Fields of study and labor market returns: evidence from Brazil

This paper investigates the differential labor market returns associated with completing distinct fields of knowledge, leveraging the Brazilian Unified Selection System (SISU, Sistema de Seleção Unificada), which allocates applicants to public higher education institutions based on their scores on the National High School Examination (ENEM) and their declared preferences over up to two ranked course options. Using the admission cutoff of the first-choice field as an instrument for completing it — while holding the second-choice field fixed as the relevant counterfactual — we estimate a 2SLS specification that identifies returns to completing the first-choice field relative to the second-choice alternative, measured eight years after SISU registration. Our results show that crossing the admission cutoff substantially raises the probability of completing the first-choice field across all combinations, with magnitudes that vary by combination and second-choice margin, suggesting that the latter carries relevant information about individuals. We also find no significant earnings differences for most field combinations, with departures concentrated in a few specific second-choice subsamples — several fields display higher returns relative to completing Natural Sciences & Mathematics, Education, or Technology, while others yield lower returns relative to Health & Welfare. The asymmetry of these returns suggests that the order of preferences differentiates individuals and may help explain the observed heterogeneity. These findings highlight the importance of considering distinct counterfactuals when assessing the returns to completing different fields of knowledge.