关键词: 教育收益率, 选择偏差, 参数估计, 半参数估计

Abstract: This paper uses CHNS data to estimate the rate of return on education in China during 1989 to 2006 based on Mincer wages function. We apply parametric and semiparametric methods to estimate the models, and further compare these two models using the Hausman test. The results indicate that the former is more efficient than the latter from the perspective of estimation efficiency, but it is opposite from the perspective of estimation accuracy. In other words, the latter is a consistent estimator while the former is a biased one. Moreover, despite the fact that the rate of return on education trends toward a gradual increase, it is still much lower than the rate of return on capital in kind. Both the education market and the labor market in China need substantial investments and substantive reforms.

Key words: Return to Education of China, Selection Bias, Parametric estimation, Semiparametric estimation