Research & Publications
Insider Trading, Earnings Quality, and Accrual Mispricing
2002, Accounting Review
Messod Daniel Beneish, Mark E Vargus
The paper provides evidence that the signal contained in insiders' trading behavior is useful in making refined assessments of earnings quality, and informative about the valuation implications of accruals. We find that income-increasing accruals and unexpected accruals have lower (higher) persistence when managers engage in abnormal selling (buying) suggesting that insider trading information is useful in assessing the quality of the non-cash components of earnings. We show that the accrual mispricing phenomenon observed in previous work is largely due to the mispricing of positive accruals. We find that investors price all positive accruals as if they were informative and that a subset of positive accruals is correctly priced. That is, (1) investors correctly price positive accruals that are likely to be informative because concurrently insiders engage in abnormal buying, and (2) investors price all positive accruals the same independently of insider trading. We find that the extent of the mispricing is greater when positive accruals occur concurrently with abnormal selling relative to cases where there is no trading. The extent of the mispricing and the magnitude of the one-year ahead returns (14.7 to 22 percent) to a trading strategy based on positive accruals and abnormal selling suggests that these accruals arise from opportunistic earnings management that is successful in misleading investors. By contrast, the smaller positive accrual mispricing when there is no insider trading is more likely related to either the complexity of the firms' accrual generating process (Thomas and Zhang (2001) or to earnings fixation (Sloan 1996). Our evidence thus suggests that opportunistic earnings management is a partial explanation for the accrual mispricing phenomenon.
Beneish, Messod D., and Mark E. Vargus (2002), "Insider Trading, Earnings Quality, and Accrual Mispricing," The Accounting Review, Vol. 4, pp. 755-791.