نقش خبر خوب در بهره کوتاه

Stocks with relatively high short interest subsequently experience negative abnormal returns, but the effect can be transient and of debatable economic significance. In contrast, relatively heavily traded stocks with low short interest experience both statistically and economically significant positive abnormal returns. These positive returns are often larger (in absolute value) than the negative returns observed for heavily shorted stocks. Thus, the positive information associated with low short interest, which is publicly available, is only slowly incorporated into prices, which raises a broader market efficiency issue. Our results also cast doubt on existing theories of the impact of short sale constraints.

On any given day, there are many relatively large and liquid stocks that could be easily and cheaply shorted, but nonetheless have few or no shorted shares (i.e., little or no short interest). The short interest in a stock is often viewed as a measure of heterogeneity of investor opinion. If this is the case, an easily shorted stock that is completely avoided by short sellers suggests unanimity among market participants that the stock is, at a minimum, not overvalued. An additional implication is that short sellers do not possess significant private negative information about the stock. Unlike previous studies on short selling, our goal in this paper is to investigate whether the absence of short selling is informative about future returns. Using NYSE, Amex, and Nasdaq short interest data from 1988 through 2005, we find that portfolios of lightly shorted stocks have economically large and statistically significant positive abnormal returns. These returns are often larger (in absolute value) than the negative returns on portfolios of heavily shorted stocks, and they are robust to issues such as portfolio weighting, the timing of portfolio formation, the risk-adjustment procedure, listing venue, and the inclusion/exclusion of recent new listings or the 1998–2000 period. Our results have significant implications for well-known models of the impact of short sale constraints on asset prices such as Miller (1977), Diamond and Verrecchia (1987), and Hong and Stein (2003). In these models, short sale constraints inhibit the incorporation of negative information in stock prices, but, because there are no constraints to going long, there is no such barrier to the incorporation of positive information or opinion. This assumption is central in, for example, Hong and Stein's (2003) explanation of why markets melt down, but don’t melt up. However, our results show that both positive and negative information apparently known to short sellers is not incorporated in stock prices, casting doubt on the critical asymmetry between the way good and bad news is revealed to market participants. Overall, we find evidence that short sellers are able to identify overvalued stocks to sell and also seem adept at avoiding undervalued stocks. Of course, our results raise the broader question of why prices only slowly adjust to reflect information from public short interest data, thereby joining a growing list of related anomalies. We have no explanation for this apparent market efficiency failure, but we can observe that the powerful “barriers to arbitrage” argument of Shleifer and Vishny (1997) does not seem to apply because the abnormal returns we identify can be captured by simple buy-only strategies. The remainder of this paper proceeds as follows. Section 2 reviews the relevant short sale literature. Section 3 presents the data. Section 4 discusses the research methods and the baseline results. Section 5 presents additional analyses and robustness results. Section 6 concludes.

We contribute to the ongoing debate regarding the role of short sellers in making prices more informative. Using short interest data on NYSE-, Amex-, and Nasdaq-listed stocks from 1988 to 2005, we examine stocks that are relatively intensively traded (as measured by share turnover), but nonetheless have little short interest. We find that these stocks are significantly undervalued, both statistically and economically. The positive abnormal returns on these stocks are often larger (in absolute value) than the negative returns on portfolios of heavily shorted stocks. They are robust to issues such as portfolio weighting, the timing of portfolio formation, the risk-adjustment procedure, the inclusion/exclusion of new lists and low-priced stocks, and the 1998–2000 period. Overall, we show that short sellers appear to be somewhat successful in identifying overvalued stocks to short, but they seem to be at least as adept at identifying (and avoiding) undervalued stocks. The positive abnormal returns earned by lightly shorted stocks are economically large and can be exploited with simple buy-only strategies. The short interest data we use are fully public, and the stocks in question are relatively intensively traded. There are no meaningful barriers to arbitrage, so our results challenge traditional views of market efficiency. The undervaluation of lightly shorted stocks is concentrated in smaller stocks, but this only makes matters more puzzling. Such stocks are usually viewed as more difficult and/or expensive to short, so a lack of short interest could be indicative of a barrier to short selling. If the absence of short selling in these stocks is due to short sale constraints, then we might expect them to be overvalued (Miller, 1977) or correctly valued on average (Diamond and Verrecchia, 1987), but no model of which we are aware suggests undervaluation. Additionally, why would the more liquid of these stocks (as measured by turnover) be the most undervalued? In the same vein, orthodox thinking is that short sale constraints inhibit the incorporation of negative information in stock prices, possibly leading to misvaluation and contributing to market crashes. In this framework, there are no barriers to the incorporation of positive information. This asymmetry is crucial in understanding how short sale constraints can explain important characteristics of market crashes, such as the tendency for stock returns to become highly correlated during meltdowns. Our findings cast doubt on this supposed asymmetry, showing that, for whatever reason, significant positive information is also only slowly reflected in certain types of stocks. And that's the good news in short interest.