A Monte-Carlo study of Dynamic Panel Data Estimators with Autocorrelated Error Terms

Abstract

Many economic relationships are dynamic in nature and one of the advantages of panel data is that they allow the researcher to understand the dynamics of adjustment. These dynamic relationships are characterized by the presence of a lagged dependent variable among the regressors. In this research, the properties of some Dynamic Panel Data estimators including Ordinary Least Squares were investigated; the Anderson-Hsiao (AH), Arellano-Bond Generalized Method of Moment (GMM) one-step, Blundell- Bond System (SYS1) one-step, M and MM estimators in the presence of serial correlation. Monte-Carlo simulations were carried out at varying sample sizes with different degrees of autocorrelation. The results showed that in small and large sample situations, irrespective of time dimension, Anderson-Hsiao estimator (AH) outperformed all other estimators. Meanwhile, Blundell-Bond System estimator has the least performance among all the estimators.