Siew Hong Teoh Yinglei Zhang Abstract: Kraft, Leone, and Wasley (2005) find that after trimming extreme return observations, there is an inverted-U relationship between Accruals or NOA and subsequent one-year abnormal returns. They therefore reject psychological explanations for Accruals and NOA anomalies. We show that this relationship is a spurious consequence of the truncation bias of Kothari, Sabino, and Zach (2005), and of inadequate asset pricing controls. Loss firms are concentrated among low Accruals and NOA portfolios, and have more positively skewed returns, intensifying truncation bias among such portfolios. Even after trimming, using appropriate asset pricing controls the NOA anomaly is monotonic; the Accruals anomaly is monotonic among profit firms. |