Association for Public Policy Analysis & Management

New York State has the most racially segregated school system in the country, and New York City remains one of the nation’s most racially and socioeconomically segregated school districts (Kucsera & Orfield, 2014). Given the persistence of residential segregation, school choice has been proposed as one remedy for integrating schools by race and family income. In this paper, we examine patterns of school segregation by race, gender, socioeconomic background, and academic achievement over a period of expanding school choice in NYC: 2005-2016. To assess the extent of segregation attributable to choice—that is, above and beyond what would be predicted by residential sorting alone—we compute segregation indices under two allocations of students to schools: their actual (observed) school, and a counterfactual allocation in which all students are assumed to attend their closest school via walking or public transport (e.g., Allen, 2007; Bifulco, Ladd, & Ross, 2009). We address capacity constraints by allocating only as many students to a school as there are seats; when a school is full, we allocate students to the next-closest school.

We find that school choice marginally increases racial segregation relative to a “closest school” assignment for most groups and grade levels. For example, black and Latino students have 5-10 percent lower exposure to white students under choice relative to their closest school assignment. In contrast, white exposure to Asian students increases under school choice, especially in higher grades. We find substantial increases in segregation by free and reduced-price lunch status under school choice, particularly in middle school. The biggest effect of choice on segregation by far appears to be on segregation by academic ability. Students in the bottom quartile of math and reading are substantially segregated from top quartile students in NYC, with segregation increasing in the higher grades. This is driven in part by the extensive use of academic screening by NYC middle and high schools. We observe only small changes in school segregation over time, although there is evidence that segregation by academic ability has increased as school choice opportunities have grown.

Finally, the fact that racial and ethnic groups appear to be no more integrated in practice than under a “closest school” assignment is not explained by students attending close to home, reinforcing residential segregation. On the contrary, a large fraction of NYC students attends a school that is not their closest school. This is increasingly true in older grades, where (for example) fewer than 20 percent of all 9^th graders attend one of their three closest high schools. Of all racial groups, black students are the least likely to attend their closest school, and are most likely to attend a school where they are the only student from their Census tract. Black students in NYC frequently travel long distances to schools that are as segregated by race and ethnicity as their “neighborhood” school.