Resource on litigation risk and market reaction to restatements

Resource on litigation risk and market reaction to restatements

Image: K Salavei Bardos Please consider as a resource for media coverage about litigation risk and market reaction to restatements, Katsiaryna Bardos, Ph.D., associate professor of finance at Fairfield University's Charles F. Dolan School of Business. Dr. Bardos recently published a paper "Litigation Risk and Market Reaction to Restatements," co-authored with Joseph Golec and John Harding, in the Journal of Financial Research. This paper investigates the extent to which market reactions to restatement announcements are explained by litigation risk. Dr. Bardos and her co-authors found that about half of the -9.2 percent average restatement announcement effect is due to expected litigation costs. They also found that the significance of the accounting error does not directly affect the magnitude of the abnormal return; it only affects abnormal return indirectly because it increases the probability of being sued.

One of the main implications of their study is that damages to shareholders calculated in class action lawsuits are overstated for firms that are more sueable (such as large firms with good cash flows). This happens because the stock loss at restatement announcement is often used to estimate damages in class action lawsuits and in academic literature that tries to establish whether class action lawsuits have merit. Their analysis shows that using the full loss overstates shareholder damages due to restatement because it implicitly includes greater litigation costs for firms that are more sueable.

Dr. Bardos and her fellow researchers' results have other implications: First, policymakers are interested in the magnitude of the shareholder harm caused by the misinformation, which firms correct in the restatement. Because they found that half of the announcement return is due to a firm's "suability," as opposed to the seriousness of its accounting errors, less stringent regulatory remedies could be appropriate. Similarly, the debate on tort reform could benefit from this evidence that litigation risk accounts for half of shareholder losses at restatement announcement. Second, financial analysts should be interested in disentangling the effects since litigation costs are one-time costs while the rest of the announcement effects may be related to ongoing operations. Lastly, short-sellers who trade in anticipation of restatements can use their results to gauge when stocks have fallen far enough to warrant covering their short sale.

Figure 1 (See below) starkly illustrates the issue of this paper. Using their sample of restating firms, Figure 1.A reproduces the typical pattern of abnormal returns observed in earlier studies around restatement announcements. The average abnormal return on the day of and the day following the restatement announcement is about -9 percent. But this masks significant differences between sued and non-sued firms. Figure 1.B shows that average two-day abnormal return of firms that were sued is much lower (about -20 percent) than that of non-sued firms (about -4 percent). A -4 percent average market reaction to an announcement is substantial, but an average -20 percent reaction for the sued firms suggests that expected litigation costs could be relatively more important than the seriousness of the accounting errors.

Figure 1.B (See below) shows that the return pattern in Figure 1.A. is largely driven by the restating firms that are sued. Compared to non-sued firms, sued firms' shareholders start to anticipate a costly problem well before the restatement announcement, and after a sharp two day drop at the announcement, the total price decline is maintained or increases somewhat. Non-sued firms' show little pre-announcement effects and they recoup their two-day loses shortly after the announcement. The fleeting announcement effect for non-sued firms implies that any firm revaluation due to their restatements is minimal.

For sued firms, investors appear to sense financial weakness up to 30 days prior to their restatements. Sued firms suffer -20 percent abnormal returns on average during the period from day -30 through day -2, and suffer another -20 percent decline at restatement announcement (days zero and plus one). Such a large negative market reaction after an already substantial decline suggests either a surprisingly large accounting problem, or that shareholders anticipate that the restatement itself could bring on large additional costs, e.g., litigation costs. Although Dr. Bardos and her fellow researchers found that sued firms commit somewhat more serious accounting errors than non-sued firms, the magnitude and endurance of restatement day abnormal returns for the sued firms is surprising.

To sort out the differential effect of litigation risk and the magnitude of the accounting error, one must recognize that firms' litigation risks and their restatement-related abnormal returns are interdependent. Lawsuits present a major expense to shareholders, therefore, more "sueable" restating firms should have more negative market reactions to their restatement announcements, all else equal. On the other hand, Jones and Weingram (1997) show that the likelihood of litigation is greater after large one-day stock price declines. Therefore, they modeled the relation between restatement announcement returns and litigation risk as a simultaneous equations system.

"Litigation Risk and Market Reaction to Financial Statement Restatements," with Joseph Golec and John Harding. Journal of Financial Research 36(1), pages 19-42, 2013.

Figure I: Cumulative Abnormal Returns Around Restatement Announcement

Figures show market model cumulative average abnormal return for 60 days surrounding restatement announcements. Market model parameters are estimated over a 250-day period starting on day -46 relative to restatement.

Figure I.A. Full Sample

Image: Bardos chart

Figure I.B. Sub-samples of Sued and Non-sued Restating Firms

Image: Bardos chart

Posted On: 09-25-2014 03:09 PM

Volume: 47 Number: 68