It is commonly believed that short sellers must possess better information than the average trader in order to overcome the costs associated with short sale restrictions. Previous studies provide evidence which is consistent with, but insufficient to prove, informed trading. Because the role of short sellers in the stock market has come under greater scrutiny by regulators, understanding the nature of the information used by these traders is crucial in determining whether there is need for additional restrictions on short sellers.
We examine short interest around large, one-day stock returns. Our methodology allows us to overcome biases that may exist in previous studies. We find significant and robust increases in short interest immediately prior to large negative returns caused by a limited subset of corporate events, but no statistically significant increases in short interest for firms with negative returns caused by commonly occurring events. Surprisingly, the level of short interest is higher for firms experiencing large, positive returns, and we find no significant declines in short interest preceding these positive returns. Thus, we find weak evidence of informed trading by short sellers, but must conclude most short selling activity appears to result from speculation or hedging and not informed trading.
We examine whether short sellers anticipate large changes in stock prices. That short sellers must possess superior information has been generally accepted since Diamond and Verrecchia (1988) and Fabozzi and Modigliani (1992). These authors show that in the presence of short sale restrictions, short sellers must have an information comparative advantage in order to overcome the costs of these restrictions. Documented stock market reactions appear to be consistent with this notion. Conrad (1986) and Senchack and Starks (1993) find that there are significantly negative stock price changes around announcements of large levels of and changes in short interest. Further, Desai et al. (2002) show that stocks listed on NASDAQ with relatively large levels of short interest experience longer-term abnormal returns. That short sellers possess more precise information is given further credence by Kadiyala and Vetsuypens (2002), who use short interest as a measure of information near stock splits. In general, these authors find that the level of short interest changes around certain splits, leading the authors to conclude that some stock splits provide information to market participants.
The question of whether short sellers do possess superior information, however, is not yet resolved. While negative returns following short sale announcements are a necessary condition to show short sellers are better informed, the negative returns alone are not sufficient to prove the superior information hypothesis. The use of ex post stock market reactions such as those in Senchack and Starks (1993) may lead to erroneous conclusions. For example, Chan (2003) documents that market participants tend to react incorrectly to news, particularly negative news. Further, it is now well known that investors suffer from numerous cognitive biases (see Hirshleifer 2001 for a complete description). It is plausible, therefore, that returns surrounding short interest announcements offer an incomplete picture of the news contained in short interest-that is, market participants may inaccurately assume that short sellers possess a better information set and adjust stock prices based simply on the observation of an increase in short sales. Indeed, recent papers suggest that short sellers may be no better informed regarding the prospects of the average firm.
Mercado-Mendez, Best and Best (2006) find that, in general, short sellers do not increase short positions in the month prior to large negative earnings surprises. Further, the levels of short interest in firms with large negative earnings surprises are similar to those for firms that have large positive earnings surprises. …