overstated to that in the post-overstatement years. This approach has the advantage that each test company serves as
its own control. We contrast overstatement periods with post- rather than with pre-overstatement periods for two
reasons. First, periods in which earnings overstatements violate GAAP are typically preceded by one to three years during
which managers manipulate earnings upward, but within the bounds of GAAP (Ettredge et al., 2010). This manipulation in
the pre-overstatement periods biases against the detection of conservatism differences between the pre-overstatement
and overstatement periods. Second, managers’ incentives to report conservatively are stronger after they have publicly
corrected prior overstatements of earnings. Given that we maintain the hypothesis that test companies’ reporting is less