Editors' Summary

Brookings-Wharton Papers on Urban Affairs presents new research on urban economics to a broad audience of interested policy analysts and researchers. The papers and comments contained in this volume, the tenth in the series, were presented at a conference on November 13–14, 2008, at the Brookings Institution. The papers examine a range of issues that are relevant to urban economics, including the effects of job location in an urban area on residential choice patterns, racial bias in mortgage lending, and the effects of urban characteristics on the development of new patents. The volume also contains three papers on urban developments outside of the United States. The topics treated include urban sprawl in Europe, rural-to-urban migration patterns in Brazil, and locational patterns of establishments across Japanese cities.

After World War II, a growing percentage of Americans moved to the nation's suburbs, and a shrinking percentage chose to live in central cities. This shift in residential patterns occurred at the same time as a shift in the location of jobs. Compared with job locations in the early postwar period, a smaller share of U.S. employment is now concentrated in central cities and a bigger share is located in the suburbs. For regional planners and urban economists, this raises an important question: Have residents of metropolitan areas flocked to the suburbs because that is where the jobs are? Or have employers followed urban migrants out into the suburbs?

In their paper "Job Decentralization and Residential Location," Leah Platt Boustan and Robert A. Margo offer a partial answer to this question using information on the residential choices of state government employees who hold jobs inside and outside of state capitals. The location of state capitals was chosen many decades ago, and no state has chosen to relocate its capital since the early twentieth century. State government employment in capital cities tends to be concentrated near the historical heart of cities, usually in or near the central business district. The persistence of state employment patterns in state capitals gives analysts evidence that can be helpful in determining the effect of the geographical distribution of employment on employees' residential choices.

There are two kinds of contrast that are useful in determining the impact of job location on residential choice. First, state government employees who work in capital cities can be compared with employees who work for other kinds of establishments in the same metropolitan areas. Since state government employees in those cities are much more likely to work in the central city, the difference between their residential choices and those of employees of other kinds of establishments sheds light on the impact of job location on the decision of workers to live inside or outside of the central cities. Second, state government employees who work outside of capital cities are considerably less likely to work in the central city of a metropolitan area. The distribution of their employment inside and outside of central cities is similar to that of employees of other kinds of establishments that are located in the same metropolitan areas. Boustan and Margo combine information for these two sets of contrasts to determine how the geographical concentration of state government employment in capital cities affects the choice of employees to live inside or outside of the central city.

The authors use individual-level information from the 1980 decennial census to assess the impact of job location on residential choice. They find that state government employees in capital cities are about 13 percent more likely than state employees in other locations to work in the central city. At the same time, the state employees who work in capital cities are 3 to 4 percentage points more likely to reside in the central city. From this the authors infer that the increased concentration of employment in the central city has a nontrivial effect on the residential choices of state employees. They estimate that the addition of 100 new employment positions in the central city would add 25 new working residents to the population of the central city. The authors attempt to verify their findings by performing some sensitivity...