Returns To Education In China

China Data Center released a survey on college students in 2012, The issue of’the college students’starting salaries are lower than that of migrant worker’ triggered a big discussion between the media and net friends. Market and media research center of Peking University joint ganji.com issued their new report:Employment Trends for Major Force Employment in 2015, the report found:The national college graduates hit a record high, and the generation after 90s’average monthly salary was 2687 RMB, in addition the proportion of self-employed was up to 15.6, but the highest proportion was in sales areas. The Chinese People’s Livelihood Development Report 2015 showed that the current income feels completely out of balance with their education level. There exists a huge income gap between urban and rural areas, men and women, east and west. In the meantime, the non-influence of individual factors (such as the education of parents or spouses) become more outstanding. Thus in recent years, education problem has always been a hot topic of society, and the issue of return on education has been catching research scholars’eyes.The rate of return on education refers to individual workers whose net economic rewards shows diversity as a result of education. As it used here is the data from China Health Nutrition Survey (CHNS) which released in 2014, the thesis goes on with the discussion about the returns to education concluding two points:One is to measure and calculate the returns to education in China. The other is to analyze the differences of the return to education among gender, urban-rural areas, regions, as well as unit &non-unit organizations. This paper chooses Mincer wage equation as a model to calculate the rate of return to education, in order to make more accurate estimates, four methods are considered. there are, Method of ordinary least squares(OLS), Quantile Regression (QR), Instrumental Variables (IV) and instrumental Variable Quantile Regression (IVQR). QR is a good method for dealing heterogeneity bias of the equation and for handling fat-tail distribution; IV is available to control the selection biases and solve the problem of omitted variables; IVQR has the excellent characteristics of both the IV and QR.This paper used four instrumental variables, expected to tackle the problem of endogeneity in the regression. And it is proved that the education years of partner is the proper instrumental variable. So it is easy to calculate China’s education yield with GMM, and the result is 6.1%. The rate in women is higher than man, but men’s average wage is much higher (up to 26%) than female; the highest yields of education is in the western region, followed by the eastern region, and finally the central region, eastern and western areas have the similar trend in upper quartiles; the rate of unit organizations take on the quality that increases in lower quartiles and decreases in upper. And for non-unit organizations, it is obvious to see the increasing trend steadily, up to 3%, and the rate in urban areas is higher than rural areas.