Abstract
This paper studies the effects of a government scholarship program for low-income college students in Brazil, the Prouni. In order to deal with selection effects, I use propensity score matching based on observable student characteristics and a proxy for previous student performance. The results are robust across different specifications, and suggest that students who received a scholarship perform better than comparable students and take less time to reach the final year of college. These effects are higher for students with full scholarships than for students with partial scholarships, and seem to be partially driven by a decrease in the proportion of students who work and an increase in time spent studying.