Research On Specification Test And Estimation Of Panel Data Fractional Response Model

With the persistent diversification of micro data,empirical research in the field of individual behavior has attracted more and more attention.However,because most research modeling is non-linear model,compared with linear model,there is greater difficulty in the research process.There are many fields still to be explored in the theoretical research.This thesis focuses on in-depth research on the specification test and estimation correction of panel data fractional response model,in order to provide more ideas and contributions in related fields.At present,the domestic theoretical research field involving panel data fractional response model started very late,and there are few related studies.This thesis first briefly summarizes the research results on the specification test and estimation inference of the score panel data model at home and abroad,which provides a certain idea for further research.Secondly,this thesis draws on Ramsey's RESET test and applies it to the panel data fractional response model.The smallsample simulation results in this thesis show that the power of the heteroscedasticity test is better,but the fixed-effects probit panel data fractional response model has a lower power for the missing variable test,and further research is needed.Then this thesis aims at the incidental parameters problem in the estimation of the scored panel data model leading to the large error of the estimator,and proposes an estimation method of error correction.Error correction,the estimated amount of error correction is proposed.The small sample results show that the error correction estimator proposed in this thesis can well correct the error of the usual quasimaximum likelihood estimator whether it is the estimation of the parameter or the estimation of the average marginal effect.At the end of the thesis,some of the proposed estimation test methods are applied to the empirical study of capital structure to study the influencing factors of the choice of corporate capital structure.