Good versus Bad Earnings Management: Is Income Smoothing a Deliverance?

This paper examines, whether market participants differently assess the information uncertainty associated with earnings management depending on the degree of income smoothing. Hypothesis tests using annual data for 850 firms listed on German stock exchanges for the years 1990 to 2003 show that firm years with low (high) degrees of earnings management display lower (higher) degrees of information uncertainty. Income smoothing produces the reverse effect. Further hypothesis tests demonstrate that within a high income smoothing subsample, the differences in the degree of information uncertainty between high and low earnings management firm years are about half of the size compared to a low income smoothing sub-sample. These effects are only partially robust to the inclusion of control variables for unexpected earnings, past returns (momentum effect) and extreme financial performance. Using a fixed effects panel regression model including these control variables, the information uncertainty decreasing effect of income smoothing remains significant, but only small evidence for the information uncertainty increasing effect of earnings management can be obtained. The regressions produce no significant indication that income smoothing information is used to differentiate between opportunistic earnings management and earnings management to achieve a more reliable measure of firm performance.