Are social inequalities being transmitted through higher education? A propensity-score matching analysis of private versus public university graduates using machine learning models

Andrea Visentin & Louis Volante


This study investigates differences in employment outcomes of students graduating from private versus public universities in Spain, and the resulting impact on employment outcomes. The methodology involves propensity score matching, utilising novel machine learning approaches. Machine learning algorithms can be used to calculate propensity scores and can potentially have advantages compared to conventional methods. Contrary to previous research carried out in Spain, this analysis found a wage premium for those pupils who attended a private university in the short and medium term, although these differences were relatively small. The discussion outlines the implications for intergenerational inequality, policy development, and future research that utilises machine learning algorithms.

Keywords: employment outcomes; intergenerational inequality; university graduates; machine learning; propensity score matching.

JEL Classification: I24, I26, J62

Download the working paper