Propensity Score Techniques to Evaluate Returns of College Education

This paper surveys the empirical evidence on causal effects of college education on earnings for China and compares alternative studies in the light of their underlying identifying assumptions. Recent data from a national survey, the Chinese Health and Family Life Survey (CHFLS), provide the empirical basis for the application of mathematical models for the return to college education. In the empirical analysis we assess the earning differential of college graduates versus high school graduates. Probit estimates show that parental kind of job, gender and urban influence positively the probability of obtaining a university degree. Then a matching method is used to estimate the effect of college achievement on earnings. Results suggest that eliminating selection bias college education has a significant positive impact on earnings. This conclusion leads to important policy implications both for the financing of college education and on individual decisions.