Using a representative longitudinal survey of U.S. teenagers, we investigate how peer racial composition in high school affects individual turnout of young adults. We exploit across-cohort, within-school differences in peer racial composition. One within-school standard deviation increase in the racial diversity index leads to a 2.3 percent increase in the probability to be registered to vote seven years later and to a 2.6 percent higher probability to vote six years later. These effects are likely due to positive interracial contact when socialization has long-lasting effects: higher racial diversity in school is linked to more interracial friendships in school and later on.