Research On Educational Rate Of Return Based On Spatial Quantile Regression

China's current economic development is in the process of transforming from factor-driven growth to innovation-driven growth,so the importance of human capital investment is increasing,and education investment is one of the important forms of human capital investment.The rate of return on education is one of the ways to measure the return on education investment.As for the research on educational rate of return,most of the existing paper take either the trend of temporal variation as the research direction,or the trend of marginal variation as the research direction,and the spatial correlation of educational rate of return as the research direction,almost none of which take the spatial correlation of educational rate of return as the research direction.The research focus of this thesis is to discuss the educational rate of return among 31 provinces and cities in China,that is,to explore the spatial correlation of educational rate of return,and to analyze the individual differences under the spatial effect.Considering the spatial effect of educational rate of return,according to the distribution characteristics of spatial data,we want to comprehensively describe how dependent variables change with the change of independent variables,so we choose the spatial quantile regression model and use the instrumental variable method to solve the parameter estimation problem of the model.In order to better verify the effectiveness of the model relative to the quantile regression model,monte carlo simulation was used in the case of limited sample size.The verification results show that compared with the quantile regression estimation,the spatial quantile regression estimation has smaller deviation and higher robustness.Separately in space quantile regression and quantile regression,the data by the mainland's 31 provinces and cities in China's mainland in 2013 the CGSS data to do empirical analysis,contrast found space by introducing instrumental variable quantile regression model more objectively measured under different quantile influenced by education fixed number of year of income,solve the problem of endogenous.This thesis also use of space differential effects on individual quantile regression model analysis,found that: first,women's education returns less than men(10%)of the whole education returns(15%),but with the increase of income and education phase of ascension,the gender discrimination in continuous improvement,even at a0.8 site on female education returns are slightly higher than male;Second,the rate of return on education(20%)brought by non-state ownership is higher than the rate ofreturn on education(5%?10%)brought by state ownership.Thirdly,the rate of return on education of non-party members is higher than that of party members(4%)at the low score point(10%),and lower than that of party members at the high score point(15%)and middle score point(15%?20%).Fourth,the educational return rate of non-agricultural household register(10%?20%)is higher than the educational return rate of agricultural household register(5%?15%),and the greater the number of points,the smaller the gap.