Changes in the predictive ability of earnings around earnings restatements

The purpose of this paper is to investigate whether earnings restatements have a larger effect on the earnings quality (proxied by persistence) of restating firms relative to similar non‐restating firms and if restated earnings are more persistent than the originally reported earnings.

Cross‐sectional earnings persistence models are used to analyze how earnings persistence changes around restatements for both the originally reported earnings and the new restated earnings numbers. The study looks at restatements from 1997 through 2006.

The findings show that restating firms exhibit a larger increase in earnings persistence from the two‐year period before to the two‐year period after the restatements. Results also show that the restated portion of earnings is incrementally persistent relative to the originally reported earnings and the incremental persistence, although mitigated, is still significant after the passage of the Sarbanes‐Oxley Act. In addition, the evidence shows that core account restatements are associated with more persistent earnings relative to non‐core restatements in the two‐year period after the most recent restatement year.

The paper presents the first study to examine earnings restatements' impact on the future earnings persistence of restating firms in the context of the restated financial period as opposed to the restatement announcement period.