PurposeThis paper assesses how discretionary accruals (DAs) affect corporate cash savings policies and the motivation behind this cash saving behavior and, also whether the linkage between DAs and cash saving affect the market-perceived cash value.Design/methodology/approachWe construct the measure of DAs using the previous five-year average information to investigate the association of DAs with the change in cash. Moreover, the Faulkender and Wang (2006) methodology is utilized to examine the market-perceived cash value in DAs.FindingsThe key finding is that firms with high DAs save significantly more cash. A one standard deviation increase in DAs saves cash by 12.59%. Furthermore, the value of cash is low for these firms. The effect is stronger in firms with poor governance but not present in financially constrained firms.Research limitations/implicationsThe empirical evidence highlights DAs have negative effect on market-perceived cash value, which underscores the insight that managers manage earnings opportunistically using DAs.Originality/valueTaken together, we provide more evidence on the literature of accruals in earnings manipulation.

Accounting researchers frequently employ industry-specific residual based models to draw inferences. Examples include discretionary accruals in Jones (1991) and modified Jones model (1995), and accruals quality in Francis et al. (2005). This paper illustrates that the interpretation of the residual terms is potentially subject to the problem that arises from industry misclassification. In an industry-specific crosssectional regression, the assumed homogeneity within the same industry is problematic because the industry classification system is noisy, and thus large magnitude residuals are potentially caused by misclassified observations. Moreover, the misclassification may not happen randomly. If firms with certain characteristics are more likely to be misclassified, directional biases rather than pure noise may emerge. In firm-specific time-series settings, the implied stationarity over time for the same firm is also questionable. Firms face external shocks and/or internal changes. The big magnitude residuals could capture those "shock" or "change" years. Given that researchers often investigate whether there exists earnings management around special events, such as M&A or equity issuance, it would be necessary to distinguish between whether abnormal accruals around the event is due to earnings management or simply non-stationarity.