Association for Public Policy Analysis & Management

School attendance boundaries are the most common student assignment mechanism in U.S. public schools. Even in an era of expanding school choice, most local school districts use student assignment plans at least partly based on student residential locations. I study how school districts set these policies, with the aim of characterizing local policymaker preferences for racially desegregated schools.

I define a school boundary desegregation policy index that evaluates existing school boundaries relative to counterfactual boundaries that imitate a minimum travel distance – "neighborhood schools" – plan. Using a novel database on 2013-14 school boundaries linked to census block demographic data, I find wide heterogeneity in boundary desegregation across districts, with the mean district having boundaries that approximately adhere to a neighborhood schools plan. The distribution is right skewed, implying that the minority of school boundaries that appear to be gerrymandered are more likely to encourage school integration than to exacerbate school segregation beyond residential segregation levels.

To explain this policy variation, I introduce a stylized model of the policymaker’s SAB drawing problem as the maximization of a heterogeneous utility function defined over school integration and aggregate daily student travel to school. Districts draw boundaries to maximize this utility subject to a trade-off between integration and travel that depends on local residential racial sorting patterns. The model suggests an estimation framework of desegregation demand as a function of local cost and preference parameters.

To measure the cost of desegregating schools by gerrymandering school boundaries, I develop a simple algorithm that estimates rate of transformation between aggregate travel and racial integration – the "price" of desegregation – conditional on residential sorting patterns. I estimate that districts facing lower desegregation costs tend to enact significantly more integrative SABs, suggesting there is demand for school integration across local school districts. After controlling for prices, I test for drivers of desegregation preference heterogeneity. I find that active court desegregation orders have a positive association with desegregation policy while active but not after they have been rescinded. This suggests that these court orders influence the way school boundaries are drawn, but that districts revert to neighborhood schools boundaries when judicial oversight is no longer in place. Importantly, I also estimate that racial intolerance of local whites is negatively correlated with integration efforts. These findings suggest that even today, racial animus continues to be an important barrier to school desegregation efforts on the part of local policymakers.