Racial Fractionalization and School Performance

Abstract

I Introduction Parents and policymakers have long been concerned about racial diversity's affect on school performance. Racial diversity's impact on African-American students' academic achievement was at the heart of the 1954 U.S. Supreme Court decision in Brown v. Board of Education (Armor 1995). White flight, where white families move to segregated communities or enroll their children in private schools, was another historical manifestation of this concern (Fairlie and Resch 2002). Most recently, policy debates surrounding ability tracking, charter schooling, open enrollment, and school vouchers have also reflected concern about how these policies will affect school racial composition and academic outcomes (Greene 1999; Renzulli and Evans 2005). In light of this concern a body of research examines the impact of racial on school performance. Following the Brown v. Board of Education decision, early social science research focused on racial isolation's harmful effects on African-Americans' academic performance. Two influential reports were Coleman et al. (1966) and the U.S. Commission on Civil Rights (1967), which found that African-American students' racial isolation in segregated schools lowered their academic achievement. Jencks et al. (1972) found that desegregation improved black children's school performance by 2 to 3 percent; and Guryan (2004) estimates that half of the decline in black dropout rates during the 1970s occurred because of desegregation. In a related line of research Hoxby (2000) finds that black third graders perform substantially worse when surrounded by other black students than when they are in classes that are primarily white. Hanushek, Kain, and Rivkin (2009) isolate the peer composition of racial diversity and find similar results, namely, that having a higher percentage of black classmates lowers black academic achievement. Positive racial peer effects are the conventional channel through which racial diversity is thought to affect school performance. However, recent research in economic development suggests that diversity might negatively affect education and education-related outcomes. Several papers in this literature find that greater diversity is associated with worse political-economic outcomes (see, for instance, Easterly and Levine 1997; La Porta et al. 1999; Zak and Knack 2001). (1) This includes those related to schooling. For example, Easterly and Levine (1997) find a negative relationship between the degree of ethnolinguistic fractionalization and the number of years of schooling in a country. Similarly, Alesina, Baqir, and Easterly (1999) examine U.S. school districts and find that racial diversity lowers school spending. This literature points to policy choices--in particular goods spending--as the channel through which diversity negatively affects education. According to this reasoning, diverse citizens have diverse and often inconsistent needs. They therefore find it more difficult to agree on the level and kinds of goods that government should provide. This disagreement in turn leads important goods, such as education, to be underprovided, if underprovided education means lower educational achievement, this work suggests a which we call the public goods channel, through which higher racial diversity could lead to lower school performance. This article empirically tests the goods channel using data on Ohio school districts. We find that moving from a completely homogenous school district to one in which two racial groups have equal population shares is associated with a 7-17.5 percentage point decline in the passage rate on the state math exam. While our results suggest that racial diversity is negatively associated with school performance, we find that that the goods channel is not the reason for this relationship. Compared to between countries, there is minimal institutional or policy variation between Ohio's school districts. …