I combine a regression discontinuity design with rich data on academic and labor market outcomes for a large sample of Florida students to identify the returns to four-year college for students on the academic margin of college admission. In addition, I develop a theoretical model of college choice with and without credit constraints that allows for intuitive tests of the importance of credit constraints within this population. I find that students who obtain high school grades just above the threshold value for admissions eligibility at a large public university in Florida are much more likely to attend a four-year college and much less likely to attend a community college than students with grades just below the threshold. The earnings returns to a year of four-year college for affected students are 8.7 percent, nearly identical to returns to college for the population of Florida high school students. Consistent with the credit constraints hypothesis, poorer students who are more likely to be credit constrained work more while in college and realize higher post-college returns.