Robust Estimation of Moments in Dynamic Panel Models with Potential Intercorrelation

In this article, we provide some robust estimation of moments of the random effects and the errors in dynamic panel data models with potential intercorrelation. By differencing the residuals over the individual and time indies, we modify the popularly used Arellano-Bond GMM estimator of the parameter coefficient and study its asymptotic properties. Based on the modified parameter estimator, we construct, respectively, some moment estimators of the random effects and the errors with no affecting each other. Their asymptotic normalities are obtained under some mild conditions. The finite sample properties are investigated by a small Monte Carlo simulation experiment.