On the Use of Instrumental Variables in Accounting Research

Instrumental variable (IV) methods are increasingly being used in earnings management, corporate governance, executive compensation, and disclosure research where there is reason to believe that regressor variables are endogenous. While IV estimation is the standard textbook solution to mitigating the inconsistency in parameter estimates caused by endogeneity, the appropriateness of IV methods in typical accounting research settings is not obvious. Drawing on recent advances in statistics and econometrics, we provide conditions under which IV methods are likely to be preferred to traditional ordinary least squares (OLS). These results raise considerable doubt about the appropriateness of typical IV applications in accounting research. We illustrate these concerns by examining the association between corporate disclosure and the cost of capital. Our results indicate that the commonly used instruments are unlikely to provide estimates that are preferable to OLS.