A rapid expansion in demand for post-secondary education triggered an unprecedented boom in higher education programs in Colombia, raising concerns about their relevance and quality. This paper shows that the penalty on student learning and labor market outcomes of attending a recently created program is large but, to a large extent, it is driven by student and program selection. Using rich administrative data-sets that match higher education school admission information, socioeconomic characteristics of the young graduates, standardized test scores pre- and post-higher education, and formal labor market outcomes, we characterize this selection process by disentangling the relative roles of demand and supply forces. The main factor behind the learning penalty is student selection in baseline ability. In the case of labor market outcomes, the penalty is due to a combination of student and program characteristics.

We present new empirical evidence on the implications of minimum wages on employment in the uncovered sector in developing countries, analyzing a unique dataset assembled from a set of micro surveys collected in 49 low- and middle-income countries. Our identification strategy exploits relative bindingness in minimum wages across labor market groups within countries and across years. We find that a higher minimum wage is associated with a larger self-employment share. The estimated impact of the minimum wage on the uncovered sector is economically significant: a 1 percentage point increase in the minimum wage ratio is associated with 0.1 percentage points increase in the self-employment rate, 0.182 when the non-compliance rate with the minimum wage is taken into account.

This paper partially identifies population treatment effects in observational data under both non-random treatment assignment and sample selection. Bounds are provided for both average and quantile population treatment effects, combining assumptions for the selected and the non-selected subsamples. We show how different assumptions help narrow identification regions, and we illustrate our methods by partially identifying the effect of maternal education on the 2015 PISA math test scores in Brazil. We find that while sample selection increases considerably the uncertainty around the effect of maternal education, it is still possible to calculate informative identification regions.

We use rich Brazilian administrative data to examine peer effects on wages for an entire local labor market, distinguishing between same-sex and opposite-sex peer effects. Both individuals’ and peers’ average ability are unobserved, so we estimate them by taking advantage of the panel structure of the data. We also control for workers’ sorting into peer groups and firms. We find that same-sex peer effects are remarkably larger (about double) than opposite-sex peer effects in the workplace for both men and women. In addition, differences between same-sex and opposite-sex peer effects are attenuated in contexts with more egalitarian gender identity norms. We argue that these patterns are consistent and best explained by economic models of behaviour and interactions that incorporate the psychology and sociology of identity.