Jean-William Laliberté (University of Calgary)

Abstract: Race-based affirmative action policies are widespread in higher education. Despite the prevalence of these policies, there is limited evidence on whether they affect students before they reach college. We exploit the 2003 Supreme Court ruling in Grutter v. Bollinger, which overturned affirmative action bans in Texas, Louisiana, and Mississippi, but not in other states, to study the effect of affirmative action on high school students’ outcomes. We analyze four data sets, including nationwide SAT data and ad- ministrative data for the state of Texas. The SAT data allow us to leverage state and time variation in difference-in-differences and synthetic control group analyses. Within Texas, variation in race, time, and ex ante ability further help us to isolate the effects of the policy change on secondary school grades, attendance, and college applications. Across data sets, outcomes, and identification strategies, the results all point toward gains for underrepresented minority students and reductions in the racial achievement gap. These gains were concentrated among students in the top of the ability distribution, who also experienced the largest increases in the returns to pre-college human capital in college admissions due to the policy change. This suggests that students increased their human capital investment in response to increases in the returns to effort.