The use of 4 years (16 quarters) of data in the estimation of the standard deviation of residuals from quarter/industry cross-sectional regressions follows from Francis et al. (2005). The Fama and French 49 industry classification is used to arrange firms into sectors for running the regressions. All regression variables (including the constant term) are scaled by total assets (#44), averaged between quarters t 1 and t. Theway our base-case quality measure AQDD is estimated suggests that the higher the standard deviation of residuals is, the more uncertain (volatile) the mapping of accruals into cash flows, thus the lower the quality of information in accruals and the more ‘informationally’ risky the firm is. The second accruals quality measure we employ is an interesting modification to the Dechow and Dichev (2002) and Francis et al. (2005) model, proposed and tested by Prakash (2009). This alternative measure, which we denote AQDD(P), is again estimated as the 4-year (i.e.,16 quarters, from quarter t up to quarter t 16) standard deviation of firm i residuals from the following regression, that is estimated cross-sectionally every quarter t and for every industry with at least 20 firmobservations in a given quarter