Are Accruals Really Mispriced? Evidence from Tests of an Intertemporal Capital Asset Pricing Model

This paper examines the anomaly, first reported by Sloan (1996), that the market misprices stocks of firms with extreme (high or low) accruals. The paper proposes a four-factor ICAPM, based on Campbell and Vuolteenaho (2004) and Fama and French (1993), and tests the model using a two-pass cross-sectional regression. Two principal findings are reported. First, the model successfully prices the cross-section of accrual portfolios with an error that is statistically indistinguishable from zero at conventional sizes. In addition, abnormal returns to a variety of hedge portfolios are statistically or economically insignificant. These results do not hold for the CAPM and the FamaFrench three-factor model. Secondly, tests based on Chan and Chen (1991) reveal that the return behavior of the low accrual portfolio mimics the return behavior of a portfolio of firms with high bankruptcy risk. In sum, the evidence suggests that (i) cross-sectional variation in average returns to high and low accrual firms is due to differences in risk rather than mispricing, and (ii) these differences in risk are not due to accruals per se, but rather, to well-known economic and financial distress characteristics that are correlated with accruals. * Joseph L. Rotman School of Management, 105 St. George St., Toronto, Ontario, M5S 3E6, Canada. Email: [email protected] Tel: 416-946-8064. This paper is based on my dissertation, undertaken at the University of Toronto. I am indebted to the members of my committee: Jeff Callen (Chair), for being an enduring source of inspiration and support, and without whom this paper would not have been possible; Raymond Kan, for always graciously sharing his time and expertise; and Gordon Richardson and Dan Segal, for their guidance. I am grateful to Professor Ross Watts for valuable discussions and support. Helpful comments were received from Gauri Bhat, Gus DeFranco, Andras Fulop, Ole-Kristian Hope, Hai Lu, Jan Mahrt-Smith and seminar participants at the University of Toronto. I appreciate valuable comments from Professor Ray Ball on an earlier draft of the paper. This paper is to be presented at the 2005 AAA FARS midyear conference. I thank an anonymous FARS reviewer for useful comments.