We provide the first experimental evidence on the effect of student loans on educational attainment. Loan amounts listed in financial aid award letters (“offers”) do not alter students’ choice sets but significantly affect borrowing. Students randomly receiving a nonzero offer were 40 percent more likely to borrow than those who received a $0 offer. Per additional borrower, loans increased by $4,000, GPA and completed credits increased by 30 percent, and transfers to four-year public colleges increased by 11 percentage points. Cost-benefit and theoretical analyses suggest nonzero offers enhance welfare, yet over five million students are not currently offered loans. (JEL D14, D91, I22, I23)