Discussion of a Market-Based Evaluation of Discretionary Accrual Models

Abstract

The use of accrual models has been pervasive in accounting research in the past ten years. This literature has sought to examine whether and why managers use discretion in financial reporting. The findings suggest that managers use their discretion for a wide range of reasons, including to increase their own compensation and protect their job security, to communicate their expectations of long-term firm performance with investors, and to create stockholder wealth at the expense of other stakeholders (such as debtholders, taxpayers, and regulatory bodies). Of course, the power and internal validity of these tests depend critically on whether the models of accruals used in the studies are well specified. Guay, Kothari, and Watts's (henceforth GKW) paper attempts to provide evidence on this question by modeling the relation between stock returns, and discretionary accruals and nondiscretionary earnings. GKW then examine whether the five most popular accrual models used in the literature produce discretionary accruals and nondiscretionary earnings that conform to their modeling predictions. The findings suggest that even the most effective models, the Jones and modified Jones models, do only a modest job of parsing earnings into discretionary and nondiscretionary components. Taken at face value the findings of the GKW paper have serious consequences for how readers should interpret the findings of earlier earnings management studies. However, such strong conclusions are

Keywords

Related Publications