Herding behavior of investors after the disclosure of individual short positions: evidence from the Japanese stock market.
During the recent financial crisis several regulators have introduced a disclosure obligation for individual short positions. Based on data from the Tokyo Stock Exchange we analyze how the market reacts to these publications and find significant price declines after the disclosure of short position increases. This effect becomes even more evident for publications of Asian investors. Overall, the results imply that disclosures of short positions lead to a herding-like behavior among investors and accelerate rather than absorb price declines. Up until now regulators considering a permanent adoption of these measures could only borrow results from research on insider trading disclosures and majority shareholder notifications.

Keywords: Herding Behavior, Short Selling, Short sale ban, Market quality, Stock returns

Article Type:
Short selling (Laws, regulations and rules)
Stock markets (Analysis)
Kampshoff, Philipp
von Nitzsch, Rudiger
Braun, Dirk
Pub Date:
Name: Journal of International Business and Economics Publisher: International Academy of Business and Economics Audience: Academic Format: Magazine/Journal Subject: Business, international; Computers Copyright: COPYRIGHT 2012 International Academy of Business and Economics ISSN: 1544-8037
Date: May, 2012 Source Volume: 12 Source Issue: 2
Event Code: 930 Government regulation; 940 Government regulation (cont); 980 Legal issues & crime Advertising Code: 94 Legal/Government Regulation Computer Subject: Stock market; Government regulation
Geographic Scope: Japan Geographic Code: 9JAPA Japan
Accession Number:
Full Text:

Since the beginning of organized trading, investors considered the basic principle of "buying low and selling high" as the central maxim of successful investments in capital markets. Short sellers, however, realize that they can also profit from applying the reverse order by initially selling high and subsequently buying low. By means of selling a borrowed security and later repurchasing the share at a lower price the short seller benefits from falling prices. Hence, short sellers can make substantial profits while the rest of the market experiences severe losses. For this reason short selling is frequently in the focus of political debates and often blamed to cause or at least fuel financial crises. The opinions of politicians, regulators, investors, and corporate managers about the benefits and the potential harm of short sales are as diverse as their representatives themselves.

During times of booming markets short sales are commonly regarded as an elementary part of efficient capital markets. Especially the informatory content, the important role in pursuing hedging strategies, and the positive effect on market liquidity are critical aspects in favor of short sales. Also short sellers are said to be among the first to discover problems with Enron, WorldCom, and Bear Stearns. During recessions, however, critics stress the risks of disorderly or malfunctioning markets as well as the abusive use of short sales to manipulate share prices. Thus, it is not surprising that during almost all downturns financial authorities put strict regulations into force and gradually relax these constraints throughout the following upswing. Even as late as summer 2007 the US authorities decided to lift the uptick-rule, a mechanism, which only allows short sales at a price equal or above the last trade. Yet, the discussions over the past months have shown that regulators have still not managed to implement a stable regulatory environment for short selling.

After the collapse of the global capital markets in September 2008 almost all major financial authorities decided to implement temporary restrictions on short sales to limit the potential risks of price manipulation and disorderly markets. In most jurisdictions these regulatory measures included a temporary ban of short sales in selected industries (e.g., insurance companies and financial service providers) as well the tightening of disclosure obligations. For example the American Securities Exchange Commission (SEC) as well as the British Financial Services Authority (FSA) implemented on September 18, 2008 a temporary ban on naked and covered short sales in selected financial institutions. Other countries like Japan, Germany, and France implemented a ban on naked short selling only. Similar to the UK, Japan also demanded that information on large short positions (exceeding 0.25% of the issued share capital) are made publically available.

By tightening the disclosure regime financial authorities aimed to achieve three critical goals to stabilize capital markets. Firstly, the private disclosure of large short positions enables regulators to detect potential cases of market abuse. Price manipulation through short selling usually requires an investor to build up a large short position. In this case the disclosure obligation functions as an early warning system, which helps regulatory authorities to focus their attention to situations with a high likelihood of abusive behavior. However, abusive short sellers are not barred from working anonymously below the threshold. Also it is difficult for regulators to ensure that all investors comply with the disclosure obligations. Secondly, regulators intend to use the information on individual short positions to prevent disorderly markets. The disclosure obligation enables authorities to identify situations where unusual short selling activity might lead to 'fails to deliver' or an overreaction of the market and where proactive interventions are necessary. Thirdly, a public disclosure of short positions enhances the transparency for market participants. If the information on the size and the holder of the position is interpreted correctly, market efficiency can be enhanced and uncertainty among investors can be reduced. However, the publication of this information can also lead to a herding-like behavior of uninformed investors, thereby reducing market stability. Currently the vast majority of regulatory authorities consider a permanent adoption of the disclosure obligation for individual short positions. However, questions about the optimal disclosure threshold and the potential effects of making the information available to the public remain to a large degree unanswered.

In this paper we study the effects of the public disclosure of short positions in Japan after its introduction in November 2008. In Japan the Tokyo Stock Exchange centrally collects the information on changes in individual short positions exceeding 0.25% of the issued capital and publishes the information ca. two days after the actual trade. Details on the exact timing of the publication and the size of the short position change as well as the time lag between the transaction and the publication provide a unique opportunity to examine the reaction of the market upon the public disclosure. More specifically, we hypothesize that publications of changes in individual short positions lead to a herding-like behavior of investors resulting in abnormal returns and an increase in trading volumes following the announcement. This is for example the case if investors belief that high volume short sellers, who further increase their risk exposure, are likely to possess negative (not yet public) information about the future performance of the respective company. As discussed in section 2 also the findings of related scientific research in the areas of short interest publications, insider trading announcements, and majority shareholder notifications support this conjecture.

Furthermore, we expect that the public disclosure obligation of individual short positions does not stabilize stock prices (as intended by the regulatory authorities) but increases the volatility of the affected shares. Especially during times of uncertainty the market is likely to overreact upon new (unexpected) information. This is particularly the case for disclosures of short positions as investors are only provided with information about the performance expectation of an individual short seller but not with details about his decision basis. Lastly, we examine if investors value information about the identity of the short seller. We hypothesize that the market reveals a stronger herding-like behavior when investors believe that an informed trader or a company with local expertise performed the short sale transaction. Overall we find strong empirical evidence in support of these hypotheses. On the day following the announcement of a short position increase of >0.01%-points stocks show statistically significant abnormal returns between -0.31% and -0.53% depending on the size of the published change. In case of announcements of Asian short sellers or investors with a clear investment focus on the Japanese market stocks displayed even higher abnormal returns of up to -1.35% on the day following the publication and 0.7% on the subsequent day.

The remains of the paper are structured as follows. Section 2 reviews the existing literature on the impact of short interest disclosures and trade publications of individuals as well as the current state of research regarding short sale regulations during the financial crisis in 2008-2010. Also the contributions of this paper to the existing literature are discussed in this section. Section 3 presents the dataset of the Tokyo Stock Exchange and provides some summary statistics. Section 0 describes the event study methodology, which we applied to calculate abnormal returns and unusual trading activity around the disclosure of individual short positions. Also the results for these parameters are provided in this section. Section 5 presents the methodology and the findings of the analysis of volatility changes after the introduction of the disclosure obligation. The impact of the short sellers' identity on the magnitude of herding-like behavior is described in section 1. Finally in section 7 we conclude our results and discuss the limitations of our analyses.


The disclosure of short positions can either refer to the aggregate disclosure of all short positions in a particular stock or to the disclosure of individual positions held by a particular investor. While we found no literature specifically addressing the effect of publishing individual short positions, extensive research has been dedicated to understanding investors' behavior with respect to aggregate short positions.

2.1 Literature on publications of aggregate short positions

Miller (1977) demonstrates on the basis of a simple theoretical model, how short sale restrictions can lead to a systematic overvaluation of shares in presence of diverging investor opinions. In his model, as in Harrison and Kreps (1978) and Morris (1996), optimistic investors own the majority of stocks as they are willing to pay a higher price for the security. Consequently, the share price does not reflect the average opinion of the market and remains overvalued as long as short sales (by more pessimistic investors) are restricted. If, however, short selling is permitted, the disclosure of changes in aggregate short positions provides the market with valuable information on the opinion of more pessimistic investors. In a modified model, which assumes that investors account for the effect of short sale restrictions when forming 'rational expectations', Diamond and Verrecchia (1987) investigate how regulations affect the speed of price adjustments. Their analysis shows that share prices react significantly slower to new information if short selling is restricted. Furthermore, they find that announcements of short-interest (defined as the total amount of shorted shares which have not been repurchased yet) lead to price adjustments if they contain unexpected (non-public) information. These results are in line with Seneca (1967) and Kerrigan (1974) who studied the reaction of the market to changes in the short interest ratio (short interest relative to average daily trading volume). Essentially, they test if an increase in the short interest ratio leads to price increases because investors might anticipate a future demand for stocks as short sellers will eventually have to close their positions or if it causes negative returns as investors believe in the right judgment of the short seller. Both come to the conclusion that investors are more likely to trust in the correct judgment of short sellers and that an increase in short interest ratios consequently leads to an overall downward adjustment of performance expectations thereby driving down the market.

In contrast to these analyses of the overall market behavior Senchack Jr. and Starks (1993) focus on the relationship between short interest and individual stock returns. In this context they find that unexpected increases in short interest lead to small negative abnormal returns. In addition they discover in a cross-sectional analysis of individual firms that the degree of unexpected high short interest is positively correlated with negative short-term abnormal returns. Focusing more on the informatory content (long term returns) of short interest publications Figlewski (1981) observes that short interest portfolios earned substantially lower returns than low short interest portfolios. However, the results of his analysis are week as they do not provide statistically significant evidence for most of the portfolios including the group with the highest level of short interest. Also Desai et al. (2002) examine the relationship between short interest and stock returns. Based on a Nasdaq dataset from June 1988 until December 1994 they show that heavily shorted stocks subsequently experience significant abnormal returns. Furthermore, they demonstrate that the severity of negative returns is correlated to the level of short interest. More recently Asquith et al. (2005) advance these approaches by incorporating the ownership share of institutional investors as an approximation for the supply of lendable shares in their analysis. In line with previous research they also find that shares with a high level of short interest show significantly lower returns than comparable stocks with low levels of short interest. Moreover, Boehmer et al. (2008) analyze short sale order flow data from the NYSE between 2000 and 2004. They show that short sellers as a group are well informed. Particularly institutional non-program short sales are found to be the most informative. They also discover that short sellers with large volume orders of more than 5000 shares are significantly better informed that investors with small size orders of less than 500 shares. In an extensive study of forty-six markets between 1990 and 2001 Bris et al. (2007) analyze the effect of short sale restrictions on market efficiency. For this purpose they measure the share of company-specific information in stock prices and compare it between stocks with and without restrictions. Their results also suggest that short sale restrictions lead to losses in information efficiency.

In summary the empirical literature confirms, with some qualifications, the informatory content of short sales. Furthermore, investors seem to utilize the information of aggregate short positions for their trading decision. To our best knowledge, however, no study has yet been dedicated to examine the reaction of investors to the disclosure of the identity of short sellers and changes in their net short positions. So far financial authorities could only borrow results of studies analyzing the reaction of investors to trade announcements of majority shareholders and corporate insiders in order to estimate the impact of a public short position disclosure obligation. Consequently, the related literature on the disclosure of trades by individuals might be important to set the results of this study in the right context.

2.2 Literature on disclosure of individual trades

Most studies focusing on the market reaction to the disclosure of individual trades are related to majority shareholder notifications and corporate insider transactions. Mikkelson and Ruback (1985) examine share price reactions after the filing of a Schedule 13D with the SEC, which indicates that a shareholder's ownership position has exceeded the 5% threshold. Based on a dataset from 1978 to 1980 they find that the target firm as well as the acquirer experience statistically significant positive returns on the day of the announcement. Furthermore, they discover that the changes in the target's share price are more severe when the acquirer is a 'frequent purchaser' suggesting that investors do not only react upon the announcement but also process information on the identity of the buyer. Focusing on corporate insider transactions Lakonishok and Lee (2001) analyze the effect of published insider trading in the US from 1975 to 1995. Overall, their analysis shows positive abnormal returns for announced purchases and negative abnormal returns for sales. The observed magnitude of these returns is, however, typically less than 0.5%. Similarly, Friederich et al. (2002) also show for the UK market between 1986 and 1994 that directors' purchases trigger positive abnormal returns and directors' sales cause negative abnormal returns. Consistent with previous evidence the size of the observed returns is small.

More recently Fidrmuc et al. (2006) discover higher abnormal returns following the announcement after controlling for releases of corporate news related to board and asset restructuring or other important business events. More specifically they observe that directors' purchases (sales) trigger immediate abnormal returns of +3.12% (-0.37%) during the two day period starting with the announcement. Although studies like Lin and Howe (1990) show that insider trades of executives close to the operations of the firm contain more information, Fidrmuc et al. (2006) find that investors do not particularly value the information about the type of corporate insider (CEO, other executive directors, nonexecutive chairman, or nonexecutive directors). Korczak and Lasfer (2008) differentiate the impact of directors' trade announcements on share prices between UK companies which are only domestically-listed and those with a cross-listing in the US. Their results support previous research but also show that abnormal returns around insider trades are mainly confined to domestically-listed firms suggesting that directors' trades in cross-listed companies are less informative. The authors attribute these results to the fact that cross-listed companies are confronted with stricter legal requirements and have a higher visibility through better analyst coverage.

In contrast to the above mentioned studies Jaffe (1974) and Seyhun (1986) find that outsiders in most cases do not profit from abnormal returns following the dissemination of insider trading information when accounting for bid-ask spreads and commission fees. Seyhun (1986) also does not discover a relation between the size of abnormal returns and the type of corporate insider. Overall the literature on market reactions upon the disclosure of individual trades remains manifold. The most recent studies about shareholder notifications and insider trades suggest that investors value this information. To what extent, however, information about the identity of the trader plays a role for the size of abnormal returns is inconclusive and requires further research.

2.3 Literature on short sale regulation during the financial crisis 2008-2010

Since September 2008 extensive research has been conducted on the effect of the temporary short sale regulations, which most countries adopted after the collapse of the global banking system. As one of the first Bris (2008) published a paper on the effect of the initial naked short sale ban for 19 financial companies in the US. His results indicate that the emergency order has led to a reduction of market efficiency in terms of company-specific information content in stock prices. His conclusions, however, are week to the extent that there are solely based on a comparison of absolute changes in his parameters (e.g., cross-autocorrelation) and no details on statistical significance are provided. Within the same regulatory context Boulton and Braga-Alves (2010) confirm Miller's hypothesis that short sale restrictions lead to an overvaluation in share prices. On September, 18 the British Financial Service Agency and the US Securities and Exchange Commission introduced a temporary ban on naked as well as covered short sales for most financial institutions. Based on data from this period Clifton and Snape (2008) document a significant reduction of liquidity in relation to the ban of short selling. Similarly to Marsh and Niemer (2008) study the effects of changes in short sale regulations in the US, UK, Germany, France. Their results, however, do not show a significant impact of short sale regulations on market efficiency. In contrast, Kampshoff and Nitzsch (2010) find significant losses in price and information efficiency by using a industry-related control group and by focusing on the effect on the content of negative information in share prices. Also Autore et al. (2011) and Gagnon and Witmer (2009) observe a severe deterioration of market quality after the prohibition of short selling in the US and Canada respectively. The vast majority of studies on short sale regulations during the financial crisis 2008-2010 addresses the impact of short sale bans on liquidity, volatility and market efficiency. To our best knowledge no study has yet addressed the market reaction to changes in disclosure regimes. The aim of this paper is to address this research gap and to provide authorities with valuable information when deciding on the design of a long-term regulatory environment for short selling.


In order to mitigate potential risks of disorderly markets and market abuse the Japanese Financial Services Agency (FSA) implemented a short position disclosure obligation for all market participants. Effective since November 7, 2008 investors holding a short position of 0.25% or more of the outstanding shares of a company are required to report these positions (after changes have occurred) to the local exchanges through their security firms or brokers. In case of (hedge) funds, the fund manager is required to file a report of the open short position. Names of the ultimate investor or beneficiaries of the fund, however, do not need to be disclosed. In addition, the local exchanges are required to publically disclose this information in a timely manner. The Tokyo Stock Exchange (TSE) usually publishes these reports 2 trading days after the initial change in the position became subject for disclosure. For administrative reasons, the TSE compiles the information throughout the day and releases the reports collectively in two tranches at approximately 4pm and 5pm (both after the closing of the exchange at 3pm). The time lag of ca. 3 days between the actual transaction and the first trading session after the disclosure provides a unique dataset, which enables the analysis of market reactions to the publication without a bias by the initial transaction.

During the 30 day observation period starting from November 12, 2008 through December 25, 2008 in total 898 reports have been published by short position holders through the TSE containing information on more than 500 listed companies. Each report comprises the date when the short position became subject to disclosure, the report date, the number of shorted shares and the short position as percentage of shorted shares relative to the outstanding shares of the company. From this dataset we exclude duplications and late reports that are simultaneously published with a newer report containing more recent updates. We also exclude the stocks with the most extreme volatility (top 5% quantile with the highest standard deviation during the control period) in order to reduce a distortion of results from large unrelated fluctuations. In addition, we control for first time publications of large positions as beforehand no reference point has been available to the market. Thus, we obtain 3272 data points for stocks where at least one investor published a change in his short position on a particular day. Of these 1715 data points refer to reductions and 1557 to increases in short positions. In order to analyze the market reaction to the identity of the short seller, short position holders (incl. funds) and security firms have been classified depending on their origin or investment focus. We hypothesize that market participants believe that local and/or specialized investors are more likely to trade upon industry knowledge/expertise and non-public information. Therefore, we expect to find stronger reactions to publications of changes in short positions by local/specialized investors. Companies, which are based in an Asian country or region (e.g., Japan, Singapore, Hong Kong) or those, which have an exclusively Asian investment focus (e.g., Japanese funds with base in Cayman Islands), are classified as Asian investors. All other short position holders and security firms are categorized as international investors. Table 1 provides a summary of descriptive statistics.

Data on daily stock prices and trading volumes has been retrieved from Datastream for the observation and the control period. In line with standard event study methodology the control period was set to 180 trading days ranging from February 13, 2008 to October 31, 2008. In addition, the volatility observation period was extended beyond the initial 30 days until June 30, 2009 to also compute the long-term volatility effects of the disclosure obligation. In order to calculate abnormal returns from market model regressions, the Nikkei 225 index was employed to determine market returns. The Nikkei 225 contains the 225 most important Japanese companies (according to the Nihon Keizai Shimbun newspaper) with listings at the TSE and is commonly regarded as the most relevant index to reflect developments at the Japanese stock market.


In the light of the severe market turbulences during the recent financial crisis most financial authorities implemented stricter short selling disclosure obligations to create more transparency for investors and to stabilize stock prices. In addition, various countries, such as Japan, introduced an obligation for market participants to publically disclose any changes in short positions exceeding 0.25% of the outstanding capital. In order to test our hypotheses about the market's reaction to these announcements a prediction error (abnormal returns) methodology is employed. Daily prediction errors are computed using a market model of the form

[r.sub.it] = [[alpha].sub.i] + [[beta].sub.i] x [r.sub.mt] - [[epsilon].sub.it] (1)

where [r.sub.it] are the daily returns of the stock; ' (at time t), [[alpha].sub.i] and [[beta].sub.i] are the estimated model parameters, and [[epsilon].sub.it] are the prediction errors. The market model parameters [[alpha].sub.i] and [[beta].sub.i] are estimated for all stocks over a 180 trading days window starting on February 13, 2008 and ending on October 31, 2008, one week before the introduction of the disclosure obligation. The test window is set to -1 day before until +2 days following the announcement. This event interval is selected because stock prices prior to -1 will be affected by the actual short sale transaction and the market is unlikely to react with more than a two days time lag to the disclosure. Also the short test window minimizes potential overlaps with subsequent announcements. To test if the publication of changes in large short positions is associated with a significant change in security holder wealth, the cross-sectional mean abnormal return (AAR) for the days surrounding the public disclosure is then calculated as

[AAR.sub.t] = [1/N] [N.summation over (i=1)] [[epsilon].sub.it] (2)

In order to test for statistical significance we employ a standard t-test and require abnormal returns to have a p-value of less than five percent on a particular day to be regarded as significant. Analogously, abnormal returns and statistical tests are computed for the two sub-groups of Asian and international Investors.

As opposed to previous studies on the effect of short interest publications the main goal of this paper is to examine the reactions to individual (not aggregated) changes in short positions. Consequently, an observation is classified as "increase (decrease) in short position" if at least one investor has published a rise (reduction) in his short position. In order to test, whether the size of the disclosed change has an effect on the magnitude of the market reaction, observations are further allocated to different portfolios ([absolute value of x] < 0.0, < 0.01, < 0.02, < 0.03, < 0.04, < 0.05 %points) if at least one investor has announced an increase (decrease) above the respective threshold on a particular day. Multiple announcements for the same stock on the same day, however, are rare. In 87% of all observations only one investor published a change of a short position for a certain stock on a particular day. Furthermore, in 11% of the cases it was only two investors who referred to the same stock on the same day. In the seldom case, where an announced increase and decrease fall together they are accounted for in both subgroups as we test the reaction to individual announcements independent of the publications of other short sellers.

Based on the results of e.g. Mikkelson and Ruback (1985), Friederich et al. (2002) or Fidrmuc et al. (2006), who found a positive (negative) market reaction after the announcement of share purchases (disposals) by majority shareholders and corporate insiders, we hypothesize that the disclosure of individual short positions also has an observable impact on share prices. If market participants believe that investors who build up large short positions possess non-public information, which is not yet reflected in stock prices, the announcement of an increase is likely to drive down stock prices. In line with this hypothesis we find a statistically significant decline in share prices right after the announcement of a short position increase. Table 2 provides an overview of the results for the time window between one day before the announcement d(-1) until two days after the announcement d(+2). The above results further show that as soon as one investor discloses an increase of more than 0.01%-points relative to the issued capital, stock prices fall abnormally by -0.31%. In addition, the findings for the portfolios where investors announced larger changes indicate more severe market reactions of up to -0.53% following the publication (statistically significant at the 0.1% level). This relationship between the size of the announced increase and the magnitude of the market reaction suggests that investors might not only react upon announced changes in general but also upon information about their size. In this context it is important to note that disclosures of individual short positions provide the market only with insights about other investors' opinions but not with information in relation to the future performance of the company as it would be the case for earnings or merger announcements. Hence, the market reaction is likely to be driven rather by a herding-like behavior of investors than by informed trading.

Interestingly, the analysis of announced reductions in short positions does not show the inverse results. As shown in Table 3 the market does neither react significantly before nor directly after the disclosure. At first these findings seem to be different from the results of Baesel and Stein (1979), Friederich et al. (2002) and Fidrmuc et al. (2006) who all discovered a stronger reaction of the market after the announcement of purchases than after the disclosure of share disposals by corporate insiders. However, these results are consistent with our findings as in both cases the market values more the disclosed information when the investor announces an increase of his risk exposure. Both, the corporate insider, who buys his own stock, as well as the short seller, who expands his open position, build up their risk exposure. Clearly, in the opposite cases it is more difficult for market participants to determine whether the investment decision is driven by changes in performance expectations or other reasons not directly related to the respective firm.

The analysis of the announcement effect on liquidity is based on differences in daily trading volumes. As trading volumes vary heavily throughout the year and are significantly influenced by macroeconomic developments the control period to determine a change in liquidity is chosen close to the announcement date. As the days prior to d(-1) are likely to be affected by the actual short sale transaction and d(1) might still be influenced by the announcement, the control variable was defined as the average trading volume of d(-1; 2; 3; 4) excluding d(0) and d(1) in between. In order to avoid any distortion from subsequent short position announcements in the same stock, only observations without a second announcement during the control period are taken into account. In total 490 observations meet this requirement.

In comparison to the average trading volume during the surrounding days d(-1; 2; 3; 4) trading activity significantly surged on the announcement day of the change in a short position. For the portfolio with changes [absolute value of x] > 0 trading volumes are on average 9.21% higher than on the days directly before and after the disclosure. The size of the increase, however, does not seem to depend on the magnitude of the announced change. Table 4 provides an overview of changes in trading activity on the publication day.


The disclosure obligation of individual short positions enables regulatory authorities to timely detect abnormal trading activities and to better investigate cases of possible abusive behavior. In addition, the Japanese FSA also required the publication of these positions to provide market participants with more transparency on the identity and the current exposure of short sellers. As the level of uncertainty among investors is a key driver of volatility in stock markets and short selling is frequently blamed to elevate the risk of severe market crashes, the FSA intended to stabilize stock prices by providing information on these short positions. In order to test if the FSA could successfully reduce the level of volatility in the market, a short and a long timeframe comparison are conducted. The former is based on a comparison between the stock price volatility during the 30 trading days prior the publication obligation (18.9.08-31.10.08) and the volatility during the 30 days after the introduction (12.11.08-25.12.08). The latter focuses on a longer period comparing the effects during the 153 trading days prior (21.3.08-31.10.08) and the 153 days post (12.11.08-30.06.2009) the regulatory change. For both periods only stocks with disclosed changes in short positions during the initial observation window are taken into account. To determine changes in volatility for both timeframes continuously compounded returns [r.sub.it] are calculated as

[r.sub.it] = ln ([p.sub.it]/[[p.sub.it]-1]) (3)

where [p.sub.it] is the price of the stock; ' at day t. The standard deviation [[sigma].sub.i] is then computed according to

[[sigma].sub.i] = [square root of [1/[N - 1]][N.summation over (t=1)][([r.sub.it] - [[bar.r].sub.i]).sup.2]] (4)

and resulting changes are tested for statistical significance. In both cases (short and long observation period) we compare identical timeframes, which renders an annualization of the volatility measures unnecessary. Global capital markets, however, underwent a period of severe market turbulences with a peak in September and October 2008 right around the bankruptcy of major financial institutions such as Lehman Brothers. We therefore expect to examine an overall decrease in volatility during the subsequent months. In order to account for this overall stabilization of the market we compare our results to volatility changes of the market as a whole using the Nikkei 225 as an approximation. Clearly, some of the companies in our sample are part of the Nikkei 225. This interdependence, however, will only weaken the measured differences but not distort the overall conclusions. The results of the short timeframe volatility comparison (30 days prior to the regulatory change vs. the 30 after the implementation) are displayed in Table 5. As expected, the standard deviation is significantly lower (between -2.66%-points and -2.54%-points) for all portfolios during the time after the introduction of the public disclosure obligation. Interestingly, the analysis of the Nikkei 225 displays a stronger decline in volatility of -2.73%-points. This comparison reveals that stocks with publications on short positions did not stabilize to the same extend as the rest of the market. In addition, the differences in relative changes between the index and the sample are even more significant. Justifiably, it is possible to argue that the initial short sale transaction can have contributed to some extend to the volatility in the publication period. We therefore also compute the standard deviation for the affected stocks for a longer observation window of 154 days prior and post the introduction of the disclosure obligation. As displayed in Table 6, the absolute and relative differences in volatility reductions are even larger for the extended observation period.

Overall, it is not possible to infer that the disclosure obligation has or has not contributed to stabilizing the market by providing investors with the guarantee of being informed about changes in significant short positions. However, we believe it is unlikely that stocks become less volatile only because no investor has published a change in a significant short position. On the contrary the abnormal returns on the announcement day and the smaller volatility decrease of stocks with publications on short positions suggest that the introduction of the disclosure obligation by the FSA was potentially counterproductive and has destabilized the prices of the affected stocks.


A wide body of empirical literature addresses the market reaction to corporate insider trading. Several of these studies, such as Baesel and Stein (1979) and Lin and Howe (1990) further focus on the information content of the identity of the short seller. Consistently, they find that transactions by top executives close to the operations of the firm best predict the future share price performance of their company. Similarly, we assume that the market not only reacts upon information on short positions in general but also processes information about the identity of the short seller. Especially in cases, where investors are likely to believe that a short seller relies on private (not yet public) information or that the short seller possesses a distinct expertise on the Japanese market, we expect to observe the largest stock price reaction. As an approximation for the attributed likelihood of informed trading or the level of assumed expertise with regard to Japanese stocks we use the origin of the investor and his investment focus. Accordingly, investors who are based in an Asian country or region such as Japan, Singapore, and Hong Kong or funds with a clear investment focus on Japanese Stocks (mostly also with known Japanese fund managers) are classified as Asian, while all other investors are denoted as international. Applying the above mentioned methodology to calculate abnormal returns for the two subsamples, we find that the market reacts strongly upon publications of Asian short sellers. As shown in Table 7 we observe abnormal returns between -0.78% and -1.35% on the day of the announcement. Also on the day after the announcement d(+1) the analysis displays statistically significant negative abnormal returns between -0.41% and -0.7%. For d(-1) and d(+2) average stock price movements are small and statistically insignificant. The magnitude of the abnormal return around the announcement, however, does not seem to depend on the size of the published short position change. This implies, that investors predominantly react upon the information that the short seller expects a share price decline. The degree to which the short seller increases his risk exposure is only processed to a limited extent. The reason for this could be that only those short sellers have to publish changes in their short positions that have already exceeded the threshold of 0.25%. With respect to the large size of these positions, the difference between an increase of 0.02%-points and 0.04 %-points might not be as relevant as the message conveyed by a general reinforcement of the performance expectation. On the contrary, the analysis of the market behavior after the announcement of short position increases by international investors does not reveal any abnormal reactions. Neither directly after the announcement on d(0) nor on the following day d(+1) we observe statistically significant abnormal returns (Table 8). Moreover, neither of the two sub-groups shows significant reactions after the announcement of a short position decrease.

These results are in line with our hypothesis that the magnitude of the herding-like behavior of investors strongly depends on the identity of the short seller. To some extend these results can be compared to investors' reactions after changes in performance expectations and purchase recommendations of brokers. For example Stickel (1995) finds that the reaction of the market upon analyst recommendations highly depends on the reputation of the respective analyst and the size of the broker house. With regard to short position disclosures investors seem to behave in a similar way by choosing individual short sellers and imitating their strategy. In order to test if the short position disclosure of an Asian investor enhances market efficiency by providing valuable information about the future performance of the stock we also compute cumulative abnormal returns (CARs) for the 30 and 90 days following the announcement day. In line with the calculations of abnormal returns following the publication of a short position change, CARs are computed using a market model regression based on continuous returns. The market model parameters [[alpha].sub.i] and [[beta].sub.i] are estimated over a 180 trading days window starting on February 13, 2008 and ending on October 31, 2008. In this context statistically significant negative abnormal returns would imply that Asian short sellers are informed traders and imitating their investment strategy can yield sizable returns. Consequently, the public disclosure of changes in short positions would be an important source of information for the market and increase overall market efficiency. If, however, initial abnormal returns from the announcement day are fully absorbed by positive returns during the post disclosure period, the gain on short position transparency only leads to an overreaction of the market, which is subsequently being corrected. Such a volatility increase would be counterproductive and opposite to the intent of the Japanese FSA to stabilize the capital market. Table 9 displays the CARs during the 30 and 90 days following the disclosure of a short position increase by an Asian investor.

Almost all portfolios show negative CARs during the time after a disclosed short position increase indicating that the announcement might have some informatory content. Yet, the size of the cumulative abnormal returns varies considerably between the portfolios and none of results is statistically significant at the 5% level. We therefore cannot infer that the public disclosure of individual short positions enhances information efficiency. Net of the transaction cost that incur in a stock sale and a subsequent repurchase, most investors are not able to achieve statistically significant abnormal returns by copying the investment strategy of short sellers. The evidence is thereby consistent with market efficiency.

7 Conclusion

In the light of the severe market turbulences during the financial crisis 2008-2010 several regulatory authorities around the world implemented a disclosure obligation for changes in significant short positions. One of the most far-reaching regulations was introduced by the Japanese FSA on November 7, 2008 requiring all investors with short positions in excess of 0.25% of the outstanding capital to publically disclose changes in these positions. To our best knowledge, no previous research has addressed the effects of this type of disclosure obligation on investors' behavior and market efficiency. In line with the results of previous studies on the implications of majority shareholder notifications and insider trading such as Mikkelson and Ruback (1985), Lakonishok and Lee (2001) or Fidrmuc et al. (2006) we find a statistically significant reaction of the market to information about individual short sale trades. Overall, we show that stocks experience negative abnormal returns of up to -0.31% and -0.53% following the announcement of a short position increase of > 0.01%-points depending on the size of published change. Also trading volumes significantly surge on this day.

The main goal of implementing short sale regulations during the financial crisis was to stabilize capital markets. From our analysis we cannot determine whether the disclosure obligation has or has not contributed to reducing volatility of the overall market by providing investors with more transparency. It is, however, likely that global markets in general stabilized after the implications of the credit crisis had become more predictable and first government emergency programs had been implemented. In addition, first studies, such as Marsh and Niemer (2008) and Kampshoff and Nitzsch (2010), have already shown that not even the complete ban on short selling in the USA and UK was able to prevent major price declines or to substantially reduce volatility. On the contrary the smaller volatility decrease of stocks with publications on short positions in comparison to the market and the negative abnormal returns on the day of the announcement imply that the disclosure obligation was rather counterproductive and led to a destabilization of stock prices.

In line with the research of Baesel and Stein (1979) and Lin and Howe (1990) on the reaction to disclosed transactions of individuals in the context of insider trading we find that the market values information about the identity of the short seller. Announcements by investors which are based in an Asian country or which have a clear investment focus on Japanese stocks, are followed by a significantly stronger market reaction than disclosures of other investors. For the former we find statistically significant negative abnormal returns of up to 1.35% on the day of the publication and 0.7% on the subsequent day. In contrast we found no abnormal returns after the announcements of non-Asian investors. The results suggest that market participants attribute a higher likelihood of informed-trading to local short sellers. Based on this competence assessment they select certain investors and imitate their investment strategy, leading to a selective herding-like behavior of the market. The analysis of cumulative abnormal returns during the 30 and 90 days after the announcement does not show statistically significant price movements. Net of the transaction costs of selling and subsequently repurchasing the stock investors are not able to achieve substantial abnormal returns. The evidence is therefore consistent with market efficiency (semi-strongform) and implies that the measure of the FSA has neither stabilized the market nor significantly contributed to enhance information efficiency. Our results are therefore different from the findings of research on the market reaction to insider trading. The vast majority of these studies observe statistically significant returns during the time after the disclosure suggesting that trade publications of corporate insiders increase market efficiency.

The public disclosure obligation is also associated with potential risks for short sellers themselves. Competitors might take advantage of the knowledge that the short seller will eventually have to close his open short positions and cause a short squeeze by purchasing the respective stock in parallel. For this reason short sellers might reduce their trading activity to stay under the threshold of 0.25%. This reduction of short selling can result in a deterioration of liquidity, which in turn can widen bid-ask spreads and decrease market efficiency. Overall, we believe that information on individual short positions can help regulatory authorities to timely investigate situations of likely market abuse. Especially price manipulation through short selling usually requires investors to build up substantial short positions. In this case the disclosure obligation can function as an early warning system, which helps authorities to focus their attention. In order to establish a sustainable regulatory environment for short selling we therefore support the implementation of a reporting obligation of individual short positions to the responsible financial authority. Based on previous experience with short sale related price manipulation, regulators should consider an adjustment of the reporting threshold of 0.25%. This threshold should be set to affect as few short sellers as possible while still allowing an early identification of abusive activities. Given the results of this study we do not recommend a public disclosure obligation for these positions.

The presented results should, however, only be regarded as a first indication for the effects of short position disclosure obligations. Especially the short observation period limits the explanatory power of the analysis. Also other factors such as the uptick rule and the implementation of the ban on naked short sales can have diluted the magnitude of the market reaction after the announcement. E.g., investors, who did not possess shares in the affected company, might not have been able to short the stock at declining prices due to the uptick-rule or simply because they were not able to locate any lendable shares. In addition, our volatility analysis does not provide detailed results on the extent to which the disclosure obligation might have contributed to the overall market stabilization. Related research, however, finds that even a complete ban of short sales was not able to significantly reduce market volatility.


Asquith, P., Pathak, P.A., Ritter, J. R., "Short interest, institutional ownership, and stock returns", Journal of Financial Economics, Volume 78, Number 2, Pages 243-276, 2005.

Autore, D. M., Billingsley, R. S., Kovacs, T., "The 2008 short sale ban: Liquidity, dispersion of opinion, and the cross-section of returns of US financial stocks", Journal of Banking & Finance, Volume 35, Number 9, Pages 2252-2266, 2011.

Baesel, J.B., Stein, G. R., "The value of information: Inferences from the profitability of insider trading", Journal of Financial & Quantitative Analysis, Volume 14, Number 3, Pages 553-571, 1979.

Boehmer, E., Jones, C.M., Zhang, X., "Which shorts are informed?", Journal of Finance, Volume 63, Number 2, Pages 491-527, 2008.

Boulton, T.J., Braga-Alves M. V., "The skinny on the 2008 naked short-sale restrictions", Journal of Financial Markets, Volume 13, Number 4, Pages 397-421, 2010.

Bris, A., "Short selling activity in financial stocks and the SEC July 15th emergency order", Working Paper, International Institute for Management Development. IMD Geneva Report, 2008.

Bris, A., Goetzmann, W. N., Zhu, N., "Efficiency and the bear: short sales and markets around the world", Journal of Finance, Volume 62, Number 3, Pages 1029-1079, 2007.

Clifton, M., Snape M., "The effect of short-selling restrictions on liquidity: Evidence from the London Stock Exchange", Working Paper, Sydney University of Technology, 2008.

Desai, H., Ramesh K., Thiagarajan S. R., Balachandran, B. v., "An investigation of the informational role of short interest in the Nasdaq market", Journal of Finance, Volume 57, Number 5, Pages 2263-2287, 2002.

Diamond, D. W., Verrecchia, R. E., "Constraints on short-selling and asset price adjustment to private information". Journal of Financial Economics, Volume 18, Number 2, Pages 277-311, 1987.

Fidrmuc, J. P., Goergen M., Renneboog L., "Insider trading, news releases, and ownership concentration", Journal of Finance, Volume 61, Number 6, Pages 2931-2973, 2006.

Figlewski, S., "The informational effects of restrictions on short sales: Some empirical evidence", Journal of Financial & Quantitative Analysis, Volume 16, Number 4: Pages 463-476; 1981.

Friederich, S., Gregory, A., Matatko, J., Tonks, I., "Short-run returns around the trades of corporate insiders on the London Stock Exchange", European Financial Management, Volume 8, Number 1, Page 7, 2002.

Gagnon, L. J., Witmer, J. L., "Short changed? The market's reaction to the short sale ban of 2008", Working Paper, SSRN, 2009.

Harrison, J. M., Kreps, D. M., "Speculative investor behavior in a stock market with heterogeneous expectations", Quarterly Journal of Economics, Volume 92, Number 2, Pages 323-336, 1978.

Jaffe, J. P., "Special information and insider trading", Journal of Business, Volume 47, Number 3, Pages 410-428, 1974.

Kampshoff, P., Nitzsch, R. v., "Banning short sales during the crisis--Blessing or curse for financial markets?", Journal of Business Economics, Volume 80, Number 11, Pages 1179-1206, 2010.

Kerrigan, T. J., "The short interest ratio and its component parts", Financial Analysts Journal, Volume 30, Number 6, Pages 4549, 1974.

Korczak, A., Lasfer, M., "Does cross-listing mitigate insider trading?", Working Paper, SSRN, 2008.

Lakonishok, J., Lee, I., "Are insider trades informative?", Review of Financial Studies, Volume 14 (1), 2001

Lin, J., Howe, J. S., "Insider trading in the OTC market", Journal of Finance, Volume 45, Number 4, Pages 1273-1284, 1990.

Marsh, I. W., Niemer, N., "The impact of short sale restrictions", Working Paper, SSRN, 2008.

Mikkelson, W. H., Ruback, R. S., "An empirical analysis of the interfirm equity investment process", Journal of Financial Economics, Volume 14, Number 4, Pages 523-553, 1985.

Miller, E. M., "Risk, uncertainty, and divergence of opinion", Journal ofFinance, Volume 32, Number 4, Pages 1151-1168, 1977.

Morris, S., "Speculative investor behavior and learning", Quarterly Journal of Economics, Volume 111, Number 4, Page 1111, 1996.

Senchack Jr., A. J., Starks L. T., "Short-sale restrictions and market reaction to short-interest announcements", Journal of Financial & Quantitative Analysis, Volume 28, Number 2, Pages 177-194, 1993.

Seneca, J. J., "Short interest: Bearish or bullish?", Journal ofFinance, Volume 22, Number 1, Pages 67-70, 1967.

Seyhun, H. N., "Insiders' profits, costs of trading, and market efficiency", Journal of Financial Economics, Volume 16, Number 2, Pages 189-212, 1986.

Stickel, S. E., "The anatomy of the performance of buy and sell recommendations", Financial Analysts Journal, Volume 51, Number 5, Page 25, 1995

Philipp Kampshoff, RWTH Aachen University, Aachen, Germany

Rudiger von Nitzsch, RWTH Aachen University, Aachen, Germany

Dirk Braun, RWTH Aachen University, Aachen, Germany

Number of observations if at least one investor reported the
respective size of change

(excluding multiple observations for one company on the same day)

Type of         Size of                         Disclosed
Investor        change                          increase

                                                Abs.   Rel.

All             [absolute value of x] > 0       1557   48%
                [absolute value of x] > 0.01    1039   32%
                [absolute value of x] > 0.02    795    24%
                [absolute value of x] > 0.03    659    20%
                [absolute value of x] > 0.04    568    17%
                [absolute value of x] > 0.05    4S1    15%

Asian           [absolute value of x] > 0       323    56%
                [absolute value of x] > 0.01    236    41%
                [absolute value of x] > 0.02    192    33%
                [absolute value of x] > 0.03    163    29%
                [absolute value of x] > 0.04    142    25%
                [absolute value of x] > 0.05    120    21%

International   [absolute value of x] > 0       1272   46%
                [absolute value of x] > 0.01    322    30%
                [absolute value of x] > 0.02    614    22%
                [absolute value of x] > 0.03    496    16%
                [absolute value of x] > 0.04    430    16%
                [absolute value of x] > 0.05    209    6%

Type of          Disclosed       Total      Number of
Investor         decrease                   affected

                Abs.   Rel.   Abs.   Rel.   Abs.

All             1715   52%    3272   100%   472
                1058   32%    2097   64%    446
                735    22%    1530   46%    418
                546    17%    1207   3?%    397
                416    13%    986    30%    368
                326    10%    807    25%    341

Asian           251    44%    074    100%   149
                171    30%    407    71%    141
                133    23%    325    56%    131
                99     17%    2G7    46%    127
                60     14%    222    39%    121
                67     12%    187    33%    114

International   1486   54%    2758   100%   388
                898    33%    1720   63%    364
                609    22%    1223   44%    343
                453    16%    949    34%    320
                341    12%    771    28%    254
                262    9%     471    17%    270


In percent        Disclosed change (increase in %-points)
                            of short position

Day          x > 0    x > 0.01   x > 0.02   x > 0.03    x > 0.04

-1           -0.06    0.05       0.12       -0.13       0.17
0            -0.15    -0.31 **   -0.42 **   -0.48 ***   -0.50 **
+1           0.01     0.07       0.06       -0.05       -0.06
+2           0.07     0.18       0.14       0.06        0.05

N (abs.)     1557     1039       795        659         568

In percent   Disclosed change
             (increase in
             %-points) of
             short position

Day          x > 0.05

-1           0.19
0            -0.53 ***
+1           -0.01
+2           0.08

N (abs.)     481

* p < .05, ** p < .01, *** p < .001


In percent           Disclosed change (decrease in %-points)
                                of short position

Day          x < 0     x < -0.01   x < -0.02   x < -0.03   x < -0.04

-1           0.03      0.07        0.11        0.05        0.14
0            -0.18 *   -0.21       -0.23       -0.19       -0.24
+1           -0.15     -0.22       -0.20       -0.30       -0.36
+2           -0.10     0.00        -0.02       0.08        0.08

N (abs.)     1715      1058        735         548         418

In percent   Disclosed change
             (decrease in
             %-points) of
             short position

Day          x < -0.05

-1           0.20
0            -0.30
+1           -0.32
+2           -0.07

N (abs.)     326

* p < .05, ** p < .01, *** p < .001


Disclosed                       [DELTA] d(0)
change                          and average
                               of d(-1;2;3;4)
                                 in percent

[absolute value of x] > 0         9.21 **
[absolute value of x] > 0.01       5.77 *
[absolute value of x] > 0.02      7.43 **
[absolute value of x] > 0.03      8.18 **
[absolute value of x] > 0.04       6.68 *
[absolute value of x] > 0.05       5.88 *

* p < .05, ** p < .01, *** p < .001


Change                         Pre-period STD     Public. Period STD
in %-points                    18.9.08-31.10.08   12.11.08-25.12.08
                               (30 days, in %)    (30 days, in %)

[absolute value of x] > 0      6.66               4.00
[absolute value of x] > 0.01   6.66               4.02
[absolute value of x] > 0.02   6.69               4.07
[absolute value of x] > 0.03   6.69               4.07
[absolute value of x] > 0.04   6.69               4.11
[absolute value of x] > 0.05   6.64               4.10

Nikkei 225                     5.90               3.17

Change                         Delta       Delta
in %-points                    absolute    relative
                                           in %

[absolute value of x] > 0      -2.66 ***   -39.91 ***
[absolute value of x] > 0.01   -2.64 ***   -39.66 ***
[absolute value of x] > 0.02   -2.62 ***   -39.17 ***
[absolute value of x] > 0.03   -2.61 ***   -39.08 ***
[absolute value of x] > 0.04   -2.58 ***   -39.63 ***
[absolute value of x] > 0.05   -2.54 ***   -38.27 ***

Nikkei 225                     -2.73       -46.22

* p < .05, ** p < .01, *** p < .001


Change                          Pre-period STD     Public. Period STD
in %-points                     21.3.08-31.10.08   12.11.08-30.06.09
                                (154 days, in %)   (154 days, in %)

[absolute value of x] > 0       3.88               3.42
[absolute value of x] > 0.01    3.89               3.43
[absolute value of x] > 0.02    3.91               3.45
[absolute value of x] > 0.03    3.90               3.46
[absolute value of x] > 0.04    3.91               3.48
[absolute value of x] > 0.05    3.89               3.48

Nikkei 225                      2.95               2.32

Change                          Delta        Delta
in %-points                     Absolute     relative
                                             in %

[absolute value of x] > 0       -0.46% ***   -11.94 ***
[absolute value of x] > 0.01    -0.45% ***   -11.69 ***
[absolute value of x] > 0.02    -0.45% ***   -11.58 ***
[absolute value of x] > 0.03    -0.44% ***   -11.24 ***
[absolute value of x] > 0.04    -0.43% ***   -10.95 ***
[absolute value of x] > 0.05    -0.40% ***   -11.40 ***

Nikkei 225                      -0.63%       -21.31

* p < .05, ** p < .01, *** p < .001


In percent   Disclosed change (increase in %-points) of short position

Day          x > 0       x > 0.01    x > 0.02    x > 0.03    x > 0.04

-1           0.03        0.07        0.10        0.14        0.44
0            -0.78 ***   -0.99 ***   -1.30 ***   -1.18 ***   -1.35 ***
+1           -0.41 *     -0.59 **    -0.55 *     -0.63 *     -0.70 *
+2           -0.06       0.20        0.25        0.16        0.22

N (abs.)     323         236         192         168         142

In percent   Disclosed change
             (increase in
             %-points) of
             short position

Day          x > 0.05

-1           0.54
0            -1.24 ***
+1           -0.48
+2           0.16

N (abs.)     120

* p < .05, ** p < .01, *** p < .001


In percent   Disclosed change (increase in %-points) of short position

Day          x > 0    x > 0.01   x > 0.02   x > 0.03   x > 0.04

-1           -0.06    0.07       0.13       0.14       0.09
0            0.00     -0.14      -0.17      -0.25      -0.19
+1           0.09     0.22       0.21       0.13       0.14
+2           0.07     0.15       0.08       0.00       -0.03

N (abs.)     1272     822        614        496        430

In percent   Disclosed change
             (increase in
             %-points) of
             short position

Day          x > 0.05

-1           0.07
0            -0.28
+1           0.14
+2           0.04

N (abs.)     209

* p < .05, ** p < .01, *** p < .001


In percent   Disclosed change (increase in %-points) of short position

Period       x > 0   x > 0.01   x > 0.02   x > 0.03   x > 0.04

CAR          -1.15   0.03       -0.30      -0.76      -1.99
CAR          -2.77   -1.26      -0.19      -1.58      -2.86

N (abs.)     323     236        192        168        142

In percent   Disclosed change
             (increase in
             %-points) of
             short position

Period       x > 0.05

CAR          -2.13
CAR          -2.68

N (abs.)     120

* p < .05, ** p < .01, *** p < .001
Gale Copyright:
Copyright 2012 Gale, Cengage Learning. All rights reserved.