The effect of structural changes on discretionary accrual estimates: Some implications for earnings management research

Abstract

This paper examines if firms' structural changes affect bias and measurement error in discretionary accrual model estimates. Understanding the bias and measurement error properties of discretionary accrual estimates is particularly important because the empirical evidence in many earnings management studies is based entirely on these estimates. The evidence in this study is consistent with structural changes causing measurement error in discretionary accrual estimates. The evidence also shows that the measurement error is correlated with the level of earnings and other variables that have commonly been used by earnings management researchers to partition samples into groups with different earnings management incentives. Such a correlation provides evidence of a strong potential for bias in these estimates. Since the empirical results from many studies are based entirely on these estimates, some earnings management studies may lack valid empirical support. © City University of Hong Kong.