4.4 Testing for test on earnings management is based on the ratio of the standard deviation of the change in net income (∆NI) to the standard deviation of the change in cash flows (∆CF).We performed this test for both sub-groups of firms based on the voluntary or mandatory adoption of IFRS. In order to capture any confounding effects by factors that are not attributable to the financial reporting setting, we followed Barth et al. (2008) and Ahmed et al.(2010).Thus, we compared the ratio of the standard deviation of the change in net income (∆NI*) to the standard deviation of the change in cash flows (∆CF*) residuals, from the following two regression models, instead of comparing the SD of the change of NI and CF directly. The relative models have the following form:
∆NIit = a0+a1SIZEit+a2GROWTHit+a3LEVit+a4AUDit+eit (3)
∆FCit = a0+a1SIZEit+a2GROWTHit+a3LEVit+a4AUDit+eit (4)
Where:
∆NI is the annual change in net income divided by lagged total assets
∆CF is the annual change in the operating cash flow divided by lagged total assets
SIZE is the natural logarithm of end year total assets
GROWTH is the percentage change in sales
LEV is the ratio of end. Year total liabilities to end year total common equity
AUD is a dummy receiving(1) if the firm is audited by PeC,KPMG, Grand Thornton,
E&Y or D&T, and (0) otherwise
Ahmed et al.(2010) declare that if managers use accruals to smooth the changes in net income should be less than the variance of the change in cash flows. So, less positive values of this ratio would indicate greater smoothing and more earnings manipulation.