Testing for asymmetric effects in the accrual anomaly using piecewise linear regressions

Abstract

PurposeThe purpose of this paper is to test the persistence and pricing of earnings, free cash flows (FCF) and accruals using Australian data. In response to arguments concerning omitted variables in the Mishkin test, it seeks to explore asymmetric effects by incorporating categoric variables capturing firm size (microcap, small, medium and large); industry (industrial/mining); profit making (profit/loss); and dividend paying (contemporaneous dividend/no contemporaneous dividend) into forecasting and pricing equations.Design/methodology/approachThe paper examines a large sample of hand‐checked Australian earnings, accruals and cash flow data. It analyses these data using a series of piecewise linear regressions.FindingsThe results indicate that asymmetry is a valid concern since the extent and nature of mispricing of earnings components vary considerably across the categories included in the model. For example, the base case firms (microcap, loss‐making, resource companies that pay no contemporaneous dividends) exhibit no evidence of significant differences between the actual and implied persistence of FCF and accruals. Conversely, for industrial firms, the implied persistence of FCF and accruals from the pricing equation significantly underestimates the persistence of both earnings components as shown in the forecasting equation.Originality/valueThe study extends the research investigating the accruals anomaly by accommodating different factors that might induce asymmetric effects. Based on the evidence, such effects represent an important consideration for work conducted in this and related accounting research areas.