The Potential Outcome of Schooling: Individual Heterogeneity, Program Risk and Residual Wage Inequality

This paper analyzes the link between individual heterogeneity, program risk, and outcome inequality within the potential outcome framework. Using standard ignorability conditions we derive identifying conditions for the variance of the treatment effect, i.e. the program risk, and relate this parameter to the variance treatment effect. Moreover, for a rather general set-up allowing for binary as well as continuous treatments, we derive upper and lower bounds for the program risk. Two applications demonstrate the approach at work. Using the Dehejia/Wahba (1999) data we check whether the estimates obtained by our approach can replicate the treatment parameters obtained from experimental data. Secondly, we estimate the returns from upper secondary graduation in Germany and the implied residual wage inequality. JEL classification: C21, J24, J31