Association for Public Policy Analysis & Management

Education is widely accepted as the most successful strategy to overcome poverty. Nevertheless, recent literature has identified substantial heterogeneity of the return of tertiary education. In this paper we use administrative records from different sources to construct a unique data set of low-income students in Colombia, which include cognitive test results and socioeconomic information at their high school final year, and information on their labor market results, many years after high school graduation. We evaluate the returns to college education by estimating the Marginal Treatment Effect (MTE) of the college investment decision. The MTE allows estimating a random parameter for the return to college, which varies with unobserved heterogeneity across workers. We find considerable heterogeneity in returns, to the extent that for a considerable mass of the population the return is negligible. Using the estimated models, we simulate two types of policies: one that increases the supply of tertiary education and another one that enhances the quality on secondary education, in a way that low-income students can increase their possibilities of enrolling to college. We find that a policy that expands college extensively has the same return of a policy that increases the average quality of high school institution in half of a standard deviation. This paper contributes to the literature by estimating the return of enhancing the quality of the secondary level of education, to the best of our knowledge a plausible causal effect of such a policy has not been reported in the literature.