Regency EF
Estimation of Semiparametric Functional-Coefficient Panel Data Models with Individual and Time Fixed Effects (304000)
*Shaymal Halder, Auburn UniversityKeywords: fixed effect, individual effect, time effect, LSDV, first difference, local linear regression
We consider the problem of estimating a semiparametric varying-coefficient panel data model with both the individual and time fixed effects. We extend Sun et. al. (2009) estimator to also accommodate unobservable unit-invariant common effects that may correlate with explanatory variables in an arbitrary way. The unknown functional coefficients of interest are estimated by (asymptotically) concentrating out both fixed effects via local-linear kernel-smoothed least squares dummy variable approach. We also extend Rodriguez-Poo & Soberon(2014) estimator to concentrate out both individual and time effects via local-linear kernel-smoothed first differencing two-stage approach. Monte Carlo simulations under different scenarios show that the proposed estimation procedures exhibit good finite-sample performance.