Can Short Sellers Anticipate Accounting Restatements?

Abstract

An announcement by corporate managers that they will restate a past financial statement due to an "accounting irregularity" provides the laboratory market for our tests of short seller expertise. For 565 firms announcing a restatement during the period 1995 - 2002, we find the level of short interest in the announcement month is substantially higher for restating firms than for control firms. A statistically significant difference begins 19 months before the announcement date, increases significantly in the six months preceding the restatement announcement, and peaks in the announcement month. In the twelve months after the announcement, the average level of short interest remains steady for the restating firms - as if short sellers are waiting to be sure all bad news has been released before unwinding their positions. Consistent with a contagion effect, the level of short interest for the control firms begins to increase after a restatement is announced by its industry- and size-matched counterpart. Using three proxies for "severe" accounting irregularities, we find that the significant difference in short interest between restatement and control firms is driven entirely by those firms restating because of severe accounting irregularities. Our results indicate a high level of short seller expertise in identifying accounting irregularities.

Keywords: restatements, short selling, short interest

JEL Classification: M41, M43, G12, G14