2005b. “Separate When Equal? Racial Inequality and Residential Segregation.” Unpublished paper

Standard intuition suggests that residential segregation in the United States should decline when racial inequality narrows. In this paper, we hypothesize that the opposite will occur, due to a process of neighborhood formation. Increases in the proportion of highly educated blacks in a metropolitan area allow new middle-class black neighborhoods to emerge, and these are attractive to blacks, especially those who are highly-educated, resulting in increases in residential segregation as households re-sort. Our primary empirical analysis, using U.S. Census data, supports this hypothesis: increased black educational attainment in a metropolitan area between 1990 and 2000 leads to a large rise in the number of middle-class black communities, expanding the neighborhood choice set. In turn, such a narrowing of racial inequality significantly increases neighborhood segregation on the basis of race, contrary to what one would expect if statistical discrimination or correlations of race with socioeconomic status were driving overall segregation patterns. The within-metropolitan area sorting mechanism we identify competes in cross-section with a reverse channel of causation, whereby the educational attainment of young black men is lower in more segregated cities: our mechanism dominates for older black men and in first differences. We also show that across-metropolitan area sorting cannot account for the significant effects we find. Of broader relevance, our analysis draws attention to a negative feedback loo