Accruals and earnings volatility mispricing: Evidence from US Stock Market 1988-2017

Academics have studied a lot of use of financial accounting information in predicting firms’ future performance, namely earnings. Firms’ net earnings are sum of period’s cash flows and period’s accruals and i.e. Sloan (1996) found out that relative magnitude of period’s cash flows and accruals predict future periods’ earnings differently. Remarkable finding of Sloan (1996) was that current periods cash flows predicts better future periods earnings than do accruals and hence there is negative association between earnings persistence and magnitude of accruals. Moreover, Sloan (1996) reported that capital markets misprice abovementioned information and abnormal returns can be yielded by following investment strategy based on magnitudes of accruals, opposite what efficient market hypothesis would forecast. Furthermore, Dichev and Tang (2009) found similar negative association between past earnings volatility and future earnings persistence. In this study empirical confirmation for following questions were to find; (i) to confirm negative association between earnings persistence and magnitude of accruals and earnings volatility (ii) to find evidence if above average return is possible to yield following investment strategy based on accruals and earnings volatility, and (iii) is there change in abnormal returns over time when following mentioned investment strategies. Study found supporting evidence for following hypotheses (i) abnormal returns have been possible to yield following accrual-based investment strategies on period 1988-1996, (ii) abnormal returns from following accruals based investment strategy have not been able to yield during 1997- 2017, (iii) abnormal returns has not been observed from earnings volatility based investment strategy during through sample data and, (iv) negative association between magnitude of earnings volatility and accruals is observable through data set. Study was conducted using previously proven robust research methods, so findings can be considered as at least indicative. However, research data set differs from other similar earlier studies and thus, findings may have different magnitudes. For example, sampling between earlier studies may be different in terms of sampling parameters and how outliers are handled producing slightly different results