Prior studies on the horizon problem of departing CEOs have considered only R&D expense as the tool for earnings management.
For better comparability with these prior studies, we repeat our analyses after replacing the variable abnormal discretionary expenses related to R&D, advertising, and SG&A, with raw R&D expense.
We also repeat our analyses with abnormal R&D expense, which is estimated using a model similar to Eq.
Table7 reports that when Early Years it is not included as an explanatory variable, the coefficients on Final Years it are insignificant in both R& D expense and abnormal R&D expense models.
This finding is consistent with those of Murphy and Zimmerman (1993) and Cheng (2004).
However, after including Early Years it as an additional explanatory variable in these models, the coefficients on Final Years it become negative and significant, consistent with the departing CEOs’ incentive to overstate earnings.
Moreover, as expected, the coefficients on Early Years it are negative and significant.