The Housing Choice Voucher program is currently the largest federally funded housing assistance program. Although the program aims to provide housing assistance, it also could affect children's educational outcomes by stabilizing their families, enabling them to move to better homes, neighborhoods, and schools, and increasing their disposable incomes. Using data from New York City, the nation's largest school district, we examine whether--and to what extent--housing vouchers improve educational outcomes for students whose families receive them. We match over 88,000 school-age voucher recipients to longitudinal public school records and estimate the impact of vouchers on academic performance through a comparison of students' performance on standardized tests after voucher receipt to their pre-voucher performance. We exploit the conditionally random timing of voucher receipt to estimate a causal model. Results indicate that students in voucher households perform 0.05 standard deviations better in both English Language Arts and Mathematics in the years after they receive a voucher. We see significant racial differences in impacts, with small or no gains for black students but significant gains for Hispanic, Asian, and white students. Impacts appear to be driven largely by reduced rent burdens, increased disposable income, or a greater sense of residential security.