Discussion of Detecting Earnings Management: A New Approach

Dechow, Hutton, Kim, and Sloan (2012) propose a novel approach to identify accruals-based earnings management. Their approach is based on the concept that accruals-based earnings management made in one period must reverse in a subsequent period. By attempting to model the dynamic nature of accruals-based earnings management, they make a significant contribution to the literature on identifying earnings management. There are, however, several concerns that limit the effectiveness of their approach. First, it relies on the traditional methods of estimating discretionary accruals. Their approach is therefore subject to the estimation problems of these traditional methods. Second, their approach does not take advantage of dynamic features of the firm’s accounting numbers arising from the reversal of manipulated accruals. It instead tests whether there are group differences in the levels of discretionary accruals based on whether firms engaged in earnings management. Third, their approach assumes that the researcher already knows the periods in which earnings management occurred and reversed. It provides no guidance in identifying these periods.