Varying Coefficient Panel Data Model with Interactive Fixed Effects

In this paper, we propose a varying coefficient panel data model with unobservable multiple interactive fixed effects that are correlated the regressors. We approximate each function by B-spline, and robust nonlinear iteration scheme based on least squares method to estimate functions of interest. also establish asymptotic theory resulting estimators under certain regularity assumptions, including consistency, convergence rate distribution. Furthermore, develop dummy variable study an important special case proposed model: additive effects. To construct pointwise confidence intervals for functions, residual-based block bootstrap is reduce computational burden as well avoid accumulative errors. Simulation studies real analysis carried out assess performance our methods.