Instrumental variables approach to correct for endogeneity in finance

Abstract

The endogeneity problem has received a mixed treatment in corporate finance research. Although many studies implicitly acknowledge its existence, the literature does not consistently account for endogeneity using formal econometric methods. This chapter reviews the instrumental variables (IV) approach to endogeneity from the point of view of a finance researcher who is implementing instrumental variables methods in empirical studies. This chapter is organized into two parts. Part I discusses the general procedure of the instrumental variables approach, including two-stage least square (2SLS) and generalized method of moments (GMM), the related diagnostic statistics for assessing the validity of instruments, which are important but not used very often in finance applications, and some recent advances in econometrics research on weak instruments. Part II surveys corporate finance applications of instrumental variables. We found that the instrumental variables used in finance studies are usually chosen arbitrarily, and very few diagnostic statistics are performed to assess the adequacy of IV estimation. The resulting IV estimates thus are questionable.