Are there neighborhood effects on young adult neighborhood attainment? Evidence from mixed-logit models

Abstract

Studies of racial residential attainment show an intergenerational transmission of racial contexts from youth to adulthood, but it is unclear why this transmission is so robust. It is possible that experiences in racial contexts during youth have lasting effects on neighborhood selection in adulthood, but evidence for this claim has come from research using statistical methods that suffer from problems of ecological dependence and conflation of other neighborhood characteristics. In this study, we address these limitations using mixed-logit models, a form of discrete choice analyses, allowing us to control for differences across metropolitan areas and for multiple characteristics of neighborhoods that may affect the selection of destination neighborhoods. Data for the analyses come from the National Educational Longitudinal Study, the 1990 and 2000 Censuses, and other sources. We find that most of the intergenerational process results from young adults moving to neighborhoods short distances from their origin ones, but the models also suggest a contextual effect of youth experiences in racial compositions on neighborhood selection. The latter finding indicates that policies promoting integration among youth can have long-lasting effects on residential attainment.