Recent court cases have ended court-order desegregation in school districts across the United States, and the end of these policies led to increases in segregation of students across schools (Clotfelter, Ladd, and Vigdor 2006, An and Gamoran 2009, Lutz 2011, Reardon et al 2012).  While the effects of the end of desegregation on schools is well studied, less is known about the potential effects of ending school desegregation policies on neighborhood composition.  This study seeks to investigate these potential effects and determine whether the end of court-ordered school desegregation led to increases in residential segregation of households across neighborhoods.  Understanding factors influencing residential segregation is extremely important, as recent research shows that intergenerational income mobility across geographic areas correlates highly with residential segregation patterns (Chetty et al. 2014).

The current analysis is unique in that it uses linked person-level records from the 2000 and 2010 censuses of the United States population.  Specifically, the study makes three contributions to the current literature.  First, using a “difference in difference” (DD) approach, it examines the effects of desegregation dismissals on district composition and residential segregation.  Next, the rich micro-data allow the study to examine heterogeneity in treatment effects along a variety of dimensions.  Last, the analysis examines differential micro-level mobility responses of households in dismissed school districts.

The results indicate no effect of dismissals on either residential composition or segregation.  These estimates are precise enough to reject even modest effects, and do not differ significantly between households with and without school-age children.    Moreover, results do not differ across income groups or region of the country.  Last, the analysis examines micro-level destinations of individuals out of dismissed districts.  These estimates are imprecise and the analysis cannot reject potentially substantial differential mobility responses as a result of dismissal, which leaves open the possibility that dismissals created differential mobility responses that were not detected in aggregate segregation statistics.

This study also makes use of plausibly exogenous variation generated by the end of desegregation policies to estimate how school racial concentrations affect residential composition. In particular, ongoing research investigates these effects by using the difference between school and neighborhood racial composition to instrument for observed changes in school composition after the end of desegregation policies.