We use regression discontinuity and regression kink designs to estimate the impact of need-based grant aid on the borrowing and educational attainment of students enrolled in a large public university system. Pell Grant aid substantially reduces borrowing: among students who would borrow in the...