This research tests the spatial mismatch hypothesis by comparing the e
mployment probabilities of central city versus suburban African Americ
an males in nine metropolitan areas. Treatment effects models are used
to control for the effects of both individual characteristics and res
idential. self-selection on the probability of employment, A positive
effect of suburban residential location on employment is found for the
residents of Dallas, Los Angeles, New York, and Washington, while no
effect is found for the residents of Cleveland, Houston, Jackson, Memp
his, and Newark. The general conclusion to be drawn is that the spatia
l mismatch effect is contingent on the particular characteristics of e
ach metropolitan area. Preliminary analysis shows that metropolitan ar
eas with a spatial mismatch effect are large in terms of total populat
ion and total land area, and have less efficient transportation system
s.