Endogeneity and the economic consequences of tax avoidance

Academic research investigating the economic consequences of tax avoidance is almost always interested in the consequences of intentional, deliberate actions undertaken to reduce taxes relative to income. Therefore, it is crucial that such research distinguishes between intentional and incidental tax avoidance, since failure to do so can create endogeneity concerns and lead to incomplete and incorrect economic inferences. In this paper, we first develop a framework that conceptually defines and distinguishes between intentional and incidental tax avoidance. We highlight that the endogeneity problem arises because intentional tax avoidance is not directly observable. We consider two approaches to mitigating endogeneity concerns and apply these approaches by reexamining two influential studies that investigate the economic consequences of tax avoidance. We show how controlling for past accounting losses eliminates the effect of tax avoidance on credit spreads (Hasan et al. 2014, Journal of Financial Economics, 113(1), 109–130) and how using an instrumental variables approach changes the sign of the relation between tax sheltering and stock price crash risk (Kim et al., 2011, Journal of Financial Economics, 100(3), 639–662). Overall, our paper punctuates the importance of both (1) conceptually distinguishing between incidental and intentional tax avoidance and (2) econometrically addressing the challenges that arise when empirical differentiation between incidental and intentional tax avoidance is important to the research question.

Duke Scholars

Author Scott D. Dyreng Fuqua School of Business