Understanding the Income Gradient in College Attendance in Mexico: The Role of Heterogeneity in Expected Returns

Differences in college enrollment rates between poor and rich students are a prevalent phenomenon, but particularly striking in Latin America. The literature suggests explanations such as differences in "college preparedness" on the one hand, in that poor students lack skills that enable them to benefit from college, and "credit constraints" on the other hand. One explanation that has been neglected in this analysis consists of differences in information sets between the poor and the rich - for example about career opportunities-translating into different perceptions of individual returns to college. Data on people's subjective expectations of returns allow to take this factor into account and to directly address the following identification problem: conditional on their information sets poor people might expect low returns and thus decide not to attend. Or they might face high (unobserved) costs that prevent them from attending de-spite high expected returns. Conventional approaches rely on strong assumptions about people's information sets and about how they form expectations to address this identification problem. Data on people's subjective expectations of returns as well as on their schooling decisions allow me to directly estimate and compare cost distributions of poor and rich individuals. I find that poor individuals require significantly higher expected returns to be induced to attend college, implying that they face higher costs than individuals with wealthy parents. I then test predictions of a model of college attendance choice in the presence of credit constraints, using parental income and wealth as a proxy for the household's (unobserved) interest rate. I find that poor individuals with high expected returns are particularly responsive to changes in direct costs, which is consistent with credit constraints playing an important role. Evaluating potential welfare implications by applying the Local Instrumental Variables approach of Heckman and Vytlacil (2005) to my model, I find that a sizeable fraction of poor individuals would change their decision in response to a reduction in direct costs. Individuals at the margin have expected returns that are as high or higher than the individuals already attending college, suggesting that government policies such as fellowship programs could lead to large welfare gains.