After a Restatement: Long-Run Market and Investor Response
Purdue University - Krannert School of Management
Devin M. Shanthikumar
University of California, Irvine - Paul Merage School of Business
July 7, 2008
We examine returns and institutional trading in a long window surrounding earnings restatements. We find significantly positive abnormal returns following negative restatement announcements, even though, on average, these announcements are accompanied by negative pre-announcement and announcement-period returns. In particular, we find significantly positive returns in the six months following the announcement, using raw returns, 3- and 4-factor calendar-time abnormal returns, firm-specific 4-factor adjusted returns, and characteristic-adjusted returns. Results suggest that these positive returns are not due to a change in traditional risk factors or cost of capital. Linking the restatements to analyst forecasts, we find that analyst forecast dispersion increases around the restatement and then decreases in the 3-6 months after the restatement, consistent with an announcement-period increase in firm-specific uncertainty and information risk, which subsequently resolves. Analyst forecast errors do not become overly negative or subsequently drift upwards, inconsistent with investor overreaction.
Given the potential variation in investors' willingness to tolerate information risk and uncertainty, we examine institutional ownership changes. We find that dedicated institutions sell restatement-firm shares in the quarters before and after the announcement while transient and quasi-indexing institutions sell before the announcement but buy significantly in the months after. The event-time trading of dedicated institutions has weakly positive predictive ability for future returns while the event-time trading of transient and quasi-indexing institutions has weakly negative predictive ability. The differences in the returns predicted by the three groups' trading are statistically and economically significant. Together, the results suggest that transient and quasi-indexing institutions are less willing to tolerate increased information risk immediately surrounding a restatement, helping drive a strong negative price reaction to the initial announcement, along with a later recovery.
Number of Pages in PDF File: 60
Keywords: restatement, financial accounting, information risk, long-run returns, institutional investing, transient institutions, dedicated institutions, quasi-indexer institutions
JEL Classification: G14, M40working papers series
Date posted: March 17, 2008 ; Last revised: January 10, 2009
© 2013 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollo3 in 0.469 seconds