RACIAL SORTING AND NEIGHBORHOOD QUALITY*

Predominantly black, high-amenity neighborhoods are scarce in most US cities, implying that neighborhood race and neighborhood quality are often explicitly bundled – to choose high-amenity neighborhoods, households must typically live with a higher fraction of white neighbors. In the presence of any form of segregating preferences, this bundling raises the utility cost of neighborhood amenities for blacks relative to whites. As a result, we hypothesize that racial sorting (whether driven by decentralized preferences or discrimination) significantly accentuates racial differences in neighborhood quality. To shed light on this hypothesis, we use detailed Census microdata to estimate an equilibrium sorting model developed in Bayer, McMillan, and Rueben (2005) and carry out two counterfactual simulations. In the first, eliminating ‘preferences’ for neighborhood race substantially reduces the racial gap in a wide range of neighborhood amenities – indeed implying that racial sorting provides almost as important an explanation for the observed gaps as underlying socioeconomic differences across race. In the second, increasing the proportion of blacks in the population results in the increased availability of predominantly black, high-amenity neighborhoods; this relaxes the bundling constraint and leads to a substantia