Credit constraints in higher education in a context of unobserved heterogeneity

•We study the credits constraint in higher education access in Chile.•We consider a context with unobserved heterogeneity.•We use several administrative datasets.•We find no credit constraints on enrollment of higher education.•However, we find constraints on graduation of higher education.

This article tests the existence of credit constraints on higher education access by estimating actual marginal returns in the context of unobserved heterogeneity. We estimate higher education returns for those who attended and compare them with those of individuals who are at the margin of attending. Following the (Carneiro and Heckman, 2002) reasoning, if the returns of the latter group are larger than those of the former one we could be in presence of unobservable barriers to higher education access, such as credit constraints. We use a rich administrative database composed of three sources: data of enrollment and graduation from the Chilean higher education system, test scores and labor market outcomes from the Chilean Unemployment Insurance database. Our results suggest that there is no evidence of credit constraints for enrolling into the Chilean Higher Education system. However, we do find some evidence of credit constraints in graduation.