The returns to college choice: Loans, scholarships and labor outcomes

To estimate causal effects of college choice, we exploit eligibility rules for student loans in a regression discontinuity design. Loan programs induce students to pursue college degrees that are more expensive and prolonged relative to technical education. Although higher education is profitable, the marginal return of college is identical to that of technical education when students are about 30 years old. The college premium seems to increase over time, possibly offsetting the initial experience gap and covering cost differences under moderate discount rates. We study the effects of debt burden on college choice using a similar cutoff rule for scholarships.