Regulators are concerned that during the process of restating financial statements, firms fail to provide timely progress updates, and delay earnings announcements and regulatory filings. To reduce these perceived lags in disclosure, an advisory group to the Securities and Exchange Commission recommends more use of catch-up adjustments rather than restatements to correct accounting errors. Some investor groups oppose the recommendations because they fear that preparers will begin to correct important errors through catch-up adjustments, which are less transparent than restatements.
We inform this debate by examining (1) the length of disclosure lags around restatements to understand the extent of the problem, and (2) the causes of disclosure lags to evaluate whether the reforms would address the root causes of the lags. We find that lengthy lags are uncommon and appear to be largely unavoidable consequences of fraud investigations. When fraud is a factor, the firm typically takes weeks or months to release restatement details, quarterly earnings, and SEC filings, likely because investigations are necessary to restore the firm's ability to produce reliable information. When fraud is not a factor, the firm typically discloses the restatement's earnings impact within a day of the initial restatement announcement, and delays the quarterly earnings announcement and SEC filing by less than a week. Although fraud is by far the most economically significant cause of lags, we also find that lags increase when a restatement involves multiple, long-standing, or large errors. Finally, we show that the restatements targeted by the reforms tend to have the shortest lags, even among non-fraudulent restatements. Thus, the proposed reforms would have a negligible effect on disclosure timeliness because the targeted restatements tend to have short lags to begin with, and because long lags appear to be caused by inherent constraints on producing reliable information.
JEL Classifications: M41, G38.
Data Availability: Data are available from sources identified in the paper.