Symposium on Decentralization and Urban Sprawl
The enormous decentralization of economic activity over the past several decades has been the central feature of metropolitan development during the last half-century. Some observers view this trend with alarm and argue that decentralization has led to a series of maladies in central cities, as well as to urban sprawl and related problems in the suburbs. Others view decentralization as a positive development, reflecting the preferences of Americans regarding residence, employment, and other factors. A full assessment of decentralization, however, requires facts documenting the extent and nature of the trend, analyses of the causes and consequences of decentralization, and a conceptual framework that integrates these factors. Accordingly, the volume begins with a symposium of three papers to explore these issues.
Decentralization: Basic Facts and Their Implications
In the traditional view, urban areas have a dense central business district with concentrated employment and production, in order to reduce costs of transportation and information sharing. In these so-called monocentric urban areas, land has the highest value in the city center because employment density is highest there. As distance from the city center increases, land and housing prices fall, lot sizes lengthen, workers’ commute times rise, and poverty declines.
This standard view has been increasingly challenged by decentralization. In their paper, economists Edward Glaeser of Harvard University and Brookings and Matthew Kahn of Tufts University document in new ways the basic facts of decentralization and analyze the implications for the economic structure of urban areas. They show that in 1940, only one of the ten largest cities had population density below 10,000 people per square mile. By 1990, seven of the ten largest cities had densities below 7,500. Although there is no formal definition of decentralization, urban areas today are highly decentralized: by 1996, only 24 percent of jobs in U. S. metropolitan areas were within three miles of a city center.
Despite the ubiquity of decentralization, there are substantial differences across urban areas. In New York, one of the few cities with concentrated employment patterns, over 45 percent of jobs are within three miles of the city center. In Los Angeles, which is known for its sprawling nature, 45 percent of employment is located within an 11-mile ring of the city center. Glaeser and Kahn also find that the dispersion of employment and population in Chicago, and in most other American metropolitan areas, bears a much stronger resemblance to the spatial patterns in Los Angeles than in New York.
After documenting these trends, the authors examine the causes. The residential preferences of workers appear to be the driving forces behind this move toward decentralization, according to Glaeser and Kahn. That is, firms have located in the suburbs in large part because that is where the workers choose to live. In addition, better access to trucking routes from suburban locations encourages manufacturing firms to locate in the suburbs. The trend toward decentralization appears to be somewhat less pronounced among cities that are more than 200 years old and somewhat more pronounced among those less than 60 years old. In addition, cities with younger suburbs are associated with more decentralization of employment. Forces offsetting decentralization include the advantage of urban areas in accelerating the communication of ideas. Idea-intensive or service-based industries (such as commercial banking) are more likely to locate in the central city, while manufacturing firms and firms that require more physical infrastructure are more likely to locate in the suburbs.
These results have striking implications. The decentralization of employment suggests that the classic stylized facts of urban economics may no longer be empirically valid. Indeed, in contrast to the standard model, Glaeser and Kahn find that cities with more decentralized employment are less likely to have housing prices fall and commute times rise as distance from the city center increases. Interestingly, the spatial patterns for these factors are still correlated with employment densities, but since employment is no longer concentrated at the city center, the spatial patterns do not follow those suggested by the traditional approach. Thus, Glaeser and Kahn present an important set of facts that disputes much traditional thinking on cities, and will challenge urban experts to develop innovative approaches for thinking about urban dynamics.
Systematic Thinking About Urban Sprawl
Extensive decentralization of cities and the resulting development of the urban fringe bring new users to roadways, reduce open space, and require cities or suburbs to extend their utility services farther. All of these factors raise concerns about urban sprawl, which has become an increasingly contentious political issue in recent years. Twelve states have enacted growth management programs, and 240 anti-sprawl initiatives appeared on November 1998 local and state ballots, with many additional proposals in November 2000.
In light of this heated public debate, University of Illinois economist Jan Brueckner provides a framework in which to analyze sprawl issues and the relative merits of alternative remedies. He defines urban sprawl as the excessive spatial growth of urban areas. The emphasis on whether growth is excessive is crucial, because the natural growth of urban areas due to increases in income or population or improvements in transportation efficiency should not be a cause for concern.
Brueckner’s paper suggests that under current policies, there are reasons to believe that too much new suburban development occurs because developers and new residents are not forced to pay the full costs they impose on others. He recommends that forcing developers and new residents to face these costs with taxes and fees is the best strategy for controlling growth. He highlights three such problems and discusses the prospects and pitfalls of the implied policy solutions.
First, each new commuter that moves to the suburbs adds to roadway congestion there and consequently imposes costs on others, including increasing the time and fuel costs needed to drive a given distance. Economists have often suggested forcing drivers to shoulder these costs by assessing congestion tolls. Brueckner points out that technological developments could make assessing such tolls simpler now than they have been previously, but notes a history of strong political resistance to such taxes.
Second, suburban development reduces the amount of open space around the perimeter of a city, which reduces the benefits that all area residents can obtain from such spaces. Brueckner discusses the possibility of using development taxes to correct this market failure, but warns that identifying the optimal level for such a tax would be extremely difficult in practice because it is nearly impossible to measure precisely the value people place on the amenity of undeveloped land outside the city.
Third, cities do not charge developers or new residents the full cost of extending infrastructure—such as water and sewer lines—to new developments. This encourages development beyond optimal levels. Impact fees for service extension could correct this pricing error.
Some metropolitan areas concerned about urban sprawl, such as Portland, Oregon, have tried another policy alternative: urban growth boundaries. These boundaries are set by governments and designate a ring beyond which urban development is prohibited or very strictly controlled. Brueckner argues that such policies are likely to be difficult to implement effectively and thus are likely to hurt urban residents, especially low-income residents. Moreover, the policies do not specifically target the market failures of road congestion, utility use, and open space. He warns that in some cases, a draconian urban growth boundary policy could be worse for society than no government intervention and its resulting sprawl.
Brueckner’s findings will not provide solace to staunch advocates on either side in the debate on urban sprawl. Rather, the paper highlights the importance of systematic thinking on the causes and consequences of sprawl and the necessity of matching both the character and level of the policy intervention to the particular problems that sprawl presents.
Employment Opportunities in Decentralized Metropolitan Areas
The exodus of jobs to the suburbs raises the possibility that people who continue to live in cities may find it more difficult to find jobs close to home. If that is the case, the ability to commute via public transportation or private automobile would become an increasingly important determinant of people’s ability to obtain and retain employment. The issue becomes especially interesting from a policy viewpoint, considering that urban residents are disproportionately minority and lower-income households, and that public transportation often provides poor service for those who live in the city and work in the suburbs.
University of California economists Steven Raphael and Michael Stoll examine the importance of car ownership for employment. They document anew the spatial mismatch between certain employment opportunities and available workers. The main focus of their paper is an empirical examination of how car ownership affects employment probabilities for the black and Latino population relative to the white population. White families are more likely to have cars; only 5 percent of white households do not own an automobile, compared to 24 percent of black households and 12 percent of Latino households. Among those who own cars, however, the probability of being employed is nearly constant across these groups. The authors find a positive relationship between owning a car and being employed, controlling for other factors. They also find that this relationship is stronger for more segregated minority groups. It is much stronger for blacks than for whites, and slightly stronger for Latinos than for whites. This effect is also stronger in cities in which black households are more geographically isolated from employment opportunities.
Raphael and Stoll then consider subsidies for minority car ownership. The authors show that raising car ownership rates for minorities to those for white households would be expected to significantly reduce the employment rate gap between these groups. They note that increased car use would worsen traffic-related externalities such as congestion and air quality, but they also suggest that many minorities with cars would be reverse commuters, traveling from homes in the central city to employment in the suburbs. Furthermore, they expect many minority workers to be employed in night shifts or at other times that would not conflict with rush hour traffic.
The other three papers presented at the conference provide new evidence on several current issues in urban economics and urban policy: the effects of housing relocation programs, the causes of the decline in food stamp participation, and the determinants of minority business ownership.
Do Housing Relocation Programs Help Children’s School Performance?
The combination of residential segregation by income and race and the suburbanization of high-income households raises many questions about the effects of neighborhood on residents and children. To what extent do peer groups influence individuals’ outcomes? Do children who grow up in high-poverty environments develop differently from other children? At what point do neighborhood effects become irreversible?
The experimental design of the Department of Housing and Urban Development’s Moving to Opportunity program, currently operating in five cities, allows researchers to study overall effects of neighborhood on children’s outcomes. The program randomly assigns public housing families that volunteer for the program to one of three groups: a group that is given Section 8 housing vouchers valid only in census tracts with a very low poverty rate; a group that is given Section 8 vouchers valid anywhere; and a group that is given no assistance in finding alternative housing.
The effect of this program on children’s educational achievements is the focus of the conference paper by economists Jens Ludwig of Georgetown University, Helen Ladd of Duke University, and Greg Duncan of Northwestern University. (Earlier research on the issue by Duncan and Ludwig was published as a Brookings Institution Children’s Roundtable Policy Brief in July 2000.) In the conference paper, the authors find that children aged five through twelve who moved to low-poverty neighborhoods demonstrated substantial improvement on standardized tests in both reading and math. For teenagers who moved to low-poverty neighborhoods, only the reading tests were available, and those scores did not differ from non-movers. However, teens who transferred to lower poverty areas were more often retained in grade and more often subject to disciplinary action than their peers who did not move. The authors suggest that differences in school standards, rather than actual declines in teens’ behavior or abilities, may account for these apparently negative implications.
The authors believe their findings are encouraging, but are only one component of a larger debate about whether expansion of relocation policies would be desirable. Which aspects of the new neighborhoods or schools are leading to the positive effects remains unclear. In some cases, reform of schools in the low-income neighborhoods might be as effective as relocation. Also, introducing students from low-income neighborhoods into low-poverty areas might have negative impacts on outcomes of host children. The best policy options for taking advantage of neighborhood effects are uncertain. Nevertheless, the findings provide important new evidence that should be part of any discussion of the costs and benefits of relocation programs.
Why Did Food Stamp Program Participation Decline?
The Food Stamp Program provides a monthly allotment of coupons or credit for low-income households to purchase food at participating stores, and is currently one of the government’s largest transfer programs. However, program participation rates have fallen dramatically in the last several years, from 27 million persons in 1994 to 20 million in 1998, with much of the decline occurring after 1996 and most of it occurring in urban areas. The conference paper by UCLA economists Janet Currie and Jeffrey Grogger examines the causes of the decline in Food Stamp Program enrollment and the implications for public policy and urban areas.
Three alternative (and not mutually exclusive) hypotheses have been put forth to explain falling enrollment: welfare reform, the strength of the economy, and changes in stigma and transaction costs attached to participation in the Food Stamp Program.
The 1996 welfare reform law allowed states to penalize food stamp participants for failure to meet work requirements, which may have substantially reduced the size of the eligible population. Currie and Grogger found that, overall, welfare reform was responsible for almost a third of the decrease. This effect was especially strong among single-head households.
The strong economy and corresponding lower unemployment rates explained about one-fifth of the overall decline in food stamp participation between 1993 and 1998.
Both of these factors, along with changes in program rules, may have increased the transaction costs and stigma associated with food stamp participation. More frequent recertification requirements under the new rules impose higher costs on households wishing to sign up for the program. Stigma may be more of a factor in states that still provide food stamp coupons, as compared with the roughly two dozen states using credit card-style electronic balance transfer methods. (All states are required to shift to electronic transfer methods by October 2002. Currie and Grogger’s results indicate that transaction costs are especially problematic for single parents and rural households, while stigma has more of an effect for married households without children.
The decline in food stamp use is particularly important for urban areas, given the concentration of the poor in cities. Much of the fall in participation rates, in fact, occurred in cities. This is especially true of the decline due to welfare reform. Currie and Grogger found that implementation of the Temporary Assistance for Needy Families (TANF) program had almost no impact on program rolls in rural areas, but was responsible for about 40 percent of the decline in food stamp enrollment in central cities. In comparison, decreases in unemployment were responsible for only 18 percent of the decline in central city locations. TANF implementation seems to play a greater role in the reduction of Food Stamp Program participation in cities while falling unemployment has less impact, suggesting needy families in central cities are experiencing more difficulty in obtaining food stamps, rather than experiencing reduced need. Under these circumstances, policy reforms intended to restore benefits to households in need should target families in central cities, especially single-head households.
Why Do Minority Self-Employment Rates Vary Across Metropolitan Areas?
The promotion of minority business ownership has been a major thrust of federal and local government policies for some time. The much lower rate of self-employment among minorities contributes to differences in employment rates between minority and white populations, and may well restrict the abilities of minority groups to accumulate wealth. About 11 percent of the overall population, but only 4 percent of the black population, is self-employed. Whites are self-employed at above-average rates of 13 percent, with Asians at 11 percent and Latinos in the middle at about 7 percent. Average earnings from self-employment are 93 percent higher for whites than for blacks. The variation in self-employment rates across metropolitan areas is also higher for all three minority groups than for whites.
In their conference paper, Syracuse University economists Dan Black, Douglas Holtz-Eakin, and Stuart Rosenthal show that characteristics, including age, immigrant status, and education, are important determinants of whether a particular person is self-employed, but they do not vary across metropolitan areas in ways that can explain variation in self-employment rates and earnings in different cities.
The authors also examine city-level variables and show that the degree of segregation of the minority population within a given metropolitan area does not have a significant impact on minority self-employment rates. What does make a difference is the purchasing power of minority groups in a city. In cities with wealthier minority populations, higher rates of minority self-employment are observed, suggesting that the economic power of the minority community is important to sustaining minority entrepreneurs. There are several possible channels through which a minority group’s economic clout could influence the self-employment rate, including various consumer discrimination theories and improved access of minorities to lenders. Distinguishing among these alternatives will play an important role in determining the policy implications of the authors’ findings.
Taken together, the papers presented at the Brookings-Wharton conference show that significant progress can be made in disentangling economic issues that affect and are affected by urban areas. Each of the papers provides new facts or models that will be useful to policymakers, academics, students, and other interested groups and individuals.
As the problems and opportunities created by urban areas continue to evolve, the findings in these papers should aid in providing new policy solutions, developing ways to exploit opportunities for enhancing urban development and economic well-being, and setting a foundation for future analysis.