Americans rarely marry outside their racial or class groups, a pattern with well-documented implications for inequality and intergenerational mobility. Limited exposure may explain the low rates of intergroup marriage. We instrument for exposure using variations in childhood neighborhoods based on whether other race and class groups have other children of the opposite sex of the same age. Exposure increased interracial marriage (height and parental height) but had no detectable effect on interracial marriage (White and Black). A spatial marriage model predicts that home separation-one of the many forms of exposure-accounts for more than one-third of marriages sorted by class but less than 5 percent by race.
The paper is from the latest AER, by Benjamin Goldman, Jamie Gracie, and Sonya R. Porter.