Heterogeneous returns to education in Italy

The estimation of the return from an additional education unit is subject to the ''ability bias'', responsible for the ''endogeneity'' of education to wages, and the ''return bias'', induced by self-selection into education levels based on individual-specific unobservable gains. We use a control functions estimator that, differently to standard instrumental variables, can account for both selection mechanisms. It adds two corrections terms to the wage equation, one for each selectivity source. We use Bank of Italy data for Italian men in the 25-60 age interval. Identification uses a major reform that in 1962 increased years of compulsory schooling, implemented through a set of cohort dummies as exclusion restrictions. We find no evidence of absolute unobservable wage advantages for the more educated. Still, they are positively self-selected: they study more because their marginal return to education is higher than the average, due to specific unobservable gains from additional schooling.