The Impact of School Racial Compositions on Neighborhood Racial Compositions: Evidence from School Redistricting

Abstract:     
I use administrative data surrounding extensive school redistricting in the Charlotte-Mecklenburg Public School District (CMS), North Carolina, to examine the impact of large changes in school racial compositions on neighborhood racial compositions over time. In the fall of 2001, CMS was ordered to dismantle the race-based student assignment plan that had been in effect for 30 years. Under this plan, school assignment zones were typically drawn to capture non-contiguous areas with vastly different racial compositions to achieve racial balance in schools. A district-wide public school choice plan was approved for the 2002-2003 school year, with school assignment zones dramatically redrawn to give each student a guaranteed seat at a school close to her residence, typically the closest (students could gain admission to other schools in the district through a lottery process). Approximately 50 percent of families were reassigned to different schools, causing large changes in school racial compositions across the district. These extensive shocks to school racial compositions make CMS an ideal place to identify the effect of school racial characteristics on neighborhood racial characteristics. I study the impact of changes in school racial compositions on neighborhood racial compositions, constructed using average student characteristics by neighborhood, from 2001, prior to the redistricting, to 2006. I show that, over the five-year period, the effect of an increase in the percent black of the assigned elementary school on the percent black of the neighborhood was positive and statistically significant. A 10 percentage point increase in the percent black of the assigned elementary school led to an approximately 1.2 percentage point increase (relative to a mean of 53.6 percent) in the percent black of the neighborhood five years after the redistricting. The impacts increased over time, consistent with predictions from a simple model of short-run dynamics in neighborhood racial compositions, where the change in neighborhood racial composition is a function of lagged school and neighborhood racial compositions. Because I have student-level administrative data from CMS, including exact residential locations for all enrolled students in each year, I also examine student-level attrition from and decisions to change residences within CMS to shed additional light on which racial groups drove the changes in neighborhood racial compositions. These results indicate that the estimated responses of neighborhood racial compositions to changes in assigned school racial compositions (at least for the longer time horizons) were consistent with non-black residents moving from neighborhoods with increases in the percent black of the assigned elementary school to neighborhoods with decreases in the percent black of the assigned elementary school, while black residents did not change residential locations in response to the school reassignments. My findings have implications for the potential effects of school racial desegregation policy changes on neighborhood racial compositions.

Number of Pages in PDF File: 43

JEL Classification: J1

working papers series