The Returns from Training in Colombia: Evidence from a Longitudinal Survey

Abstract

In this paper we explore the wage returns from training in Colombia, using data from the "The Social Longitudinal Survey" between 2009 and 2010. Using the longitudinal component of the data, we control for time invariant individuals' unobserved heterogeneity and, following Wooldridge (1995), we correct for sample selection in panel data. Our results suggest that there is a positive return from training. The magnitude of the estimates suggests an increase in wages between 7% and 9%. We find that OLS overestimate the effects of receiving training. This suggests that individuals who receive training are the ones with high unobserved skills. We also find that trainees increase their probability of being continuously employed, suggesting an improvement in individuals' working conditions.