USC FBE FINANCE SEMINAR presented by: Adam Reed FRIDAY, Sept. 10, 2010 10:30 am - 12:00 pm, Room: JKP-202

HOW ARE SHORTS INFORMED? SHORT SELLERS, NEWS, AND INFORMATION PROCESSING*

Joseph E. Engelberg Kenan-Flagler Business School, University of North Carolina [email protected] Adam V. Reed Kenan-Flagler Business School, University of North Carolina [email protected] Matthew C. Ringgenberg Kenan-Flagler Business School, University of North Carolina [email protected]

JULY 15, 2010†

ABSTRACT Combining a database of short sellers’ trading patterns with a database of news releases, we show that short sellers’ trading advantage comes largely from their ability to analyze publicly available information. Specifically, the prior finding that short sellers’ trades predict future negative returns (e.g., Boehmer, Jones, and Zhang (2008) and Asquith, Pathak, and Ritter (2005)) is more than twice as strong in the presence of news stories. Further, the most profitable short sales do not appear to come from market makers, but from clients, and these client short sales are particularly profitable in the presence of news. We find no evidence that short sellers anticipate news events, as the ratio of short sales to total volume is nearly constant around news periods, and when we do find differences between the timing of short sellers’ trades and the overall market, relative to other types of trading there is a significant increase in short selling after news stories. We also find no evidence to support the idea the short selling around news events is more profitable because of liquidity effects; in fact, we find an increase in transaction costs on news days. Finally, short sellers’ ability to predict returns appears to be concentrated in many of the news categories in which short sellers trade relatively late, a finding consistent with the idea that short sellers’ advantage arises from their ability to process publicly available information. *

The authors thank Dow Jones for providing access to their news archive and Paul Tetlock for assistance with the Dow Jones news archive. We have benefited from comments from Greg Brown, Jennifer Conrad, Wayne Ferson, Charles Jones, Günter Strobl, and Robert Whitelaw. We also thank seminar participants at the University of North Carolina and the 2010 Utah Winter Finance Conference. This paper was previously titled “Buy on the Rumor, [Short] Sell on the News: Short Sellers, News and Information Processing.” † Comments welcome. © 2010 Joseph E. Engelberg, Adam V. Reed, and Matthew C. Ringgenberg.

There is overwhelming evidence that short sellers are informed traders. In particular, a number of empirical papers find that short selling predicts future returns (Asquith and Meulbroek (1995), Senchack and Starks (1993), and Boehmer, Jones, and Zhang (2008)).

Return

predictability, however, tells us little about how short sellers obtain an informational advantage over other traders. In this paper we address this question by combining a database of public news events with a database of all short sale trades, a unique combination that allows us to comprehensively examine the relation between short selling and the release of public information. One aspect of the relation between short sales and news that has received a lot of attention in the literature is timing.

Short sellers have been shown to trade before public

information is released. For example, Karpoff and Lou (2009) show that short selling increases before the initial public revelation of firms’ financial misrepresentation. Similarly, Christophe, Ferri, and Angel (2004) find evidence of informed short selling in the five days before earnings announcements. The financial crisis has also been linked to the timing of short sellers’ trades, with the Securities and Exchange Commission suggesting that short sellers had spread “false rumors” in an effort to manipulate firms “uniquely vulnerable to panic.”1 Using a sample of news events in the U.S. over the 2005 to 2007 period, we examine whether short sellers’ informational advantage is due to timing by looking for evidence of abnormal short selling preceding the news events, a pattern that would be consistent with anticipation. We find no such pattern. In fact, we find that the ratio of short sales to total volume is nearly constant around news events. Further, when we do find differences between

1

“What the SEC Really Did on Short Selling,” by Chairman Christopher Cox, 24 July 2008, The Wall Street Journal.

1

the timing of short sellers’ trades and the overall market, we observe that, relative to other types of trading, there is a significant increase in short selling after the news event. This result indicates that, on average, short sellers trade on publicly available information, that is, they do not trade on information before it becomes public. Having established that short sellers tend to trade on or after news events, we next ask whether news events present profitable trading opportunities for short sellers. The answer to this question is not obvious from the theoretical literature. On the one hand, a number of papers argue that news reduces information asymmetry (e.g., Korajczyk, Lucas, and McDonald (1991) and Diamond and Verrecchia (1987)). For example, if a firm announces a merger, investors who knew that the merger was likely no longer have an informational advantage over those who did not.

The news announcement therefore reduces the information asymmetry between the

informed and uninformed investors. Under this view, news announcements should make the trades of informed traders (short sellers) less profitable on news days. On the other hand, several papers suggest that public news events may lead to differential interpretations by traders based on variation in the traders’ skill (e.g., Kandel and Pearson (1995)). Rubenstein (1993) puts it succinctly: “In real life, differences in consumer behavior are often attributed to varying intelligence and ability to process information. Agents reading the same morning newspapers with the same stock price lists will interpret the information differently.” This view therefore holds that public information events present profitable trading opportunities for skilled information processors, which can explain not only high volume around news events (Kandel and Pearson (1995)) but also evidence of return predictability from “soft” information in news announcements (e.g., Engelberg (2008), Demers and Vega (2008)). Under

2

this view, news announcements should make the trades of informed traders (short sellers) more profitable on news days. When we take both of these theories to the data, we find evidence in support of the second view. Several papers find that abnormal short selling unconditionally predicts lower future returns (see, e.g., Asquith and Meulbroek (1995), Senchack and Starks (1993), Boehmer, Jones, and Zhang (2008)). We find that abnormal short selling does indeed lead to lower future returns, but this effect is largely concentrated around news events: short selling’s predictive effect on future returns is more than twice as strong in the presence of news stories. Thus, a short seller’s most informative trades appear to be those in response to newly released public news, consistent with short sellers being good processors of information. An alternative explanation for the above result may be that some buyers make systematic mistakes around news events (Antweiler and Frank (2006)), and that these buyers’ mistakes are reflected in market makers’ offsetting short sales. To determine whether short sellers’ trades are due to superior information processing or to offsetting positions, we exploit a unique feature of the short selling data, namely, exempt versus non-exempt trade marking, which allows us to distinguish market makers from non-market makers (or clients). We find that clients’ trades are particularly well informed, and that these trades are much more profitable in the presence of news events. In contrast, market makers’ trades are not particularly well informed, and there is no differential impact in the presence of news. Overall, we conclude that the most informed short sales are from clients, and that these shorts are particularly well informed in the presence of recent news. We interpret this finding that the profitability of short sales is higher on news days as evidence that short sellers are relatively good at processing news announcements.

3

Another alternative explanation for our result is that short sales are profitable on news days because news days provide short sellers increased liquidity. This explanation requires that the costs of short selling be lower around news announcements. We find the opposite, however. For example, we find that bid-ask spreads actually rise by approximately 5% around news announcements. Furthermore, when we include measures of transaction costs in our main regressions, they have no effect on our core findings. This evidence lends further support to the view that short sellers’ information advantage is due to their superior information processing ability. Our next set of tests seeks to identify which types of information are associated with short sellers’ advantage. To do so, we use the news subject classification in the Dow Jones archive to sort stories into various categories ranging from analyst comments to earnings announcements to new debt issues.

We find that short sellers’ most informative trades

concentrate in nine categories: Corporate Restructurings, Divestitures or Asset Sales, Earnings, Earnings Projections, Initial Public Offerings, Management Issues, New Products & Services, Research and Development, and Stock Ownership.

Interestingly, many of these categories

correspond to the categories in which short sellers’ trades are measurably later than other investors’ trades.

This result thus lends additional support to the idea that short sellers’

advantage stems from superior ability to process publicly available information rather than an ability to uncover information before it becomes publicly available. Finally, we examine the economic significance of traders’ ability to respond to news by implementing a portfolio approach. Recognizing that the presence of news is likely correlated with firm characteristics, and that some categories of news may be more relevant for some firms than for others, we conduct an experiment in which each firm’s response to a news event is 4

matched by a similar firm’s response on the same day. We find that across all news categories, short sellers’ advantage in predicting returns concentrates in firms with news. The remainder of this paper proceeds as follows. Section I discusses related literature. Section II describes the databases used in this study and presents our main hypotheses. Section III presents our analyses and findings. Finally, Section IV concludes.

I.

Related Literature

The ideas in this paper relate to three distinct branches of the existing literature. First, this paper relates to an extensive literature on the behavior of short sellers relative to other traders. Second, our paper contributes to a growing literature on how market participants respond to public news. Finally, this paper sheds light on an emerging debate on whether news increases or decreases information asymmetry. In this section, we first discuss prior papers that connect news to short selling. We then provide an overview of the relevant literature in each of these three branches. Several extant papers look at short selling behavior in the context of a specific type of corporate news event. As such, these studies shed light on a subset of this paper’s sample of news events. Karpoff and Lou (2009), for example, examine short sellers’ positions in firms that are investigated for financial misconduct and find that short sellers generally anticipate public announcements of investigations. Christophe, Ferri, and Angel (2004) and Christophe, Ferri, and Hsieh (2005) focus on short sellers’ trades around earnings announcements and analyst downgrades, respectively, and find that short sellers do not tend to trade before these events. Similarly, Daske, Richardson, and Tuma (2005) look at short selling around earnings 5

announcements and management forecast announcements and find no evidence that short sale transactions concentrate prior to bad news events. Finally, Nagel (2005) looks at the cash flow news implied by a vector auto regression and finds an asymmetric effect on returns, indicating that short sellers help incorporate news into prices when short selling is not constrained. While the above papers identify patterns in short selling around a handful of specific corporate new events, the current paper aims to uncover patterns in short sellers’ trades around all types of corporate news events. Doing so allows us to speak more generally about short sellers’ behavior around new releases of public information. In particular, using a list of all corporate news events, we can sort the universe of trading days into those with and without news and examine the differential performance of short sellers surrounding news events.

A. Short Sellers’ Trading Patterns Several papers compare the trades of short sellers to the trades of other market participants. There are several dimensions over which trades can be compared. Much of the recent literature focuses on the profitability of trades, which, roughly speaking, can be measured using the performance of a stock’s price after the initiation of a short sale. One of the most widely cited results in this vein of the literature is found in Asquith and Muelbroek (1995), who show that high short interest precedes negative future returns, consistent with informed trading. Similarly, Asquith, Pathak, and Ritter (2005) show that when short selling is constrained and there are relatively diverse opinions, abnormally high short interest can precede negative future returns. Using transaction data at a higher frequency, Boehmer, Jones, and Zhang (2008) find that heavily shorted stocks significantly underperform lightly shorted stocks, and Diether, Lee, 6

and Werner (2008) show that not only do prices follow short selling, but short selling also follows prices, that is, short sellers tend to short after price run ups. These results further indicate that short sellers may have an informational advantage.2 In sum, the above work establishes that the performance of short sellers’ trades indicates that short sellers are informed traders. Our paper contributes to this literature by asking how short sellers come to enjoy an informational advantage in the first place.

B. Public News While a large literature examines volume and return phenomena around specific news events (e.g., earnings announcements, mergers, and dividend initiations and omissions), a more recent literature considers such phenomena around any corporate news event. Categorizing all Wall Street Journal stories between 1973 and 2001, Antweiler and Frank (2006) find that return responses vary widely across news categories, although they find evidence of overreaction (return reversal) on average. Also using a database of all news events, Tetlock (2008) finds evidence of even stronger return reversal following repeated news events, consistent with the idea that investors overreact to “stale” news stories. Several studies using comprehensive news databases examine whether well-known asset pricing anomalies are related to news. Chan (2003) considers the momentum anomaly among stocks with and without recent news and finds evidence of price momentum only among news stocks. Similarly, Vega (2006) finds more earnings momentum among stocks with high differences of opinion on news days. 2

A closely related dimension of research is whether short sellers’ trades reveal information to other market participants. In other words, are short sellers’ trades news worthy in and of themselves? Senchack and Starks (1993) show that abnormally large short interest announcements have small but significant negative returns. Similarly, Aitken et al. (1998) show that short sales are followed by price declines within 15 minutes on the Australian Stock Exchange.

7

More recently, researchers have asked whether the content of news stories contains value-relevant information. Tetlock, Saar-Tsechansky, and Macskassy (2008) and Engelberg (2008) show that, indeed, the qualitative content of the information contained in news stories can predict both earnings surprises and short-term returns. These findings support the idea that there is value-relevant or “soft” information in news stories that is not immediately impounded into prices. To summarize, this literature highlights the importance of looking at more than one news category when assessing the behavior of short sellers. Moreover, it shows that the information content of news leaves room for traders with different information processing abilities to arrive at different conclusions about the value relevance of the news. Our work builds on these findings by analyzing the universe of corporate news events in the U.S. over our sample period, and by asking whether, in our sample, information processing ability plays a role in the performance of short sellers’ trades.

C. Public News and Informed Trading There are two views regarding the relation between the trading patterns of skilled investors and the release of public news items such as the articles contained in the Dow Jones archive. Under the first view, public information does not provide traders with an information advantage, that is, managers who rely on public information (rather than generate private information) are low-skilled. Consistent with this view, Kacperczyk and Seru (2007) estimate managers’ reliance on public information (RPI) as the R-squared of a regression of percentage changes in fund managers’ portfolio holdings on changes in analysts’ past recommendations and 8

find that fund managers with low RPIs (low reliance on public information) perform better than fund managers with high RPIs (high reliance on public information). Under the alternative view, the public release of information presents trading opportunities for skilled processors of information, that is, when news is released, traders with superior information processing skills can convert this news into valuable information for trading. Earnings announcements, for example, are often accompanied by lengthy documents and conference calls that are scrutinized by information processors. Those traders who show exceptional skill in converting such data into value-relevant information are rewarded with superior returns on event-driven trades. Evidence consistent with this view comes from studies that attempt to look at the textual content of news and firm announcements. Specifically, Tetlock et al. (2008), Engelberg (2008), Demers and Vega (2008), and Feldman et al. (2009) all show that the content of corporate news predicts returns, which is consistent with the view that information processing skills can generate superior returns. Our paper sheds light on the above debate by finding additional evidence in support of the second view, that is, by showing that trades occurring after the release of news stories can be more profitable than trades in non-news periods.

II. Hypotheses & Methodology A. Hypothesis Development In this section, we formalize many of the ideas introduced in the beginning of the paper. Our first set of hypotheses concerns the timing of trades while the second set concerns the

9

profitability of trades. Finally, we have two sets of hypotheses that aim to explore the source of short sellers’ profitability. The timing of trades is one of the areas in which short sellers may differ from other traders. Prior research finds some evidence that short sellers trade before public information is released (e.g., Karpoff and Lou (2009) and Christophe, Ferri, and Angel (2004)). Similarly, the Securities and Exchange Commission has suggested that short sellers spread “false rumors” in an effort to manipulate firms. Our first set of hypotheses thus seeks to empirically test this view. Formally: H1a: In the presence of news events, short sellers trade before other traders. This hypothesis is an alternative to the null hypothesis that there is no difference in timing: H10: In the presence of news events, the timing of short sales is the same as the timing of other trades. We next turn to the profitability of short sellers’ trades around news events.

The

literature is split as to whether news events increase or decrease asymmetric information, thereby increasing or decreasing the profitability of informed trades. On the one hand, many papers model news events as points in time associated with reduced information asymmetry (e.g., Korajczyk, Lucas, and McDonald (1991) and Diamond and Verrecchia (1987)). If news events do indeed reduce asymmetric information, the trades of informed traders (e.g., short sales) should be less profitable on news days. On the other hand, other papers suggest that public news events are subject to differential interpretations by traders (e.g., Rubenstein (1993) and Kandel and Pearson (1995)). Under this view, public information events present profitable trading 10

opportunities for skilled information processors, and thus the trades of informed traders (e.g., short sellers) should be more profitable on news days. This discussion leads to the following set of hypotheses: H2a: Short sales are less profitable around news announcements. H2b: Short sales are more profitable around news announcements. These hypotheses rest against the backdrop of the null hypothesis which states short sales are as profitable around news events as they are at other times: H20: Short sales are no more or less profitable around news events. Since our empirical work finds that short sales are more profitable around news events, we also explore why profitability increases. While the literature finds that news events create trading opportunities for informed traders (e.g., Engelberg (2008), Demers and Vega (2008)), other potential explanations exist. The first alternative explanation posits that some buyers make systematic mistakes around news events (e.g., Antweiler and Frank (2006)), and that these mistakes are reflected in market makers’ offsetting short sales. We formalize this idea in our third set of hypotheses: H3a: The profitability of short sales comes from market makers’ hedging trades. H30: The profitability of short sales comes from market maker and non-market maker trades. Another alternative explanation relates to liquidity. Given the increase in volume around news events, news events may provide a trading opportunity for those traders for whom liquidity is an important factor in a trade’s profitability (e.g., Kyle (1985) or Admati and Pfleiderer 11

(1988)). As a result, the perceived profitability of short sales around news events may have nothing to do with information; rather, short sellers may simply be trading around news events because news events create liquidity, which leads in turn to profitable trades. This relation between news events and liquidity is the basis for our fourth and final set of hypotheses: H4a: The profitability of short sales around news events is due to the increased liquidity that news events provide. H40: The profitability of short sales around news events is not a result of the liquidity that news events provide.

B. Data To test the hypotheses developed above, we employ two main databases. The first database contains information on short sales while the second contains news articles from the Dow Jones News Service.

B.1. Short Sales Information on short sales transactions comes from the NYSE TAQ Regulation SHO database. Regulation SHO was adopted by the SEC in June of 2004 to establish new rules governing short sales in equity transactions and to evaluate the effectiveness of price test restrictions on short sales. As one consequence of regulation SHO, transaction-level short sales data were publicly disclosed for the period January 3, 2005 through July 6, 2007. The NYSE TAQ regulation SHO database therefore contains data for all short transactions that were 12

reported on the NYSE during this period. Specifically, the database contains the stock ticker, the date and time of the transaction, the number of shares traded, the execution price, and an indicator that denotes whether the transaction was exempt from price test rules. One of the reasons a short sale transaction could be classified as exempt is that it was made by market makers engaged in bona fide market making activity. The exempt indicator has thus been used to separate trading by market makers from trading by non-market makers (e.g., Evans et al. (2009), Christope, Ferri, and Angel (2004), Boehmer, Jones, and Zhang (2008), Chakrabarty and Shkilko (2008)).3 However, when regulation SHO was implemented, a group of randomly selected stocks was selected to be part of a pilot study for which the exempt/non-exempt classification was no longer required. We exclude these pilot firms when using the exempt indicator variable in our analyses (i.e., Tables VI and VII).4 For the purposes of our analysis, we aggregate the transaction data at the daily level, and we use the TAQ master files to add CUSIPs to the database. We then use the CRSP Daily Stock Event file to add PERMNOs to the database. Finally, we add returns, bid price, ask price, total volume, and shares outstanding from CRSP. Using this data, we calculate the Amihud (2002) Illiquidity measure defined as |retit| / volumeit, where volumeit is the dollar volume, and we calculate the daily bid-ask spread as a percentage of the closing mid-price. In addition, we also add information on the daily volume weighted rebate rate for equity loans in each stock over the sample period. The rebate rate for an equity loan is the rate at which interest on collateral is rebated back to the borrower. This rate is related to the cost of shorting a

3

For example, NASD NTM 06-53 notes that “Rule 5100(c)(1) provides an exception to the bid test for short sales by a market maker registered in the security in connection with bona fide market making activity.” 4 Details regarding the regulation SHO pilot study, including a list of firms involved, are available on the SEC website: http://www.sec.gov/rules/other/34-50104.htm. Our results are robust to the inclusion of the regulation SHO pilot firms.

13

stock. Our data on rebate rates come from a proprietary database on equity loan transactions as described in Kolasinski, Reed, and Ringgenberg (2010). The data are compiled by a third-party provider that is both a market maker in the equity loan market and a data aggregator for major equity lenders.

B.2. Dow Jones Archive To compile our sample of news events, we use the Dow Jones archive as in Tetlock (2009). This archive contains all Dow Jones News Service stories and Wall Street Journal stories over our 2005 to 2007 sample period. The Dow Jones database also contains subject codes that identify the information content of each news article; for example, there is a code to indicate whether an article contains information about insider stock sales. We adopt Dow Jones’ subject categorizations. Starting with the database described in Tetlock (2009), we have 71 news categories. However, many of these subject codes are general codes that do not provide valuable information about the content of a news article. For example, nearly every article in the database has the code Company News assigned to it, in addition to a more specific news code. We remove these general codes from our analysis to obtain a final list of subject codes that contains 39 different news categories.5 The resulting news database contains a unique firm identifier, subject codes, a dummy variable that takes the value of one if a story was released in multiple pieces over the news day, and two sentiment score variables that indicate whether a story contains negative words in the 5

Specifically, after computing the correlations between subject codes, we exclude subject codes if their correlation with a more specific news category exceeds 80%. We also drop news categories that are associated with fewer than 1,000 news events over the entire sample (see Table I for the frequency of each news event in our sample).

14

headline and body of the text. The first sentiment variable is constructed using the Harvard-IV-4 dictionary as in Tetlock (2007) and Engelberg (2008) while the second sentiment measure uses the negative word list developed by Loughran and McDonald (2009).

In both cases, we

construct the sentiment score as the sum of the number of negative words in an article’s headline and body divided by the sum of the total number of words in the headline and body. We use the unique firm identifier to match the news data to the short sales database. The resulting database has 1,888,868 observations over the period January 3, 2005 to July 6, 2007. Table I contains summary statistics for the combined database. The mean number of articles per firm-day is 1.10. However, there is substantial cross-sectional variation in this number, and larger firms typically have more news articles on a given day. Certain news categories also appear much more often than others. For example, the category High Yield Issuers appears 173,357 times in the database while the category 10K appears only 1,320 times. To address the potential issue of news clustering, when we conduct category-specific analyses we remove stories that are within 30 days of another story in the same category.

III. Analyses and Results In this section we explore how short sellers differ from other traders. We begin by asking whether short sellers respond to news before other market participants. We find that short sellers tend to trade at the same time as other traders, and when they do not, they tend to trade less than other market participants in the days leading up to a news event while they tend to trade more after news events occur. These results suggest that short sellers do not anticipate news. Next, we ask whether short sellers’ trades are more profitable than other trades, consistent with a 15

superior ability to process news, and we find evidence in favor of this view. In a third set of tests we analyze which types of information are associated with short sellers’ profitability. Finally, we conduct a matched sample portfolio approach to shed additional light on the economic impact of news-based short sales strategies.

A. Do Short Sellers Anticipate News? One way in which short sellers may differ from other traders is in the timing of their trades. There is some evidence that short sellers anticipate bad news announcements (e.g., Angel, Ferri, and Christophe (2004) and Karpoff and Lou (2009)). However, these findings correspond to specific types of corporate events. Here we seek to shed light on short sellers’ timing behavior around all types of news events in our sample period. To determine the extent of short sales timing around news events, in Figure 1 Panel A we plot daily short sales volume (solid line), total volume (dashed line), and the ratio between the two (dotted line) in calendar time around our universe of news events. The basic result is readily apparent: short sellers trade when other traders do. In other words, the graph provides some visual evidence in support of the null hypothesis, H10 and against the alternative H1a. Clearly, all traders respond to news, as there is a significant increase in volume on the news event day and on surrounding days. However, the ratio of short sales to total volume is nearly constant over the news period, with no significant change in the ratio around news events. This result suggests that short sellers do not uncover and trade on information before it becomes publicly available.

16

Of course, in line with the prior research above, it may be the case that short sellers respond more to certain types of news, particularly bad news. Thus, in Panels B and C of Figure 1, we focus only on negative news events, where negative news events are defined using the Harvard-IV-4 dictionary (see Section II.B) and the Loughran and McDonald (2009) negative word list. The results are largely unchanged, indicating that the timing of short sellers’ response to news does not depend on whether the news is bad. Next, we formally examine the timing of short sellers’ trades around news events. We begin by regressing short sales volume on an indicator variable that takes the value one if any news event occurs, and zero otherwise. To control for the well-documented response of short sellers to past returns (e.g., Diether, Lee, and Werner (2008)), we include two lags of daily returns. In order to understand the timing of short sales transactions around news events, we run six different specifications by varying the timing of the dependent variable relative to the news event. The results, shown in Table II, suggest that short sellers do trade more before, during, and after news events (Panel A), however relative to all investors (Panel B), they actually trade less before and during news events. Following news events, short sellers do trade more however their trading volume is not statistically different from the trading volume of all investors. We next examine the trading patterns of short sellers around specific news categories. Table III presents results from a regression of short sales volume on two lags of daily returns and a set of indicator variables representing each of the news categories. The results indicate that for the majority of news categories, short sellers respond at the same time as other traders. More specifically, for a given news event, when we compare the coefficient estimates for regressions on short selling after the news event to estimates for regressions on short selling before the news

17

event, we find that the coefficient estimates are largely the same. In other words, we find statistical evidence rejecting the alternative hypothesis H1a in favor of the null hypothesis H10. However, there are a number of interesting exceptions. For both types of earnings news stories, Earnings and Earnings Projections, there is more short selling after the news event than before the news event, a result largely consistent with the findings of Angel, Ferri, and Christophe (2004). The statistically significant estimate of 0.0049 in the t+2 specification for Earnings indicates that there is a 0.49% increase in short selling as a percentage of total volume two days after news of this type is reported. This late response is also apparent in news stories about joint ventures and product distribution. In contrast, stories about Leveraged Buyouts show the opposite pattern. The estimate of 0.0145 in the t-1 specification indicates that the short selling ratio increases 1.45% on the day before news stories about leveraged buyouts, and the statistically significant coefficient estimate on After Minus Before indicates that, relative to the two-day period before the news event, the short selling ratio decreases 2.2% in the two-day period after the news event. Table IV presents results from a similar setup, but here the dependent variable is raw daily short sales volume rather than short sales volume scaled by total volume. The change in raw sales volume can be interpreted as a direct measure of the average increase or decrease in the number of shares traded in response to news events. Our findings are qualitatively unchanged. For instance, in the most extreme case, news days that contain Money Market News are associated with a statistically significant increase of 291,721 additional shares sold short. Despite the natural interpretation of these results, however, they are not as meaningful as the scaled results discussed above. The reason is that over our sample period there is a strong

18

market-wide trend towards greater short volume, and thus in this analysis After Minus Before is statistically significant in approximately half of the news categories.6 Overall, the evidence suggests that short sellers generally trade at the same time as other traders, and in those instances in which they show different timing, short sellers tend to trade after other traders, not before. In other words, the results suggest that the previously documented information advantage of short sellers (e.g., Boehmer, Jones, and Zhang (2008) and Asquith, Pathak, and Ritter (2005)) does not stem from an ability to anticipate news.

B. Do Short Sellers Have Superior Information Processing Ability? Given our finding above, in this subsection we ask whether short sellers’ informational advantage derives from an alternative source, namely, a superior ability to process the information contained in publicly available news. To answer this question as directly as possible, we begin by replicating Table IV of Boehmer, Jones, and Zhang (2008), shown in our Table V below. Specifically, we compute 20day rolling returns (i.e., t+1 to t+21) from January 3, 2005 through July 6, 2007 and we regress these returns on the Short Volume Ratio on day t, which is defined as daily short volume divided by total volume. The Boehmer, Jones, and Zhang (2008) result comes through strongly in these results: in each of the specifications, Short Volume Ratio is negative and statistically significant, indicating that when there is an increase in short sales, future prices decrease. However, given our previous results, we might expect this pattern to be stronger among firms for which news is 6

Mean daily short volume increases from 120,910 shares in 2005 to 150,681 shares in 2007, and the increase is statistically significant. This steady increase in short volume through time offers an explanation for the fact that abnormal volume in Figure 1 is generally slightly above 1.

19

released when short volume is high. To test for this effect, we include the indicator variable News Event, which takes the value one if there is news on day t and zero otherwise. We also include contemporaneous returns as a control for the information content in the news, and we include two days of lagged returns to control for the tendency of short sellers to trade following recent price increases as documented by Diether, Lee, and Werner (2008). Note that because news coverage is correlated with firm characteristics such as size and institutional ownership (e.g., Chan (2003), Vega (2006), Engelberg (2008), and Fang and Peress (2009)), our empirical design is meant to estimate the effect of news within firms rather than across firms. We thus follow Skoulakis (2005) and apply the Fama-Macbeth approach to firms: we first run a timeseries regression for each firm; we then take the average of the coefficients and use the standard deviation to estimate standard errors. The results, shown in Table V, provide strong evidence on the informational advantage of short sellers. Specifically, the coefficient estimate of -0.0053 on Short-News Interaction in Model 5 is negative and statistically significant, indicating that among stocks with high short volume, those with news have significantly more negative future returns than those without news, providing support for hypothesis H2b.7 Even after controlling for the contemporaneous effect of returns, the coefficient on Short Volume Ratio is still negative at the 5% level, in other words, the Boehmer, Jones, and Zhang (2008) finding that short volume leads negative returns continues to hold. Our findings thus provide new insight into the source of short sellers’ informational advantage. In particular, we find that the previously documented relation between

7

In unreported results, we add the number of negative words (a measure of the sentiment in the article) as a control variable. This variable does not change the general magnitude or statistical significance of the results, indicating that the findings are not driven by either very good or very bad news events.

20

short volume and returns is more than twice as strong for those stocks that have a public news event.8

B.1 Alternative Interpretations We interpret our evidence that the profitability of short sales is higher (lower) on (off) news days as evidence that short sellers process news announcements very well. However, there are some possible alternatives that also explain the evidence: (1) short sales may be more profitable on news days because they provide liquidity to traders who make systematic mistakes around news events (i.e., Hypothesis H3a) and (2) short sales may be more profitable on news days because the cost of shorting is lower on news days (i.e., Hypothesis H4a). The following results will show that neither alternative is supported by the evidence. One drawback to using total short sales volume as a measure of short selling is that some short sales are generated as a result of market making – to the extent that some buyers make systematic mistakes, the corresponding short sales are simply offsetting positions, not informed trades. Thus, with the aggregate measure of short sales volume used in Table V, we cannot distinguish the effect of short sales that arise in response to counterparty purchases from the effect of shorts that arise for the purpose of gaining negative exposure. This raises the question of whether our results in Table V can be attributed to informed trading. To address this concern,

8

The Boehmer, Jones, and Zhang (2008) result can be thought of as a high-frequency analog of the results in Asquith and Muelbroek (1996) and Asquith, Pathak, and Ritter (2005). This second set of papers measures short trading with short interest instead of short volume, and they use future returns that are measured over longer periods. Although we would like to examine the relation between news and short sellers’ advantage in the context of these short interest-based findings, there is an econometric challenge in making a direct comparison. Specifically, news in our database is marked with daily time stamps, so either we would have to aggregate news to match the monthly frequency of short interest or we would have to throw out much of our news data. It is not clear how a reduction in the frequency of the news variable would change expectations about the short positions.

21

we take advantage of a unique feature of the data, namely, the exempt versus non-exempt classification of trades. This classification allows us to separate shorts into market making and non-market making (i.e., client) trades.9 Tables VI and VII report the results for non-exempt and exempt trades, respectively. In Table VI the statistically significant coefficient estimate of −0.0075 in Model (5) indicates that high short volume is a significant predictor of low future returns. Moreover, the magnitude on short volume is 44% larger than the corresponding coefficient for total short sales volume in Table V, which suggests that the ability of short sales to predict future returns is particularly strong for non-market-making trades. We also see that the short-news interaction estimate of −0.0056 is significantly negative, indicating that non-market makers’ shorts are 75% (i.e., −0.0056/−0.0075) more profitable in the presence of news events than at other times.

In

contrast, the results in Table VII indicate that market makers’ trades are not particularly well informed: the short volume ratio loads as a positive predictor of price, indicating that market makers’ shorts are actually associated with positive future returns on average.10 Further, there is no differential effect of market makers’ trading in the presence of news. Overall, the evidence in this subsection suggests that the most informed short sales are made for the purpose of gaining negative exposure, and that these trades are particularly well informed in the presence of recent news events. In other words, we find evidence rejecting H3a in favor of H30.

9

Anecdotal evidence suggests that the exemption is sometimes abused, but only in one direction: trades may be inappropriately marked as exempt when they are not. Since the exemption removes potential restrictions, it is unlikely that exempt trades would ever be inappropriately marked as non-exempt. In other words, exempt trades may include client trades, but non-exempt trades are unlikely to include market maker trades. 10 Even though the magnitudes of the coefficients vary between Table V and Table VI, the economic impact is of the same order of magnitude. Specifically, a one standard deviation increase in short volume among non-exempt trades leads to a 0.200% decrease in future returns, while a one standard deviation increase in short volume among nonexempt trades is associated with a 0.154% increase in future returns.

22

Finally, in Tables VIII, IX and X we consider the possibility that short sales are profitable on news days because news days provide short sellers an opportunity to trade at a lower cost. Under this view, short sellers have an information advantage well before a news announcement. However, they cannot execute their trades if transaction costs are high or liquidity is low. Since news days tend to have higher than normal volume, it may be that these days are low cost and/or high liquidity days on which short sellers can execute their trades. As a result, if news events provide opportunities to transact at lower costs then short sellers may appear to have more profitable trades around news announcements even if the news events themselves are not the source of an information advantage. This story, however, requires that the costs of short selling are lower around news announcements. Figure 2, which presents a plot of rebate rates, the Amihud (2002) Illiquidity measure, and bid-ask spreads around news events, suggests that this is not the case. Both rebate rates and the Amihud measure show no decrease in costs on news days and moreover, bid-ask spreads actually increase on news days. More specifically, Table VIII tabulates the mean values of these variables around news announcements and both rebate rates and the Amihud Illiquidity measure have no statistically detectable change around news announcements while bid-ask spreads actually rise by approximately 5% on news days. Moreover, the increase in bid-ask spreads, which results in higher transaction costs on news days, is consistent with existing models of market maker behavior in the presence of informed traders (e.g., Glosten and Milgrom (1985) and Kyle (1985)). Although there is no evidence of decreasing transaction costs on news days, we further test this alternative by including measures of transaction costs in our main regressions. These results, shown in Table IX (Amihud illiquidity) and Table X (bid-ask spread), provide additional 23

evidence that short selling remains twice as profitable on news days even after controlling for changing transaction costs on news days. In other words, we find evidence rejecting H4a in favor of H40.

C. Price Responses by News Category In an extension to the above analysis, in this subsection we ask whether short sellers’ information processing ability is uniformly strong across news categories.

To get at this

question, we repeat the analysis in Table V separately for each news category. Specifically, for each news category we run a regression in which the dependent variable is the compound return from the first to the twentieth trading day after the news event, and the main independent variable is short vol / market vol, which is the amount of short selling relative to total volume on the day of the news event. Since the type of news (good or bad) may have some effect on future returns (e.g., Bernard & Thomas (1989)), we attempt to control for whether the news is good or bad by including the event-day return on the right-hand side of the specification. The results, shown in Table XI, indicate that short sellers have some ability to identify trades that are likely to be profitable around certain news events. Specifically, we find that the coefficient estimates on short vol / market vol are significantly negative for 12 of the 39 news categories; of these, five are statistically significant at the 1% level (Corporate Restructurings, Earnings, Earnings Projections, New Products & Services, and Stock Ownership), and nine are significant at the 5% level (Corporate Restructurings, Divestitures or Asset Sales, Earnings, Earnings Projections, Initial Public Offerings, Management Issues, New Products & Services, Research and Development, and Stock Ownership). As a further test of statistical significance, 24

we conduct a Fisher test of combined probability to determine whether the cross-sectional distribution of the p-values from each regression differs significantly from a uniform zero-one distribution. The Fisher test rejects this null at the 1% level of significance across all news categories, suggesting that the coefficient on short volume is statistically different from zero for the cross-section. We also find that many of the categories that are statistically significant are the same categories identified in Subsection A as the categories in which short sellers’ trades are measurably later than other investors’ trades (e.g., Earnings, and Earnings Projections). Taken together, these results indicate that when short selling predicts future returns, short sellers appear to be making profitable trades. This evidence lends further support to the idea that short sellers’ informational advantage stems from superior ability to process publicly available information.

D. Matched Sample Portfolio Approach So far we provide evidence that short selling is more informative on news event days. In this subsection we shed light on the economic impact of news-based short selling strategies using a portfolio approach. This approach recognizes that the presence of news is likely correlated with firm characteristics and that certain categories of news may be more relevant for some firms than for others. In other words, since news is strongly related to several firm characteristics, we cannot simply sort on news. Moreover, news coverage is highly persistent: firms that have many news articles in the Dow Jones archive in one year are likely to have many articles in following years. Thus, in order to conclude that news, rather than a particular firm characteristic, is driving the differential returns we observe we need to compare two firms that are identical apart from the 25

fact that one firm has a news event while the other firm does not. We do this using a matched sample portfolio approach. Our approach is based on forming portfolios of stocks around news events. Because previous research indicates that firm characteristics may affect future returns, we implement a control sample methodology to control for these previously documented effects. Specifically, for every stock with a news event, we identify a control stock that is the closest match in the following four dimensions: bid-ask-spread, institutional ownership, market capitalization, and number of news events over the previous month.

We match by selecting the stock that

minimizes the sum of the rank differences in each of these categories. Furthermore, to eliminate potentially contaminating competitive effects (e.g., Slovin, Sushka, and Bendeck (1991), Chen, Ho, and Ik (2005), and Hsu, Reed, and Rocholl (2009)), we require that control firms and sample firms be members of different Fama-French 48 industries.11 The analysis yields results for each of the 39 news categories. Figure 3 presents the results for three categories as examples. In the Dividends category, we see that among firms with dividend news, firms with high short volume have significantly lower returns than firms with low short volume. This difference is approximately 4.39% at the one-year point. In contrast, the control sample shows similar returns across high short volume stocks and low short volume stocks. Table XII summarizes the detailed results of this analysis. The economic significance of news-based short sales becomes apparent when we compare differences in portfolio returns. For example, if an investor were to sell a portfolio of stocks with high short selling and buy a 11

In unreported results, we use the Fama-French 12 industry classifications instead of the Fama-French 48 classifications. The results are not qualitatively different.

26

portfolio of stocks with low short selling on the day that Product Distribution news is released, that investor would earn an excess annual return of 6.50%. The same strategy for a matched portfolio of no-news stocks would return

5.74% over the period, yielding a difference of

12.24% annually between the two strategies. In fact, this strategy yields positive excess returns in 34 out of our 39 news categories, with some news categories yielding annualized excess returns of over 10%. The statistically significant 2.89% return for the mean excess return difference indicates that not only do short sellers have a significant advantage over other traders, but their advantage comes largely from their ability to process and trade on news events. To summarize, this analysis shows that the inverse relation between short volume and future returns is strongest around news events, whereas during non-news events this relation may be insignificant or even go in the other direction. These results lend additional support to our main finding that the previously documented informational advantage of short sellers is driven in large part by short sellers’ superior ability to process information contained in publicly available news.

IV. Conclusion Previous research documents that short sellers are informed traders (e.g., Boehmer, Jones, and Zhang (2008) and Asquith, Pathak, and Ritter (2005)). Yet we know little about the source of short sellers’ informational advantage. This paper seeks to fill this gap by investigating the following questions: To what extent are short sellers able to anticipate news events? Are short sellers better able to process and react to news? And, are short sellers’ trades particularly profitable around specific categories of news? 27

To address these questions, we combine a

database of all public news events in the U.S. with a database of short sale trades over the same sample period. We find that, in general, short sellers trade at the same time as other traders. Specifically, for most news categories the ratio of short sales to total volume is nearly constant over news periods, with no significant change in the ratio around news events. However, we do find some differences between the timing of short sellers’ trades and the overall market: for news stories about analysts’ comments and ratings, earnings, earnings projections, joint ventures, and stock ownership, there is a significant increase in short selling after the news story. This finding suggests that, like other traders, short sellers trade on publicly available information, and hence their informational advantage is not due to an ability to uncover or anticipate information before it becomes public. Given the result that short sellers’ advantage is not due to timing, we next ask if it could be due to a superior ability to process the information available in public news stories. We find supportive evidence. In particular, we find that across all types of news, short selling predicts future returns: even after controlling for the unconditional relation between short selling and news (e.g., Boehmer, Jones, and Zhang (2008)), short selling’s predicative effect on future returns is more than twice as strong in the presence of news. This result is not a reflection of persistent mistakes by buyers, as the most informed short sales are not from market makers but rather from clients, and these client shorts are particularly well informed in the presence of news. Moreover, we find no evidence that short selling around news events is more profitable because of the liquidity that news events provide, as the bid-ask spread is actually found to increase (by approximately 5%) on news days. We further find that the predicative effect is strongest for nine categories of news (Corporate Restructurings, Divestitures or Asset Sales, Earnings, Earnings 28

Projections, Initial Public Offerings, Management Issues, New Products & Services, Research and Development, and Stock Ownership), and that many of these categories are the same categories for which short sellers’ timing follows the overall market. Finally, recognizing that the presence of news is likely to be correlated with firm size, and that certain categories of news may be more relevant for some firms than for others, we conduct an experiment in which each firm’s response to a news event is matched by a control firm’s response on the same day. We find that across all news categories, short sellers’ advantage in predicting returns concentrates in firms with news. In sum, we show that, on average, short sellers’ advantage is not due to an ability to influence the public’s perception of value, as recently suggested by the Securities and Exchange Commission.12 Rather, we find that short sellers generally trade when other traders do, and to the extent that the timing of their trades differs, short sellers actually trade after other traders. We further find that short sellers’ ability to predict future negative returns is concentrated around news events. Thus, by connecting short sellers’ trading patterns with news releases, we show that short sellers’ trading advantage derives primarily from their superior ability to analyze publicly available information.

12

Short sellers were accused of “distort and short” schemes in “What the SEC Really Did on Short Selling” by Chairman Christopher Cox, 24 July 2008, The Wall Street Journal.

29

REFERENCES Admati, Anat R. and Paul Pfleiderer, 1988, A Theory of Intraday Patterns: Volume and Price Variability, The Review of Financial Studies, 1, 3-40. Antweiler, Werner and Murray Z. Frank, 2006, Do U.S. stock markets typically overreact to corporate news stories?, Working Paper, University of British Columbia. Asquith, Paul, and Lisa Meulbroek, 1995, An empirical investigation of short interest, Unpublished Working Paper, M.I.T. Asquith, Paul, Parag A. Pathak, and Jay R. Ritter, 2005, Short Interest, Institutional Ownership, and Stock Returns, Journal of Financial Economics 78, 243-276. Aitken, Michael J., Alex Frino, Michael S. McCorry, and Peter L. Swan, 1998, Short sales are almost instantaneously bad news: Evidence from the Australian Stock Exchange, Journal of Finance 53, 2205-2223. Bernard, Victor L. and Jacob K. Thomas, 1989, Post-earnings-announcement drift: delayed price response or risk premium?, Journal of Accounting Research 27, 1-36. Boehmer, Ekkehart, Charles M. Jones, and Xiaoyan Zhang, 2008, Which Shorts are Informed?, Journal of Finance 63, 491-527. Chakrabarty, Bidisha, and Andriy Shkilko, 2008, Information Leakages in Financial Markets: Evidence from Shorting around Insider Sales, Working Paper. Chan, Wesley S., 2003, Stock price reaction to news and no-news: drift and reversal after headlines, Journal of Financial Economics 70, 223-260. Chen, Sheng-Syan, Kim Wai Ho, and Kueh Hwa Ik, 2005, The Wealth Effect of New Product Introductions on Industry Rivals, Journal of Business 78, 969-996. Christophe, Stephen E., Michael G. Ferri, and James J. Angel, 2004, Short-Selling Prior to Earnings Announcements, Journal of Finance 59, 1845-1875. Christophe, Stephen E., Michael G. Ferri, and Jim Hsieh, 2009, Informed trading before analyst downgrades: Evidence from short sellers, Journal of Financial Economics 95, 85-106. Daske, Holger, Scott A. Richardson, and A. Irem Tuma, 2005, Do Short Sale Transactions Precede Bad News Events?, Working Paper. Diamond, Douglas W. and Robert E. Verrecchia, 1987, “Constraints on short-selling and asset price adjustment to private information,” Journal of Financial Economics, 18, 277–311. Diether, Karl B., Kuan-Hui Lee, and Ingrid M. Werner, 2008, Short-sale Strategies and Return Predictability, Review of Financial Studies 22, 575-607. 30

Engelberg, Joseph, 2008, Costly Information Processing: Evidence from Earnings Announcements, Working Paper, University of North Carolina. Evans, Richard, Chris Geczy, David Musto and Adam Reed, 2009, “Failure is an Option: Impediments to Short-Selling and Options Prices”, The Review of Financial Studies 22(5), 2009. Fang, Lily and Joel Peress, 2009, Media Coverage and the Cross-Section of Stock Returns, Forthcoming in the Journal of Finance. Feldman, Ronen, Suresh Govindaraj, Joshua Livnat, and Benjamin Segal, 2008, The incremental information content of tone change in management discussion and analysis, Working paper, INSEAD. Fox, Merritt B., Lawrence Glosten, and Paul Tetlock, 2009, Short Selling and the News: A Preliminary Report on an Empirical Study, Working Paper. Gervais, Simon, Ron Kaniel, and Dan Mingelgrin, 2001, The High-Volume Return Premium, Journal of Finance 56, 877-919. Glosten, Lawrence R. and Paul R. Milgrom, 1985, Bid, ask and transactions prices in a specialist model with heterogeneously informed traders, Journal of Financial Economics 14, 71100. Hsu, Scott H.C., Adam V. Reed, and Jörg Rocholl, 2009, The new game in town: competitive effects of IPOs, Forthcoming in the Journal of Finance. Kacperczyk, Marcin and Amit Seru, 2007, Fund Manager Use of Public Information: New Evidence on Managerial Skills, Journal of Finance 62, 485-528. Kandel, Eugene and Neil D. Pearson, 1995, Differential interpretation of public signals andtrade in speculative markets, Journal of Political Economy 103, 831-872. Karpoff, Jonathan M. and Xiaoxia Lou, 2009, Short sellers and financial misconduct, Working Paper. Kolasinski, Adam C., Adam V. Reed, and Matthew C. Ringgenberg, 2010, A Multiple Lender Approach to Understanding Supply and Search in the Equity Lending Market, Working Paper. Kyle, Albert S., 1985, Continuous auctions and insider trading, Econometrica 53, 1315-1336. Loughran, Tim and Bill McDonald, 2009, When is a Liability not a Liability? Textual Analysis, Dictionaries, and 10-Ks, Working Paper, University of Notre Dame.

31

Nagel, Stefan, 2005, Short sales, institutional investors and the cross-section of stock returns, Journal of Financial Economics 78, 277-309. Rubinstein, Ariel, 1993, On Price Recognition and Computational Complexity in a Monopolistic Model, Journal of Political Economy 101, 473-84. Senchack, A.J., Jr., and Laura T. Starks, 1993, Short-sale restrictions and market reaction to short-interest announcements, Journal of Financial and Quantitative Analysis 28, 177194. Skoulakis, Georgios, 2005, Assessment of Asset-Pricing Models using Cross-Sectional Regressions, Working paper, Northwestern University. Slovin, Myron B., Marie E. Sushka, and Yvette M. Bendeck, 1991, The Intra-Industry Effects of Going-Private Transactions, Journal of Finance 46, 1537-1550. Tetlock, Paul C., 2007, Giving content to investor sentiment: The role of media in the stock market, Journal of Finance 62, 1139-1168. Tetlock, Paul C., 2008, All the News That’s Fit to Reprint: Do Investors React to Stale Information?, Working Paper. Tetlock, Paul C., 2009, Does Public Financial News Resolve Asymmetric Information?, Working Paper. Tetlock, Paul C., M. Saar-Tsechansky, and S. Macskassy, 2008, More Than Words: Quantifying Language to Measure Firms' Fundamentals, Journal of Finance 63, 1437-1467. Vega, Clara, 2006, Stock price reaction to public and private information, Journal of Financial Economics 82, 103-133.

32

Figure 1 Volume around News Events Figure 1 displays short volume, total volume, and the ratio of short volume to total volume for the 15 days before and after news events. Short volume and total volume are scaled by their mean values over the period -16 to -30. Panel A displays volume around all news events. Panels B and C display volume for negative news events only. Version 1 (Panel B) uses a negative sentiment variable that was constructed using the Harvard-IV-4 Dictionary as in Tetlock (2007) and Engelberg (2008) while version 2 (Panel C) uses a sentiment measure that employs the negative word list developed by Loughran and McDonald (2009). Panel A: All News Events

Volume

All News Events 1.8000 1.6000 1.4000 1.2000 1.0000 0.8000 0.6000 0.4000 0.2000 0.0000 -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +10+11+12+13+14+15

Short Volume

Total Volume

Short Vol / Total Vol

Panel B: Negative News Events – version 1

Volume

Negative News Events - Version 1 1.6000 1.4000 1.2000 1.0000 0.8000 0.6000 0.4000 0.2000 0.0000 -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +10+11+12+13+14+15

Short Volume

Total Volume

Short Vol / Total Vol

Panel C: Negative News Events – version 2

Volume

Negative News Events - Version 2 1.6000 1.4000 1.2000 1.0000 0.8000 0.6000 0.4000 0.2000 0.0000 -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +10+11+12+13+14+15

Short Volume

Total Volume

33

Short Vol / Total Vol

Figure 2 Market Quality around News Events Figure 2 displays Rebate Rates, Amihud Illiquidity, and Bid-Ask Spread measures for the 15 days before and after news events. Panel A contains the Rebate Rate which is the rate at which interest on collateral is rebated back to the borrower in an equity loan transaction. Panel B contains the daily Amihud (2002) Illiquidity measure defined as |retit| / volumeit where volumeit is the dollar volume. Panel C contains the Bid-Ask Spread measured as a percentage of the closing mid-price on each day. Panel A: Rebate Rates

Rebate Rate around All News Events Rebate Rate (Percentage Points)

3.89 3.88 3.87 3.86 3.85 3.84 3.83 -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +10+11+12+13+14+15

Panel B: Amihud Measure

Amihud Measure around All News Events

Amihud (% of volume)

0.01200 0.01000 0.00800 0.00600 0.00400 0.00200 0.00000 -15-14-13-12-11-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8 +9+10+11+12+13+14+15

Panel C: Bid-Ask Spread

Bid-Ask Spread (% of closing mid-price)

Bid-Ask Spread around All News Events 0.00166 0.00164 0.00162 0.00160 0.00158 0.00156 0.00154

0.00152 0.00150 -15-14-13-12-11-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8 +9+10+11+12+13+14+15

34

Figure 3 Example Short Volume Portfolio Returns following News Events Figure 3 displays buy and hold portfolio returns for a 12 month period following news events. Each day for each news event, two portfolios are formed: the first portfolio consists of those firms that had a specific news event and had low short volume as a percentage of total volume; the second portfolio consists of those that had the news event and had high short volume as a percentage of total volume. We then form control portfolios using a sample of firms that did not experience a news event but were similar in terms of bid-ask-spread, institutional ownership, market capitalization, and the number of news events over the previous month. The detailed results are shown in Table XI and three example results are shown below. Panel A displays portfolio returns following dividend news and the returns for the matched control sample. Panel B displays portfolio returns following earnings news and the associated control returns and Panel C contains returns following news about insider stock sales and the associated control returns. Panel A: Dividends News Sample

Control Sample Low Short Volume

High Short Volume

5.0%

5.0%

4.0%

4.0%

Buy and Hold Return

Buy and Hold Return

Low Short Volume

3.0% 2.0% 1.0%

0.0%

3.0% 2.0% 1.0% 0.0% -1.0%

-1.0% 0

1

2

3

4

5

6

7

8

9

10

11

0

12

1

2

Panel B: Earnings News Sample Low Short Volume

Buy and Hold Return

Buy and Hold Return

5.0% 4.0% 3.0% 2.0%

1.0% 0.0% -1.0% 2

3

4

5

6

7

6

7

8

9

10

11

12

8

9

10

11

11

12

11

12

High Short Volume

8.0% 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% -1.0%

0

12

1

2

3

4

5

6

7

8

9

10

Months after Portfolio Formation

Months after Portfolio Formation

Panel C: Insider Stock Sales News Sample Low Short Volume

5

Control Sample

6.0%

1

4

Low Short Volume

High Short Volume

7.0%

0

3

Months after Portfolio Formation

Months after Portfolio Formation

Control Sample Low Short Volume

High Short Volume

4.0%

8.0%

3.0%

7.0%

Buy and Hold Return

Buy and Hold Return

High Short Volume

2.0% 1.0% 0.0%

-1.0% -2.0%

High Short Volume

6.0%

5.0% 4.0% 3.0% 2.0% 1.0% 0.0%

-3.0% 0

1

2

3

4

5

6

7

8

9

10

11

0

12

1

2

3

4

5

6

7

8

9

Months after Portfolio Formation

Months after Portfolio Formation

35

10

Table I Summary Statistics The database has 1,888,868 observations over the period January 3, 2005 through July 6, 2007. Panel A provides summary statistics at the firm level. News articles may be reissued throughout the day as more information becomes available; in such situations we consider all of the related article updates to be one unique news event and we keep track of the number of articles that are rolled-into this unique news event. News Articles per Firm-Day is a count of all news articles including reissued (updated) articles while Unique News Events per Firm-Day is a count of the unique stories, excluding subsequent updates to an article. Short Vol. / Total Vol. is the short volume from the NYSE TAQ Regulation SHO database as a percentage of total volume; Exempt and Non-Exempt denote market maker short sales (exempt) from non-market maker short sales (non-exempt), see section II.A for details. Market Capitalization is from CRSP. Panel B contains summary statistics on the frequency of each news category in the database as well as the mean number of negative words as a percentage of total words in the headline and body text of each article. News articles may be classified into more than one category and we adopt two methods for counting the number of negative words in the headline and body text of each article: Version 1 uses the Harvard-IV-4 Dictionary as in Tetlock (2007) and Engelberg (2008) while Version 2 uses the negative word list developed by Loughran and McDonald (2009). Mean

Median

1st Percentile

99th Percentile

Standard Deviation

News Articles per Firm-Day

1.10

0.00

0.00

18.00

4.09

Unique News Events per Firm-Day

0.81

0.00

0.00

12.00

2.77

Short Vol. / Total Vol.

19.60%

17.47%

0.52%

62.46%

27.22%

Short Vol. / Total Vol. – Exempt

3.61%

1.44%

0.01%

32.26%

8.11%

Short Vol. / Total Vol. – Non-exempt

17.63%

15.66%

0.39%

55.37%

26.76%

Market Capitalization ($ mm)

$5,856

$1,228

$32

$80,360

$19,329

Panel A – Firm Level Statistics

Panel B – News Categories

10K 8K Acquisitions, Mergers, Takeovers Analysts' Comments & Ratings Annual Meetings Antitrust News Bankruptcy-Related Filings Bond Ratings & Comments Buybacks Contracts, Defense Contracts, Government (not defense)

N

1,320 10,803 56,993 49,508 4,041 5,217 6,258 15,343 6,269 4,734 3,321

Mean Negative Headline Words (% of total) Version 1 Version 2 6.65% 4.83% 5.09% 5.00% 6.63% 7.40% 6.88% 6.39% 4.94% 6.03% 5.92%

36

4.07% 1.80% 1.36% 2.12% 1.32% 3.88% 3.30% 1.62% 1.03% 1.85% 1.44%

Mean Negative Body Text Words (% of total) Version 1 Version 2 3.35% 2.91% 3.18% 3.16% 3.01% 3.99% 4.47% 3.29% 3.01% 3.24% 3.20%

1.79% 1.10% 1.00% 1.08% 0.93% 1.72% 1.99% 1.15% 0.93% 1.07% 1.01%

Table I (continued)

Panel B – News Categories

N

Contracts, Nongovernment Corporate Governance Corporate Restructurings Divestitures or Asset Sales Dividend News Earnings Earnings Projections Financing Agreements High-Yield Issuers Initial Public Offerings Insider Stock Buys Insider Stock Sells Joint Ventures Labor Issues Lawsuits Leveraged Buyouts Management Issues Market News Money Market News New Products & Services Personnel Appointments Point of View Product Distribution Research & Development Spinoffs Stock Options Stock Ownership Stock Splits

19,102 4,981 5,631 11,587 24,731 40,705 37,432 6,919 173,357 9,351 21,489 54,868 9,081 12,376 15,351 2,286 13,840 14,068 1,298 24,583 29,994 17,316 2,440 5,323 1,874 5,679 25,567 2,230

Mean Negative Headline Words (% of total) Version 1 Version 2 5.96% 7.00% 6.04% 4.85% 7.64% 5.84% 5.57% 5.01% 5.24% 3.70% 1.65% 1.68% 5.71% 7.12% 7.23% 6.07% 5.58% 6.16% 6.61% 7.08% 6.71% 6.09% 5.90% 6.80% 3.89% 5.42% 2.69% 4.23%

37

1.16% 2.49% 2.16% 1.25% 0.60% 1.09% 1.39% 1.27% 1.68% 1.01% 0.45% 1.29% 1.20% 2.49% 3.85% 1.46% 1.71% 2.01% 1.99% 1.24% 1.36% 1.84% 1.52% 1.88% 1.17% 1.95% 0.70% 0.56%

Mean Negative Body Text Words (% of total) Version 1 Version 2 2.93% 3.96% 3.78% 3.21% 2.64% 3.12% 3.26% 2.95% 2.92% 2.59% 1.15% 1.34% 3.03% 3.89% 4.37% 3.79% 3.38% 3.77% 4.00% 3.02% 3.19% 3.88% 3.21% 3.76% 2.68% 3.38% 2.38% 2.13%

0.82% 1.55% 1.43% 1.03% 0.54% 0.89% 0.99% 0.98% 0.90% 0.76% 0.40% 0.34% 0.88% 1.44% 2.40% 1.16% 1.18% 1.31% 1.37% 0.77% 0.86% 1.26% 0.99% 1.21% 0.85% 1.27% 0.34% 0.35%

Table II Regression Analysis of Short Volume around News Events Table II contains the results of twelve regressions that examine short sales volume around news events. In each regression the dependent variable is aggregate short volume and the independent variable of interest is an indicator variable that takes the value one if a news story occurs and zero otherwise. In Panel A the dependent variable is raw short volume (i.e., not scaled by total volume) and in Panel B the dependent variable is short volume as a percentage of total volume. In each panel we examine six different regressions that vary the timing of the dependent variable relative to the news event in order to examine short volume changes around news. For example, t-2 indicates that the dependent variable is observed two days prior to the news event. After Minus Before indicates that the dependent variable is the difference in short volume after the event relative to before the event. To control for the documented response of short sellers to past returns, we include two lags of daily returns. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. News Events

t-2

Event Time of the Dependent Variable t-1 t=0 t+1

t+2

After Minus Before

PANEL A: Dependent Variable - Raw Short Volume (Not Scaled by Volume) Mean of the fixed effects Return (1 day lag) Return (2 day lag) News Event Indicator

141,554 351,964*** 160,355*** 5,254***

137,446 353,430*** 160,263*** 13,188***

128,667 337,458*** 158,165*** 47,034***

121,433* 335,642*** 153,859*** 31,311***

144,869 342,201*** 152,638*** 12,332***

-86,425** 483,615*** -632,172*** 26,369***

0.1844*** 0.3882*** 0.2735*** -0.0014***

0.1837*** 0.3893*** 0.274*** -0.0001

0.185*** 0.3903*** 0.2743*** 0.0004

-0.0092 0.606*** -0.4527*** 0.0047***

PANEL B: Dependent Variable - Short Volume Ratio (Scaled by Volume) Mean of the fixed effects Return (1 day lag) Return (2 day lag) News Event Indicator

0.1824*** 0.3877*** 0.2714*** -0.0018***

0.1855*** 0.3854*** 0.2729*** -0.002***

38

Table III Regression Analysis of Short Volume Ratio around News Events Table III contains the results of six regressions of short sales volume on a set of indicator variables representing specific categories of news events. In particular, the dependent variable is aggregate short volume as a percentage of total volume and the independent variables are indicator variables that take the value one if there is a news story in a particular news category and zero otherwise. We vary the timing of the dependent variable relative to the news event in order to examine short volume changes around news. For example, t-2 indicates that the dependent variable is observed two days prior to the news event. After Minus Before indicates that the dependent variable is the difference in the short volume ratio after the event relative to before the event. To control for the documented response of short sellers to past returns, we include two lags of daily returns. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. News Events Mean of the fixed effects Return (1 day lag) Return (2 day lag) 10K 8K Acquisitions, Mergers, Takeovers Analysts' Comments & Ratings of Stocks Annual Meetings Antitrust News Bankruptcy-Related Filings Bond Ratings & Comments Buybacks Contracts, Defense Contracts, Government (not defense) Contracts, Nongovernment Corporate Governance Corporate Restructurings Divestitures or Asset Sales Dividend News Earnings

t-2 0.1823*** 0.3878*** 0.2715*** -0.0007 -0.0011 0.0023 0.0017 -0.0084** -0.0041 0.0035 0.0005 -0.0038 -0.0021 -0.0059 0.0008 -0.0008 0.0033 -0.0006 0.0015 0.0008

Event Time of the Dependent Variable t-1 t=0 t+1 0.1855*** 0.3854*** 0.2729*** -0.0036 -0.0019 0.0006 0.0023 -0.0022 -0.0059 0.0011 -0.0005 -0.0052* -0.0028 -0.0084* 0.0014 -0.0026 0.0021 -0.0011 -0.0034** -0.0027*

39

0.1844*** 0.3882*** 0.2735*** -0.0039 -0.0003 0.0001 0.0123*** -0.0016 -0.0028 -0.0004 -0.0005 -0.0017 0.0078 -0.0039 -0.0012 -0.0004 0.0004 -0.0010 -0.0005 -0.0025

0.1835*** 0.3893*** 0.2738*** -0.0031 -0.0003 0.0000 0.006*** -0.0037 -0.0046 -0.0035 -0.0018 -0.0041 0.0025 -0.0010 0.0003 -0.0001 -0.0024 -0.0018 0.0003 0.0028*

t+2

After Minus Before

0.185*** 0.3904*** 0.2744*** -0.0018 -0.0003 0.0006 0.0063*** -0.0013 -0.0053 -0.0013 -0.0012 -0.003 0.0009 -0.0015 -0.0006 0.0008 -0.0001 0.0000 0.0005 0.0049***

-0.0091 0.6060*** -0.4527*** -0.0009 0.0032 -0.0003 0.0088** 0.0058 0.0006 -0.0104 -0.0036 0.0016 0.0081 0.0117 -0.0021 0.0048 -0.0077 0.0004 0.0047 0.0127***

Table III (continued)

News Events Earnings Projections Financing Agreements High-Yield Issuers Initial Public Offerings Insider Stock Buys Insider Stock Sells Joint Ventures Labor Issues Lawsuits Leveraged Buyouts Management Issues Market News Money Market News New Products & Services Personnel Appointments Point of View Product Distribution Research & Development Spinoffs Stock Options Stock Ownership Stock Splits

t-2 -0.0036* -0.0017 -0.0004 -0.0030 -0.0024 -0.0058** -0.0046 -0.0013 0.0053* -0.0031 -0.0036 -0.0030 -0.0018 0.0021 -0.0012 -0.0011 -0.0054 -0.0019 -0.0057 -0.0030 -0.0045*** 0.0017

Event Time of the Dependent Variable t-1 t=0 t+1 0.0004 -0.0035 0.0019 0.0015 -0.0029 -0.0059** -0.0051* -0.0006 -0.0014 0.0145*** -0.0024 -0.0037 -0.0019 -0.0025 0.0024 0.0016 -0.0035 -0.0003 -0.0037 -0.0031 -0.0042*** 0.0000

40

-0.0007 -0.0038 0.0057** 0.0009 -0.0021 -0.0066** -0.0017 -0.0016 -0.0053* -0.0119** -0.0047* -0.0057* -0.0105 -0.0020 -0.0002 -0.0038 -0.0042 -0.0058 -0.0069 -0.0040 -0.0041*** 0.0047

0.0031 -0.0006 0.0035 -0.0025 -0.0014 -0.0040 0.0028 0.0009 -0.0003 -0.0025 0.0000 -0.0042 0.0006 0.0001 0.0013 -0.0028 -0.0012 -0.0051 -0.0059 -0.0038 -0.0023* 0.0018

t+2

After Minus Before

0.0025 -0.0030 0.0024 -0.0009 -0.0023 -0.0011 -0.0004 -0.0008 0.0011 -0.0063 -0.0020 -0.0056* -0.0083 0.0017 0.0034* -0.0020 -0.0028 -0.0021 -0.0087 -0.0035 -0.0010 0.0102***

0.007** 0.0016 0.0056 -0.0027 0.0009 0.0068 0.0103* 0.0029 -0.0030 -0.022** 0.0030 -0.0044 -0.0011 0.0033 0.0035 -0.0050 0.0042 -0.0029 -0.0025 0.0009 0.0057** 0.0090

Table IV Regression Analysis of Raw Short Volume around News Events Table IV contains the results of six regressions of short sales volume on a set of indicator variables representing specific categories of news events. In particular, the dependent variable is the raw aggregate short volume (not scaled) and the independent variables are indicator variables that take the value one if there is a news story in a particular news category and zero otherwise. We vary the timing of the dependent variable relative to the news event in order to examine short volume changes around news. For example, t-2 indicates that the dependent variable is observed two days prior to the news event. After Minus Before indicates that the dependent variable is the difference in short volume after the event relative to before the event. To control for the documented response of short sellers to past returns, we include two lags of daily returns. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. News Events Mean of the fixed effects Return (1 day lag) Return (2 day lag) 10K 8K Acquisitions, Mergers, Takeovers Analysts' Comments & Ratings of Stocks Annual Meetings Antitrust News Bankruptcy-Related Filings Bond Ratings & Comments Buybacks Contracts, Defense Contracts, Government (not defense) Contracts, Nongovernment Corporate Governance Corporate Restructurings Divestitures or Asset Sales Dividend News Earnings

t-2 141,452 351,936*** 160,030*** -23,762*** 7,775** 1,946 4,658 3,465 3,173 15,192** 7,069* 19,459*** -31,532*** -13,446 4,944 20,671*** 806 4,527 6,770*** -8,809***

Event Time of the Dependent Variable t-1 t=0 t+1 137,444 353,335*** 160,302*** -28,923*** 7,171* -1,256 31,158*** 15,713*** -7,773 22,822*** 29,597*** 21,194*** -27,023** -18,100** 2,579 29,455*** 10,367* 9,350** 2,951 1,668

41

128,403 335,762*** 157,685*** 23,824*** 14,523*** 11,297*** 77,550*** 20,810*** 26,145*** 90,670*** 77,503*** 92,991*** -11,273 -3,401 14,985*** 44,600*** 57,111*** 38,324*** 12,491*** 63,417***

120,962* 331,588*** 152,383*** 24,974*** 14,654*** 8,393** 26,301*** 12,863** 3,301 42,514*** 28,790*** 53,651*** -10,146 1,168 12,979*** 27,269*** 37,131*** 10,396** 17,614*** 65,849***

t+2

After Minus Before

144,973 342,662*** 151,151*** 1,230 3,666 6,162* 11,636*** 8,636 10,656* 42,438*** 5,213 19,574*** -10,117 -3,031 4,351 13,420** -7,468 2,118 8,102*** 25,914***

-85,952** 482,389*** -631,964*** 79,052*** 7,103 14,526** 2,304 3,689 17,826 44,289*** -2,544 33,614*** 38,378* 26,244* 10,634 -10,846 21,411* -3,029 16,076*** 101,051***

Table IV (continued)

News Events Earnings Projections Financing Agreements High-Yield Issuers Initial Public Offerings Insider Stock Buys Insider Stock Sells Joint Ventures Labor Issues Lawsuits Leveraged Buyouts Management Issues Market News Money Market News New Products & Services Personnel Appointments Point of View Product Distribution Research & Development Spinoffs Stock Options Stock Ownership Stock Splits

t-2 -4,336 6,598 -3,626 17,781*** 14,979*** 18,020*** 10,298** 6,204 4,035 25,052*** -4,362 6,757 89,479*** -2,304 2,044 19,505*** -14,420** 21,134** -1,126 6,848 2,128 -4,078

Event Time of the Dependent Variable t-1 t=0 t+1 3,161 2,358 -2,538 53,697*** 17,307*** 5,991 -528 14,395*** 14,461*** 56,619*** 1,217 22,521*** 135,472*** -6,230 -711 24,110*** 6,252 -6,816 77,486*** 17,998*** 7,942*** 27,491***

42

53,246*** 29,615*** 37,108*** 60,989*** 8,147** -1,482 8,540* 65,368*** 16,622*** 172,693*** -6,665 199,417*** 291,721*** -3,791 13,598*** 30,348*** 6,156 3,107 66,522*** 97,709*** 9,474*** 14,689**

44,373*** 13,609*** 4,242 19,461*** 4,510 4,197 10,657** 32,374*** 15,239*** 107,072*** -820 87,374*** 133,469*** 2,751 6,932** 15,518** 6,571 -9,898 42,935*** 47,563*** 8,505*** 433

t+2

After Minus Before

16,240*** -147 1,661 9,143 1,752 6,098 4,212 22,333*** 1,014 46,789*** -4,285 22,841*** 64,114*** 3,359 3,177 12,959* -17,414** -3,898 29,173*** 42,051*** 3,436 -3,889

62,340*** 6,244 12,827* -36,373*** -27,033*** -13,076 7,565 36,419*** 60 66,930*** -2,832 85,646*** -27,562 14,803* 9,226 -11,607 -3,157 -33,872** -9,998 65,122*** 2,952 -26,376**

Table V Cross-Sectional Relation between Returns, Short Sales, and News Table V contains the results from Fama-MacBeth (1973) type regressions using daily observations over the period January 3, 2005 through July 6, 2007. The regressions are done firm by firm, and the dependent variable is the buy and hold (compound) return over the next 20 trading days. Panel A is calculated using raw returns as the dependent variable while Panel B uses characteristic adjusted returns as in Daniel, Grinblatt, Titman, and Wermers (1997), however we omit the book to market factor due to missing Compustat data for some firms. The Short Volume Ratio is daily short volume / total volume. News Event is an indicator variable that takes the value one if a news event occurs for a particular stock, and Short-News Interaction is the product of Short Volume Ratio and the News Event indicator. Returnt=0 is the return on each stock on the day that short volume and news are observed. T-statistics are below the parameter estimates in italics and are calculated using Newest-West (1987) standard errors with 20 lags. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. Panel A: Raw Returns

Intercept Short Volume Ratio

(1)

(2)

0.0171*** (4.70) -0.0053** (-2.19)

0.0173*** (4.78) -0.0053** (-2.20) -0.0010 (-1.37)

0.0054*** (5.32) -0.0069*** (-3.27)

0.0056*** (5.16) -0.0068*** (-3.27) -0.0009* (-1.72)

News Event Short – News Interaction

Model (3)

(4)

(5)

0.0171*** (4.73) -0.0044* (-1.86) 0.0000 (0.02) -0.0050** (-2.41)

0.0169*** (4.82) -0.0047** (-2.01) 0.0000 (0.05) -0.0053*** (-2.62) 0.0213 (1.29)

0.0168*** (4.99) -0.0052** (-2.24) 0.0000 (0.06) -0.0053*** (-2.72) 0.0227 (1.35) 0.0287* (1.86) 0.0356** (2.38)

0.0054*** (4.97) -0.0058*** (-2.85) 0.0003 (0.35) -0.0059*** (-2.75)

0.0052*** (4.62) -0.0061*** (-3.00) 0.0003 (0.42) -0.0063*** (-2.98) 0.0216 (1.37)

0.0051*** (4.07) -0.0065*** (-3.23) 0.0003 (0.44) -0.0063*** (-3.12) 0.0226 (1.41) 0.0303** (2.03) 0.0333** (2.33)

Returnt=0 Returnt=-1 Returnt=-2 Panel B: DGTW Returns Intercept Short Volume Ratio News Event Short – News Interaction Returnt=0 Returnt=-1 Returnt=-2

43

Table VI Cross-Sectional Relation between Returns, Short Sales, and News for Non-Exempt Trades Table VI contains the results from Fama-MacBeth (1973) type regressions using daily observations over the period January 3, 2005 through July 6, 2007. The sample only includes those short sales transactions that were classified as non-exempt as discussed in Section II.A of the text. The regressions are done firm by firm, and the dependent variable is the buy and hold (compound) return over the next 20 trading days. Panel A is calculated using raw returns as the dependent variable while Panel B uses characteristic adjusted returns as in Daniel, Grinblatt, Titman, and Wermers (1997), however we omit the book to market factor due to missing Compustat data for some firms. The Short Volume Ratio is daily short volume / total volume. News Event is an indicator variable that takes the value one if a news event occurs for a particular stock, and Short-News Interaction is the product of Short Volume Ratio and the News Event indicator. Returnt=0 is the return on each stock on the day that short volume and news are observed. T-statistics are below the parameter estimates in italics and are calculated using Newest-West (1987) standard errors with 20 lags. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. Panel A: Raw Returns

Intercept Short Volume Ratio

(1)

(2)

0.0179*** (4.87) -0.0077*** (-2.98)

0.0181*** (4.97) -0.0076*** (-2.98) -0.0012* (-1.48)

0.0061*** (5.45) -0.0092*** (-3.85)

0.0063*** (5.27) -0.0091*** (-3.84) -0.0011** (-1.53)

News Event Short – News Interaction

Model (3)

(4)

(5)

0.0179*** (4.92) -0.0067*** (-2.67) -0.0002 (-0.26) -0.0049* (-1.95)

0.0177*** (5.02) -0.0072*** (-2.84) -0.0002 (-0.20) -0.0055** (-2.25) 0.0395** (2.16)

0.0176*** (5.16) -0.0075*** (-2.94) -0.0002 (-0.17) -0.0056** (-2.33) 0.0366** (2.12) 0.0236** (2.34) 0.0221 (1.54)

0.0061*** (5.08) -0.0081*** (-3.44) 0.0001 (0.11) -0.0063** (-2.51)

0.0060*** '(4.80) -0.0085*** '(-3.59) 0.0002 '(0.21) -0.0069*** '(-2.84) 0.0390** '(2.23)

0.0059*** (4.56) -0.0088*** (-3.66) 0.0002 (0.23) -0.0070*** (-2.94) 0.0351** (2.10) 0.0258*** (2.70) 0.0223 (1.65)

Returnt=0 Returnt=-1 Returnt=-2 Panel B: DGTW Returns Intercept Short Volume Ratio News Event Short – News Interaction Returnt=0 Returnt=-1 Returnt=-2

44

Table VII Cross-Sectional Relation between Returns, Short Sales, and News for Exempt Trades Table VII contains the results from Fama-MacBeth (1973) type regressions using daily observations over the period January 3, 2005 through July 6, 2007. The sample only includes those short sales transactions that were classified as exempt as discussed in Section II.A of the text. The regressions are done firm by firm, and the dependent variable is the buy and hold (compound) return over the next 20 trading days. Panel A is calculated using raw returns as the dependent variable while Panel B uses characteristic adjusted returns as in Daniel, Grinblatt, Titman, and Wermers (1997), however we omit the book to market factor due to missing Compustat data for some firms. The Short Volume Ratio is daily short volume / total volume. News Event is an indicator variable that takes the value one if a news event occurs for a particular stock, and Short-News Interaction is the product of Short Volume Ratio and the News Event indicator. Returnt=0 is the return on each stock on the day that short volume and news are observed. T-statistics are below the parameter estimates in italics and are calculated using Newest-West (1987) standard errors with 20 lags. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. Panel A: Raw Returns

Intercept Short Volume Ratio

(1)

(2)

0.0149*** (4.32) 0.0202** (5.14)

0.0148*** (4.31) 0.0196*** (4.89) 0.0039* (1.90)

0.0033 (1.53) 0.0204*** (5.75)

0.0032 (1.49) 0.0200*** (5.55) 0.0030 (1.62)

News Event Short – News Interaction

Model (3)

(4)

(5)

0.0147*** (4.32) 0.0203*** (4.77) 0.0025 (0.14) 0.2099 (0.75)

0.0140*** (4.26) 0.0200*** (4.48) 0.0006 (0.04) 0.2189 (0.83) -0.0845** (-2.26)

0.0138*** (4.34) 0.0191*** (4.13) -0.0051 (-0.20) 0.3374 (0.84) -0.1198*** (-3.03) 0.1459*** (2.91) -0.1224*** (-3.20)

0.0031 (1.47) 0.0205*** (5.54) 0.0048 (0.29) 0.158 (0.64)

0.0025 (1.19) 0.0201*** (5.15) 0.0028 (0.17) 0.1806 (0.73) -0.0815** (-2.25)

0.0022 (1.09) 0.0191*** (4.72) -0.0017 (-0.08) 0.2748 (0.77) -0.1198*** (-3.20) 0.1602*** (3.58) -0.1407*** (-3.67)

Returnt=0 Returnt=-1 Returnt=-2 Panel B: DGTW Returns Intercept Short Volume Ratio News Event Short – News Interaction Returnt=0 Returnt=-1 Returnt=-2

45

Table VIII Market Quality around News Events Table VIII displays the cross-sectional means of the Rebate Rate, Amihud Illiquidity, and BidAsk Spread measures in event time before, during, and after news events. t-15 represents the value 15 days before a news event, t=0 is the value on the day of a news event, and t+15 represents the value 15 days after a news event. Panel A contains the cross-sectional mean value and Panel B contains the results of a t-test of means relative to the news date (t=0), The Rebate Rate for an equity loan is the rate at which interest on collateral is rebated back to the borrower. The Amihud (2002) Illiquidity measure is the daily illiquidity measure defined as |retit| / volumeit where volumeit is the dollar volume. Bid-Ask Spread is measured as a percentage of the closing mid-price on each day. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. Panel A: Mean Event Time

Rebate Rate

Amihud Illiquidity

Bid-Ask Spread

t-15 t-10 t-5 t=0 t+5 t+10 t+15

3.8767 3.8806 3.8589 3.8731 3.8629 3.8502 3.8657

0.00916 0.00921 0.00949 0.00969 0.01031 0.00926 0.00996

0.00157 0.00157 0.00157 0.00165 0.00158 0.00156 0.00157

t=0 vs. t-15 t=0 vs. t-10 t=0 vs. t-5

-0.0036 -0.0075 0.0142

0.00053 0.00047 0.00020

0.00008*** 0.00008*** 0.00008***

t=0 vs. t+5 t=0 vs. t+10 t=0 vs. t+15

0.0102 0.0229* 0.0074

-0.00063 0.00043 -0.00027

0.00007*** 0.00009*** 0.00008***

Panel B: T-test

46

Table IX Cross-Sectional Relation between Returns, Short Sales, and News including Amihud Table IX contains the results from Fama-MacBeth (1973) type regressions using daily observations over the period January 3, 2005 through July 6, 2007 where the dependent variable is the buy and hold (compound) return over the next 20 trading days. Panel A is calculated using raw returns while Panel B uses characteristic adjusted returns as in Daniel, Grinblatt, Titman, and Wermers (1997), however we omit the book to market factor due to missing Compustat data for some firms. Short Volume Ratio is daily short volume / total volume. News Event is an indicator variable that equals one if a news event occurs for a particular stock and Short-News Interaction is the product of Short Volume Ratio and the News Event indicator. Amihud is the daily Amihud Illiquidity measure and Amihud-News Interaction is the product of Amihud and the News Event indicator. Returnt=0 is the return on each stock on the day that short volume and news are observed. T-statistics are below the parameter estimates in italics and are calculated using Newest-West (1987) standard errors with 20 lags. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. Panel A: Raw Returns

Intercept Short Volume Ratio

(1)

(2)

0.0171*** (4.70) -0.0053** (-2.19)

0.0172*** (4.74) -0.0090*** (-3.67) -0.0005 (-0.70) 0.0427*** (8.82)

0.0054*** (5.32) -0.0069*** (-3.27)

0.0055*** (5.11) -0.0104*** (-4.96) -0.0005 (-0.88) 0.0405*** (9.74)

News Event Amihud Short – News Interaction Amihud – News Interaction

Model (3)

(4)

(5)

0.0171*** (4.70) -0.0083*** (-3.45) 0.0002 (0.29) 0.0427*** (8.44) -0.0046** (-2.24) 0.0428*** (2.98)

0.0169*** (4.80) -0.0086*** (-3.58) 0.0003 (0.32) 0.0426*** (8.40) -0.0049** (-2.45) 0.0428*** (2.94) 0.0184 (1.13)

0.0168*** (4.88) -0.0089*** (-3.72) 0.0003 (0.34) 0.0426*** (8.54) -0.0050** (-2.57) 0.0415*** (2.88) 0.0198 (1.20) 0.0263* (1.74)

0.0054*** (4.94) -0.0096*** (-4.68) 0.0005 (0.63) 0.0407*** (9.26) -0.0053** (-2.49) 0.0389*** (3.34)

0.0052*** (4.60) -0.0099*** (-4.79) 0.0005 (0.69) 0.0406*** (9.25) -0.0056*** (-2.69) 0.0389*** (3.29) 0.0194 (1.25)

0.0051*** (4.29) -0.0101*** (-4.92) 0.0005 (0.71) 0.0406*** (9.37) -0.0057*** (-2.83) 0.0374*** (3.17) 0.021 (1.32) 0.0281* (1.93)

Returnt=0 Returnt=-1 Panel B: DGTW Returns Intercept Short Volume Ratio News Event Amihud Short – News Interaction Amihud – News Interaction Returnt=0 Returnt=-1

47

Table X Cross-Sectional Relation between Returns, Short Sales, and News including Bid-Ask Table X contains the results from Fama-MacBeth (1973) type regressions using daily observations over the period January 3, 2005 through July 6, 2007 where the dependent variable is the buy and hold (compound) return over the next 20 trading days. Panel A is calculated using raw returns while Panel B uses characteristic adjusted returns as in Daniel, Grinblatt, Titman, and Wermers (1997), however we omit the book to market factor due to missing Compustat data for some firms. Short Volume Ratio is daily short volume / total volume. News Event is an indicator variable that equals one if a news event occurs for a particular stock and Short-News Interaction is the product of Short Volume Ratio and the News Event indicator. Bid-Ask is the spread as a percentage of the closing mid-price and Bid-Ask-News Interaction is the product of Bid-Ask and News Event. Returnt=0 is the return on each stock on the day that short volume and news are observed. T-statistics are below the parameter estimates in italics and are calculated using Newest-West (1987) standard errors with 20 lags. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significance at the 10% level. Panel A: Raw Returns

Intercept Short Volume Ratio

(1)

(2)

0.0171*** (4.70) -0.0053** (-2.19)

0.0148*** (4.03) -0.0055** (-2.35) -0.0001 (-0.08) 1.0055*** (6.94)

0.0054*** (5.32) -0.0069*** (-3.27)

0.0033*** (3.41) -0.0070*** (-3.44) -0.0001 (-0.13) 0.9459*** (8.83)

News Event Bid-Ask Short – News Interaction Bid-Ask – News Interaction

Model (3)

(4)

(5)

0.0146*** (3.92) -0.0047** (-2.01) 0.0013* (1.72) 1.0321*** (6.83) -0.0052** (-2.54) -0.1641 (-0.79)

0.0144*** (4.01) -0.0049** (-2.14) 0.0013* (1.71) 1.0285*** (6.90) -0.0053*** (-2.64) -0.1608 (-0.77) 0.0219 (1.38)

0.0143*** (4.07) -0.0052** (-2.29) 0.0013* (1.73) 1.0251*** (6.95) -0.0052*** (-2.66) -0.1739 (-0.84) 0.0239 (1.49) 0.0339** (2.32)

0.0030*** (3.12) -0.0060*** (-3.01) 0.0017** (2.50) 0.9827*** (8.45) -0.0062*** (-2.90) -0.2276 (-1.22)

0.0028*** (2.88) -0.0062*** (-3.14) 0.0017** (2.53) 0.9791*** (8.60) -0.0063*** (-3.04) -0.2269 (-1.22) 0.0242 (1.62)

0.0027*** (2.64) -0.0065*** (-3.30) 0.0017** (2.55) 0.9763*** (8.76) -0.0062*** (-3.06) -0.2361 (-1.27) 0.0261* (1.71) 0.0316** (2.27)

Returnt=0 Returnt=-1 Panel B: DGTW Returns Intercept Short Volume Ratio News Event Bid-Ask Short – News Interaction Bid-Ask – News Interaction Returnt=0 Returnt=-1

48

Table XI Equity Returns Following Specific News Events Table XI examines equity returns following news events according to the model:

where the dependent variable is the compound excess return from day 1 to day 20 following the news event, ret0 is the excess return on the day of the news event, and Size is measured using the market capitalization for each firm. Regressions are run individually for each news event and only when a news event occurs. Firm fixed effects are included and the intercept is the average of the fixed effects. T-statistics are reported below. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significant at the 10% level. News Events 10K 8K Acquisitions, Mergers, Takeovers Analysts' Comments & Ratings Annual Meetings Antitrust News Bankruptcy-Related Filings Bond Ratings & Comments Buybacks Contracts, Defense Contracts, Government (not defense) Contracts, Nongovernment Corporate Governance Corporate Restructurings Divestitures or Asset Sales Dividend News Earnings Earnings Projections Financing Agreements High-Yield Issuers Initial Public Offerings

Intercept Estimate t-stat 0.0119 0.0223 0.0383 0.0372 0.0188 0.0118 0.0144 0.0189 0.0186 0.0080 0.0247 0.0323 0.0416 0.0355 0.0325 0.0272 0.0412 0.0445 0.0215 0.0358 0.0107

0.03 0.33 0.19 0.86 0.96 -0.38 -1.55 0.37 -0.17 0.33 0.21 0.35 -1.27 0.07 0.40 2.04** 0.36 0.07 0.58 1.40 -1.41

Return (t=0) Estimate t-stat -0.5421 -0.1218 -0.0333 0.0499 -0.3073 0.3671 -0.1910 -0.0117 -0.0337 -0.2923 -0.0249 -0.1014 -0.4701 -0.0316 0.1591 -0.1339 0.1145 0.0594 -0.1209 0.0226 0.4056

49

-2.31** -2.20** -0.61 1.74* -1.93* 3.34*** -1.73* -0.28 -0.55 -1.01 -0.11 -1.24 -3.05*** -0.44 2.24** -4.02*** 5.25*** 2.28** -1.37 0.67 2.91***

Short Volume Estimate t-stat 0.0675 -0.0079 -0.0194 -0.0232 0.0037 0.0051 0.0323 -0.0168 -0.0060 -0.0012 -0.0300 -0.0229 -0.0207 -0.0953 -0.0467 -0.0047 -0.0271 -0.0435 -0.0213 -0.0137 0.0592

0.57 -0.45 -1.26 -1.90* 0.14 0.20 0.86 -1.11 -0.29 -0.02 -0.72 -1.20 -0.61 -3.24*** -2.29** -0.61 -3.08*** -3.76*** -0.88 -1.09 2.16**

Size Estimate

t-stat

-0.0024 -0.0025 -0.0067 -0.0077 -0.0012 -0.0007 -0.0016 -0.0012 -0.0009 -0.0007 -0.0008 -0.0035 -0.0027 -0.0009 -0.0022 -0.0033 -0.0080 -0.0073 -0.0016 -0.0116 -0.0013

-1.17 -4.92*** -5.48*** -5.89*** -2.29** -1.42 -2.60*** -2.50** -2.39** -1.08 -0.95 -4.39*** -3.64*** -1.65 -3.49*** -7.01*** -9.85*** -7.48*** -2.81*** -5.68*** -2.05**

Table XI (continued) News Events Insider Stock Buys Insider Stock Sells Joint Ventures Labor Issues Lawsuits Leveraged Buyouts Management Issues Market News Money Market News New Products & Services Personnel Appointments Point of View Product Distribution Research & Development Spinoffs Stock Options Stock Ownership Stock Splits

Intercept Estimate t-stat 0.0194 0.0279 0.0318 0.0372 0.0417 0.0070 0.0382 0.0416 0.0375 0.0340 0.0325 0.0294 0.0260 0.0310 0.0315 0.0178 0.0318 0.0234

0.13 -1.49 1.06 0.32 0.50 -0.25 0.25 0.08 1.16 -0.01 0.22 1.55 -1.43 2.34** -0.59 0.26 0.14 -0.95

Return (t=0) Estimate t-stat -0.2852 -0.2876 0.0844 0.0647 -0.0004 -0.2145 0.1289 -0.0841 -0.6654 -0.1535 -0.0694 0.0633 -0.3118 -0.2311 -0.2707 0.0024 -0.0153 -0.0146

Fisher Stat Fisher P-Value

-3.03*** -2.18** 0.69 1.07 0.00 -1.26 1.62 -1.68* -2.71*** -1.31 -1.19 0.43 -1.87* -1.25 -1.67* 0.03 -0.35 -0.13

Short Volume Estimate t-stat -0.0012 -0.0208 -0.0189 -0.0267 -0.0442 -0.0018 -0.0439 0.0054 -0.0742 -0.0819 -0.0170 0.0251 -0.0397 -0.0964 -0.0325 0.0144 -0.0225 -0.0410

-0.08 -0.90 -0.75 -1.25 -1.78* -0.04 -2.06** 0.22 -1.22 -4.13*** -1.24 0.65 -1.11 -2.35** -0.40 0.37 -3.25*** -1.74* 160.37*** 0.00%

50

Size Estimate

t-stat

-0.0017 -0.0043 -0.0028 -0.0034 -0.0037 -0.0003 -0.0040 -0.0040 -0.0006 -0.0021 -0.0052 -0.0040 -0.0013 -0.0008 -0.0011 -0.0015 -0.0042 -0.0006

-3.57*** -2.94*** -3.77*** -4.16*** -3.78*** -0.43 -5.12*** -5.36*** -1.54 -2.36** -6.20*** -2.53** -2.16** -1.31 -1.34 -2.47** -8.09*** -1.04

Table XII Short Volume Portfolio Returns following News Events

Table XII displays buy and hold portfolio returns for a 12 month period following news events. Each day for each news event, two portfolios are formed: the first portfolio consists of those firms that had a specific news event and had low short volume as a percentage of total volume; the second portfolio consists of those that had the news event and had high short volume as a percentage of total volume. In addition, we form control portfolios using a sample of firms that did not experience the news event but were similar in terms of bid-ask-spread, institutional ownership, market capitalization, and the number of news events over the previous month. Difference is the return of the High portfolio less the Low portfolio, and Difference in Difference is the Difference value of the Control Sample less the Difference value of the Event Sample. *** indicates significance at the 1% level, ** indicates significance at the 5% level, and * indicates significant at the 10% level. News Events 10K 8K Acquisitions, Mergers, Takeovers Analysts' Comments & Ratings Annual Meetings Antitrust News Bankruptcy-Related Filings Bond Ratings & Comments Buybacks Defense Contracts Contracts, Defense Contracts Government (not defense) Corporate Governance Corporate Restructurings Divestitures or Asset Sales Dividend News Earnings Earnings Projections Financing Agreements High-Yield Issuers Initial Public Offerings Insider Stock Buys Insider Stock Sells

Event Sample: 12 Month Returns Low High Difference 3.29% 3.56% 4.59% 4.61% 1.62% 2.70% 5.52% 3.05% 3.77% 8.83% 3.25% 3.27% 4.33% 4.19% 2.15% 4.06% 6.47% 4.64% 2.59% 3.25% 5.41% 0.25% 1.52%

2.23% -1.25% 3.59% -0.51% -1.25% 0.81% 6.84% 2.17% 0.75% 4.37% 5.41% 2.49% 2.54% 4.22% 2.18% -0.33% 1.02% 1.33% 2.04% -0.07% 6.55% -0.73% -1.76%

-1.06% -4.81% -0.99% -5.12% -2.87% -1.90% 1.32% -0.89% -3.02% -4.47% 2.16% -0.78% -1.79% 0.03% 0.03% -4.39% -5.45% -3.31% -0.54% -3.32% 1.14% -0.98% -3.29%

51

Control Sample: 12 Month Returns Low High Difference 3.07% 5.52% 4.45% 2.33% 5.43% 9.31% 8.42% 2.30% 5.32% 3.43% 5.68% 3.70% 1.21% 0.34% 2.27% 3.55% 5.84% 4.89% 4.23% 6.04% 5.68% 5.29% 4.59%

3.99% 3.18% 4.97% 4.26% 5.57% 7.51% 1.18% 7.29% 3.16% 9.59% -2.73% 5.63% 9.48% 9.86% 6.88% 2.90% 2.18% 4.02% 5.62% 7.14% 7.80% 4.54% 1.69%

0.91% -2.35% 0.52% 1.93% 0.14% -1.80% -7.24% 4.99% -2.16% 6.15% -8.41% 1.92% 8.27% 9.52% 4.61% -0.65% -3.66% -0.86% 1.40% 1.10% 2.11% -0.76% -2.90%

Difference in Difference 1.98% 2.46% 1.51% 7.05% 3.01% 0.09% -8.56% 5.88% 0.86% 10.62% -10.57% 2.71% 10.06% 9.49% 4.58% 3.74% 1.79% 2.45% 1.94% 4.42% 0.97% 0.23% 0.39%

Table XII (continued) News Events Joint Ventures Labor Issues Lawsuits Leveraged Buyouts Management Issues Market News Money Market News New Products & Services Personnel Appointments Point of View Product Distribution Research & Development Spinoffs Stock Options Stock Ownership Stock Splits

Event Sample: 12 Month Returns Low High Difference 4.59% 2.09% 7.01% 1.97% 3.90% 3.49% 5.73% 5.22% 5.98% 8.11% 4.03% 3.69% 1.80% 7.49% 2.68% 4.66%

3.62% 2.91% 3.50% 2.24% -0.63% 0.95% 7.34% -0.17% 0.64% 0.57% -2.46% 1.72% 2.96% 5.31% 0.20% 3.18%

-0.97% 0.82% -3.51% 0.27% -4.53% -2.54% 1.61% -5.38% -5.33% -7.54% -6.50% -1.98% 1.15% -2.18% -2.48% -1.48%

Mean Median

Control Sample: 12 Month Returns Low High Difference 3.70% 5.60% 4.65% 5.75% 3.99% 6.37% 3.68% 7.95% 3.32% 4.09% 0.67% 7.91% 4.78% 7.76% 5.65% 2.40%

5.59% 6.45% 6.34% 7.88% 5.30% 4.44% 19.59% 4.99% 3.36% 4.91% 6.41% 5.19% 2.74% 4.43% 3.72% 2.04%

1.89% 0.86% 1.70% 2.14% 1.31% -1.93% 15.91% -2.97% 0.03% 0.83% 5.74% -2.71% -2.04% -3.33% -1.94% -0.35%

Difference in Difference 2.86% 0.04% 5.21% 1.87% 5.84% 0.61% 14.29% 2.42% 5.37% 8.36% 12.24% -0.74% -3.19% -1.15% 0.54% 1.12% 2.89% 2.42%

T-statistic Wilcoxon Z Score

3.73*** 3.91***

52