Estimation And Application Of Linear Quantile Regression Model For Panel Data With Endogenous Variables

The panel data quantile regression model has the advantages of panel data,but also retains the advantages of quantile regression,and has been widely used in econometrics and other fields.In the research of practical problems,we often encounter endogenous variables.If the endogenous problem of variables is ignored,the estimation results will inevitably be biased.Therefore,this article proposes an improved two-stage quantile regression estimation for the panel data regression model with endogenous variables.Its main work is as follows:First,introduce instrumental variables to eliminate the endogenous problem of the model,and use compound quantile regression to estimate the endogenous explanatory variables.Secondly,the individual fixed effects of the panel data are processed by withingroup centralization,and the regression coefficients are estimated by quantile regression and compound quantile regression respectively.Under some regular conditions,the asymptotic nature of the proposed estimation is proved,and the error term is simulated under the three different distributions of normal distribution,t distribution and Cauchy distribution.The simulation study shows that the method proposed in this paper is not only The advantages of quantile regression are retained,and the endogenous problem of the model is effectively dealt with.Finally,the proposed method is used to analyze the impact of foreign trade openness,labor and human capital stocks on economic growth,and geographic distance is used as an instrumental variable for foreign trade openness.The results show that foreign trade openness,labor and human capital stocks are Different quantiles are promoting economic growth,that is,there is a positive correlation with economic growth.