The price effects of large short position disclosures in the Netherlands

(1)

The Price Effects of Large Short Position

Disclosures in the Netherlands

Name: Fabian Ruinard Student number: 10269487 Date: July 1, 2014

Place: Amsterdam Number of ECT’s : 12

Abstract

Economists have discussed the effects of short interest on stock prices but not extensively because of a lack of data to work with. On November 1st_{2012 several European countries adopted a new}

regulation that publicized a register with short positions which made it possible for research to obtain information about large short positions taken and the date when they were taken. Using this new data, I test the hypothesis that large short positions are followed by negative abnormal returns on the underlying stock. The two possible explanations for a negative price effect could be a negative disclosure effect following the disclosure or superior information the firm taking the short position has. The main findings are that in the first two days following the disclosure there is on average no negative abnormal return. On the longer run as of ten, twenty and thirty days after the disclosure, the stocks show on average significantly negative abnormal returns. The results could imply that there is no significant disclosure effect but short sellers who take a large short position are on average better informed than the market about the underlying’s stock price.

Keywords: Short position disclosures, Price effects Data: Euronext Security Prices Database

Specialization: Finance and Organization Supervisor: Ieva Sakalauskaite

(2)

I. Introduction

Short selling is a type of transaction investors can engage in to earn money if the price of the

underlying stock drops after they established the position. The practice of short selling brings in a lot of questions for possible research. An important aspect that has been investigated a lot is the effect of short selling on price efficiency. Because in a short sale the stock is supplied, the short sale should have a negative price effect on the underlying stock price through the mechanism of supply and demand. For an investor who engages in a short sale, the stock price needs to drop after he made the sale in order to make a profit because the investor has to buy back the stock in the future. Because of the expected negative effect on stock prices, short sales have been regulated heavily in the past. Regulations like short selling bans were introduced during the financial crisis. Several studies (Miller (1977) and Beber and Pagano (2009)) showed that constraints and bans on short selling were however detrimental for stock price efficiency and liquidity. On the Dutch financial market there has been implemented a new regulation regarding short selling. The Authority of Financial Markets in the Netherlands and several other European countries implemented a new rule November 1st_{2012 which consisted out of a disclosure policy regarding large short positions in the} Netherlands. The policy affects companies and investors in the Netherlands with a short position larger than 0,5% of the issued share capital of the firm in which the short position is established. When a short position larger than 0,5% of issued share capital has been taken, firms now need to report it to the Dutch Authority of Financial Markets and the AFM will disclose it in a register. This register gives opportunities for further research to short positions. Now that such a register has been implemented in several European countries, short position disclosure effects are easier to measure and provide an opportunity for a study on the effects. Short positions are expensive to take because of fees and risks to short squeezing which will be explained more extensively in section III. Because establishing such a large short position comes with a certain amount of risk, there is reason to believe short sellers will only establish such a position if they are better informed than the market. Earlier data on short interest was not widely available and contained only the total short interest a

(3)

firm had on a certain moment. The new register of the AFM provides opportunities to measure the price effect on a stock on which a large short position has been taken immediately after the position has been established. The main research question I will answer in this thesis is:

Are there significantly low abnormal returns after the disclosure of a large short position in the Netherlands?

The main findings of this thesis show that in the first two days following the disclosure of a large short position in the Netherlands the targets stock price shows no significant negative abnormal returns. The abnormal returns on ten days, twenty days and thirty days after the disclosure are significantly lower than zero. An explanation for the negative returns could be that short sellers taking a large short position are better informed than the market about the fundamental values of the stock prices of the firms in which they are taking the short position.

The remainder of this thesis is structured as follows. Section II provides general information and regulations regarding short selling. Section III discusses related literature, section IV describes the methodology used in this thesis and the data and its origin. Section V discusses the results and an interpretation and section VI gives a conclusion. The thesis ends with section VII which gives limitations of this study and possibilities for future research.

II. General information regarding short selling and regulations regarding short selling Short selling

A short sale is a practice that investors who do not own a stock can engage in in order to anticipate a price decline of a stock. The practice of short selling has existed since the 17th_{century when investors} could go short in the Dutch Vereenigde Oost-Indische Compagnie (VOC) (NRC, 2008). There are two ways to go short as an investor: the covered short and the naked short position. The covered short means that an investor borrows a stock from a stock broker and sells it with the promise to buy the stock back later and return it to the broker. A naked short is a way of going short in which the investor has not been able to borrow the stock or has not ensured that the stock is borrowable.

(4)

Naked short selling has been under heavy regulation after the financial crisis of 2007-2008 and has even been banned in the USA by the SEC (SEC, 2009-172).

The most important aspect of short selling to understand this thesis is the fact that the investor who takes the short position makes a profit if the price has declined when he buys back the stock. Say the investor borrows a stock from a broker and sells it to another investor for price P1. After a certain amount of time the investor buys back the stock for price P2 and delivers the stock to the broker. The investors cash flows are P1-P2 and therefore if P2<P1, the investor makes a profit. A short sale has a negative effect on the price of the underlying stock because the stock will be supplied on the stock market. Through the mechanism of supply and demand the supply of a good negatively influences the price of a good.

Earlier short selling regulations

Because of the negative price effect of short sales on stock prices short selling has been regulated in the past. During the financial crisis for example many institutions were taking large short positions in financial institutions which would cause their stock prices to drop and could harm the total financial system. Therefore several governments around the world banned short selling on financial

institutions during the 2007-2008 financial crisis to prevent negative speculation (SEC Emergency Order Release No. 34-58592). Bans on short selling turned out to be detrimental for the liquidity of the underlying stocks and caused overpricing of stock because of reduced market efficiency (Beber & Pagano, 2009).

Short selling disclosure

After the short-selling bans were dissolved, the Authority of Financial Markets (AFM) in the

Netherlands made new regulations November 1st_{2012 in order to prevent abusive short selling. The} first regulation was the prohibition of naked short selling. An investor can only take a short position if he can show with a high probability that he will be able to deliver the stock to the buyer. Another regulation is the regulation on which I will focus this thesis: the short-selling disclosure requirements.

(5)

Starting from November 1st_{2012 investors and investing firms need to publicly disclose their net} short position if it becomes 0,5% or larger from the issued share capital of the target firm. Also every 0,1% change above the 0,5% short position needs to be publicly disclosed in the AFM register. This register supplies data on the exact date a short position has been established an therefore the register gives opportunities to measure price effects after such an disclosure. Before the disclosure regulation had been adopted, these data were scarce. In the United states for example, the only available information regarding short position is the average level of shorting activity on an individual stock. The average shorting activity is made available in the financial press once a month and

therefore it is hard to measure the effect after the short position has been taken.

III. Related literature

In this section I will discuss related literature regarding short selling and short interest effects. I will start the section by discussing several articles about the theoretical background around short selling. Then I continue discussing short selling and its impact on market efficiency and the section will end with empirical articles about short interest and its effects on stock prices and other variables. The theoretical background behind short selling and the reasons why investors decide to establish a short position in a stock can contribute to the interpretation of the results in this paper. Diamond and Verrechia (1987) suggest that investors could believe in overvaluation by the market of the underlying stock and therefore take a short position in the stock. If an investor believes the stock is overvalued he can sell the stock short and buy it back with a profit if the market has corrected the overvaluation. Another reason to take a short position could be to hedge yourself against potential negative stock price movements (Woolridge & Dickinson, 1994). A combination of long and short positions or short positions and other derivatives on the same stock could reduce the maximum amount the investor is able to lose on its investment because he will receive a profitable payoff no matter what the stock price does. Brent et al. (1990) confirm the theory of Woolridge and Dickinson and show that an increase in short interest goes along with an increase in option trading which

(6)

would imply that most of the increase in short interest can be related to arbitrage and hedging opportunities. Another reason to establish a short position could be to try to exploit arbitrage opportunities during mergers (Mitchell et al., 2004). They found that a large proportion of the negative reaction to stock merger announcements is caused by short sellers trying to exploit merger arbitrage opportunities. In their investigated sample the median increase in short interest for acquiring firms increases with 40 percent. The short interest in the data sample of this thesis could be taken because of all the reasons mentioned above except for the merger arbitrage because no target firm has been merged or acquired in the time frame of the research.

Dechow et al. (2001) discuss reasons behind short selling regulation. A reason why the US imposed regulations on short selling is according to Dechow et al. that short sellers can cause stock prices to drop downwards. They claim as well that short selling is more expensive and has more risk than taking a long position in a stock because of possibilities of short squeezing or a borrowing term fee. Short squeezing means that the lender of the stock wants to sell the stock itself and the borrower has to repurchase it. There are possibilities for a borrower to ensure a certain term but that often requires payment of an additional fee. The expensiveness of going short gives another reason to believe there may be superior knowledge or a disclosure effect that ensures firms that go short heavily to have a better chance of making a profit. The disclosure effect could be a negative return following the disclosure because the market can interpret a short position as a negative speculation on the stock.

A lot of research has been done to the efficiency effect of short sellers on stock prices. Miller (1977), Harrison and Kreps (1978) and Saffi and Sigurdsson (2011) find that when short selling is constrained, stock returns are abnormally high and above their fundamental value. They name the unability of investors to react to overpricing as the reason for the abnormally high returns. Jones and Lamont (2002) find that stocks which are expensive or difficult to short have relatively high market-to-book ratio’s which could imply that constraints on short selling lead to overpricing and confirm Miller’s theory. Beber and Pagano (2009) did research to short selling on price efficiency by

(7)

measuring abnormal positive returns during the short selling ban on financial stocks during the 2007-2009 financial crisis. Their findings matched earlier research by showing that stocks in the US during the ban realized significantly positive abnormal returns. Next to abnormal returns, Beber & Pagano showed as well that the ban on short selling has reduced market liquidity significantly and slowed down price discovery.

Several published articles discuss the effect of short-interest on stock prices. Aitken et al. (1998) focused on intraday shorting activity on the Australian stock market and its returns in a fifteen minute interval after the short position has been taken on a transaction-to-transaction basis. The main findings in this article are that within twenty trades after a short sale in their market setting, stock prices show a significant 0,2 percent negative abnormal return. This paper does only use intraday data on short positions and does not look for lagged effects. My research differs from this paper that I will use a time frame of one day after the short position has been taken until an eventual time frame of thirty days after the short position has been taken. I will not only look at the immediate stock price effects on a short position but also at the development of the stock’s price in the longer run. Senchack and Starks (1993) did research on short position effects in the US by using the financial press. The United States have not adopted the same regulations regarding short position disclosures as several European countries have done. In the USA therefore, every fifteenth of the month the two major stock exchange’s members construct a list of all the common stocks in which a short position has been taken. Senchack and Starks measured the stock price effect of firms in which the short interest has increased substantially. The time span which they used to measure the stock price effect was from fifteen days prior to the publication until fifteen days after the publication of the short positions. Their findings are that some significant negative reaction occurs in the period around the publication which differs from the results I found. I found no significant negative abnormal returns on average in the first two days following the disclosure of the short position. Woolridge & Dickinson (1994) did a similar research as Senchack and Starks (1993) on 50 random companies from the NYSE and NASDAQ which had short position information available every

(8)

15th_{of the month. Their research failed to show negative price reactions on a month-to-month basis} on the shorted stocks and concluded that short sellers do not earn abnormally high or low returns. The difference between the work of Woolridge and Dickinson (1994) and Senchack and Starks (1993) on the short positions publicized in the financial press differs from the research in this thesis because this thesis focuses on the disclosure of the short position at the moment is has been taken. The positions which they used in their sample were already taken and therefore they could only measure how firms that were targets of shorting activity were performing and not have a possible larger abnormal return due to disclosure effects.

Jones et al.(2012) did a research on the effects of large short position disclosure requirements in France, Spain and the United Kingdom. The disclosure requirements are the same in France, Spain and the United Kingdom as they are the Netherlands. They did not try to measure effects of the new disclosure requirements but used the data register that has been made available with the new requirements to measure price effects. Their main findings are that there are no significant negative abnormal returns for short positions disclosed in stocks that are not involved in rights issues. However, when a stock is involved in rights issues, Jones et al. found that in the first thirty days after the announcement, the average daily return was -0,26% and the cumulative abnormal return was a significant -8,11%. The benchmark Jones et al. used to measure abnormal returns was the stock’s industry benchmark. Next to the abnormal returns, they found that there was significant follow-on shorting which means that a disclosure of a large short position is very likely to be followed by another large short position on the same stock within a month. There is however no evidence that the negative returns are an effect of the disclosures because non disclosed stocks involved with rights issues showed no significant difference in return with the disclosed ones. The interpretation of the effects is according to Jones et al. (2012) that a negative disclosure effect causes the earnings to be abnormally negative. Jones et al. speak of possibilities to manipulate stock prices by exploiting the new disclosure requirements. My thesis does not conclude that results are in fact disclosure effects because there could also be a possibility of luck or superior knowledge that the

(9)

firms have that take the short position.

Asquith, Pathak and Ritter (2005) discuss the effects of short interest and selling constraints on stock prices. The main findings in this article are that stocks imposed by a short-selling constraint show lower abnormal returns than stocks who do not have this constraint. They also show that short-interests are on average only able to predict returns on smaller stocks that have low institutional ownership. So there was no significant negative price effect on larger stocks with high institutional ownership while there was on the smaller stocks.

To conclude this section I will discuss disclosure effects regarding long positions. Large long positions could be the exact opposite of the short positions I am investigating in this thesis. Brav et al. (2008) did research to hedge funds in the USA who had to disclose their position in firms if their position was substantial. They showed that in a time span of twenty days before the announcement of the stock acquisition until twenty days after the announcement, the markets react in a favorable way with average abnormal returns between 7% and 8%. They explain their results by the fact that investors see hedge fund activism and intervention as a positive thing and therefore the returns on the stock become abnormally high after a intervention announcement. This could relate to the disclosure of large short positions in a way that investors see short interest as negative and the stock price of the underlying could be earning abnormally low returns after the disclosure.

IV. Data and methodology

This section describes the methodology that will be used in this thesis and explains the data and its origin.

The two hypotheses that will be tested are:

1. Stock prices will show negative abnormal returns on average in the first two days after the position is disclosed in the AFM register.

2. Stock prices will show negative abnormal returns on average in the first ten, twenty and thirty days after the position is disclosed in the AFM register.

(10)

I believe these hypotheses to be true because of two reasons. The first reason is that because of the extra risk and costs associated to going short as described by Dechow et al. (2001) I believe firms would not take a short position of an amount larger than 0,5% of a target firms shared capital if they had no superior knowledge or are sure that the share price has a good probability on going down. Another reason is that there could be a disclosure effect leading the public disclosure of the large short position. If investors see that a well-known investment firm is going short heavily on a firm they might think that the shorting investment firm has superior knowledge and they short the stock as well or sell their shares that they already own which would lead to a negative stock price effect. As mentioned in the previous section, Brav et al. (2008) showed that this disclosure effect is significantly positive with long positions established by hedge funds because the market sees hedge fund activism as a positive thing and this reflects in the share price. I believe the market sees large short positions as a negative thing and therefore it could reflect the opposite way in this case as well on the share price.

In order to test the hypotheses I will perform a t-test on the Average Abnormal Returns and test if they are significantly lower than zero at the 10%, 5% and the 1% significance level. The hypotheses to use to test this are:

H0: AAR = 0 H1: AAR <0

The formula for the t-statistic is:

t =

𝐴𝐴𝐴𝐴𝐴𝐴−0_{𝜎𝜎/√𝑛𝑛}

With n = 142 The denominator 𝜎𝜎

√𝑛𝑛 is called the standard error of the estimate and it will be called so in the

remainder of this thesis.

For the first hypothesis I will do the above t-test for the average abnormal returns of the first day and the average abnormal returns of the second day. The second hypothesis will be tested by looking at the average abnormal returns of the first ten days, twenty days and thirty days after the disclosure. I calculated the returns for each time frame using the same formula. For

(11)

example the formula to calculate the ten day return for each short position was:

R

10

=

𝑃𝑃10−𝑃𝑃1_𝑃𝑃1

To then arrive at the abnormal return for ten days I used:

AR

10

= R

10

– E(R)

10

To arrive at the average abnormal return for a certain time frame I cumulated all the abnormal returns of each short position and divided them by the number of short positions which is 142. A limitation of the practice of using average abnormal returns is that average abnormal returns do only take the stock prices at the end of a time frame into account. So if the average abnormal return in the time frame t=[0,10] is -2%, it could be that the average abnormal return in the time frame t=[0,5] is 0% for example. To avoid this problem, cumulative abnormal returns could be used but for the Dutch stock market. I had an absence of Dutch stock price data so I focused on the average returns which all could be calculated with the data I could obtain.

The data sample for this research consists of 142 short positions taken between November 1st_{2012 and April 8}th_{2014 on the Dutch stock market. The data has been retrieved from} the public register composed by the AFM which keeps track of short positions larger than 0,5%. The AFM register is downloadable of the AFM website and can be retrieved by anyone who would like to know current large short positions on the Dutch stock market. The 143 short positions were taken by 63 different investing firms on 28 different firms. All firms in the sample are listed on the NYSE Euronext and operating in the Netherlands. In order to test the hypotheses I need to collect stock price data of all target firms one day, two days, ten days, twenty days and thirty days after the short position has been taken. Because all of my firms of interest are listed on the NYSE Euronext, I am able to collect my data from the NYSE Euronext database. A sample of 20 short positions has been double checked with the Datastream database and I found no discrepancies. For every short position I will retrieve the stock prices on the day the position was taken, one day, two days, ten days, twenty days and thirty days later and I will calculate returns within these time frames.

(12)

The benchmark I will use to identify abnormal returns will be the Capital Asset Pricing Model. The Capital Asset Pricing Model is an expected or required return theorem which consists out of the risk-free return and systematic market risk. I choose this model as a benchmark for the

expected return because according to Berk and Demarzo (2011) the model as the basic model for asset pricing. Next to that, Fama and French (2004) find the model a fundamental theorem in the asset pricing and portfolio theory.

f

)

I deduct the risk free rate of the return on the underlying stock as my dependent variable and regress that on the market risk premium.

The alpha should not be significantly different from zero according to the Capital Asset Pricing Model and the beta will be the sensitivity of the underlying security’s returns to the market returns. The OLS regressions will be done in the statistical software SPSS. Table 1 in the appendix provides the estimated beta’s and alphas for each firm. Only one estimated alpha was significantly different from zero but that is the historical out- or underperformance of the stock to the Capital Asset Pricing Model estimation so I do not need to take it into account when estimating returns.

For the risk free rate I used the interest on US Treasury bonds. The reason why I chose US Treasury bonds is that Standard & Poor’s downgraded the rating on Dutch sovereign debt to AA instead of AAA+. US treasury bonds still have the AAA+ rating and therefore I believe the return on treasury bonds is risk free. The thirty day risk free rate is available for anytime in the register of the SEC. To calculate the risk free rate for ten days for example I took the thirty day risk free rate to the power of 1

3 To get the other time frame risk free rates I took them to the power of 1

𝑡𝑡 where t is the

number of days you need to divide by to get to the number of days the time frame lasts.

IV. Results

In this section I will discuss the results of the statistical analysis step by step for each time frame of interest. At the end of each paragraph per time a table with summary statistics has been given for each time frame. For the ten, twenty and thirty days return a histogram of the abnormal returns has been supplied after the statistics table. These are not included for the one and two day abnormal returns because the one and two day returns are all very close around zero and do not add anything relevant to my disclosure of results. Bold abnormal returns in the tables indicate they are

significantly lower than zero at the 1% level. No boldness means that they are not significant at the 10%, 5% or 1% level. There are no results which are not significant at the 1% level and are significant

(14)

at the 5% or 10% level. Therefore I do not make a difference between these levels and therefore boldness means significant at the 1% level and no boldness not significant at the 1%, 5% or 10% level.

Results t = [0,1]

The first day after the disclosure of the short position the average return has been 1,50% with a standard error of 1,522% The Abnormal return comparing it to the expected return of the Capital Asset Pricing Model has been on average 1,04% with a standard error of 1,53%. Because the Average Abnormal Return is positive it only has meaning to look if the AAR is larger than zero. The t-statistic following from the t-test is 0,67989 which is not significant at the 10% level.

Statistics for t=[0,1]

Stock Return Expected Return Abnormal Return

Average 1,504% 0,464% 1,040%

StdDev 18,202% 1,325% 18,292%

StdErr 1,522% 0,111% 1,530%

Results t = [0,2]

In the first two days following the disclosure of the short position the average return has been 0,5134% with a standard error of 1,603%. The abnormal return was on average 0,207% with a standard error of 1,6%. In the time frame t=[0,2] the abnormal return has been positive on average as well. The t-statistic of the t-test is 0,12936 and is not significantly different from zero at the 10% level.

Average 0,513% 0,307% 0,207%

StdDev 19,174% 1,842% 19,135%

StdErr 1,603% 0,154% 1,603%

Results t = [0,10]

In the first ten days after the disclosure of a short position the average return has been -2,2998% 14

(15)

with a standard error of 0,739%. The Abnormal returns were on average -3,015% the first ten days with a standard error of 0,654%. The t-statistic following from the t-test is -4,608 which means the average abnormal returns were significantly lower than zero at a 1% significance level. As is able to see in the frequency histogram, the abnormal returns are very much all around zero but the most are a little negative. The histogram shows that on average the abnormal returns are negative for

t=[0,10].

Frequency histogram of abnormal returns for t=[0,10]

Average -2,298% 0,718% -3,015%

StdDev 8,842% 2,877% 7,824%

StdErr 0,739% 0,241% 0,654%

(16)

Results t =[0,20]

In the first twenty days after the disclosure of a short position the average return has been -3,792% with a standard error of 0,957%. The Abnormal returns were on average -5,365% with a standard error of 0,911%. The t-statistic following from the t-test is -5,889 which means the average abnormal returns were significantly lower than zero at a 1% significance level. For this time frame also holds that the histogram shows that on average the abnormal returns are less than zero.

Average -3,792% 1,573% -5,365%

StdDev 11,446% 3,596% 10,896%

StdErr 0,957% 0,301% 0,911%

(17)

Results t = [0,30]

In the first thirty days after the disclosure of a short position the average return has been –4,868% with a standard error of 1,18%. The Abnormal returns were on average -7,229% with a standard error of 1,18%. The t-statistic following from the t-test is -6,1786 which means the average abnormal returns were significantly lower than zero at a 1% significance level. The histogram shows that the thirty day abnormal returns are on average a little lower than zero and gives a better view about how I came to my results. The distribution of the abnormal returns for this time frame shows more skewness to the left than the histograms in the [0,10] and the [0,20] time frame which explains why the thirty day abnormal return is more negative than the abnormal returns in the other two long run time frames.

Average 4,868% 2,362% -7,229%

StdDev 14,142% 4,123% 14,074%

StdErr 1,183% 0,345% 1,177%

(18)

Interpretation of results

As mentioned before I suggested two possible explanations for negative abnormal returns. The first was a possible disclosure effect through which the market sees a large short position as negative and which would cause the stock price to drop after such an announcement. The second explanation could be that firms that establish such a large short position could have superior knowledge or are sure that a firm’s stock is overpriced. That could be the case because it is very costly and not without risks to establish a large short position. Regarding my test results, it is interesting to see that the abnormal returns on the first two days after the disclosure of the short position are nonnegative and not significantly different from zero. Therefore I conclude that an immediate disclosure effect is not very likely. However the abnormal returns in the time frames t=[0,10], t=[0,20] and t=[0,30] show large negative abnormal returns which are all significantly lower than zero at the 1% level. That could mean that firms who take such a large short position are on average right about overpricing of the target firm’s stock price and better informed than the market.

VI. Conclusion

The disclosure of large short positions has become a new regulation in several European countries. The register in which the short positions are being disclosed gives opportunities for research to short interest announcements. One question that came into mind after hearing about the disclosure regulation was if the disclosure of a large short position would lead to abnormal negative returns. Two explanations for such an effect could be that short sellers have knowledge about possible overvaluation of the stock or that the disclosure of such a position would cause a negative price shock due to a negative perception of the market. The main findings in this thesis are that in the first two days following the disclosure of the short position, target firms show a non-significant average abnormal return of 1,04% for the first day and 0,207% for the second day. Therefore a disclosure effect looks unlikely. In the longer run the target firms experience in all three time frames significant

(19)

average abnormal returns lower than zero. -3,015% the first ten days; -5,365% the first twenty days and -7,229% the first thirty days following the disclosure. These results appear to show that short sellers taking a large position are better informed than the market on the long run.

VII. Limitations and possibilities for future research

The main focus of this thesis is the price effects of large short position disclosures in the Netherlands. The research has been an quantitative event study and only measured a price effect. The reasons behind the price effect could not be explained by the research done in this thesis. I discussed two possible explanations of the effect and tried to underpin them with related literature and arguments. There are possibilities for future research to maybe look at firm characteristics and look at the reasons behind the large short positions. Next to that there could be a size effect as well on the negative abnormal returns. A regression on the abnormal returns with a size of the short position variable could show if the size of the short position taken has a significant effect on the abnormal returns realized. There are also opportunities to make this research more externally valid. This thesis only focused on the Dutch stock market which does not imply that these results are valid for other European countries which have adopted the disclosure regulation.

Bibliography

Aitken, M., Frino, A., McCorry, S., Swan, P. (1998), Short Sales Are Almost Instantaneously Bad News: Evidence From The Australian Stock Market, The Journal of Finance, pp. 2205-2222.

Asquith, P., Pathak, P., Ritter, J. (2005) Short Interest, Institutional Ownership and Stock returns, Journal of

Financial Economics.

Beber, A., Pagano, M. (2013), Short Selling Bans Around the World: Evidence from the 2007-09 Crisis, The

Journal of Finance.

Berk, J., Demarzo, P. (2011), Corporate Finance, Pearson Global Edition, pp 350-357.

(20)

Brav, A., Jiang, W., Partnoy, F., Thomas, R. (2008), Hedge Fund Activism, Corporate Governance and Firm Performance, The Journal of Finance.

Boehmer, E., Jones, Charles M., & Zhang, X., (2008), Which shorts are informed?, The Journal of Finance. Brent, A., Morse, D., Stice, E. (1990), Short Interest: Explanations and Tests, Journal of Financial and

Quantitative Analysis, pp. 273-289.

Diamond, D., Verrechia, R., (1987) Constraints on Short Selling and Asset Price Adjustment to Private Information, Journal of Financial Economics.

Dechow, P., Hutton, A., Meulbroek, L., Sloan, R. (2001), Short Seller, Fundamental Analysis and Stock Returns,

Journal of Financial Economics, pp. 77-106.

Economic Staff, (2008), Dutch Invented Short Selling in 1609, NRC Handelsblad. Retrieved from: http://vorige.nrc.nl/international/article1993052.ece/Dutch_invented_short_selling_in_1609 Fama, E., French, K., (2004), The Capital Asset Pricing Model, The Journal of Economic Perspectives.

Harrison, J., Kreps, D. (1978), Speculative Investor Behavior in a stock market with heterogeneous expectations,

The Quarterly Journal of Economics.

Jones, C. , Reed, A. & Waller, W. (2012), Revealing shorts: an examination of large short position disclosures,

SSRN.

Jones, Charles M. (2002), Shorting Restrictions, Liquidity, and Returns, Working paper, Columbia University.

Miller, E. (1977) Risk, Uncertainty, and Divergence of Opinion, Journal of Finance, pp. 1151-1168.

Mitchell, M., Pulvino, T., Stafford, E. (2004), Price Pressure Around Mergers, The Journal of Finance, pp.31-36. Saffi, P., Sigurdsson, K. (2011), Price Efficiency and Short Selling, Review of Financial Studies.

SEC, (2009), SEC Takes Steps to Curtail Abusive Short Sales and Increase Market Transparency.

Senchack, J., Starks, L. (1993), Short Sale Restrictions and Market Reaction to Short Interest Announcements,

Journal of Financial and Quantitative Analysis, pp. 177-194.

(21)

Woolridge, R., Dickinson, A., (1994), Short Selling and Common Stock Prices, Financial Analysts Journal.

Appendix

Table 1. Beta’s and Alpha’s of target firms

Beta alpha

Accell Group N.V 0,5309 -0,014

AkzoNobel NV 1,47333 -0,01

AMG Advanced Metallurgical Group NV 1,707 0,003

Aperam SA 1,85 0,04 ASM International NV 1,315 0,02 Binckbank NV 0,7302 -0,01 Corbion NV 0,6228 0 Core Laboratories NV 0,78 0,03 Corio NV 1,03 -0,045 Fugro NV 1,4112 -0,02 Gemalto NV 0,299 -0,01 Heijmans NV 1,3116 -0,016

Koninklijke BAM Groep NV 1,399 0,014

Koninklijke KPN NV 0,267 0,07 Koninklijke Vopak NV 0,301 0,03 NSI NV 0,76 0,03 Pharming Group NV 0,809 0,05 PostNL NV 1,8231 -0,01 Royal Imtech NV 1 0,002 SBM Offshore NV 1,223 -0,004 SNS Reaal NV 1,66 0,024 TNT Express NV 1,554 0,01 TomTom NV 1,42 0,07 Unibail-Rodamco SE 0,998 -0,03 USG People NV 1,498 0,01 Wereldhave NV 0,61 -0,03 Wolters Kluwer NV 0,82 0,02 Ziggo NV 0,6377 -0,043

Bold alpha’s are significantly different from zero at the 10% significance level.