Post date: Dec 03, 2012 5:53:16 PM
"Selection Bias, Comparative Advantage and Heterogeneous Returns to Education: Evidence from China in 2000 (http://onlinelibrary.wiley.com/doi/10.1111/j.1468-0106.2004.00242.x/abstract)" by James Heckman and Xuesong Li.
This paper studies the heterogeneity in education seriously. I've also learnt many useful econometric terms today, such as treatment effect, ATE, MTE, TT (treatment on the treated), TUT (treatment on the untreated). When treatment effect presents, the usual IV estimator is invalid and hence constructing propensity scores is an alternative.
It seems like labor people usually consider returns to education in terms of increases of wage for an extra year of education. However, the quality of education has significant heterogeneity, especially in an emerging economy like China. The effect of this heterogeneity can be decomposed to selection bias and comparative advantage. The latter refers to personal choice of education comparing the disutility of education and gain from expected wage rate. The former is more likely associated with family's choice over the kid. Personally, I think it's a smart way to use parents' income to capture this selection bias in this paper.
As usual, I'm more interested in the social return of education. If the matching between education level and personal comparative advantage is positive sortative, would separate equilibriums be better than a pooling equilibrium that everyone gets college degree?