Racial Segregation and Poverty by

This paper examines the relationship between racial segregation and poverty. Using data primarily from the 2000 Census, I control extensively for metropolitan area characteristics and other factors. In addition, I use metropolitan area physical geography characteristics as instruments for racial segregation as one approach to address the problem of endogeneity bias. I find significant and positive effects of black (Latino)-white segregation on black and Latino poverty, respectively, that remains evident across a variety of model specifications. Evidence from two-stage least squares estimates reveals that the effect on Latino poverty seems more plagued by endogeneity bias than that on black poverty. Simulations show that if black-white levels of segregation were reduced from hyper segregation (actual) levels to low levels of segregation the black-white poverty rate gap would close by fifty percent. The influence of segregation on poverty varies across important metropolitan characteristics such as region, metro size and percent of the population that minority.