefi t from the bridge should bear its costs—with effi - ciency considerations. The distortions arising from the underutilization of the bridge would need to be compared with the distortions associ- ated with alternative ways of raising revenues (e.g., taxes) to fi nance the bridge. Finally, the government might build the bridge even if the maxi- mum revenue it could obtain from the tolls was less than the cost of the

BRIDGES: HOW A USER FEE CAN RESULT IN

UNDERCONSUMPTION

If the capacity is large enough, the bridge is a non-rival good.

Although it is possible to exclude people from using the

bridge by charging a toll, p, this results in an underconsumption of the good, Qe, below the non-

toll level of consumption, Qm.

FIGURE 5.1

Bridge capacity

Demand curve for trips

Number of trips taken

Price (toll)

Qe

p

Qm Qc

105Public Goods

bridge, as it recognizes that there is some consumer surplus from the bridge: the amount that at least some individuals might be willing to pay for the bridge may be considerably greater than even the amount raised by the revenue-maximizing toll.

THE FREE RIDER PROBLEM

Many of the most important publicly provided goods—such as pub- lic health programs and national defense—have the property of non- excludability, making rationing by the price system unfeasible. For instance, the international vaccine program against smallpox virtually wiped out the disease, to the benefi t of all, whether they contributed to supporting the program or not. National defense has the property of non-excludability and zero marginal cost, but there are a few goods that have the property of at least high costs of exclusion, even though the marginal cost of using the good is positive. Congested urban streets are an example: under current technology, it is expensive to charge for the use of the street (someone could collect tolls at each corner, but the cost would be extremely high) but the throughput of the street may be limited, so if one more person uses it, another is displaced—indeed, in some cases, as more people attempt to use the street, the total through- put of the street may even be decreased, as gridlock sets in.

The infeasibility of rationing by the price system implies that the competitive market will not generate a Pareto efficient amount of the  public good. Assume that everyone valued national defense, but the government did not provide for it. Could a private firm enter to fill this gap? To do so, it would have to charge for the services it pro- vided. However, because all individuals would believe that they would benefit from the services provided regardless of whether they con- tributed to the service, they would have no incentive to pay for the services voluntarily. That is why individuals must be forced to support these goods through taxation. The reluctance of individuals to con- tribute voluntarily to the support of public goods is referred to as the free rider problem.

An example will help to illustrate the nature of this problem. In many communities, fi re departments are supported voluntarily. Some individu- als refuse to contribute to the fi re department, yet, in an area where build- ings are close together, the fi re department will usually put out a fi re in a noncontributor’s building because of the threat it poses to adjacent con- tributors’ structures. Knowing that they will be protected even if they do not pay induces some people to be free riders.

106 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

Clearly, if it is not possible to use price to ration a particular good, the good is not likely to be provided privately. If it is to be provided at all, gov- ernment will have to take responsibility.

In a few cases, non-excludable public goods are provided privately. Usually this is because there is a single, large consumer whose direct ben- efi ts are so large that it pays for the consumer to provide it for himself or herself. This consumer knows that there are free riders benefi ting from these actions, but in deciding how much to supply, the consumer looks only at his or her own direct benefi t, not at the benefi ts that accrue to others. For instance, a large shipowner might fi nd it worthwhile to install a lighthouse and light buoys, even if others cannot be excluded from enjoying the benefi ts. But in deciding how many lighthouses and buoys to construct, the shipowner looks only at the benefi ts that accrue to his or her own ships. The total benefi t of an additional buoy—including the benefi ts both to the shipowner’s ships and to others, for instance—might be considerable, even though the direct benefi t to its own ships might not warrant the additional cost. In that case, the shipowner would not put the additional buoy into place. Thus, even if there is some private provision of public goods, there will be an undersupply.

ECONOMISTS AND THE FREE RIDER PROBLEM

T he free rider problem is just a refl ection of an important incentive problem that arises in the case of public goods: If the good is going to be provided anyway, why should I pay? What I would contribute would be negligible, and would hardly alter the aggregate supply. To be sure, if everyone reasoned the same way, the good would not be supplied. That is one of the arguments for government’s providing these goods, because gov- ernment has the power to compel people to con- tribute (through taxes).

In many instances, though, public goods— volunteer fi re departments, local charities, hospitals, public TV, and many others—are being supported

voluntarily. How do we explain these? Is it possible that economists have overemphasized the “selfi sh” nature of humanity? Several recent studies in exper- imental economics have suggested that this might be the case. These experimental situations are designed to make individuals face situations in which they could free ride if they wanted to; alternatively, they could cooperate in providing for a public good. Systematically, more cooperative behavior and less free riding are observed than economists’ analysis of selfi sh incentives would suggest. Interestingly, when economists participate in these experiments, their behavior systematically seems much closer in accord with the predictions of their theory.

107Public Goods

PURE AND IMPURE PUBLIC GOODS

A pure public good is a public good for which the marginal costs of pro- viding it to an additional person are strictly zero and for which it is impos- sible to exclude people from receiving the good. National defense is one of the few examples of a pure public good. Many public goods that govern- ment provides are not pure public goods in this sense. The cost of an addi- tional person using an uncrowded interstate highway is very, very small, but not zero, and it is possible, though relatively expensive, to exclude people from (or charge people for) using the highway.

Figure 5.2 compares examples of goods that are often publicly provided with the strict defi nition of a pure public good. It shows the ease of exclu- sion along the horizontal axis and the (marginal) cost of an additional individual’s using the good along the vertical axis. The lower left-hand corner represents a pure public good. Of the major public expenditures, only national defense comes close to being a pure public good. The upper right-hand corner represents a good (ordinary health services) for which the cost of exclusion is low and the marginal cost of an additional indi- vidual using the good is high. It is easy to charge each patient for health services, and it costs a doctor twice as much to see two patients as to see one—there are signifi cant marginal costs of providing health services to additional individuals.

Many goods are not pure public goods but have one or the other prop- erty (non-rivalrousness or non-excludability) to some degree. Fire protec- tion is like a private good in that exclusion is relatively easy—individuals

PUBLICLY PROVIDED GOODS

Pure public goods are char- acterized by non-rival con- sumption (the marginal cost of an additional individual’s enjoying the good is zero) and non-excludability (the cost of excluding an individual from enjoying the good is prohib- itively high). Goods provided by the public sector differ in the extent to which they have these two properties.

FIGURE 5.2

Pure public good: National defense

Congested highway

Pure private good: Health services, education

Fire protection

Ease of exclusion

Marginal cost of use

108 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

who refuse to contribute to the fi re department could simply not be helped in the event of a fi re. However, fi re protection is like a public good in that the marginal cost of covering an additional person is low. Most of the time, fi refi ghters are not engaged in fi ghting fi res but are waiting for calls. Protecting an additional individual has lit- tle extra cost. Only in that rare event when two fi res break out simultaneously will there be a signifi cant cost to extending fi re protection to an additional person. Even here, though, mat- ters are more complicated: if we want to protect the building next door, which has paid for fi re protection, it may be necessary to put out the fi re in the building that has not paid for protection— exclusion may not really be feasible. Similarly,

although the main benefi ciary of a vaccination may be the individual protected, and there is a signifi cant marginal cost of vaccinating an addi- tional individual, the public health benefi ts from universal vaccination— the reduced incidence of the disease, possibly its eradication—are benefi ts from which no one can be excluded.

Sometimes the marginal cost of using a good to which access is easy (a good that possesses the property of non-excludability) will be high. When an uncongested highway turns congested, the costs of using it rise dramatically, not only in terms of wear and tear on the road but also in terms of the time lost by drivers using the road. It is costly to exclude by charging for road use—as a practical matter, this can be done only on toll roads, and, ironically, the tollbooths often contribute to the conges- tion. New technologies, which automatically bill regular users of the road, have radically reduced these costs.

COSTS OF EXCLUSION For many goods, the issue is not so much the feasibility of rationing, but the cost. Thus, TV and radio provided over the airwaves has one of the two properties of a public good: consumption is non-rival. However, it is now feasible to exclude some consumers, as in the use of cable TV and satellite radio, although there is a cost to exclu- sion without any benefi t to society from doing so. In other cases, such as a slightly crowded highway, there is a cost to exclusion (the cost of collect- ing tolls), and some benefi t (less congestion).

There are, of course, costs associated with exclusion for private goods as well as for public goods. Economists call these transactions costs. For example, the salaries of checkout clerks at grocery stores and collec- tors of tolls along toll highways and at toll bridges are transactions costs,

PUBLIC GOODS • Pure public goods have the properties of perfectly

non-rival consumption and non-excludability.

• With non-rival consumption, it is not desirable to exclude anyone from the benefi ts. With private provision, there will be underconsumption and/ or undersupply.

• With non-excludability, it is not feasible to exclude anyone from the benefi ts of the good. There will be a free rider problem. Such goods typically cannot be provided by the market, and when they are privately provided, they will be undersupplied.

109Public Goods

part of the administrative costs associated with operating a price mecha- nism. Although the costs of exclusion are relatively small for most private goods, they may be large (prohibitive) for some publicly provided goods.

EXTERNALITIES AS IMPURE PUBLIC GOODS Pure public goods have the property that if one individual purchases more of a good, all individuals’ consumption of that good increases by the same amount. (Individuals may, of course, diff er in how they value the increased con- sumption.) Pure private goods have the property that if one individual purchases more of a good, others are (at least directly) unaff ected. Goods for which there are externalities in consumption have the property that others are aff ected, but not necessarily in the same amount. Externalities can thus be viewed as a form of impure public goods (or, perhaps better stated, public goods can be viewed as an extreme form of externalities).1

GLOBAL AND LOCAL PUBLIC GOODS The benefi ts of some public goods are enjoyed only locally—by those living in a particular community. These are called local public goods, and include, for instance, local police security. However, more and more public goods are global public goods, the benefi ts of which accrue to anyone in the world. The most important global public goods include the global environment (everyone is aff ected by global warming, though in diff erent ways), global health (everyone can be aff ected by the spread of a pandemic like the bird fl u or other infectious diseases that can move easily across borders), global knowledge (everyone can benefi t from developments in modern science), and global security.

Steep declines in transportation and communication costs have cre- ated a world in which everyone is more interdependent. Everyone can be aff ected by threats to global economic security (we worry about the con- tagion of fi nancial crises) and physical security through global terrorism, cyber warfare, and increased nuclear proliferation.

Whenever there is a public good, there is a need for collective action; that is, individuals and fi rms acting privately will result in an undersupply of the public good. Within a country, the national government provides for national public goods, and at the local level, local governments pro- vide for local public goods; however, there is no global government that provides for global public goods. There are, in particular, no taxes that are levied globally to fi nance global public goods. The locus of decision making remains largely within the nation-state.

1�This is sometimes expressed by saying that for private goods, J’s individual utility depends only on his own purchases, XJ. For a pure public good, J’s utility depends on the sum of the purchases of all individuals: X1 1 X2 1 · · · 1 XJ 1 · · ·. When there is an externality, J’s utility may depend more heavily on his own purchases, but it may also depend on others’ more weakly; for instance, it might depend on aX1 1 aX2 1 · · · 1 XJ 1 · · · aXn, where a is a small number.

110 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

PROPERTY RIGHTS, EXCLUDABILITY, AND EXTERNALITIES

Some problems of excludability arise not from the feasibility of exclusion, but from imperfec-tions in the legal structure that make exclusion diffi cult. Some economists, such as Nobel Prize– winning economist Ronald Coase, have argued that many public-good and externality problems would be resolved if property rights, which allow exclu- sion, were established.

Consider a crowded room. The air in the room is a public good: all persons in the room breathe essentially the same air. When any individual smokes, he or she creates an externality on oth- ers. In the absence of property rights, the smoker would fail to take that into account in his or her decision to smoke. If property rights were assigned, though, the problem would (so the argument goes) be resolved. Assume the “rights” to the air were given to a particular individual. (Coase argues that, apart from distribution, to whom it is given makes no difference.) That individual would then conduct an auction: he or she would ask the smokers how

much they would be willing to pay to allow smoking, would ask the nonsmokers how much they would be willing to pay to prohibit smoking, and would then offer the air for sale to those for whom the value was highest. This would be effi cient.*

This “solution,” however, ignores the free rider problem: individual smokers have an incentive not to reveal the full value of the right to smoke (if they might actually have to pay that amount); and, simi- larly, individual nonsmokers have an incentive not to reveal the full value of clean air.

Although in some important cases, assign- ing property rights would reduce or eliminate externalities or public good problems, in some of the most important cases, assigning property rights is either impracticable or would not resolve the underlying problems. Alternative policy responses are required, such as government regulation or cre- ative solutions that utilize local social capital and traditional community norms (see Chapter 4).

*The resource allocation would be Pareto effi cient, given the assignment of the property right of the air to a single individual. Of course, however, if the smokers bid more than the nonsmokers, the smoker who winds up having to pay for the right to smoke is worse off relative to the initial situation where he or she could smoke without paying anything. Although in this example the smokers could have compensated the nonsmokers, such compensation may well not be made.

Still, the international community has come together to address many of these global public goods. International agreements and treaties, for instance, limit nuclear proliferation and attempt to protect some aspects of the global environment. In some cases, there are even enforcement mechanisms, although they tend to be relatively weak. Countries that did not comply with the agreement to reduce their utilization of gases that destroyed the vital layer of ozone in the atmosphere were threatened with trade sanctions. Typically, though, it is pressure from other countries that generates compliance and cooperation. In some cases, such as with the World Health Organization, most countries see the overwhelming bene- fi ts of cooperation, so they do so voluntarily.

111Publicly Provided Private Goods

International organizations like the World Health Organization and the broader United Nations have made important advances in global cooperation for provision of global public goods (and control of global public bads), but there is a long way to go. Developing more eff ective insti- tutional arrangements, with more enforcement, and with the ability to raise funds to fi nance global public goods, is one of the main challenges facing the world today.2

PUBLICLY PROVIDED PRIVATE GOODS

Publicly provided goods for which there is a large marginal cost asso- ciated with supplying additional individuals are referred to as publicly provided private goods. Although the costs of running a market provide one of the rationales for the public supply of some of these goods, it is not the only or even the most important rationale. Education is a publicly provided private good in the sense defi ned previously—if the number of students enrolled doubles, costs will roughly double (assuming that qual- ity, as refl ected in class size, expenditures on teachers and textbooks, and so on, are kept roughly the same). One of the usual explanations given for public provision of education is concerned with distributive consider- ations; many feel that the opportunities of the young should not depend on the wealth of their parents.

Sometimes when the government provides a private good publicly (like water), it simply allows individuals to consume as much as they want

2�The concept of local public goods was developed by J. E. Stiglitz, “Theory of Local Public Goods,” in The Economics of Public Services, ed. M. S. Feldstein and R. P. Inman (New York: MacMillan, 1977), pp. 274–333; and that of global public goods by J. E. Stiglitz, “The Theory of International Public Goods and the Architecture of International Organizations,” Background Paper No. 7, Third Meeting, High Level Group on Development Strategy and Management of the Market Economy, UNU/WIDER, Helsinki, Finland, July 8–10, 1995.

For an extensive discussion of knowledge as a global public good, see J. E. Stiglitz, “Knowledge as a Global Public Good,” in Global Public Goods: International Cooperation in the 21st Century, ed. I. Kaul, I. Grunberg, and M.A. Stern, United Nations Development Programme (New York: Oxford University Press, 1999), pp. 308–325 . For a discussion of the role of the international fi nancial institutions, such as the World Bank and the IMF, in the provision of international public goods, see J. E. Stiglitz, “IFIs and the Provision of International Public Goods,” European Investment Bank, Cahiers Papers 3, no. 2 (1998): 116–134.

For a good summary of the challenges of global public goods, see the other chapters in the Kaul et al. and W. Nordhaus, “Some Foundational and Transformative Grand Challenges for the Social and Behavioral Sciences: The Problem of Global Public Goods,” paper submitted to the National Science Foundation as part of its SBE 2020 planning activity, September 2010; for application of these general concepts to global environmental public goods, see R. Arriagada and C. Perrings, “Paying for Inter- national Environmental Public Goods,” Ecosystem Services Economics Working Paper No. 17, United Nations Environment Programme, October 2011.

112 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

without charge. Remember: for these goods, there is a marginal cost asso- ciated with each unit consumed. It costs money to purify water and to deliver it from the source to the individual’s home. If a private good is pro- vided freely, there is likely to be overconsumption of the good. Because individuals do not have to pay for the good, they will demand it until the point at which the marginal benefi t they receive from the good is zero, in spite of the fact that there is a positive marginal cost associated with pro- viding it. In some cases, such as water, satiation may be quickly reached, so the distortion from overconsumption may not be too large (Figure 5.3A). In other cases, such as the demand for certain types of medical services, the distortion may be very large (Figure 5.3B). The welfare loss can be measured by the diff erence between what the individual is willing to pay for the increase in output from Qe (where price equals marginal cost) to

DISTORTIONS ASSOCIATED

WITH SUPPLYING GOODS FREELY

(A) For some goods, such as water, supplying the good

freely rather than at marginal costs results in relatively little

additional consumption. (B) For other goods, such as certain

medical services, supplying the good freely rather than at mar- ginal costs results in extensive

overconsumption.

FIGURE 5.3

Demand curve

Price

Marginal cost

0

Welfare loss from excessive consumption

QuantityQe Qm

Welfare loss from excessive consumption

Demand curve

Marginal cost

Price

QuantityQe Qm 0

A

B

113Publicly Provided Private Goods

Qm (where price equals zero) and the costs of increasing production from Qe to Qm. This is the area of the shaded triangles in Figure 5.3.

RATIONING DEVICES FOR PUBLICLY PROVIDED PRIVATE GOODS

When there is a marginal cost associated with each individual using a good, if the costs of running the price system are very high, it may be more effi cient simply to provide the good publicly and to fi nance the good through general taxation, even though providing the good publicly causes a distortion. We illustrate this in Figure 5.4, in which we have depicted a good with constant marginal costs of production, c. (It costs the fi rm c dollars to produce each unit of the good.)3 However, to sell the good entails transactions costs, which raise the price to p*. Assume now that the government supplies the good freely. This eliminates the transactions costs, and the entire area in rectangle ABCD is saved. There is a further gain as consumption increases from Qe to Qo, as individuals’ marginal val- uations exceed the marginal costs of production. The shaded area ABE measures the gain. On the other hand, if individuals consume the good until the marginal value is zero, in expanding consumption from Qo to

3�We assume, moreover, that the demand curve does not shift signifi cantly as we raise taxes.

TRANSACTIONS COSTS

When transactions costs are suffi ciently high, it may be more effi cient to supply the good publicly than to have the good supplied by private markets.

FIGURE 5.4

Transactions costs

Demand curve

c, production costs

Quantity

Price

A

E

B

D

C

p*

F

Qe QmQo

114 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

Qm, the marginal willingness to pay is less than the cost of production. This is obviously ineffi cient. To decide whether the good should be pro- vided publicly, we must compare the savings in transactions costs plus the gain from increasing consumption from Qe to Qo with (1) the loss from the excessive consumption of the good (the shaded area EFQm in Figure 5.4) and (2)  the  loss from the distortions created by any taxes required to fi nance the provision of the good publicly.

The high costs of private markets’ providing insurance has been used as one of the arguments for the public provision of insurance. For many kinds of insurance, the administrative costs (including the selling costs) associ- ated with providing the insurance privately are more than 20 percent of the benefi ts paid out, in contrast with the administrative costs associated with public insurance, which (ignoring the distortions associated with the taxes required to fi nance the administrative costs of the social insurance programs) are usually less than 10 percent of the value of the benefi ts.

Given the ineffi ciencies arising from overconsumption when no charges are imposed for publicly provided private goods, governments often try to fi nd some way of limiting consumption. Any method restricting consump-

tion of a good is called a rationing system. Prices provide one rationing system. We have already discussed how user fees may be employed to limit demand. A second commonly employed way of rationing publicly provided goods is uniform provision: supplying the same quantity of the good to everyone. Thus, we typically provide a uniform level of free education to all individuals, even though some indi- viduals would like to have more and some less. (Those who would like to purchase more may be able to purchase supple- mental educational services like tutoring on the private market.) This, then, is the major disadvantage of the public provision of goods; it does not allow for adaptation to diff erences in individuals’ needs and desires as does the private market.

This is illustrated in Figure 5.5, in which the demand curves for two dif- ferent individuals are drawn. If the good was privately provided, Individual 1, the high demander, would consume Q1, while Individual 2, the low demander, would

THREE METHODS OF RATIONING PUBLICLY PROVIDED GOODS 1. User charges

Advantage: Those who benefi t bear the costs.

Disadvantages: Results in underconsumption.

Administering pricing system adds transactions costs.

2. Uniform provision

Advantage: Saves on transactions costs.

Disadvantages: Leads some to underconsume, others to overconsume.

High demanders may supplement public consumption, increasing total transactions costs.

3. Queuing

Advantage: Goods (like health care) allocated not necessarily on basis of who is wealthiest.

Disadvantages: Alternative basis of allocation (who has time to spare) may be undesirable.

Time is wasted.

115Publicly Provided Private Goods

consume the much smaller quantity Q2. The government chooses to sup- ply each individual with a quantity that is somewhere in between, Q*. At this level of consumption, the high demander is consuming less than he or she would like; the high demander’s marginal willingness to pay exceeds the marginal cost of production. On the other hand, the low demander is consuming more than the effi cient level; his or her marginal willingness to pay is less than the marginal cost. (Because the low demander does not have to pay anything for it, and still values the good positively, he or she, of course, consumes up to Q*.)

For certain types of insurance, say, Social Security for retirement, the government provides a basic, uniform level. Again, those who wish to pur- chase more can do so, but those who wish to purchase less cannot. The dis- tortion here may not be very great, however; if the uniform level provided is suffi ciently low, then relatively few individuals will be induced to consume more than they would otherwise, and the savings in administrative costs that we referred to earlier may more than off set the slight distortion associ- ated with the uniform provision of the basic level of insurance. On the other hand, the system of combining public and private provision may increase total transactions (administrative) costs over what they would be if only the public sector or private sector took responsibility.

A third method of rationing that is commonly employed by the gov- ernment is queuing: rather than charging individuals money for access to the publicly provided goods or services, the government requires that they pay a cost in waiting time. This allows some adaptability of the level of supply to the needs of the individual. Those whose demand for medical services is stronger are more willing to wait in the doctor’s offi ce. It is claimed that money is an undesirable basis upon which to ration medical

DISTORTIONS ASSOCIATED WITH UNIFORM PROVISION

When the publicly provided private good is supplied in equal amounts to all individuals, some get more than the effi cient level and some get less.

FIGURE 5.5 Demand curve of high demander

Demand curve of low demander

Marginal cost of production

Quantity

Price

Q2

c

Q1Q*

116 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

services: Why should the wealthy have a greater right to good health than the poor? Queues, it is argued, may be an eff ective device for discriminat- ing between the truly needy (who are willing to wait in line) and those who are less needy of medical care. But queues are far from a perfect way of determining who is deserving of medical care, as those who are unem- ployed or retired may not be so needy of medical care but more willing to wait than either the busy corporate executive or the low-paid worker holding down two jobs. In eff ect, willingness to pay has been replaced as a criterion for allocating medical services by willingness to wait in the doctor’s offi ce. There is, in addition, a real social cost to using queuing as a rationing device—the waste of time spent queuing is a cost that could be avoided if prices were used as a rationing device.

EFFICIENCY CONDITIONS FOR PUBLIC GOODS

A central concern is how large the supply of public goods should be. What does it mean to say that the government is supplying too little or too much of a public good? Chapter 3 provided a criterion that enables us to answer this question: a resource allocation is Pareto effi cient if no one can be made better off without someone else becoming worse off . There we established that Pareto effi ciency in private markets requires, among other criteria, that the individual’s marginal rate of substitution is equal to the marginal rate of transformation.

In this section of the chapter, we characterize what is required for Pareto effi ciency in the supply of pure public goods, and in particular, goods for which the marginal cost of an additional individual enjoying them is zero. Pure public goods are effi ciently supplied when the sum of the marginal rates of substitution (over all individuals) is equal to the marginal rate of transformation. The marginal rate of substitution of private goods for public goods tells how much of the private good each individual is willing to give up to get one more unit of the public good. The sum of the marginal rates of substitution thus tells us how much of the private good all the members of society together are willing to give up to get one more unit of the public good, which will be consumed jointly by all. The mar- ginal rate of transformation tells us how much of the private good must be given up to get one more unit of the public good. Effi ciency requires, then, that the total amount individuals are willing to give up—the sum of the marginal rates of substitution—must equal the amount that they have to give up—the marginal rate of transformation.

117Efficiency Conditions for Public Goods

Let’s apply this effi ciency condition to national defense. Assume that when we increase production of guns (national defense) by one, we must reduce production of butter (basic consumer products) by one pound (the  marginal rate of transformation is unity). Guns used for national defense are a public good. We consider a simple economy with two individ- uals: Crusoe and Friday. Crusoe is willing to give up one-third of a pound of butter for an extra gun, but his one-third pound alone does not buy the gun. Friday is willing to give up two-thirds of a pound of butter for an extra gun. The total amount of butter that this small society would be willing to give up, were the government to buy one more gun, is 1/3 1 2/3 5 1.

The total amount they would have to give up to get one more gun is also one. Thus, the sum of the marginal rates of substitution equals the marginal rate of transformation; their government has provided an effi cient level of national defense. If the sum of the marginal rates of substitution exceeded unity, then, collectively, individuals would be willing to give up more than they had to; we could ask each of them to give up an amount slightly less than the amount that would make them indiff erent, and it would still be possible to increase the production of guns by one unit. Thus, they could all be made better off by increasing the production of the public good (guns) by one.

DEMAND CURVES FOR PUBLIC GOODS

Individuals do not buy public goods. We can, however, ask how much indi- viduals would demand if they had to pay a given amount for each extra unit of the public good. This is not a completely hypothetical question, for as expenditures on public goods increase, so do individuals’ taxes. We call the extra payment that an individual has to make for each extra unit of the public good the tax price. In the following discussion, we shall assume that the government has the discretion to charge diff erent indi- viduals diff erent tax prices.

Assume that the individual’s tax price is p; that is, for each unit of the public good, the individual must pay p. Then, the total amount the indi- vidual can spend, the budget constraint, is:

C 1 pG 5 Y,

where C is the individual’s consumption of private goods, G is the total amount of public goods provided, and Y is the individual’s income. The budget constraint shows the combinations of goods (here, public and private goods) that the individual can purchase, given his or her income and tax price. The budget constraint is represented in Figure 5.6A by the line BB. Along the budget constraint, if government expenditures are

118 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

lower, consumption of private goods is obviously higher. The individual wishes to obtain the highest level of utility he or she can, consistent with the budget constraint. Figure 5.6A also shows the individual’s indiff er- ence curves between public and private goods. The individual is willing to give up some private goods to get more public goods. The quantity of pri- vate goods the individual is willing to give up to get one more unit of pub- lic goods is his or her marginal rate of substitution. As the individual gets more public goods (and has fewer private goods), the amount of private goods he or she is willing to give up to get an extra unit of public goods becomes smaller; that is, the individual has a diminishing marginal rate of substitution. Graphically, the marginal rate of substitution is the slope

INDIVIDUAL DEMAND CURVE FOR

PUBLIC GOODS

(A) The individual’s most preferred level of expenditure is the point of tangency between

the indifference curve and the budget constraint. (B) As the tax price decreases (the

budget constraint shifts from BB to BB9 ), the individual’s

most preferred level of public expenditure increases.

FIGURE 5.6

Budget constraints

Consumption of public goods

Private consumption

Indifference curves

B

E

E¿

B¿B G1 G2

Quantity of public goods

Tax price

Demand for public goods

E¿

Ep1

p2

G1 G2

A

B

119Efficiency Conditions for Public Goods

of the indiff erence curve. Thus, as the individual consumes more public goods and fewer private goods, the indiff erence curve becomes fl atter.

The individual’s highest level of utility is attained at the point of tan- gency between the indiff erence curve and the budget constraint, point E in Figure  5.6A. At this point, the slope of the budget constraint and the slope of the indiff erence curve are identical. The slope of the budget con- straint tells us how much in private goods the individual must give up to get one more unit of public goods; it is equal to the individual’s tax price. The slope of the indiff erence curve tells us how much in private goods the individual is willing to give up to get one more unit of public goods. Thus, at the individual’s most preferred point, the amount that he or she is willing to give up to get an additional unit of public goods is just equal to the amount he or she must give up to get one more unit of the public good. As we lower the tax price, the budget constraint shifts out (from BB to BB9 ), and the individual’s most preferred point moves to point E9. The individual’s demand for public goods will normally increase.

By raising and lowering the tax price, we can trace out a demand curve for public goods, in the same way that we trace out demand curves for private goods. Figure 5.6B plots the demand curve corresponding to Figure 5.6A. Points E and E9, from panel A, show the quantity of public goods demanded at tax prices p1 and p2. We could trace more points for panel B by shifting the budget constraint further in panel A.

We can fi rst use this approach to trace out the demand curves for public goods of two individuals, Crusoe and Friday, and then add them vertically to derive the collective demand curve in Figure 5.7. Vertical summation is appropriate because a pure public good is necessarily provided in the same amount to all individuals. Rationing is infeasible and is also undesirable, as one individual’s usage of the public good does not detract from any other individual’s enjoyment of it. Therefore, for a given quantity we add up every- one’s willingness to pay to calculate the total willingness to pay; by calculat- ing this amount at every quantity, we trace out the collective demand curve.4

The demand curve can be thought of as a “marginal willingness to pay” curve. That is, at each level of output of the public good, it says how much the individual would be willing to pay for an extra unit of the public good. (Remember, the tax price for the public good faced by the individ- ual is set equal to his or her marginal rate of substitution, which simply gives how much of the private good the individual is willing to give up for one more unit of the public good.) Thus, the vertical sum of the demand curves is just the sum of their marginal willingness to pay; that is, it is the

4�The collective demand curve is also sometimes referred to as the aggregate demand curve (not to be confused with the same term used in macroeconomics), and is the analog to the market demand curve for private goods. When constructing the market demand curve for private goods, we add up the quan- tities demanded for a given price, because all individuals face the same price but may consume dif- ferent amounts. The market demand curve is thus derived by adding up the individual demand curves horizontally.

120 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

total amount that all individuals together are willing to pay for an extra unit of the public good. Equivalently, because each point on the demand curve of an individual represents the individual’s marginal rate of sub- stitution at that level of government expenditure, by adding the demand curves vertically we simply obtain the sum of the marginal rates of sub- stitution (the total marginal benefi t from producing an extra unit). The result is the collective demand curve shown in Figure 5.7.

We can draw a supply curve just as we did for private goods; for each level of output, the price represents how much of the other goods must be forgone to produce one more unit of public goods; this is the marginal cost, or the marginal rate of transformation. At the output level at which the collective demand is equal to the supply (Figure 5.8), the sum of the mar- ginal willingness to pay (the sum of the marginal rates of substitution) is just equal to the marginal cost of production or the marginal rate of trans- formation. Because, at this point, the marginal benefi t from producing an extra unit of the public good equals the marginal cost, or the sum of the marginal rates of substitution equals the marginal rate of transformation, the output level described by the intersection of the collective demand curve and the supply curve for public goods is Pareto effi cient.

Although we constructed each individual’s demand curve for public goods in a manner analogous to the manner in which we could construct his or her demand curve for private goods, there are some important dis- tinctions between the two. In particular, although market equilibrium occurs at the intersection of the demand and supply curves, we have not provided any explanation for why the equilibrium supply of public goods

COLLECTIVE DEMAND FOR PUBLIC GOODS

Because the price is equal to the marginal rate of substitution

at each point on the demand curve, by adding the demand

curves vertically we obtain the sum of the marginal rates of

substitution, the total amount of private goods that the individu- als in society are willing to give up to get one more unit of the public good. The vertical sum

thus can be thought of as gen- erating the collective demand

curve for the public good.

FIGURE 5.7

Collective demand curve

$

Tax price paid by Crusoe

Tax price paid by Friday

Crusoe’s demand curve for public goods

Friday’s demand curve for public goods

Quantity of public goods

G

121Efficiency Conditions for Public Goods

should occur at the intersection of the demand curve we have constructed and the supply curve. We have only established that if it did, the level of production of the public good would be Pareto effi cient. Decisions about the level of public goods are made publicly, by governments, and not by individ- uals; hence, whether production occurs at this point depends on the nature of the political process, a subject discussed at length in Chapter 9.

Moreover, whereas in a competitive market for private goods all individ- uals face the same prices but consume diff erent quantities (refl ecting diff er- ences in tastes), a pure public good must be provided in the same amount to all aff ected individuals, and we have hypothesized that the government could charge diff erent tax prices for the public good. One way of thinking about these prices is to suppose that each individual is told beforehand the share of public expenditures that he or she will have to bear. If some indi- vidual must bear 1 percent of the cost of public expenditures, then an item that costs the gov- ernment $1 costs this individual 1 cent, whereas if an individual has to bear 3 percent of the cost of public expenditures, then an increase in public expenditures by $1 costs that individual 3 cents.

Finally, we should emphasize that we have characterized the Pareto effi cient level of expenditure on public goods corresponding to a particular distribution of income. As we see in the next section, the effi cient level of expen- diture on public goods generally depends on the distribution of income.

EFFICIENT PRODUCTION OF PUBLIC GOODS

An effi cient supply of public goods occurs at the point of intersection of the demand curve and the supply curve. The collective demand curve gives the sum of what all individuals are willing to give up, at the margin, to have one more unit of public goods (one more gun), whereas the supply curve gives the amount of other goods that must be given up to obtain one more unit of the public good.

FIGURE 5.8

Supply curve

Collective demand curve

Quantity

Price

EFFICIENCY CONDITION FOR PURE PUBLIC GOODS • Effi cient production requires that the sum of the

marginal rates of substitution equal the marginal rate of transformation.

• Effi cient production occurs at the intersection of the collective demand curve, formed by verti- cally adding the demand curves for each individ- ual, with the supply curve.

122 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

PARETO EFFICIENCY AND INCOME DISTRIBUTION

As discussed in Chapters 3 and 7, there are many Pareto effi cient resource allocations; any point on the utility possibilities schedule is Pareto effi - cient. The market equilibrium in the absence of market failures corre- sponds to just one of those points. By the same token, there is not a unique Pareto optimal supply of public goods. The intersection of the demand and supply curves in Figure 5.8 is one of these Pareto effi cient levels of supply, but there are others as well, with diff erent distributional implications.

To see how the effi cient level of public goods depends on the distribution of income, assume the government transferred a dollar of income from Crusoe to Friday. This would normally shift Crusoe’s demand for public goods (at any price) down and Friday’s up. In general, there is no reason why these changes should exactly off set each other, so the aggregate level of demand will normally change. With this new distribution of income, there is a new effi cient level of public goods. However, effi ciency is still characterized by the sum of the mar- ginal rates of substitution equaling the marginal rate of transformation. To put it another way, each point on the utility possibilities schedule may be charac- terized by a diff erent level of public goods, but at each point the sum of the marginal rates of substitution equals the marginal rate of transformation.

The fact that the effi cient level of public goods depends, in general, on the distribution of income has one important implication: one cannot sep- arate effi ciency considerations in the supply of public goods from distribu- tional considerations. Any change in the distribution of income, say, brought about by a change in the income tax structure, will thus be accompanied by corresponding changes in the effi cient levels of public-goods production.5

LIMITATIONS ON INCOME REDISTRIBUTION AND THE EFFICIENT SUPPLY OF PUBLIC GOODS

Governments, in evaluating the benefi ts of a public program, often seem to be particularly concerned with the question of who benefi ts from the program. They seem to weigh benefi ts that accrue to the poor more heav- ily than benefi ts that accrue to the rich. However, the previous analysis

5�Some economists have suggested that decisions concerning the effi cient level of public-goods produc- tion and distribution of income can be separated; for instance, there is a view that decisions about the distribution of income should be refl ected in tax schedules and welfare programs, but decisions about the supply of public goods can and should be made independently of such considerations. In some cases, the decisions can be separated, but these are indeed special. See Atkinson and Stiglitz, Lectures in Public Economics (New York: McGraw-Hill, 1980) or L. J. Lau, E. Sheshinski, and J. E. Stiglitz, “Effi ciency in the Optimum Supply of Public Goods,” Econometrica 46 (1978): 269–284.

123Efficiency Conditions for Public Goods

suggested that one should simply add up the marginal rates of substitu- tion, the amounts that each individual is willing to pay at the margin for an increase in the public good, treating the rich and the poor equally. How can these approaches be reconciled?

Recall our example of the Robinson Crusoe economy, in which, in the process of transferring oranges from Crusoe to Friday, some of the oranges are lost. In the U.S. economy, we use primarily the tax system and welfare system to redistribute resources. Not only are the adminis- trative costs of running these systems large, but they also have important incentive eff ects, for instance, on individuals’ savings and work decisions. The fact that redistributing resources through the tax and welfare sys- tems is costly implies that the government may look for alternative ways to achieve its redistributive goals; one way is to incorporate redistributive considerations into its evaluation of public projects.

DISTORTIONARY TAXATION AND THE EFFICIENT SUPPLY OF PUBLIC GOODS

The fact that the revenue raised to fi nance public goods is raised through distortionary taxes, such as the income tax, has some important implica- tions for the effi cient supply of public goods. The amount of private goods that individuals must give up to get one more unit of public goods is greater than it would be if the government could raise revenue in a way that did not entail distortionary incentive eff ects and that was not costly to administer.

We can defi ne a feasibility curve, giving the maximum level of private-goods consumption consistent with each level of public goods, for our given tax system. The tax system introduces ineffi ciencies, so this feasi- bility curve lies inside the production possibilities schedule, as in Figure 5.9.

The amount of private goods we must give up to obtain one more unit of public goods, taking into account these extra costs, is called the marginal economic rate of transformation, as opposed to the marginal physical rate of transformation we employed in our earlier analysis. The latter is completely determined by technology, whereas the marginal economic rate of transformation takes into account the costs associated with the taxes required to fi nance increased public expenditure. Thus, we replace the earlier condition, that the marginal physical rate of transformation must equal the sum of the marginal rates of substitution, with a new con- dition: that the marginal economic rate of transformation must equal the sum of the marginal rates of substitution.

Because it becomes more costly to obtain public goods when taxation imposes distortions, normally this will imply that the effi cient level of pub- lic goods is smaller than it would have been with nondistortionary taxation.

124 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

Indeed, it appears that much of the debate about the desirable level of pub- lic goods provision centers around this issue. Some believe that the distortions associated with the tax system are not very great, whereas others contend that the cost of attempting to raise additional revenues for public goods is great. They may agree on the magnitude of the social benefi ts that may accrue from additional government expenditures, but disagree on the costs.

EFFICIENT GOVERNMENT AS A PUBLIC GOOD

One of the most important public goods is the management of the govern- ment: we all benefi t from a better, more effi cient, more responsive gov- ernment. Indeed, “good government” possesses both of the properties of public goods we noted earlier: it is diffi cult and undesirable to exclude any individual from the benefi ts of a better government.

If the government is able to become more effi cient and reduce taxes without reducing the level of government services, everyone benefi ts. The politician who succeeds in doing this may get some return, but this return is only a fraction of the benefi ts that accrue to others. In particular, those who voted against the politician who succeeds in doing this gain as much as those who worked for the politician’s election, and the individual who did not vote, who attempted to free ride on the political activities of others, benefi ts as much as either.

THE FEASIBILITY CURVE

The feasibility curve gives the maximum output (consumption)

of private goods for any level of public goods, taking into

account the ineffi ciencies that arise from the taxes that must be imposed to raise the

requisite revenue. The feasibility curve lies below the production

possibilities schedule.

FIGURE 5.9

Production possibilities

curve

Feasibility curve

Public good

Private good

125Review and Practice

SUMMARY

1. This chapter has defi ned an important class of goods, pure public goods. They have two critical properties:

a. It is impossible to exclude individuals from enjoying the benefi ts of the goods (non- excludability).

b. The marginal cost of an additional individual enjoying the good is zero (non-rival consump- tion). It is undesirable to exclude individu- als from enjoying the benefi ts of the goods, since their enjoyment of these goods does not detract from the enjoyment of others.

2. Although there are a few examples of pure public goods, such as national defense, for many publicly provided goods exclusion is possible, although frequently costly. Charging for use may result in the underutilization of public facilities. For many publicly provided goods, there is some marginal cost of an individual enjoying the good. Although, for example, the marginal cost of an individual using a completely uncongested road may be neg- ligible, if there is some congestion, the marginal cost may be more signifi cant.

3. Private markets either will not supply or will pro- vide an inadequate supply of pure public goods.

4. The problem with voluntary arrangements for providing public goods arises from individuals trying to be free riders and enjoying the benefi ts of the public goods paid for by others.

5. Developing effi cient and eff ective mechanisms for the provision of global public goods is one of the greatest challenges now facing the interna- tional community.

6. For many publicly provided goods, consump- tion is rivalrous; consumption by one individual reduces that of another; or the marginal cost of supplying an extra individual may be signifi - cant, equal to, or even greater than, the average cost. These are called publicly provided private

REVIEW AND PRACTICE

goods. If they are supplied freely, there will be overconsumption.

7. For publicly provided private goods, some method of rationing other than the price system may be used; sometimes queuing is used, whereas at other times the good is simply provided in fi xed quantities to all individuals. Both of these entail ineffi ciencies.

8. Pareto effi ciency requires that a public good be supplied up to the point at which the sum of the marginal rates of substitution equals the mar- ginal rate of transformation. Diff erent Pareto effi cient levels of consumption of the public good will be associated with diff erent distributions of income.

9. The basic rule for the effi cient level of supply of public goods must be modifi ed when there are costs (distortions) associated with raising reve- nue and redistributing income.

10. Effi cient management of the government is a pub- lic good in itself.

KEY CONCEPTS

Collective action

Collective demand curve

Exclusion

Feasibility curve

Free rider problem

Global public goods

Leftover curve

Marginal economic rate of transformation

Marginal physical rate of transformation

Non-rival consumption

Publicly provided private goods

Pure public goods

Queuing

Rationing system

126 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

Rival consumption

Tax price

Transactions costs

Underconsumption

Undersupply

Uniform provision

User fees

QUESTIONS AND PROBLEMS

1. Where should each of the following goods lie in Figure 5.2? Explain why each is or is not a pure public good. Where applicable, note instances where the good is both publicly and privately provided:

a. College education

b. A local park

c. Yosemite National Park

d. Sewage collection

e. Water

f. Electricity

g. Telephone service

h. Retirement insurance

i. Medicine

j. Police protection

k. TV

l. Basic research

m. Applied research

2. What happens to the effi cient allocation between public and private goods as an economy becomes wealthier? Can you think of examples of public goods, the consumption of which would increase more than proportionately to the increase in income? Less than proportionately to the increase in income?

3. The government rations a variety of publicly provided private goods and impure public goods in which there is congestion in a variety of ways. Discuss how each of these is rationed, and con- sider the eff ect of alternative rationing systems:

a. Public higher education

b. Health services in the United Kingdom

c. Yellowstone National Park

What happens to a publicly provided good in which congestion can occur (e.g., a highway or swimming pool on a hot, sunny day), but no direct rationing system is employed?

4. To what extent do you think diff erences in views between advocates of less spending on public goods and advocates of more spending can be attributed to diff erent assessments of the mar- ginal cost of public goods, including the increased distortions associated with the additional taxes required to fi nance public goods? What are other sources of disagreement?

5. What implications might the fact that effi cient government is a public good have for the effi - ciency with which governments function?

6. Discuss the issue of vaccination from the per- spective of public goods/externalities. Why might individuals not consent voluntarily to be vaccinated?

7. There has been increasing concern about increased atmospheric concentrations of green- house gases, such as carbon dioxide, which are likely to lead to global warming. Discuss the world’s atmosphere as an “international” public good. What are some of the problems of ensur- ing that individuals and countries take actions to reduce emissions of greenhouse gases?

8. Discuss how changes in income, technology, or other changes in the economic environment may lead to changes in the balance between public and private provision. Illustrate, for instance, by a discussion of the role of public parks.

127

APPENDIX: THE LEFTOVER CURVE

In this appendix we provide an alternative, diagrammatic exposition for the basic effi ciency condition for public goods discussed in the chapter:

Sum of marginal rates of substitution 5 Marginal rate of transformation.

In Figure 5.10A we have superimposed Crusoe’s indiff erence curve on the production possibilities schedule. If the government provides a level of public goods G, and wishes, at the same time, to ensure that Cru- soe attains the level of utility associated with the indiff erence curve U1 drawn in the fi gure, then the amount of private good that is “left over” for Friday is  the vertical distance between the production possibilities schedule and the indiff erence curve. Accordingly, we call the (vertical) diff erence between the two the leftover curve. (This curve is plotted in Figure  5.10B.) We  now superimpose Friday’s indiff erence curves on Figure 5.10B. The highest level of utility he can attain, consistent with the production possibilities schedule, and consistent with the prespecifi ed level of utility of Crusoe, is at the point of tangency between his indiff er- ence curve and the leftover curve.

There is a simple way to express this tangency condition. Because the leftover curve represents the diff erence between the production possi- bilities schedule for the economy and the fi rst individual’s indiff erence curve, the slope of the leftover curve is the diff erence between the slope of the production possibilities schedule and the slope of the fi rst individual’s indiff erence curve. The slope of the production possibilities schedule is, as we just saw, the marginal rate of transformation, whereas the slope of the fi rst individual’s indiff erence curve is his or her marginal rate of sub- stitution. If G is the optimal level of public goods, the leftover curve must be tangent to the second individual’s indiff erence curve. Hence, Pareto effi ciency of the economy requires that the slope of the leftover curve be equal to the slope of the second individual’s indiff erence curve—that is,

MRT 2 MRS1 5 MRS2

or

MRT 5 MRS1 1 MRS2,

where MRT stands for the marginal rate of transformation, MRS1 for the fi rst individual’s marginal rate of substitution, and MRS2 for that of the  second individual. The marginal rate of transformation must equal the sum of the marginal rates of substitution.

128 CHAPTER 5 PUBLIC GOODS AND PUBLICLY PROVIDED PRIVATE GOODS

It should be clear that if we chose a diff erent (say, higher) initial level of utility for Crusoe, the leftover curve would be shifted (down), and there would be a new point of tangency of Friday’s indiff erence curve with the new leftover curve. At the new point of tangency, the level of expenditure on public goods may be higher, lower, or the same as in the initial situa- tion. This illustrates the point made in the text: there is not necessarily a single “effi cient” level of expenditure on public goods; there are many Pareto effi cient levels of expenditures, depending on the distribution of income (welfare). Issues of distribution and allocation cannot, in general, be separated.

DETERMINATION OF THE EFFICIENT LEVEL OF PRODUCTION OF

PUBLIC GOODS

(A) If the level of public goods is G, and Crusoe is to get level of utility U1, then the distance AB represents the amount of pri-

vate goods left over for Friday. (B) Friday’s welfare is maximized

at the point of tangency of his indifference curve and the

“leftover” curve.

FIGURE 5.10

Production possibilities

curve

Crusoe’s indifference

curve

Friday’s indifference

curve

Leftover curve

Public good

Public good

Private good

Private good

A

B

C

G B

A

U1

U1

U2

U2

U2¿

U2¿

A

B

129

EXTERNALITIES AND THE ENVIRONMENT

The federal government has long had an interest in environmental policy—the earliest federal action appears to have been the Refuse Act of 1899, designed to rid navigable waters of debris. The modern era of federal environmental regulation began with the Water Pollution Control Act of 1948, the fi rst of a series of laws to protect the water we drink and the lakes and rivers in which we swim and fi sh.

Government activity on behalf of the environment has clearly had some benefi cial eff ect. The quality of air in major industrial cities such as Gary, Indiana and Pittsburgh has improved noticeably since passage of the Clean Air Act of 1963. Lakes such as Lake Erie, which once faced the prospect of becoming so polluted that much marine life would be extin- guished, have been saved.

Although it is increasingly agreed that government actions are required to preserve our environment, the extent and form of those actions remain a subject of debate. This chapter describes the economic rationale for government intervention in the environment and reviews the major government programs and policy issues related to environmen- tal intervention.

6 1. What are externalities?2. How do private markets respond to externalities? What are the limitations of these private remedies?

3. What are the principal ways by which the public sector attempts to deal with externalities? What are the advantages and disadvantages of these alternative approaches?

4. What currently are the major environmental public policy issues? What policies regarding these issues have succeeded and what policies have failed? What are some of the current controversies in environmental public policy, and what insights does economic analysis provide into these controversies?

FOCUS QUESTIONS

130 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

THE PROBLEM OF EXTERNALITIES

Air and water pollution are two examples of a much broader range of phe- nomena that economists refer to as externalities, one of the market fail- ures discussed in Chapter 4. Whenever an individual or fi rm undertakes an action that has an eff ect on another individual or fi rm for which the latter does not pay or is not paid, we say there is an externality. Markets aff ected by externalities result in ineffi cient resource allocations. Levels of produc- tion, as well as expenditures directed at controlling the externality, will be incorrect. For instance, consider a fi rm that could, by expending resources, reduce its level of pollution. Although there would be a large social benefi t, there is no private incentive driving the fi rm to spend the money.

In some cases, the actions of an individual or fi rm confer (uncompen- sated) benefi ts on others; these are called positive externalities. A home- owner who maintains his or her property, including planting attractive fl owers in front, provides a positive externality. Actions that adversely aff ect others are called negative externalities.

The level of production of negative externality-generating commodi- ties will be excessive. Figure 6.1 shows conventional demand and supply curves. We argued earlier that, in the absence of externalities, the result- ing market equilibrium, Qm, was effi cient. The demand curve refl ected the individual’s marginal benefi ts from the production of an extra unit of the commodity, and the supply curve refl ected the marginal costs of

FIGURE 6.1

EXCESSIVE PRODUCTION OF GOODS YIELDING

NEGATIVE EXTERNALITIES

The presence of a negative externality means that marginal

social costs exceed marginal private costs, and the market

equilibrium will entail an excessive production of the

commodity. Qm is market equilibrium, Qe is the effi cient

level of output. Demand curve

(marginal benefit)

Supply curve (marginal private cost)

Marginal social cost

Quantity of steel

Price

Qe Qm

131The Problem of Externalities

producing an extra unit of the commodity. At the intersection of the two curves, the marginal benefi ts just equal the marginal costs. Now, with externalities, the industry’s supply curve will not refl ect marginal social costs, only marginal private costs—those borne directly by the producers. If the expansion of steel production increases the level of pollution, there is a real cost to that expansion in addition to the costs of the iron ore, labor, coke, and limestone that go into the production of steel. However, the steel industry fails to take the cost of pollution into account. Thus, Figure 6.1 also shows the marginal social cost curve, giving the total extra costs (private and social) of producing an extra unit of steel. This cost curve lies above the industry supply curve. Effi ciency requires that marginal social cost equal the marginal benefi t of increasing output: production should occur at Qe, the intersection of the marginal social cost curve and the demand curve. The effi cient level of production is lower than the market equilibrium level.

An important class of externalities arises from what is referred to as common resource problems. Their central characteristic is that they pertain to a pool of scarce resources to which access is not restricted. Consider a lake in which the total number of fi sh caught increases with the number of fi shing boats, but less than proportionately, so that the number of fi sh caught per boat decreases as the number of boats increases. Each additional boat reduces the catch of other boats. This is the externality. The marginal social benefi t of an additional boat is thus less than the average catch of each boat, as shown in Figure 6.2; some of the fi sh that the additional boat catches would have been caught by some other boat.

COMMON RESOURCE PROBLEM LEADS TO EXCESSIVE FISHING

The extra output of an additional boat is less than the average output. There will be an excessive number of boats.

FIGURE 6.2

Average return to a boat

Marginal social return

Efficient number of

boats

Market equilibrium

Cost of a boat

Number of boats

Output per boat

132 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

The private return to an additional individual deciding whether to purchase a boat is simply the average return (once they are on the lake, all boats catch the same number of fi sh), which is much more than the marginal social return. Thus, whereas the private market equilibrium entails average returns equal to the cost of a boat (assumed to be constant), social effi ciency requires that the marginal social return be equal to the cost of a boat.

In general, when there are externalities, the market equilibrium will not be effi cient.

PRIVATE SOLUTIONS TO EXTERNALITIES

Under some circumstances, private markets can deal with externalities without government assistance.

INTERNALIZING EXTERNALITIES

The simplest way this can be done is by internalizing externalities— forming economic units of suffi cient size so that most of the consequences of any action occur within the unit. Consider, for instance, any community, whether a group of neighboring houses or a set of apartments in the same or neighboring buildings. The quality of life in the neighborhood is aff ected by how each household maintains its property. If people plant fl owers, they confer a positive externality; if they let their houses run down, they confer a negative externality. Even when each family owns its own apartment, the households may collectively decide that maintenance of the facilities that aff ect them all—including the external appearance—should be undertaken collectively. They form a cooperative or a condominium association.

There must be, of course, some way of enforcing the collective agree- ment that those who purchase a condominium or an apartment in a coop- erative sign. A member of the condominium association might prefer to be a free rider, not paying his or her share of the cost of the maintenance of the common facilities; or the member might refuse to maintain his or her apartment in ways that are collectively agreed upon, and that may

EXTERNALITIES Externalities arise whenever an individual or fi rm undertakes an action that has an effect on another individual or fi rm for which the latter does not pay or is not paid. The consequences are:

1. Overproduction of goods generating negative externalities

2. Undersupply of goods generating positive externalities

133Private Solutions to Externalities

adversely aff ect neighboring apartments. There must be recourse to the legal system, which ensures that the terms of the agreement—by which those living near each other attempt to deal with some of the externalities they impose on each other and to provide what are “public goods” to the group—are adhered to.

THE COASE THEOREM

As we have noted, externalities arise when individuals do not have to pay for the full consequences of their actions. There is excessive fi shing in a common pool because individuals do not have to pay for the right to fi sh. Frequently, externalities can be dealt with by the appropriate assignment of property rights. Property rights assign to a particular individual the right to control some assets and to receive fees for the property’s use.

Consider the problem of oil pools. Oil is usually found in large pools beneath the ground. To obtain access to a pool, all one needs to do is to buy enough land to drill a well and equipment for the drilling. The more oil that one well takes out of the pool, the less there is for others to take.1 The total extra oil obtained as a result of drilling an extra well—the marginal social benefi t—is thus less than the amount obtained by the additional well. Too many wells will be drilled.

The reason for this is that no one has the property right to the entire pool of oil. When the oil pool is controlled by a single individual, that indi- vidual has an incentive to make sure that the correct number of wells is drilled. Because economic effi ciency is enhanced by having a single fi rm control the entire pool, any fi rm could buy the land over the pool from its present owners (at what they would have received from selling the oil) and wind up with a profi t. In this view, no outside intervention would be required to ensure that an effi cient pattern of property rights emerged.

Even when property rights for a common resource are not assigned to a single individual, the market may fi nd an effi cient way of dealing with the externality. Owners of oil wells frequently get together to unitize their production, thus making it less likely that too many wells will be drilled.2 Fishermen using the same grounds may get together to devise mutually agreed-upon restrictions to prevent excess fi shing.

The assertion that whenever there are externalities the parties involved can get together and make some set of arrangements by which

1�There is another externality: as oil is removed, the costs of pumping out additional oil rise because under- ground pressure is reduced. Additional wells may actually reduce the total amount that will be extracted. 2�Under unitization, the development of an oil or gas reservoir is put under a single management, with proceeds distributed according to a formula specifi ed in the unitization agreement. This unitization is not done to reduce competition (it occurs even among small oil companies that take the price of oil as given, unaff ected by their actions), but to increase effi ciency.

134 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

the externality is internalized and effi ciency is ensured is referred to as the Coase theorem.3

For instance, when there are smokers and nonsmokers in the same room, if the loss to the nonsmokers exceeds the gains to the smokers, the nonsmokers might get together and “bribe” (or, as economists like to say, “compensate”) the smokers not to smoke. Or, say the smokers are in a non- smoking compartment of a train, and the restriction on smoking (which can be viewed as an externality imposed on the smokers by the nonsmok- ers) takes away more from their welfare than the nonsmokers gain. Then the smokers might get together and “compensate” the nonsmokers in order to allow themselves to smoke.

Of course, the determination of who compensates whom makes a great deal of diff erence to the distributive implications of the externality. Smokers are clearly better off in the regime in which smoking is allowed unless they are paid not to smoke, compared to the regime in which smok- ing is banned unless they compensate nonsmokers.

USING THE LEGAL SYSTEM

Even when property rights are not perfectly defi ned, the legal system can provide protections against externalities. Our system of common law does not allow one party to injure another, and “injury” has been interpreted to include a variety of economic costs imposed on others. Implicitly, courts have given individuals some property rights—say, in the waters that they rely on for fi shing. Those who have been injured have increasingly turned to courts to enforce those property rights.

When the Exxon tanker Valdez spilled oil into Alaska’s Prince William Sound in 1989, those damaged by the spill—the fi shermen whose catch was diminished, as well as those in the tourist industry who depend on sports fi shermen—successfully sued Exxon. Many Americans believed that by spoiling one of the relatively pristine environments in the coun- try, the spill hurt them too. They valued the existence of these natural resources, even if they did not immediately enjoy the benefi ts by visiting Alaska; to that extent, the Valdez oil spill had an externality eff ect upon them too. Courts have recognized these existence values—in the Valdez case, the state of Alaska, acting as trustee, collected more than a billion dollars in compensation.

Similar issues are now being litigated after the April 2010 explosion and fi re on the BP-licensed Transocean drilling rig Deepwater Horizon

3�R. H. Coase, “The Problem of Social Cost,” Journal of Law and Economics 3 (1960): 1–44.

135Private Solutions to Externalities

in the Gulf of Mexico and the subsequent worst oil spill in U.S. history. Although BP set up an independently administered $20 billion escrow fund while oil was still gushing into the gulf, the U.S. government saw this as a down payment toward compensation for victims of the oil spill. It fi led suit in December 2010 against BP and several of its partners

THE EXXON VALDEZ OIL SPILL

Oil tankers have long been a major source of ocean pollution. The potential for dam-age was forcefully brought home with the grounding of the Exxon Valdez in Alaska’s Prince William Sound in March 1989. Nightly pictures on TV depicted graphically the massive death of wild- life, including sea otters, salmon, birds, and seals. How long the devastation would last—or whether nature would ever fully recover—was not clear.

Exxon was made to pay more than $1 billion, most of which was to be spent on correcting the environmental damage; and the company claimed to have spent more than $2 billion beyond that in the months immediately after the spill, trying to limit the extent of damage. Even so, there was debate over whether the amount paid by Exxon was adequate: How much should Americans be com- pensated for the damage of the spill?

To answer this question, a study was done in which individuals were asked questions about how much they would be willing to pay to preserve a natural hab- itat, such as that which was harmed by the Valdez spill. Just as opinion polls, by sampling a thousand indi- viduals, can provide an accurate forecast about how the entire population will vote, so, too, a sample of individuals can provide an accurate estimate for the value that would be assigned by the entire popula- tion. Some individuals will assign a high value, others a relatively low value, but these differences will be refl ected in the sample. By projecting the distribution of values in the sample to the entire population, one can calculate the total value for the nation. In the case of the Valdez oil spill, the value estimated in this way

was about $3 billion. This methodology for assessing existence values is called contingent valuation.

The more fundamental question was how to pre- vent such disasters or, more accurately, how to make their occurrence less likely and the consequences less severe. As long as oil is transported, there is some risk of a spill, and no one has contemplated a complete ban on shipping oil. Shippers may not have the appropriate incentives to avoid a spill, though, because they do not bear the full consequences. This is a particularly severe problem, as many shipping companies are poorly capi- talized, and in the event of an accident they would sim- ply go bankrupt. Only a company as large and strong as Exxon could pay out $3 billion, yet almost any large oil tanker could do comparable damage.

To rectify this problem, Congress passed the Oil Pollution Act of 1990. This combined a system of incen- tives with regulations. Vessels had to be double-hulled, for example, thus reducing the likelihood of spillage.

One criticism that economists have raised is that the funds paid in compensation for damage, in gen- eral, must be used for cleanups. This constraint induces an ineffi ciency. The amount that the owners of vessels that have spills are required to pay should be designed to provide the corrective incentive to avoid spills. This may be more than the amount that is appropriately spent on cleanup. For instance, suppose that the con- sequences of a spill would be rectifi ed by nature on its own in a year, and that it would cost an enormous amount to speed up the restoration process. We still would want to penalize fi rms that spill, but we might not want to spend the money to speed the restoration, as there would be little benefi t to doing so.

136 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

over violation of safety regulations, seeking unlimited damages to cover the cost of the oil cleanup, long-term environmental damage, and local business losses.

To reduce the uncertainty about these often imperfectly defi ned property rights, govern- ment has tried to clarify them and to specify more precisely the nature and amount of dam- ages that can be collected. Thus, more recent legislation and regulation have recognized the

importance of existence values; the government—as “trustee” for the coun- try’s natural resources—has the right to sue for damages, although under current legislation the amount recovered must be used for restoration.

FAILURES OF PRIVATE SOLUTIONS

If the arguments asserting that private markets can internalize externali- ties are correct, is there any need for government intervention, other than to establish clear property rights? Furthermore, if these arguments are correct, why have cooperative agreements failed to take care of so many externalities?

There are several reasons why government intervention is required. The fi rst has to do with the public goods problem discussed in Chapter 5. Many (but far from all) externalities involve the provision of a public good, such as clean air or clean water; in particular, it may be very costly to exclude anyone from enjoying the benefi ts of these goods. If nonsmok- ers get together to compensate smokers for not smoking, it pays any indi- vidual nonsmoker to claim that he or she is almost indiff erent to letting others smoke. This individual will attempt to be a free rider on the eff orts of other nonsmokers to induce the smokers not to smoke.

The problems of arriving voluntarily at an effi cient solution are exac- erbated by the presence of imperfect information. Smokers will try to per- suade nonsmokers that they require a lot of compensation to induce them not to smoke. In any such bargaining situation, one party may risk the possibility of not arriving at a mutually advantageous agreement in order to get more out of any bargain that might be made.

Problems may arise even when markets are well established. Consider the problem of an oil pool, the land above which is owned by several indi- viduals. Effi ciency can be obtained by bringing all the land covering an oil pool under a single unitized management and control—called unitization.

PRIVATE SOLUTIONS TO EXTERNALITIES 1. Internalize externality.

2. Assign property rights (Coase theorem).

3. Use the legal system.

137Private Solutions to Externalities

However, if all but one of the landowners unit- ize, it may not pay for the last owner to join. This landowner knows that production on the unitized portion will be reduced, thus enabling him or her to increase production. The last owner will join only if he or she receives more than a proportion- ate share of the revenues. Then, all small owners may believe they can gain by holding out to be the last to join the unitization agreement (or selling to a large fi rm attempting to purchase all the small owners). States have therefore found it necessary to pass legislation requiring unitization.

Another reason for government interven- tion concerns transactions costs. The costs of getting individuals together to internalize externalities voluntarily are signifi cant. The provision of those organizational services itself is a public good. Indeed, the government may be looked upon as precisely the mechanism that individuals have set up to reduce the welfare losses from externalities.

Transactions costs are a major disadvantage of dealing with externali- ties through judicial processes. For many externalities, the losses involved may simply be too small to justify undertaking litigation. Because those generating externalities know that litigation is expensive, they may be inclined to generate their externality just up to the point at which it pays the injured party to sue—giving rise to considerable ineffi ciencies. One way of partially dealing with this is to charge anyone shown to have imposed an externality on another a multiple of the estimated value of the damages. However, this gives rise to a countervailing danger of unwar- ranted lawsuits, with defendants settling claims simply to avoid the enor- mous litigation costs.

Uncertainty about the extent of the injury frequently compounds the problem of transactions costs, and there is also some ambiguity about the outcome of most suits. If litigation costs are large, the uncertainty acts as a further deterrent to individuals contemplating using the court system to deal with externalities.

The high litigation costs and uncertain outcome of the litigation pro- cess imply that there is, in eff ect, diff erential access to legal remedies— poor people may not be able or willing to bear the risks of litigation—a situation that confl icts with our usual notion of justice in a democracy. Because the legal system and the other private methods of addressing externalities so often work so poorly and so inequitably, there has been increasing reliance on public remedies.

FAILURES OF PRIVATE REMEDIES FOR EXTERNALITIES • Public good (free rider) problems

• Compounded by imperfect information problems

How much does the individual need to be compensated for externality?

Incentive not to reveal truth • Transactions costs

• Additional problems with litigation

Uncertainty about outcomes Differential access

138 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

DOUBLE DIVIDEND

Some have argued that there is a double div-idend from imposing taxes (or fi nes) on pol-lution. Not only does it discourage pollution, but it also raises revenue, so the government has to rely less on distorting taxes. Those who believe that the tax system is distorting—with taxes on capital discouraging savings and taxes on labor

discouraging work—emphasize this double divi- dend. Not only will national output as convention- ally measured go up, but true output—which takes account of the pollution and degradation of the environment—goes up even more than convention- ally measured output, as the tax or fi ne discourages pollution.

PUBLIC SECTOR SOLUTIONS TO EXTERNALITIES

Public sector solutions to environmental externalities fall into two broad categories: market-based solutions and direct regulation. Market-based solutions attempt to infl uence incentives to ensure economically effi - cient outcomes. For instance, fi nes for polluting can be used to present fi rms with the true social costs of their actions, thereby diminishing their incentive to pollute. By contrast, government has used direct regulations to limit externalities, as in the case of mandatory emissions standards for automobiles.

Before comparing the merits of these diff erent approaches, we should fi rst dispel the common fallacy that asserts that an individual or fi rm should never be allowed to impose a negative externality on others. For example, it is sometimes asserted that a fi rm should never be allowed to pollute the air and water. In the view of most economists, such absolut- ist positions make no sense. There is indeed a social cost associated with pollution (or any other negative externality), but the cost is not infi nite; it is fi nite. There is some amount of money that people would be willing to receive in compensation for having to live in a community with dirtier air or dirtier water. Thus, we need to weigh the costs and benefi ts associated with pollution control, just as we need to weigh the costs and benefi ts associated with any other economic activity. The problem with the market is not that it results in pollution; there is, indeed, a socially effi cient level

139Public Sector Solutions to Externalities

of pollution. The problem rather is that fi rms fail to take into account the social costs associated with the externalities they impose—in this case, pollution—and as a result, the level of pollution is likely to be excessively high. The task of the government is to help the private sector achieve the socially effi cient level of pollution, to make individuals and fi rms act in such a way that they are induced to take into account the eff ects of their actions on others.

In the ensuing discussion the focus will be on pollution externalities. The arguments, however, extend in a straightforward way to other cate- gories of externalities.

MARKET-BASED SOLUTIONS

Even when markets themselves do not lead to effi cient resource allocations—as when there are externalities—market-like mechanisms can often be used to ensure effi cient behavior. Market-based solutions to environmental externalities take three forms: fi nes and taxes, subsidies for pollution abatement, and marketable permits. We now consider each of these solutions in turn.

FINES AND TAXES�The simplest form of market-based solution involves levying fees or taxes in proportion to the amount of pollution emitted. In general, whenever there is an externality, there is a diff er- ence between the social cost and the private cost, and between the social benefi t and private benefi t. A properly calculated fi ne or tax presents the individual or fi rm with the true social costs and benefi ts of its actions. Fines of this sort—designed to make marginal private costs equal mar- ginal social costs, and marginal private benefi ts equal to marginal social benefi ts—are called corrective taxes, or sometimes Pigouvian taxes, after A. C. Pigou, a great English economist of the fi rst half of the twen- tieth century.4

Consider the example, discussed earlier, of steel producers polluting the air. We showed that because fi rms were concerned only with pri- vate marginal costs, not the social marginal costs (the two diff ering by the marginal costs of pollution), the output of steel would be excessive. By charging each fi rm an amount equal to the marginal cost of pollution, though, the marginal private costs and marginal social costs are equated.

4�Pigou argued persuasively for the use of corrective taxes in his book, The Economics of Welfare (London: Macmillan, 1918).

140 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

In Figure 6.3 we have assumed that the amount of pollution is propor- tional to the level of output, and the marginal cost of each unit of pollution is fi xed; hence, by imposing a fi xed charge per unit of output, equal to the marginal social cost of pollution, each fi rm will be induced to produce the socially effi cient level of output. In the fi gure, the distance EA represents the pollution tax per unit output, and the area EABC represents the total pollution taxes paid.

Firms can reduce pollution by producing less, or by changing produc- tion methods. Changes in production methods may entail direct expen- ditures for pollution control devices, or changes in the input mixes and other alterations in the production process. Fines related directly to the amount of pollution ensure that fi rms will undertake the pollution abatement in the least costly—most effi cient—manner possible. Assume that there is a given, known marginal social cost imposed on others by each unit of pollution (measured, say, by the number of particles added to the air per unit of time). It is costly to reduce pollution—and we assume that at any given level of production, it costs more to reduce pollution more. In other words, the marginal cost of pollution control is rising. This is depicted in Figure 6.4, in which we measure along the horizontal axis the reduction in pollution (from what it would be if the fi rm spent nothing on pollution abatement). Effi ciency requires that the  marginal social benefi ts associated with further pollution abate- ment expenditures just equal the marginal social costs, point P* in the diagram. If the fi rm is charged a fi ne, f*, equal to the marginal social cost of pollution, the fi rm will undertake the effi cient level of expendi- ture on pollution abatement.

MARKET EQUILIBRIUM WITH AND WITHOUT

FINES

In the absence of a tax on pollution, fi rms will set price

equal to marginal private cost. There will be excessive

production (Qm). By setting a tax equal to the marginal pollution

cost, effi ciency is obtained.

FIGURE 6.3

Demand (= marginal social benefit)

Pollution tax per unit = marginal cost of pollution

Marginal private cost

Marginal social cost (including marginal

cost of pollution)

Qutput of steel

Price ($)

Qe

B

C E

A

Qm 0

141Public Sector Solutions to Externalities

SUBSIDIZING POLLUTION ABATEMENT Because a fi rm is likely to receive a negligible direct benefi t from pollution abatement (most of the benefi ts accruing to those who live in the vicinity of the plant), absent a fi ne on pollution, it has little incentive to spend money on pollution abatement. There is, from a social point of view, too little expenditure on pollution abatement. Rather than taxing pollution, the government could subsidize pollution abatement expenditures. By providing a sub- sidy equal to the diff erence between the marginal social benefi t of pollu- tion abatement and the fi rm’s marginal private benefi t, the effi cient level of pollution abatement expenditures can be attained. This is illustrated in Figure 6.5. The marginal cost of pollution depicted in Figure 6.4 is directly related to the marginal benefi t of pollution abatement depicted in Figure 6.5. Whereas in Figure 6.4 we assumed a fi xed marginal social cost of pollution, and hence a fi xed marginal social benefi t from pollution abatement, in Figure 6.5, as pollution decreases, the marginal social bene- fi t from further pollution abatement decreases. Similarly, in Figure 6.4 we have assumed rising costs of pollution abatement, whereas in Figure 6.5 marginal costs are constant. Either case may hold in a real situation.

This remedy, however, does not attain a socially effi cient resource allo- cation. The reason is simple: the total marginal social costs of producing steel include the costs of the government subsidies for pollution abatement. Firms fail to take this into account in deciding on the level of production. Thus, as before, the marginal social cost of steel production exceeds the marginal private cost. The pollution abatement subsidy reduces the mar- ginal social cost of output (from the dashed line to the solid black line in Figure 6.6). But it also reduces the marginal private costs. There is

EFFICIENT CONTROL OF POLLUTION

The effi cient level of pollution can be attained either by charging fi rms a fi ne of f * per unit of pollution (say, measured by the number of particles added to the air) or by imposing a regulation that fi rms have a pollution abatement level P*.

FIGURE 6.4Marginal cost of pollution abatement

Marginal social cost of pollution

(marginal benefits of pollution abatement)

Level of pollution abatement

Price for pollution

(fine)

P*

f*

142 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

still an excessive level of production of steel, as illustrated by point Qs in Figure  6.6.5 Qm is the output before subsidy, which is markedly greater than Qe, the effi cient level of output with the subsidy. A well-designed subsidy lowers the total marginal social cost of production—there is less pollution. Although there is a cost of the abatement subsidy, including the distortions arising from the taxes required to raise the revenues to fi nance it, the benefi ts from the lower pollution exceed these costs; hence, the optimal level of output with the subsidy is greater than the optimal level of output when fi rms have no incentives to reduce pollution. Thus, Qo is greater than Qe. On the other hand, if the pollution abatement equip- ment confers some ancillary benefi ts to the fi rm, it may simultaneously reduce the fi rm’s marginal private costs of production, as indicated by the light gray line. Thus, the fi rm’s output level also increases from Qm to Qs. However, because the main benefi t of the pollution abatement equipment is to reduce pollution, presumably, the distortion—the magnitude of the excess production—is reduced.

The reason why polluters prefer subsidies over fi nes for pollution abatement is clear: profi ts are higher under the former system than under the latter. The distributional consequences are not limited to the pollut- ing fi rms and their shareholders. Because output will be smaller under the system of fi nes, prices will be higher, and consumers of the polluting fi rm’s products will be worse off . On the other hand, those who have to

5�If the level of pollution of a fi rm cannot be directly monitored, a desirable policy would entail a subsidy for expenditures on pollution abatement combined with a tax on output. The tax on output, if set at the appropriate rate, reduces the level of output to the socially effi cient level.

POLLUTION ABATEMENT SUBSIDIES

By subsidizing the purchase of pollution abatement equipment (by the difference between mar-

ginal social benefi t of pollution abatement and marginal private

benefi t), an effi cient level of expenditure on pollution abate-

ment can be attained.

FIGURE 6.5

Subsidy

Marginal private benefit of pollution

abatement

Marginal costs of pollution abatement

Marginal social benefit of pollution abatement

Quantitiy of pollution abatement

equipment per unit output

Price

Qe

143Public Sector Solutions to Externalities

pay the taxes to fi nance the subsidies for pollution abatement are clearly better off under the system of fi nes. It should be emphasized, however, that the choice between subsidies and fi nes is not just a distribution issue. As we have seen, under the pollution abatement subsidy scheme, producers do not face the true social cost of their production; there is an ineffi ciency. By contrast, with an appropriately designed system of fi nes, producers do face the true social costs.

MARKETABLE PERMITS An increasingly popular market-based solu- tion involves marketable permits, commonly referred to as tradable per- mits, which operate under a cap and trade system. A limit, or cap, is placed on the total amount of a pollutant that may be emitted, and this limit is either allocated or sold to fi rms in the form of emissions permits. These limit the amount of pollution that any single fi rm may emit. For instance, each fi rm may be allowed to emit 90 percent of the amount it emitted the previous year. Thus, a fi rm is granted a permit to emit so many units of pollutants. Because what the government cares about is the total amount of emission reduction, it allows fi rms to trade permits. A company that cuts its emissions in half could sell some of its permits to another com- pany that wants to expand production (and hence increase its emission of pollutants).

Under this system, fi rms will be willing to sell permits as long as the market price of the permit is greater than the marginal cost of reducing pollution, and fi rms will be willing to buy permits as long as the marginal cost of reducing pollution is greater than the market price of the permit.

MARKET EQUILIBRIUM WITH POLLUTION ABATEMENT SUBSIDIES

Even after the pollution abate- ment subsidy, the equilibrium level of output of steel is still ineffi cient; the fi rm fails to take into account the extra costs of public subsidies for pollution abatement associated with increased output of steel as well as the marginal social cost of any remaining pollution.

FIGURE 6.6 Marginal social costs

before subsidy Marginal social costs

after subsidy

Marginal private costs before subsidy

Marginal private costs after subsidy (for

pollution abatement)

Marginal benefits (demand curve)

Output after

subsidy

Output before

subsidy

Efficient level of output with

no subsidy

Efficient level of output

with subsidy

Output of steel

Price

Qe Qo Qm Qs

144 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

Thus, in equilibrium, each fi rm will reduce pollution to a level such that the marginal cost of pollution reduction is equal to the market price of the permit. Like fi nes, marketable permits use the market mechanism to ensure economic effi ciency in the reduction of pollution: the marginal cost of reducing pollution is the same for every fi rm.

In the absence of uncertainty, the two systems are essentially equiva- lent: setting a price (say, for emissions) leads to a particular quantity, and setting a quantity leads to the corresponding price. In the presence of uncertainty, both about the costs of pollution and of pollution abatement, the two may not be fully equivalent. In setting a price, the quantity of pollution that emerges will be variable; in setting a quantity, the price that polluters will have to pay is uncertain.

Consider the problem of global warming, with which we know that there is a high social cost of carbon emissions, but we are uncertain exactly what that cost is. At the same time, we are uncertain about how fi rms and consumers will respond to any price of emissions imposed. Those who feel confi dent that we know much more about what we need to do—by how much emissions should be reduced—argue for marketable permits and for forcing fi rms to bear the risk of the uncertain price for emissions that will emerge. Advocates of a carbon tax (a price system) argue that, in practice, adjustments will have to be made to either prices or quantities over time. What matters is the level of concentration of car- bon dioxide in the atmosphere, and we can adjust the fl ow (the level of emissions) as we see the level of concentration build up and as we observe the consequences, either by adjusting the level of marketable permits or by adjusting the price of carbon.

There are three problems with tradable permits. The fi rst is making the initial assignments. Even though the system of assigning fi rms a frac- tion of their current levels of pollution might at fi rst glance seem reason- able, it causes a major equity problem: marketable permits are an asset that can be traded, so giving pollution permits is equivalent to giving away money. Why should fi rms that have been polluters in the past be entitled to receive a bigger gift from the government? There is even a perversity in doing so: “good” fi rms that have spent large amounts on pollution control are given fewer permits and, if they have already installed state-of-the- art technology, will have a harder time reducing pollution. An alternative system bases permits simply on the level of production. This system is basically the one used when tradable permits were introduced to con- trol acid rain. When Los Angeles introduced tradable permits in 1994, its assignments took into account both the levels of output and pollution and the state of the fi rm’s current technology. There is a third alternative, par- ticularly attractive in an era of budget stringency: auction off the emission permits. Firms that have been polluting will be worse off than before, but

145Public Sector Solutions to Externalities

the question is: Why should they receive money for not harming others? (Remember the reason that pollution imposes costs on others.) As we have noted, with such auctions there is little diff erence between market- able permits and taxes/fi nes.

This leads to the second problem, which traditionally was viewed as an advantage of tradeable permits: fi rms could be “bribed” to support the legislation curbing emissions, because they have less (or little) to lose. Indeed, the fi rms that believe that they can reduce emissions consider- ably become supporters, because selling their emission permits can be a new source of revenue. In Europe, though, we have seen the downside: governments come under pressure to issue more emission permits (it is like giving away money, without a budget constraint). The result is less reduction in pollution.

The third problem is more subtle. Tradable permits work well only when the location of the pollutant makes no diff erence. In many situa- tions this is not the case, thus air pollution is much more serious near large cities. Moreover, with prevailing winds blowing from west to east, pollution along the East Coast may not be much of a problem, as most of it gets blown immediately out to sea; but pollution in the Midwest may have adverse eff ects on all the eastern states. The marketable permits that were introduced to control acid rain did not fully take this problem into account.

REGULATION

Economists have usually argued that market-based solutions provide the most promise for curbing environmental externalities, but government traditionally has relied on direct regulation. It has set emission standards for automobiles; put forth detailed regulations relating to the disposal of toxic chemicals; outlawed smoking on domestic airline fl ights; imposed laws requiring oil companies with wells in the same oil pool to unitize their production; imposed restrictions on fi shing and hunting to reduce the ineffi ciencies associated with excessive utilization of these common resources. These examples illustrate the myriad forms that regulation may take.

Advocates of regulations argue that they provide greater certainty: if fi rms are prohibited from emitting more than a given level of pollution into the water, then one knows the maximum level of pollution; with fi nes, the level of pollution depends on the costs of reducing the pollution level. However, advocates of fi nes argue that one can easily adjust fi nes to induce fi rms to lower pollution to the desired level. Moreover, marketable permits provide a market-based way to attain effi cient pollution reduction

146 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

and certainty of outcome. Indeed, a major criticism of regulations is that they do not reduce pollution in the most effi cient way: diff erent fi rms may face diff erent marginal costs of further pollution abatement. Furthermore, regulations typically provide little or no incentive for fi rms to reduce pol- lution below the standard that has been set, regardless of how low the cost of doing so.

In the case of pollution, we should distinguish between two import- ant classes of regulations. Recall that the market-based mechanisms dis- cussed earlier focus on the amount of pollution; to pollute more, a fi rm must pay more in fi nes or buy more permits. This is a performance-based system, since the government only cares about the fi nal outcome— how much pollution is produced. There are many performance-based regulations, such as regulations on automobile emissions, that also focus on the fi nal outcome. However, much of pollution regulation has focused on standards, practices, and inputs, rather than performance. For instance, the government may prohibit the use of certain grades of coal, or it may require fi rms to employ scrubbers and other pollution abatement devices, or to construct smokestacks to specifi c heights. These are called input regulations. Market-based mechanisms may also focus on inputs and practices in this way; for example, a tax may be levied on high-sulfur coal, rather than on the pollution emitted.

When feasible, it is preferable to focus on performance, either for reg- ulations or for market-based mechanisms. The one argument for focusing on inputs and practices is that they may be more easily monitored. Thus, it may be diffi cult to measure the amount of pollution coming out of a smoke- stack, but it is certain that if scrubbers (devices that reduce the amount of sulfur being emitted by a coal-burning electric power plant) are used, the amount being emitted will be less than if scrubbers are not used.

Although there may be good reasons for these policies, in some cases, politics rather than policy has dominated the decision. In the case of coal, had a performance-based standard been used, eastern coal produc- ers would have been disadvantaged relative to western coal producers, because eastern coal contains more sulfur. To attain the same level of sul- fur, fi rms using eastern coal would have had to use scrubbers, whereas those using western coal would not. Eastern coal producers successfully lobbied for the universal imposition of the requirement to use scrubbers.

INNOVATION

One of the reasons for performance-based regulations (as opposed to input standards) and pollution-based taxes (as opposed, say, to subsidies

147Public Sector Solutions to Externalities

for particular forms of abatement equipment) is that they directly address what is of concern—the level of pollution—and they may induce innova- tions, such as new ways of producing that generate less pollution or new techniques for abating pollution at lower costs. Advances in technology have improved the ability to monitor some kinds of pollution; ongoing innovation should enhance this further.

There has been considerable controversy over the best way to stimu- late innovation and about the scope for innovation. Some environmental- ists are less convinced than economists of the power of normal economic forces. Many believe that industry must be forced to innovate. Thus, by imposing extremely stringent standards—for instance, that cars get at least 40 miles per gallon—they will force industry to develop a product meeting these standards. Implicitly, they believe that the benefi ts of the innovation would outweigh the costs, but that the incentives that could be provided by the price system—charging car companies taxes in pro- portion to the amount of pollution—simply do not suffi ce to warrant their attention on this area.

In practice, the success of this strategy has been mixed. In some cases, rather than inducing innovation, stringent regulations have induced liti- gation: it may appear cheaper to a fi rm to try to persuade a court that the regulation is unreasonable than to spend the money to meet the standards imposed by the regulation. In some cases, fi rms have played a game of chicken, gambling that if they fail to meet the standards, the government will not shut them down, for fear of a political backlash from workers who are put out of a job. In some cases, however, industry unity has been broken by an innovative fi rm that showed that the standards are indeed attainable, or can even be surpassed. For instance, the great commer- cial success of the Toyota Prius has spurred its American competitors to develop hybrid cars of their own to compete in a rapidly growing market.

Environmentalists who doubt the eff ectiveness of market incen- tives by themselves in inducing innovation often point to the large gap between best practices—which often seem to be the most cost-eff ective practices—and what actually occurs. They point out, for instance, that there are energy-effi cient light bulbs that more than pay for their higher costs in terms of reduced usage costs. Sometimes there is a coordination failure that government action can help remedy: no one wants to install fi xtures that use energy-effi cient light bulbs if it is going to be diffi cult to replace the bulbs when they burn out, and stores will not carry such light bulbs if there is no demand; and there won’t be a demand if builders do not put them in houses they construct and lamp manufacturers do not put them in the lamps they make. Sometimes there are information barriers and other types of barriers to the adoption of cost-effi cient energy-saving

148 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

technologies, and government can help promote the diff usion of these technologies by disseminating information.

Behavioral economics has explained how sometimes there can be inertia in changing customs. Government can help change norms, though. There have been large changes, for instance, in norms concerning smok- ing or the use of plastic bags, induced in part by government regulations. Of course, prices too can help in changing practices. One of the reasons that energy-effi cient light bulbs are more prevalent in Japan than in the United States is that the price of electricity is higher there. Between these two camps are those who argue that simply providing information and small price signals will not lead to the large changes in behavior that are needed; changes are likely to be modest and slow. Today, labels on many electric products identify energy usage and costs, helping buyers make more intelligent decisions about lifetime costs of diff erent products.

Thus, there may be win–win regulations with which effi ciency is enhanced at the same time that environmental costs, especially those associated with the use of energy, are reduced; slight modifi cations in construction practices—the color of roof shingles or the planting of trees, for example—can have a noticeable eff ect on energy consumption.

Critics of approaches focusing on inputs rather than performance argue that such approaches are not only ineffi cient, but also stifl e inno- vation and push it in the wrong direction. For instance, rather than seeking the most eff ective way of reducing emissions from coal-burn- ing power plants, research is focused on making cheaper scrubbers. Moreover, research directed at improving the ability to monitor outputs accurately—thus reducing the necessity of relying on input regulations—is not encouraged.

INFORMATION DISCLOSURE

In some areas, governments have been experimenting with another approach, focusing on public pressure rather than the heavy hand of government. Government’s role would be limited to requiring fi rms to disclose, for instance, the potentially cancer-inducing chemicals that they discharge into the water or emit into the air. Government would not even comment on the extent of scientifi c evidence concerning the impact of the chemicals on humans. Critics of this approach often argue that the costs of such information disclosures can be high, but their real concern is that government would be encouraging a scare campaign. Most people would simply assume that if a chemical is listed as dangerous, it must be dangerous—or, in any case, why risk it? People in the neighborhood would

149Public Sector Solutions to Externalities

put enormous pressure on the fi rm to eliminate the chemical, without any assessment of the costs or benefi ts of doing so. There could be enormous adverse economic eff ects. Particularly troubling is the evidence of one study by the Environmental Protection Agency (EPA) that showed the magnitude of popular misconcep- tions about environmental risks. Scientists and nonscientists were asked to rank a number of dif- ferent potential environmental health hazards, and there was little correlation between the two. Among the risks rated most highly by the non- scientists were several that were ranked at the bottom by the scientists, and vice versa.

COMPENSATION AND DISTRIBUTION

So far most of our discussion has focused on the effi ciency of alternative ways of controlling pol- lution (externalities). But much of the debate is about distribution—who bears the costs. Diff er- ent systems of controlling pollution may have markedly diff erent distributive consequences. Subsidies for pollution abatement equipment may result in a less effi cient resource allocation than a system of fi nes for polluting; but fi rms will clearly prefer subsidies. Greater effi ciency means that, in principle, the overall gains to society from using the more effi cient system are such that the gainers could compensate the losers.

Why then do governments so often resort to ineffi cient systems like abatement subsidies? The reason is that the compensation is typically not paid. Partly this is because it is often diffi cult to measure the gains and losses to each individual— the information required to implement the desired compensations is simply not available—and partly it is because those who benefi t from the ineffi cient system are more politically organized. The losers from a system of fi nes are clear—both the owners of the polluting fi rms,

ASSESSING ALTERNATIVE APPROACHES TO CONTROLLING EXTERNALITIES Performance-based versus input-based • Performance-based is more effi cient when

performance can be measured.

• Costs of monitoring inputs may be lower.

Regulation versus fi nes • Fines related to costs of pollution provide

appropriate incentives.

• However, there may be greater uncertainty about the actual level of pollution.

• Regulations provide greater certainty, strong incentives to meet the regulatory standards, but no incentive to reduce pollution below the standard, regardless of how low the cost is to do so.

Marketable permits versus fi nes • Both can result in effi cient reductions in

pollution levels.

• Marketable permits provide greater certainty about the level of pollution.

• Marketable permit systems face diffi cult problems in allocating initial permit rights.

• There are further problems if the costs of pollution depend on where the pollution occurs.

Subsidies for pollution reductions versus fi nes for pollution • Both can induce reductions in pollution levels

and even achieve effi cient levels of pollution abatement. The level of production of the pollution-inducing industry will be too high, however, because the fi rm will not take into account full costs—including costs of pollution abatement. Firms prefer pollution abatement subsidies.

150 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

who see their profi ts decreased, and the consumers of their products, who see their prices increased. The gainers are more dispersed—all the tax- payers who bear the burden of the taxes used to pay the subsidies, and all the consumers of all the other products who might see their prices rise slightly as production shifts slightly toward the subsidized industry. As is so often the case, the losers are much easier to identify than the gainers; thus, it is much easier for the losers to get together and use the political process to argue for a system that, though ineffi cient, makes them bear less of the cost of reducing pollution.

One of the reasons that governments may resort to regulation rather than the market-based systems described earlier is that the distribu- tive consequences may be less; for instance, individuals and fi rms may respond only to a limited extent to a small price of emission permits. To induce the large changes in behavior that are required to avoid global warming (say, to avoid an increase in the world’s temperature by more than 2 degrees Celsius, a goal of the international environmental commu- nity) might require a very large increase in the price of emissions, which would translate into large increases in energy prices, with signifi cant dis- tributive consequences. It may be possible to “force” the adjustments the economy needs with less adverse distributional eff ects by imposing regu- lations on electricity producers and car manufacturers.

Recent discussions of environmental policy have also focused on the dis- tributional consequences of environmental hazards. For instance, the poor are more likely to live near toxic waste sites, and exposure to these haz- ards imposes large costs not only on parents (on their health and life expec- tancy), but also on their children. There is a greater likelihood that their children will be born with low birth weight, which, in turn, is associated with lifelong consequences, including lower average lifetime incomes.

PROTECTING THE ENVIRONMENT: THE ROLE OF GOVERNMENT IN PRACTICE

We now look more closely at the actual policies that the government undertakes to protect the environment. For convenience, we divide them into three categories: those directed primarily at air, water, and land. There are, of course, important interactions among these pieces of our environment, so several of the policies aff ect two or more of these categories.

151Protecting the Environment: The Role of Government in Practice

AIR

The air we breathe has been taken for granted since the beginning of time, but by the middle of the twentieth century this was no longer pos- sible in many major cities. London became famous for its pea-soup fog generated by pollution; Los Angeles for its life-threatening smog; Gary and Pittsburgh for their brilliant red overcast skies, a product of the steel mills on which those cities’ economy depended. People with weak respi- ratory systems knew the dangers of living in these cities, but the fact that all individuals faced greater health risks was recognized only slowly.

There are several aspects of the nation’s attempt to control air pol- lution. Two have been marked by considerable success; there is heated controversy over a third one; and in the fourth, progress remains slight.

The most marked success is associated with ozone depletion and chlorofl uorocarbons.

OZONE DEPLETION The Earth’s atmosphere has a thin layer of ozone, which shields us from harmful solar radiation. In the late 1980s it became clear that a hole was appearing in the ozone layer over Antarctica, and that the cause of the hole was chlorofl uorocarbons (CFCs) and other ozone-depleting substances (ODSs). The nations of the world responded in 1987 with a treaty signed in Montreal, Canada, called, appropriately, the Montreal Protocol. It initially required the production and consump- tion levels of CFCs to be cut in half by 1999, but was strengthened in 1990 with a commitment to phase out CFCs and halons entirely by 2000 (by 2010 for less developed countries). Since then, ODS atmospheric concen- tration has steadily declined, although complete recovery of the Antarctic ozone layer is not expected until 2050 at the earliest.

ACID RAIN�The control of sulfur dioxide (SO2), which gives rise to acid rain and is emitted especially by coal-burning power plants, is another success story. In the 1970s, we became aware that the leaves on the trees in many of our forests were turning yellow and many of our lakes seemed to be devoid of fi sh. The Acid Precipitation Act of 1980 and the Clean Air Act Amendments of 1990 began a program to control these emissions, and a national program of tradable permits was introduced. These are estimated to have signifi cantly reduced the overall cost of bringing down the level of pollution. Regional cap and trade markets for nitrogen oxide (NOx)—another pollutant—have been established as well.

An especially problematic aspect of acid rain is that SO2 and NOx emissions cross state lines. In 2011, the U.S. Environmental Protection Agency issued the Cross-State Air Pollution Rule (CSAPR) to strengthen

152 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

its court-challenged legal authority to regulate interstate acid rain emis- sions. The CSAPR focuses on pollution in the eastern United States, with the goal of reducing power plant SO2 emissions by 73 percent and NOx emissions by 54 percent from 2005 levels by 2014 in aff ected regions.6

PARTICULATE AIR POLLUTION Small particulates in the air can pose serious health problems for certain individuals. The federal govern- ment’s involvement in addressing the challenges of particulate air pol- lution has evolved over the past half-century from purely research (Air Pollution Control Act of 1955), to setting of standards (Clean Air Act of 1970), to enhanced authority for pollution control (1990 Clean Air Act Amendments). Although the EPA estimates that the direct benefi ts from the 1990 Clean Air Act Amendments will reach an annual economic value of almost $2 trillion by 2020, much less that the estimated $65 billion annual cost of public and private eff orts to comply with the requirements of this law, there is still considerable debate about the assumptions under- lying this cost–benefi t analysis.7 The controversy concerns the cold calcu- lations of the costs of reducing the level of particulates versus the health benefi ts. As hard as it is to quantify these benefi ts, there are standard pro- cedures by which this is done; such calculations are made routinely in evaluating how much to spend to make a safer highway, a safer car, or a safer airplane. These changes can result in a slightly smaller probability of an accident from which one can calculate, on average, how many lives will be saved. The government must have a systematic way of deciding whether the benefi ts exceed the costs, by doing so, it places a value on life—in this case, the value of reducing the risk of fi ne particle-related pre- mature mortality.

GLOBAL WARMING The one area in which little progress has been made is global warming and greenhouse gases. The Swedish chemist Svante Arrhenius explained as early as 1896 that carbon dioxide emis- sions from the burning of coal would enhance Earth’s natural greenhouse eff ect (retention of solar radiation by Earth’s atmosphere) and thus lead to global warming, but it was not until the 1980s that the world’s attention turned to the issue. The scientifi c community was able to show that the current level of carbon dioxide concentration in the atmosphere due to the burning of carbon—from coal, gas, and oil—is, indeed, substantially greater than it was at the beginning of the Industrial Revolution, and

6�U.S. Environmental Protection Agency, Cross-State Air Pollution Rule (CSAPR), accessed July 20, 2011, http://www.epa.gov/crossstaterule. 7�For the complete cost–benefi t analysis, see U.S. Environmental Protection Agency, Offi ce of Air and Radiation, The Benefi ts and Costs of the Clear Air Act from 1990 to 2020 (Washington, DC: U.S. EPA, March 2011).

153Protecting the Environment: The Role of Government in Practice

continues to grow. Scientists have also established that there was over- whelming evidence that these substantial increases were leading to signifi cant increases in Earth’s temperature. Since 1990, the Intergovern- mental Panel on Climate Change (IPCC) has issued periodic assessments on the state of global warming, and, together with former vice president Al Gore, was awarded the Nobel Peace Prize for contributions to interna- tional environmentalism.

To many, the world seems embarked on a risky experiment with our planet as we continue to add carbon dioxide and other greenhouse gases to the atmosphere; it is leading not just to a warming of Earth, but also a rise in the sea level. There will be severe adverse eff ects from this warm- ing, especially on the tropics, and the increase in sea level will obviously have adverse eff ects on low-lying islands and countries, such as Bangla- desh, Vietnam, and the Netherlands. Other predicted eff ects include an increase in the variability of weather. Even though there is clear evi- dence for the increase in greenhouse gases, and there is a general (but not universal) consensus on the long-run eff ects, there is more contro- versy over whether the eff ects are already being felt. There is evidence, for instance, of a marked increase in losses from weather—far greater than the increased losses from nonweather events such as earthquakes. Although there is some disagreement among economists about the mag- nitude of the overall costs—with some countries in cold climates actually benefi ting—the consensus, refl ected in an agreement made in Copenha- gen, is that the costs will be large.

As the scientifi c evidence has mounted, the issue has taken on greater urgency over the past two decades. In 1997 collective global concern resulted in adoption of the Kyoto Protocol, under which industrialized countries pledged to reduce their emissions of greenhouse gases. Most countries have ratifi ed the Kyoto Protocol, with the notable exception of the United States. A key U.S. objection is the protocol’s exclusion of devel- oping countries such as China and India, who will be the largest future sources of CO2 emissions.

In response, developing countries have argued that most of the prob- lem has been caused by the profl igacy of the advanced countries, their own contributions to the greenhouse gases thus far have been relatively small, their projected future contributions to greenhouse gases will still be much lower than high-income countries on a per capita basis, and they are too poor to devote much of their resources to reducing their emissions below the level that economics dictates.

To reduce overall costs, the countries agreed at Kyoto to explore more market-based mechanisms—tradable permits and a variant called “the clean development mechanism” or joint implementation—because

154 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

one country would “buy” the greenhouse gas reduction from another; in eff ect, by paying for it, they could be thought of as “jointly” imple- menting the greenhouse gas reduction. Joint implementation can be thought of as a limited form of marketable permits. Some critics sug- gested that the United States was advocating joint implementation not out of a commitment to economic effi ciency, but because it could not or would not take measures that would reduce greenhouse gases within its own borders.

Today, there are many active cap and trade systems for greenhouse gases. The European Union Emission Trading Scheme (EU ETS) was established by the EU to help meet its Kyoto Protocol commitments by the time the protocol expired at the end of 2012, and is now the world’s largest CO2 emissions trading system.

Negotiations for a new global agreement on greenhouse gases are continuing, with very slow progress. In Copenhagen, countries agreed to draw up their own action plans to reduce emissions. There was peer pressure to make signifi cant reductions, but no commitments were made. Subsequently, parties to the United Nation’s Framework Convention on Climate Change concluded their December 2012 climate conference in Doha, Qatar, with agreement to continue negotiations on a successor to the Kyoto Protocol, which they hope to approve during their 2015 confer- ence in France.

The easiest and most cost-eff ective way of reducing emissions in the United States would be a carbon or energy tax, but this option has been adamantly opposed by the powerful gas, coal, and oil interests. Currently, gasoline taxes in the United States are markedly lower than in Europe and most other industrialized countries, as shown in Table 6.1.

TABLE 6 .1 GASOLINE TA XES AROUND THE WORLD

COUNTRY TAX PER LITER IN U.S. DOLLARS

United Kingdom 1.32

Germany 1.26

France 1.19

Italy 1.14

Spain 0.90

Japan 0.77

Canada 0.38

United States 0.11

SOURCE: International Energy Agency, End-Use Petroleum Product Prices and Average Crude Oil Import Costs, Table 3 (Average end-use prices) and Table 5 (Average end-use taxes), March 2011.

155Protecting the Environment: The Role of Government in Practice

A central problem in any international agreement is enforcement. Global warming is a global public good, and there is an incentive for each country to be a free rider; that is, each would like to continue to pollute, but also to enjoy the benefi ts of a world that is not subject to global warming as a result of the eff orts of others. In the Montreal Protocol, discussed ear- lier, countries that did not comply were threatened with trade sanctions. Some have suggested that similar sanctions might be desirable, or even necessary, to deal with global warming. Indeed, many European coun- tries that were making large eff orts to reduce their emissions thought that the failure of the United States to do anything was giving U.S. fi rms an unfair competitive advantage. European steel producers, for instance, had to pay for their carbon emissions, while U.S. steel producers did not. It was as if the U.S. producers were being subsidized, because they did not bear the full costs of their production. Some have suggested, accordingly, that it is only appropriate that those who do impose a carbon price (either through a carbon tax or tradable permits) impose a tax on imports from the United States or any other country refusing to do similarly.

WATER

The debate over clean air today centers not around whether pollution should be controlled, but rather how and at what levels. Likewise, in the case of water, there is consensus that the controls that have been put in place for drinking water make sense, but controversy remains over the ben- efi ts relative to costs of stringent regulations attempting to reduce pollu- tion in streams and rivers. Much of today’s water pollution comes not from factories, which can be more easily controlled, but from diffi cult-to-control sources such as runoff from farms. Controlling such pollution would require controlling the use of fertilizers and pesticides. Although price mechanisms (taxes) might discourage the use of fertilizers and pesticides, it would be virtually impossible to diff erentiate between usages that con- tribute to pollution and those that do not. Moreover, there is controversy over some of the benefi ts. How worried should we be about the pollution of groundwater that will almost surely never be used for drinking and that is unlikely to seep into wells or springs? Some environmentalists believe that we should never spoil a part of our natural heritage; keeping groundwater clean has nothing to do with the use to which groundwater might be put. Others take a very risk-averse stance: How can we be sure that the ground- water will never be used for drinking? Controversy over these issues has prevented reauthorization of the Clean Water Act, although many limited, specifi cally targeted clean water bills have been enacted in recent years.

156 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

LAND

TOXIC WASTES Newspapers have presented graphic stories of rivers, canals, and land that chemical companies have turned into toxic waste sites, subjecting those who come into contact with them on a regular basis to increased risk of cancer. Americans had nightmares about discovering that their homes had been built over toxic waste sites. In response, in 1980 Congress passed the Comprehensive Environmental Response, Compen- sation, and Liability Act (CERCLA), commonly known as Superfund, after the trust fund it established with a tax on chemical and petroleum companies, to help clean up these sites.

The law was badly designed, however. It was based on the principle that those who contributed to the pollution should pay for its cleanup, but it provided that anyone who had contributed at all to the particular site was liable for the entire cleanup costs. This enabled the government to go after large corporations—those with “deep pockets”—forcing them to pay for the cleanup and letting them sue the other polluters to recover their shares. (This is called the system of joint and several liability.) Furthermore, it provided for perpetual liability: even after the site was seemingly cleaned up, it was always possible that on the site another chemical with adverse eff ects would be discovered, so the guilty party is never completely free from liability. To make matters worse, the polluters’ insurance companies argued that their general liability policies did not cover pollution other than that which originated as a result of an accident. In several states, the courts supported the insurance companies, whereas in other states, the courts said that the insurance companies were liable. The upshot was litigation—between the government and the polluters, among the diff erent polluters, and between the polluters and their insur- ance companies. The lawyers made out like bandits, but the toxic wastes did not get cleaned up. Over 70 percent of insurance company expendi- tures went to legal fees, and more than a quarter of what the polluters spent went to lawyers. Worse still, property owners had a new nightmare to worry about: that toxic waste would be discovered on their property that might not hurt their health but would defi nitely hurt their pocket- books. They would be responsible for cleaning it up, and the EPA often set standards for which the costs simply could not be justifi ed by the ben- efi ts. In some places, the result was “gridlock” in the land market: no one would buy potentially polluted sites, which hampered eff orts to redevelop inner-city areas. Banks would not make loans, lest they wind up holding the property (and being responsible for the cleanup) in case of a default. America’s landscape was scarred with such “brownfi elds” (as they came to be called), and America’s fi rms had to get out to fi nd green fi elds to

157Protecting the Environment: The Role of Government in Practice

build their new factories. A law intended to preserve and protect the land had led to opposite results.

Thirteen years after the bill was passed, only one of seven major sites that had been identifi ed had been cleaned up. Four years later, the administration was claiming that two-thirds of the sites had been, or were on the way to being, cleaned up. The situation was similar in 2010, when more than half the work had not yet been completed for the reme- dial construction phase of cleanup for over 60 percent of the 239 nonfed- eral National Priorities List (NPL) sites with unacceptable or unknown human exposure.

Despite this modest progress over the thirty years since enactment of the Superfund legislation, the other problems, such as the brownfi elds and the inequities associated with joint and several liability, persisted. Even though there was universal agreement that reforms were needed, the dif- fering perspectives of environmentalists, insurance companies, and pol- luters made resolution of these problems diffi cult. There were confl icts about both the standards of cleanup and about who should pay. Although forcing those who pollute to pay provides strong incentives not to pollute, there was controversy over whether it was right to force people to pay for actions that were not illegal at the time they occurred, and whose con- sequences might not even have been apparent. Indeed, the worst eff ects of the powerful chemicals that dry cleaners had used was probably not on the land onto which some of these chemicals spilled but on the own- ers and workers who spent their lives working in these cleaning plants, unaware of the health eff ects. They had already paid a high price—and given the competitive nature of the industry, the benefi ts of their using the chemicals were received by their customers: had they been required to dispose of the chemicals in another way, they would have done so and passed the costs on to their customers. And in any event, in many cases, it was not the polluters that would actually pay but their insurance com- panies, thus undermining the moral argument that polluters should pay. However, environmentalists point out that fi rms are in the best position to judge the risks posed by the chemicals they use, and that insurance fi rms are in a good position to put pressure on those they insure to look carefully at the risks of their products. The debate continues, although the EPA claims, “Over the past 201 years, we’ve located and analyzed tens of thousands of hazardous waste sites, protected people and the envi- ronment from contamination at the worst sites, and involved others in cleanup.”8

8�U.S. Environmental Protection Agency, Superfund: Basic Information, accessed July 20, 2011, http:// www.epa.gov/superfund/about.htm.

158 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

ENDANGERED SPECIES�There has been concern not only to protect the environment from pollution, but also to preserve it. As populations expand, they crowd out nature. Throughout the world, a multitude of species are threatened with extinction. At the global level, international treaties have been signed to combat these threats; 1992 was marked not only by a treaty on global warming, but also by a treaty intended to preserve the world’s biodiversity— plants as well as animals. Particularly powerful in this debate was the recog- nition that within this diverse biological heritage, there might live cures for a myriad of diseases. Other international treaties are directed at preventing the extinction of whales and eliminating trade in ivory (which might encour- age the extinction of elephants) and rhinoceros horn (highly valued in certain parts of the world for its alleged powers in enhancing sexual potency).

In 1973, the United States passed the Endangered Species Act. The legislation has been highly controversial because of its potentially strong economic impact. For instance, logging in large parts of Washington and Oregon was halted because of a concern over the destruction of the habi- tat of the spotted owl, an endangered species; and in Texas, development of areas near Austin was halted over fear of destroying the habitat of some endangered species of spiders.

Critics argue that the preservation of these species is a public good, but a public good that owners of these particular parcels of land are made to pay for. If the public wants these species to be preserved, it should buy the land. Prohibiting owners from developing the land is almost tantamount to seizing it. Indeed, many argue that any restrictions on usage represent a “taking” of property; just as the government cannot simply take away your property without compensation, it should not be allowed to take away the uses to which you can put your property without compensation. There is a fundamental diff erence between laws that stop a person from imposing an externality on others, and laws that require a person to provide a public good (the protection of an endangered species) to others.

Supporters of endangered species legislation, even when they recog- nize these arguments, say that there simply is not enough money available to provide compensation to property owners; the choice is a pragmatic one—allow the species to become extinct, or impose these mandates on property owners. Besides, the longer the law is on the books, the less these arguments on “takings” become relevant: those who buy property know that their use may be encumbered by the Endangered Species Act, and this is refl ected in the purchase price they pay. The cost was eff ectively borne by the owners of the land at the time the law was enacted; if anyone should be compensated, it is the former owners, not the current ones.

Protecting endangered species is only one of several pieces of leg- islation designed to protect our natural environment. There is also, for

159Review and Practice

instance, important legislation protecting the wetlands and coastal land. President Theodore Roosevelt protected 230 million acres of some of the country’s most treasured areas while serving from 1901 to 1909 by estab- lishing 5 national parks, 150 national forests, 4 national game preserves, and 51 federal bird reservations.

CONCLUDING REMARKS

Although the specifi cs of environmental legislation—how best to improve the environment and how high to set the standards—are likely to remain contentious, there is a growing awareness that the government has at its disposal a wide range of instruments and a growing consensus on a set of general principles: the environment is of critical importance; markets alone will not provide effi cient outcomes because of important externali- ties; some form of government action is required; when possible, interven- tions should be performance based and market oriented; the government must be sensitive to the distributional consequences both of environmen- tal degradation and the policies that are implemented to ensure the pro- tections of the environment; and the environment is so important that we cannot make the perfect the enemy of the good.

SUMMARY

1. Externalities are actions of an individual or fi rm that have an eff ect on another individual or fi rm for which the latter does not pay or is not paid.

2. Sometimes economic effi ciency can be attained without resorting to government intervention

a. By establishing suffi ciently large economic orga- nizations, the externalities can be internalized.

b. By establishing clear property rights, private parties can bargain toward an effi cient solu- tion, as suggested by Coase.

c. By using the legal system, imposers of exter- nalities can be forced to compensate victims.

 3. There are important limitations to each of these private remedies. For instance, public goods problems and transactions costs impede effi cient bargaining solutions in the manner suggested by Coase. These failures necessitate a greater role for government in remedying the problems of externalities.

 4. There are four methods by which the government has attempted to induce individuals and fi rms to act in a socially effi cient manner: fi nes and taxes, subsidies, tradable permits, and regulation.

 5. When there is good information about the mar- ginal social cost of the externality (as with pollu- tion), and the fi nes can be adjusted to refl ect those

REVIEW AND PRACTICE

160 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

costs, then a fi ne system can attain a Pareto effi - cient outcome. Subsidies to pollution abatement, while enabling the effi cient level of pollution abate- ment to be attained, will result in excessive pro- duction of the pollution-generating commodity. In principle, the gainers under the fi ne system could more than compensate the losers, but in practice these compensations are seldom made. Thus, the choice of the system for controlling externalities has important distributional consequences.

 6. Tradable permits (cap and trade systems) can also result in effi cient pollution abatement.

 7. Regulations focusing on inputs or standards are likely to result in ineffi ciency.

 8. The Clean Air Act has greatly reduced the level of pollution in the air. There is increasing concern about greenhouse gas emissions, which may lead to global warming.

 9. The Clean Water Act has greatly reduced water pollution. Controversy remains over whether standards are excessively stringent, so that at the margin, costs exceed benefi ts.

10. The Superfund program, which is intended to clean up toxic waste sites, faces several problems, including excessive litigation costs and slow clean- ups. Remedies include reforms in the legal system and the cleanup standards, and must address the problem of who should bear the costs of cleanup.

11. There has been increasing interest in preserving biodiversity and protecting endangered species. There is concern, however, that restrictions on land usage required to protect endangered spe- cies constitute an unfair “taking” of property.

12. There are several global environmental problems, the most important of which is global warming. There is broad consensus among the scientifi c community that increases in the concentration of greenhouse gases in the atmosphere will lead to higher temperatures, an increase in sea level, and increased weather variability, with signifi - cant economic consequences. The international community has so far not been able to reach an agreement to curb emissions of carbon dioxide and other greenhouse gases.

KEY CONCEPTS

Cap and trade

Coase theorem

Common resource problems

Corrective taxes (Pigouvian taxes)

Existence values

Externalities

Input regulations

Internalizing externalities

Joint implementation

Marginal social cost curve

Marketable permits

Performance-based regulations

Property rights

Superfund

System of joint and several liability

Unitize

QUESTIONS AND PROBLEMS

1. Make a list of the positive and negative externali- ties that you generate or that aff ect you. For each, discuss the advantages and disadvantages of each of the remedies discussed in the text.

 2. An important class of externalities to which atten- tion has recently been directed is called informa- tion externalities. The information produced by one individual or fi rm generates benefi ts for others. The success of an oil well on one tract of land increases the likelihood of oil’s being found on an adjacent tract, and hence increases the value of that tract. Can you think of other examples of information externalities? What are the likely consequences of information externalities for the effi ciency of resource allocations? Discuss the possibilities of private market solutions to these problems.

 3. Explain why subsidies for pollution abatement equipment, even if they result in an effi cient level of pollution abatement, will not result in an effi - cient resource allocation.

 4. Assume that there is uncertainty about the value of pollution control as a result of, for instance,

161Review and Practice

uncertainty about the costs of pollution. Draw two diff erent “demand curves” (or benefi t curves) for pollution, showing the marginal benefi t of reducing pollution by one more unit decreas- ing as the level of pollution reduction increases. Assume that the marginal cost of pollution abate- ment increases as the level of pollution abatement increases.

a. Assume the government can regulate the amount of pollution control after it knows what the benefi ts are. Show what the level of pollution control will be in each situation.

b. Assume the government can impose a tax on pollution after it knows what the bene- fi ts are. Show what the level of tax (or fi ne) will be in each situation. Is there any diff er- ence between regulations and fi nes in these circumstances?

c. Now assume that the government must set the level of allowable pollution before it knows what the benefi ts are. How will it set the level of allowable pollution? (Hint: there is a cost to society due to allocative ineffi ciency: from allowing too much pollution, if it turns out the benefi ts of pollution reduction are high; from being too restrictive, if it turns out the bene- fi ts are low.) How do you minimize the sum of these two social welfare losses?

d. Now assume the government must set the level of fi ne before it knows what the benefi ts are. How will it set the fi ne? Is there a diff er- ence between fi nes and regulations here?

e. Now assume that marginal benefits of pol- lution abatement are unknown. Assume also that when marginal benefits are high, the marginal costs are also high, and sim- ilarly, costs are low when benefits are low. Contrast a system of fines and regulations under these circumstances, where the level of fine or regulation must be set before costs are known.

 5. Assume there are two types of communities in the United States, those in which there is a high benefi t of pollution control and a high cost of

pollution control, and those in which there is a low benefi t of pollution control and a low cost of pollution control. Assume that the government must set either uniform regulations (a uniform level of pollution control) or a uniform fi ne for pollution. Show diagrammatically that a regula- tory scheme may be preferable to a system of fi nes. How does your answer change if communities in which there is a high marginal cost of pollution control happen to be communities in which there is a low marginal benefi t; and communities with a low marginal cost of pollution control happen to be communities in which there is a high marginal benefi t?

 6. The impact of some externalities is very local, such as noise from airplanes landing and tak- ing off at an airport. Such externalities depress the value of the immediately surrounding real estate. We say that the cost of the externality is capitalized in the value of the property. Assume that poor individuals are more willing to accept the high level of noise pollution, in return for the much lower rents they have to pay for housing. Describe the incidence of a regulation lowering the noise level surrounding the airport; that is, who benefi ts? (Hint: What will happen to land values? To rents?)

 7. Zoning laws, which restrict how individuals can use their land, are sometimes justifi ed as a means of controlling externalities. Explain. Discuss alternative solutions to these externalities.

 8. What is the externality associated with an addi- tional individual’s driving on a congested road? How do tolls help alleviate this externality? How should the toll be set?

 9. Explain why a system of joint implementation for reducing greenhouse gases is more effi cient than a system whereby each country must reduce its pollution by a fi xed amount.

10. Many economists are worried that unless all countries are required to reduce their levels of greenhouse gas emissions, reductions in emis- sions in one country may be partially off set by increases in another. Explain how this might occur.

162 CHAPTER 6 EXTERNALITIES AND THE ENVIRONMENT

11. Global warming is related to the concentration of greenhouse gases in the atmosphere. Once in the atmosphere, gases remain there for long peri- ods of time (centuries). Greenhouse gases include carbon dioxide and methane.

a. Assume the eff ect on global warming of a given amount of carbon dioxide is four times that of methane. What should be the relative fi ne (tax) on emissions of the two gases?

b. How should the tax vary over time?

c. A carbon tax is a tax related to the amount of carbon dioxide that burning gas, oil, or coal adds to the atmosphere. With a carbon tax, coal is taxed very heavily (relative to the amount of energy put out) and natural gas relatively lightly. A BTU (British thermal unit) tax is a tax related to the amount of energy produced, for instance, by burning gas, oil, or coal. If one is concerned about greenhouse gas warming, why is a carbon tax preferable to a BTU tax?

12. Two diff erent strategies are debated for reduc- ing greenhouse gas emissions. One is that all the countries in the world should adopt common measures, such as a carbon tax. The other is that all countries in the world should adopt common goals, such as reducing the level of emissions to the levels of 1990. Explain the distribution and effi ciency aspects of these two strategies.

13. Corporate Average Fuel Economy (CAFE) standards stipulate the average fuel effi ciency (miles per gallon) of cars produced by each manufacturer. That is, the average fuel effi ciency

of cars sold by GM, Ford, Chrysler, Toyota, and so forth must be at least equal to the standard set by the government. Explain why such a sys- tem introduces both ineffi ciencies and inequi- ties among diff erent automobile manufacturers. (Hint: Consider a company that specializes in small cars versus one that specializes in large cars.)

14. Discuss some of the problems with CAFE stan- dards. How might a system of “tradable CAFE standards” be designed, and why might such a system improve effi ciency?

15. To resolve the controversy over insurance cover- age of Superfund sites, it has been proposed that the insurance industry be taxed to create a fund out of which claims would be paid. What diff er- ence does it make if the tax is levied on the basis of:

a. Insurance premiums as of 1980?

b. Current insurance premiums?

Which insurance companies might be expected to prefer each way of levying the tax? (You can use a supply and demand diagram to illustrate the answers.) How does each form of tax aff ect the supply curve of insurance?

16. One proposal to reduce automobile emissions involves “pay at the pump insurance,” under which individuals would pay, say, 25 cents per gallon of gasoline, with the proceeds going toward an insurance fund. What might be the environmental eff ects of such a proposal? Can you think of other grounds on which such a pro- posal might be attractive?

163

EFFICIENCY AND EQUITY

Chapter 3 took up Pareto effi ciency, the condition in which no one can be made better off without making someone else worse off . It showed that in the absence of market failures, a free market would be Pareto effi cient. Even if the competitive economy is effi cient, however, the distribution of income to which it gives rise may be viewed as undesirable. One of the main consequences, and main objectives, of government activity is to alter the distribution of income.

The evaluation of a public program often entails balancing its consequences for economic effi ciency and for the distribution of income. A central objective of welfare economics is to provide a framework within which these evaluations can be performed systematically. This chapter shows how economists conceptualize the trade-off s between effi ciency and equity.

7 1. How do economists think systematically about how to make social choices when there are trade-off s; that is, when after fi nd- ing all possible Pareto improvements, gains to the welfare of one indi- vidual must come at the expense of the welfare of others? What is the social welfare function, and why do economists fi nd this concept useful?

2. How do economists think systematically about the trade-off s between effi - ciency and inequality? How do they measure poverty or inequality? How do they measure effi ciency?

3. As a practical matter, how do governments translate these general principles into a form that can actually be used in decision making?

4. Can we still make improvements that will increase both effi ciency and equity? Are there market distortions whose benefi ts accrue primarily to the rich and whose costs fall predominantly on the poor, whose miti- gation will improve both productivity and fairness?

FOCUS QUESTIONS

164 CHAPTER 7 EFFICIENCY AND EQUITY

EFFICIENCY AND DISTRIBUTION TRADE-OFFS

Consider again a simple economy with two individuals, Robinson Crusoe and Friday. Assume initially that Crusoe has ten oranges, while Friday has only two. This seems inequitable. Assume, therefore, that we play the role of government and attempt to transfer four oranges from Crusoe to Friday, but in the process, one orange gets lost; hence, Crusoe ends up with six oranges and Friday with fi ve. We have eliminated most of the inequity, but in the process, the total number of oranges available has been diminished. There is a trade-off between effi ciency—the total num- ber of oranges available—and equity.

The trade-off between equity and effi ciency is at the heart of many discussions of public policy. Two questions are debated. First, there is dis- agreement about the nature of the trade-off . To reduce inequality, how much effi ciency do we have to give up? Will one orange or two be lost in the process of transferring oranges from Crusoe to Friday?

Second, there is disagreement on the relative value to be assigned to a decrease in inequality compared to a decrease in effi ciency. Some people claim that inequality is the central problem of society, and society should simply minimize the extent of inequality, regardless of the consequences to effi ciency. Others claim that effi ciency is the central issue. They argue that even if one wishes to help the poor, in the long run, the best way to do that is not to worry about how the pie is to be divided but to increase the size of the pie—to make it grow as rapidly as possible—so that there are more goods for everyone.

These disagreements relate to social choices between equity and effi - ciency. We now take a closer look at these choices.

ANALYZING SOCIAL CHOICES

When economists analyze consumer choice, the opportunity set is defi ned by the consumer’s budget constraint, and the consumer’s preferences are described by indiff erence curves (see Chapter 3). The individual chooses the point on the budget constraint that is tangent to an indiff erence curve, which puts him or her on the highest indiff erence curve feasible, given the budget constraint.

Economists have tried to use the same framework for analyzing social choices. The utility possibilities curve, introduced in Chapter 3, describes the opportunity set. It gives the highest level of utility (or welfare)

165Analyzing Social Choices

attainable by one individual, given the levels of utility attained by others. An economy is Pareto effi cient if and only if it is operating along the utility possibilities schedule. The fi rst fundamental theorem of welfare econom- ics says that competitive economies are always on the utility possibili- ties schedule. The second fundamental theorem of welfare economics says that every point on the utility possibilities schedule can be attained through a competitive market process if the government redistributes ini- tial endowments accordingly.

How does society select a point along the utility possibilities curve? Just as indiff erence curves for individuals describe how they make trade- off s between diff erent goods, social indiff erence curves describe how society might make trade-off s between utility levels of diff erent individ- uals. A social indiff erence curve gives the combinations of utility of, say, Crusoe and Friday, between which society is indiff erent.

The two central questions of welfare economics can now be restated in terms of this social choice framework. Assume the current competitive market equilibrium is represented by the point A on the utility possibili- ties schedule depicted in Figure 7.1. Suppose society decides to move, say, from point A to point B along the utility possibilities schedule, represent- ing an increase in Friday’s utility and a reduction in Crusoe’s utility. The fi rst question is: What is the trade-off ? The utility possibilities curve gives

SOCIAL INDIFFERENCE CURVES

The social indifference curves describe how society evaluates trade-offs between Friday and Crusoe; it gives the combinations of utilities between which society is indifferent. Society is better off on a higher social indifference curve, just as an individual is better off on a higher individual indifference curve. Just as the individual chooses the point on the budget constraint at which the indifference curve is tangent to the budget constraint, society’s preferred point on the utility possibilities curve is the point at which the social indifference curve is tangent to the utility possibilities curve.

FIGURE 7.1

Utility of Crusoe

Utility of Friday

UC1

A

B

UF0

UF1

S1

S0

UC0

166 CHAPTER 7 EFFICIENCY AND EQUITY

the answer by showing the increase in Friday’s utility from UF0 to U�F1 and the decrease in Crusoe’s utility from U�C0 to U�C1. The second question con- cerns social preferences: How does society evaluate the trade-off ? The slope of the social indiff erence curves gives the trade-off s for which soci- ety is indiff erent. Point B is on the social indiff erence curve S1, which is tangent to the utility possibilities curve, and lies on a higher indiff erence curve than S0. Point B is therefore preferred by society.

The next two sections take a closer look at each of these questions regard- ing trade-off s and the economist’s framework for analyzing social choice.

DETERMINING THE TRADE-OFFS

As we saw in Figure 7.1, the utility possibilities schedule shows us the trade- off s of transferring utility from Crusoe to Friday. The shape of the utility possibilities schedule tells us something more about those trade-off s. Con- sider the utility possibilities schedule shown in Figure 7.2. Assume that the economy lies at point A, at which Crusoe enjoys much more utility than Friday. Moving up and to the left along the schedule increases Friday’s util- ity and decreases Crusoe’s. Suppose we transfer oranges from Crusoe to Friday by moving in two steps, from point A to B to C. Clearly, this makes Crusoe worse off . As depicted in the fi gure, the decreases in Crusoe’s utility are small in comparison to the increases in Friday’s utility.

CRUSOE’S AND FRIDAY’S UTILITY POSSIBILITIES

CURVE

As oranges are transferred from Crusoe to Friday, Crusoe’s utility is decreased and Friday’s increased.

In moving from point A to B, the gain in Friday’s utility appears much greater than the loss in

Crusoe’s utility. That is because Friday is so much worse off than Crusoe. In moving from B to C, the gain in Friday’s utility is still larger than the loss in Crusoe’s

utility, but the trade-off has changed so that Friday’s gain is

smaller than the gain from A to B.

FIGURE 7.2

Utility of Crusoe

Utility of Friday

UCC U C B U

C A

A

B

CUFC

UFB

UFA

167Analyzing Social Choices

Utility theory helps explain this outcome. Economists use the term utility function to describe the relationship between the number of oranges and Friday’s level of utility; the extra utility Friday gets from an extra orange is called his marginal utility. These are shown in Figure 7.3. At each point, marginal utility is the slope of the utility function—the change in utility from a unit change in orange consumption. Notice that as more oranges are consumed, utility rises more slowly, and marginal utility falls. (Thus, the slope of the utility function at point C is less than

THE UTILITY FUNCTION AND MARGINAL UTILITY

(A) Shows the utility function: as we give Friday more oranges, his utility increases, but each additional orange gives him less extra utility. (B) Shows marginal utility: the extra utility Friday gets from an extra orange decreases as the number of oranges increases, corresponding to the decreasing slope of the utility function.

FIGURE 7.3

Oranges

Friday’s utility

A

20 21 22 23 24

24

C

B

MU21, marginal utility of the

increase from 21 to 22 oranges

Utility function

A

B

MU20, marginal utility of the

increase from 20 to 21 oranges

Oranges

Friday’s marginal

utility

20 21 22 23

Marginal utility

MU20

MU21

168 CHAPTER 7 EFFICIENCY AND EQUITY

the slope at A or B.) This is because Friday enjoys the fi rst orange very much, the next one a little less, and additional oranges still less. Finally, he becomes satiated and derives very little additional enjoyment from an additional orange. As an individual consumes more of any good, the extra gain from having one extra unit of that good becomes smaller. This phe- nomenon is referred to as diminishing marginal utility.1

By the same token, as we take away oranges from Crusoe, his utility decreases; and as we take away more and more oranges, the extra utility he loses from each additional loss of an orange increases. That is why with diminishing marginal utility, the utility possibilities schedule has the shape depicted in Figures 7.1 and 7.2. This shape says that when Friday has very little income (few oranges), we can increase his utility a great deal with a small decrease in Crusoe’s utility, but when Friday is much better off , we can increase his utility only a little with even a large decrease in Crusoe’s utility.

A second important determinant of the shape of the utility possibilities schedule is the effi ciency with which we can transfer resources from one indi- vidual to another. In our society, the way we transfer resources from one group (say, the rich) to another (say, the poor) is by taxing the rich and subsidizing the poor. The way we do that normally interferes with economic effi ciency.

1��We write the utility function as U 5 U(C1, C2, . . . , Cn�), where C1, C2, . . . , Cn represent the quantities of consumption of the various goods. Marginal utility of, say, C1, is then simply the increase in U (utility) from an increase in consumption of C1. Diminishing marginal utility implies that successive increments in C1 yield successively smaller increments to U.

With costless transfers

With costly transfers

Crusoe’s utility

Friday’s utility

C

FIGURE 7.4

UTILITY POSSIBILITIES SCHEDULE WITH

COSTLY TRANSFERS

The set of points we can achieve through redistribution, when

transfers are costly, lies within the utility possibilities curve,

given costless transfers.

169Analyzing Social Choices

The rich may work less hard than they would otherwise because they reap only a fraction of the returns for their eff ort, whereas the poor may work less hard because by working harder, they may lose eligibility for benefi ts. The magnitude of these disincentives—a subject of considerable controversy— aff ects the entire shape of the utility possibilities schedule. In Figure 7.4, the red line represents the utility possibilities schedule assuming that it is cost- less to transfer resources. The black line, which lies far below the previous locus, except at point C—the point that occurs without any redistribution— represents the schedule when transfers are very costly.

EVALUATING THE TRADE-OFFS

The second basic concept used in analyzing social choices is the social indiff erence curve. As described in Chapter 3, an indiff erence curve gives those combinations of goods which give the individual the same level of utility. Just as individuals derive utility from the goods they consume, we can think of society as deriving its welfare from the utility received by its members. The social welfare function gives the level of social welfare corresponding to a particular set of levels of utility attained by members of society. The social indiff erence curve is defi ned as the set of combi- nations of utility of diff erent individuals (or groups of individuals) that yields equal levels of welfare to society—for which, in other words, the social welfare function has the same value.

The social welfare function provides a basis for ranking any allocation of resources: we choose the allocations that yield higher levels of social welfare. The Pareto principle says that we should prefer those allocations in which at least some individuals are better off and no one is worse off . It says that if some individuals’ utility is increased and no one else’s utility is decreased, social welfare increases. Thus, in Figure 7.5, the combinations to the north- east of A make everyone better off , and hence satisfy the Pareto principle.

Unfortunately, most choices involve trade-off s, with some individuals being made better off and others worse off . At point B the second group is better off than at A, but the fi rst group is worse off . We thus need a stronger criterion, and this is what the social welfare function provides. The social indiff erence curves provide a convenient diagrammatic way of thinking about the kinds of trade-off s society faces in these situations. Thus, in Figure 7.5, all combinations of the utilities of Groups 1 and 2 that are on the social indiff erence curve labeled W2 yield a higher level of social welfare than those combinations on the curve labeled W1. This shows that B is preferred to A.

Social welfare functions can be thought of as a tool economists use to summarize assumptions about society’s attitudes toward diff erent

170 CHAPTER 7 EFFICIENCY AND EQUITY

distributions of income and welfare. If society is very concerned about inequality, it might not care that Crusoe has to give up seventy oranges for Friday to get one orange, “since Crusoe has so many to begin with.” As long as Friday is poorer than Crusoe, any sacrifi ce on Crusoe’s part that makes Friday better off would be justifi ed. On the other hand, society might not care at all about inequality; it could value an orange in the hands of Friday exactly the same as an orange in the hands of Crusoe, even though Friday is much poorer. In that case, it would focus only on effi ciency—the number of oranges available. No redistribution of oranges from Crusoe to Friday would be justifi ed if, in the process, a single orange was lost.

UTILITARIANISM�Social welfare functions—and the associated social indiff erence curves—can take a variety of shapes; Figure 7.6 illustrates three diff erent cases. In Figure 7.6A, the social indiff erence curve is a straight line, implying that no matter what the level of utility of Friday and Crusoe, society is willing to trade off one “unit” of Friday’s utility against one unit of Crusoe’s. The view represented by this social indiff er- ence curve has a long historical tradition. Jeremy Bentham was the leader of a group, called utilitarians, that argued that society should maximize the sum of the utilities of its members; in our simple example with two individuals, the social welfare function is

W 5 U1 1 U2.

SOCIAL INDIFFERENCE CURVES

Society is willing to trade off some decrease in the utility

of one group for an increase in that of another group. A social

indifference curve gives the combinations of utilities of

Group 1 and Group 2 between which society is indifferent.

Points on the social indifference curve labeled W2 yield a higher

level of social welfare than do points on the social indifference

curve labeled W1.

FIGURE 7.5

Social indifference

curves

North

East

Utility of first group

Utility of second

group

W2

W1

W1 W2

A

B

171Analyzing Social Choices

ALTERNATIVE SHAPES OF SOCIAL INDIFFERENCE CURVES

(A) A utilitarian is willing to give up some utility for Crusoe as long as Friday gains at least an equal amount of utility. The social indifference curves are straight lines. (B) Some argue that society requires more than an equal increase in the utility (U2) of a rich individual to compensate for a decrease in the utility (U1) of a poor individual. (C) Rawls maintains that no amount of increase in the welfare of the rich can com- pensate for a decrease in the welfare of the poor. This implies that the social indifference curves are L-shaped.

FIGURE 7.6

Utilitarian social indifference curve

A

Friday’s utility (U1)

Crusoe’s utility (U2)

Rawlsian social indifference curve

B

Friday’s utility (U1)

Crusoe’s utility (U2)

Change in U2

U*2

U*1

Change in U1

C

Friday’s utility (U1)

Crusoe’s utility (U2)

172 CHAPTER 7 EFFICIENCY AND EQUITY

It is clear that with this social welfare function, the social indiff erence curve has the shape depicted in Figure 7.6A.

It is important to emphasize that with a utilitarian social welfare function, society is not indiff erent to an increase of one orange (or one dollar of income) for Individual 1 and a decrease of one orange (or one dollar of income) for Individual 2. If Individual 1 has a lower level of income (fewer oranges) than Individual 2, then the increase in utility of Individual 1 from one more orange (one more dollar) will be greater than the decrease in utility for Individual 2. What the utilitarian social wel- fare function says is that the utility of any individual should be weighted equally to the utility of any other individual.

Many would argue that when one individual is worse off than another, society is not indiff erent to a decrease in the utility of the poorer (Individual 1) matched by an equal increase in the utility of the richer (Individual 2). Society should be willing to accept a decrease in the util- ity of the poor only if there is a much larger increase in the utility of the rich. The social indiff erence curve refl ecting these values is drawn in Figure 7.6B, where it appears not as a straight line but as a curved one; as the poorer individual becomes worse and worse off , the increment in util- ity of the richer individual that makes society indiff erent must be larger and larger (i.e., the slope of the social indiff erence curve becomes steeper and steeper).

RAWLSIANISM�An extreme position of this debate was taken by John Rawls, a former professor of philosophy at Harvard University. Rawls argued that the welfare of society depends only on the welfare of the worst-off individual. So society is better off if you improve that individ- ual’s welfare, but gains nothing from improving the welfare of others. In his view, there is no trade-off . If Friday is worse off than Crusoe, then anything that increases Friday’s welfare increases social welfare. As oranges are transferred from Crusoe to Friday, it makes no diff erence how many are lost in the process—how ineffi cient the transfer process is—as long as Friday gets something. To put it another way, no amount of increase in the welfare of the better-off individual could compensate society for a decrease in the welfare of the worst-off individual. Diagram- matically, this is represented by an L-shaped social indiff erence curve, as in Figure 7.6C.2

2�The social welfare function is written:

W 5 min {U1, . . . , Un}.

Social welfare refl ects only the utility of the worst-off member of society.

173Analyzing Social Choices

TWO CAVEATS

Many public sector economists have made extensive use of the concepts of social welfare functions and the utility possibilities curve, but these concepts have also been extensively criticized, on several grounds.

INTERPERSONAL COMPARISONS�We assume that when an individual consumes more, his or her utility rises. But we cannot measure the level of utility or the change in utility. Social welfare functions seem to assume not only that there is a meaningful way of measuring an indi- vidual’s utility,3 but that there is a meaningful way of comparing the utility of diff erent individuals. For example, with the utilitarian social welfare function, we add up the utility of the dif- ferent members of society. Because we add Crusoe’s and Friday’s utility together, we are assuming that somehow we can compare, in a meaningful numerical way, their levels of utility. But when we transfer an orange from Robinson to Friday, how can we compare in an objective way the value of Friday’s gain and Robinson’s loss?

The same problem arises with a Rawlsian social welfare function, where we are told to maximize the welfare of the worst-off mem- ber of society. To judge who is worst off, we must somehow compare utilities.

Many economists believe that these interpersonal utility compari- sons cannot be made in any meaningful way. I may claim that although I have a much higher income than my brother, I am less happy; not only that, I may claim that I know how to spend income much better, so that the extra increment in my utility from a dollar given to me is much greater than the extra increment in utility that he would get from receiving an extra dollar. How could anyone prove that I was wrong (or right)? Because there is no way of answering this question, economists argue that there can be no scientifi c basis for making welfare comparisons. And because there is no “scientifi c” basis for making such welfare comparisons, many economists believe they should limit themselves to describing the con- sequences of diff erent policies—only pointing out who are the gainers and who are the  losers—and that should be the end of their analysis.

3�In some situations, it may be possible to use the amount of money an individual would be willing to pay for an object as a measure of the utility of that object. However, this does not resolve the problem of comparing utilities across individuals.

SOCIAL CHOICE IN THEORY 1. Construct the opportunity set. The utility

possibilities schedule describes how much one person’s utility must be decreased when another’s is increased.

2. Defi ne preferences. Social indifference curves describe how much society is willing to decrease one person’s utility to increase another’s by a given amount.

3. Adopt programs that increase social welfare. Find the programs that put society on the highest social indifference curve.

174 CHAPTER 7 EFFICIENCY AND EQUITY

They believe that the only circumstances in which economists should make welfare judg- ments are those in which the policy change is a Pareto improvement. Unfortunately, as we have said, few policy changes are Pareto improvements; hence, without making inter- personal comparisons of welfare, economists have little to say regarding policy.

WHENCE SOCIAL WELFARE FUNC- TIONS? The second set of objections con- cerns the very nature of social welfare functions. Individuals have preferences; they can decide whether they prefer some com- bination of apples and oranges to another combination. Society consists of many indi- viduals, but society itself does not have pref- erences. We can describe the preferences of

each individual, but whose preferences does the social welfare function represent? If there were a dictator, the answer to that question would be easy: the social welfare function would refl ect the preferences of the dic- tator. But in a democratic society, there is no easy answer to the question. Some individuals (particularly the rich) may care little for redistribution, whereas others (particularly the poor) may argue that greater weight should be placed on redistribution.

As a descriptive matter—as part of a positive analysis—societies seldom exhibit consistency. One of the results to be described in Chapter 9 explains why this is not unexpected. Most economists think of the concepts we have described—as part of a normative analysis—as tools that help us think sys- tematically about the trade-off s society constantly must face. As we noted earlier, the systematic analysis of these trade-off s actually constitutes an important part of the process by which decisions get made.

SOCIAL CHOICES IN PRACTICE

In practice, government offi cials do not derive utility possibilities sched- ules, nor do they write down social welfare functions. But their approach to deciding whether, say, to undertake any particular project does refl ect the concepts we have introduced.

First, they attempt to identify and measure the net benefi ts (benefi ts minus costs) received by diff erent groups. Second, they ascertain whether

COMPARING INDIVIDUAL AND SOCIAL CHOICES

INDIVIDUAL CHOICES SOCIAL CHOICES

Step 1: Defi ne Opportunity Set

Budget constraint Utility possibilities curve

Step 2: Defi ne Preferences

Individual indifference curve

Social indifference curve

Step 3: Choose Preferred Point

Tangency between individual indifference curve and budget constraint

Tangency between social indifference curve and utility possibilities curve

175Social Choices in Practice

the project is a Pareto improvement—that is, whether everyone is bet- ter off . If so, clearly the project should be undertaken (this is the Pareto principle).

If the project is not a Pareto improvement, matters are more diffi cult. Some gain, some lose. The government needs to make an overall judgment. One commonly used approach looks at two summary statistics, describing “effi ciency” and “equity” eff ects. Effi ciency is measured by simply summing the gains or losses for each individual (which are calculated in a manner to be described shortly). Equity is measured by looking at some overall mea- sure of inequality in society. If a project has net positive gains (positive effi - ciency eff ects) and reduces measured inequality, it should be undertaken. If a project has net positive losses and increases measured inequality, it should not be undertaken. If the effi ciency measure shows gains but the equality measure shows losses (or vice versa), there is a trade-off , which is evaluated using a social welfare function: How much extra inequality is society willing to accept for an increase in effi ciency?

There are numerous examples in which choices between equality and effi ciency must be made. For instance, in general, the more a tax system redistributes income, the greater the ineffi ciencies it introduces. There is a trade-off between equality and effi ciency. There are, of course, import- ant instances of poorly designed tax systems; such tax systems put the economy below its utility possibilities schedule. In such cases, it may be possible to increase both equality and effi ciency.

We now take a closer look at how economists measure effi ciency and inequality.

MEASURING BENEFITS

The fi rst problem is how to measure the benefi ts of some program or proj- ect to particular individuals. In the earlier discussion of utility theory, we described how giving Friday more oranges increased his utility. But how do we measure this?

The standard way this is done is in terms of willingness to pay. We ask how much an individual would be willing to pay to be in one situation rather than another. For example, if Joe likes chocolate ice cream more than vanilla, it stands to reason that he would be willing to pay more for a scoop of chocolate ice cream than for a scoop of vanilla. Or if Diane would rather live in California than in New Jersey, it stands to reason that she would be willing to pay more for the West Coast location.

Notice that how much a person is willing to pay is diff erent from how much that individual must pay. Just because Joe is willing to pay more for

176 CHAPTER 7 EFFICIENCY AND EQUITY

chocolate ice cream than for vanilla does not mean he will have to pay more. What he has to pay depends on market prices; what he is willing to pay refl ects his preferences.

Using willingness to pay as our measure of utility, we can construct a diagram like of Figure 7.7A, which shows the level of utility Mary receives from sweatshirts as the number of sweatshirts she buys increases. This  information is also given in Table 7.1. Here we assume that Mary is willing to pay $200 for fi ve sweatshirts, $228 for six sweatshirts, $254 for seven sweatshirts, and so on. Thus, fi ve sweatshirts give her a utility

UTILITY AND MARGINAL UTILITY

(A) shows that utility increases continually with consumption

but tends to level off as consumption climbs higher. (B) explicitly shows marginal utility; notice that it declines

as consumption increases.

FIGURE 7.7

Number of sweatshirts

Willingness to pay

Utility curve

A

B

0 5 10 15 20

100

200

300

400

500

Number of sweatshirts

Marginal utility

Marginal utility curve

0 5 10 15 20

10

20

30

40

50

177Social Choices in Practice

of 200, six a utility of 228, and seven sweatshirts a utility of 254. Mary’s willingness to pay increases with the number of sweatshirts, refl ect- ing the fact that additional sweatshirts give her additional utility. The extra utility of an additional sweatshirt measured here by the additional amount she is willing to pay, is the marginal utility. The numbers in the third column of Table 7.1 give the marginal, or extra, utility she received from her last sweatshirt. When Mary owns fi ve sweatshirts, an additional sweatshirt yields her an additional, or marginal, utility of 28 (228 – 200); when she owns six sweatshirts, an additional one gives her a marginal utility of only 26 (254 – 228). Figure 7.7B traces the marginal utilities of each of the increments.4

4�As marginal utility is the extra utility from an extra unit of consumption, it is measured by the slope of the utility curve in Figure 7.7A.

TABLE 7.1 UTIL IT Y AND MARGINAL UTIL IT Y

NUMBER OF SWEATSHIRTS MARY'S WILLINGNESS TO PAY (UTILITY) MARGINAL UTILITY

0 0 50

1 50 45

2 95 40

3 135 35

4 170 30

5 200 28

6 228 26

7 254 24

8 278 23

9 301 22

10 323 21

11 344 20

12 364 19

13 383 18

14 401 17

15 418 16

16 434 15

17 449 14

18 463 13

19 476 12

20 488

178 CHAPTER 7 EFFICIENCY AND EQUITY

ORDINARY AND COMPENSATED DEMAND CURVES

We can use the concept of willingness to pay to construct a demand curve. We have already asked how much Mary is willing to pay for each addi- tional sweatshirt. If the price of sweatshirts is $29, then she will buy fi ve sweatshirts. She would have been willing to pay $30 for the fi fth sweat- shirt, so clearly, the marginal benefi t of the fi fth sweatshirt exceeds its cost; but she is willing to pay only $28 for the sixth sweatshirt, so the mar- ginal benefi t is less than the cost. Thus, the marginal utility curve drawn in Figure 7.7B can also be thought of as the demand curve.

However, this is a special demand curve, called the compensated demand curve, which diff ers slightly from the ordinary demand curve. Recall that we constructed the compensated demand curve by asking how much Mary would be willing to pay for each additional sweatshirt; thus, as we give her more sweatshirts, we are always keeping her at exactly the same level of utility.

To construct the ordinary demand curve, we need to know how many units of the commodity Mary would buy at each price. As the price is low- ered, Mary not only demands more, but is made better off . As prices are lowered, individuals substitute the cheaper good for others goods. If the price of sweatshirts is lowered, Mary will substitute sweatshirts for sweat- ers. This is called the substitution eff ect. Because of the lower price, Mary is better off ; if she bought exactly the same amount of goods that she did

Ordinary demand

curve

Compensated demand curve

Quantity

Price FIGURE 7.8

COMPENSATED VERSUS UNCOMPENSATED DEMAND CURVES

The compensated demand curve gives the demand for

a good, assuming as price is changed that money is taken

away, or given to the individual to leave that individual just as well off as he or she was

before the price change. It thus measures only the substitution

effect associated with the price changes. Because, as price is

lowered individuals are better off, and as a result buy slightly more of (normal) commodities,

the ordinary demand curve is slightly fl atter than the compen-

sated demand curve.

179Social Choices in Practice

before, she would have money left over. She spreads this money around. Some of it is spent on buying sweatshirts. The increase in demand for sweat- shirts as a result of the fact that Mary is better off —it is as if she had more income—is called the income eff ect. If we take away this extra money, we have the compensated demand curve; we eliminate the income eff ect. Thus, the compensated demand curve refl ects only the substitution eff ect. In most cases, the diff erences between the two are negligible. If Mary spends one-tenth of 1 percent of her income on sweatshirts, taking away the extra income has almost no eff ect on her demand for sweatshirts, or any other commodity. Thus, Figure 7.8 shows the ordinary and compensated demand curves as being almost the same, with the ordinary demand curve being slightly fl atter (lowering the price from its current level results in a slightly greater increase in the quantity demanded, and raising the price from its current level results in a slightly greater decrease in quantity demanded).

CONSUMER SURPLUS

The diff erence between what an individual is willing to pay and what he or she has to pay is called the consumer surplus. Mary would have been willing to pay $50 for the fi rst sweatshirt, $45 for the second, $40 for the third, and so on. If the market price is $29, however, that is all she has to pay for each sweatshirt. Thus, on the fi rst sweatshirt, she gets a surplus of $21 ($50, what she was willing to pay, minus $29, what she actually pays); on the second sweatshirt, she gets a surplus of $16; on the third sweat- shirt, she gets a surplus of $11, and so on. The total consumer surplus is thus the sum: $21 1 $16 1 $11 1 $6 1 $1 5 $55.

Diagrammatically, the consumer surplus is depicted in Figure 7.9 as the shaded area under the compensated demand curve and above the price line. Of course, because the compensated and uncompensated demand curves are almost the same, typically, we calculate the consumer surplus simply by looking at the area under the ordinary demand curve above the price line.

USING CONSUMER SURPLUS TO CALCULATE THE BENEFITS OF A GOVERNMENT PROJECT The compensated demand curve can be useful for measuring the benefi ts of government projects. For instance, constructing a bridge on which no toll will be charged can be thought of as lowering the price from “infi nity” (one simply cannot buy trips across a nonexistent bridge) to zero. The welfare gain is just the total consumer surplus, the area under the demand curve in Figure 7.10. This measures the maximum individuals could pay and still be as well off with the bridge as they were without it. Clearly, if the consumer surplus is less than the

180 CHAPTER 7 EFFICIENCY AND EQUITY

cost of the bridge, it does not pay to construct it, whereas if the consumer surplus is greater than the cost of the bridge, it does pay to build it.

There are several ways that economists go about trying to measure consumer surplus and willingness to pay. For many goods, there are data with which economists can construct the demand curve (the quantity that individuals are willing to purchase at each price) and the compensated

GRAPHICAL REPRESENTATION OF CONSUMER SURPLUS

An individual’s surplus is the difference between what the

individual is willing to pay (rep- resented by the area beneath

the demand curve) and what he or she actually pays (the area

under the price line). The con- sumer surplus here is indicated

by the shaded region.

MEASURING THE BENEFITS OF A GOVERNMENT

PROJECT: BUILDING A BRIDGE

The benefi ts of a bridge for which no tolls will be charged can be measured by the total

area under the demand curve— the total consumer surplus.

FIGURE 7.9

F IGURE 7.10

9 Sweatshirts

Price

1 2

50

45

40

35

30

25

3 4 5 6 7 8

Compensated demand curve

P = $29.00

Compensated demand curve

Trips

Toll ($)

Consumer surplus

181Social Choices in Practice

demand curve.5 In that case, willingness to pay can be calculated simply as the area under the compensated demand curve. For some goods, such as the Grand Canyon, there is no market demand curve, yet the government still might want to know how much citizens are willing to pay to preserve it in its pristine condition. Econo- mists have designed elaborate survey techniques to elicit meaningful answers from individuals concerning their willingness to pay. These meth- ods are discussed at greater length in Chapter 11.

MEASURING AGGREGATE SOCIAL BENEFITS

We have now described how we can measure the benefi ts that an individual receives. Social benefi ts are typically measured by adding up the benefi ts received by all individuals. The numbers obtained represent the total will- ingness to pay of all individuals in society. The diff erence between the total willingness to pay and the total costs of a project can be thought of as the net “effi ciency” eff ect of the project. It is a dollar value of the net benefi ts.

MEASURING INEFFICIENCY

In assessing alternative policies, economists have put particular emphasis on economic effi ciency. Taxes are criticized for discouraging work eff ort, monopolies for restricting production and driving up prices. To measure the dollar value of an ineffi ciency, economists use exactly the same meth- odology they use to measure the dollar value of a new project. There, we calculated the consumer surplus associated with the project. Here, we calculate the consumer surplus associated with the elimination of the ineffi ciency. That is, economists ask: How much would an individual be willing to give up to have the ineffi ciency eliminated? Consider the inef- fi ciency caused by a tax on cigarettes. We ask each individual how much he or she would be willing to pay to have the tax on cigarettes eliminated. Say one answer is $100; thus eliminating the cigarette tax and imposing in its place a $100 lump-sum tax—that is, a tax that the individual would

5�As was noted previously, for most goods, the compensated and uncompensated demand curves are very similar. If the income elasticity (the percentage increase in the demand for the good when income increases by 1 percent) is known, one can calculate the compensated demand curve from the uncom- pensated demand curve.

CONSUMER SURPLUS • Measured by the area under the (compensated)

demand curve.

• Used to measure the value of a government project or assess the magnitude of an ineffi ciency.

182 CHAPTER 7 EFFICIENCY AND EQUITY

DRAWING A POVERTY LINE

T he offi cial poverty line determines how many people the government counts as poor. But what determines the poverty line itself? In the early 1960s, Mollie Orshansky, an offi cial

at the Social Security Administration, developed a method of measuring poverty from a survey of household expenditures. She found that a typical family spent one-third of its income on food. She then gathered information on minimum food bud- gets for families of various sizes, and multiplied that number by 3 to get an estimate of the poverty line for the different family sizes. With minor changes, Orshansky’s poverty line was offi cially adopted in 1969 and it has been increased by the overall rate of infl ation since then.

One can ask a number of questions about how poverty is measured. Here are three:

First, the survey Orshansky relied on to fi nd that households spent one-third of their income on food was taken in 1955. Since then, household expendi- tures have shifted. Households now spend a much lower percentage of income on food, perhaps one- fourth or one-fi fth. If the minimum food budget were accordingly multiplied by 4 or 5, the poverty line would be much higher.

Second, the poverty line does not take in-kind benefi ts into account. In-kind benefi ts include any benefi ts that are not received in cash form, such as Medicaid, food stamps, and subsidized school lunches. If those benefi ts are measured as

additional income, the number of people below the poverty line falls by about 20 percent.

Finally, some critics have proposed that pov- erty should be thought of as a relative rather than an absolute concept. They argue that those at the bottom of society—say, the bottom 5 or 10 or 20  percent—are poor relative to everyone else. Poverty is more appropriately viewed as an extreme case of inequality.

For many, this last criticism goes too far. They fear that a relative concept of poverty could reduce the moral urgency of fi ghting poverty. There is broad social support for efforts to ensure that people have basic levels of food, housing, clothing, and medical care, even if defi ning those amounts is controversial.

In 1995, a National Academy of Sciences (NAS) study proposed major revisions in how we mea- sure poverty. Although there was agreement about including noncash income, the diffi cult problems of how best to include health care expenditures were not fully resolved. Should a sick, poor person who receives $150,000 for a kidney transplant have that added to his or her income, in which case he or she now appears to be in an upper income bracket? The study proposed an adjustment in the poverty level that went beyond just taking into account infl ation, but it did not propose increasing the poverty level in proportion to increases in average income, which would have made poverty a purely relative phenom- enon. However, even this compromise generated

have to pay regardless of what he or she did—leaves this individual’s wel- fare unchanged. The diff erence between the revenue raised by the cig- arette tax (say, $80) and the lump-sum tax that the individual would be willing to pay is called the deadweight loss or excess burden of the tax. It is the measure of the ineffi ciency of the tax. Taxes, other than lump- sum taxes, give rise to a deadweight loss because they cause individuals to forgo more-preferred consumption in favor of less-preferred consump- tion to avoid payments of the tax. Thus, even a tax that raises no gov- ernment revenue—because individuals completely avoid purchasing the taxed commodity—can have a substantial excess burden.

183Social Choices in Practice

We can calculate the deadweight loss using compensated demand curves. Assume the cost of producing a cigarette is c0, and the tax raises the price from c0 to c0 1 t, where t is the tax per pack. We assume the individual consumes q0 packs of cigarettes with the tax, and q1 after the tax has been removed (but replaced by a lump-sum tax that leaves the individual no bet- ter or worse off than when there was a cigarette tax). We have drawn the resulting compensated demand curve in Figure 7.11. The deadweight loss is measured by the shaded area ABC, the area under the compensated demand schedule and above c0, between the output with and without the tax.

The triangle ABC is sometimes called a Harberger triangle,6 in honor of University of Chicago and UCLA economist Arnold Harberger, who used such triangles not only to measure the ineffi ciencies associ- ated with distortionary taxation but also to measure other ineffi ciencies, such as those associated with monopoly. Why does the Harberger trian- gle provide a measure of deadweight loss? The price tells us the value of the last unit consumed; that is, at q0, the individual is willing to trade off p0 5 c0 1 t units of “income” (with which he or she could have purchased other goods) for one more pack of cigarettes. Of course, when the individ- ual has q0 1 1 packs of cigarettes, he or she will value an additional pack of cigarettes less than when he or she has q0 packs, and so the price the individual is willing to pay will fall.

6�See, for instance, A. Harberger, “Taxation, Resource Allocation and Welfare,” in The Role of Direct and Indirect Taxes in the Federal Revenue System, ed. J. Due (Princeton, NJ: Princeton University Press, 1964), reprinted in A. Harberger, Taxation and Welfare (Chicago: University of Chicago Press, 1974).

SOURCES: Joyce E. Allen and Margaret C. Simms, “Is a New Yardstick Needed to Measure Poverty?” Focus (February 1990): 6–8; Measuring Poverty: A New Approach (Washington, DC: National Academy of Sciences, 1996); Gordon M. Fisher, “The Development and History of the U.S. Poverty Thresholds—A Brief Overview,” GSS/SSS Newsletter (Winter 1997): 6–7; and Kathleen Short, “The Research Supplemental Poverty Measure: 2011,” Current Population Reports, P60-244, United States Census Bureau, November 2012.

a strong dissent from one of the members of the academy’s panel.

The U.S. Census Bureau tested a number of experimental poverty measures based on recom- mendations of the 1995 NAS report, and, in 2009, the Interagency Technical Working Group on Developing a Supplemental Poverty Measure was formed. In 2011, the Census Bureau, with support from the Bureau of Labor Statistics, released its fi rst report presenting a newly developed supplemen- tal poverty measure (SPM) that incorporated many NAS-based measures. According to the Census Bureau, the SPM will not replace the offi cial pov- erty measure and will not be used to determine

eligibility for government programs. Instead, it is designed to provide an alternative perspective of poverty that better refl ects contemporary social and economic conditions, including government policies that signifi cantly alter resources avail- able to familities, and hence, their poverty status. When compared with the offi cial poverty measure in 2011, the SPM estimates a slightly higher over- all poverty rate (16.1 versus 15.1 percent), but a substantially lower poverty rate for those under 18 years old (18.1 versus 22.3 percent) and almost double the poverty rate for those 65 years and older (15.1 verus 8.7 percent).

184 CHAPTER 7 EFFICIENCY AND EQUITY

Assume that initially consumption is 100 packs, and consumption increases by 10 packs when the tax is removed; the tax is 10 cents, and the cost of production is $1 per pack. (Tax revenue is 100 packs times 10 cents per pack, or $10.) The individual is willing to pay $1.10 for the fi rst additional pack, $1.09 for the second, $1.08 for the third, and so on. If the tax were elim- inated, and the price fell to c0, the cost of production ($1 a pack), the total amount that the individual would be willing to pay would be 10 cents times 100 packs 5 $10 (the amount saved on the fi rst 100 packs the individual has purchased, which is equal to the tax revenue) plus 10 cents for the 101st pack (the diff erence between how much the individual values the 101st pack and what he or she must pay), 9 cents for the 102nd pack, and so on. Remember, we are calculating how much more the individual would be willing to pay beyond the $1 that he or she will have to pay for each pack. The total that the individual would be willing to pay is thus $10.50. Because the tax raised revenue of $10, the deadweight loss is 50 cents, which is, of course, just the area under the compensated demand curve and above c0, between q0 and q1.

QUANTIFYING DISTRIBUTIONAL EFFECTS

Assessing the distributional eff ects of a project or a tax is often far more complex than assessing the effi ciency eff ects. There are many groups in a society, and each may be aff ected diff erently. Some poor individuals may be hurt, some helped; some middle-income individuals may be helped, others hurt. In some cases, the rich may be helped the most, the poor helped moderately, and the middle class made only slightly worse off .

MEASURING INEFFICIENCIES

The area ABC measures the deadweight loss, the effi ciency

loss as a result of a cigarette tax. A lump-sum tax that would

have the same effect on the individual’s welfare as the

cigarette tax would raise an additional revenue of ABC.

FIGURE 7.11

Consumption

Price

q0 q1

A

B

Cc0

c0 + t Deadweight loss

Compensated demand schedule

185Social Choices in Practice

THE GREAT GATSBY CURVE

R ising inequality in the United States since the 1970s has been well documented; examples include the dramatic contrast in real after-tax income growth between the top 1 percent of fami- lies and everyone else, the steadily shrinking middle class, and the substantial rise in the Gini coeffi cient.

However, less attention has been given to the relationship between current levels of inequality and intergenerational income mobility. Recent studies have found that your parents’ income is a good predictor of your subsequent income.

Alan Krueger, Chairman of the Council of Eco- nomic Advisers under President Obama, created what he calls the “Great Gatsby Curve,” named after F. Scott Fitzgerald’s fi ctional character known for his lavish lifestyle during the Roaring Twenties. The Great Gatsby Curve plots the Gini coeffi cient— current income distribution—on the x-axis and intergenerational income elasticity (IGE)—relation between parents’ and children’s income—on the

y-axis. It shows a disconcerting link between IGE and income inequality at a given point in time: countries that have a high degree of inequality also tend to have less economic mobility across genera- tions. For example, OECD countries that had more inequality across households in the 1980s also had more persistence in income from one generation to the next.

Not only did the United States have greater income stagnation than most other OECD countries in this study, but one cannot help but be concerned that already low income mobility across generations has been exacerbated by the rise of inequality in the United States. Using the Great Gatsby Curve, Krueger estimates that the IGE for the United States will increase from 0.47 to 0.56 for the next gener- ation. In other words, lack of equality threatens equality of opportunity in the United States—the fortunes of one’s parents are increasingly important in determining the fate of their children.

In practice, governments focus on a few summary measures of inequality. Because the poor are of particular concern, they receive spe- cial attention. The poverty index measures the fraction of the population whose income lies below a critical threshold; below that threshold, indi- viduals are considered to be in poverty. In 2010, the poverty threshold for a family of four was $22,113.7

Another measure is the poverty gap. The poverty index only counts the number of individuals who are below the poverty threshold; it does not look at how far below that threshold they are. The poverty gap asks: How much income would we have to give to the poor to bring them all up to the poverty threshold?

Two other measures are briefl y discussed in the appendix to this chapter.

7�U.S. Census Bureau, Poverty Thresholds, http://www.census.gov.

186 CHAPTER 7 EFFICIENCY AND EQUITY

THREE APPROACHES TO SOCIAL CHOICES

We now have the basic tools for describing social choices in the diffi cult cases in which the project does not constitute a Pareto improvement. There are three approaches, which we shall refer to as the compensation principle, trade-off s across measures, and the weighted benefi ts approach.

THE COMPENSATION PRINCIPLE

What happens if the total willingness to pay exceeds the total costs, but the costs borne by some individuals exceed their willingness to pay? Should the project be undertaken? The compensation principle says that if the aggregate willingness to pay exceeds the cost, the project should be undertaken. Most economists criticize this principle, for it ignores dis- tributional concerns. Only if the compensation is actually paid to those adversely aff ected can we be sure that the project is desirable, for then it is a Pareto improvement.

Because the compensation principle does not pay adequate attention to distributional concerns, economists have turned to two other approaches.

TRADE-OFFS ACROSS MEASURES

With a measure of effi ciency (net benefi ts) and a measure of inequality, public decision making—conceptually, at least—should be easy: one sim- ply evaluates whether the increase in effi ciency is worth the increase in inequality, or vice versa.

The previous two sections have described how we measure total effi ciency and inequality. These are just statistics, numbers that help to summarize the impacts of a project or program. Such summary statistics, though useful, often submerge some of the detailed information that is important in public decision making. Ideally, we would look at the impacts on each individual, and then use the social welfare function to add up the eff ects. In practice, the government does not attempt to identify impacts on every individual, but it does attempt to ascertain the eff ects on each major group. For instance, it may look at the impact on individuals in diff erent income categories—say, families with incomes below $10,000, between $10,000 and $20,000, and so forth.

187Three Approaches to Social Choices

WEIGHTED NET BENEFITS

This may be all the information required for policy makers to make a decision. If the aggre- gate net benefit (the sum of the willingnesses to pay minus costs) is positive, and if the poor are net beneficiaries and the rich are net los- ers, then the project increases both efficiency and equity and should be adopted. Often, however, matters are more complicated. For instance, the poor and the rich may be worse off, but middle-income individuals better off. How do we assess such a change? Again, we turn to our social welfare function to add up the effects. The weighted net benefits approach assigns weights to the net gains of different groups to summarize the impacts in a single number. The social welfare func- tion tells us how to do that. Because of the concern for equity, effects on higher-income groups are weighted less heavily; how much less heavily may determine whether it is desir- able to undertake a project. For instance, a project that helps the middle class but hurts the poor and the rich might not be undertaken if we weight the losses of the poor much more heavily than the gains to the middle class.8

The use of weights can be thought of as based on three assumptions: fi rst, that there is diminishing marginal utility; second, that diff erent individuals have the same relation between utility and income; and third, that society is concerned with total utility—the sum of the utilities of all individuals (the utilitarian social welfare function). Although each of these assumptions may be questioned, we can also think of these proce- dures as simply a convenient way to summarize data that decision makers often fi nd helpful.

8�Given the importance of these weights in social decision making, economists have looked for a rational basis for assigning weights. One way is to think about how rapidly marginal utility diminishes with increased income. Inferences about this can be made from observing individual behavior in risky situ- ations: if marginal utility diminishes very rapidly, individuals will be very averse to undertaking large risks, and will be willing to pay large premiums to divest themselves of risk. On the basis of this evi- dence, most economists argue that a doubling of income will lower the marginal utility of income by a factor of between 2 and 4, so that a change in the income of a middle-class individual with an income of $30,000 should be weighted half to a quarter of the same change in income of a poor individual with an income of $15,000.

SOCIAL CHOICE IN PRACTICE • Identify Pareto improvements.

• If some individuals are better off while others are worse off, identify groups of individuals who are better off and groups that are worse off (by income, region, age), and gains and losses of each major group:

Ascertain whether aggregate net benefi ts are positive (compensation principle).

Look at change in measure of effi ciency and measure of inequality, and evaluate trade-offs.

Calculate weighted net benefi ts, weighting gains and losses to the poor more heavily than those to the rich, according to the social welfare function.

188 CHAPTER 7 EFFICIENCY AND EQUITY

THE TRADE-OFF BETWEEN EFFICIENCY AND FAIRNESS REVISITED

The trade-off between effi ciency and fairness is based on two notions: the economy is (Pareto) effi cient, so no one can be made better off with- out making others worse off , and transferring income from one person to another is costly. As we noted in Chapter 3, competitive economies (with- out other market failures) are Pareto effi cient. However, the conditions required are highly restrictive; in general, there is room for improvement. There are many Pareto effi cient outcomes, one corresponding to each set of initial endowments (of, say, wealth), but, typically, taxing rich individu- als to help those less well-off imposes costs, as such taxes may discourage work or savings.

However, increasingly around the world, and especially in the United States, it has become clear that it is possible to have more equality and greater effi ciency at the same time. The obvious example is a monopoly (like Microsoft), which raises prices and impedes economic effi ciency, but garners huge wealth for the monopolist. More eff ective enforcement of antitrust laws would have led a more effi cient economy with less inequal- ity. This is an example of a wide range of activities, referred to as rent seeking, by which individuals increase their income, not so much by add- ing to the size of the economic pie, but by increasing their share of the economic pie. Other examples include CEOs who use defi ciencies in cor- porate governance (the rules that govern corporations) to pay themselves outsized bonuses; those in the fi nancial sector who engage in predatory lending and other activities that do not improve economic performance, and likely hinder it, but garner for the bankers enormous incomes; cor- porations that use their political infl uence to get government contracts that overpay them for what they sell to the government; or corporations that obtain natural resources from the government, paying only a frac- tion of what those resources are worth. Tax laws that tax capital gains at a low rate, sought after by the rich—a form of income that overwhelm- ingly accrues to those at the top—distorts the economy. They encourage more resources to go into speculative activities, which can destabilize the economy, rather than into, say, real research that would enhance growth and effi ciency. Reforming these laws would enhance both effi ciency and equity.

Effi ciency and equity can be simultaneously enhanced in other ways. In the United States, a child’s life prospects are more dependent on the

189Review and Practice

income and education of his or her parents than in almost any of the other advanced countries, which means that a large fraction of the children born to poor families are not living up to their potential. Better public schools would create more equality and more equality of opportunity, and improve overall economic performance.

In short, a combination of market and government failures produces an economy characterized by widespread rent seeking and other distor- tive behavior that benefi ts a very small number of extremely wealthy citi- zens at the expense of everyone else. In such circumstances, we can have both a more effi cient and productive economy, as well as increased equal- ity. In fact, it is inequality that is reducing the economy’s effi ciency and lowering its productivity.9

9�For an in-depth examination of the relationship between economic effi ciency and social equality in the United States today, see J. E. Stiglitz, The Price of Inequality: How Today’s Divided Society Endangers Our Future (New York: W. W. Norton & Company, 2012).

SUMMARY

1. Welfare economics—or normative economics—is concerned with criteria for evaluating alterna- tive economic policies. In general, it takes into account both effi ciency and equity.

2. The social welfare function provides a framework within which the distributional consequences of a policy may be analyzed. It specifi es the increase in utility of one individual that is required to com- pensate for a decrease in utility of another.

3. In the utilitarian social welfare function, social welfare is equal to the sum of the utilities of the individuals in society. In the Rawlsian social wel- fare function, social welfare is equal to the utility of the worst-off individual in society.

4. The concept of consumer surplus—how much individuals are willing to pay for a project or pro- gram in addition to what they have to pay—is used to measure the aggregate benefi ts of a project

or  program. The concept of deadweight loss is used to measure the ineffi ciency of a tax; it asks how much extra revenue could have been gener- ated by a lump-sum tax that would have left indi- viduals just as well off as the tax that was imposed.

5. As a practical matter, in evaluating alterna- tive proposals we do not detail the impact each proposal has on each individual in society, but rather we summarize its eff ects by describing its impact on some measure of inequality (or on some well-identifi ed groups) and describing the effi ciency gains or losses. Alternative proposals often present trade-off s between effi ciency and distribution; to get more equality, one must give up some effi ciency. Diff erences in views arise concerning the nature of the trade-off s (how much effi ciency one needs to give up to get some increase in equality), and values (how much effi - ciency one should be willing to give up, at the margin, to get some increase in equality).

REVIEW AND PRACTICE

190 CHAPTER 7 EFFICIENCY AND EQUITY

6. Three approaches for making social choices when there is not a Pareto improvement are:

a. the compensation principle;

b. trade-off s across measures of effi ciency and equality; and

c. the weighted benefi ts approach.

7. The poverty index measures the fraction of the population whose income lies below some threshold.

KEY CONCEPTS

Compensated demand curve

Compensation principle

Consumer surplus

Dalton–Atkinson measure of inequality

Deadweight loss (excess burden)

Diminishing marginal utility

Gini coeffi cient

Harberger triangle

Income effect

Interpersonal utility comparisons

Lorenz curve

Lump-sum tax

Marginal utility

Poverty gap

Poverty index

Rawlsian social welfare function

Social indifference curves

Social welfare function

Substitution effect

Trade-off

Utilitarians

Utility function

Weighted net benefi ts approach

QUESTIONS AND PROBLEMS

1. Assume that Crusoe and Friday have identical utility functions described by the following table.

Utility Functions for Friday and Crusoe

NUMBER OF ORANGES UTILITY MARGINAL UTILITY

1 11

2 21

3 30

4 38

5 45

6 48

7 50

8 51

Draw the utility function. Fill in the marginal utility data in the table above, and draw the mar- ginal utility function.

2. Assume that there are eight oranges to be divided between Friday and Crusoe. Take a utilitarian view—assume that social welfare is the sum of the utilities of the two individuals. Using the data from Problem 1, what is the social welfare corre- sponding to each possible allocation of oranges? What allocation maximizes social welfare? Show that it has the property that the marginal utility of an extra orange given to each individual is the same.

3. Now take a Rawlsian view and assume that the social welfare function is the level of utility of the individual with the lowest utility level. Using the data from Problem 1, and again assuming there are eight oranges, what is the social welfare associated with each allocation of oranges? What allocation maximizes social welfare?

4. Draw the utility possibilities schedule based on the data from Problem 1. Mark the points that maximize social welfare under the two alterna- tive criteria from Problems 2 and 3.

5. Assume that Crusoe’s and Friday’s utility func- tions are described in Problem 1. Assume now, though, that initially Crusoe has six oranges and Friday two. Assume that for every two oranges taken away from Crusoe, Friday gets only one, an orange being lost in the process. What does the utility possibilities schedule look like now? Which of the feasible allocations maximizes

191Review and Practice

social welfare with a utilitarian social wel- fare function? With a Rawlsian social welfare function?

6. If marginal utility did not decrease at all for both Friday and Crusoe, what would the utility possi- bilities schedule look like?

7. Consider an accident in which an individual loses a leg. Assume that it lowers the individual’s utility at each level of income but increases his or her marginal utility (at each level of income), though only slightly. Show diagrammatically the utility functions before and after the accident. Show that if you were a utilitarian, you would give more income to the individual after the accident, but that even after the transfer, the individual who had the accident is worse off than before. Show the compensation that a Rawlsian would pro- vide. Is it possible for a utilitarian to give more to the individual who had experienced the acci- dent than a Rawlsian? Under what circumstances would a utilitarian give nothing to an individual who had experienced an accident?

8. For each of the following policy changes, explain why the change is or is not likely to be a Pareto improvement:

a. Building a park, fi nanced by an increase in the local property tax rate.

b. Building a park, fi nanced by the donation of a rich philanthropist; the city acquires the land by exercising the right of eminent domain.10

c. Increasing medical facilities for lung cancer, fi nanced out of general revenues.

d. Increasing medical care facilities for lung cancer, fi nanced out of an increase in the cig- arette tax.

e. Replacing the system of agricultural price sup- ports with a system of income supplements for poor farmers.

f. Protecting the automobile industry from cheap foreign imports by imposing quotas on the importation of foreign cars.

g. Increasing Social Security benefi ts, fi nanced by an increase in the payroll tax.

h. Replacing the primary reliance at the local level on the property tax with state revenues obtained from an income tax.

i. Eliminating rent control laws.

In each case, state who the losers (if any) are likely to be. Which of these changes might be approved under the compensation principle? Which might be approved under a Rawlsian social welfare function?

9. Assume you are shipwrecked, and there are ten of you in a lifeboat. You know that it will take ten days to reach shore and that there are rations for only ten person-days. (The ration is the minimum amount needed for survival.) How would a utilitarian allocate the rations? How would a Rawlsian? Some people think that even Rawlsian criteria are not suffi ciently egalitarian. What might an extreme egalitarian individual advocate? What does Pareto effi ciency require? In each case, state what assumptions you need to make to make the decision.

10�The right of eminent domain gives public authorities the right to take property, with compensation, for public uses.

192

APPENDIX: ALTERNATIVE MEASURES OF INEQUALITY

In the text, we introduced the two most commonly used measures of inequality. These measures are criticized, however, for focusing exclu- sively on the impact on the very poor. In this appendix, we discuss two more-inclusive measures.

THE LORENZ CURVE

Economists often represent the degree of inequality in an economy by a diagram called the Lorenz curve, shown in Figure 7.12. The Lorenz curve shows the cumulative fraction of the country’s total income earned by the poorest 5 percent, the poorest 10 percent, the poorest 15 percent, and so on. If there were complete equality, then 20 percent of the income would accrue to the lowest 20 percent of the population, 40 percent to the lowest 40 percent. The Lorenz curve would be a straight line, as depicted in Figure 7.12A. On the other hand, if incomes were very concentrated, then the lowest 80 percent might receive almost nothing, and the top 5 percent might receive 80 percent of total income; in this case, the Lorenz curve would be bowed, as illustrated in Figure 7.12B. When there is a great deal of inequality, the shaded area between the 45-degree line and the Lorenz curve is large (Figure 7.12B). When there is complete equality, this area is zero (Figure 7.12A).

The ratio of the area that lies between the line of perfect income equality and the Lorenz curve over the total area under the line of perfect income equality is a commonly employed measure of inequality, called the Gini coeffi cient. The lower the Gini coeffi cient, the more equal the distribution of income: 0 indicates complete equality (every person has the same income) and 1 indicates complete inquality (one person has all the income).

Figure 7.13 shows Lorenz curves for the United States, both before and after the government tax and transfer programs have had their eff ect. The after-tax curve is decidedly inside the pre-tax, indicating that the combined eff ect of government redistribution programs is to make incomes more equal than the market would have made them. Thus, while the effi ciency costs are less clear-cut, the redistributive gains are undeniable.

193Appendix: Alternative Measures of Inequality

THE LORENZ CURVE

(A) Shows a Lorenz curve for an economy in which income is evenly distributed. The bottom 20 percent of the economy has 20 percent of income, the bot- tom 40 percent has 40 percent of income, and so on. (B) Shows a Lorenz curve for an economy in which income is unequally distributed. The curvature of the line indicates that now the bottom 20 percent has less than 20 percent of income, the bottom 40 percent has less than 40 percent of income, and so on.

FIGURE 7.12

Percent of population

Percent of population

Percent of income

Lorenz curve with perfect

income equality

A

0 10 20 30 40 50 60 70 80 90 100

100

90

80

70

60

50

40

30

20

10

Percent of income

Lorenz curve with income

inequality

B

0 10 20 30 40 50 60 70 80 90 100

100

90

80

70

60

50

40

30

20

10

194 CHAPTER 7 EFFICIENCY AND EQUITY

THE DALTON–ATKINSON MEASURE

There is another measure, first introduced by Sir Hugh Dalton, a pro- fessor of public finance at the London School of Economics who went on to become the Chancellor of the Exchequer for the United Kingdom. This measure was based on the premise that societies prefer more egal- itarian distributions. Figure 7.14 shows two distributions of income. In distribution B, more of the income is concentrated at the center, and for societies that value equality, this is the preferred distribution. We can ask: If  society could move from its current distribution to a situation in which income was completely equally distributed, what fraction of its total income would it be willing to give up? This frac- tion is called the Dalton–Atkinson measure of inequality. Of course, different individuals might have different views on the amounts that society should be willing to give up (this says nothing about how much they would have to give up to accomplish the redistribution). The amount society would be willing to give up depends on its social wel- fare function. With a Rawlsian social welfare function, the amount would be much larger than with a utilitarian social welfare function.

INEQUALITY MEASURES

Taxes and subsidies affect the distribution of income. The

fi gure shows two Lorenz curves for the United States in 1995,

income before and after taxes have been levied and gov-

ernment transfers have been received. Clearly, some redistri- bution does take place through

these mechanisms, as they move the Lorenz curve toward

greater equality.

FIGURE 7.13

SOURCE: U.S. Census Bureau, Current Population Survey, Table E, March 1996.

Percent of population

Percent of income

Before tax and transfer

programs

After tax and transfer

programs

0 20 40 60 80 100

100

80

60

40

20

195

Anthony Atkinson, of Nuffield College, Oxford, argued that the amount was significant, often between a quarter and a third of total income in more developed economies. Changes in the Dalton–Atkinson measure can be used to assess the impact on inequality of any proposed govern- ment program.11

11�Formally, the Dalton–Atkinson measure can be defi ned as follows. Assume a utilitarian social welfare function

W 5 U(Y1) 1 U(Y2) 1 U(Y3) 1 · · ·

and let Y be the average income. Then the Dalton–Atkinson measure D is given by

U[(1 2 D)Y] 5 U(Y1) 1 U(Y2) 1 U(Y3) 1 · · · .

The measure clearly depends on the utility function. Atkinson, in his analysis, focused on constant- elasticity utility functions, which have the form

U 5 Y�(1 2 a)

1 2 a .

He used values of a between 1 and 2.

SOURCE: U.S. Bureau of Labor Statistics and Bureau of the Census, Annual Demographic Survey, Supplement, March 1997.

TOWARD A MORE EGALITARIAN INCOME DISTRIBUTION

Curve A represents the actual income distribution of full-time, permanent U.S. workers in 1995. Curve B represents what the income distribution might look like under a more egalitarian setting.

FIGURE 7.14

Level of income

(1000’s)

Number of workers (millions)

A B (actual) (equitable)

0 0–5 20–25 40–45 60–65 80–85 100+

12

10

8

6

4

2

Appendix: Alternative Measures of Inequality

PUBLIC EXPENDITURE THEORY

Part Three covers the basic theory of public expenditures. Chapter 8 is concerned with the government as a producer. It asks, for instance, whether there are reasons one might expect the government to be less efficient than private firms in the production of goods and services. It  also explores the trade-offs entailed in privatization, as well as alternative institutional models for public goods production such as government corporations, performance-based organizations, and public– private partnerships. Chapter 9 focuses on public choice theory in explaining how the level of expenditures on public goods is determined, with particular emphasis on the consequences of majority voting.

PART THREE

199

PUBLIC PRODUCTION OF GOODS AND SERVICES

This chapter is concerned with the role of government in production. Chapter 4 identifi ed several market failures. When there is a market failure, some form of government intervention is required. For instance, Chapter 5 explained why there will be an undersupply of public goods, and described the effi cient level of provision of public goods. But government does not have to produce these goods; all it has to do is pay for them. There are many public goods that are privately produced. There are also many private goods that are publicly produced, such as postal services and utilities.

Even though market failures provide a rationale for some form of gov- ernment intervention, they do not, by themselves, provide a rationale for government production; yet there are certain areas in which government production dominates, and others in which it is very commonly used. For instance, with few exceptions, governments have not relied on mercenary armies. In most countries, governments run the school systems, and, in almost all countries, the postal system. Until a few years ago, in most coun- tries governments also ran public utilities such as telecommunications.

Two common threads run through these examples. First, in many of these cases, competition is not viable. Remember, markets result in

8 1. What is the role of gov-ernment in production? What is the rationale behind government production of goods and services such as electric utilities and water? Why are these called natural monopolies, and what are alternative ways by which abuses of monopoly power might be prevented?

2. What are the causes of government failure? Why is production in the public sector often less effi cient than production in the private sector? What are the dangers and limits of privatization? What are the alternatives to conventional public pro- duction or purely private production of goods and services?

FOCUS QUESTIONS

200 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

effi ciency when they are competitive. Historically, only one fi rm has pro- vided postal services and one company has provided telephone services. Without government intervention one fi rm, or a few fi rms, would be able to exercise market power and exploit consumers. Governments have inter- vened in two ways. In the past, most governments have chosen to take charge of the industry directly, providing telephone services or electric- ity. As an alternative to producing the goods itself, however, government can regulate private fi rms—for example, by controlling their prices—to ensure that they do not exercise their monopoly power. In the last several years, there has been a shift away from public production toward private production with regulation. This process of privatization has been par- ticularly widespread in the market-based economies of Japan and West- ern Europe, as well as the transitional economies of Eastern Europe and Asia, especially in utilities (gas, electricity, and telecommunications) and transportation (railroads and airlines). Although some privatizations have been successful, many others have fallen far short of expectations.

The other common thread in many of the examples of government pro- duction is that the public interest has many dimensions. Will the actions of profi t-maximizing fi rms refl ect these broader public interests? There is often no simple way government can intervene to ensure that they do. This is why the government does not contract out to private fi rms to run the national defense system. It does contract out specifi c activities, such as building ships or airplanes, but it does not, for example, say to a private fi rm, “Run our defense establishment in Europe.” Similarly, some believe that schools serve a variety of social functions, which go beyond conveying skills and knowledge. They transmit national values and help form a sense of nationhood. There is concern that a private school system, as eff ective as it might be in imparting skills, might not work as eff ectively in advancing this broader set of public objectives. There has also been a resurgence of leaders, particularly in Latin America, who believe that the state can best achieve its social objectives by exerting direct control over key economic sectors. This has resulted in government takeover, or nationalization, of a large num- ber of corporations, and in some cases, of entire industries, many which had been previously privatized. For example, in 2006 Bolivian President Evo Morales nationalized fi rms in oil and gas, telecommunications, and power; and in 2007, Venezuelan President Hugo Chávez nationalized oil, cement, steel, rice processing and packaging, and supermarket businesses.1

1�Deliberate nationalization as part of a government’s development strategy should not be confused with unintended, de facto nationalization that often takes place amidst an economic crisis—for example, nationalization of some Swedish banks in 1992 and nationalization of several severely distressed fi nancial institutions in the United States in 2008, such as the Federal National Mortgage Association (Fannie Mae) and the Federal Home Mortgage Corporation (Freddie Mac). These government takeovers are meant to be temporary, as indicated by the terms and form of the takeover. There is often a timetable for the public sale of the government’s stake, and use of the conservatorship model or purchase of preferred (nonvoting) shares signal the government’s desire to refrain from engagement in the business’s day-to-day operations.

201Natural Monopoly: Public Production of Private Goods

The global fi nancial crisis of 2008 brought about another wave of nationalizations around the world, especially in fi nance, as it became clear that without government help, fi rms would collapse and unemploy- ment would increase still further. In the United States, the government took over the major mortgage fi nancing companies, the country’s larg- est insurance company, the largest auto fi nancing companies, and two of the three major auto companies. In addition, it provided massive funds for banks, although it did not take them over. (Some critics, however, say it should have.) By contrast, in many other countries, including the United Kingdom, the government took over the major banks. There is an important distinction between these nationalizations and earlier ones: for the most part, these were viewed explicitly as temporary measures to be taken in an emergency, until the private economy got back on its feet. However, when these measures were taken, it became clear how depen- dent the private sector was on the public sector.

Although market impediments (such as limited competition) and con- cerns about broader objectives provide motivations for public production, there is one compelling argument against public production: often, but not always, governments seem to be ineffi cient producers. There are exam- ples showing that government enterprises can be effi cient: Chile’s pub- licly owned copper mines have been (many believe) just as effi cient as the private ones, but provide far higher revenues to the public; Korea’s pub- licly owned steel company was one of the most effi cient in the world—far more effi cient than America’s privately owned steel companies. However, critics of government ownership suggest that they are the exceptions that prove the rule. Thus, identifying when government should engage in pro- duction and when it should use private fi rms involves a balance. To better understand the nature of the balance, we begin by taking a look at gov- ernment production of private goods, for which issues of “social values” enter in a much more limited way. The primary concern is that the private good—electricity, mail service, or telephone service—be provided in the most effi cient way at the lowest possible cost to consumers.

NATURAL MONOPOLY: PUBLIC PRODUCTION OF PRIVATE GOODS

The most important market failure that has led to public production of private goods arises when markets are not competitive. This provides at least part of the explanation for government production in postal

202 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

services, telecommunications, water, harbors, and electricity. As we saw in Chapter 4, a common reason why markets may not be competitive is the existence of increasing returns to scale; that is, the average costs of production decline as the level of production increases. In that case, eco- nomic effi ciency requires that there be a limited number of fi rms. Indus- tries for which increasing returns are so signifi cant that only one fi rm should operate in any region are referred to as natural monopolies.

Water is a good example. The major cost associated with delivering water is the network of pipes. Once pipes have been installed, the addi- tional costs of supplying water to one extra user are relatively insignifi - cant. It would clearly be ineffi cient to have two networks of pipes, side by side, one delivering to one home, the next to a neighbor’s. The same is true of electricity, cable TV, and natural gas.

THE BASIC ECONOMICS OF NATURAL MONOPOLY

The average cost curve and the demand curve for a natural monopoly are represented in Figure 8.1. Because the average costs of production decline as the level of production increases, it is effi cient to have only one fi rm. In the case depicted, there is a whole range of viable outputs for which the

NATURAL MONOPOLY

With no sunk costs and poten- tial entry, a natural monopolist

would operate at Q1, the lowest price consistent with at least

breaking even. With sunk costs, the price will be higher. The

monopolist unconcerned with the threat of entry operates at

Q*, where marginal revenue equals marginal cost.

FIGURE 8.1

Average cost Marginal

cost

Profit per unit = difference between price and average cost

Demand

Marginal revenue

Output Monopoly output

Monopolist’s profit

Price

Q*

P

Q1 Q0

203Natural Monopoly: Public Production of Private Goods

fi rm makes a profi t. The maximum viable output (without subsidies) is Q1, where the demand curve intersects the average cost curve.

In these situations, we cannot rely on the kinds of competitive forces that we discussed earlier to ensure that the industry operates at the effi - cient level. Effi ciency requires that price equal marginal cost (at quantity Q0). If the fi rm charges a price equal to marginal cost, however, it will suf- fer a loss, as marginal cost is lower than average cost for industries with declining average cost.

One common recommendation in this situation is for the government to provide a subsidy to the industry and insist that the fi rm charge a price equal to the marginal cost. This policy, however, ignores the question of how the revenues required to pay the subsidy are to be raised; in particu- lar, it assumes that there are no costs associated with raising this revenue. Moreover, it assumes that the government knows the magnitude of the subsidy that will enable the fi rm to be viable.

In practice, most governments have attempted to make such indus- tries pay for themselves. (They may also be concerned with the equity of making general taxpayers pay to subsidize a private good that is enjoyed by only a portion of the population, or enjoyed by diff erent individuals to diff erent extents.) Thus, they have insisted on government-managed nat- ural monopolies operating at the intersection of their demand curves and their average cost curves (Q1 in Figure 8.1). This is called the zero profi t point.

The zero profit point is precisely the point at which natural monopolies may operate, under the assumption that there is effec- tive potential competition. Assume a firm tried to charge a price that exceeded the average cost of production. If it were easy to enter (and exit) from an industry, a firm that tried to capture a profit for itself would instantaneously be threatened with entry by other firms willing to provide the given service or commodity at a lower price. New firms could come in and provide the services or commodities at a profitable price, without worrying unduly about the reactions of the original firm.2 Thus, the presence of a single firm in an industry does not, in itself, imply that the firm can exercise monopoly power. As long as there are potential entrants, that single firm must charge a price equal to average cost.

2��In the literature on industrial organization, markets with decreasing average costs but no sunk costs, in which price is maintained at a level equal to average costs, are referred to as contestable. See W. J. Baumol, J. Panzar, and R. Willig, Contestable Markets and the Theory of Industrial Organization (New York: Harcourt Brace Jovanovich, 1982). For a simple exposition of the theory of contestable markets, see W. J. Baumol, “Contestable Markets: An Uprising in the Theory of Industry Structure,” American Economic Review 72 (1982): 1–15.

204 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

EFFECTS OF SUNK COSTS All this changes when there are sunk costs. Sunk costs are costs that are not recoverable upon the exit of the fi rm. Most research and development expenditures represent sunk costs—but a building that can be converted costlessly for another use does not represent a sunk cost. An airplane, for example, which can easily be sold to another airline, does not represent a sunk cost.

Why are sunk costs so important? They create an essential asymmetry between a fi rm that is established in an industry and one that is not. The potential entrant is not in the same position as the fi rm already in the industry, for the fi rm already in the industry has expended funds that it cannot recover. In deciding whether to enter, a fi rm does not look at the level of current profi ts and prices, but at what prices and profi ts will be after entry. Even if prices currently are considerably above average costs (so profi ts are large), a potential entrant may well believe that the fi rm already in the industry will respond to entry not by exiting the industry but by lowering its price; at the lower price, entry no longer is profi table. Moreover, when sunk costs are signifi cant, an entrant worries that it will not be able to recover all the expenditures it makes upon entry. Thus, it will be reluctant to gamble that the current fi rm will either exit or leave its prices at their currently high levels. Accordingly, sunk costs act as a barrier to entry and allow the established fi rm a degree of monopoly power that it could not exercise otherwise.

Because virtually all natural monopolies entail important sunk costs, the government cannot simply rely on the threat of potential competition. The fact that a single fi rm controls consumers’ water or electricity gives rise to concern: the monopolist is in a position to exploit its consumers. The monopolist that is unconcerned about entry by other fi rms charges a price that maximizes its profi ts, the price at which the marginal revenue it gets from selling an additional unit is equal to the marginal costs (out- put Q* in Figure 8.1). Its profi t per unit of output is the diff erence between the price it charges and the average costs.

MULTIPRODUCT NATURAL MONOPOLIES So far, we have focused on a natural monopoly producing a single commodity. If the industry is not to be subsidized, it must charge a price in excess of marginal cost.

On what principle should prices be set when a natural monopoly pro- duces several commodities? Prices, on average, will still need to exceed marginal cost if the fi rm is to break even. Should the ratio of the price to marginal cost be the same for all the fi rm’s products? Should higher charges on some services be used to subsidize other services?

For example, the U.S. Postal Service imposes uniform charges for delivering mail, even though the marginal cost of delivering mail to a rural household in North Dakota may be much higher than the cost of

205Natural Monopoly: Public Production of Private Goods

delivering a letter in Chicago. If the post offi ce is to break even, there must be a cross-subsidy, a subsidy from one user (product) to another user (product).

The issue is obviously very political; the elimination of cross-subsidies will aff ect some groups adversely. When pricing decisions are made polit- ically, these groups will attempt to persuade those in charge to lower the prices to them, implicitly raising prices to others.

The analysis of pricing decisions involves both efficiency and dis- tributional considerations. Economists have been particularly con- cerned with the efficiency costs of politically determined pricing policies. When prices are raised on some service, the consumption of that service declines, but a 1 percent price increase reduces demand more for some goods than for other goods. Goods for which demand is more sensitive to price increases are said to have price elasticity. Figure 8.2A shows an inelastic demand curve, for which a change in price does not result in a very large change in consumption, whereas in Figure 8.2B, the demand is very elastic: a change in price results in a large change in consumption.

If a natural monopoly is to break even (without government subsidies), it obviously must charge a price in excess of marginal cost. If the govern- ment increased price above marginal cost by the same percentage for all commodities, clearly consumption of goods with elastic demand would be reduced by more than consumption of goods with inelastic demand. Under some circumstances it can be shown to be desirable to charge prices such that consumption of every good is reduced by the same percentage (from what it would be if price equaled marginal cost). If the government wishes to do this, it should increase the price (above marginal cost) more for commodities whose demand is inelastic than for commodities whose demand is elastic.3

3��This policy will minimize the deadweight loss resulting from price exceeding marginal cost. The prob- lem of how to set prices for a multicommodity public monopoly was fi rst solved in 1956 by Marcel Boi- teux, who served as the director of Electricité de France, the government agency in France responsible for producing electricity. For an English translation, see “On the Management of Public Monopolies Subject to Budgetary Constraints,” Journal of Economic Theory 3 (1971): 219–240. This question of the determination of prices for diff erent commodities turns out to be equivalent to a similar question posed some twenty-fi ve years earlier by the great British economist Frank Ramsey: If the government must raise a given amount of revenue by distortionary taxation, how should it raise the revenue? Should it, for instance, charge a uniform tax on all commodities, so the ratio of the price to marginal cost would be the same for all commodities? Would there then be no relative distortions? Ramsey showed that, as plausible as that might seem, a uniform tax was not the correct answer; it was preferable to charge a higher tax on a commodity whose demand was inelastic. Both Ramsey and Boiteux ignored the distributional issues that are central to most of the political debate. These were introduced into the analysis by M. Feld- stein, “Distributional Equity and the Optimal Structure of Public Prices,” American Economic Review 62 (1973): 32–36. In the context of taxation, see P. Diamond and J. Mirrlees, “Optimal Taxation and Pub- lic Production,” American Economic Review 61 (1971): 261–278; A. B. Atkinson and J. E. Stiglitz, “The Structure of Indirect Taxation and Economic Effi ciency,” Journal of Public Economics (1972): 97–119; and “The Design of Tax Structure: Direct versus Indirect Taxation,” Journal of Public Economics 6 (1976): 55–75. Also see Chapter 21 for more on optimal taxation.

206 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

REGULATION AND TAXATION (SUBSIDIES)

When there is a natural monopoly with sunk costs, there is a danger that the monopolist will take advantage of its position and charge a high price. One way of addressing this concern, as we have seen, is for government to take over production. However, there has been increasing concern that governments do not do a good job at managing production. Rather than attempting to produce the good itself, the government may leave production to the private sector, but regulate prices to ensure that the fi rm does not take

PRICING IN A MULTICOMMODITY

NATURAL MONOPOLY

(A) With an inelastic demand, an increase in the price above marginal cost results in a rela- tively small decline in output.

(B) With an elastic demand, an increase in price above

marginal cost results in a large decline in output.

FIGURE 8.2

Marginal cost

A

B

Demand curve (inelastic commodity)

Output

Price

Price increase

Q*2 Q*1

Marginal cost

Demand curve (elastic commodity)

Output

Price

Price increase

Q*2 Q*1

207Natural Monopoly: Public Production of Private Goods

advantage of its monopoly position. Moreover, it can use subsidies to encour- age the fi rm to provide services that might not be profi table privately but are viewed as socially desirable, such as providing postal service to rural areas.

Those who advocate regulation and subsidies (or taxes) as remedies for market failures believe that they have three major advantages over pub- lic ownership. First, they allow for a more consistent and effi cient national policy. Assume that it is desirable to locate fi rms in areas with high unem- ployment, and because of this, government-run fi rms are told to locate to such areas. However, it is better to provide a general subsidy—to fi rms for which the move has the least cost taking advantage of the subsidy—than to simply impose the burden on fi rms that happen to be government run.

Second, the utilization of tax and subsidy schemes allows a clearer esti- mate of the costs associated with pursuing a given objective. It may be desir- able to reduce the level of pollution, but how much is it worth? It may be desirable to locate a fi rm in an area where there is high unemployment, but how much is it worth? It is often diffi cult to ascertain the additional costs of government enterprises’ pursuing alternative objectives; providing direct government subsidies brings the costs more into the open and thus allows a more rational decision concerning whether the costs are worth the benefi ts.

Third, there is a widespread belief that incentives for effi ciency are greater with private fi rms, even with regulation.

In the case of natural monopolies, the United States has long relied on regulation (although there is some government production, partic- ularly of hydroelectric power and water), in contrast to Europe, which, until recently, relied on government ownership. Regulation is not without its problems: there are signifi cant costs to administering the regulations, prudent implementation requires substantial government capacity, and almost any regulatory scheme gives rise to distortions—that is, devia- tions in their behavior from what effi cient, competitive fi rms would do— as private fi rms try to maximize their profi ts, given the regulatory rules. Thus, if regulations allow a particular return on capital, there may be an incentive to invest too heavily in capital; and if they allow more gener- ous depreciation allowances for one type of capital than another, this may also distort the investment decisions.

In spite of these regulatory problems, beginning in the 1970s and 1980s, there was a major movement throughout the world toward privat- ization: selling off government-owned enterprises and subjecting them to regulation. By and large, the advantages of gains in effi ciency seemed to overwhelm any disadvantages associated with regulation.

In the subsequent years, however, understandings have become more nuanced, as the diffi culties of successful regulation and privatizations have become clearer. Special interests try to infl uence the regulatory

208 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

structure—there have been marked advances in the ability of markets to take advantage of defi - ciencies in regulations—to the detriment of the public; markets on their own often evolve in a way that undermines competition; and there have been notable failures in privatization.

Deregulation of electricity in California proved to be the quintessential failure: Enron, one of the companies that championed dereg- ulation, managed to manipulate prices, caus- ing massive shortages before it itself went bankrupt—the largest corporate bankruptcy up to that time. Once regulations were restored, the shortages, which conservatives had blamed on excessive regulation, disappeared.

Failures often have been due to irreconcil- able diff erences between public policy objec- tives and private sector priorities. Private prisons, for instance, were more interested in saving on costs than in rehabilitation, because they made more money if those released soon returned to prison. In some cases, such as the privatization of British Rail, there is a broad consensus that consumers have not been served well—prices have increased while the quality

of service has deteriorated. Some privatized airports seem more focused on selling goods than in providing facilities for effi cient and comfortable arrivals and departures.

In some cases, such as with highways in Argentina, Colombia, and Chile, privatizations cost the government much more than anticipated because contracts were continuously renegotiated and public funds were used repeatedly to bail out franchise holders. In other cases, the privatized enterprises had to receive public assistance, sometimes surreptitiously, such as the U.S. Enrichment Corporation, which enriches uranium.

Many privatizations were unsuccessful because of a failure to under- stand the potential for abuse of monopoly power. Telecom deregula- tion did not produce the competitive market that many had hoped, and although airline deregulation resulted in strong competition in some segments of the market, in others, monopolies and duopolies led to high prices. Especially in economies making the transition from communism to the market, monopoly rents were simply transferred from the public to the private sector, many times to former state enterprise managers

GOVERNMENT PRODUCTION OF PRIVATE GOODS

Examples: Postal Service, utilities

Rationale: Market failure—lack of competition associated with natural monopoly

Problems: Ineffi ciencies associated with government production

Cross-subsidization issues

Intervention of political concerns

Alternatives: Regulation

Ignoring problem (all known cures worse than disease)

Current trends: Privatization

Focusing government involvement on natural- monopoly core, encouraging competition where feasible

Public-private partnerships

209Natural Monopoly: Public Production of Private Goods

who claimed they were now private sector entrepreneurs. The transfer of monopoly rents also occurred in market-based economies—Mexico’s pri- vatization of its telecom industry helped create one of the world’s richest men. Multiple studies have demonstrated the diff erential impact of pri- vatization to insiders versus outsiders: those with political or economic connections (insiders) have gained at the expense of those without spe- cial access to information or decision makers (outsiders).

To enhance competition, it has become common in the privatiza- tion process to diff erentiate among activities—for instance, the pro- duction, transmission, and sale of electricity; the provision of tracks (railroad infrastructure); and the provision of services. Coordination among these pieces often proves diffi cult, and if one segment is a natural monopoly, such as electricity transmission (the grid), it can exert undue infl uence in the workings of the system as a whole. Performance has been undermined further by the noncompetetitive and nontransparent process by which the privatizations have been undertaken, as well as by incomplete privatization contracts, with unclear metrics for determin- ing achievement of performance benchmarks and ineff ective provisions for dispute resolution, resulting in extensive litigation. For example, pri- vate water companies have threatened to pull out if what they receive is not increased.4

NO GOVERNMENT INTERVENTION

Some economists, including Harold Demsetz of the University of California at Los Angeles and the late George Stigler of the University of Chicago, ques- tion whether it might be better in most cases to just let the private sector alone, even with natural monopoly. Monopolists are effi cient; the only prob- lem with them is that they charge too high a price, and, accordingly, produce too little. Arnold Harberger, however, in a famous calculation, estimated that the loss from monopoly pricing is relatively small (less than 3 percent of the value of output). Monopolies reduce production relative to the effi - cient level, but the resources not used by the monopolist go elsewhere in the economy. The loss is the diff erence in the marginal values of the two uses.

4�For more on privatization of social services, see R. M. Blank, “When Can Public Policy Makers Rely on Private Markets? The Eff ective Provision of Social Services,” The Economic Journal 110 (2000): 34–49. Two good studies of privatization to insiders versus outsiders are S. Estrin, J. Hanousek, E. Kocenda, and J. Svejnar, “The Eff ects of Privatization and Ownership in Transition Economies,” Journal of Eco- nomic Literature 47 (2009): 699–728; and R. Frydman, C. Gray, M. Hessel, and A. Rapaczynski, “When Does Privatization Work? The Impact of Private Ownership on Corporate Performance in the Transi- tion Economies,” The Quarterly Journal of Economics 114 (1999): 1153–1191. A good general literature review of privatization evaluations is W. Megginson and J. Netter, “From State to Market: A Survey of Empirical Studies on Privatization,” Journal of Economic Literature 39 (2001): 321–389.

210 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

The  cumulative loss in effi ciency from either regulation5 or government production, these economists believe, may be much greater.

Most economists remain skeptical. They see larger losses from natural monopolies, partly because managers in industries not sub- ject to competition (and not subject to the scru- tiny of regulators) have a tendency to become lax, and partly because, in the absence of com- petition, incentives to innovate may be limited.

Whether for these or for other reasons, popular support for eliminating regulations of natural monopolies remains limited. However, there is still considerable interest in narrowing the scope of regulation. It is now recog- nized that there is more scope for competition than had been previously realized—for instance, several fi rms now provide cellular telephone service, and there is a multitude of generators of electricity. There is still some natu- ral monopoly—only a single fi rm provides local landline telephone service or electricity to most homes. Today, regulation is focused on ensuring that there is competition when competition is viable, and that the parts of the system in which there is a natural monopoly do not abuse their monopoly power, either by leveraging their monopoly power to gain further control or by raising prices to enable high rates of return. Major reforms in telecommunications and in electricity regulation in 1996 refl ected these changed perspectives.

GOVERNMENT FAILURES

Although market failures led to the institution of major government pro- grams in the 1930s and 1960s, in the 1970s and 1980s the shortcomings of many such programs led economists and political scientists to investigate government failures. Under what conditions would government programs not work well? Were the failures of government programs accidents, or did they follow predictably from the inherent nature of governmental activity? Were there lessons to be learned for the design of programs in the future?

There are four major reasons for the systematic failures of the govern- ment to achieve its stated objectives: the government’s limited information,

5�Stigler has argued, further, that regulation may be ineff ective, as the regulated group “captures” the regulators, partly because the only people who are well informed on the highly technical matters of regulation are the regulated parties, and partly because the regulators often get lucrative jobs in the regulated industry after they leave their regulatory positions. Although the regulations may be ineff ec- tive in limiting the regulated entity’s profi t, the regulations may still be highly distortionary. In recent years, there have been problems on the other side: elected regulators may follow populist policies, driv- ing down prices to the point that the utility has little incentive to invest further. George J. Stigler, “Free Riders and Collective Action: An Appendix to Theories of Economic Regulation,” Bell Journal of Eco- nomics and Management Science 5 (Autumn 1974): 359–365; and “The Theory of Economic Regulation,” Bell Journal of Economics and Management Science 2 (Spring 1971): 3–21.

NATURAL MONOPOLY • Effi cient production entails only a single fi rm.

• Market equilibrium will be characterized by a lack of competition.

• It provides rationale for government production or regulation.

211Natural Monopoly: Public Production of Private Goods

RENT CONTROL AND AGRICULTURAL PRICE SUPPORTS: CASE STUDIES IN GOVERNMENT FAILURE

Creation of Excess Demand

In the aftermath of the Great Depression and World War II, a housing shortage developed in New York City. The failure to expand supply to keep pace with demand led to an increase in prices, as any economist would have predicted. The polit- ical response did not, however, take into account these underlying forces. When lawmakers passed rent control legislation, they failed to anticipate its full consequences, overlooking the fact that apart- ments were supplied by those who could turn else- where for better investment opportunities if the return to investments in housing fell. Advocates of rent control thus failed to anticipate that the supply of rental housing would decrease, and that the qual- ity of services provided by landlords would deteri- orate. Even though the government attempted to control this deterioration by imposing standards on landlords, these attempts were only partially suc- cessful, and indeed exacerbated the decline in the supply of rental housing. There was little the city government could do to stop this, short of repeal- ing the rent control statutes for new housing, which it eventually did, although numerous older build- ings remain under rent control. Many more remain under “rent stabilization” legislation, which controls the rate of increase in rents.

Creation of Excess Supply After World War II, Japan experienced substan- tial rice shortages. To ensure food security, the government controlled rice production and mar- keting until 1995. The result was predictable by

economists but unanticipated by Japan’s polit- ical leaders. Farmers generated large surpluses as they responded to these incentives, while per capita rice consumption dropped dramatically because of higher prices and a more diversifi ed diet. Rather than cut price supports in response to this rice glut, which would have been extremely unpopular with the ruling party’s farmer political base, the government decided to reduce rice pro- duction and maintain high rice prices. It initiated the gentan acreage reduction policy in 1970, which pays rice farmers not to grow rice by either leav- ing paddies uncultivated or by switching to crops such as wheat and soybeans (40 percent of rice paddies are now subject to acreage reduction). Rice production has been reduced but chronic sur- pluses persist; the government buys them to use for animal fodder and food aid. Rice prices have remained high but at the expense of the Japanese taxpayer, consumer, and potential rice exporter: acreage reduction has cost more than ¥7 trillion (more than $80 billion) cumulatively, excluding additional direct income support; the price of rice in Japan is much higher than the world price; and competition from imports has been minimized by imposing a 778 percent tariff. Government incen- tives have also encouraged ineffi cient part-time farmers, preventing full-time farmers from expand- ing their rice paddies to help lower their high production costs. The government is now contem- plating abandonment of its production adjustment program and targeting its income support system to full-time farmers.

212 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

its limited control over private responses to its actions, its limited control over the bureaucracy, and the limitations imposed by political processes.

1. Limited information. The consequences of many actions are compli- cated and diffi cult to foresee. The government did not anticipate the precipitous increase in expenditures on medical care by the aged that followed the adoption of the Medicare program. Often, government does not have the information required to do what it would like to do. For instance, there may be widespread agreement that the government should help the disabled, but that those who are capable of working should not get a free ride at public expense. However, limited infor- mation on the part of government may preclude it from distinguishing between those who are truly disabled and those who are pretending.

2. Limited control over private market responses. The government has only limited control over the consequences of its actions. For exam- ple, we noted earlier that the government failed to anticipate the rapid increase in health care expenditures after the adoption of the Medi- care program. One reason for this is that government did not directly control the total level of expenditures. Even when it set prices—such as for hospital care and doctors’ services—it did not control utilization rates. Under the fee-for-service system, doctors and patients deter- mine how much and what kinds of services are provided.

3. Limited control over bureaucracy. Congress and state and local legisla- tures design legislation, but delegate implementation to government agencies. An agency may spend considerable time writing detailed reg- ulations; how they are drafted is critical in determining the eff ects of the legislation. The agency may also be responsible for ensuring that the regulations are enforced. For instance, when Congress passed the Envi- ronmental Protection Act, its intent was clear—to ensure that industries did not pollute the environment. However, the technical details, such as determining the admissible level of pollutants for diff erent indus- tries, were left to the Environmental Protection Agency (EPA). During the administration of President George W. Bush, there were numerous controversies over whether the EPA had been lax in promulgating and enforcing regulations, thus subverting the intentions of Congress. Two of the most contentious EPA rulings were failure to fi nalize standards for the allowable concentration of toxic arsenic in drinking water and weak enforcement of the Clean Air Act, particularly for violations by aging electric-generating stations.6

6�For a detailed review of EPA enforcement during the administration of George W. Bush, see J. A. Mintz, “‘Treading Water’: A Preliminary Assessment of EPA Enforcement during the Bush II Administration,” Environmental Law Institute News & Analysis 34 (October 2004): 10933–10953.

213Comparison of Efficiency in the Public and Private Sectors

In many cases, the failure to carry out the intent of Congress is not deliberate, but rather a result of ambiguities in Congress’s intentions. In other cases, problems arise because bureaucrats lack appropriate incen- tives to carry out the will of Congress. For instance, in terms of future job prospects, those in charge of regulating an industry may gain more from pleasing members of the industry than from pursuing consumer interests.7

4. Limitations imposed by political processes. Even if government were perfectly informed about the consequences of all possible actions, the political process through which decisions about actions are made would raise additional diffi culties. For instance, representatives have incentives to act for the benefi t of special interest groups, if only to raise funds to fi nance increasingly expensive campaigns. The elector- ate often has a penchant to look for simple solutions to complex prob- lems; their understanding of the complex determinants of poverty, for instance, may be limited. The revolving door—through which both politicians and bureaucrats move almost seamlessly from the private to the public sector—creates obvious confl icts of interest. Too often, when in public service, they seem to be serving the private interests from which they came and to which they are likely to return.

Critics of government intervention in the economy, such as Nobel laureate Milton Friedman, best known for his advocacy of free markets while at the University of Chicago, believe that the four sources of government failure are suffi ciently important that the government should be restrained from attempting to remedy alleged or demonstrable defi ciencies in markets. Most countries, however, have devised political processes—rules and regulations governing democracy—that can enhance the performance of the public sec- tor, and that at least ameliorate the problems posed by “government failures.”

COMPARISON OF EFFICIENCY IN THE PUBLIC AND PRIVATE SECTORS

Anyone who follows the news media has encountered shocking stories of government ineffi ciency, often citing misguided procurement policies featuring $1000 toilet seats, $400 hammers, and the like. Similar accounts have been issued from the private sector, such as the quest of cigarette companies to create a smokeless, safe cigarette.

7�This view has been particularly argued by George Stigler. See, for instance, his “Theory of Regulation,” Bell Journal (Spring 1971): 3–21.s

214 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

NATIONAL PERFORMANCE REVIEW

T he problems of government ineffi ciency, particularly arising out of procurement and personnel policies, were the subject of a major review under the direction of Vice President Al Gore. A report was issued in September 1993; its basic theme was that procurement and personnel policies had to be revised substantially in the pro- cess of reinventing government. The report empha- sized greater reliance on performance measures, greater use of market incentives, and greater use of commercial procurement practices:

Our government, built around a complex cluster of monopolies, insulates both man- agers and workers from the power of incen- tives. We must force our government to put the customer fi rst by injecting the dynamics of the marketplace. The best way to deal with monopoly is to expose it to competition.

Elsewhere, the report argued that “competition is the one force that gives public agencies no choice but to improve.”

Competition is enough to force the private sector to constantly “reinvent” itself and look for

better and more effi cient ways of doing what it does. However, it is political pressures—including the pressure of a huge government defi cit—that provide the incentive to reinvent government, and it is often political pressure from special interest groups that prevents the process of reinvention from succeed- ing. There are many examples of this. For more than a hundred years, government has been responsible for meat inspection. Americans want to know that the meat that they are eating is safe. However, when the system of government inspectors was estab- lished following Upton Sinclair’s graphic description of Chicago’s stockyards,* all that was available was a system of visual inspection: Did the meat look rot- ten? Today, we know that most problems arise from microbes that are invisible to the eye, and there are advanced ways of both detecting and dealing with these microbes. However, despite the existence of systems that could do a better job at lower costs— and of private fi rms ready and able to perform these tasks—the meat inspectors’ union worries that its workers will lose their jobs, and politicians worry that the union will mount a scare campaign, so the ineffi - cient system of visual inspection survives.

*Upton Sinclair, The Jungle (New York: Doubleday, Page & Company, 1906).

The misallocation of resources attributable to America’s fi nancial sec- tor in the years before the crisis was enormous; the full costs of the crisis that it helped bring on are in the trillions—greater than any “waste” from the public sector.

Comparing effi ciency of public and private enterprises is typically dif- fi cult, because they often work in diff erent parts of the economy and face diff erent constraints: public post offi ces have an obligation to serve every part of the country, while private companies can select the most profi t- able ones to serve. When they do produce similar commodities like edu- cation, it is hard to measure both the inputs (the quality of students) and the outputs (tests adequately capture only some dimensions of student

215Comparison of Efficiency in the Public and Private Sectors

achievement; creativity and citizenship values, both of which schools strive to promote, are typically left out). The result is that hard evidence on government ineffi ciency is diffi cult to come by.

Although the weight of the available evidence, both in the United States and abroad, suggests that government enterprises are less effi cient than their private counterparts, some evidence shows that this need not be the case. The French-run public enterprises have long been held up as models of effi ciency. For instance, the French electricity company developed a single design of a nuclear power plant, which they replicated throughout the country, pushing costs signifi cantly below those associated with U.S. nuclear power plants, where there used to be many diff erent designs. Within the United States, state-run liquor stores charge prices that are 4 to 11 percent lower than those charged by private retailers.8 Administrative costs of the Social Security Administration are less than 1 percent of the benefi ts paid, but private insur- ance companies frequently spend as much as 30 to 40 percent of the amount provided in benefi ts in administrative and sales costs.

Admittedly, it is diffi cult to measure the productivity of many govern- ment workers, those who are engaged in administrative activity; there is no good measure of their output. However, there are some indirect indicators. Since 1992, the number of government public employees in the United States has been brought down dramatically, to the level of the early 1960s, and as a percentage of the civilian labor force, to a level comparable to the early 1930s; in the same period, there has been a huge increase in government service and an increase in the populations served—suggesting an increase in productivity.9

The U.S. Postal Service provides an example in the United States of how diffi cult it is to draw general conclusions. In areas of direct compe- tition, such as overnight mail and parcel post, the Postal Service has not fared well in recent decades. On the other hand, in its main line of busi- ness, it has shown remarkable increases in productivity in the past twenty years—three times the pace of business sector productivity.

Of all the comparative productivity studies, one of the most telling was that between private and public railroads in Canada. One of the two major Canadian rail systems is private, the other public. The study con- cluded that there was no signifi cant diff erence in the effi ciency of the two systems. Evidently, competition between the two provides strong incen- tives for effi ciency in both.10

8�Sam Peltzman, “Pricing in Public and Private Enterprises: Electric Utilities in the United States,” Journal of Law and Economics 14, no. 1 (1971): 109–147. This article discusses both electric utilities and liquor stores. 9�Prior to the 1990s, productivity growth in the government may also have been greater than in the private sector. See Nancy Hayward and George Kuper’s detailed study over the period of 1967 to 1978, “The National Economy and Productivity in Government,” Public Administration Review 38 (1978); and U.S. Offi ce of Personnel Management, Measuring Federal Productivity, February 1980. 10�D. W. Daves and L. R. Christensen, “The Relative Effi ciency of Public and Private Firms in a Compet- itive Environment: The Case of Canadian Railroads,” Journal of Political Economy 88 (1980): 958–976.

216 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

SOURCES OF INEFFICIENCY IN THE PUBLIC SECTOR

There are several reasons why we might expect public enterprises to be systematically less effi cient than private enterprises. These have to do with incentives and restrictions at both the individual and organizational levels.

ORGANIZATIONAL DIFFERENCES

ORGANIZATIONAL INCENTIVES Because public enterprises are not driven by the profi t motive, they have little incentive to maximize pro- ductivity. Indeed, they are often driven by political concerns that work against productivity, such as providing jobs, especially in regions where there is an unemployment problem. In some countries, public enterprises may not even worry about taking a loss, as they cannot go bankrupt, and any losses are made up out of government revenues. They face, in other words, a soft budget constraint, and they often operate in an environ- ment with limited competition.

PERSONNEL RESTRICTIONS Worry that public employees might abuse their position and power—to the detriment of taxpayers who might have to pay more for the services than they should—has resulted in the imposition of numerous constraints. Private fi rms can hire whomever they like and pay whatever salaries they like. The owners suff er if a fi rm pays someone more than they are worth, while the taxpayer suff ers if a govern- ment agency pays someone more than they are worth. We fi nd it particularly objectionable when a public offi cial does not act fairly; equity is an essential part of public trust. Thus, we have imposed strong civil service rules, which are designed to ensure that the government hires and promotes the most qualifi ed individuals and that their pay is appropriate. Although such rules serve an important function, they introduce rigidities: it is diffi cult for a gov- ernment agency to fi re an incompetent worker, and this attenuates incen- tives. It is diffi cult for the government to compete with private companies for the best brains; these often command a high wage premium, well beyond the civil service scales for someone with the same qualifi cations.

PROCUREMENT RESTRICTIONS Similarly, to prevent abuses in the government’s purchase of billions of dollars of goods and services every year, procedures have been designed to ensure that the government is not taken for a ride, but their eff ect is often to raise costs. In buying a jet engine, for instance, these procedures reportedly result in cost increases of as much

217Sources of Inefficiency in the Public Sector

as one-third. In many areas, the government insists on competitive bidding. But to do this, the government must specify in minute detail what it is pur- chasing. A T-shirt may take thirty pages of fi ne print, detailing the quality of thread, the shape, and so on. Because the specifi cations required by the gov- ernment typically diff er in several ways from those the T-shirt companies make for the private market, fi rms will have to have separate production runs to meet the government specifi cations. Relatively few companies will fi nd complying with all the government regulations worthwhile; competi- tion will thus be restricted, and the prices bid will refl ect these high costs of complying with the government specifi cations and regulations. As a result, the government may end up paying substantially more than it would have to pay for comparable products already available to the public. Off -the-shelf purchasing—a well-informed consumer using the discipline of the market combined with product testing—may save considerable money. Procure- ment reform enacted in 1994 along these lines has saved billions of dollars.11

BUDGETING RESTRICTIONS Another way government agencies dif- fer from private fi rms is in budgeting, particularly in making long-term investments. It took the airlines many years and hundreds of millions of dollars to develop their airline reservation systems, but they could easily budget for what they knew was an important capital expenditure. The air traffi c control system—which makes sure that airplanes do not crash into each other—is run by the Federal Aviation Administration. Keeping up with the immense increase in air traffi c and updating the controllers’ obsolete computers will require investments of billions of dollars. Con- gress makes appropriations only on an annual basis, however, and given the tight budgetary situation, it never appropriates enough to fund fully the rapid modernization. It does not look at the matter as a business would, assessing the return on the investment.

INDIVIDUAL DIFFERENCES

Many of these organizational diff erences have immediate impacts on individuals. Because individuals cannot be fi red, and cannot be rewarded for good performance with the kinds of bonuses that private fi rms pay, there are neither the carrots nor the sticks to provide as strong individual incentives. Because public agencies have less incentive for effi ciency or

11�For instance, within two years of the reform’s passage, the Defense Department reported saving $4.7 billion from the new procurement programs (U.S. Department of Defense, “Defense Acquisition Pilot Programs Forecast Cost/Schedule Savings of Up to 50 Percent from Acquisition Reform,” News Release No. 138–96, March 14, 1996).

218 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

for ensuring that they are attentive to their “customers,” what powers to provide incentives they do have often are not directed at those objectives, but rather at more political goals.

There is a whole set of traits that are normally associated with bureaucratic behavior. Bureaucrats may not receive larger paychecks or bigger dividends from increased effi ciency, but they often seem to act as if they enjoy the power and prestige associated with being in charge of a larger  organization. They thus may try to maximize the size of their bureaucracy; if the demand curve for their services has less than uni- tary elasticity, by reducing effi ciency—increasing the price per unit ser- vice provided—they actually can increase total expenditures on their agency, and its size (see Figure 8.3). What stops bureaucrats from doing this is competition—competition between bureaucracies. W.A. Niskanen, a member of the Council of Economic Advisers in the Reagan admin- istration and chairman emeritus of the Cato Institute, a conservative think tank, has argued that the increasing centralization of government bureaucracies—in the attempt to ensure that two government agencies do not perform duplicative functions—though intended to enhance effi ciency, has reduced competition, thereby giving bureaucrats more scope to pur- sue their interests at the expense of effi ciency and the public interest.

The problem to which Niskanen called attention—that government bureaucrats may act in their own interests, and not necessarily in the inter- ests of the citizens whom they are supposed to serve—is an example of a general class of problems called principal–agent problems. The principal– agent problem is simply the familiar problem of how one person gets another to do what he or she wants. Here, the problem is, how do citizens (the “principals”) get their employees, public servants (the “agents”), to act in their interests? The analogous problem in the private sector is, how do shareholders (the principals) get their employees, the managers and workers in the fi rms they own (the agents), to act in the shareholders’ interests?

Principal–agent problems arise in all organizations, whether public or private. Managers always face the problem of ensuring that their employ- ees’ behavior conforms with their wishes; unless the fi rm is owned by managers, the owners always have a problem in ensuring that managers act in their interests. The problems of controlling employees are particu- larly acute in large organizations, and the problem of controlling manag- ers is perhaps as serious in a large corporation in which there is no large shareholder as it is in a government enterprise. When British Petroleum was government owned, it acted little diff erently from any other large oil company with diff use ownership, such as Texaco (although in some ways, both may have acted diff erently from a large oil company controlled by a single family, such as Getty). What diff erence does it make whether there are private shareholders or a single shareholder—the government? Some

219Sources of Inefficiency in the Public Sector

contend that managers of public enterprises may behave in much the same way that managers of large private enterprises do. In both cases, manag- ers have a large amount of discretion, allowing them to pursue their own interests, often at the expense of the public interest (in the case of public enterprises) or shareholder interests (in the case of private enterprises). Payoff s, called “green-mail,” out of the corporate purse to those attempt- ing to take over a fi rm, have confi rmed these views; such payments, as well as the provisions that management has attempted to put into their corporate charters making takeovers more diffi cult, have preserved man- agement’s prerogatives, but at the shareholders’ expense.

MAXIMIZING A BUREAUCRACY’S SIZE

(A) Shows a demand curve for a bureaucracy’s service. As the price of the service (cost per unit) declines, the quantity of services demanded increases. (B) The bureaucrat can calculate the total expenditures—price times quantity—at each level of price. In the absence of competition, the bureaucrat can choose the price—and the bureaucrat will choose the price that maximizes total expenditures (the size of the bureaucracy). At p*, expenditures are maximized.

FIGURE 8.3

Demand for bureaucracy’s

service

Price

p*

Q*

Total expenditure

Price

p* E

Quantity of services

provided

A

B

220 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

In large organizations, principal–agent problems are never fully resolved. Incentive structures—rewards for “good” performance (often fi nancial rewards) and punishments for “bad” performance (being fi red)—represent the most eff ective ways of aligning incentives. Although both public and private sectors face problems in designing incentives that fully resolve agency problems, these problems seem more formidable within the public sector, partly because of restrictions on how public agencies can compensate their employees.

BUREAUCRATIC PROCEDURES AND RISK AVERSION

Bureaucrats’ desire to increase the size of their budgets seems to provide an explanation of many aspects of bureaucratic behavior. Other aspects of bureaucratic behavior can best be explained by another important aspect of the incentives bureaucrats are given. Even though bureaucrats’ pay may not be closely and directly related to their performance, in the long run,

their promotion depends at least partially on observed performance. Bureaucrats can absolve themselves of responsibility for mistakes by

following certain bureaucratic procedures that ensure that all their actions are reviewed by others. Although this process of group decision making also reduces the claims an individual can make for any success, bureau- crats seem willing to make this trade-off . We say they are risk averse. This is what gives rise, in part, to the nature of bureaucracies: everything must pass through the appropriate channels (red tape). The emphasis is on procedural compliance rather than quality of results.

Two other factors contribute to the prevalence of bureaucratic pro- cedures. First, many of the costs associated with engaging in risk-averse activities are not borne by bureaucrats themselves. Rather, they are borne by society as a whole, through the taxes required to pay the extra person- nel. Further costs are imposed on those dealing with the bureaucracy in the form of delays, paperwork, and so on. (Indeed, there are those who claim that bureaucrats may actually enjoy the bureaucratic process.)12

12�There are alternative theories (more psychologically or sociologically based) as to why bureaucrats behave bureaucratically.

EXPLANATIONS OF INEFFICIENCY IN THE PUBLIC SECTOR Organizational differences

• Soft budget constraints (government subsidies, no bankruptcy)

• Role of political concerns

• Absence of competition

• Additional restrictions

On personnel (civil service—hard to fi re, pay competitive wages)

On procurement On budgeting (hard to do long-term budgeting

required for large capital investments)

Individual differences

• Absence of incentive pay, diffi culty of fi ring reduces incentives (removes carrots and sticks)

• Principal–agent problems

Pursuit of bureaucratic objectives—maximizing size of organization

Excessively high levels of risk aversion—leading to a focus on following procedures (red tape)

221Corporatization

Second, the prevalence of set routines that must be followed entailing the approval of any proposal by several diff erent individuals has a posi- tive aspect as well; it is not just a consequence of bureaucrats’ pursuing their own self-interest. It follows naturally from the fi duciary relation- ship between government bureaucrats and the funds they allocate. That is, government bureaucrats are not spending their own money; they are spending public resources. It is generally accepted that an individual should have more discretion—should take greater care—in spending the money of others than in spending his or her own money. Again, what is implied by taking greater care is that certain routines are fol- lowed; these ensure that the funds are spent not according to the whim of any single individual. They also reduce the possibility of corruption. Because many individuals must give approval, it is usually not within the power of any individual to give a contract at an above-market price and thus receive a kickback.

It should be clear, however, that both government and large corpo- rations rely on bureacratic procedures. There may thus be a greater dif- ference between small enterprises and large ones, than between publicly owned and privately owned large enterprises.

Two examples of bureaucratic routines commonly used in the pub- lic sector are the use of cost–benefi t analyses and environmental impact statements. The intent behind such procedures is clear. On the other hand, because the data on which an assessment can be fi rmly established are rarely available, the studies often become pro forma exercises with predictable outcomes. Occasionally they serve as the basis for attempts by opponents of a project to delay the project and thereby to increase the costs to the point at which the project is no longer economically feasible. There is a social loss in these delaying practices.

CORPORATIZATION

The last section explored a number of the reasons why government enter- prises are often less effi cient than their public counterparts. We have seen that a host of regulations and restrictions, on hiring and promotion, on procurement, and on budgeting for long-term investment—for all of which there may be good reasons—inhibit effi ciency. But enterprises do not have to be turned over to the private sector—privatized—to address these concerns. The United States and other countries have experimented with forms of organization that lie between conventional public agen- cies  and private companies, including government corporations and performance-based organizations.

222 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

There are many examples of government corporations. Among the most widely known are the Postal Service; the Tennessee Valley Authority, the nation’s largest producer of electricity; the Saint Lawrence Seaway; and, until it was privatized in 1998, the U.S. Enrichment Corporation (USEC), which converts natural uranium into enriched uranium used in nuclear reactors and atomic bombs. These corporations are owned by the U.S. government, so typically their boards of directors and president are appointed by the President; however, they are intended to be apoliti- cal, with terms of offi ce that do not necessarily coincide with that of the president. Like ordinary fi rms, they raise their revenues by selling their products or services. Most importantly, they are freed from most of the restrictions imposed on government agencies—they can borrow and lend, and they have considerable discretion in pay and procurement. By and large, they act much like private corporations.

Typically, before a government enterprise is privatized, it goes through the intermediate stage of corporatization. This often entails moving the enterprise off budget to increase transparency and accountability. Rather than commingling revenue and expenses with general departmental fi nances, the enterprise produces its own fi nancial statements to better refl ect its true fi nancial position. Subsidies for loss-making government enterprises are also more transparent, as they are explicit on-budget expenditures. Most of the effi ciency gains seem to occur in this stage, although there is controversy as to why. Some argue that the freedom from government personnel, procurement, and budget restrictions is all that is required; under corporatization, eff ective incentive schemes can be put into place. Others argue that without the profi t motive—derived from pri- vate ownership—these gains could not be sustained. Often the managers of government enterprises do well after privatization, becoming highly paid executives in the new private company and/or receiving hefty shares or options in the newly privatized company—and it is these economic returns that drive them to improve effi ciency during the corporatization stage.

Performance-based organizations (PBOs) are government agencies that remain more fi rmly within the public sector, but in which agency offi cials are rewarded on the basis of performance. In the United King- dom the Patent Offi ce became a PBO, and a similar proposal has been made in the United States.

Many countries are involved in a lively debate: When should a govern- ment agency be privatized, corporatized, or converted into a PBO? In all three cases, the agency’s output must be measurable. In the case of privat- ization or corporatization, the agency has to produce a product that can be sold, or there at least has to be a source of revenue related to its activities: the TVA sells electricity; the Postal Service, stamps. Many countries have

223Corporatization

privatized their air traffi c control systems, and the Clinton administra- tion proposed corporatizing it; it could easily be fi nanced either through a ticket tax or through landing or takeoff fees.

Why not privatize everything, or at least everything for which a charge could be imposed? The answer is that there are public objectives that may not be well addressed by a private fi rm, and that cannot be well addressed through regulation. One might not want a profi t-maximizing fi rm in the business of producing enriched uranium that could be used to make atomic bombs—although, curiously, the Bush administration proposed doing precisely that, and the Clinton administration actually did it in 1998. If the Postal Service were privatized, concern would arise that it  might raise prices in an attempt to exercise its monopoly power and that it might not serve rural areas as well. (Although the government might require that it serve those areas, stipulating how it should do so, and then monitoring the quality of the service, might be diffi cult.)

Issues of public interest are often complex, hard to measure, and dif- fi cult to refl ect adequately in the design of PBOs. One wants the Patent Offi ce to be not only fast, but also accurate; that is, one does not want it to deny patents that should be granted, or to grant patents that will be over- turned by the courts. Unfortunately, the appeal process often takes years; accordingly, it may take years to know how good a job the Patent Offi ce has done. Moreover, there are fundamental public issues associated with, for instance, the scope of a patent: Should a patent for a new genetically altered tomato be limited to that tomato breed, or to all genetically altered tomatoes, or to all genetically altered plants? Typically, such decisions are made now on a case-by-case basis. There is concern that if the Patent Offi ce became a PBO, it might not make these decisions in a way that best refl ects the national interest.

Privatization has been pushed even further at the subnational level. Some states already have privatized prisons (see case study, “Privatizing Prisons”), and Texas proposed privatizing the administration of welfare programs. The federal government said that Texas could not do so, on the grounds that the private welfare agency might have an incentive to deny benefi ts to those who were really eligible for benefi ts. Critics of the federal government’s decision argue that an incentive scheme could be found that would address these concerns—for instance, by imposing a suffi ciently large fi ne on the private welfare agency when its denial of benefi ts was overturned upon appeal. Texas subsequently privatized the state’s food stamp eligibility program in 2005, resulting in what the U.S. Department of Agriculture called an unwarranted fi ve-year slide in processing food- stamp applicants that left thousands of people without benefi ts for which they were eligible.

224 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

PRIVATIZING PRISONS

P risons may seem among the least likely can-didates for privatization. After all, a for-profi t company would have every incentive to reduce services, such as the quality of food, to the bare minimum; the “customers” might complain, but they really have nowhere else to go. There is also a potential confl ict of interest in turning over prison construction and management to corpo- rations whose business model relies on growing prison populations to ensure high occupancy rates in their existing and prospective private prisons. Indeed, although private prisons are not new— they date back to colonial times—they have had a somewhat spotty history. Privatization of prisons has grown dramatically in the United States over the past twenty-fi ve years, so that by the end of 2009, more than 120,000 inmates were held in privately owned or operated federal or state prisons.

It is not surprising that these private prisons have been the subject of considerable controversy. Some claim that costs have been contained, even as services, such as educational training and drug

treatment programs, are improved. The leading corporation, Corrections Corporation of America, has done this partly by “reinventing” prisons—for instance, redesigning them so that fewer guards can supervise more people. They also claim that with less tension in the prison, not only are prison- ers better off, but fewer guards are required, and absenteeism rates are far lower than in conventional prisons. Critics, however, argue otherwise: that costs are actually higher, violence is increased, and that there is a lower rate of rehabilitation.

Of particular concern is the perverse incentive to which private prisons give rise: they benefi t from having legal frameworks with long mandatory jail sentences even for minor infractions.

Despite current levels of overcrowding and pro- jected continued growth in the prison population, severe fi scal constraints will limit the construction of new public institutions. In spite of the controversy over private prisons, it is anticipated that the short- fall will be met largely by privately built or managed prisons.

SOURCES: Charles Thomas, Dianne Bolinger, and John Badalamenti, “Private Adult Correctional Facility Census,” Private Corrections Proj- ect, Center for Studies in Criminology and Law, University of Florida, March 1997, p. iv; Management & Training Corporation, “Privatization in Corrections: Increased Performance and Accountability Is Leading to Expansion,” MTC Institute, December 2009; Lauren E. Glaze, “Cor- rectional Populations in the United States, 2009,” Bureau of Justice Statistics Bulletin (December 2010); and American Civil Liberties Union, “Banking on Bondage: Private Prisons and Mass Incarceration,” November 2, 2011.

Public–private partnerships (PPPs) have evolved over the past two decades as an alternative to purely public, purely private, or corpo- ratized public provision of infrastructure and services, particularly in the power, water, and transportation sectors. PPPs try to allocate risks and rewards between the public and private sectors based on respective mis- sions and capabilities. For example, governments often have land, or the authority to assemble land (when diff erent parcels are owned by diff erent individuals), but are typically short of the resources required to design, fi nance, construct, and operate extremely large and complex infrastruc- ture improvements, commonly refered to as megaprojects. In contrast,

225A Growing Consensus on Government’s Role in Production

private fi rms do not have the legal or regulatory authority of government, but can mobilize large pools of funds and expertise.

The most common types of PPPs range from relatively simple service, management, and lease contracts for public services, such as operation of public sports and recreation venues, to very intricate fi nancing and operational arrangements for large-scale public infrastructure such as build–operate–transfer (BOT) concessions and public–private joint ven- tures. The PPP model has been used around the world for the provision and operation of power plants, toll roads, and health facilities, as well as for the upgrading of muncipal water, sewage, and solid waste systems.

PPPs have produced mixed results. In some electricity projects, the prices paid by consumers have been exorbitant. In some cases, the part- nership seems to involve the government bearing the downside risk and the private party getting the upside profi ts, with the private party threat- ening to withdraw if its returns turn out too small. The weaker the check of the market on abuses, the less transparent the bidding process, the greater the chances that the PPP will not succeed.13

A GROWING CONSENSUS ON GOVERNMENT’S ROLE IN PRODUCTION

The role of the government in production will remain an area of active debate. Today, there is growing consensus that government should not be involved in the production of ordinary private goods, and there is a consensus that gov- ernment should not privatize national defense (although the United States did hire mercenaries during the Revolutionary War). Even in these areas, however, government can purchase many goods and services from private contractors. It has long purchased airplanes and tanks, but it is turning more to private fi rms for housing and other logistical support services. There also has been increasing reliance (especially during the Iraq and Afghanistan confl icts) on contractors such as Xe Services LLC (formally Blackwater USA and Blackwater Worldwide) for broadly defi ned security services, sometimes blurring the line between combat and noncombat support.

Some of the arguments against further privatization are political. Opponents of privatizing the TVA and the other government hydroelectric

13�For an analysis of eff orts to evaluate PPPs, particularly in respect to PPP eff ectiveness and value for money, see G. Hodge and C. Greve, “PPPs: The Passage of Time Permits a Sober Refl ection,” Economic Aff airs 29 (2009): 33–39.

226 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

projects fear that prices would rise to market level, and a hidden subsidy would be elimi- nated. Others oppose based on the principle of “if it’s not broken, don’t fi x it”: many of these government enterprises seem to be reasonably effi cient, and there is no clear case that privat- ization would improve effi ciency. Opposition to privatizing—or even corporatizing—the air traffi c control system arises from owners of corporate jets and from general aviation, which currently receive these services with- out paying their fair share of the costs—they receive a subsidy estimated at as much as $2 billion a year.

In other cases, however, there is real con- cern that with privatization, the broader range of public objectives would not be pursued—and it is in this arena that the debate is likely to be most lively: Should prisons, welfare agencies, schools, or the production of the material to make nuclear bombs be privatized? Or can most

of the gains in effi ciency be achieved simply by corporatizing, rewarding performance, and encouraging public agencies to think about those they serve as customers and clients?

To what extent can government, by imitating the private sector—for instance, by making more extensive use of incentive pay—achieve compara- ble effi ciencies? And to what extent are observed ineffi ciencies in the public sector an inherent consequence of the nature of what the public sector does? It is, for instance, diffi cult to measure performance in many areas of admin- istrative work; certainly, one does not want to measure performance by the number of pages of paper produced or processed. In many areas, there can- not be competition, or competition might be feasible but undesirable. Do we want two competing armed forces? Two competing judicial systems? To be sure, in some instances competition can be introduced, especially when the government has maintained a state monopoly—the examples in broadcast- ing, in telephone service, in health, and in education have been promising. Even after competition is introduced, however, there remain vexing ques- tions: How do we ensure that the public and private fi rms are on a level playing fi eld, that there are not hidden subsidies in the government’s oper- ation? The recent fi nancial crisis brought out hidden subsidies even within the private sector: the large (“too big to fail”) banks can obtain capital at a low rate, partly because providers of capital feel that it is very likely that, in the event of troubles, they will be bailed out.

ALTERNATIVE ORGANIZATIONAL FORMS Private fi rms

Government corporations

Performance-based organizations

Conventional government agencies

Public-private partnerships

• Criteria for privatization:

Source of revenue (related to its functions) Possible to deal with “externalities” and other

public interest issues such as safety and abuse of monopoly power in satisfactory manner, as through regulation

• Criteria for performance-based organization:

Possible to measure performance Possible to deal with public interest issues in

satisfactory manner

227Review and Practice

There are other concerns besides effi ciency. Although the Social Secu- rity system has lower transactions costs than private companies provid- ing annuities, this public system also provides a product that the private sector does not—insurance against infl ation. In many areas of fi nance, it has been hard to ensure that the private sector provides the products that help ordinary families manage the risks they face, and that the sector not take advantage of those who are fi nancially less sophisticated. Some call for increased regulation, whereas others believe that government enter- prises can, and have, provided fi nancial products that are less exploitive and more eff ective in meeting needs of ordinary citizens.

Thus, although the presumption remains that for ordinary commodi- ties, the private sector should be relied on, there are a wide range of goods and services for which the debate about how best to provide these services rages: Should there be more reliance on private provision, with strong regulation to prevent abuses, knowing full well that the regulation will be imperfect and may itself introduce distortions? Should there be more reliance on public provision? Should we turn to hybrid models, PPPs, or corporatizations? Each institutional arrangement has its successes and failures. Today, the challenge is to fi nd the best institutional arrangement for a given country in a given instance, at a given moment of time.

SUMMARY

1. In the United States, the government has played an important role in production in several sec- tors, although its role is far more limited than in most other countries. Market failures provide an explanation for government intervention, but not an explanation for government production.

2. Natural monopolies are industries in which it is effi cient to have only one producer; in such situations, there is unlikely to be eff ective com- petition in the market equilibrium. This lack of competition provides a major reason for govern- ment production of private goods. An alternative to public production is government regulation. Whereas the United States has, for the most part, addressed the problem of natural monopolies

through regulation, until recently, in Europe, telephone and other natural monopolies were government controlled. More recently, many have been privatized, with mixed results.

3. There is some limited evidence that governments, on average, are less effi cient than private enter- prises in providing comparable services. There are important exceptions, though, suggesting that government enterprises are not necessarily less effi cient than their private counterparts.

4. The government, however, is not necessarily the solution to private sector failures. The failure of many public programs can be attributed to four factors:

a. The consequences of any action by the govern- ment are complicated and diffi cult to foresee.

REVIEW AND PRACTICE

228 CHAPTER 8 PUBLIC PRODUCTION OF GOODS AND SERVICES

b. The government has only limited control over these consequences.

c. Those who design legislation have only limited control over the implementation of the gov- ernment programs.

d. Politicians may act to further special private interests; more generally, political processes are complicated and need not yield effi cient outcomes.

5. Government enterprises diff er from private enterprises in several respects: whereas private enterprises maximize profi ts, government enter- prises may pursue other objectives; government enterprises often face soft budget constraints and limited competition; and they face additional constraints, in personnel policy (pay and fi ring), procurement, and budgeting. Although there may be good reasons for these restrictions, they nonetheless interfere with economic effi ciency.

6. These diff erences lead to diff erences in individ- ual incentives. Bureaucrats often try to maximize the size of their organization and to avoid risk.

At the same time, public organizations share with private fi rms the principal–agent problem, the problem of ensuring that their employees act in the interests of the organization, or more broadly, that managers and workers in fi rms act in ways that are congruent with the interests of shareholders, and that public servants act in ways that are congruent with the interests of citizens.

7. A number of organizational forms lie between conventional public agencies and private corpo- rations, including government corporations and performance-based organizations. They may be able to achieve many of the effi ciency benefi ts of private organizations, and at the same time pur- sue public interests more eff ectively than purely private fi rms subject to regulations. Much of the debate in the future will be about the extent of utilization of these organizational forms and whether private fi rms should enter into areas that were previously thought of as core govern- ment functions, such as prisons and social ser- vices. A hybrid public–private partnership model has emerged as a way to take advantage of the

relative strengths of government and business while mitigating their respective weaknesses.

KEY CONCEPTS

Bureaucracy

Corporatization

Cross-subsidy

Government corporations

Government failures

Nationalization

Natural monopolies

Performance-based organizations

Price elasticity

Principal–agent problems

Privatization

Public–private partnerships (PPPs)

Regulation

Risk averse

Soft budget constraint

Sunk costs

Zero profi t point

QUESTIONS AND PROBLEMS

1. In the past three decades there has been exten- sive privatization of public enterprises. The U.S. government sold Conrail, the French government sold off many of its banks, and the British gov- ernment partly sold off its telephone services. In each of these cases, outline the major arguments in favor of and against privatization. Do you feel diff erently about the three cases? Why?

2. Under the Reagan administration, the Interior Department greatly increased the rate at which it leased off shore oil and gas. This had the eff ect of signifi cantly reducing the prices that the gov- ernment received for these leases. (Although the leases are sold by auction, there was only a single bidder for more than two-thirds of the tracts.) Discuss the distributional and effi ciency conse- quences of this policy.

229Review and Practice

3. The Postal Service claims that one reason why it cannot provide services as cheaply as private fi rms is that it is required to provide services to rural areas but cannot charge them more than the urban areas. The private companies “skim” the low-cost markets. (Eff ectively, urban areas are subsidizing rural areas.) Discuss the effi - ciency and equity consequences of this kind of cross-subsidization.

Some have argued that if it is desirable as a matter of national policy to subsidize rural post offi ces, the subsidies should be paid out of general tax revenue, not by the other users of the postal system. Discuss the advantages and problems of such an alternative subsidy scheme.

4. There are many private security fi rms, and many large housing developments have police protec- tion provided by such private fi rms. Few towns, though, contract out their police departments. Why do you think this is so? What would be the advantages and disadvantages of doing so? Recently, however, many communities have con- tracted to have their prisons run by private fi rms. What advantages or problems might you antici- pate from this?

5. The military buys most of its equipment from private contractors but does not use private con- tractors to staff its ships or fl y its airplanes. What diff erences in the nature of the services provided might account for these diff erences?

6. There have been recurrent proposals for educa- tion voucher schemes, in which the government provides a voucher that can be used to pur- chase education from either a public provider

(the  local  town) or a private provider. The GI Bill eff ectively provided such vouchers for veter- ans of the Korean War and World War II. In the 1996 presidential campaign, Senator Robert Dole proposed that the federal government fi nance a limited number of such vouchers. What do you see as the advantages and disadvantages of these voucher schemes? Are there some circumstances (some kinds of educational services) for which vouchers seem more attractive?

7. Discuss what organizational form (e.g., private fi rm, government corporation, or normal govern- ment production) you think might be appropriate for each of the following. In each case, discuss problems of designing appropriate incentives and eff ective regulatory systems.

a. Public housing

b. Production of enriched uranium to be used in atomic bombs or nuclear power plants

c. Production of helium, sometimes used by the government for military purposes

d. Air traffi c control system

e. The Patent and Copyright Offi ce

f. Prisons

g. Job placement services

h. Administering the welfare program

i. Administering the food stamp program

8. Recently, there have been several proposals to privatize the Social Security system; some coun- tries have actually privatized part or all of their social security systems. What arguments might be put in favor or against doing this?

230

PUBLIC CHOICE

9

Unlike expenditures on conventional private goods, which are deter- mined through the price system, expenditures on public goods are deter- mined through a political process. This chapter examines some models of that political process, bringing us to the border between political science and economics.

PUBLIC MECHANISMS FOR ALLOCATING RESOURCES

Chapter 3 explained how the market economy depends on the price sys- tem to arrive at effi cient resource allocations in the production of private goods. The price system provides incentives for fi rms to produce goods that are valued and a basis for allocating among consumers the goods that are produced. We often speak of the important role prices play in convey- ing information, from consumers to producers concerning the value they

231Public Mechanisms for Allocating Resources

attach to diff erent commodities, and between producers and from pro- ducers to consumers concerning the costs of production and the scarcity of these commodities.

Equilibrium in private markets is determined at the intersection of the demand curve and the supply curve. When, for one reason or another, the demand for a commodity increases, its demand curve shifts up, the price rises, and this induces fi rms to produce more. Thus, the price system con- veys information about a change in individuals’ tastes to fi rms. Similarly, when, for one reason or another, a commodity becomes less costly to pro- duce, its supply curve will shift down, the price will fall, and individuals will be induced to shift their consumption toward the now-cheaper com- modity. Again, the price system has conveyed information about the change in technology from fi rms to consumers. Indeed, one of the central results of modern welfare economics, as was pointed out in Chapter 3, is that in a competitive economy, the resulting resource allocations are effi cient.

Decisions about resource allocations in the public sector are made in quite a diff erent manner. Individuals vote for elected representatives; these elected representatives, in turn, vote for a public budget; and the money itself is spent by a variety of administrative agencies. Thus, there is a major diff erence between how an individual decides to spend his or her own money and how, for example, Congress decides to spend the public’s money. The vote of a member of Congress is supposed to refl ect the views of constituents, not just the representative’s own views. In deciding how to vote, members of Congress face two problems: fi rst, they must ascertain the views of their constituents, and second, because these views are likely to diff er, they must decide how much weight to assign to various positions.

THE PROBLEM OF PREFERENCE REVELATION

Individuals may express their views about the desirability of one private good versus another by a simple action—by buying the good or not—but there is no comparably eff ective way that individuals can express their views about the desirability of one public good versus another. Elections of public offi cials convey only limited information about voters’ attitudes toward specifi c public goods; at best, they convey a general notion that voters prefer more or less government spending.

Even if individuals were asked directly about their preferences, would they truthfully and meaningfully reveal them? In recent years, politicians have increasingly turned to polls to ascertain voters’ preferences, and although they have found the polls to be useful, they have come to treat the results with extreme caution. For instance, in today’s environment of

1. In what ways does collective decision making, such as determining the level of public goods, dif- fer from standard decision making within a house- hold? What is the problem of eliciting preferences? When individuals diff er in what they want, say, about the level of expenditures on a public good, how are those diff erences re- solved? What is meant by the problem of “aggregat- ing preferences”?

2. Why may there not be a well-defi ned outcome when majority voting is used to resolve diff erences in views? Is there any vot- ing procedure that yields a well-defi ned outcome in all situations? When there is a well-defi ned outcome, is it necessarily effi cient?

3. What are alternative ways for determining the level of public goods expen- ditures? Are there ways that ensure an effi cient level of expenditures on public goods?

4. What are some of the ways in which politics aff ects the outcomes of public decision making about resource allocation?

FOCUS QUESTIONS

232 CHAPTER 9 PUBLIC CHOICE

exploding government defi cits, polls consistently show strong sentiment toward reducing the defi cit. However, no consistent picture emerges con- cerning the trade-off s: some polls suggest that voters would be willing to pay higher taxes or accept expenditure cuts to reduce the defi cit, but other polls (confi rmed to some extent by voting patterns) suggest other- wise. Although in polls voters consistently say they believe that the gov- ernment should spend less on assistance to foreign countries, when asked how much should be spent, they give a number considerably in excess of what the United States is currently spending.

Unless individuals are faced with concrete trade-off s, under which they actually have to give up something to get more, say, of another good, it is diffi cult to get them to think hard about their choices. In addition, there are special problems in inducing individuals to reveal truthfully their preferences concerning public goods. If what they have to pay does not depend on their answer, then there is a tendency to ask for more of the good: one would normally like more public goods as long as one does not have to pay for them. However, if what an individual says aff ects how much he or she has to pay, there is an incentive for the individual to pre- tend that he or she enjoys the good much less than he or she really does— the individual knows that the answer will have a negligible eff ect on the total amount supplied, and he or she would like to be a free rider.1

In private decisions, the decision maker knows his or her own pref- erences. In public decisions, however, the decision maker must ascertain the preferences of those on whose behalf he or she is making the decision. This is the fi rst important diff erence between public and private resource allocations.

INDIVIDUAL PREFERENCES FOR PUBLIC GOODS

Collective decision making is diffi cult because diff erent individuals have diff erent views, for instance, about how much should be spent on public goods. They diff er for three reasons. Sometimes there are simply diff er- ences in tastes. Just as some individuals prefer chocolate ice cream and others vanilla, some people prefer public parks and education, whereas others prefer private goods, such as video games and cars. This book will not have much to say about these matters of tastes.

The other two sources of diff erences are incomes and taxes. Richer individuals have higher incomes, so normally they prefer to spend more on all goods, both public and private. When the government spends more on

1�Economists have devised some clever ways of getting around these problems in principle, but in prac- tice, no government has actually implemented them. See the appendix to this chapter.

233Public Mechanisms for Allocating Resources

public goods, however, richer individuals often have to pay a relatively large share of the additional cost. In the case of private goods, rich and poor indi- viduals typically pay the same price; with public goods, in eff ect, richer indi- viduals typically have to pay a higher price. The tax price is the additional amount an individual must pay when government expenditures increase by one dollar. The tax price multiplied by total government expenditures equals the individual’s tax payment. A higher tax price by itself means that richer individuals would want a lower level of expenditures on public goods. With an income eff ect leading to a higher desired demand, however, and a price eff ect leading to a lower desired demand, the net eff ect is ambiguous.

To observe these eff ects, consider fi rst a situation in which there are N people and each must pay the same amount, regardless of income. Under this system of uniform taxation, the tax price is just 1/N and the tax payment is G/N.2 With proportional taxation, everyone pays the same percentage of income. The tax price can be easily calculated. If Y is aver- age income, N�Y is total income; if t is the tax rate, then tN�Y is total gov- ernment revenue. This must equal government expenditures:

G 5 tN�Y.

Thus, the tax rate is

t 5 G

NY �.

The tax payment of an individual with income Yi is

tYi 5 GYi N�Y

�.

If government expenditures increase by a dollar, the individual’s incremental tax—the tax price—is just Yi /NY. Thus, an individual with average income (Yi 5 Y�) faces a tax price of 1/N, an individual with above-average income (Yi . Y�) faces a higher tax price, and an individual with below-average income faces a lower tax price.

A progressive tax system is one in which tax payments increase more than proportionately with income; a regressive tax system is one in which they increase less than proportionately. Accordingly, the tax price for a high-income individual under a progressive tax system is typically greater than Yi /NY.

2�No government today imposes uniform taxation—that is, taxes that do not depend on income. However, many “clubs,” which can be thought of as voluntary collective organizations, do charge uniform fees, and there are some dedicated taxes (e.g., gasoline taxes that are used for road construction) that do not depend directly on income.

234 CHAPTER 9 PUBLIC CHOICE

INDIVIDUAL’S CHOICE OF “MOST PREFERRED”

LEVEL OF GOVERNMENT EXPENDITURES

The individual’s most preferred level of government expendi- tures occurs at the tangency

between the budget constraint and the indifference curve.

(A) As shown, with proportional taxation, individuals with lower incomes face a lower tax price (fl atter budget constraint). The

income and substitution effects work in opposite directions,

so it is ambiguous whether the most preferred level of gov-

ernment expenditure is higher or lower. (B) As shown, with

uniform taxation, all individuals face the same tax price, so there

is only an income effect. Rich individuals prefer higher levels

of expenditure. (In this example, the rich and poor are assumed to have the same indifference

curves—preferences—and differ only with respect to the

budget constraints.)

FIGURE 9.1

Indifference curves

Quantity of public goods

Private goods

YP

YR

GPGR

Rich person’s budget constraint

Poor person’s budget

constraint

NY

Quantity of public goods

Private goods

GP GR

Indifference curves

Poor person’s budget constraint

Rich person’s budget constraint

A

B

Given the individual’s tax price, we can derive his or her preferred level of public goods expenditure, as illustrated in Figure 9.1. Individuals with diff erent incomes face diff erent budget constraints; the preferred levels of public goods expenditure are at the tangencies of the indiff erence curves with the budget constraints.

Diff erent individuals will diff er with respect to their preferred level of expenditures. With proportional taxation, poorer individuals face

235Public Mechanisms for Allocating Resources

lower tax prices, and on that account, their preferred level of expen- ditures, GP, is higher. Poorer individuals have lower incomes, however, and with lower incomes they demand less public as well as private goods. The net eff ect is ambiguous. Figure 9.1A illustrates the case in which the substitution eff ect (lower tax price) dominates the income eff ect, so the poorer individual does prefer a higher level of public goods than the richer person.

With uniform taxation, there is only an income eff ect, so high-income individuals will prefer higher levels of public expenditure (Figure 9.1B); with progressive income taxation, lower-income individuals will face a lower tax price than with proportional taxation, so their preferred level of expenditures will be even higher than with proportional taxation.

Figure 9.2 shows how utility depends on the level of government expenditures. The individual’s most preferred level of expenditures occurs at G*, but utility is maximized under a budget constraint, at the point of tangency with the indiff erence curve. The further away the actual level of expenditures is from the preferred level of expenditures, G*, the lower the level of utility.

Figure 9.3 indicates the relationship between the level of utility and the level of public goods expenditure for three diff erent groups—the rich, the poor, and the “middle,” assuming a particular tax system. Each has its own preferred outcome, and utility decreases as expenditures deviate either above or below that level. For expenditures above the middle, the marginal benefi ts of increased public expenditure are less than the mar- ginal costs the individual bears in additional tax payments, whereas the converse holds for expenditures below the preferred level.

UTILITY DEPENDS ON LEVEL OF GOVERNMENT EXPENDITURE

Utility is maximized at the point of tangency between the indif- ference curve and the budget constraint. The farther away the actual level of expenditures from the preferred level of expenditures, G*, the lower the level of utility.

FIGURE 9.2

Quantity of public goods

Utility

G*

236 CHAPTER 9 PUBLIC CHOICE

THE PROBLEM OF AGGREGATING PREFERENCES

In the private market, the fi rm does not have to balance the claims and interests of one group against those of another. If an individual is will- ing to pay a price for a commodity that exceeds the marginal cost of production, it pays for the fi rm to sell the commodity to the individual. Decisions are made on an individual basis. In the public sector, on the other hand, decisions are made collectively. For example, when a poli-

tician votes to increase expenditure on some public good, it is not as if the politician has to pay for the good himself or herself. The politician’s vote is intended to represent the interests of his or her constituents, but their opinions are not likely to be unanimous: some individuals would like more military spend- ing, others less; some individuals would like more expenditures on welfare, others less.

The problem of reconciling diff erences arises whenever there must be a collective decision. Popular political discussions often refer to what “the people” want. Because diff erent people want diff erent things, how- ever, how can a social decision be made from these divergent views? In a dictatorship, the answer is easy: the dictator’s preferences dominate. There is no such easy resolution in

UTILITY AS A FUNCTION OF G

The fi gure refl ects the preferences of three groups in

the population: the poor, the middle class, and the rich. In this example, the rich prefer

higher levels of expenditure to the middle class, who prefer higher levels than the poor.

FIGURE 9.3 Utility of rich

Utility of poor

Quantity of public goods

Utility with

uniform taxation

GrGmGp

Utility of those with middle income

CENTRAL PROBLEMS OF PUBLIC CHOICE Preference Revelation:

• Ascertaining the desired level of public goods of each individual.

Aggregating Preferences:

• Different individuals have different preferred levels of public expenditure.

• Preferred level depends on both individuals’ income and the tax system.

• Other things being equal, rich typically prefer higher levels.

• However, the cost of increased public expenditure may be higher for the rich.

237Public Mechanisms for Allocating Resources

a democracy, though. A number of diff erent voting rules have been sug- gested, among them unanimity voting, simple majority voting, and two- thirds majority voting. Of these, perhaps the most widely employed rule for decision making in a democracy is simple majority voting.

MAJORITY VOTING AND THE VOTING PARADOX

We all know how majority voting works. Suppose you and two friends are trying to decide whether to go to a movie or a basketball game. You take a vote: if the movie gets two votes, you go to the movie; if the basketball game gets two votes, you go to the game. Sometimes majority voting does not lead to such a clear outcome, though, when there are more than two alterna- tives. A majority voting equilibrium requires that there is one alternative that can win a majority in a contest against any alternative. As early as the eighteenth century, the famous French philosopher Nicolas de Condorcet noted that there may not exist any majority voting equilibrium. The problem may be seen in the following simple example, in which there are three vot- ers and three alternatives, denoted A, B, and C. A could be going to a movie, B to a basketball game, C renting a video; or A could be spending more money on health care for children, B reducing the defi cit, C cutting taxes.

Voter 1 prefers A to B to C.

Voter 2 prefers C to A to B.

Voter 3 prefers B to C to A.

Assume we vote on A versus B. Voters 1 and 2 vote for A, so A wins. Now we vote on A versus C. Voters 2 and 3 prefer C to A, so C wins. It appears that C should be the social choice. C wins against A, which wins against B. But let us now have a direct confrontation between C and B. Both Voter 1 and Voter 3 prefer B to C. This is referred to as the voting paradox, or the paradox of cyclical voting. There is no clear winner. B beats C and C beats A but A beats B. Assume we began by saying we were going to fi rst vote on B versus C, and put the winner against A. B beats C, and then A beats B. But just to check that we had made the right decision (A), we decide to put A against C. C beats A. So we think C is the winner. But then we check that by challenging C with B. B beats C—which was our original vote. B again appears to be the winner. But just to check, we again challenge it with A. A again beats B, as we knew from our earlier vote. The voting process thus goes on and on.

Often, to avoid these voting cycles, democracies organize their decision making as a sequence of votes—for instance, fi rst they will

238 CHAPTER 9 PUBLIC CHOICE

vote A against B, and the winner of that vote will be put against C, with the fi nal determination depending on the outcome of that vote, with no further contests. In that case, it may be very important to control the agenda—the order in which the votes occur. For instance, as just depicted, we know that A would defeat B, and C would defeat A, so C would be the decision. But suppose, instead, we structure the election as fi rst a con- test between B and C, and then the winner of that contest against A. A  would win that election. Finally, suppose we structured the elec- tion as fi rst a contest between A and C, and the winner of that election against B. Then clearly B would win. Thus, the winner of each of these elections is determined solely by the order in which the pairwise com- parisons were made.

Note, too, that if individuals realize there is going to be a particular sequence of votes, they may wish to vote strategically. That is, in the fi rst round of the vote, Voter 1 may not vote his or her true preferences on, say, A versus B, but think through the consequences of that for the eventual equilibrium. The voter may vote for B, even though he or she would prefer A, knowing that in a contest between C and B, B will win, whereas in a contest between A and C, C might win. Because this voter prefers B to C, he or she votes initially for B.

This analysis leads to two questions. First, are there voting rules that will ensure a deter- minate outcome for any vote? Second, are there any circumstances under which simple major- ity voting will yield a determinate outcome? We take up these two questions in the next two sections. It turns out that the voting par- adox cannot be resolved through voting rules, but there are indeed circumstances in which majority voting yields clear decisions.

ARROW’S IMPOSSIBILITY THEOREM

An endless cycle of voting is clearly an unsat- isfactory state of aff airs. It is natural to ask, then, whether there is any other political

RANK-ORDER VOTING FAILS: INDEPENDENCE OF IRRELEVANT ALTERNATIVES • John and Jim prefer a swimming pool to a library

to a tennis court.

• Tom prefers a tennis court to a swimming pool to a library.

• Lucy and Jill prefer a tennis court to a swimming pool to a library.

Majority voting: Three prefer a tennis court to a swimming pool.

Three prefer a tennis court to a library.

Five prefer a swimming pool to a library.

SWIMMING POOL WINS

Rank order (lowest sum of ranks wins): Tennis, nine; swimming pool, eight; library, thirteen.

TENNIS COURT WINS

Rank order in choice between tennis court and swimming pool (library not an option): Tennis, seven; swimming pool, eight.

SWIMMING POOL WINS

Adding “irrelevant choice”—library—alters the outcome.

239Public Mechanisms for Allocating Resources

mechanism—any other set of rules for making social decisions—that eliminates this problem. An ideal political mechanism should have four characteristics:

1. Transitivity. If the rule shows that A is preferred to B, and B is pre- ferred to C, then A should be preferred to C. As we have seen, simple majority voting lacks this essential property. Without this property, we can get into cyclical voting.

2. Nondictatorial choice. There is a simple way of avoiding voting cycles: give all decision-making powers to a dictator. As long as the dictator has con- sistent preferences, then there will never be a voting cycle. In a democratic society, however, a meaningful political mechanism must ensure that the outcomes do not simply refl ect the preferences of a single individual.

3. Independence of irrelevant alternatives. The outcome should be inde- pendent of irrelevant alternatives; that is, if we have to make a choice between, say, a swimming pool and a tennis court, the outcome should not depend on whether there is a third alternative, such as a new library.

4. Unrestricted domain. The mechanism must work no matter what the set of preferences and no matter what the range of alternatives over which choices are to be made.

In looking for a system that would satisfy all four of these properties, a number of alternative rules have been examined, but each fails one or more of the requirements. For instance, rank-order voting in which indi- viduals rank the alternatives, then the ranks assigned by all individuals are added together, and the alternative with the lowest score wins, does not satisfy the “independence of irrelevant alternatives” criterion.

The quest for an ideal system came to an end with the fi ndings of Nobel Laureate Kenneth Arrow of Stanford University. He showed that there was no rule that would satisfy all the desired characteristics. This theorem is referred to as Arrow’s impossibility theorem.3

Arrow’s impossibility theorem has one further interesting and import- ant implication. We often hear expressions such as “The government seems to be acting in an inconsistent manner,” or “Why doesn’t the gov- ernment determine its priorities and then act on them?” This language personifi es the government, treating the government as if it were an indi- vidual. Language is important: although we all know that the government is not a single individual, speaking of it as if it were frequently leads us to think of it in such facile terms. We come to expect that government

3�Kenneth Arrow, Social Choice and Individual Values, 2nd ed. (New York: John Wiley & Sons, 1963).

240 CHAPTER 9 PUBLIC CHOICE

should act consistently like a rational individual. Arrow’s impossibility theorem, however, suggests that, unless some individual is granted dic- tatorial powers, the government should not be expected to act with the same degree of consistency and rationality as an individual.

In the earlier example in which there was no majority voting equilib- rium, we saw the importance of control of the agenda. We also saw that it can be benefi cial for individuals to vote strategically—that is, to vote not according to their true preferences, but to take into account how the out- come of the current vote will aff ect the fi nal outcome. Just as Arrow estab- lished that there does not exist any way of adding together the preferences of diff erent individuals to satisfy all the desired characteristics of a choice mechanism, it has been shown that there does not, in general, exist any vot- ing system4 in which individuals will always vote their true preferences.

SINGLE-PEAKED PREFERENCES AND THE EXISTENCE OF A MAJORITY VOTING EQUILIBRIUM

Although Arrow’s impossibility theorem shows that there is no voting rule that always satisfi es the desirable properties of a social choice mech- anism described earlier, under some conditions, the simple system of majority voting yields a determinate outcome.

Figure 9.3 showed the level of utility as a function of the level of expen- diture on public goods. There, each individual has a single peak to his or her preference profi le. This property of single-peaked preferences is enough to guarantee the existence of a majority voting equilibrium. Note that the peak need not be “interior” but may lie on the “end,” so that preferences such as those in Figure 9.4A are also consistent with single-peakedness.

On the other hand, preferences such as those illustrated in Figure 9.4C are not consistent with single-peakedness. Both 0 and G1 are (local) peaks. Unfortunately, such examples arise naturally in considering many public choice problems.

For instance, consider the problem of an individual’s attitudes toward expenditures on public education. If the level of expenditure on public education is below a certain minimum level, a rich individual may pre- fer sending his or her children to private schools. If so, any increase in expenditure on public schools simply increases his or her taxes; the rich individual gets no direct benefi ts. Thus, his or her utility decreases with

4�A voting system is any set of voting rules by which a group of individuals tries to reach a decision—for instance, by dropping from consideration the alternative with the lowest number of votes or by giving individuals several votes and allowing them to assign as many as they like to each alternative.

241Public Mechanisms for Allocating Resources

SINGLE-PEAKED AND DOUBLE-PEAKED PREFERENCES

(A and B) With single-peaked preferences, there always exists a majority voting equilibrium. (C) Without single-peakedness, there may not exist a majority voting equilibrium.

FIGURE 9.4

Single-peaked preference

Quantity of public goods (G)

Utility (U)

0

Single-peaked preference

G

U

0

Double-peaked preference

G

U

0 G1

A

B

C

242 CHAPTER 9 PUBLIC CHOICE

government expenditures up to a critical level at which the individual decides to send his or her children to public school. For increases beyond that level, the individual derives some benefi t. Of course, beyond some point, the increases in taxes more than off set the benefi ts. For this indi- vidual, a high level of expenditure is preferred to no expenditure, but no expenditure is preferred to an intermediate level of expenditure. There may be no majority voting equilibrium in this case.

Although preferences for a single public good (with no private good option, unlike education) are usually single-peaked, when we have to rank choices involving more than one public good, those rankings are seldom single-peaked.5 To obtain single-peakedness, we have to restrict ourselves to voting on one issue at a time.6

Equally important, for most distribution issues there will not be a majority voting equilibrium.7 This can be seen most clearly in consider- ing the structure of income taxation. Suppose we are voting among three income tax schedules that are designed to raise the same amount of reve- nue. For simplicity, we assume that there are three groups of individuals with equal numbers—the poor, the middle class, and the rich—and that they vote in solid blocks. The three tax schedules are denoted as A, B, and C in Table 9.1. Tax schedule A is strictly proportional; it takes the same fraction of income from each individual. The poor and the middle class then get together and propose tax schedule B. This reduces the taxes they pay but taxes the rich much more heavily. Clearly, tax schedule B will win a majority over A. Now, the rich propose to the poor: “Because you are more needy, why don’t we lower your taxes somewhat more? At the same time, we’ll adjust the tax schedule at the upper end, to reduce the inequi- ties associated with excessive taxation.” Thus, they propose tax schedule C, which, relative to B, lowers the taxes on low and high income and raises them on middle income, so that now both the middle- and upper-income individuals pay a larger proportion of their income in taxes than do the poor.

5�See G. Kramer, “On a Class of Equilibrium Conditions for Majority Rule,” Econometrica 41 (1973): 285–297. 6�See S. Slutsky, “A Voting Model for the Allocation of Public Goods: Existence of an Equilibrium,” Jour- nal of Economic Theory 11 (1975): 292–304. 7�See D. K. Foley, “Resource Allocation and the Public Sector,” Yale Economic Essays 7 (1967): 45–98.

TABLE 9 .1 A LTERNATIVE TA X SCHEDULES

PAID IN TAXES

FRACTION OF INCOME

A B C

Poor 20% 18% 17%

Middle 20% 18% 21%

Rich 20% 23% 22%

243Public Mechanisms for Allocating Resources

Clearly, tax schedule C wins a majority over B. Now, however, the middle class proposes going back to straight proportional taxation. Because both the upper- and middle-income individuals prefer schedule A, A defeats C. We again get a cyclical pattern of voting.8

THE MEDIAN VOTER

When preferences are single-peaked, we asserted that there is a well-defi ned majority voting equilibrium. What does it look like? And how does it correspond to the Pareto effi cient equilibrium that we described in Chapter 5?

When preferences are single-peaked, we can rank individuals by their preferred levels of, say, expenditure on the public good, from the individual who prefers the least to the individual who prefers the most. The median individual is the individual such that half prefer less and half prefer more than he or she does. In Table 9.2 Jim is the median voter. The outcome of majority voting corresponds to the preferences of the median voter. In this case, it is Jim’s preferred level ($1000) that wins. The rea- son is simple: if any level of expenditure below $1000 is voted on against $1000, Jim plus all of those who want more than $1000 vote for $1000; because Jim is the median voter, a majority cast their vote for $1000. If any level of expenditure above $1000 is voted on against $1000, Jim and all of those who want less than $1000 vote for $1000. Again, $1000 wins.

THE INEFFICIENCY OF THE MAJORITY VOTING EQUILIBRIUM

Because the median voter determines the level of expenditure on public goods, to ascertain whether there is too much or too little expenditure we

8�If we restrict the set of tax schedules over which voting occurs, for instance, to tax schedules with an exemption level and a fi xed marginal tax rate (called fl at-rate tax schedules), there may be a majority voting equilibrium. See T. Romer, “Individual Welfare, Majority Voting, and the Properties of a Linear Income Tax,” Journal of Public Economics 4 (1975): 163–185.

TABLE 9 .2 PREFERRED LE VEL S OF E XPENDITURE ON PUBLIC GOODS

LUCY TOM JIM JOHN JILL

$600 $800 $1000 $1200 $1400

Jim is the median voter.

244 CHAPTER 9 PUBLIC CHOICE

need only examine how the median voter votes, and contrast that with the conditions for effi ciency discussed in Chapter 5. The median individ- ual is assumed to compare only the benefi ts he or she receives with the costs that he or she bears. The median voter’s benefi ts are lower than total social benefi ts (which includes all the benefi ts that accrue to others), but so are his or her costs. Whether there is too much or too little expenditure on public goods thus depends on whether the median voter’s share of total (marginal) costs is less or greater than his or her share of total (marginal) benefi ts.

For a wide variety of public goods, with proportional or progressive taxation majority voting will result in an oversupply of public goods. To see this, assume there are N individuals. With uniform taxation, the tax price for each individual would be 1/N; with proportional taxation, it would be Ym/�YN, where Ym is the income of the median voter and Y is average income; and with progressive taxation, the tax price would be still lower. With a symmetric distribution of income, as illustrated in Figure 9.5A, the  income of the median individual equals the average income—that is, Ym 5 Y. However, most income distributions are skewed, as shown in Figure 9.5B. There are a few very rich individuals, which increases the average income. As a result, average income exceeds the income of the median individual, Ym , Y, so that with proportional taxa- tion the tax price is even less than 1/N.

The question then is: What fraction of the total marginal benefits accrues to the median voter? If the marginal benefits are uniform, then the median voter gets 1/N of the total marginal benefits; so, with uniform taxation, the median voter would get 1/N of the total social benefits and bear 1/N of the total costs. Therefore, the median voter would vote for an efficient level of expenditure. With proportional or progressive taxation, however, the median voter’s share of the cost would be smaller than his or her share of the benefits, and he or she would vote for excessive expenditures—that is, for a level of expendi- ture at which the sum of the marginal benefits is less than the total marginal cost to society.

Some forms of public expenditures are actually enjoyed dispro- portionately by the poor: the rich may make less use of public parks because they have large yards of their own. According to the median voter theory, there is an especially large tendency for an oversupply of such goods. In many cases, it is difficult to ascertain the balance: the median voter gets a smaller share of the benefits and bears a smaller share of the costs.

In the examples discussed so far, the median voter corresponds to the voter with the median income. This is often, but not necessarily,

245Public Mechanisms for Allocating Resources

INCOME DISTRIBUTIONS

(A) Shows a symmetric income distribution; the median income, Ym, equals the average income. (B) Shows a skewed income distribution more typical of the United States. Because there are a few very rich individuals, the average income exceeds the median income.

FIGURE 9.5(f )

Ym = Y

(f )

YYm

A

B

the case. Higher-income individuals normally demand more of all goods, including public goods, if they face the same prices; but the tax price faced by higher-income individuals is typically higher, so in some cases, very–high-income individuals may actually prefer lower levels of public expenditures on certain goods than do middle-income individuals.

Median voter theory says that to understand collective choices, we should focus our attention on the median voter and how a particular policy aff ects that individual. Consider what the median voter theory says about support for, and consequences of, abolishing the deductibility of state and local income and property taxes for purposes of the federal income tax, a proposal that is frequently raised in the context of tax reform. Most indi- viduals do not itemize; that is, in calculating their taxable income, they

246 CHAPTER 9 PUBLIC CHOICE

subtract a “standard deduction” rather than listing all the allowable deductions, including local and state taxes. Accordingly, they would be unaff ected directly, although the increased federal revenue, which could be used to reduce the defi cit or taxes, would be a benefi t. Thus, the median voter should support this reform. However, it has been defeated repeatedly, sug- gesting that there are other considerations that go into determining outcomes in political pro- cesses, as we note later in this chapter.

One of the worries of state governors has been that if state taxes are made nondeduct- ible, the “tax price” for state expenditures will go up. Now, in effect, the federal govern- ment subsidizes the state. If a New York voter who pays 28 percent of his or her income to the federal government in income taxes, pays $1000 more in state taxes, and can deduct that expenditure from income, his or her fed-

eral tax liability goes down by $280, and the net cost to him or her is only $720. With a higher tax price, though, demand for state-provided goods and services would normally be expected to decrease. The median voter theory says “not to worry”: tax deductibility has no effect on the level of state-provided goods and services because the median voter in the state does not itemize.

THE TWO-PARTY SYSTEM AND THE MEDIAN VOTER

We noted earlier that an elected representative bears a negligible frac- tion of the costs of, and receives a negligible fraction of the benefi ts from, an increase in government expenditure. What can economic theory say about how he or she should vote? A natural supposition is that the politi- cian wishes to stay in offi ce and that, accordingly, this individual wishes to maximize his or her votes, given the position taken by his or her rival. A vote-maximizing voting strategy can easily be defi ned as follows: Assume there are two parties, “R” and “D.” Party R takes the position of Party D as a given. Focusing on a single issue, the level of expendi- ture, denote by GR the position of Party R (that is, the level of public expenditure advocated by the party) and by GD the position of Party D.

MA JORITY VOTING Proposal that gets a majority against all other wins.

When majority voting equilibrium exists, it refl ects preferences of the median voter.

• In a two-party system, both parties will converge to the position of the median voter.

• Majority voting equilibrium is not, in general, Pareto effi cient.

Majority voting equilibrium may not exist.

• Proposal A defeats B, C defeats A, but B defeats C.

• Majority voting equilibrium exists if preferences are single peaked.

Arrow’s Impossibility Theorem: There does not exist an alternative voting mechanism that avoids prob- lems of majority voting (and satisfi es certain other desired properties).

247Public Mechanisms for Allocating Resources

For each value of GD there is an optimal (that is, vote-maximizing) posi- tion for GR.

Under the hypothesis that each party seeks to maximize its vote given the position of its rival, what will each party do? Let Gm be the preferred level of expenditure of the median voter. Suppose Party D chooses GD . Gm. Then, if Party R takes a position between Gm and GD, it will get all the voters who prefer an expenditure level less than or equal to Gm, and some who prefer slightly more; thus Party R gets over 50 per- cent of the votes and wins. In response, Party D will choose a position, GD9, between Gm and GR, which wins against GR. Then, however, Party R chooses a position, GR9, between GD9 and Gm. The process continues until both parties stand for the same position: that of the median voter (Gm). (See Figure 9.6.)

This result is consistent with the widely observed allegation that with our two-party system, voters get no choice: both parties take a “middle- of-the-road” position. This is precisely what the theory predicts.9

Some important limitations of the theory, however, need to be borne in mind. First, we noted earlier that in general, there may not be a major- ity voting equilibrium. One does exist if individuals have single-peaked preferences. In the present context, this requires that we should be able to arrange issues along a single dimension—for example, conservative– liberal. If, however, there are a variety of dimensions—some individuals are liberal on some issues and conservative on others—then the median voter is not well defi ned, and there may be no equilibrium to the political process.

Second, we have ignored questions of participation in the political pro- cess. There are, for instance, costs associated with becoming informed

9�This general theory is due to Harold Hotelling, a pioneer in mathematical economics who taught at Columbia and North Carolina State universities. See H. Hotelling, “Stability in Competition,” Economic Journal 39 (March 1929): 41–57. For a comparative portrait of the median voter over time and across countries, see: J. E. De Neve, “The Median Voter Data Set: Voter Preferences across 50 Democracies,” Electoral Studies 30 (December 2011): 865–887.

TWO-PARTY SYSTEM

If both parties in a two-party system try to maximize their votes, taking the position of the rival as fi xed, in equilibrium both parties will adopt the position of the median voter.

FIGURE 9.6

Gm GRGR¿ GD¿ GD

248 CHAPTER 9 PUBLIC CHOICE

and voting. These costs are suffi ciently great relative to the perceived ben- efi ts that, for example, even slight changes in the weather, which makes it slightly less pleasant to go outside to vote, have signifi cant eff ects on voter participation. In particular, voters whose preferences are near the median have little incentive to be politically active, particularly if they believe that the political process will refl ect their preferences anyway. Thus, it may be in the interests of those who are more extreme to attempt to pull their party away from the center. This tendency for greater political activism at the

SOCIAL CHOICE THEORY

T he hypothesis that government actions can be explained as the outcome of individuals’ acting rationally in their own self-interest, in response to the political “rules of the game,” is a central tenet of social choice theory. James Buchanan, who taught at George Mason University until his death in 2013, received a Nobel Prize in 1986 for his pioneering work in this area. In this view, designing the rules of the game—the constitution—is critical. An important part of the rules of the game is the imposition of constraints on government (such as limits on defi cit fi nancing). Without some form of constraints, the incentive of the majority to redistribute income in their favor, away from the minority, or of special interest groups to try to induce elected representatives to act in their interests, rather than in the general interest, is too great to be resisted.

Even acts of Congress that seem to go against the special interest groups appear in a different light when viewed from this perspective. Consider the 1986 tax reform, which attempted to strip out of the tax law many of the favorable provisions that special interest groups had successfully lobbied for inclusion. Professor Buchanan argued that this action should be viewed from the self-interest of Congress: the tax system had become so overladen with special provisions that the cost of granting

further special benefi ts was rising sharply. The tax reform enabled Congress to start with a clean slate: now there was greater opportunity for intro- ducing new special benefi ts. The greater opportu- nity for Congress to dispense special favors was of immense benefi t to congressional representatives.*

One issue addressed by Nobel Laureate Gary Becker, of Chicago and Stanford universities, is how to explain the seemingly disproportionate infl uence of certain small interest groups. Although they con- stitute less than 2 percent of Americans, farmers have succeeded in getting huge subsidies from the federal government. The answer Becker suggests is that with small groups, the free rider problem that we encountered in Chapter 5 is smaller. “Bribing” representatives to support one’s special interest is a public good: all wheat farmers benefi t from a wheat subsidy; all steel or car producers benefi t from trade barriers that keep out less expensive foreign steel or cars. The smaller the group, the easier it is to persuade all members to contribute to the cost of lobbying. Each of these programs has losers, and not only are the losers far more numerous, but their aggregate losses exceed the benefi ts of the special interest groups—although each of the losers loses a little. Opposing the special interest groups is also a public good, and each opponent has an incentive to be a free rider.

*James M. Buchanan, “Tax Reform as Political Choice,” Journal of Economic Perspectives 1 (Summer 1987): 29–35.

249Alternatives for Determining Public Goods Expenditures

extremes may partially off set the median-directed tendencies noted earlier. There is a tension between the centripetal force of the median voter model and the centrifugal force of the base voter mobilization model.

ALTERNATIVES FOR DETERMINING PUBLIC GOODS EXPENDITURES

We have identifi ed several major problems with majority voting, the most commonly employed way of making collective decisions: there may not be a determinate outcome, and even when there is, it may not be effi cient; and voters may vote strategically, not revealing their true preferences. Even if there is no ideal system, are there perhaps systems that resolve one or the other of these problems? Economists have looked for such alternative systems.

LINDAHL EQUILIBRIUM

The most famous alternative system is called the Lindahl solution, after the great Swedish economist Erik Lindahl, who fi rst proposed it in 1919.10 He was looking for a system that would yield effi ciency; he paid little attention to the other problems listed earlier. Lindahl’s system attempts to mimic, as far as possible, the way that the market works in providing pri- vate goods. Remember, market equilibrium for private goods is described by the intersection of the demand and supply curves. All individuals face the same price. The sum of the quantities they demand is equal to the sum of the quantities supplied by fi rms.

One way we can characterize the effi cient level of public goods is the intersection of the “collective’’ demand curve (formed by adding verti- cally each individual’s demand curve) with the supply curve. The demand curves are generated by asking the individual how much of the public good he or she would demand if he or she were to pay so much for each unit produced; that is, in Figure 9.7A if the fi rst individual faced a tax price of, say, p1, he or she would demand G*.

10�E. Lindahl, “Positive Lösung, Die Gerechtigkeit der Besteuerung,” translated as “Just Taxation— A Positive Solution” in Classics in the Theory of Public Finance, ed. R. A. Musgrave and A. T. Peacock (New York: St. Martin’s Press, 1958). Application of the Lindahl equilibrium for determining public goods expenditures is still debated today, see for example, J. Dávila, J. Eeckhout, and C. Martinelli, “Bargaining over Public Goods,” Journal of Public Economic Theory 11 (December 2009): 927–945; and A. van den Nouweland and M. Wooders, “Share Equilibrium in Local Public Good Economies,” Journal of Mathematical Economics 47 (May 2011): 376–381.

250 CHAPTER 9 PUBLIC CHOICE

Government expenditure

Tax price

A Individual 1

B Individual 2

C Collective demand curve— the vertical sum of individual demands

p1

p1

G*

Government expenditure

Tax price

p2

p2

G*

G*

p

Supply curve

p1 + p2

FIGURE 9.7

LINDAHL EQUILIBRIUM

(A and B) Demand curves for public goods; demand

depends on the tax price faced by the individual. (C) The two

demand curves are added vertically—that is, at each level

of government expenditures, the tax prices of Individuals 1

and 2 are added together. The Lindahl equilibrium is the level

of expenditures G*, at which the collective demand curve thus constructed intersects

the supply curve. At the intersection, the sum of

the tax prices equals the supply price.

251Alternatives for Determining Public Goods Expenditures

COMPARISON OF PUBLIC CHOICE MECHANISMS Majority voting

• Equilibrium may not exist.

• When equilibrium exists, in general, it is not Pareto effi cient.

Lindahl equilibrium

• Equilibrium always exists.

• Equilibrium is always Pareto effi cient.

• Individuals do not have incentive for truthful revelation of preferences.

The Lindahl equilibrium is illustrated in Figure 9.7C. We add verti- cally the demand curves for Individuals 1 and 2 illustrated in Figure 9.7A and 9.7B, obtaining the collective demand. The Lindahl equilibrium occurs at the intersection of this curve with the supply curve. Price, along the supply curve, measures the marginal cost of production. p1 measures the fi rst individual’s marginal benefi t (marginal willingness to pay for an additional unit of government expenditure) at G*, and what he or she has to pay, and p2 measures the same thing for the second individual. At G*, the sum of the marginal benefi ts (the total marginal willingness to pay) just equals the marginal cost. The Lindahl equilibrium is thus a set of tax prices (the amount each individual has to pay if one more unit of the public good is produced) adding up to the marginal cost of production, such that, given those tax prices, every individual prefers the same level of expenditures, G*. Because at the Lindahl equilibrium the sum of the marginal benefi ts equals the marginal cost, the Lindahl equilibrium is Pareto effi cient.

We noted earlier that, in fact, there were a whole range of Pareto effi cient resource allocations, with one individual better off in some and another better off in others. Almost by defi nition, there cannot be una- nimity about which of these points is preferred. The Lindahl equilibrium picks one of the Pareto effi cient points, but individuals who are disadvan- taged by this particular Pareto effi cient point will not agree to the use of this mechanism for determining the allocation of public goods; indeed, they would prefer Pareto ineffi cient allocations as long as the level of util- ity they obtain is higher.

The most telling criticism of the Lindahl solution is that individuals do not have an incentive to tell the truth because their tax price increases as their stated demand does—that is, the higher their stated demand (given the demand state- ments of others), the higher the equilibrium expenditures on public goods will be. Higher expenditures on public goods, of course, necessitate higher equilibrium tax prices. The demand curves that are used in the Lindahl analysis were drawn under the hypothesis that individuals face a given tax price; they believe that nothing they say will alter what they have to pay per unit of public expenditure. If they understand the Lindahl mechanism, however, they will realize that what they say does alter what they have to pay per unit of public expen- diture, and thus they will not truthfully reveal their demands.

252 CHAPTER 9 PUBLIC CHOICE

Let us briefl y review the two sets of processes by which collective deci- sions concerning public goods could be determined, majority voting and the Lindahl equilibrium. Voting, as we have seen, may not yield an equilib- rium; when it does, it is, in general, not Pareto effi cient. We saw earlier that no democratic mechanism fully resolved the problem of the nonexistence of a voting equilibrium. The Lindahl equilibrium will be effi cient, but indi- viduals do not have an incentive to be honest in providing the informa- tion required to implement it. The next natural question is: Are there more general plausible ways of organizing collective decision making that yield effi cient outcomes, and that do not have the problems we have encoun- tered in the case of majority voting and the Lindahl equilibrium? Although economists have devised complicated schemes that have some desirable properties, or that may work in some circumstances, it appears that there is no perfect solution to the problem of collective decision making in dem- ocratic societies. The appendix to this chapter describes one scheme that, although it induces individuals to reveal honestly their preferences, has other problems that perhaps explain why it is not commonly used.

POLITICS AND ECONOMICS

The preceding discussion of the political process diff ers markedly from the kind of analysis one might typically fi nd in a political science course. There, one might discuss the roles of special interest groups and political institutions. A full discussion of the relationship between economic the- ories of the political process and other theories of the political process would take us beyond the scope of this book. In the following pages, we touch on some economic interpretations of certain political phenomena.

WHY DO INDIVIDUALS VOTE?

In many elections, voter participation rates are low and sensitive to such chance occurrences as changes in the weather. The reason for this, as we have noted, is that the benefi ts of voting for the individual are low—there is little chance of one person aff ecting the outcome. The alternatives may diff er so little that the outcome is inconsequential, and, even though the costs of voting are relatively low, they are not low in relation to the ben- efi ts. Indeed, in a fully rational calculation, no one would vote: the prob- ability that an individual’s vote would make a diff erence to the outcome (because, in most cases, the individual cares only about whether his or

253Politics and Economics

her candidate wins or loses, not the magnitude of the win or loss) is essen- tially zero. However, individuals do vote.11

This paradox is resolved, in a somewhat tautological manner, simply by assuming that individuals get utility out of voting, or, more generally, out of participation in the political process. More to the point, considerable time and energy are devoted to inculcating into our children notions of civic responsibility, and among these civic responsibilities is the respon- sibility to be an informed voter.

The same considerations imply that when the individual votes, he or she may not act in the narrowly self-interested manner that we have assumed in our discussion so far. Individuals may support representa- tives who vote for closing some loophole in the income tax system because it would result in a more equitable distribution of the tax burden, even though their personal tax liability might thereby be increased.

ELECTIONS AND SPECIAL INTEREST GROUPS

The political models we have discussed thus far have assumed that all individuals are well informed about the consequences of all alternatives under consideration, that all individuals vote, and that they cast their votes on the basis of the implications that each alternative has for their own (private) welfare.

Many believe that this does not provide an adequate description of the political process. Although constitutionally each person has one vote, some votes seem more eff ective than others. The outcome of the polit- ical process, in this view, refl ects the political power of special interest groups.

Assessing the validity of these views is beyond the scope of this chap- ter. Instead, we focus on a more narrow set of questions: What can eco- nomic theory say about the kinds of interest groups that are likely to be eff ective? How can we reconcile the eff ectiveness of special interest groups with the fact that each individual in the United States has only one vote?

The answer to these questions is related to our discussion in Chapter 5 of the public interest as a public good. We saw there that the effi cient man- agement of the public sector was a public good. Similarly, when others choose elected offi cials who are competent and who refl ect values similar

11�In U.S. presidential elections since 1932, between 49 and 63 percent of the voting-age population has voted. In elections of U.S. representatives to Congress, participation has been somewhat lower, at 33 to 59 percent (U.S. Census Bureau, The 2011 Statistical Abstract, Table 418). In local school board elections, often less than 10 percent of the eligible voters vote.

254 CHAPTER 9 PUBLIC CHOICE

to our own, we benefi t as well. This may reduce our incentive to become informed voters, and encourage us instead to “free ride.”

At the same time, we should note that the free rider problem may not be as serious in small groups as it is in large ones. It is easier, for instance, to form an interest group of a small number of steel producers to attempt to persuade Congress to restrict steel imports than it is to form an interest group of the large number of steel users who would be adversely aff ected by such restrictions. Each of the producers has more to gain than each of the consumers has to lose, even though the aggregate gains of producers may, in fact, be less than the aggregate losses to consumers.

Trade unions, recognizing the nature of the free rider problem, have long sought closed shops, in which all workers must support the activities the union believes to be in their collective interests. Once they have this power, they can attempt to use it not only at the bargaining table but also in the political arena, where they act as a special interest group.

THE POWER OF SPECIAL INTEREST GROUPS

How are interest groups able to exercise power? There appear to be at least three mechanisms. First, as we noted earlier, individuals have little incentive to vote or to become informed concerning the issues. Interest groups can attempt to lower the costs of becoming informed and vot- ing, particularly for voters who are likely to support them. They do so by making information (supporting their own views) readily available and providing transportation, child care, and other assistance on poll- ing day. Interest groups mobilize their constituencies through a combi- nation of targeted information campaigns and aggressive “get out the vote” eff orts.

Second, we noted the diffi culty that politicians have in obtaining information about the preferences of their constituents. There is no sim- ple demand-revealing mechanism for public goods as there is for private goods. Interest groups attempt to provide such information. Politicians may lack the technical information required to make informed political decisions—for example, they may not know the consequences of contin- ued imports of cheap foreign cars or clothing. Interest groups are a pri- mary source of information, and it is through providing information that they often exercise infl uence.

The third mechanism is through direct and indirect bribery of the politician. Direct bribery does not occur often, at least in most jurisdic- tions in the United States. (Presumably, this may not be due to the purity of our politicians so much as to the costs associated with being caught.)

255Politics and Economics

Indirect bribery, however, is important: special interest groups provide fi nancial and other forms of support for politicians who support their interests; this support is viewed to be essential in running a successful political campaign.

Since Congress enacted the fi rst federal campaign fi nance legislation in 1867 prohibiting federal offi cials from requesting contributions from navy yard workers, there have been many attempts at campaign fi nance reform. Key laws over the past century include the Tillman Act of 1907, which prohibited corporations and national banks from contributing money to federal campaigns; the Federal Corrupt Practices Act of 1925, which strengthened disclosure requirements and increased expenditure limits; and the Federal Election Campaign Act of 1971 (together with the 1971 Revenue Act and many subsequent amendments) to provide stricter limits and greater transparency in campaign fi nancing, an institutional framework for enforcing federal campaign fi nance laws, and federal fi nancing of presidential elections.

Nonetheless, the impact of many of these statutes has been steadily eroded by legislative loopholes, poor enforcement, and court challenges, usually based on freedom of speech. For example, the Supreme Court rul- ing in Citizens United v. Federal Election Commission, 558 U.S. 310 (2010), determined that the First Amendment’s right to free speech applies to corporations, permitting them to advocate expressly for or against can- didates without any government-imposed limits on corporate campaign expenditures (see case study, “Campaign Finance Reform”).12

OTHER ASPECTS OF THE POLITICAL PROCESS

Several other aspects of the political process aff ect outcomes. In the United States, each of the two parties holds primary elections to deter- mine who represents them in the general election. Voter turnout in each refl ects disproportionately the “extremes” in the party: more conserva- tive Republicans or more liberal Democrats. Although each party has an interest in nominating a candidate who will win—and thus one who is a “centrist”—there is an off setting centrifugal tendency in the primaries.

This has been exacerbated by the process of gerrymandering— designing congressional districts in which one or the other party is dom- inant. Winning the primary in such a district is almost tantamount to winning the election.

12�Federal Election Commission, The FEC and the Federal Campaign Finance Law, February 2004 (updated February 2011), and The Federal Election Campaign Laws: A Short History, www.fec.gov; and National Public Radio, A Century Of U.S. Campaign Finance Law, www.npr.org.

256 CHAPTER 9 PUBLIC CHOICE

One consequence of gerrymandering is that a party can take control of, say, the House of Representatives even if it gets far less than a majority of votes. This happened in the 2012 election, where the Democrats gar- nered more than one million more votes, but the House was controlled by Republicans. Other countries, concerned that such outcomes are undem- ocratic, have introduced systems of proportional representation, in which

CAMPAIGN FINANCE REFORM

Behind the power of many special interest groups is the power of money—especially money to support candidates they favor. Tra- ditionally, the Democratic Party has had less access to money, and its members have been particularly concerned that the power of money has created an unlevel playing fi eld, where the voices of rich indi- viduals and corporations are heard more clearly than the voices of the poor and middle class. How- ever, incumbents of any party, who typically have better access to fi nancial support, are usually less enthusiastic about reform than those trying to get elected. This helps explain in part why Congress, at least in the past, has had only limited enthusiasm for strong reforms.

Some reforms have focused on limiting the amount of money that any organization can give to any candidate. Such reforms typically have been subverted: if a corporation cannot give money directly, it may still encourage its employees to con- tribute. If they cannot give money to a candidate, they give money to the candidate’s party. Even if they were restricted in their contributions to the candidate’s party, they could give money to some group identifi ed with a position of the candidate. If the National Rifl e Association were limited in the amount it gave to pro-gun representatives, for instance, it could still mount a campaign against their opponents who advocate gun control.

Some critics of campaign fi nance reform worry that it will create an unlevel playing fi eld of a

different sort. Groups that can mobilize thousands of volunteers and huge letter-writing campaigns, such as senior citizen advocates or labor unions, will still be able to exercise political infl uence.

There is also a worry that any restrictions on contributions to parties or causes would be, in effect, a restriction on First Amendment (free speech) rights—in a democracy, individuals should not be restricted in what they say, how they say it, or to whom they say it. Campaign fi nance mainly goes to support the dissemination of views; although, to be sure, questions may be raised at the extent to which a thirty-second spot ad on TV promotes intel- ligent decision making.

One set of reforms that has already been imple- mented entails public disclosure of campaign con- tributions. The hope is that candidates who receive large amounts of money, say, from the fi nancial sec- tor might feel embarrassed supporting legislation that is excessively favorable to that sector. Even this disclosure requirement has been undermined recently, though, with the creation of secretive “SuperPacs” that provide support for candidates and positions.

The second reform entails trying to create a more level playing fi eld by partially matching with public funds money raised privately by the cham- pion money raisers. Such actions do not circum- scribe the right of any individual to say what he or she wants, but ensures that others with differing views can also get their perspectives across.

257Politics and Economics

the number of parliamentarians is proportional to votes received, rather than a “winner-takes-all” system.

We have noticed the problems posed by low voter turnout. Some coun- tries have tried to address this by requiring everyone to vote (or, more accurately, go to the voting booth). Even with small fi nes, much higher voter turnout is achieved, with the hope that the electoral outcomes are more refl ective of the views of the citizenry.

We noted earlier that, in the American system, politicians can be partially “bought” by campaign contributions. There are other ways by which their behavior gets distorted to refl ect the interests of special inter- ests. Of particular concern is the “revolving door,” in which politicians obtain jobs after their public service in the industries over which they passed legislation while in offi ce. There is a clear confl ict of interest, and although there are rules in places to limit the extent to which this occurs, they have proved imperfectly eff ective.

Finally, the media play an important role in forming opinions, both of political leaders and ordinary citizens. In some countries, the media are controlled by a small elite, representing particular views (say, of the very wealthy). In such circumstances, news and information can get distorted. Although in most countries politicians believe that they are unfairly treated, in many, there is a real problem of lack of media diver- sity. Some developing countries have responded by providing public sup- port for community radio stations; many developed countries provide broad-based support for various forms of media. Information—especially accurate information—is a public good, and, as we have repeatedly seen, private markets undersupply public goods.

THE ALTRUISTIC POLITICIAN?

An alternative view holds that many politicians do not behave in as self-interested a way as we have assumed throughout this chapter. Just as individuals behave altruistically as private citizens and give to charity, so too do they behave as public citizens, in their capacity as elected offi cials. In our society, considerable status and respect are accorded to public offi - cials and public service. Eff ective government depends on the quality of these public servants.

However, being disinterested does not suffi ce to ensure that politi- cians will make wise decisions. As has been noted, even a well-intentioned public offi cial may have a diffi cult time ascertaining the preferences of his or her constituents. Further, even if the electorate would like to choose altruistic representatives who vote exclusively on principle, attempting

258 CHAPTER 9 PUBLIC CHOICE

to ensure that there is an “effi cient” supply of public goods, they face a diffi cult problem: as long as not all individuals running for offi ce are disinterested, the voters must select between those who are and those who are not. If voters believe that it is better to be a “disinterested” pub- lic servant than a selfi sh politician, then self-interested politicians will all attempt to resemble disinterested public servants. How are voters to choose among them on the basis of the limited information they nor- mally have available?

In numerous instances, politicians seem to act “on principle,” voting in a way that is inconsistent not only with their own narrowly defi ned self-interest, but also with the wishes of their voters. They thus risk not getting elected (even though the electorate often respects such shows of “independence” and “principle”).

Although altruistic behavior on the part of politicians is to be pre- ferred to corruption, or even to lesser forms of self-aggrandizement, economists have long worried about the reliability of seemingly altruistic behavior in the pursuit of the common interest. Indeed, it was the seem- ing capriciousness of the actions taken by political leaders (often allegedly in the public interest) that led Adam Smith to suggest that there was a bet- ter way that the public interest might be served: by each individual’s pur- suing his or her private interest. Unfortunately, although Adam Smith’s invisible hand may work well for most goods, it does not work well for public goods. Still, at least to a limited degree, self-interest on the part of politicians—their desire to get re-elected—serves an important function: it induces them to try to elicit accurately the preferences of their (voting) constituents and to vote for measures that refl ect those preferences. It is this form of self-interest that lies behind the analysis of the two-party model presented earlier.

THE PERSISTENCE OF INEFFICIENT EQUILIBRIUM

In recent years, the prices of commodities—from cars to corn to aluminum to uranium to tomatoes—have been higher as a result of trade restrictions, or the threat of imposition of trade restrictions. When special interest groups manage to impose trade restrictions or to obtain subsidies for themselves, the resulting resource allocations not only violate generally accepted standards of equity and fairness but also are frequently ineffi - cient. There are alternative allocations that could make everyone bet- ter off . Why, in the face of this, don’t individuals get together and propose one of these Pareto dominating alternatives, to which, presumably, all

259Review and Practice

would agree? There is no universally agreed-upon answer to this puzzle. Several “partial” answers may be suggestive.

First, as we have already seen, the public interest is a public good. Because the eff orts to maintain a good government must come from pri- vate individuals, there will be an undersupply of this (as any other) pri- vately provided public good.

Second, many of the distributive implications of public programs undertaken at the behest of special interest groups are far from obvious— and this is deliberately so. It is unlikely, for instance, that the American voters would deliberately vote to transfer resources (give a public gift) to rich rice farmers. For these individuals to receive a transfer at the pub- lic expense, they must be included in a broader-scale program, of which they appear to be almost accidental benefi ciaries. Thus, rich rice farmers become advocates of federal aid to rice farmers, singling out, in their public rhetoric, the benefi ts that would accrue to poor rice farmers from such a program. A Pareto improvement might, for instance, entail giving each rice farmer a fi xed sum, which would leave the farmer free to move into some other occupation where productivity might be higher. Although such direct grants could be structured to make everyone better off , they would expose the true distributive implications of the program—that is, that most of the benefi ts accrue not to poor rice farmers, but to rich ones. Because the likely result would be weakened political support for subsidies to rice farmers, this Pareto improvement would not receive the backing of rich rice farmers.

SUMMARY

1. Collective decision making, such as determin- ing the level of public goods, diff ers from stan- dard decision making within a household in two important ways. First, there is a problem of elic- iting preferences. If the amount that individuals have to pay depends on their statements, they may tend to understate their true preferences. If the amount that individuals have to pay does not depend at all on their statements, they may tend to overstate their true preferences. Second, there is a problem of resolving diff erences in prefer- ences: How much should be provided if diff erent

individuals desire that the government should, for instance, spend diff erent levels on providing public goods?

2. Majority voting is the simplest way by which such diff erences are resolved. Unfortunately, a major- ity voting equilibrium may not exist.

3. Arrow’s impossibility theorem demonstrates the impossibility of fi nding an alternative, nondic- tatorial political mechanism resolves this prob- lem of majority voting and satisfi es certain other properties that one would desire of any political mechanism, such as independence of irrelevant alternatives.

REVIEW AND PRACTICE

260 CHAPTER 9 PUBLIC CHOICE

 4. The majority voting equilibrium exists if prefer- ences are single-peaked.

 5. Preferences for a single public good will usually be single-peaked. Preferences will not be single- peaked if:

a. There is more than one public good, and the vote is taken over packages, rather than over a single good at a time.

b. Voting is over a publicly provided private good, for which there exists a private alternative, such as education.

c. Voting is over distributional questions, such as the structure of the income tax schedule.

 6. The majority voting equilibrium, when it exists, refl ects the preferences of the median voter.

 7. In a two-party system, there will be a conver- gence of positions of the two parties toward that of the median voter.

 8. The majority voting equilibrium does not, in general, result in an effi cient supply of public goods; there may be either an undersupply or an oversupply.

 9. The Lindahl equilibrium is the level of pub- lic goods provision in which the sum of the tax prices equals the marginal cost of production. Although the Lindahl equilibrium is Pareto effi - cient, there is no incentive for individuals to tell the truth concerning their preferences.

10. In many elections, voter participation is low. In fact, economists are puzzled why individ- uals bother to vote at all, as the expected pri- vate benefi t—given the low probability that they aff ect the outcome—is typically less than the private cost.

11. Special interest groups often exercise strong infl uences over the outcome of political processes.

KEY CONCEPTS

Aggregating preferences

Arrow’s impossibility theorem

Lindahl equilibrium

Median voter

Preference revelation

Proportional taxation

Single-peaked preferences

Tax price

Uniform taxation

Voting paradox

QUESTIONS AND PROBLEMS

1. Assume that some individual’s marginal valu- ation of public goods increases. What does this do to the Pareto effi cient level of public expendi- tures? If this individual is not the median indi- vidual, what will happen in a two-party system to the equilibrium level of expenditure on public goods? If the equilibrium was originally Pareto effi cient, will it still be?

2. Assume that all individuals have identical pref- erences but some individuals are wealthier than others. Assume there is a single public good and a single private good.

a. Show diagrammatically how you derive the demand curve for the public good, as a func- tion of the tax price charged the individual.

b. Assume that the demand function is of the form

G 5 kY p ,

where k is a constant (less than 1), Y is income, and p is the tax price. This says that when income doubles the demand for public goods doubles, but when the tax price doubles the demand is cut in half. If the tax price is proportional to the indi- vidual’s income (as with proportional taxation), how will demand for public goods diff er among those with diff erent incomes?

3. Assume instead there is uniform taxation, so all individuals face the same tax price. Recall that along each individual’s demand curve, the price equals the marginal rate of substitution. Thus,

MRS 5 p 5 kY G �,

261Review and Practice

the marginal rate of substitution is proportional to income. Assume that income is symmetri- cally distributed, so that mean income equals the median. Explain why the majority voting equilibrium will be Pareto effi cient. Now, assume that income is not symmetrically distributed, but rather is skewed toward higher incomes, as in Figure 9.5B. Will the majority voting equilibrium still be effi cient? Will there be an under- or over- supply of public goods?

4. Demand curves are said to be income elastic if the demand for the good increases more than pro- portionately with income. For instance, with the demand curve

G 5 kY�2

p �,

the demand for public goods increases with the square of income. Draw the marginal rate of sub- stitution as a function of income for a fi xed level of expenditure on public goods. Assume income is symmetrically distributed. What is the rela- tionship between the average value of the mar- ginal rate of substitution and the marginal rate of substitution of the median individual? What does this imply about the equilibrium supply of public goods under majority voting with uniform taxation?

5. In the text, we suggested that for well-off indi- viduals with uniform taxation, preferences for education were not single-peaked. Why might preferences for local parks and urban public transportation systems (e.g., buses and subways) also not be single-peaked?

6. Is the median voter always the voter with the median income? Give examples.

7. How might the majority voting model be used to explain the growth of government expenditures?

a. Should changes in median or average income better explain increases in the demand for government services?

b. What should be the eff ect of an increase in the costs of producing public good caused by government ineffi ciency? Would it make a dif- ference if the increase in cost is a result of gov- ernment paying above-market wages (wages higher than those paid comparable workers in the private sector)? (Does your answer to the last question depend on whether the median voter is a government employee?)

c. Why might you expect that if income per cap- ita remains the same but the number of indi- viduals in the economy increases, the demand for public goods would increase?

8. One popular voting scheme is rank-order voting, in which individuals assign a rank (1, 2, 3) to the possible alternatives; the assigned ranks are then added up, and the alternative with the lowest sum wins. Consider a choice among four alter- native ways of spending public funds: a library, a ski slope, a swimming pool, and a garbage dump. Can you construct an example in which the out- come (the most preferred alternative) is, say, a library, if the vote is among the fi rst three alter- natives, whereas the outcome is a ski slope, if the vote is among all four alternatives? This voting scheme thus violates the principle that the cho- sen outcome should be independent of irrelevant outcomes (the garbage dump was not chosen in either situation).

9. Median voter theory says that to predict changes in collective decision making, one should focus on the median voter. Between 1973 and 1993, average incomes in the United States increased, whereas the income of the median family remained roughly stag- nant. (Since then, the median income has increased slightly, but not enough to erase the increased gap of the previous two decades.) How might an econ- omist focusing on median voter behavior and an economist focusing on average incomes diff er in their predictions concerning changes in the level and composition of public expenditures?

262

APPENDIX: NEW PREFERENCE- REVELATION MECHANISMS

This appendix describes a simple procedure that induces individuals to reveal their demands truthfully, provided there is no collusion among individuals. Everyone is asked to give his or her demand curve for public goods, just as in the Lindahl equilibrium. As before, the equilibrium will be at the intersection of the collective demand (formed by adding verti- cally the demand curves of each individual) and the supply curve. For simplicity, we assume that the marginal cost of production of the public good is constant, so the supply curve is horizontal. Now, however, there is a diff erent rule for determining the individual’s tax liability.

We fi rst add up the demand curves for all other individuals (vertically). The collective demand curve of all others intersects the supply curve at G0 in Figure 9.8. This is what the level of public goods would be if an individual said that he or she got no value out of the public good. This individual is told that for each unit beyond G0 that the government produces, he or she will have to pay the diff erence between the marginal costs of production and the collective valuation (demand) of all others. If the equilibrium entailed an output of G0 + 1, the individual would have to pay AB, the distance between the marginal cost curve and the others’ collective demand curve.

The individual is in a position to determine the level of public goods simply by announcing how much they are valued to him or her. Clearly, this individual will try to increase G to the point at which the marginal cost of increasing G is equal to his or her marginal benefi t. This can be seen in two alternative ways. First, in Figure 9.8B, we have plotted the marginal cost to the individual from each additional unit of production beyond G0�, given others’ demands. Because the individual’s marginal cost is the diff erence between the cost of production and others’ demand, the marginal cost of the G 1 1st unit is equal to AB. In Figure 9.8B, we have also drawn the individual’s demand curve; the individual will wish point G* to be chosen, where his or her demand curve intersects his or her marginal cost curve.

We now show that each individual has an incentive to reveal honestly his or her demand for public goods, and that the equilibrium is Pareto effi cient. To see this, we look at the individual’s budget constraint. The individual faces a budget constraint as depicted in Figure 9.9. The extra amount that the individual has to give up for each extra unit of public goods beyond G0 is the marginal cost minus the others’ collective demand (marginal valuation). Thus, the individual sets his or her marginal rate of substitution equal to the marginal cost minus the others’ collective

263Appendix: New Preference-Revelation Mechanisms

Quantity of public goods

Price

G0 G0 + 1

Collective demand of all others

Marginal cost of the public goodA

B

Quantity of public goods

Price

G0 G0 + 1 G*

Marginal benefit to individual (his or her

demand curve)

Marginal cost to individual

(given demand curve of others)

A

B

A

B

NEW PREFERENCE- REVELATION MECHANISM

(A) Shows the collective demand of all but one individual for the public good (the sum of their marginal valuations) and the marginal cost of production. If the last individual placed no value on the public good, the level of production of the public good would be G0, where the sum of the marginal valuations equals the marginal cost. As the level of expenditure increases beyond G0, the last individual is required to pay, for each additional unit, the difference between the collective (marginal) valuation of all others and the marginal cost. Thus, if G0 1 1 is produced, this individual must pay the amount denoted by AB. (B) Shows, at each level of output, the marginal cost that the (last) individual must pay for each extra unit of output. Thus, to have the economy go from producing G0 to producing G0 1 1 requires that this individual pay AB. It also shows the last individual’s marginal valuation of the public good (his or her demand curve). The individual’s most preferred level of expenditure is where his or her marginal benefi t from increased expenditures (given by the demand curve) exactly equals his or her marginal cost— that is, G*. If the government sets public expenditures at G*, the individual will be induced to reveal his or her demand truthfully.

FIGURE 9.8

264 CHAPTER 9 PUBLIC CHOICE

demand—point E in Figure 9.9. It is clear that the individual has no incen- tive to misrepresent his or her preferences. If the individual asked for any level of public goods other than G*, he or she would be worse off .

Assume now that each individual honestly announces his or her demand curve. Recall that in constructing the demand curve, the tax price for each individual (the slope of the budget constraint) was set equal to the individual’s marginal rate of substitution. Hence, when the demand curves are added vertically, the sum of the marginal rates of substitution are just the sum of the tax prices, and at the Pareto effi - cient allocation that equals the marginal cost (the marginal rate of transformation):

MRS1 1 MRS2 1 · · · 5 MC.

In other words, each individual’s marginal rate of substitution is equal to the marginal cost of the public good minus the sum of the marginal rates of substitution of others (the sum of their tax prices). For instance, for the fi rst individual,

MRS1 5 MC 2 (MRS2 1 MRS3 1 · · ·).

This is exactly the point we described earlier, at which the marginal cost to the individual of further increases in government expenditure (which equaled the marginal cost of production minus the sum of others’ demand prices at the given quantity) equaled the marginal benefi t to the individual (his or her marginal rate of substitution). We have just shown that by

Government expenditure

Private goods

G*G0

Individual’s indifference curve

At E, slope of indifference curve = marginal rate of substitution = marginal cost of public good – sum of others’ demand = slope of individual’s budget constraint

Individual’s budget

constraint

E

CHOICE OF OPTIMAL G BY INDIVIDUAL

If the individual must pay the difference between marginal

cost and others’ demands, and others have honestly revealed

their demands, the level of public goods will be Pareto effi - cient. From Figure 9.8, the price

the individual has to pay for each increment in public goods

expenditure increases. That is why the individual’s budget con-

straint has the shape depicted.

FIGURE 9.9

265

honestly revealing his or her demand curve, the individual maximizes his or her own utility, and the allocation of resources to public goods will be Pareto effi cient.

In spite of the attention that revelation mechanisms such as the one we have just described have received from economic theorists, there is considerable controversy about their relevance. There are several objec- tions to them, besides their possibly high administrative costs and the fact that they are susceptible to collusion (two individuals could get together, agree to distort what each said, and each be better off ).

Like the Lindahl equilibrium we described earlier, these mechanisms ensure that the condition for a Pareto effi cient allocation—that the sum of marginal rates of substitution equals the marginal rate of transformation— is satisfi ed. However, some individuals might prefer another Pareto inef- fi cient allocation that gives them a higher level of utility. Hence, it is not obvious that they would agree to a decision to adopt this mechanism, knowing that they would thereby be disadvantaged. Finally, the mecha- nisms do not, in general, guarantee a balanced budget. Although the sum of the marginal valuations (marginal rates of substitution) does equal the marginal cost, the total amount paid may well diff er from the total costs of the public good.

Appendix: New Preference-Revelation Mechanisms

PUBLIC EXPENDITURE IN PRACTICE

In this part, we show how the theoretical models we developed in pre- vious chapters can be used to analyze public expenditure in practice, focusing on national defense, health care, education, welfare, and social insurance. These particular programs were chosen for two reasons. First, they are among the most important programs for many countries, includ- ing the United States. In terms of dollars spent, they account for approx- imately three-fourths of federal expenditures, as well as of total public expenditures in the United States. Second, the examination of these particular programs brings out most of the critical issues in expenditure analysis; other programs can be analyzed using the basic framework and tools of analysis that we develop here. The first two chapters are devoted to explaining our basic approach to the analysis of public expenditures: Chapter 10 develops a general framework, and Chapter 11 shows how the benefits and costs of dif ferent government programs can be quantified both before and after the money has been spent. Chapters 12 through 16 then apply these concepts to specific sectors.

PART FOUR

269

FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

At least since the beginning of the 1990s, there have been calls for change in major segments of the American economy: its health care and edu- cation systems; major reforms in welfare and Social Security; cutbacks in some programs like defense; and expansions in others such as those aimed at developing new technologies. This chapter provides a frame- work for thinking systematically about such policies: questions that need to be asked, and methods that can be employed to help answer them.

Policy makers need such a framework to address increasingly com- plex issues. Indeed, the complexity of most government programs is so great that Congress delegates responsibility for working out most of the details (within guidelines Congress has set up) to the executive branch. In a process called rule making, agencies of the executive branch (such as the Environmental Protection Agency or the Department of Transpor- tation) spell out these details, and the public is given time to comment. The Offi ce of Management and Budget (OMB) within the Offi ce of the President provides guidance to the agencies in how to go about this pro- cess. Currently, OMB guidance closely refl ects the framework discussed in this chapter.

10 1. What are the major steps in the analysis of a public expenditure program?2. What are some of the reasons why the actual eff ects of a government program are diff erent from those that are intended or apparent at fi rst sight? What is meant by the incidence of a program?

3. Why are some programs said to be ineffi cient?

4. How in practice are the distributional impacts of a program assessed?

5. What is meant by the trade-off between equity and effi ciency? Is there always such a trade-off ?

6. Why might an under- standing of the political process be relevant for an understanding of the design of government programs?

FOCUS QUESTIONS

270 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

This framework for analyzing public expenditures provides guidelines for applying the effi ciency and equity criteria presented in Chapters  3 and 7. It is not a simple formula that can be applied blindly to all prob- lems, but rather a list of considerations that should be raised. Some may be more relevant to certain government programs than to others. The kinds of questions we are ultimately interested in addressing are:

• Why is there a government program in the fi rst place? • Why does the government program take on the particular form that it

takes? • How does the government program aff ect the private sector? • Who gains and who loses as a result of the government program? Are

the gains greater than the losses? • Are there alternative programs that are (Pareto) superior to current

government programs (that is, in which some individuals can be made better off without adversely aff ecting anyone else)? Are there alterna- tive programs that have diff erent distributional consequences but at the same time achieve the program’s primary objectives? What are the impediments to the introduction of these alternative programs?

We begin by breaking down the analysis of public expenditures into ten steps: (1) the need for a program; (2) market failures addressed by the program; (3) alternatives to the program; (4) particular design features of the program; (5) private sector responses; (6) effi ciency consequences; (7) distributional consequences; (8) equity–effi ciency trade-off s; (9) public policy objectives; and (10) the political process.

NEED FOR A PROGRAM

It is often useful to begin the analysis of a public program by investigating the program’s history and the circumstances under which it arose. Who were the individuals or groups who pressed for its passage, and what were the perceived needs that it supposedly addressed?

For instance, when the bill establishing the Social Security pro- gram was passed in 1935, the United States was in the midst of the Great Depression. Up to that time, few employers provided adequate pensions for their employees, and the private market for annuities (insurance poli- cies that provide individuals with a given annual income from retirement until death, regardless of how long they live) was undeveloped; many individuals had failed to save adequately for their retirement, and many who had saved had found their savings wiped out by the stock market

271Market Failures

crash in 1929. The failure to have adequate savings was not as irrational and improvident as it appears to us today; in those days, many individuals continued to work until they died. They needed life insurance to provide for their families after their demise, but did not need pensions for them- selves. In the Great Depression, however, many of these individuals lost their jobs and had no unemployment insurance. It was widely felt that society had to make some provision for them and that it was preferable to do so on a systematic basis, rather than just to solve the immediate prob- lems of the time.

MARKET FAILURES

The second step in the analysis of public programs is to attempt to relate the need, the source of demand, to one or more of the market failures dis- cussed in Chapter 4: imperfect competition, public goods, externalities, incomplete markets, imperfect information, and macroeconomic dis- equilibrium. In addition, we saw in Chapter 4 that even if the economy is Pareto effi cient, there are two further arguments for government inter- vention: fi rst, that the distribution of income emerging from the market economy may not be socially equitable; and second, that an individual’s own perceptions of his or her welfare may be an inappropriate or inad- equate criterion for making welfare judgments. There are merit goods, which the government should encourage, and merit bads, which the gov- ernment should discourage or prohibit.

In some cases, the nature of the market failure is obvious: national defense is a pure public good, and, as we argued earlier, in the absence of public provision, such goods will always be in undersupply. In other cases the answers are not so obvious, and economists may not agree about the nature of the market failure. Some economists believe that education is a public good, for example. However, most economists argue that it is essentially a private good (in the technical sense defi ned in Chapter 5) and that to fi nd an explanation for its public provision one must look else- where; for instance, at capital market imperfections, at the distributive consequences of public provision, or at education as a merit good, essen- tial for the functioning of a democratic society.

The fact that there is a demand for the public provision of some good or service does not in itself imply that there has been a market failure. Some demands for public provision arise from an inadequate understand- ing of the market and of the government’s capabilities for making things better. Identifying whether there is or is not a market failure is an essen- tial step in identifying the appropriate scope for government action.

272 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

HIGHER EDUCATION IN THE UNITED STATES

I n the United States, higher education illustrates several of the alternative forms of government involvement. It is publicly produced: every state has its own system of universities, colleges, and junior colleges. Although direct aid to private universities is limited in the United States, in other countries like Canada, it is common, and is granted on the basis of the number of students enrolled. In the United States the federal government provides considerable aid to research universities through a variety of programs of support for basic and applied research.

Most federal support to higher education, how- ever, takes the form of support to the consumers— the students. Even though there have been no general programs of support, there have been fi ve major selective programs:

1. Since World War II, a large number of veterans have attended colleges and universities at gov- ernment expense, most under the original GI Bill, and later under the Montgomery GI Bill and Post– 9/11 GI Bill. The federal government also offers merit-based Army, Air Force, and Navy Reserve Offi cers’ Training Corps (ROTC) scholarships, as well as Iraq and Afghanistan service grants for students whose parent or guardian died as a result of military service in either country.

2. Government provides targeted grant programs, such as Pell grants for low-income students, and Teacher Education Assistance for College and Higher Education (TEACH) grants for those studying to become elementary and secondary school teachers.

3. Government offers a variety of subsidized and unsubsidized student loans using either federal funds channeled through schools (Direct Loan Program) or funds from private lenders that are guaranteed by the federal government (Federal Family Education Loan Program).

4. Federal Work-Study Program provides part-time employment for enrolled students to help them pay for their educational expenses.

5. Federal government offers tax incentives for education-related expenses: the American opportunity tax credit for students pursuing an undergraduate degree or other recognized edu- cation credential, and the lifetime learning tax credit for those taking postsecondary education or vocational courses.

ALTERNATIVE FORMS OF GOVERNMENT INTERVENTION

Once a market failure has been identifi ed, a variety of government actions might address the problem. The four major categories of govern- ment action are: (1) public production; (2) private production with taxes and subsidies aimed at encouraging or discouraging certain activities; (3) private production with government regulation aimed at ensuring

273Alternative Forms of Government Intervention

that fi rms act in the desired way; and (4) some form of public–private partnership (PPP).

If the government decides to bear responsibility for production, it must decide on how the output is to be allocated. It can charge for the good at market prices; it can charge for the good at something approximating the cost of production, as it typically does for electricity; it can charge for the good, but the charges can be much less than the cost of production, as it typically does for higher education; it can provide the good free of charge and uniformly, as it does for elementary school and secondary school edu- cation; or it can allocate the good or service in some way corresponding to a perceived need or benefi t. In countries like Britain, where medicine is free, it is obviously not provided equally to all individuals. Needs diff er, so the decision as to who gets how much of the available supply of medical services is left to doctors (operating within guidelines set up by the gov- ernment, in consultation with them).

Similarly, if the good is to be produced privately, the government must decide whether to (a) contract directly for the commodity but retain responsibility for distributing it; (b) provide a subsidy to produc- ers, with the hope that some of the benefi ts will be passed on to consum- ers through lower prices; or (c) provide a subsidy to consumers. If some form of subsidy is desired, government must decide whether it should be provided through the tax system or through a direct grant. If a subsidy is granted, the terms have to be decided upon, for example, how restrictive eligibility standards should be. Cur- rently, governments engage in all these possible forms of action.

If the government opts for a hybrid PPP approach, then it must decide on the nature of its relationship with the private sector, particu- larly regarding mutual roles and responsibilities as well as allocation of risk. Arrangements can be as simple as service, lease, and management contracts with the private sector, whereby the government retains asset ownership but out- sources daily operations. They also can be as complicated as build–operate–transfer (BOT) contracts, whereby the private sector builds, for example, a toll road, operates it for an extended period of time to recoup its initial investment plus a market return to the capital it expended— perhaps 20 to 30 years—and then transfers the asset to the government.

ALTERNATIVE FORMS OF GOVERNMENT INTERVENTION 1. Public Production

Free distribution

Distribution at below cost of production

Distribution at cost

2. Private Production

Government subsidies to (taxes on) producers

Government subsidies to (taxes on) consumers

Direct government distribution

Government regulation

3. Public–Private Partnership

Outsourcing procurement of goods and services

Service, lease, and management contracts for operations and maintenance

Concessions such as BOTs for capital investment

Joint ventures (fi nancial and in-kind)

274 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

The importance of identifying alternative programs is increasingly recognized. Frequently, new programs can be devised that attain the objectives of older programs at less cost and more eff ectively. “Social innovation” is no less important than technological innovation. Today, there is increasing emphasis on the use of markets and market-like mechanisms.

THE IMPORTANCE OF PARTICULAR DESIGN FEATURES

The detailed provisions of a program—for instance, the precise state- ments concerning eligibility standards—are often crucial in determin- ing the effi ciency and equity consequences of the program. Fairness and effi ciency require making a number of distinctions that, though clear in principle, are diffi cult to apply in practice. The distinction between those who are hungry and those who are not is an important one, but devising a program to provide food for the hungry requires some easy way of iden- tifying who the hungry are. Too narrow a defi nition will result in many of those who are needy not receiving aid. Too broad a set of eligibility standards will result in many individuals who are not needy receiving aid, much to the objections of other taxpayers who are having to contribute to these individuals’ support. Thus, because of the impossibility of iden- tifying perfectly those who are truly deserving of aid, there is a trade-off , when designing regulations, between two types of errors: denying aid to those who are deserving and granting aid to those who are not deserving (see Figure 10.1). Diff erent individuals may judge the importance of these two kinds of errors diff erently.

The design of eligibility standards has further eff ects, as individuals may alter their behavior to gain eligibility or to receive larger benefi ts. There has been concern, for instance, that welfare programs that provide funds only to single mothers discourage marriage. The Supplemental Nutrition Assistance Program (SNAP; formerly the Food Stamp Pro- gram),1 which provides assistance to people with low available income, off ers another example of altered incentives. To calculate the amount of income available for spending on food, expenditures on housing are subtracted from the individual’s take-home pay, so the individual who spends more on housing receives more in food assistance. This may alter

1�Not only has the name of the program changed, but paper coupons (“food stamps”) have been replaced by the electronic benefi ts transfer (EBT) issuance system, similar to a debit card system.

275Private Sector Responses to Government Programs

behavior: because of the particular way it has been designed, SNAP— a program intended to encourage better nutrition among the poor—may encourage more expenditure on housing. There is little hard evidence to suggest that either of these concerns is signifi cant.

PRIVATE SECTOR RESPONSES TO GOVERNMENT PROGRAMS

One of the central features of a mixed market economy like that of the United States is that the government has only a limited degree of con- trol over it. The private sector, for instance, may react to any government program in such a way as to undo many of its alleged benefi ts. For exam- ple, when the government increases Social Security benefi ts, the welfare of the aged may not increase in the long run by the full corresponding amount; individuals may be induced to reduce their own savings for retirement, and children may be induced to provide less support for their aging parents. Public support may thus crowd out private support, erod- ing the impact of the program. Again, emperical observation to date pro- vides little evidence that this has occurred in a signifi cant way. In fact, there might actually be crowding in, for example, when government basic research encourages private sector applied research.

THE TRADE-OFF IN DESIGNING REGULATIONS

When eligibility standards are loose, many individuals who are undeserving will qualify for aid. When standards are tight, many deserving individuals will not qualify for aid.

FIGURE 10.1

Number of undeserving individuals receiving aid

Number of deserving

individuals not receiving aid

Tighter regulations

Looser regulations

276 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

In considering the consequences of a government program, one needs to look at the long-run consequences, after all producers and consumers have adjusted their behavior, as well as the immediate impact. One of the major impacts of rent control, for instance, is that the supply of new hous- ing dries up, the eff ects of which are felt only gradually.

Calculating the full private sector responses is often one of the most diffi cult and contentious aspects of analyzing a government program. To what extent, for instance, will a government subsidy to builders of lower-income housing result in higher profi ts, thus benefi ting the build- ing industry? To what extent will competition in the industry bid these profi ts away, lowering the price and increasing the supply, thus benefi t- ing the intended benefi ciaries? The answers depend on views concerning the housing and construction markets. How competitive is the building industry? If it is competitive, what is the elasticity of supply? What is the elasticity of demand?2

As we have already noted, the eff ects of a government program may hinge critically on seemingly innocuous design features. Economists look for marginal incentive eff ects. Two programs giving the same average subsidy can have quite diff erent marginal incentive eff ects. For example, SNAP, with a $200 cap on benefi ts per month, may have no marginal eff ect for those individuals who spend more than $200 on food. An analy- sis of the magnitude of the demand and supply responses from the private sector—and thus of the eff ects on price and quantities—must pay careful attention to these marginal incentives.

Recent research in economics has suggested that the economist’s stan- dard model may not always work as well as traditionally thought. Individ- uals may pay less attention to some forms of incentives than the standard theory would have predicted, or react in an unexpected way. Many econ- omists have thus tried to understand unanticipated responses (or no response at all) to incentives through the lens of behavioral economics, a relatively new discipline that explores the psychology underpinning household and fi rm responses to government policies and programs. The hope is that by exploring public fi nance through the lens of behavioral economics, we can more accurately predict private sector reactions to government initiatives spanning the domain of public sector economics, including public expenditure, taxation, social insurance, retirement sav- ings accruals, social safety nets, and income redistribution.3

2�Elasticity of supply is defi ned as the percentage change in quantity supplied as a result of a 1 percent change in price; elasticity of demand is defi ned as the percentage change in quantity demanded as a result of a 1 percent change in price. 3�See, for example, J. King, S. Mullainathan, and W. Congdon, Policy and Choice: Public Finance through the Lens of Behavioral Economics (Washington, DC: Brookings Institution Press, 2011).

277Efficiency Consequences

EFFICIENCY CONSEQUENCES

The next steps in expenditure policy analysis entail identifying the effi - ciency and distributional consequences of each alternative program and assessing the extent to which alternative programs can meet the objec- tives of public policy.

Government programs may result in ineffi ciencies both in the pro- duction of a good or service and in levels of consumption. In Chapter 8 we suggested that the government’s decision to produce a good or service itself, to purchase the good or service from private fi rms but distribute it itself, or to have private fi rms produce it and market it subject to govern- ment regulation may signifi cantly aff ect the costs associated with produc- ing and delivering the given good or service.

We also suggested that when consumers had an element of choice, the competition among providers would likely increase the effi ciency with which the goods or services were provided, as well as make what was produced more responsive to the needs and desires of consumers. These arguments are less persuasive if consumers have limited information con- cerning the product they are purchasing like medical care, or if consumer concern about costs is reduced because the government pays all (or a sub- stantial part of�) the costs, again, as in the case of medical care.

INCOME AND SUBSTITUTION EFFECTS AND INDUCED INEFFICIENCY

For many programs, it is useful to distinguish between substitution effects and income effects. Whenever a government program lowers the price of some commodity, there is a substitution effect: the indi- vidual substitutes the cheaper good for other goods. For example, with tuition subsidies for higher education, individuals substitute educa- tion for other goods on which they might have spent their money. On the other hand, grants to individuals that make them better off but do not alter the relative prices of different commodities result in an income effect: an individual changes his or her expenditure pattern after becoming better off. In many cases, there are both an income effect and a substitution effect, and both alter the individual’s behav- ior. Normally, however, it is only the substitution effect that we asso- ciate with inefficiency.

To illustrate, assume that the government gives an individual SNAP assistance to buy $10 worth of groceries every week. Prior to this, the

278 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

individual’s budget constraint was the lower one in Figure 10.2, which shows a substitution eff ect: by giving up $1 of groceries, the individual could acquire $1 more of other goods. SNAP shifts budget constraint to the right, so if the individual now wants to consume more than $10 worth of groceries, he or she still must give up $1 of other goods for each extra dollar of groceries consumed. There is no substitution eff ect, but there is an income eff ect: the individual now has an extra $10 to spend. The eff ect on food consumption is the same as giving the individual an equivalent amount of income (except when the individual would prefer to consume less than $10 worth of food each week). SNAP has altered the individual’s behavior to consume more food (B) than previously (A). Notice, however, that the individual does not increase food consumption by the full $10; the individual spreads this extra income between food and all other goods, just as he or she would have with $10 more of income. Because there is no substitution eff ect, there is no ineffi ciency associated with SNAP.

To see the substitution eff ect, assume, in contrast, that the govern- ment has agreed to pay for 30 percent of food purchases. This lowers the cost of food. The new budget constraint is shown in Figure 10.3. Now there is a substitution eff ect. Food is cheaper relative to all other goods, so the budget constraint rotates as shown. Note that in Figure 10.2, by con- trast, the new budget constraint associated with SNAP is parallel with the original budget constraint. (There is also an income eff ect in Figure 10.3: because with cheaper food, the individual not only consumes more food, but can also consume more of all other goods.)

INCOME EFFECT

Giving free food has an income effect but no substitution effect: its effects are identical to giving

an individual extra income.

FIGURE 10.2

Groceries consumed

Other goods consumed

Budget constraint

before free good

$10

Budget constraint after free food

Indifference curves

BA

279Efficiency Consequences

It is important to distinguish between income and substitution eff ects. In some cases, the government may want to mitigate a market distortion, such as monopoly power; or it may wish to encourage or discourage a par- ticular economic activity, in which case, it may want a large substitution eff ect. For example, if there is a belief that poor individuals do not attach suffi cient importance to housing, and the government wishes to improve the quality of housing they purchase, then a program in which the gov- ernment pays a fraction of housing expenditures (which has, as a result, a substitution eff ect) will be more eff ective than a fl at housing grant, which (unless it is very large) has only an income eff ect. The same rationale would apply to address the government’s concern that parents might not put a strong enough emphasis on the nutrition of their children.

On the other hand, if the government is primarily concerned with how well off diff erent individuals are if there are no market failures, then pro- grams that do not alter marginal incentives are preferable; such programs do not cause the ineffi ciencies associated with the substitution eff ect.

Returning to the SNAP case, we can see how a change in the design of the program avoids the ineffi ciency generated by the substitution eff ect. When the program was established in 1964, participants purchased food stamps at a discount from their face value, so the government paid a frac- tion of the costs of stamps: food stamps worth $100 might cost a poor person $70, and he or she may be allowed to buy, say, up to $2000 of food stamps. If he or she purchases $1000 worth of food stamps, the total subsidy would be $300. Today, the government simply gives a low-income individual a

SUBSTITUTION EFFECT

When government pays a part of the costs of food, there is a sub- stitution effect. The slope of the budget constraint changes. In this fi gure, the government pays a fi xed fraction of the cost of food, regardless of the amount the individual consumes.

FIGURE 10.3

Groceries consumed

Other goods consumed

Budget constraint

before food subsidy

Budget constraint after food subsidy

Indifference curves

280 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

fi xed amount of assistance (in the form of a credit on the recipient’s elec- tronic benefi ts transfer [EBT] card rather than in paper coupons); as long as the amount given is equal to or less than the amount the individual would spend on food anyway, this is equivalent to an income grant.

As we have seen, this version of food subsidies has only an income eff ect, whereas the earlier version has a substitution eff ect as well. The substitution eff ect introduces an ineffi ciency: the true cost of groceries— the amount of other goods that society must give up to obtain an extra unit of food—remains unchanged. For each additional dollar of food con- sumed, society must give up $1 worth of other goods. Under the original version of the food stamp program, however, individuals had to pay only 30 cents for a dollar’s worth of groceries.

Figure 10.4 shows how the new form of the program can cost the gov- ernment less, and leave poor individuals receiving the subsidy just as well off as before. BB represents the budget constraint before any food subsi- dies. The line BKB9 represents the budget constraint under the original form of the food stamp program, under which the government pays for a fi xed fraction of the costs of groceries, up to some limit. After that limit is reached (represented by the point K) individuals must pay the full price of groceries. The individual chooses the point E, at which his or her indif- ference curve is tangent to the budget constraint. The magnitude of the subsidy is the diff erence between what the individual has to pay and what society has to forgo; it corresponds to the vertical distance between the before-subsidy budget constraint and the after-subsidy budget constraint at the equilibrium level of consumption of groceries, the distance AE.

INEFFICIENCY ASSOCIATED WITH OLD-STYLE FOOD STAMP PROGRAM

Under the original form of food stamps, the government

paid a fi xed fraction of the costs of groceries, up to some

limit, generating the budget constraint BKB9. The new

form (BLB0), under which the government pays for a fi xed

amount of food, can make individuals just as well off,

but cost less. The “savings” is represented by the distance EG.

FIGURE 10.4

Groceries

Other goods

B

B B– B¿

F

A

K

E

G

H L

281Distributional Consequences

Figure 10.4 also includes a budget constraint for assistance given in the form of a fi xed amount to be spent on food. How do we know that the individual is just as well off ? Because the income grant was set so that the budget constraint it produced would be tangent to the indiff erence curve through E. By defi nition of indiff erence curves, then, the individual is just as satisfi ed at H as at E. But the income grant costs the government less. Again, the cost of the program is represented by the vertical distance between the before-subsidy and after-subsidy budget constraints. The size of the grant, HF, is smaller than the size of the 30 percent subsidy, AE.4 The reason for this is simple enough: when individuals must pay the full price of food at the margin (that is, when they have to pay $1 for $1 more worth of groceries), they value the increased consumption of gro- ceries by precisely what they have to forgo in other consumption goods. But when individuals are given a 30 percent subsidy, they then purchase groceries up to the point at which they value $1 worth of groceries at 70 cents, which is the cost to them of the $1 worth of groceries.

Note, however, that under the new form of assistance, individuals con- sume less food than under the old form. If the purpose of the program is to encourage food consumption—because, for instance, government believes that individuals, in maximizing their own utility, will not con- sume enough food—then the old form, under which the government in eff ect lowers the price of food to the poor, is more eff ective.

DISTRIBUTIONAL CONSEQUENCES

It is not always easy to ascertain who really benefi ts from a given gov- ernment program. Consider, for instance, the Medicare program, under which government fi nances most medical care for the aged. The aged clearly benefi t greatly from the program, but to some extent, the federal aid substitutes for money that families of the elderly would have contrib- uted (public expenditures thus crowd out private expenditures), and to that extent, the true benefi ciaries of the program are not the elderly but their children. With this sort of analysis, economists seek to identify a property they call the incidence of a government expenditure program or tax; that is, they seek to answer the question of who ultimately benefi ts from, is hurt by, or bears the burden of the program or tax.

4�To see this, note that the budget constraint segment LB0 is parallel to the budget constraint BB—the individual has to pay the full marginal cost of food, so the trade-off between food and other goods is unchanged. Thus, AG 5 FH (the vertical distance between two parallel lines is the same everywhere). The ineffi ciency associated with the 30 percent subsidy is thus measured by EG—the government must spend extra that amount to leave the individual just as well off as he or she would have been with an income grant. EG is the deadweight loss associated with the ineffi cient subsidy.

282 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

Government programs often induce a variety of responses from the private sector that result in changes in prices. Thus, a program’s eff ects can extend well beyond the people aff ected directly, and often the bene- fi ciaries are diff erent from those intended. There has been considerable concern that, at least in the short run, federal subsidies for private hous- ing for the poor increase the price of housing, making the true benefi cia- ries the slum landlords, not the poor.

The eff ect of a government subsidy is illustrated in Figure 10.5, which shows the demand and supply curves for housing. In the short

SHORT-RUN AND LONG-RUN INCIDENCE OF

EXPENDITURE PROGRAM

(A) In the short run, a subsidy may increase price more than quantity. Thus, landlords may

benefi t from a housing subsidy given to help the poor acquire better housing. (B) In the long

run, the output response will be larger and the price

response smaller.

FIGURE 10.5

Output (quantity of housing)

Price of housing

Q Q¿

Demand curve after subsidy

Short-run supply curve

p¿

p

Demand curve before subsidy

Output (quantity of housing)

Price of housing

Q Q¿

Demand curve after subsidy

Long-run supply curve

p¿ p

Demand curve before subsidy

A

B

283Distributional Consequences

run (Figure 10.5A), the supply of housing is assumed to be very inelastic because it takes some time for new housing to be constructed. Assume the government has passed a general subsidy for housing, the eff ect of which is to increase the demand for housing (the demand curve shifts up). In the fi gure, almost the entire subsidy is refl ected in the increased price of housing; the actual level of housing services provided increases very little. In the long run, of course, the supply response is likely to be larger; hence, in Figure 10.5B, the long-run supply curve is fairly fl at, showing that a small percentage increase in the price, given enough time, elicits a fairly large increase in the supply of housing. In the short run, the ben- efi ciaries of housing subsidies are the current owners of houses; renters fi nd that virtually their entire subsidy is refl ected in higher rents (the shift from p to p9). In the long run, however, renters are better off , as the increase in the quantity of housing supplied (from Q to Q9) serves to limit the price increase.

Mass transit subsidies provide another example: Who benefi ts from a new subway system? At fi rst glance, the answer seems obvious: sub- way riders. However, this may be incorrect. Those who own houses or apartments near the subway will fi nd that their houses and apartments are more sought after; the increased demand for these residences will be refl ected in the rents that the owners can charge (and in the market value of the houses and apartments). The commuter who owns no real estate fi nds that he or she is better off because of the better subway service, but worse off because of the higher rents, and the two eff ects are likely to cancel out. The true benefi ciaries are the property owners near the subway lines.

The subway example illustrates a general principle: the benefits of government programs are often capitalized in the value of scarce assets associated with obtaining those benefits (land near the subway stops). In  that case, the true beneficiaries are those who owned the asset at the time the program was announced (or passed, or when it came to be believed that the program would pass). By the same token, the costs of a program are often capitalized, so a tax on land is reflected in the value of the land; the true costs are borne by those who owned the asset at the time the tax was announced (or passed, or when it came to be believed that the tax would be passed). When those who benefit from a government program are different from those whom the program was intended to help, we say that the benefits have been shifted, or that the actual incidence (those on whom the benefits actu- ally fall) is different from the intended one. Considerable research in recent years has been devoted to determining the actual incidence of government programs.

284 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

EVALUATING THE DISTRIBUTIONAL CONSEQUENCES

As we have noted, diff erent individuals receive diff erent benefi ts from a given government program. Although it is obviously not possible to iden- tify how much each individual benefi ts, it may be important to know how diff erent groups in society are aff ected diff erentially. Which groups we focus on may vary from program to program, and benefi ts may vary within a particular income group. Thus, a program of rebates for heating-oil expenditures for people whose income falls below a particular level obvi- ously benefi ts the poor more than the rich, but it benefi ts some poor (those who consume a lot of heating oil, those who live in the Northeast) more than others (those who live in the Sun Belt). If the variability of consump- tion of heating oil among the poor is very large, this rebate program may be viewed as an unfair way of helping the poor, unless those who consume a lot of heating oil are viewed as particularly deserving of assistance.

INCIDENCE OF EDUCATION TAX CREDITS

In the 1996 election campaign, President Clin-ton proposed a $1500 tuition tax credit for the fi rst two years of college for students who obtained a B average. The B average requirement was intended to encourage students to work hard. With the tuition tax credit, a middle-income family that spent $2000 in college tuition on their fresh- man daughter would be able to reduce their tax bill by $1500. It was as if they paid their full taxes, and then the federal government sent them a check for $1500. By subtracting the $1500 directly from the tax payment, this “round trip”—money going to the government and then coming back—is avoided.

Although the intent of the tuition tax credit was to help middle-class families with children in college, to increase enrollment, and to encourage better school performance, there was considerable controversy about the true incidence.

• Many states charged less than $1500 for com- munity colleges. There was a concern that they

would (perhaps gradually) raise their tuition— after all, with $1500 coming from the federal gov- ernment in tax credits, individuals could afford to pay more for tuition.

• Colleges could offset the increased tuition costs for those whose parents did not receive the tuition tax credit by giving them scholarships. However, in cases they do not, students would fi nd college less, not more, affordable. Because most A and B students already go to college, the program might not increase enrollment much among them, but it might reduce enrollment among C students. Alternatively, teachers might worry that by giving a low grade—even a C—they might shut off a student’s chance of staying in college by cutting off the tuition tax credit, so there may be grade infl ation. With grade infl ation, though, better stu- dents may have less incentive to work hard.

• If states did raise their tuition, then at least some of the benefi ts would accrue to state governments,

285Distributional Consequences

rather than to the taxpayers. This is the “true” incidence. Education still might be helped if the state spent most of the increased tuition revenue on education; on the other hand, the state might reduce its educational expenditures, and give a tax cut to its taxpayers. Then the true incidence would be on the states’ taxpayers in general, not just on those who had children in college. If the state reduced taxes at the top, then what had appeared to be a middle-class tax break would become a tax break for upper-income individuals as a result of shifting.

• Similarly, private schools might be induced to raise their tuitions, enabling them to pay higher faculty salaries or support more research. Again, the inci- dence is markedly different from that intended.

As a result of concern about some of these perverse incentive effects—as well as complica- tions in implementation—the B requirement was dropped when the Hope and Lifetime Learning tax credits were enacted as part of the 1997 Taxpayer Relief Act.* Since 1998, use rates of these two tax credits have increased steadily, with the following results: while both enrollments and tuition fees have gone up, there is no evidence such increases have been due to the tax credits; and in 2008, although 10 percent of tax credit claimants had an adjusted gross income (AGI) of less than $25,000 or more than $100,000, the remaining 90 percent of tax benefi ts went to the intended target population of middle-income taxpayers with an AGI between $25,000 and $100,000.

*The Hope tax credit was expanded and renamed the American Opportunity Tax Credit in 2009.

SOURCES: Department of Education, Funding Education Beyond High School: The Guide to Federal Student Aid 2009–10, www.FederalStudentAid .ed.gov; Internal Revenue Service, IRS Publication 970—Tax Benefi ts for Higher Education, www.irs.gov; and Benjamin Rue Silliman, “College Tuition Tax Credits: An Examination of the Impact of the American Opportunity, HOPE, and Lifetime Learning Tax Credits Since 1998”, Proceedings of ASBBS 18, no. 1 (February 2011): 279–288.

In other cases, we may attempt to identify how producers are aff ected diff erentially. This typically is the focus of analysis in the evaluation of programs, such as agricultural price supports, aimed at aiding particular industries. In still other cases, such as the Social Security program, we may be concerned with the diff erential impact on the present elderly versus the impact on the young—the elderly of the future. We refer to these impacts as the program’s intertemporal distribution eff ects— distribution eff ects over time. In still other cases, we may wish to iden- tify the regional impact or the impact on cities versus suburbs, or urban versus rural areas.

When a program’s benefi ts accrue disproportionately to the poor (they receive more than their contribution to the costs of the program through the tax system), we say that its distribution eff ect is progressive. If the benefi ts accrue disproportionately to the rich, we say that the program’s distribution eff ect is regressive.

There are often controversies about who are the real benefi ciaries of a program, and one’s perspective on its distributive impact is determined in large part by the group on which one is focusing. For instance, govern- ment support for higher education is often viewed as enabling the chil- dren of the poor to go to college, and thus is viewed to have a positive

286 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

redistributive impact. However, children of the middle and upper-middle classes are more likely to avail themselves of a higher education. Thus,  general subsidies—such as reduced tuition for all students— disproportionately benefi t children of middle- and upper-income families. Indeed, by some calculations, they benefi t more than their share of taxes—educational subsidies to higher education are thus regressive. This is in contrast to targeted subsidies—such as scholarships for children from low-income families. Even then, it is not clear that parents’ income pro- vides the appropriate focus of attention: the benefi ciaries of education are not the parents but the children, since it is they who will receive higher wages as a result of their increased level of education.5 Those who hold to this view often favor student loan programs.

Let us contrast the distributional consequences of direct state sup- port for universities (allowing them to charge a low tuition) with the distributional consequences of a student loan program. Those who avail themselves of higher education will, on average, have a much higher income than those who do not. A loan program may thus be more pro- gressive than the current system, in which even low-wage high school dropouts are called on to provide some support for higher education. Loan programs introduced in 1993, which allowed repayments to be related to students’ incomes, could have increased progressivity still further, as students who wind up making higher incomes in eff ect pay more than those who receive low incomes, but in practice, relatively few student loans are of this type. As this example makes clear, one’s view of the distributional impact of a government program depends not only on what groups one focuses on but also on the available alternatives to a given program. The relevant choice is seldom one program versus no program, but one type of program versus another. Thus, the present state system of aid to higher education may be more progressive than a totally private education system, but its distributional impact may look less favorable when contrasted with a system of income-contingent loans for higher education.

FAIRNESS AND DISTRIBUTION

Political discussions commonly focus on the equity of various propos- als, with each side claiming that its proposals are more fair. Notions of fairness, unfortunately, are not always well defi ned, and diff erent

5�With middle- and upper-income parents who would have sent their children to college anyway, the true benefi ciaries may be the parents who save the money they otherwise would have spent, but to the extent that parents use this money to increase the bequest they leave to their children, it is the children who really benefi t.

287Equity–Efficiency Trade-Offs

individuals may have confl icting views of what is fair. A middle-class couple who love children but have decided, for fi nancial reasons, to limit the number of children they have to two may feel that it is unfair for them to have to support a child from a family of ten children whose par- ents do not want to use modern birth control methods and cannot aff ord to send their chil- dren to college without government assistance. A couple who have saved $40,000 to put a child through college may feel that it is unfair that they are not entitled to receive a government grant or loan when their next-door neighbors, with the same income but who have put nothing aside for their children’s education, enjoy expen- sive vacations every winter and are entitled to a government grant, which depends on assets as well as family income.

An unmarried person and a family with both spouses working may both think it unfair that their expected returns from Social Security are so much lower than those of an individual whose spouse does not have a job outside the home. However, an individual whose spouse does not work outside the home may feel that it is fair that he or she receive more, because the family has not had the benefi t of a second income.

EQUITY–EFFICIENCY TRADE-OFFS

Because of the ambiguities associated with using the term “fair,” economists try to avoid it in their analysis; rather, they focus on identifying the impact of programs. Economists begin their analysis of any program by looking for Pareto or near-Pareto improvements: changes in the program that make someone (or some groups) better off without making anyone (or almost any- one) worse off . Rent control, it is argued, fails to benefi t renters in the long run, as the supply of housing dries up. There are better ways of helping low-income individuals obtain housing. Welfare programs would serve the benefi ciaries better if taxpayers invested a little more money in training and education; in the long run, and the tax burden resulting from supporting the welfare popu- lation might actually be reduced. There may be alternative market-based ways of dealing with pollution, such as fi nes and tradable permits (see Chapter 6),

CONSEQUENCES OF PUBLIC PROGRAMS 1. Government programs may crowd out private

actions; equivalently, private actions may largely offset public actions, resulting in a small net effect.

2. Government programs give rise to income and substitution effects. The substitution effect is related to the magnitude of the marginal incentives.

3. Ineffi ciencies in public programs are related to the magnitude of the substitution effect.

4. The incidence of a program describes who ulti- mately benefi ts from, or is hurt by, the program. The actual incidence is often markedly different from the intended or apparent incidence.

5. The benefi ts of a program may be capitalized, in which case the true benefi ciaries are those who own the asset in which they are capitalized at the time the program was started (or announced).

288 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

which can achieve higher levels of pollution reduction than a system of strict regulation at lower costs—benefi ting both the environment and the economy.

Unfortunately, although there is considerable scope for such Pareto or near-Pareto improvements, in many expenditure programs trade-off s exist between the objectives of effi ciency and equity (redistribution of income or benefi ts to the needy). It may be possible to design a more progressive expenditure program, but only at some cost. An increase in Social Security benefi ts may be desirable from the perspective of cer- tain distributional goals, but the increased benefi ts may lead to earlier retirement, and the higher taxes required to fi nance them may decrease work incentives. Higher unemployment compensation may provide increased income to some who are among the most needy, but unem- ployment insurance may make some individuals feel disinclined to fi nd another job.

Disagreements about the desirability of diff erent programs often arise from  disagreements not only about values—the relative importance of equity versus effi ciency considerations—but also about the nature of the trade-off s, how much loss of effi ciency would result from an attempt to change a program’s structure of benefi ts to make its distributional impact more progressive.

Figure 10.6 shows the equity–effi ciency frontier for a hypothetical program and the indiff erence curves for two individuals. In Figure 10.6A, Scrooge is much less willing to give up effi ciency for a gain in equity than is his brother, Spendthrift. E1 represents the point on the trade-off curve that is optimal as Spendthrift sees it, whereas E2 is optimal from the point of view of Scrooge. Not surprisingly, Scrooge chooses a point with higher effi ciency but lower equity than does Spendthrift. Thus, the source of the disagreement about policy is a diff erence in the values held by the two individuals.

On the other hand, Figure 10.6B depicts a situation in which the dif- ferences about policy arise from diff erences in judgments concerning the nature of the trade-off . Scrooge thinks that to get a slight increase in equity, one must give up a lot of effi ciency. On the other hand, Spendthrift thinks that one can get a large increase in equity with just a slight loss in effi ciency.

For instance, if the main reason that unemployed individuals do not obtain jobs is that there are no jobs available, then the size of unem- ployment insurance may have little eff ect on their search. If unem- ployment insurance has little eff ect on job search, there is not much trade-off between effi ciency and equity, and the frontier is consistent with Spendthrift’s perceptions. However, if job search is very sensitive to unemployment compensation, there is a signifi cant trade-off , and the equity–effi ciency frontier is consistent with Scrooge’s perceptions.

289Equity–Efficiency Trade-Offs

The equity–effi ciency trade-off is encountered repeatedly in the evaluation of the detailed provisions of any government program. The decision to charge tolls on a bridge means that those who benefi t from the bridge (that is, those who use it) have to bear its costs. To many peo- ple, this is desirable for equity reasons; it is unfair to make someone who does not drive over the bridge pay for it. However, there is an effi - ciency cost in money and time: the wages of toll collectors and the time of motorists. Moreover, if some drivers are discouraged from using the bridge (when it is below capacity), there is a further effi ciency loss from underutilization.

SOURCES OF DIFFERENCES IN VIEWS CONCERNING PUBLIC PROGRAMS

(A) Scrooge and Spendthrift have the same perceptions concerning trade-offs but differ in values (indifference curves). (B) Scrooge and Spendthrift differ in their perception of the nature of the equity–effi ciency trade-off.

FIGURE 10.6

Equity

Efficiency

Scrooge’s indifference

curve

Equity– efficiency

frontier

Spendthrift’s indifferene

curve

E2

E1

Equity

Efficiency

Spendthrift’s perception of trade-off

Scrooge’s perception of trade-off

A

B

290 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

When programs are ineffi cient and not well designed, we can get more of both equity and effi ciency. Later in this book, for instance, we describe attempts to help more Americans go to college by providing tax credits. Tax credits are, however, not typically the best way of accom- plishing those objectives. Furthermore, they do not benefi t the very poor—those from fami- lies that are so poor that they pay little, if any, taxes. Thus, there are alternative ways of pro- viding assistance that are both more equitable and more effi cient.

This example illustrates another theme: there is no single measure that captures all the dimensions of equity. Some programs may help more those of the middle class; others help those at the bottom. Both may improve “equity,” but in two quite diff erent senses, and these pose some of the hardest political choices.

PUBLIC POLICY OBJECTIVES

The discussion so far has focused on two bases for evaluating public programs: their eff ect on economic effi ciency and their eff ect on distribution. However, government policy may be concerned with a broader range of objectives. For instance, government may be concerned with the extent to which individu- als of diff erent racial, ethnic, and class backgrounds are mixed together in schools. It may be concerned not just with the income of the poor, but also with the physical appearance of the housing in which they live. When these alternative objectives are fairly well defi ned, the government can still make use of a variety of instruments for attaining them. It can, for instance, still make use of private producers by imposing regulations on them or setting standards that must be met for individuals or fi rms to be eligible to receive subsidies. Thus, the government has specifi ed that institutions receiving fed- eral grants must comply with certain affi rmative action regulations.

In some cases, however, it may be diffi cult for the government to specify clearly (and in advance) all of its objectives, or to articulate them in the form of a set of regulations or standards. There is widespread belief that private producers, in the absence of well-articulated and enforced regulations, will simply pursue profi t-maximizing behavior, regardless of alternative objec- tives that they may affi rm. In such circumstances, there is an argument for the government to assume direct responsibility for the activity. To the extent

EQUITY–EFFICIENCY TRADE-OFF 1. Sometimes programs can result in Pareto

improvements, making some people better off without making anyone worse off.

2. More typically, there are trade-offs between equity and effi ciency; more progressive tax systems reduce marginal incentives to work.

291Political Process

that this is true, however, it may be diffi cult for Congress or the executive branch to clearly specify the objectives it wishes the bureaucrats responsible for implementing its programs to pursue. In that case, the bureaucrats will be left with considerable discretion, and the discrepancies between how they exercise that discretion and the intent of, say, Congress may be signifi cant.

Similarly, there is concern that whenever the government fi nances an activity, it will almost inevitably impose a set of regulations, some of which may have adverse eff ects, particularly on economic effi ciency; thus, many of the alleged effi ciency advantages of private production may be lost. These concerns have been raised, for instance, in discussions of school voucher programs, which would provide students with funds that could be used at any school, private or public.

POLITICAL PROCESS

In a democracy, the design and adoption of any public expenditure pro- gram involves many individuals and groups with various objectives and various beliefs about how the economy works. The program that eventu- ally is adopted represents a compromise among their views; it probably will not conform to the views of any one individual and may seem to be inconsistent with any single set of objectives. If two chefs, say, disagree about the appropriate liquid to add to a sauce, one arguing for lemon juice and the other for cream, the compromise solution of adding a little of both may be disastrous, with results inconsistent with any culinary objective.

The study of the political process by which a particular expenditure program was adopted may be insightful for several reasons. First, we may be able to understand why the program looks the way it does. Con- sider the government program to stabilize farmers’ prices. This program addresses a market failure: the inability of individuals to obtain insurance for many of the important risks they face, including the risks associated with the variability of prices.6 A closer examination of the price stabili- zation program, however, suggests that if that were the only objective, it would be designed in a quite diff erent way. What farmers care about is income risk—not just the variability of price, but the variability of all the factors that go into determining net income, including output and costs. In some cases, the price stabilization program may actually increase income variability. In fact, reducing risk is probably not the true objective: the

6�Futures markets now enable farmers to divest themselves of some of the risks associated with price variability.

292 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

real objective is to transfer resources (income) to farmers from the rest of the population. If that is the objective, there are more effi cient ways of transferring resources to the farmers; outright grants would be preferable to the present program. However, if that objective were made explicit—if the transfers were made conspicuous—they might not get approved. Vot- ers in urban districts might strongly oppose them, even though they do not oppose the present form of ineffi cient subsidies, simply because they are not fully aware of the nature of the transfers.

This example also highlights the dilemma that there are many dif- ferent views of fairness. Some believe that anything they have earned is theirs, and it is unfair for the government to take it away; others recognize, perhaps more reasonably, that no one could have gotten to where he or she is without a vast array of help from others—we all depend on a variety of public services, including the rule of law, without which businesses could not fl ourish. Diff erences in these perspectives—and the diffi culties of each group to understand others’ perspectives—are a major impediment to reaching political consensus around a variety of issues.

Particular provisions of public programs are likely to have strong dis- tributional consequences for particular groups in the population. If one group can be suitably organized, it will attempt to induce the political process to adopt provisions that are to its benefi t. In Chapter 6, we dis- cussed the regulations that provide for scrubbing the smoke emitted from burning coal. These regulations may have an enormous eff ect on the rel- ative demand for hard (or western) coal and bituminous coal, and hence on the incomes of both miners and coal producers in diff erent parts of the country. The shape of environmental legislation and regulation may be aff ected as much by these particular distributional consequences as by overall effi ciency considerations.

A second reason why it may be helpful to study an expenditure pro- gram’s adoption process is that in democracies, programs respond at least in part to the desires and perceptions of voters. Because programs have to be explained and “sold” to voters, there is a premium on simplicity. In addition, programs often look diff erent from the way that economists think they should be designed because voters often do not understand the true incidence of a program. For instance, most voters think that half the cost of Social Security is paid for by contributions from the employer; most economists believe that the true incidence is the same as it would be if Social Security contributions were paid entirely by workers. In this case, the confusion over incidence has few consequences, but in many other programs, this confusion can have signifi cant impacts, as we shall see in later chapters.

293Political Process

Equally important, however, the political process is shaped by special interests, and the programs that emerge refl ect those special interests typically more than the rhetoric that is used to justify them. Consider America’s farm programs discussed earlier. Such programs are often justifi ed as helping poor family farmers—but the way that they are designed provides most of the subsidy to large farms, many of which are corporate farms, not family farms at all. Many ways of rede- signing the programs would target more of the money on small, family farms, but such reforms are not likely to get the political support they need to be adopted.

Indeed, typically special interests try to enlist others in their cause, trying to persuade them that they are the real benefi ciaries. The most egregious example of this was the corn ethanol subsidy, which was mainly of benefi t to the ethanol producers, including Archer Daniels Midland (ADM), a large agro-business. ADM persuaded corn farmers that they were the benefi ciaries, and with their support, this costly subsidy was adopted in 1978. In fact, though, in the early days of ethanol, the amount of corn used was so small that it had a negligible eff ect on corn prices and producers. ADM also tried to persuade environmentalists that it was good for the environment, but they were not so easily duped. The subsidy was supposed to be short-lived, just enough to help a new industry get started, but, in fact, it lasted for more than thirty years before being abol- ished at the end of 2011.

Finally, the design of programs may aff ect the extent to which they are subjected to political pressures or corruption. Corruption is an increas- ing concern in many countries. It can take a variety of forms. In mod- ern democracies, special interests contribute to campaigns, often in an attempt to “buy” legislation that favors them; in many countries, bureau- crats use their discretionary powers to extend favors in return for bribes. In New York City, there have been extensive reports of bribes to build- ing inspectors, more to ensure that they inspect the building in a timely way (so there will not be costly interruptions to construction) than to give approval to a substandard building. The more discretion that is left to bureaucrats, the more potential there is for the exercise of political infl u- ence and corruption.

Accordingly, in evaluating alternative policies, one needs to take into account the political process, what the legislation might look like after it has been subjected to the political process, and what the consequences of the program will be, knowing that it will be administrated by bureau- crats, probably not unlike those administering other government pro- grams, and subject to the same kinds of incentives.

294 CHAPTER 10 FRAMEWORK FOR ANALYSIS OF EXPENDITURE POLICY

SUMMARY

There are ten major elements in the analysis of public expenditure programs:

1. Identifying a need, the source of demand for the government program

2. Identifying a market failure (if it exists) and ascertaining whether what is at issue is a con- cern for (the consequences of) the distribution of income or the provision of a merit good

3. Identifying alternative programs that might address the perceived problems

4. In ascertaining and evaluating the impacts of alternative programs, paying attention to the importance of particular design features

5. Identifying private sector responses

6. Identifying the effi ciency consequences of alter- native programs

7. Identifying the distributional consequences of alternative programs

8. Identifying the trade-off s between equity and effi ciency considerations

9. Identifying the extent to which alternative pro- grams achieve public policy objectives

10. Identifying how the political process aff ects the design and implementation of public programs

KEY CONCEPTS

Capitalization

Crowding in

Crowding out

Incidence

Income effects

Intertemporal distribution effects

Progressive

Regressive

Shifting

Substitution effects

QUESTIONS AND PROBLEMS

1. Explain how the following actual design features have an important eff ect on the consequences of government programs:

a. The income ceiling for eligibility for SNAP benefi ts is reduced by expenditures on housing.

b. Until recently, whether an individual between ages sixty-fi ve and seventy was eligible for Social Security benefi ts depended on his or her income calculated on a month-by-month basis.

c. An ex-spouse becomes eligible for Social Secu- rity benefi ts only if the marriage lasted at least ten years.

Can you think of other instances in which partic- ular design features have seemingly unintended consequences?

2. Who may be the actual benefi ciaries of the following government program or proposed programs? That is, taking into account how indi- viduals respond to the government program, who is actually better off as a result of the program?

a. Medicare

b. Housing subsidies for the poor

c. Education loans

Can you think of other programs whose actual benefi ciaries may diff er from those the program seemingly intended?

3. In Chapters 12 to 16, we will use the framework we have discussed in this chapter to analyze several diff erent government programs. Before reading those analyses, see if you can answer the following questions for each program:

a. What were the original sources of demand for the program? What perceived need was it intended to address?

b. What are the market failures that gave rise to the program?

REVIEW AND PRACTICE

295Review and Practice

c. What are the possible forms of government intervention? Are there particular design features that have had, or currently have, an important impact on the program’s eff ec- tiveness? How do private sector responses weaken, or reinforce, the intended eff ects of the program? What is the true incidence of the program?

d. What are the major effi ciency consequences of the program?

e. Does the program entail any eff ective redistri- bution of income?

f. Are there important instances of trade-off s between equity and effi ciency in the program’s design?

g. What are some alternatives for meeting the program’s objectives? To what extent might they do a better job, for example, by reducing distortions and increasing the equity of the programs?

h. How has the political process aff ected the nature of the present program?

4. Draw the budget constraint between housing and “other consumption” for an individual on SNAP, where the amount of food assistance the individ- ual receives depends on his or her income net of housing costs. Does it make a diff erence whether the individual is consuming an amount of food equal to, less than, or greater than the SNAP allotment?

5. State governments eff ectively subsidize tuition in state universities and colleges. How might this aff ect the amount of education that indi- viduals get? Is there a substitution eff ect? Is there a market failure that this program might be addressing? Are there alternative ways of addressing it?

296

EVALUATING PUBLIC EXPENDITURE

The preceding chapter set out the basic framework for analysis of govern- ment expenditure policies. In many cases, government wants more than a qualitative analysis; it needs a quantitative analysis. It needs to know not only that there is a rationale for government action; it also needs to know whether the benefi ts of the particular government action (project, regula- tion) will exceed the costs. For example, should the government:

• Build a bridge—and, if so, of what size? • Construct a dam—and, if so, of what size? • Institute more stringent regulations for fl ammability of mattresses? • Institute more stringent regulations for licensing drugs? • Extend the Washington, DC subway system? • Declare certain portions of the Cape Cod seashore a national park?

After expenditures are authorized, government also needs to know whether public funds are expended effi ciently and eff ectively. It needs to assess past government expenditure to improve future government performance.

This chapter describes how the government goes about making these evaluations. First, however, it is instructive to consider how a private fi rm makes decisions concerning which projects to undertake.

11

297Private Cost–Benefit Analysis

PRIVATE COST–BENEFIT ANALYSIS

Private fi rms continually must decide whether to undertake investments. The procedures they follow can be characterized in four steps:

1. Identify the set of possible projects to be considered. If a steel fi rm wishes to expand its production capacity, there may be a number of ways it can do this. There may be alternative technologies available for smelting iron ore, and there may be a number of alternative specialized forms of steel that can be produced. The fi rst stage, then, is to list the various major alternatives.

2. Identify the full consequences of each alternative. The fi rm is concerned primarily with its inputs and outputs. Thus, it will determine the labor, iron ore, coal, and other materials required for each production alter- native; it will assess the quality of steel that will be produced under each alternative; it will determine the quantity of various wastes that will be produced.

3. Assign a value to each input and output. The fi rm will estimate the costs of various kinds of labor (with various skills) over the lifetime of the plant; the costs of other inputs, such as coal and iron ore; the prices at which it can sell the steel (which will depend on the quality of the steel produced, which may in turn vary from project to project); and the costs of disposing of wastes.

4. Add up the costs and benefi ts to estimate the total profi tability of the proj- ect. The fi rm will undertake the project with the highest profi t (the maximum diff erence between benefi ts and costs)—provided, of course, that profi ts are positive (taking appropriate account of the opportunity costs, the return the fi rm’s resources could obtain elsewhere). If profi ts for all contemplated projects are negative, the fi rm will undertake no project; it will invest its funds elsewhere.

PRESENT DISCOUNTED VALUE

The procedures just described seem simple and straightforward. Only one part requires some elaboration. The benefi ts and costs of the steel mill occur over an extended period of time. Surely the fi rm is not indiff erent

1. What is cost–benefi t anal- ysis, and why is it useful? What are the basic steps in cost–benefi t analysis?

2. How does private cost– benefi t analysis diff er from social cost–benefi t analysis?

3. What is consumer surplus and what role does it play in cost–benefi t analysis?

4. How does the govern- ment value nonmarketed benefi ts of a project, such as time or lives saved?

5. What discount rates should be used for valuing future benefi ts and costs in social cost–benefi t analysis?

6. How should risks be treated in cost–benefi t analysis? How should distributional concerns be brought into the analysis?

7. How can government assess past expenditure effi ciency and eff ective- ness? In what ways can post-expenditure evalua- tion improve future public sector performance?

FOCUS QUESTIONS

298 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

when it comes to choosing between receiving a dollar today and receiving one in twenty-fi ve years. How are the benefi ts and costs that accrue at diff erent dates to be valued and compared?

The method used is based on the premise that a dollar today is worth more than a dollar tomorrow. Suppose the interest rate is 10 percent. If the fi rm receives $1 today, it can take that dollar down to the bank, deposit it, and have $1.10 at the end of the year. Thus, $1 today is worth $1.10 next year. The fi rm is just as well off receiving $1 today as $1.10 next year. If the fi rm invests the $1.10, it will have $1.21 at the end of the following year. Accordingly, the fi rm is indiff erent between receiving $1 today and $1.21 in two years’ time.

To evaluate projects with receipts and expenditures in future years, the fi rm multiplies those receipts and payments by a discount factor, a number (less than 1) that makes the future receipts and payments equiv- alent to current receipts and payments. The discount factor is smaller the further into the future the benefi t is received. The discount factor for pay- ments in one year is just 1/1 1 r, where r is the rate of interest1 (in our example, r 5 0.10, so the discount factor is 1/1.1 5 0.9); for payments in two years’ time, it is 1/(1 1 r)2 (in our example, it is 1/1.21). The value today of $100 to be received two years in the future is thus $100/1.21 5 $82.64. We then add up the value of what is to be received (or paid out) in each year of the project. The sum is called the present discounted value (PDV) of the project. If Rt is the net receipts from the project in period t, and r the rate of interest, then if the project lasts for N years, its PDV is given by

PDV 5 R0 1 R1

1 1 r 1

R2 (1 1 r)2

1 · · · 1 Rt

(1 1 r)t 1 · · · 1

RN (1 1 r)N

�.

Table 11.1 provides an illustration of how this might be done for a hypo- thetical steel mill lasting fi ve years. (Most steel mills last much longer than that; this makes the calculations more complicated, but the principle is the same.) For each year, we multiply the net receipts of that year by the discount factor for that year. Notice the large diff erence between undis- counted profi ts ($1000) and discounted profi ts ($169). This diff erence is likely to be particularly large for long-lived projects entailing large initial investments; the benefi ts for such projects occur later in time (and there- fore have lower value) than the costs, which occur earlier in time.

�1 To see this, compare what the fi rm has at the end of the year if it receives $100 3 1/(1 1 r) today. It takes the $100 3 1/(1 1 r) and invests it, receiving a return of r. Thus, at the end of the year, it has

$100 3 1/(1 1 r)���the original amount

1 r 3 $100 3 1/(1 1 r)� �the interest

(1 1 r) 3 [$100 3 1/(1 1 r)] 5 $100.

Therefore, the fi rm is indiff erent between receiving $100 3 1/(1 1 r) today and $100 next year.

299Social Cost–Benefit Analysis

SOCIAL COST–BENEFIT ANALYSIS2

The government goes through basically the same procedures as the fi rm in evaluating a project. There are, however, two critical diff erences between social and private cost–benefi t analysis.

First, whereas the only consequences of a project that are of concern to the fi rm are those that aff ect its profi tability, the government may be concerned with a much broader range of consequences. For example, it may be concerned with the ecological eff ects of a dam and with the impact of the dam on the river’s recreational uses.

Second, whereas the fi rm uses market prices to evaluate what it has to pay for its inputs and what it receives for its outputs, there are two instances in which the government might not use market prices in evaluating projects:

1. When the outputs and inputs are not sold on the market, market prices do not exist. Market prices do not exist, for instance, for clean air, for lives saved, or for the preservation of wilderness in its natural state.

2. When there is a market failure, market prices do not represent a proj- ect’s true marginal social costs or benefi ts. The prices the government

2 There is a substantial body of literature on social cost–benefi t analysis. A collection of 32 seminal articles written from 1963 to 1999 are compiled by two of the leading authorities on this topic in A. Harberger and G. Jenkins, eds., Cost–Benefi t Analysis, vol. 152, International Library of Critical Writ- ings in Economics (Cheltenham, UK, and Northampton, MA: Edward Elgar, 2002). Four classic books are: E. Mishan, Cost–Benefi t Analysis: An Introduction (New York: Praeger Publishers, 1971); I. Little and J. Mirrlees, Project Appraisal and Planning for Developing Countries (New York: Basic Books, 1974); L. Squire and H. van der Tak, Economic Analysis of Projects (Baltimore: Johns Hopkins University Press, 1975); and M. Roemer and J. Stern, The Appraisal of Development Projects: A Practical Guide to Proj- ect Analysis with Case Studies and Solutions (New York: Praeger Publishers, 1975). Three more recent works are: E. Gramlich, A Guide to Benefi t–Cost Analysis, 2nd ed. (Long Grove, IL: Waveland Press, 1997); R. Zerbe and A. Bellas, A Primer for Benefi t–Cost Analysis (Cheltenham, UK, and Northampton, MA: Edward Elgar, 2006); and A. Boardman et al., Cost–Benefi t Analysis: Concepts and Practice, 4th ed. (Upper Saddle River, NJ: Pearson–Prentice Hall, 2010).

TABLE 11.1 HYPOTHETICAL CALCUL ATION OF PROFITABILIT Y FOR A FIVE-YE AR S TEEL MILL

BENEFITS NET DISCOUNTED YEAR (RECEIPTS) COSTS NET PROFITS DISCOUNT FACTOR PROFITS

1 3000 23000 1 23000 2 1200 200 1000 1/1.1 = .909 909 3 1200 200 1000 1/(1.1)2 = .826 826 4 1200 200 1000 1/(1.1)3 = .751 751 5 1200 200 1000 1/(1.1)4 = .683 683

Total 4800 3800 1000 169

300 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

uses to evaluate its projects must refl ect the market failure. (Recall from Chapter 3 that in the absence of market failures, market prices do refl ect marginal social costs and benefi ts; accordingly, in the absence of market failure, the government should use market prices in evaluating its projects.)

Social cost–benefi t analysis is concerned with developing systematic ways of analyzing costs and benefi ts when market prices do not refl ect social costs and benefi ts. In the following sections, we will look at how the government values bene- fi ts that are typically not monetized—such as the value of the environment, or of lives—and how the

government values marketed goods and services when there are reasons to believe that important market failures exist, such as massive unemployment, which result in market prices that do not refl ect social benefi ts and costs.

CONSUMER SURPLUS AND THE DECISION TO UNDERTAKE A PROJECT

Before turning to these issues, however, there is one other set of situa- tions in which cost–benefi t analysis plays an important role. Even when the price system is working well, so prices refl ect marginal benefi ts and costs, a project may not break even—and thus would not be provided by the market—and yet total benefi ts exceed costs, so the project should be under- taken. Typically, these are projects that have large fi xed costs, such as a bridge, or, more generally, projects that are large enough to have an eff ect on prices. Thus, market prices can be used for valuing projects only when projects are suffi ciently small that they have a negligible eff ect on prices. In the case of a bridge, although it may leave prices in general unchanged, the “price” of crossing the river at that particular place can be thought of as being reduced from infi nite (the good simply is not available) to zero.

Figure 11.1 shows the demand curve for a bridge. Even at a price of zero, only a certain number of trips across the bridge will be taken— denoted by point E. Assume that the capacity of a minimal-size bridge is C, which exceeds E, and that the marginal cost of using the bridge is zero. Then  effi cient utilization of the bridge requires a zero toll (price); any

MA JOR DIFFERENCES BET WEEN SOCIAL AND PRIVATE COST–BENEFIT ANALYSIS 1. Social cost–benefi t analysis takes into account a

wider range of impacts, not just profi ts.

2. In social cost–benefi t analysis, market prices may not exist for many benefi ts and costs, and mar- ket prices may not be used because of market failures (so market prices do not refl ect marginal social benefi ts and costs).

301Consumer Surplus and the Decision to Undertake a Project

higher price will restrict usage, when the marginal cost of usage is zero. Clearly, though, at a zero price, no private fi rm would build the bridge.

Even though the marginal value of a trip is zero, however, the total value of the bridge is clearly positive. The question is: Is the total value large enough to off set the costs of the bridge? To fi nd the total value of the bridge, we ask a simple question: How much, in total, would individuals be willing to pay to have the bridge (with a toll of, say, zero)? As we saw in Chapter 7, the total amount that individuals would be willing to pay in excess of what they have to pay is called the consumer surplus. There, we showed that consumer surplus is measured as the diff erence between the area under the compensated demand curve and what consumers actually have to pay for the good.3

3 Recall that along the compensated demand curve, the individual’s welfare (utility) is constant. The compensated demand curve tells us the quantity of the good that the individual demands at each price if, as the price is lowered, we take away just enough income to leave the individual no better off as a result of the price decrease. If the individual consumes relatively little of the good, then the amount we have to take away is relatively small. The diff erence between the compensated and uncompensated demand is the result of the “income eff ect”—the change in demand (here, for trips) from taking away this small amount of income. Accordingly, the diff erence between the compen- sated and uncompensated demand curves for an item like a bridge is typically small. See R. Willig, “Consumer’s Surplus without Apology,” American Economic Review 66 (1976): 589–597. Obviously, in other cases, such as the supply of labor (demand for leisure), the diff erence could be large. See J. Hausman, “Exact Consumer’s Surplus and Deadweight Loss,” American Economic Review 71 (1981): 662–676. Whether or not economists should ignore the income eff ect, in practice, they frequently do because of the diffi culties in quantifying its magnitude. (As a matter of terminology, the area under the ordinary demand curve is often called the consumer surplus, as opposed to the exact consumer’s surplus, which is the area under the compensated demand curve. The exact consumer’s surplus is what is relevant for project evaluation.)

EFFICIENT UTILIZATION OF A BRIDGE

If the minimum scale capacity for a bridge, C, exceeds the demand at a zero price, E, then effi ciency requires that no toll be charged, but it still may be worth constructing the bridge.

FIGURE 11.1Price

Demand curve for a bridge

CE Number of trips

302 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

Figure 11.2 shows the compensated demand curve—everywhere along this line, each individual’s welfare is the same as it would have been if no bridge had been constructed. If no tolls were charged, the consumer surplus would be the entire area under the demand curve, the area of the triangle AFE. This would measure the entire amount that all individuals would be willing to pay to have the bridge—say, the value of the savings in time and driving costs of using that bridge rather than using the bridge one mile downstream. If a toll P is charged, then the total people are will- ing to pay still exceeds the amount actually paid by the amount of the triangle AGB.

The decision to build the bridge is then a simple one: Do the total bene- fi ts (revenues plus consumer surplus) exceed the total costs, including any costs incurred in raising the revenue to fi nance the bridge?�4

Sometimes economists look at the ratio of benefi ts to costs. The crite- rion for undertaking any project for which benefi ts, B, exceed costs, C,

Undertake a project if B . C,

can be rewritten: Undertake any project for which the ratio of benefi ts to costs exceeds unity:

Undertake a project if B C

. 1.

4 There is one subtlety: the total costs should include not only the expenditures on the bridge, but also the additional costs associated with raising the required tax revenues to fi nance it.

CALCULATION OF CONSUMER SURPLUS

The consumer surplus is the area under the (compensated) demand curve. If a zero toll is

charged, the bridge should still be constructed if the consumer surplus exceeds the cost of the bridge. (If a toll of P is charged,

then the consumer surplus is the area AGB, and the bridge

should be constructed if the consumer surplus, which is now

only AGB, plus the revenues raised, FGBQ, exceed the

cost of the bridge.)

FIGURE 11.2

Number of trips

Price

Demand curve

CEQF

P

A

B G

303Measuring Nonmonetized Costs and Benefits

Often, governments must choose which project among several to undertake. There may be a dam site on which only one dam can be constructed. Diff erent dams may have diff er- ent benefi ts and costs. In this particular case, the government should undertake the project in which the total net benefi t from the project, the diff erence between the benefi ts and costs, is largest.

Choosing the project that maximizes the diff erence between benefi ts and costs is not the same as choosing the project that maximizes the cost–benefi t ratio. A very small project with a small benefi t and an even smaller cost could have a very high cost–benefi t ratio, yet yield rel- atively small net benefi ts.5

MEASURING NONMONETIZED COSTS AND BENEFITS

For many of the costs and benefi ts associated with government projects and regulations—such as lives saved, or environments protected—there are no market prices. Economists, therefore, have developed system- atic procedures for estimating these values. In some cases, such as the value of time, we can make inferences about individuals’ evaluations from market data and from their observed behavior in other contexts. In other cases, such as the value of the Grand Canyon, survey tech- niques have been employed. Many of these valuation techniques remain controversial.6

5 The problem is that we have not included in our cost measure the opportunity cost of the dam site. If we correctly calculate the opportunity cost of the dam site, then there will be only one project for which the benefi ts exceed the total costs; this is, of course, the same project that we identifi ed before as maximizing the net benefi ts from the dam site. 6 For more on determining the benefi ts of environmental goods, see M. Cropper and W. Oates, “Envi- ronmental Economics: A Survey,” Journal of Economic Literature 30 (June 1992): 675–740; “Symposium on Contingent Valuation,” Journal of Economic Perspectives 8, no. 4 (Fall 1994): 3–64; “Economic Valu- ation of the Environment: A Special Issue,” Environmental Science & Technology 34, no. 8 (April 2000): 1381–1461; and L. Wainger and M. Mazzotta, “Realizing the Potential of Ecosystem Services: A Frame- work for Relating Ecological Changes to Economic Benefi ts,” Environmental Management 48 (October 2011): 710–733.

COST–BENEFIT ANALYSIS: CRITERION FOR ACCEPTING PROJECTS • A project should be undertaken if its total bene-

fi ts exceed total costs, or if the cost–benefi t ratio exceeds unity. Total benefi ts include the consumer surplus, the difference between what individuals would have been willing to pay and what they have to pay.

• If the government must choose one from among a set of projects (e.g., several alternative designs for a dam to construct), it should choose the project with the highest net benefi ts, not the highest cost– benefi t ratio.

304 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

VALUING TIME

The old adage, “time is money,” describes how most economists evaluate the savings in time resulting from an improved transportation system, such as a better subway system or road network. The typical approach is to attempt to ascertain the wage rate of those who use the transportation system; under certain ideal conditions, the wage provides a measure of an individual’s evaluation of his or her own time. In simple economic mod- els, an individual is pictured as making a choice between the amount of leisure and the amount of work undertaken. As a result of giving up one more hour of leisure, the individual gets an increase in consumption goods equal to his or her hourly wage. In equilibrium, the individual is indiff erent when choosing whether to give up one more hour of leisure and increase consumption by an amount equal to his or her hourly wage, or to reduce work (increase leisure) by an hour and decrease consumption by an amount equal to his or her hourly wage. Thus, the individual’s wage provides a mon- etary valuation of his or her time. If a faster subway reduces commuting time by twenty minutes, and the wage is $9 an hour, the value of the time saved is $3. We calculate the value of time saved by each individual and add the values together to obtain the total value of time saved.

Some claim that this overestimates the value of time: many individu- als would like to work more at their wage rate but are unable to fi nd addi- tional employment at that wage; the job restricts the number of hours that they can work. The individual’s valuation of his or her leisure is thus fairly low; the compensation that would be required for reducing an individual’s leisure by one hour is, in this view, much less than the wage that the indi- vidual receives for the work that he or she is able to obtain.

Others claim that the wage may underestimate the value of leisure for some individuals and overestimate the value of leisure for other individuals. They point out that professors, for instance, have chosen a comparatively low- wage job relative to other options available to them because of the great non- monetary benefi ts associated with the job. The value of their leisure exceeds the wage they receive. On the other hand, the wage of the coal miner or the garbage collector includes some compensation for the unattractive features of those jobs and hence represents an overestimate of the value of leisure.

VALUING LIFE

Probably no subject in public cost–benefi t analysis has engendered as much emotional discussion as economists’ attempts to place a monetary value on life. As distasteful as such a calculation may seem, it is necessary,

305Measuring Nonmonetized Costs and Benefits

CHILDREN, CAR SAFETY, AND THE VALUE OF LIFE

*For surveys of the value of life, see pages 713–715 in M. Cropper and W. Oates, “Environmental Economics: A Survey,” Journal of Economic Literature 30, no. 2 (June 1992): 675–740; Peter Dorman, Markets and Mortality: Economics, Dangerous Work, and the Value of Human Life (Cambridge, UK: Cambridge University Press, 1996); and W. Kip Viscusi, “The Value of Life: Estimates with Risks by Occupation and Industry,” Economic Inquiry 42, no. 1 (January 2004): 29–48. For a detailed review of both the theory and application of VSL and VSLY, see L. Robinson, “How US Government Agencies Value Mortality Risk Reductions,” Review of Environmental Economics and Policy 1 (Summer 2007): 283–299.

E ven though life is priceless, economists have developed several alternative methodologies to put a dollar number on it (see the “Valuing Life” section of this chapter). Different studies pro- duce different results, with a range of $1 to $25 mil- lion, but most studies show numbers at the lower end of that range, between $3 and $9  million (in 2012 dollars).*

The U.S. government has debated whether to use a single number for cost–benefi t analyses in all agencies. So far, different agencies use differ- ent numbers—with the Environmental Protection Agency typically using numbers considerably larger than those of most outside studies to estimate the value of a statistical life (VSL).

Those who argue against using a single num- ber suggest that a number of other factors should go into the analysis, such as whether the death resulted from an action voluntarily engaged in (such as driving).

One of the most diffi cult questions in valuing lives is whether the life of a child should be valued differently from the life of middle-aged adult, or of an 80-year-old. The issue comes up repeatedly: How should we allocate money between two cancer research programs, one targeting a cancer typically found in children and the other a form of cancer that typically only shows up among the elderly?

The Department of Transportation (DOT) is responsible for imposing regulations to ensure the safety of cars; in imposing these regulations, it looks at the costs and benefi ts. In the mid-1990s, DOT  addressed whether to strengthen the stan- dard for car frames that would largely affect deaths in the rear seat in side collisions. A dispro- portionate fraction of those saved would be chil- dren (as children more often ride in the rear seat). This raised the issue of whether to employ a differ- ent value of life instead of the one used elsewhere, when those saved were more typically adults. When a child’s life is saved, more “life-years” are saved than when the life of an 80-year-old is saved.

DOT has continued to use the traditional method for valuing life, which treats all lives the same. Even so, an alternative methodology, which focuses on life-years saved rather than lives, was sanctioned in new federal guidelines established in 2003 with the issuance of Offi ce of Management and Budget (OMB) Circular A-4, “Regulatory Anal- ysis.” OMB allows use of the value per statistical life-year (VSLY) approach for adjusting VSL esti- mates to refl ect differences in remaining life expec- tancy. This entails calculating the value of each year of life extension, which adjusts VSL to refl ect age differences.

306 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

in a variety of circumstances, for governments to face up to this problem. There is virtually no limit to the amount that could be spent to reduce the likelihood of an accident on a road, or of death from some disease. At some point, however, a judgment must be made that the gain from fur- ther expenditures is suffi ciently small that additional expenditures are not warranted. An individual may die as a result of this decision—yet we cannot spend 50 percent of our national income on transportation safety, or 50 percent of our national income on health.

Two methods have been used for estimating the value of life. The fi rst is the constructive method—that is, we estimate what the individual would have earned had he or she remained alive (until the “normal” age of death). To do this, we extrapolate the individual’s employment history, comparing it to that of those in similar positions.

This method, however, fails to distinguish between the value of life and the livelihood that goes with it. It thus suggests that after retirement, an individual’s life has zero value, as there would be no loss of earnings. This seems clearly wrong. (It confuses means and ends: income is earned to provide consumption; producing income is not presumably the object of life, and therefore not the basis of valuing it.)7

There is an alternative, indirect method that does recognize the natural desire to live longer. In some occupations, there is a much higher chance of death than in others. For instance, the accident rates for coal miners are higher than those for college professors, and the death rates for workers in asbestos factories and jackhammer operators are much higher than those for clerical workers. Individuals who undertake riskier occupations nor- mally require compensation for assuming these additional risks. By choos- ing the riskier occupation, they are saying that they are willing to face a higher chance of death for a higher income while they are alive.

The second method calculates the value of life by looking at how much income individuals are willing to expend to avoid potentially fatal risks, or need to compensate them for an increase in the chance of death. This is the approach used by the federal government, referred to as the “value of a statistical life” (VSL). For example, the Environmental Protection Agency uses an estimate of $7.4 million (in 2006 dollars), updated to the year of analysis, for how much people are willing to pay for small reduc- tions in their risks of dying from adverse health conditions that may be caused by environmental pollution, regardless of the age, income, or other population characteristics of the affected population.8 This is the same

7 For an early critique of this method and one of the fi rst developments of the second, indirect method, see T. Schelling, “The Life You Save May Be Your Own,” reprinted in T. Schelling, Choices and Conse- quences (Cambridge, MA: Harvard University Press, 1984). 8 For a detailed explanation of EPA’s approach to mortality risk valuation, see http://yosemite.epa.gov/ ee/epa/eed.nsf/pages/MortalityRiskValuation.html.

307Measuring Nonmonetized Costs and Benefits

approach used by Joseph Stiglitz and Linda Bilmes in their analysis of the full budgetary and economic costs of the Iraq war—they assigned a VSL of $7.2 million (in 2007 dollars) to each death of a U.S. service member, and a prorated fraction of that amount to each serious injury.9

There is considerable controversy about this second method, however, just as there is about the fi rst. Some believe that it grossly underesti- mates the value of life; they argue that individ- uals are not well informed concerning the risks they face.10 Also, for well-known psychological reasons, individuals attempt to ignore the information they do have con- cerning the riskiness of their jobs.

As controversial as the estimates of the value of life may be, there appears to be no alternative to using them in the evaluation of projects that aff ect the likelihood of death.

VALUING NATURAL RESOURCES

A question of increasing concern is how to value impacts on the environ- ment. That issue was raised forcefully by the 1989 Exxon Valdez oil spill. Network news reports on the millions of dying otters, salmon, and birds brought into every American home the impact of the spill in a relatively remote area. If people had died, it is clear that Exxon would have owed the families huge amounts of money in compensation. Obviously, the relatives of the animals that had been killed had no standing in court to demand compensation. Many Americans, however, felt that Exxon should pay something, both to deter others from taking actions that might dam- age the environment and to compensate them for their perceived loss from the environmental injury. Using a relatively new technique called contingent valuation, courts valued the compensation that Exxon would have to pay at approximately $1 billion. This was compensation that went beyond the direct economic injury, for instance, to fi shermen who lost their livelihood.

9�J. Stiglitz and L. Bilmes, The Three Trillion Dollar War: The True Cost of the Iraq Confl ict (New York: W. W. Norton & Company, 2008). 10�Several studies have attempted to estimate the magnitude of workers’ misperceptions and suggest that they may not be too large. See, for instance, W. K. Viscusi, Risk by Choice: Regulating Health and Safety in the Workplace (Cambridge, MA: Harvard University Press, 1983). For a general analysis of technical challenges in applying VSL, see O. Ashenfelter, “Measuring the Value of a Statistical Life: Problems and Prospects,” Institute for the Study of Labor (IZA) Discussion Paper No. 1911, January 2006.

ALTERNATIVE METHODS OF VALUING LIFE 1. Constructive method: What would the individuals

have earned had they remained alive?

2. Revealed preference method: How much extra income do individuals need to compensate them for an increase in the chance of death, as refl ected in market wages for riskier jobs?

308 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

In contingent valuation, individuals are asked a series of questions intended to elicit how much they value the environmental damage or the preservation of some species. Many (but not all) individuals seem will- ing to pay something, for instance, to preserve whales or the spotted owl or other endangered species, or the Arctic National Wildlife area, even if they themselves do not directly come into contact with the species or do not visit the preservation area. These values are referred to as existence values. Even if each individual is willing to pay only a little, say, $5 or $10, when added up over all Americans, the values may be signifi cant—in excess of $1 billion. This is what the court found in the Valdez case. Simi- lar issues are now being litigated in respect to the 2010 BP oil spill in the Gulf of Mexico.

Although there is considerable controversy over the accuracy of these methods, a special panel set up by the National Oceanic and Atmospheric Administration, which included distinguished Nobel Prize winners Ken- neth Arrow of Stanford and Robert Solow of MIT, recommended cautious use of the methodology by the government. In 1994, the government pro- posed new regulations for implementing contingent valuation, which are still being debated.11

SHADOW PRICES AND MARKET PRICES

Whenever there is a market failure, market prices may not refl ect true marginal social costs or benefi ts. In such circumstances, economists attempt to calculate the true marginal social costs or benefi ts—for instance, of hiring an additional worker, or of importing or exporting additional goods. They call these “social prices” or “shadow prices.” The term shadow prices reminds us that although these prices do not really exist in the market, they are the true social costs and benefi ts, refl ected imperfectly in the market price.

11� See M. Common, I. Reid, and R. Blaney, “Do Existence Values for Cost–Benefi t Analysis Exist?” Journal of Environmental and Resource Economics 9, no. 2 (1997): 225–238; J. Duffi eld, “Nonmarket Val- uation and the Courts: The Case of Exxon Valdez,” Contemporary Economic Policy 15, no. 4 (October 1997): 98–110; K. Arrow et al., “Is There a Role for Benefi t–Cost Analysis in Environmental, Health, and Safety Regulation?” Environment and Development Economics 2, no. 2 (May 1997): 196–201; P. Portney, “The Contingent Valuation Debate: Why Economists Should Care,” Journal of Economic Perspectives 8, no. 4 (Fall 1994): 3–17; and three articles in Journal of Economic Perspectives 26 (Fall 2012): 3–56, namely, C. Kling et al., “From Exxon to BP: Has Some Number Become Better than No Number?” R. Carson, “Contingent Valuation: A Practical Alternative When Prices Aren’t Available,” and J. Haus- man, “Contingent Valuation: From Dubious to Hopeless.”

309Discount Rate for Social Cost–Benefit Analysis

In the absence of a market failure, the price of something equals its opportunity cost, what is forgone in alternative uses. In an economy in which there is massive unemployment, the market wage exceeds the opportunity cost—indeed, what is forgone is the individual’s leisure—but when workers are unemployed involuntarily, the market wage exceeds the value of this forgone leisure, often by a considerable amount. The shadow price of labor when there is massive unemployment is the low value of the forgone leisure, not the market wage.

Similarly, in an economy in which capital markets work very imper- fectly, and fi rms cannot raise additional capital at the “market rate of interest,” the shadow cost of capital—what is forgone by using capital for one purpose rather than another—may exceed the market rate of interest by a considerable amount.

DISCOUNT RATE FOR SOCIAL COST–BENEFIT ANALYSIS

Our discussion of private cost–benefi t analysis noted that a dollar next year or the year after is not worth as much as a dollar today. Hence, income to be received in the future or expenses to be incurred in the future must be discounted. In deciding whether to undertake a project, we look at its

DIFFERENCES BET WEEN MARKET PRICES AND SHADOW PRICES Shadow prices refl ect true marginal social costs. When there are market failures, market prices may not fully refl ect social costs.

Examples are shown in the following table:

MARKET DIFFERENCE BETWEEN MARKET AND SHADOW PRICES EXPLANATION

Labor Shadow wage is less than market wage when there is unemployment.

No loss in output elsewhere when an individual is hired; hence, marginal cost of hiring a worker is less than wage.

Capital Shadow interest rate exceeds market interest when there is rationing in capital markets

Firm’s expected return exceeds interest rate (fi rm would like to borrow more at given interest rate, but cannot). Thus, opportunity cost of funds is greater than the interest rate.

Steel Shadow cost of production exceeds market cost.

Steel producer fails to value marginal social cost of pollution resulting from increased production.

310 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

present discounted value. The discount rate private fi rms use is r, the rate of interest the fi rm has to pay. What discount rate should the government use? The discount rate used by the government is sometimes called the social discount rate. The central question of concern is the relationship between this rate and the interest rate faced by consumers, on the one hand, and the rate faced by producers, on the other.

For evaluating long-lived projects, such as dams, the choice of the dis- count rate is crucial: a project that looks very favorable using a 3 percent interest rate may look very unattractive at a 10 percent rate. If markets worked perfectly, the market interest rate would refl ect the opportunity cost of the resources used and the relative evaluation of income at dif- ferent dates. However, there is a widespread belief that capital markets do not work well. Moreover, taxes may introduce large distortions, with large diff erences between before- and after-tax returns. Thus, it is not clear which of the various market rates of interest, if any, should be used: for instance, should it be the rate at which the government can borrow, or the rate at which the typical taxpayer can borrow?

If the individuals who benefi t from the project are the same as those who pay the costs, we can simply use their marginal rate of substitution, how they are willing to trade off the reduction in current consumption for gains in future consumption. Because their marginal rate of substitution will be directly related to the rate of interest at which they can borrow and lend, in this case we can use that market rate of interest for evaluating costs and benefi ts in diff erent periods. Often, though, the project has fur- ther ramifi cations—a public project may, for instance, displace a private project—and we then have to look at all the consequences, the net change in consumption.

If a public project displaces a private project of the same size, then the net reduction in consumption today from the project is zero. If both the public and private projects yield all their returns in the same period, then we can easily decide whether to undertake the public project: we should undertake it if its output exceeds that of the private project; or, equiva- lently, if its rate of return exceeds that of the private project. In this view, which, not surprisingly, is called the opportunity cost view—because the private project is the opportunity cost of the public project—it is the producer’s rate of return that should be used in project evaluation.

Focusing on opportunity costs and focusing on consumers’ marginal rates of substitution yield exactly the same result in economies in which there are no market failures, for then the marginal rate of substitution (which equals the rate of interest facing consumers) equals the rate of return on capital, or the producers’ rate of interest (the opportunity cost). Problems arise when there are market failures or taxes and/or when

311Discount Rate for Social Cost–Benefit Analysis

those who benefi t from a project are diff erent from those who pay for them.12 Today, many economists argue that the appropriate rate of inter- est for government discounting may be none of the observed market rates of interest.

More generally, it is recognized that choosing the appropriate inter- est rate is an exceedingly complex matter, with very important conse- quences, especially when it comes to decisions that have eff ects long into the future. The eff ects of increases in greenhouse gases that most scien- tists believe will result in global warming, a rising of the sea level, and climate change will not be fully felt for decades. A discount rate of, say, 7 percent means that a dollar of cost 100 years ago is worth less than one cent now—we can essentially ignore it. So too, we will be dealing with the highly radioactive wastes from nuclear reactors for decades, or even cen- turies, but if we use a 7 percent discount rate, we would essentially ignore most of those future costs.13

In the more general case, there is no presumption that the ratio of the marginal valuation of an increase in consumption by one generation to that of another is related to any interest rate. One approach, in that case, is to use social welfare functions. We fi rst introduced the concept of a social welfare function in Chapter 7 as a way of formalizing how con- sumption or income of diff erent individuals could be compared. Exactly the same principles apply in comparing individuals over time as in com- paring individuals at the same point of time, with one diff erence. In both cases, there is diminishing marginal utility, so if future generations have higher incomes than the current generation, the marginal valuation of a dollar of consumption to them is lower. But some economists believe that the welfare of individuals of future generations at the same level of

12 In a few special cases, the fact that the benefi ts may accrue to diff erent generations poses no problem. If the government has engaged in optimal intergenerational redistribution of income, then the mar- ginal value of a dollar to every generation will be equal to the market rate of interest, and as long as the project is relatively small, we can evaluate the marginal benefi ts received by diff erent generations using market rates of interest, just as we can when the impacts are felt by a single individual.

Similarly, if society consists of a set of family dynasties, in which each family optimally redistrib- utes income from the current generation to succeeding generations, then the marginal value of a dollar to every generation should equal the consumer rate of interest. [See R. Barro, “Are Government Bonds Net Wealth?” Journal of Political Economy 82 (1974): 1095–1117.] The validity of this model has been strongly questioned. It implies, for instance, that when the government ran huge defi cits in the 1980s, individuals increased their savings in a fully off setting way. Thus, it implies that in the absence of these defi cits, personal savings would have been negative. For an extensive discussion of these issues, see the symposium in the Journal of Economic Perspectives 3, no. 2 (Spring 1989); and K. Arrow et al., “How Should Benefi ts and Costs Be Discounted in an Intergenerational Context? The Views of an Expert Panel,” Resources for the Future Discussion Paper No. 12-53, 2012. 13�In a few special cases, there is a simple and clear solution—for instance, if the only imperfection in the market is optimally chosen taxes (in later chapters, we will describe in detail what is entailed by optimal taxes, but for now, we simply note that actual tax systems seldom comport even closely with optimal tax structures), then the producer’s rate of return should be used in project evaluation. [See P. Diamond and J. Mirrlees, “Optimal Taxation and Public Production,” American Economic Review 61 (1971): 261–278.] For a discussion showing how even slight changes in assumptions can lead to markedly diff erent conclusions, see J. E. Stiglitz and P. Dasgupta, “Diff erential Taxation, Public Goods, and Eco- nomic Effi ciency,” Review of Economic Studies 39 (1971): 151–174.

312 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

CLIMATE CHANGE AND DISCOUNT RATES

*J. Bruce, Hoesung Lee, and E. Haites, eds., Climate Change, 1995: Economic and Social Dimensions of Climate Change, Contribution of Working Group III to the Second Assessment Report of the Intergovernmental Panel on Climate Change (Cambridge, UK, and New York: Cambridge University Press, 1996). More information on the IPCC can be found at http://www.usgcrp.gov/ipcc/.

A s discussed in Chapter 6, one important policy issue facing the world over the com-ing decades will be how to respond to the threat of global warming resulting from increased concentrations of greenhouse gases such as carbon dioxide in the atmosphere. The effects go beyond just an increase in temperature: there are concerns about the rise in the level of seawater and increased weather variability.

In spite of international agreements to reduce emissions of greenhouse gases, within the United States there has been controversy about how much should be spent to reduce these emissions. The controversy arises in part because of discounting: most of the effects of global climate change will not be felt for a hundred years—and at a 7 percent dis- count rate, the value of $100 a hundred years from now is about 12 cents. It clearly will not be worth spending much today to avert even large costs in the future. On the other hand, at a 1 percent dis- count rate, $100 a hundred years from now would be worth more than $35.

A special working group of economists of the International Panel on Climate Change (IPCC),

including Nobel Prize winners Kenneth Arrow and Joseph Stiglitz, argued that the appropriate methodology implied using a low interest rate for purposes of discounting for costs and benefi ts associated with climate change.* Future genera- tions would be adversely affected if actions mitigat- ing the pace of emissions of greenhouse gases were not undertaken, and there were no ethical grounds for valuing their welfare substantially less than the welfare of the current generation.

Critics of this view argued that future gener- ations could be made whole “simply” by setting aside money today, investing it at the market rate of interest, and letting the amount accumulate to be used to address the costs of climate change.

The worry, however, was not only that estimates of future damage repair costs might be too low, and, more fundamentally, that there might be some dam- age that was irreparable at any cost, but also that countries would not set aside the funds. If they did not, then the appropriate trade-off analyzed by the IPCC committee was between consumption of the current generation and the welfare of future generations that would be adversely affected by climate change.

income should be weighted less than the welfare of the current gener- ation, simply because it exists in the future. The rate at which future generations’ welfare should be discounted is referred to as the pure discount rate. Other economists, such as the distinguished Cambridge economist Frank Ramsey, argued that all generations should be given equal weight.

To see what is implied by this approach, assume that per capita income is increasing at the rate of 1.5 percent, and the elasticity of marginal utility is 1.

313Discount Rate for Social Cost–Benefit Analysis

(The elasticity of marginal utility is the per- centage decrease in marginal utility from a 1  percent increase in consumption. As we saw in Chapter 7, economists usually assume the elasticity of marginal utility is between 1 and 2.) Then if the pure discount rate is zero, the social rate of discount is 1.5 percent, roughly equal to the real interest rate on safe (government) securities, but considerably below the opportu- nity cost of capital.

More recently, the rate of increase in per capita income has been less than 1.5 percent. If it is 1 percent, then the discount rate is also smaller at just 1 percent. (This discussion assumes that we can ignore the distribution of income, but we should not do this. In recent years, the median family—the family such that half have an income that is higher and half lower—has seen its income essentially stagnate. If we value future income based on their marginal utility, the discount rate is close to zero.)

The question of the appropriate social rate of discount has become a hotly contested political issue. Those who are concerned about the envi- ronment and who see environmental impacts stretching out over decades, for example, believe strongly in low discount rates. For instance, in their view, simply because the eff ects of nuclear waste can be postponed for fi fty or a hundred years is no reason to essentially ignore them—which a 10 percent discount rate eff ectively tells us to do.

Today, the federal government uses a real discount rate of 7 percent in cost–benefi t analyses, except for the evaluation of projects whose impacts are felt over the very long run, like global warming, where lower discount rates are used. This partially refl ects the opportunity cost view—the fed- eral government uses this rate because it “approximates the marginal pretax rate of return on an average investment in the private sector in recent years.”14

It is not surprising that the discount rate should be a subject of such political controversy. But why can’t economists agree among them- selves? Our discussion has highlighted several sources of disagree- ment about the economy and about the government, summarized in Table 11.2.

14�This quotation is taken from Offi ce of Management and Budget Circular No. A-94 Revised (October 29, 1992) which establishes guidelines and discount rates for cost–benefi t analysis of federal pro- grams still applicable today. Subsequent guidelines issued by the federal government in 1995 allowed for the use of lower discount rates for long-lived projects with impacts over many generations.

THREE VIEWS ON THE SOCIAL DISCOUNT RATE 1. Refl ects consumers’ rate of time preference (the

consumers’ borrowing rate).

2. Refl ects opportunity cost of capital (the producers’ borrowing rate).

3. May refl ect neither: in long-lived projects affecting different generations, when social marginal valua- tion of consumption of different generations may have nothing to do with observed interest rates, and in the absence of optimal intergenerational redistribution of income.

314 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

TABLE 11.2 SOURCES OF DISAGREEMENT IN DISCOUNT R ATES

HIGH DISCOUNT RATE (OPPORTUNITY COST OF CAPITAL) LOW DISCOUNT RATE (SOCIAL RATE OF DISCOUNT)

1. Government investment tends to displace private investment.

1. Assessing net impacts is typically far more complicated than just assuming that a dollar of public investment displaces a dollar of private investment.

2. Even in a world with distortions, everyone could be made better off if effi ciency is maintained—this entails the rate of return on public projects’ equaling that on private projects.

2. Assessing the desirability of a project must take into account intergenerational distributional effects as well as effi ciency effects. (a) Programs’ benefi ciaries are often

different from those who bear the costs.

(b) Even if the government could in principle make everyone better off, the required compensations (for instance, to those who are adversely affected) are seldom made. Moreover, there may be high costs associated in providing compensation.

3. Even in the absence of government intergenerational redistributions, if parents leave bequests to their children, marginal valuations of consumption of different generations will be equalized (the dynastic model).

3. In the absence of optimal intergenerational redistribution, market rates of interest do not refl ect marginal social valuations of dollars to different generations. Further, the dynastic model is implausible.

4. When market distortions are caused by optimal taxes, then effi ciency is still desirable, so the government should use the opportunity cost of capital.

4. With market distortions, marginal rates of substitution (how individuals value a marginal dollar in different years) and marginal rates of transformation (the trade-offs facing fi rms) may differ markedly.

With distortionary taxes, effi ciency—as exemplifi ed by using the private sector’s opportunity cost of capital in the public sector—is desirable only under highly restrictive conditions.

THE EVALUATION OF RISK

The most common mistake in trying to cope with the uncertainties of the benefi ts and costs of a project is to argue that in the face of risk, the gov- ernment should use a higher discount rate. Recall that the discount rate relates the value of a dollar at one date to its value at a later date. To see how increasing the discount rate may lead to absurd results, consider a

315The Evaluation of Risk

project that, at termination, requires an expenditure (say, an automobile must be towed to the junkyard). Assume that there is some uncertainty about the magnitude of that cost. We would normally think that this uncertainty would make the project less attractive than if we knew for sure what the termination costs were. However, consider what happens if we use a higher discount rate to off set the risk: the discount factor is lower, the present value of those costs is reduced, and the project looks more, not less, attractive. To use a higher discount rate confuses the eval- uation of income at diff erent dates with the evaluation of risk; these are two separate issues.

To evaluate risks, economists introduce the concept of certainty equivalents. Assume that there is some risky project. Next year, the output of the project may be worth $0 or $100; there is a fi fty-fi fty chance of each outcome. The average value is just $50 ( 12 3 $100 1 12 3 $0 5 $50). If we dislike risk, however, we would clearly prefer a project whose return was a certain $50. In fact, we would prefer a project with a smaller average value, as long as the risk was smaller. If we would be indiff erent in choos- ing between the risky project with an average value of $50 and a perfectly safe project with a value of $45, we would say that $45 is the certainty equivalent of the risky project with an average value of $50. To evaluate risky projects, then, we simply take the present discounted value of the certainty equivalents.15

Thus, to be acceptable, risky projects must earn a higher return than safe projects with the same certainty equivalent. The extra amount a risky project must earn to compensate is its risk premium.

We illustrate the procedure in Table 11.3, for a fi ve-year project. We have assumed that the initial investment in the fi rst period is certain. The benefi ts that accrue in years 2, 3, and 4 are increasingly uncertain,

15�This methodology is not perfectly general. It requires that we be able to separate the analysis of risk at one date from that of other dates. For most practical purposes, however, it is suffi ciently general.

TABLE 11.3 EX AMPLE OF COST–BENEFIT ANALYSIS FOR RISK Y INVESTMENT

CERTAINTY TIME DISCOUNT DISCOUNTED VALUE EXPECTED NET EQUIVALENT FACTOR (10 PERCENT OF CERTAINTY YEAR BENEFIT NET BENEFIT INTEREST RATE) EQUIVALENT NET BENEFIT

1 $2100 $2100 1 $2100 2 100 90 .91 81.90 3 100 80 .83 66.40 4 100 75 .75 56.25 5 250 275 .68 251

Total 150 70 53.55

316 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

refl ected in the certainty equivalents. In the fi nal year, the project is scrapped; there are large costs associated with the termination of the proj- ect. (Consider the problem of what to do with a nuclear power plant when its useful life has come to an end.) However, these costs are uncertain; hence, its certainty equivalent exceeds the $50 expected cost. (In  con- trast, had we employed a higher time discount rate to take account of risk, these uncertain scrapping costs would not have weighed very heavily in our cost–benefi t calculation.)

To obtain the present discounted value of the certainty equivalent net benefi t at any date, we multiply it by the time discount factor. To obtain the present discounted value, we add up the discounted certainty equiva- lent net benefi ts for the life of the project.

How should the government evaluate the risks associated with various projects? In some cases, such as the risks associated with the generation of electricity, it can look to how private markets value risks. However, for risks for which there is no comparable private project, matters are more diffi cult. Some, such as a fl ood control project, serve to reduce the risks individuals face, and for these projects, the risk premium is negative. Indi- viduals are willing to pay something to reduce the risk of fl ood. Because the government can spread risks over the entire population, when the project neither serves an insurance function (reducing the risks individ- uals would otherwise face) nor provides a return that is correlated with income from other sources (that is, the return to the project is neither particularly high nor particularly low when the economy is, say, healthy), the government should employ no risk premium.

Sometimes, government projects essentially provide insurance; that is, they reduce the risks that society would otherwise face. A dam reduces the risk of fl oods. Individuals would have been willing to pay a high pre- mium for the reduction of such risks. In this case, the risk premium is negative; that is, the project should be undertaken, even if the expected return is lower than the return on a perfectly safe project.

Expenditures to reduce the risk of global warming provide an import- ant instance for which public expenditures should arguably be under- taken even if the average return is lower than on a safe project—that is, the government should use a discount rate lower than that for a safe project. One way of thinking about this is the following: such investments yield a high return when it turns out that the eff ects of (and costs of) global warming are high—when sea level rises more than we had expected, or the increase in temperature or the variability in climate is greater than we had anticipated. In those instances, our standards of living will be low- ered, as we will have to spend a great deal of money dealing with the con- sequences of global warming and making investments to help us adapt to

317The Evaluation of Risk

these changes. However, that means that the marginal utility of a dollar will be higher. In short, when the eff orts at reducing global warming pay off , we value those dollars especially highly.

RISK ASSESSMENT

An area of increasing scrutiny—and controversy—in risk analysis is the risks to health, safety, and indeed life, that are posed, for instance, by haz- ardous wastes, pesticides, and fungicides. Chemicals in the water and air increase the likelihood of cancer and a variety of other ailments, often life-threatening. About this there is little doubt. The debate has focused on risk assessment—on how the magnitude of these risks is assessed and how priorities for reducing these risks should be established.

For instance, many risks are related to exposure. A chemical in dirt that is sealed under a thick layer of concrete is unlikely to impose sig- nifi cant risks; there would be a much higher risk if that same dirt were ingested directly by a child. In assessing the overall risk, one must take into account the probability of diff erent levels of exposure, as well as the risks associated with each level of exposure.

The Environmental Protection Agency, in setting its priorities and its standards—for instance, for cleaning up hazardous wastes—has been criti- cized on several grounds. Rather than determining “compounding of prob- abilities” eff ects the way students are typically taught in modern statistics courses, the EPA uses a “worst case scenario analysis,” which looks at the risks associated with the worst possible case. For example, what would be the risk assuming that the concrete seal around the dirt cracked, and a child wandered into the site? There have been some famous stories of the EPA insisting on cleanups to the standard that a child could eat the dirt for a six- week period without having any signifi cant increase in health risk.16 In set- ting priorities, there has been concern that the government has not gone after the highest risks, but rather the risks that have the most “popular appeal.” The risks addressed by the EPA are often far lower than the risks that individ- uals take in their day-to-day lives—for instance, from drinking alcohol mod- erately, let alone from smoking. There is, however, one critical distinction: the risks on which the EPA focuses are those (like air and water pollution) over which individuals have no choice; they are incurred involuntarily, as opposed to the risks associated with smoking and drinking. Still, the fact that indi- viduals seem willing to incur certain risks reveals information about their

16� See Stephen G. Breyer, Breaking the Vicious Circle: Toward Eff ective Risk Regulation (Cambridge, MA: Harvard University Press, 1993), p. 12.

318 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

valuation of the risks, a fact that government should presumably take into account when adopting environmental risk standards. Recent government regulations have put greater emphasis on assessments of comparative risks; there is a reluctance to impose costly regulations to reduce risks that are of the size that individuals seem willing to accept in ordinary circumstances.

There has been increasing concern that environmental risks are borne disproportionately by the poor, who often live in industrial areas with heavier pollution. This is not surprising, as land in such areas typically is less valuable, so lower-income individuals can obtain housing at lower costs. President Clinton signed an executive order on environmental justice in 1994, instructing agencies to ascertain the distributional impact of environmental measures they might undertake. After more than a decade of federal government inertia in advancing environmental justice, in 2011 President Obama strengthened eff orts to implement the 1994 executive order by issuing a memorandum of understanding signed by the heads of seventeen federal agencies commiting each agency to “identify and address, as appropriate, any disproportionately high and adverse human health or environmental eff ects of its programs, policies and activities on minority populations and low-income populations.”17

DISTRIBUTIONAL CONSIDERATIONS

The benefi ts of any given public project are not uniformly distributed across the population. Some projects, such as a dam, have benefi ts that are limited geographically. Other projects, such as bilingual education programs and jobs retraining programs, are directed mainly at particular groups (e.g., immigrants and the unemployed). The government is clearly concerned about the impact of its programs on the distribution of income.

Should these distributional eff ects be taken into account in cost– benefi t analysis? If so, how can they be quantifi ed?

The issue of whether government should take distributional eff ects into account is analogous to the issue of whether, in choosing a social discount rate, the government needs to be concerned with the impact on diff erent generations. If the “social” marginal value of a dollar to all indi- viduals is the same, we can simply add up the dollar value of the impacts on consumption of diff erent individuals. However, there is a widespread presumption that the social marginal value of a dollar to a poor individual is greater than it is to a rich individual.

17�Memorandum of Understanding on Environmental Justice and Executive Order 12898.

319Cost Effectiveness

The fi rst step in any distributional analysis is to ascertain as precisely as possible how the program aff ects individuals in diff erent circum- stances. Typically, the focus is placed on individu- als of diff erent incomes, although regional impacts are also frequently taken into account. Of two programs with similar overall impact, the one in which more of the benefi ts and fewer of the costs accrue to poor individuals would presumably be preferred, if society cares about distribution.

Often, however, there is a desire to go beyond simply enumerating the impacts on diff erent groups, to obtain a broader picture. This is done in two diff erent ways. The fi rst uses the social wel- fare function approach referred to earlier. It rec- ognizes that the marginal valuation of a dollar is greater to a poor person than to a rich person, and uses the concept of the elasticity of marginal util- ity to quantify the extent to which this is so. For example, using an elasticity of unity (1), and giv- ing a weight of unity to those at median income, impacts on those with half the median income receive a weight of 2, whereas impacts on those with twice the median income receive a weight of 12 �. Using these weights, the total “weighted bene- fi t” is calculated, and of two programs with the same costs, the one with the highest weighted benefi t is undertaken.

The second approach looks at the impact on the overall distribution of income or wealth.18 However, this approach lends itself only to major programs with the capacity for substantial distributional eff ects, such as changing the welfare system or the tax system. Most projects undertaken by the government are smaller in scale.

COST EFFECTIVENESS

In some cases, it is diffi cult to compare costs and benefi ts. The benefi ts may be improved health; the costs are dollars expended. Although we have emphasized the necessity of making hard—monetary—judgments concern- ing life and health, the political process often tries to avoid making such

18 This approach relies on measures like the Gini coeffi cient, discussed in the appendix to Chapter 7.

KEY ISSUES IN MEASURING A PROJECT’S BENEFITS 1. Measuring consumer surplus

2. Measuring nonpecuniary benefi ts:

Valuing time Valuing life Valuing the environment

3. Valuing marketed goods in the presence of market failure:

Using shadow prices to measure marginal social costs when market prices do not accurately measure it

4. Valuing consumption (output) at different dates:

Choosing the right discount rate

5. Valuing risk

6. Valuing distributional considerations:

How are impacts on different groups to be compared?

320 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

judgments, when possible. Cost eff ectiveness (CE) analysis provides a way of doing this by looking at programs with the same (or similar) benefi ts, and asking which produces those benefi ts at the least cost.

Assume that we wish to avoid the problems associated with valuing lives while helping the government assess a variety of ways of reducing highway deaths. We could calculate the costs associated with each of sev- eral methods of accomplishing the same goal. Or we might simply show the marginal costs associated with incremental reductions in the death rate under each method, and leave it to the legislators to determine which point along the curve should be chosen (and, therefore, what method of improving traffi c safety should be chosen).

When the Occupational Safety and Health Administration considered standards for noise pollution, it did a CE study, calculating how many extra workers would be protected from hearing loss as a result of alterna- tive standards. It then calculated the cost associated with each standard. From this information, it calculated the marginal gross and net costs (taking into account the fact that hearing losses reduce productivity) associated with diff erent levels of protection, as depicted in Figure 11.3.

COMPARISON OF ALTERNATIVE STANDARDS FOR OCCUPATIONAL

NOISE EXPOSURE

Higher standards cost more and protect more workers

from hearing loss.

FIGURE 11.3

Number of workers protected (thousands)

Marginal costs

(thousands of dollars)

High

Marginal gross costs

Very high standard

Marginal net costs (taking account of productivity gains from lower noise

level)

MediumLow 10 20

180 190 200 210 220 230 240 250 260

300

200

100

50SOURCE: J. R. Morrall III, “Exposure to Occupational Noise,” in Benefi t– Cost Analyses of Social Regulation,

ed. James C. Miller III and Bruce Yandle (Washington, DC: American

Enterprise Institute for Public Policy Research, 1979).

321Cost Effectiveness

The curve shows that there are signifi cant extra costs of trying to protect additional individuals from hearing loss. On the basis of this, one study concluded, “an eff ectively administered hearing-protector program could provide most of the benefi ts at much lower cost in comparison with an industrywide engineering-only noise standard. . . . [A]n 85-decibel hearing-protector standard [has] the relatively reasonable marginal cost of about $23,000 per hearing impairment avoided.” In ordinary English, the study recommended the use of earplugs rather than the drastic changes in plants and equipment that would be required to implement the same level of hearing protection.

Table 11.4 shows another example of CE studies, this time compar- ing the cost eff ectiveness of alternative medical interventions. There is an enormous range of cost eff ectiveness ratios, from $2158 per life-year saved for administering a low dose of the drug lovastatin to reduce choles- terol for males between ages 55 and 64 who had prevalent coronary heart disease, to $41,000 for annual breast examination and mammography for females aged 55 to 65, to $88,000 for a coronary artery bypass graft for someone with a single-vessel disease with moderate heart weakness, to $335,000 for the use of an exercise cardiogram as a screening test for heart disease for 40-year-old females.

The use of CE analysis to guide resource allocation in health care has grown substantially in Europe over the past decade, but there has been con- siderable resistence to explicit consideration of costs in the policy coverage criteria of major U.S. health schemes such as Medicare. More attention, though, is now being given in the United States to the eff ectiveness compo- nent of CE analysis through support of comparative eff ectiveness research (CER), exemplifi ed by the provision of $1.1 billion under the American Recovery and Reinvestment Act of 2009 (ARRA) for CER. However, the primary objective of CER is to improve the quality of health care by provid- ing information on which medical interventions work best for whom under what circumstances, without direct reference to relative costs.19

Although CE analysis is simpler than cost–benefi t analysis because it avoids all the problems of measuring and valuing benefi ts, most of the issues discussed in measuring and valuing benefi ts remain, scaled down, for mea- suring and valuing costs. For instance, shadow prices for inputs may diff er from market prices; a social discount rate must be used to value costs incurred at diff erent dates; and there is considerable uncertainty—for example, we might be unsure of the exact degree to which hearing loss hurts productivity, or how much it will cost to bring a new weapon to completion, or how much it will cost fi rms to comply with stricter environmental standards.

19 For an update on the study cited in Table 11.4, see Alan M. Garber and Charles E. Phelps, “Future Costs and the Future of Cost-Eff ectiveness Analysis,” Journal of Health Economics 27 no. 4 (2008): 819–821.

322 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

Cost–benefi t and CE analysis are important tools used by policy mak- ers throughout the world. They provide discipline to the decision-making process. Although critics complain that they reduce everything to cold calculations, these tools can be used to bring systematically into the anal- ysis not only economic costs and benefi ts, but also concerns about the

TABLE 11.4 ES TIM ATED COS T EFFECTIVENES S OF COMMONLY USED MEDICAL INTERVENTIONS (ALL INTERVENTIONS COMPARED TO “USUA L CARE” UNLES S OTHERWISE NOTED)

INTERVENTION COST/LIFE-YEAR ($1993)

Low-dose lovastatin for coronary heart disease prevention a Males, age 55–64, cholesterol level $ 250 2,158 Males, age 55–64, cholesterol level , 250 22,929 Female nonsmokers, age 35–44 2,023,440

Exercise electrocardiogram as screening testb

40-year-old males 124,374 40-year-old females 335,217

Hypertension screening c

40-year-old males 27,519 40-year-old females 42,222

Breast cancer screening d

Annual breast examination and mammography, females age 55–65 41,008

Physician advice about smoking cessation e

1% quit rate, males age 45–50 3,777

Pap smear starting at age 20, continuing to 74f Every 3 years, versus not screening 24,011

Coronary artery bypass graft g Left main coronary artery disease 8,768 Single-vessel disease with moderate angina 88,087

Neonatal intensive care units h

Infants 1000–1500 g 10,927 Infants 500–999 g 77,161 a L. Goldman et al., “Cost-Effectiveness of HMG-GoA Reductase Inhibition for Primary and Secondary Prevention of Coronary Heart Disease,“ Journal of the American Medical Association 265 (1991): 1145–51. b H. C. Sox Jr. et al., “The Role of Exercise Testing in Screening for Coronary Artery Disease,“ Annals of Internal Medicine 110 (1989): 456–69. c B. Littenberg et al., “Screening for Hypertension,“ Annals of Internal Medicine 112 (1990): 192–202. d D. M. Eddy, “Screening for Cervical Cancer,“ Annals of Internal Medicine 113 (1990): 214–26. e S. R. Cummings et al., “The Cost-Effectiveness of Counseling Smokers to Quit,“ Journal of the American Med- ical Association 261 (1989): 75–79. f D. M. Eddy, “Screening for Breast Cancer,“ Annals of Internal Medicine 111 (1989): 389–99. g M. C. Weinstein, “Economic Assessment of Medical Practices and Technologies,“ Medical Decision Making 1 (1981): 309–30. h M. H. Boyle, G. W. Torrance, J. C. Sinclair, and S. P. Horwood, “Economic Evaluation of Neonatal Intensive Care of Very-Low-Birth-Weight Infants,“ New England Journal of Medicine 308 (1983): 1330–37.

SOURCE: Alan M. Garber and Charles Phelps, “Economic Foundations of Cost-Effectiveness Analysis,“ Journal of Health Economics 16, no. 1 (1997): 1–31.

323Post-Expenditure Evaluation: Assessing and Improving Government Performance

environment, health, and distribution. Although there never will be com- plete precision, especially in these hard-to-quantify areas, judgments will be made weighing these various considerations, and quantifi cation can be a helpful step in resolving the complicated trade-off s that must be faced.

POST-EXPENDITURE EVALUATION: ASSESSING AND IMPROVING GOVERNMENT PERFORMANCE

Thus far, this chapter has been devoted to pre-expenditure project appraisal. It has explored the use of cost–benefi t analysis to compare expected future benefi ts with anticipated costs to produce these benefi ts, as well as the use of CE analysis to estimate the cost eff ectiveness of alter- native potential future interventions.

However, it is also important to undertake post-expenditure evalua- tion to assess and improve government performance: Did the government spend its funds wisely? Did the public get good value for its tax payments? How can expenditure policies and practices be improved?

In addition to evaluating the economic effi ciency and distributional eff ects of public expenditures (see Chapter 10), we can also evaluate pub- lic expenditures in terms of meeting policy objectives. The broadest form of this type of evaluation is a national public expenditure review (PER), commonly conducted by the World Bank. PERs provide an overview of the level, composition, and results of public expenditures, and usually entail examination of (1) aggregate spending levels in the context of a country’s macroeconomic policies; (2) the impact of sectoral, geographic, and demographic allocation of spending on poverty reduction; (3) the respective roles of government and the private sector in the fi nancing, production, and provision of public infrastructure and services, with a focus on the need for public sector mitigation of market failures while not crowding out the private sector; (4) the balance between capital invest- ments to increase future capacity and recurrent expenditures to operate and maintain existing infrastructure; and (5) quality of the expenditure process in terms of transparency and accountability. PERs are expen- sive and complex endeavors, so they are usually undertaken periodicially (about every fi ve years) rather than annually.20

20� For a detailed description of PERs, see Sanjay Pradhan, “Evaluating Public Spending: A Framework for Public Expenditure Reviews,” World Bank Discussion Paper 323, Washington, DC, 1996, and the World Bank public expenditure website www.worldbank.org.

324 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

We can also evaluate public expenditures made to deliver specifi c goods and services. The most common methodology for doing this is comparative performance measurement (CPM), which is a tool to determine expenditure effi ciency by assessing the total, unit, or relative costs of inputs and measure expenditure eff ectiveness by evaluating the quantity and quality of both short-term outputs (intermediate results) and long-term outcomes (achievement of policy objectives). The key to CPM’s usefulness is to place these measurements in a comparative perspective—that is, to track them over time, as well as between sec- tors, institutions, or jurisdictions. This allows a unit of government to see how well it is performing by comparing itself to similar agencies or localities; communicate past performance to its constituents and prior- itize future expenditure needs in consultation with these constituents;

TAKING A BITE OUT OF CRIME IN THE BIG APPLE

SOURCES: Eli Silverman, “CompStat’s Innovation,” in Police Innovation: Contrasting Perspectives, ed. David Weisburd and Anthony Braga (Cambridge, UK: Cambridge University Press, 2006); and Robert D. Behn, “Designing PerformanceStat,” Public Performance and Manage- ment Review 32, no. 2 (December 2008): 206–235.

A lthough traditional indicators of police per-formance, such as response times and arrest rates, measure operational effi ciency, they do not reveal whether police departments are actu- ally making their communities any safer. Thus, New York City Police Commissioner William Bratton and Deputy Commissioner Jack Maple created CompStat in 1994 as a management tool to both measure and improve performance in crime reduction.

The centerpiece of CompStat is frequent crime control strategy meetings between all precinct commanders and New York Police Department (NYPD) top management to discuss the latest crime statistics. The key is to creatively compile and ana- lyze crime data, and then make these data accessi- ble in a timely manner to front-line police offi cers so they can use it to achieve NYPD policy objec- tives. CompStat generates a considerable amount of information over time and by police precinct, so

all can see the latest trends, try to understand the underlying causes of these trends, assess results of follow-up responses to previously identifi ed prob- lems, and devise strategies for coping with per- formance shortcomings. CompStat also produces “commander profi le reports,” so senior NYPD exec- utives can evaluate the internal management per- formance of their fi eld leaders.

CompStat is widely credited with helping to dramatically reduce crime in New York City, and has been adopted by police forces throughout the United States and abroad—for example, in Austra- lia. CompStat has also been adapted to other public agencies and even entire jurisdictions, giving rise to the term “PerformanceStat”—for example, JobStat (New York City), EdStat (Rhode Island), FEMAStat (Federal Emergency Management Agency), Citi- Stat (Baltimore, Somerville, Atlanta), and StateStat (Maryland).

325Post-Expenditure Evaluation: Assessing and Improving Government Performance

and revise its policies, management, and operations to improve future performance.21

For example, if a city wants to evaluate its expenditures on secondary edu- cation, it can analyze how much it spent on this in total for the past fi ve years (aggregate costs), how much it spent per teacher or per student (unit costs), or how much it spent versus primary education or public health expenditures (relative costs), all compared with expenditures of similar cities. Although such an analysis does not indicate how well this money is being spent (value for money), it can reveal whether production costs seem to be in line with established norms—that is, with what is going on elsewhere.22

A city could go further and also measure expenditure eff ectiveness. Again using secondary education as an example, it can measure, in a com- parative time series, short-term results such as the number of students who attend school or who graduate (aggregate outputs), as well as teacher qualifi cations or student–teacher ratios (which are supposed to be prox- ies for quality). It can also try to determine whether students are actu- ally getting a good education, the primary policy objective of secondary education expenditures, by using performance standards or benchmarks such as standardized test scores or college enrollment rates. The diffi culty, however, is that even these measures do not really answer the question in which we are really interested: Have our schools made their students more productive, with higher incomes, more creative, better citizens, contributing to rising living standards? It typically takes decades before we know the answers to such questions, and then, there are many other factors aff ecting outcomes. There is also considerable debate on whether measuring basic skills distracts from investments in more important outcomes, such as cognitive development. Still, there is a broad concen- sus that these indirect indicators, as imperfect as they are, provide some insight into the extent to which schools are succeeding in their mission.

CPM is a simple but powerful tool for public institutions to gener- ate and analyze data in a form that facilitates learning from both their own experience and the experiences of others with similar missions and challenges.

21 For a detailed explanation of CPM and examples of the application of CPM, see Elaine Morley, Scott P. Bryant, and Harry P. Hatry, Comparative Performance Measurement (Washington, DC: The Urban Institute Press, 2001). 22 A more sophisticated application of CPM to assess expenditure effi ciency is data envelope analysis (DEA). This is used most extensively to assess effi ciency in the provision of municipal services, such as street repair, building maintenance, fl eet management, solid waste collection, water and sewerage provision, and public transit. DEA uses a nonparametric measurement of technical effi ciency, defi ned as production of the most outputs from the fewest inputs. For a detailed explanation of DEA and exam- ples of the application of DEA, see Adrian Moore, James Nolan, and Geoff rey F. Segal, “Putting Out the Trash: Measuring Municipal Service Effi ciency in U.S. Cities,” Urban Aff airs Review 41, no. 2 (2005): 237–259. For an international adaptation of DEA, see António Afonso, Ludger Schuknecht, and Vito Tanzi, “Public Sector Effi ciency: An International Comparison,” European Central Bank Working Paper No. 242, Frankfurt am Main, 2003.

326 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

Considerable progress has been made around the world over the past two decades in the quantitative measurement of public sector perfor- mance under the general rubric of “new public management” as govern- ments continue to search for ways to do things “better, faster, cheaper.”23 For example, widespread adoption of performance-based budgeting (PBB) refl ects the desire to allocate resources based on documented results rather than previous levels of input—PBB allocates resources based on the achievement of specifi c, measurable short-term outputs and long-term outcomes. The growth of PerformanceStat initiatives takes the concept further, as they incorporate performance measurement into integrated performance management—leadership uses data on how well people are doing their jobs as part of an integrated management informa- tion system to improve overall institutional performance (see case study, “Taking a Bite Out of Crime in the Big Apple”). The hope is that a public leadership strategy based on the utilization of objective and transparent performance measurement will make government more accountable to the public for the way it spends tax revenue, and provide the public sector with incentives to replicate private sector effi ciency and eff ectiveness in service of the public interest.

23 This is the name of a popular website for state and local governments that provides “innovative ideas for government that deliver public value and lower the cost of government services,” http://www .governing.com/blogs/bfc.

SUMMARY

1. Cost–benefi t analysis provides a systematic set of procedures by which a fi rm or government can assess whether to undertake a project or program and, when there is a choice among mutually exclu- sive projects or programs, which one to undertake.

2. Private cost–benefi t analysis entails determining the consequences (inputs and outputs) associated with a project, evaluating these using market prices to calculate the net profi t in each year, and

fi nally, discounting profi ts in future years to cal- culate the present discounted value of profi ts.

3. Social cost–benefi t analysis involves the same pro- cedures as private cost–benefi t analysis,  except that a broader range of consequences is taken into account, and the prices at which inputs and out- puts are evaluated may not be market prices, either because the inputs and outputs are not marketed (so market prices do not exist) or because market prices do not accurately refl ect marginal social costs and benefi ts, due to a market failure.

REVIEW AND PRACTICE

327Review and Practice

4. When the government makes available a good or service that was not previously available (e.g., constructs a bridge across a river), the value of the project is measured by the consumer sur- plus it generates; this is the area under the (com- pensated) demand curve.

5. The government has to make inferences (based on market data or observed behavior) concern- ing the valuation of nonmarketed consequences, such as lives and time saved or impacts on the environment.

6. The rate of discount used by the government to evaluate projects may diff er from that used by private fi rms.

7. To evaluate risky projects, the certainty equiv- alent of the benefi ts and costs needs to be calculated.

8. Distributional considerations may be introduced into evaluations, either by weighting the bene- fi ts accruing to diff erent groups diff erently or by assessing the impact of the project on some mea- sure of inequality.

9. It is also important to undertake post-expendi- ture evaluation to assess and improve govern- ment performance.

KEY CONCEPTS

Certainty equivalents

Comparative performance measurement (CPM)

Consumer surplus

Contingent valuation

Cost–benefi t analysis

Cost effectiveness (CE)

Discount factor

Environmental justice

Existence values

Opportunity costs

Opportunity cost view

Present discounted value (PDV)

Public expenditure review (PER)

Pure discount rate

Risk assessment

Risk premium

Shadow prices

Social discount rate

QUESTIONS AND PROBLEMS

1. Consider a project that costs $100,000 and yields a return of $30,000 for fi ve years. At the end of the fi fth year, there is a cost of $20,000 to dispose of the waste from the project. Should the project be undertaken if the discount rate is 0? 10 per- cent? 15  percent? The interest rate at which the net present discounted value of the project is zero is referred to as the internal rate of return of the project.

2. Assume there is uncertainty about the costs of disposing of the waste: there is a fi fty-fi fty chance that they will be $10,000 or $30,000. Discuss how this uncertainty aff ects the cost– benefi t calculation, if the government is risk neutral (i.e.,  that it requires no risk premium to compensate it for bearing risk) or very risk averse (i.e., that it requires a large risk premium to com- pensate it for bearing risk).

3. Assume now that there are two groups in the population. Each contributes equally to the cost of the project, but two-thirds of the bene- fi ts accrue to the richer group. Discuss how this alters the cost–benefi t calculation. Under what circumstances will the decision to undertake the project be altered?

4. Assume that the government now has a choice between undertaking the project described in problem 1 and undertaking a larger project. If it spends an additional $100,000, returns will be increased by $25,000 per year and disposal costs in the fi nal year will increase by $20,000. Which project should be undertaken if the discount rate is 0? 10 percent? 15 percent? If there are two groups in the population, how are your answers aff ected if two-thirds of the incremental benefi ts go to the poor (with the incremental costs being shared equally, as before)?

328 CHAPTER 11 EVALUATING PUBLIC EXPENDITURE

5. Discuss why, under each of these circumstances, a social cost–benefi t analysis might diff er from a private cost–benefi t analysis:

a. The unemployment rate is 10 percent.

b. The government has imposed a tariff on the importation of textiles.

c. The government has imposed a quota on the importation of oil.

d. The government has imposed a tax on interest income.

e. The government has imposed price controls on natural gas.

f. The government has regulated airlines, so prices exceed the competitive levels.

6. For each of the following projects, what benefi ts or costs might be included in a social cost–benefi t analysis that might be excluded from a private cost–benefi t analysis:

a. A hydroelectric project

b. A steel mill

c. A chemical plant

d. A project to improve car safety

e. A training program to improve the skills of minority workers in a fi rm

How might your answers be aff ected by changes in legislation (e.g., concerning manufacturers’ liabilities for automobile accidents, legislation imposing fi nes on polluters)?

7. How might the techniques used to analyze the distributional consequences in cost–benefi t anal- ysis be employed to ensure that concerns about environmental justice are incorporated into the analysis?

8. How are issues of incidence analysis and capital- ization (discussed in the previous chapter) incor- porated into cost–benefi t analysis? For each case below, does it make a diff erence to your answer whether poor individuals own or rent their houses? In particular, if you wished to incorporate the distributional consequences of the following policies and programs, how might you do so?

a. A government regulation that reduces the allowable level of noise for aircraft. (Assume that those who live in the neighborhood of the airport are relatively poor.)

b. A subway line intended to make it less expen- sive for those in low-income neighborhoods to get to jobs in the center city.

c. The Superfund program is intended to clean up toxic waste sites. Currently, a dispropor- tionate number of poor people live near toxic waste sites.

9. The government is debating whether to spend $100 billion to reduce global warming dam- age 100 years from now. It is estimated that $800  billion of damage will be averted. A critic of the expenditure says that it would be far better to take the $100 billion, invest it in the stock mar- ket, earning an average return of 6 percent per year, and use the proceeds of the investment in 100 years to repair the damage. Should the proj- ect be undertaken?

329

The end of the Cold War, marked by the breakup of the Soviet Union in 1991, had profound consequences. At the time, defense had long been the largest single item of public expenditures in the United States. In 1953, the last year of the Korean War, defense spending reached 14.7 percent of GDP. After the Vietnam War, defense expenditures fell to under 6 per- cent of GDP, but during the 1980s, as part of the Cold War, they increased again, peaking at 7.4 percent of GDP in 1986 and 1987.

After the collapse of the Soviet empire, there was excitement in the United States about a “peace dividend.” Defense expenditures dropped to half of their 1980s share of GDP—by 2000 they were just 3.7 percent of GDP. These spending reductions were critical to the successful eff ort to reduce the federal budget defi cit during the Clinton administration.

The end of the Cold War did not bring global peace, however. There were confl icts in the Balkans and sub-Saharan Africa. There were also continuing worries about nuclear proliferation. Then, with the attack on the World Trade Center on September 11, 2001, the United States became engaged in a new kind of confl ict, called the global war on terrorism. Although it centered on Iraq and Afghanistan, it was a war that knew no

DEFENSE, RESEARCH, AND TECHNOLOGY

12 1. How do we go about deciding how much to spend on national defense? What is the role of marginal analysis?

2. What are the current key issues concerning defense strategy and their implications for the level and allocation of defense expenditures?

3. What are some of the key ways by which the Defense Department is currently attempting to increase its effi ciency?

4. What is the rationale for government actions to promote research and new technologies?

5. What are intellectual property rights? How well do they address the prob- lem of underinvestment in research?

6. What are the other ways by which government encourages the private production of knowledge? Why does government provide direct support to research?

FOCUS QUESTIONS

330 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

clear boundaries and in which there was no clear victory. By 2010, defense expenditures had risen to 5.6 percent of GDP, and at $717 billion (in 2005 dollars), they were at their highest point in constant dollars since World War II (see Figure 12.1).

In 2010, U.S. defense expenditures comprised the greatest share of GDP and the greatest share of total central government expenditures of any of the OECD countries, and were almost twice the OECD average for both these public expenditure measures. Although no low-income coun- tries had similar spending fi gures in 2010, several middle income coun- tries spent comparable amounts on national defense in terms of share of GDP and share of central government expenditures (see Table 12.1).

U.S. expenditures amounted to nearly half of all global military expen- ditures, raising questions about whether spending had become excessive. Expenditures were forecast to remain high even as the United States exited from the wars in Afghanistan and Iraq, although budgetary pres- sure limited the growth in these expenditures beginning in 2013.

Another consequence of the end of the Cold War related to the federal government’s expenditures on research and development (R&D). For dec- ades, defense spending and R&D spending had been closely intertwined. Indeed, in 2010, 57 percent of the federal government’s expenditures on R&D were defense related. There had been great spin-off s to the commercial sector—from advances in computer technology to advances in ceramics— but U.S. commercial interests almost surely did not benefi t from these expenditures to the same extent that they would have if more of govern- ment R&D expenditures had been directed at commercial uses. Nondefense R&D expenditures lagged behind those of the country’s competitors, and in the long run this could have signifi cant adverse eff ects on U.S. competitive- ness. There was thus a need for a new U.S. research and technology policy.

This chapter discusses these two important areas of government expenditure.

DEFENSE EXPENDITURES

Even though defense has long been the major item of expenditure at the federal level, it has traditionally received little attention from economists. Instead, military experts have led the way in determining how to achieve the country’s defense objectives. However, defense spending is fundamen- tally a question of resource allocation—the country wishes to receive the best defense for the money spent—and therefore hinges on economic reasoning.

As the defense budget is being formulated, the fundamental question is: How much? Conventional economic analysis—focusing on marginal

331Defense Expenditures

U.S. DEFENSE EXPENDITURES, 1929–2010

(A) Shows federal defense expenditures in 2005 dollars. (B) Shows federal defense expenditures as a percentage of GDP. Defense expenditures include both government con- sumption and gross investment.

FIGURE 12.1

Cold war peakVietnam war

peak

World War II peak

Korean war peak

Iraq and Afghanistan

conflicts

0

200

1925 1935 1945 1955 1965 1975 1985 1995 2005

1400

1200

1000

800

600

400

0%

5%

10%

1925 1935 1945 1955 1965 1975 1985 1995 2005

50%

40%

45%

35%

30%

20%

25%

15%

Federal Defense Expenditures, Fiscal Years ($ billions 2005)

A

Federal Defense Expenditures as a Percentage of GDP

B

SOURCE: Bureau of Economic Analysis, U.S. Department of Commerce, National Income and Product Accounts, Tables 1.1.5, 1.6.4, and 3.9.5.

332 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

*Share of central government expenditure **2009 ***2008

SOURCES: World Bank, World Development Indicators; and Stockholm International Peace Research Institute (SIPRI), Yearbook: Armaments, Disarmament and International Security.

TABLE 12.1 DEFENSE E XPENDITURES IN COMPAR ATIVE PERSPECTIVE

COUNTRY/COUNTRY GROUP % OF GDP

(2010) % OF EXPENDITURES*

(2009)

HIGH-INCOME COUNTRIES

Australia 1.9** 7.3

Canada 1.4 7.6

France 2.3 5.3

Germany 1.4 4.5

Greece 3.1 6.3

Japan 1.0 na

South Korea 2.7 13.4

Saudi Arabia 10.4 na

United Kingdom 2.7 5.7

United States 4.8 18.0

High Income, OECD 2.8 10.0

High Income, Non-OECD 5.5 na

MIDDLE-INCOME COUNTRIES

Brazil 1.6 6.3

Chile 3.2 15.4

China 2.0 16.1***

Colombia 3.7 18.9

Ghana 0.4 2.4

India 2.4 16.4

Indonesia 1.0 5.5

Pakistan 3.2 19.3

Russian Federation 4.0 14.0

South Africa 1.2 4.1

Thailand 1.5 9.5

Turkey 2.4 9.7

LOW-INCOME COUNTRIES

Bangladesh 1.2 10.2

Burkina Faso 1.5 10.4

Kenya 1.9 9.1

Middle and Low Income 2.0 na

Lower Middle Income 2.1 12.0

333Defense Expenditures

benefi ts and costs—is essential to answering this question, yet the most diffi cult questions go beyond simple economics. How do we assess the magnitude of the benefi ts from defense expenditures, and how should an appropriate military strategy be selected? The next two sections take up these questions. Later sections consider how to increase the effi ciency of our defense expenditures and how to address the problems associated with downsizing the military.

THE VALUE OF MARGINAL ANALYSIS

In allocating a given defense budget, one needs to consider the eff ect of the expenditures on various defense objectives. In evaluating whether we should spend more on defense, we similarly need to know how much extra “protection” we get from an extra expenditure of $1 billion.

The following example, provided by Charles Hitch, who was assistant secretary of defense in the Kennedy–Johnson administration, illustrates the role of marginal analysis.1 Assume that each missile has a 50 percent probability of success in killing its target. We have 100 targets that we would like to destroy. One hundred missiles sent at the targets would “achieve an expectancy of 50 kills, 200 missiles—75 kills, 300 missiles—87 kills,” as depicted in Figure 12.2. Clearly, there are very strong diminishing returns.

1 From C. J. Hitch, Decision Making for Defense (Berkeley: University of California Press, 1966), pp. 50–51.

THE ROLE OF MARGINAL ANALYSIS IN DEFENSE 

The relevant question is not whether we should have 500 missiles or no missiles, but how many extra kills we get from each additional missile. There may be sharply diminishing returns.

FIGURE 12.2

Total return

Marginal return 25

100 200 300 400 500 Number of Missiles

Number of kills

100

75

50

334 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

Each target can be destroyed only once, and some of the additional mis- siles will land on a target that had already been destroyed. Whereas the fi rst 100 missiles give us 50 kills, increasing the number of missiles from 400 to 500 increases the number of kills by only 3. We need to ask our- selves not whether it is worth the cost of 500 missiles to get 97 kills but whether it is worth the cost of 100 additional missiles to get 3 extra kills.

This kind of analysis is not easy. By relating expenditures to objectives and showing what one gets from additional expenditures, though, one can hope to make more rational decisions concerning how much is enough.

In making these assessments, however, one set of considerations is particularly hard to evaluate: deterrence.

DEFENSE STRATEGY

In the 1960s and 1970s, defense focused on deterrence—establishing a strong enough force that no one would contemplate attacking. Under President Ronald Reagan, another approach was advocated: the Strategic Defense Initiative (SDI), which attempted to provide a fail- safe defense capability. Today, the focus of debate is on several other issues: the capability of fi ghting two wars simultaneously, nuclear prolif- eration, terrorism, and chemical and biological warfare.

DETERRENCE The existence of strong capacities for retaliating serves to deter others from “misbehaving.” This thinking was central to the defense strategy of the United States and the Soviet Union after World War II. It was believed that with a large enough capacity to destroy the opponent, there would be little incentive for each country to attack. The capability of retaliating depended on the number of missiles that survived the fi rst strike, known as the second strike capability. Because the num- ber that survived depended on the size of the fi rst strike, each side had a strong incentive to create a larger arsenal; thus, the arms race was born.

However, with the end of the Cold War and the dismantling of much of Russia’s nuclear capabilities, no country had the capacity to maintain a sustained attack against the United States. This led many to question the relevance of the deterrence strategy, and, if it was still needed, whether the size of the U.S. deterrence capacity could be smaller. The United States reduced its nuclear arsenal, but critics suggested that there was scope for signifi cant further reductions.

TWO-THEATRE CAPABILITY With the threat of massive nuclear war subsiding, attention has focused on regional wars—a confl ict in the

335Defense Expenditures

Persian Gulf or on the Korean peninsula. There is a widespread view among military experts that we should have the capability of fi ghting at least two such localized wars simultaneously.

The problem is analogous to larger communities facing how large a fi re department to have. Assume that it takes twenty men to fi ght a moderate- sized fi re. Most of the time, there will be no fi res. Occasionally, there will be a fi re, and it is surely worthwhile to have a fi re department ready for that emergency. For simplicity, we assume it takes two hours to contain a fi re. Assume that the community has 10,000 buildings, and the chance of any building catching fi re in any two-hour span is one in a thousand. There is a small probability that in any two-hour span more than one fi re will occur. The question is: Is it worth having a second fi re department—an additional twenty fi refi ghters—on reserve for this small-probability event? It may or may not be. If it is, we then need to ask, is it worth having a third fi re depart- ment on reserve for the very slight chance that there will be three or more fi res? The probability of using each additional fi re department—and thus the expected benefi t—diminishes markedly.

In one respect, the fi re analogy is inappropriate. Typically, we do not worry about an arsonist waiting to light a fi re until the fi re department is busy putting out a fi re elsewhere. In the area of defense, however, this is a real concern: potential enemies or troublemakers may wait until the United States is occupied with fi ghting in one theatre to start trouble in another. Many worry that without a two-theatre capacity, the United States would be inviting trouble. From this perspective, the major role of the two-theatre capability is its role as a deterrent.

Currently, the dominant view—reaffi rmed in the 2010 Quadrennial Defense Review Report2—is that the United States should have the capac- ity to fi ght two localized wars of moderate scale simultaneously. It is the ability to fi ght the second war that is compromised when the defense budget is cut.

Of course, matters are seldom clear-cut. One of the main bottlenecks in fi ghting a war is logistics—in particular, getting troops and equipment to the scene of battle. These bottlenecks are likely to be particularly binding in the early stages of a war, and at times when troops and equipment must be deployed rapidly. Given a span of a few months, ships being used for other purposes can be converted into transports for military equipment and personnel. This can be done even when troops need to be deployed quickly, provided there is willingness to force the domestic economy to undertake greater short-run costs: the military could, for instance, order civilian aircraft to transport troops. Thus, greater expenditures

2 U.S. Department of Defense, Quadrennial Defense Review Report, February 2010, p. vi.

336 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

for defense readiness reduce the costs should a crisis occur. The question then is: How much of an “insurance premium” are we willing to pay?

The issue, then, is not (as it is sometimes put) simply whether the United States could fi ght on two fronts simultaneously, but rather, what would be the additional costs and risks on the second front? Almost all the serious confl agrations that the United States has faced in the past fi fty years have lasted months, if not years. The Iraq and Afghanistan wars have, indeed, been the longest-lasting wars in America’s history. If this is the likely pattern in the future, then the need for a full capacity for rapid deployment on a second front may not be a high imperative.

However, increased threats from nonconventional warfare—whether from forces representing failed states or nonstate entities, or from new weapons such as those of mass destruction and those operating in cyberspace—have required the Defense Department to enhance capacity elsewhere. New priorities include unmanned aircraft systems for intelli- gence, surveillance, and reconnaissance; rapid deployment special opera- tions forces; and counterinsurgency and counterterrorism. Skills and

GAME THEORY, THE ARMS RACE, AND THE THEORY OF DETERRENCE

T he 1994 Nobel Prize in economics was awarded to three economist–mathematicians who pioneered in the development of game theory: John Nash of Princeton, John Harsanyi of the University of California at Berkeley, and Rein- hold Selten of Rheinische Friedrich-Wilhelms-Uni- versität in Bonn, Germany.

Most of game theory is based on the postulate of rationality. For example, Bradley not only plays to win but also believes Boris, his opponent, plays to win; and believes Boris believes that he is playing to win, and so forth. Thus, each player, in deciding what to do, puts himself in his rival’s shoes: If I do X, what will my rival do? But in putting himself in Boris’s shoes, Bradley reasons that Boris will be thinking about what he would do, assuming Boris is rational.

Much of the development of game theory was supported by the Department of Defense, in an

attempt to understand better how the Soviet Union would respond to what the United States did. Game theory was used both as the basis of the theory of deterrence, which underlay American defense strat- egy, and to explain the arms race.

Although there is some debate about the ade- quacy of the theory of deterrence and the arms race that game theory provided, it has subsequently found wide use throughout economics. For instance, game theory has been used extensively in the analysis of patent races, to understand when and whether pat- ents actually spur innovation. Some models suggest, for example, that in some circumstances, there may be little impetus for research, as the lead fi rm gets suffi ciently ahead of its rivals and then rests on its lau- rels. The fi rm’s rivals know that should they accelerate their research, the fi rm will respond, and thus, their prospects of winning the patent race are minimal.

337Defense Expenditures

technology required for success in these arenas are quite diff erent from those that have marked the military of the past. A current concern is that expenditures are unbalanced—too much directed at dealing with the “old” threats, too little on the new.

ARMS PROLIFERATION After the fall of the Soviet empire, atten- tion focused on denuclearizing Ukraine and Kazakhstan (the two for- mer Soviet republics, besides Russia, that inherited nuclear weapons), reducing the number of nuclear warheads in Russia, and countering arms proliferation in general (see case study, “Converting Swords into Plowshares”).

The arms industry is profi table, and it is one of the few industries in which Russia is globally competitive. With many small countries interested in buying weapons, there is real danger that the arms will feed the ethnic and political struggles that seem to be proliferating around the world. As the United States has itself been a major arms sup- plier to the rest of the world, it is diffi cult for this country to criticize others for doing as it does, even though it claims to be more responsible in its arms sales.3

The United States and seventy-nine other countries signed a nonpro- liferation treaty in 1968, attempting to restrict access to nuclear weap- ons and the methods of delivering them.4 The United States has used the threat of economic sanctions to enforce the agreement since it passed its fi rst nonproliferation sanctions law in 1974, targeted at India for what India called its “peaceful nuclear explosion.” Nonproliferation laws impose sanctions against individuals, private entities, and governments that engage in proliferation activities.

For example, in 1993 and again in 1996, in retaliation for exporting missiles to Pakistan, China was threatened with losing its most favored nation status—under which its exports to the United States faced the lowest tariff s that the United States imposes on any country (other than Canada and Mexico). In May 1998, India and Pakistan both successfully tested nuclear bombs, joining the exclusive nuclear club. In response, the United States and other countries imposed sanctions on both coun- tries, cutting off nonhumanitarian assistance and restricting loans to the Indian and Pakistani governments.

3�From 2003 to 2010, the United States exported $170.8 billion worth of weapons (in constant 2010 dollars), which was higher than any other country and accounted for 39 percent of the world’s total. The second largest arms exporter for the same period was Russia, whose $81.1 billion worth of weapons sales was 18 percent of the world’s total. (Source: Richard F. Grimmett, Conventional Arms Transfers to Developing Nations, 2003–2010, Congressional Research Service, September 2011, Table 31.) 4 By 2010, 190 countries were members of the nonproliferation treaty. [Source: United Nations Offi ce for Disarmament Aff airs, “Treaty on the Non-Proliferation of Nuclear Weapons (NPT),” http://www .un.org/disarmament/WMD/Nuclear/NPT.shtml.]

338 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

The alleged reason that the United States went to war in Iraq was over concerns that the country was building nuclear weapons. Even though UN inspectors had contended that this was not the case, the United States argued otherwise. After the invasion, it turned out that the UN inspectors had been correct. Reports concerning Iraq’s nuclear capacities were based on fl awed and fabricated evidence, throwing into question the competency of America’s intelligence services and the credibility of its public assurances.

One of the most heavily invoked nonproliferation laws at present is the Iran, North Korea, and Syria Nonproliferation Act of 2006. Military confrontation between North Korea and the United States has been avoided periodically over the past two decades when standoff s between the two countries have ended with the signing of agreements committing the United States to provide aid in return for North Korea’s dismantling of its nuclear potential, although these agreements have invariably broken down amid mutual recriminations of noncompliance.

Subsequently, the major concern over nuclear proliferation has been with Iran. Iran claims that the nuclear enrichment facilities it wishes to

CONVERTING SWORDS INTO PLOWSHARES

T he end of the Soviet Union posed a challenge: how to make sure that the highly enriched uranium in the nuclear warheads in the mis- siles in Russia, Ukraine, and Kazakstan did not fall into the wrong hands. The United States proposed an innovative approach: it would buy the nuclear material, bring it to the United States, and de-enrich it so that it could be used in nuclear power plants.

There was, however, one obstacle: the United States was in the process of privatizing the govern- ment agency responsible for enriching uranium. The U.S. Enrichment Corporation (as it was called) would make most of its money from selling enriched uranium to power plants, and it did not want com- petition. Opponents of privatization argued that this was just one of the reasons that one should not

privatize the company. The same processes that are used for making enriched uranium for nuclear power plants can further enrich it, to the level required for nuclear weapons. It seemed reckless to turn over responsibility for making enriched uranium to a pri- vate fi rm, whose incentives were to sell the enriched uranium to the highest bidder.

In the end, the privatization went ahead. It slowed the process of bringing the enriched uranium into the United States, but it could not stop it. On the other hand, the alleged benefi ts of privatization—greater effi ciency—never mate- rialized. The fi rm has remained perilously close to bankruptcy, with occasional proposals being dis- cussed in Congress to renationalize it. This is one privatization that no one considers a real success.

SOURCES: The early story of the battles over privatization of USEC are told in J. Stiglitz, “Unfair Trade Laws and Other Mischief,” in Globaliza- tion and Its Discontents (New York: W. W. Norton & Company, 2002), pp. 166–179. The saga is updated in D. Koplow, “Subsidies to Uranium Enrichment,” in Nuclear Power: Still Not Viable without Subsidies (Cambridge, MA: Union of Concerned Scientists, February 2011), pp. 62–72.

339Increasing the Efficiency of the Defense Department

construct are for peaceful purposes (nuclear energy), but others are not convinced, setting the stage for a number of confrontations between Iran and other countries committed to limiting nuclear proliferation.

CHEMICAL AND BIOLOGICAL WEAPONS AND TERRORISM Nuclear nonproliferation sanctions have now been expanded to prevent the proliferation of chemical and biological weapons as well, collectively referred to as weapons of mass destruction (WMD).

The 1991 Gulf War focused attention more broadly on the threat of rogue states, and their use of terrorism and chemical and biological weapons. That a small vial containing a deadly virus could cause massive death is a frightening prospect. Even if Iraq had not engaged in biologi- cal warfare, the numerous maladies plaguing those returning from the Gulf War were a constant reminder of the threat of chemical weapons (although it subsequently turned out that many of these maladies may have been related to spent uranium used in many U.S. military weapons). A bipartisan eff ort under both President Bush and President Clinton led to an international agreement to ban chemical weapons.

Establishing enforceable international agreements and cooperation to combat biological and chemical weapons and terrorism is one part of the U.S. strategy for dealing with these threats. Research aimed, for instance, at building the capacity to develop antidotes and antibodies quickly, is another. Given the magnitude of the danger, there is concern that the United States is spending too little on such research.

INCREASING THE EFFICIENCY OF THE DEFENSE DEPARTMENT

No matter how much is spent on defense, it should be spent well. Recent attention has been focused on improving the Defense Department’s pro- curement policies and on reorganization.

DEFENSE PROCUREMENT

During the height of the Cold War, thousands of fi rms focused on sup- plying the research and weapons required for the expanding military. Producing for the military was diff erent in many ways from producing for the civilian sector. Civilian aircraft had many potential buyers. The various aircraft manufacturers produced the plane that they thought best

340 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

met the aviation needs of the world, and then tried to persuade each of the airlines that their aircraft was better than their rivals’. If an aircraft man- ufacturer produced an advanced military plane, though, there was usually only one customer: the Department of Defense. Usually, the Department of Defense had only one supplier of that particular plane; it was simply too costly to have two or more fi rms producing exactly the same plane. Many of the procurement problems described in the following sections apply across all branches of government, but some are particularly acute in the Defense Department because of the unique nature of its purchases.

STANDARD PROCEDURES To ensure that it obtains the best price, the government usually resorts to competitive bidding: diff erent contrac- tors tell the government the price at which they are willing to deliver, say, one thousand tanks of a given specifi cation, and the government pur- chases the tanks from the lowest bidder. Frequently, however, there are major cost overruns—that is, the costs exceed the producer’s original estimate. Sometimes the contract calls for these costs to be shared by the government and the private contractor; such contracts are called cost- sharing contracts. Even when the contract does not explicitly call for sharing of cost overruns, the government may absorb all or a signifi cant fraction of the additional costs. The contractor may claim that the cost overruns are a result of changes in the design specifi cation; such changes almost always accompany the development of a new weapon, particularly when the development occurs over a period of several years, and it is often diffi cult to ascertain to what extent the cost overruns are, in fact, a result of the design changes. In other cases, the private contractor simply says that it cannot complete the contract without further funds; the govern- ment then has the choice of losing all that it has already spent or negotiat- ing a settlement with the contractor. Even if the government were to sue the contractor for breach of contract, the delays in the development and deployment of the weapons could be very costly.

The prevalence of cost overruns means that the public seldom has an accurate view of the cost of a ship, a defense system, a tank, and so on at the time a commitment is made to purchase them. It also means that the government seldom knows whether it has, in fact, let out the contract to the lowest-cost producer5; all that it knows is that it let out the contract to the lowest bidder.

What are the reasons for these cost overruns? In the case of new weapons, errors in estimating costs are common—but why should there

5 Notice that at the time the cost overruns occur, what limited competition there was before the con- tract was let out no longer exists: it would, in general, be costly or impossible for the government at that point to turn to other potential suppliers.

341Increasing the Efficiency of the Defense Department

be a bias in these errors? That is, why should there be a systematic tendency to underestimate the costs? Part of the reason has to do with the competitive bidding process: potential contractors know that they have to produce a low bid to win. The system of cost sharing (implicit or explicit) means that there is relatively little penalty associated with bidding too low. There is, however, a penalty associated with bidding too high, par- ticularly when other fi rms are bidding low (using, say, their most optimis- tic estimates of costs): the high-bidding fi rm will fail to get the contract.

The system of cost sharing has a further disadvantage in addition to reducing the penalty for underbidding: the winner of the contract has little incentive to be effi cient. Cost-plus contracts further reduce the incentive to be effi cient; for example, a contract where the government pays whatever it costs to develop weapons plus, say, 10 percent, actually provides incentives to be ineffi cient—the more the fi rm spends, the more it gets from the government.

Why does the government engage in cost sharing, with all its obvi- ous disadvantages? Part of the reason for cost sharing is the uncertainty inherent in the development of a new weapons system. The best that a fi rm can do, as we have said, is to provide an estimate of these costs. If there were fi xed-fee contracts (contracts under which the contractor was paid a fi xed amount, regardless of the eventual cost), the contractor would have to bear considerable risk; even if it were very effi cient, there is some chance that it would encounter diffi culties in the development of the system that would increase its costs way beyond the fi xed fee it would receive, in which case it might incur an enormous loss. If fi rms (or their managers) are risk averse and insist on being compensated for bear- ing risks, they will all put in high bids, representing their estimate of the actual costs plus a fee for bearing the risk. The government is in a better position to bear the risk. By agreeing to a cost-sharing contract, it absorbs much of the risk, but at the same time reduces the incentives for effi ciency.

Even though this provides an important rationale for cost-sharing contracts, some critics of the Pentagon argue that other forces are at work when the government agrees to pay all or part of a cost overrun. A large number of military offi cers, upon their retirement from the armed forces, take up positions in private industry—in particular, with defense con- tractors. Critics say that this provides an incentive for these offi cers to be accommodating to the requests of defense contractors.

Supporters of the current system, though admitting that it is far from perfect, point out that there is a healthy level of competition among defense contractors and that this competition provides at least some limit to the extent of ineffi ciency. Any contractor that performed consistently worse than other fi rms would fi nd itself having diffi culty obtaining contracts.

342 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

EXCESSIVE MONITORING AND PROCUREMENT REFORM Many believe that the most important problem with defense contracting is not cost overruns, but the detailed procedures that are instituted to ensure that the government is not cheated. These require excessive monitoring and record keeping, and reduce the fl exibility that fi rms need to produce in a cost-eff ective manner. Indeed, when the government buys exactly the same equipment that the private sector buys, such as a jet engine, as a result of government procurement procedures the government winds up paying substantially more—because it costs the fi rm substantially more to produce the item. In 1994, the Department of Defense reformed its pro- curement procedures, which had added substantially to its cost of doing business. It sought to make its procedures conform more closely to those that were standard in the private sector. Because, with military down- sizing, fewer fi rms would be able to produce exclusively for the military, fewer fi rms would be willing to go through the hassle required to conform to government procurement rules.

There were several aspects to procurement reform. One was the pro- motion of dual-use technologies—that is, technologies that could be used by both civilian and military customers. This would have two advan- tages. For many of the items, the larger market would reduce the cost of production and enhance the degree of competition. Another aspect to procurement reform was to emphasize “off -the-shelf” products—products that are available commercially. Thus, rather than buying made-to-order T-shirts, the Defense Department might choose among T-shirts already in the market. There is no compelling reason why defense T-shirts need to be designed diff erently from T-shirts used elsewhere.

Although this reform improved general defense procurement effi - ciency, the Defense Department’s weapon acquisition programs have remained on the list maintained by the Government Accountability Offi ce (GAO; formerly the General Accounting Offi ce) of areas in which the government is at high risk for overpaying since 1990. For exam- ple, the GAO’s seventh annual assessment of selected weapon pro- grams (FY 2008) found that research and development costs were 42 percent higher than originally estimated, total acquisition cost was $296  billion over budget (FY  2009 dollars), and the average delay in delivering initial capabilities had increased to 22 months.6 In response, the Senate and the House of Representatives unanimously passed the Weapon Systems Acquisition Reform Act of 2009. This law creates a new Offi ce of Cost Assessment and Program Evaluation (CAPE) at the Defense Department, reporting directly to the Secretary of Defense.

6�U.S. Government Accountability Offi ce, Defense Acquisitions: Assessments of Selected Weapon Programs, Report to Congressional Committees, March 2009, p. 6.

343Defense Conversion

It  also calls for more extensive testing of new weapons before they enter production to ensure that they perform as designed, and gives fi eld commanders a greater voice in formulating the technical specifi - cations for new weapons.

DEFENSE CONVERSION

The end of the Cold War brought with it both a downsizing of the military and a restructuring. This process was referred to as defense conversion. Any change of this magnitude—a redeployment of even a couple of percent of GDP—puts strains on the economy, and strains on the economy get translated quickly into political pressures. Although the major rationale for defense expenditures is protection of the United States, political support for the Department of Defense also arises from the fact that defense expenditures generate so many jobs. A sound mac- roeconomic policy should be able to maintain the economy at or near full employment, but individuals in defense-related jobs worry that defense cutbacks will cost them their jobs, and that even if they could get other jobs, they would not be as good as their current jobs.

These concerns are exemplifi ed by the controversy over base closures. As the size of the military forces was reduced, fewer bases were required. However, closing a base means a loss of jobs. Congressional representatives see maintaining jobs in their districts as one of their key responsibilities.

The politics of base closure were so intense that in 1988, Congress cre- ated the Commission on Base Realignment and Closure (BRAC) to recom- mend which bases should be realigned or closed. The commission depoliticized the process (and took the heat off Congress), because although Congress could veto the commission’s entire recommendation, it could not remove individual bases from the list. This process was repeated in 1991, 1993, and 1995 under the auspices of the Defense Base Realignment and Closure Act of 1990. The fi rst four BRAC commissions recom- mended the closing or realignment of approxi- mately 450 installations, eliminating 20 percent of the Defense Department’s 1988 pre-BRAC capacity. This included closing 97 out of 495 major domestic installations. All recommendations

DOWNSIZING DEFENSE • Defense conversion entails redeploying resources.

There may be signifi cant costs to redeployment.

• It is important that resources, such as bases, be redeployed effi ciently.

• Even though, in the long run, society benefi ts from redeployment, particular individuals and communities may be adversely affected. Markets fail to provide insurance against economic dislo- cation, and accordingly there may be grounds for government assistance.

344 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

were accepted by Congress. These four BRAC rounds produced net sav- ings of approximately $16.7 billion through 2001, including the cost of environmental cleanup, with recurring savings beyond 2001 estimated to be about $6.6 billion annually.7 In 2005, a fi fth BRAC commission recom- mended an additional 182 closures or realignments, including twenty-two major closures and thirty-three major realignments, estimated to save taxpayers $15 billion over twenty years. Again, all recommendations were accepted by Congress.8

Defense downsizing led to high levels of unemployment and other serious adverse economic impacts in many communities that relied heavily on defense expenditures. Some have argued that the federal government has special responsibility to aid these communities. There is a market fail- ure rationale for such aid: the inability to obtain insurance against events. In eff ect, by providing transitional help in the face of large shocks, society is providing a form of insurance not available through the market. Critics of such aid argue that this is simply part of capitalism: communities are buff eted by all sorts of shifts in demand. They ask, why should one kind of shift in demand—that arising from a change in the demand by govern- ment for defense services—warrant preferential treatment?

In practice, the government has tried to soften the adverse eff ects by slowing down the pace of transition and giving substantial advance warn- ing. Disengagement from the confl icts in Iraq and Afghanistan, coupled with large budget defi cits and a rapidly mounting public debt, means that the United States is facing several more rounds of defense downsizing. As a fi rst step, in 2009, the Secretary of Defense proposed canceling or sig- nifi cantly curtailing weapons programs with a projected cost of at least $126 billion.

ACCOUNTING AND THE DEFENSE DEPARTMENT

The Iraq and Afghanistan wars brought out a new set of problems in the conduct of our defense: the process of budgeting and appropriation. When the Bush administration went to war, it estimated that the costs would be in the range of $60 billion. The eventual tab to the government will turn out to be somewhere in excess of $2 trillion.

For a long time, it was diffi cult even to ascertain how much the gov- ernment was spending, partly because some of the spending was not fully

7 U.S. Army Base Realignment and Closure Division, “BRAC Frequently Asked Questions,” http://www .hqda.army.mil/acsim/brac/faq.htm. 8 Defense Base Closure and Realignment Commission, 2005 Defense Base Closure and Realignment Com- mission Report, pp. iii–viii.

345Research and Technology

disclosed (in eff ect, it was partially hidden) and partly because the money for the war was dribbled out in a series of small “emergency” appropria- tions that became almost regular in nature.

Part of the problem, too, arose because of the way that the government does accounting, which is diff erent from that used by most private sector fi rms. It is based on cash receipts and outlays, rather than accrual—that is, if a fi rm undertakes an obligation today for which it pays tomorrow, it must account for the obligation today. For instance, many fi rms pay a part of a worker’s salary in the form of deferred compensation—that is, pay that the worker will receive some time later. However, the fi rm must treat those future costs as part of today’s production costs. This is not so for the government, which does not have to account for those future costs until they actually occur.

In the case of the military, these costs can be enormous. For instance, close to half of the troops returning from the Afghanistan and Iraq wars have some form of disability, and many have multiple disabilities. Future government expenditures for disability pay and health care costs are con- servatively estimated at around a trillion dollars.

Worse still, misleading accounting systems can give rise to distorted incentives. Most important, if true total costs are understated, the level of current spending may be beyond what citizens would choose if they knew the true costs.

RESEARCH AND TECHNOLOGY

The United States government (like governments in other countries) has long played an important role in basic research and the diff usion and development of new technologies.

For example, a key factor in the remarkable increase in agricul- tural productivity in the last century was establishment of land-grant colleges by the Morrill Land–Grant Acts of 1862 and 1890, followed by the creation of the Cooperative Agricultural Extension Service in 1914 by the Smith–Lever Act to disseminate knowledge generated by the land-grant colleges. The nuclear age was a direct outgrowth of govern- ment research, largely conducted in government-run laboratories. Other marvels of the twentieth century, such as the jet engine and supercom- puters, were largely by-products of defense R&D expenditures. The Internet and the biotech revolution are based on government funded research as well.

346 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

The federal government continues to invest heavily in R&D— in 2010, it spent $141 billion on research and development. While the government is currently increasing expenditures on nonde- fense R&D, defense R&D still accounts for 57 percent of total R&D expenditures—about the same as the average for 1960–2010, but down from the peak of 89 percent in 1954 (see Figure 12.3). Moreo- ver, although the United States accounts for one-third of global R&D spending (see Table 12.2), when public and private expenditures are combined as a percentage of GDP, the United States spends substan- tially less on nondefense R&D (2.3 percent) than Japan (3.3 percent), South Korea (3.2 percent), or Taiwan (2.9 percent).9 It is perhaps not surprising that the United States has not been doing as well at innovating, as evidenced, for instance, by patent applications.

SOURCE: Executive Offi ce of the President, Offi ce of Management and Budget, Budget of the United States

Government, Fiscal Year 2013, Historical Tables, Tables 9.8 and 10.1.

GOVERNMENT EXPENDITURES ON

RESEARCH AND DEVELOPMENT,

FISCAL YEARS 1949–2010

The fi gure shows defense, nondefense, and total federal expenditures on research and development, in 2005 dollars.

The U.S. government spent $125 billion on research and

development in 2010, more than half of it on defense projects.

FIGURE 12.3

Total government R&D expenditures

Government nondefense R&D expenditures

Government defense R&D expenditures

0

10,000

20,000

30,000

19551945 1965 1975 1985 1995 2005

($ millions 2005)

80,000

90,000

100,000

110,000

120,000

130,000

140,000

60,000

70,000

40,000

50,000

9 National Science Foundation, Science and Engineering Indicators, 2012, Table 4-20. (The percentages noted are for 2009, except for South Korea, which is for 2008.)

347Research and Technology

In both Japan and South Korea, the rate of patent applications, adjusted for population, is far higher than in the United States.10

Considerable concern has been expressed in recent years at the slow rate of growth in the United States—productivity growth from 1973 to 1995 averaged less than 0.5 percent, and although it rebounded to 1.4 percent from 2000 to 2007, it subsequently dropped to 0.7 percent from 2008 to 2010. This is in contrast to the two decades after World War II, when it averaged slightly less than 3 percent.

The three sources of growth in productivity are increases in capital, improvements in human capital (the quality of the labor force) through educa- tion and experience, and technological change. The major source of the slow- down in productivity growth since 1973 has been the slowdown in the pace of technological change—hence the concern with the underinvestment in R&D.

TABLE 12.2 RESE ARCH AND DE VELOPMENT E XPENDITURES IN COMPAR ATIVE PERSPECTIVE (2010)

COUNTRY % OF GDP

(PPP, $US BILLIONS) GERD*

(PPP, $US BILLIONS) SHARE OF TOTAL GLOBAL

R&D SPENDING (%)

TOP 10

United States 2.8 415.1 33.2

China 1.5 149.3 11.9

Japan 3.4 148.3 11.8

Germany 2.8 82.9 6.6

South Korea 3.4 49.0 3.9

France 2.2 47.4 3.8

United Kingdom 1.8 39.3 3.1

India 0.8 32.5 2.6

Canada 2.0 25.9 2.1

Brazil 1.1 23.9 1.9

SELECTED OTHERS

Russian Federation 1.0 22.9 1.8

Taiwan 2.3 18.9 1.5

Sweden 3.6 12.9 1.0

Israel 4.3 9.4 0.8

Singapore 2.5 7.4 0.6

Finland 3.9 7.2 0.6

*GERD: Gross expenditures on research and development, includes public and private, current and capital expenditures.

SOURCE: Battelle, 2012 Global R&D Funding Forecast, December 2011.

10 World Intellectual Property Organization (WIPO), World Intellectual Property Indicators, 2011 edition, Table P1.

348 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

MARKET FAILURES

The slowdown by itself might not be a rationale for government interven- tion, at least from an economist’s perspective. The question is: Is there a market failure—a reason why the market, left to its own, would under- invest? The answer is yes: knowledge is, in many respects, like a public good, and as we saw in Chapter 5, the private sector will underinvest in public goods.

Recall from that discussion that there are two critical properties of public goods: the undesirability of exclusion (the zero marginal cost of providing the good to an additional individual)11 and the impossibility of exclusion. Research and development (or more accurately, knowledge, the product of research) has the fi rst property, and often has the second as well. Giving information to additional individuals does not detract from the total amount of knowledge available.12 However, if knowledge were provided freely, it would not pay for anyone to produce it. Thus, either the government must provide for the production of knowledge, through direct support of R&D, or it must ensure that individuals or fi rms that produce knowledge are compensated for doing so.

There are many ways that innovators receive rewards for their invest- ments in research. With the “fi rst mover advantage,” the fi rst fi rm in the market with a new product can often get loyal customers for itself, and is in a better position to make further advances. In many sectors, much of the knowledge associated with innovation is “tacit,” hard to convey from one fi rm to another. In sectors like metallurgy, little reliance is placed on intel- lectual property. In fact, to get a patent, one must disclose a considerable amount of information, and many fi rms would rather maintain their advan- tage through trade secrets. For instance, the formula for Coca-Cola is not patented: its discoverers chose instead to keep the formula in a bank vault.

INTELLECTUAL PROPERTY RIGHTS In some sectors, however, to enable individuals to reap rewards from their knowledge-creating activi- ties, others must be legally excluded from using it, or at least from using it without compensating the creator. The government does this by cre- ating “property rights” in knowledge; that is, it grants a patent, which gives the discoverer of the knowledge exclusive use of the knowledge, including the right to license others to use it, for a limited period of time.

11 The observation that knowledge was like a candle, which, even as it lights another candle, does not have its own light diminished, was made forcefully by Thomas Jeff erson. 12 This fact should not be confused with the fact that the return that an individual can obtain from a particular piece of knowledge may depend on how many other people know that piece of information. A monopolist of a piece of information may be able to obtain a return that would not be possible if the information were made freely available.

349Research and Technology

A copyright gives the author of a work the exclusive right to use and market his or her own writings. Patents and copyrights establish intellectual property rights, which ensure that inventors and authors can appropriate some of the fruits of their labor. The framers of the U.S. Constitution recognized the importance of intellectual property rights when they gave the newly established federal government the right to grant copyrights and patents to encourage creative activity.

Not all ideas and discoveries are patentable—mathematical theorems are not, whereas algorithms may be—and even when a particular discov- ery is patentable, it is often possible to invent around the patent. Thus, whereas it is possible to patent a drug, it is often easy to devise a slight modifi cation of the same drug with the same medicinal properties.

In determining the life of a patent, the government faces a trade- off . By extending the life of the patent, it provides greater incentives for private fi rms to engage in R&D; on the other hand, the knowledge pro- duced will not be used effi ciently for a longer period of time. Assume, for instance, that a fi rm has discovered a new, less expensive way of mak- ing a product—so much less expensive that the fi rm can undercut all its rivals. By patenting the discovery, the fi rm will be in a monopoly position; therefore, less of the product will be produced than if the knowledge were freely disseminated.

The loss resulting from the patent can be seen in Figure 12.4, in which we have drawn the market demand curve for a medicine. The cost of production

EFFECTS OF PATENTS

A patent system results in a lower output than with free dissemination of knowledge. The deadweight loss is given by the triangle BDF.

FIGURE 12.4

Demand curve

Quantity of medicine

Q0

P0 = C0

C1 A

E

B

D

F

Q1

Price

350 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

prior to the invention was C0, and the competitive equilibrium was thus (P0, Q0) at point D. The fi rm that makes the small invention has a lower cost of production, C1. It charges a price just below P0, getting the entire market; its profi ts are thus ABDE, and its sales are Q0.13 If the information about the new innovation were made freely available, the price would fall to C1 and the quan- tity would rise to Q1. Giving the fi rm with the patent a monopoly on its knowl- edge has resulted in output being smaller than it otherwise would be, and a deadweight loss of BDF. When the patent expires, the price will fall to C1, but the return to the innovator will drop to zero. Thus, the longer the life of the patent, the greater the deadweight loss associated with the ineffi ciency of giv- ing the fi rm a monopoly over the use of the information—but the greater the return to the innovators and hence the greater the incentives for innovation.

13 The fi rm is not, of course, free to off er any price and quantity, because at any price on or above C0�, the fi rm loses its monopoly position. In other words, the fi rm maximizes its profi ts when subject to the constraint that P1 # C0.

THE SCOPE OF THE PATENT: CAN THE HUMAN BODY BE PATENTED?

Intellectual property rights enable inventors to appropriate some part of the fruits of their innovative activity. However, there are impor- tant issues in defi ning intellectual property. For instance, the traditional life of a patent is seventeen years. The longer the life, the greater the returns to the innovator, but the longer the period over which the innovator exercises monopoly power, the longer the period during which the production of the innovation is restricted.

The life of the patent is not the only issue in defi ning intellectual property. An equally important issue is scope: How broad should the patent be? Consider recent advances in genetic engineering. Should the fi rst person who develops a genetically engineered tomato be granted a patent for (a) all genetically engineered plants and animals, (b) all genetically engineered plants, (c) all genetically engineered tomatoes, or (d) this particular variety of tomato?

Controversy over the scope of a patent has recently been brought home forcefully by research on human genes. Decoding the millions of genes that make a human being is an immensely important task; it should, for instance, enable the development of drugs that can address innumerable diseases and health conditions. For these purposes, though, it is not necessary to decode the entire gene; most of the relevant “information” is contained in only a fraction of the gene. Some fi rms claim that if they can decode that information, they should be granted a patent on the gene. Others fi nd the entire idea of patenting a human gene repugnant.

Most of the important decisions on intellectual property are made in the courts, on a case-by-case basis. Prior to the Diamond v. Chakrabarty case, 447 U.S. 303 (1980), life forms were considered a part of nature and were not patentable. Since then, genetically engineered bacteria have been deemed patentable because they do not occur naturally.

351Research and Technology

Intellectual property rights may enhance the private provision of knowledge, but they typically interfere with its effi cient utilization. Giving a fi rm a monopoly on the production of a new product will result in too small a level of production—this is true even if there is a rationale in pro- viding the monopoly. There is a trade-off between “dynamic effi ciency,” which provides fi rms with an incentive to innovate, and “static effi ciency,” which requires that fi rms produce up to the point at which price equals marginal cost.

More recently, there has been a concern that intellectual property rights, especially if they are not well designed, may actually impede innova- tion. The most important input into any research is knowledge; by making access to prior knowledge more diffi cult or expensive, follow-up research is discouraged. Moreover, in certain industries, innovators have been faced with a patent thicket, a complex web of claims on intellectual property; the innovator faces the threat of a suit from any one of these who might claim that his or her intellectual property rights have been infringed.

The Patent Offi ce has also issued patents on whole genes whose function is known, and has developed guidelines on dealing with DNA sequences of less than a whole gene (“fragments”), as they do not have a known function.

In 2009, the American Civil Liberties Union (ACLU), on behalf of medical associations, doctors, and patients, sued Myriad Genetics, an American molecular diagnostic company, before the U.S. District Court for the Southern District of New York, to declare the patents on two genes related to breast cancer invalid (Association For Molecular Pathology et al. v. United States Patent and Trade- mark Offi ce). Women who have the two genes have more than 50 percent chance of getting cancer; knowledge that they have the gene is important to maintain close monitoring to prevent the develop- ment of the disease. In this case, the social value of the patent was especially questionable, as there was a successful global project to decode the entire

human genome; in the absence of Myriad’s research, it would have been discovered shortly after. The cost of the patent, however, has been enormous, as the patent holder has insisted on remaining the monopolist for the supply of tests—even when bet- ter tests were available at lower costs elsewhere. Patients who were given false assurances that they did not have the genes because they had the less accurate tests have subsequently faced deaths that might have been avoided or delayed. In 2013, the Supreme Court upheld that naturally occurring genes cannot be patented.

Knowledge is one of the essential inputs in the production of further knowledge. Broadening the scope of patents can increase the cost of the “knowledge” input into follow-on research, and thus actually reduce the overall pace of innovation. The academic community, in particular, has been concerned that excessively broad patents could stifl e their research.*

*The private sector has argued (following on the ideas of Coase, discussed in Chapter 6) that property rights can never be excessively broad. It would always pay a fi rm to grant a license to a follow-on inventor who might make use of its innovation. As we noted there, however, Coase’s analysis assumed no transactions costs, perfect information, and competitive markets. In practice, these assumptions are not satisfi ed, and excessively broad patents can stifl e innovation.

SOURCE: “Genetics and Patenting,” Human Genome Project Information, http://www.ornl.gov/sci/techresources/Human_Genome/ elsi/patents.shtml.

352 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

Complex products require putting together a multitude of ideas, and America’s intellectual property regime puts impediments to doing this. Even worse, some fi rms specialize in getting patents that they can use to “hold up” others, demanding large compensation for the use of an even minor patent, while knowing that the cost and delay in inventing around the patent may be very expensive. BlackBerry was subject to such a holdup, and had to pay more than $600 million, lest it be shut down. A small company threatened similarly to shut down the imports of most of Intel’s advanced microprocessors—and the Apple and HP computers that used them. The Supreme Court has recently demonstrated concern that America’s intellectual property rights regime, including the way that it has been enforced, is unbalanced and may be impeding innovation.

As we have noted, many discoveries are not protected by intellectual property rights, or are protected only imperfectly. The key lesson is that intellectual property rights, as important as they are, only partially solve the market failures associated with the production of knowledge. Other government interventions are required.

OTHER MECHANISMS FOR ENCOURAGING PRIVATE PRODUCTION OF KNOWLEDGE Governments also encourage the private production of knowledge by subsidizing a critical input into R&D, scientifi c man- power, through its support of education programs (see Chapter 14), and through a tax credit, the incremental research and experimentation tax credit (R&E credit).14 Firms that increase their research expendi- ture are allowed to deduct 20 percent of their incremental expenditures from their tax; the government pays, in eff ect, 20 percent of the marginal costs. The reason the credit is on incremental expenditures is to lower the marginal cost of research—the government does not want to subsidize research that would otherwise have been undertaken anyway—and thus to maximize the “bang for the buck.” The advantage of a tax credit over a direct subsidy is that the government does not have to choose among applicants: those who believe that their projects are worth undertaking put up their own money, matched by the government (on a one-to-four basis) at the margin. The disadvantage of a tax credit is that it does not distinguish research projects which have large externalities from those which have none. Indeed, there is even some question about whether some of what gets supported is real research—overhead and marketing research may be passed off as R&D expenditures.

14 The credit has not been a permanent feature of the tax code. It was originally introduced as part of the Economic Recovery Act of 1981, and has been renewed almost continuously over the past three decades. There is concern that the incentive eff ects have been attenuated because companies cannot rely on the credit—it was temporarily extended for the fourteenth time in December 2010 (after it expired on December 31, 2009), which also marked the ninth retroactive extension in a row.

353Research and Technology

GOVERNMENT DIRECT SUPPORT

Government supports research not only indirectly, such as through the patent system and tax credits, but also directly. Such direct sup- port is particularly important for basic research. Basic research adds to our underlying store of fundamental knowledge (as opposed to applied research, which is intended to result more immediately in a new prod- uct or manufacturing technique). The results of basic research are more likely to be inappropriable; even if they are appropriable, the social cost of restricting the utilization of this knowledge may be particularly high, as basic research is such an important input into the production of other ideas. As a result, it is generally agreed that some form of direct support for basic research is required if there is to be an effi cient allocation of resources to it.

Far more controversial, however, is government support to applied research. Sometimes, governments pick out particular industries to encourage through supporting research in those areas. Such funding or other policies, such as protection from foreign imports, aimed at promoting particular industries are referred to as industrial policies. Advocates of these policies argue that there is no clear line between basic and applied research; that much of applied research has huge spillovers—externalities to others that are not appropriated by the inno- vator—and that, hence, there will be underinvestment. The transistor was invented in Bell Labs, for example, and was of immediate benefi t to AT&T, which paid for the research, but the benefi ts to the world clearly extended far beyond.

Critics have argued that the government has a terrible record of pick- ing winners, because of its lack of “profi t motive,” and that the government should not be in the business of directing the economy. Michael Boskin, chairman of the Council of Economic Advisers under the fi rst President Bush, was quoted as saying, “It makes no diff erence whether the economy produces potato chips or computer chips.”

Advocates of government support for technology reply that the gov- ernment actually has a remarkably good record of picking winners.15 More importantly, they argue that the objective of government support is not to push the economy in particular directions, but simply to identify areas

15 Government attempts at coal gasifi cation and other synthetic fuel projects, as well as renewable energy projects, are usually cited as examples of these failures. Interest in industrial policies grew in the 1980s when Japan’s economic success was often contrasted with the problems facing the U.S. econ- omy. Japan had long used industrial policies, for instance, to encourage the development of the com- puter chip industry. The U.S. government provided support for the computer chip industry: it helped establish a consortium of U.S. producers, called Sematech, which it then subsidized. The industry has enjoyed a resurgence in the United States to the point at which the country once again dominates the world market. Government subsidies to Sematech have now ceased.

354 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

in which there appear to be large spillovers, so that without government support, there will be underinvestment. Several examples could be cited. For instance, the telecommunications industry owes much to the federal government, from the fi rst telegraph line between Baltimore and Washington in 1842 to the development of the Internet. In the nineteenth and twenti- eth centuries, government support for agricul- ture research and its dissemination was key to the enormous increase in productivity in that sector—which used to be the core of the economy—whereas in the twentieth century, government was central in high-technology industries, from computers to biomedicine.

In its recent technology programs, the government has tried to learn from its past mistakes. To improve incentives, it has focused on part- nerships, in which the private sector puts up at least 50 percent of the costs; it has provided grants on a competitive basis, with evaluations by outside experts, and with projects in a wide variety of areas competing against each other; and it has instituted a system of ongoing review to terminate unsuccessful projects. (In the past, the reluctance to termi- nate such projects has been a major problem; congressional representa- tives from districts in which projects are located typically argued that success was just around the corner.) Ironically, although these reforms seem to have worked, they have also undermined support for technol- ogy research programs: special interests no longer see them as cash cows and therefore devote few resources to lobbying for them.

KNOWLEDGE, PUBLIC GOODS, AND EXTERNALITIES Knowledge is, to a large extent, a public good. Through patent protection (or other forms of intel- lectual property rights), inventors can appropriate some of the returns to their inventive activity. Still, there are likely to be externalities. Moreover, the appropriation interferes with the effi cient diffusion and utilization of knowledge. This provides the rationale for government support. It is particularly cogent for basic research.

REVIEW AND PRACTICE

SUMMARY

1. In determining “how much is enough” in defense expenditures, marginal analysis—looking at the incremental benefi t from increased expenditures— is essential.

2. Alternative views of overall defense strategy are important determinants of defense needs. During the Cold War, attention was focused on deterrence; President Reagan initiated the Strategic Defense Initiative; more recently, debate has focused on the desirability of having

355Review and Practice

a two-theatre capability. There is increasing con- cern over nuclear proliferation and cyberspace threats, and over the use by rogue states and non- state entities of chemical and biological warfare and terrorism.

3. The Defense Department has attempted to increase its effi ciency by reforming its procure- ment policy, with greater reliance on perfor- mance standards and off -the-shelf purchases and on the development of dual-use technologies. Traditional procurement systems, including cost- plus contracts, may have contributed signifi cantly to high costs by providing attenuated or perverse incentives. Despite procurement reforms, most major weapon acquisition programs are still well over budget and behind schedule.

4. Base closures and other reductions in defense spending will cause economic hardship in some areas. There is disagreement on what and how much the government should do to help aff ected communities. The absence of insurance markets insuring against these economic risks provides a rationale for government action.

5. The government has long had a role in the support of research and the development of technology. Knowledge has both of the properties of a public good, so without some government intervention, there will be underinvestment in research.

6. The government encourages innovation by estab- lishing intellectual property rights (through pat- ents). Although patents thus allow innovators to appropriate some of the returns to their innova- tion, and thus provide incentives for the produc- tion of knowledge, they interfere with its effi cient use. But the patent system may actually impede innovation.

7. The government also provides direct support for research, especially basic research, and indirect support through tax credits and the support of education—producing the scientists who are the critical input into research. Government pro- grams in support of technology are aimed at iden- tifying areas in which there are large spillovers. Although such industrial policies have been con- troversial, the government has a credible record

of picking winners. Recent government policies emphasize public–private partnerships, with the private sector providing a signifi cant fraction of the costs.

KEY CONCEPTS

Copyright

Cost overruns

Cost-plus contracts

Cost-sharing contracts

Defense conversion

Deterrence

Dual-use technologies

Fixed-fee contracts

Incremental research and experimentation tax credit (R&E credit)

Industrial policies

Intellectual property rights

Patent

Peace dividend

Second strike capability

Strategic Defense Initiative (SDI)

Two-theatre capability

Weapons of mass destruction (WMD)

QUESTIONS AND PROBLEMS

1. Assume the government has decided to install a missile system designed to provide a second strike capability with 100 missiles. It is now con- sidering whether to increase the number to 110. Assume you are on the congressional committee that must approve the increased expenditure. List some of the questions you might ask to ascer- tain whether the increased expenditures are desirable.

2. Should military offi cers and Defense Depart- ment offi cials be proscribed from working for private defense contractors for a period of sev- eral years after termination of their government service?

356 CHAPTER 12 DEFENSE, RESEARCH, AND TECHNOLOGY

3. In what ways is the purchase of a hammer or of labor services by the military diff erent from the purchase of an MX missile system? How does this aff ect government procurement policies in these two areas?

4. Consider the following proposed system of bid- ding: the contract would be awarded to the low- est bidder at the price bid by the second-lowest bid. Why does such a system encourage bidders to bid their true costs of production honestly?

5. In some cases, the government runs duplicative projects, particularly at early stages of develop- ment. What do you think are the advantages of doing this? The disadvantages?

6. Consider the following proposal for reducing cost overruns. Two contractors would be given a contract to produce a tank of a given specifi - cation. Producer A would be reimbursed for the actual costs incurred by producer B, and vice versa. Explain why this system might induce each fi rm to produce effi ciently. If the two fi rms were essentially identical, what risk premium would they require in bidding on the contract? Under what conditions might the risk premium be large? What are other possible pitfalls in this system?

7. Imagine that Congress is considering a bill to reduce the current seventeen-year life of patents to eight years. What negative eff ects might this change have on the rate of innovation? What pos- itive eff ect might it have for the economy?

8. Suppose that many years ago, one inventor received a patent for orange juice, and then another inventor came forward and requested a patent for lemonade. The fi rst inventor main- tained that the orange juice patent should be inter- preted to cover all fruit juices, whereas the second inventor argued that the original patent included only one particular method of producing one kind of juice. What trade-off s does society face in set- ting rules for deciding cases such as these?

9. Why might a company invest in R&D even if it does not believe it will be able to patent its discovery?

10. Some public goods are “local”—that is, they pro- vide services only to those living in a particular locality. What might be meant by a “global” pub- lic good? Why might knowledge be thought of as a global public good? What implications does this have for government policy in a small country? Can a country be a “free rider” in basic research?

11. In what sense is international security, including eff orts to stymie nuclear proliferation, a global public good?

357

HEALTH CARE

13

Three separate, and somewhat confl icting, concerns about the U.S. health care system continue to inform the debate over health care: excessive costs, limited insurance coverage, and the fi scal strains that providing health care imposes on government.

Health expenditures as a percentage of GDP are higher in the United States than in any other country, amounting to $2.6 trillion in 2010—about 18 percent of GDP, or $8400 per capita. Even so, health status, as recorded by such measures as life expectancy and infant mortality, is actually lower here than in many countries that spend considerably less. Further- more, expenditures are rising rapidly, partly due to increased quantity of services, and partly due to prices increasing faster than the price level in general, as indicated in Figure 13.1.1

1 Changes in the quality of health care services pose serious measurement problems. Most economists agree that conventional price measures overstate the rate of infl ation in health care costs, because they do not adequately refl ect improvements in what is being provided, although there is little consensus over the magnitude of the measurement errors.

1. What are the fundamental problems facing the health care system today?

2. What role does the gov- ernment play in the health care sector today?

3. What are the reasons for government action? What are the market failures? What are the ways in which the market for health care diff ers from markets for other commodities? Why do concerns about distribu- tion play a particularly large role in health care?

4. What are some of the problems that arise from the fact that a large frac- tion of health care costs are covered by insurance? What problems confront insurance providers?

5. What are the key public policy issues today? What are some ways in which costs can be contained or insurance coverage extended? What are some of the major proposed reforms to Medicare, a public program that provides health insurance to the aged, and Medicaid, a public program that pro- vides health care to those who otherwise might not be able to aff ord it?

FOCUS QUESTIONS

358 CHAPTER 13 HEALTH CARE

A second problem is that, at least before the Patient Protection and Aff ordable Care Act (PPACA) of 2010, many individuals lacked health insur- ance. Health insurance coverage declined in the United States in the years before the pasage of the PPACA. Slightly more than one of eight people were uninsured in 1987, and that number had risen to one in six before passage of the new health care law. Lack of coverage has become a major political issue because of the anxiety that it imposes, especially on middle-class workers. Because most health care insurance is provided by employers, workers who lose their jobs lose their insurance. The poor are usually covered by pub- lic programs, such as Medicaid. It is typically workers in low-paying jobs, workers without regular employment, and the self-employed who are left

SOARING HEALTH CARE COSTS AND

EXPENDITURES, 1960–2010

(A) Health care expenditures have been rising rapidly— faster than GDP. (B) This is partly because the volume

of services has increased, and partly because health care

prices (measured by the medical consumer price index [medical

CPI]) have been rising faster than prices in general

(measured by the CPI).

FIGURE 13.1

$ bi

lli on

s

% G

D P

0 1960 1965 1970 1975 1980 1985 1990 1995 2000 20102005

3000

2500

2000

Total health expenditures ($ billions)

Total health expenditures (% of GDP)

1500

1000

500

0

20

18

16

14

12

10

8

6

4

2

B as

e pe

ri od

: 1 98

2– 19

8 4

= 1

0 0

0 1960 1965 1970 1975 1980 1985 1990 1995 2000 20102005

450

400

350

300

250

CPI

Medical CPI

200

150

100

50

A

B

SOURCES: Centers for Medicare and Medicaid Services, Offi ce of the

Actuary, National Health Statistics Group, National Health Expenditure Accounts, 1960–2010; and Bureau of

Labor Statistics, Department of Labor, http://www.bls.gov/data.

359Health Care

uncovered. Although those without coverage can get access to emergency health care services—under the Emergency Medical Treatment and Labor Act (EMTALA) of 1986, hospitals must provide emergency services regard- less of ability to pay—these individuals often delay getting adequate treat- ment until their condition qualifi es as an emergency, thereby raising the total costs of the treatment. Because those without coverage do eventually get treatment, for which they may not pay, lack of coverage also results in cost shifting, a shift of the costs of unpaid bills to others.

Soaring health costs put a strain on government budgets, pushing up government expenditures for Medicare and Medicaid. As the fed- eral budget defi cit has soared, these rising government expenditures for health care, illustrated in Figure 13.1A, and for the two principal govern- ment programs, Medicare and Medicaid, in Figure 13.2, have become an increasing source of concern. Federal government expenditures on Medicare and Medicaid, which in 2010 accounted for 5.5 percent of GDP, have been projected to almost triple as a share of the economy over the next sixty years, growing to 14.9 percent of GDP by 2070. This growth is due to an increase in the elderly population, greater quantities of ser- vices being used by the aged, and health care prices that are anticipated to rise faster than the rate of infl ation. Combined with Social Security, total federal expenditures on these programs have been projected to grow to 23.6 percent of GDP by 2085.2 (To put this into perspective, total federal

2�See Congressional Budget Offi ce, Historical Budget Data—January 2012 Baseline, Publication 42911, and CBO’s 2011 Long-Term Budget Outlook, Publication 41486. The health expenditure fi gures after 2010 include Medicare, Medicaid, Children’s Health Insurance Program (CHIP), and exchange subsidies.

MEDICARE AND MEDICAID EXPENDITURES, 1966–2010

Expenditures on Medicare and Medicaid, the two major government health programs, have been rising rapidly and are imposing large fi scal strains, which will continue to grow in coming decades.

FIGURE 13.2

SOURCE: Centers for Medicare and Medicaid Services, Offi ce of the Actuary, National Health Statistics Group, National Health Expenditure Accounts, 1960–2010.

0 1965 1970 1975 1980 1985 1990 1995 2000 20102005

600

500

400

300

Medicaid

Medicare

200

100

$ billions

360 CHAPTER 13 HEALTH CARE

expenditures have historically amounted to about 20 percent of GDP.) If such expenditures do grow unchecked, then either taxes will have to be raised, the defi cit will soar, or other government programs will have to be cut back. However, from 2010 to 2012, health spending grew at an annual rate of just 1.1 percent in real per capita terms, which is the lowest rate in the fi fty years such data have been collected and much lower than the 6 percent growth rate from 2000 to 2010. Much of the slowdown in health care expenditure growth is attributed to PPACA, signed into law in March 2010, and this encouraging trend has dramatic ramifi cations for the nation’s long run fi scal soundness.

This chapter provides an overview of the U.S. health care system. It examines how diff erent market failures have shaped the role of govern- ment in the system; tracks the sources of the problems of excessive costs, limited insurance coverage, and fi scal strains; and discusses the diff erent eff orts to reform the system in recent years.

THE HEALTH CARE SYSTEM IN THE UNITED STATES

Governments may be involved in the health care system in a variety of ways: directly paying for health care, subsidizing individual purchases of health care and health insurance, providing health care services, fi nanc- ing and conducting research, preventing the spread of communicable dis- eases, and regulating drugs and medical devices. In the United States, the government is involved in each of these areas, but to a lesser extent than in many other countries. For instance, in Great Britain, the major health care delivery system is run by the government.

The U.S. health care system, with its mix of public, private, and non- profi t providers, is also one of the most complex systems in the world. Most health care consumers in the United States fall into one of four groups: the poor, who receive medical care through the federal government’s Medic- aid program; the aged, whose basic medical costs are paid by Medicare, another program of the federal government; employed individuals covered by employer-provided health insurance; and the uninsured, who some- times purchase health care services directly, but often receive uncompen- sated care, that is, the providers of the services are not compensated and the cost is passed on to others. In addition, a small number of those who do not enjoy employer-provided insurance purchase their own insurance.

For virtually everyone, a substantial part of health care is paid not by the individual receiving the treatment, but by a third party—the government

361The Health Care System in the United States

or a health insurance provider. As Figure 13.3 shows, consumers prepay (directly or through their employers) 34 percent of the cost of personal health care in the form of private insurance premiums. Another 48 percent of health care costs are paid by the government. Only 14  percent of the money spent on medical care comes from direct payments by consumers.3 The percentage of total expenditures which individuals bear themselves has declined steadily over the past half century. Total third-party payments

3 These fi gures on the overall sources of funds disguise the fact that the share of out-of-pocket costs varies a great deal among consumers (depending on their insurance plan) and among kinds of health expenditures. For example, in 2010, only 3.2 percent of hospital costs were paid directly by consumers, compared to 9.6 percent of physician and clinical expenditures and 28.3 percent of nursing home and continuing care retirement community costs.

What matters, of course, is not just the average amounts paid by consumers, but also their mar- ginal costs. At the margin, what fraction of incremental costs do consumers bear? On average, the share of marginal costs borne by individuals is likely to be less than these numbers indicate. See Centers for Medicare and Medicaid Services, Offi ce of the Actuary, National Health Statistics Group, National Health Expenditure Accounts, 1960–2010.

SOURCES OF FUNDS FOR PERSONAL HEALTH CARE EXPENDITURES, SELECTED YEARS

Americans pay a relatively small percentage of health care costs out of pocket, and the share of such payments has been falling in recent decades. “Other” includes spending by charitable organizations and industrial on-site health services.

FIGURE 13.3

Out-of-pocket (55.9%)Private

health (21.1%)

Government (21.4%)

Other (1.6%)

1960

Out-of- pocket

(26.9%)

Private health

(28.3%)

1980

Government (40.5%)

Other (4.3%)

Out-of- pocket

(17.3%)

Private health

(34.9%)

Government (42.5%)

Other (5.3%)

2000

Out-of-pocket (13.7%)

Private health

(34.1%)

Government (47.9%)

2010

Other (4.3%)

SOURCE: Centers for Medicare and Medicaid Services, Offi ce of the Actuary, National Health Statistics Group, National Health Expenditure Accounts, 1960–2010.

362 CHAPTER 13 HEALTH CARE

(by government and health insurance) in 1960 amounted to 43 percent of total personal health care expenditures; by 2010, they were 82 percent.4

As indicated in Table 13.1, U.S. health expenditures are a third higher than the OECD average as a share of GDP, and are the highest among OECD countries as a share of total government expenditures; in contrast, out-of-pocket expenditures as a percentage of total health expenditures in the United States is among the lowest of OECD countries. As a share of GDP and share of total government expenditures, middle- and low- income countries spend, on average, half and one-third of the amount spent by OECD countries on health, respectively; out-of-pocket expendi- tures in lower–middle-income and low-income countries account for half of total health expenditures.

Not only are U.S. health care expenditures larger (both as a share of GDP and per capita) than in any other advanced industrial country, but the numbers have been increasing as well. The complaint, though, is not so much with how much is spent, but with how poor the outcomes have been—that is, with the eff ectiveness of the health care system. With the aging of the population and advances in technology, it might be natural for the United States to increase its expenditures on medicine, but the concern is that outcomes—measured, for instance, by life expectancy or infant mortality rates—are far poorer than those in other advanced countries that spend but a fraction of the amount that the United States spends. In some respects, U.S. outcomes are comparable to those in poor developing countries, particularly when disaggregated by location (high-income versus low-income communities) and demographics (age, race, ethnicity).5

Some of this is attributable to lifestyle decisions—contributing, for instance, to the high incidence of obesity or the existence in parts of the country of “food deserts,” where it is diffi cult to get access to fresh fruits and vegetables. Some of this is attributable to the fact that the United States has a higher level of inequality than other countries, and has not provided easy access to health care for those who cannot aff ord it.

There are two implications of these observations: (a) There is ample room for improvement in both the effi ciency and eff ectiveness of America’s health care sector. (b) Given the concern about health, it makes sense to focus not just on the delivery of medical services, but also on other aspects of our economy and society that aff ect health, such as smoking and drink- ing, general eating habits, and exercise regimes.

4 Centers for Medicare and Medicaid Services, Offi ce of the Actuary, National Health Statistics Group, National Health Expenditure Accounts, 1960–2010. 5 National Center for Health Statistics, Centers for Disease Control and Prevention, U.S. Department of Health and Human Services, Health, United States, 2011 with Special Feature on Socioeconomic Status and Health, May 2011; and Summary Health Statistics for the U.S. Population: National Health Interview Survey, 2011, December 2012.

363The Health Care System in the United States

TABLE 13.1 HEALTH EXPENDITURES IN COMPARATIVE PERSPECTIVE (2009)

COUNTRY/COUNTRY GROUP

PUBLIC AND PRIVATE HEALTH EXPENDITURES

(% OF GDP)

PUBLIC HEALTH EXPENDITURES

(% OF GOVERNMENT EXPENDITURES*)

OUT-OF-POCKET EXPENDITURES

(% OF TOTAL HEALTH EXPENDITURES)

HIGH-INCOME COUNTRIES

Australia

Bahrain

Canada

France

Germany

Greece

Japan

South Korea

United Kingdom

United States

High Income, OECD

8.5

4.5

10.9

11.7

11.3

10.6

8.3

6.5

9.3

16.2

12.1

17.1

10.9

17.0

16.0

18.0

13.0

17.9

12.3

15.1

18.7

16.6

18.6

18.1

15.5

7.2

11.4

35.3

14.9

34.8

10.4

12.4

13.7

MIDDLE-INCOME COUNTRIES

Brazil

Chile

China

Columbia

Ghana

India

Indonesia

Pakistan

Russian Federation

South Africa

Thailand

Turkey

Middle Income

Lower–Middle Income

9.0

8.2

4.6

6.4

6.9

4.2

2.4

2.6

5.4

8.5

4.3

6.7

5.6

4.3

6.1

15.6

10.3

17.9

9.2

4.1

6.9

3.6

8.5

9.3

14.0

12.8

na

5.8

31.0

34.4

41.2

7.9

43.2

50.0

35.2

56.8

28.8

17.7

16.5

16.0

36.5

49.6

LOW-INCOME COUNTRIES

Bangladesh

Burkina Faso

Kenya

Low Income

3.4

6.4

4.3

5.1

7.5

16.3

5.4

na

65.9

35.6

51.2

50.1

*Share of central and local government expenditures

SOURCES: World Bank, World Development Indicators; and World Health Organization, National Health Account Database.

364 CHAPTER 13 HEALTH CARE

THE PRIVATE SECTOR

Private insurance providers play a major role in the U.S. health care system. There are two dominant forms of private insurance.

In one, called fee-for-service plans, doctors are reimbursed on the basis of the service provided—this is, the traditional insurance plan. Typ- ically, insurance reimburses a fi xed amount of the total bill, up to some limit. Often, the insurance company pays only an amount in excess of a certain level, called the deductible; the fraction of the excess over this deductible that the individual pays is called the co-payment. Insurance policies also have caps on what they will pay for certain services.

In the other form of private insurance, called managed care, a third- party payer (such as an insurance company, a corporation, or the govern- ment) mediates between physicians and patients, specifying the treatments that they will cover and negotiating fees for these medical services. The most common managed care institution is called health maintenance organization (HMO), which oversee a system under which the patient pays a fi xed annual amount, called a capitation fee, to cover medical costs. Members of an HMO must go to the doctors in that HMO. In addition, to see a specialist, they must be referred by their HMO family doctor.6 HMOs have grown rapidly in the past few years, especially among those who are covered by employer-provided health plans, mainly because they have helped contain cost increases through a combination of expenditure effi - ciency and an emphasis on the prevention and early detection of disease.

THE ROLE OF GOVERNMENT

We have already noted the large role of government in the U.S. health care system, and the huge growth over the past two decades of the two most important government programs, Medicare and Medicaid. Both pro- grams were motivated largely by a concern that two major segments of the population, the poor and the aged, did not have health insurance coverage (and, in many cases, could not have obtained it, or at least not at premiums that could generally be regarded as aff ordable). The larger of the two pro- grams is Medicare, which provides medical care for everyone over 65 and for certain disabled persons. Medicare has three components—hospital insurance (Part A) and supplementary medical insurance, which pays for

6�There are also hybrids of the two systems. An example is preferred provider plans that pay for all the costs of visits to its preferred providers. If the member goes to another provider, the plan pays for only a fraction of the costs. Similarly, many HMOs allow patients to go to other doctors, but reimburse only a fraction of the costs.

365The Health Care System in the United States

physicians’ services (Part B) and prescription drugs (Part D).7 Recipients must make some contributions to supplementary insurance, but their contribution covers only about a quarter of the costs of even this part of Medicare. The rest comes from general tax revenues. Hospital costs are covered by a 2.9 percent payroll tax (paid by working people as part of the Social Security tax, although half is nominally paid by employers). The money from the 2.9 percent payroll tax goes into a trust fund, which has been near exhaustion a number of times. Each time, quick fi xes have enabled the fund’s life to be extended.

Medicaid provides medical care for certain low-income families with dependent children, and for most poor aged, blind, and disabled persons. Unlike Medicare, Medicaid is administered by the states. The eligibility standards are set by each state within federal guidelines, and the states are required to pay for between 17 and 50 percent of the benefi ts, depend- ing on the per capita income of the state. The federal portion is paid out of general tax revenues. States also pay 50 percent of administrative costs.

The rapid growth in the costs faced by states under Medicaid has imposed serious problems for states, just as the rapid costs for health care services under Medicare and Medicaid poses one of the most serious fi scal problems for the federal government. It is worth noting, though, that reform proposals that simply shift the burden to individuals will not improve standards of living: they simply change how services are paid for. By contrast, improvements in the effi ciency of the health care delivery system could result in lower costs (whoever pays) without adverse eff ects on health. By some estimates, if the United States had as effi cient a health care system as several of the European countries that have better out- comes, the government’s fi scal defi cit would be eliminated entirely.

OTHER EXPENDITURE PROGRAMS

Three other categories of direct expenditures should be mentioned: the Public Health Service, VA hospitals, and expenditures on medical research and teaching. The Public Health Service, which traces its roots to 1798 with the establishment of the Marine Health Service, was created in part to address important externalities associated with communicable diseases. (If individuals have a contagious disease, they impose an exter- nality on others by increasing the likelihood that the disease will spread.) The Public Health Service played a pivotal role in reducing or eliminating contagious diseases such as smallpox and measles.

7��There is also a Medicare Part C, which is a combination of Parts A and B, provided by Medicare- approved private insurance companies.

366 CHAPTER 13 HEALTH CARE

Another major government medical program is the VA hospitals, run by the Veterans Administration. Dating back to the establishment of the Veterans Administration in 1930, the VA hospitals provide medical care for those injured while serving in the armed forces and for other veter- ans. As those who fought in World War II have grown older, requiring more medical care, increasing demands have been imposed on the sys- tem. VA medical costs reached $46 billion in 2010, almost half of total VA expenditures.

Expenditures for medical research and teaching are another impor- tant category of expenditures. The government runs some very success- ful research laboratories (the National Institutes of Health) and supports extensive research in universities and medical schools. Medical research expenditures amounted to $49.3 billion in 2010.8

TAX EXPENDITURES

For more than four decades, there has been concern about the uninsured. Without insurance, a serious illness or accident can cause a huge fi nancial burden on an individual. Accordingly, the government has encouraged employers to provide insurance by giving their employees a signifi cant tax advantage: employees do not have to pay taxes on the value of their employer-provided health benefi ts. In 2011, the value of this tax break was estimated to be $163 billion.

Government provides a second substantial tax break for health expen- ditures: medical expenses in excess of 10 percent of income are deduct- ible from income. The estimated value of this tax break in 2011 was $8.3 billion.9 The rationale for this provision is that taxes should be based on some measure of ability to pay, and that large medical expenses—to the extent that they are not voluntary—reduce an individual’s ability to pay.

These tax expenditures encourage both health insurance purchases and medical expenditures. They eff ectively lower the price an individual must pay for insurance, and health insurance lowers the price individuals have to pay for medical care.

There is concern not only about the effi ciency consequences of our tax expenditures—that is, the excessive consumption of medical ser- vices that is induced—but also about their equity. Tax benefi ts are clearly larger for those at higher marginal tax rates (wealthier individuals). The unemployed and those at low-paying jobs with few or no benefi ts

8�Centers for Medicare and Medicaid Services, Offi ce of the Actuary, National Health Statistics Group, National Health Expenditure Accounts, 1960–2010; and National Center for Veterans Analysis and Statistics, Department of Veterans Aff airs. 9 Offi ce of Management and Budget, Analytical Perspectives, Fiscal Year 2013 Budget, Table 17-1.

367Rationale for a Role of Government in the Health Care Sector

obviously do not enjoy the tax expenditures associated with employer- provided medical insurance.

Prior to 1981, these tax expenditures were so regressive that total fed- eral expenditure, per capita on high-income individuals, including tax expenditures, actually exceeded those on middle-income individuals, even though direct expenditures through Medicare and Medicaid were considerably lower on a per capita basis for upper-income groups. In fact, expenditures per capita on the middle-income group were lower than those on any other group. Changes to the tax code in the 1980s reduced marginal tax rates and increased the minimum expenditure required for tax deductibility of medical expenses. These changes made health expen- ditures less regressive.�10

RATIONALE FOR A ROLE OF GOVERNMENT IN THE HEALTH CARE SECTOR

At the beginning of this chapter we noted that rising costs are a major concern about the U.S. health care system. Typically, an increase in the price of some commodity is not by itself taken as grounds for government intervention. Market prices change in response to demand and supply, resulting in changes in scarcity.�11 Similarly, the fact that Americans spend more on health care (as a percent of GDP) than other countries may sim- ply be a refl ection of preferences. (We would not infer from the fact that Americans prefer larger cars that there is something wrong with the U.S. automobile market requiring government intervention.) The health care market, though, is rife with imperfections, and some of these do lead to excessive expenditures. In this section we begin with general theory, and then focus on two special aspects of the U.S. health care market.

We will fi rst discuss four market failures. The fi rst two, imperfect information and limited competition, were discussed in general terms in

�10 See G. Wilensky, “Government and the Financing of Health Care,” American Economic Review 72, no. 2 (May 1982): 205; J. Gruber, “The Tax Exclusion for Employer-Sponsored Health Insurance,” Work- ing Paper 15766, National Bureau of Economic Research, February 2010; and M. Pauly, T. McGuire, and P. Barros, eds., Handbook of Health Economics, vol. 2 (Waltham, MA: North Holland, 2012). �11 In competitive markets, increases in expenditures can result from a shift in the demand curve along a given supply curve and/or from an upward shift in the supply curve along an inelastic demand curve. Both of these may well have occurred, with demand shifting as Americans have become more health conscious, and the supply curve shifting as health care costs have been rising faster than the costs of goods and services in general. As we comment below, however, the concern is that the increases in health care expenditures do not refl ect the normal workings of competitive markets.

368 CHAPTER 13 HEALTH CARE

Chapter 4. The two others relate to the large roles of nonprofi t institu- tions and of the insurance industry in the health care sector.

Even if there were no market failures, there might be a role for gov- ernment: some people might have such a low income that they could not afford or would choose not to get adequate health care. The final section addresses these concerns about the provision of health care to the poor.

IMPERFECT INFORMATION

When consumers go to a doctor, in large measure they are buying the doctor’s knowledge and/or information. As a patient, the consumer must rely on the doctor’s judgment as to what medicine is required or whether an operation or other procedure is advisable. Because they lack medical expertise, patients generally cannot eff ectively assess and eval- uate their doctors’ advice. They may not even be able to tell whether a doctor is qualifi ed. This explains why government has long taken a role in licensing doctors and regulating the drugs they can administer to their patients.�12

These information problems are far more severe than those faced by consumers in other areas. In the case of repeat purchases, like groceries, consumers either are able to judge the quality of the products themselves or come to rely on a grocery store (say, for the freshness of its vegetables). Typically, however, individuals do not have repeat purchases of medical procedures, such as kidney transplants, heart bypass surgery, or even ulcer treatments. In the case of products like cars, independent rating agencies, such as Consumers Union (which publishes Consumer Reports), test the product and describe its strengths and weaknesses. However, there are simply too many doctors and hospitals and too many procedures for that to be feasible in the health care sector: a hospital may be good at one procedure and weak at another. Success may depend on subjective factors like a doctor’s personal manner, which may work well with some individuals and not with others. Even “report cards”—say, the fraction of heart bypass patients who survived one year—may not be fully informa- tive, because hospitals in one region may have an intake of sicker patients than hospitals in another.

Insurance companies also encounter information problems relating to doctors and patients. Like patients, they must rely largely on doctors

�12 Kenneth Arrow has emphasized the importance of imperfect information in medical markets. See K. J. Arrow, “Uncertainty and Welfare Economics of Medical Care,” American Economic Review 53 (1963): 941–973.

369Rationale for a Role of Government in the Health Care Sector

to determine what procedures are necessary and useful. Imperfect infor- mation about patients creates problems in the market for insurance, as discussed later (see the section on adverse selection).

LIMITED COMPETITION

Imperfect information decreases the eff ective degree of competition.�13 A fi rm selling a standard commodity, like a Sony television, knows that it can attract customers away from other stores by lowering its price. Customers can easily ascertain where they are getting the best value for their money.

By contrast, potential patients who see a doctor with lower prices than competitors may infer that this doctor is not in great demand and is there- fore trying to attract more customers; but the apparent lack of demand for the doctor’s services may also suggest to them that he or she is not a good doctor.

By the same token, the heterogeneity of medical services makes price and quality comparisons diffi cult and thus inhibits the eff ective dissem- ination of information. My neighbor may be pleased with the medical treatment that he obtained from his doctor, but if his medical problems are diff erent from mine, his satisfaction is no assurance that I will be pleased if I go to the same doctor. And if I hear that one doctor charges more than another doctor, to evaluate whether one is a better buy, I would have to know precisely what services were performed.

The practices of the medical profession may compound the inevi- table limitations of competition resulting from imperfect information. In the past, doctors were not allowed to advertise. In other contexts, restrictions on advertising have been shown to raise prices (because they inhibit competition). Thus, for instance, several states now allow advertising for eyeglasses, and in those states, there has been a dramatic decrease in the price of eyeglasses. In many states, doctors and hospitals have been allowed to advertise, but the eff ects on competition are not yet clear. In particular, much of advertising is not informative. Given the potentially huge variations in quality, simply knowing about price is only of limited value.

Probably of more value will be government assessments of hospitals, for example, in terms of the numbers of particular procedures they per- form and the outcomes, and accessible information about the relationship between certain procedures and health outcomes. Some costly procedures have been shown to have little statistical relationship to health outcomes.

13 See, for instance, S. Salop, “Information and Monopolistic Competition,” American Economic Review (May 1976): 240–245.

370 CHAPTER 13 HEALTH CARE

One of the objectives of the PPACA was to undertake more assessments of the benefi ts and costs of diff erent medical practices to help increase the effi ciency of the health care system.

There are other measures by which doctors can attempt to restrict price competition. It has been suggested, for instance, that “lowering fees might provoke one’s colleagues to deny a surgeon hospital privileges or seek to damage his reputation.”14

Finally, in many small communities, there are few doctors from whom to choose. The grouping together of doctors into HMOs and preferred provider plans may actually be signifi cantly lessening competition in these areas.

Moreover, there is also limited competition among hospitals. Most smaller communities have at most only a few hospitals. In the event of an emergency, an individual seldom is in a position to choose from among many hospitals. Even when there is time to make a choice, the choice is made not by the individual but by his or her doctor.

ABSENCE OF PROFIT MOTIVE

Another important diff erence between medical markets and standard competitive markets is the large role of not-for-profi t organizations in the provision of health care. For a long time, not-for-profi t hospitals vastly outnumbered their for-profi t counterparts; even today, the majority of hospitals in the United States are not-for-profi t institutions. Such institu- tions do not view their objective as simply minimizing the cost of deliver- ing medical care, or maximizing profi ts.

In contrast, in theory, for-profi t hospital chains respond more clearly to incentives for effi ciency, and at times, some of the large for-profi t hos- pital chains seemed both effi cient and innovative.15 More recently, the concern has shifted to perverse incentives for lowering quality, espe- cially in circumstances under which quality is hard to judge, and much of the money the for-profi t institutions collect goes to pay high salaries of administrators and dividends, not to provide better health care. For-profi t blood providers, for instance, earned a reputation for gathering blood from drug addicts and others whose blood was more likely to carry dis- ease. As a result, that market has been dominated by not-for-profi t fi rms.

Incentive problems encountered with both for-profi t and not-for-profi t providers arise in large measure from imperfect information. If consumers

14 For a detailed examination of these issues, see M. Gaynor and R. Town, “Competition in Health Care Markets,” Chapter 9 in Handbook of Health Economics, vol. 2, ed., M. Pauly, T. McGuire, and P. Barros (Waltham, MA: North Holland, 2012), pp. 499–638. �15 In 1986, Humana, a large private health care chain, received considerable attention for the successful implanting of an artifi cial heart in its Louisville, Kentucky, hospital.

371Rationale for a Role of Government in the Health Care Sector

could easily ascertain the quality of what they purchased, both types of pro- viders would have stronger incentives to achieve effi ciency and high quality.

SPECIAL CHARACTERISTICS OF THE U.S. MARKET

Two characteristics of the U.S. health sector exacerbate the problems identifi ed thus far: the prevalence of third-party payments and the fee- for-service system.

THIRD-PARTY PAYMENT�For an ordinary commodity, for which the consumer directly pays the full price, it can be taken for granted that the consumer believes that the benefi ts of the commodity are at least as great as its cost. Health care diff ers from an ordinary commodity in that consumers are insulated from cost considerations at the point of consumption, partly through private insurance and partly through government programs. As we have noted, individuals pay for less than 14 percent of all personal health care costs, an even smaller fraction of doctor costs, and a negligible fraction of hospital costs (around 3 percent). Because so much of medical expenditures are paid for by third parties, consumers have little incentive to be cost conscious. The force of the price system is greatly diminished.

FEE FOR SERVICE The fee-for-service system that is prevalent in the United States may also exacerbate the problems arising from imperfectly informed consumers. The patient goes to the doctor for advice about what medical services are necessary and appropriate. However, the doctor, like any purveyor of services, has a vested interest in selling more services: the more services the doctor provides, the higher his or her income. The threat of malpractice suits (see case study, “Medical Malpractice”) also encour- ages doctors to be very cautious. In addition, patients tend to be risk averse and demand any test or treatment that might protect their health. As a result, there is a tendency for doctors to recommend, and for patients to accept, high levels of service. These problems are exacerbated by the fact that a third party (the insurance company or the government) pays the costs.

HMOs may resolve some of these problems by altering the incentives faced by doctors and patients. Evidence shows that there are important dif- ferences in the care provided and utilized under the fee-for-service and HMO systems. Generally, utilization of health care services is reduced when doctors are compensated with salaries or capitation systems, as is common in HMOs.16

16�L. DeBrock and R. J. Arnould, “Utilization Control in HMOs,” Quarterly Review of Economics and Finance 32, no. 3 (Autumn 1992): 31–53. See also Laurence Baker, “HMOs and Fee-for-Service Health Care Expen- ditures: Evidence from Medicare,” NBER Working Paper 5360, November 1995, p. 27; and David M. Cutler, “A Guide to Health Care Reform,” Journal of Economic Perspectives 8, no. 3 (Summer 1994): 13–29.

372 CHAPTER 13 HEALTH CARE

One of the concerns with HMOs, though, is that their incentives may lead to underprovision of services. This concern has grown as competi- tion among HMOs has intensifi ed and profi t margins have been cut.

THE ROLE OF THE HEALTH INSURANCE INDUSTRY

Individuals are risk averse; that is why they buy insurance. They would rather pay a certain amount every year to the health insurance company than go one year with few expenditures because they are lucky and have no illness or accident, and another year with high expenditures when they are less lucky. This is all the more so because when expenditures are very high, income may be particularly low because those with health problems may not be able to work. Indeed, if, on average, individuals expect to pay $4000 in medical bills, most individuals are willing to pay an insurance company more—a “risk premium”—to avoid the risk. Thus, even if there

MEDICAL MALPRACTICE

C onsumers are frequently disappointed with what they purchase. In the context of health care, people who believe they have received poor medical treatment often fi le malpractice suits against physicians and health care providers.

Mixed trends over the past two decades have drawn considerable attention to this contentious issue: between 1992 and 2001, median trial awards increased 70 percent, but the number of medical malpractice jury trials remained stable; further- more, the cost of medical malpractice insurance rose from $5.4 billion in 1991 to $11.2 billion in 2008, although total losses incurred fell 45 percent from 2003 to 2008 and most medical malpractice claims were closed without any compensation pro- vided to those claiming a medical injury.

Some are concerned that malpractice claims have led to unreasonably high malpractice insurance premiums, which have in turn driven up health care costs and caused doctors to abandon high-risk spe- cialties such as obstetrics. There is also alarm at the

rise of “defensive medicine,” whereby doctors order unnecessary diagnostic tests and prescribe inef- fective medication to avoid claims of medical neg- ligence. Defensive medicine is practiced because a doctor can be sued for failing to administer some test, even if the costs relative to  the information it yields are high, or for failing to prescribe some drug, even if the probability of its having a benefi cial effect is low. Because third parties bear most of the costs of such drugs and tests, doctors and their patients have every incentive to make use of them. The prob- lem is further exacerbated as the defensive practices become widely used, and thus become the standard of practice against which malpractice is judged.

What trade-offs would arise from limiting mal- practice suits? On the one hand, economic effi - ciency would probably be increased through a reduction in needless expenditures. Excessive legal expenditures, which now equal the amount received by plaintiffs in such cases, could be reduced, as might excessive expenditures designed to forestall

373Rationale for a Role of Government in the Health Care Sector

are substantial transactions costs (the costs of fi ling and processing insur- ance claims), insurance may be desirable.

There has been considerable dissatisfaction with the insurance pro- vided by the market. The complaints are that (a) some people buy too much insurance, and insurance induces excessive expenditure on health care; (b) many people cannot obtain insurance (there is too little coverage) or can obtain it only at an excessive cost; (c) transactions costs, including profi ts, are excessive; (d) competition among the health insurance providers is too lim- ited; and (e) insurance fi rms have restricted consumer choice (e.g., in their choice of doctors) and, in their attempt to keep costs down, have squeezed doctors and often denied individuals access to the care that they need and that their doctors think is appropriate. The problems interact: lack of com- petition leads to higher profi ts, and the high profi ts provide the insurance companies with the resources and incentives to lobby Congress against many reforms that would increase the overall effi ciency of the health care system.

malpractice suits. Limiting malpractice liability costs could also reduce physician shortages in high- demand medical specialties that are also considered high-risk from a legal perspective—many physicians are especially unwilling to go into fi elds in which there is a higher chance of a malpractice suit.

On the other hand, a system that limits malprac- tice suits (or the magnitude of rewards) might result in some individuals not being adequately compen- sated for their injuries. The Institute of Medicine esti- mates that between 44,000 and 98,000 people die in hospitals each year as a result of medical errors that could have been prevented, and that prevent- able medical errors in hospitals cost between $17 billion and $29 billion annually as a result of addi- tional care necessitated by the errors, as well as lost patient income and productivity. Limits might also fail to signifi cantly reduce medical expenditures.

There are, however, ways of providing compen- sation for those who suffer as a result of inappro- priate health care (whether an honest mistake or the negligence of a doctor or a hospital) other than through malpractice. New Zealand, for instance, has instituted such a scheme.

The evidence suggests that increases in mal- practice payments are not a major factor driving either increases in insurance premiums or overall health care expenditures, although the threat of malpractice lawsuits has contributed to increased expenditures on diagnostic procedures. Moreover, although many states have amended their laws to discourage liability suits, states with tort reform (as such changes are referred to) do not have lower medical malpractice premiums or defensive medi- cine than states without such reforms.

SOURCES: National Association of Insurance Commissioners, Countrywide Summary of Medical Malpractice Insurance, Calendar Years 1991–2008 and Statistical Compilation of Annual Statement Information for Health Insurance Companies in 2009; T. H. Cohen and K. A. Hughes, Medical Malpractice Insurance Claims in Seven States, 2000–2004, NCJ Report 216339, Bureau of Justice Statistics, Department of Justice, March 2007; Daniel Kessler and Mark McClellan, “Do Doctors Practice Defensive Medicine?” Quarterly Journal of Economics 111, no. 2 (1996); Institute of Medicine, To Err Is Human: Building a Safer Health System, November 1999; U.S. Government Accountability Offi ce, Medical Malpractice Insurance: Multiple Factors Have Contributed to Increased Premium Rates, GAO-03-702, June 2003; and K. Baicker and A. Chandra, “Defensive Medicine and Disappearing Doctors,” Regulation (Fall 2005).

374 CHAPTER 13 HEALTH CARE

INSURANCE AND EXCESSIVE EXPENDITURES ON HEALTH CARE

When individuals buy insurance, they no longer pay the full costs of health care. If illnesses simply happened, unaff ected by the actions of the individual, and if doctors faced no choices in how much to spend to deal with the illness, then there would be no problem. However, individuals can often aff ect their likelihood of needing health care by taking preven- tive actions. (As a result of insurance their incentives to do so are reduced, although this eff ect may not be too signifi cant, as most of the “costs” of an illness are probably the discomfort costs and risks borne by the individ- ual, not the doctors’ and hospitals’ medical bills.)

More important, however, is the fact that there is some discretion in treatment. There may be some benefi t from staying in a hospital an extra day, but the benefi t may be far less than the cost. A very expensive drug may represent a very slight improvement over a much cheaper drug. A patient might not be willing to pay the cost, but would not turn it down if off ered the more expensive drug for nothing, or for pennies.

The demand curve for health services, like the demand curve for any other commodity, is downward sloping—and because it is downward sloping, lowering the marginal cost paid by the patient increases the utilization. With many medical plans, patients pay only 20 percent of the true marginal costs; with some, they pay essentially none of the costs. (Demand is still limited, because there is a time and discomfort cost to spending time in the hospi- tal or going to the doctor.) Figure 13.4 shows a downward-sloping demand curve, and demonstrates how consumption is increased from what it would be if individuals paid the full marginal cost. The magnitude of the increase depends on the elasticity of demand. One estimate puts the demand elasticity for medical care at 0.14, so that with a 20 percent co-payment (the individual pays 20 percent of the marginal cost) demand is increased by 11.2 percent (price is lowered by 80 percent, so demand is increased by 0.14 3 0.80).17

The issue is not so much the increase in the health care expenditures, but the fact that at the margin the social benefi ts of the extra expenditures are less than the costs. Indeed, at the margin, the marginal benefi t (with a 20 percent co-payment) is only equal to 20 percent of the marginal cost.

The fact that, with insurance, individuals spend too much on health care is called the moral hazard problem. The insurance industry has long wor- ried that insurance would increase the likelihood of whatever was insured against; it thought of this as a moral problem—with excessive fi re insurance, say, an owner of a property might even be induced to  burn  down  his  or

17�Joseph P. Newhouse, Free for All? Lessons from the RAND Health Insurance Experiment (Cambridge and London: Harvard University Press, 1993), p. 120.

375Rationale for a Role of Government in the Health Care Sector

her building. Economists, however, view the issue simply as one of incen- tives. With insurance, incentives to maintain health and control health expenditures are attenuated. There is a trade-off : the more insurance, the less risk the individual faces, but the weaker the incentives and thus the greater the overall costs. Optimal insurance balances these trade-off s.18

The signifi cance of the moral hazard eff ect depends, of course, on the particular issue being addressed. Individuals are not likely to chose to have quadruple heart bypass surgery simply because it is less expensive if they have insurance. They will do what the doctor recommends. Further- more, because a very large fraction of overall medical expenditures is for such “catastrophic” events, including the treatment of cancer, accidents, heart attacks, and a variety of diseases of old age, many question the rele- vance of moral hazard for these expenditures.

There is a concern that government policies have led to excessive insur- ance, and because insurance is excessive, expenditures on health care are excessive. The tax system subsidizes insurance. Employer-provided insur- ance premiums are essentially a tax-free form of compensation. The marginal cost of providing health care through employer-provided insurance is thus markedly lower than if the individual buys the health care services directly (or buys insurance directly). These tendencies for excessive insurance were

18�That is, with optimal insurance, individuals will spend more on health care than they would with no insurance, but this is the “price” that must be borne to reduce the risks they face.

INSURANCE AND THE UTILIZATION OF HEALTH CARE SERVICES

Insurance lowers the price paid by the individual and thus increases the quantity of health care services consumed, from Q0 to Q1. For these additional expenditures, marginal costs (refl ected in the market price, p) exceed the marginal benefi t to the individual (refl ected in will- ingness to pay, as evidenced by the individual’s demand curve). The area ABC measures the deadweight loss from excessive consumption.

FIGURE 13.4

Deadweight loss

II

Marginal cost

80% of cost paid

by insurance

20% paid by individual

Demand for medical services

Quantity of medical service

Price of medical service

BA

EC p

Q0 Q1

I

376 CHAPTER 13 HEALTH CARE

exacerbated by the way that employers used to allow employees to choose among insurance plans. The employer would pay a fraction of the additional costs of the more expensive plan (which traditionally provided more ser- vices). Thus, although the marginal benefi t of the extra insurance to the indi- vidual could be less than the marginal cost, the individual would still choose the more expensive plans. As many fi rms have worried about the soaring costs of their health programs, they have required individuals to pay the full marginal costs if they choose a plan costing more (and typically providing more extensive coverage) than the basic plan off ered by the fi rm.

ADVERSE SELECTION Another problem that arises in many insurance markets is adverse selection. Consider a simple situation in which the insur- ance fi rm could tell nothing about an individual, other than that he or she was willing to purchase an insurance policy at the premium being off ered. At higher premiums, those who are least likely to need medical care—say, healthy young individuals—decide it is not worth paying that premium, or they may decide to buy a policy that covers only very large medical expenses. Thus, as premiums increase, there is an adverse selection eff ect: the best risks decide not to purchase the policy. This is illustrated in Figure 13.5A. With the best risks dropping out of the market, however, the average cost per policy issued increases. Figure 13.5B illustrates the competitive market equi- librium, which occurs when the premium equals the expected payout per policy. Because there are few highly risk-averse individuals who are willing to pay a large amount in excess of their expected costs, a small fraction of the population obtains insurance. These individuals pay a high price for not hav- ing insurance. Individuals buy insurance because they are risk averse; that is, they do not like the volatility of what they would have left over to spend on other things if they faced a big health care bill. The result of a market econ- omy in which individuals feel that premiums are too high to buy insurance is that these individuals must bear the costs of the risks themselves.

Insurance fi rms do not, of course, sit idly by, accepting anyone who applies. They actively attempt to increase the “quality” of those they insure. They do this in several ways, each of which has adverse eff ects on the eff ectiveness of insurance. Two of the most common tactics are now banned by PPACA. First, they often did not provide insurance for “pre-existing conditions”—for instance, illnesses whose onset occurred before the purchase of insurance. In some cases, insurance fi rms simply refused to insure those with a bad health history. These provisions had a particularly adverse eff ect on the ability of indi- viduals to switch jobs. If anyone in the worker’s family had an illness, the worker could become locked into his or her current job, because if the worker moved, the insurance at the new company often did not cover the medical costs of that illness. Second, insurers imposed limits on the extent of coverage. Although insurance is supposed to cover big losses—those losses which individuals are

377Rationale for a Role of Government in the Health Care Sector

least able to bear—in fact, insurance markets simply failed to cover them. Even insurance designed to cover major medical losses typically put caps on claims— for instance, at $1 million. However, even under PPACA, insurers may concen- trate their selling eff orts in low-risk communities. Health status varies greatly across diff erent communities; the incidence of malnutrition, violence, drugs, and alcoholism—all of which can lead to high medical costs—is much higher among some socioeconomic groups than others.

ADVERSE SELECTION

(A) As the premium rises, the percentage of the population purchasing insurance decreases, with the best risks—those with the least chance of needing insurance—dropping out of the market fi rst. As a result, the average payout increases as the premium increases. (B) Shows the market equilibrium, at which the premium equals the average payout. In equilibrium, either relatively few or relatively many individuals may remain uncov- ered. (C) Shows that indeed there may be multiple equilibria: in the high-price equilibrium, relatively few people are cov- ered; in the low price, most people are covered.

FIGURE 13.5A

B

C

Premium

% Population that buys policy

Premium

Premium

Average cost per policy

Average cost per policy

378 CHAPTER 13 HEALTH CARE

These attempts by insurance companies to improve the mix of those covered are sometimes referred to as cherry picking or cream skimming. The important point is that insurance fi rms may be able to increase their profi ts more by engaging in such activities than by increasing the effi - ciency with which they provide insurance services. Although the drive to increase profi ts by reducing costs is clearly socially benefi cial, the drive to increase profi ts by cream skimming is more problematic: the gains to one fi rm occur largely at the expense of other fi rms, which fi nd that they have a worse mix of insured individuals.19

Concern about the lack of coverage for certain high-risk groups in the population—in particular, the elderly—has formed one of the strongest motivations for the expansion of the government’s role in health care.

The costs that insurance companies bear to try to make sure that they insure the best risks contributes to the ineffi ciency of America’s private health insurance industry. Moreover, there is a moral issue: Should someone who has the misfortune of having been born with a heart condition, which will shorten his or her life, also be asked to pay more for insurance? Most other advanced industrial countries have answered that question in the negative, and have provided some system of universal health insurance, sometimes run by the government, sometimes run by private or not-for-profi t entities.

TRANSACTIONS COSTS The costs of buying and selling goods, includ- ing the cost of running markets, are referred to as transactions costs. These are the costs associated with making the economic system work.

The costs of running the insurance system appear to be high. In a competitive world without transactions costs, all the money going into an insurance company in the form of premiums would be paid out in the form of benefi ts. However, there are costs of selling insurance policies and paying the benefi ts. For instance, insurance fi rms spend large amounts in an eff ort to identify good and bad risks. These transactions costs are relatively low for policies provided through large fi rms—administrative expenses amount to about 5.5 percent of the claims paid—but small fi rms face a much heavier burden, with administrative expenses amounting to 40 percent of the claims paid. Private health insurance companies spent an average of 12 percent on administrative expenses in 2010.20

These direct transactions costs, however, are not the entire transac- tions costs of the system. Doctors and hospitals must fi ll out vast numbers

19�The issue is somewhat more complicated because there can be effi ciency gains from sorting individ- uals according to their risks. For instance, it may be useful to ascertain that some individuals are bad drivers in order to discourage them from driving. Charging higher premiums for health insurance to those who smoke or drink may similarly have benefi cial incentive eff ects. However, this is not the case for most of the sorting that occurs in health insurance markets. 20� Centers for Medicare and Medicaid Services, Offi ce of the Actuary, National Health Statistics Group, “National Health Expenditure Accounts, 1960–2010.”

379Rationale for a Role of Government in the Health Care Sector

of forms. By one estimate, 20 percent of expendi- tures on hospitals actually goes to administrative costs. Still more costs are borne by the insured as they try to collect on their insurance. Critics of the current system argue that standardization of forms and other reforms of the health care sys- tem could substantially reduce those costs.

Even with standardization, however, trans- actions costs would likely be substantial. Insur- ance fi rms need to monitor doctors and hospitals to make sure that claims are legitimate. Critics of government insurance programs argue that in spite of their high transactions costs, they have done an ineff ective job of monitoring. The estimated cost of fraud and abuse in public and private health care spending is more than $100 billion annually.21

Because of the incentive eff ects and large transactions costs typically associated with insur- ance, most economists believe that insurance markets should focus on large losses, those that the individual would fi nd diffi cult to bear. Unfortunately, many individuals feel that they are not getting their “money’s worth” from insurance unless they regularly get at least something back; insurance com- panies have thus found that customers are attracted to covering small losses, such as, in automobile insurance, towing costs.

Modern behavioral economics has helped explain why individuals act in such seemingly irrational ways—and how insurance companies have learned how to exploit these irrationalities to induce individuals to buy insurance products for low-probability events at prices well above those based on actuarial odds.

CONSEQUENCES OF INEFFICIENCIES IN HEALTH CARE MARKETS

Failures in health care markets lead to ineffi ciencies and may contrib- ute to the rising costs that have fueled recent public policy debates. Two important examples of such ineffi ciencies are the excessive provision of health care services and the provision of inappropriate care.

21�For a discussion of these estimates, see M. K. Sparrow, Testimony at “Criminal Prosecution as Deter- rent to Health Care Fraud” before Senate Committee on Judiciary, Subcomittee on Crime and Drugs, May 20, 2009; and W. J. Rudman et al., “Healthcare Fraud and Abuse,” Perspectives in Health Informa- tion Management 6 (Fall 2009).

PROBLEMS WITH MARKET- PROVIDED HEALTH INSURANCE • Moral hazard—reduced incentive to avoid insured-

against event.

Trade-off between risk and incentives. Insurance fi rms try to limit through use of

co-payments and deductibles.

• Adverse selection—those who choose a particular policy may have risk characteristics that differ from those of the population as a whole.

As price of insurance increases, best risks may decide not to apply.

Insurance fi rms have incentive to try to insure only low-risk individuals—“cherry picking” or “cream skimming.”

• Transactions costs.

380 CHAPTER 13 HEALTH CARE

SUPPLY CREATES ITS OWN DEMAND Conventional competition theory predicts that an increase in the supply of doctors will lower the price of medical services. However, during the 1970s, when the number of doctors almost doubled—expansion far greater than the increase in popu- lation or income—prices for medical services did not fall. One explanation for this was discussed earlier: patients judge quality partly on the basis of price because of their limited information, so doctors are reluctant to cut prices, lest they get a reputation as being second-rate.

For a conventional commodity, if supply increases and price does not fall, there will be excess supply. There is some evidence of this occurring in medical practice; one detailed study in a suburban New York community showed surgeons having a workload only one-third of what experts viewed as a full schedule. These results were consistent with similar fi ndings based on aggregate statistics of the number of surgeons and operations.

In health care, though, there is another possible response: doctors may increase the demand for their services, because patients do not know what care is necessary and appropriate and pay little of the cost of their care. Supply creates its own demand. There is some evidence, in fact, that increasing the number of surgeons leads to an increase in the number of operations even when prices do not change at all.22

Indeed, there is even evidence of what might be viewed as a back- ward-bending supply curve. In Canada, as the government has attempted to ratchet health care costs down, it has reduced the fees paid to doctors for each service performed, to which doctors responded by performing more services. (To some extent, the doctor may not actually be perform- ing more services, but rather will be billing more of the services actually performed. Thus, a doctor who might have treated a patient for two mal- adies at the same time, and billed for one offi ce visit, might ask the patient to come back to take a more careful look at the second malady.)

Questions of whether an operation is advisable or necessary are, of course, debatable. Most doctors do not recommend operations simply to increase their own income. In making a professional judgment about whether an operation is desirable, however, time constraints and the demands of other patients have important eff ects—and patients with lim- ited information are likely to have surgery if their surgeon recommends it.23

22�See V. Fuchs, Who Shall Live?; W. McClure, “Buying Right: The Consequences of Glut,” Business and Health (September 1985): 43–46; V. R. Fuchs, “The Supply of Surgeons and the Demand for Operations,” in The Health Economy (Cambridge, MA: Harvard University Press, 1986), pp. 126–147; and M. Gaynor and R. Town, “Competition in Health Care Markets,” Chapter 9 in Handbook of Health Economics, vol. 2, ed. M. Pauly, T. McGuire, and P. Barros (Waltham, MA: North Holland, 2012), pp. 499–638. 23�Although seeking second opinions before undergoing surgery may improve matters, it is far from a solution. In a community in which there is an increased supply of physicians, diagnostic criteria may change, so a second opinion would merely refl ect the same general inclination to recommend surgery more frequently.

381Rationale for a Role of Government in the Health Care Sector

INAPPROPRIATE CARE There is considerable evidence that much of medical care expenditures may, in fact, be inappropriate. Perhaps the most telling are the data comparing practices in diff erent locations. For instance, the average length of a community hospital stay in South Dakota is more than twice that in Oregon. This large diff erence cannot be accounted for by diff erences in the populations.24 Huge practice diff erences occur across states as well as countries. For example, although hospital stays are rela- tively infrequent and short in the United States when compared with those in other OECD countries, hospital spending per discharge is much greater than in other OECD countries, at almost triple the OECD median.25

These variations lead to huge cost diff erences. Total personal health care spending per capita in Massachusetts is almost double that in Utah.26 There is also a wide variation in long-term cost trends. For example, in 1992, Boston, San Francisco, and eastern Long Island (New York) had almost identical per capita Medicare spending, but their expenditures grew at very diff erent annual rates over the next fourteen years: 2.4, 3.0, and 4.0 percent, respectively. By 2006, per capita spending in east- ern Long Island was $2500 more than in San Francisco, equal to about $1 billion in additional annual Medicare spending. It is generally agreed that diff erences in age and other demographic characteristics of the pop- ulation fail to explain most of the variation in practices.27

One of the objectives of the PPACA was to improve the overall per- formance of the health care system by systematically studying these diff erences in practices, and ensuring that the best practices are used everywhere.

POVERTY, INCOMPLETE COVERAGE, AND THE ROLE OF GOVERNMENT

Most of our discussion so far has focused on market failures, such as inef- fi ciencies in the market arising from imperfections of information and competition. One of the principal reasons for government action in health care, however, has nothing to do with effi ciency; even if markets were perfectly effi cient, there would be a concern that those who are very poor do not receive adequate health care.

24�National Center for Health Statistics, Centers for Disease Control and Prevention, U.S. Department of Health and Human Services, Summary Health Statistics for the U.S. Population: National Health Inter- view Survey, 2011, December 2012, Table 3.2. 25�D. Squires, “The U.S. Health System in Perspective: A Comparison of Twelve Industrialized Nations,” The Commonwealth Fund Pub. 1532, Issues in International Health Policy 16 (July 2011). 26 S. Cliff , “Our Health Care Spending, In One Map,” Washington Post Wonkblog, December 8, 2011. 27 E. Fisher et al., “Slowing the Growth of Health Care Costs—Lessons from Regional Variation,” New England Journal of Medicine 360 (February 2009): 849–852.

382 CHAPTER 13 HEALTH CARE

Many believe that no individual, regardless of income, should be denied access to adequate medical care. If choices must be made, they should be made on the basis not of wealth but of other attributes, such as age, the likelihood of success of a procedure, or, perhaps, random selection. This view holds that medical services are diff erent from clothes, movies, auto- mobiles, and most other commodities. Just as the right to vote should not be subject to the marketplace (individuals are not allowed to buy and sell their votes), and just as when there was a draft, individuals were not allowed to buy their way out of their military obligations (except during the Civil War), the right to live—access to medical services—should not be controlled by the market. The view that there are goods and services like health care, whose availability to diff erent individuals should not depend solely on their income is known as specifi c egalitarianism.28

Not all economists agree about whether medical services should be treated diff erently from other commodities. Many hold that they should not: those who have more money and want to spend it on getting health care should be allowed to do so. Those who hold this view often point out that the relationship between medical care and life (death) is very weak; other factors, such as smoking, drinking, and eating patterns (consump- tion of vegetables and fruits, for example) have greater eff ects on longev- ity. If one wanted to improve the health status of the poor, one could do so in a more cost-eff ective way by waging campaigns against smoking and drinking, raising taxes on tobacco and alcohol, and encouraging the con- sumption of fruits and vegetables.

Still a third view—toward which many Western democracies seem to be gravitating—is that everyone should have the right to a certain mini- mal level of care. The provision of Medicaid can be thought of as refl ect- ing that view.

REFORMING HEALTH CARE

Recent eff orts in health care reform have targeted the big issues identifi ed at this chapter’s outset: high costs, limited health insurance coverage, and growing fi scal strains. This section describes some of the most import- ant attempts to counter these problems, including the PPACA enacted in 2010, usually referred to as the Aff ordable Care Act, or just Obamacare (see case study, “Comprehensive Health Care Reform”).

28 J. Tobin, “On Limiting the Domain of Inequality,” Journal of Law and Economics 13 (1970): 263–277.

383Reforming Health Care

COST CONTAINMENT

Economists look to improved incentives as the primary way in which costs can be contained.

ELIMINATING TAX DISTORTIONS Eliminating the tax incentives associated with employer-provided insurance has met with strong resis- tance from labor unions, largely because they have made more extensive health care a major thrust of labor negotiations over the past several decades, accepting better health benefi ts in lieu of higher wages. (Even though total compensation, including health benefi ts, has increased roughly in line with productivity, real wages have increased more slowly— in some cases, actually falling, especially for unskilled workers.) There is also worry that without such tax incentives, the number of uninsured would increase, exacerbating other problems in the health care system. A compromise proposal is to cap the tax deductibility, so that at the mar- gin, individuals and fi rms face the correct incentives, and to make it a condition for eligibility for tax deduction that if employees are given a choice of plans, they must face the full marginal costs of the more expen- sive plans. Some go further and argue that there should be a tax on “Cadil- lac plans,” because such plans raise excessively the usage of health care. However, most economists would argue that if individuals choose to buy such plans (or their unions choose to give up wage increases to get them) they should be allowed to do so. Of course, the government should not subsidize such plans through preferential tax treatment.

MORE EXTENSIVE USE OF MANAGED CARE As we noted, man- aged care has grown extensively, to the point at which the vast majority of those with employer-provided coverage are in HMOs. Ironically, the debate about managed care has shifted. Originally, there was concern about whether it would simply reduce costs by, say, 15 percent, leaving the rate of increase unchanged, or whether it could reduce the rate of increase. The experience in the 1990s, though it left this longer-range question unsettled, demonstrated clearly that HMOs could reduce costs markedly.29 As they reduced costs, however, concerns were expressed about whether they had too strong incentives for cost containment, unbal- anced by commitment to the quality of care. The issue became cystallized in a debate over “drive-through deliveries,” in which HMOs allowed new mothers only a 24-hour hospital stay after the delivery of a baby. Other practices, such as a reluctance to pay for expensive new experimental

29�Critics noted that some of the cost savings came from the fact that HMOs that owned their own hos- pitals managed to reduce the amount of uncompensated care they provided.

384 CHAPTER 13 HEALTH CARE

COMPREHENSIVE HEALTH CARE REFORM

SOURCES: Offi ce of the Legislative Counsel, U.S. House of Representatives, Compilation of Patient Protection and Affordable Care Act, May 2010; H.  Chaikind, C. W. Copeland, C. S. Redhead, and J. Staman, PPACA: A Brief Overview of the Law, Implementation, and Legal Challenges, Congressional Research Service, March 2011; and Henry J. Kaiser Family Foundation, Timeline: History of Health Reform in the U.S., www.kff.org.

Since the adoption in 1965 of Medicare and Medicaid as part of President Johnson’s “Great Society,” all attempts at compre- hensive reform of the health care sector over the subsequent 45 years had been abject failures until enactment of the Patient Protection and Affordable Care Act (PPACA) in March 2010, often referred to as “Obamacare.”

PPACA is an extraordinarily complex, 974-page law to be phased in over several years. The primary objectives of PPACA are to broaden the cover- age and reduce the cost of health insurance and improve access to health care more generally.

Coverage and access is increased by comple- menting current employer-provided insurance with an expansion of Medicaid eligibility and extension of fund- ing for the Children’s Health Insurance Program (CHIP), subsidized private insurance premiums for qualifying low-income individuals, and creation of individual and small-business health insurance exchanges. In addition, health plans will not be allowed to deny coverage for most pre-existing conditions regardless of age, cancel coverage because of changes in health status, or place annual and lifetime limits on essential health benefi ts.

PPACA costs are to be offset by a combination of increased revenue (taxes, fees, and penalties) and reforms designed to improve the effi ciency and effectiveness of the health care system, including measures to combat fraud and abuse in the Medi- care and Medicaid programs; administrative simpli- fi cation and creation of uniform electronic standards and operating rules to promote the spread of elec- tronic health records; more coordinated care for patients with chronic conditions; creation of the Patient-Centered Outcomes Research Institute to

provide physicians and patients with information on the effectiveness of medical technologies and inter- ventions; creation of a complementary Innovation Center in the Centers for Medicare and Medicaid Services to develop, test, and evaluate new policies and programs to enhance the quality of care; and establishment of the Independent Payment Advi- sory Board, a panel of medical experts tasked with improving Medicare’s payment system.

The Congressional Budget Offi ce and the Joint Committee on Taxation estimate that PPACA will reduce federal defi cits by $143 billion over a ten- year period (2010–2019), and by 2019, will result in 94 percent of the non-elderly, legally present U.S. population having some form of coverage. How- ever, passage of the PPACA was extremely conten- tious, with all Republicans voting against it in both houses of Congress. The legislation then faced many legal challenges, culminating in the June 2012 Supreme Court decision that affi rmed the author- ity of Congress, under its taxing power, to require, beginning in 2014, that most individuals have health insurance or pay a penalty (known as the individual mandate). However, the Supreme Court also ruled that in the complicated domain of federal authority versus states’ rights in the health sector, the provi- sion requiring states to comply with new eligibility criteria for Medicaid or risk losing their funding was constitutional, only as long as the states lose just the new funds for noncompliance, rather than los- ing all their funding. The result is that some states, including populous states with a signifi cant percent- age of low-income residents and limited Medicaid coverage, could continue to have large numbers of uninsured people.

385Reforming Health Care

procedures, also drew widespread media attention. Drive-through deliv- eries were banned, and many states proposed a stronger role of govern- ment in regulating HMOs.

HEALTH SAVINGS ACCOUNTS Legislation passed in 2003 estab- lished health savings accounts (HSAs).30 These are intended to encour- age individuals to buy insurance policies with larger deductibles and co-payments, and thus contain fewer incentives for moral hazard—they off er tax-preferred treatment of money saved for medical expense for individuals covered by high-deductible health plans. Funds allocated to these accounts are tax deductible, but annual contributions are limited and excess contributions are not tax deductible. HSAs eliminate one of the distortions of the existing system under which health care paid for through insurance companies is, in eff ect, tax deductible, but direct pay- ments are not. Both HSAs and their predecessor, medical savings accounts (MSAs), have been very popular: in 2011, deductions for HSAs and MSAs cost an estimated $1.9 billion in tax expenditures.31

One criticism of HSAs relates to the relative strength of the incen- tive eff ects for reducing moral hazard, compared to the adverse selection eff ects. Clearly, those who think it unlikely that they will need much med- ical care, and rich individuals better able to bear the risk of paying a large deductible or co-payment and for whom the tax break is particularly sig- nifi cant, will be much more attracted to the HSAs. If healthier individuals are attracted to the HSAs, then those not having HSAs will face higher insurance premiums.

Thus, for some individuals, HSAs might prove very attractive; their overall medical costs—insurance plus out-of-pocket expenses—might well go down. The unanswered question is, to what extent are those lower costs a refl ection of greater effi ciency (a reduction in the distortions arising from moral hazard) and to what extent are they a refl ection of cream skimming?

EXTENDING INSURANCE COVERAGE

As we noted at the beginning of this chapter, many people in the United States have no health insurance. A number of reform eff orts have attempted to address the problem of the uninsured.

30 HSAs were preceded by medical savings accounts (MSAs), established on a pilot basis for self-employed individuals and employees of qualifying small business employers as part of the Health Insurance Portability and Accountability Act of 1996. The Treasury Department extended the MSA pilot program periodically through 2007, after which no new MSAs could be opened because they had been superseded by HSAs, included as part of the Medicare Prescripton Drug, Improvement, and Modernization Act of 2003. See Internal Revenue Service, Department of the Treasury, Health Savings Accounts and Other Tax-Favored Health Plans, Publication 969. 31�Offi ce of Management and Budget, Analytical Perspectives, Fiscal Year 2013 Budget, Table 17-1.

386 CHAPTER 13 HEALTH CARE

MANDATED HEALTH INSURANCE Extending health insurance is central to the PPACA. Today, most employers provide health insurance to their employees; beginning in 2014, the PPACA establishes an individual mandate, under which most individuals will be required to have health insurance or pay a penalty.

Many advocates of universal health insurance favor a single payer system like Canada’s, which works much like Medicare, with the gov- ernment paying medical bills. However, fi nancing such a system would require new taxes—a political impossibility at present. Instead, the individual mandate constitutes a hidden tax, as ultimately the mandate imposes insurance costs on fi rms and consumers.

PRE-EXISTING CONDITIONS As noted earlier, many insurance fi rms have refused to cover health conditions existing prior to the beginning of the insurance coverage. Thus, a worker with a kidney problem would fi nd it impossible to change jobs because the new insurance policy would not cover the costs of continued treatment. Legislation passed in August 1996—called the Kennedy–Kassebaum bill after the senators who wrote it—was devised to prevent insurers from excluding pre-existing condi- tions from coverage. A number of questions were raised with the passage of the legislation: How much would insurance fi rms have to raise pre- miums to extend this coverage? Would employers, concerned that hiring a worker with a pre-existing condition would lead to soaring premiums, discriminate against job candidates they suspected of having such con- ditions? So far, it appears that the bill’s impact on either the insurance or labor market has been limited. Thus, the PPACA strengthens these provi- sions by prohibiting exclusions for pre-existing conditions.

FILLING IN THE GAPS When the PPACA was enacted in 2010, approx- imately 50 million Americans did not have health insurance. Roughly 30 million people are expected to gain coverage under the new law, and half of these will gain coverage through an expansion of Medicaid. (Begin- ning in 2014, state Medicaid programs must provide coverage to all eli- gible non-pregnant, non-elderly legal residents with incomes up to 133 percent of the federal poverty level.) Other signifi cant gaps to be fi lled under PPACA are health insurance for employees of small businesses, via establishment of federally subsidized health insurance exchanges; health insurance for retirees not yet eligible for Medicare benefi ts (ages 55 to 64), through the Early Retiree Reinsurance Program; and health insurance for young adults, by extending coverage under their parents’ plan until the age of 26. Other gaps had been fi lled by previous legislation, such as health insurance for low-income children who are ineligible for Medicaid

387Reforming Health Care

but cannot aff ord private insurance�32 and for employees who were fi red, changed jobs, or got divorced.33

MEDICARE REFORM: EASING LONG-TERM FISCAL STRAINS

The long-run fi scal strains arising from the health care system are largely the result of growing Medicare expenditures. Eff orts to reform Medicare focus on three issues: improved incentives, better management of health care provision, and more competition.

IMPROVED INCENTIVES The Medicare program itself has large co-payments and deductibles, but most elderly buy, at quite moderate costs, Medigap insurance—that is, insurance that covers most of these out-of-pocket costs. For individuals with Medigap insurance, there are few incentives to economize on the use of medical care. Those who buy Medigap insurance do not pay the full incremental cost associated with the policy, and because of this reduced incentives to economize on health care, those with Medigap insurance make more demands on the Medi- care system itself. Medigap premiums cover the direct costs incurred by the insurance companies, but not the additional costs imposed on the government because of increased utilization. One set of reform proposals thus would either restrict the coverage that can be provided by such poli- cies or force the policies to pay the full incremental costs.

One of the fastest-growing parts of the Medicare system is home health care. Many individuals do not need to stay in hospitals or long-term care facilities, as long as they have some limited assistance at home. Home health care provides that assistance. However, home health care has been subject to extensive abuse. Unlike doctors, nurses, and other traditional health care providers, who must pass stringent standards, relatively few skills are required of home health care providers, and entry into the business appears easy. Furthermore, although the demand for heart surgery may be rela- tively insensitive to price (and would be limited even at a zero price), almost anyone off ered free assistance would fi nd it useful. Ironically, however, although the Medicare system has deductibles and co-payments associated with most medical treatments, it has not imposed any co-payments on home

32 The Children’s Health Insurance Program (CHIP) was created by the Balanced Budget Act of 1997 as Title XXI of the Social Security Act and extended under the CHIP Reauthorization Act of 2009. 33�Health benefi t provisions in the Consolidated Omnibus Budget Reconciliation Act (COBRA) of 1986 require most employer group health plans to provide temporary continuation of group health cover- age that otherwise might be terminated, and the Health Insurance Portability and Accountability Act (HIPAA) of 1996 allows the right to enroll in another group health coverage program without waiting until the next open season for enrollment.

388 CHAPTER 13 HEALTH CARE

health services since 1972. In 2011, the Medicare Payment Advisory Com- mission (MEDPAC)—an independent congressional agency established by the Balanced Budget Act of 1997 to advise Congress on Medicare fi nancing and related issues—recommended co-payments for home health services, but there has been strong opposition because of claims that it is ineffi cient and regressive. Critics say co-payments would discourage home health care in favor of more expensive nursing facilities, and would fall most heavily on the poorest and oldest Medicare benefi ciaries.

IMPROVED MANAGEMENT As costs soared soon after the Medi- care program was begun, attention focused on providing better incen- tives to the providers. Initially, Medicare had simply reimbursed hospitals for whatever costs they incurred. Such a system clearly pro- vides perverse incentives. In 1983, Medicare switched to a system in which hospitals were reimbursed a fi xed amount for the treatment of a particular illness.

Overall, the system called diagnosis-related groups (DRGs) has worked well. Initially, there was concern that hospitals might refuse to provide treatment in areas where the reimbursement rate was too low, or that they would be tempted to cheat on the diagnoses or to cut corners on the services provided, adversely aff ecting the quality of the care. (Of course, the intent of the system was to cut unnecessary costs, such as hos- pital stays that were longer than needed. The problem is that it is often diffi cult to distinguish between cutting corners and cutting unnecessary expenditures.) Neither of these fears has come true.

One ongoing concern, however, is the ballooning of out-of-hospital costs. Hospitals have tried to push patients out of the hospital as quickly as possible. Thus, although hospital costs have been contained, overall health costs have not performed as well. This has prompted some to sug- gest an extension of the system, called bundling, to include all the treat- ment associated with an illness, both in hospital and out of hospital.

Conservatives have long argued that private insurance companies could manage Medicare better than the government. In fact, Medicare benefi ciaries have had the option to receive their benefi ts through private health plans since the 1970s, mainly through HMOs, as an alternative to the traditional federally administered Medicare program. However, this option was greatly expanded under the Bush administration through cre- ation of Medicare Advantage: by 2012, 27 percent of Medicare benefi cia- ries were enrolled in a Medicare Advantage plan. It turned out, though, that to provide services comparable to those provided by the traditional Medicare program actually increased costs, by as much as 20 percent. In response, the PPACA mandates the gradual reduction of federal payments

389Reforming Health Care

to Medicare Advantage plans to bring them closer to the average costs of care under the traditional Medicare program. It also provides for bonus payments to Medicare Advantage plans based on quality ratings, and restricts the share of premiums that these plans can use for administra- tive expenses and generation of profi ts.34

STRENGTHENED COMPETITION Although there are many doctors and hospitals in the country as a whole, competition within a given locale may be very limited. Providers have sought immunity from (or at least a loosening of) antitrust provisions, which would reduce competition even further. So far, these pressures have been resisted. Such immunity might, for instance, allow all the X-ray labs in an area to get together to set prices, or a single health provider—HMO or preferred provider organization (PPO)—to sign up all the bone specialists in a community.

Some have argued that the government should strengthen competitive forces by having competitive bidding for the supply of Medicare services. The winning bidder would, in eff ect, become a preferred provider, with the government paying 100 percent of its bid. Other providers might be reimbursed a fi xed amount up to a fi xed percentage of the winning bid. Critics of such plans worry about ensuring quality and about the devel- opment of a two-tier system, in which poor individuals would go to the preferred provider, while wealthier individuals would go to higher-priced providers.

The measures just discussed are not the only cost saving measures that have been proposed. For instance, there was a provision in the law establishing the prescription drug benefi t under Medicare (Medicare Part B) restricting the ability of the government to negotiate with the drug companies, a provision that was estimated to increased govern- ment spending over ten years by as much as half a trillion dollars. Repeal of this provision and the adoption of standards to ensure cost-eff ective selection of drugs would accordingly save a considerable amount of money.

FINANCIAL REFORMS The reforms just described are directed at improving the efficiency with which Medicare services are provided. There is another set of reforms focused more narrowly on lowering the financial drain on the government, either by reducing benefits—such as by increasing the age of eligibility from 65 to, say, 67, just as the normal age of retirement has been increasing under Social Security— or by imposing more of the costs on recipients. Medicare subsidies are, for instance, extended to all participants, regardless of their

34�The Henry J. Kaiser Family Foundation, Medicare Advantage Fact Sheet, November 30, 2012.

390 CHAPTER 13 HEALTH CARE

income, which means that a millionaire retiree is being subsidized by a $20,000-a-year worker. Pressure to increase payments, at least for higher-income beneficiaries, has been mounting in recent years, par- ticularly as some recent studies suggest that the Medicare program overall is regressive. With taxes proportional to income, and all indi- viduals having the same entitlement, it had been thought that Medi- care was highly progressive. But when account is taken of the fact that richer individuals tend to live longer and are better capable of extract- ing services out of the Medicare system, it turns out that the rich get more, so much more that it more than offsets their higher Medicare tax contributions.

REFORMING MEDICAID

Medicaid provides medical assistance to the poor and the disabled and long-term nursing care to the aged. Each of these comprises about a third of the program. The long-term nursing program has been a particular source of concern. Although Medicaid began as a program for the des- titute, today about one-third of all nursing home and home health care expenditures are paid by Medicaid. Some elderly individuals transfer wealth to their children so they can qualify to receive these benefi ts. Although the government has attempted to limit these transfers (typi- cally, funds transferred within three years of entering a nursing home are treated as if they were available to pay for nursing home care), it has had only limited success. Nursing home expenditures are expected to soar as the number of very old Americans—those over 85—increases. Further- more, whereas just 5 percent of the general population over 65 is in nurs- ing homes, a quarter of those over 85 are, and half of the nursing home population is 85 or older.

The private insurance market for long-term nursing care was slow to develop. In 2010, private insurance paid only 9 percent of total expen- ditures for nursing homes and continuing care retirement communities (barely higher than the 7 percent fi gure for 1994).

There have been three major groups of proposals to reduce government costs, all based on the premise that most individuals should make provi- sion for themselves, but if they do not, it is diffi cult for the government not to provide assistance. One proposal is to mandate insurance: every individual would have to purchase, either privately or through the gov- ernment, long-term nursing care insurance. Providing it through the gov- ernment would entail eff ectively expanding the Social Security program. To fi nance long-term nursing care, the government would require an

391Review and Practice

increase in the Social Security tax of around 1.0 to 1.5 percentage points. The second proposal is to increase incentives—both carrots and sticks— to purchase insurance by providing tax preferences and asset protection (under “asset protection,” individuals who purchased a minimal amount of nursing home insurance could keep a certain amount of assets to pass on to their heirs or for other expenses without losing eligibility for Med- icaid when their insurance ran out), and by reducing the magnitude of public support for nursing care. The third proposal is to assist the pri- vate sector in providing meaningful long-term nursing care insurance by providing reinsurance against large losses (associated with long stays) or against hard-to-anticipate long-term trends, either in longevity or overall nursing home care expenses.

SUMMARY

1. Although decisions about health are diffi cult, resource allocations—choices among alternative uses of funds—must be made. Economic analysis may be useful in making those decisions in a sys- tematic and consistent way.

2. U.S. health expenditures are a third higher than the OECD average as a share of GDP, and are the highest among OECD countries as a share of total government expenditures—the government now pays almost half of all personal health care expen- ditures, whereas out-of-pocket expenditures as a percentage of total health expenditures in the United States is among the lowest of the OECD countries. In spite of the high level of health expenditures, health outcomes are poorer than in other OECD countries. The four major pub- lic programs in the United States are Medicare, Medicaid, health care for veterans, and public support for research and development. In addi- tion, there are two major categories of tax expen- ditures: employer-fi nanced health insurance and tax deductibility of medical expenses exceeding a certain level.

3. Three fundamental problems facing the health care system in the past couple of decades before the passage of PPACA have been:

a. A large and growing population of uninsured

b. High and rising health care costs

c. Growing federal health care expenditures

4. Among the reasons why markets may not provide an effi cient allocation of resources in the health care sector are:

a. Consumers and insurers have limited information.

b. Competition is limited.

c. Only a small fraction of health care costs are paid directly by the consumer.

5. Health insurance is associated with the prob- lems of adverse selection and moral hazard. Insurance reduces individuals’ incentives to take care of themselves and to economize on health care services, and insurance fi rms may increase their profi ts more by ensuring that they cover only the most healthy than by increasing their effi ciency.

REVIEW AND PRACTICE

392 CHAPTER 13 HEALTH CARE

6. The tax system encourages excessive purchase of insurance and excessive consumption of health care services. Tax expenditures for health amount to billions of dollars. Most economists believe that these tax benefi ts should be capped, so that individuals will more nearly pay the full marginal costs of services.

7. Concern that the fee-for-service system leads to excessive consumption of health care has led to rapid growth of health maintenance organiza- tions. More recently, there has been a concern that HMOs have been too zealous in controlling costs, sacrifi cing quality and service.

8. Prior to the enactment of comprehensive health care reform in 2010, including an individual mandate for medical insurance, attention was focused on incremental reforms, such as extend- ing insurance coverage to the unemployed and to children, eliminating restricted coverage of pre-existing conditions, and providing medical savings accounts to encourage individuals to purchase policies with larger deductibles and co-payments.

9. Proposals to reform Medicare focus on improv- ing incentives, introducing better management of health care providers, ensuring more eff ective competition, and placing more of the fi nancial burden on benefi ciaries.

10. The major concern today about Medicaid is the growing burden of nursing home care.

KEY CONCEPTS

Adverse selection

Bundling

Capitation fee

Co-payment

Cost shifting

Deductible

Diagnosis-related groups (DRGs)

Fee-for-service plans

Health maintenance organization (HMO)

Health savings accounts (HSAs)

Individual mandate

Malpractice suits

Managed care

Medicaid

Medicare

Medigap

Moral hazard

Preferred provider

Single payer system

Specifi c egalitarianism

Tax expenditures

Third-party payments

Uncompensated care

QUESTIONS AND PROBLEMS

1. List the various distortions in incentives that arise in the health care sector. (Be sure to include those that aff ect the purchase of insurance as well as the purchase of health care services directly.)

2. In what ways is the purchase of medical services similar to the purchase of a car? In what ways is it diff erent?

3. We have noted that there is extensive disagree- ment on what should be done about the way med- ical care is provided in the United States. To what extent is this disagreement due to diff erences in judgments concerning how the market for med- ical services functions? Be specifi c. To what extent is the disagreement due to diff erences in values?

4. Consider the “market failures” that arise in med- ical markets and current proposals for altering the way medical care is provided in the United States. Discuss the extent to which each of the proposals is aimed at remedying particular mar- ket failures.

5. During the past fi fty years there has been a decline in community-run hospitals and an increase in private (for-profi t) hospitals. Are there reasons why hospitals should be partic- ularly suited or unsuited to being run publicly? What do you think accounts for these trends?

393Review and Practice

6. Critics of malpractice suits claim that they have contributed signifi cantly to the rise of medical costs and want legislation that would limit the size of awards and lawyers’ fees, or otherwise discourage such suits. Many lawyers are con- cerned that any such legislation would impair the rights of victims of malpractice to be justly compensated for the damages they have suff ered. Discuss the equity–effi ciency trade-off s. What do you think should be done�?

7. Assume that medical expenditures are fully deductible from taxes. Show diagrammatically the eff ect on the demand for medical services. If the elasticity of demand with respect to price is 0.7, what is the eff ect of deductibility on an individual in the 15 percent marginal tax bracket; in the 28 percent marginal tax bracket; in the 40 percent marginal tax bracket?

8. Some states have proposed requiring com- munity rating—that is, insurance companies would not be allowed to charge individuals different premiums, regardless of their health,

age, or sex; everyone in the community would pay the same premium. Discuss the possible consequences of community rating. Evaluate such proposals from the perspective of equity and efficiency.

9. Many employers off er their employees a choice of plans, paying a fi xed share of the cost of each. What ineffi ciencies does this introduce�? Some employers, such as Stanford University, have instead off ered a fi xed payment, regardless of the plan chosen, and have insisted that all programs off er identical coverage. Within three years, the cost of providing this standard coverage fell by 20 percent in real terms. Explain why this may have happened.

10. Economists have criticized the tax treatment of health insurance. Why have unions resisted changing this tax treatment�? What are the equity and effi ciency consequences of capping the tax deductibility at some number, say, $5000, repre- senting the average costs of health benefi t plans in unionized fi rms�?

394

EDUCATION

In the United States, education has long been recognized as a responsibil- ity of government. The Land Ordinance of 1785, enacted even before the Constitution, set aside land in the newly established Western territories to fund public schools. Although the locus of responsibility remains at the state and local levels—education is the single largest expenditure at those levels—today the federal government fi nances about 15 percent of all public current and capital education expenditures, using these funds for specifi c purposes, such as providing education for the disadvantaged, encouraging the establishment of educational standards, promoting sci- ence education, and providing fi nancial assistance to enable more people to attend college.

In U.S. national politics, bipartisan consensus on the importance of education is the norm. However, major controversies exist over how best to attain a high-quality education system that provides equality of oppor- tunity, even to the poor.

Two major economic issues—a slowdown in productivity and increas- ing inequality—have informed much of the recent concern about edu- cation in the United States. During the decades between World War II

14

395Education

and the early 1970s, output per worker grew at a rate of 2.8 percent per year, dropped to 1.1 percent per year from 1973 to 1979, and then slowly rebounded to reach 2.5 percent annually from 2000 to 2007. Although explanations of this long interlude of productivity slowdown and sub- sequent substantial increase over the past two decades remain uncer- tain, many economists and policy makers agree that improvements in human capital—the skills and experience of workers—may hold the key to improving productivity growth, and strong education is seen as critical to that goal.1

The early 1970s also marked an important change in the pattern of income distribution in the United States. The fruits of economic growth were shared more equitably during the two decades prior to 1973 than during recent decades.2 Much of this increase in inequality can be related to education. During the 1980s, the diff erence between the income of a college graduate and that of a high school graduate increased enormously3 as the market placed a greater premium on skilled (educated) workers. Thus, providing better education and ensuring that a larger fraction of the population went to college were seen as ways of enhancing oppor- tunity and reducing inequality. Not only would those receiving a college education enjoy higher incomes, but wages would also rise in the market for unskilled labor as the number of unskilled laborers dropped.

For these reasons, and others, economists and policy makers have placed a high priority on strong education—but how strong is the educa- tion system in the United States? One of the most impressive achievements of the U.S. system has been the growth in college attendance throughout the latter half of the twentieth century. The GI Bill of Rights, which pro- vided an opportunity for World War II veterans to attend college, trans- formed higher education from a privilege of the elite to an expectation of the broad middle class. That expectation carried through to another gen- eration as college enrollment rates soared in the 1970s and 1980s,4 while high school dropout rates fell.

1. What are the reasons that government plays such a big role in education?

2. What are the key prob- lems with education in the United States today?

3. What are the major pro- posed solutions?

4. What is the relationship between educational expenditures and outcomes?

5. What are vouchers, and what are the arguments for and against them?

6. What other initiatives are there for changing school governance?

7. Why is there controversy over school standards?

FOCUS QUESTIONS

1 U.S. Bureau of Labor Statistics, “Productivity Change in the Nonfarm Business Sector, 1947–2011,” Labor and Productivity Costs, http://data.bls.gov. 2 One way of seeing this is to divide the country into quintiles by income, distinguishing the poor- est fi fth of the population from the next poorest fi fth, and so on. Before 1980, all quintiles saw their incomes growing, and the poorest quintiles saw their incomes growing fastest, so their share of aggre- gate income increased. From 1980 to 2000, the growth rate of the poorest quintile lagged behind that of the richest quintile, and their income has actually declined in constant 2009 dollars over the past decade, while the richest quintile has continued to enjoy income growth. Thus, since 1980, the lowest quintile’s share of aggregate income has declined from 4.2 to 3.4 percent, whereas the highest quintile’s share has increased from 44 to 50 percent. Rising inequality is even more striking when looking at the relative income growth of America’s richest households over the past three decades: since 1980, the top 5 percent’s share of aggregate income has increased from 17 to 22 percent. (Source: U.S. Census Bureau, The 2012 Statistical Abstract, Table 694). 3 Economic Report of the President, 1997 (Washington, DC: Government Printing Offi ce, 1997), Chart 5-4, p. 169. 4 Economic Report of the President, 1996, Chart 7-1, p. 193.

396 CHAPTER 14 EDUCATION

Even so, in recent decades questions have been raised about the caliber of the U.S. education system. International comparisons inform some of these questions. The United States, despite 89 percent of its population aged 25 to 64 having graduated from high school and 32 percent hav- ing graduated from college by 2010—among the highest rates of OECD countries—has not fared well on standardized math and science tests administered around the world.5 For example, in 1995, eighth graders in the United States ranked twenty-eighth in math—behind Singapore, South Korea, Bulgaria, Russia, and most of the developed countries— and seventeenth in science. Even though America’s performance had improved considerably by 2011, when it was ranked ninth in math and tenth in science, it was still behind much of East Asia and several Euro- pean countries. Inequality in the United States may help explain such results. Students outside America’s poor urban schools achieve results comparable to their peers in other advanced countries, and although the median math and science scores of American students sometimes rank below those of other high-income countries, the best American students (in the ninety-ninth percentile) rank above their peers from many of those countries.6

This brings us to a major issue in recent policy debates: education of the economically disadvantaged, particularly in America’s inner cities and impoverished rural areas. Indigent students receive inadequate pri- mary and secondary education, leaving them unprepared for colleges and universities. Without adequate skills, these people earn low wages, and are thus forced to continue to live in poor areas, continuing the cycle of poverty.

The problem has been exacerbated by some other trends in the econ- omy. During the 1980s, while real wages of the poorest declined, real tui- tion at public two- and four-year institutions of higher learning increased markedly, so the indigent students who were qualifi ed for higher education faced increasingly insurmountable fi nancial hurdles. Thus, even though the increased returns to education noted earlier did lead to an increase in college enrollments, children of more affl uent families were better able to respond to these market signals, and the gap between enrollment rates of the disadvantaged and other Americans actually increased.

All these issues have made education a critical concern for the U.S. public sector. This chapter describes the structure of the U.S. education

5 Organization for Economic Cooperation and Development, Education at a Glance, 2012. 6 International Association for the Evaluation of Educational Achievement (IEA), Trends in Interna- tional Mathematics and Science Study (TIMSS), 1995 and 2011, available from the website of the National Center for Education Statistics (NCES), U.S. Department of Education, http://nces.ed.gov. Research conducted by the OECD yields even worse comparative results for the United States; see OECD, Pro- gram for International Student Assessment (PISA), 2009.

397The Structure of Education in the United States

system, explores the rationale for public fi nance and provision of educa- tion in the United States, and analyzes some of these current issues in U.S. education policy.

THE STRUCTURE OF EDUCATION IN THE UNITED STATES

Traditionally, elementary and secondary school education has been the responsibility of local communities. They fi nanced it (usually with prop- erty taxes), hired the teachers, and determined the curriculum. Figure 14.1 shows the trends in local, state, and federal spending on elementary and secondary education. From 1945 to 1972, the state percentage of educa- tion funds remained steady at 39 percent. During the mid-1970s, state expenditures began to increase at a faster rate than local expenditures,

FIGURE 14.1

REVENUE SOURCES OF PUBLIC ELEMENTARY AND SECONDARY SCHOOLS

All sources of support have increased markedly in recent years.

SOURCE: U.S. Department of Education, National Center for Education Statistics, Digest of Education Statistics 2010, Table 180.

State

Local

Federal

0

50,000

Revenue receipts

(millions of dollars)

200,000

250,000

300,000

150,000

100,000

19 47

–19 48

19 51

–19 52

19 55

–19 56

19 59

–19 60

19 63

–19 64

19 67

–19 68

19 71

–19 72

19 75

–19 76

19 79

–19 80

19 83

–19 84

19 87

–19 88

19 91

–19 92

19 93

–19 94

19 95

–19 96

19 97

–19 98

19 99

–2 00

0

20 01

–2 00

2

20 03

–2 00

4

20 05

–2 00

6

20 07

–2 00

8

398 CHAPTER 14 EDUCATION

so that by the end of the decade, states actually paid for a higher per- centage of total costs. Today, states fi nance almost half the cost of public elementary and secondary schools. Federal support for education, mostly for special programs such as aid to schools with a large number of disad- vantaged children, increased in the 1960s and 1970s, and by 1980 repre- sented 10 percent of total fi nancing for public elementary and secondary schools. Federal support declined from that level to under 7 percent from the mid-1980s to the mid-1990s, but it has risen to between 8 and 9 per- cent of total expenditures for the past decade.

There is, however, a great deal of variability, both in the federal govern- ment’s role and that of the states. For example, whereas in recent years the federal government has provided approximately one-sixth of the funds for primary and secondary schools in Louisiana and Mississippi, it has pro- vided less than 5 percent of the funds in Connecticut and New Jersey. On average, states provided slightly more than half of the nonfederal funds, but this varied, from over 90 percent in Hawaii (where there are no local school districts) and Vermont (where public education is funded primar- ily by a state-determined share of the residential property tax), to only one-third in Illinois and Nevada. Revenue from private sources, while still comprising a small share of total funding for public elementary and sec- ondary schools, is rapidly increasing in importance, providing 6 percent of nonfederal funds in Tennessee and 5 percent in Oklahoma.7

At the same time, the states have taken an increasingly active role in setting certain minimum standards. They set the number of days stu- dents must be in school per year; the lowest age at which children may drop out of school; and the minimum education requirements for teachers employed in public schools, often specifying, for instance, not only that the teachers have a college degree, but also that they have taken a speci- fi ed number of courses in education. States also play a role in determining curricula. For example, many states require courses in American his- tory in the eleventh grade and adopt certain approved lists of textbooks, from which the local community must select the books it will use. Several states have made provisions for state takeovers of local school districts when there is “academic bankruptcy”—that is, when the local school dis- tricts fail to meet certain minimal standards.

One consequence of local control is that many rich communities spend far more on education than most poor communities. The range in edu- cational expenditures per student has been enormous, even within indi- vidual states, with rich districts spending two, three, or more times the amount spent in poor districts.

7 U.S. Department of Education, National Center for Education Statistics, Digest of Education Statistics 2010, Table 181.

399The Structure of Education in the United States

About 90 percent of all elementary and secondary school students attend public schools. Of those who go to private schools, about 40  percent go to Catholic schools and another 40 percent to other parochial schools, with the remaining one-fi fth attending nonsecta- rian schools. Private institutions play a much more important role in higher education. About 40 percent of bachelor’s, master’s, and doctoral degrees are earned at private institutions, almost all of which are not- for-profi t, and approximately 60 percent of fi rst professional degrees are earned in private schools.8

Because the principle that states and localities should be responsible for making educational decisions is held so dear, the federal government has relied on a variety of indirect ways to try to infl uence states and locali- ties. In 1957, after the Soviet Union launched Sputnik, the fi rst space sat- ellite, and thus appeared to be surpassing the United States in science, the federal government supported eff orts to improve the science curric- ulum in elementary and secondary schools. In the early 1990s, when it became apparent that many of those graduating from high school who did not go on to college were having a hard time entering the labor force, the federal government sponsored a school-to-work initiative. This encour- aged the states to design apprenticeship-like programs (so successful in other countries, such as Germany) that would facilitate the movement from school to work. When, again in the early 1990s, the federal govern- ment wanted to encourage higher standards in elementary and secondary schools, it did not impose national standards, but rather provided fund- ing that helped the states establish standards for themselves, and several commissions were set up to help in the formation of national standards in the various disciplines.

FEDERAL TAX SUBSIDIES TO PRIVATE AND PUBLIC SCHOOLS

The personal income tax may have important eff ects on the demand for public and private education. Expenditures by state and local communi- ties for education (as well as for other purposes) are implicitly subsidized by the federal government because state and local taxes are deductible on federal income tax returns, and interest on state and local bonds is exempt from federal taxation. This means that if my community taxes me $1000 to support public schools, the cost to me is far less than $1000. If I am in

8 First professional degrees include those for medicine, dentistry, pharmacy, veterinary medicine, law, and theological professions. See U.S. Department of Education, National Center for Education Statis- tics, Digest of Education Statistics 2010, Tables 35, 62, and 287.

400 CHAPTER 14 EDUCATION

the 31 percent marginal tax bracket (so I pay 31 cents of each additional dollar of taxable income to the federal government), then by deducting $1000, my federal taxes are reduced by $310. The net cost to me of $1000 on public education is thus only $690. In contrast, if I spend $1000 on private education, it costs me $1000 that I could have spent elsewhere.9

The total value of tax expenditures specifi cally for education in 2011 was estimated to be approximately $48 billion, as Table 14.1 shows. Most of these expenditures came from the reduction of taxable income because of a variety of deductions, exclusions, and exemptions. Because the value of these income adjustments is greatest for higher-income individuals, this form of support for education is, in eff ect, regressive; that is, it benefi ts higher-income indi- viduals and higher-income communities more than lower-income ones.

At the same time, the tax system serves to discourage private expen- ditures on education.10 If I spend $1000 in tuition to send my child to a

TABLE 14.1 ES TIMATED FEDER AL TA X E XPENDITURES FOR EDUCATION, FISCA L YE AR 2011 (MILLIONS OF DOLL ARS)

REVENUE LOSS

CORPORATE INDIVIDUAL

American Opportunity Tax Credit $13,060

Lifetime Learning Tax Credit 2,800

Exclusion of Scholarship and Fellowship Income 3,060

Exclusion of Interest on Student-Loan Bonds and on Bonds for Private Nonprofi t Educational Facilties

750 1,830

Parental Personal Exemption for Students Age 19 and Over

5,600

Qualifi ed Tuition Programs 1,610

Deductibility of Charitable Contributions for Educational Purposes

650 2,870

Deductibility of Student-Loan Interest 1,400

Deduction of Non-Business State and Local Taxes (Total = $41,060)

11,907

Other 260 1,850

Total $1,660 $45,987

SOURCE: Calculated from Executive Offi ce of the President, Offi ce of Management and Budget, Budget of the United States Government, Fiscal Year 2013, Analytical Perspectives, Table 17-2. The expenditure resulting from the deductibility of state and local taxes assumes state and local jurisdictions spend 29 percent of revenues on education (calculated from “State and Local Governments—Revenue and Expenditures by Function: 2007 and 2008,” Statistical Abstract of the United States, 2012, Table 436).

9 This example assumes that I itemize my deductions rather than claiming the standard deduction. Only individuals with large deductible expenses such as interest on a home mortgage itemize. See Chapter 22 for a more detailed discussion. 10 However, the tax deductibility of gifts to private schools, which are treated as charities, strongly encourages private education.

401Why Is Education Publicly Provided and Publicly Financed?

private school, not only do I pass up the already-paid-for public education, but the expenditure is not even tax deductible. Later, the return to the education, which occurs in the form of higher wages, is taxed. The tax system thus discourages private spending on education.

WHY IS EDUCATION PUBLICLY PROVIDED AND PUBLICLY FINANCED?

In the United States, the public role in the provision of education has been so pervasive that it is generally taken for granted. In some other countries, however, although the government may provide funds to edu- cational institutions, much of the education itself is provided by private, particularly religious, schools.

IS THERE A MARKET FAILURE?

Education is not a pure public good. The marginal cost of educating an addi- tional child is far from zero; indeed, the marginal and average costs are (at least for large school districts) approximately the same. Furthermore, there is no technical diffi culty in charging individuals for use of this service.

Those who seek to justify public education in terms of a market failure focus on the importance of externalities. It is often claimed, for instance, that there are important externalities associated with having an edu- cated citizenry. A society in which everyone can read functions far more smoothly than a society in which few can read. There is also a large pri- vate return to being able to read, and even in the absence of government support, many individuals would learn this and other basic skills. Indeed, most individuals would go far beyond that. The question is: Given the level of education that individuals would choose to undertake privately if there were no government subsidy, would further increases in education generate any signifi cant externalities? There is no agreement concerning the answer, but the case for government support based on these kinds of externalities—at least for advanced industrial economies like the United States—seems, at best, unproved.11

11 See, for instance, D. M. Windham, “Economic Analysis and the Public Support of Higher Education: The Divergence of Theory and Policy,” in Economic Dimensions of Education, A Report of a Committee of the National Academy of Education, May 1979.

402 CHAPTER 14 EDUCATION

There may be other important externalities associated with educa- tion. Public education may have played an important role in integrating new immigrant groups into American culture. Public education may have been essential in making the melting pot work. The benefi ts of this accrued not only to individuals, but also to the nation as a whole.

There is another indirect form of externality to investments in science and technology education: people with these scarce skills are the key to technological progress and, typically, innovators capture only a fraction of their overall contribution to the increase in productivity.

However, there may be a reason why individuals do not invest as much in education as they would like, even to the point at which the private return equals the cost of capital: they may lack access to funds to fi nance their edu- cation. Private lenders are not, for the most part, willing to lend to fi nance education; understandably, banks are concerned about getting repaid. Hence, those lacking funds of their own (or funds from their families) would be denied access to education without some assistance from the government.

There is another market failure: investments in education are very risky. Individuals do not really know what the payoff is going to be—but there is no way that individuals can divest themselves of this risk. Many individuals will be reluctant to invest in their education, even if the aver- age returns exceed the cost of capital, simply because it would impose too much risk. Matters have become worse in recent years. It used to be that individuals whose income turned out low could discharge their debts through bankruptcy—and they can for debts other than student debts. Since 2005, however, it has become extremely hard for students to dis- charge their debts. Therefore, even if the school they attend does not pre- pare them for a better paying job, they are saddled with paying the loan.

There is still another reason why there may be underinvestment in edu- cation. At the elementary and secondary levels, parents make decisions on behalf of their children. Although most parents may view expenditures on their children altruistically, investing in their children as long as the returns are suffi ciently high, some parents may not do so. Under a system of privately fi nanced education, the children of such parents might receive an insuffi cient education. Thus, a rationale for public support of primary and secondary education is provided on distributional grounds: there is a widespread belief that children’s access to education should not depend on their parents’ fi nancial ability or their sense of altruism. Indeed, such distributional concerns provide the strongest motivation for a public role in education more generally—for the fi nancing of elementary and second- ary education and for the extensive fi nancial support of higher education.

Although these arguments provide a rationale for government fi nancing of education, they do not provide an argument for government

403Issues and Controversies in Educational Policy

production. Why is it that at the elementary and secondary levels, govern- ment production (public schools) dominates, whereas at the college level, private not-for-profi t education plays a far more important role? In many other countries, private elementary and secondary schools are more dom- inant than they are in the United States. We examine the controversy over private schools later in this chapter.

THE FEDERAL ROLE

The arguments discussed earlier explain why some governmental entity should provide fi nancing for education, but they do not explain the level— local, state, or federal—at which the support should be provided. Increas- ingly, many arguments for a public role in education are arguments for a federal role. Largely this is because to the extent that there are exter- nalities, given the huge migrations that occur throughout the country, those externalities need to be addressed at the federal level. Inadequate education is associated with a variety of social problems and antisocial behaviors. Those who grow up in one state and receive an inadequate education there have a good chance of practicing their antisocial behav- iors in another state.

Moreover, distribution concerns can be addressed fully only at the national level, given the huge disparities in per capita incomes (and corre- sponding disparities in education expenditures per pupil) across states.12 States that spend less on education per pupil tend to generate lower per- formance, as measured by earnings.13

ISSUES AND CONTROVERSIES IN EDUCATIONAL POLICY

Education has long been a source of political and intellectual debate. There are diff erent perspectives on the basic impact of education on stu- dents, how education should be produced and fi nanced, and whether the

12 For instance, the state with the highest per capita income (Connecticut) had a per capita income almost twice that of the lowest (Mississippi) in 2008; total education expenditures per elementary and secondary pupil in Connecticut were also almost twice that of Mississippi. 13 D. Card and A. B. Krueger, Wages, School Quality, and Employment Demand (New York: Oxford University Press, 2011).

404 CHAPTER 14 EDUCATION

current system results in too much inequality. The following sections take up some of these issues and controversies.

EDUCATION OUTCOMES

There are diff erent perspectives on the outcomes of the education pro- cess. One view, implicit in the foregoing discussion, holds that education increases the skills of individuals, and thereby wages. Called the human capital view, this perspective sees investment in people as akin to capi- tal investment. The greater the investment, the greater the productivity.14 However, there is controversy over why, and to what extent the higher wages earned by skilled workers refl ect an increase in their productivity resulting from education.

One variant of the human capital view, emphasized by Samuel Bowles and Herb Gintis of the University of Massachusetts, focuses on the social- izing role of education. Education teaches people how to perform well in the workplace by teaching how to obey orders, follow directions, and work in teams. When successful, this socialization teaches punctuality and reliability. In this perspective, those who go to school longer learn more of these social skills, or, in any case, have demonstrated a greater ability or willingness to cope with the demands of the school system. These social abilities (or drives) make these individuals more valuable in the workplace.15

Another view of education, called the screening view, argues that one of the important functions of education is to identify the abilities of dif- ferent individuals. Those who go to school longer receive higher wages and are observed to be more productive. This is not because the schools have increased their productivity, but rather because the schools have identifi ed those individuals who are the most productive, or who have the necessary drive and ambition. The school system is viewed as a screening device, separating the very able and highly motivated from the less able and less motivated.16

14 See G. Becker, Human Capital: A Theoretical and Empirical Analysis with Special References to Edu- cation, 2d ed. (New York: National Bureau of Economic Research, Columbia University Press, 1975). 15 A. Weiss, “Human Capital versus Signaling Explanations of Wages,” Journal of Economic Perspec- tives 9, no. 4 (Fall 1995): 133–154. Weiss closely studied low-skilled workers in a manufacturing plant. Long-run success depended not on any particular skill but on social characteristics such as reliability (low levels of absenteeism) and punctuality. 16 This view has been put forward by J. E. Stiglitz, “The Theory of Screening Education and the Distribution of Income,” American Economic Review 65 (1975): 283–300; A. Michael Spence, “Job Market Signaling,” Quarterly Journal of Economics 87 (1973): 355–374; and K. J. Arrow, “Higher Educa- tion as a Filter,” Journal of Public Economics 2 (1973): 193–216.

405Issues and Controversies in Educational Policy

In the screening perspective, the social returns to education are far less than the private returns. The private returns can be signifi cant: those who go to college receive a substantially higher income than those who do not. If all that was going on were the identifi cation of who was more able and who less, though, the total production of society would be unaff ected; the social returns would be zero. In fact, education also identifi es diff er- ences in skills, enabling a better matching of individuals and jobs, and this does increase overall productivity: there can be signifi cant social returns to screening. There is general agreement today that some of the returns to education are the result of increases in skills and some are a result of screening; the controversy arises over their relative importance.17

DO EXPENDITURES MATTER?

The signifi cance of the ways in which education diff ers from other com- modities is posed most starkly by the controversy over whether increased educational expenditures lead to increased education performance. In the production of a standard commodity, an increase in input would necessar- ily lead to an increase in output. Earlier, we noted the low performance of American students compared to those abroad, as indicated by certain test scores. Figure 14.2 shows clearly that this is not because the United States is spending less—in fact, on a per-pupil basis, United States is spend- ing more than almost any other country. Similarly, expenditures per pupil have increased markedly over the past fi fteen years, with only a modicum of eff ect on test scores. The weak link between expenditures and perfor- mance had been demonstrated earlier, in a classic study by James Cole- man of the University of Chicago and his co-authors.18 Coleman’s study spawned a huge body of literature, with his results still in dispute.19

17 For instance, studies showing that wages do not depend closely on the subjects studied suggest that content (skill formation) matters little. Moreover, if skill formation were predominant, one would expect that the increase in returns would be a relatively steady process, so the increment in wages per year would be the same; in fact, though, the increase in wages from the completion of the fourth year of college is much greater than the increase in wages from the completion of the preceding three years. 18 J. S. Coleman et al., Equality of Educational Opportunity (Washington, DC: Department of Health, Education, and Welfare, 1966). 19 See the survey by Eric A. Hanushek, “School Resources and Student Performance,” in Does Money Matter? The Eff ect of School Resources on Student Achievement and Adult Success, ed. G. Burtless (Washington, DC: Brookings Institution, 1996), pp. 43–73; and Hanushek’s earlier arti- cle, “The Economics of Schooling: Production and Effi ciency in Public Schools,” Journal of Economic Literature 24, no. 3 (September 1986): 1141–1177. See also H. Wenglinsky, When Money Matters: How Educational Expenditures Improve Student Performance and How They Don’t (Princeton, NJ: Educa- tional Testing Service, Policy Information Center, 1997); A. B. Krueger, “Experimental Estimates of Educational Production Functions,” Quarterly Journal of Economics 113 (May 1999): 497–532; and E. N. Wolff , W. Baumol, and A. N. Saini, “A Comparative Analysis of Education Costs and Outcomes: The United States vs. Other OECD Countries,” Working Paper, June 2013.

406 CHAPTER 14 EDUCATION

Many economists believe that increased expenditures do make a diff erence—a greater diff erence than Coleman’s earlier study suggested. A set of studies that looked at the performance of identical twins (so natural diff erences were fully accounted for) showed that levels of educa- tion made a systematic diff erence in earnings. Another study compared states with diff erent rules for when students could drop out of school; those who were forced to stay in school longer (controlling for other vari- ables) seemed to do better.

The problem with interpreting the consequences of diff erences in expenditure levels is that much—and an increasing proportion—of edu- cation expenditures in the United States goes to purposes not directly related to teaching (such as administrative expenses) and to addressing the requirements of those with special needs. There are only limited data looking at, for instance, the consequences of smaller class sizes. The most famous experiment occurred in Tennessee, where students were ran- domly assigned to diff erent class sizes; the results of that experiment,

DIFFERENCES IN EDUCATIONAL

EXPENDITURES AMONG COUNTRIES

The United States spends more per pupil on primary education than almost any other country, and 40 percent more than the

OECD average.

SOURCE: Organization for Economic Cooperation and Development,

Education at a Glance 2011: OECD Indicators, Table B1.1a.

FIGURE 14.2

$0

$2,000

$4,000

$6,000

$8,000

$10,000

$11,077

$9,982

$9,080 $8,758

$7,491 $7,208 $7,184 $7,153 $6,723

$6,267 $5,929

$5,420 $5,314

$2,246

$12,000

No rw

ay

Un ite

d S ta

te s

Sw ed

en

Un ite

d K ing

do m

Ja pa

n

Ne th

er lan

ds Sp

ain

OE CD

A ve

ra ge

Au str

ali a

Fra nc

e

Ge rm

an y

Ko re

a Isr

ae l

M ex

ico

Expenditures per pupil

for primary education (2008 $)

407Issues and Controversies in Educational Policy

though not conclusive, support the view that smaller class size (entailing greater expenditures) does lead to improved student performance.20

Another set of critiques of Coleman’s fi ndings argues that the real output of educational expenditures should not be higher test scores but higher productivity, leading to higher wages. Several studies have estab- lished a clear link between expenditures and earnings.21

Those who believe that schools have relatively little impact on earn- ings believe that home background is critical. Even in that view, it is still possible that by increasing expenditures on the disadvantaged, pub- lic schools can help off set defi ciencies in home background. This raises fundamental questions about the allocation of resources within schools: How much should go to helping the academically gifted, the average stu- dents, and those at the bottom? The country’s technological leadership depends on having the best scientists in the world, and this argues for putting resources at the disposal of the scientifi cally gifted. On the other hand, without adequate skills, those at the bottom will see their wages fall behind, as they have been doing in the last two decades (see Figure 14.3). Increasing inequality is likely to give rise to increasing social problems in the decades ahead. Education advocates argue that we should spend more on both, but given the current overall limitations on expenditures, the issue is: Should we direct more to the top or the bottom, or do we now have just about the right balance? The appendix to this chapter presents a framework for thinking about this issue.

SCHOOL VOUCHERS: CHOICE AND COMPETITION

Perhaps the most heated recent debate in the political sphere has con- cerned the question of school choice. Should parents be given more choice about where their children go to school? The simplest proposals for pro- viding school choice entail school vouchers: each child would be given a

20 See E. Word et al., “Student/Teacher Achievement Ratio (STAR)—Tennessee’s K-3 Class Size Study,” Tennessee Department of Education, 1994. This study corroborates other fi ndings. In an analysis of many studies, Gene Glass concluded that “small classes were very much better than large classes; large classes were hardly any better than very large classes.” See G. V. Glass, School Class Size: Research and Policy (Beverly Hills, CA: Sage Publications, 1982), p. 47. In a review of eleven studies, Robert Slavin found that the largest (but still modest) eff ects of class size on test performance were observed in the study with the largest reduction in class size (from twenty-three to fi fteen). See R. E. Slavin, “School and Classroom Organization in Beginning Reading,” in Preventing Early School Failure: Research, Policy, and Practice, ed. R. E. Slavin et al. (Boston: Allyn and Bacon, 1994), pp. 122–130. See also K. Akerhielm, “Does Class Size Matter?” Economics of Education Review 13, no. 3 (September 1995): 229–241; D. Card and A. Krueger, “School Resources and Student Outcomes,” Annals of the American Academy of Political and Social Science 559 (September 1998): 39–53; and G. J. Whitehurst and M. M. Chingos, Class Size: What Research Says and What It Means for State Policy (Washington, DC: Brookings Institution, May 2011). 21 See, for example, C. R. Belfi eld and H. M. Levin, eds., The Price We Pay: Economic and Social Conse- quences of Inadequate Education (Washington, DC: Brookings Institution, 2007).

408 CHAPTER 14 EDUCATION

SOURCE: T. D. Synder and S. A. Dillow, Digest of Education Statistics 2010,

Table 391. The ratios are computed from median incomes of full-time workers

aged 25 and older, in 2009 dollars.

DIFFERENCES IN INCOME ACROSS LEVELS

OF EDUCATION

These charts plot median incomes of men and women

with higher levels of education relative to high school dropouts.

Since 1990, the earnings of minimally educated men and

women have fallen behind their better-educated peers.

FIGURE 14.3

Ratio of bachelor’s degree to high school dropouts

Ratio of high school grads to dropouts

Ratio of some college to high school dropouts

0.0

0.4

0.2

0.6

1.0

1.4

1990 1995 2000 2005

Income ratio

Males

1.6

1.8

2.0

2.2

2.4

1.2

0.8

1990 1995 2000 2005

Ratio of bachelor’s degree to high school dropouts

Ratio of high school grads to dropouts

Ratio of some college to high school dropouts

0.0

0.4

0.2

0.6

1.0

1.4

Income ratio

Females

1.6

1.8

2.0

2.2

2.4

1.2

0.8

409Issues and Controversies in Educational Policy

coupon—worth, say, $5000—to be used at the school of the parent’s choice. Public schools would, under this proposal, have to compete directly with private schools; they would have to raise their revenue by persuading stu- dents to attend, just as private schools do. If parents valued the kinds of pro- grams provided by the public schools, then the public schools would do well.

Implicit in much of the discussion favoring vouchers is a critique of the role of government in education: today it both fi nances education and produces it. Many believe, as we saw in Chapter 8, that, for a variety of reasons, the government is not an effi cient producer, and that in the com- petition between schools, private schools would win out. Moreover, they say, when parents choose their children’s school, they become more com- mitted to that school and more involved in their children’s education, and this contributes to school performance.

Interestingly, and contrary to the claims of critics, advocates of choice argue that private schools not only produce a higher-quality education at lower costs, but actually promote equality. In another famous study, James Coleman and his co-authors argued that America’s public schools were actually more segregated, both racially and socioeconomically, than were its private schools. The segregation resulted because the public school system, especially at the primary level, is based on neighborhood schools, and neighborhoods are, in eff ect, segregated. They also argue that private schools are more eff ective in educating disadvantaged stu- dents (for instance, as measured by test scores).22

Finally, advocates of vouchers argue that America already has a sys- tem of choice, but constrained choice. Those with enough resources can choose to move to the suburbs or put their children in private school. It is the poor whom the current system deprives of choice. There is competi- tion, but only for the children of the affl uent, not for the poor. America believes in competition. It may be natural for those in any industry to argue that competition in their sector would be less desirable, and the postulates of perfect competition may not be satisfi ed perfectly. Still, as a country, we believe that the advantages of competition outweigh the dis- advantages. Why should education be any diff erent?

Voucher programs have been attacked on several grounds. The condi- tions that make competition work in conventional markets do not exist in education. Parents, particularly less well educated parents, often are not well informed; they are ill prepared to judge the eff ectiveness of schools in providing key skills. Indeed, this information problem is a key rationale for the establishment of national standards.

22 J. Coleman, T. Hoff er, and S. Kilgore, Public and Private Schools: An Analysis of High School and Beyond  (Washington, DC: U.S. Department of Education, National Center for Educational Statistics, November 1981).

410 CHAPTER 14 EDUCATION

This problem is even true for trade schools and colleges. Many with poorer backgrounds have been taken advantage of by schools that do not provide a good education and that engage in deceptive advertising—schools where comple- tion rates are low and job placements are poor. Government attempts to regulate these eff ec- tively for-profi t schools, or even to limit govern- ment loan programs for such schools, have been beaten back. The schools are highly profi table, and it pays for these schools to lobby heavily and make substantial campaign contributions. These for-profi t schools, and their ability to manipulate the political system for their own advantage—even at the cost of those that they are supposed to serve—have become a major national scandal.

Moreover, in many areas, the number of schools is limited; many par- ents are inclined to choose schools on the basis of convenience rather than educational excellence. Competition is further limited by the fact that once children are enrolled in a school, transferring is likely to be dif- fi cult; and the less standardized the curriculum, the more diffi cult adjust- ment will be.

Critics of choice argue, furthermore, that the alleged superior perfor- mance of private schools is a mirage. Part of the seemingly superior perfor- mance arises from a selection eff ect: those who choose to send their children to private schools are more committed to education, and it is this commit- ment, rather than what the schools do, that accounts for any measured dif- ferences in performance. Second, there is a discipline eff ect: private schools reject or expel those who are disruptive or who fail to perform adequately in other ways. Public schools cannot do this. Although it may be true that today performance in private schools exceeds that of public schools for indi- viduals of seemingly similar backgrounds, this would not be the case if there were a major expansion of private schools, especially if they were required to take all students, including those with discipline problems.

Moreover, say many critics of vouchers, the scheme would lead to a more socially and economically stratifi ed society, with children of wealthy and well-educated parents going to one set of schools and chil- dren of poor and less well-educated parents going to another. Although regulations to prevent racial discrimination might be easily enforced, reg- ulations to ensure the absence of socioeconomic stratifi cation would be diffi cult to implement.

THE VOUCHER DEBATE PRO CON

Competition promotes effi ciency.

Conditions that make competition work are not satisfi ed.

Choice promotes commitment.

Will result in more social and economic stratifi cation.

Risk of economic stratifi cation.

Evidence on greater effi ciency is questionable: What will happen to problem children?

411Issues and Controversies in Educational Policy

Voucher schemes remain very much on the political agenda. In part, this refl ects the widespread dissatisfaction with American elementary and secondary education—although curiously, whereas most parents are critical of the state of education overall, they believe their own children are receiving a good education. The concerns over quality are especially strong in America’s inner cities, which have high dropout rates.

Those seeking to introduce innovation and more competition within public schools have looked to alternatives to vouchers. One is contracting out: the school district contracts outside managers to run their schools. Advocates claim that the competition between management teams will lead to higher performance. School districts can make more informed decisions concerning management teams than many parents can con- cerning schools. Several school districts have experimented with such arrangements, and despite suffi ciently ambiguous evidence of signifi cant cost savings or improved student achievement, the contracting-out busi- ness has continued to grow.23

Still another initiative involves charter schools, which are self- managed or managed directly by the children’s parents, rather than by a larger school district. A variation of this in England is “academies,” pub- lic schools removed from local council control and given more autonomy in their staffi ng and teaching methods. There are now a large number of charter schools throughout the United States, and parents of children attending them generally are highly supportive. Supporters of the char- ter schools hope that their innovations will be imitated elsewhere, and that they will thereby spur an overall improvement in educational quality. Critics are more skeptical, however. Currently, charter schools, like pri- vate schools, cater to those most committed to education and most desir- ous of an alternative to the current public school system. Without wider scope for school choice, however, how will their impact be felt on pub- lic education more broadly? Studies examining charter schools suggest that, on average, they have not improved performance, although the best schools have done very well.24

Other reforms emphasize the broader role of incentives, such as merit pay for teachers. How eff ective such incentives can be depends to a large extent on how well teacher performance can be measured. The problem is

23 For an extensive discussion of contracting out principles and practices, see P. Hill, L. Pierce, and J. Guthrie, Reinventing Public Education: How Contracting Can Transform America’s Schools (Chicago: University of Chicago Press, 1997). For an analysis of contracting out results, see P. Burch, “After the Fall: Educational Contracting in the USA and the Global Financial Crisis,” Journal of Education Policy 25 (November 2010): 757–766. 24 For assessments of charter school performance, see M. Berends and H. Walberg, Charter School Out- comes (Hillsdale, NJ: Lawrence Erlbaum Associates, 2008); R. Zimmer et al., Charter Schools in Eight States: Eff ects on Achievement, Attainment, Integration, and Competition (Santa Monica, CA: Rand Cor- poration, 2009); and S. Imberman, “Achievement and Behavior in Charter Schools: Drawing a More Complete Picture,” The Review of Economics and Statistics 93 (May 2011): 416–435.

412 CHAPTER 14 EDUCATION

VOUCHERS: THE SAN JOSE AND MILWAUKEE EXPERIMENTS

*G. Bridge, “Citizen Choice in Public Services; Voucher System,” in Alternatives for Delivering Public Services, ed. E. S. Savas (Boulder, CO: Westview Press, 1977).

†C. E. Rouse, “Private School Vouchers and Student Achievement: An Evaluation of the Milwaukee Parental Choice Program,” Quarterly Journal of Economics 113, no. 2 (May 1998): 553–602; and J. Greene et al., Expanding Choice in Elementary and Secondary Education: A Report on Rethinking the Federal Role in Education (Washington, DC: Brown Center on Education Policy at Brookings, February 2010).

Education vouchers would seem to be an ideal subject for experimentation: communities could make vouchers available and see what happens. In fact, a few communities have done just that. One of the earliest experiments, involving choice only among public schools, occurred in San Jose, California. One study concluded, after exam- ining the results of that experiment, that

education vouchers stand little chance of succeeding in American elementary and secondary schools. [Parents eventu- ally] learn about their alternatives and the rules governing choice. But there is still some question about the social conse- quences of parents’ or students’ program choices. Specifi cally, there is some con- cern that parents pick programs which reinforce their class-related social values, so that poor children have little opportu- nity to acquire the beliefs, attitudes, and social competencies necessary for social mobility to the middle class.*

One of the few experiments to provide vouch- ers that could be used in public or private schools was begun in Milwaukee in 1990. Curiously, as the experiment proceeded, both supporters and critics of vouchers claimed that the Milwaukee experiment confi rmed their views concerning the effectiveness of vouchers.

Milwaukee provided a limited number of vouch- ers that parents could use to send their children to the school of their choice. In 1990, the fi rst year of the voucher program, the number of vouchers was limited to 1 percent of the public school enrollment (approximately 950 students); the program was expanded to 1.5 percent in 1994.

Critics pointed out that relatively few parents availed themselves of the opportunity. Between 1990 and 1993, fewer than 600 parents applied each year, with an average of under 400 selected and 350 actually enrolled. To set this number in perspective, Milwaukee public school enrollment was 95,000, of which 42 percent came from families below the pov- erty level; any family with an income of up to 1.75 times the poverty level was eligible to apply. Even more telling, of those who elected to go to private schools, a large fraction transferred back into pub- lic schools within a year or so. By 1994, only 92 of the 354 students enrolled in a choice school in 1990 remained in the program. There seemed to be lit- tle effective competition with—and therefore spur to—public schools, and little spur to the expansion or creation of new private schools.

Advocates pointed to the higher performance of those who did switch to private schools and remained there, compared to that of peers with similar backgrounds who had remained in public schools. Statisticians clashed over the extent to which these differences could be attributed to the special nature of the children who had elected to go to and remain with the private schools.†

413Issues and Controversies in Educational Policy

that it is diffi cult to assess the teacher’s value added: that is, if a teacher has students that enter the class poorly prepared, educational attainment is likely to remain weak at the end of the year—even if the teacher did a good job. A second problem is that teachers are typically poorly paid. Many, perhaps most, enter the profession not because of the money but because they love teaching. Small pittances paid as part of incentives reminds them of how poorly paid they are. This may actually lead to poorer per- formance, and those who have opportunities elsewhere move into other better paid jobs. In many contexts, it has been shown that incentive pay systems backfi re.25

Many of the critics of America’s education system like to blame unions— but the role of unions is more ambiguous than the rhetoric often suggests. Some of the highly unionized states have far better outcomes than the less unionized states. Unions have fought hard to maintain public support for public education, and to ensure that there are suffi cient funds—including, not surprisingly, funds to pay teachers adequately. Without their actions, sal- aries and working conditions might have been much worse, and thus school systems would not have been able to recruit good teachers, and education outcomes would have suff ered. Still, in some places and in some areas, teach- ers unions have resisted changes and reforms that might have had positive benefi ts. What is clear is that without the buy-in of those who are responsible for education—the teachers—it is hard to see how one can get educational reforms that would result in substantial improvements in performance. That is why some suggest that the best approach is a “deal”: more fl exibility on the part of unions in return for better support for public education.

SCHOOL DECENTRALIZATION

A widely discussed school reform, school decentralization, aims to address problems associated with the control of schools by large educa- tional bureaucracies.26 Decentralization has been an issue in large urban school districts, where such bureaucracies have been seen as unrespon- sive to parents. The eff ect of decentralization is to shift authority to

25 There is a large and growing literature discussing these reforms, as well as many fi eld experiments testing diff erent incentives. See, for instance, E. Hanushek and D. Jorgenson, eds., Improving America’s Schools, The Role of Incentives (Washington, DC: National Research Council, 1996); D. Neal, “The Design of Performance Pay in Education,” Chapter  6 in Handbook of the Economics of Education, vol. 4, ed. E. Hanushek, S. Machin, and L. Woessmann (Waltham, MA: North-Holland, 2011); M. Springer et al., Teacher Pay for Performance: Experimental Evidence from the Project on Incentives in Teaching (POINT) (Evanston, IL: Society for Research on Educational Eff ectiveness, 2011); and E. Hanushek and L. Woessmann, “Symposium on Performance Pay for Teachers,” Economics of Education Review 30 (2011). 26 For an extensive discussion of decentralization, see J. Hannaway and M. Carnoy, eds., Decentraliza- tion and School Improvement (San Francisco: Jossey-Bass, 1993).

414 CHAPTER 14 EDUCATION

individual schools and their principals, giving greater infl uence to par- ents. Advocates argue that school decentralization creates stronger incen- tives for parents to monitor teachers and schools, to the benefi t of school performance. Teachers, too, the argument goes, have stronger incentives to perform better because they have greater infl uence in the educational process at their schools.

Criticisms of decentralization have been far more muted than those of the voucher programs; a number of important decentralization initiatives are already in place, including those in two of the country’s largest cities: New York and Chicago. However, there is concern that unless union rules and attitudes are changed, little headway will be made in fundamental reform. Another worry is that if there is too much discretion at the level of schools, too much inequality will be generated: some schools will be far better than others, and this seems unfair. Supporters of decentralization point out that this, of course, is already true. Thus, any attempt to improve quality will entail some temporary inequality, until the poorer schools are lifted to the level of the better schools, but it makes little sense to keep the best schools down simply to avoid a disparity.

PERFORMANCE STANDARDS: NO CHILD LEFT BEHIND AND RACE TO THE TOP

Many believe that before decentralization, choice, or any other reform can be eff ective, clear performance measures for schools must be established. Without such measures, parents, teachers, and educational reformers cannot begin to determine whether schools are performing well or poorly, and therefore have no basis for educational reforms.

Establishing clear performance goals was the main objective of the No Child Left Behind (NCLB) Act of 2001, a bill to overhaul the Elementary and Secondary Education Act of 1965 as well as to replace the Goals 2000: Educate America Act of 1994. Under this legislation, states are required to conduct annual student assessments linked to state standards and partici- pate in National Assessment of Educational Progress reading and mathe- matics tests in fourth and eighth grades, develop adequate yearly progress (AYP) standards and take specifi c steps when AYP is not met, and ensure that their teachers are “highly qualifi ed.” NCLB also provides grants for the development and implementation of standards-based assessments, curriculum development in priority subject areas, targeting of disadvan- taged students, and the promotion of educational innovations.

Even though there is widespread support for the establishment of per- formance goals, the legislation has encountered considerable opposition,

415Issues and Controversies in Educational Policy

for several reasons. Some voice concern that certain skills are easier to test than others. For example, there are good ways of testing basic skills, like phonics, but testing creativity and what are called “higher order” cognitive (thinking) skills is far more diffi cult. One of the strong points of American education has been how it has encouraged creativity. Will the tests divert teachers’ attention toward teaching basic skills, and away from encouraging creativity and thinking skills? Some critics also claim that teachers are likely to gear their teaching in response to the tests (“teaching to the test”), and that high-stakes testing has led some schools to intentionally manipulate student test scores. (Several scandals have suggested that this has actually happened.)

On the other hand, advocates claim that our schools are failing to teach basic skills to many youngsters, and getting them to do that would be a major advance.

Still others are worried about potential inequalities. Schools with disadvantaged children worry that they will be criticized for their poor performance, and they are not persuaded by reassurances that in being judged, the background of their students will be taken into account. There is also a concern that children from disadvantaged backgrounds will likely perform relatively poorly on these standardized tests, and that employers will then use these tests as a simple and acceptable screen for determining whom to hire, putting already disadvantaged groups in an even worse position.

Most importantly, critics of NCLB argue that parents already know that schools are not performing as well as they would like. They believe that without more fundamental changes in the organization of schools, information itself will do little but increase parents’ sense of helplessness.

Critics also say that the resources required to achieve the goals of “leaving no child behind” have not been provided—and are unlikely to be provided in the current era of budget stringency.

Congress is now debating whether to “end or mend” NCLB. This legislation has indeed com- pelled states to design and implement school accountability systems and take concrete action to improve performance based on annual stu- dent assessments. However, the results on stu- dent achievement thus far are mixed. NCLB has generated substantial and almost universal gains in mathematics, but has had virtually no positive eff ect on reading skills, and has gener- ated only modest impacts among disadvantaged cohorts in math, thus making little progress in

ENHANCING THE EFFICIENCY OF SCHOOLS 1. Choice and competition: school vouchers,

contracting out, charter schools

2. Incentives: merit pay

3. Reorganization: decentralization

4. Performance standards: No Child Left Behind and Race to the Top

416 CHAPTER 14 EDUCATION

closing achievement gaps. There is also concern that although NCLB has focused attention on measuring and improving educational performance for all students, its narrowly defi ned and infl exible parameters for success have created perverse incentives to do well, such as narrowing the cur- riculum, setting artifi cially low standards, and excluding good teachers who do not qualify under such restrictive assessment metrics.27

To address the shortcomings of NCLB, the Obama administration launched the Race to the Top program under the American Recovery and Reinvestment Act of 2009. Race to the Top is a $4.35 billion com- petitive grant program to assist states in implementing innovative edu- cation reform in four areas: upgrading of standards and assessments, improvement of data collection and utilization, enhancement of teacher eff ectiveness and equity in teacher distribution, and turnaround of lowest-performing schools.

INEQUALITY

One of the underlying purposes of public fi nance and provision of educa- tion is to promote equality of opportunity. However, there is a consider- able concern that today the public school system does not come close to living up to that ideal. This is refl ected in recent statistics showing that the United States has become the country with the least (or at least one of the worst) equality of opportunity among the advanced industrial countries— that is, a child’s lifetime prospects are more dependent on the income and education of his or her parents than in other similar countries.28

The major reason for the shortfall is not hard to fi nd. Education has remained primarily a local responsibility. The resources avail- able in some communities, such as wealthy suburban communities or towns in which there is a large factory, are a multiple of those available in others. Demands for public services in general are greater in some communities than in others: urban communities typically require far higher expenditures for maintaining public safety than do rural com- munities. Finally, diff erent communities have expressed diff erent “preferences” for education. Communities with the same resources may spend diff erent amounts.

27 For a rigorous analysis of the academic impact of NCLB, see T. Dee and B. Jacob, “The Impact of No Child Left Behind on Student Achievement,” National Bureau of Economic Research Working Paper 15531, November 2009. For a summary of the NCLB reauthorization policy debate, see R. R. Skinner, “The No Child Left Behind Act: An Overview of Reauthorization Issues for the 111th Congress,” Congressional Research Service Report RL33749, May 2009. 28 See J. E. Stiglitz, The Price of Inequality: How Our Divided Society Endangers Our Future (New York: W. W. Norton & Company, 2012).

417Issues and Controversies in Educational Policy

Whatever the reasons, large diff erences in support for public edu- cation can have signifi cant consequences. Those lucky enough (or rich enough) to live in communities that provide good schools have better prospects than those who live in communities that do not, and because America is becoming more segregated on the basis of income—with rich families increasingly living in communities in which other rich people live—this means that educational outcomes, and lifetime prospects, are increasingly diff erentiated.

The system of local funding has been tested in the nation’s courts. A number of state supreme courts have found that relying on local prop- erty taxes to fi nance education violates provisions of state constitutions ensuring equal access to public education. At the federal level, the U.S. Supreme Court ruled, in San Antonio Independent School District v. Rodri- guez, 411 U.S. 1 (1973), that local funding in Texas did not violate the “equal protection clause” of the U.S. Constitution (the Fourteenth Amendment) even though it resulted in large variability in expenditures.

Such decisions raise several important questions. Should spending be required to be the same in every community within a state? Or should the provision of a minimal level of education in every community be the only requirement? Mandating equality would seem to preclude communities’ spending more on their children. Several states have, in fact, put caps on the levels of expenditures, thereby attaining equality not only by raising the minimum level but also by lowering the maximum. If only a minimum standard is required, how is it to be determined? Obviously, if the stan- dard is set low enough, it will have no eff ect at all. If it is set too high, how will poorer communities fi nd the funds? Higher standards necessitate a greater role for the state in fi nancing education. If equality is insisted upon, what adjustments should be made for diff erences in the costs of education in diff erent communities, and in the nature of the student bod- ies? Is equality of spending enough? Some communities might use more of their funds to develop better athletic facilities, others for basic-skills development, still others for special educational programs. The result is diff erent treatment of similar individuals who happen to reside in diff er- ent communities. Ensuring complete equality would require eliminating community control and establishing a centralized educational system within each state.

At issue are not only the basic trade-off s of equity and effi ciency; most people, to some degree, believe that all parents should have the right to make decisions concerning their children’s education. With local control of education, parents at least feel that they can have some infl uence over the outcomes. Thus, local autonomy of schools has become for many a princi- ple in its own right; some individuals might still favor local control even if

418 CHAPTER 14 EDUCATION

it could be shown that central control was both more equitable and more effi cient. To others, the trade-off at issue is between the rights of parents (to decide about their children’s education) and the rights of children (to equality of opportunity, regardless of who their parents are).

Both state and federal governments have responded to these concerns by attempting to increase equality while retaining at least some degree of local autonomy, especially by providing funds to communities with low incomes and with many disadvantaged children. New Jersey, for instance, has broadened its role in fi nancing edu- cation, with poorer districts getting much more aid than better-off districts; set minimal educa- tional standards that all districts must attain;

and, in some cases, placed ceilings on the amount that higher-spending school districts can spend.

LIMITATIONS ON EQUALITY IMPOSED BY PARENTAL CHOICE Some attempts to attain greater equality, such as the ceilings on expendi- tures by richer school districts, have been criticized as misguided and self- defeating, as the underlying problem is not just the degree of inequality within public schools but also the total extent of inequality in our society. As long as the government is not willing to prohibit individuals from going to private schools, any attempt to introduce too much equality into the public educational system will result in individuals’ transferring to private schools.

In England, there have been periodic proposals for the government to actively discourage private schools on the grounds that private educa- tion leads to social stratifi cation (only the upper and upper-middle classes send their children to private schools). In the United States, however, restrictions on private schools might well be unconstitutional. In any case, although there has been controversy about whether private schools should receive public support, no one has suggested that private schools be actively discouraged.

AID TO HIGHER EDUCATION

Interestingly, many of the controversies that plague elementary and sec- ondary education have been resolved at the higher-education level. There has even been a voucher plan: the GI Bill of Rights, which spurred the

APPROACHES TO ENHANCING EQUALITY • Provide minimum level of expenditures for

all districts.

• Cap expenditures of high-spending districts.

• Federal and state fi nancial assistance to low income communities and communities with large numbers of disadvantaged children.

• Key issue: how to maintain local responsibility and autonomy while providing equality of opportunity.

419Aid to Higher Education

growth of higher education in the United States after World War II, in eff ect gave veterans a voucher they could use at any university, public or private. Private schools, as we have already noted, play a large role at the college and university level, and there is a high level of decentralization.

The federal role dates from the setting aside of land for a university in the Land Ordinance of 1785, to the founding of the land grant universities in 1862, to the extensive support of research universities through grants and contracts in the post–World War II era.

A fundamental diff erence between higher education and elementary and secondary education is that the students have reached an age at which they can make decisions for themselves. It is the students who should judge whether the returns to further education warrant further invest- ment. At least that would be the case if individuals had good information, and had the skills required to assess whatever information is presented. However, many private for-profi t schools do not provide the kind of infor- mation that individuals would need to make an informed decision—and it is precisely the skills in assessing such information that a college educa- tion is intended to hone. In many other areas, we have learned that there is a need for consumer protection, and this seems one for which the need, given the evidence of market exploitation, is especially strong.

Beyond this, government has a key role to ensure access, so students have the fi nancial resources to go on to college. Currently, this is accom- plished in fi ve ways. First, states greatly subsidize higher education, typi- cally charging tuitions in their public colleges and universities that are a fraction of the total costs. Second, the federal government provides grants, such as Pell and TEACH grants, for students who meet eligibility criteria based on fi nancial need. Third, the federal government sponsors loan pro- grams for higher education, such as Perkins and Ford loans. Fourth, it supports a limited work-study program, called the Federal Work-Study Program. Fifth, it provides tuition tax deductions and credits.29

The state subsidies have been widely criticized as being untargeted (or  mistargeted). Indeed, because enrollment rates typically are higher among the children of the affl uent, on average, it is the more affl uent who benefi t from such subsidies. Moreover, to the extent that the subsidies can be thought of as directed at the children themselves, rather than at the parents, lifetime incomes, even of children from poorer families, will be higher than the average income of the population. Therefore, critics say, government should focus its attention on loans or grants to children from poor families.

Federal tuition tax credits were introduced in 1997 (eff ective in 1998). There are now two federal tax credits to help pay higher education costs:

29 For a summary of federal student aid programs, see U.S. Department of Education, Funding Your Education: The Guide to Federal Student Aid, 2012–13.

420 CHAPTER 14 EDUCATION

the American Opportunity Credit, up to $2500 per eligible student annu- ally for the fi rst four years of an undergraduate degree program, and the Lifetime Learning Credit, up to $2000 per eligible student annually for all years of post-secondary education, including nondegree and skills improvement courses. These, too, have been criticized as untargeted: the money will not go to the very poor (although 40 percent of the Ameri- can Opportunity Credit is refundable, so even those who do not pay taxes can receive up to $1000). Critics worry that the tuition tax credits may simply induce colleges to raise tuitions, especially in states that charge very low tuition for their community colleges.30 They worry that enroll- ment rates of the very poor may even go down, as poor people who cannot take advantage of the tuition tax credit fi nd college less aff ordable, unless states increase scholarship funds to off set any tuition increases. (Propo- nents argue that even if tuitions increase, the increased availability of funds for education will be helpful.)

Critics of the tax deduction and credit program argued that it would have cost far less and been far more eff ective in increasing enrollment rates, particularly among the disadvantaged, if the government had announced a broader policy of guaranteed fi nancial access, a combination of loans, grants, and work to ensure all Americans access to higher education.

Federally guaranteed loans were fi rst authorized as part of the Higher Education Act of 1965. Although today there is little controversy about the value of student loans, there has been recent debate over how best to provide them. The federal government now has two major loan pro- grams: Federal Perkins Loans, in which private institutions lend to stu- dents while the government bears the risk of default; and William D. Ford Federal Direct Loans, in which the federal government lends directly to students. The latter began in 1994 under the Clinton administration, which claimed that the federal government could administer the pro- gram more eff ectively and at lower cost than private sector institutions— a claim that the subsequent years bore out. Part of the rationale for the program is that the division of risk bearing and administration under the existing program gave rise to insuffi cient incentives for screening and debt enforcement (collection) on the part of private institutions. Equally troublesome, the private fi nancial sector was charging students as if they were bearing the risks, when they were not. It became a lucrative source of fi nance—so lucrative that there were scandals involving bribes to col- leges to use particular lenders. Moreover, the government would be able to introduce more fl exible loan instruments, which indeed it has done,

30 The credit is 100 percent of the fi rst $1000 in tuition and 50 percent of the second $1000. Raising tui- tion from $500 to $1000 will thus cost the student nothing—what the student pays will be unchanged, although what the federal government pays will increase substantially.

421Aid to Higher Education

including repayment plans that are graduated (stepwise increase in pay- ment size over the loan term), extended (up to 25 years rather than the standard 10 years), income-based (maximum monthly payment is 15 per- cent of discretionary income), pay as you earn (maximum monthly pay- ment is 10 percent of discretionary income), income contingent (payments based on a capacity-to-pay formula), and income sensitive (payments based on annual income). Borrowers who work in the public sector for a decade might also be eligible for the Public Service Loan Forgiveness Program after making 120 full, on-time loan payments.31 Since 1994, the federal government has become the major lender, issuing 93 percent of all student loans in academic year 2011–2012.

As the cost of higher education continues to increase while both state aid for education and median household income continue to decline, stu- dent loan debt has grown dramatically. This trend has persisted through- out the Great Recession, despite a substantial reduction in other types of consumer debt; at the end of 2012 it was approximately $1 trillion, far exceeding total household credit card debt. Not only is the magnitude of student loan debt worrisome, but the rise in arrears is also troublesome—at over 10 percent, student loan balances 90 or more days delinquent for the fi rst time exceed the “serious delinquency” rate for credit card debt. These problems are exacerbated by U.S. bankruptcy laws. Unlike most other types of consumer debt, student loans are extremely hard to discharge.

With incomes of most Americans stagnating and tuition costs soaring, America faces a bind: if current trends continue, young Americans (and their families) will face increasing debts to fi nance their education. This could have severe macroeconomic consequences, just as other aspects of household debt did, in the run up to the global fi nancial crisis. However, it will also mean that many Americans will decide that it is just too risky and will not get the education they need to be competitive in the twenty- fi rst century. This will obviously harm their own economic prospects, but if (as seems likely) enough people make that decision, it could aff ect America’s overall economic prospects.

One way out of the dilemma is further government support to higher education, which would enable a lowering of tuition, or at least grants to more individuals so they could take on less debt. With budget cutbacks at both the state and national levels, however, the prospects of this are bleak.

Since the late 1980s, several countries have adopted income- contingent loans (ICLs) as a possible solution; the United States intro- duced ICLs in 1994. The key distinguishing characteristic of ICLs is that collection of the debt depends on the borrower’s capacity to pay.

31 For a detailed explanation of these federal student loan products, see http://studentaid.ed.gov/ repay-loans/understand/plans.

422 CHAPTER 14 EDUCATION

For example, in 1989 Australia introduced a risk-sharing ICL, the Higher Education Contribution Scheme (HECS). Under this program, student debt is repaid through the tax system, with the amount due dependent on personal income: there is no repayment obligation unless income exceeds a certain threshold, and then the repayment amount increases as income rises.32 The program not only reduces the risks associated with invest- ing in education, but it also signifi cantly lowers transactions costs. It has been an important ingredient in Australia’s successful attempt to extend access to a college education: the percentage of Australia’s population 25 to 64 years old who have graduated from college rose from 18 percent in 1999 to 27 percent in 2010. This brought Australia closer to the U.S. graduation rate. Although the United States started from a better position of 28 percent in 1999, the U.S. graduation rate did not grow as quickly, reaching 32 percent in 2010.33

REVIEW AND PRACTICE

SUMMARY

1. The past fi fty years have seen changes in the structure of education in the United States, including an increase in the fraction of funds provided by states and the federal government. Although direct federal expenditures for educa- tion are low relative to state and local spending, federal tax expenditures provide large implicit subsidies. In addition, federal involvement in guiding education policy is growing.

2. Education is not a pure public good. Important externalities are associated with education, but these externalities do not provide a fully

32 For further information about Race to the Top, see the U.S. Department of Education website www .ed.gov. For more data on student loan debt, see the Federal Reserve Bank of New York’s quarterly reports on household debt and credit. For a detailed explanation of ICLs, see B. Chapman, “Income Con- tingent Loans for Higher Education: International Reforms,” Chapter 25 in Handbook of the Economics of Education, vol. 2, ed. E. Hanushek and F. Welch (Amsterdam: Elsevier, 2006), pp. 1435–1503. HELP is the Australian government loan program to assist students fi nance their tuition fees (FEE-HELP or VET FEE-HELP), student contributions (HECS-HELP), and overseas study expenses (OS-HELP). In 2012–2013, the HELP repayment threshold was A$49,096, above which the repayment rate rose in 0.5 percent increments from 4 to 8 percent of HRI. More information about HELP can be found at http://studyassist.gov.au and www.ato.gov.au. 33 Organization for Economic Cooperation and Development, Education at a Glance, 2012.

persuasive justifi cation for the role of the govern- ment. The major justifi cation for public support of elementary and secondary education is the belief that the quality of education obtained should not depend solely on the resources of the child’s par- ents. There are important market failures, includ- ing imperfections of capital and risk markets.

3. There are concerns about both the effi ciency and equity of the U.S. educational system. Propos- als to improve the effi ciency of the system are (a) a voucher program to enable more choice and promote competition; (b) decentralization reforms, to enable teachers and parents to exert more control over schools; (c) charter schools to provide more

423Review and Practice

scope for innovation; (d) contracting out, to provide more competition for school management, without the disadvantage of social stratifi cation that may arise from voucher plans; and (e) national stand- ards with clearer goals and performance meas- ures, both for schools and teachers. Most of these “experiments” have had, at best, mixed results.

4. The heavy reliance on local fi nance for schools results in a great deal of inequality in expendi- tures. Reforms include providing more state aid.

5. Although education is not the only determinant of an individual’s future wages, there is a systematic correlation between the level of education and wages. There is controversy, however, concerning the explanation of this correlation. Some claim that it is primarily due to the increased skills that children obtain at school (the human capital view), whereas others claim that it is due to the schools’ identifying the very able and diff erentiat- ing them from the less able (the screening view).

6. The poor performance of American students on standardized international exams in spite of heavy expenditures is related in part to the high level of inequality in the United States. In addi- tion, large amounts of expenditures go to admin- istration and special programs. The overall link between expenditures and performance has been hotly debated.

7. The government has long played an active role in higher education, although its dominance is not as great as at the elementary and secondary school levels. Some believe that the eff ects of government aid to higher education are regressive, as those who benefi t from college are likely to have higher incomes. They believe that direct subsidies should be replaced by loan programs. With tuitions soar- ing and incomes of most Americans stagnating, however, student debt is already soaring. Australia has pioneered an alternative with low transactions costs called income-contingent loans.

8. The federal government has long played an important role in supporting higher educa- tion. Student loans and federal grant programs have increased accessibility to higher educa- tion, but stagnating incomes in the middle and

rising tuitions now threaten that accessibility. There is considerable controversy over the extent to which tuition tax credits will increase enroll- ments. Most agree, however, that such tax credits are not well targeted.

KEY CONCEPTS

Charter schools

Compensatory education

Contracting out

Human capital

Income-contingent loans (ICLs)

School vouchers

Screening

QUESTIONS AND PROBLEMS

1. a.  Discuss the equity and effi ciency arguments for raising tuition at state universities. To what extent do your answers depend on whether there is a good college loan program available?

b. Discuss the equity and effi ciency arguments for providing college loans at subsidized inter- est rates.

2. List some characteristics of our educational institutions or outcomes of our educational sys- tem that seem to be more consistent with the screening view of education than with the human capital view.

3. The property tax base per student is often as high in industrial centers as it is in the suburbs. Why might you still expect expenditure per pupil to be lower in industrial centers than in the suburbs?

4. Discuss the trade-off s involved in deciding on the appropriate level and form of decentralization/ centralization within education. Bear in mind that diff erent aspects—fi nance, control of curric- ulum, control of hiring—can be decentralized to diff erent extents, and that the issues of centrali- zation and decentralization relate not only to the division of responsibility among the federal gov- ernment, the state, and the local community, but

424 CHAPTER 14 EDUCATION

also to the division of responsibility within the school district, among central offi ce administra- tors, school principals, and teachers.

5. Discuss the trade-off s between parental choice and equality of opportunity. To what extent should the principle of consumer sovereignty extend to parental rights to choose the amount and form of education for their children?

6. Provide an economic analysis of the issue of tracking (putting students of similar abilities in the same classes). What evidence would you like to have to decide whether tracking is desirable? What are the trade-off s? (How does your answer depend on whether teachers fi nd it easier to teach classes in which students have relatively sim- ilar abilities, and the extent to which less able students learn from having more able students inside the classroom?)

7. A community is considering how to allocate expenditures between education and other goods. Draw the budget constraint, putting “edu- cation” on the horizontal axis, and “other goods” on the vertical.

a. Contrast the budget constraint of a poor com- munity with that of a richer community. Using indiff erence curves, explain why a poorer community is likely to spend less on education than a richer community.

b. Show how the deductibility of property taxes, used to fi nance public education, aff ects the budget constraints of the two communities. How does it aff ect choices? If the price elastic- ity of the demand for education is unity, what should be the eff ect on the demand for educa- tion in a community in which the median voter is in the 33 percent marginal tax bracket? Or in a 15 percentage tax bracket?

c. Assume now that the poor community is sub- sidized (and the rich community taxed) so that the extra educational expenditures it gets from a 1 percent tax is the same as that of a

rich community. Draw the budget constraints. Will the poorer community continue to spend less on education than the richer community? On what does your answer depend?

d. Whose indiff erence curve is relevant for the analysis? (Recall the discussion of median vot- ers from Chapter 9.)

8. List the various concerns about inadequacies of the U.S. educational system. Evaluate various reform initiatives in terms of the extent to which they address these concerns.

a. If you thought that the primary problem was education of the disadvantaged, what reforms would you emphasize?

b. If you thought that the primary problem was the lack of an adequate supply of highly trained scientists and engineers, what reforms would you emphasize?

c. If you thought that the primary problem was inadequate performance of elementary and secondary schools, which reforms would you emphasize?

d. If you thought that the primary problem was the lack of enrollment in universities, what reforms would you emphasize?

9. Assume that spending more on the education of a student increases his or her productivity. Draw the relationship between productivity/wages (assum- ing that wages increase in tandem with produc- tivity and that there are diminishing returns). Assume that more able students have, for each level of education, a higher level of productivity.

a. Assume that one wanted to maximize total output for a given level of expenditure. How would you allocate educational expenditures? Would you necessarily spend more on the more able students?

b. Assume that you wanted to reduce inequalities in income. How would this aff ect patterns of educational expenditure?

425

APPENDIX: HOW SHOULD PUBLIC EDUCATIONAL FUNDS BE ALLOCATED?

Every school district faces the problem of allocating its educational budget. It can allocate more funds for special education, for remedial classes for the disadvantaged, or for accelerated classes for the gifted. As more funds are allocated to any one individual, there is some increase in that individual’s productivity. This is the return to education.

If we wished to maximize national output, and if effi ciency alone were our goal, we would allocate funds so the increase in productivity from spending an extra dollar on one individual would be the same as the increase in productivity from spending an extra dollar on another. If very able individuals not only reach a higher level of productivity than others at each level of education but also benefi t more from education, so the mar- ginal return to education is higher, such a policy entails spending a greater amount on the education of the able than on that of the less able. Some would say this is unfair; the government should ensure equality of expen- ditures in public education. However, when educational expenditures are equalized, those who are more able—or who have home backgrounds that give them an advantage—will still be better off . Accordingly, some believe that government should engage in compensatory education: it should attempt to equalize not input (expenditure), but output (achievement). It should attempt to compensate for the background disadvantages facing some groups in our society. One of the major federal programs is directed specifi cally at encouraging local communities to provide such compensa- tory education.

As more and more funds (of a fi xed budget) are allocated to the less able, and less and less to the more able, total output falls, as the marginal return to education (under our assumption) for the less able is smaller than for the more able. Under these assumptions, there is a trade-off between effi ciency and equality, as depicted in Figure 14.4A. What point one chooses on this locus depends on one’s values, on how one is willing to trade off effi ciency versus equality.

Some maintain, however, the trade-off curve does not look like Figure 14.4A but like Figure 14.4B—that is, some movement toward com- pensatory education may actually increase national output. In this view, those who are advantaged have a higher output than the disadvantaged at each educational level, but the marginal return to further education for the more able is actually lower than for the less advantaged. This implies that

426 CHAPTER 14 EDUCATION

FIGURE 14.4

Compensatory education— equalized productivity

Output maximum

Equal expenditure

Efficiency (e.g., national output)

AEquality (measured

by some inequality

index) EQUITY–EFFICIENCY

TRADE-OFFS IN EDUCATION EXPENDITURES

(A) Poorer individuals with poorer home backgrounds have, at any level of educa-

tional expenditures, a lower marginal return to educa-

tion. Accordingly, maximizing output requires spending

more on richer individuals. By spending an equal amount,

output is reduced but equal- ity is increased. Compensa- tory education—equalizing

productivity—requires spending even more on poorer

individuals, reducing output but increasing equality. (B) Here, poorer individuals may have

a lower level of income at any given level of educational

expenditures, but a higher marginal return. Then output

maximization (equalizing marginal returns) requires spending more on poorer individuals. Spending an

equal amount on each will then lower equality and effi ciency. Compensatory education—

equalized productivity

Output maximum

Equal expenditure

Efficiency

Equality B

427Appendix: How Should Public Educational Funds Be Allocated?

we can get both more effi ciency (higher output) and more equality with at least some degree of compensatory education.

The diff erences in the education–productivity relationship between one individual and another may result either from diff erences in innate ability or from diff erences in environment (home background). There is a long-standing controversy about the relative contribution of these two factors in explaining performance. In the case of two individuals with the same innate ability but diff erent home backgrounds, the nature of the education–productivity relationships may depend on whether edu- cation in the home (home background) is a substitute for or a comple- ment to schooling. If home background is a complement to schooling, it means that it increases the return to education. If it is a substitute, the more education that occurs in the home, the smaller the return to formal education.

428

WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

All societies have made some provision for the poor and destitute. We call programs that transfer cash and consumption goods to the poor public assistance or welfare programs. The manner in which the very poor are taken care of has changed dramatically over time. In medieval Europe the church took responsibility, often establishing almshouses. In modern soci- eties, governments play a major role. In this century, there have been two dramatic changes in the manner in which government fulfi lls that role. The fi rst was in 1935, when the federal government fi rst took on a major respon- sibility for welfare under the Aid to Families with Dependent Children (AFDC) program. The second was on August 22, 1996, when President Clinton signed a historic bill, the Personal Responsibility and Work Oppor- tunity Reconciliation Act of 1996, which even changed the name of the assistance program from AFDC to Temporary Assistance for Needy Families (TANF). This act culminated a movement for reforming the wel- fare system that had been the subject of increasing national attention.

In his presidential campaign four years earlier, Clinton had accom- panied his promise to “end welfare as we know it” with the slogans “a hand up, not a handout” and “making work pay.” Welfare, it was felt, had

15

429Welfare Programs and the Redistribution of Income

created a kind of dependency; there was widespread agreement that the welfare system had to be restructured to help those on welfare get off welfare and become productive members of the labor force. This required economic incentives—it had to pay to work rather than to be on welfare. But it was felt that the carrot of increased income would not suffi ce: there had to be a stick—the termination of welfare payments after a fi xed period of time (two to fi ve years). However, that carrot and stick might not be enough; there needed to be complementary systems and training as well if people were to move from welfare to work.

The antipathy to welfare was partly based on its seeming failure, but it was also largely driven by misperceptions. The federal defi cit had become a major source of concern in the 1990s. There was a widespread impres- sion that welfare was largely responsible, and that many on welfare were taking the system for a free ride. In fact, however, total welfare expendi- tures have never amounted to a large fraction of total government expen- ditures. In 1996 total welfare expenditures were less than 10 percent of total federal government outlays, and welfare expenditures excluding Medicaid were only 4 percent of total outlays. Even complete elimination of welfare expenditures—a far more Draconian measure than anyone was proposing—would not have eliminated the defi cit. Indeed, the magnitude of the spending cuts enacted under the new legislation did not signifi - cantly alter the status of the federal defi cit.

Ironically, shortly after the passage of the bill, not only did the fi scal defi cit cease to be an immediate problem—the federal government turned a large surplus for the fi rst time in three decades—but welfare rolls also declined precipitously, partly because of the booming economy that had brought overall unemployment rates down to 4.3 percent, a level that also had not been seen in three decades.

Critics of these reforms worried, however, what would happen when the country went into a downturn and jobs became scarce. Would the new welfare system work, especially to protect children? The Great Recession that began in 2008 provides telling evidence of defi ciencies, with child poverty in particular increasing—children under 18 years living in poor families rose from 18 to 22 percent from 2006 to 2010. Alternative meas- ures of poverty that include some assessment of the value of in-kind ben- efi ts show a somewhat less dramatic increase in overall poverty, but still very high levels of childhood poverty.

This chapter provides a brief review of the history of U.S. welfare poli- cies and programs, a look at the United States in comparative perspective, a summary of the major analytic issues, and a discussion of the 1996 wel- fare bill, explaining why it was so controversial and describing ongoing eff orts at further reforms.

1. What are the major welfare programs? How have they grown over time? To what extent did the growth of welfare programs account for the growth in the defi cit in the 1980s? How much of the budget do they con- sume today?

2. What are the dimensions of the poverty problem in the United States? How has it changed over time?

3. What is the eff ect of welfare programs on labor supply? What other distortions are associated with welfare programs?

4. In what forms are welfare benefi ts received? What is the distinction between categorical and broad- based assistance? What are the effi ciency and equity issues associated with each of these forms of assistance?

5. What has been the impe- tus for welfare reform? What have been the major reform proposals? What were the major compo- nents of the 1996 welfare reform bill, and what have been the consequences?

FOCUS QUESTIONS

430 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

A BRIEF DESCRIPTION OF MAJOR U.S. WELFARE PROGRAMS

In the United States, although states and localities had long provided some form of general assistance to the needy, supplementing church and other voluntary programs, the federal government took on major responsibility with the New Deal in the 1930s. The Social Security Act of 1935 established Aid to Families with Dependent Children to provide assistance to families without a major breadwinner and Supplemental Security Income (SSI) to provide funds to aged and disabled individuals with low incomes (supplementing Social Security payments).

The next major expansion of federal welfare programs occurred when President Johnson declared a “War on Poverty” in the 1960s. As part of that war, a number of programs were introduced, including Medicaid, to provide medical assistance; Medicaid has subsequently become the larg- est assistance program in dollar terms. The following sections describe briefl y the major programs.

AFDC AND TANF

Since 1935, AFDC had been the primary cash program in the U.S. welfare system. The program was a combination of federal and state programs. The states not only administered AFDC, but also set benefi t levels and had some discretion over rules. The federal government provided a portion of the funds, which varied from approximately one-half to three-fourths, depending on the state’s per capita income. Programs in which federal outlays depend on state expenditures are called matching programs. The federal matching subsidy presumably resulted in the states’ provid- ing higher levels of benefi ts than they would have if they had had to pay the full (marginal) costs themselves.

Starting in 1997, TANF replaced AFDC. TANF represented a marked departure from the earlier system in two ways. First, it replaced the old sys- tem of matching grants with block grants—a fi xed amount of money—with states given considerable discretion as to how that money could be spent, including discretion in determining the eligibility of needy families and the benefi ts and services those families receive. Second, TANF focused on mov- ing individuals from welfare to work. The states were given broad fl exibility in the design and operation of their welfare-to-work programs, but the use of TANF funds had to be consistent with federal priorities of strong work requirements, time limits to receiving assistance, a reduction in welfare

431A Brief Description of Major U.S. Welfare Programs

dependency, and the encouragement of two-parent families. Welfare expen- ditures under TANF have been rising only slightly, and TANF has indeed succeeded in substantially reducing the number of people on welfare. How- ever, TANF has not had a signifi cant impact on reducing poverty levels, and during the 2008 recession, TANF did not perform its countercyclical role of safety net during hard times. Food stamps/SNAP (see below) have only par- tially fi lled the gap, with approximately one out of seven Americans receiv- ing these benefi ts, but with an increasing focus on the federal defi cit, many in Congress would like to dramatically downsize this program.

Under both AFDC and TANF, there has been considerable variation in the amount of benefi ts provided by the states. For example, in 2009 the highest benefi ts per resident, in New York, were more than ten times the lowest level of benefi ts, in Idaho. Expenditures have varied greatly over time as well. Total expenditures (in constant 2009 dollars) increased from $35 billion in 1987 to $44 billion in 1995, fell to $30 billion in 2006 in the aftermath of the Clinton welfare reform, and then rose again to $34 billion in 2009 in response to the economic crisis. The total number of benefi ciar- ies has also fl uctuated widely, with recipients of cash assistance more than doubling, from 6.7 million in 1969 to 14.2 million in 1993, then dropping 72 percent to 4.0 million in 2008, before climbing to 5.0 million in 2010.1

Over time, the details of the federal program have varied. Any means- tested program of assistance—that is, any program whose benefi ts are targeted to those with low incomes—must reduce benefi ts as income rises. Before 1979, welfare recipients’ benefi ts were cut $67 for every increase of $100 in earned income beyond a certain minimal amount, an eff ec- tive marginal tax rate of 67 percent. After 1979, the eff ective tax rate was raised to 100 percent, although some allowance was made for child care expenses and the adverse eff ects were somewhat mitigated by the Earned Income Tax Credit (see next section).

EARNED INCOME TAX CREDIT

The Earned Income Tax Credit (EITC) program supplements the income of low-income families with children2 by an amount that depends

1�U.S. Census Bureau, Statistical Abstract of the United States, 2010, Table 12; and U.S. House of Rep- resentatives, Committee on Ways and Means, Background Material and Data on Programs within the Jurisdiction of the Committee on Ways and Means: 2011 Green Book. The latter reference is a report pub- lished every few years by the House Ways and Means Committee, which provides the most complete compilation on the welfare programs. It is generally referred to as simply the Green Book (refl ecting the color of its cover). Much of the data for this chapter are drawn from the 2011 report, which will be referred to simply as the 2011 Green Book in the following references. 2�The 1993 law added a small credit for childless workers. The maximum amount was $464 in 2011, but all EITC fi gures are indexed for infl ation, so currently, the maximum credit is somewhat greater.

432 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

on their income and number of children. Although this amount was small in the 1980s, and was intended to off set Social Security contribu- tions, in 1993, the EITC was greatly expanded to encourage people to work by ensuring that their after-tax earned income would be greater than their untaxed income from welfare payments. By 2011, a worker with three dependents could have his or her income supplemented by as much as 45 percent. Thus, a worker receiving $5 an hour has his or her wage increased, in eff ect, to $7.25 an hour. The EITC gradually increases as earned income rises, plateaus at the maximum benefi t level ($5751 in 2011), and then is phased out as income continues to rise.

Under the old AFDC program, EITC benefi ts were not counted as income for the purposes of determining benefi t levels, so individu- als were still better off as a result of working. For example, if a worker receiving the maximum benefi t at the time earned an extra $100 of income, although his or her AFDC benefi ts would go down by approxi- mately $100, the EITC payment would go up by $40, so he or she faced an eff ective tax rate of “only” 60 percent. TANF has no provision regarding treatment of earned and unearned income. States set their own income limits and make their own rules governing the treatment of earnings and other income, so the interaction between TANF and EITC may vary from state to state.

The number of families receiving EITC benefi ts grew from 6.2 million at its inception in 1975 to 26.5 million in 2009; the average credit per family grew from $200 to $2200, and the total expenditures grew from $1.25 billion to $58.8 billion.

FOOD STAMPS/SNAP

TANF is the principal cash welfare program. Most assistance to the poor is directed at in-kind benefi ts like medical care, or to fi nance par- ticular categories of expenditure like food and energy.

Food stamps, fi rst introduced in 1964 on a nationwide basis as part of President Johnson’s War on Poverty, are designed to assist poor indi- viduals in buying food. The federal government bears essentially all the costs, and sets uniform benefi t levels. The benefi ts depend on a measure of income, which allows a variety of adjustments, the most important of which is that housing expenditures are deductible.

The Food Stamps Program (FSP) has been modifi ed considerably over the past fi fty years, mostly regarding eligibility requirements, benefi t levels, and administrative procedures. The latest round of reforms was enacted under the 2008 farm bill (formally, the Food,

433A Brief Description of Major U.S. Welfare Programs

Conservation, and Energy Act of 2008). These included changing the name to the Supplemental Nutrition Assistance Program (SNAP) in an eff ort to counter the stigma associated with food stamps. After two decades of pilot projects and phased implementation, the 2008 farm bill also made electronic benefi t transfer (EBT) the program’s standard issuance vehicle to improve effi ciency and reduce fraud. Paper coupons were replaced by benefi t cards—SNAP benefi ts are deposited electroni- cally into the recipient’s account every month; the recipient has access to this account when paying for food, using the benefi t card like a bank debit card.

Participation in FSP/SNAP peaked at more than 27 million per month in 1994, before gradually dropping to about 17 million in 2000 after the 1996 welfare reform bill limited the benefi ts of working-age adults with- out children. Such individuals cannot receive SNAP benefi ts for more than three months in a thirty-six-month period if they have not worked twenty hours a week, completed a job training program, or participated in a workfare program. (Workfare programs provide assistance in exchange for work.) However, participation rates and program costs have risen steadily since then as a result of a combination of outreach eff orts and poor economic conditions, reaching almost 45 million recipients per month and $72 billion in benefi ts in 2011.3

SNAP is the most important “food” program, but two others should be noted. Thirty-two million children receive free or subsidized lunches, and twelve million children receive free or subsidized breakfasts. These programs were started when, in the nationwide draft for World War II, the extent of malnutrition in poor children was fi rst widely recognized; moreover, educators have long argued that inadequate nutrition adversely aff ected learning.

An additional nine million pregnant women, infants, and children under age 5 who are at risk of inadequate nutrition receive food sup- port under the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). Inadequate nutrition at these points can have lifelong eff ects, particularly on a child’s ability to learn. While the other programs described so far are entitlement programs—that is, anyone meeting the criteria is entitled to receive the benefi ts; thus, total federal expenditures depend simply on the number of individuals who are eligible and who decide to apply—the federal government provides a fi xed amount of money to support the WIC program.

3�The SNAP gross monthly income eligibility limit is set at 130 percent of the poverty threshold (poverty line), a measure of the minimum subsistence level required for, say, a family of four. In 2011, the amount of the threshold for a family of four was $2422 per month. See the website of the Food and Nutrition Service, United States Department of Agriculture, for a detailed history, full eligibility requirements, and program data for SNAP. http://www.fns.usda.gov/snap.

434 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

MEDICAID

Established in 1966, this program provides medical assistance to the poor, especially poor children; medical care to the disabled; and nursing home care to a large proportion of the aged. Medicaid is a federal-state govern- ment matching program, with the federal government paying between 50 and 83 percent of the costs, depending on the state’s per capita income. The states were given considerable discretion in determining eligibility requirements and coverage. Medicaid, together with the Children’s Health Insurance Program (CHIP), provides coverage for 54 million low-income individuals, including more than 43 million children, com- prising half of all low-income children in the United States.4

Historically, families receiving benefi ts under AFDC were eligible for Medicaid. Although TANF families are not automatically eligible, states are still required to provide Medicaid to children and family members who would have been eligible for AFDC using the program’s terms as of July 16, 1996 (making adjustments for infl ation). Eligibility for Medicaid is thus based on a threshold test: those with incomes above the threshold (essentially, the cutoff level of AFDC) are not eligible. Because many employers do not provide medical benefi ts to low-income workers, many of those on welfare fi nd themselves in a bind: even if they would like to work, they would lose eligibility for Medicaid benefi ts if they accept a job. This is particularly important for those with children requiring medical attention. As a result, they are reluctant to move off welfare. This situation is referred to as welfare lock. States can remove adults from Medicaid rolls if they refuse to comply with the work requirements of TANF.

An important part of President Obama’s attempt to increase access to health care was to expand Medicaid eligibility to adults with income up to 138 percent of the federal poverty level (FPL), extending benefi ts to 21.3 million additional individuals by 2022. Because Medicaid is a joint federal/state program, the new law had strong fi scal incentives for states to participate. In 2012, when the Supreme Court ruled that the individual mandate under the Aff ordable Care Act was constitutional, it also ruled that the act’s provision requiring states to comply with new eligibility cri- teria for Medicaid or risk losing their funding was constitutional only as long as states would lose just new funds for noncompliance, not all their funding (see Chapter 13). As this book goes to press, approximately thirty

4 CHIP was established as part of the Balanced Budget Act of 1997. It was the largest single expansion of health insurance coverage for children in more than thirty years. CHIP, like Medicaid, is jointly funded by the federal government and the states, although the federal matching rate for state CHIP programs is usually about 15 percentage points higher than the Medicaid matching rate. The purpose of CHIP is to provide health coverage for children in families with incomes too high to qualify for Medicaid but too low to aff ord private coverage.

435A Brief Description of Major U.S. Welfare Programs

states have decided to expand Medicaid under the Aff ordable Care Act, with the remaining states declining to increase Medicaid coverage at this time but retaining the option to do so in the future.

HOUSING

In 2008, 5.1 million housing units (providing homes for 9.6 million people) received a total of $35.8 billion in subsidies from numerous Department of Housing and Urban Development (HUD) programs. Forty-four percent of those assisted received help under the Housing Choice Voucher program, allowing them to fi nd and lease housing in the private market using the vouchers. Housing vouchers are like food stamps, they give the recipient a fi xed amount to be spent on hous- ing chosen by the participant. Multifamily assistance, subsidies pro- vided to private landlords for their low-income rental housing, made up one-third of available units. Public housing assistance, whereby local housing authorities receive HUD funding to build, operate, or make improvements to housing owned by local agencies, comprised the remaining one-fourth of available units. The Internal Revenue Service- administered Low-Income Housing Tax Credit (LIHTC), which subsi- dizes the acquisition, construction, or rehabilitation of property rented to low-income households, also supported 1.7 million housing units, at an estimated cost of $5.4 billion in tax expenditures.5

This emphasis on tenant-based and multifamily assistance, which relies on subsidies for housing provided by the private sector, contrasts sharply with the government’s original direct provision of housing for the poor. Public housing was greatly criticized on several grounds. In many cases, costs were high for the quality of housing that was pro- vided—the government seemed an inefficient producer and manager of housing. Worse, government housing projects isolated the poor; they became “warehouses of the poor.” The government was unable to maintain them adequately, and they became infested with crime, drugs, and rats.

By the mid-1990s, the problems had become so manifest that the fed- eral government set about tearing down the worst of these projects, end- ing “public housing as we know it.” In one famous incident, the city of St. Louis leveled its Pruit-Igoe project with dynamite rather than trying to maintain the facility. Thus, much public housing assistance has been

5�Some LIHTC properties also receive HUD subsidies. (Sources: Offi ce of Policy Development and Research, Department of Housing and Urban Development, A Picture of Subsidized Households 2008; and Joint Committee on Taxation, Estimates of Federal Tax Expenditures for Fiscal Years 2008–2012, October 2008, Table 3.)

436 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

replaced by these other programs, designed to provide improved incen- tives and greater social integration. It is worth noting that government assistance for housing the rich and middle class—tax deductibility of mortgage interest—far exceeds government assistance to the poor.

OTHER PROGRAMS

There are many other means-tested programs: the Low-Income Home Energy Assistance Program (LIHEAP); and a variety of education pro- grams such as Head Start, which provides preschool education for chil- dren of low-income families, and Pell grants, which help pay for college education for children in low-income families. Still other programs pro- vide job training to unskilled and economically disadvantaged individuals.

Figure 15.1 shows overall expenditures on means-tested programs. Much of the increase is accounted for by Medicaid, which almost dou- bled, from 27 percent of welfare expenditures in 1975 to 52 percent in 2009. Figure 15.2 compares U.S. spending on means-tested programs with spending by other OECD countries. In contrast to countries like Australia, the United Kingdom, and Canada that have well-developed

UNITED STATES EXPENDITURES ON

MEANS-TESTED PROGRAMS

Overall expenditures on means-tested programs

have quintupled as a share of GDP since 1965.

SOURCE: Offi ce of Management and Budget, Budget of the U.S.

Government, Fiscal Year 2013, Historical Tables, Table 8.4.

FIGURE 15.1 % GDP

0.0 1965 1970 1975 1980 1985 1990 1995 2000 20102005

4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2

Expenditures

437Rationale for Government Welfare Programs

social insurance systems (see Chapter 16), means-tested programs in the United States account for a relatively small share of the total public social expenditures.6

RATIONALE FOR GOVERNMENT WELFARE PROGRAMS

Chapters 3 through 7 set forth the basic theoretical rationale for govern- ment programs: markets may produce effi cient outcomes (ignoring the market failures), but they do not necessarily produce a distribution of income that is socially acceptable. Welfare programs focus on one aspect of the distribution of income: those at the very bottom.

6 The OECD comparison includes general means-tested expenditures such as cash and in-kind pay- ments for income maintenance and social assistance, means-tested support payments to the elderly and disabled, and means-tested spending on the unemployed; it does not include specifi c housing subsi- dies, spending to improve job or earnings prospects (“active labor market programs”), or means-tested medical expenditures. For a detailed explanation of the calculations underlying this comparison, see W. Adema, P. Fron, and M. Ladaique, Is the European Welfare State Really More Expensive? Indicators on Social Spending, 1980–2012; and “Manual to the OECD Social Expenditure Database (SOCX),” OECD Social, Employment and Migration Working Papers No. 124, Paris, 2011.

SOURCE: OECD Social Expenditure Database (SOCX), www.oecd.org/els/ social/expenditure.

Percent of public social

expenditures

0%

5%

10%

15%

20%

25%

30%

35% 34.9%

21.6%

14.5%

3.5%

13.1%

5.0% 3.0%

10.2% 11.7%

4.0%

24.1%

7.5% 10.6%

Au str

ali a

Ch ile

Ca na

da

Fra nc

e Ita

ly

Ge rm

an y

Ja pa

n

M ex

ico Ko

re a

Sw ed

en

Un ite

d K ing

do m

OE CD

A ve

ra ge

Un ite

d S ta

te s

40% OECD EXPENDITURES ON MEANS-TESTED PROGRAMS IN 2007

The United States spends less than the OECD average on means-tested programs, in contrast to other Anglophone countries such as Australia, the United Kingdom, and Canada.

FIGURE 15.2

438 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

Government programs are often thought of as safety net pro- grams. There is a view, for instance, that a civilized society cannot allow individuals to starve, or to die as a result of inadequate health care. These programs are also often thought of as a form of social insurance. Know- ing that adversity can strike anyone, individuals may think, “There but for the grace of God go I.” The knowledge that there is a safety net adds to their sense of economic security, and thus to their overall well-being. The welfare programs are not formally part of the social insurance system, in which individuals pay explicit premiums to obtain, say, unemployment insurance against the risk of losing their jobs, but they still perform this insurance function—they are intended to provide funds to allow individ- uals to get back on their feet. Safety net and social insurance programs are increasingly grouped together under the general category of social protection.

Similarly, programs aimed at children have been justifi ed both as a refl ection of basic values and as an investment in the future. As a basic value, it is argued that all children should, to the extent possible, have the opportunity to live up to their full potential. Children with inadequate nutrition and health care cannot do this. Further, children growing up in poverty are less likely to fi nish high school and more likely to end up in a life of crime: in fi scal year 2009, more than half of all convicted federal off enders had not graduated from high school, in contrast to the 5 percent who were college graduates.7 Thus, the overall returns to society from reducing childhood poverty appear to be large.

DIMENSIONS OF THE PROBLEM

The poverty rate is the fraction of the population whose income lies below a threshold that is intended to measure the minimal level required to maintain a subsistence living standard. The threshold clearly needs to be adjusted each year to refl ect increasing prices (infl ation). Although there is some controversy about the measure (what does it really mean to be in poverty?), movements in the poverty measure do track what is happening at the bottom of the income distribution. The overall pov- erty rate, which was more than 22 percent in 1960, was reduced to a much lower level of about 12 percent in the 1970s. Since then, however, the poverty rate has averaged 13.4 percent, fl uctuating from a low of 11 percent in 2000 to peaks of more than 15 percent in 1983, 1993, and 2010. The primary reason for the increase in poverty since the 1970s has

7 M. Motivans, Bureau of Justice Statistics, U.S. Department of Justice, Federal Justice Statistics 2009— Statistical Tables, Table 4.4.

439Rationale for Government Welfare Programs

been reduction in the real wages of those at the bottom of the income distribution since that time—even those in the middle have not fared well. However, the poverty rate in the United States is an absolute num- ber based on an assesment of a family’s basic needs, whereas many Euro- pean countries base their measure of poverty on relative deprivation, say, half of median income. When poverty is viewed as a relative concept, as inequality increases, typically poverty also increases. Given the large increase in inequality in the United States since 1980, this would mean that, using such measures, the United States would have seen an almost steady increase in poverty rather than the steady average with modest fl uctuations described above.

There is considerable controversy about the factors contributing to the decline in the well-being of those at the bottom. Weaker unions have played a part. So too has globalization—competition, in eff ect, from unskilled, low-paid workers abroad. Those at the bottom are dispropro- rtionately from minority groups, and there has been remarkably little progress in reducing discrimination against these groups. Technology, too, has played a role: new technologies seem to be biased in favor of higher skills, and the quality and quantity of education provided at the bottom has not kept pace. This is why most experts believe that combat- ing poverty requires a comprehensive agenda, of which welfare programs such as TANF and SNAP are an important part, but which also need to include EITC, increase of the minimum wage, programs fi ghting dis- crimination, and provision of better access to high quality education and health care for low-income families.

Particularly disturbing is the rapid rise in the number of children in pov- erty. The poverty rate for children rose from 15 percent in 1970 to 22 percent in 2010. Much of this arose from an increase in the number of children being raised in households with single earners: more than half of all children in poverty resided in households headed by single women in 2010.

The problems of poverty in the United States are seen most strongly among the black population. In 2010, 27.4 percent of blacks lived below the poverty line, almost triple the 9.9 percent poverty rate for non-Hispanic whites.8 High poverty, combined with inadequate access to health, contribute to the high infant mortality and low life expectancy rates for blacks compared to those for whites. In 2007, whites experi- enced an infant mortality rate of 5.6 per 1000 live births, whereas blacks had a rate of 13.2; life expectancy in 2007 was 78.4 years for whites, but only 73.6 for blacks.9

8 U.S. Census Bureau, Historical Poverty Tables, Tables 2, 3, and 10. 9 U.S. Census Bureau, Statistical Abstract of the United States: 2012, Tables 104 and 116.

440 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

ANALYTIC ISSUES

Over the years, policy debates surrounding the welfare programs have focused on a set of incentive and equity issues: Do welfare programs dis- courage work? How can people most eff ectively be moved from welfare to work? Do noncash programs have other adverse incentive eff ects? On what grounds can they be justifi ed? Do welfare programs contribute to the long-run welfare problem by causing dependency and encouraging out-of-wedlock births?

LABOR SUPPLY

Welfare programs use income as a basic criterion for determining eligi- bility. As income rises, benefi ts are reduced; if income rises above a given threshold, a family may become ineligible for Medicaid benefi ts. So individ- uals care about their total income—what they earn plus what they receive from the government. The total net income of poor individuals thus rises far more slowly than their before-subsidy income. It is as if poor individuals face very high marginal tax rates. (Recall that the marginal tax is the extra tax an individual pays as a result of earning an extra dollar of income.) Thus, prior to the 1996 reforms, a very–low-income individual on welfare who earned an extra $100 would have AFDC benefi ts cut by $100, but have EITC increase by $40. The increased EITC payments, however, would reduce food stamps by about $10. Hence, the net income would increase only by $30—in eff ect, there was a 70 percent tax rate. Such high tax rates discour- age work, or at least discourage reporting the income earned from work.

Because the marginal return—the extra net income received from work- ing an extra hour—is reduced, individuals on welfare have less incentive to work. A survey of studies shows a midpoint estimate of work reduction for AFDC and food stamp recipients of 30 percent attributable to AFDC and of 10 percent attributable to food stamps.10,11 The main eff ect seems to be a

10 R. Moffi t, “Incentive Eff ects of the U.S. Welfare System: A Review,” Journal of Economic Literature (March 1992): 1–61. There is a wide range of estimates. See, for instance, J. Hausman, “Labor Supply,” in How Taxes Aff ect Economic Behavior, ed. H. Aaron and J. Pechman (Washington, DC: Brookings Institution, 1981), pp. 27–72; Hausman obtains estimates that are thirty-seven times as large as those obtained by R. Moffi t in “An Economic Model of Welfare Stigma,” Rutgers University mimeo, 1980. For more recent studies, see B. L. Wolfe, “Incentives, Challenges, and Dilemmas of TANF,” Institute for Research on Poverty Discussion Paper No. 1209-00, University of Wisconsin—Madison, May 2000; J. Liebman, “The Optimal Design of the Earned Income Tax Credit,” in Making Work Pay: The Earned Income Tax Credit and Its Impact on American Families, ed. B. D. Meyer and D. Holtz-Eakin (New York: Russell Sage Foundation, 2002), pp. 196–234; and N. Eissa and H. W. Hoynes, “Behavioral Responses to Taxes: Lessons from the EITC and Labor Supply,” in Tax Policy and the Economy, vol. 20, ed. J. M. Poterba (Cambridge, MA: MIT Press, 2006), pp. 73–110. 11� These numbers may exaggerate the eff ects, because there may be a larger reduction in reported income and work than in actual income and work.

441Analytic Issues

reduction in labor force participation (whether the recipient works or not); for those who work, there appears to be little eff ect on hours.12

The EITC was designed to provide greater incentives for individuals to participate in the labor force—it increased the overall return to work. (At the same time, for many individuals, it decreased the marginal return to working an additional hour, as benefi ts were cut as income rose; but the earlier cited studies suggest that this eff ect is not signifi cant.) One of the aims of TANF was to complement the “carrot” that the EITC provided for those with children to participate in the labor force. Eligibility for receiv- ing welfare had strict time limits and work requirements.

Welfare programs, like Medicaid, that employ thresholds may have even more dramatic adverse eff ects on work. The extra return to earning a dollar that pushes a family over the threshold is very negative. Loss of eligibility of medical benefi ts is cited as one of the main impediments to moving people from welfare to work, especially because many employers of low-wage workers do not provide health care benefi ts to their employees.

Although many workers may not work because of the loss of benefi ts, many others work but do not report their income. For instance, a welfare recipient may be employed as a household worker, performing services such as housecleaning or child care, but not report the income. Given the inherent lack of data, economists remain uncertain about the prevalence of this prac- tice, and therefore about the full eff ects on labor supply of welfare programs.

DIAGRAMMATIC EXPOSITION We can use standard budget constraints and indiff erence curves to illustrate the adverse eff ects of welfare on work in the standard economist’s model. In Figure 15.3A, BB gives the before- tax budget constraint, showing how the consumption (after-tax income) of a worker, Alfred, increases as work increases (leisure decreases). The slope of the budget constraint is the wage: if Alfred’s wage is $6 an hour, an extra hour of work increases consumption by $6. The indiff erence curves have the shape illustrated because Alfred does not like to work (or at least does not like the work available to him)—he prefers leisure. Alfred requires extra consumption to compensate him for working more, and because the more he works the less leisure he has, the more valuable leisure is to him at the margin. Furthermore, because the more he works the higher his consumption, the less valuable is his marginal increase in consump- tion. Accordingly, the extra consumption he requires to compensate him for an extra hour of work—the marginal rate of substitution—increases the more he works. That is why the indiff erence curve not only is upward sloping, but becomes steeper as he works more. The original equilibrium

12�R. Moffi t and A. Rangarajan, “The Work Incentives of AFDC Tax Rates: Reconciling Diff erent Estimates,” Journal of Human Resources (Winter 1991): 165–179.

442 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

EFFECT OF WELFARE PROGRAMS ON LABOR SUPPLY 

Welfare programs can adversely affect labor supply. (A) Stylized

version of welfare program before 1979, when payments

were reduced as the individual earned more. Both the income

and substitution effects led to reduced work. (B) Stylized

version of welfare program after 1979, when benefi ts

were reduced dollar for dollar, beyond a certain minimal

amount. No one worked beyond this level. (C) A simplifi ed ver-

sion of the current system, with TANF, EITC, and food stamps.

FIGURE 15.3

Original budget

constraint

$500

New budget

constraint

A

B

C

Work

Consumption

B

B¿ E¿

E

D

B

WorkB

B¿

D B

Work

Consumption

Consumption

O

B–

D F

G H J K

L

443Analytic Issues

is at point E, where the slope of the budget constraint equals the slope of the indiff erence curve: the wage equals the marginal rate of substitution.

B'DB is the new budget constraint under a welfare system in which Alfred gets a fi xed payment—say, $500 per month, which decreases the higher his income becomes. (At point D, the benefi t is completely phased out.) We assume—as was the case prior to 1979—that for every $3 of income Alfred earns, he loses $1 of benefi ts. Thus, the new budget con- straint is fl atter; his after-tax wage is not $6 an hour, but $4 an hour. The new equilibrium is at E', with a lower level of work than before. Work is reduced for two reasons. First, because Alfred is better off , he takes some extra leisure. This is the income eff ect; normally, as incomes rise, individ- uals consume more of any good. Second, because the return for working an extra hour is reduced, Alfred’s incentives for working are reduced; this is the substitution eff ect. Both have the eff ect of reducing work.

Figure 15.3B shows the budget constraint under a welfare system in which Alfred loses a dollar for each dollar he earns beyond a certain mini- mum amount. Not surprisingly, Alfred chooses to work only enough to generate that income; he would be foolish to work more than that, as he earns no marginal return.

Figure 15.3C shows a simplifi ed version of the current system, com- bining TANF, EITC, and food stamps. In this stylized version, there are seven segments to Alfred’s budget constraint. There is a certain basic ben- efi t, given by OB''. There is a certain minimal amount that he can earn without losing benefi ts. Thus, the slope of the budget constraint in the segment B"D is just the wage ($6 an hour), augmented by the EITC, but reduced by Social Security taxes. Beyond this point, benefi ts are reduced the more Alfred works; but the EITC supplements his wage by 40 percent, so his net tax rate is about 70 percent—the return per hour worked is $1.80. Thus, in the interval DF, the budget constraint is relatively fl at. At F, TANF benefi ts are eliminated, but Alfred still has food stamps, pays Social Security taxes, and receives the EITC. Thus, the slope of the budget constraint in the interval FG is much steeper than in DF. Then, however, at an only slightly higher income, Alfred starts paying income taxes. This is refl ected in a fl atter budget constraint in the interval GH. At a slightly higher income, EITC benefi ts start being cut (segment HJ). This greatly reduces the marginal return to working. Soon thereafter, eligibility for food stamps is exhausted. In the interval JK, the slope of the budget constraint is thus slightly steeper. Eventually, EITC benefi ts also are exhausted, and the individual has to pay Social Security and income taxes (segment KL).

From the budget constraint, one can see that, depending on the part of the budget constraint at which an individual is located, there

444 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

can be a variety of eff ects on labor supply. In general, the marginal return to working is less than the wage—in some cases, far less. This is the substitution eff ect; the large substitution eff ect implies that there are large distortions, and that labor supply will be reduced. In addi- tion, for most poor individuals, their after-tax and transfer-payment income is greater  than it would be without taxes and subsidies; the  income eff ect too induces less work. The income and substitution eff ects thus reinforce each other.

The standard economic model often suggests bigger labor supply eff ects than are observed. This is partly because it omits important consid-

erations, both economic and noneconomic. Individuals know that if they are unemployed for an extended period of time, future job prospects will be jeopardized. Thus, after searching for a while, they may accept a job even if it is far below their qualifi cations. Moreover, there is a considera- ble body of research suggesting that the economist’s model of labor supply is, at best, incomplete. In the economist’s model, if one gave an individual the same money that he would have received on the job, but left him with- out employment, he would be better off because consumption would be unchanged and leisure increased. In practice, individuals value mean- ingful employment. It gives them a sense of worth and provides impor- tant source of social connectedness. The loss of welfare especially from extended periods of unemployment is far greater than can be accounted for by the loss of income.

The economic eff ects of limiting unemployment insurance or other sources of income support are also more ambiguous than this standard analysis would suggest. An important aspect of job search is fi nding the right job—a job that uses the individual’s skills as fully as possible. If peo- ple are forced to terminate search prematurely, there will be a less perfect “match,” and economic productivity will suff er as a result.

CASH VERSUS IN-KIND REDISTRIBUTION

More than 70 percent of welfare benefi ts are not unrestricted cash, but are directed to the purchase of food, housing, energy, or medical care. Today, Medicaid and other medical care alone account for more than half of all welfare expenditures. Two dollars out of every ten transferred

INCENTIVE EFFECTS OF WELFARE PROGRAMS 1. The EITC provides positive incentives to partici-

pate in the labor force.

2. Welfare programs provide disincentives to work longer hours; effectively, there is a high marginal tax rate on working longer hours.

3. Benefi ts with thresholds—which suddenly disappear when incomes exceed a certain level— have particularly adverse incentive effects near the cutoff level.

445Analytic Issues

through welfare are provided in the form of subsidized food, housing, and energy. The present system is criticized on three grounds:

1. It introduces ineffi ciencies in resource allocations when there are sub- stitution eff ects; and when there are no substitution eff ects, the conse- quences are not diff erent from those of a direct transfer of income.

2. It is inappropriate for the government to attempt to distort individuals’ consumption decisions; that is to say, in-kind benefi ts are paternalistic.

3. It is administratively costly: each program must be run separately; sev- eral diff erent agencies must determine the eligibility of each individual for each program. (Eligibility standards to determine who is qualifi ed to receive aid under each program are based primarily on income, but adjustments for family size and other circumstances are generally made, and may diff er markedly from program to program.)

The sections that follow discuss the fi rst two criticisms in greater detail. The third criticism applies broadly to most focused welfare pro- grams, whether they are directed at particular groups of individuals (categorical programs) or particular forms of assistance (in-kind ben- efi ts). The  eligibility standards for the various programs could clearly be coordinated and simplifi ed, but there will always remain signifi cant administrative costs associated with such programs. The broader ques- tion, addressed in the following sections, is whether there are benefi ts from such noncash and categorical programs that justify these addi- tional costs.

INEFFICIENCIES FROM IN-KIND BENEFITS

In-kind benefit programs often distort individuals’ choices, typically because they reduce the cost of obtaining the good—and thus induce individuals to consume more of the good than they would otherwise. In some cases, the programs reduce the marginal cost of the good to zero—up to the limit provided by the government. Clearly, if the gov- ernment is giving away a good, individuals will consume it (as long as it has a positive benefit), whether or not the value to them is less than the cost of production. Of course, one of the reasons for in-kind benefits is precisely to promote the consumption of certain goods. Whether government should do that is another matter, to which we shall turn shortly.

446 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

SNAP The SNAP food assistance program provides an excellent illus- tration of how in-kind benefi ts can distort behavior, as we saw earlier in this chapter. Three aspects of behavior are aff ected:

1. Because (under the current system) an individual with a given income receives a fi xed amount of SNAP assistance, in principle SNAP should have the same eff ect on behavior as would a comparable transfer of income, as long as the individual consumes at least as much in food as provided by the SNAP assistance—which is the case for 85 to 90 percent of participat- ing households.13 For these households, it appears that SNAP has a slightly larger eff ect on consumption than one might have expected, with $100 of assistance leading to an increase in food expenditures of between $20 and $45. As one of the intents of food stamps is to improve the nutrition of the poor, advocates of food stamps view this as a marked success.14

2. Because the value of SNAP assistance given decreases as an individu- al’s income increases, SNAP has an adverse eff ect on labor supply.

3. Because the value of SNAP assistance given increases the smaller the individual’s income net of housing expenditures, SNAP encourages housing consumption. It is ironic that whereas SNAP is intended to encourage food consumption, its major eff ects may be to encourage housing consumption and to discourage work.

Although these are theoretical possibilities, there is little empirical evidence suggesting that such eff ects are signifi cant. By contrast, mod- ern behavioral economics suggests that labeling assistance as “food” or “housing” will lead people to consume more food or housing than they otherwise would; the evidence is consistent with that hypothesis.

MEDICAID AND THRESHOLDS All means-tested programs can have adverse eff ects on labor supply, as we have noted, but the eff ects of certain in-kind benefi ts can be particularly dramatic. These are programs under which the government provides a certain free benefi t to those who are

13 J. Currie, “Welfare and the Well-Being of Children: The Relative Eff ectiveness of Cash and In-Kind Transfers,” in Tax Policy and the Economy, vol. 8, ed. J. M. Poterba (Cambridge, MA: MIT Press for the National Bureau of Economic Research, 1994), pp. 1–43. 14 B. Devaney and T. M. Fraker, “The Eff ect of Food Stamps on Food Expenditures,” American Journal of Agricultural Economics (February 1989): 99–104. These results contrast with estimates of 5 to 13 cents additional food expenditures for an additional dollar of cash, reported by Thomas M. Fraker in The Eff ects of Food Stamps on Food Consumption: A Review of the Literature (Alexandria, VA: U.S. Department of Agriculture, Offi ce of Analysis and Evaluation, Food and Nutrition Service, 1990), and are consistent with experimental fi ndings that show that when individuals receive the equivalent in cash instead of food stamps, food consumption actually decreases by between 18 and 28 cents for each dollar of food stamps cashed out. See T. M. Fraker, A. P. Martini, and J. C. Ols, “The Eff ect of Food Stamp Cashout on Food Expenditures: An Assessment of the Findings from Four Demonstrations,” Journal of Human Resources 30, no. 4 (Fall 1995): 633–649; and M. K. Fox, W. Hamilton, and B. Lin, eds., “Literature Review,” Eff ects of Food Assistance and Nutrition Programs on Nutrition and Health, vol. 3 (Washington, DC: U.S.Department of Agriculture, Economic Research Service, Food and Nutrition Research Program, 2004).

447Analytic Issues

eligible—that is, for those whose income is below a certain threshold—and nothing for those above it. The most important program having this feature is Medicaid. There is an obvious administrative reason for such provisions: it makes it much easier to determine whether an individual qualifi es for a pro- gram. An alternative would require the individual to face a fee schedule that depended on his or her income. Now, a doctor simply has to know whether the individual qualifi es for the program. Under the alternative, either the doctor or the government would have to send poor individuals a bill for each service rendered, which would depend on the individual’s income.

Let’s say Figure 15.4 shows the budget constraint for David, who con- sumes a fi xed amount of medical care services. Labor is shown on the hori- zontal axis, and consumption, including Medicaid benefi ts, on the vertical. As David works beyond a certain level, his income, L*, exceeds the thresh- old for Medicaid, so his total consumption actually falls. Clearly, David will work an amount either just short of L* or considerably greater than L*.

HOUSING Public housing programs raise a number of complicated issues associated with in-kind programs. Direct provision increases the quantity of housing supplied at any price, and thus benefi ts not only those who receive public housing, but also others, as it drives down the equilib- rium rent. On the other hand, government has proven itself an ineffi cient producer of housing—with the costs per square foot considerably higher than in comparable privately provided housing. That is one of the motiva- tions for vouchers and tax credits.

EFFECT OF THRESHOLDS ON LABOR SUPPLY

Programs like Medicaid, where there is a given level of income above which individuals lose eligibility, discourage work.

FIGURE 15.4

Labor

Total consumption

L*

448 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

Tax credits also operate directly on the supply side. Those who pro- duce low-income housing receive a subsidy through the tax system. More housing is supplied at each price, lowering the equilibrium price both for those living in subsidized housing and for others. Programs that require housing receiving such special tax treatment to be mixed-income—that is, that the apartment buildings not just have low-income units—have reduced the isolation of the poor. However, these programs have also been criticized: much of the tax benefi t goes to the brokers who package the deals. The “bang for the buck” may be relatively low.

Housing advocates, accordingly, have argued that money should be given directly to the poor, rather than to real estate developers. This is what housing vouchers do: they subsidize the rent that individuals pay. Figure 15.5 shows vouchers as shifting the demand curve for housing to the right. In the short run, when the supply is relatively inelastic (Figure 15.5A), the main eff ect may be on equilibrium rents—then those not receiving the vouchers are actually made worse off . In the long run, when the supply is relatively elastic (Figure 15.5B), supply responds—just as it would if the gov- ernment had directly provided the housing itself. Still, unless the long-run supply curve is horizontal, those not receiving the subsidy are hurt.

One particularly troublesome feature of the design of U.S. housing pro- grams arises from the fact that there is a limited budget for assisting the poor in acquiring housing. The government faced a dilemma: it could pro- vide a little subsidy to all poor people, or a large subsidy to a few. It chose the latter strategy. Specifi cally, it required that public housing meet a minimum standard. (The government did not want to be accused of being a slumlord.) However, the minimum standard it chose was quite high. As a result, only about a quarter of those eligible for housing assistance receive it. There is concern that the result is worse than a random lot- tery, with most poor people unaff ected, and those lucky enough to get the subsidy receiving a big windfall. The reason that it is worse than just a lottery is that it inhibits those lucky enough to get the subsidy from mov- ing from welfare to work. Thus, public housing represents another form of welfare lock. America is a mobile country; every year about 16 percent of Americans move from one community to another, often in search of a job, or a better job. For poor individuals in public housing, though, moving becomes very unattractive. An individual who moves to another commu- nity goes to the bottom of the queue of applicants for housing assistance. In most cases, this individual will lose his or her benefi ts by moving. If the value of the housing subsidy is, say, $300 per month, the loss represents almost 40 percent of the income of a minimum-wage worker. When com- bined with other costs of moving—and other taxes, implicit and explicit— moving to another community for work becomes decidedly unattractive.

449Analytic Issues

FIGURE 15.5

HOUSING MARKET EQUILIBRIUM WITH VOUCHERS

Vouchers shift the demand curve to the right. (A) In the short run, supply is inelastic. The main effect is on price. Those not receiving the vouchers are worse off, as rents increase. (B) In the long run, supply is more elastic. However, unless the long-run supply curve is per- fectly horizontal, there will be some increase in rents.

Quantity

Rent

Demand curves

After vouchers

Before vouchers

Supply curve

R1

R0

Quantity

Rent

Demand curves

After vouchers

Before vouchers

Supply curve

R1 R0

A

B

450 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

ARE IN-KIND BENEFITS PATERNALISTIC?

We have noted that in-kind benefi ts are typically distortionary—that is, there is a substitution eff ect associated with them. Individuals would be better off if they were just given the same amount in cash; alternatively, the government could make them just as well off , and reduce its total expenditures. Ironically, the only circumstances in which in-kind benefi ts are not distortionary are when they have no substitution eff ect, only an income eff ect; then they sim- ply increase administrative costs, but the eff ect on consumption of food (or whatever the benefi t being provided) is the same as a cash benefi t.

Most advocates of in-kind benefi ts are bothered more by those cases in which it turns out that the benefi ts are ineff ective, than by those in which they are distortionary. In their view, the reason for government’s providing in-kind benefi ts is that it wants to ensure that the money trans- ferred is spent on “good” uses—on housing, food, and medicine. In this perspective, what society cares about is not just the perceived well-being of the recipients, but also the outward manifestations of poverty—the slums, malnutrition, and the like that result from it. Some economists fi nd this paternalistic view of the government objectionable; they argue that it violates the principle of consumer sovereignty.

Research in behavioral economics and the economics of information has shown, however, that many individuals, especially those of limited income and education, may not be fully cognizant of the eff ects on them- selves and their children of inadequate nutrition or preventive health

care, and may be prone to exploitation by, for instance, underregulated banks and merchants. As we noted, these in-kind programs may lead to better nutrition and health than would be the case with a cash grant.

Moreover, in Chapter 13, we encountered the view, called specifi c egalitarianism, that society should concern itself not only with the distribu- tion of purchasing power in general but also with access to particular goods, services, and rights. The right to a minimal level of medical care, food, and shelter, in this perspective, ought to be viewed as a basic right.

There is another rationale for in-kind bene- fi ts: the government is concerned not only with what gets consumed, but also with who con- sumes it. Some in-kind assistance is intended to target particular benefi ciaries, especially

IN-KIND BENEFITS Arguments in favor:

• Targets aid where it is most needed

• Specifi c egalitarianism

• Political expediency

Arguments against:

• High administrative costs

• Ineffi cient (distortionary): same gain in well-being of recipients could be achieved at lower costs

• Rules exacerbate distortions: SNAP encourages consumption of housing

• Ineffective (often do not signifi cantly increase expenditures on desired items, like food)

• Paternalistic

451Analytic Issues

children and those needing medical care. In a world of perfect informa- tion, the government might be able to target these groups, and in that case, cash payments would clearly be preferred. However, in a world of imper- fect information, restricting transfers to in-kind payments increases the proportion of funds going to the intended groups.

A study of this in the United States found that in-kind transfers have stronger eff ects on the well-being of children than do cash transfers, and that more narrowly targeted programs, such as Medicaid and Head Start, have greater impacts than more broadly targeted programs like SNAP. In contrast, a study of the impact of cash transfers of equal value to the in-kind Food Support Program (Programa Apoyo Alimentario) in rural Mexico found that although both forms of transfer had a large positive impact on food and total consumption, and led to a signifi cant reduction in poverty as well, the magnitude of impact was the same for both.15

Finally, there are political arguments for in-kind benefi ts. Diff erent congressional committees oversee each of the programs. SNAP, for exam- ple, is under the jurisdiction of the congressional agricultural committees. One of the reasons why the committees supported expansion of the SNAP food assistance program was that they viewed the program as benefi ting their special interest—the agricultural sector—just as the housing indus- try is one of the strongest advocates of housing programs. In this perspec- tive, overall support for welfare benefi ts would be reduced if these benefi ts were converted into cash payments. There is also a political dimension to the economic paternalism discussed previously: it is easier for politicians to convince their constituents to support in-kind benefi ts because of the widespread concern that cash assistance will be abused—for example, used to buy alcohol or cigarettes rather than nutritional food staples.

CATEGORICAL VERSUS BROAD-BASED AID

Should aid be given to all poor people or only to the poor who fall into cer- tain categories? A number of programs raise this question. For instance, the Supplemental Security Income program transfers income only to the aged or disabled poor.

The most basic criticism of categorical aid programs (whether in-kind or cash) is their relatively high administrative costs, as noted in the discussion

15 J. Currie, “Welfare and the Well-Being of Children: The Relative Eff ectiveness of Cash and In-Kind Transfers,” in Tax Policy and the Economy, vol. 8, ed. J. M. Poterba (Cambridge, MA: MIT Press, 1994), pp. 1–44; and E. Skoufi as, M. Unar, and T. Gonzalez-Cossio, “The Impacts of Cash and In-Kind Trans- fers on Consumption and Labor Supply: Experimental Evidence from Rural Mexico,” Policy Research Working Paper No. 4778, Washington, DC, World Bank, 2008. See also C. Blackorby and David Donaldson, “Cash versus In-Kind, Self-Selection and Effi cient Transfers,” American Economic Review 78, no. 4 (September 1988): 691–700.

452 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

of in-kind benefi ts. The greater expense of these programs arises primarily from the costs associ- ated with ascertaining eligibility. For instance, administrative costs per benefi ciary for the SSI program are now three times more per benefi - ciary and six times more per dollar of benefi t pay- ments than administrative costs for the Old-Age and Survivors Insurance Program.16

EFFICIENCY AND EQUITY ISSUES Besides considerations of administrative costs, both effi - ciency and equity issues arise in comparing cat- egorical versus broad-based aid. Categorical aid may have the eff ect of inducing individuals to fall

into the benefi ted category; it may have a distortionary eff ect. This is not true of Social Security in the sense that individuals do not become older faster simply to take advantage of the program, although people can choose to retire at an earlier age to collect Social Security. However, there were allegations that AFDC had contributed to the breakup of families, since in many states, eligibility could be aff ected by the presence of a man in the household.

The major advantage of categorical aid over broad-based programs is that under certain circumstances it can provide more eff ective redistribu- tion, with less loss in effi ciency. It can enable the targeting of aid to the most needy, who, at the same time, will not have adverse incentive responses. We have repeatedly emphasized the trade-off s between equity and effi ciency considerations in the design of redistribution programs. Providing a basic level of income through transfer payments that decline as the individual’s income increases (from wages or other sources) may discourage work. This will normally imply that a lower level of redistribution is more desirable for individuals whose response to incentives is large, than for individuals whose response to wage incentives is small (e.g., those over age 70).

There is an equity argument against categorical programs. Some believe that the government should not discriminate in favor of or against any particular groups. Two individuals who are equally poor should receive the same amount from the government, whether they are young or old. There should not be “favored” categories, such as single-parent households or the aged. Admittedly, older individuals may have more medical expenses, and one might want to adjust the transfers to take this into account; however, this is already eff ectively done through Medicare.

16�S. Szymendera, “Social Security Administration (SSA): Budget Issues,” Congressional Research Ser- vice Report R41716, August 2011.

CATEGORICAL AID Argument in favor:

• Targets aid where it is most needed, reducing overall distortionary effects and costs

Arguments against:

• Unfair to treat different poor people differently

• Distortionary effects in meeting eligibility stand- ards (effect of TANF on family dissolutions)

• High administrative costs

453Analytic Issues

IS MEANS TESTING OBJECTIONABLE IN ITS OWN RIGHT?

A far more fundamental set of questions is raised by means testing. Should benefi ts be targeted at the poor alone? Any program like Medicare that provides benefi ts for all individuals, irrespective of their income, has a dis- proportionate eff ect on the poor. One way to help the poor is to provide equal benefi ts to everyone in society. Such programs have several distinct advantages. First, they reduce the distortions associated with the phaseout of means-tested benefi ts. Second, means-tested programs are often seen as carrying a stigma. That is one reason why Social Security and Medicare were not means tested. Not all those who are eligible for public assistance participate. The administrative complexity may provide part of the expla- nation, but another reason may be that some of those eligible view means- testing programs as demeaning. Finally, there is a political argument, which is sometimes put thus: “Means-tested programs are mean”—that is, because they lack political support, they tend to be stingy.

In recent years, as the government has faced perceptions of increas- ing fi nancial stringency, these arguments have been overwhelmed by the desire to target the limited funds to those most needy. Thus, there has been increasing support for making wealthier individuals pay a higher fraction of their Medicare costs. (Note that whether there is “fi nancial stringency” is largerly a matter of perception and politics. Countries far poorer than the United States have provided broad social benefi ts without means testing.)

OTHER DISTORTIONS

So far, we have discussed how welfare programs can adversely aff ect labor supply, and how in-kind benefi ts can distort consumption decisions. Critics of welfare programs argue that they have even more fundamental eff ects on behavior, such as that they have contributed to family breakdown, increased teenage pregnancy, and concentrated welfare dependency in the states that pay the highest benefi ts, as the poor migrate to those states.

It is clear that welfare has, or has had, features that might result in those behavioral eff ects. For instance, because welfare benefi ts typi- cally depend on household income, the departure of a low-wage father may increase total “family” income by making the family eligible for assistance.17 It is also true that a disproportionately large number of

17 According to one study, the combined AFDC and food stamp benefi t for an AFDC family was higher than the reported income of 62 percent of black men aged 20 to 24 and of 29 percent of black men aged 25 to 34. See F. Levy and R. Michel, “Work for Welfare: How Much Good Will It Do?” American Economic Review 76 (May 1986): 399–404.

454 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

CONDITIONAL CASH TRANSFER PROGRAMS

O ne of the most important innovations in welfare programs over the past decade is conditional cash transfers (CCTs), some- times referred to as contingent welfare programs. The objective of CCTs is to offer basic income support to low-income families, while at the same time promote investments in human capital to reduce long-term poverty and inequality. CCTs are incentive-based social welfare programs that pro- vide direct cash transfers to poor households con- ditional upon behavioral change that will improve development prospects of the next generation, such as sending children to school and bringing children to health centers.

Brazil’s Bolsa Família (Family Allowance) is the world’s largest CCT. As of 2010, it had 12.7 million ben- efi ciary households (more than 49 million individuals) and an annual budget of $5 billion. Bolsa Família was created in 2003 by combining and redesigning pre- viously existing cash transfer programs for education (Bolsa Escola), food and nutrition (Cartão Alimentção and Bolsa Alimentção), and cooking gas (Auxilio Gás). Its key components are monthly transfers per child or pregnant/breastfeeding mother, conditional on a school attendance rate of at least 85 percent for chil- dren age 7 to 15, vaccinations for children 0 to 6 years, and prenatal care visits for pregnant women.

Bolsa Família has had remarkable education and health impacts to date. It has signifi cantly increased school attendance, rates of grade progression, school retention of children after the age of 14, pre- natal care visits, the probability that a child receives all seven vaccines required by the age of 6 months, and healthy body mass index (BMI) among children.

Bolsa Família’s success is not unique. Several other CCT programs have also helped both to reduce

current poverty and to increase future opportunities for low-income families, such as Bolsa Família’s predecessor, Mexico’s widely studied Progresa/ Oportunidades program. Approximately thirty countries now have CCTs. Most large-scale CCTs to date have been in Latin America, which amplifi es the potential impact of CCTs, given that this region has the world’s highest degree of income inequality.

Three distinguishing characteristics of CCTs underlie their success and worldwide growth. First, CCTs creatively mitigate the trade-off between humanitarian relief and economic effi ciency by mov- ing beyond a focus on just short-term poverty allevia- tion and income redistribution to include promotion of long-term economic growth through investment in human capital development. This also addresses the criticism of unconditional cash transfers, such as that they produce perverse incentives that reduce labor supply, crowd out private transfers, and encour- age dependency. Second, CCTs use incentives to encourage utilization of public services, particularly health and education services, designed to offer a pathway out of poverty. Third, CCTs are designed and implemented in the context of “evidence-based public policy,” in contrast to traditional safey net pro- grams; especially in developing countries, CCTs have been closely studied, exhaustively documented, and rigorously evaluated.

However, a word of caution is in order. CCTs have been most successful in middle-income countries with both a strong civil service and a relatively wide- spread network of local schools and health centers. It is diffi cult for conditional transfers to work if the speci- fi ed conditions cannot be met easily by low-income families because of weak public sector institutional capacity and insuffi cient public infrastructure.

SOURCES: L. Aber and B. Rawlings, “North–South Knowledge Sharing on Incentive-Based Conditional Cash Transfer Programs,” Special Protection & Labor Discussion Paper No. 1101, Washington, DC, World Bank, January 2011; L. Brenzel, “Brazil—Bolsa Familia Project : Results-Based Financing (RBF) for Health—Country Snapshot,” Washington, DC, World Bank, November 2009; A. de Brauw et al., “The Impact of Bolsa Família on Education and Health Outcomes in Brazil,” paper by the International Food Policy Research Institute presented at the Second Generation of CCTs Evaluations Conference, World Bank, October 2011; and F. Veras Soares, R. Perez Ribas, and R. Guerreiro Osório, “Evaluating the Impact of Brazil’s Bolsa Família: Cash Transfer Programs in Comparative Perspective,” Latin American Research Review 45 (2010): 173–190.

455Analytic Issues

welfare recipients live in one-parent or no-parent families.18 However, the causal link has been questioned. Out-of-wedlock births also increased dramatically among those not on welfare during the 1970s and 1980s (although they declined somewhat among both groups in the 1990s), and the incidence of out-of-wedlock births does not seem to be higher in states off ering larger benefi ts. Indeed, in many states in which there has been a marked rise in such births, benefi ts come nowhere near covering the additional costs of a child. Although there may be some eff ect, welfare benefi ts simply cannot account quantitatively for the magnitude of the change.19 Indeed, studies have not found eff ects of welfare on illegitimacy large enough to explain the observed increases in female-headed fami- lies.20 To be sure, poverty itself contributes to family breakdown and wel- fare dependency.21

Although there has been extensive migration, especially from the South to northern urban areas, since AFDC was fi rst adopted in the midst of the Depression, there is debate about the extent to which that migration was induced by AFDC, in say, northern states. Other factors were prob- ably more important. Jobs were more available, pay was higher, and there was less racial and class discrimination. Although, from an economist’s perspective, if one is going to be unemployed, simply receiving checks, it makes sense to do so where the checks are largest, migration does not seem to be particularly sensitive to changes in welfare benefi ts. This may be because most welfare recipients plan to return to work; job opportuni- ties and other factors, such as proximity to friends, dominate. These, plus the costs of moving, more than off set the diff erences in welfare benefi ts among the states.

18 In 2010, 53 percent of TANF cash assistance went to one-parent families and 42 percent went to no- parent families; only 5 percent went to two-parent families. (Source: House Ways and Means Commit- tee, 2011 Green Book, Table 7-10.) 19 One interesting theory attributes the increased incidence of out-of-wedlock birth to the increased availability of abortions. Previously, a signifi cant fraction of marriages were shotgun marriages—as evidenced by the number of fi rstborn children arriving, say, seven months or less after marriage. This number has decreased dramatically; the availability of abortions has meant that men no longer feel the moral obligation to marry, as the woman has a choice. See G. A. Akerlof, J. L. Yellen, and M. L. Katz, “An Analysis of Out-of-Wedlock Childbearing in the United States,” Quarterly Journal of Economics 111, no. 2 (May 1996): 277–317. 20 See R. Moffi t, “Incentive Eff ects of the U.S. Welfare System”; David Ellwood and Lawrence Summers of the Harvard Kennedy School, found that variations in benefi t levels across states were not associated with variations in divorce rates, illegitimacy, or percentages of children in families headed by single parents. See D. Ellwood and L. Summers, “Poverty in America: Is Welfare the Answer or the Problem,” in Fighting Poverty: What Works and What Doesn’t, ed. S. Danziger and D. Weinberg (Cambridge, MA: Harvard University Press, 1986), pp. 78–105. 21 See J. Currie, “Welfare and the Well-Being of Children”; according to Moffi t, “The results show consistent evidence of the strong correlations between parental welfare receipt and later behavior of daughters. Daughters of welfare families are much more likely to participate in the welfare sys- tem themselves at a later date, and are more likely to have births in general and premarital births in particular.” See “Incentive Eff ects of the U.S. Welfare System,” p. 36.

456 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

WELFARE REFORM: INTEGRATION OF PROGRAMS

This chapter began by noting that some of the dissatisfaction with wel- fare programs was misplaced; although they are blamed for the increas- ing defi cit, they really had little to do with it. However, the discussion in the preceding section has uncovered several other complaints, about actual or perceived distortions in behavior induced by the system. In recent years, there have been major initiatives for reform—one aimed at simplifying and integrating the welfare programs, and another aimed at trying to “make work pay.”

As we have seen, the current welfare system consists of a potpourri of programs. Some programs, such as TANF, originated to deal with particular problems. Others grew in part in response to special interest groups—for example, the food stamp program gave the agriculture indus- try an opportunity to increase the demand for their products, and thus the income of farmers.

Regardless of the origins of programs within the welfare system, their lack of integration has two negative eff ects. First, it greatly increases administrative costs and burdens. Each program has, for instance, a dif- ferent eligibility form. The bureaucratic complexity discourages some who could really benefi t from the programs from taking advantage of them. It has been estimated that only about two-thirds of those eligible for AFDC received it, only 60 percent of those eligible for food stamps received food stamps, and only 83 percent of those eligible for EITC actu- ally collected it.22

The second negative eff ect of the lack of integration is that the pro- grams interact in ways that increase the magnitude of the distortions. Each means-tested program has a phaseout, an interval of incomes over which benefi ts decline. The decline in benefi ts is equivalent to a tax. Even though legislators may be aware of this eff ect, as they consider each pro- gram in isolation they seldom think of the combined eff ects of all of these taxes. As we have seen, the combined eff ect may result in a total marginal tax rate well in excess of 60 percent. If all the programs were consolidated into a single cash program, not only would overall administrative costs be reduced, but attention would also be paid to the combined eff ects.

At various times, there has been support for integrating all of the welfare programs together, and “cashing them out”—allowing those on

22 J. K. Scholz, “Tax Policy and the Working Poor: The Earned Income Tax Credit,” Focus (Winter 1993–1994): 1–12.

457Welfare Reform: Integration of Programs

welfare to spend the money however they wish. Such proposals amount to an expanded EITC. They are sometimes referred to as negative income tax proposals.

In this system, all individuals would be required to fi le tax returns, but just as only individuals above a critical threshold level would have to pay income taxes, those below that critical threshold level would receive a check from the government. Consider a tax regime in which everyone receives a check from the government of, say, $1000 per year, but then pays to the government, say, one-third of each dollar earned. Those with an income of less than $3000 would receive something net from the government; those with an income greater than $3000 would pay to the government more than they receive. Those who favor this sys- tem argue that it is not only administratively simpler, but it is also less demeaning than the present system, which forces individuals to present evidence to several agencies concerning their low income.

Advocates of a negative income tax, however, have argued for more than just the consolidation of current welfare programs and a conversion to cash benefi ts. They have argued for a change in the structure of total benefi ts, for two key reasons. First, the high marginal tax rates on the poor discourage work eff ort among people whose attachment to the labor force is already weak; second, the total benefi ts under the current system amount to less than what is required to remain out of poverty.

However, this gives rise to one of the fundamental dilemmas in welfare policy. If we want to provide benefi ts that are large enough to ensure a minimal, say, poverty, level of income even for those who do not work, and if we want the phaseout to be slow so the implicit mar- ginal tax rate is low, then the benefi ts must extend high up into the income distribution. For instance, if the minimal benefi t for a family of three were $12,000, and the marginal tax rate, including the full Social Security contributions of approximately 15 percent, were kept to 30 percent (ignoring state taxes), then benefi ts would extend up to families with an $80,000 income; if the marginal tax rate were increased to 55 percent—more than one out of two dollars earned by the individual thus going to the government—benefi ts would still extend up to $40,000. The costs of such an arrangement would be high—too high to be politically acceptable. Thus, either initial ben- efi ts must be below the poverty level or eff ec- tive marginal tax rates must be very high.

BASIC TRADE-OFF IN THE DESIGN OF NEGATIVE INCOME TAX Higher benefi ts for the very poor entail either greater expenditures (raising tax rates, and reducing incentives for higher-income individuals) or faster phaseouts, implicitly entailing higher marginal tax rates for lower-income individuals.

458 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

THE WELFARE REFORM BILL OF 1996

As we have noted, the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 marked a major departure from the past in two important ways. First, it replaced the AFDC system, in which the federal government paid a fraction of the costs, depending on the per capita income in the state, with the block grants of TANF. Second, it imposed a number of stringent requirements designed to encourage movement from welfare to work. Two beliefs underlay the reforms: one that the programs were too costly (partly based on the misperception noted earlier concerning the size and growth of welfare programs and their role in the growing defi cit), or another that the programs had failed in their primary purpose—rather than being a safety net, for too many people, welfare had become a way of life.

BLOCK GRANTING

Although the states administered the AFDC program, many of its basic features were dictated by the federal government. One of the most impor- tant features of the 1996 welfare reform was to end the federal welfare entitlement. State matching grants—in which the federal government contributed, say, a dollar for every dollar spent by the state—were converted to block grants—fi xed amounts of money—giving the states much more discretion in how they administered the funds. Critics worried, however, that converting to block grants would lead to reductions in expenditures. The 50 percent matching rate had eff ectively lowered the “price” of wel- fare. A state could get two dollars of welfare benefi ts for its citizens by spending only one dollar. Welfare, it was argued, was like any other good: lowering the price raised the quantity demanded; raising the price would lower the quantity demanded.

Critics worried further that there would be a “race to the bottom.” Lowering a state’s welfare benefi ts relative to those of neighboring states would provide incentives for those on welfare to move; each state would have an incentive to encourage those free riding on the welfare system to leave. The upshot of these incentives, critics argued, would be competi- tion among the states to drive out those dependent on welfare.

Advocates of block grants argued that there was not strong evidence of a large elasticity of demand—that is, of a large responsiveness to price—so that fears of welfare cuts were exaggerated. Moreover, there were incen- tives to drive out welfare recipients under the old regime: the states still

459The Welfare Reform Bill of 1996

paid half of the cost of the benefi ts, although to be sure, these incentives would be enhanced if the states had to pay, at the margin, 100 percent of the benefi ts. Nonetheless, it was hard to see evidence of a race to the bottom; even though there were diff erences in benefi t levels among states, they were related more to the states’ income and demographics. Moreover, advocates of block grants argued, the increased effi ciency resulting from state control and the elimination of federal bureaucratic requirements would increase the benefi t received per dollar spent—at least partially off setting the fact that the federal government was no longer picking up 50 percent of the tab.

To address concerns that there would be an excessive reduction in welfare support, proponents of block grants agreed to a “maintenance- of-eff ort provision,” under which states receiving the block grants would agree to continue spending at least 75 percent of the amount that they had previously spent on welfare. Advocates argued that with the greater effi ciency resulting from the devolution of responsibility to the states, the eff ective level of welfare benefi ts would not be cut at all. But critics raised two objections: Advocates of block grants exaggerated the negative impact of federal regulations and bureaucracy, so that any major cut in expenditures would be felt.

Moreover, with time, the adverse impact of block grants would worsen, particularly in states in which for one reason or another there was a marked increase in the welfare load (for instance, because of an economic downturn). If, under the old regime, benefi ts would have expanded by only 15 percent, then a 75 percent maintenance-of-eff ort requirement would be equivalent to more than a one-third cut in benefi ts, relative to what they otherwise would have been.

Furthermore, hidden in the 1996 bill were provisions that allowed funds spent on other social programs directed at the poor to be counted toward the maintenance-of-eff ort requirement, substantially lowering the eff ective maintenance-of-eff ort requirement.

ANALYTICS OF STATE RESPONSES TO BLOCK GRANTS

We can use some simple diagrams to illustrate why economists worried that converting from a matching system for welfare to a block grant would reduce overall support for welfare.

Figure 15.6 shows how the states might be anticipated to respond. BB gives a state’s budget constraint between welfare and other expenditures before the federal program, and shows a hypothetical state indiff erence

460 CHAPTER 15 WELFARE PROGRAMS AND THE REDISTRIBUTION OF INCOME

curve, depicting the trade-off between welfare and other expenditures. The state chooses point E, maximizing its (economic) welfare, subject to its budget constraint.

BB' gives a state’s budget constraint under the former system of fed- eral matching funds. The state is induced to spend more on welfare, both because of an income eff ect (the state is better off as a result of the fed- eral subsidy) and a substitution eff ect (for each dollar of increased welfare expenditure, the state has to give up only 50 cents of nonwelfare expendi- tures). The state now chooses point E'. The increased funding for welfare and the lower price both lead to more expenditure on welfare than before. There are positive income and substitution eff ects.

B"B" now gives the budget constraint when the federal subsidy is con- verted into a lump sum, not dependent on the amount spent. The state is better off , but unless the income eff ect is large, expenditures on wel- fare are decreased; although, because of the income eff ect, they remain higher than with no subsidy. The fi gure shows what has been the con- cern of critics of block granting: a large decrease in welfare expenditures. (However, note that, as depicted, the indiff erence curve in the new equi- librium is at a higher level, refl ecting the general proposition that nondis- torting subsidies are more effi cient.)

EFFECT OF BLOCK GRANTING

Switching from a matching grant system to a block grant

costing an equal amount of money is likely to reduce

welfare expenditures.

FIGURE 15.6

Welfare expenditures

Other expenditures

B ¿

B

B ¿¿ W ¿¿