Stealth Disclosure of Accounting Restatements

In our paper, Stealth Disclosure of Accounting Restatements, which was recently accepted for publication in the Accounting Review, we investigate whether the prominence of the disclosure of a restatement is correlated with the market reaction and the likelihood of litigation. In our sample, we observe and categorize firms into three levels of disclosure. Some companies disclose their restatement prominently in the headline of a press release, usually one that is dedicated to the accounting misstatement (high prominence). Other firms provide less prominent disclosure, typically citing an earnings release in the headline, but still discussing the misstatement in the body of the press release (medium prominence). Most of the remaining firms simply restate prior-period comparative balances in an earnings release, with a footnote briefly explaining that the financial figures for the prior year have been changed (low prominence).

We investigate whether companies providing medium or low prominence disclosure of their restatement benefit from a less negative market reaction and/or a reduced likelihood of litigation. Our first finding is that the magnitude of the market response to a restatement announcement is related to press release format. Three-day returns differ substantially across the three categories of disclosure prominence, averaging -8.3 percent, -4.0 percent, and -1.5 percent for high, medium, and low prominence, respectively. Returns for the high prominence group are statistically different from those for the medium and low prominence groups. Next, we extend the return window to 20 days after the announcement to investigate post- announcement responses to restatements. We find returns of -7.9 percent, -6.4 percent, and -3.2 percent for the high, medium, and low prominence firms, respectively. These returns are considerably less dispersed than the short-window returns, and the 1.5 percent return difference between high and medium prominence is not statistically significant. Apparently, market participants initially underestimate the seriousness of some misstatements disclosed without a headline but subsequently correct their underreaction.

Next, we find that the average return for firms that have post-restatement news items is not significantly different from zero. In contrast, we find a statistically significant drift of -3.7 percent for firms with no news items in the 20-day period, which suggests that investors further evaluate the original press release information. Analysts appear to play an important role in this evaluation because most of the drift is in companies covered by three or more analysts. In addition, once we control for the seriousness of the accounting misstatement, we find that the press release remains highly significant in explaining announcement period returns (-1, +1), but not significantly associated with returns over the longer window (-1, +20).

Lastly, we find that the frequency of lawsuits declines monotonically across the three categories of disclosure prominence (27 percent, 16 percent, and 0 percent for the high, medium, and low prominence firms, respectively). The 16 percent litigation rate for medium disclosure suggests that some managers use medium prominence disclosure for an accounting misstatement that plaintiff attorneys view as serious. We estimate a logistic regression model of the likelihood of litigation in our sample, and find that the prominence index coefficient is positive and significant in the model (even after controlling for endogeneity), suggesting that the likelihood of litigation rises with disclosure prominence. Reducing disclosure prominence by one level (e.g., medium instead of high) reduces the odds of a lawsuit by about half.

The full paper is available for download here.