Worksheet_Economics Reading

What Happens When Development Regulations Constrain

Housing Supply?

A Review of the Literature on Housing, Employer, and Labor Market Responses

Prepared for: U.S. Department of Housing and Urban Development

Washington, D.C.

Prepared by:

Michael Carliner Newport Partners LLC

David Rodda

Abt Associates

Eric Belsky and Daniel McCue Joint Center for Housing Studies

May 2007

This is a working paper distributed for review. It should be treated as confidential and not reproduced without the express permission of Newport Partners.

Workforce Housing Final Report Table of Contents i

Table of Contents

Overview ............................................................................................................................. 1

Introduction ........................................................................................................................ 8

Chapter One: Housing Market Effects ............................................................................10

Defining, Measuring and Comparing Regulations ......................................................11

Regulations and House Prices ...................................................................................16

Direct Effects on House Prices and Construction Levels ............................................17

Regulations, Supply Elasticities, and House Prices ....................................................19

House Prices and Labor Markets ...............................................................................20

Regulations and Labor Markets .................................................................................22

Regulations and Industrial Compositions ...................................................................25

Summary and Research Extensions ..........................................................................26

Chapter Two: Business Perspective ...............................................................................28

Industries and Employment ........................................................................................29

Business Location and Agglomeration Effects ...........................................................31

Business Dynamics ....................................................................................................37

Chapter Three: Metropolitan Labor Supply ....................................................................40

Identifying Labor Supply Problems .............................................................................41

Housing and Net Migration .........................................................................................44

U.S. Migration Studies ...............................................................................................46

Foreign Migration Studies ..........................................................................................47

Evident Impacts of Housing........................................................................................49

Needed Research ......................................................................................................50

Chapter Four: Labor Productivity ...................................................................................51

Wage Models .............................................................................................................53

Turnover ....................................................................................................................56

Education and Training ..............................................................................................61

Returns to Skill, Creative Occupations, and Knowledge Spillovers.............................62

Chapter Five: Spatial Mismatch and Job Search ...........................................................70

Spatial Mismatch ........................................................................................................71

ii Table of Contents Workforce Housing Final Report

Job Search ................................................................................................................ 75

Regional Adjustment Models ..................................................................................... 78

Chapter Six: Commuting and Transportation Networks ............................................... 82

Trade-off Between Housing Costs and Commuting Costs ......................................... 83

Commuting Impact on Pay ......................................................................................... 90

Jobs-Housing Balance ............................................................................................... 91

Congestion Pricing and Measurement ....................................................................... 94

Chapter Seven: Recommended Research Directions ................................................. 100

Bibliography .................................................................................................................... 104

Workforce Housing Final Report 1

Overview

While the impacts of residential development regulations and of economic growth on

housing markets have been studied extensively, the impact of housing supply on labor

markets and local economic growth (other than the direct effect of construction-related jobs)

has not. Moreover, only one paper traces the influence of development regulations on

housing outcomes through to labor outcomes at the metropolitan level (Saks 2005). In light

of this fact, the U.S. Department of Housing and Urban Development (HUD) commissioned

a paper to review the literature on the influences of residential development regulation on

housing markets, and through them, on businesses, labor markets, and regional economic

competitiveness.

Previously, HUD commissioned detailed reviews of the literature on development regulation

influences on housing market outcomes (see for example Schill, 2004, and Quigley and

Rosenthal, 2005). Those reviews did a thorough and comprehensive job of critically

evaluating this literature and drawing findings from it. Rather than repeat this exercise, this

paper is intended to provide a conceptual overview of the impact of development regulations

on housing markets and how these impacts may influence labor markets, business location

decisions, the economic competitiveness of metropolitan areas, the industrial composition of

metropolitan areas, and labor productivity. Due to the lack of studies that attempt to model

the impact of development regulations on employment outcomes in response to labor

demand shocks, this paper reviews a large body of related literature on employment location

dynamics, such as agglomeration effects and labor pooling, and the factors that influence

labor migration. Some of this literature at least tangentially relates to the impact of higher

housing costs and other regional quality of life factors on labor migration. Perhaps most

importantly, this paper provides a conceptual framework for thinking about how development

regulations intended to govern residential construction might have secondary impacts on

labor supply and demand. It also highlights the data constraints that make it so difficult to

study the impacts of residential development regulations directly on housing markets, let

alone on overall economic activity, as well as the research questions most urgently in need

of attention.

Regulations that govern residential construction result in housing market outcomes that

could have labor market consequences. Specifically, development regulations constrain the

supply response to demand by 1) imposing costs and restrictions on construction that shift

2 Workforce Housing Final Report

the housing supply curve to the left and 2) constraining the type of housing that can be built

and where it can be built. The former adds to the housing costs, lifts prices, and limits the

capacity to supply housing over the long run. The restrictions on supply in the immediate

area may result in a sprawling pattern of development and what has been characterized as

a spatial mismatch between jobs and housing.

Affordable housing built at high densities is not permitted or is severely constrained relative

to market demand in many jurisdictions. Many places that allow industrial and commercial

development to strengthen fiscal capacity do not allow for construction of housing that is

affordable to the workers employed in these economic activities or to essential municipal

service workers such as police, fire fighters, administrators, and teachers. The distance

between job locations and housing thus increases. This separation of jobs and housing

induces longer commutes, as does the rent gradient that usually makes housing on the

periphery of metropolitan areas less expensive than housing near employment nodes.

Of course, this stylized view of the influence of development regulations on housing market

outcomes oversimplifies a complex subject, does not pertain in all places, and ignores a

host of legitimate methodological issues that have been raised with much of the literature

from which this stylized picture is drawn. Still, it is widely accepted that development

regulations often have just these sorts of outcomes. More in dispute is the magnitude of the

effects and the transmission mechanisms from regulations to market outcomes. The

question of moment for this investigation is how these development regulation-driven

housing market outcomes could influence labor markets, business location decisions,

regional economic competitiveness, and the industrial composition of employment at the

metropolitan level.

By constraining the supply response to housing demand, development regulations drive up

long-run housing costs which should in turn place pressure on local employers to increase

wages to attract and retain workers. It should also dampen employment growth in response

to a positive labor demand shock because housing supply cannot keep pace with demand

and workers don‘t come to fill jobs. This is precisely what Saks tested for and found in the

one study that traces the impact of development regulations, through housing markets, on

labor markets using econometric models.

Workforce Housing Final Report 3

At a more micro level, the impact of higher housing costs and longer commutes on labor

markets depends importantly on the type of labor. Employers and workers engaged in the

production of goods or services for export from the region in which a firm is located are

under potential pressure from competition from producers located elsewhere that can recruit

workers of comparable quality but lower cost because housing costs are less or commutes

times shorter. These firms are often called ―basic industries‖ or ―export-producing

industries‖ even if some of their goods and services are consumed locally. Employers and

workers in firms that serve strictly local demand do not face competitive pressures from

firms in other locations in the conventional sense. Retail stores, for example, mostly

compete directly with firms exposed to the same or similar housing market costs and labor

market conditions (though they may face competition from non-store retailers such as web-

based providers).

Local greengrocers, gas stations, beauty salons, drycleaners and the like as well as local

public services constitute the far extreme of firms that serve a local market. They too may

face difficulties recruiting or retaining talent as a result of long commutes and high housing

costs. But all their competitors must struggle with these same problems. Despite

difficulties, these firms will continue to provide their products and services as long as

customers are willing and able to cover the costs of producing these goods and services

and supply a competitive risk-adjusted return on their capital. Basic industries, on the other

hand, can only remain competitive if the advantages of remaining or starting up in a high-

cost area (or one with long commutes and congestion) outweigh any disadvantages.

In general, higher housing costs can increase the compensation it takes to attract and retain

workers. The higher wages that employers must pay will not be seen by workers as higher

real wages unless the wages more than offset the higher cost of living. Unless these higher

costs for employers are justified by agglomeration effects (positive effects of businesses

locating together for reasons of productivity, lower costs to recruit labor because of a

concentration of workers with similar skills, reduced input and delivery costs, greater

information), basic industries may eventually exit. Over time, the only basic industries that

remain should be those that enjoy higher revenues or cost reductions to offset the higher

wage rates. The non-basic, local service employers will respond by charging higher prices,

adding to the cost of living pressures directly from housing.

4 Workforce Housing Final Report

Despite these expected outcomes, we could find no studies that actually attempt to

empirically investigate or rigorously identify econometric equations to examine the direct or

indirect impacts of development regulations on the composition of employment or the costs

of other locally provided goods and services. There is a theoretical literature that touches

on it but little empirical work (citations). Instead, the focus has been on the impact of these

regulations on wage and employment growth in places with tight regulations relative to those

with looser regulations (citations).

There is a relatively well-developed literature that explores the influences of various

conditions on labor migration. While, of course, labor migration is strongly influenced by

labor demand, labor is motivated to migrate from one metropolitan area or another for a host

of reasons. Among them are housing costs, commute times, and other costs of living and

quality of life factors. To the extent that high cost areas make it more difficult to attract and

retain workers, positions may get filled by inferior personnel that reduce productivity.

Spatial patterns of residential development that result in longer commutes, as well as higher

housing costs, can influence the productivity of labor and the nature of labor supply.

Workers that take long commutes may have to leave for family or other commitments at

designated times and have less flexibility to work late, also reducing productivity. There is

widespread anecdotal evidence of these effects, but no comprehensive studies.

People are willing to endure longer commutes and higher housing costs in certain

metropolitan areas with tighter development regulations because they value the amenities

and employment opportunities these places may offer. Indeed, some of the most costly

markets with the most elaborate and expensive development regulations are considered

highly desirable places to live and work, such as San Francisco, Hawaii, Boston, New York

and Los Angeles. Like employers, employees may value agglomeration of similar firms

because it increases their employment opportunities. But even if highly desirable areas,

restrictions on housing supply may take an economic toll.

Unfortunately, the scholarly literature is largely silent on many of these issues. As noted,

the influence of development regulations on the composition of employment has not been

studied. The influence of housing costs on business location decisions has also not been

well studied, nor the influence of these on labor productivity. The impact of higher housing

costs or longer commutes on recruiting costs, employee turnover, or labor productivity also

Workforce Housing Final Report 5

have not been studied much. Without such studies, the influence of development

regulations on labor markets, business decisions, and regional economic competitiveness

remains very uncertain.

Moreover, despite greater attention to the direct impact of regulations on housing supply and

cost, measures of development regulations and their administration remain underdeveloped

and difficult to develop. Furthermore, the fact that these development regulations and their

administration vary dramatically across jurisdictions within metropolitan areas is largely

glossed over in most empirical studies. It is important to understand how these intra-

metropolitan differences in regulatory intensity influence development patterns and traffic as

well as how the aggregated effects of these regulations affect labor market conditions in one

metropolitan area compared with another.

In the absence of fundamental understanding of many of these questions, it is difficult to

judge what public policies might improve regional economic competitiveness or for business

to recognize the effects of development regulations on their costs. In short, this

investigation into the impact of development regulation on labor markets and businesses

raises far more important questions than the existing literature permits it to answer. You will

find these research questions in each of the core subject chapters. However, the final

chapter describes four (4) major research directions that should be pursued to further the

understanding of the linkages between housing and employment and enable improved

policy decisions. These broad initiatives are developed from multiple research needs and

are not listed in a priority order.

1) Update and develop better measures for land use regulations and create a

national database.

Measurements of regulatory restrictiveness vary greatly, and are often unreliable or out of

date. Updated and improved measurements of regulatory restrictiveness are vital to the

establishment of a comprehensive understanding of regulatory restrictiveness.

In addition to updated and improved measurements of regulatory restrictiveness, the

authors suggest creating a comprehensive database containing up to date information on

local regulations across the nation. Ideally, this effort would be accompanied by a national

survey to collect a consistent set of specific elements on a regular basis.

6 Workforce Housing Final Report

2) Quantify how the diversity of regulations within a metro area affects the

performance of entire metro area in order to: (a) gain finer resolution of impact

by industry, sub-metro geography, and type of development structures; and, (b)

determine if it is the quantity or the distribution of workforce housing that matters

with respect to house prices, labor outcomes, and business growth, stability,

and/or composition.

Within a metropolitan area, varying levels of restrictiveness may create an inaccurate overall

impression of the restrictiveness of the entire metro area. For example, the efforts of a few

low-restriction jurisdictions to encourage development may make a highly-restrictive metro

area appear to be less restrictive than it actually is overall. Determining the effects of

diverse regulatory restrictiveness within a metro area will provide a more accurate picture of

what is happening on the local level. This study should also include an analysis of how

regulations affect industry mixes and entry and exit of industries within the metro area.

Metropolitan area empirical data is needed in order to understand the localized effects of

regulatory restrictiveness on industry.

When studying the impacts of diversity of regulatory restrictiveness within a metro area, the

authors suggest examining the effects on housing and labor market behavior within cross

sectional studies.

The other two recommended research initiatives combine identified or suggested research

projects primarily from chapters two through six into multi-faceted research initiatives. In

doing so, the reader should be aware that each of these chapters identifies additional

research ideas and needs that are missed in this summation.

3) Determine how housing costs and congestion affect industries and firms,

including the following:

the rate of overall economic growth;

changes in industry composition;

differentiation by occupation, level of education, and income;

identification of industry winners and losers as a result of rising house prices;

and,

Workforce Housing Final Report 7

identification of industry winners and losers as a result of congestion

(commuting cost, delay, or uncertainty).

Industries that rely heavily on mid-skill workers and have to locate downtown or in dense

employment centers may be most vulnerable to high housing costs and congestion effects.

Firms locating downtown may have the benefit of better public transportation service, but

much more expensive housing and parking. The goal of this research effort would be to

quantify the degree to which industries vary in their sensitivity to housing costs and

congestion effects.

4) Develop qualitative detail on how rising house prices affect employers through

market research to address the following:

recruitment strategies, e.g., longer search, wider search, subsidize

transportation/housing, training;

productivity measures, e.g., lateness, absenteeism, telecommuting, job

shopping on the job; and,

retention / turnover , e.g., shorter tenure, higher wage and promotion.

To better appreciate the connection between house prices and labor supply, a deeper

understanding of how employers respond is needed. Although this research project

included case studies of selected localities, more data needs to be collected through

interviews and other means to be able to fill in additional information on employer‘s behavior

in a variety of economic circumstances.

Further information on each of these suggested research directions can be found within the

body of the document and additional detail is provided in Chapter Seven: Recommended

Research Directions and additional research questions are contained within each chapter.

8 Workforce Housing Final Report

Introduction

The purpose of the Workforce Housing Project is to develop a research agenda that shows

the impact of regulatory barriers on employers‘ ability to recruit and retain employees. This

paper contributes to that effort by reviewing the literature and offering recommendations for

additional data collection and research. Presentation of this literature review to an advisory

panel of experts provides them with the opportunity to provide their own feedback. The

authors of this literature review are certainly aware that it is not comprehensive. However,

the review constitutes a starting point for identifying the most important gaps in the existing

research and the most promising new research directions.

The root problem is a shortage of workforce housing that is affordable given existing wages

for a substantial range of working households. Beyond affordability, the issue of workforce

housing encompasses the jobs-housing mismatch and the role that land use regulation

plays. If businesses have difficulty recruiting workers because housing is so expensive or

distant, the businesses cannot grow or effectively compete. Zoning is designed to preserve

residential areas for adequate housing at different cost levels. However, moderate cost

housing has often been squeezed out of locations near employment centers by more

profitable and higher tax generating commercial uses. Workers face the tradeoff of less

expensive housing with a long commute or a shorter commute from expensive and crowded

housing. Employers must increase wages either to compensate workers for the long

commute or more expensive housing. At some point, the lack of workers or the high cost of

workers is enough to hurt the profitability of businesses, which may be forced to relocate

where labor is cheaper.

High housing cost burdens for workers can reflect strong labor demand as much as limited

housing supply. It is analogous to the commuter congestion problem. When a local

economy is booming, the roads are usually congested with commuter traffic as demand for

access outstrips supply. Although high housing prices and congestion delays often

accompany growth, they may ultimately constrain continued growth. Some firms can

sustain higher wages as long as the increased employment density results in better

matching of employees to jobs and higher productivity, but productivity may actually suffer

from other activities, and high wages may drive away firms that do not enjoy offsetting

revenue gains or cost savings. Moreover, technology keeps advancing, though not

Workforce Housing Final Report 9

necessarily smoothly, and both firms and workers can move to better situations, though

often with difficulty. If cities adopt too many regulations that block growth and change,

workers and firms will eventually move to places where they get a better deal. The issue of

workforce housing is fundamentally about urban efficiency in creating a productive

environment that can compete for growth. Businesses should care about workforce housing

because their growth prospects are tied to the availability of labor, which in turn, is limited by

the availability of moderately priced housing and adequate transportation networks.

The following literature review is organized into seven chapters. The first chapter on

Housing Market Effects establishes the links between land use regulations, an inelastic

housing market, labor supply, spatial mismatch, and the commuting congestion that results

when workers bridge the long gap between work and home. The second chapter highlights

the business perspective in which businesses attempt to fill their vacancies and hang onto

their existing workers either by moving to labor pools or enticing the labor to migrate to the

business. Literature on the differing demands by industry and the choice of business

location are followed by papers on agglomeration effects, business dynamics, migration,

and new data measures of labor activity and regulatory stringency. Chapter three looks at

metropolitan labor supply issues resulting from a constrained housing supply. In the fourth

chapter, labor productivity is subdivided into sections on wage models, turnover, training

and returns to skill. A fifth chapter on spatial dimensions considers the topics of spatial

mismatch, jobs search, migration, commuting costs and the jobs-housing balance. The sixth

chapter deals with issues of commuting and transportation. Our recommendations for the

workforce housing research agenda are summarized in the final chapter. Those

recommendations are broad and should be read as more general directions rather than

specific research projects. Many other research questions and suggested research topics

are contained throughout the document.

10 Workforce Housing Final Report

Chapter One: Housing Market Effects

The recent boom in housing prices and widespread decay of affordability across many of the

nation‘s largest metropolitan areas have increased concern over the effects of housing

affordability on the future economic growth and well-being of these areas. Of primary

concern is that local land use regulations, intended to shape the physical growth of an area,

may in fact be hindering the area‘s potential for economic growth by contributing to high

housing costs that repel would-be workers and employers. Indeed, the effect of land use

regulations on house prices and affordability has been well studied. A recent literature

review by Quigley and Rosenthal (2005) lists no less than forty such empirical studies in the

last four decades, the majority of which indicate that factors which restrict development are

associated with inelastic housing supply responses to increasing demand and high housing

prices. Fewer studies have looked at the potential effect of housing supply and housing cost

on metropolitan economic growth. Even fewer of these studies, only a single one in fact,

has traced through the effect on metropolitan employment and wage levels from regulations

and the regulation-induced inability to build housing quickly enough during periods of

economic growth.

We review the literature on the effects of development regulations on housing market

fundamentals, the limited empirical findings on the effects of housing market fundamentals

on labor market fundamentals (jobs and wages), and the even more limited empirical

findings on the effects of development regulations on labor market fundamentals at the

metropolitan level. To detect an influence of development regulation on metropolitan area

markets, a variety of different measures of regulation and regulatory restrictiveness have

been used. Given their importance and variety, we begin with a review of regulations

themselves. The second section reviews the large literature on the relationship between

development regulations and housing markets. We divide this literature into two types of

studies, those that test for a direct relationship between regulations and high house prices,

and those that model the effects of regulations on lowering supply elasticities, which are

then related to high house prices in the long term. The third section reviews the more

limited empirical findings on the effects of housing supply on labor market fundamentals

(jobs and wages) at the metropolitan level. We conclude with a discussion of the most

significant gaps in the literature with respect to the possible impacts of development

regulations—via higher housing costs and constrained housing supply responses—on

Workforce Housing Final Report 11

metropolitan economic activity. Importantly, the literature is largely silent on whether

development regulations influence the industrial mix of a metropolitan area over time.

Defining, Measuring and Comparing Regulations

There is no single, commonly accepted definition of regulatory restrictiveness. The term can

be loosely defined as the net limiting effect that local land use regulations have on

development in a given area. Land use regulations are defined as the set of controls that a

jurisdiction places over growth and development in that area. These controls may include

zoning regulations, environmental laws, historic preservation mandates, development fees,

procedural requirements for developers, or any of a number of government-induced

elements that affect land use and development. Though regulation is often seen as a term

that is interchangeable with restriction, regulations are not always negative, and as Pendall

et al. (2006) describes, regulations may be enacted as part of an area‘s overall growth

management framework, which may encompass both restrictions and incentives. Under this

broader framework, regulatory restrictiveness would be the net result of the balance

between land use restrictions and countervailing efforts to promote development such as

government incentives and affordable housing programs.

Economists, however, describe regulations more negatively as influences outside of what

would be imposed by an entirely free market which work to raise development costs or

reduce the quantity of development from levels that would otherwise be possible (Mayer and

Somerville, 2000). As they stated succinctly, regulations ―work locally to restrict

development by adding explicit costs, uncertainty, or delay to the development process.‖

Although federal and state governments impose regulations that influence the cost and

nature of residential development through environmental and other dictates, most regulatory

authority is delegated to and exercised by the city, town, or county governments. The

diffusion of regulatory power without any standardized set of controls across areas, has

made it difficult for economists and planners to define and identify similar degrees of

regulatory restrictiveness across metropolitan areas. This, of course, seriously hampers an

attempt to use cross-sectional analysis to determine the effect that such regulations may

have on the supply and cost of housing. Even if local governments across the country had

the same set of regulatory tools in their toolboxes, application of these tools varies across

the country. As Pendall (2006) finds, regions have different preferences for different types

12 Workforce Housing Final Report

of regulations, which limits the ability to compare restrictiveness across regions that use

different methods to achieve similar goals. Furthermore, thousands of cities, towns, and

counties regulating land use around the country have diverse attitudes toward development,

interpretations of model codes, and operating procedures that may not be entirely reflected

within the text of the local regulations themselves. In other words, two locations with the

same set of regulations may have entirely different administration of these regulations that

result in different actual limitations on development. Many development regulations grant

considerable discretion to local planning boards and zoning boards of appeal in how they

are ultimately implemented. Even if local governments across the country had the same set

of regulatory tools in their toolboxes, therefore, it would still be difficult to draw conclusions

from cross-sectional analysis.

With these difficulties in mind, several papers have made attempts at categorization of land

use regulations in an effort to develop serviceable measures of local or metropolitan area

regulatory restrictiveness for econometric modeling. Most studies focus narrowly on the

regulatory restrictions themselves, or what Schill (2004) calls the ―barriers to development.‖

One of the most concise and explanatory categorizations of these barriers is offered by

Quigley and Rosenthal (2005) who refer to the five types of regulatory restrictions proposed

by Deakin (1989) as follows:

1) Limits and geographic preference on the quantity or density of development

allowed (i.e. basic zoning)

2) Design and performance standards for lots and buildings (i.e. design guidelines)

3) Costs shifting from the locality to the developer (i.e. utility fees)

4) Reduction of the amount of land available for development (i.e. environmental

regulations)

5) Direct and indirect controls on growth, applied against buildings and population

(i.e. growth boundaries)

Pendall, et al. use a more simple categorization of land use regulation to differentiate

between simple zoning versus more involved limitations on the pace, location, or extent of

development.

Branching out from the taxonomical study of regulatory restrictions, studies of the effect of

regulations on house prices use either single measure or composite measure. Single

Workforce Housing Final Report 13

measure studies look at a single regulation thought to be important in order to determine its

effect on house prices or other economic measures. Findings from these studies are varied

(Fischel 1989). In his 2004 study, Schill notes that ―for some regulations, such as building

codes and environmental regulations, the literature barely exists. For others, such as land

use regulations and impact fees, many studies exist, but the results are often contradictory

and difficult to interpret.‖ Difficulty in measuring the effects of a single regulation stem partly

from the fact that regulations rarely appear alone, and are more commonly put in place

where other types of regulation already exist (Pollakowski & Wachter 1990, Shuetz 2005).

For instance, Dain & Shuetz (2005) find that of 187 communities in Eastern Massachusetts,

97 percent have subdivision by-laws, 70 percent have wetlands by-laws, and 60 percent

have local septic regulations. With the tendency for regulations to appear in groups, looking

at specific land use regulations as if they are independent of each other would lead to

underestimates of the impact of growth controls on local economics. (Pollakowski &

Wachter 1990). Therefore, regulatory influence on local economics must also include

effects of statewide regulations, and following Pollakowski and Wachter (1990) it is apparent

that regulatory influences on house prices also include extra-territorial spillover effects.

The need expressed by Pollakowski and Wachter to view specific land use controls ―in the

context of overall land use policy‖ suggests that the use of a composite-measure type of

study may be more appropriate. (Peterson 1974, White 1988, Linneman 1990, Malpezzi

1996, Saks 2005). Composite measures are built as indexes of regulatory activity.

Malpezzi (1996) offers a good example of how measures from surveys are built into

composite regulatory indexes. His index is as follows: (1) the change in approval time

(zoning and subdivision) for single family projects between 1983 and 1988 (1=shortened

considerably, 2=shortened somewhat, 3=no change, 4=increased somewhat, 5=increased

considerably); (2) estimated number of months between application for rezoning and

issuance of permit for a residential subdivision less than 50 units (1=less than 3 months,

2=3-6 mos., 3=7-12 mos., 4=13-24 mos., 5=GT 24 mos.); (3) Similar to (2) but time for

single family subdivision greater than 50 units; (4) how does the acreage of land zoned for

single family compare to demand? (1=far more than demanded, 2=more than demanded,

3=about right, 4=less than demanded, 5=far less); (5) how does the acreage of land zoned

for multifamily compare to demand? (1=far more than demanded, 2=more than demanded,

3=about right, 4=less than demanded, 5=far less); (6) percent of zoning changes approved

(1=90-100%, 2=60-89%, 3=30-59%, 4=10-29%, 5=0-10%); (7) a scale for adequate

14 Workforce Housing Final Report

infrastructure (roads and sewers), (1=much more than needed, 2=slightly more, 3=about

right, 4=less than needed, 5=far less than needed).

Although indexes like this provide a framework for gauging the relative regulatory

restrictiveness of different places, models that use them to derive definitive conclusions

about any specific effect of regulations themselves may still run into some difficulty. The first

such difficulty is that a highly regulated development environment may also proxy an area‘s

overall anti-development attitude which may induce impacts that have little to do with the

specific regulations incorporated in the index, but cannot be separated from the effect of the

regulations themselves within a model. Second, composite indexes have limited ability to

control for variations in regulations within the cities and towns that comprise a metropolitan

area. This is an important factor when such indexes are used to compare restrictiveness

across metropolitan areas. Third, building an index of regulatory restrictiveness requires a

certain number of subjective determinations in terms of the regulations included in the index

and the relative weight given to each measurement as a percentage of the total index value.

This is especially noteworthy for cross-metropolitan studies given that recent findings

suggest different regions prefer different types of regulations. (Pendall 2006)

Whether considered separately or part of an index, data on regulations are difficult to collect.

Data may be gathered through first hand research of documents from individual towns

detailing the regulations themselves. The time intensity of such an effort usually limits these

surveys to a small number of cities or towns within a metropolitan area. For instance, Dain

and Shuetz (2005)‘s ambitious compilation of regulations in 187 communities in Eastern

Massachusetts was used by Glaeser, Schuetz, and Ward (2006) to study whether

regulations, not land supply, have caused low levels of new construction and high house

prices in Greater Boston. Similarly, Greene (1999) looks at 37 suburbs of Milwaukee to find

the effects of six specific land use restrictions on house prices, and Pollakowski (1990) looks

at time series data from 17 planning area groups in Montgomery County MD to study the

influence of spillover effects of restrictions in neighboring jurisdictions on local house price

changes.

To extend the geographical range, studies may utilize second-hand surveys of local

planning or regulatory authorities. Pendall et al. (2006), use this method to gather

regulatory information from nearly 1500 cities, towns, and counties in the top 50

metropolitan areas. Other studies such as Saks (2004) and Malpezzi (1996) have used a

Workforce Housing Final Report 15

combination of cross metropolitan surveys. Unfortunately, aside from Pendall and the

recently updated Wharton study (Gyourko, Saiz, and Summers, 2006), the most recent of

these surveys were published over ten years ago and are based on data collected several

years before that. Some of the more widely used cross-metropolitan studies are:

1) The Wharton Land Use Control Survey 1990. This has been perhaps the most

widely used survey in recent studies of regulatory restrictiveness. This is a

survey of local officials on the various aspects of their zoning and other policies

that may limit development. Among the most widely used indicators from this

survey are time-to-approval measures, such as the average length of time for

subdivision permits to be approved or the change in zoning approval time from

1983-88.

2) Srinivasan and the Regional Council of Governments (1985) survey local officials

to estimate of the amount of otherwise developable land rendered un-buildable

by regulations. While appealing, the results are ultimately subjective.

3) The American Institute of Planning in1976 published its Survey of State Land

Use Planning Activity, which indicates the presence of various land use policies

for each of the 50 states.

In sum, there are six principal problems in measuring development regulations for the

purposes of evaluating their individual or collective impact on housing markets. First, the

complexity of regulations and variations in administration among thousands of local

regulatory agencies means there is no shared standard among jurisdictions or metropolitan

areas. Second, the diffusion of regulatory authority and diversity of regulations also makes

data collection difficult, and leads most studies to rely on a small number of dated surveys,

each with its own inherent flaws. Third, there is an inherent difficulty disentangling effects of

specific regulations and constellations of regulations. Fourth, once a bundle of relevant

measures of regulatory restrictiveness are chosen for use within a model, there remains an

inability to conclude whether any measured effects are caused by the specific regulations

themselves or a broader opposition to development they may represent. Fifth, by looking to

model whether areas with similar regulations have similar changes in house prices, models

fail to capture differences in enforcement of regulations across jurisdictions or admit the

possibility that jurisdictions with similar restrictions may apply them differently. And lastly,

composite models of regulatory restrictiveness rarely focus on intra-metro variations in

restrictiveness.

16 Workforce Housing Final Report

Regulations and House Prices

The theoretical link between regulations and house prices and construction levels is well-

established, though at times complicated. Nearly all empirical studies find that regulatory

restrictions add to development costs and reduce construction levels either directly or

indirectly through one or more of the following actions:

restricting land available for development;

reducing unit densities of development;

adding delay and risk to the development process;

adding directly to costs.

Two general types of study test either for a direct relationship between regulations and high

house prices, or for a direct relationship between regulations and lower supply elasticity,

which translates into high house prices over time. Studies that model regulatory effects on

house prices directly tend to focus on factors that limit supply levels or impose undue costs

on development. Elasticity models, while also focusing on factors that impose costs, add

emphasis to factors that delay the development process and therefore impact supply

responses to shocks in demand. Both types of study focus on long-term impacts and may

include both cross-sectional or cross-year survey data. Elasticity models, however,

incorporate short-term supply reactions to demand and therefore take the secular increase

in high-housing costs as the cumulative impact of depressed supply elasticities in a

metropolitan area subject to constant demand shocks.

Figure 1: Modeling Regulatory Effects on Housing: Direct vs. Elasticity-based

(A) Direct Link

(B) Dynamic Depressed

Supply

Levels

Regulations

Low Supply

Elasticities

High House Prices

High House Prices

Regulations

Demand

Shock

Workforce Housing Final Report 17

Direct Effects on House Prices and Construction Levels

Most studies on regulations and the housing market look to establish evidence of a direct

relationship between restrictive regulations and house prices and/or construction levels. In

general, higher housing prices and lower construction levels are attributed to restrictions on

land available and unit densities, delays in the development process, and development fees

in the following manner:

Restrictions on land available and unit densities. Steady-state models of regulatory

effects on house prices share greatest agreement in showing how restricting unit densities

or land available for development increase long-term house price and construction levels by

constraining total supply to levels below that which would otherwise satisfy demand in that

location. Studies incorporating these types of regulations model the steady-state house

price and construction levels to show long-term impacts on metropolitan area housing

market. Representative studies include Segal and Srinivasan (1985), who used their survey

of local officials to find that house price inflation had a nonlinear relationship with

percentages of land removed from development by regulation, resulting in significantly

greater house price and appreciation levels for the most restrictive cities when compared to

those unrestricted cities. Glaeser (2006) uses a collection of regulatory and house price

data from Eastern Massachusetts municipalities to find that high house prices in Boston are

not based on a shortage of developable land, but on the rights associated to the land that is

developable. Density based regulations are also linked to high house prices within Glaeser,

Schuetz and Ward (2006) where median house prices in municipalities near Boston

correlate to minimum lot size allowed in their zoning by-laws.

Models on house prices are also commonly paired with similar models associating

regulations to reductions in steady-state construction levels. For instance, Glaeser et al.

(2006) also uses models to show that regulatory environments reduce steady-state

construction levels in Manhattan and Greater Boston to far below what would have been

constructed if they had been less regulated areas. Similarly, Katz and Rosen (1987) study

house prices in several San Francisco municipalities in 1979 and find house price in areas

with growth control plans or growth moratoria, which simply do not allow development in

certain areas, were 17-38 percent higher than in those without.

18 Workforce Housing Final Report

Delays in the Development Process: Several studies look at the effect of delay

variables on steady state price levels. The basic measure of delay within most recent

studies is taken from the Wharton Land Use Control Survey, which asks for the average

time between applications for rezoning and issuance of building permits. This variable is

indexed in Gyourko and Glaeser (2003), as a measure for overall regulatory restrictiveness,

and is found to have significant positive relationship with house prices. Delay variables are

also used as part of a more comprehensive regulatory restrictiveness index constructed by

Malpezzi (1996), which is also used in Malpezzi Chun and Green (1998), which as part of an

index is found to correlate with higher prices, but singular effects of delay itself are not

identified. Malpezzi (2002) uses the same index and finds regulation effects on prices not

only significant, but substantial, and that a one unit increase in the regulation index (which

ranges from 15-30) amounts to an 8.5 percent increase in house prices. Mayer and

Somerville (2000) find that delays in the approval process have a significant negative effect

on the level of construction in a metropolitan area, lowering ―steady-state‖ level of housing

starts by up to 45 percent in a MSA, with greatest impact given to those regulations that

lengthen development times.

Development Fees: The impact of development fees has a complex relationship to house

prices. As stated by Evans (2003) theory states that absent close substitute markets,

builders will respond to high impact fees by passing the cost on to the consumer through

higher house prices, or possibly ignoring lower-income households and focusing on more

expensive housing where the impact fee can be more easily passed on (Huffman et al.

1988). Empirical studies largely show that development fees increase house prices, but

differ on who bears the burden of the fee and how fees translate to added costs to

consumers in the form of higher house prices. Two studies identified in the Evans (2003)

review separate effects of development fees using time series data on areas before and

after impact fees are introduced. The first study is Delaney and Smith (1989a), who build a

constant-quality index to find that impact fees in one community had a positive impact on

new home prices relative to two neighboring communities, and the second is Singell and

Lillydahl (1990), who find an increase in prices and the pace of house price appreciation in

Loveland, Colorado, after impact fees are introduced. A more recent study by Ihlanfeldt and

Shaughnessy (2002) finds that a $1.00 increase in development fees increases the price of

both new and existing housing by about $1.60, an amount attributed to the tax savings

expected by homeowners in a fee-based system. The study also finds that land prices are

Workforce Housing Final Report 19

reduced by the amount of the development fee, suggesting burdens of the fee are born

primarily by owners of undeveloped land. (Anderson 2003)

Models on the relationship between development levels and impact fees have resulted in

conflicting findings. Looking across metropolitan areas at the effect of impact fees on

development, however, Mayer and Somerville (2000) conclude that the presence of

development fees do not show a significant effect on construction levels given the existence

of long approval processes and other growth management techniques in an area. In

contrast, Skidmore and Peddle (1998), who study a sample of all municipalities in DuPage

County, Illinois, from 1977 through 1992 find that impact fees reduce residential

development by over 25 percent.

Regulations, Supply Elasticities, and House Prices

In contrast to studies looking to establish evidence of a general, long-term relationship

between regulations and other housing market factors, a second type of study tests for a

more direct relationship between regulations and the elasticity of the housing supply.

Supply elasticity models explore the relationship between regulatory restrictiveness and

construction levels given changes in demand fundamentals, such as would occur in periods

of rapid economic expansion. Within elasticity models, there is a focus on regulations that

impose additional costs or delay to the development process. This follows the theory that

new regulations that add direct costs to the development process initially depress

development because in the short run, these costs are typically borne by the developer

(Huffman et al. 1988). Regulations that add delay or risk to the development process

similarly reduce production responses by increasing the reaction time builders need to

respond to any increases in demand. These factors are then indirectly associated as supply

side determinants of house price dynamics to show positive shocks in demand are forced to

translate into positive shocks in house price given an inelastic supply. Theoretically, while

low supply elasticity translates into higher house prices in the short-term, it may also affect

prices in the long term as a metropolitan area experiences multiple demand shocks to which

its supply levels never fully respond. Models confirm expectations that if regulations lower

steady state construction levels in a metropolitan area, they will also inhibit an area‘s ability

to respond quickly and completely to shocks in demand fundamentals with additional

housing.

20 Workforce Housing Final Report

Example studies modeling the effect of regulations on housing supply elasticities include

Quigley & Rosenthal (2004) who model this supply elasticity for cities in California,

differentiating between those minimally regulated, with less than two growth control

regulations, and the highly regulated with two or more. Given exogenous predictors of

increased housing demand, responsiveness of the housing stock via new construction is

weaker in more regulated cities relative to less regulated cities. Mayer & Somerville (2000)

also model the elasticity of housing supply and find similarly that regulation lowers elasticity

of housing supply by up to 20 percent in a highly regulated MSA, and appropriately find that

elasticities are especially low for metros that have regulations lengthening development

times. Glaeser, Gyourko and Saks (2006) also state how limited supply response to

demand for housing in Manhattan ―primarily is the consequence of an increasingly restrictive

regulatory environment.‖

House Prices and Labor Markets

Though not modeled in the Quigley and Rosenthal nor the Somerville papers, the fact that

restrictive regulations have a significant negative effect on metropolitan area elasticities of

housing supply suggests that regulations may also have a significant negative effect on job

growth in highly regulated areas during economic boom times. The same holds true for

population growth. In other words, it follows from these studies that slowed growth in the

housing stock means actual employment and population growth in response to growth in

demand for employment and population is also depressed, and that both are associated

with restrictive regulations. However, while many studies look at the effect that regulations

have on house price measures, a much smaller subset of these studies link regulatory

restrictiveness to a metropolitan area‘s ability to grow its economy.

The primary link between regulations and economic growth is through the area‘s ability to

grow its housing stock, and therefore its population of workers, in response to increases in

positive housing demand shocks created by expansion of employment in growing industries.

Comparative studies have associated reduced levels of metropolitan-level immigration,

population, and employment growth with high housing costs, but none have also included

links to regulatory restrictiveness – a major contributor to the high housing costs.

The link between high housing costs and reduced immigration is established in several

studies. Case (1991) finds the housing boom in Massachusetts 1984-1987 created a

Workforce Housing Final Report 21

significant increase in demand for labor, but also contributed to a slowdown in the growth of

the labor force. Case gets his findings by isolating the effect of high home prices on labor

force growth empirically and demonstrating that high home prices discourage labor force

entry. Referenced in the Case study are results from Gabriel (1991), who concludes in his

study that house price differentials between metropolitan areas offset incentives to migrate

to regions with tight labor markets and deter migration from lower cost to higher cost

regions. 1

Figure 2: Models Linking House Prices and Labor Markets Generally Do Not Include

Regulations

House Prices

Migration Rates

Employment Levels

Wages

A related set of studies explores the relationship between metropolitan-level housing costs,

employment levels, and wages. These studies find that, given similar predicted values of

employment and wage changes due to shifting fundamentals, areas with high housing costs

have lower job growth and higher wage increases in response to increases in demand. An

1 See also Bluestone (2006) who looks for nationwide data to support the theory that high house

prices are negatively impacting job growth and migration in Massachusetts, and using a steady

state OLS model of house prices on employment and migration with data from 300 metropolitan

areas, finds some evidence that housing costs play a role in employment growth and that after

controlling for the strength of the labor market, there is an independent and statistically robust

effect of housing costs on metropolitan migration patterns.

22 Workforce Housing Final Report

example study is Johnes (1999) who examined interactions between local housing and labor

markets in the Hartford, Houston, Fort Lauderdale and Milwaukee areas using quarterly data

for the 1980s. The study looks at dynamics, and specifies an error correction model with

reduced-form equations explaining the average wage, the unemployment rate, the labor

force and the average house price in an urban area 2 and find some evidence that

unemployment and labor force changes affect house prices and that house prices have a

significant effect on the size of the labor force. Bover (1989) looks at the effect of high

house prices on migration and wages in the South East UK. Relating growth in the area

house-price-to-earnings ratio as a proxy for a positive shock to employment demand, which

one would think would increase supplies of local immigration to these areas to meet

demand, Bover instead adds to several studies that report less intuitive negative correlations

between these two measures. Bover‘s major contribution is in finding that areas with the

lowest supply elasticities saw the lowest net migration levels, the highest wage levels, and

the greatest increases in price to earnings ratios given nationwide increases in labor

demand. Bover briefly mentions planning and zoning constraints as an example of an

institutional distortion that would theoretically raise returns on owner occupied housing, but

stops short of modeling relationships between these constraints and the regional trends in

employment and wages.

Regulations and Labor Markets

Given the wealth of studies finding that regulatory restrictiveness constrains metro-area

housing supplies and increases costs, and the additional body of literature on the

relationship between high housing costs and reduced inter-metropolitan migration and flows

of labor, reduced employment growth, and higher wages, it follows that there would be a

relationship between regulatory restrictiveness and labor dynamics. This is mentioned

briefly in Bover, but only a single study, Saks (2004) empirically connects the path from

regulatory restrictiveness to labor markets. The following section details this study.

2 (The reduced form house price equation is modeled with explanatory variables of local wage

level, interest rates, and the stock of houses available in the previous period.)

Workforce Housing Final Report 23

Figure 3: Regulations and Labor Market Dynamics: Connecting the Link Category 1: Modelling Effects of Regulations on House Prices

(A) Long Term

(B) Dynamic

Category 2: Modelling Effects of High House Prices on Labor Markets

Category 3: Modelling The Effect of Regulations on House Prices and Labor Markets

Regulations

Supply Elasticities

House Prices

House Prices

Migration Rates

Demand

Shock

House Prices Regulations

Employment Levels

Wages

House Prices Regulations

Employment Levels

Wages

Depressed

Supply

Levels

Supply Elasticities

Demand

Shock

Depressed

Supply

Levels

Saks (2004) is the first study to empirically determine a relationship between regulations and

labor supply through a series of modeled relationships. Ultimately, the study finds that the

inability to respond to demand shocks with housing fast enough or completely enough, such

as occurs in areas heavily constrained by regulations, leads to a suppression of labor

demand, (i.e. fewer actual jobs resulting from labor demand shocks) in both the short and

long term.

The study begins with a simple model relating variation in the elasticity of housing supply to

local regulatory constraints through estimating parallel regressions of annual changes in the

logarithm of housing prices and the housing stock on the labor demand shocks and an

interaction with the index of housing supply regulation. Two key variables in this model are

the regulatory restrictiveness index and the predicted labor demand. The restrictiveness

measure is created for this study as a single, comprehensive, composite index of

restrictiveness according to variables from six other indexes, including the Wharton survey.

The second key variable is that of predicted change, or shock to labor demand, which

24 Workforce Housing Final Report

follows the methodology of Bartik (1991) and is based on state and national employment

trends for the industries heavily represented in a metropolitan area.

The resultant model is as follows:

pit = θ hit + π rihit + xi + dt + Є it

where i indexes metropolitan areas and t indexes time, p is the log of house prices (or

housing stock in the parallel model), h is the annual changes in housing demand (labor

demand shocks), r represents regulatory restrictiveness, x and d represent location and time

specific fixed effects, and Є represent the error term.

Results from the study find that, on average, a 1 percent increase in labor demand is

associated with a 0.25 percent increase in the housing stock and a 0.8 percent increase in

housing prices. The interactions show that where a 1 percent increase in demand would

lead to a 0.35 percent increase in the housing stock in an area with an average amount of

housing supply regulation, the effect of same demand shock would be 17 percent smaller in

an area with a 1 standard deviation higher value of the housing supply regulation index.

The second stage of this study focuses on the effect of housing on labor markets. The

model used is based on the labor dynamics model of Blanchard and Katz (1992) with added

variables to explicitly indicate housing market characteristics such as the regulatory index,

the elasticity of housing supply, and fixed city-specific factors that may cause persistent

differentials in house prices across metropolitan areas. The resultant three VAR models

include the change in the logarithm of employment, the logarithm of wages, and the

logarithm of housing prices each as a function of two of its own lags, two lags of the other

endogenous variables, and the contemporaneous labor demand shocks with the following

form:

To examine how labor and housing market dynamics vary with the elasticity of housing

supply, variables for wages, employment levels, and house prices are interacted with the

Workforce Housing Final Report 25

index of housing supply regulation. Interactions result in findings that although demand

shocks increase employment, wages, and house prices on average, regulations have a

significant impact on above average wage responses and below average employment

responses to demand shocks in a metropolitan area.

The model also shows regulations may have a lasting impact on wages as long as housing

prices are also permanently higher. The long-run impact of a 1 percent demand shock in a

highly constrained market results in only a 0.9 percent increase in employment, instead of

the 1 percent increase found in less constrained areas. Additionally, measured over a

twenty-year period, in New York, which was estimated to have the most inelastic housing

supply, a 1 percent increase in labor demand leads to only a 0.65 percent increase in

employment.

Lastly, the study models aggregate output indicators on regulatory restrictiveness and finds

that constrained employment growth due to housing regulations has an aggregate effect on

the economy lowering the gains from migration by about 18 percent in constrained areas

relative to those with fewer constraints.

Though literature on the effects of regulations on labor markets is sparse, Saks (2004)

covers much ground, especially in noting that findings refute assumptions made in models

such as Blanchard and Katz (1992) that assume relative wages across local areas tend to

converge over time. Following this work, regulatory restrictiveness can be shown to inflate

house prices in long and short term, inflate wages to compete for labor in the face of higher

costs of living, suppress demand (quantity of employment) given similar shocks to the

economy, and increase sprawl (land consumption), longer commutes, and lower than

optimal metro area productivity.

Regulations and Industrial Compositions

While Saks details the effect of regulatory restrictiveness on aggregate employment and

wages, no known studies detail the effect of restrictiveness on industry composition.

There is established theory of how relative wage differentials affect industry dynamics and

inter-metropolitan industry composition. (Blanchard and Katz 1992) Several non-empirical

studies point to the implied relationship between high housing costs and industry

26 Workforce Housing Final Report

composition of certain metropolitan areas. Cheshire (2004) refers to Flammang (1979,

1990) and states that rising incomes ―imply rising costs and a continuing loss of lower value

added activity,‖ and lists ―travel agencies, retail, real estate‖ and other services as

disproportionately located in large cities as a sign of transaction economies (Stein 2002), but

actual industry composition is not tested in relation to high-cost metros or regulatory

restrictiveness. Other studies give statements that industries leave high-cost areas (Glaeser

2006), but limited evidence pointing to this relationship.

Malpezzi (2002) comes closest to the topic with an empirical study of the relationship

between aspects of the ―new economy,‖ land use and development regulation, and housing

prices. Malpezzi tests whether a single industry, high-tech, is associated with regulations,

given the assumption that high-tech industries are footloose and high-paying and therefore

are attracted to either a) high-amenity areas that are ―the foundation of local regulation‖ or b)

traditional high-growth areas that have put regulations in place to stem this growth. Using

the regulatory restrictiveness index developed in Malpezzi, Chun, and Green (1998), and

the DeVol (1999) instrument for the area‘s high-tech quotient, Malpezzi fails to find any

support for a relationship between high regulation and high-tech industries.

Summary and Research Extensions

What we know: much literature has linked restrictive development regulations to higher

prices, and other studies have linked impaired supply responses to lower migration levels

and employment gains during times of economic expansion, but few study empirically

connect housing regulations and labor markets, and only one study links restrictive

development to lower employment gains and higher wages during expansions. Saks (2004)

also gives direct support that regulatory restrictiveness also has significant long-term

impacts on steady-state employment and wage levels that had previously been buried within

metropolitan fixed effects of other models.

This review finds that there are several questions that remain at the macro level. The first is

how regulatory restrictiveness affects metro area industry composition and change. While

there is an established body of knowledge on how wage differentials affect inter-

metropolitan industry compositions, and Saks (2004) has linked how regulatory

restrictiveness contributes to these wage differentials, an elaboration of how regulations

Workforce Housing Final Report 27

affect industry mixes and entry and exit of industries within a metro have not been subject to

rigorous empirical study.

The second question involves how the diversity of regulatory restrictiveness within metro

areas affects housing and labor market behavior within cross-sectional studies. Indexes on

a metropolitan scale are forced to agglomerate effects of both restrictive and non-restrictive

regulatory jurisdictions with diverse attitudes toward new development. This approach

encompasses, for instance, how the existence of a few low-restriction areas eager to attract

new development may impact the overall dynamics of an otherwise highly restrictive

metropolitan area in terms of both housing supply, as suggested by Pollakowski and

Wachter (1991) and labor force.

A third remaining question involves the age and reliability of available data on regulatory

restrictiveness. As mentioned in Pendall (2006), the absence of a detailed national

database with up-to-date information on local regulations inhibits cross-metropolitan study of

the impacts of regulations. In order to obtain a greater degree of rigor and significance in

regulatory studies, there remains a need to update and improve measurements of regulatory

restrictiveness that are both recent and expansive; standardized or normalized across metro

areas; and reflective of differences in strictness of enforcement across jurisdictions.

Creation of such a database would require a substantial effort to compile and maintain. The

alternatives would be a continuation of the numerous limited one-off surveys every few

years, or the unlikely scenario of achieving greater standardization of the regulations

themselves in terms of creation and implementation on the state and local level. Truly the

relationship between regulations, housing, and labor markets is complex and much remains

to be known.

28 Workforce Housing Final Report

Chapter Two: Business Perspective

High housing costs and limited housing availability logically have an adverse impact on

businesses and jobs, but that impact is difficult to demonstrate, in part because there are such

strong influences in the opposite direction. Rapid growth in output and employment in a local

area tends to drive up housing prices and create supply shortages, so that there is generally a

positive correlation between changes in employment and housing cost. Also, places where

housing is limited and expensive also tend to have other characteristics, such as highly

educated populations, that have been favorable for growth.

Saks (2005) has provided unprecedented evidence that job growth in areas with highly

regulated housing markets would have been even greater without the housing constraints.

The local area industry mix plays a central role in Saks‘ analysis, since the estimate of

unconstrained ―shocks‖ to local employment is calculated as the employment growth that

would have occurred if each industry in the area matched growth for that industry elsewhere.

But the local industries that failed to reflect national industry growth are not identified.

Knowing which industries and occupations are most affected by housing supply problems can

be useful in several ways. The businesses that are most vulnerable to adverse risks from

housing may be persuaded to become involved in efforts to address shortages. Specific

housing needs may be anticipated in local planning. Officials in areas with favorable housing

supply conditions can select promising industry targets for development.

Certain characteristics of industries suggest greater sensitivity to housing costs. For example,

industries with a greater share of total costs attributable to employee compensation and a

larger share of workers in less-skilled, lower-paid occupations are likely to be more vulnerable.

Firms may locate in an area in order to serve the local market, to have access to natural

resources or transportation facilities in the area, or because of other attractions, including

the presence of other firms.

Businesses that primarily serve local populations tend to be distributed in proportion to

population or labor force. Those that serve broader markets, however, will be less evenly

distributed among areas. The ones serving broader regional, national or global markets are

Workforce Housing Final Report 29

referred to as ―export base,‖ ―economic base,‖ or ―basic‖ industries. (See Schaffer 99;

Fujita, Krugman, and Venables 2001, Ch. 3)

Businesses serving regional, national, or global markets may be bound by ties to the labor

force, local suppliers, or other factors, but those ties could become frayed if housing and

other costs increase, and those activities are potentially mobile. Businesses involved in

such activities could relocate to other areas, but more likely there would be a shift in activity

to other areas as competitors develop or grow elsewhere and local businesses contract or

close.

To the extent that activity shifts elsewhere, it will affect not only businesses in footloose

industries but also the businesses serving the local market. Distinguishing between

activities that are strictly local and those ―exporting‖ to broader markets will, however,

contribute to understanding the impacts of housing supply.

The sections that follow provide some background, preliminary analysis, and research

suggestions related to several topics relevant to the understanding of the impacts of housing

supply conditions. The topics include:

Industries and Employment

Business Location and Agglomeration Effects

Business Dynamics

Metropolitan Labor Supply

Evident Impacts of Housing

Industries and Employment

From 1955 to 2005, the share of (full-time plus part-time) jobs in goods-producing industries

(including construction) fell from 37 percent to 17 percent, with manufacturing industry jobs

declining from 27 percent to 10 percent. The non-construction goods share of real GDP

actually increased, but disproportionate growth in productivity, along with the fact that more

of the inputs used by factories and other goods producers consisted of services rather than

materials, changed the mix of employment.

30 Workforce Housing Final Report

Table A-2 shows the total number of workers and value added by industry in 2005, along

with the average share of value added devoted to employee compensation and to taxes (net

of subsidies) and gross operating surplus. Value added and compensation per worker are

also shown. In capital-intensive industries, such as utilities, employee compensation

represents a smaller share, but for many industries employee compensation represents over

two-thirds of value added.

For some of the industries with high labor costs, employee compensation per worker is less

than lavish, but there are many workers. Other industries face high labor costs with

relatively fewer workers, but with more of those workers possessing high skills and receiving

high pay.

Table A-3 shows the distribution of employment for industries among major occupational

categories. Rows showing average annual wages and salaries and total employment for

each occupation, and columns showing similar aggregates for each industry are also

included in the table. For industries such as computer manufacturing, data processing, and

professional and technical services, more than half of all jobs involve management and

professional occupations, with average wages/salaries correspondingly high. The industries

with most jobs in low-skilled occupation may be the ones especially sensitive to housing

costs. 3

All industries involve a mix of occupations, and one response to high housing costs in

particular locations is to separate support activities from those involving key skilled workers.

The support jobs may be done in areas with lower wage rates and housing costs, such as

South Dakota or Bangalore. This type of arrangement is facilitated by improvements in

communications. (See Cohen 2000). Elvery (2006) finds the mix of occupations within

industries is skewed toward higher skills in larger metro areas, although the analysis does

3 The data in the table come from the BLS Occupational Employment Statistics survey, which

collects information from over a million businesses on a rotating 3-year cycle. Information on

number of workers, hourly wages, and annual wages are gathered for 801 detailed occupations.

Although data for occupation by industry are also collected from households in the CPS, ACS,

and decennial Census, the OES data are probably more reliable, and have finer industry and

occupation breakdowns.

Workforce Housing Final Report 31

not specifically look for whether there is a link to living costs or whether the same firms place

less-skilled work elsewhere.

Business Location and Agglomeration Effects

Industries tend to be clustered in particular areas to a degree that goes beyond the influence

of proximity to markets or proximity to raw materials or transportation. This clustering has

been attributed to the advantages of ―agglomeration.‖ The idea is that the presence of

other, similar firms in an area reduces costs or creates other advantages and attractions.

The term ―agglomeration‖ has also been applied to concentrations that occur because of the

presence of a natural advantage such as mineral deposits, favorable climate, or a body of

water, but recent emphasis has tended to fall on the benefits of being around other firms

and the entourage of workers, institutions, and infrastructure their presence adds to the

locality. To distinguish the specific effects of co-location, the term ―agglomeration

externalities‖ will be used here.

Analyses and conjecture extending back at least to Alfred Marshall (1920 {first edition

1890}) have identified agglomeration externalities such as the presence of a cadre of

appropriately-skilled workers, the presence of specialized suppliers, and the flow of

information derived from the presence of similar firms. Agglomeration externalities have

been attributed to concentrations of similar firm (localization economies) and/or to the

overall size of the local area (urbanization economies). 4

Many attempts have been made to explain the micro-foundations of urban agglomeration

economies. Duranton and Puga (2004) provide a good review of the micro-foundation

literature, which essentially extends along three dimensions: sharing, matching, and

learning. Sharing refers to the sharing of indivisible goods and facilities that would be too

expensive for an individual or single firm to support, but these facilities are feasible when the

fixed costs are spread over many users. Athletic stadiums, markets, and airports are

common examples. Another form of sharing is the sharing of risk though labor pooling.

Matching refers to the productivity gain when a worker has the right experience to meet the

specialized needs of the employer. Matches should improve because there are more

4 The argument for urbanization economies is typically attributed to Jacobs (1969), and such

effects have also been termed ―Jacobs externalities.‖

32 Workforce Housing Final Report

agents trying to make a match (Helsley and Strange 1990). A better match between worker

and employer reduces the need for turnover and re-training. Firms devote substantial

resources to learning and innovation. Interaction with like-minded researchers, whether

private, academic, or government employees, promotes the diffusion process.

The presence and relative importance of localization economies and urbanization

economies has been debated in the literature (Strange 2005; Rosenthal and Strange

2003,2004; Shefer 1973; Moomaw 1981; Tabuchi 1986; Ciccone and Hall 1996, Wheaton

and Lewis 2002).

As an empirical matter it may be difficult to distinguish the effects of agglomeration

externalities from the advantages of a location that would be present even if there were no

other similar firms present. The hypothesis that there are benefits from agglomeration

independent of the natural resources becomes, after all, that many firms are there because

many firms are there. Some attempts have been made, however, to distinguish between

natural advantage and agglomeration externalities (e.g., Ellison and Glaeser 1999;

Chatterjee 2003).

A common measure used to identify industry concentration is the ―location quotient.‖ A

metropolitan area‘s location quotient for an industry may be calculated as the industry‘s

share of total employment in the area divided by the industry‘s share of national

employment. Thus, if 25 percent of workers in an area worked in manufacturing, and 10

percent of national employment was in manufacturing, the location quotient for

manufacturing in that area would be 2.5. Location quotients substantially above 1.0 suggest

an export base industry, although relatively high location quotients could just reflect high

local demand, and some industries with low location quotients may primarily serve

customers outside the local area.

Table B-1 shows the range of location quotients for the 100 largest metropolitan areas for

various industry categories. Location quotients could not be calculated for every industry in

every MSA, because employment was not reported uniformly. The column labeled

―reporters‖ shows how many MSAs had data. The table includes the lowest value, the 10 th

percentile, median, 90 th percentile, and highest for those that reported. As expected, the

range of values is narrower for industries like retail trade that primarily serve the local

Workforce Housing Final Report 33

population, while industries that serve broader regional, national, or global markets have a

wide range of location quotients.

The column labeled ―combined‖ reflects total large-metro industry employment as a share of

total overall employment, based only on the MSAs reporting employment for that industry.

For example, among the largest 100 MSAs, data for employment in Fabricated Metals

Manufacturing were reported for 88 MSAs, which collectively had 882,000 private

Fabricated Metals jobs among 81,592,000 total non-agricultural payroll jobs in 2005, or 1.08

percent. Fabricated Metals represented 1.15 percent of national employment, so the

combined location quotient was 0.94 (1.08%/1.15%).

The 12 MSAs for which employment in Fabricated Metals was not reported probably had

smaller shares of employment in the industry than those for which data were available, so

the combined LQ in the table is probably biased upward. Even so, the location quotients

shown for manufacturing industries and for other goods-producing industries (except

construction) are generally less than 1.0. Goods-producing industries are more likely to be

located in smaller metropolitan areas or in nonmetropolitan areas. For most manufacturing

industries, location quotients for large metropolitan areas were already less than 1.0 in

1990, and the shares of industry jobs in large metros declined further between 1990 and

2005. Notable exceptions include apparel and computers, where substantial employment

declines in large metropolitan areas were exceeded by even greater reductions in

nonmetropolitan areas.

Location quotients are a simple and popular measure of industry concentration, but are

crude statistics. Extreme values will be more common for areas with less total employment

and/or for industries where individual establishments are large (Ellison and Glaeser 1997;

Naude 2006). For the large MSAs and aggregated industry groups shown here, that should

not make a huge difference.

Location quotients measure the degree to which a particular area has a disproportionate

share of an industry, not the degree to which an industry is concentrated in a few areas.

The distribution among areas of location quotients provides some indication of the degree of

concentration of an industry, but there are other, possibly superior, measures. One

possibility is to use GINI coefficients, similar to the way such statistics are used in analysis

of income inequality (Krugman 1991). More sophisticated methods incorporating

34 Workforce Housing Final Report

adjustments for plant size have been developed by Ellison and Glaeser (1997) and Maurel

and Sedillot (1999).

The concentration of people with high income and wealth, and perhaps skill, in ―Superstar‖

cities such as San Francisco and Boston where housing supply has been constrained has

been documented by Gyourko, Mayer, and Sinai (2006). They find that in-migration to such

places is heavily skewed toward high-income people. They don‘t find equally compelling

evidence that out-migration is skewed toward low-income people. The relationship of

housing supply to the concentration of high-skilled people and to income distribution among

areas was not found to be significant in studies such as Berry and Glaeser (2005). But

Glaeser elsewhere (2006) expresses the opinion that there is an effect, such that ―in the

long run, firms generally leave high-cost areas.‖

Concentrations of industries, and indeed, the locations of cities, reflect history. There are a

variety of factors such as investments in fixed capital and moving costs that would tend to

perpetuate industry concentrations even after the influences or happenstance that created

them are gone. But compared to location choices based on local service functions or

immovable resources, the clusters caused by agglomeration externalities may be more

changeable.

Several of the key theorists of agglomeration economics and ―New Economic Geography‖

indicate the possibility of ―catastrophic bifurcations‖ (Fujita and Mori 2005) under which ―the

agglomeration collapses suddenly‖ (Strange 2005). These descriptions generally refer,

however, to the properties of abstract models, rather than to any particular real-world

experience (Martin and Sunley 1996).

The New Economic Geography literature is a variant on agglomeration economies in that

firms are assumed to have increasing returns to scale. Based on a Journal of Political

Economy article by Paul Krugman (1991), New Economic Geography (NEG) models are

distinguished by five essential ingredients (Head and Mayer 2004):

1) Increasing returns to scale based on fixed overhead that are internal to the firm.

2) Imperfect competition, usually some form of monopolistic competition with limited

pricing power.

3) Trading costs for transport and sale of inputs and outputs.

Workforce Housing Final Report 35

4) Endogenous firm locations, entry and exit based on expected profitability with

increasing returns to scale favoring large plants serving their customers from a

distance.

5) Endogenous location of demand, consumption is mobile either because workers

consume where they work or firms use the output of other firms as intermediate

inputs.

The net effect of these five ingredients, especially the endogenous location of demand, is

that agglomeration effects build up some regions and pass over other regions largely

independent of the original distribution of natural resources. Natural advantages and human

capital externalities are alternative explanations for uneven economic development. The

emphasis of NEG is on economic development through trade built up from production based

on increasing returns and distribution limited by trading costs. Land use regulation fits into

the NEG perspective in that regulation can add significantly to the fixed cost of a project,

whether commercial or residential. Zoning not only restricts where firms can locate

production and distribution facilities, the requirements for linkage fees, extensive impact

assessments, and protracted negotiations mitigate against small-scale development.

One of the implications of NEG models is that wages should vary by region according to the

market potential of that region especially in trade with surrounding areas. Hanson (1998)

estimates a log wage equation in which wages for a given county are determined by the

trade potential with surrounding counties as captured by their personal income, wage rate,

and housing stock. Essentially, if people in other counties had more income to spend on

non-housing consumption, their demand would increase the own county wages. Although

the model fits the spatial variation in wages reasonably well and validates the NEG theory,

the expenditure share on the traded good is estimated between 0.91 and 0.97. These

values leave too little room for expenditure on housing, the non-traded good.

Hanson‘s market potential model is estimated assuming that trade costs are a power

function of distance, but this implies that demand shocks disproportionately affect wages

and house prices in nearby regions. Hanson (1997) estimated Mexican wages following

trade liberalization with the U.S. That study shows that a 10 percent increase in distance

from employment centers in Mexico City reduces wages by 1.92 percent and the same

increase in distance from the U.S. border reduces wages by 1.28 percent.

36 Workforce Housing Final Report

Kim (1995) estimates the relationship between industry concentration and scale economies

using panel data on manufacturing over the very long run, 1860 to 1987. Regional

specialization rose from 1860 to the turn of the century and remained stable before declining

after the Second World War. Kim regresses Gini indexes for twenty 2-digit industries on the

number of production workers per plant (as a proxy for internal scale economies) and the

ratio of raw materials to value added (as a proxy for resource intensity) along with industry

and year fixed effects. The positive and significant coefficient on the proxy for scale

economies supports the NEG model and shows that increasing scale industries are

associated with spatial concentration. However, when examining European economies,

Brulhart (2001) finds no significant relation and Haaland, Kind, and Midelfart-Knarvik (1999)

find a negative impact of scale on concentration.

Ottaviano and Thisse (2004) provide another view of NEG models in that small shocks can

have permanent impacts on the economic landscape. The putty-clay nature of economic

geography seems to capture a fundamental aspect of modern economies and planning

decisions. The researchers write (p. 2603), ―the steady fall in transport costs seems to allow

for a great deal of flexibility on where particular activities can locate, but once spatial

differences develop, locations tend to become quite rigid. ...Nonetheless, we have also

seen that such an extreme agglomeration may give rise to various forms of price differentials

that can trigger a process of redispersion, or that more sophisticated migration behavior may

prevent the emergence of a single core.‖ The net effect is that it can be very difficult to

predict long-run land use patterns even though they are stable during intervening periods.

The industrial revolution brought a major shift toward urbanization and specialization, but

following that revolution there have been long cycles of convergence and divergence. The

dominant expectation in the mid-twentieth century was probably that regional economies

would become more alike, rather than more specialized. With advances in transportation

and communications, and with decreased shares of employment tied to immovable natural

resources, some analysts documented and/or predicted regional convergence (Hoover and

Giarratani 1984, Ch. 11; Garnick and Friedenberg 1982, Barro 1991). Regional industrial

structures generally became more similar from the 1930s until the 1970s (Kim 1995), and

income differences among regions narrowed. Since about 1980, however, there has been

no clear overall trend toward convergence or divergence in incomes or industrial

composition (Bernard 1996; Gutierrez 2001; Bernat 2001; Holmes and Stevens 2004).

Workforce Housing Final Report 37

Business Dynamics

Net growth in local area employment in each industry is the result of much larger gross

flows, with increases in employment due to the creation of new establishments and

expansion of some existing establishments, offset by reductions from contractions or

closures of other existing establishments. Relocations of businesses into or out of the

locality could also contribute.

Recently, with the creation of several new data bases (Parker 2006; Neumark et al. 2006a,

2006b; Acs and Armington 2005) there has been a surge of research regarding

establishment and employment dynamics. Except for work using a private data set

constructed from Dun and Bradstreet records (e.g. Neumark), this research has been

conducted by employees of the Bureau of Labor Statistics or Census Bureau, or by

researchers granted special access to the confidential microdata.

Most of the recent work has not focused on differences among local areas. Research using

the new Census and BLS data has most commonly been at the national level (Butani 2005;

Davis 2006a, 2006b; Figura 2006; Knaup 2005; Spletzer 2004; Clayton 2006; Pinkston

2004; Yashiv 2006). Some studies have documented dynamics for specific areas, but have

not compared areas across the nation (Neumark 2006a, 2006b; Faberman 2001, 2002;

Spletzer 1998). Of those that have compared geographic areas below the national level,

most have used states rather than metropolitan areas (Abowd 2003; Benedetto 2007; Foster

2006; Rigby and Essletzbichler 2000).

During short-term periods, expansions and contractions of existing establishments account

for most employment change, but measured over several years, the creation of new

establishments and closure of existing establishments is the main influence.

Failure rates are high among young establishments. Survival rates rise with age, at least

after the first few years (Caves 1998; Figura 2006; Dunne 1989, 2005; Headd 2001; Nucci

1999). Despite high closure rates among young establishments, high entry rates and high

growth among surviving young firms are typically key factors in local areas with high

employment growth. (Faberman 2006; Neumark 2006a.)

38 Workforce Housing Final Report

Overall closure rates tend to be as high, or higher, in growing areas as in declining areas

(Greene 1982; Faberman 2006). High closure rates of young establishments may obscure

relative closure rates across among metropolitan areas, because areas with healthy

economies and high growth have more failure-prone young firms, possibly offsetting high

survival rates among older area firms. Faberman (2006) not only finds that the presence of

young firms raises the closure rate for an area, but that in the fastest-growing areas the

failure rate among young firms is higher than for young firms in less-dynamic areas where

growth is slower.

Although businesses may move to new quarters within a metropolitan area, relocations

between areas do not appear to be common. Most studies of business dynamics have not

considered relocations, but the studies that have measured moves conclude that inter-area

relocations represent a very small share of changes in employment (Neumark 2006a,

2006b; Greene 1982). Perhaps, for multi-establishment firms, there are some stealth

relocations, where a new establishment is opened in another area and at some later point

the original establishment is closed down. Even if these situations were included, however,

most of the net growth in an area, and most of the transfer of activity among areas, would

appear to be related to entry and exit of new firms or branches.

Many of the studies of business and employment dynamics have been limited to

manufacturing (e.g., Davis, Haltiwanger, and Schuh 1996; Rigby and Essletzbichler 2000;

Dumais, Ellison, and Glaeser 2002; Foster 2006). For example, Dumai, Ellison, and

Glaeser (2002) use data from the Census Bureau‘s Longitudinal Research Database on

U.S. manufacturing industries. The researchers find that plant births reduce agglomeration

while plant closures reinforce agglomeration. Another curious finding is that concentrated

industries are just as mobile as un-concentrated industries. The agglomeration of an

industry does not seem to inhibit movement through the balance of new plants and closing

firms.

The dynamics in non-manufacturing industries may be quite different (Acs and Armington

2005; Anderson and Meyer 1994; Foote 1998). With manufacturing industry jobs

representing only about 11 percent of national employment and an even smaller share in

major metropolitan areas, studies of manufacturing may appear to be of little interest.

Certainly, information based on all types of businesses is preferable, but manufacturing

plants tend to be part of the export base of local economies, even where location quotients

Workforce Housing Final Report 39

are not high. Moreover, manufacturing is generally more labor-intensive and uses more

locally sourced inputs than some other export base industries such as mining or long-haul

transportation. Thus, manufacturing may perhaps deserve special attention.

Business data are generally collected and reported on either an establishment basis or a

firm/enterprise basis. An establishment is a single location where business is conducted or

services provided. A firm or enterprise may consist of a single establishment or of multiple

establishments. Most firms consist of a single establishment. Multi-establishment firms

represent less than 5 percent of employer firms and include only about a quarter of

establishments, but they provide the majority of jobs. (Butani 2005; Acs and Armington

1999, 2005.)

40 Workforce Housing Final Report

Chapter Three: Metropolitan Labor Supply

Adverse impacts of constrained housing supply and high housing costs on employers will

take the form of labor supply problems. Local labor supply problems are not easily

identified, however. In contrast to the mountains of data and extensive analyses with

respect to whether workers can find jobs, there is a dearth of information about employers‘

ability to find workers and their costs for doing so.

Housing-related constraints on labor supply will occur largely through effects on net

migration. This is both because housing supply has greater impact on net migration than on

the labor supply response of the existing population (in the form of changes in the labor

force participation rate or unemployment rate) and because net migration is the key

determinant of overall changes in labor supply for metropolitan areas, at least in the long

run.

Blanchard and Katz (1992) examined the effects of employment shocks and reported that

―by five to seven years, the employment response consists entirely of the migration of

workers.‖ That may be an extreme view, but Bartik (1993) reviews a number of other

studies and finds that most estimates of the migrants‘ share of new jobs range from 60

percent to 90 percent in the long run. Reduced unemployment rates and higher labor force

participation rates in areas with job growth are partly attributable to in-migrants with higher

labor force participation rates and employment rates than incumbents.

An alternative source of labor, rather than migration or increased employment among local

residents, could be commuting from other metropolitan areas or from nonmetropolitan areas.

In the compact geography of European countries, a trade-off between migration and

commuting has been given considerable attention (e.g., Ommeren 1999; Cameron 1998,

2006; Elliason 2003; Gordon 1998; Jackman 1992).

In the U.S., there has been recent growth in ―extreme‖ commuting, defined by the Census

Bureau as traveling 90 minutes or more getting to work (Census Bureau Press Release

March 30, 2005; Pisarski 2006; Paumgarten 2007). Data from the 2005 American

Community Survey show that among people employed in Metropolitan Statistical Areas

(MSAs), 91 percent also lived in those MSAs, with most of the remainder coming from

adjacent MSAs, often within a broader ―Combined Statistical Area.‖

Workforce Housing Final Report 41

Although MSAs are defined as counties or groups of counties based on commuting patterns,

indicating that they represent integrated labor and housing markets, congestion and

mismatched jobs and housing may adversely affect the supply of labor even from within the

area. That topic is explored in a separate section.

Identifying Labor Supply Problems

The symptoms of labor supply problems encountered by employers may include higher

wages, higher non-wage compensation, unfilled jobs, and higher costs for recruitment.

Productivity may suffer, but the negative effects on productivity may be obscured if labor-

intensive lower-skilled tasks are shifted to other locations or if employers invest in more

equipment to substitute for labor. In that case labor productivity may appear to be better

than average, even though total factor productivity is substandard.

Of the various indications of adverse local labor supply conditions, the most visible may be

wage rates that are higher than elsewhere. Wages for seemingly-comparable workers and

jobs are generally higher in places with high house prices. The higher wages have been

interpreted, in some analyses, as reflections of higher productivity generated by

agglomeration externalities, or as due to unmeasured but valuable characteristics of

workers, especially highly educated white-collar workers (Glaeser-Mare 2001, Berry-Glaeser

2005, Peri 2002). Such interpretations imply that the higher wages do not represent higher

marginal unit labor costs, and therefore the absence of adverse impacts on employers.

Information about non-wage compensation is generally available as well. The principal

components of non-wage compensation are health insurance and retirement benefits.

Some benefit costs are related to wage levels, or to local costs. In general, however, they

may be less influenced than wage rates by constraints on housing and labor supply.

To the extent that employer-assisted housing benefits are an element in non-wage

compensation, the cost of those benefits could represent a significant impact of housing

supply conditions on employers. Housing-related benefits are rare, however, except

perhaps as an element in relocation packages. The incidence of such benefits is unknown,

but their infrequency is implied by the fact that they do not even garner a footnote among

42 Workforce Housing Final Report

the benefits enumerated in the BEA National Income and Product Accounts or the detailed

catalog of benefits from the BLS National Compensation Survey.

Labor supply constraints could be manifest in unfilled jobs. There is less information about

unfilled jobs, however, than about employers‘ costs for wages and benefits. There is also

much less information about unfilled jobs than about unemployed workers.

The availability of information about job vacancies, as well as about hiring and separations,

has been enhanced by the creation of the BLS Job Openings and Labor Turnover Survey

(JOLTS). Information is collected on a monthly basis from about 16,000 establishments

regarding total employment, job vacancies at the end of the month, employees added to

payrolls during the month, and the number who quit or are terminated. So far, JOLTS has

only been presented at the national level and analyzed mainly in terms of changes over

time. Overall totals (without industry detail) are also reported for four broad regions.

Since JOLTS information is not reported below the broad regional level, it has been

impossible to use for analyzing labor supply and job vacancies in metropolitan areas. The

principal excuse given for this lack of geographic detail is that the sample is too small.

Although the sample includes only about 0.2 percent of eligible establishments, the sample

establishments employ over 12 million workers—nearly 10 percent of the relevant

employees. The Current Population Survey, used (along with some other information) to

produce annual benchmarks for local area labor force and unemployment estimates,

collects information from a sample of about 60,000 households, which include fewer than

100,000 workers.

It may be possible for BLS staff or for researchers granted special access to use

unpublished JOLTS microdata for analysis. Several interesting studies that go beyond the

information in the published data have been conducted on that basis (e.g., Davis 2007).

Even if local vacancy data were available, it may be difficult to identify labor shortages or

other problematic labor supply conditions.

In analyzing job vacancies, there may be benefits to applying some of the analysis that has

been conducted regarding housing vacancies. It is likely, for example, that equilibrium job

vacancy rates for an area or occupation would be lower where the average job tenure is

longer.

Workforce Housing Final Report 43

Although the primary research direction for assessing and documenting the effect of housing

supply on labor supply should involve looking at relationships between wages or vacancies

and housing supply, there is also a need from both a housing market and labor market

standpoint to better understand the symptoms of problematic labor supply from the

employers‘ standpoint. One possibility would be to develop comprehensive survey data to

measure subjective labor supply concerns and relate those measures to the data on wages,

vacancies, turnover, etc.

The number of vacant positions will generally be higher when the unemployment rate is

lower. The relationship between vacancies and unemployment is represented by the

Beveridge Curve (Blanchard and Diamond 1989; Bleakley and Fuhrer 1997; Valletta 2005,

2006; Abraham 1987).

In the absence of vacancy data prior to the availability of the JOLTS, a number of analysts

constructed estimates of the Beveridge Curve using the Conference Board Help-Wanted

Advertising Index and other fragmentary data. The results indicated that the curve shifted

outward from the 1960s until the mid-1980s, with more unemployment relative to the level of

vacancies, and vice-versa. The curve then apparently shifted inward, suggesting more

efficient matching of workers with jobs.

Explanations for the outward shift of the Beveridge Curve in the 1960s and 1970s and the

subsequent inward shift beginning in the mid-1980s include entry into the labor force of

young and inexperienced workers. Bleakley and Fuhrer (1997) cite adjustments following

absorption of the baby boom and of increased female labor force participation as plausible

causes for increased matching efficiency and decreased level of churning.

In addition, divergence in regional economic conditions, followed by more uniform

employment growth and unemployment rates among localities, have been identified as

major factors (especially by Abraham 1987 and Valletta 2005). It is obviously harder to fill

vacant jobs with unemployed workers if they are in different places. Housing market

constraints to migration can be expected to impede the process of matching workers to jobs.

The relationship between job vacancies and unemployment represented by the Beveridge

Curve is related to the concept of a natural unemployment rate and a non-accelerating

44 Workforce Housing Final Report

inflation rate of unemployment (NAIRU). If unemployed workers cannot be efficiently

matched to jobs, market friction and anti-inflation monetary policy will mean higher overall

unemployment rates (Brauer 2007, Katz 1999).

Housing and Net Migration

Although areas with favorable climate and other amenities will attract migrants from less-

blessed locations, those attractions are relatively static ―fixed effects‖ and the driving force

for changes in migration of the working-age population will generally be wages and

employment opportunities, offset by living costs, of which the cost of housing is the most

important.

Despite the theoretical and empirical significance of housing supply/cost in explaining

migration, it is not uncommon for models of migration to ignore housing cost or to assert that

housing costs are captured by crude measures of overall costs of living, such as those

constructed from regional versions of the consumer price index. In other cases, housing

costs were tested and found to be insignificant. Migration studies have often been based on

states or regions, and housing supply effects may be weaker at that level of geography.

According to Cameron et al. (2005), however, ―leaving out housing market effects typically

results in misspecified models in which labour market effects are estimated as weak or even

perverse in direction.‖

One of the problems with estimating the effect of housing cost on migration is that housing

costs reflect the quality of life in an area (Ezzet-Lofstrom 2004; Clark 2004; Fu 2005b;

Roback 1882, 1988), as does migration. In a cross-sectional analysis, this could mean that

people appear to prefer to migrate to high-cost areas. There are two main approaches to

dealing with this. One is to measure the effects of changes in house prices (and/or wages)

assuming that quality of life doesn‘t change much over the short term. The other is to

include variables that may be measures of quality of life, such as climate, crime rates, or air

quality.

Some migration models are built on the idea that, given differences in area amenities, there

will be adjustments in local area wages, house prices, etc., to produce equilibrium where

migration is minimized. This approach is more common in Europe, where mobility rates

average roughly half of those in the U.S. and adjustments in the housing supply occur more

Workforce Housing Final Report 45

slowly. Such European studies typically include envious references to U.S. conditions that

are seen as more responsive to regional economic imbalances and better able to eliminate

regional differences in unemployment rates and to allocate resources productively (e.g.,

Vermeulen 2005; Cannari 2000; OECD 2005). The ideas of equilibrium versus

disequilibrium models for migration are described in Goetz (1999).

In considering the effects of housing supply on net migration, there is some question as to

the appropriate measure of housing cost or availability. Rents may perhaps be most

appropriate, since movers (both local and inter-metropolitan) tend to be younger than non-

movers and (even after adjusting for age) are less likely to be homeowners. 5 Migration

models using rents or implicit rents include Berger (1992).

One complication in the use of rents is that the average rent paid by current residents may

not accurately reflect the rents available to in-migrants. That distinction exists in private

rental markets, but is most significant where rental housing is subsidized or publicly owned.

Most recent migration studies incorporating housing cost have used measures of the cost of

owner-occupied housing, rather than rents. In some cases (e.g., Cannari 2000) house

prices were used because of perceived deficiencies in data for rents. The price of owner-

occupied housing has been portrayed by price levels (with or without quality adjustments),

prices relative to incomes, and/or user cost (either explicitly or implicitly through inclusion of

appreciation rates). Gabriel et al. (1992) found house prices more statistically significant

than user cost, interpreting this as due to cash flow or liquidity constraints. Murphy (2006)

found both price levels relative to income and appreciation rates important to regional UK

migration, but found that in Greater London, the absolute supply (relative to population) was

more important than cost.

The perceived user cost facing migrants or other home buyers largely consists of

expectations of house price changes. User cost variables in models of migration typically

assume that expected appreciation is equal to actual appreciation in the preceding one to

three years. The potential in-migrants from areas with stagnant prices may, however, have

5 According to the March 2005 Current Population Survey, among persons aged 5 and over who

did not move between 2000 and 2005, 85.7 percent lived in owner-occupied homes, compared to

52.5 of all movers and 51.8 percent of domestic interstate movers.

46 Workforce Housing Final Report

formed different expectations of house price changes, for both origin and destination areas,

than existing residents in possible destinations. Murphy (2006) include expected house

price change for each UK region based on a ―semi-rational‖ equation incorporating

incomes, stock market prices, and interest rates, as well as recent changes in the region, in

contiguous regions, and in Greater London.

U.S. Migration Studies

Berger and Blomquist (1992) studied county-level migration during 1975 to 1980. They

calculated separate estimates of the probability that an individual would leave a county and

the probability that movers would choose specific destination counties. Housing costs were

represented by quality-adjusted rents and imputed rents. The models also included

demographic characteristics of the individual, county quality of life measures, and wages. In

the estimate of the probability of moving away from the initial county of residence, the rent

measure had the right sign but was not statistically significant. In the choice of destination

county, however, rents were highly significant. This is consistent with other studies

indicating that housing costs affect gross in-migration more than gross outmigration.

Gabriel, Shack-Marquez, and Wascher (1992, 1993) used migration data for nine divisions

(New England, South Atlantic, Pacific, etc.) between 1980 and 1981 and between 1986 and

1987. Since there are widely divergent housing supply conditions within such broad areas,

it is difficult to assess the impact of local housing supply constraints, but they still found that

house prices had a significant impact on migration between those regions, with migration

from low-cost regions to high-cost regions, in particular, constrained. They tried data for the

level of prices in the origin and destination regions, as well as a measure of user cost

incorporating price changes in the preceding 3 years. The level of prices in the destination

region showed the strongest and most consistent effects.

Frey and Liaw (2005) analyzed gross interstate migration from 1995 to 2000. They

represent housing cost as the average of median value in 1990 and 2000. A slope dummy

distinguishes housing value for individuals with and without college educations. The results

show a significant impact on the probability of leaving a state, and an even greater influence

on the likelihood of moving to a state—but only for those without a college education.

Workforce Housing Final Report 47

Foreign Migration Studies

Bover, Muellbauer, and Murphy (1989) describe how housing price differentials between the

high-cost South East and the rest of the U.K. impede migration (p. 130):

―The ‗mobility trap‘ caused by an upswing in aggregate housing demand operates as

follows. As the relative appreciation of their house prices gathers pace, households in

the South East (UK) initially become more reluctant to move to other areas. They fear

that they would miss out on further relative appreciation and that they may not be able to

bridge the house price gap should they subsequently wish to return. Thus, relatively few

housing slots are freed for potential migrants to the South East, and this causes further

relative appreciation. As it continues, households outside the South East become

increasingly unable to bridge the gap between whatever equity stake they may already

have in housing and the price of a house in the South East.‖

As the housing price/earnings differential nears a peak, out-migration from the high-cost

area picks up, housing supply is catching up, speculative investment is increasing, and in-

migration is stalling from credit constraints. At this point, the market is vulnerable to an

adverse housing demand shock. If such a shock occurs, as it did in 1973-75, the premium

for South East housing could fall as speculative expectations reverse. At the peak and post-

peak points of the housing price cycle, firms have a particularly difficult time keeping or

hiring workers. House prices are either too high or projected to soften, so workers are either

unable or reluctant to buy into such a market. And many existing workers want to sell out of

the market before it starts to fall. The historical record shows that in 1973, there was a 25-

year peak net outflow from the South East of 69,000 with subsequent large net out-

migrations in 1974 and 1975. Bover, Muellbauer, and Murphy suggest that firms also

participated in the exodus, frustrated by high wages, labor shortages, and high land costs.

There was concern in the UK about the ability to respond to regional labor requirements

because of immobility of social/council (public) housing residents, as well as the less-

extreme immobility of homeowners, particularly following the conjecture by Oswald

(1996,1997) that ownership increased unemployment. House price differentials were seen

as exacerbating transactions costs for owners. Cameron and Muellbauer (1998) claimed

there is a solid body of research showing that high relative earnings and employment

opportunities encourage in-migration while high relative house prices discourage in-

48 Workforce Housing Final Report

migration, writing ―As owner-occupation has risen, the evidence is that the influence of

relative house prices on net migration rates has risen also.‖ Expensive housing could block

in-migration, leaving a long commute or unemployment as the alternatives. More recently,

Murphy, Muellbauer, and Cameron (2006) claim that ―high house prices choke off migration

caused by strong labour market conditions.‖

Hamalainen and Bockerman (2004) discovered that an increase in the internal turnover of

jobs in regions of Finland was associated with an increase in net-migration. Because of

housing market constraints to in-migrants, the response to strong labor markets was

primarily through less out-migration.

Cannari (2000) considered the effects of housing costs on migration from the south of Italy

to the north. Despite continued superior incomes and employment opportunities in the

north, migration slowed sharply after the 1960s. Growing differences in quality-adjusted

house prices were found to be a significant explanation.

Vermeulen (2004, 2005) looked at several European countries and concluded that housing

cost differences offset the incentives to migrate to areas with lower unemployment, allowing

differences in unemployment rates to persist for extended periods. They argue that lower

house prices provide such significant compensation for higher unemployment that E.U.

subsidies to economically depressed areas are not justified.

Workforce Housing Final Report 49

Evident Impacts of Housing

In most U.S. markets, housing supply conditions do not have major, obvious impacts on

labor supply and the ability of local businesses to compete with firms located elsewhere in

the U.S. Under more extreme conditions, the role of housing as critical infrastructure

becomes manifest.

Although surveys have found that housing is a factor in business location decisions, it has

rarely been identified as one of the most important factors (Gottlieb 1994; Salvesen 2003;

Gambale/Area Development 2006a, 2006b; Czohara 2004) and is often not even mentioned

(Expansion Management 2007). The only surveys where housing has gotten major

attention have been in the most expensive areas, such as Silicon Valley or Boston

(Gerston/Silicon Valley Leadership Group 2006; Boston Globe survey cited by Bluestone

2006). Although housing in much of California is only marginally more affordable and

plentiful than in the Bay Area, statewide surveys have not placed housing among the

biggest concerns of business (Baldassare 2007; Bain/California Business Roundtable

2004). In fact as part of this research we conducted interviews to systematically assess the

prevalence of employer problems suggested to be caused by the high cost and limited

availability of housing. None of the businesses interviewed identified that the cost of

housing was a primary consideration in location selection, or of primary concern.

The first real involvement of the U.S. Federal government in housing supply came during the

mobilization for World War I. The president of Newport News Shipbuilding told Congress

―You cannot get ships unless houses are provided for workmen.‖ (NY Times 1/10/18.)

Within a few weeks Congress appropriated funds for the acquisition of housing and loans to

developers around shipyards, and President Wilson was authorized to commandeer lumber

and move tenants who were not working in shipyards out of nearby hotels and boarding

houses. The Ordnance Department separately diverted funds out of their production budget

to ―industrial housing‖ located ―near isolated explosive and bag-loading plants.‖ (Colean

1940.)

In other countries where housing markets were recently less fluid than in the U.S., the

impacts of housing supply on labor supply and economic efficiency have been clearly

demonstrated. Several studies analyzed conditions in Poland in the 1970s and 1980s, and

found that housing shortages and rent controls (influenced by Marxist theory considering

50 Workforce Housing Final Report

housing to be unproductive) impeded labor mobility and productivity (Mayo and Stein 1988;

Pogodzinsky 1995; Hacker 1999).

While not as severe as in Poland, sclerotic housing markets in the U.K. have been identified

as a constraint to matching labor and jobs, with a lack of housing in the booming South East

affecting labor supply and business locations, and contributing to national unemployment

(TZ Consulting 2006; Murphy, Muellbauer, and Cameron 2006; Barker 2004; Hughes and

McCormick 1987). Overall, this topic has gotten more attention in the U.K. than in the U.S.,

although some of the recent attention has been related to the contention by Oswald

(1996,1997) that homeownership adds to unemployment.

Other studies from Europe have indicated that high housing prices in areas of labor

shortage, and low prices in areas of labor surplus, have impeded labor market adjustments

through migration. These include Vermeulen (2005), Hamalainen (2004), Cannari (2000),

and Duffy (2005), with regard to the Netherlands, Finland, Italy, and Ireland, respectively.

Hurricane Katrina in 2005 destroyed more housing than any previous natural disaster.

Estimates of the number of units destroyed are in the range of 200,000, compared to less

than 30,000 from Hurricane Andrew in 1992 or the San Francisco earthquake and fire in

1906. This tragedy provides a basis for more fully exploring the link between housing

supply and economic potential and efficiency. So far, most of the evidence has been

anecdotal. (See stories regarding Oreck in NY Times 1/15/07; PBS 2/7/07; ―Ship Shape‖

N.O. Times-Picayune 12/3/06.)

Needed Research

Direct measure of export base/non-local customers

Labor shortage measures – metro area vacancies, recruiting costs

Compilation of an additional survey data determining whether employers in

different areas and industries perceive labor supply problems, and relate those

results to statistics on wages and vacancies.

New research on the effects of housing on workforce migration, concentrating on

metropolitan areas and types of workers

Documentation of the impacts of extreme housing supply conditions (e.g., New

Orleans) on businesses

Workforce Housing Final Report 51

Chapter Four: Labor Productivity

Business executives are concerned about the effects of housing labor productivity,

recruitment, and retention. When local housing is expensive, firms must pay workers

nominal wages high enough so that the real wages (net of local living costs and amenities)

are comparable to other employment centers. Workers who can find higher wages or lower

housing costs are likely to leave their current job for the higher real wages. Therefore, the

literature review on labor productivity begins with wage models and measures of turnover.

Labor is differentiated by education and training, and the skilled workers (college-educated)

are often considered the driving force in business growth. To attract skilled workers,

businesses seek locations where skilled workers already live and work. A robust literature

has developed around spillover effects, agglomeration effects, and the density of cities. All

of these factors affect labor productivity and the co-location choices of firms and workers.

Another set of literature focuses on the spatial mismatch and job search. Rising house

prices may force some workers to seek affordable housing far from their employment site.

The spatial mismatch literature grew out of concern that minority workers lived far from job

centers, which would reduce their opportunities to find employment. The suburbanization of

jobs has reduced the spatial mismatch for some non-minority workers, but increased the

spatial mismatch for many minority workers. Spatial mismatch research goes beyond

racial/ethnic differences to include income class differences, which generally align with

education and wages. If employment cannot be found within commuting distance, workers

can migrate to other metropolitan areas seeking workers. The review of the migration

literature considers the differences in search patterns between skilled and unskilled workers

as well as the labor market dynamics as workers respond to market disequilibrium.

An important factor in worker productivity and job match is the commuting time and distance.

Long commutes are more liable to delays or breakdowns that can lead to worker absences

and lost productivity. Traffic congestion has led some firms to relocate away from downtown

properties, which has created a challenge for public transit systems and workers without

cars. Although there is a vast literature on transportation, this review focuses on how

accessibility has affected the size of the ―labor-shed‖ and the jobs-housing balance.

The size and shape of a metropolitan area depends on the growth controls and land use

regulations. Tight land use regulations that reduce the supply elasticity of housing channel

52 Workforce Housing Final Report

the demand into higher house prices rather than more houses and workers. Short-run

adjustments are made through house prices and rents. In the medium term, businesses

must respond with higher wages to retain their workers. In the long term, growth occurs in

less restricted areas that can accommodate more houses and more jobs. A related

literature, nexus studies, has developed from California law, which attempts to measure how

a new development will affect the availability and affordability of housing. Nexus studies

often are used to justify linkage fees charged to developers by local governments. The fees

can channel development and may offset some of the externalities associated with new

development.

The logical extension of studying land use patterns is the debate over sprawl and smart

growth. Although sometimes portrayed as a conflict between developers and planners, the

debate attempts to weigh the costs and benefits of density on commuting, pollution, energy

use, agricultural land consumption, income segregation, infrastructure, fiscal expenditure,

and personal health. The goal in this literature review is not to settle the debate, but rather

point out how it touches on the accessibility, availability, and affordability of workforce

housing.

A number of papers take a broad view of economic development at the regional level,

considering employment, housing, transportation, migration, and fiscal impacts. This

research provides case studies of specific regions recognizing the historical patterns of

development and projecting how the region could be improved in the future. Although it is

more difficult to translate the findings of these regional case studies to other places, the

richness of the detail highlights the forces at work.

The Labor Productivity section is divided into four sections: wage models, turnover,

education and training, and returns to skill. Productivity has traditionally been measured in

the context of a wage demand equation derived from the production function. Thus, wage

models are a logical place to start. Efficiency wage models are based on the idea that

employers can save the cost of hiring and training new workers by paying their existing

workers slightly above the wage at which those workers could easily find another job. One

approach to examining the effects of land use regulation is to see if those regulations are

related to either the wages paid or the rate of labor turnover. Much of the recent labor

literature has focused on the increased differentiation between skilled and unskilled labor

often distinguished by the degree of education. Cities with a high share of skilled workers

Workforce Housing Final Report 53

seem to be places where new industries grow and all workers benefit through higher wages.

One explanation for the higher wages is the knowledge spillovers among workers and firms.

Each section concludes with some ideas about further research that would highlight the role

of land use regulation on worker productivity.

Wage Models

In neoclassical economic models, labor demand is derived from the production function in

which profits are maximized by paying labor its marginal product. Labor productivity can be

increased by increasing the amount of capital per worker or by improving how the workers

are combined with capital and technology or total factor productivity. Shapiro and Stiglitz

(1984) suggest another way, i.e., paying the workers a higher wage or efficiency wage.

Workers being paid no more than what they would receive at their next best job are not too

concerned about losing their job. Assuming that the worker has some knowledge that has

been customized for his or her current employer, that knowledge is lost when the worker

leaves the firm. Another worker must be trained as a replacement and the overall

productivity of the firm suffers during the retraining period. By paying the worker a slightly

higher wage, the firm maintains higher productivity to justify the higher wage, and the worker

does not shirk his responsibilities in order to keep his job.

Akerlof and Yellen (1990) developed the fair wage-effort hypothesis, which is a variant of the

efficiency wage model. Workers form a notion of the fair wage based on the pay and effort

of their co-workers and neighbors. If the actual wage paid is lower than what the worker

perceives is a fair wage, the workers withdraw effort in proportion. Depending on the wage-

effort elasticity and the costs to the firm of low effort, the fair wage becomes a factor in the

wage bargaining. Krueger and Summers (1988) explain the variation in wage scales across

industries in the context of fair wages. If firms must pay high wages to certain workers,

either because they are in short supply or the sensitivity of total output to their efforts, then

demands for fairness will lead to a compression of the pay scale. Wages of other workers in

the firm will be paid more than in other firms, and similar patterns develop throughout an

industry.

Another explanation for wage differentiation is provided by Diamond and Simon (1990) who

claim that workers demand higher wages in more specialized cities. Worker productivity is

increased by the greater specialization, but that specialization also increases risk from a

54 Workforce Housing Final Report

drop in demand for the output of that industry. To compensate for that risk, firms in the

specialized city must pay the workers a premium. Workers in a large city with diversified

industries face a lower risk of job loss and a better chance of finding alternative employment

during a slump.

To test these theories about wages and labor productivity, Hellerstein, Neumark and Troske

(1999) combine data on individual workers and their employers to estimates of marginal

productivity differentials for different types of workers. Generally productivity is inferred from

a wage equation based on the assumption that workers are paid their marginal product.

Hellerstein et al. have independent measures of productivity and wages to test whether

higher wages correspond to higher productivity. The data come from the Worker

Establishment Characteristics Database (WECD), which matches the long-form responses

of the 1990 Decennial Census to the data on their employers in the Longitudinal Research

Database (LRD). In addition, the researchers have demographic information on workers in

a sample of plants with information on plant-level inputs and outputs. The plant-level data

provide the measures of productivity by demographic group, which can be compared to

wages paid to those demographic groups according to the Census. They found that the

higher wages paid to prime-aged workers (aged 35 to 54) and older workers (aged 55+) are

justified by their higher relative marginal product. However, the lower relative earnings of

women are not reflected in lower relative marginal products.

Leonard (1987) provides another test of the shirking vs. turnover versions of efficiency

wage, and finds little empirical support for either. The shirking version emphasizes a trade-

off between self-supervision and external supervision, whereas the turnover version

assumes turnover is costly to the firm. Comparing across firms, variation in wages paid to a

selected set of homogeneous workers should be explained by variations in the cost of

monitoring/shirking or turnover. Even for narrowly defined occupations within one industrial

sector of one state, Leonard found widely dispersed wages. This finding indicates there are

more factors or a more complex process in wage determination than the relatively simple

efficiency wage models.

One of the complicating factors may be the business cycle, which affects industries

differently as they respond with adjustments in output, investment, and labor demand.

Glosser and Golden (2005) looked at fluctuations in labor demand across U.S.

manufacturing firms. They found that after 1979 hours have become more flexible and

Workforce Housing Final Report 55

employment considerably less flexible, especially during expansion phases. To reduce

turnover costs, firms are keeping a stable workforce and adjusting output by adjusting the

hours of the workers. On the one hand, more stable employment may help workers, but the

demand risk has been transferred to workers in the form of fluctuating incomes. One

explanation for the higher wages of risky industries is the increased uncertainty about the

number of hours paid. Multiple workers per household and even multiple jobs per worker

can offset the risk from uncertain hours. Cities facilitate these multiple employment

arrangements. The urban population is more productive, and the households receive a

higher income.

An alternative view is that labor market pooling improves the job-worker match. Costa and

Kahn (2001) write about ―Power Couples‖ in which both people have college degrees or

advanced degrees and want to have fulfilling careers. The competing demands for

specialized employment, lengths and costs of commutes, and quality of housing can only be

satisfied in a large city. Costa and Kahn estimate that 36 percent of the increase in

concentration of power couples in large cities is due to the dual career hypothesis. If the

highly educated are critical to productivity gains, large cities will benefit from attracting these

power couples. Pingle (2006) adds that couples‘ migration propensity is substantially lower

when the income of the couples is more nearly equal. Once each worker in the couple has

a good job, the couple is less likely to move because it is difficult to find two good job

matches at the same time and place.

Smoluk and Andrews (2005) follow the tradition of estimating a wage equation loosely

derived from a constant elasticity of substitution (CES) production function. The novelty is to

estimate the equations at the state aggregate level and then make comparisons across the

lower 48 states to determine which factors are most important to long-term economic

prosperity. They found that education and density are positively related to productivity, and

taxes are negatively related to productivity.

Research Ideas and Extensions

Smoluk and Andrews (2005) have made a first step in adding the spatial dimension to

productivity by looking at the state level. A logical extension would be to look at the MSA or

county level where most land use decisions are made. Then measures of land use

regulations could be incorporated into the wage equations. A major obstacle is how to deal

56 Workforce Housing Final Report

with the endogeneity that higher wages may cause as well as be caused by restrictive land

use regulations that make housing relatively scarce. Panel data may help by using

predetermined or lagged measures of land use. Following Krueger and Summers (1988),

variation by industry may also be helpful in that industries with increasing productivity are

better able to compete for labor with higher wages. Cities with elastic housing supply may

be able to draw more workers and grow faster with a smaller increase in wages than highly

regulated cities with inelastic housing supply. A test of this hypothesis is to look across

cities at the highly productive industries and compare the wage and employment responses.

The opportunities for expansion may depend on the relative productivity of the other

industries in the city (draw workers away from other firms) and the restrictiveness on

housing (draw workers in from other cities). Industries with close forward and backward

trade linkages within the city, may benefit from generally higher productivity. Whereas,

industries not tightly integrated with local suppliers may benefit from productivity gains that

make it easy for the productive industry to draw workers away from other firms.

An extension of the Glosser and Golden (2005) analysis would be to investigate whether

reduced employment flexibility is related to housing supply inflexibility and/or the ownership

trap. Workers may not be willing to risk losing their house because purchasing another one

is so much more expensive or house values have fallen and the owners are reluctant to sell

at a loss. This reluctance to move by the employee may give the employer more leverage to

cut back hours without increasing turnover. Easier credit terms make it possible for the

worker to smooth consumption despite fluctuations in income. According to this hypothesis,

inflexible housing supply facilitates fluctuating labor demand and maybe encourages

multiple worker households to diversify their income sources.

Turnover

After an opening anecdote about labor turnover at Sam‘s Club vs. Costco, this section

presents the factors related to labor turnover. The section also reviews some of the papers

on job creation, job destruction, and job flows, ending with some suggestions on follow-up

research.

Wal-Mart‘s Sam‘s Club and Costco target a similar consumer market, but have a very

different approach when it comes to employee wages and benefits. According to Cascio

(2006), Costco has 338 stores and 67,600 full-time employees, while Sam‘s Club has 551

Workforce Housing Final Report 57

stores and 110,200 employees. Wages and benefits are much higher at Costco than Sam‘s

Club, but Costco offsets those higher labor costs with very low turnover. The average

annual cost of replacing the employees at Sam‘s Club is nearly three times greater than for

Costco. Here is one example where higher wages buy employee loyalty and lower turnover

costs.

One of the stylized facts established by Farber (1999) is that low-skilled workers experience

more job turnover than others, though that rate of turnover declines with age and length of

tenure at a job. Low-skilled workers are also more likely to experience layoffs or involuntary

discharges rather than quits. And those terminations are more likely to result in a gap in

employment rather than a direct movement into another job, as is more common among

skilled workers.

More recently Farber (2005) analyzed the job loss experience using the Displaced Worker

Survey, covering the period 1981-2003. The Displaced Worker Survey is the most

comprehensive source on the incidence and costs of job loss. Since 1984, the survey has

been administered every two years as a supplement to the Current Population Survey.

Several issues are worth noting relative to the Displaced Worker Survey. The survey does

not capture multiple job losses or termination for cause. The recall period is based on the

last three years. Farber reports that 35 percent of the job losers are not employed as of the

subsequent survey date, and about 13 percent are re-employed at part-time jobs. Of those

job losers re-employed at full-time jobs, their income averages 13 percent less on the new

job, and counting foregone earnings increases enjoyed by the job non-losers, the average

loss is 17 percent. Even though the economy may benefit from these adjustments, job

losers suffer significantly in terms of income, health benefits, pension benefits, and self

esteem. There are also negative ramifications for the community with substantial numbers

of long-term unemployed in terms of reduced consumer spending, loss of house value,

foreclosures, and out-migration.

Although more common among unskilled workers generally, certain skilled workers

experience rapid turnover, such as programmers and software engineers. Fallick,

Fleischman, and Rebitzer (2005) write about ―Job Hopping in the Silicon Valley,‖ and report

that computer industry workers have higher mobility rates in Silicon Valley than outside of

California. If thick labor markets in cities allow for better matches, then we might expect

lower turnover in industrial clusters or large cities. Apparently the opportunities for better

58 Workforce Housing Final Report

matches may offset the quality of the original match such that labor turnover is not

significantly different in large cities. Also, industry-specific shocks and rapid technological

development may lead to higher turnover despite good firm-specific matches.

Specialization, which enhanced the productivity initially, may be displaced by new

technology and exacerbate unemployment.

Metropolitan areas are sometimes segregated by race, which leads employers to be

concerned about diversity and isolation when mixing workers across race, gender, and age.

Leonard and Levine (2006) report on a large case study of 70,000 workers conducted in

1996-98 from over 800 similar workplaces owned by a single corporation. The focus is on

the turnover of new hires, who are mostly young and 50 percent quit within four months.

Lateness or absence is the most common cause for worker dismissal. Although women

disliked gender diversity, the authors find no consistent evidence that diversity increased

turnover. Also (p. 566), although whites are almost twice as likely as blacks (14 vs. 7.6

percent) to live in the same Zip code as their workplace, proximity to workplace had no

effect on turnover rates. Moreover, retention rates were higher for minorities when they

worked in neighborhoods with many residents of their race.

More broadly, Holzer (1996) reports that average annual turnover rates across

manufacturing, retail, and service sectors are approximately 21 percent or about 3 percent

per month. Here turnover counts quits and discharges, but not layoffs. Stewart (2002) uses

the sequence of March CPS data from 1975-2001 to investigate trends in job stability and

job security. He reports that employment-to-unemployment transitions declined dramatically

representing an increase in job security. For men the improvement occurred largely in the

1990s, whereas for women the improvement spread throughout the entire timeframe. There

was an equally dramatic increase in the employment-to-employment transitions (defined as

job changes with two weeks or less of unemployment between jobs). For individual workers,

the reduction in employment losses was paired with a reduction in stability – more job

switching. The same pattern holds for married couples. The loss in stability would have

been smaller but for the increase in proportion of couples for which the wife contributed a

significant share toward the total family earnings. Two-earner families can handle job

changing and are less susceptible to job loss. The increase in female labor force

participation has a powerful impact on productivity, but complicates the household decision

on housing location.

Workforce Housing Final Report 59

Shen (2001) focuses on Boston in 1990 and reports that about 95 percent of all job

openings are related to turnover. The ramification of this fact is that most job openings

come from existing jobs and relatively little from job growth. The spatial distribution of job

openings follows the location of existing jobs. Thus, even though the new jobs may be

occurring in suburban employment centers, which may be higher paying jobs, the

geographic shift away from the central business district takes a long time.

There is a related literature to turnover looking at job creation, job destruction, and job flows.

Acs, Armington, and Robb (1999) use the Longitudinal Establishment and Enterprise

Microdata (LEEM) to measure the annual job flows for both manufacturing and non-

manufacturing. This research, under contract for the Small Business Administration, finds

that gross job flow rates decrease with firm age and establishment size (controlling for age)

whether the size is measured by the initial size or the average size over a period. However,

the relationship of net job growth to business size is very sensitive to the choice of average

size vs. initial size of the business.

Quintin and Stevens (2005) analyze French turnover data. They find that among the

surviving establishments, worker turnover is positively correlated to industry-level exit rates

even after controlling for employee tenure and establishment size. This empirical finding is

consistent with turnover models featuring firm-specific human capital because workers are

more likely to develop firm-specific human capital if the firm survival is above average

relative to the industry.

Davis, Haltiwanger, and Schuh (1996) wrote the book on Job Creation and Destruction,

which has been updated in Davis, Faberman, and Haltiwanger (2006). For any single

business, the net employment change (hires less separations) equals the job flows (creation

less destruction). When aggregated over employers for a region or industry, there are

typically large values for both job creation and job destruction. The new data source is the

Job Openings and Labor Turnover Survey (JOLTS), which is a monthly survey of 16,000

establishments beginning in December 2000. However, JOLTS does not capture

establishment entry and exit. An alternative is the Business Employment Dynamics (BED)

data, which cover quarterly job flows for virtually all businesses.

Another new combination of data is the Longitudinal Employer-Household Dynamics

(LEHD), which is an innovative program within the U.S. Census Bureau. State

administrative (primarily unemployment insurance) data on employers and employees are

60 Workforce Housing Final Report

combined with federal economic and demographic data while protecting the confidentiality of

firms and workers. The Bureau of Labor Statistics coordinates with the states to develop

consistent coding standards covering 98 percent of all private, non-agricultural employment

with quarterly earnings reports. It is possible to reconstruct complete employee rosters at a

point in time and then following the changes for the firm over time or following the worker

between employers over time. Demographic data on the workers are linked by the Census

Bureau according to date of birth, place of birth, and sex of worker. Information on the firm

from the Economic Censuses (taken every 5 years) and the Annual Survey of Manufactures

or Business Expenditures Survey (for non-manufacturing) is linked to the unemployment

insurance records. Wage equations can also be estimated that exploit the longitudinal

nature of the data by controlling for worker and firm fixed effects.

Based on the LEHD data for the private sector, average job creation and destruction

approaches 8 percent of employment per quarter. The worker flows in the form of hires and

separations are more than twice as large. Most (two-thirds) of the job destruction occurs at

establishments that shrink by more than 10 percent in the quarter, and more than 20 percent

occurs at establishments that close. The research describes cyclical movements in the

layoffs-separation ratio. The contribution of job-loss and job-finding rates to the

unemployment rate depends on whether the employment downturn is shallow or deep.

Over the long run, the magnitude of job flows and private sector job creation has been

declining. A second trend has been that industries vary greatly in their reliance on layoffs

for ―right-sizing‖ employment. The new data set, with its match between Bureau of Labor

Statistics data and Census Bureau data has the potential to tease out the spatial dimensions

of these job flows.

Research Ideas and Extensions

A basic research question is: Do high housing costs or shortages in available housing affect

the rate of labor turnover? The new matched BLS-Census data could be used to compare

turnover rates by place and time. Turnover at a firm will be influenced by many factors from

the industry and local economy including the availability of housing. One hypothesis is that

high housing costs lead to long commutes and higher turnover rates as workers change jobs

to fit scarce housing. An alternative view is that scarce housing has a lock-in effect whereby

workers are reluctant to move, so workers are willing to accept lower turnover and lower

wages to keep their living unit. The net impact on productivity may depend on the added

Workforce Housing Final Report 61

stress from long commutes, the quality of the job match, and the depth of capital and

technology. Ultimately the firm and urban area face competition from other urban areas

such that job creation will shift toward those areas most successful in providing a good

match to workers‘ needs and the capacity for additional growth.

Another hypothesis, following Shen‘s work, is that unskilled workers are more sensitive to

job openings and skilled workers are more sensitive to job growth.

Education and Training

One way to raise the productivity and wages of workers is to increase their education or

training. The Department of Labor has funded a number of training programs for

disadvantaged workers. Heckman, Lalonde, and Smith (1999) review some of those

programs and they find fairly limited cost effectiveness. More recent work (Heckman and

Smith, 2004) decomposes the participation process for a typical training program to show

that personal choices and awareness of eligibility can substantially affect the participation

rates and ultimate effectiveness of the program.

An alternative to general training programs is on-the-job training or work-first approaches,

which are often tied to the relative returns of job-staying and job-leaving. Holzer (2004) and

Andersson, Holzer, and Lane (2005) provide an in-depth analysis of advancement for

workers in the low-wage labor market using LEHD data for five states: California, Florida,

Illinois, Maryland, and North Carolina. They found that (p.12), ―Working in a higher-wage

industry (such as manufacturing, construction, and wholesale trade), working in a larger firm,

and working in a firm with low turnover are all associated with better pay for initial low

earners; working for a firm that pays positive wage premia is especially important.‖ The

workers who were able to transition early to better jobs at higher-paying employers and then

build up tenure at that employer did the best in the long run. Although there is great

variation among firms even at the local level, the firms that provide opportunities for

advancement of workers do so persistently. Training programs are most successful when

they match workers to firms, which provide those workers further opportunities for

advancement. Newly developed data tools by Census, such as On the Map and the

Quarterly Workforce Indicators (QWI) from the LEHD program can help to identify which

types of employers are hiring which types of workers in a local area. See Lane et al. (2006)

for the relationship between workforce development and rental housing policy.

62 Workforce Housing Final Report

Public school quality is a more fundamental way to provide skills for future workers and one

which is related to house prices. Brasington and Haurin (2006) use a hedonic house price

model and a median voter model to estimate the income and price elasticities of demand for

educational quality. They estimate the own price elasticity of demand for schooling is about

–0.5 and the income elasticity of demand is about 0.5. Using cross-price elasticities, the

researchers find that a community‘s income level, percentage white households, and level of

public safety are estimated to be substitutes for higher school quality. Not surprisingly, high-

income communities provide better schools and a safer environment in which students can

acquire skills needed for good jobs. The challenge is for firms in low-income communities to

support the school system so that graduates have the skills needed for firm advancement.

The issue of business support for local education systems parallels the issue of business

support for workforce housing. Businesses benefit indirectly by getting more productive

workers.

Research Ideas and Extensions

There are a number of research questions that build on the data and work already done:

Does the lack of affordable rental housing have a differential impact on labor

turnover and search relative to owner-occupied housing?

Are employers more likely to provide training and internal advancement when

high housing costs make it difficult to recruit replacement workers?

Are firms moving to the suburbs because suburban school systems do a better

job of preparing workers with the skills the firms need?

Returns to Skill, Creative Occupations, and Knowledge Spillovers

This section is divided into two subsections: firm level returns to skill and city level benefits

from skilled workers. The traditional labor literature provides examples whereby firms seek

out skilled workers because the firm‘s output and profitability is closely linked to the

ingenuity of workers with specific skills. Since 2000, there have been a number of papers

that show cities also benefit from skilled workers. Cities with high shares of college-

educated workers tend to grow faster and offer more opportunities for advancement to other

workers through knowledge spillovers. However, cities catering to the creative occupations

Workforce Housing Final Report 63

may also be associated with higher income inequality and higher housing costs, especially

when there are housing supply restrictions such as land use regulations.

Firm Level Returns to Skill

Juhn, Murphy, and Pierce (1993) documented the increase in wage inequality for U.S.

workers using March Current Population Survey (CPS) data from 1963 to 1989. Wages for

unskilled workers at the 10 th percentile of the wage distribution declined by 5 percent, while

wages for skilled workers at the 90 th percentile increased by 40 percent. The researchers

found that much of the increase in wage inequality was associated with returns to skill other

than schooling and years of experience. Apparently increases in demand for high-skilled

workers drove up their wages.

These findings on wage inequality have led to the study of technology‘s role. Skilled

workers can increase their productivity with advanced technology, particularly computer

technology. Abowd et al. (2005) look at the relationship between technology and a variety

of skill levels using LEHD data. Labor demand is derived from the production function by

way of a cost minimization for a given output yjt for firm j and time t:

k

jtjt

l

Hjtljtlkjtkbjt ywwZS ln)/ln(

3210

where Sbjt is the share of type b workers, b=1,...,B, and wbjt is the appropriate shadow wage

rate for type b workers, and Z measures different types of information technology capital.

The researchers have developed new measures of human capital by which they can index

the composition of firms‘ workforces. Firm data reported in the Economic Census on inputs

and outputs allow the separate identification of the technology component and multiple skill

levels. The purpose is to understand how differences in technology affect the demand for

different skill levels. One major finding is that firms that use a high degree of technology

also use high ability workers, but are less likely to use high experience workers. As stated

on page 25: ―...the capital intensity of a business, the computer investment of a business,

and the computer software expenditure intensity of a business are all positively related to

the level of human capital at the business. ...businesses that upgrade their technology are

also observed to upgrade their skills.‖

64 Workforce Housing Final Report

There appears to be great potential for this research with LEHD data because the locations

of the firms are known. The local land mix and skill composition of the employees can be

merged into a panel of data. With these data a number of interesting research questions

could be potentially answered:

Is there a connection between the high-technology firms, demand for high-skilled

workers, high wages, high housing prices and restrictive land use regulation?

Are the housing location choices of high-skilled workers with low experience

different from the choices of high-skilled workers with long experience?

Do firms that use high technology and moderately-skilled workers pay a wage

premium because the technology increases the productivity of those workers?

Is that high-technology wage premium sufficient to offset the higher housing

costs in the city?

Where do firms locate that use high tech to complement unskilled labor as

opposed to high technology to substitute for unskilled labor?

Another recent example of the interaction between technology and productivity is the work

by Jorgenson, Ho, and Stiroh (2006). They predict labor productivity growth for the U.S.

economy at 2.25 percent for the next decade based primarily on (p. 8): ―The continued

strength of the technological progress and the rising importance of investment in IT

equipment and software imply higher trend productivity growth rates.‖ However, there is a

countervailing force from the aging of the labor force, which slows the gains from education

and experience. The average labor productivity (ALP) can be decomposed into three

sources: capital deepening (capital services per hour worked), labor quality (labor input per

hour worked), and total factor productivity growth (associated with improvements in

technology among other possibilities). The researchers classify workers by sex,

employment class, age and education, then weight the hours supplied for each labor type by

their labor compensation. The labor quality growth is measured as the difference between

the growth rate of the compensation-weighted index of labor input and the index of hours

worked. Despite the impending retirement of the baby boom generation, the researchers

predict that technology-enhanced labor will continue to improve labor productivity.

Workforce Housing Final Report 65

City Level Benefits from Skilled Workers

Glaeser and Mare (2001) estimate that workers in cities earn on average 33 percent more

than non-urban workers. Some of that premium reflects a selection effect whereby highly

skilled workers are attracted to the city. However, Glaeser and Mare provide evidence that

cities make workers more productive and that a portion of the urban wage premium is paid

in higher wage growth. In fact, workers who leave the city carry with them a wage premium,

but the workers who remain in the city enjoy higher wage growth. Evidence from Census,

NLSY, and PSID confirm that older urban workers have a larger wage premium than young

workers with 0 to 5 years of experience. Peri (2002) provides additional evidence from the

1990 Census that urban workers enjoy a $2 wage premium over rural workers. Peri

suggests that highly educated workers are attracted to cities where they benefit from

learning externalities and gradually acquire a larger experience premium.

Recent work by Berry and Glaeser (2005) shows that college-educated urban workers get a

wage premium because their efforts drive city growth. In their model (p. 1), ―the clustering of

skilled people in metropolitan areas is driven by the tendency of skilled entrepreneurs to

innovate in ways that employ other skilled people.‖ In the 1970s, cities grew by adding

unskilled workers. But in the 1990s (p. 8), ―skilled cities grew by attracting skilled workers.‖

As the demand for skilled labor increases, wages to skilled workers increase. Moreover,

wage-driven demand and regulation-constrained supply lead to increases in house prices.

The net effect for unskilled workers is that they are facing higher housing prices in order to

live near skilled people in the cities.

Drennan (2005) provides further evidence at the city level that wage divergence is linked to

the share of college-educated workers. However, Gabe (2006) claims that an initial strong

presence in the creative economy is not a prerequisite for future growth. Shapiro (2006)

uses Census data from 1940 to 1990 to estimate a neoclassical growth model. He finds that

a 10 percent increase in the concentration of college-educated residents is associated with

a 0.8 percent increase in subsequent employment growth. The model (a dynamic variation

on the Roback 1982 model) also shows that about 60 percent of the employment growth

effect of college graduates is due to higher productivity growth. The rest of the employment

growth is associated with improvements in the quality of life. Endogeneity is an econometric

challenge throughout the urban growth literature. Shapiro addresses the issue by using the

presence of a land grant university as an instrumental variable for the share of college

66 Workforce Housing Final Report

graduates. Also the 1989 Wharton data on land use constraints are used in a subsample as

a robustness test. The productivity component remains the same (0.63), but the effect of

human capital operates through employment in low-regulation areas, whereas the human

capital seems to increase wages and rents in high-regulation areas. The key point is that

not all of the job growth can be attributed to the higher productivity from better human

capital. A significant portion of the job growth derives from consumption amenities. Shapiro

writes (p. 3), ―Though preliminary, this exercise suggests the effect may be operating

through the expansion of consumer amenities such as bars and restaurants (Glaeser, Kolko,

and Saiz, 2001) rather than through the political process.‖ One implication is that housing

prices are increasing through a combination of higher demand (driven by population, wages,

and desire for amenities) as well as supply constraints (land use regulation designed to

restrict housing development).

Another dynamic model of the determinants of employment growth is provided by Owyang,

Piger, Wall, and Wheeler (2006). They use a Markov-switching model to separate a city‘s

growth path into a recession phase and an expansion phase. Growth during expansions is

related to human capital, industry mix, and average firm size, whereas growth during

recessions is related to industry mix, especially the relative importance of manufacturing.

The frequency of recession appears to be associated with only the non-education measure

of human capital. Wheeler‘s previous work (2005) verified that metropolitan areas have

higher returns to skill and greater income inequality. Wheeler and La Jeunesse (2006)

introduce the spatial effect by showing that the degree of productivity or knowledge spillover

from skilled to unskilled workers depends on how the college graduates are distributed over

the metropolitan area. Using a number of segregation metrics, they show that there has

been an increase in separation from 1980 to 2000, which seems to be associated with more

income inequality during that time.

Knowledge Spillovers

One common explanation for the urban wage premium is the idea that there are human

capital externalities or knowledge spillovers that make urban workers more productive. The

challenge is to separate out the aggregate effects of human capital from the private returns

to education and experience. Lucas (1988) provides estimates that each year of schooling

is associated with 8 to 12 percent increase in earnings. See also Card (1999). In other

words, are there social returns to education above and beyond the individual returns such

Workforce Housing Final Report 67

that other workers in the local geographic area benefit from higher productivity and wages?

Moretti (2004a) presents a unified equilibrium framework with productivity spillovers.

Externalities are identified by comparing the wages of workers in cities with different levels

of human capital controlling for individual worker characteristics. Alternatively, externalities

can be identified by comparing house prices across cities controlling for individual house

characteristics.

Both endogeneity and unobserved heterogeneity complicate such econometric estimations.

The selection process is ongoing. Both firms and workers are selecting locations based on

expectations of high productivity and externalities. Workers are seeking high wages and

amenities. Firms are seeking high productivity and suitable infrastructure. The alignment

between firms and workers makes it difficult to determine whether the spillovers from

education caused productivity and agglomeration effects or high wages make possible high

education, strong institutions, and more efficient government. Moretti cites Rauch (1993) in

which Census data is used to show a strong correlation between average wages and the

percent of college graduates, but those estimates are presumed biased by sorting on

unobserved factors. Moretti (2004b) improves on the Rauch approach by using two

instrumental variables: age structure of the city and presence of a land grant college in the

city. Unobserved individual ability is handled by using panels of the National Longitudinal

Survey of Youth (NLSY). The instrumental variable estimates show that a one percent

increase in the share of college graduates in a city raises the average wages by 0.6 to 1.2

percent beyond the private return on education. Complementarity estimates of education

groups show that the least educated workers (high school dropouts) benefit more than high

school graduates from the presence of college graduates.

The spillover effects estimated by Moretti are larger than what Acemoglu, Daron, and

Angrist (2000) found. They used the state variation in compulsory schooling laws to

instrument for average schooling. The education externality was about 1 percent, but it was

not statistically significant. Moretti claims that his higher and significant estimates are

attributed to focusing on college graduates at the city level, rather than high school dropouts

at the state level. The theory states that less-educated workers should benefit from

increased wages due to both the substitution and spillover effects. Ciccone and Peri (2002)

follow a constant-composition approach. When the estimates constrain college graduates to

be perfect substitutes with less-educated workers, the externality estimates are in line with

68 Workforce Housing Final Report

Moretti and Rauch. However, when imperfect substitution is allowed, the human capital

spillovers become insignificant.

An aside, the same challenge of separating substitution from spillover effects exists for

house prices. Some of the increased demand for low-cost housing may be due to an

increased demand for unskilled workers who complement skill workers. Alternatively,

substitution of land uses away from low-cost housing to high-cost housing (teardowns or

non-residential uses) to meet the demands of skilled workers may also drive up house

prices.

For a broad literature review of innovations, spillovers and agglomeration effects, see

Feldman (1999). A prolific branch of the spillover literature focuses on patent citations

started by Jaffe, Trajtenberg, and Henderson (1993). They found that citations to domestic

patents are more likely to come from papers in the same state and SMSA. The localization

effects fade slowly over time and there is no evidence that more ‗basic‘ inventions diffuse at

a different rate than others. Breschi and Lissoni (2006) claim that spatial distance is a proxy

for social distance. Workers‘ switching between firms or social networks increases the

sharing of knowledge and the pace of learning in a city. Carlino, Chatterjee, and Hunt

(2005) attribute the higher patent density in a city to the fact that workers can be more

selective in their job matches and therefore more productive in dense cities. Many more

cites on density and agglomeration effects will be presented in the next section.

Research Ideas and Extensions

The labor market literature has provided an active research in the returns to skills and

knowledge spillovers. The logical extensions would relate those findings to the spatial

distribution of workers at their firms and of workers in their residences.

Is the degree of knowledge spillover affected by residential proximity or only

employment proximity?

Do cities with stringent regulations favor professional services and a high share

of college-educated workers?

Is the income inequality (based on residences) greater in high cost cities?

Does the high cost of housing create a barrier for unskilled workers and firms

needing to employ unskilled workers?

Workforce Housing Final Report 69

Do land use regulations exacerbate the income inequality and segregation by

income, which undercuts the spillover benefits for unskilled workers?

Is it possible to distinguish the substitution effect from spillover effect from land

use regulation effect?

The last bullet refers to the idea that less-educated workers should benefit from higher

demand (substitution effect) as firms react to high wages of skilled workers and take

advantage of spillover effect (unskilled more productive by working with the skilled workers

and benefiting from their innovations). It is conceivable that the higher house prices driven

by increased demand for skilled workers are not completely offset by the productivity and

wage gains of unskilled workers such that unskilled workers are better off (despite the

increase in wage dispersion and increase in housing costs). However, it seems more likely

that unskilled workers would be worse off because increases in housing costs exceed their

increases in employment and wages.

70 Workforce Housing Final Report

Chapter Five: Spatial Mismatch and Job Search

The NY Times (McGeehan, 1/26/2007) announced, ―Building Boom May Mean Jobs Can‘t

Be All Filled, Report Says.‖ The construction boom in New York City is having difficulty

finding enough workers to fill the 275,000 positions in the building industry. Mr. Anderson,

president of the New York Building Congress, said that contrary to conventional wisdom the

typical construction worker is not a white suburbanite who commutes, but ―a younger city

resident who is almost as likely to have been born abroad as in America.‖ The boom has

bolstered strong city finances, but the rapid development of expensive housing has created

an affordable housing crisis driving out municipal workers and middle-class families. This

chapter focuses on the differences in distribution between employment and residences that

spawn frequent relocation and migration.

The interface between the labor and housing markets is most apparent when considering

where workers choose to live relative to their employment, where they look for new jobs, and

where they choose to migrate when commuting is no longer feasible. In short hand, this

chapter is about three topics: spatial mismatch, job search and migration. The spatial

mismatch grew out of the work by John Kain (1968) who pointed out that employment

opportunities were primarily in the suburbs, which were a long way from minority households

living in the center cities. Although the research has broadened to consider the metropolitan

distributions of employment and housing, there continue to be large spatial gaps that require

a car for workers to get to many suburban jobs.

The second section addresses job search and worker displacement. A key finding is that

high-skilled workers search over a broader geographical area (often nationally), while low-

skilled workers confine their search within the local commuting sphere. As a result, high-

skilled workers move more and are unemployed less than locally bound workers. Mobility

interacts with homeownership in that high-skilled workers are more likely to be homeowners.

Moving costs reduce the job search effort, and it seems to diminish the search effectiveness

of low-skilled workers more than high-skilled workers.

Migration entails many of the same trade-offs as commuting but on a grander scale. The

traditional economic view is that firms relocate to take advantage of relative wage

differentials. Both firms and workers are expected to keep moving until they reach an

equilibrium in wages controlling for the cost of living. Another perspective is that firms

Workforce Housing Final Report 71

relocate to service growing population demands, and amenities compensate for regional

differences. Regional convergence literature has given way to models justifying regional

divergence. Given the endogenous relationship between migration and firm relocation, it

may not be possible to resolve this chicken-and-egg relationship, but it is clear that high

housing prices can counteract the in-migration to a strong labor market. And, expected

housing gains and expected wage growth can offset high levels of house prices, feeding an

upward spiral in house prices. Census reports that people are moving away from areas of

high-cost housing and slow growth, which almost certainly ties in with land use restrictions

that slow the development of workforce housing. And, circling back to the spatial mismatch

concerns, most new jobs seem to be captured by in-migrants rather than local residents.

Spatial Mismatch

John Kain was instrumental in launching the spatial mismatch literature (1968) as well as

providing a comprehensive review of its development twenty-four years later (1992). The

issue began with a focus on housing market discrimination, i.e. segregation in housing kept

minorities from relocating in the suburbs where jobs were plentiful. The high unemployment

of young black men in the center cities was associated with spatial mismatch. The young

black men living in center city ghettos did not have access to housing, transportation, or the

connections that would make it possible to get a higher-wage job in suburban businesses.

The solutions proposed by Kain and others entailed:

integration of housing to allow minorities full access to suburban housing,

job training and placement with suburban employers,

wage subsidies to encourage employment of long-term unemployed,

integration and improvement in public schooling.

Briggs (2005) provides an update on the discrimination and segregation issues.

The modeling of spatial mismatch has been updated by Brueckner, Thisse, and Zenou

(2002) who link the skill space of workers with the physical space of firms and cities. In their

model the local labor market is an oligopsony in which firms compete for heterogeneous

workers and firms have some market power over the workers who live in their vicinity. The

degree of monopsony power exerted by firms depends on the labor supply elasticity in the

firms‘ labor pool area. High costs in acquiring skills and commuting to work are associated

72 Workforce Housing Final Report

with an inelastic supply of labor. Workers bear the training cost that brings their innate skills

in line with the firm‘s needs. Low-skilled workers are assumed to commute the greatest

distance because their low wages correlate with a low value of time and toleration of a long

commute. The authors conclude that socioeconomic ghettos develop because workers with

low skill matches also incur the highest commuting costs.

Morrison (2005) also emphasizes the importance of heterogeneous labor in synthesizing the

conflicting views on localized sub-markets with varying unemployment rates. In one model,

labor demand is confined to labor sub-markets within the metropolitan market. High

unemployment in those pockets is due to a shortage of jobs within a short commuting

distance of workers living there. The solution is to increase labor demand by attracting

private businesses or government offices into those local labor markets. A second model

views the entire city as a single market for labor and capital. Localized unemployment is

due to housing market clustering by income. Low-cost housing is associated with low-

income families and high unemployment. In this view, spurring labor demand might

increase overall output of the city, but the high-wage jobs will mostly go to people

commuting from high-income neighborhoods. The long-run effect is that unemployment

remains high in the low-income neighborhoods. Morrison claims that both models are

relevant, but for different types of workers. Skilled, high-income workers serve the entire

metropolitan market, whereas unskilled, low-income workers limit their job search to local

labor sub-markets. If demand is slack in a worker‘s own neighborhood but tight in another

neighborhood, the unskilled worker will not be able to commute to the opening as readily as

the skilled worker.

As a case in point, Bruce Katz (2007) testified that the District of Columbia accounts for 34

percent of the region‘s high-wage jobs (paying over $75,000 per year), but only 20 percent

of the low-wage jobs (paying under $35,000 per year). Low wage jobs are twice as

dispersed as high-wage jobs.

Concerns about the decline in demand for unskilled workers during the 1980s (Freeman,

1991) were somewhat ameliorated during the protracted expansion of the 1990s (Freeman

and Rodgers, 1999). Based on the 2000 Census, Raphael and Stoll (2002) report that

during the 1990s the spatial mismatch for blacks improved by 13 percent. The correlation

remains that metropolitan areas with a high degree of segregation also had a high degree of

spatial mismatch. The residential movement of blacks toward the suburbs and suburban

Workforce Housing Final Report 73

jobs helped reduce the spatial mismatch. Martin (2004) used the decennial Census results

for 1970 to 2000 and determined that both the jobs and housing distributions were gradually

suburbanizing. Ironically, residents tended to move away from areas gaining jobs.

Employers moved to the suburbs to be closer to the growing population centers, but that

enabled many suburban workers to locate even farther from downtown. The net effect is

that jobs are following the general population and blacks are following the suburban jobs.

This pattern lowers the commuting times for whites, but it raises the commuting times and

distances for blacks.

Using the American Community Survey (ACS 2005) and the 2000 Census, Berube and

Kneebone (2006) report on another trend in America‘s 100 largest metropolitan areas. By

2005, poverty in the suburbs exceeded poverty in the cities by 1 million. However, poverty

rates were twice as high in the cities as the suburbs (18.8 vs. 9.4 percent) and poverty rate

increases were predominantly in the Midwest and South. The manufacturing cities of

Cleveland, Toledo, Detroit, and Columbus were hit hardest, and those cities have some of

the most affordable housing in the country. The annual ACS data make it possible to track

changes in spatial mismatch for large cities, which continue to expand even when growth in

the core of the city has stalled.

Vermeulen and Ommeren (2004) have calculated from Dutch data that house prices are

10.4 percentage points lower and rents are 2.4 percentage points lower when regional

unemployment is one percent point higher. Their conclusion is that workers are

compensated by cheaper housing in high unemployment areas. That explanation sounds

more appealing than workers being paid less because their productivity is low or land is

cheap because people don‘t want to live there.

Some research has called into question earlier attempts to ―solve‖ the spatial mismatch

problem. It seemed clear enough to California planners in the 1990s that there should be a

balance between jobs and housing. Counties that favored businesses for their high tax

revenues and low burdens on municipal services forced their neighbors to supply housing

for workers. However, Cervero (1996) examined the San Francisco area and found little

association between the jobs-housing balance and self-containment to reduce commuting.

Even in balanced towns, there was a high degree of crisscross commuting as workers

sought out the best match for their skills.

74 Workforce Housing Final Report

A spatial analysis of job openings in Boston by Shen (2001) showed that job openings

suitable for less-educated job seekers is relatively concentrated in the center city. Most

openings are from turnover rather than the creation of new positions. For jobs requiring few

skills or experience, there was actually an advantage to living in the city where more

openings occurred. However, the most important aspect in determining access to

employment was access to a car. Suburban jobs are very difficult to reach by transit.

Where a person lived is less important than whether the person could use a car to reach

potential employers.

Finally, Bayer and Ross (2006) have determined that the net effect of neighborhood quality

on labor market outcomes is actually small. A fundamental challenge in labor market

studies is separating the neighborhood sorting from the individual skills. The researchers

have devised a very clever instrumental variable strategy. They instrument for observed

neighborhood attributes of an individual with the average neighborhood attributes of

observationally equivalent individuals. This approach eliminates the portion of the variation

in neighborhood attributes that is due to sorting on unobserved individual attributes and

produces unbiased estimates of the neighborhood impact on employment and wages. This

strategy seems well suited for other spatial studies where endogeneity is common.

Research Ideas and Extensions

The main concern in spatial mismatch is that workers may have more difficulty in obtaining

employment at good wages because they live so far from work. The concerns of workforce

housing are similar in that the lack of housing at a reasonable cost and distance from work

will make it difficult for workers to find employment and firms to find sufficient workers. The

logical extension of the spatial mismatch literature is to determine whether land use

restrictions have played a role in increasing commutes or weakening the matching process

between workers and firms. For example, in cities where stringent land use has reduced the

availability of moderate-cost housing:

Do workers have longer commutes?

Are there sub-markets of high unemployment and low wages?

Do workers without cars have lower wages controlling for age and education?

Are minority workers living in the center city less likely to have jobs with suburban

employers than minority workers living at an equal distance in the suburbs?

Workforce Housing Final Report 75

Do unskilled workers have longer commutes because there is not moderate-cost

housing close to their employer, or shorter commutes because they only take

jobs close to their home?

Do land use restrictions exacerbate racial or income segregation, which, in turn,

increases the problems of spatial mismatch for workers and firms?

Job Search

There is a vast literature in labor economics on job searching and wage determination

including many recent developments in game theory and strategic bilateral bargaining.

Covering these topics is beyond the scope of this review, but fortunately Mortensen and

Pissarides (1999) have written a comprehensive review in the Handbook of Labor

Economics, Vol 3b. We focus on more recent articles that consider the role of the housing

market and homeownership. In addition, there are several interesting articles on worker

displacement from plant closings, which force the worker to search for an alternative job

within local commuting or migrate to a new city.

Thomas Dohmen (2000) provides a theoretical model designed to fit a number of empirical

regularities that he sees:

Gross flows in migration are much larger than indicated by net migration flows.

Most of the rise in unemployment is due to long-term unemployment rather than

higher inflow rate.

Rates of homeownership are positively correlated with unemployment suggesting

that impediments to mobility play a crucial role in unemployment.

Moving costs are higher for owners, but mobility increases with income.

The Beveridge Curve has shifted outward meaning the number of unemployed

has increased relative to each vacant position.

Skilled workers are unemployed less often and for shorter durations than

unskilled workers.

Skilled workers move more, search more, and have a preference for search while

still employed (Pissarides and Wadsworth, 1994).

Dohmen cites work by McCormick (1997) that finds little evidence that manual laborers

move to low unemployment markets as predicted by the neoclassical labor model.

76 Workforce Housing Final Report

Moreover, Evans and McCormick (1994) provide evidence that non-manual labor is flexible

and geographically integrated, while manual labor is spatially rigid. If mobility and search

behavior are at least partially affected by the housing market, then the wage equation and

matching models should include housing market measures. Therefore, Dohmen creates a

labor market model with differing degrees of mobility based on moving costs by class of

worker. The theoretical model shows that higher moving costs reduce mobility and increase

unemployment. Also, indirectly the higher moving costs raise the reservation wage, which

reduces job acceptance rates and prolongs unemployment. Asymmetric shocks increase

the number of workers moving to the booming city and reduce the number leaving that city.

Mobility increases the most for high-skilled workers with higher wages. Those skilled

workers are the first to leave the depressed region and the first to take advantage of the

higher wages in the booming city. As a result, skilled workers are unemployed less than

unskilled workers.

One view of land use restrictions is that they increase moving costs because inelastic supply

of housing is less responsive to fluctuations in demand from labor. Furthermore, zoning that

restricts rental housing may disproportionately affect unskilled workers, raising their search

costs and reducing their mobility especially if the positive demand shocks for labor occur in

the restricted city. The model does not address commuting issues, nor how an increase in

house prices is related to moving costs, but presumably moving costs are correlated with

house prices and wages.

Several recent papers have provided empirical tests for Dohmen‘s model. Bound et al.

(2004) find that the production of college graduates in an MSA is only weakly related to the

share of graduates living in the MSA. This evidence supports Dohmen‘s idea that skilled

workers are mobile. Munch, Rosholm, and Svarer (2006) examine data for Denmark. They

conclude that homeownership has a negative impact on job-to-job mobility both in terms of

finding a new job in the local labor market and finding a job in a different labor market. The

research also identifies a negative relationship between homeownership and unemployment

risk, but a positive relationship between homeownership and wages.

McQuaid‘s (2006) research on job search found that professional qualifications, ‗soft‘ verbal

skills, and the willingness to send unsolicited applications to employers were significantly

associated with success in job searches. McQuaid‘s econometric models verified that

length of unemployment, age, and manual occupation in the last job could reduce re-

Workforce Housing Final Report 77

employment success by as much as thirty percent. High academic qualifications and the

expectation of promotion raise reservation wages and hurt the chances of finding acceptable

new jobs. Particularly germane, McQuaid found that geographical accessibility to local jobs

was a significant and positive factor in job search success.

Wheeler (2006) drew a sample of young men from the National Longitudinal Survey of

Youth (NLSY79) to study how men experiment with job changes as they establish a career.

He found for the first job change that the likelihood that a worker changes industries

increases with the size and diversity of his local labor market. However on successive job

changes, the association with size and diversity wanes. By the fourth change, the likelihood

of changing industries actually has a negative relation to the scale and diversity of the local

market. Wheeler concludes that city size and the opportunities offered by local market scale

play an important role in the job matching process.

Job search is often tied to job loss and worker displacement. Hamermesh (1989) provides a

good review on worker displacement facts updated by Farber (1997). Rephann, Makila, and

Holm (2005) use that background to formulate a microsimulation model of the local labor

market impact from an automotive plant shutdown. The model is benchmarked to

households in Sweden for 1985 to 1995 with Monte Carlo simulation for stochastic

transitions. The migration module allows for both intra- and interregional moves. A

conditional logit model predicts the destination of the move based on measures of distance

between origin-destination labor markets, population of working-age residents with earnings

at the origin and destination labor markets. The model includes spatial, social, and

economic factors aggregated at different geographical scales to impact the workers‘ location

choices and employment prospects.

The microsimulation model was tested against the results of a Saab plant shutdown and

accurately predicted that reemployment for older (50+) people is low, men get reemployed

more quickly than women, more educated and native workers have an easier time finding

new jobs than other workers. Net out-migration from the region of the plant is reversed after

several years as workers return (perhaps due to low house prices). Average wages decline

and many workers get more education following the plant shutdown. Many laid off workers

choose early retirement rather than leave their houses, though the model does not detail

housing market dynamics.

78 Workforce Housing Final Report

An obvious extension of the model would be to incorporate a housing market so that

migration decisions reflected the availability and regional cost of housing. The model

assumes infinitely elastic regional labor demand, and no constraints on housing supply. The

spatial detail of the microsimulation model allows for land use regulations to impinge on the

housing supply elasticity, which would affect migration decisions. A transportation module

could also be added to predict the impact on commuting patterns and congestion.

Presumably, a more extensive transportation system would allow workers to search and

obtain employment in the local market. Such job replacement would avoid the cost of

migration to another city, but also could reduce the worker‘s wage profile. The ability of a

microsimulation model to accommodate spatial detail seems like an ideal platform for testing

the impacts of land use restrictions in the housing market.

Regional Adjustment Models

Regional migration models can be viewed as either disequilibrium or equilibrium models.

According to Carruthers and Vias (2005), the disequilibrium view of migration is that people

move to areas with more jobs and higher wages. Spatial equilibrium will gradually develop

when no one can improve his/her utility by moving. In this view, people follow the firms,

which create the demand for workers (Greenwood, 1985). The equilibrium view of migration

is that compensating differentials in wages satisfy an equilibrium condition all along.

Differences in amenities, cost of living, employment security, and growth prospects

compensate individuals for the difference in wages, but net of those effects people receive

the same compensation at least in terms of utility. This equilibrium view corresponds to the

population deconcentration view that people are dispersing to less urban locations to take

advantage of reduced transportation and communication costs. Firms follow this outward

dispersion in order to provide products and services to that population.

Carruthers and Vias consider these two views of migration in the context of the Rocky

Mountain states. They have data for 8 states from 1982-1997 in 5-year increments. During

this time the total amount of developed land in the Rocky Mountains grew by two million

acres, or about half an acre per person. Low-density urbanization has emerged as the

dominant mode of growth. The two-equation model estimates the log change in population

density and the log change in employment density using two-stage least squares (2SLS).

The data are collected from the Census of Governments, County Business Patterns,

Economic Research Service, National Resources Inventory, and the Regional Economic

Workforce Housing Final Report 79

Information System. In their model, Carruthers and Vias hypothesize that growth has

diminishing spatial effects at the development margin. Zoning predetermines the land

development pattern rather than a free-market process. The authors write (p. 31), ―To be

clear: low-density zoning commits large tracts of open space to eventual urbanization, even

if significant time passes before they are completely filled in. Zoning is well known to work

against the underlying market conditions (Fischel, 1999) and has been cited as a factor that

may disrupt the free-market process that adjustment models simulate.‖ The first developed

plots have a large spatial impact, which gradually tapers off as more people and firms add

on. Increases in density are smaller and smaller. Lagged log of employment is used to

instrument for log employment density and lagged log of population is used to instrument for

log population density. In other words, the model expects and finds concurrent population

density and employment density to have positive impacts on one another, but the lagged

own effects will be negative as denser places realize smaller proportional changes. Thus

the changes are measured in logs rather than first differences.

Carruthers and Vias conclude that what is needed is more active participation by state and

local governments to preserve open space so that the existing economic development is

sustainable. Environmental quality is vital to continuing economic prosperity, though it is not

clear how the estimated regressions prove their point. The logic is that jobs follow people

and people will stop migrating in when the natural amenities are overrun by too many

people. The mathematical relationship verified by the regressions is that the first hundred

people have a bigger impact on density than each succeeding hundred.

One of the underlying ideas embedded in the disequilibrium view of migration is that there

are regional ripple effects as waves of migrants move between linked housing markets.

Jones and Leishman (2006) test this notion on a smaller scale by studying the linkages

between local housing market areas using private house sales for Strathclyde, a sub-region

of Scotland, with Glasgow, the leading housing market in the region. The researchers find

that the extent of common price dynamics and the length of lagged response in the

secondary markets depend on the rate of migration from the primary housing market to the

secondary markets. Tolbert, Blanchard, and Irwin (2006) make a related point that

commuting zones, rather than political boundaries, are more useful for measuring the

degree of intra-urban vs. inter-urban moves.

80 Workforce Housing Final Report

As a final point on the impact of migration, Ottaviano and Peri (2006) determine that there is

a net positive effect of cultural diversity on productivity of native workers. In cities where the

share of foreign-born residents increased between 1970 and 1990, U.S.-born citizens living

in those cities experienced significant increases in wages and rents. The result holds up

through various attempts to cope with omitted variable bias and endogeneity bias. It would

be interesting to see what happens to these results in the cities where house prices have

risen so high relative to wages that people are moving out to lower cost cities. Soureli,

Pelletiere, and Wardrip (2006) find a counter-movement since 2000 of people leaving high-

cost cities. Are the out-migrant households coming from the below-average wage portion of

the high-cost cities? Their move could raise average wages in the cities they are leaving

and increase wages and rents in the cities to which they are moving. It is necessary to

separate the sorting effect of migration from the complementary effects of raising native

productivity implied by Ottaviano and Peri.

Research Ideas and Extensions

Following on Dohmen‘s research, it would be useful to know whether the land use

regulations of a city affect the job search and mobility for different classes of workers.

Metropolitan areas could be ranked by degree of regulation and a mobility model could be

benchmarked on the patterns in the middle third of metro areas. Then the predictions of the

model could be compared to the actual patterns in the most regulated versus the least

regulated to see what kinds of differences are associated with high and low degrees of

regulation. Alternatively, the regulation index could be incorporated into the model to see if

land use regulations have a significant impact. Our prior understanding is that regulation

increases moving costs and decreases mobility, especially for less-skilled workers who have

fewer choices in affordable housing. For example,

Do regulations dampen the degree to which higher-skilled workers have higher

rates of mobility and lower rates of unemployment than less-skilled workers?

Do regulations increase the moving cost, frequency of moves, or distance of

moves?

Is there an indirect impact of regulations through homeownership in that:

- a shortage of rental housing makes the labor force less flexible?

- high homeownership rates and high housing prices reduce the mobility of

owners?

Workforce Housing Final Report 81

Are there differences in commuting patterns by race that are exacerbated by land

use regulations (after controlling for income, education, neighborhood, car

ownership, etc.)?

Do regulations affect the expectations of owners for future house price gains or

the expectations of workers for wage gains?

Census data including the ACS data provide some information on the frequency and

distance of moves as well as education, industry, occupation, ownership status, immigration

status, and employment status of workers. It seems likely that mobility patterns are

influenced by the industry and occupation of the worker, which must be controlled for to get

an unbiased measure of the regulation effect. Moving cost information may be available

from the Consumer Expenditure Survey or the Survey of Consumer Finances.

82 Workforce Housing Final Report

Chapter Six: Commuting and Transportation Networks

This chapter focuses on the challenge of workers commuting from their homes to their jobs.

In most cities there is a trade-off between house prices and commuting costs. More

affordable house prices at the suburban fringe are offset by longer and more expensive

commutes. Relatively inexpensive car ownership and operation makes it possible for

workers to live a considerable distance from their workplace. However, Rouwendal and

Nijkamp (2004) point out that traditional urban economic models predicted that higher

income individuals would not be willing to make long commutes. In fact, there seems to be

―excess‖ commuting time, which suggests that commuting cost is only one factor in a

household‘s choice of how far to live from work. Also, cheap transportation has not

diminished the importance of space. The first section of the chapter examines the combined

burdens of housing and transportation costs as well as the trade-off facing workers and

employers as they figure out how to co-locate.

The second section considers the pay advantages from commuting. One motivation for a

long commute is to obtain less expensive housing. Another motivation for extending the

commute is to gain higher pay.

From the urban planners viewpoint, a separation of employment centers from residences

creates a daily challenge for the transportation network. Congested highways create

barriers between bedroom communities and downtown offices. One response to the

congestion is to require a balance between jobs and housing so that workers could live

closer to work and reduce the amount of commuting needed. Though office parks make

better neighbors than factories, requiring housing to be in the vicinity of employment centers

does not guarantee that workers will lessen their commute. Even in well-balanced

communities, there persists a great deal of criss-cross commuting as workers attempt to

optimize their job match. Furthermore, more compact cities may shorten commuting

distances, but exacerbate the congestion as the same number of cars try to pass one

another in more confined spaces.

The final section focuses on the measurement of congestion and the prospect for

congestion pricing to reduce the delays during rush hours. Schrank and Lomax (2005)

estimated for 85 urban areas that the annual delay for commuters stuck in traffic jams has

risen from 16 hours in 1982 to 47 hours in 2003. Road pricing, time-of-day tolls, or a fuel tax

Workforce Housing Final Report 83

have the potential of reducing congestion and promoting telecommuting, but it is less clear

how these solutions will affect worker productivity or the spatial distribution of housing.

Downs (2004) points out that congestion is fundamentally a reflection of our prosperity. The

more productive and higher paid the workers are, the more likely they will choose larger cars

and larger houses in the more distant suburbs.

Trade-off Between Housing Costs and Commuting Costs

The spatial mismatch literature emphasizes that spatial separation of housing from

employment makes it more difficult for workers to find and commute to good jobs, so they

are more likely to be marginally employed or take lower paying, local jobs. High housing

costs for units close to employment centers force workers to occupy less expensive housing,

which is more distant from their jobs. Workers trade-off high housing costs near to work for

lower housing more distant from work, but higher commuting costs. In 20 of the 28 MSAs

examined, Haas, Makarewicz, Benedict, Sanchez, and Dawkins (2006) find that housing

affordable to moderate-income households is more distant from jobs. The researchers

gathered data from the 2000 Census and the Consumer Expenditure Survey on 29,607

census tracts to measure the incomes, housing costs, transportation costs, job accessibility,

and commuting characteristics of working families. The analysis was done on households

that moved within the last five years. On average, working families spend 28 percent of

their income on housing and 29 percent on transportation for a combined housing plus

transportation cost burden of 57 percent. The combined cost burden is remarkably constant

across the metropolitan areas with a low of 54 percent in Seattle to 63 percent in Chicago.

Places in which housing is relatively less expensive, like Cincinnati and Detroit, are places

that spend a larger share on transportation. Whereas, workers in high-cost places, like

Boston and Los Angeles, spend more on housing but less on transportation. Working-class

families, defined as households with income between $20,000 and $50,000, can save on

housing costs by living farther away from their jobs, but the added cost of commuting

generally increases their combined cost burden above 50 percent.

Haas et al. (2006) categorize each census tract into one of four categories according to

share of income spent on housing (H) and transportation (T) with percent of households in

parentheses:

1) Low percent on H and T: wealthy suburban community (38%)

84 Workforce Housing Final Report

2) Low percent on T, but high percent on H: mixed income urban community (16%)

3) High percent on H and T: lower income urban/inner-suburban community (26%)

4) Low percent on H, but high percent on T: moderate-income exurb (20%)

Commuting by auto dominates in all four quadrants with more than 85 percent of working-

class families driving to work (as summarized by Lipman, 2006). The highest share of

transit use (27 percent) is in the low T/ high H quadrant for which commuting time is an

average 46 minutes compared to auto commute times of 27 minutes. In fact, the auto

commute times are nearly the same (28 minutes) in the opposite quadrant of low H / high T.

The three-percent transit riders from this quadrant have the longest commute averaging 64

minutes.

In a related study using Census 2000 data by Cervero, Chapple, Landis and Wachs (2006),

seven large MSAs are selected as case studies to study how working families trade off

housing costs and commuting times. Researchers estimated bid-rent functions on the log

housing cost per room for different household types and controlling for commute times with

various interactions. Strictly speaking, commute times are jointly chosen with housing, so

school test scores by PUMA are used to construct an instrument. For married-with-children

households, a one-way commute of 18 to 20 minutes is associated with a housing cost

premium of $500 to $570 per room per month. Other household types have flatter bid-rent

curves indicating a smaller premium for a short commute.

In terms of housing elasticity of demand with respect to commute time, married households

with children decreased their housing costs by 2.6 percent for each 10 percent increase in

commuting time. Other household types had elasticities about half of that. Upper-income

households, who have the most choice in housing locations, decrease their commute time

by 8.9 percent for each 10 percent increase in income. Working families in the lowest third

of the income distribution have an elasticity of 3.5 percent, which is just above homeowners

at 3.2 percent. By comparison, renters, who are generally more mobile than homeowners,

have an elasticity of 8.9 percent and recent movers (within last 5 years) have an elasticity of

15.1 percent. This suggests that recent movers are the most sensitive to commute times.

The elasticities are expressed as a percent decrease in housing costs relative to a 10

percent increase in commute time. People who move from a city apartment to a house in

the suburbs would normally be increasing their commute times for the same job unless they

switched from transit to car mode or switched to a suburban job. One interpretation of these

Workforce Housing Final Report 85

results is that the suburbanization of jobs has facilitated a further suburbanization of

workers. Another view is that workers may be willing to pay a little more for workforce

housing that is substantially closer to their job, but low-income workers and homeowners

seem the least likely to move for that reason.

Cervero et al. (2006) also estimated discrete choice models in which the choice of commute

mode is nested within or conditioned by the choice of residential neighborhood. Separate

multinomial logit estimates are made for each of seven MSAs and six different household

types within each MSA. The choice of neighborhood (measured by PUMA) is a function of:

household income, household size, tenure, condominium, sex and age of household head,

average high school test scores by PUMA, and auto accessibility to work by occupation.

The choice of commute mode conditional on choice of neighborhood is a function of: auto

availability, vehicles per worker, auto accessibility to work by occupation, sex and age of

worker. Although referred to as choice models, the estimated probabilities reflect the

availability of housing units, jobs, and transportation networks built up in previous rounds as

much as the preferences of workers in the current round. In other words, the short term

choices are conditional on the long term patterns of infrastructure and development.

The discrete choice models show that (p.31), ―Working family households consisting of

married-couples with children, for example, are 86 percent less likely to live in one of

Atlanta‘s central city PUMAs ... and 49% more likely ... to live in one of its outer suburban

PUMAs than are wealthier married-couple-with-children families.‖ Dallas-Ft. Worth and

Washington, D.C. follow a similar pattern of working families living toward the outside of the

city. Working families in Chicago and New York are more likely to live near secondary

employment centers rather than new suburbs. Los Angeles and San Francisco are more

likely to have working families living in older, center city neighborhoods.

The commuting mode choices for working families are similar to wealthier households

because cars dominate across the board. Working families in Dallas and L.A. are more

likely to carpool to work from their center city houses. Also, working families in New York

are more likely to walk to work. When the neighborhood types are sorted by the ratio of

transit-to-auto commute time, Chicago and New York follow the expected pattern that outer

suburbs and suburban fringe communities have much longer commutes than central cities.

However of Atlanta‘s neighborhoods, the central city transit commute times are relatively

longer and the suburban fringe has the least disparity relative to auto commute times. So

86 Workforce Housing Final Report

many of Atlanta‘s jobs are near the Perimeter Freeway, that working families living

downtown take twice as long to commute via transit as by car.

While it is difficult to summarize the results overall, Cervero et al. (2006) can characterize

the housing burdens and commuting burdens of working-class families by MSA. For

example, in Atlanta working families face high house cost burdens in fringe communities

and high commuting burdens in downtown neighborhoods. In Chicago, the housing burdens

are more uniform, and all working families except those living downtown face significant

commuting burdens. The number of MSAs are too few to estimate patterns for the typical

MSA or to reach a deeper understanding of the forces creating these patterns. The data is

from a single cross-section of the 2000 Census. Extensions over time and explanatory

variables could help highlight the role that land use restrictions played in shaping the

residential and transportation choices made by workers.

Nelson (2004) projects workforce housing needs by household size and income bracket

using the Nationwide Household Transportation Survey and then assigns the demand by

occupation using the BLS report Occupational Forecast 1998-2008. Census data provide

the geographic distribution by occupation. The embedded assumption is that these

distributions by income, occupation, and location will not change much during the forecast

period of thirty years. Although this approach is unlikely to produce highly accurate

estimates, it does convey the point that workforce housing needs are not currently being met

and the disparities are likely to increase in the future.

The advantages of less time-consuming transportation connections can be measured in

house prices. Sichelman (2007) reports on NAHB research showing that housing near

public transit stops is higher by 12 percent. Similarly, Mikelbank (2004) shows that past,

current and approved (but not yet begun) road projects have significant impacts on house

values.

For our purposes, an ideal study would relate the accessibility of workforce housing to labor

productivity. Prud‘homme and Lee (1999) use French and Korean data to show that the

elasticity of labor productivity with respect to commuting speed is +0.30. Workers with

shorter commutes are somewhat more productive, which implies a city with an efficient

transportation network is more productive. Cervero (2001) tests the idea of efficient

urbanization on a cross-section of 47 cities from 1990 and specifically the San Francisco

Workforce Housing Final Report 87

Bay area. Although there appears to be some economic advantages from larger laborsheds

and good accessibility between jobs and housing, Cervero concludes (p. 1668),

―Statistically, the elasticities between labour productivity and both labour-marketshed size

and labour accessibility were fairly small and were not highly significant.‖ He holds out the

hope that the measures might prove more significant in a model with many more MSAs.

Cervero‘s research also supports a ―peculiar‖ relationship between economic performance

and congestion (p. 1668), ―Across metropolitan areas, more crowded freeways appear to be

a consequence, in part, of rapid economic growth. Accordingly, an inverse statistical

relationship between average travel speeds and productivity levels was found.‖ The author

notes that during the recession of the early 1990s, the problem of traffic congestion greatly

dissipated.

There appears to be back pressure such that productive cities grow and in so doing

generate so much traffic that the roadways become congested. As Downs (2004) is fond of

touting, congestion is a reflection of urban growth and success. The broader point may be

that the availability of workforce housing must be considered in the context of the existing

transportation infrastructure. Expensive worker housing, just like congested highways, may

reflect economic growth and success as the urban system reaches capacity as well as

signals future slower growth as companies cannot retain or recruit more workers.

Alternatively, expensive workforce housing could result from under-provision either due to

regulation frustrating new supply or competition from higher-valued alternative building

projects.

Another measure of accessibility‘s importance to economic outcomes is whether workers,

given better access to employment, are more likely to find a job. Sanchez, Shen, and Peng

(2004) analyzed data on Temporary Assistance for Needy Families (TANF) recipients in six

selected MSAs. The ordered multinomial logit model regressed employment on measures

of access to public transportation and employment centers. From the spatial mismatch

literature, the authors expected that good access was significant in TANF recipients finding

employment. However, the results showed virtually no association. The lack of recent work

experience may weaken the application of this research to broader populations.

The shift from a monocentric city to a polycentric form can affect commuting and

accessibility. Cervero and Wu (1998) estimate that the subcentering of employment means

23 percent more commuting as workers travel farther to get to their jobs, which are less

88 Workforce Housing Final Report

centrally located. Eighty percent of the additional vehicle miles are attributed to longer

distances between home and work. However, the travel times are actually shorter, in part

due to less traffic congestion. In another interesting study on dispersing employment using

AHS data, Crane and Chatman (2003) find that a 10 percent increase in employment is

associated with a 3 percent reduction in the average commuting distance. If the growth in

employment occurs in the same direction as where people live, growth does not necessarily

increase commuting distance.

For a number of years, commuting times remained fairly stable, and there developed the

rational locator hypothesis to explain that phenomenon (Levinson and Kumar, 1994;

Levinson and Wu, 2005). The idea is that a worker has a personal commuting budget and

will arrange work-housing combinations that fall within an acceptable range of time.

Examining Washington, D.C. data, Levinson and Wu find a stable journey-to-work time from

1957 to 1988 during a long period of suburbanization. Commuting distances and

congestion increased, but average durations remained stable or declined modestly as

workers shifted from slow trips downtown to faster suburban routes. However, using more

recent data from 1994 and a larger geographical measure of the metropolitan area, the

researchers found increasing commute times. Also, a comparison to Twin Cities data

provided evidence that the spatial structure of the metropolitan area influences commuting

times. Thus, Levinson and Wu reject the theory of personal commuting budget by noting

that commuting times seem to change over both time and place.

Two theoretical models belong in this section on the joint determination of residential

location and commuting mode choice. The first is by DeSalvo and Huq (2005), which

extends the Alonso (1964) and Muth (1969) monocentric models by including mode choice

and a total time budget constraint. Anas (1982, 1999) is credited with integrating a discrete

and stochastic mode choice into an urban residential location model. The commuting mode

is defined in terms of average speed (mode cost is linear in time), and the mode choice

variable is continuous. The comparative static results show that commuters should choose

higher speed modes for longer commutes, at higher wage rates, with greater tastes for

housing, and with lower housing prices. Assuming housing is a normal good, workers with

higher non-wage income should live farther from the CBD. The model can also show that

marginal commuting cost rises with an exogenous increase in housing price. Land use

regulations that lead to higher housing prices would presumably raise marginal commuting

Workforce Housing Final Report 89

costs, though this is not treated explicitly in the model. Some of these results become

ambiguous if mode cost is allowed to be nonlinear in time.

A second theoretical model of interest is by Rouwendal (2004) on search theory and

commuting choices, which incorporates spatial considerations into a standard model of job

search. The model is used to develop isochrones or acceptable commuting distances such

that the decrease in house prices is not more than offset by an increase in commuting costs.

Extending the model for repeated search helps resolve the empirical finding of apparently

excessive commuting. Long commutes may be followed by short commutes on the next

round of searching. Both draws may be from the same distribution, which will approach a

more moderate commuting distance on average after multiple draws. Also, even if the firm

makes the same offer to all workers, the spatial dispersion of workers means the wage net

of commuting cost varies considerably. The spatial dispersion explains the tradeoff between

a higher offered wage and a longer duration of vacancy even when the workers have the

same reservation wage.

Rouwendal and Nijkamp (2004) provide a critique of commuting models. The monocentric

city model assumes workers do not like to commute, but not all workers can live close to

their jobs clustered in the central business district. Therefore, housing prices are distributed

so that closer houses are more expensive than distant ones but the difference is

compensated by the distance of the commute. Increases in income or decreases in

transport costs lead to greater suburbanization. Allowing for some heterogeneity in the

workforce, the model predicts that high-wage workers would prefer short commutes, but the

empirical evidence indicates that the rich tend to live in the suburbs.

One explanation is that the rich want new housing and less dense neighborhoods. The

durability of existing housing means the land for new housing is only available on the urban

fringe despite the long commute. In the long run equilibrium, suburban houses are cheaper

because they are farther from work (further out on the bid-rent curve) and workers don‘t like

to commute. But in the short run, changes in the stock may run counter to the long run

equilibrium. Income sorting and exclusive zoning are other obstacles to achieving the

archetypal equilibrium.

A related feature of the monocentric model is that commuting distances are minimized,

although the reality often is that workers commute much longer distances than predicted.

Part of the ―excess commuting‖ can be explained by the measurement of commuting cost in

90 Workforce Housing Final Report

time rather than distance. Another aspect of reality is that labor and housing markets are

not perfect and job matching appears much closer to random than minimizing commuting

distances. Dual-earner households further complicate the residential location process of

heterogeneous workers. Urban consumption amenities provide another justification for

urban dwellers even when their jobs are outside of the center city.

Given these apparent empirical exceptions to the monocentric model, Rouwendal and

Nijkamp point to the value of time, which Brownstone and Small (2004) estimate to be about

half of the wage rate. The range of values is wide and valuations of hypothetical situations

are much lower than the prices drivers do pay in actual road pricing situations. Some of the

heterogeneity can be explained by differences in gender and household responsibilities.

Women prefer shorter commutes and usually have primary responsibility for taking care of

children outside of school. The authors remained convinced that space plays an important

role in the job matching process, but commuting costs or time do not seem sufficient to

explain the wide range of commuting patterns found in the empirical studies.

Commuting Impact on Pay

The prime motivation for workers to commute, especially long distances, is to increase their

pay. Three studies are sufficient to make the point that workers can increase their pay by

commuting to markets where employers offer higher wages. Imerman, Orazem, Sikdar, and

Russell (2006) study the pay and turnover of Iowa nurses based on license information and

a supplemental survey. The authors found that the exit rate from nursing is extremely small

– 80 percent of nurses reach 30 years without letting their license lapse. Those that do

leave nursing usually leave the workforce as well. On average, nurses commute 21 minutes

to work, but rural nurses are most likely to commute more than 40 minutes. Commuting has

a significant impact raising pay by 5 percent for an additional 20 minutes of commuting.

Rural hospitals and clinics have to compete with city hospitals, which offer higher wages and

benefits.

Police officers in eastern Massachusetts can transfer between local police forces and keep

their civil service benefits accruing. This information comes from a personal communication

with a newly elected selectman from a suburban town. The towns with high-cost housing

often have a difficult time keeping new recruits because the recruits establish some

experience and then transfer to a town closer to where they can afford to buy a home. High-

Workforce Housing Final Report 91

cost towns tend to subsidize the recruiting and training for lower-cost towns. The City of

Boston has a residency requirement that police officers have to live in the city to work on the

city police force. Given the high cost of housing in Boston, the city must either pay higher

wages or face frequent turnover and training costs to offset the drain of officers to other

towns. The issue of workforce housing must be keenly felt by local police chiefs and the

licensing information may provide a good approach for measuring the turnover of officers.

Another study from the Baltic countries shows how commuting has helped reduce the local

disparities in wages. Hazans (2004) demonstrates that rural workers are drawn to work in

the city for higher wages. The rural to urban commuting pattern reduces not only the wage

differential, but also the unemployment rate differential, and increases the overall output of

the economy.

One way to facilitate higher wages for low-income workers may be to subsidize their

transportation. Margy Waller (2005) counters concerns that car ownership is expensive with

evidence that workers with a car are better able to obtain a job with good wages. Public

transit saves money on ownership and operating cost, but the opportunity cost is that

workers are more limited in their search to jobs along transit lines. Waller cites a Vermont

study on subsidized car ownership by Lucas and Nicholson (2003). They found that an

individual‘s income increased by $124 to $127 per month after obtaining a car and was 19

percent more likely to have earned income. Waller concludes (p. 1), ―when all costs are

considered along with benefits of private vehicles, it makes sense to press for more

assistance and policies that reduce car ownership costs for poor workers.‖

Jobs-Housing Balance

One apparent way to solve the problem of spatial mismatch and traffic congestion is to

create more housing near the employment centers. According to Nowlan and Stewart

(1991), Toronto managed to avoid serious traffic problems by following their office building

boom in the 1970s and 1980s by accelerated downtown housing construction. However,

this ―solution‖ must overcome very substantial challenges. Businesses gain agglomeration

advantages by clustering together, which motivates them to outbid residential uses for the

land. Local governments enjoy the tax advantages of businesses who generally pay much

more to the government than they demand in public services. And once the infrastructure

has been established for commercial purposes, it is expensive to retrofit for mixed use,

92 Workforce Housing Final Report

especially land for schools and recreational facilities. Other obstacles include frequent job

turnover, two-earner households, and exclusionary zoning policies. Cervero (1996)

examined 23 San Francisco Bay area cities, and found that the jobs-housing balance

worsened in 8 of the 10 most job-rich cities from 1980 to 1990. During that time, commutes

increased by 30 percent even while employers and retailers were following the labor to the

suburbs. In fact, 14 of the 28 Bay Area cities were more balanced in 1990 than 1980 in

terms of jobs to employed residents. The exception was the wealthy communities, which

blocked commercial development, and job-surplus business centers, mostly in Silicon

Valley.

One city, Pleasanton, switched from a bedroom community to a jobs-surplus city during the

decade after the development of a massive office park. Ironically, Pleasanton is close to

balanced between jobs and housing, but far from self-contained because most workers live

elsewhere and most residents work elsewhere. As Cervero explains (p. 7):

―According to the 1990 journey-to-work statistics, 35 percent of Pleasanton‘s employed

residents worked in San Francisco, the Silicon Valley, or the dense Oakland-Fremont

corridor paralleling Interstate 880. Thus, as new jobs were created, most new workers

found that Pleasanton‘s housing was already occupied by traditional suburban

households whose workers commuted to downtown jobs. The housing units added were

too few to accommodate many new workers. While Pleasanton‘s workforce grew from

7,161 in 1980 to 33,325 in 1990, or 365 percent, housing increased from 11,665 to

19,356 units, or only 66 percent, over that decade.‖

The balance would have been better, and perhaps more self-contained, if the developers of

the Hacienda Business Park had been allowed to build over 2,000 apartments adjoining the

office park, but a citizen backlash against growth blocked those moderate-cost residences.

Fewer, high-end houses were built instead, which were too expensive for most of the clerical

workers in the business park.

Cervero (1996) estimated a single-destination gravity model in which commuting time from

neighboring cities to Pleasanton is a function of the number of housing units in the source

city, the median single-family house price in the source city, and the straightline distance

from the source city to Pleasanton. Using data from 1990 Census STF 3A, the simple log-

linear model had a goodness-of-fit (R 2 ) of 0.836. The long commutes can be largely

Workforce Housing Final Report 93

explained by the availability of less expensive housing in other cities. Cervero concluded

that NIMBY opposition and planning failures are at least as much to blame as market

failures in providing workforce housing. Congestion tolls and parking fees can be justified

as an offset to these planning failures and may force commuters to internalize the

externalities generated by their commuting.

A more recent paper by Cervero (2006) continues the investigation by seeing whether the

jobs-housing balance or the retail-housing mix is more important in reducing the vehicle

miles traveled (VMT). Using 2000 travel-diary data from the San Francisco Bay area, the

trips are divided between work trips and shopping trips, and the trips are measured in both

distance and time. Cervero found that a 10 percent increase in the number of jobs within

four miles of the worker‘s residence was associated with a 3.29 percent decrease in the

vehicle miles traveled. Moreover, every 10 percent increase in the number of retail and

service jobs within four miles of the worker‘s residence was associated with only a 1.68

percent reduction in shopping and personal-service trips. The results based on time were

similar. Cervero concluded that the jobs-housing balance had a bigger role in reducing auto

traffic than having retail stores close to housing.

Other studies have considered the impact of better jobs-housing balance. The Institute of

Transportation Engineers (2001) Trip Generation Handbook estimated that there was a 2

percent reduction in the number of trips to work when housing is mixed with office parks, but

38 percent fewer trips when retail and housing mixed together. Ewing (1998) studied 500

communities in Florida, and found the share of people commuting within their home

community was significantly higher when the number of jobs and number of housing units

were approximately in balance. Krisek (2003) used longitudinal panel data from the Puget

Sound area and found that shortened commutes meant fewer miles traveled, but more

frequent trips. It was easier to go home and come back to work during the day when the

employee lives close to work.

Levine (1998) found that low-moderate income, single worker households benefit the most

from policies that promote job-housing balance, because those households are more willing

to move to workforce housing projects near work. Dual earner households have more

obstacles to moving. Finally, Schwanen, Dieleman, and Dijst (2004) examined Dutch data,

and they found that both commuting times and distances were longer in polycentric cities,

but the metropolitan structure can only explain a small part of the commuting behavior.

94 Workforce Housing Final Report

Given the variety of results and the prevalence of criss-cross commuting in a polycentric

city, it seems safe to conclude that the jobs-housing balance is far from sufficient to reduce

commuting miles and congestion, let alone self-containment.

Congestion Pricing and Measurement

Even if commute times have remained stable in some cities over extended periods, the level

of congestion delays continues to worsen. Shrank and Lomax (2005) report that the annual

average delay in 85 urban areas has increased from 16 hours in 1982 to 47 hours in 2003.

According to the 2005 Urban Mobility Report by the Texas Transportation Institute (p. 1),

―congestion caused 3.7 billion hours of travel delay and 2.3 billion gallons of wasted fuel ...

to a total cost of $63 billion.‖ The worst delays are in very large metropolitan areas. In

terms of annual delay per traveler, the top three metro areas are Los Angeles, San

Francisco, and Washington, D.C.

Tony Downs (1992, 2004) boils down the traffic congestion to four causes. First, most

people work during the day because it is more productive for businesses to interact that

way. But the infrastructure does not and cannot have the capacity for everyone to be

moving at the same time. So, some people have to wait, and in an efficient system the less

productive should wait for the more productive. Second, rising incomes allow more people

to own cars and move to low density suburban neighborhoods, which increases the traffic

load on roads to work. Similarly, rising population combined with rising income accelerates

the demand for road travel. Finally, the high volume of traffic spawned by prosperity and

growth inevitably leads to traffic accidents. As we have all witnessed, even a small accident

or breakdown can generate a big traffic jam when the volume of road use is high. The main

point from Downs is that traffic is the result of economic success, not policy failure.

A number of researchers have tried to model the relationship between congestion and land

use, particularly density and compactness. Malpezzi (1999) found that population density

was negatively related to commute times, while concentration (ratio of center city population

to MSA population) was positively related to commute times. Malpezzi used 1990 Census

data for US MSAs and instrumented for transit supply with a predicted value of a separate

transit supply equation. Ewing et al. (2003) modeled 83 MSAs in cross-section equations

for both 1990 and 2000 Census data. For 2000, the authors found that the land use mix

(jobs-housing balance) was negatively related and the street accessibility mix (block length)

Workforce Housing Final Report 95

was positively related to commute time. Gordon et al. (2004) estimated the effect of land

use measures on commute times in 77 large metropolitan areas in 1990 and 2000. Their

measures included demographics, and transport supply and demand, such as the proportion

of commuters using transit which could be reflecting congestion delays as much as causing

them. They found that suburbanization and population density reduced commute time in

2000.

The distribution of residences seems bound to affect congestion, but the actual relationship

depends on the congestion measure. Sarzynski, Wolman, Galster, and Hanson (2006)

have conducted a careful study of the influence of land use patterns in 1990 on the

measures of traffic congestion in 2000 for a sample of 50 large U.S. urban areas. Multiple

regression is used to control for prior levels of congestion, changes in an area‘s

transportation network, and demographic features (particularly population growth).

Correlation and principal components factor analyses are used to combine 14 land use

indices into seven factors. The factors are scaled such that higher values indicate less

sprawl. Density measures both housing units and jobs per square mile. Continuity

measures the degree to which land has been developed in an unbroken fashion throughout

the metropolitan area. The combination density/continuity index was found to be positively

related to subsequent roadway ADT/lane (average number of vehicles per freeway lane)

and delay per capita. A second finding was that housing centrality (the degree to which land

use is located nearer the core of the urban area) was positively related to subsequent delay

per capita. A third finding was that the housing-job proximity is inversely related to

subsequent commute time. Of these measures, the first two suggest compact cities are

more likely to have traffic jams, while the third shows that housing close to employment

centers could reduce a worker‘s commute time.

Rodriguez, Targa, and Aytur (2006) study the transport implications of urban containment

policies using panel data for 25 major U.S. metropolitan areas between 1982 and 1994.

The measures of local containment policies are provided by Pendall (1999 and 2000) using

a national survey of 1,510 cities and counties. Growth management efforts at the state level

are provided by: Burby and May, 1997; Burby, Nelson, Dennis, and Handmer (2001);

Carruthers, 2002; Nelson, 1999; and Zovanyi, 1998. As with the earlier research, the key

problem is how to handle the endogenous relationships between land use and travel

demand. The endogenous variables are density (population per land area of the

metropolitan area) and VMT (vehicle miles traveled per capita). Density is estimated on

96 Workforce Housing Final Report

personal income, land supply, farmland value, local urban containment and state growth

management. The predicted value of density is incorporated into the VMT equation along

with personal income, transportation infrastructure supply, and fuel cost. An IV-2SLS model

with fixed effects is estimated on the pair of equations, density and VMT. The model results

show that urban containment policies increase the density of cities and increase the miles

traveled, which exacerbates congestion unless there are offsetting policies such as a fuel

tax or expanded public transit.

Three recent papers represent the literature on congestion pricing. Winston and Langer

(2006) describe the effect of government spending on commuting costs and present the

case for road pricing as the most efficient way to allocate those costs according to users‘

willingness-to-pay. Walls, Safirova, and Jiang (2006) consider the implications of

congestion pricing on telecommuting. Based on a 2002 survey of about 5,000 Southern

California residents, they found that a worker‘s propensity to telecommute increased with

age and educational attainment as well as particulars of the job and firm, which are at least

as important as the demographics. This research fits into the picture that more productive

employees are less willing to commute, but it begs the question of how telecommuting

propensity will affect the distribution of housing relative to employment. Certainly

congestion pricing would increase the cost of commuting and increase the relative

advantages of telecommuting.

The final paper addresses congestion pricing from a long run perspective of gradual

changes in land use. Safirova, Houde, Lipman, Harrington, and Baglino (2006) have

created a computable general equilibrium (CGE) model with a spatial component. The

emphasis is on the impact of road pricing rather than land use regulation, but it looks like

adapting the model for investigating regulation would be feasible. The CGE model is well

suited for considering the social welfare aspects of land use regulation because it brings

together the housing, labor, and transportation markets with the machinery for summing up

the long-run welfare gains and losses. A key assumption is that markets move toward an

equilibrium. The CGE model results for a cordon toll were modest long-term welfare gains

on the order of 0.05 percent of annual income, but it was several times larger than a short-

run transportation model. The difference is due to the prediction by the CGE model that the

unemployment rate would increase and employers would have to increase wages to get the

needed labor supply. A second finding was that higher skilled workers benefited from the

Workforce Housing Final Report 97

wage increases, while the less skilled workers benefited from the toll redistribution. The

magnitude of the welfare gains is sensitive to the redistribution mechanism.

Parking costs are closely connected to congestion pricing schemes as shown in the book by

Arnott, Rave, and Schob (2005). They make the point that underpricing of street parking

increases traffic congestion as does the cruising for parking spaces. By the authors‘

estimates, during peak commuting hours in large cities, 30 percent of the traffic comes from

cruising for parking with an average cruise time of 7.8 minutes. A model of labor productivity

considers the impact of staggered work hours. There is a trade-off between travel-time

elasticity with respect to the flow-capacity ratio and the labor productivity with respect to the

proportion of the workforce at work at the same time. Unfortunately, there is no empirical

consensus on those elasticities, but the model does frame a fundamental issue. Workers

can reduce commute times with staggered hours, but the loss in productivity through

reduced agglomeration effects may hurt the economy more than it helps the workers.

Research Ideas and Extensions

The work of Haas et al. (2006) and Cervero et al. (2006) have added to the Census data a

great deal of housing market, labor market, and commuting information. However, the

patterns of location and the complexity of joint decision-making defy simplified

representation. Certainly the monocentric model with steadily downward sloping bid-rent

curves does not adequately represent real cities. Here are several ideas extending their

analysis that would advance our knowledge at the local level.

1) The elasticities of house prices with respect to commuting time reported by

Cervero et al. (2006) could be extended to control for age, education, occupation,

or industry. It seems reasonable to expect differences by type of worker with

young workers being more mobile and willing to relocate. Some occupations,

such as marketing, which already entail a fair amount of travel, may be less

sensitive to house location, but more sensitive to airport and freeway access.

Similarly, industries may vary in their pattern of elasticities. The goal is to

understand how important commuting is in the decision by workers of where to

live relative to where they work. The firms and industries for which their workers

are most sensitive to commuting time should be the most eager to ensure

98 Workforce Housing Final Report

sufficient workforce housing is available within a reasonable commuting time

from their business.

2) Does the local land use regulation affect either the lengths of commutes,

variation in delays, or elasticities of house prices with respect to commuting

time? Zoning that makes housing supply less elastic may also be associated

with workers who are less willing or able to adjust their commuting patterns to

meet the fluctuating demands of businesses.

3) Commuting time and cost can influence the net wages that workers are willing to

accept as well as the housing prices those workers are willing to pay. Therefore,

a set of wage equations controlling for commuting times and costs could

generate a set of elasticities that are as informative about workers‘ choices as

the elasticities for house prices with respect to commuting times.

4) Labor catchment area or laborshed – Haas et al. looked at the average

commuting patterns by location of residence. This approach could be extended

by seeing whether restrictive zoning neighborhoods have systematically different

laborsheds. The alternative view is to look at the commuting patterns by location

of the firm. Where are its workers coming from? What are the turnover rates by

commuting distances or times? What is the availability of affordable workforce

housing within that catchment area? It may be that firms with more local workers

have lower turnover and higher growth potential. Knowing the typical size of a

laborshed for an industry, an economic development board could tout the

number of workforce housing units available within the laborshed. A related

project would be to compare growth rates of firms by the radius or average

commuting time of the laborshed. Do firms that can meet their labor needs within

a small laborshed have greater potential for future growth?

5) The fact that young and relatively unskilled workers seem to limit their job search

to areas relatively close to their home seems to run counter to the idea that low-

paid workers should be willing to commute longer distances because their value

of time is lower. It may be that search areas are related to word-of-mouth

networks that increase with the experience of the worker. The commuting

patterns of younger workers may reflect the small search area, with wider

commuting areas by age, education, experience, and frequency of search.

Therefore the research approach would be to measure the commuting patterns

for workers along those dimensions as a way to better understand a worker‘s

search areas. This information could be used by firms to improve their recruiting

Workforce Housing Final Report 99

through targeted outreach to the areas where suitable workers live. Alternatively,

firms may choose to advocate for more workforce housing close to their business

so there are more workers to meet their needs.

6) The discrete version of Cervero et al. (2006) uses nested multinomial logit

estimation in which the choice of commute mode is conditional on the choice of

neighborhood (PUMA type). The models were estimated on the set of recent

movers. A straightforward extension would be to include measures of land use

restrictiveness by neighborhood to see if that influenced the choice of

neighborhood. An alternative modeling approach would be to assume that there

is a 3-way joint estimation of house location, work location, and commute mode.

Professional workers may secure the job first then move to obtain a nice

neighborhood and reasonable commute, which would justify a nested logit

approach. Less skilled workers may move to a location with moderate cost

housing and then begin looking for a job within commuting distance of his/her

house. A two or three-way multinomial logit may be more appropriate for this

group of workers.

7) Switching regression for mode choice – Another modeling alternative well suited

for the choice of commute mode would be switching regression. A worker

attempts to minimize the commuting cost by choosing between driving his/her car

or taking public transit. The car takes less time, but the worker may run into a

traffic jam and have to pay for parking. The train requires a fare and more rigid

schedule, but the rider can work on the train and avoid the traffic congestion.

The worker considers both modes and chooses the one with less cost on

average. A switching regression could use information on the revealed

preference of each worker to determine the optimization rule for switching from

one mode to the other.

100 Workforce Housing Final Report

Chapter Seven: Recommended Research Directions

Throughout the chapters there are many possible research projects. In this final chapter, we

present four research directions that deserve consideration. The order does not indicate

priority. Factors such as cost and plausible data sources will need to be evaluated when

setting priorities for these or other projects that will form the research agenda.

6) Update and develop better measures for land use regulations

a. go beyond composite index at metro level

b. include measures for strictness in implementation

c. cope with multidimensionality (time, cost, uncertainty) and collinearity

d. allow for differentiation at sub-metro level

e. identify data needed and most likely sources

The existing measures of land use regulation are mostly based on the 1989 Wharton study.

This study has recently been updated, but it still has the limitations of being based on a

modest number of big city planners. It is difficult to discern from those responses the

degree of regulation enforcement or the extent of fluctuation over the housing cycle.

Typically a regression analysis uses a composite index at the metropolitan level because

the composite index avoids the difficult issues of multidimensionality (time, cost, and

uncertainty) as well as collinearity among the selected variables. The metro index also

glosses over differences in land use regulation that occur by town or neighborhood within

the laborshed. To acquire the necessary data, it may be necessary to collect the results of

many local efforts or initiate a national survey to collect a consistent set of specific elements

on a regular basis. The power of econometric analysis is clearly hampered without a panel

of regulation measures. Once constructed, however, indicators of labor supply, housing

market, and business activity can be related to the regulatory measures, and hypothesis

testing can be conducted in a statistical framework.

7) Quantify how diversity of regulations within metro area affects performance of

entire metro area

a. gain finer resolution by industry, sub-metro geography, type of development

structures

Workforce Housing Final Report 101

b. is it the quantity or distribution of workforce housing that matters with respect

to

i. house prices?

ii. labor outcomes?

iii. business growth, stability, or composition?

The impact of land use regulations in one section of a city depends greatly on the

alternatives offered by nearby towns. If a wealthy neighborhood has exclusive regulations

but surrounding neighborhoods accommodate with more permissive land use, the net

impact on business and labor markets may be insignificant. On the other hand, if competing

municipalities engage in a regulatory ―arms race,‖ the housing prices could become extreme

as builders have little land where the construction of moderate cost housing is permitted.

Another strategic pattern could be that the regulations are stringent, but the enforcement

varies much more in some municipalities than others. Also, regulations may be put in place

to slow development during a period of rapid growth, but eased when the demand returns to

a more normal pace.

Industries vary in their reliance on workers with moderate skills, and so the impact of

regulations is likely to vary by industry. Also, the agglomeration effects may develop with

advances in technology such that co-location becomes relatively more important to certain

industries. Municipalities willing to accommodate the growth of an industry could witness

very rapid expansion that could not have been predicted five or ten years before. Such

growth spurts may be more obvious when the analysis is done by municipality and industry

rather than viewed metropolitan-wide.

8) Determine how housing costs and congestion affect industries and firms,

including

a. rate of overall growth

b. changes in compositions by industry

c. differentiation by occupation, level of education, and income

d. identification of which industries are the winners and losers as a result of

rising house prices

e. identification of which industries are winners and losers as a result of

congestion (commuting cost, delay, or uncertainty)

102 Workforce Housing Final Report

Industries that rely heavily on mid-skill workers and have to locate downtown or in dense

employment centers may be most vulnerable to high housing costs and congestion effects.

Firms locating downtown may have the benefit of better public transportation service, but

much more expensive housing and parking. The goal of this research effort would be to

quantify the degree to which industries vary in their sensitivity to housing costs and

congestion effects. The most vulnerable industries are expected to suffer from labor

shortages and limited growth. Over time those industries stagnate or shrink leading to

changes in the industrial composition. Industrial and transportation companies move out of

center cities being replaced by finance, insurance and high-end retail companies. Higher

wages and public transportation subsidies are used to offset the rising house prices and

traffic congestion. Documenting these patterns will clarify the impact that a loss of workforce

housing has on the composition and location of industries in cities.

The Bureau of Labor Statistics vacancy survey (JOLTS) could be an excellent source of

information to determine which industries are most affected by labor supply problems. To

make full use of that information, researchers would need access to the location of firms in

the survey. In the past, such access has been very difficult due to confidentiality concerns.

Properly motivated, BLS and the Department of Labor may be willing to partner in this

research.

9) Fill in qualitative detail on how rising house prices affect employers through

a. recruitment - longer search, wider search, subsidize transportation/housing,

training

b. productivity - lateness, absenteeism, telecommuting, job shopping on the job

c. retention / turnover – shorter tenure, higher wage and promotion

To better appreciate the connection between house prices and labor supply, we need a

deeper understanding of how employers respond. The most direct response to lingering

vacancies would be to increase the wages offered, but employers have many alternatives.

They could intensify their recruiting with more advertising, a broader geographical search,

lower standards, higher signing bonus, subsidized relocation, etc. It would be helpful to

identify the strategies used and in what circumstances they are more effective. Another

dimension is to determine how labor productivity suffers when workers commute long

distances or telecommute. Does flextime make the scheduling of meetings more difficult?

Do workers spend more of their work-time dealing with child care issues or looking for a

Workforce Housing Final Report 103

replacement job? Are workers less willing to work overtime or do training sessions in the

evening? Is the retention of distant workers significantly different? Do workers in high

housing cost markets demand higher wages and faster promotion? A challenge throughout

this research is to identify housing costs as the proximate cause of the labor supply

difficulties as compared with big city competition.

104 Workforce Housing Final Report

Bibliography

Abadie, Alberto and Sofia Dermisi (2006) ―Is Terrorism Eroding Agglomeration Economies in

Central Business Districts? Lessons from the Office Real Estate Market in Downtown

Chicago,‖ NBER Working Paper No. 12678, November.

Abowd, John M. and Francis Kramarz (1999) ―The Analysis of Labor Markets using Matched

Employer-Employee Data,‖ Ch. 40 in Orley C. Ashenfelter and David Card (eds.),

Handbook of Labor Economics, Vol. 3b, Amsterdam: Elsevier, 2629-2710.

Abowd, John M., John Haltiwanger, Julia Lane, Kevin L. McKinney, Kristin Sandusky (2005)

―Technology and Skill: An Analysis of Within and Between Firm Differences,‖ Mimeo,

May 25.

Abowd, John M., Paul A. Lengermann, and Kevin L. McKinney (2003) ―The Measurement of

Human Capital in the U.S. Economy,‖ Census Bureau LEHD Program Technical paper

TP-2002-09.

Abraham, Katherine G. (1987) ―Help-Wanted Advertising, Job Vacancies, and

Unemployment‖ Brookings Papers on Economic Activity 1:207-248

Acemoglu, D. (1996) ―A Microfoundation for Social Increasing Returns in Human Capital

Accumulation,‖ Quarterly Journal of Economics, 779-804.

Acemoglu, Daron and Joshua Angrist (2000) ―How Large are Human Capital Externalities?

Evidence from Compulsory Schooling Laws,‖ NBER Macroeconomic Annual 15.

Acs, Zoltan, Catherine Armington, and Alicia Robb (1999) Small Business Administration,

January.

Acs, Zoltan, and Catherine Armington (2005) ―Using Census BITS To Explore

Entrepreneurship, Geography, and Economic Growth,‖ Small Business Administration

Office of Advocacy.

Workforce Housing Final Report 105

Akerlof, George A. and Janet Yellen (1990) Efficiency Wage Models of the Labor Market,

Cambridge: Cambridge University Press.

Alonso, William (1964) Location and Land Use, Cambridge: Harvard University Press.

Anas, Alex (1982) Residential Location Markets and Urban Transportation, New York:

Academic Press.

Anas, Alex (1999) ―Congestion, Land Use, and Job Dispersion: A General Equilibrium

Model,‖ Journal of Urban Economics, 45: 451-473.

Anderson, Patricia M. and Bruce D. Meyer (1994) ―The Extent and Consequences of Job

Turnover,‖ Brookings Papers on Economic Activity, pp. 177-248.

Andersson, Fredrik, Harry J. Holzer, and Julia I. Lane (2005) Moving Up or Moving On: Who

Advances in the Low-Wage Labor Market, Russell Sage Foundation Publications.

Andersson, Fredrick, Simon Burgess, and Julia I. Lane (2004) ―Cities, Matching and the

Productivity Gains of Agglomeration,‖ CEP Discussion Paper No. 648, September.

Arnott, Richard, Tillman Rave, and Ronnie Schob (2005) Alleviating Urban Traffic

Congestion, The CESifo Book Series, Cambridge and London: MIT Press.

Audretsch, D.B. and M. P. Feldman (1996) ―R&D Spillovers and the Geography of

Innovation and Production,‖ American Economic Review, 86(3): 630-640.

Bain and Company (2004) ―California Competitiveness Project: Assessment of California

Competitiveness,‖ California Business Roundtable.

Baldassare, Mark (2007) ―Californians and their Government,‖ Public Policy Institute of

California.

Ballou, Brian and Dan L. Heitger (2006) ―Tapping a Risky Labor Pool,‖ Harvard Business

Review, 84 (12): 28-30.

106 Workforce Housing Final Report

Bar-El, Raphael (2006) ―Inter-regional Labor Market Equilibrium: Another Pattern of Spatial

Mismatch,‖ Annals of Regional Science, 40(2): 393-405.

Barker, Kate (2004) Review of Housing Supply, Final Report (Norwick, HM Stationery

Office).

Barro, Robert J. and Xavier Sala-i-Martin (1991) ―Convergence Across States and Regions,‖

Brookings Papers on Economic Activity, no. 1, 107-158.

Barro, Robert J. and Xavier Sala-i-Martin (1995) Economic Growth, New York: McGraw-Hill.

Bartik, T. J. (1991) Who Benefits from State and Local Economic Development Policies?

Kalamazoo: W.E. Upjohn Institute for Employment Research.

Bartik, T. J. (1993) "Who Benefits from Local Job Growth: Migrants or the Original

Residents," Regional Studies, 27: 297-311.

Bartley, Katherine F. (2006) ―Technology and the Convergence of the U.S. Urban Migration

Patterns: 1970-2000,‖ Growth and Change, 37(1): 82-106.

Brauer, David (2007) ―The Natural Rate of Unemployment‖ Congressional Budget Office

Working Paper 2007-06

Bayer, Patrick and Stephen L. Ross (2006) ―Identifying Individual and Group Effects in the

Presence of Sorting: A Neighborhood Effects Application,‖ University of Connecticut

Working Paper 2006-13, April.

Been, Vicki L. (2004) ―Impact Fees and Housing Affordability‖ (paper presented for HUD

Conference on Regulatory Barriers to Housing, April 22, 2004).

Benedetto, G., Haltiwanger, J., Lane, J. and McKinney, K. (2007) "Using Worker Flows in the Analysis of Firms." Journal of Business and Economic Statistics, forthcoming.

Berger, Mark C., and Glenn C. Blomquist (1992) ―Mobility and Destination in Migration

Decisions: The Role of Earnings, Quality of Life, and Housing Prices‖ Journal of Housing

Economics 2: 37-59

Workforce Housing Final Report 107

Berliant, Marcus and Fan-chin Kung (2006) ―Can Information Asymmetry Cause

Agglomeration?‖ Munich Personal RePec Archive, October 31.

Bernard, Andrew B. and Charles I. Jones (1996) ―Productivity and Convergence across U.S.

States and Industries,‖ Empirical Economics, 21(3): 113-135.

Bernat, G. Andrew, Jr. (2001) ―Convergnece in State Per Capita Personal Income, 1950-

99,‖ Survey of Current Business, June.

Berry, Christopher R. and Edward L. Glaeser (2005) ―The Divergence of Human Capital

Levels across Cities,‖ Harvard Institute of Economic Research Discussion Paper 2091.

Berube, Alan and Elizabeth Kneebone (2006) "Two Steps Back: City and Suburban Poverty

Trends, 1999-2005," The Brookings Institution, Census Series, December.

Blanchard, Oliver J., and Peter Diamond (1990) ―The Cyclical Behavior of the Gross Flows

of U.S. Workers‖ Brookings Papers on Economic Activity 2: 85-143

Blanchard, Olivier and Lawrence Katz (1992) ―Regional Evolutions,‖ Brookings Papers on

Economic Activity, 1992(1): 1-75.

Bleakley, Hoyt and Jeffrey C. Fuhrer (1997) ―Shifts in the Beveridge Curve, Job Matching,

and Labor Market Dynamics,‖ New England Economic Review, Sept./Oct.

Blomquist, Glenn C., Marck C. Berger, and John P. Hoehn (1988) ―New Estimates of Quality

of Life in Urban Areas,‖ American Economic Review, 78(1): 89-107.

Bluestone, Barry (2006) ―Sustaining the Mass Economy: Housing Costs, Population

Dynamics, and Employment,‖ Boston Federal Reserve Bank.

Bound, John, Jeffrey Groen, G. Gabor Kezdi, and Sarah Turner (2004) ―Trade in University

Training, Cross-State Variation in the Production and Use of College-Educated Labor,‖

Journal of Econometrics, 121(1-2): 143-173.

108 Workforce Housing Final Report

Bover, Olympia, Muellbauer, John and Anthony Murphy (1989) ―Housing, Wages and the

UK Labour Market,‖ Oxford Bulletin of Economics and Statistics, 51:97-136.

Brasington, David and Donald R. Haurin (2006) ―The Demand for Educational Quality:

Combining a Median Voter and Hedonic House Price Model,‖ Louisiana State University,

Department of Economics Working Paper, 2006-07.

Breschi, Stefano and Francesco Lissoni (2006) ―Mobility of Inventors and the Geography of

Knowledge Spillovers. New Evidence on US Data.‖ CESPRI Working Paper 184,

October.

Briggs, Xavier de Souza, ed. (2005) The Geography of Opportunity: Race and Housing

Choice in Metropolitan America, Brookings Institution Press.

Brown, Nestor Duch (2006) ―The empirics of spatial competition: Evidence from European

regions,‖ University of Barcelona, Economics Working Paper No. 150.

Brownstone, David and Kenneth A. Small (2004) ―Valuing Travel Time and Reliability:

Assessing the Evidence from Road Pricing Demonstrations,‖ University of California

Irvine Working Paper.

Brueckner, Jan K., Jacques-Francois Thisse, and Yves Zenou (2002) Local Labor Markets,

Job Matching, and Urban Location,‖ International Economic Review, 43(1): 155-171.

Brulhart, M. (2001) ―Evolving Geographical Specialisation of European Manufacturing

Industries,‖ Weltwertschaftliches Archiv, 137(2): 215-243.

Brunello, Giorgio and Maria De Paola (2006) ―Training and Economic Density: Some

Evidence from Italian Provinces,‖ Marco Fanno Working Paper No. 30, November.

Burby, R.J. and P.J. May (1997) Making Governments Plan: State Experiments in Managing

Land Use, Baltimore, MD: Johns Hopkins University Press.

Workforce Housing Final Report 109

Burby, R.J., A.C. Nelson, P. Dennis, and J. Handmer (2001) ―Urban Containment Policy and

Exposure to Natural Hazards: Is There A Connection?‖ Journal of Environmental

Planning and Management, 44: 475-490.

Butani, Shail, George Werking, and Vinod Kapani (2005) ―Employment dynamics of

individual companies versus multicorporations,‖ Monthly Labor Review, December: 3-13.

Cameron, Gavin and John Muellbauer (1998) ―The Housing Market and Regional

Commuting and Migration Choices,‖ Scottish Journal of Political Economy, 45(4), 420–

446.

Cameron, Gavin, John Muellbauer, and Anthony Murphy (2005) ―Migration Within England

and Wales and the Housing Market‖ Economic Outlook 29(3): 9-19

Cameron, Gavin, John Muellbauer, and Anthony Murphy (2006) "Housing Market Dynamics

and Regional Migration in Britain" London: Center for Economic Policy Research

Discussion Paper No. 5832

Cannari, Luigi, Francesco Nucci, and Paolo Sestito (2000) "Geographic Labour Mobility and

the Cost of Housing: Evidence from Italy" Applied Economics, 32: 1899-1906

Card, D. (1999) ―Education in the Labor Market,‖ in O. Ashenfelter and D. Card (eds). The

Handbook of Labor Economics, Amsterdam: North-Holland.

Carlino, Gerald, Satyajit Chatterjee, and Robert Hunt (2005) ―Matching and Learning in

Cities: Urban Density and the Rate of Invention,‖ Federal Reserve Bank of Philadelphia,

April.

Carruthers, J.I. (2002) ―The Impacts of State Growth Management Programmes: A

Comparative Analysis,‖ Urban Studies, 39: 1959-1982.

Carruthers, John I. and Alexander C. Vias (2005) ―Urban, Suburban, and Exurban Sprawl in

the Rocky Mountain West: Evidence from Regional Adjustment Models,‖ Journal of

Regional Science, 45(1): 21-48.

110 Workforce Housing Final Report

Cascio, Wayne F. (2006) ―The High Cost of Low Wages,‖ Harvard Business Review, 84

(12): 23.

Case, K. E. (1991) ―The Real Estate Cycle and the Economy: Consequences of the

Massachusetts Boom of 1984-97,‖ New England Economic Review (Sept/Oct.): 37-46.

Cavailhes, Carl Gaigne, and Jacques-Francois Thisse (2006) ―Trade and the Structure of

Cities,‖ Kyoto Institute of Economic Research, Discussion Paper No. 623, September.

Caves, Richard E. (1998) ―Industrial Organization and New Findings on the Turnover and

Mobility of Firms,‖ Journal of Economic Literature, 36(4): 1947-1982.

Cervero, Robert (1996) ―Jobs-Housing Balance Revisited: Trends and Impacts in the San

Francisco Bay Area,‖ Journal of the American Planning Association, 62(4): 492-511.

Cervero, Robert (2001) ―Efficient Urbanisation: Economic Performance and the Shape of the

Metropolis,‖ Urban Studies, 38(10): 1651-1671.

Cervero, Robert (2006) ―Alternative Approaches to Modeling the Travel-Demand Impacts of

Smart Growth,‖ Journal of the American Planning Association, 72(3): 285-295.

Cervero, Robert, Karen Chapple, John Landis, and Martin Wachs (2006) ―Making Do: How

Working Families in Seven U.S. Metropolitan Areas Trade Off Housing Costs and

Commuting Times,‖ The Center for Housing Policy and the National Housing

Conference, June 15.

Cervero, Robert and Michael Duncan (2006) ―Which Reduces Vehicle Travel More: Jobs-

Housing Balance or Retail-Housing Mixing?‖ Journal of the American Planning

Association, 72(4): 475-490.

Cervero, R. and K.L. Wu (1998) ―Subcentering and Commuting: Evidence from the San

Francisco Bay Area,‖ Urban Studies, 35(7): 1059-1076.

Chatterjee, Satyajit (2003) ―Agglomeration Economies: The Spark that Ignites a City?‖

Business Review (Federal Reserve Bank of Philadelphia) Q4: 6-13.

Workforce Housing Final Report 111

Ciccone, Antonio and Giovanni Peri (2002) ―Identifying Human Capital Externalities: Theory

with Application to U.S. Cities,‖ IZA Discussion Paper No. 488.

Ciccone, Antonio and Robert E. Hall (1996) ―Productivity and the Density of Economic

Activity,‖ American Economic Review, 86(1): 54-70.

Cohen, Natalie (2000) ―Business Location Decision-Making and the Cities: Bringing

Companies Back,‖ The Brookings Institution, Center on Urban and Metropolitan Policy.

Colean, Miles L. (1940) Housing for Defense, The Twentieth Century Fund.

Costa, D. L. and M.E. Kahn (2001) ―Power Couples.‖ Quarterly Journal of Economics, 116:

1287-1315.

Clark, Kelly A. and Mary Anne Phillips (2002) ―A Comparison of Job Openings Surveys:

Concepts and Survey Design,‖ Bureau of Labor Statistics, Office of Survey Methods

Research.

Clayton, Richard L. and David M. Talan (2006) ―Business Employment Dynamics:

Measuring Gross Job Gains and Losses,‖ Bureau of Labor Statistics.

Crane, R. and D. Chatman (2003) ―As Jobs Sprawl, Wither the Commute?‖ Access, 23: 14-

19.

Cutrini, Eleonora (2006) ―The Balassa Index Meets the Dissimilarity Theil Index: a

Decomposition Methodology for Location Studies,‖ Universita Politecnica Delle Marche,

Working Paper No. 274, November.

Czohara, Laura, Julia Melkers, and Kouassi Dagawa (2004) ―Firm Location Decisions and

Information Needs,‖ Georgia State University, Fiscal Research Center Report No. 93.

Dalmazzo, Alberto and Guido de Blasio (2005) ―Where Do Human Capital Externalities End

Up?‖ European Regional Science Association Conference proceedings, August.

112 Workforce Housing Final Report

Davis, Steven J., John C. Haltiwanger and Scott Schuh (1996) Job Creation and

Destruction, MIT Press.

Davis, Steven J., John Haltiwanger, Ron S. Jarmin, C.J. Krizan, Javier Miranda, Alfred

Nucci, and Kristin Sandusky (2006a) ―Measuring the Dynamics of Young and Small

Businesses: Integrating the Employer and Nonemployer Universes,‖ Census Bureau

Center for Economic Studies, CES 06-04.

Davis, Steven J., John Haltiwanger, Ron Jarmin, and Javier Miranda (2006b) ―Volatility and

Dispersion in Business Growth Rates: Publicly Traded versus Privately Held Firms,‖

Census Bureau Center for Economic Studies, CES 06-17.

Davis, Steven J., R. Jason Faberman, and John Haltiwanger (2006) ―The Flow Approach to

Labor Markets: New Data Sources and Micro-Macro Links,‖ Journal of Economic

Perspectives, 20(3): 3-26.

Davis, Steven J., R. Jason Faberman, and John Haltiwanger (2007) ―The Establishment-

Level Behavior of Vacancies and Hiring‖, Presentation to Chicago Fed/NBER Workshop

on Macroeconomic Perspectives, March 2-3

Deakin, Elizabeth (1989) Growth Controls and Growth Management: A Summary and

Review of Empirical Research. In Understanding Growth Management: Critical Issues

and a Research Agenda, edited by David J. Brower, David R. Godschalk, and Douglas

R. Porter. Washington, DC: The Urban Land Institute.

DeSalvo, Joseph S. and Mobinul Huq (2005) ―Mode Choice, Commuting Cost, and Urban

Household Behavior,‖ Journal of Regional Science, 45(3): 493-517.

Diamond, C.A. and C.J. Simon (1990) ―Industrial Specialization and the Returns to Labor,‖

Journal of Labor Economics, 8(2): 175-201.

Dohmen, Thomas J. (2000) ―Housing, Mobility, and Unemployment,‖ IZA Discussion Paper

No. 210, November.

Workforce Housing Final Report 113

Downs, Anthony (1992) Stuck in Traffic: Coping with Peak-Hour Traffic Congestion,

Washington, D.C.: The Brookings Institution.

Downs, Anthony (2004) Still Stuck in Traffic, Washington, D.C.: The Brookings Institution.

Drennan, Matthew P. (2005) ―Possible Sources of Wage Divergence among Metropolitan

Areas of the United States,‖ Urban Studies, 42(9): 1609-1620.

Duffy, David, John FitzGerald, and Ide Kearney (2005) ―Rising House Prices in an Open

Labour Market‖, The Economic and Social Review, 36(3) 251-272

Dumais, Guy, Glenn Ellison, and Edward L. Glaeser (2002) ―Geographic Concentration as a

Dynamic Process,‖ Review of Economics and Statistics, 84(2): 193-204.

Dumond, M. J., Hirsch, B.T., and D.A. Macpherson (1999) ―Wage Differentials Across Labor

Markets and Workers: Does Cost of Living Matter?‖ Economic Inquiry, 37 (4): 577-98.

Dunne, Timothy, Mark J. Roberts, and Larry Samuelson (1989) ―Plant Turnover and Gross

Employment Flows in the U.S. Manufacturing Sector,‖ Journal of Labor Economics 7(4):

48-71.

Dunne, Timothy, Shawn D. Klimek, Mark J. Roberts (2005) ―Exit from Regional

Manufacturing Markets: The Role of Entrant Experience,‖ International Journal of

Industrial Organization, 23(5-6): 399-421.

Duranton, Giles and Diego Puga (2004) ―Micro-Foundations of Urban Agglomeration

Economies,‖ in J. V. Henderson and J-F. Thisse, eds., Handbook of Urban and Regional

Economics, Vol. 4, Amsterdam: North-Holland.

DTZ Consulting and Research (2006) ―Housing, Economic Development, and Productivity:

Literature Review,‖ Report for The Department of Trade and Industry (Reading,

England) January.

Eliasson, Kent, Urban Lindgren, and Olle Westerlund (2003) "Geographic Labour Mobility:

Migration or Commuting?" Regional Studies 37(8): 827-837

114 Workforce Housing Final Report

Ellison, Glenn and Edward L. Glaeser (1997) ―Geographic Concentration in U.S.

Manufacturing Industries: A Dartboard Approach,‖ Journal of Political Economy, 105(5):

889-927.

Ellison, Glenn and Edward L. Glaeser (1999) ―The Geographic Concentration of Industry:

Does Natural Advantage Explain Agglomeration,‖ American Economic Review, 89(2):

311-316.

Elvery, Joel A. (2006) ―City Size and Skill Intensity of Production,‖ Typescript, October.

Engelhardt, Gary V. (2002) ―Nominal Loss Aversion, Housing Equity Constraints, and

Household Mobility: Evidence from the U.S. Housing Market,‖ Syracuse University

Working Paper 2002-024.

Evans, P. and B. McCormick (1994) ―The New Pattern of Regional Unemployment: Causes

and Policy Significance,‖ Economic Journal, 104: 633-647.

Ewing, R. (1998) Transportation and Land Use Innovations, Chicago: APA Press.

Ewing, Reid, Rolf Pendall, and Don Chen (2003) ―Urban Sprawl and Transportation,‖ Paper

presented at the Annual Meeting of the Transportation Research Board.

Expansion Management (2007) ―America‘s 50 Hottest Cities: Perception Is a Critical Factor

in Attracting Expanding Companies,‖ by Ken Krizner, Managing Editor, February 15

(www.expansionmanagement.com/smo/newsviewer ).

Ezzet-Lofstron, Roxanne (2004) ―Valuation of Metropolitan Quality of Life in Wages and

Rents‖ Cityscape 7(1): 23-39

Faberman, R. Jason (2001) ―Job creation and destruction within Washington and Baltimore,‖

Monthly Labor Review, September.

Faberman, R. Jason (2002) ―Job flows and labor dynamics in the Rust Belt,‖ Monthly Labor

Review, September.

Workforce Housing Final Report 115

Faberman, R. Jason (2006) ―New Evidence on the Labor Dynamics and Establishment

Distributions Underlying Urban Growth,‖ Bureau of Labor Statistics, June.

Fallick, Bruce, Charles A. Fleischman, and James B. Rebitzer (2005) ―Job Hopping in the

Silicon Valley: Some Evidence Concerning the Micro-Foundations of a High Technology

Cluster,‖ Federal Reserve Board, Finance and Economics Discussion Series, 2005-11.

Farber, Henry S. (1997) ―The Changing Face of Job Loss in the United States, 1981-1995,‖

Brookings Papers on Economic Activity: Microeconomics, 55-128.

Farber, Henry S. (1999) ―Mobility and Stability: The Dynamics of Job Change,‖ in

Ashenfelter and Card (eds.) Handbook of Labor Economics, Vol. 3b, Amsterdam:

Elsevier, 2439-2483.

Farber, Henry S. (2005) ―What Do We Know about Job Loss in the United States?

Evidence from the Displaced Workers Survey, 1984-2004,‖ Princeton University,

Industrial Relations Section, Working Paper #498, January 5.

Feldman, Maryann P. (1999) ―The New Economics of Innovation, Spillovers and

Agglomeration: A Review of Empirical Studies,‖ Economics of Innovation and New

Technology, 8: 5-25.

Figura, Andrew (2006) ―Why are Plant Deaths Countercyclical: Reallocation Timing or

Fragility?‖ Census Bureau Center for Economic Studies, CES 06-24.

Fischel, W.A. (1985) The Economics of Zoning Laws, Baltimore, MD: The Johns Hopkins

University Press.

Fischel, W.A. (1990) Do Growth Controls Matter: A Review of Empirical Evidence on the

Effectiveness and Efficiency of Local Government Land Use Regulation. Cambridge:

Lincoln Land Institute.

Fischel, W.A. (1997) Comment on ―The Portland Region: Where City and Suburb Talk with

Each Other—and Often Agree.‖ Housing Policy Debate 8(1):65–73.

116 Workforce Housing Final Report

Foote, Christopher L. (1998) ―Trend Employment Growth and the Bunching of Job Creation

and Destruction‖ Quarterly Journal of Economics, 113(3): 809-834.

Foster, Lucia, John Haltiwanger, and Namsuk Kim (2006) ―Gross Job Flows for the U.S.

Manufacturing Sector: Measurement from the Longitudinal Research Database,‖ Census

Bureau Center for Economic Studies CES 06-30, December.

Frech, H.E., III, and R. N. Lafferty, (1984) ―The Effect of the California Coastal Commission

on Housing Prices,‖ Journal of Urban Economics 16 (4): 105–123.

Freeman, Richard (1991) ―Labour Market Tightness and the Mismatch Between Demand

and Supply of Less-Educated Young Men in the United States in the 1980s,‖ in F.P.

Schioppa (ed.) Mismatch and Labour Mobility, Cambridge, UK: Cambridge University

Press, pp. 360-381.

Freeman, Richard B. and William H. Rodgers, III (1999) ―Area Economic Conditions and the

Labor Market Outcomes of Young Men in the 1990s Expansion,‖ Cambridge, MA:

National Bureau of Economic Research.

Frey, William H. (1993) ―The New Urban Revival in the United States,‖ Urban Studies, 30:

741-774.

Frey, William H. and Kao-Lee Liaw (2005) ―Interstate Migration of Hispanics, Asians and

Blacks: Cultural Constraints and Middle Class Flight‖ Population Studies Center

Research Report 05-575

Fu, Shine (2005) ―Smart Cafe Cities: Testing Human Capital Externalities in the Boston

Metropolitan Area,‖ AREUEA Conference, January 6.

Fu, Shine (2005b) ―What Has Been Capitalized into Property Values: Human Capital, Social

Capital, or Cultural Capital?‖ U.S. Census Bureau, Center for Economic Studies, CES

05-25

Workforce Housing Final Report 117

Fujita, Masahisa and Jacques-Francois Thisse (2002) Economics of Agglomeration: Cities,

Industrial Location, and Regional Growth, Cambridge University Press.

Fujita, Masahisa and Tomoya Mori (2005) ―Frontiers of the New Economic Geography,‖

Institute of Developing Economies, Japan External Trade Organization, Working paper

No. 69, April.

Fujita, Masahisa, Paul Krugman, and Anthony Venables (2001) The Spatial Economy:

Cities, Regions, and International Trade, MIT Press.

Gabe, Todd M. (2006) ―Growth of Creative Occupations in U.S. Metropolitan Areas: A Shift-

Share Analysis,‖ Growth and Change, 37(3): 396-415.

Gabriel, S. A., and J. R. Wolch (1980) ―Local Land Use Regulation and Urban Housing

Values,‖ Center for Real Estate and Urban Economics Working Paper No. 80–18.

Berkeley, CA: CREUE.

Gabriel, S.A., J. Shack-Marquez, and W.L. Wascher (1991) ―Regional house-price

dispersion and interregional migration,‖ USC School of Business Administration Working

Paper 90-41.

Gabriel, S.A., J. Shack-Marquez, and W.L. Wascher (1992) ―Regional House Price

Dispersion and Interregional Migration,‖ Journal of Housing Economics, 2(3): 235-256.

Gabriel S.A., J. Shack-Marquez, and W.L. Wascher (1993) ―The Effects of Regional House

Price and Labor Market Variability on Interregional Migration: Evidence from the 1980s,‖

in Housing Markets and Residential Mobility, edited by G. T. Kingsley and M. Turner,

Washington, DC: The Urban Institute Press.

Gabriel, Stuart A. and Stuart S. Rosenthal (2004) ―Quality of the Business Environment

versus Quality of Life: Do Firms and Households Like the Same Cities?‖ The Review of

Economics and Statistics, 86(1): 438-444.

Gallin, Joshua Hojvat (2004) ―Net Migration and State Labor Market Dynamics,‖ Journal of

Labor Economics, 22: 1-21.

118 Workforce Housing Final Report

Gambale, Geraldine (2006a) ―21 st Annual Corporate Survey,‖ Area Development,

(www.areadevelopment.com ).

Gambale, Geraldine (2006b) ―3 rd Annual Consultants Survey,‖ Area Development,

(www.areadevelopment.com ).

Garnick, Daniel H. and Howard L. Friedenberg (1982) ―Accounting for Regional Differences

in Per Capita Income Growth, 1929–79,‖ Survey of Current Business, 62 (September):

24–34.

Genesove, David and Christopher Mayer (2001) ―Loss Aversion and Seller Behavior:

Evidence from the Housing Market,‖ Quarterly Journal of Economics, 116(4): 1233-1260.

Gerston, Larry N. (2006) ―CEO Business Climate Survey‖ Conducted by the Silicon Valley

Leadership Group.

Gibbons, Robert and Michael Waldman (1999) ―Careers in Organizations: Theory and

Evidence,‖ Chap. 36 in Orley C. Ashenfelter and David Card (eds.), Handbook of Labor

Economics, Vol. 3b, Amsterdam: Elsevier, 2373-2438.

Glaeser, Edward L. (2006) ―The Economic Impact of Restricting Housing Supply,‖ Harvard

University Rappaport Institute for Greater Boston Policy Brief PB 2006-3.

Glaeser, Edward L. and David C. Mare (2001) ―Cities and Skills,‖ Journal of Labor

Economics, 19(2): 316-342.

Glaeser, Edward L., and Joseph Gyourko (2003) ―The Impact of Building Restrictions on

Housing Affordability,‖ Federal Reserve Bank of New York Economic Policy Review 9

(2): 21–39.

Glaeser, Edward L., Joseph Gyourko, and Raven E. Saks (2006) ―Why is Manhattan So

Expensive? Regulation and the Rise in House Prices,‖ Journal of Law and Economics.

Workforce Housing Final Report 119

Glaeser, Edward L., J. Kolko, and A. Saiz (2001) ―Consumer City,‖ Journal of Economic

Geography, 1: 27-50.

Glaeser, Edward L., J. Schuetz, and B. Ward (2006) ―Regulation and the Rise of Housing

Prices in Greater Boston”, Pioneer Institute for Public Policy Research and Rappaport

Institute for Greater Boston, Kennedy School of Government, Harvard University.

Glaeser, Edward L. and Matthew E. Kahn (2001) ―Decentralized Employment and the

Transformation of the American City,‖ in W. Gale and J. Pack, eds., Brookings-Wharton

Papers on Urban Affairs, Washington: The Brookings Institution Press, pp. 1-63 or

NBER Working Paper No. 8117.

Glosser, Stuart and Lonnie Golden (2005) Is Labour Becoming More or Less Flexible?

Changing Dynamic Behaviour and Asymmetries of Labour Input in US Manufacturing,‖

29(4): 535-557.

Goetz, Stephan J (1999) ―Migration and Local Labor Markets‖, in Scott Loveridge, ed., The

Web-Book of Regional Science, , Morgantown, WV, WVU Regional Research Institute,

Available at http://www.rri.wvu.edu/WebBook/Goetz/contents.htm

Gordon, P., B. Lee, and H.W. Richardson (2004) Travel Trends in U.S. Cities: Explaining the

2000 Census Commuting Results, Los Angeles, CA: University of Southern California.

Gordon, Ian, and Ian Molho (1998) ―A Multi-stream Analysis of the Changing Patterns of

Interregional Migration in Great Britain, 1960-1991‖ Regional Studies, 32(4): 309-323

Gottlieb, Paul D. (1994) ―Amenities as an Economic Development Tool: Is there Enough

Evidence?” Economic Development Quarterly, 8(3): 270-285.

Green, R. K., S. Malpezzi and S. K. Mayo (1999) ―Metropolitan-Specific Estimates of the

Price Elasticity of Supply of Housing, and their Sources,‖ Manuscript (Dec).

Green, R. K. and Malpezzi, S. (2003) A Primer on U.S. Housing Markets and Housing

Policy. Washington, D.C.: The Urban Institute.

120 Workforce Housing Final Report

Greene, Richard (1982) ―Tracking Job Growth in Private Industry,‖ Monthly Labor Review,

September: 3-9.

Greenwood, Michael J. (1985) ―Human Migration: Theory, Models, and Empirical Studies,‖

Journal of Regional Science, 25: 521-545.

Griliches, Zvi (1992) ―The Search for R&D Spillovers,‖ Scandanavian Journal of Economics,

94, Supplement, 29-47.

Gutierrez, Rose (2001) ―Income by State,‖ Housing Economics, February.

Gyourko, Joseph, Christopher Mayer, and Todd Sinai (2006) ―Superstar Cities,‖ NBER

Working Paper No. 12355.

Gyourko, Joseph, Albert Saiz, and Anita Summers (2006) ―A New Measure of the Local

Regulatory Environment for Housing Markets: The Wharton Residential Land Use

Regulatory Index,‖ Wharton Working Paper, October 22.

Gyourko, Joseph and Joseph Tracy (1991) "The Structure of Local Public Finance and the

Quality of Life," Journal of Political Economy, 91(4): 774-806.

Haaland, J.I., Kind, H.J., and Midelfart-Knarvik, K.H. (1999) ―What Determines the Economic

Geography of Europe?‖ CEPR Discussion Paper No. 2072.

Haas, Peter M., Carrie Makarewicz, Albert Benedict, Thomas W. Sanchez, and Casey J.

Dawkins (2006) ―Housing and Transportation Cost Trade-offs and Burdens of Working

Households in 28 Metros,‖ Center for Neighborhood Technology and Virginia Tech, July.

Hacker, R. Scott (1999) ―The Effect of Residential Crowding on Labor Productivity with

Evidence from the Twilight of Polish Socialism,‖ Real Estate Economics, 27(1): 135-167.

Hamalainen, Kari and Petri Bockerman (2004) ―Regional Labor Market Dynamics, Housing,

and Migration,‖ Journal of Regional Science, 44(3): 543-568.

Workforce Housing Final Report 121

Hamermesh, Daniel S. (1989) ―What Do We Know about Worker Displacement in the U.S.?‖

Industrial Relations, 28: 51-59.

Hanson, Gordon H. (1997) ―Increasing Returns, Trade, and the Regional Structure of

Wages,‖ Economic Journal, 107(440): 113-133.

Hanson, Gordon H. (1998) ―Market Potential, Increasing Returns, and Geographic

Concentration,‖ National Bureau of Economic Research Working Paper #6429, Revised

Dec. 2001.

Harris, Timothy F. and Yannis M. Ioannides (2000) ―Productivity and Metropolitan Density,‖

Discussion Paper Series, Department of Economics, Tufts University, Number 16, May

15.

Hazans, Hihails (2004) ―Does Commuting Reduce Wage Disparities?‖ Growth and Change,

35(3): 360.

Hayward, Steven F. (undated) ―An Overview of Research on the Costs of Housing

Regulation,‖ American Enterprise Institute.

Head, Keith and Thierry Mayer (2004) ―The Empirics of Agglomeration and Trade,‖ in J.

Vernon Henderson and Jacques-Francois Thisse (eds.) Handbook of Regional and

Urban Economics, Vol. 4, Amsterdam: Elsevier.

Headd, Brian (2001) ―Business Success: Factors leading to surviving and closing

successfully,‖ U.S. Census Bureau, Center for Economic Studies Working Paper 01-01.

Heckman, James, J., Robert Lalonde, and Jeffrey A. Smith (1999) ―The Economics and

Econometrics of Active Labor Market Programs,‖ in O. Ashenfelter and D. Card (eds.)

Handbook of Labor Economics, Vol. III, Amsterdam: North-Holland.

Heckman, James J. and Jeffrey A. Smith (2004) ―The Determinants of Participation in a

Social Program: Evidence from a Prototypical Job Training Program,‖ Journal of Labor

Economics, 22: 243-298.

122 Workforce Housing Final Report

Hellerstein, Judith K., David Neumark, and Kenneth R. Troske (1999) ―Wages, Productivity,

and Worker Characteristics: Evidence from Plant-Level Production Functions and Wage

Equations,‖ Journal of Labor Economics, 17(3): 409-446.

Henderson, J. Vernon (1997) ―Externalities and Industrial Development,‖ Journal of Urban

Economics, 42: 449-470.

Holmes, Thomas J. and John J. Stevens (2004) ―Spatial Distribution of Economic Activities

in North America,‖ in J. V. Henderson & J. F. Thisse (ed.) Handbook of Regional and

Urban Economics, Chapter 63, pages 2797-2843 Elsevier.

Holzer, Harry J. (1996) What Employers Want? Job Prospects for Less-Educated Workers,

New York: Russell Sage Foundation.

Holzer, Harry J. (2004) ―Encouraging Job Advancement Among Low-Wage Workers: A New

Approach,‖ The Brookings Institution Policy Brief, Welfare Reform and Beyond #30,

May.

Hoover, Edward M. and Frank Giarratani (1984) An Introduction to Regional Economics, 3 rd

Edition, Alfred A. Knopf (http://www.rri.wvu.edu/WebBook/Giarratani/main.htm ).

Horner, Mark W. (2006) ―A Multi-Scale Analysis of Urban Form and Commuting Change in a

Small Metropolitan Area (1990-2000),‖ The Annals of Regional Science, published

online, October 21.

Hughes, Gordon and Barry McCormick (1987) ―Housing Markets, Unemployment and

Labour Market Flexibility in the UK,‖ European Economic Review, 31(3), 615-41.

Ihlanfeldt, Keith R. and Shaughnessy, Timothy M. (2004) ―An Empirical Investigation of the

Effects of Impact Fees on Housing and Land Markets,‖ Regional Science and Urban

Economics. Elsevier, 34 (6): 639-661.

Imerman, Mark D., Peter Orazem, Shiva Sikdar, and Gina Russell (2006) ―Wages, benefits,

commuting time, and license renewal for Iowa Registered Nurses,‖ Iowa State University

Dept. of Economics Working Paper Series, #06030, September.

Workforce Housing Final Report 123

Institute of Transportation Engineers (2001) Trip Generation Handbook, Washington, DC:

Institute of Transportation Engineers.

Jackman, Richard, and Savvas Savouri (1992) ―Regional Migration versus Regional

Commuting: The Identification of Housing and Employment Flows‖, Scottish Journal of

Political Economy 39(3): 272-287

Jacobs, Jane (1969) The Economy of Cities, New York: Vintage Press.

Jaffe, A. B., M. Trajtenberg, and R. Henderson (1993) ―Geographic Localization of

Knowledge Spillovers as Evidenced by Patent Citations,‖ Quarterly Journal of

Economics, 10, 577-598.

Johnes Geraint and Thomas J. Hyclak (1994) ―House prices, migration and regional labor

markets,‖ Journal of Housing Economics, 3:312–329.

Johnes, Geraint and Thomas J. Hyclak (1999) ―House Prices and Regional Labor Markets,‖

The Annals of Regional Science, 33(1): 33-49.

Jones, Colin and Chris Leishman (2006) ―Spatial Dynamics of the Housing Market: An

Interurban Perspective,‖ Urban Studies, 43(7): 1041-1059.

Jorgenson, Dale W., Mun S. Ho, and Kevin J. Stiroh (2006) ―Potential Growth of the U.S.

Economy: Will the Productivity Resurgence Continue?‖ Business Economics, January.

Jud, G. D. and D.T. Winkler (2002) ―The Dynamics of Metropolitan House Prices,‖ Journal of

Real Estate Research, 23: 29-45.

Juhn, Chinhui, Kevin M. Murphy, and Brooks Pierce (1993) ―Wage Inequality and the Rise in

the Return to Skill,‖ Journal of Political Economy, 101(3): 410-442.

Kain, John F. (1968) ―Housing Segregation, Negro Employment, and Metropolitan

Decentralization,‖ The Quarterly Journal of Economics, 82: 175-197.

124 Workforce Housing Final Report

Kain, John F. (1992) ―The Spatial Mismatch Hypothesis: Three Decades Later,‖ Housing

Policy Debate, 3(2): 371-460.

Kamarianakis, Yiannis and Julie Le Gallo (2006) ―Exploratory Spatial Data Analysis and

Spatial Econometric Modeling for the study of Regional Productivity Differentials in

European Union, from 1975 to 2000,‖ University of Crete, Dept. of Economics, Working

Paper No. 401.

Katz, Bruce (2007) ―Remaking Transportation and Housing Policy for the New Century,‖

Congressional testimony before the House Appropriations Subcommittee on

Transportation, Housing and Urban Development, and Related Agencies, U.S.

Congress. February 28.

Katz, Lawrence F. and Alan Krueger (1999) ―The High-Pressure U.S. Labor Market of the

1990s,‖ Brookings Papers on Economic Activity 1: 1-65.

Katz, L., and K. Rosen (1987) ―The Interjurisdictional Effects of Growth Controls on Housing

Prices,‖ Journal of Law and Economics, 30:149–60.

Kiel, K.A. (1994) ―The Impact of House Price Appreciation on Household Mobility,‖ Journal

of Housing Economics, 3(2): 92-108.

Kiel, K. A. (2004) ―Environmental Regulations and the Housing Market: A Review of the

Literature,‖ (paper prepared for HUD Conference on Regulatory Barriers to Housing,

April 22, 2004).

Kim, Sukkoo (1995) ―Expansion of Markets and the Geographic Distribution of Economic

Activities: The Trends in U.S. Regional Manufacturing Structure, 1860-1987,‖ Quarterly

Journal of Economics, 110(4): 881-908.

Knaup, Amy E. (2005) ―Survival and Longevity in the Business Employment Dynamics

Data,‖ Monthly Labor Review, May: 50-56.

Koo, Jun (2005a) ―Technology Spillovers, Agglomeration, and Regional Economic

Development,‖ Journal of Planning Literature, 20(2): 99-115.

Workforce Housing Final Report 125

Koo, Jun (2005b) ―Agglomeration and Spillovers in a Simultaneous Framework,‖ The Annals

of Regional Science, 39:35-47.

Krisek, K. (2003) ―Residential Relocation and Changes in Urban Travel: Does

Neighborhood-Scale Urban Form Matter?‖ Journal of the American Planning

Association, 69(3): 265-281.

Krueger, Alan B. and Lawrence H. Summers (1988) ―Efficiency Wages and the Inter-

Industry Wage Structure,‖ Econometrica, 56(2): 259-293.

Krugman, Paul (1991) Geography and Trade, MIT Press.

Krugman, Paul (1991) ―Increasing Returns and Economic Geography,‖ Journal of Political

Economy, 99: 483-499.

Lane, Julia, Ned English, Fredrik Andersson, and Patrick Park (11/14/2006) ―Workforce

Development and Rental Policy: How Can What We Know Inform Next Steps?‖ paper

presented at the conference ―Revisiting Rental Housing: A National Policy Summit,‖

Joint Center for Housing Studies, Harvard University.

Leonard, Jonathan S. (1987) ―Carrots and Sticks: Pay, Supervision, and Turnover,‖ Journal

of Labor Economics, 5(4): S136-S152.

Leonard, Jonathan and David Levine, (2006) ―The Effect of Diversity on Turnover: A Large

Case Study,‖ Industrial & Labor Relations Review, 59(4): 547-572.

Levine, Jonathan (1992) ―Decentralization of Jobs and Emerging Suburban Commute,‖

Transportation Research Record, 1364: 71-80.

Levine, Jonathan (1998) ―Rethinking Accessibility and Jobs-Housing Balance,‖ Journal of

the American Planning Association, 64(2): 133-149.

Levinson, David and A. Kumar (1994) ―Rational Locator: Why Travel Times Have Remained

Stable,‖ Journal of the American Planning Association, 60: 319-332.

126 Workforce Housing Final Report

Levinson, David and Yao Wu (2005) ―Rational Locator Reexamined: Are Travel Times Still

Stable?‖ Transportation, 32(2): 187-202.

Lipman, Barbara J. (2006) ―A Heavy Load: The Combined Housing and Transportation

Burdens of Working Families,‖ Center for Housing Policy, October.

Listokin, D. and D. Hattis (2004) ―Building Codes and Housing‖ (paper prepared for HUD

Conference on Regulatory Barriers to Housing, April 22, 2004).

Loury, Linda Datcher (2006) ―Informal Contacts and Job Search Among Young Workers,‖

Tufts University, Dept. of Economics Working Paper.

Lucas, Robert E. (1988) ―On the Mechanics of Economic Development,‖ Journal of

Monetary Economics, 22: 3-42.

Luger, M. I. and K. Temkin (2000) Red Tape and Housing Costs, New Brunswick: Center for

Urban Policy Research.

MacGillis, Alec (2007) ―Fairfax Shines In Jobs Report: County Challenges D.C.‘s Leadership

of Area Economy,‖ Washington Post, January 30, A01.

Magrini, Stefano (2004) ―Regional (Di)convergence,‖ Chapter 62 in J.V. Henderson and J.F.

Thisse (eds.) Handbook of Regional and Urban Economics, Vol. 4, New York: Elsevier.

Malcomson, James M. (1999) ―Individual Employment Contracts,‖ Chapter 35 in Orley C.

Ashenfelter and David Card (eds.), Handbook of Labor Economics, Vol. 3b, Amsterdam:

Elsevier, 2291-2372.

Malpezzi, Stephen (1996) ―Housing Prices, Externalities, and Regulation in U.S.

Metropolitan Areas,‖ Journal of Housing Research, 7 (2): 209–242.

Malpezzi, Stephen (1999) ―Estimates of the Measurement and Determinants of Urban

Sprawl in U.S. Metropolitan Areas,‖ Center for Urban Land Economics Research,

University of Wisconsin.

Workforce Housing Final Report 127

Malpezzi, Stephen (2002) ―Urban Regulation, the ‗New Economy,‘ and Housing Prices‖

Housing Policy Debate, 13(2) 323-349.

Marshall, Alfred (1920) Principles of Economics, 8 th edition, Macmillan

(http://www.econlib.org/library/Marshall/marPContents.html ) Book IV:X.

Martin, Richard W. (2004) ―Spatial Mismatch and the Structure of American Metropolitan

Areas, 1970-2000,‖ Journal of Regional Science, 44(3): 467-488.

Martin, Ron and Peter Sunley (1996) ―Paul Krugman‘s Geographical Economics and Its

Implications for Regional development Theory: A Critical Assessment‖ Economic

Geography, 72(1): 259-292.

Mas, Alexandre and Enrico Moretti (2006) ―Peers at Work,‖ NBER Working Paper No.

12508, September.

Maurel, F. and Sedillot, B. (1999) "A Measure of the Geographic Concentration in French

Manufacturing Industries," Regional Science and Urban Economics, 29(5): 575-604.

May, Peter J. (2004) ―Regulatory Implementation: Examining Barriers from Regulatory

Processes‖ (paper prepared for HUD Conference on Regulatory Barriers to Housing,

April 22, 2004).

Mayer, C. J., and C. T. Somerville (2000) ―Land Use Regulation and New Construction,‖

Regional Science and Urban Economics 30: 639–662.

Mayo, Stephen K. and James I. Stein (1988) ―Housing and Labor Market Distortions in

Poland: Linkages and Policy Implications,‖ World Bank Policy Planning and Research

Staff Discussion Paper, June.

McCormick, B. (1997) ―Regional Unemployment and Labour Mobility in the UK,‖ European

Economic Review, 41: 581-589.

128 Workforce Housing Final Report

McGeehan, Patrick (2007) ―Building Boom May Mean Jobs Can‘t All Be Filled, Report Says,‖

NY Times, January 26.

McQuaid, Ronald W. (2006) ―Job Search and Employability in Local Labor Markets,‖ Annals

of Regional Science, 40(2): 407-421.

Mitra, Arup and Hajime Sato (2006) ―Agglomeration Economies in Japan: Technical

Efficiency, Growth and Unemployment,‖ Institute of Developing Economies, Discussion

Paper No. 48, February.

Moomaw, R.L. (1981) ―Productivity and City Size: A Critique of the Evidence,‖ Quarterly

Journal of Economics, 96: 675-688.

Moretti, Enrico (2004a) ―Human Capital Externalities in Cities,‖ in J. Vernon Henderson and

Jacques-Francois Thisse (eds.) Handbook of Regional and Urban Economics, Vol. 4,

New York: Elsevier.

Moretti, Enrico (2004b) ―Workers‘ Education, Spillovers and Productivity: Evidence from

Plant-Level Production Functions,‖ American Economic Review, 94(3): 656-690.

Morrison, Philip S. (2005) ―Unemployment and Urban Labour Markets,‖ Urban Studies,

42(12): 2261-2288.

Mortensen, Dale T. and Christopher A. Pissarides (1999) ―New Developments in Models of

Search in the Labor Market,‖ Ch. 39 in Orley C. Ashenfelter and David Card (eds.)

Handbook of Labor Economics, Vol. 3b, Amsterdam: Elsevier, 2567-2628.

Munch, Jakob Roland, Michael Rosholm, and Michael Svarer (2006) ―Home Ownership, Job

Duration and Wages,‖ Institute for the Study of Labor (IZA) Discussion Paper No. 2110,

May.

Murphy, Anthony, John Muellbauer and Gavin Cameron (2006) ―Housing Market Dynamics

and regional Migration in Britain,‖ University of Oxford, Department of Economics

Discussion Paper series No. 275, August.

Workforce Housing Final Report 129

Muth, Richard F. (1969) Cities and Housing, Chicago: University of Chicago Press.

Naude, C.M. (2006) ―Measures of Manufacturing Industry Concentration – Implications for

South Africa,‖ TIPS (Trade and Industry Policy Strategies) Forum.

Nelson, Arthur C. (1999) ―Comparing States With and Without Growth Management

Analysis Based on Indicators with Policy Implications,‖ Land Use Policy, 16: 121-127.

Nelson, Arthur C. (2004) ―Workforce Housing Balance for the Atlanta Regional Commission:

A Spatial Distribution Assessment,‖ mimeo, March 25.

Neumark, David, Junfu Zhang, and Brandon Wall (2006a) ―Employment Dynamics and

Business Relocation: New Evidence from the National Establishment Time Series‖ UC

Irvine, Department of Economics, Working Paper 050624, March.

Neumark, David, Junfu Zhang, and Jed Kolko (2006b) ―Interstate Business Relocation: An

Industry-Level Analysis,‖ Public Policy Institute of California, Occasional Papers, June.

New York Times (1918) "The Backward Shipbuilding Program," p. 12.

(See www.fairview.ws/beginnings/planning.html ).

Nickrell, S.J. (1976) ―Wage Structures and Quit Rates,‖ International Economic Review, 17:

191-203.

Nowlan, David and Greg Stewart (1991) ―Downtown Population Growth and Commuting

Trips,‖ Journal of the American Planning Association, 57(2): 165-182.

Nucci, Alfred R. (1999) ―The Demography of Business Closings,‖ Small Business

Economics, 12(1): 25-39.

van Ommeren, Jos, Piet Rietveld, and Peter Nijkamp (1999) "Job Moving, Residential Moving, and Commuting: A Search Perspective" Journal of Urban Economics 46: pp 230- 253

130 Workforce Housing Final Report

Oskamp, Anton (1995) ―LocSim: A Microsimulation Approach to Household and Housing

Market Modelling,‖ PDOD-Paper No. 29, Department of Planning and Demography,

University of Amsterdam.

Oswald, Andrew (1996) ―A Conjecture on the Explanation for High Unemployment in the

Industrialized Nations: Part I,‖ University of Warwick Economic Research Papers No.

475.

Oswald, A.J. (1997) Theory of Homes and Jobs, University of Warwick.

Ottaviano, Gianmarco and Jacques-Francois Thisse (2004) ―Agglomeration and Economic

Geography,‖ in J. Vernon Henderson and Jacques-Francois Thisse (eds.) Handbook of

Regional and Urban Economics, Vol. 4, New York: Elsevier.

Ottaviano, Gianmarco and Giovanni Peri (2006) ―The Economic Value of Cultural Diversity:

Evidence from U.S. Cities,‖ Journal of Economic Geography, 6(1): 9-44.

Owyang, Michael T., Jeremy M. Piger, Howard J. Wall, and Christopher H. Wheeler, (2006)

―The economic performance of cities: a Markov-switching approach,‖ Federal Reserve

Bank of St. Louis Working Paper 2006-056, October.

Parker, Robert P. (2006) ―Employment Dynamics: BLS and Census Bureau Use

Administrative Records to Provide New Data,‖ Business Economics, April: 55-61.

Paumgarten, Nick (2007) "There and Back Again: The Soul of the Commuter" The New

Yorker, April 16

Pendall, R. (1999) ―Do Land-Use Controls Cause Sprawl?‖ Environment and Planning B,

26:555-571.

Pendall, R. (2000) ―Local Land Use Regulation and the Chain of Exclusion,‖ Journal of the

American Planning Association, 66: 125-142.

Workforce Housing Final Report 131

Pendall, R., R. Puentes, and J. Martin (2006) ―From Traditional to Reformed: A Review of

the Land Use Regulations in the Nation‘s 50 largest Metropolitan Areas,‖ Research Brief,

Washington, D.C., Brookings Institution Press.

Peri, Giovanni (2002) ―Young Workers, Learning, and Agglomeration,‖ Journal of Urban

Economics, 52(3): 582-607.

Perry, Marc J. (2006) ―Domestic Net Migration in the United States: 2000 to 2004,‖ U.S.

Census Bureau, Economics and Statistics Administration, Current Population Reports.

Pingle, Jonathan F. (2006) ―The Relocation Decisions of Working Couples,‖ Federal

Reserve Board Finance and Economics Discussion Series, 2006-33, August.

Pinkston, Joshua C. and James R. Spletzer (2004) ―Annual measures of gross job gains

and gross job losses,‖ Monthly Labor Review, November: 3-13.

Pisarski, Alan (2006) Commuting in America III, Washington: Transportation Research

Board

Pissarides, C.A. and J. Wadsworth (1994) ―On-the-Job Search: Some Empirical Evidence

from Britain, European Economic Review, 38(2): 385-401.

Pogodzinski, J.M. (1995) ―The Effect of Housing Market Disequilibrium on the Supply of

Labor: Evidence from Poland 1989-1990,‖ Real Estate Economics, 23(4): 497-527.

Pogodzinski, M. and T. Sass (1991) ―Measuring the Effects of Municipal Zoning

Regulations: A Survey,‖ Urban Studies 28 (4) 597–621.

Pollakowski, H. O., and S.M. Wachter (1990) ―The Effects of Land-Use Constraints on

Housing Prices,‖ Land Economics 66 (3): 315–324.

Potepan, M.J. (1994) ―Intermetropolitan Migration and Housing Prices: Simultaneously

Determined?‖ Journal of Housing Economics, 3(2): 77-91.

132 Workforce Housing Final Report

Prud‘homme,R. and G. Lee (1999) ―Sprawl, Speed and the Efficiency of Cities,‖ Urban

Studies, 36: 1849-1858.

Quigley, John M. and L.A. Rosenthal (2004) ―The Effects of Land-Use Regulation on the

Price of Housing: A Literature Review and Research Agenda‖ (paper prepared for HUD

Conference on Regulatory Barriers to Housing, April 22, 2004).

Quigley, John M. and L.A. Rosenthal (2005) ―The Effects of Land Use Regulation on the

Price of Housing: What Do We Know? What Can We Learn?‖ Cityscape, 8 (1): 69-137.

Quintin, Erwan and John J. Stevens (2005) ―Raising the Bar for Models of Turnover,‖

Federal Reserve Board, Finance and Economics Discussion Series, 2005-23, April 26.

Raphael, Steven and Michael A. Stoll (2002) "Modest Progress: The Narrowing Spatial

Mismatch Between Blacks and Jobs in the 1990s," The Brookings Institution, Census

Series, December.

Rauch, J.E. (1993) ―Productivity Gains from Geographic Concentration of Human Capital:

Evidence from the Cities,‖ Journal of Urban Economics, 34: 380-400.

Rephann, Terance J., Kalle Makila, and Einar Holm (2005) ―Microsimulation for Local Impact

Analysis: An Application to Plant Shutdown,‖ Journal of Regional Science, 45(1): 183-

222.

Rigby, David L. and Jurgen Essletzbichler (2000) ―Impacts of Industry Mix, Technological

Change, Selection and Plant Entry/Exit on Regional Productivity Growth,‖ Regional

Studies, 34(4): 333-342.

Roback, J. (1982) ―Wages, Rents and the Quality of Life,‖ Journal of Political Economy, 90,

1257-1278.

Roback, J. (1988) ―Wages, rents, and amenities: Differences among workers and regions,‖

Economic Inquiry, 26(1): 23-41.

Workforce Housing Final Report 133

Rodriguez, Daniel A., Felipe Targa, and Semra A. Aytur (2006) ―Transport Implications of

Urban Containment Policies: A Study of the Largest Twenty-five US Metropolitan Areas,‖

Urban Studies, 43(10): 1879-1897.

Rosenthal, Stuart S. and William C. Strange (2001) ―The Determinants of Agglomeration,‖

Journal of Urban Economics, 50: 191-229.

Rosenthal, Stuart S. and William C. Strange (2003) ―Geography, Industrial Organization,

and Agglomeration,‖ Review of Economics and Statistics, 85(2): 377-393.

Rosenthal, Stuart S. and William C. Strange (2004) ―Evidence on the Nature and Sources of

Agglomeration Economies,‖ in Henderson, V. and J.F. Thisse (eds.) Handbook of Urban

and Regional Economics: Cities and Geography (Vol. 4), New York: Elsevier.

Rouwendal, Jan (2004) ―Search Theory and Commuting Behavior,‖ Growth and Change,

35(3): 391-418.

Rouwendal, Jan and Peter Nijkamp (2004) ―Living in Two Worlds: A Review of Home-to-

Work Decisions,‖ Growth and Change, 35(3): 287-303.

Rubinfeld, D. L., (NEED YEAR) ―Suburban Employment and Zoning: A General Equilibrium

Analysis,‖ Journal of Regional Science, 18(1): 33-44.

Safirova, Elena, Sebastien Houde, D. Abram Lipman, Winston Harrington, and Andrew

Baglino (2006) ―Congestion Pricing: Long-Term Economic and Land-Use Effects,‖

Resources for the Future Discussion Paper 06-37, September.

Saks, Raven E. (2004) ―Job creation and housing construction: Constraints on employment

growth in metropolitan areas,‖ Harvard University Mimeograph.

Saks, Raven E. (2004) ―Housing supply restrictions across the United States,‖ Wharton

Real Estate Review.

134 Workforce Housing Final Report

Saks, Raven E. (2005) ―Job Creation and Housing Construction: Constraints on

Metropolitan Area Employment Growth‖ Federal Reserve Board, Finance and

Economics Discussion Series, No. 2005-49.

Salvesen, David and Henry Renski (2003) ―The Importance of Quality of Life in the Location

Decisions of New Economy Firms‖ University of North Carolina, Center for Urban and

Regional Studies.

Salop, J. and S. Salop (1976) ―Self-selection and Turnover in the Labor Market,‖ Quarterly

Journal of Economics, 90: 619-628.

Sanchez, Thomas W., Qing Shen, and Zhong-Ren Peng (2004) ―Transit Mobility, Jobs

Access and Low-Income Labour Participation in US Metropolitan Areas,‖ Urban Studies,

41(7): 1313-1331.

Sarzynski, Andrea, Harold L. Wolman, George Galster, and Royce Hanson (2006) ―Testing

the Conventional Wisdom about Land Use and Traffic Congestion: The More We

Sprawl, the Less We Move?‖ Urban Studies, 43(3): 601-626.

Schaffer, William A. (1999) Regional Impact Models, Regional Research Institute, West

Virginia University.

Schill, M. H. (2004) ―Regulations and Housing Development: What We Know and What We

Need to Know,‖ Prepared for the U.S. Department of Housing and Urban Development‘s

Conference on Regulatory Barriers to Affordable Housing.

Schrank, D. and T. Lomax (2003) Urban Mobility Study, Texas Transportation Institute,

College Station, TX.

Schrank, David and Tim Lomax (2005) ―The 2005 Urban Mobility Report,‖ Texas

Transportation Institute, May.

Schwanen, Tim, Frans M. Dieleman, and Martin Dijst (2004) ―The Impact of Metropolitan

Structure on Commute Behavior in the Netherlands: A Multilevel Approach,‖ Growth and

Change, 35(3): 304-333.

Workforce Housing Final Report 135

Scully, G.W. (1969) ―Interstate Wage Differentials: A Cross Section Analysis,‖ The American

Economic Review, 59(5): 757-773.

Segal, D.and P. Srinivasan (1985) ―The Impact of Suburban Growth Restrictions on US

Housing Price Inflation, 1975-1978,‖ Urban Geography 6(1): 1-26.

Seidel, S. (1978) Housing Costs and Government Regulations. New Brunswick, NJ: The

Center for Urban Policy Research.

Shapiro, Jesse M. (2006) ―Smart Cities: Quality of Life, Productivity, and the Growth Effects

of Human Capital,‖ The Review of Economics and Statistics, 88(2): 324-335.

Shapiro, C. and J. Stiglitz (1984) ―Equilibrium Unemployment as a Worker Discipline

Device,‖ American Economic Review, 74: 433-444.

Shefer, D. (1973) ―Localization Economies in SMSA‘s: A Production Function Analysis,‖

Journal of Regional Science, 13: 55-64.

Shen, Qing (2001) ―A Spatial Analysis of Job Openings and Access in a U.S. Metropolitan

Area,‖ Journal of the American Planning Association, 67(1): 53-68.

Sichelman, Lew (2007) ―Public Transit is a Major Support for House Prices,‖ Chicago

Tribune, February 11.

Smoluk, Herbert J. and Bruce H. Andrews (2005) ―A Prerequisite for Meaningful State

Economic Performance Comparisons: Adjusting for Population Density,‖ The Annals of

Regional Science, 39: 253-272.

Soureli, K., Pelletiere, D., and Wardrip, K. (2006) ―Leaving Employment and Unaffordable

Cities: Where are the Low-Income People Migrating To?‖ National Low Income Housing

Coalition Research Note #06-02.

136 Workforce Housing Final Report

Spletzer, James R. (1998) ―The Contribution of Establishment Births and Deaths to

Employment Growth,‖ BLS Office of Employment Research and Program Development

Working Paper 310, March.

Spletzer, James R. (2000) ―The Contribution of Establishment Births and Deaths to

Employment Growth,‖ Journal of Business and Economic Statistics, 18(1): 113-126.

Spletzer, James R., R. Jason Faberman, Akbar Sadeghi, David M. Talan, and Richard L.

Clayton (2004) ―Business employment dynamics: new data on gross job gains and

losses,‖ Monthly Labor Review, April: 29-42.

Stewart, Jay (2002) ―Recent Trends in Job Stability and Job Security: Evidence from the

March CPS,‖ Bureau of Labor Statistics, Office of Employment and Unemployment

Statistics, Working Paper 356, March.

Strand, Jon (2002) ―Wage Bargaining and Turnover Costs with Heterogeneous Labor and

Perfect History Screening,‖ European Economic Review, 46: 1209-1227.

Strange, William C. (2005) ―Urban Agglomeration‖ forthcoming in S. Durlauf and L Blume,

eds., New Palgrave Dictionary of Economics, 2 nd

Edition, London:Macmillan.

Stull, W. J., (1974) ―Land Use and Zoning in an Urban Economy‖ The American Economic

Review, 64 (3): 337-347.

Tabuchi, T. (1986) ―Urban Agglomeration, Capital Augmenting Technology, and Labor

Market Equilibrium,‖ Journal of Urban Economics, 20: 211-228.

Tabuchi, T. and A. Yoshida (2000) ―Separating Agglomeration Economies in Consumption

and Production,‖ Journal of Urban Economics, 48:70-84.

Tolbert, Charles M., Troy C. Blanchard, and Michael D. Irwin (2006) ―Stability and Change in

Individual Determinants of Migration: Evidence From 1985-1990 and 1995-2000,‖ U.S.

Bureau of the Census, Center for Economic Studies, Working Paper CES 06-27,

November.

Workforce Housing Final Report 137

U.S. Census Bureau (2005) "Americans Spend More Than 100 Hours Commuting to Work

Each Year" Press Release CB05-AC.02, March 30

Valletta, Robert G. (2005) ―Why Has the Beveridge Curve Shifted Back? New Evidence

Using Regional Data,‖ Federal Reserve Bank of San Francisco Working Paper 2005-25,

December.

Valletta, Rob and Jaclyn Hodges (2006) ―Job Matching: Evidence from the Beveridge

Curve‖ FRBSF Economic Letter, Number 2006-08, April 21

Vermeulen, Wouter and Jos van Ommeren (2004) ―Are Workers Compensated by Cheaper

Housing in Regions Where Unemployment is High? Theory and Evidence from a

Housing Demand Survey,‖ European Regional Science Association 2005 Conference in

Amsterdam, ersa05p204.

Vermeulen, Wouter and Jos van Ommeren (2005) ―Compensation of Regional

Unemployment in Housing Markets‖ Tinbergen Institute Discussion Paper TI 2005-093/3

Waller, Margy (2005) ―High Cost or High Opportunity Cost? Transportation and Family

Economic Success,‖ Policy Brief, The Brookings Institution, December.

Walls, Margaret, Elena Safirova, and Yi Jiang (2006) ―Long-Term Consequences of

Congestion Pricing: A Small Cordon in the Hand Is Worth Two in the Bush,‖ Resources

for the Future Discussion Paper 06-41, October.

Wardrip, Keith E., Danilo Pelletiere, and Konstantina Soureli (2006) ―Variability of Rents

Relative to Wages for Low-Income Workers May Provide Incentive for Migration‖

National Low Income Housing Coalition Research note #06-01.

Warner, K. and H. Molotch (2001) Building Rules: How Local Controls Shape Community

Environments and Economies, New York: Westview Press

Wharton Land Use Control Survey. Prepared by Anita A. Summers and Peter Linneman.

Philadelphia: University of Pennsylvania, Wharton School, 1989.

138 Workforce Housing Final Report

Wheaton, W.C. and M. J. Lewis (2002) ―Urban Wages and Labor Market Agglomeration,‖

Journal of Urban Economics, 51: 542-562.

Wheeler, Christopher H. (2005) ―Cities, Skills, and Inequality,‖ Growth and Change, 36(3):

329-353.

Wheeler, Christopher H. (2006) ―Local Market Scale and the Pattern of Job Changes Among

Young Men,‖ Upjohn Institute Working Paper No. 06-131, November 7.

Wheeler, Christopher H. and Elizabeth A. La Jeunesse (2006) ―Trends in the distributions of

income and human capital within metropolitan areas: 1980-2000,‖ St. Louis Fed Working

Paper 2006-055A, October.

Winston, Clifford and Ashley Langer (2006) ―The Effect of Government Highway Spending

on Road Users‘ Congestion Costs,‖ AEI-Brookings Joint Center for Regulatory Studies,

Working Paper 06-11, May (forthcoming in the Journal of Urban Economics).

Yashiv, Eran (2006) ―U.S. Labor Market Dynamics Revisited,‖ IZA Discussion Paper No.

2455, November.

Yerokhin, Oleg (2006) ―Unobserved Heterogeneity and the Urban Wage Premium‖ Iowa

State University Working Paper, November.

Zovanyi, G. (1998) ―Growth Management for Sustainable Future: Ecological Sustainability

as the New Growth Management Focus for the 21 st Century,‖ Westport, CT: Praeger.

D ra

ft W o

rk fo

rc e

H o

u s in

g L

ite ra

tu re

R e v ie

w

1

3 9

Full-Time and Part-Time Employees by Industry

1955 1965 1975 1985 1995 2005

All industries 59,218 69,713 85,069 105,874 124,783 141,218

Private industries 48,036 55,598 67,069 86,079 102,707 117,090

All industries 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%

Private industries 81.1% 79.8% 78.8% 81.3% 82.3% 82.9%

Agriculture, forestry, fishing, and hunting 3.8% 2.5% 1.9% 1.3% 1.1% 1.0%

Mining 1.3% 0.9% 0.8% 0.8% 0.4% 0.4%

Utilities 0.8% 0.7% 0.7% 0.7% 0.5% 0.4%

Construction 4.9% 4.9% 4.4% 4.7% 4.4% 5.4%

Manufacturing 27.1% 24.6% 20.4% 17.0% 13.7% 10.1%

Durable goods 15.6% 14.5% 12.2% 10.5% 8.3% 6.4%

Nondurable goods 11.5% 10.1% 8.2% 6.5% 5.5% 3.8%

Wholesale trade 4.4% 4.5% 4.6% 4.8% 4.4% 4.1%

Retail trade 9.2% 9.8% 11.0% 11.4% 11.2% 11.2%

Transportation and warehousing 4.8% 3.8% 3.2% 2.8% 3.1% 3.1%

Information 2.6% 2.3% 2.4% 2.4% 2.3% 2.2%

Finance, insurance, real estate, rental, and leasing 3.9% 4.4% 5.0% 5.8% 5.6% 5.9%

Professional and business services 3.6% 4.6% 5.8% 8.5% 11.3% 12.3%

Management of companies and enterprises 1.1% 1.1% 1.1% 1.2% 1.2% 1.2%

Education, health care, and social assistance 3.4% 4.7% 7.1% 8.7% 11.0% 12.7%

Educational services 0.9% 1.2% 1.4% 1.5% 1.7% 2.1%

Health care and social assistance 2.5% 3.4% 5.6% 7.2% 9.3% 10.6%

Arts, recreation, accommodation, & food services 4.8% 5.2% 5.9% 7.4% 8.4% 9.2%

Other services, except government 6.5% 7.0% 5.6% 4.9% 4.8% 4.9%

Government 18.9% 20.2% 21.2% 18.7% 17.7% 17.1%

Federal 10.5% 9.2% 7.1% 6.0% 4.5% 3.6%

State and local 8.4% 11.1% 14.0% 12.7% 13.2% 13.5%

Addenda:

Private goods-producing industries 37.1% 32.8% 27.5% 23.8% 19.7% 16.9%

Private services-producing industries 44.0% 46.9% 51.3% 57.5% 62.6% 66.0%

Source: BEA Gross Domestic Product by Industry Accounts

Release date: December 11, 2006

Exhibit A-1a

140 Draft Workforce Housing Literature Review

Billions of Dollars Billions of 2000 Dollars Price Index

1955 1995 2005 1955 1995 2005 1955 1995 2005

GROSS DOMESTIC PRODUCT 415 7,398 12,456 2,213 8,032 11,049 18.71 92.12 112.74

Personal consumption 259 4,976 8,742 1,386 5,434 7,841 18.68 91.58 111.49

Gross private investment 69 1,144 2,057 256 1,134 1,866 26.65 100.94 110.28

Fixed investment 64 1,113 2,036 251 1,110 1,842 25.49 100.29 110.54

Nonresidential 39 810 1,266 129 763 1,224 30.41 106.24 103.43

Structures 15 207 339 107 247 252 14.21 83.88 134.65

Equipment and software 24 603 927 55 523 985 43.17 115.22 94.13

Residential 25 303 770 155 353 608 16.17 85.77 126.71

Net exports - goods+services 1 -91 -717 ..... -71 -619 ..... ..... .....

Exports 18 812 1,303 64 778 1,196 27.67 104.38 108.95

Imports 17 904 2,020 78 849 1,815 22.00 106.41 111.27

Gov't consumption/investment 87 1,369 2,373 641 1,550 1,958 13.50 88.36 121.18

Federal 55 519 878 379 580 728 14.48 89.50 120.73

National defense 47 349 589 329 389 484 14.29 89.60 121.86

Nondefense 8 171 289 54 191 244 14.67 89.35 118.61

State and local 32 850 1,494 251 968 1,230 12.60 87.78 121.46

Goods 205 2,661 3,887 662 2,639 3,881 30.79 100.86 100.16

Services 158 4,098 7,220 1,202 4,655 6,129 13.17 88.05 117.81

Structures 52 638 1,349 352 754 1,048 14.71 84.69 128.72

Share of GDP:

Personal consumption 62.4% 67.3% 70.2% 62.6% 67.7% 71.0%

Gross private investment 16.6% 15.5% 16.5% 11.6% 14.1% 16.9%

Fixed investment 15.4% 15.0% 16.3% 11.4% 13.8% 16.7%

Nonresidential 9.4% 10.9% 10.2% 5.8% 9.5% 11.1%

Structures 3.7% 2.8% 2.7% 4.8% 3.1% 2.3%

Equipment and software 5.8% 8.1% 7.4% 2.5% 6.5% 8.9%

Residential 6.0% 4.1% 6.2% 7.0% 4.4% 5.5%

Net exports - goods+services 0.1% -1.2% -5.8% -0.9% -5.6%

Exports 4.3% 11.0% 10.5% 2.9% 9.7% 10.8%

Imports 4.1% 12.2% 16.2% 3.5% 10.6% 16.4%

Gov't consumption/investment 20.9% 18.5% 19.0% 29.0% 19.3% 17.7%

Federal 13.2% 7.0% 7.1% 17.1% 7.2% 6.6%

National defense 11.3% 4.7% 4.7% 14.9% 4.8% 4.4%

Nondefense 1.9% 2.3% 2.3% 2.4% 2.4% 2.2%

State and local 7.6% 11.5% 12.0% 11.3% 12.1% 11.1%

Goods 49.3% 36.0% 31.2% 29.9% 32.9% 35.1%

Services 38.2% 55.4% 58.0% 54.3% 58.0% 55.5%

Structures 12.5% 8.6% 10.8% 15.9% 9.4% 9.5%

Source: BEA Release of 1/31/07

Table A-1b

GDP Components

Workforce Housing Final Report 141

$Billions Percent of Value Added Employment Per Full-time Equiv

Gross

Output

Value

Added

Employ

Compen-

sation

Wage &

Salary

Taxes -

Subs

Gross

Oper.

Surplus FT+PT

Full-

time

Equiv

Value

Added

Employee

Compen-

sation

Gross domestic product/All Ind 22,857 12,456 56.5 45.5 6.9 36.6 141,218 126,865 $98,182 $55,465

Private ind 20,256 10,892 52.3 43.1 8.1 39.6 117,090 106,871 $101,919 $53,289

Agriculture, forest, fishi& hunt 312 123 34.0 30.2 -11.3 77.3 1,473 1,279 $96,247 $32,709

Farms 253 96 25.3 22.3 -15.7 90.3 779 668 $143,500 $36,365

Forestry, fishing, & rel actvt 59 27 64.4 58.0 4.0 31.6 693 611 $44,586 $28,712

Mining 396 233 21.6 17.4 8.2 70.2 564 557 $418,905 $90,429

Oil & gas extraction 248 160 11.9 9.3 8.6 79.5 128 126 $1,266,484 $150,119

Mining, except oil & gas 64 32 48.1 38.1 13.4 38.6 215 212 $148,675 $71,476

Support actvt for mining 83 42 38.6 32.6 3.0 58.4 222 219 $192,845 $74,434

Utilities 410 248 22.2 16.8 16.6 61.2 554 545 $455,031 $101,136

Construction 1,175 611 64.0 52.2 1.3 34.7 7,567 7,315 $83,543 $53,471

Manufacturing 4,502 1,513 61.7 46.6 3.4 35.0 14,328 14,044 $107,698 $66,414

Durable goods 2,364 854 71.7 54.4 2.1 26.2 9,003 8,864 $96,377 $69,103

Wood prods 105 39 63.1 50.8 1.5 35.4 579 564 $69,078 $43,598

Nonmetallic mineral prods 112 53 52.3 41.7 2.2 45.4 508 496 $107,488 $56,264

Primary metals 194 61 52.4 39.6 3.5 44.1 465 459 $133,026 $69,763

Fabricated metal 271 131 62.9 49.6 1.6 35.5 1,525 1,504 $86,769 $54,576

Machinery 287 111 70.5 54.8 1.7 27.8 1,166 1,148 $96,761 $68,181

Computer & electronic 381 135 95.7 74.7 2.9 1.4 1,311 1,296 $104,404 $99,891

Electr equip, appliances 109 48 62.9 43.2 2.3 34.7 436 429 $111,476 $70,156

Motor vehicles & parts 483 95 89.8 62.1 2.7 7.5 1,100 1,093 $87,298 $78,425

Other transportation equip 192 71 78.1 60.4 1.5 20.4 673 669 $106,278 $82,999

Furniture & rel prods 85 37 74.2 51.0 1.0 24.9 569 556 $66,696 $49,477

Miscellaneous manf 145 73 54.2 41.3 1.1 44.7 670 651 $111,582 $60,447

Nondurable goods 2,138 658 48.6 36.5 5.1 46.3 5,324 5,180 $127,069 $61,812

Food, beverage & tobacco 659 176 47.1 35.8 11.7 41.2 1,687 1,627 $107,974 $50,819

Textile mills & textile prods 69 24 67.3 53.0 2.9 29.9 389 376 $63,420 $42,670

Apparel & leather 36 17 77.7 55.9 2.1 20.1 312 301 $55,934 $43,475

Paper prods 155 55 62.4 45.5 3.3 34.3 484 469 $116,330 $72,544

Printing & related 90 47 73.9 57.5 1.4 24.7 664 644 $72,884 $53,851

Petroleum & coal prods 398 63 20.9 14.2 3.1 76.1 111 109 $582,220 $121,413

Chemical prods 539 209 41.2 29.8 2.6 56.2 876 862 $242,734 $99,934

Plastics & rubber prods 193 68 59.5 47.3 2.7 37.8 802 791 $85,549 $50,942

Wholesale trade 1,074 743 52.3 43.5 22.0 25.6 5,850 5,652 $131,492 $68,832

Retail trade 1,289 824 56.8 47.8 21.5 21.7 15,763 13,723 $60,011 $34,096

Transport & warehouse 712 345 65.2 51.8 5.4 29.3 4,379 4,164 $82,766 $53,990

Air transportation 135 41 86.2 64.8 13.1 0.8 500 475 $86,303 $74,368

Rail transportation 58 32 54.9 39.9 -2.4 47.5 198 188 $171,644 $94,250

Water transportation 36 9 49.8 39.1 1.3 48.9 60 57 $158,474 $78,965

Truck transportation 251 114 61.1 48.4 2.2 36.7 1,420 1,350 $84,540 $51,683

Transit & ground pass 29 17 71.7 58.7 4.0 24.3 417 397 $43,116 $30,909

Pipeline transportation 39 9 42.9 34.8 27.4 29.7 38 36 $259,000 $111,222

Other transportation 121 89 62.5 51.7 9.1 28.4 1,159 1,102 $80,827 $50,517

Warehousing & storage 44 33 78.1 64.7 0.7 21.2 586 557 $58,704 $45,874

Information 1,161 555 43.2 34.9 7.6 49.1 3,079 2,866 $193,724 $83,761

Publishing/software 268 150 50.0 39.6 1.5 48.5 939 849 $176,966 $88,511

Motion picture & sound 87 41 57.8 49.4 2.8 39.3 382 323 $125,430 $72,542

Broadcasting & telecom 688 304 34.4 27.6 12.5 53.2 1,323 1,292 $235,362 $80,865

Info & data processing 118 60 61.3 50.2 1.6 37.1 436 401 $150,539 $92,279

Table A-2

GDP by Industry 2005

142 Draft Workforce Housing Literature Review

$Billions Percent of Value Added Employment Per Full-time Equiv

Gross

Output

Value

Added

Employ

Compen-

sation

Wage &

Salary

Taxes -

Subs

Gross

Oper.

Surplus FT+PT

Full-

time

Equiv

Value

Added

Employee

Compen-

sation

Gross domestic product/All Ind 22,857 12,456 56.5 45.5 6.9 36.6 141,218 126,865 $98,182 $55,465

Finance, insur, real estate 3,991 2,536 25.2 21.1 10.4 64.4 8,308 7,872 $322,163 $81,330

Finance & insurance 1,690 958 56.2 46.7 4.5 39.4 6,101 5,861 $163,400 $91,796

Credit intermediation 683 475 41.4 34.2 3.5 55.1 2,899 2,783 $170,576 $70,579

Securities & investments 321 167 94.6 81.2 2.8 2.5 822 789 $212,152 $200,801

Insurance 593 296 58.0 47.7 6.9 35.2 2,291 2,203 $134,425 $77,962

Funds, trusts, & other 94 19 58.7 39.6 4.8 36.5 89 86 $226,163 $132,698

Real estate & leasing 2,301 1,578 6.5 5.5 13.9 79.6 2,207 2,011 $784,872 $50,827

Real estate 2,053 1,473 5.1 4.3 14.4 80.5 1,535 1,410 $1,044,367 $52,966

Rental & leasing svcs 248 106 26.0 22.0 7.3 66.7 673 601 $176,073 $45,809

Professionl & business svc 2,318 1,459 69.9 59.4 1.9 28.2 17,384 16,257 $89,732 $62,725

Professional & technical 1,359 864 66.6 56.5 1.7 31.7 7,497 7,068 $122,257 $81,448

Legal svcs 245 181 59.4 51.1 3.1 37.5 1,331 1,255 $144,112 $85,578

Computer systems 180 141 83.3 68.5 1.8 14.9 1,201 1,132 $124,343 $103,604

Miscellaneous prof & tech 934 542 64.7 55.1 1.2 34.1 4,964 4,680 $115,918 $74,998

Management of enterprises 368 226 78.2 66.0 1.6 20.1 1,748 1,724 $130,990 $102,468

Administrative & waste mgt 591 369 72.5 62.1 2.5 25.0 8,139 7,465 $49,408 $35,819

Administrative & support 525 337 74.0 63.4 2.0 24.0 7,800 7,140 $47,136 $34,864

Waste management 66 32 57.2 48.5 7.6 35.2 339 325 $99,302 $56,785

Educ, health, & social asst 1,578 975 78.5 66.7 1.2 20.3 17,932 16,148 $60,400 $47,439

Educational svcs 192 116 88.7 77.0 1.2 10.1 2,911 2,582 $44,841 $39,770

Health care & social assist. 1,386 860 77.2 65.3 1.2 21.7 15,021 13,566 $63,362 $48,898

Ambulatory health care 649 442 67.7 56.7 0.9 31.3 5,245 4,722 $93,592 $63,394

Hospitals & res care 616 342 88.6 75.4 1.5 9.9 7,181 6,606 $51,801 $45,900

Social assistance 121 75 80.6 69.8 0.9 18.5 2,595 2,238 $33,702 $27,163

Arts, ent, rec, accom, & food 815 445 62.3 53.2 11.6 26.1 13,008 10,613 $41,895 $26,080

Arts, ent, & rec 183 114 59.3 50.3 10.5 30.3 1,981 1,655 $68,938 $40,853

Perf arts & rel actvt 82 54 60.2 50.8 7.5 32.3 500 418 $129,136 $77,677

Amusements, gambling 101 60 58.5 49.9 13.1 28.4 1,481 1,237 $48,597 $28,411

Accom & food svcs 633 331 63.3 54.2 12.0 24.7 11,027 8,958 $36,899 $23,351

Accommodations 171 105 53.3 45.1 14.8 31.9 1,837 1,684 $62,140 $33,123

Food & drink 462 226 67.9 58.3 10.8 21.3 9,190 7,274 $31,055 $21,089

Other svcs, except gov't 522 283 70.5 61.5 7.2 22.4 6,901 5,839 $48,431 $34,127

Government 2,601 1,564 85.8 62.5 -0.9 15.1 24,128 19,994 $78,204 $67,099

Federal 872 499 81.1 52.3 0.0 18.9 5,091 4,091 $121,927 $98,917

General government 782 437 78.7 50.4 0.0 21.3 4,208 3,365 $129,774 $102,074

Government enterprises 90 62 98.5 65.8 0.0 1.5 883 726 $85,552 $84,285

State & local 1,729 1,065 88.0 67.3 -1.4 13.4 19,037 15,903 $66,957 $58,914

General government 1,532 986 88.4 67.6 0.0 11.6 17,977 14,870 $66,322 $58,659

Government enterprises 197 79 82.2 63.9 -18.9 36.6 1,060 1,033 $76,099 $62,585

Private goods-producing 6,385 2,480 57.1 44.4 2.6 40.3 23,932 23,195 $106,922 $61,051

Private svcs-producing 13,871 8,412 50.9 42.7 9.7 39.4 93,158 83,679 $100,529 $51,136

Table A-2

GDP by Industry 2005

W o

rk fo

rc e

H o

u s in

g F

in a l R

e p o rt

1

4 3

1 4

4

W

o rk

fo rc

e H

o u

s in

g F

in a l R

e p o rt

W o

rk fo

rc e

H o

u s in

g F

in a l R

e p o rt

1

4 5

1 4

6

W

o rk

fo rc

e H

o u

s in

g F

in a l R

e p o rt

Workforce Housing Final Report 147

148 Workforce Housing Final Report