Semiparametric Estimation of Fixed E ects Panel Data Varying Coe cient Models

We consider the problem of estimating a varying coe cient panel data model with xed effects using a local linear regression approach. Our proposed estimator can consistently estimate the regression model with an additive intercept term, while the conventional estimation method based on a rst di erence model fails to do so. The computed estimator is shown to be asymptotically normally distributed. A modi ed least-squared cross-validatory method is used to select the optimal bandwidth automatically. Moreover, we propose a test statistic for testing the null hypothesis of a random e ects against a xed e ects varying coe cient panel data model. Monte Carlo simulations show that proposed estimator and test statistic have satisfactory nite sample performance.