Panel Data: A Methodology for Model Specification and Testing

Abstract

Abstract In this article, we provide guidance to financial economists on the choice of the proper panel data estimator for research. We first explain why the panel data methodology should be used if the unobservable heterogeneity problem arises. Additionally, we describe the alternative estimators that address the unobservable heterogeneity problem and that control for the endogeneity problem. Given that the latter problem is mitigated using an IV method, we explain the strategy to choose the right instrument set. The article also explains why the panel data methodology is particularly well suited to model specification and testing. Finally, we compare the different panel data estimators and offer researchers the key to choosing the most suitable one.