Accrual Anomaly and Idiosyncratic Risk: International Evidence

ABSTRACTIn this study, we show that accrual abnormal returns are positively correlated to idiosyncratic risk in international equity markets. In addition, we find that idiosyncratic risk has less impact on accrual abnormal returns for developed countries than emerging countries. Our results are robust to different model selections, such as portfolio approach and regression analysis, across countries. Our results support the mispricing explanation of accrual anomaly around the world.JEL: G12, G15KEYWORDS: Accrual Anomaly, Idiosyncratic Risk, International Equity Market, Limits of Arbitrage(ProQuest: ... denotes formulae omitted.)INTRODUCTIONSloan (1996) demonstrates strong and robust evidence that taking long (short) position in stocks with low (high) accruals generates significant abnormal returns the following year. Since this influential study, accrual anomaly has been extended and applied in researches in both financial economics and accounting. Understanding accruals and its impacts has become increasingly important in asset allocation, security analysis, and other applications. Although accruals have been extensively studied in the U.S. capital market, they are less explored internationally. In this study, we examine accrual anomaly and its correlation with idiosyncratic risk around the world.We begin by examining abnormal returns of zero-cost trading strategy formed on accrual anomaly in 43 countries during 1989-2009 time period. We first divide all stocks in each country into five quintiles based on the ranking of accruals. We calculate equal-weighted monthly abnormal returns as longing the quintile with the lowest accruals and shorting the quintile with the highest accruals every month. Our results show that zero-cost trading strategy produces significant average monthly abnormal returns in 10 countries, including U.S. To examine if accrual abnormal returns can be explained by some well-known risk factors, we apply the Fama-French three-factor model plus the momentum factor (Fama and French, 1992, 1993, Carhart, 1997) to examine risk-adjusted returns (i.e. the a's).When applying factor models in a global context, a natural question is whether securities are priced locally or globally (Karolyi and Stulz, 2003). Griffin (2002) examines country-specific and global versions of the Fama-French model. The author finds that domestic factor model explains much more time-series variation in returns and generally has lower pricing errors than global model. In a recent study, Hou et al. (2011) also find that local and international versions of their multifactor models have lower pricing errors than global versions, and it is particularly true for emerging markets. Ferreira et al. (2006) use both domestic and international models to evaluate mutual fund performance around the world. Their results show that domestic and international models provide similar fund performance measures. Following these studies, we apply domestic model in this study since no significant contributions are present from the foreign components of international model. Our risk-adjusted a's indicate that most of the significant abnormal returns from zero-cost strategy still exist after controlling for the risk factors. It indicates that accrual anomaly exists around the world. Our next question is what factors could contribute the existence of accrual anomaly. Many studies have provided various explanations of the existence of anomalies. It has been suggested that anomalies indicate either market inefficiency (mispricing) or inadequacies in the underlying asset-pricing models. In this study, we do not attempt to address inadequacies of existing asset-pricing models or seek additional risk factors. We focus on the test of mispricing explanation by examining the impact of idiosyncratic risk on stock abnormal returns.Shleifer and Vishny (1997) and Pontiff (1996, 2006) argue that idiosyncratic risk represents a significant cost for risk-averse arbitrageurs, who cannot hedge it completely. …