Research On The Rate Of Return To Higher Education In China

Since 1978,the Chinese economy has entered a new stage of development.Under the context of a rapid economic development,education for different age groups has also been fully developed.Among them,the direction of higher education has shifted from elite to popular education.The Chinese government also pays special attention to the development of higher education and regards "rejuvenating the country through science and education" as an important development strategy.As everyone knows,higher education is at the core of the education system.It imports economic talents,scientific research talents and technological innovation talents for the development of a country,which is related to the level and quality of a country's human resources and affects the level of a country's economic and social development.At the same time,higher education is the most important way to invest in human resources,so what is the effects? Whether the value of higher education is worthy of investment has become a research topic of interest to scholars.The measurement of the rate of return to education is a social science issue.Due to the lack of counterfactual status,research on the rate of return to education needs to overcome a series of problems such as selection bias and individual educational heterogeneity.Most of the previous scholars estimates of the rate of return to education were based on the traditional Mincers income equation using ordinary least squares(OLS).However,this method cannot control unobservable variables,which can lead to endogenous problems due to variable omissions.For the missing variables,the instrument variable method(IV)is usually used to deal with it.However,people will compare the expected returns based on unobservable factors to choose whether to receive education.There is heterogeneity in the individual education return rate.At this time,the instrument variable method fails.This paper uses the local instrument variable method(LIV)with more loose assumptions and more robust results proposed by Heckman et al.to solve the problem of selection bias and heterogeneity.In this paper,we first use the latest survey data from CGSS,CHIPS,CLDS three microscopic databases,and use ordinary least squares(OLS)and instrumental variable method(IV)for regression analysis,then,with generalized Roy model used of heterogeneous treatment effect evaluation,the rate of return to higher education is estimated.In this part,Probit,Logistic,LDA,QDA,Random Forest and Support Vector machine were used to analyze the three microscopic databases,and the optimal propensity score estimation model was selected according to the results of ROC curve and standardized difference.On this basis,semi-parametric local instrumental variable method(LIV)was used to estimate the parameters of marginal treatment effect(MTE),and ATT ATE and ATU were calculated on the basis of MTE.Finally,these estimated values were compared with OLS and IV results.The results show that theestimated annualized returns in the CGSS2015,CHIPS2013,and CLDS2016 databases are 13.65%,12.1%,and 13.7%,respectively.Estimated the rate of return to education in the three databases are not different,confirming the accuracy of the estimation results.At the same time,the estimated results of the three databases are compared with the results of OLS and IV,all of which are IV> ATE> OLS.OLS underestimates the rate of return to higher education,and IV overestimates the rate of return to higher education.The classification effects are shown as positive values in all three databases,and the selection deviations are all negative.According to the results of the semi-parametric local instrument variable(LIV)estimation,it can be found that there is a significant heterogeneity in the rate of return to education among individuals,indicating that individuals will choose the level of education they receive based on comparative advantage.