A model of spatial mismatch: Evidence from U.S. cities in 1980 and 1990

The idea that restricted residential choices and the flight of suitable employment from city centers have significantly reduced the employment and earnings of low-skilled central city residents was introduced by Kain (1968). Now referred to as the spatial mismatch hypothesis, it has generated a healthy stock of mostly empirical studies over the past 30 years. This dissertation develops a rigorous economic model of spatial mismatch, anchored firmly to the familiar basic model of residential choice in the tradition of Alonso and Mills, and tests its empirical implications for the employment, wages, and labor force participation of low-skilled urban residents with data from the 1980 and 1990 Census Public Use Microdata Sample.;This research uses an equilibrium land use model of a closed city with employment opportunities in both central and suburban locations. Low-skilled workers are limited to residential locations adjacent to the city center. Those entering the labor market offer their services to either central or suburban employers after weighing relative wages and transportation costs. In a spatial mismatch equilibrium, low-skilled residents who choose to sell their labor to employers in the central business district will populate that part of the central city residential zone immediately surrounding that job site. Meanwhile, those low-skilled residents who sell their labor to employers in the suburbs will, since they are prohibited from relocating to the suburbs, populate the outer part of the central city zone and commute through the suburbs to the job site.;The model is first used to test the relevance of spatial mismatch to the employment probability of low-skilled, male residents seeking central employment. Assuming that wages are not fully flexible (as in the presence of a wage floor), any shocks to the central city labor market must be absorbed, at least partially, through changes in employment. Determinants of employment probability are tested through maximum likelihood estimation of a reduced form solution to the theoretical model of the urban labor market. The results support the spatial mismatch hypothesis by indicating that the concentration of low-skill residents and the stock of capital in the city center, as well as commuting costs, are significantly related to employment probability.;Under the assumption that wages in the central city labor-market are flexible, a similar reduced form is next used to identify the determinants of the wages of low-skilled centrally employed males using linear regression. Estimated over a sample of those employed full-time, the model provides almost no evidence that restricted accessibility is a significant factor in the determination of wages. Finally, to examine the possibility that a lack of suitable nearby employment makes a low-skilled male more likely to drop out of the labor force, a similar equation is used to estimate the probability of participation in the labor force. In both 1980 and 1990, the results of maximum likelihood estimation indicate limited support for spatial mismatch.;For each year, the sample was split to separately estimate the impact of accessibility on the employment probability, wages, and labor force participation probability of centralized, low-skilled populations of blacks and youth. Regardless of the sample, the models of wages and labor force participation provide little support for spatial mismatch. The model of employment probability suggests that spatial mismatch generally has a greater impact on youth and that blacks are slightly more affected than non-blacks.