University of Groningen
Faculty of Economics and Business
Thesis MSc Finance
Dutch auction tender offer: Pre and post crisis wealth effects
around the announcement of preliminary results.
Abstract
The main objective of this paper is to examine the wealth effects during a Dutch auction tender offer using a sample of 114 Dutch auction tender offers taken place during 2002-2013. Previous studies have thoroughly examined the wealth effects around the announcement date. To better understand the wealth effects during the Dutch auction tender offer process we examine several points of interest with a focus on the announcement of the preliminary results of the Dutch auction tender offer. Furthermore, we classify a pre and post crisis sample to investigate the effect of the crisis on the wealth effects. We find a positive AAR around the announcement date and a negative effect around the announcement of the preliminary results which can partly be explained by the position of clearing price within the price range. The crisis has diminished the overall wealth effects and enlarged the drop in share price after the announcement of the preliminary results.
Key Words: Share repurchase, Dutch auction tender offer, Share price development, Crisis.
Name Student: N.F. ten Doesschate Student ID number: 1722859 Study Program: MSc Finance
2
I. Introduction
On July 17, 2013, the Dutch listed company Corbion (formally CSM N.V.) announced an unusual share buyback method, a Dutch auction tender offer, which caught my attention. Corbion sought to return €250 million of the proceeds of the divestment of their bakery supplies businesses to the shareholders. The company returned cash via dividend, a regular open market share repurchase and the major part (€145 million) via this Dutch auction tender offer. The share repurchase on its own is not that special as share repurchases by large and small companies have grown in popularity in recent years. The total amount spent on repurchases was $197 billion in 2004 and $590 billion in 2007 (Yook & Gangopadhyay, 2010). However, the Dutch auction tender offer method has only been used twice in the Netherlands in the past ten years and is one of the least used methods of returning capital to shareholders in the world.
Firms may repurchase their own shares for different motives such as altering capital structure, cumulating shares for employee stock option plans, fending off taking over threats, distributing excess capital, taking advantage of undervaluation etc. (Zhang, 2005). There are essentially three ways that companies repurchase shares: (1) the fixed price tender offer, (2) the Dutch-auction tender offer, and (3) the open market repurchase program (Grullon & Ikenberry, 2000). According to Zhang (2005) open market share repurchases are one of the common ways for corporations to distribute cash flows to their shareholders as these programs represent 90% of all announced repurchase programs in the US.
Next to open market share repurchases there is the option of a share repurchase via a tender offer. A tender offer is a tax efficient way to return a large amount of capital to the shareholders of a company. Under the Dutch auction share repurchase method, the firm sets a range of prices from which each shareholder who elects to tender must select a single price at which to tender. At the end of the offer period, the firm repurchases shares according to the ascending order of shareholder bids until the specified number of shares has been assembled. However, rather than paying each shareholder his bid, all shareholders are paid the same price which is the clearing bid or highest bid accepted (Gay et al., 1996).
3
The objective of this study is to examine the wealth effects during a Dutch auction tender offer. This study differs from previous studies in several important ways. Most literature on Dutch auction tender offers has investigated the wealth effects around the announcement date. However, as several other points during the tender offer might be important for the wealth effects, this study includes an examination of the wealth effects during the tender offer with a focus on the announcement of the preliminary results. There are five points of interest during a Dutch auction tender offer; the announcement date, the start of the tender period, the end of the tender period, the announcement of the preliminary results and the announcement of the final results. In three of these points of interest new information is conveyed to the market; the announcement of the Dutch auction, the preliminary results and the announcement of the final results. The type of information that is conveyed in these announcements will be elaborated on further in this study. Graph 1 shows a timeline of these events during a Dutch auction tender offer. The numbers in the graph represent the mean of the number of days between the several points of interest during the Dutch auction tender offer.Graph 1; Timeline of a Dutch auction tender offer.1
The announcement of the preliminary results is chosen as the event date because this is the first time after the announcement (date) of the Dutch auction that new information is conveyed to the market. The announcement of the preliminary results is chosen over the announcement of the final results as in more than 95% of the firms used in the sample, the final results do not differ from the preliminary results and therefore do not convey new information. By taking the announcement of the preliminary results as the event date, this study develops an opportunity to examine the wealth effects of the Dutch auction tender offer from a new perspective. Since the outbreak of the financial crisis in 2007, the risk appetite of
1
4
shareholders has changed due to the crisis as the risk appetite fell and investors sought to reduce the size of their exposures to risky trades (Kirkpatrick, 2009). As cash is used to repurchase the outstanding shares in a Dutch auction the leverage ratio of the company will increase. With the change in risk appetite, it is most likely that the shareholders are not as happy with the increase in leverage due Dutch auction share repurchase as before and the wealth effects are likely to diminish. There is almost no major research paper which focuses on the effects of the crisis on the wealth effects during the Dutch auction tender offer. Therefore, this thesis implements a research period including not only the years before but also the years after the outbreak to investigate the effects of the crisis on the wealth effects during a Dutch auction tender offer. Finally, the position of the clearing price within the price range of the Dutch auction is tested in order to find if this has an effect on the wealth effect of the announcement of the preliminary results. As the main objective of this study is to examine the wealth effects during a Dutch auction tender offer I have developed the following research question;Research Question; What are the wealth effectsduring a Dutch auction tender offer period?
The thesis is organized as follows. First, I will review prior literature concerning returning capital, Dutch auction tender offers, its impact on the share price development and the rationale behind the process. Second, the methodology and the data set are discussed. The empirical results are presented in the following section. Next, I will elaborate on the findings of the event study and the paper will end with a brief conclusion and discussion.
5
II. Literature Review
A. Returning Capital
One of the main principles in corporate finance is maximization of the market value of the outstanding securities by managers (Vermaelen, 1984). Companies can return capital to their shareholders via several methods. By far the most commonly used methods are dividend payout (in addition to their regular dividend policy) and share repurchases (Lee & Rui, 2007) (see graph 2). The appropriate method to return capital is very much dependent on the size of the envisaged return and any other special circumstances. For decades, U.S. corporations have overwhelmingly preferred to pay out cash in the form of dividends rather than share repurchases (Grullon & Michaely, 2002). The most straightforward method for returning capital is via dividend and in that case, in addition to their dividend cycle for regular payments, special dividend can be used. A special dividend is a larger dividend compared to the regular dividends paid out by firms and the special dividend ensures the passing on of the benefits of the excess cash to the shareholders. This method is Earnings per Share (EPS) dilutive as no additional earnings are acquired. A major hiccup of the special dividend method is the (potential difference in) dividend withholding tax for foreign shareholders. The dividend withholding tax is different in most countries and as most companies have foreign shareholders, the benefits of special dividend are not equally divided to all shareholders. The special dividend method is therefore not the preferable capital return option.
Graph 2; Options for companies to return capital to its shareholders.
In contrast to the dominance of dividend payout, as stated by Grullon and Michaely (2002), share repurchases have increased over time and now constitute an economically important source of payouts
6
(Lee & Rui, 2007). The total amount spent on repurchases has increased from $ 197 billion in 2004 to $590 billion in 2007 (Yook&Gangopadhyay, 2010). Several authors have provided reasons for the increase in share repurchases over dividend payout. Jagannathan et al. (2000) propose a flexibility hypothesis, which posits that dividends represent an ongoing commitment and are used to distribute permanent cash flows, while share repurchases are used to pay out cash flows that are potentially temporary for the firm. Share repurchases thus preserve the company’s financial flexibility relative to dividends because they do not implicitly commit the firm to future payouts (Lee & Rui, 2007). Rau and Vermaelen (2002) present regulatory changes , such as the softening of restriction on buyback prior to eraning releases, as the main determinant for the increase in repurchases. According to Bohman (2006) the rise in share repurchases can be the result of learning effects by managers and shareholders. Bohman (2006) argues that the probability for a firm to initiate a repurchase program is heightened if the firm has a board connection to other firms that already initiated a repurchase program during the past 365 days. Therefore the managers and shareholders learn and gather information from other firms. Brav et al. (2005) state that many managers now favor repurchases because they are can be used in an attempt to time the equity market or to increase earnings per share. Share repurchases are earnings per share enhancing (EPS) because the company’s profit is allocated over less outstanding shares of common stock which increases the earnings per share. Patgiri et al. (2004) give the decreasing time horizons of shareholders as an explanation as they state that buybacks are used if managers want to appease short-term oriented shareholders. Furthermore, from a tax perspective, there is an obvious incentive for corporations to substitute share repurchases for dividends because capital gains are taxed at more favorable rates than ordinary income (Grullon & Michaely, 2002). Therefore, the share repurchase is a cheaper way to return capital than via dividend. This is called the substitution hypothesis. The combined trend of a decreasing reliance on dividend payment and the increasing reliance on repurchases also implies that nowadays, a more appropriate tool of valuation is total payout rather than dividend pay-out (Grullon & Michaely, 2002).7
causes the firm’s stock to be mispriced. When the stock is undervalued, the lower cost for the shares repurchase will translate into profits when these shares are reissued in the future. Moreover, Peyer and Vermaelen (2009) state that signaling effects could result from a deliberate attempt to signal positive information or could be unintentional, i.e., firms repurchase shares for a variety of reasons, but only when the stock is undervalued. The authors find evidence that share repurchases are a response to market overreaction to bad news: significant analyst downgrades, combined with overly pessimistic forecasts of long-term earnings. Second, Jun et al. (2009) state that repurchasing equity reduces a firm’s assets which produces a decrease in the value of claims protecting bondholder’s interests. Therefore the repurchase can transfer wealth from bondholders to equity holders (wealth transfer hypothesis). The free cash flow hypothesis presents the repurchase of shares as a way to distribute a firm’s excess cash flow to prevent wasteful investments. According to Liang et al. (2013) agency theory argues that when a firm’s free cash flow is in excess of investment opportunities, managers have an incentive to engage in poor performing (or high risk) projects, which increases their private expected benefit but reduces firm value. Grullon and Michaely (2004) state that the reduction of free cash flows by repurchasing shares helps to improve the firm’s market return. This can be explained by the fact that they find that the market return is significant with companies that are more likely to overinvest. Shareholders therefore highly value the share repurchase. Both the free cash flow hypothesis and the wealth transfer hypothesis are increasing leverage and are in that sense methods to increase the leverage position of a firm. Empirical evidence is generally supportive of the signaling hypothesis (Jun et al., 2009).A share repurchase can essentially be executed in three ways: (1) the fixed price tender offer, (2)
8
buy back a maximum amount of the average daily volume of the shares according to the market abuse directive. The average daily volume must be based on the average daily volume traded in the month preceding the month of public disclosure of the buyback program and fixed on the bases for the authorized period of the program. If a company buys up to 25% of its average daily volume, this will have a distorting impact on the share price development and arbitrage opportunities arise. This is a major hiccup for companies that use the open market repurchase program. Moreover, the program does not ensure that all shareholders benefit from the capital return.Finally, there is the option of the tender offer which can be executed in two ways; via the fixed-price tender offer and the Dutch auction tender offer. In the next section I will elaborate on the differences between the two tender offers and the pro’s and con’s of both share buybacks.
B. Tender offers
9
these shares. When a firm decides to cancel the shares the nominal share capital will decrease proportionally with the number of cancelled shares. The cancellation of purchased shares provides a clear and strong signal to investors and the cancelled shares are not taken into consideration anymore for the EPS calculation. In any case, those shares lose their voting rights and rights to cash flows (Allen & Michaely, 2003).In a fixed-price tender offer, the corporation, through an investment bank, offers to purchase a portion of its shares at a fixed price (Allen & Michaely, 2003). Typically a premium is set to the prevailing market price before the announcement date. The tender offer remains open either for a pre-determined period or until the desired number of shares is tendered. If an insufficient number of shares are submitted the tendering firm can withdraw the tender offer, extend the offer beyond the originally specified time period or purchase shares in excess of the number specified. However, the firm usually reserves the right to increase the number of shares repurchased if the tender offer is oversubscribed, and/or to buy shares from the tendering shareholders on a pro-rata basis. If the offer is not fully subscribed, the company has the right to either buy the shares tendered or to cancel the offer altogether.
The second tender offer possibility is through a Dutch auction tender offer. With a Dutch auction tender offer a company announces a buyback program up to a certain price (the premium). Shareholders have to submit the number of shares they are willing to tender at which price. The tendering company maintains the flexibility to decide on the total amount to be repurchased by way of communicating an “up to” amount prior to the start of the subscription period. The lowest price is chosen at which the book covers the desired amount to be repurchased (reverse book-building). For avoidance of doubt, the set price is the highest price in the allocated book so that even shareholders who have tendered below the set price will receive the set price for their allocated shares to assure an equal treatment of all shareholders. Retail investors are offered the choice to participate at strike (“bestens”) or to submit a price at which they tender their shares. If the book is oversubscribed all tendered shares below the set price are fully allocated and the shares tendered at the set price will be allocated pro-rata. Throughout the process, no insight in the book building process will be communicated to the market.
Pro’s and Con’s of a Dutch auction tender offer vs. fixed price
10
signals of undervaluation (“signal-oriented”) while Dutch auction tender offers are more effective as takeover deterrents (“control-oriented”). Also, Erwin and Miller (2001) find that firms adopting Dutch auctions are more likely to be takeover targets prior to the repurchase program than are firms that undertake fixed-price offers and use the Dutch auction tender offer as take-over deterrence. According to (Ahlqvist, 2004) this can be explained by the fact that a repurchase could further result in an increased price on the company’s shares, possibly deterring a potential buyer since the cost of the target would amplify. Comment and Jarrell (1991) however provide evidence that Dutch auctions are less effective as signals of stock undervaluation because they typically expose the personal wealth of managers to less risk than fixed-price offers. This is caused by the fact that the maximum potential loss to the owner-manager’s wealth if the buyback fails as a signal is lower than with a fixed price tender offer as the minimum premium paid with a Dutch auction tender offer is substantially lower. Furthermore, the authors argue that consistent with the theory of rational choice, Dutch auctions are used more frequently by large (widely followed) firms, compared with fixed price offers, suggesting that Dutch auction users have relatively lower demand for signaling stock undervaluation.Fixed price tender offer
11
Dutch auction tender offer
Equal to the fixed price tender offer, a Dutch auction tender offer has several advantages that all tender offers have. In contrast to the fixed price tender offer it provides an equal treatment of the shareholders. Moreover, an increase of shareholder loyalty can deter future aggressors. A Dutch auction tender offer is the best share repurchase method to enhance shareholder loyalty. The Dutch auction tender offer leads to optimal pricing as the book building dynamics will lead to the lowest price. It allows market forces to work within a price range that is acceptable to the company and at the same time minimizes the required cash outlay. This will ensure support and commitment from non-tendering (and thus loyal) shareholders because they know that, due to optimal pricing, the company pays the least amount possible for the Dutch auction tender offer and therefore this is in the best interest of the non-tendering shareholders. Moreover, the loyal shareholders know that via a Dutch auction the largest possible amount of shares is repurchased which ensures optimal earnings per share enhancement. In case of under-subscription, this can be explained as a sign of strength as current shareholders are apparently not prepared to sell within the range of the premium (e.g. existing shareholder commitment). In case of over subscription, the company can announce that is has completed a successful transaction and it offered shareholders a good deal and opportunity to sell their shares with a premium. Therefore, the fact that positive communication is possible on whatever outcome is an incentive for a company to make use of a Dutch auction tender offer. Furthermore, the Dutch auction is cheaper than a fixed price tender offer. The amount of shares sought in a Dutch auction can be announced in two ways, by announcing the value of the shares a company wants to repurchase or by announcing the number of shares a company wants to repurchase. If a company states that it wants to buyback a specific number of shares at a given price range; it provides a near certainty of achieving the target value of shares. In contrast, an important drawback of a Dutch auction for a company is the uncertainty regarding the number of shares repurchased. If a company seeks to by a value of shares, the number of shares ultimately repurchased is uncertain. Another drawback is the exposure of the company to market conditions during the tender offer period.
12
Table 1. Overview of the pro’s and con’s of the different tender offer methods for both the company and its shareholders. Description Company ShareholdersPro’s Con’s Pro’s Con’s
Fixed price tender offer
- EPS enhancing
- Efficient method to repurchase large block of shares
-Expiration date is set in advance -Avoids prolonged market risk
-Certainty about the size of the transaction (in case of success)
- If share price fails, company does not benefit
- Option is offered to all shareholders, but size limits participation of all shareholders - Requires premium relative to prevailing share price pus execution costs - Setting the price too low will lead to insufficient tendered shares, Setting the price too high will lead to “rationing”
-EPS enhancing - Tax friendly distribution - Fast execution
-Firm commitment about the value of the company by the management of a firm
- Allocation results in unequal treatment for shareholders
- Possible negative effect on liquidity
-Program cannot be adjusted during execution
Dutch auction tender offer
-EPS enhancing
- Efficient method to repurchase large block of shares
- Cheaper than a fixed price offer
- Near certainty of achieving the target number of shares
-Uncertainty regarding the size of the pay-out
-Higher transaction costs than in open market
-EPS enhancing - Tax friendly distribution - Fast execution
-More complex transaction to understand -Possible negative effect on liquidity
As can be seen in table 1 there are multiple benefits, both for company and shareholders, which arise from using the Dutch auction over the fixed price tender offer. However, the Dutch auction is one of the least used capital repayment program all over the world. In the Netherlands this type of tender offer has only been used twice over the past ten years. This might be so due to the fact that a Dutch auction tender offer is a complex transaction to understand for the shareholders and many considerations (tax, time etc.) have to be taken into account.
C. Dutch auction tender offer process and its wealth effects
The date on which the Dutch auction tender offer is announced (announcement date) is the starting point of the Dutch auction tender offer process. The Dutch auction tender offer ends when the final results of the tender offer have been announced. During the Dutch auction tender offer process there are several structuring considerations which play a role such as the price range, the timing of the process, the impact on the shareholder wealth and legal aspects. In order to explain the wealth effects during a Dutch auction, I will elaborate on the considerations and the rationale behind the tender process during the three events in a Dutch auction tender offer process that convey new information; the announcement date, the
announcement of the preliminary results and the announcement of the final results. These considerations
13
Announcement date of the Dutch auction
The announcement of the Dutch auction conveys information on the range, the start, length and end date of the tender bid period and the amount of shares sought to be repurchased. To assure equal treatment of all shareholders, the tender period must be sufficient to reach all investors. As it is a complex tender offer to understand, smaller (retail) shareholders must be able to take notice of the Dutch auction and to prepare. Usually the tender period is around 20 working days. Some companies announce at the date of the start of the tender bid period; others do this way in advance. Next to transparency issues, a company may not withhold information to the market, the fact that several firms announce way in advance is due to the fact that they have to ask for approval by the shareholders for the Dutch auction tender offer., This is a legal issue which is company and firm specific. The literature on the wealth effects of Dutch auction tender offers is not extensive as most existing studies test for Dutch and fixed price tender offers combined. In the following paragraph I will elaborate on the wealth effects found by several empirical studies in which some use a combination of the two. This will help me substantiate the rationale behind the wealth effects around the announcement date of the Dutch auction tender offer.
14
they are credible managerial signals that the offering firm’s stock is undervalued. These authors find an 8% average excess return around the announcement of a Dutch auction tender offer. Yook & Gangopadhyay (2011) find a positive abnormal return of 7.43% after the announcement date for a Dutch auction tender offer. Overall, the authors find that Dutch auction tender offers experience positive wealth effects around announcement date.Announcement of the preliminary results
What happens to the share price when the tender offer is completed? After the end of the tender bid period nothing should happen as no new information is conveyed to the market. The first signal of the course of the Dutch auction is given in the preliminary results which are announced a day after the end of the Dutch auction tender offer (see graph 1; table 2). The preliminary results convey the (preliminary) clearing price, the number of shares acquired, the total value of the shares acquired and sometimes the acceptance rate and in most case the proration factor. The proration factor is the percentage of properly tendered shares that is remitted if the Dutch auction is overtaken. In case the tender offer was oversubscribed the share price might drop as investors are holding a longer position than intended. In case the tender offer was undersubscribed the share price might increase as investors think that the company is worth more than the offered price. Bagwell (1991) states that in case of under subscription, this can be explained as a sign of strength as current shareholders are apparently not prepared to sell within the range of the premium (e.g. existing shareholder commitment). The oversubscription rate is however rarely announced as the management of a company feels that a huge oversubscription shows that shareholders want to leave the company for less than the offered price. Whether a Dutch auction tender offer is oversubscribed can\therefore not be tested for. Another aspect that influences the share price after the announcement of the Dutch auction is the fact that the premium is omitted. Since tendering forces the realization of capital gains, the tender price must include a premium relative to intrinsic value to compensate for the additional tax burden. The larger the capital gain of the marginal tenderer, the greater is the clearing tender price (relative to intrinsic value). This premium causes a wealth transfer from non-tendering to non-tendering shareholders, inducing a negative stock return at offer expiration (Kadapakkam & Seth, 1997). Therefore, at the end of the tender bid period, the share price will face a sharp decline as the premium paid by the company is omitted at the end of the tender offer. There is, however, another aspects that influence the share price after the event and which can be tested for; the position of the clearing price within the price range of the tender offer.
15
‘floor’ for the share price. There are several possible price range premiums possible: the smaller premium and larger premium. The smaller premium is up to 10 % and the larger premium is 10% or larger versus the closing share price prior the tender period. A smaller premium is a suitable solution if the share price has increased sharply due to announcement of the tender offer and limits the market discussion on whether the tender offer premium is an indication for a potential public offer premium. Moreover, it is an optimal EPS accretion, relative to a larger premium, for loyal shareholders as the non-tendering (and thus loyal) shareholders will not be harmed. A larger premium comes with a higher chance of a full take up of the tender offer and supports share price optimization. The negative sides of a larger premium are the possibility of a buyback at a high price, and this higher price per share limits the number of shares being repurchased and thus limits the increase of EPS for the loyal shareholders. Furthermore, there could be a potential drop of share price after completion because a higher premium leaves more room for shareholders to tender their shares at a potentially higher price, i.e. there is a higher chance of oversubscription. Oversubscription leads to a sharp drop in the share price as investors might think that the company is less worth than the offered price.In the preliminary results the company announces the clearing price for which they have bought the shares they were willing to buy. The clearing price is offered to all shareholders as offering shareholders different prices at the same moment in time would result in discrimination between the different shareholders. Legal advisors feel that shareholder discrimination by firms is not appropriate and it would open a discussion between shareholders on a non-transparent process. The final range in which the company ultimately sells the shares is an important issue for the companies which perform the Dutch auction tender offer. The Dutch auction is designed to minimize the price at which a deal can be completed. It allows market forces to work within a price range that is acceptable to the company and at the same time minimizes the required cash outlay. If the price is at the downside of the price range this can be seen as a sign of under confidence in the firm as people are first of all willing to sell their shares and second of all at the lowest price possible. Even the lowest clearing price within the price range is above the prevailing market price as a premium is offered by the company. Shareholders perceive this premium as a good deal by the company , do not agree with the managers’ view on the price range, and are offered a way out with a profit.
16
50 percent of the price range, this can be seen as a sign of confidence in the management of the company and its future growth potential and this will lead to a less dramatic drop in the share price.Announcement of the final results
The final results of the Dutch auction tender offer convey the final (and rarely adjusted) numbers which are already announced with the preliminary results. The same argumentation used for the wealth effects around the announcement of the preliminary results should hold for the wealth effects around the announcement of the final results as in most cases the results are not adjusted. The results are only adjusted because of legal requirements. Also, when there is a dispute by shareholders on the allocation of the shares the numbers can be adjusted. However, as the results are rarely adjusted, the announcement of the preliminary results is leading for the shareholders and it is expected that the shareholders will not react to the announcement of the final results.
Long term wealth effect of a Dutch auction tender offer
The positive effect on shareholder wealth after the announcement date is followed by a negative effect after the announcement of the preliminary results. Zhang (2005) researches stock price performance after actual share repurchases. He finds that on average, repurchasing firms do not exhibit strong superior abnormal performance over the end of the tender bid period. Dann (1981) states that the positive value changes in common shares are permanent in that share prices do not return to their pre-announcement date levels following expiration of the opportunity for stockholders to tender shares. Yook & Gangopadhyay (2010) examine abnormal returns for up to 6 months after the announcement of share repurchases and find that repurchase announcements are followed by a long-run return drift but do so for share repurchases without isolating the effects of a Dutch auction tender offer. Although several authors find positive long-term abnormal returns after share repurchases, this study will specify the long term wealth effects specifically for a Dutch auction tender offer.
D. Change in pre and post crisis wealth effects during a Dutch auctions
17
on a company's equity is expected to rise. Thus, financial leverage is used in various circumstances as a means of altering the cash flow and financial position of a company (Bhatti et al. 2010). As the conditions in the global financial markets changed, investors risk appetites sharply decreased due to a widespread re-pricing of risk (González-Hermosillo, 2008). González-Hermosillo argues that the investors’ degree of risk aversion reflects underlying preferences and, as such, it is expected to change infrequently over time. In contrast, risk appetite is likely to change more often as investors respond to changing levels of uncertainty in the macroeconomic environment due to, for example, a crisis. Thus, risk appetite depends on the subjective degree to which investors are willing to bear uncertainty and the overall level of uncertainty about the fundamental factors which drive asset prices (González-Hermosillo, 2008). The risk appetite of shareholders is driven by the crisis and has changed over time. Kirkpatrick (2009) confirms this idea and states that as losses arise in the equity portfolios due to the 2007 crisis, it is not surprising that a deleveraging occurred in portfolios. Risk appetite fell and investors sought to reduce the size of their exposures to risky trades.With the change in risk appetite after the crisis, the leveraging of firms via share repurchase is probably not valued as highly as before the crisis. Before the crisis leveraging went hand in hand with a perceived increase in shareholder maximization, while after the crisis shareholders have become more risk averse. Both the signaling by the management and the premium put in place by the management of a firm during a Dutch auction tender offer should result in an increase in shareholder wealth after an announcement of the Dutch auction. As these aspects stay equal in the sample, this allows for testing the effect of the pre and post crisis shareholders perception on the increase in leverage. Due to the change in risk appetite it is most likely that the shareholders are not as happy with the increase in leverage before , compared to , after the crisis and the long-term wealth effects and the wealth effects around the announcement are likely to diminish. Furthermore, as the overall wealth effects will diminish, the expected drop after the preliminary results is larger in the post crisis sample.
Hypotheses
18
Hypothesis 1; There is a positive effect on shareholder wealth after the announcement of a Dutch auctiontender offer.
This study will put the emphasis on the wealth changes around the announcement of the preliminary results which is the day after the end of the tender bid period. As stated before the preliminary results convey the first information about the course of the Dutch auction tender offer. With a Dutch auction tender offer, the preliminary results convey information on the clearing price. Overall, the share price is expected to drop as a cause of the fact that the premium has ceased to exist. Therefore;
Hypothesis 2; There is a negative effect on shareholder wealth after the preliminary results of a Dutch
auction tender offer have been announced.
Several factors are important for the wealth changes around the announcement of the preliminary results. The final clearing price is very important for the firm and its shareholders. To further investigate the cause of the wealth effects due to the announcement of the preliminary results the position of the final clearing price within the range is examined. A clearing price which lies in the bottom 50 percent of the price range shows under confidence of the shareholders in the management of the company. Shareholders want to get rid of their shares as soon as possible and at the lowest price possible. Therefore, I have developed the following hypothesis;
Hypothesis 2A; If the clearing price lies within the bottom 50 percent of the price range, the negative
effect on the shareholder wealth is the largest.
The final results do not ,in more than 95% in the sample, convey new information on the course of the Dutch auction process. As the preliminary results are therefore the leading results for the shareholders, the following hypothesis has been developed;
Hypothesis 3; There is no wealth effect after the announcement of the final results of the Dutch auction
tender offer.
19
Hypothesis 4; There is a positive long term effect on shareholder wealth due to a Dutch auction tenderoffer.
Since the outbreak of the financial crisis in 2007, there are almost no papers focusing on the effects of the crisis on Dutch auction tender offer share repurchases. This thesis implements a research period including not only the years before but also the years after the outbreak to investigate the effects of the crisis on the Dutch auction tender offer share repurchases. The increase in leverage that comes with the Dutch auction tender offer is assessed less valuable by the shareholders due to their change in risk appetite. This change in risk appetite will diminish the wealth effects of the Dutch auction after the 2007 crisis compared to the period before the crisis. Therefore;
Hypothesis 5; The positive effect on shareholder wealth due to the Dutch auction tender offer, is larger in
bull markets (before 2007) than bear markets (after 2007).
Hypothesis 6; The positive effect on shareholder wealth after the announcement date of the Dutch
auction tender offer is larger in bull markets (before 2007) than bear markets (after 2007).
Hypothesis 7; The negative effect on shareholder wealth after the announcement of the preliminary
20
III. Methodology
In order to test the wealth effects during the Dutch auction tender offer process this study will use the event study methodology of Brown and Warner (1980; 1985) and MacKinlay (1997). An event study tests the presence of abnormal returns in the determined event window, when an event takes place. An event is an unexpected happening within a certain organization or market segment that may influence the value of the concerning organization(s). To test the presence of abnormal returns, Brown and Warner (1980; 1985) make use of three kinds of models: The Mean Adjusted Returns Model, The Market Adjusted Returns Model and the Market and Risk Adjusted Returns Model. To apply these three models, the actual returns must be calculated first, using continuously compounding returns.
(1)
This formula calculates the actual return of stock i at time t, taking the logarithm of the percentage change in the share price of a particular organization, where indicates the return of the stock and represents the price of the stock.
The Mean Adjusted Returns model calculates the abnormal return by deducting the average return of the
estimation window from the actual return. This leads to the following equation;
𝐴𝑅𝑖,𝑡 = 𝑅𝑖,𝑡− 𝐸𝑖,𝑡 (2)
where 𝑅𝑖,𝑡 indicates the return of stock i at time t and 𝐸𝑖,𝑡 the average return of stock i at time t. 𝐴𝑅𝑖,𝑡 indicates the abnormal return of stock i at time t.
The market adjusted returns model calculates the abnormal returns by deducting the actual market return
from the actual return. There are two possibilities in which the actual market returns can be used. The first option is to use a stock index for all stocks. The problem with this index is that stock markets are included where some of these companies do not even trade. This might influence the results. This is why the second option is used, which is the stock market index at which stock i is being traded. This gives:
𝐴𝑅𝑖,𝑡= 𝑅𝑖,𝑡− 𝑅𝑚,𝑡 (3)
where 𝑅𝑚,𝑡 is the stock market return at time t.
21
The market and risk adjusted returns model takes the alpha and beta of the certain stocks into account
when calculating the abnormal return. Alpha indicates the volatility of the stock, caused by other factors than the market. Beta stands for the intensity of correlation between the stock and the market (Grinblatt and Titman, 2002). This means that the calculation of the abnormal return using the market and risk adjusted method takes the systematic risk of every stock into account. This model gives the following equation:
𝐴𝑅𝑖,𝑡 = 𝑅𝑖,𝑡− 𝛼� − 𝛽𝚤 � .𝑅𝚤 𝑚,𝑡 (4)
where 𝛼� and 𝛽𝚤 � are regression parameters estimated over the lockup contract period. 𝛼𝚤 � is calculated as 𝚤 follows: 𝛼� = µ𝚤 𝑖− µ𝑚. 𝛽� 𝚤 (5) is calculated as: i t i i
Var
Cov
,ˆ
=
β
(6)where is the covariance between stock i and the stock market return and represents the variance of stock i.
When the abnormal returns of the different models are detected, two tests can be used to define whether or not the abnormal returns in the event window significantly differ. Brown and Warner (1980, 1985) make a distinction between a parametric test and a non-parametric test. For the parametric test, the Student T-test can be performed. This test finds out if the Average Abnormal Returns (AAR’s), in the event window, significantly differ from zero. The AAR’s can be calculated by dividing the total amount of the Abnormal Returns by the number of events N:
(7)
The Student T-test can be used to test the Cumulative Average Abnormal returns for significance as well. To look for persistency, it is important to calculate the Cumulative Average Abnormal Return (CAAR’s). CAAR’s can be used to determine whether the effect of the announcement of a particular event can be recovered (Thomsen and McKenzie, 2001).
(8) i
βˆ
AAR
i,t=
AR
i,t i=1 N∑
N
22
When the Jarque-Bera test of the sample data indicates values higher than 5.99, the non-parametric test must be carried out, as the data does not come from a normal distribution (Brown and Warner, 1985). When examining the Hypotheses, the outcomes of the parametric test and the non-parametric test will be discussed to decide whether or not the hypotheses will be rejected. According to Brown and Warner (1980, 1985) and MacKinlay (1997) the outcomes can be tested at a significance level of 1%, 5% and 10%.Multivariate Analysis
To further investigate the wealth effects around the announcement of the preliminary results a cross-sectional analysis is conducted. The event window around the preliminary results (0, +5) is regressed on three independent variables. The independent variables are the position of the clearing price in the price range, market value of equity (firm size) and the return on assets (profitability of company). The empirical model for the regressions:
𝐶𝐴𝐴𝑅𝑖 = 𝑎0+ 𝛽1𝐶𝐿𝐸𝐴𝑅𝐼𝑁𝐺𝑃𝑅𝐼𝐶𝐸𝑖+ 𝛽2𝑀𝑉𝐸𝑖+ 𝛽3𝑅𝑂𝐴𝑖+ 𝜀𝑖 .
24
IV. Data
As Dutch auctions are not very common in general and especially not in the Netherlands (only Hunter Douglas and Corbion in the past ten years) I have changed the focus of the research to companies which were listed on stock exchanges in the United States and Canada. In the period 2001 to end 2013 I have found 121 companies, which performed a Dutch auction tender offer, by hand on various web pages and through the Lexis Nexis database. This period is also convenient to test for the effect of the 2007 crisis on the wealth effects of Dutch auction tender offers as the sample can easily be split into a pre and post crisis sample. The data set includes Dutch auction tender offers announced by firms that are listed on the NYSE, AEX, TSX or the NASDAQ. Next, I have cross-checked date of the announcement of the Dutch auction, the announcement of the preliminary results, the announcement of the final results and the start and end date of the tender bid period via the official press releases of the companies which are included in the dataset. The announcement date is the date that the company’s board of directors first authorizes the repurchase program via an official press release. The announcement dates of the preliminary and final results are the dates on which the preliminary and final results are announced via an official press release.
To test for abnormal returns, I make use of DataStream to acquire time-series of the stock return of the listed companies. The required data of 4 out of the 121 events could not be retrieved from DataStream and 3 other events happened too recent to be able to obtain plus 50 data set. Therefore, 117 Dutch auction tender offers have been used to test the share price developments during the announcement period and after the tender bid period. Specifically, in this study the market model parameters are estimated over 150 days prior to the event window from t-200 through t-51 where t is the date of the announcement of the preliminary results. This estimation is in accordance with the article of MacKinlay (1997). Furthermore, I make use of the same event window as Yook and Gangopadhyay (2010) who use an 21-day event window; consisting of 10 pre-event days, the event day, and 10 post-event days to focus on the wealth effects of the announcement of the preliminary results per day. Next to the focus on the daily AAR around the announcement of the preliminary results in the 21-day event window, six other event windows are determined; (-50, +31), (-30, +11), (0, +5), (+11, +50) and (-50, +50). These event windows are developed to test for the different events in which new information is conveyed.
Table 2. Mean of number of days between each point in the Dutch auction tender offer process
Mean (Days)
Announcement date -27
Start tender offer -21
End tender bid period -1
Preliminary results 0
25
Table 2 shows the mean of the number of days between each different points of interest (see graph 1) during a Dutch auction tender offer with the announcement of the preliminary results as day “0”. Next, I will elaborate on the different even windows and their link to the different hypotheses. The mean of the number of days from announcement of the Dutch auction to the announcement of the preliminary results is 27. Therefore, the event window (-30+11) is developed to test for the wealth effects of the announcement date of a Dutch auction tender offer which coincides with hypothesis 1. In order to test for the run-up period toward the Dutch auction tender offer and to take the outliers into account the estimation window (-50,-31) is developed. To isolate the main event of this study, the announcement of the preliminary results, an event window of (0, +5) is chosen to examine the wealth effects. As the announcement of the final results is made around 11 days after the preliminary results the estimation window (+11, +50) allows us to examine the wealth effects for this event. Finally, in order to test for the long term wealth effects of the Dutch auction itself, an event window (-50, +50) is chosen. This allows us to test whether a Dutch auction has long-term wealth effects as it is more or less 80 days after the announcement of the tender offer.Table 3 summarizes the descriptive statistics of the event results of the total sample of 117 Dutch auction tender offers within the 150 days of the estimation period. Table 2 summarizes the descriptive statistics for the sub samples used in this thesis.
Table 3. Descriptive analysis of the average returns of the total sample within estimation window [-200;-51].
Mean adjusted returns(A.D.)
Market adjusted returns(A.D.)
26
To measure whether the adjusted returns are normally distributed we will test the adjusted returns for skewness and kurtosis. Skewness measures the asymmetry of the probability distribution of the variable. The kurtosis measures the extent to which observations cluster around a central point. The value of the kurtosis statistic should be zero for a normal distribution (Huizingh, 2007). The variables are normally distributed when the absolute values of the skewness and kurtosis are not beyond the critical threshold of about 4. Furthermore, the Jarque-Bera statistic is used to test for normal distribution.Table 4. Descriptive analysis of the average returns of the total sample within estimation window [-200;-51]
As can be seen in both tables 3 and 4, the values of the skewness and kurtosis of all samples are within the thresholds. Moreover, the values of the Jarque-Bera statistics confirm that the data sets are normally distributed. Therefore, it is not necessary to make use of a non-parametric test.
Market and Risk
Adjusted Return Obs. Mean Median SD Minimum Maximum Skewness Kurtosis
Jarque-Bera Total Sample (Dutch
27
V. Results
Tables 5, 7 and 8 report the CAAR for various windows relative to the announcement (of the preliminary result) date. These windows are chosen to examine the wealth effects of the Dutch auction from the perspective of different points of interest during the Dutch auction process. As stated before, the (-30,-11) window is created to test for the effect of the announcement of the Dutch auction (Hypothesis 1). The (0, +5) and the (+11, 50) window are created to isolate, respectively, the effect of the announcement of the preliminary results (hypothesis 2) and the final results (hypothesis 3). To test for the overall wealth effect of a Dutch auction tender offer, the CAAR of 50 days before till 50 days after the event date is examined (-50,+50) (hypothesis 4). Table 6 displays, in detail, the average abnormal returns (AAR) surrounding the announcement of the preliminary results. This consists of a 21-day event window, of the three models, described by Brown and Warner (1980; 1985). The 21-day event window tables for the crisis sample and the sample of the position of the clearing price within the range can be found in the appendix. All tables also show the corresponding p-values of the parametric T-test.
Wealth effects total sample
Table 5 shows us that, as expected, the CAAR shows a large positive effect for the period “-30,-11” and is highly significant (significant at 1% level) for all three models. It can be concluded that there is a positive abnormal return on the “-30,-11” event window due to the announcement of the Dutch auction tender offer. Hypothesis 1 cannot be rejected. These results are consistent with many empirical studies which document significantly positive stock price reactions to (Dutch auction) stock repurchase tender offer announcements (Yook & Gangopadhyay, 2010; Dann, 1981; Lakonishok & Vermaelen, 1990).
Table 5. Cumulative Abnormal Average Returns and results of Student T-Test of the total sample; Preliminary Result Mean Adjusted
Model
Market Adjusted Model
Market and Risk Adjusted Model
Days CAAR P-value
Student T-test CAAR P-value Student T-test CAAR P-value Student T-test (-50,-31) 0.0128 <.0001*** 0.0050 0.2478 0.0101 <.0001*** (-30,-11) 0.0421 <.0001*** 0.0460 <.0001*** 0.0434 <.0001*** (-10,+10) 0.0008 0.4594 -0.0001 0.4933 -0.0050 0.0089*** (0,+5) -0.0099 0.0003*** -0.0235 0.0090*** -0.0120 <.0001*** (+11,+50) 0.0076 0.0022*** 0.0063 0.1960 -0.0007 0.3675 (-50,+50) 0.0653 <.0001*** 0.0407 <.0001*** 0.0476 <.0001***
Note: ***=significant at 1%, **=significant at 5%, *=significant at 10%
28
we cannot reject hypothesis 2. The CAAR’s for the three models in the event window (-50, -31) show mixed results and only the CAAR for the mean adjusted model is significant at the 5% level. This confirms the belief that shareholders only react to the announcement of the preliminary results. Hypothesis 3 cannot be rejected. Furthermore, it is interesting to see that during the Dutch auction tender process the share price shows a sharp increase after the announcement day and that the share price does not revert back to its mean within 50 days after the preliminary results have been announced as can be seen in table 5 (event window (-50,+50). Therefore, it can be said that a Dutch auction tender offer creates lasting long term wealth effects for shareholders (up till 50 days after the announcement of the preliminary results) and hypothesis 4 can therefore not be rejected. The fact that companies use a Dutch auction share repurchase to optimize their capital structure might contribute to these lasting wealth effects.Table 6. Average Abnormal Returns and results of Student T-Test of the total sample; Preliminary Result Mean Adjusted
Model
Market Adjusted Model
Market and Risk Adjusted Model Event Day AAR P-value Student T-test AAR P-value Student T-test AAR P-value Student T-test -10 -0.0019 0.2007 -0.0104 0.0807* -0.0005 0.4079 -9 0.0032 0.0755* 0.0094 0.1011 0.0013 0.2651 -8 -0.0002 0.4594 0.0079 0.1436 -0.0002 0.4682 -7 0.0022 0.1608 -0.0017 0.4065 0.0011 0.3013 -6 0.0022 0.1608 -0.0029 0.3426 0.0026 0.1103 -5 0.0006 0.3872 0.0186 0.0063*** 0.0020 0.1686 -4 0.0008 0.3596 -0.0206 0.0029*** 0.0016 0.2261 -3 0.0018 0.2161 0.0018 0.4037 0.0006 0.3794 -2 0.0006 0.3981 0.0012 0.4348 0.0002 0.4666 -1 0.0032 0.0782* 0.0197 0.0043*** 0.0022 0.142302 0 -0.0046 0.0204** -0.0169 0.0115* -0.0069 0.0006*** 1 -0.0027 0.1161 0.0101 0.0859* -0.0020 0.1647 2 -0.0037 0.0501* -0.0229 0.0011*** -0.0033 0.0599* 3 -0.0038 0.0456** -0.0005 0.4730 -0.0040 0.0304** 4 0.0029 0.0972* -0.0017 0.4075 0.0030 0.0781* 5 0.0027 0.1189 0.0085 0.1254 0.0012 0.2776 6 -0.0010 0.3281 -0.0011 0.4385 -0.0012 0.2808 7 -0.0028 0.1052 -0.0027 0.3562 -0.0032 0.0631* 8 0.0024 0.1428 -0.0009 0.4525 0.0013 0.2660 9 -0.0009 0.3487 -0.0113 0.0643* -0.0010 0.3179 10 -0.0002 0.4598 0.0165 0.0133** 0.0000 0.4872
29
The result in the event window (0, +5) is confirmed when we further investigate the days surrounding the announcement of the preliminary results. As can be seen in table 6, which shows the daily AAR of our main event window, the t-test shows significance on both day “0” and “+2” for all three models. On day “0” the preliminary results are announced to the public and this results in a negative AAR up till day “2”. The negative average abnormal returns last up till day “+3” and shows positive average abnormal returns on day “+4”. I assume that the company needs 3 days to recover from the announcement of the preliminary results as the market adequately need time to process the new information supplied. The results on day “+3”and “+4” are significant for only 2 models.Because the three different methods have similar results and the market and risk model shows most significance in the different event windows, the remaining research will focus on the market and risk adjusted method. The market and risk adjusted method include the volatility of the stock, caused by other factors than the market, and the intensity of correlation between the stock and the market. First, I will elaborate on the pre and post crisis wealth effects followed by the position of the clearing price. Second, the results of the multivariate analysis will be shown in which the position of the clearing price within the price range will be regressed on the CAR (0, +5) and controlled for the market value of equity and return on assets.
Pre and post crisis
30
Graph 3; Market and Risk adjusted AAR for the event window (-10, +10); Total sample, Pre Crisis and Post Crisis.However, the CAAR in post crisis event window (0, +5), table 7, is negative compared to the CAAR of the total sample. It can be stated that post-crisis the negative wealth effects are larger after the announcement of the preliminary results than the total sample. Therefore, it can be said that the during and after the 2007 crisis, the Dutch auction follows a much steeper decline in shareholder wealth after the preliminary results have been announced and hypothesis 8 cannot be rejected.
Table 7. Cumulative Abnormal Average Returns and results of Student T-Test of the total sample; Preliminary Result
Total Sample Pre-Crisis Post-Crisis
Days CAAR P-value
Student T-test CAAR P-value Student T-test CAAR P-value Student T-test (-50,-31) 0.0101 <.0001*** 0.0125 0.3441 0.0108 0.1863 (-30,-11) 0.0434 <.0001*** 0.0698 0.0141** 0.0284 0.0104*** (-10,+10) -0.0050 0.0089*** 0.0003 0.4965 -0.0084 0.2420 (0,+5) -0.0120 <.0001*** -0.0040 0.4482 -0.0179 0.0696** (+11,+50) -0.0007 0.3675 -0.0033 0.4571 0.0048 0.3461 (-50,+50) 0.0476 <.0001*** 0.0792 0.0067*** 0.0354 0.0021***
Note: ***=significant at 1%, **=significant at 5%, *=significant at 10%
Another very interesting result is the significant difference in the effect of the announcement of a Dutch auction tender offer on shareholder wealth. As can be seen by looking at the results of the CAAR in table 7 of the event window (-30,-11), the share price is increasing twice as much after announcement in the pre-crisis sample compared to the post-crisis sample and hypothesis 7 can therefore not be rejected. Furthermore, the overall CAAR over the estimation period (-50, +50) is twice as high in the pre-crisis
-0,01 -0,008 -0,006 -0,004 -0,002 0 0,002 0,004 0,006 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
31
sample. These results indicate that the wealth effects of the Dutch auction tender offer are impressively larger before the crisis struck. This result shows that the positive wealth pre-crisis wealth effect is lasting and the post-crisis wealth effect will slowly revert back to its mean after the announcement of the preliminary results. The results for this event window are significant at the 1% level and hypothesis 6 can also not be rejected. The difference in pre and post wealth effects is in line with the expectations. Kirkpatrick (2009) states that as losses arise in the equity portfolios due to the 2007 crisis, it is not surprising that a deleveraging occurred in portfolios. Risk appetite fell and investors sought to reduce the size of their exposures to risky trades. Graph 4 illustrates the significant large differences in shareholder wealth. Shareholders assess the increase of leverage that is caused by the Dutch auction share repurchase less valuable and the long term wealth effects and the wealth effect around the announcement date have diminished. As stated before, there are several possibilities for the wealth effects of the Dutch auction tender offer; the premium or signaling. The astonishing notion is the fact that because all other things are held equal, it seems that the effect of leverage is prevalent over other effects such as signaling.Graph 4; Market and Risk adjusted CAAR for the event window (-50, +50); Total sample, Pre Crisis and Post Crisis.
Position of clearing price within the range
Table 10 (see appendix I) shows us highly significant and negative average abnormal returns on day “0” up till day “+3” for the Dutch auction tender offers with a clearing price in the bottom 50 % of the price range. Firms within the bottom 50 % of the price range will be called lower range and firms within the top 50 % of the price range will be called upper range. These results are in line with the expectation that if shareholders are in the lower range they have no confidence in the firm and want to bail out even for the
-0,02 0 0,02 0,04 0,06 0,08 0,1 -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 35 40 45 50
CAAR Market and Risk Return Total Sample
CAAR Market and Risk Return Post Crisis
32
lowest price possible. This shareholder sentiment causes a sharp decline in the share price after the announcement and is an explanation for the share price to plummet. Hypothesis 2A can therefore not be rejected and the results show that the decline in shareholder wealth can be explained by the position of the clearing price within the final range. This can also be seen in graph 5, which clearly shows that the share price of the firms in the lower range decline steeper as a result of the announcement of the preliminary results of the Dutch auction.Graph 5; Market and Risk adjusted AAR for the event window (-10, +10); Total sample, Upper range and Lower range.
Table 8. Average Abnormal Returns, Cumulative Abnormal Average Returns and results of Student T-Test of the total sample;
Preliminary Result
Total Sample Upper Range Lower Range
Days CAAR P-value
Student T-test CAAR P-value Student T-test CAAR P-value Student T-test (-50,-31) 0.0101 <.0001*** 0.0109 0.3182 0.0074 0.3926 (-30,-11) 0.0434 <.0001*** 0.0433 0.0320** 0.0436 0.0570** (-10,+10) -0.0050 0.0089*** 0.0113 0.3129 -0.0551 0.0243** (0,+5) -0.0120 <.0001*** 0.0005 0.4921 -0.0501 0.0358** (+11,+50) -0.0007 0.3675 -0.0026 0.4559 0.0047 0.4309 (-50,+50) 0.0476 <.0001*** 0.0629 0.0038*** 0.0005 0.4918
Note: ***=significant at 1%, **=significant at 5%, *=significant at 10%
If we closely examine the wealth effects for the other event windows in table 8, we find that the position of the clearing price within the range does not have any effect on the abnormal returns after the announcement of the Dutch auction tender offer (-30,+11). This is obvious as the position of the clearing price within the range is not known till the announcement of the preliminary results. Another interesting
-0,04 -0,03 -0,02 -0,01 0 0,01 0,02
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 AAR Upper Range
33
event window to focus on is the (0, +5) window surrounding the announcement of the preliminary results. As expected by the results of the AAR’s in the (-10, +10) window the CAAR in the (0, +5) window plummets. However, the fact that the share price reverts almost back to its level before the announcement of the Dutch auction is truly spectacular (see graph 6). Therefore it can be stated that whether or not a Dutch auction creates lasting wealth effects depends on the position of the final clearing price within the range and ultimately in how the shareholders value the company. These results are backed by the highly significant and positive 6.29% CAAR over the (-50, +50) event window. Moreover, these results show that if “signaling” is the company’s motivation to perform a Dutch auction, this will only work if the shareholders believe that the management is right about the undervaluation of the stock.Graph 6; Market and Risk adjusted CAAR for the event window (-50, +50); Total sample, Upper range and Lower range.
Multivariate analysis
The event window around the preliminary results (0, +5) is regressed on three independent variables. Table 9 presents the regression results of the variables lower range, market value of equity and return on assets. The coefficient of the lower range variable is significantly negative in all the regressions even after controlling for firm size and profitability. The outcome of the regression strengthens the notion that the position of the clearing price within the price range has a significant impact on the negative wealth effect after the announcement of the preliminary results. However, the adjusted R-squared is only 0.0677 which means that only 6.77% of the variability of the CAR (0, +5) can be explained by the three independent variables. Moreover, it is interesting to see that the regression coefficients of the lower range variable in the three samples are more or less the same. This means that, although the overall drop in share price is
-0,04 -0,02 0 0,02 0,04 0,06 0,08 -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 35 40 45 50
CAAR Upper Range CAAR Lower Range
34
steeper after the crisis, the effect of the position of the clearing price on the wealth effects does not change.Table 9. Results multivariate analysis. T-statistics are in the parentheses.
Preliminary Results : Total
Sample CAR(0,+5)
Preliminary Results : Pre-Crisis CAR(0,+5)
Preliminary Results : Post Crisis CAR(0,+5)
Constant -0.0014 0.0051 -0.0033
(-0.2049) (-0.469) (-0.3351)
Lower Range -0.0363*** -0.0384** -0.0318**
(-2.72) (-1.9151) (-1.7149)
MVE 1.55E-07 1.84E-07 -3.20E-07
-0.7889 -0.9343 (-0.4301)
ROA 0.0164 -0.0022 0.0381
-1.0972 (-0.1061) -1.5742
35
VI. Conclusion
The main objective of this study is to examine the wealth effects during a Dutch auction tender offer with a focus on the announcement of the preliminary results of the tender offer. The results in the study confirm the findings of other research that there is a positive effect on shareholder wealth after the announcement of a Dutch auction tender offer. The positive cumulative average abnormal returns of up to 4.6% in the event window (-30,-11) are as expected. Previous research on the Dutch auction tender offer is limited to the wealth effects after the announcement date. This study contributes to the existing literature by examining other events that convey new information; the announcement date of the preliminary and final results. After the peak in the beginning of the Dutch auction process the sample shows a drop in the share price after the announcements of the preliminary results. The final results usually do not convey new information on the course of the Dutch auction tender offer and the results of this study have shown us that shareholders only take notice of the preliminary results and will react accordingly. In our sample, the drop in shareholder wealth after the announcement of the preliminary results does not balance out the rise after the announcement date of the Dutch auction tender offer. The wealth effects of a Dutch auction are lasting till around 50 days after the announcement of the preliminary results for the total sample. Graph 7 shows the share price development during the Dutch auction tender offer.
