A0711
Title: Estimation of dynamic panel models with interactive effects
Authors: Wenting Wang - University of York (United Kingdom) [presenting]
Jia Chen - University of York (United Kingdom)
Yongcheol Shin - University of York (United Kingdom)
Abstract: An internal instrument variable estimation method is presented for dynamic panel data models with unobserved common factors. The main idea involves using a cross-sectional average of regressors to project out common factors potentially emerging in both regressors and dependent variables. The defactorized regressors and their lags are subsequently utilized as instruments, followed by the implementation of the two-stage least squares (2SLS) on the defactorized outcome equation. Theoretical findings include the derivation of consistency and asymptotic normality of individual and mean group estimators in heterogeneous slope model, as well as the pool estimator in homogeneous slope model. These findings are supported by Monte Carlo simulations, which demonstrate the satisfactory performance of the proposed estimators in finite samples.
