Endogeneity and the Dynamics of Corporate Governance

Abstract:     
Research in corporate finance is complicated by the endogenous relation between the control forces operating on a corporation and its financial decisions. In this paper we adapt a dynamic panel GMM estimator to deal with endogeneity in corporate finance research. The estimator incorporates the dynamic nature of corporate finance relationships to provide valid and powerful instruments that control for unobserved heterogeneity and simultaneity. The estimator is straightforward to implement and is more theoretically and practically appealing than many recent attempts to deal with endogeneity. Further, it may have significant advantages over commonly used fixed-effects estimators. We then demonstrate the estimator empirically by re-examining the relation between board structure and performance in a panel of more than 6,000 firms between 1991 and 2003. We find that when we control for past performance, simultaneity and unobservable heterogeneity, there is no causal relation between board structure and current firm performance. We illustrate why existing research that has found a causal relation is likely to be biased. We discuss situations where using this estimator will provide the greatest benefits.

Keywords: corporate boards, inside ownership, dynamic panel data estimation, regulation, endogeneity

JEL Classifications: G32, G34

Working Paper Series