Floris Reurts
Kabir, Cantrijn and Jeunink (1996) claim that there is a difference between the intentions of anti-takeover provisions implemented by Dutch firms and those implemented by U.S. ones. This difference exists because the disciplinary mechanisms at play in the Netherlands are different than those within the United States. In the former, concentrated ownership and supervisory boards exert a relatively more important role while for the U.S., the takeover market is an important disciplinary mechanism. When the biggest owner owns a bigger stake of the firm he will have more controlling power, therefore agency costs should decline. Because of this controlling power, hostile takeovers will be less necessary as an incentive for the managers to perform well. Therefore, in the Netherlands, anti-takeovers are more likely to be implemented as an assurance for not being taken over by a tender offer directly to common shareholders. A Dutch discount for Dutch shares might arise because of the limited power for common shareholders, since U.S. firms use anti-takeover provisions more often to increase the power of shareholders. For instance, to extract a higher takeover bid when a tender offer occurs. This study only finds a significant evidence for a more negative effect on shareholder value by classified boards established in the Netherlands. Additionally the results show that the effect on Tobin’s Q changes from negative in the year before the Troubled Asset Relief Program (TARP) became law, to positive in the year that the TARP was implemented
Content
Introduction 3
Literature review 6
Hostile takeovers 6
Anti-takeover provisions and how they affect shareholder value. 7
Managerial entrenchment hypothesis 7
The stockholder interests hypothesis 9
Anti-takeover provisions 9
The importance of ownership concentration 11
Empirical studies and their various results 11
Relations between anti-takeover provisions and shareholder value could be biased. 14
Insider trading 14
Multiple factors 15
Methodology 16
Hypotheses 16
The Event 16
Techniques and models 17
Dataset 19 Results 24 Evidence on Hypothesis 1 25 Evidence on Hypothesis 2 28 Conclusion 31 Literature 33
Introduction
In 2010 the Dutch telecommunications company KPN resisted a hostile takeover by América Móvil, a company that tried to acquire a majority share through buying shares directly from common shareholders. KPN used the anti-takeover provision called pre-emptive rights whereby preferred shareholders had the option to buy a massive number of shares against a price heavily below the market. This option diluted the shares acquired by América Móvil and made it more expensive and less attractive for the company to carry through the takeover attempt.
The takeover defense of KPN fended off the hostile takeover, but it also interfered with the market of corporate control. In countries like the U.S and the U.K. the takeover market is an important disciplinary mechanism for the good performance of management; in countries like the Netherlands–as shown before–anti takeover provisions destroy this disciplinary mechanism. Other, compensating, disciplinary factors also play a role. As Walsh and Seward (1990) argue, the failure of one mechanism triggers the existence of another. Therefore, within the Netherlands, concentrated ownership and supervisory boards exert a relatively more important role as disciplinary mechanism (Kabir, Cantrijn and Jeunink, 1996).
Corporate governance matters, since the control mechanisms used differ between countries. The structure of corporate laws and of ownership concentration can make one mechanism more attractive than another (La Porta et al., 1999). Several studies concerning differences in ownership concentration among countries have been carried out. For instance, La Porta et al. (1999) argue that firms in poor countries with lower shareholder protection and less regulation seem to have concentrated ownership. In addition, Tirole (2006), Franks, Mayer and Rossi (2007) and Becht and DeLong (2005) argue that ownership is more dispersed in the U.S. than elsewhere. Cramers and Nair (2005) and Gompers, Ishii and Metrick (2003) studied how internal disciplinary mechanisms like block-holdings or boards of directors, and external mechanisms like hostile takeovers work together in a system to control governance. They found that that take-over vulnerability and internal governance generate abnormal returns.
Because of the different disciplinary mechanisms among countries, Kabir, Cantrijn and Jeunink (1996) argue that anti-takeover provisions are used for different reasons. Where U.S companies mainly use anti-takeover provisions to protect shareholder interest, Dutch ones use them to limit the power of common shareholders (Kabir, Cantrijn and Jeunink, 1996). Anti-takeover provisions established by Dutch firms are claimed to be used as an assurance against a takeover while the U.S. firms would merely establish them to generate a
higher offer price when a tender offer occurs (Kabir, Cantrijn and Jeunink, 1996).1 The different disciplinary mechanisms, and therefore also the different intentions for using anti-takeover provisions, could have divergent effects on shareholder value.
There are two different theories concerning the wealth effects of anti-takeover provisions. The managerial entrenchment hypothesis by Cary (1969) and Williamson (1975) posits that strategies to prevent hostile takeovers reduce shareholder wealth because they increase management decision-making privileges and the costs of dismissing badly-performing managers. Also, the shareholder interests hypothesis of DeAngelo and Rice (1983), establishes that anti-takeover provisions enable shareholders to contract the managers more efficiently and receive a higher control premium when a takeover bid occurs. Researchers have not agreed between these two hypotheses, while evidence on the effect of anti-takeover provisions on wealth is mixed. Some researchers like Malatesta and Walking (1985), DeAngelo and Rice (1983), Jarrell and Poulsen (1988), Mahoney and Mahoney (1993), Gompers, Ishii and Metrick (2003), Bebchuck, Cohen and Ferrell (2004) and Kabir, Cantrijn and Jeunink, (1996) find significant negative share price effects, while others like Linn and McConnell (1983), McWilliams (1990), DeAngelo and Rice and Jarrell and Poulsen (1987) find a significant positive or insignificant price effect.
The abovementioned points combined show that differences in regulations and ownership concentration exist between countries, and that these differences affect the use of anti-takeover provisions, which in turn affect shareholder value in different ways. However it remains unmeasured how the intentions of anti-takeover provisions, claimed to be different between counties, affect the share prices of firms that are listed in those countries. By answering the following question, the intentions of using the provisions that are most beneficial for shareholders might become clearer.: To what extent is there a relation between the share prices of Dutch companies and their use of anti-takeover provisions compared to U.S. companies and how does this relation evolve when U.S. companies are forced to amend their use of a provision? This study examines a possible discount on Dutch shares, compared to U.S. shares, and whether the implemented anti-takeover provisions cause the discount. In addition, a shock will be used, namely the prohibition of golden parachutes through the Troubled Asset Relief Program (TARP), to measure the effect change of golden parachutes on
shareholder value before and after the TARP came in to law. 2 The TARP component of interest restricts the execution of golden parachutes for U.S. firms participating in the program in 2008.
The intention of this paper is to prove that there is a relation between the stock price of Dutch companies and their use of anti-takeover provisions. Thus, it will be determined whether Dutch stocks have been traded against a discount because of the excessive use of anti-takeover provisions that limit the power of common shareholders. Findings in the existing literature about the effect of anti-takeover provisions on shareholder value are mixed. The main contribution of this study to existing scientific literature will be introducing the different intentions of the anti-takeover provisions in explaining their effect on shareholder value. Additionally, an exogenous shock will be used to prove a relation between golden parachutes and the shareholder value of a firm.
To measure the possible Dutch discount, this thesis compares share prices and the effect of anti-takeover provisions on the share prices of Dutch and American companies. The dataset, extracted from WRDS and DataStream between 2004 and 2012, consists of 1350 U.S. firms and 111 Dutch ones. Various datasets within WRDS where merged and the Dutch and U.S data were manually combined.
To compare possible different effects of anti-takeover provisions on share prices a panel regression with fixed effects is used. To measure a possible “Dutch Discount” U.S. firms will function as a control group where Dutch firms are the group of interest. To measure the indirect effect of the TARP on shareholder value through the change in golden parachutes usage, the whole of the U.S. will function as a control group while TARP participants are the group of interest. Tobin’s Q will be used as the dependent variable as a measure of shareholder value and various interaction dummy variables are constructed to compare Dutch and U.S. firms and their use of anti-takeover provisions.
In this thesis significant evidence is found for a Dutch discount on Dutch shares compared to shares of US firms, but it remains a question whether this discount is mainly caused by the different intentions in the use of anti-takeover provisions. Compared to the U.S., only the use of classified board within the Netherlands shows a more negative effect on the Tobin’s Q, while supermajority voting rights and golden parachutes show a more positive effect. Furthermore, the shock used shows that the effect on Tobin’s Q changes from
2 Golden parachutes are substantial compensation packages paid to top executives when they are terminated after a hostile take-over has taken place
negative– in the year before the TARP came into law–to positive– in the year that the TARP was implemented.
The remainder of the paper is organized as follows. Section 2 provides an overview of the existing literature on the possible effect of anti-takeover provisions of interest on shareholder value and the effect of country characteristics on the use of anti-takeover provisions. Section 3 introduces the data and methodology while section 4 presents the empirical findings. Section 5 contains the conclusion of this research.
Literature review
Hostile takeovers
A hostile takeover is a situation in which the hostile bidder goes directly to the shareholders to acquire the company while the management does not want the deal to go through. To succeed, the acquirer needs to buy enough shares to obtain control of the company so it is able to replace the management and the board of directors. The upper level management and the board of directors of the target company will generally fight the takeover attempt. According to Berk and Demarzo (2007), the most important reasons for the board of directors to resist a takeover is that they believe the offer price is to low. There are two main reasons for the management to resist a takeover: when an acquirer offers a stock-swap for which the targets management believes that the acquirers’ stock is overvalued or, when management is afraid of losing their jobs. The latter will especially be the case when the primary motivation of the takeover is efficiency gains, as the management of the target firm could be seen as superfluous in most cases (Berk and Demarzo, 2007).
There are seven types of anti take-over provisions that managers and the boards of directors could use to prevent a hostile takeover: (1) poison pill, (2) staggered and classified board, (3) white knights, (4) golden parachutes, (5) recapitalization, (6) voting right restrictions, and (7) regulatory approval. Anti–takeover provisions increase the cost of establishing a successful hostile bid for control of the company. A possible bidder is encouraged to directly approach the target firm’s management and board of directors instead of placing a tender offer or starting a proxy fight (Maletesta and Walkling, 1987). The provisions mentioned before can be used in favor of the management; this is known as the managerial entrenchment hypothesis (Cary, 1969 and Williamson, 1975). They can also be used in favor of the board of directors and the shareholders, also called the shareholder interest hypothesis (DeAngelo and Rice, 1983).
The next part of this study only displays the anti-takeover provisions of interest and discusses the mixed results of previous studies on the effect of these anti-takeover provisions on shareholder value. Not all will be discussed because Bebchuck, Cohen and Ferrell (2004) only find evidence that six anti-takeover provisions are significantly correlated with shareholder value. The next part discusses literature on the effect of corporate structure, ownership structure, country regulations and other factors on the use of anti-takeover defenses. There, the assumption made by Kabir, Cantrijn and Jeunink (1996) that especially U.S companies use anti takeover provisions to protect shareholders interests while Dutch companies mainly use them to limit the power of common shareholders, is further elaborated.
Anti-takeover provisions and how they affect shareholder value.
Managerial entrenchment hypothesis
The managerial entrenchment hypothesis developed by Cary (1969) and Williamson (1975) holds that anti-takeover amendments act mainly to improve the existing management’s job certainty and decision-making privileges at the expense of current shareholders. It recognizes that shareholders must pay a price to prevent managers from engaging in actions that harm shareholder interests and that the costs of managerial displacement through shareholder vote– a method for shareholders to constrain managers–increase when anti-takeover provisions are implemented. Thus, according to the entrenchment hypothesis, managers support and propose anti-takeover defenses for improving their job certainty and income level (DeAngelo and Rice, 1983).
The hypothesis takes as given that there is a separation of ownership and control. DeAngelo and Rice (1983) argue that disciplinary mechanisms are needed to reduce the agency costs that come along with the separation of ownership. There are several internal and external corporate governance mechanisms like managerial labor market forces, management shareholder contracting, but the most important mechanism is the threat of managerial replacement through shareholder vote (DeAngelo and Rice, 1983). According to Manne (1965) shareholder-voting mechanism can operate in several ways. For instance, a third party, like another corporation, can use a tender offer to buy a large part of the voting shares with the intention of replacing the management when control is obtained. Alternatively, once some shareholders recognize inefficiencies, they may apply for votes from others to face the underperforming management in a proxy fight for control of the firm. According to the Managerial Entrenchment Hypothesis, disciplinary mechanisms do not eliminate agency
problems because it is assumed that all mechanisms develop costs of their own, for example, gathering information from the management and subsequently contracting and monitoring them (DeAngelo and Rice, 1983). Borokhovich, Brunarski and Parrino (1997) show that on average a CEO’s compensation increases for firms that implement anti-takeover provisions and Bertrand and Mullainathan (2000) find a comparable result. They measured the effect of new state takeover laws on CEO compensation and plant-level inefficiency and found that the new laws drove up compensation and lowered efficiency. The latter variable was measured by return on capital or total factor of productivity. The theory developed by Baumol (1959) that already holds for decades posits that managers might undertake inefficient projects so that they can extract private profits. The more managers are entrenched and the better they can resist a takeover, the more difficult it is for shareholders to replace them. Thus, the new anti-takeover provisions will keep increasing the capital expenditures (CAPEX), which will keep having a more negative effect on firm value. Gompers, Ishii and Metrick (2003) therefore use these CAPEX to measure the effect of anti takeover provisions on firm performance. They find that firms within a group that excessively use defense mechanisms have a significant higher CAPEX than those that use considerably less. Note that this only proves a relation. To prevent the management from making these excessive expenditures, shareholders could threaten with replacing the management by a shareholder vote. Although the execution is a costly process, DeAngelo and Rice (1983) argue that shareholder vote is the least costly disciplinary mechanism to address some forms of management inefficiency, since managers see the possible shareholder vote as a threat to their tenure before any costs are made. However, as a reaction to this threat, innovative managers establish ways to minimize the possibility of being replaced by a shareholder voting process. One way of increasing the costs of displacing an incumbent manager are anti-takeover amendments to the corporate charter and bylaws, but since shareholders must vote for anti-takeover defense mechanisms, the managerial entrenchment hypothesis needs an explanation of why shareholders approve these amendments. DeAngelo and Rice (1983) mention three possible explanations. The first is that the behavior of shareholders is simply irrational. Another is that shareholders, like large institutional holders, vote for anti-takeover amendments to maintain a proper relationship with the existing management regardless of their negative effect on shareholder value. The final and most likely explanation is that fighting anti-takeover amendments requires information and transaction costs that are disproportionally large to rule out proper opposition. Because of the high costs compared to the benefits of gathering information, DeAngelo and Rice (1983)
argue that there will be more uninformed traders than informed ones. Additionally, informed shareholders will vote in a manner that goes with shareholder value maximization while the uninformed assume that they are voting for just another managerial issue that will result in greater shareholder value. This means that the majority of uninformed shareholders will support management on an anti-takeover amendment even when informed shareholders are aware of its unfavorable effect. Informed traders could attempt a proxy fight to supply the information of the detrimental effect to the uninformed, but this will only be beneficial when the benefits of expropriation outweigh the costs of such opposition (DeAngelo and Rice, 1983). In sum, the same transaction and information costs that make it difficult for shareholders to replace inefficient managers also prevent shareholders from resisting the managers entrenching themselves by implementing anti-takeover provisions.
The stockholder interests hypothesis
The stockholder interests hypothesis claims that the adoption of anti-takeover provisions increases current shareholder value (DeAngelo and Rice, 1983), since there are private level incentives for the shareholder as an individual to tender at a price that will be lower than what a combined group of shareholders could extract from a potential bidder. For the target shareholders, this collective resource problem negatively affects the total expected gain from relinquishing corporate control. Therefore, DeAngelo and Rice (1983) argue that existing shareholders may benefit from mechanisms that enforce a collective response to a tender offer for corporate control. Individual shareholders will have private incentives–a higher premium-to reject the tender offer and premium-to hold the spremium-tock until the bidder offers a higher price. Then, they argue that anti-takeover provisions could benefit shareholders because these could function as a form of long-term contract for existing managers. With these contracts, managers might have more incentives to invest in longer-term projects that in turn can be beneficial for existing shareholders. Finally they claim that anti-takeover provisions could be beneficial for shareholders because they prevent more costly ways of managers trying to secure their position within a firm
Anti-takeover provisions
Previous studies on the effect of anti take-over provisions on shareholder value used different types of provisions. The used provisions and the findings of these studies are discussed in the following section.
DeAngelo and Rice (1983) use corporate charter and bylaw amendments like supermajority voting provisions, staggered board and classified board, to measure the effect
of anti-take over provisions on shareholder wealth. When voting provisions are implemented, acquirers need a certain percentage of voting rights to get the merger approved and to take control of the company. Staggered board provisions split the board of directors into different classes of which only one can be reelected every year, usually for a three-year period. Classified boards are inherently staggered, but directors within these boards serve for different term lengths, depending on their specific classification. DeAngelo and Rice (1983) claim that staggered and classified boards make the company less attractive for hostile takeovers, since the effective transfer of control can be delayed for several years after the acquirer obtains the majority-voting shares.
Poison pills give the target stockholders the right to buy shares at a deeply discounted price in either the target or the acquirer. According to Maletesta and Walkling (1987), the poison pill is one of the most common and most effective defenses. The four principle types are: preferred stock plans, flip-over plans, back-end plans, and voting plans. Under the first poison pill, preferred stock method, a dividend of convertible preferred stock is distributed to the firms’ common shareholders that can convert the preferred stock to common stock after a takeover is completed. This strategy increases the costs of the merger and dilutes the ownership of the acquiring company. Under flip over plans, the firm implements a common stock dividend in the form of rights to purchase some of its own securities after the merger, which is generally common stock. In most cases, shareholders are able to buy these shares at a discount. This makes it possible for existing investors to get instant profits and, more importantly, they dilute the shares bought by the acquirer, which makes the takeover extremely costly (Maletesta and Walkling, 1987). The third poison pill type is the back-end plan, which is similar to flip-over plans. This is an anti takeover approach that gives existing shareholders the right to debt or cash securities at a price defined by the firms’ board at any time before a block-holder, the acquirer, attempts to obtain a substantial block of the existing firms’ shares. Finally, the most straightforward poison pill: voting plans, which are designed to prevent a single party from obtaining voting control of the company (Maletesta and Walkling, 1987). If a hostile bidder acquires a substantial block of the targets’ voting stock, preferred shareholders other than the block holder would be given super voting rights. This would dilute the voting power of the hostile bidder, and make the take-over less attractive.
Knoeber (1986) studies the effect of “shark repellants” and especially golden parachutes on the shareholder and managers’ wealth. Shark repellants are hostile take-over defense mechanisms in general while golden parachutes are substantial compensation
packages paid to top executives when they are terminated after a hostile take-over. Knoeber (1986) argues that tender offers from an outside party affect the contractual relation between shareholders and managers and that it may be in the best interest for both parties to restrict the possibility of this relation being disrupted.
The importance of ownership concentration
Kabir, Cantrijn and Jeunink (1996) argue that in some countries hostile takeovers serve as an external mechanism for corporate control, but that this role is diminished within the Netherlands since there are other internal mechanisms for control like concentrated ownership and supervisory boards. They investigate under what circumstances the hostile takeover mechanism is more or less effective and therefore examine the relationship between the firm’s ownership structures, takeover defenses and stock returns. Kabir, Cantrijn and Jeunink (1996) distinguish two types of corporate systems: the network-oriented and the market-oriented systems. The former features relatively developed financial markets with active markets for control by corporations that have widely dispersed ownerships. The latter is characterized by group membership of corporations, closely held corporations and an important role of banks for corporate control and financing. Kabir, Cantrijn and Jeunink (1996) argue that the former is prevalent in, for instance, France, Germany, Japan and the Netherlands while the latter is found mostly in the U.S. and the U.K. Furthermore they argue that differences in governance systems make countries adopt different defense devices and that there are two ways to classify defense anti-takeover provisions: technical or structural. The first type of takeover provisions like limiting voting power or the issuance of preferred defense shares are mainly directed to hinder hostile takeover bids while the second, like cross holdings and the importance of debt financing, arise from existing structures of equity ownership and stock market ownership. Kabir, Cantrijn and Jeunink (1996) claim that the Dutch situation –which uses technical measures– offers firms possibilities for defending themselves against hostile takeovers, many of which do not exist in the U.S. This would mean that preferred shareholders and the managers of Dutch firms were better able to entrench themselves and defend their company against hostile takeovers than those in U.S. firms. The differences between the defenses used can entail different effects on the shareholder value.
Empirical studies and their various results
DeAngelo and Rice (1983) used a sample of 100 NYSE and AMEX companies that announced supermajority provisions and staggered-board amendments between 1974 and
1979. With the use of different methodologies and time periods for estimating abnormal returns, they find an insignificant negative effect of these provisions on shareholder wealth.
Maletesta and Walkling (1987) studied the amendments in the use of abovementioned poison pills at 132 American firms. They calculated abnormal stock returns after announcements about poison pill amendments and they found, a statistically significant number pointing to the fact that poison pill defenses implemented by management only, have a negative effect on stock prices. Note that there was no discernable effect on share prices, when shareholders explicitly approved the amendments, unlike when the management circumvented the voting process . Thereby, when firms abandon their plans for implementing anti-takeover provisions or when these are prevented from being implemented by court rulings, they receive abnormal returns that are positive and statistically significant. Finally, Maletesta and Walkling (1987) claim that unprofitable firms adopted poison pill defenses more often as they compared the profitability of these firms with the profitability of the average firm in their industry.
Knoeber (1986) develops an empirical framework for 331 U.S. firms in which he shows that golden parachutes are in the best interests for both parties because they improve contracting possibilities. Golden parachutes, when the amounts are sufficiently large, could be advantageous because they are designed to provide assurance for managers against the opportunism of shareholders when a tender offer occurs. Because of an assurance and other shark repellants, managers are more certain that compensation agreements will not be reneged. Therefore, managers will be more eager to accept deferred contracts. Secondly, shareholders also benefit from golden parachutes, as they prefer differed contracts instead of immediate compensation because it is easier for them to measure the manager performance afterwards and reward him based on performance.
Kabir, Cantrijn and Jeunink (1996) used a manually gathered sample of 177 firms listed on the Amsterdam stock exchange to determine the use of anti-takeover provisions, and provisions affected by ownership concentration and whether anti-takeover amendments affect shareholder value. First, they found that firms with low ownership concentrations adopt more takeover defense mechanisms than firms with high ownership concentrations and this finding is statistically significant. This finding is consistent with Bergström and Rydqvist (1990), who empirically studied the effect of ownership concentration on the use of takeover defenses in Sweden. They found that firms with highly concentrated ownership seldom adopt anti-takeover defenses. They state that widely held firms, like those in the U.S., developed several
anti-takeover mechanisms incorporated in the corporate charter like forms of corporate restructuring and various types of poison pills. Additionally, they argue that when ownership concentration is high enough to ensure that the shareholder with the largest block can hold the controlling fraction, these anti-takeover defenses are rarely used. When these companies use anti-takeover provisions, the preferred shareholders mostly establish them to make sure no hostile bidder would succeed in a hostile takeover.
Then, Kabir, Cantrijn and Jeunink (1996) used event studies to calculate the effect of anti-takeover provisions on shareholder value and found that stock prices reacted positively on the announcement of the first step towards the possibility of preferred share issue. This indicates that shareholders think of making profit from higher possible takeover premiums, but this opportunism disappears when the share issue is actualized and the takeover attempt seems to be eliminated, because a preferred share issue announcement is associated with a strong decline in abnormal stock returns.
The abovementioned studies all used event-study methodology, where announcements of defense mechanisms amendments are used to analyze abnormal stock returns. Gompers, Ishii and Metrick (2003) argue that such event studies face the problem that defense amendments may be caused by simultaneous conditions at the firm. Newly implemented anti-takeover provisions might both provide a signal about possible anti-takeover bids and a change in governance structure. Gompers, Ishii and Metrick (2003) avoid these difficulties by creating a governance index combining a large set of governance provisions. The index is a simple measure. It does not reflect the relative impact of each provision individually but groups them together. This index is a proxy for the strength of shareholder rights and makes it possible to measure the effect of 24 different anti-takeover provisions on shareholder value in the long-term. The more provisions implemented by the firm, the less rights shareholders are assumed to have. Gompers, Ishii and Metrick (2003) used data from the Investor Responsibility Research Center (IRRC) drawn from mainly S&P 500 companies. They found that the dictatorship portfolio–firms with a low index–underperformed the democracy portfolio–firms with a high index–by 8.5 percent. In addition they found that the lowest index value in 1999 was negatively associated with an 11.5 percent difference in firm value as measured by Tobin’s Q.
Bebchuck, Cohen and Ferrell (2004) continue the work of Gompers, Ishii and Metrick (2003) and study which of the 24 IRRC provisions really matter and how they are correlated with Tobin’s Q. They find that 6 of the 24 provisions play a significant role in the correlation
between the provisions, in the aggregate, and shareholder wealth. Four of the provisions of interests determine the limits on shareholder voting rights: staggered boards, supermajority requirements for mergers and charter amendments, and limits to shareholder amendments of bylaws. The other two are golden parachute arrangements and poison pills, which are the most striking and well known measures taken to prepare against a hostile offer. With the six provisions Bebchuck, Cohen and Ferrell (2004) construct an index that they call the “entrenchment index” and they find that the provisions–both as aggregate and individually– have a significant negative effect on Tobin’s Q values while they find no evidence for the other 18 provisions.
Relations between anti-takeover provisions and shareholder value could be biased.
Insider trading
There might be a relation between anti-takeover provisions and firm value, operating performance and market returns, but it is difficult to imply a causal role since there could be several other plausible explanations than higher agency costs, to explain why the abovementioned anti-takeover provisions seem to affect shareholder value.
The theory by Gompers, Ishii and Metrick (2003) holds that managers who forecast their firm to perform badly could implement provisions to protect themselves from losing their jobs. This would mean the existence of inverse causality. Anti-takeover measures do not have to be the reason for bad performance and it could be that managers, especially at bad performing firms, choose to protect their jobs with provisions. Then when there is poor operating performance, the managers are not surprised but the market is. Hence shareholder value will diminish.
It is well known that returns can be forecast by checking insider trading. This means that firms whose shares were intensively bought by insider traders are about to over-perform in relation to comparable shares in the following period (Seyhun, 1998). If insiders forecast bad performance it is expected that they develop ways to protect themselves. They could do this by implementing anti-takeover provisions. When this happens, weak shareholder rights would be caused by superior information, but this does not have to be the reason for bad performance in the subsequent period. Gompers, Ishii and Metrick (2003) study the relation between the use of anti-takeover provisions and insider trading and argue that when managers forecast bad performance, insider selling should be negatively correlated with the governance
index they constructed. Within their U.S dataset Gompers, Ishii and Metrick (2003) do not find any significant relations between insider trading and governance, making this theory less plausible.
Multiple factors
Multiple factors could influence the potential performance difference between firms that use a different number of anti-takeover provisions. Gompers, Ishii and Metrick (2003) find that the use of anti-takeover provisions is correlated with several other factors like institutional ownership, S&P membership, past sales growth and trading volume. If the model designed does not capture any of these characteristics, then any type of omitted variable bias might influence the measured relation between anti-takeover provision and firm performance.
Gompers, Ishii and Metrick (2003) ran several regressions with all possible influences mentioned above and others, to check whether there remained any significant effects for the usage of anti-takeover defenses on firm performance. After controlling for all characteristics they gave the remains a negative effect but this turned out to be insignificant. Controlling for all characteristics, when they compare firms that use provisions excessively with those that make considerably less use of those measures they find that the latter generate a higher return for a significance level of ten percent. Although this finding is significant, there is no strong evidence as large standard errors could mean that the results have less statistical inference.
Most studies like Kabir et al. (1996), Maletesta and Walkling (1987), DeAngelo and Rice (1983), Gompers, Ishii and Metrick (2003) and Bebchuck, Cohen and Ferrell (2004), discussed within this literature review, are experiments without random assignment. Therefore it is difficult to say anything about the causal relationship between anti-takeover provisions and firm performance, which will have an effect on shareholder value. As mentioned before, Bebchuck, Cohen and Ferrell (2004) find that 6 out of the 24 provisions used by Gompers, Ishii and Metrick (2003) play a significant role in the correlation between the provisions, in the aggregate, and shareholder wealth. This study elaborates on this finding to research the different intentions of implementing anti-takeover provisions on shareholder value. Furthermore, it tries to lay a foundation for using a shock to say something about the relation between the usage of anti-takeover provisions and shareholder value. In so doing, there is full awareness about the existence of all other possible explanations and omitted variable bias.
Methodology
Hypotheses
In the previously discussed literature it is claimed that Dutch companies use anti-takeover provisions with other intentions than U.S. ones. Kabir, Cantrijn and Jeunink (1996) establish that the intention of Dutch companies to limit the power of shareholders has a more negative effect on shareholder value than the U.S companies that use anti-takeover provisions to increase the power of shareholders. This negative effect would mean that Dutch shares are traded at a 'Dutch discount' compared to U.S. companies. To examine whether there is a Dutch discount and whether it is caused by the use of anti-takeover provisions the following hypotheses are constructed.
H1: The anti-take-over provisions poison pills, golden parachutes, classified boards and supermajority voting requirements established by Dutch firms have a more negative effect on Tobin’s Q than those established by U.S. Firms.
H2: The TARP reduces the exercise–or the change of exercise–of golden parachutes by TARP participants within the U.S. and therefore reduces the negative effect on Tobin’s Q by golden parachutes established by these TARP participants.
The Event
The Troubled Asset Relief Program (TARP) is now used to show that a decrease in the use of an anti-takeover provision due to an “exogenous shock” will influence the effect of this provision on Tobin’s Q. Here the prohibition of the golden parachutes as of the year 2008 for firms that were supported by the U.S. government, forced participating companies within the U.S to dilute their golden parachute provisions.
The TARP was a program of the United States government to purchase equity and assets from several institutions to strengthen the financial sector. U.S. President George W. Bush signed the program into law on 3 October, 2008. The U.S. government implemented several regulations to address the mortgage crisis and the restriction on the exercise and establishment of golden parachutes was one of the components
Institutions who used an auction process to sell more than $300 million in assets were not allowed to sign new golden parachute contracts with any future executives. Thus, a limit was placed on annual tax deductions of half a million for golden parachute payments to executives.
Institutions at which the government acquired equity through a direct purchase had to meet tougher agreements. They had to eliminate compensation structures that encouraged excessive and unnecessary risk-taking by executives. They were also forced to let executives repay bonuses that were paid on the basis of incorrect data. Most importantly for this research, payments of golden parachutes established previously were prohibited.
Techniques and models
To examine the effect of anti-takeover provisions on the share price of Dutch companies in comparison with U.S. ones, a panel regression with fixed effects is used. In the first model, U.S. firms function as a control group for Dutch firms. Tobin’s Q, used as the dependent variable, allows for comparing share prices.3 The variables of interest are interaction terms between dummy variables. Gompers, Ishii and Metrick (2003) used a comparable method to measure the different effects of being a firm with a low or high governance index on the return of a firm. The dummy variables, for the interaction variables, are: poison pills, staggered boards, supermajority voting rights, golden parachutes and Dutch for firms that are listed at the Dutch stock exchange. Here, Gompers, Ishii and Metrick (2003) used dummy variables for anti-takeover provision groups and a dummy for being a firm with a high or low governance index.
Given the abovementioned information, the following model is constructed to measure a possible Dutch discount caused by anti-takeover provisions, claimed to be mainly used by preferred shareholders as an assurance for not being taken over.
1 𝑄 = 𝛽!+ 𝛽! 𝑃𝑟𝑜𝑣𝑖𝑠𝑖𝑜𝑛!" + 𝛽! 𝑃𝑟𝑜𝑣𝑖𝑠𝑖𝑜𝑛!"∗ 𝐷𝑢𝑡𝑐ℎ!" + 𝛽! 𝐷𝑢𝑡𝑐ℎ!" + 𝛽! 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 + 𝛽! 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑓𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠
+ 𝛽! 𝑌𝑒𝑎𝑟 𝑓𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝜀 !"
Where Q stands for Tobin’s Q. Provision stands for poison pill, golden parachute, classified board or supermajority requirements and for limits to shareholder amendments. These
3 Tobin’s Q is calculated by the market value of total assets divided by the book value of total assets. Here market value is calculated as total book value of assets plus year-end share price times
variables are dummy variables and are 1 if a firm uses the particular anti-takeover provision. The Dutch dummy variable is 1 for Dutch stock exchange listed companies.
Within the second model, U.S. firms will function as a control group for TARP firms. New interaction variables are constructed to compare TARP-participating firms with the rest of the data set within the years of interest. Controlling dummy variables for the year and TARP participants are added to the dummy golden parachutes. The interaction dummy, which consists of the dummies: year, tarp participants and golden parachutes is a rough indicator for the effect of the shock (TARP) on the part of the Tobin’s Q that is explained by TARP participants that use golden parachutes in year 2007 and 2008. Notice that this effect is in comparison with the whole U.S. data set or only with firms active within the financial sector. The TARP is expected to reduce the exercise–or the threat of exercise–of golden parachutes by TARP participants. This might influence the part of the Tobin’s Q that is explained by golden parachutes established by these participants. If it could be assumed that this shock is exogenous, then whether golden parachutes have a causal relationship with the Tobin’s Q could have been measured. The “exogenous shock” could influence the explanation of golden parachutes established by TARP participants on Tobin’s Q. Note again that the explanation is measured before and after the shock and in comparison with both the whole U.S. dataset and only the financial sector. However, it is very unlikely that the TARP is an exogenous shock since in this year the subprime mortgages crises started and mostly poor performing financial institutions participated the program. Therefore it could never be seen as a random selection. Although it cannot be seen as an exogenous shock, this thesis uses it as one– being aware of the omitted variable bias– and tries to capture some of the effects of the golden parachutes on Tobin’s Q.
Model 2 will be used to measure the effect of the “exogenous shock” on the relation between golden parachutes established by TARP participants and Tobin’s Q, compared to the rest of the U.S. dataset and as well only with firms active within the financial sector.
2 𝑄 = 𝛽! + 𝛽! 𝐺𝑃!" + 𝛽! 2007 ∗ 𝑇𝐴𝑅𝑃 ∗ 𝐺𝑃!"
+ 𝛽! 2008 ∗ 𝑇𝐴𝑅𝑃 ∗ 𝐺𝑃!" +𝛽! 𝑇𝐴𝑅𝑃 + 𝛽! 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠
+ 𝛽! 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑓𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝛽! 𝑌𝑒𝑎𝑟 𝑓𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝜀 !"
Where Q stands for Tobin’s Q. GP for Golden parachute, 2007 and 2008 for the years 2007 and 2008 and TARP for TARP participants. These variables are dummy variables and are all
1 when it concerns a TARP participant that established a golden parachute in the years 2007 and 2008. .
To control for firm-specific effects, various control variables like firm size, market capitalization, capital structure, capital intensity, and trading volume are added to both models. The control variables audit committee independence, stock based compensation and board size, are added since they are correlated with the use of anti-takeover provision. With the use of country, industry and year dummies country, industry and year fixed effects are respectively controlled for. The abovementioned control variables and fixed effects in model 1 and additionally the event in model 2 are used to address the omitted variable bias problem insofar that is possible within these constructed models.
Dataset
After eliminating firms without an identification variable, the dataset used for the research consists of 111 Dutch firms and 1350 U.S. firms that are active within all possible industries. The derived data concerns panel data for the period 2004-2012. All firms traded on either the U.S or the Dutch stock market. The data of U.S firms was gathered from different datasets within WRDS, namely, Compustat North America, Risk Metrics and Exucomp. The database of RiskMetrics provides data of firms listed at the S&P 1500 and their use of anti-takeover provisions. No data was available on anti-takeover provisions in the year 2005 within the U.S. database. The RiskMetrics dataset records whether a firm uses the anti-takeover provisions— poison pill and pre-emptive rights, golden parachutes, classified boards and supermajority voting rights. Although supermajority-voting rights could have been made stronger or weaker, as it could vary between 50 and 100 percent, no strength distinctions were made. Compustat North America and Compustat Exucomp are databases of U.S. and Canadian market and fundamental information on inactive and active companies that are held publicly. These different datasets where merged by the use of gvkey, cusip and ticker. The data on Dutch firms were manually derived from DataStream and manually prepared for bundling with the U.S dataset. The control variable firm size, trading volume, market cap, and total stock-based compensation were reconstructed into log variables to make them more fitted. Here firm size is defined by total sales, and trading volume is calculated by taking total traded shares within a given year for each firm. Market capitalization is constructed by multiplying total shares outstanding by year-end market price. Total stock based compensation stands for the total
payouts in stock by a firm during the year. To reduce the influence of large outliers, Tobin’s Q and board size are winsorized for 5 and 1 percent of the observations respectively.
Table 1 presents the correlation between the dependent variable Tobin’s Q, the anti-takeover provision variables and the control variables. It shows that all anti-takeover provisions are negatively correlated with Tobin’s Q. This is consistent with the findings of Bebchuck, Cohen and Ferrell (2004) who found that the limits on shareholder voting rights, staggered boards golden parachute and poison pills had a significant negative effect on Tobin’s Q. The anti-takeover provisions are correlated with each other. To deal with multi-collinearity each provision is run separately. It also becomes clear that sales and market capitalization are highly correlated with Tobin’s Q, hence they can be seen as worthy control variables. The governance variables are strongly correlated with the anti-takeover provision variables. This indicates that these variables are good for controlling any omitted variable bias originating from other governance influences on share price.
Table 1
Pairwise Correlation Matrix
This table presents the pairwise correlations—with its standard errors in parentheses—between the variables presented on the left side and the topside of the table. The first column gives the correlations of each of these variables with Tobin’s Q. The remaining columns give the correlations between anti-takeover provisions mutually and the used control variables.
Graph 1 displays the usage of anti-takeover provisions over the years within the whole dataset. This graph shows the percentage of firms that used the mentioned provision in a given year. 2004 and 2005 are excluded from the graph, as no data is available for the period. It is remarkable how the usage of golden parachutes changes in the year 2008. In this year the TARP came into law, restricting the payments of existing golden parachute agreements and the establishment of new ones. Therefore this implementation could be an explanation for a part of the decline in usage of golden parachutes.
Firm%Charactaristics Tobin's%Q Poison%pills%&Pre7 emptive%rights Golden% Parachutes Classiefied% Boards Supermajority% Voting%Rights 70.043 1.000 (0.000) 70.080 0.267 1.000 (0.000) (0.000) 70.062 0.484 0.4657 1.000 (0.000) (0.000) (0.000) 70.081 0.409 0.555 0.537 1.000 (0.000) (0.000) (0.000) (0.000) 70.390 0.191 0.241 0.211 0.287 (0.000) (0.000) (0.000) (0.000) (0.000) 70.0093 0.1641 0.2723 0.241 0.306 0.0329 (0.000) (0.000) (0.000) (0.000) 70.183 0.225 0.263 0.231 0.312 (0.000) (0.000) (0.000) (0.000) (0.000) 0.065 0.092 0.083 70.020 0.125 (0.000) (0.000) (0.000) (0.020) (0.000) 70.010 70.002 70.003 70.003 70.003 (0.028) (0.580) (0.428) (0.399) (0.408) 70.173 70.078 0.047 70.024 0.119 (0.000) (0.000) (0.000) 0.030 (0.000) 0.008 0.109 70.180 70.163 70.114 (0.528) (0.000) (0.000) (0.000) (0.000) Audit%Committee% Independence Poison%pills%&%Pre7 emptive%Rights Golden%Parachutes Classified%Boards Supermajority%Voting% Rights Firm%Size Trading%Volume Market%Capatilization %Stock%Based% Compensation Debt%of%Capital Board%Size
Graph 1
Provisions Usage
The data used for Graph 2 is derived from the database DataStream only since information about ownership concentration is difficult to require from WRDS. The diagram shows that Dutch firms tend to have higher ownership concentration. On average, the biggest voter and owner of listed Dutch firms own a considerably bigger stake of the firm than the biggest voter and owner of U.S. firms listed at the S&P 1500. This statistics are consistent with the finding and claim by Kabir, Cantrijn and Jeunink (1996) that Dutch firms have higher ownership concentrations than U.S. firms.
0 .05 .1 .15 Pe rce n ta g e 2006 2007 2008 2009 2010 2011 2012 Year Pct. usage GP Pct. usage CB
Pct. usage SMVR Pct. usage PP & PrER
This graph shows what percentage of firms within the entire dataset use the anti-takeover provisions in question. With on the y-axis the percentage and on the x-axis the year. GP stands for golden parachutes; SMVR stands for supermajority voting requirements; CB stands for classified boards and PP & PrER stands for poison pills and pre-emptive rights. The vertical line drawn in the year 2008 indicates the year in which the TARP came in to law.
Graph 2
Ownership Concentration
Graph 3 presents the course of Tobin’s Q on average of the different groups of interest. The groups are U.S. Firms, Dutch Firms, TARP-participating firms and firms active in the financial sector in the U.S. The graph shows that, overall, there is a considerable difference between the Tobin’s Q of the different groups. Firms active in the financial market of the U.S. have a lower Tobin’s Q than the average of all firms active in the U.S.; Dutch Firms on average have a lower Tobin’s Q than all U.S. firms taken together while TARP firms on average have the lowest Tobin’s Q of all groups. The course of all lines in the year 2008 is interesting, as in this year the subprime mortgage crises got the worst of numerous companies. The Tobin’s Q of all groups except the TARP participants was negatively affected. This is consistent with hypothesis 2 that the restriction of golden parachutes would reduce the negative effect of golden parachutes on Tobin’s Q.
0 5 10 15 20 25 30 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 P er ce n ta ge Year
Dutch Biggest owner U.S. Biggest owner Dutch Biggest voter U.S Biggest voter
This graph presents the ownership percentages of the largest owner and voter of firms listed at the S&P 1500 and of firms listed within the Netherlands. With on the y-axis the average percentage of shares or voting share the biggest owner owns of a firm. On the y-axis the year of observation is presented.
Graph 3
Tobin’s Q
Results
This section discusses the model outcomes and examines the evidence for both hypotheses mentioned in the methodology. Table 2 shows the results of annual regressions for the performance measure Tobin’s Q on the anti-takeover provisions used by Dutch firms. These regressions show whether there is an existing “Dutch Discount” for Dutch traded stocks compared to U.S. firms. Model 1 is run separately for every provision type to prevent any chance that they affect each other, since in the data description it is shown that the provisions are inter-correlated. Table 3 shows the result of annual regressions for the performance measure Tobin’s Q on the golden parachutes used by TARP participants compared to the total U.S. dataset and the financial sector only. These regressions will show whether there is a relation between the use of golden parachutes and Tobin’s Q. Model 2 is run for both U.S.
.5 1 1.5 2 2.5 T o b in 's Q 2004 2005 2006 2007 2008 2009 2010 2011 2012 Year
Avg. Q U.S. Firms Avg. Q TARP Participants Avg. Q Financial Sector Avg. Q Dutch Firms
This graph shows the course of Tobin’s Q for the period from 2004 and 2012 for the three different groups mentioned in the legend. With on the y-axis Tobin’s Q and on the x-axis the year. The presented Tobin’s Q’s are the averages of al the firms within the mentioned different groups. The vertical line drawn in the year 2008 indicates the year in which the TARP came in to law.
firms and financial sector ones. The financial sector is the more interesting group to compare with, since TARP participants are mainly financial institutions.
Evidence on Hypothesis 1
H1: The anti-takeover provisions poison pills, golden parachutes, classified boards and supermajority voting requirements established by Dutch firms have a more negative effect on Tobin’s Q than those established by U.S. Firms.
Regressions 1 and 2 show that poison pills and pre-emptive rights established by Dutch firms have a positive effect on Tobin’s Q, where U.S. poison pills and pre-emptive rights exert a negative effect on Tobin’s Q. Poison pills and pre-emptive rights established by U.S. firms exert a negative effect of 0.006 on Tobin’s Q, but it turns out to be insignificant. The variable ‘Dutch poison pills and pre-emptive rights’ show that, compared to supermajority voting rights implemented by U.S firms, supermajority-voting rights by Dutch firms exert a more positive effect—of 0,128—on Tobin’s Q. However, this turns out to be insignificant as well. Therefore, it is impossible to say whether poison pills and pre-emptive rights established by Dutch firms have a more positive effect on Tobin’s Q than those implemented by U.S. firms.
Regressions 3 and 4 show that golden parachutes established by Dutch firms have a more positive effect on Tobin’s Q while U.S. golden parachutes exert a negative effect on Tobin’s Q. Golden parachutes established by U.S. firms exert a significant negative effect— of 0.203—for a significance level of 0.01 percent on Tobin’s Q. The variable ‘Dutch golden parachute’ shows that golden parachutes by Dutch firms exert a more positive effect—of 0,476—on Tobin’s Q. This has a significance level of 0.01 percent and the effect is more than three times the standard deviation. Therefore, golden parachutes implemented by Dutch firms seem to have a significant more positive effect on Tobin’s Q than those implemented by U.S. firms.
Regressions 5 and 6 show that classified boards at Dutch firms have a higher negative effect on Tobin’s Q than those established by U.S. firms. The variable ‘classified boards’ shows that classified boards established by U.S. firms exert a positive effect—of 0.0337—,but this result is insignificant. However, the same variable in the Dutch case shows that classified boards established by Dutch firms exert a more negative effect—of 0.561—on Tobin’s Q. This is significant at a level of 1 percent and the effect is more than three times the standard
deviation. This indicates that classified boards established by Dutch firms seem to have a more negative effect on Tobin’s Q than those established by U.S. firms.
Regressions 7 and 8 show that supermajority voting requirements established by Dutch firms have a positive effect on Tobin’s Q while U.S. supermajority voting requirements exert a negative effect on Tobin’s Q. Supermajority voting requirements implemented by U.S. firms exert a significant negative effect of 0.0823 for a significance level of 1 percent on Tobin’s Q. The variable ‘Dutch supermajority requirements’ shows that supermajority-voting rights by Dutch firms exert a more positive effect—of 0.260—on Tobin’s Q. This is for a significance level of 1 percent and the effect is more than two times the standard deviation. Therefore, supermajority-voting rights implemented by Dutch firms have a more positive effect on Tobin’s Q than those implemented by U.S. firms.
Regression 9 shows that the results do not differ drastically when all variables of interest are all run together. Thus, 41.5 percent of the model is explained by the variables used, as the R-squared is 0.415. Regressions 10, 11 and 12 demonstrate the value of fixed effects when estimating the model.
The variable Dutch Firm shows that Dutch firms are traded against a discount and this result is highly significant. This finding is consistent with graph 3 discussed in the section on datasets. Despite this high discount whether it is mainly caused by different reason that Dutch firms use anti-takeover provisions remains a question, because only classified boards exert a more significant negative effect on Tobin’s Q of Dutch firms while golden parachutes and supermajority voting rights exert a more positive return on the Dutch Tobin’s Q. Therefore, hypothesis 2 does not hold for all provisions of interest, but there is a strong suspicion that classified boards established by Dutch firms have a more negative effect on shareholder value than those established by U.S. firms.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
VARIABLES PP DPP GP DGP CB DCB SMVR DSMVR All;FE ALL;NO;FE NO;YR;FE NO;IN;FE Ppill;&;PreEemptive;Rights E0.006 E0.006 E0.002 0.033 0.004 0.024
(0.024) (0.025) (0.025) (0.026) (0.024) (0.027) Dutch;Poison;Pill;&;PreEemptive;Right 0.128 0.052 0.004 0.057 0.001 (0.081) (0.083) (0.084) (0.083) (0.081) Golden;parachute E0.185*** E0.203*** E0.202*** E0.190*** E0.199*** E0.198***
(0.027) (0.030) (0.028) (0.027) (0.025) (0.030) Dutch;Golden;Parachute 0.476*** 0.430*** 0.478*** 0.425*** 0.485***
(0.102) (0.103) (0.098) (0.103) (0.093) Classified;Board 0.037* 0.034 0.049** 0.050** 0.0511** 0.0465**
(0.021) (0.021) (0.0214) (0.024) (0.021) (0.024) Dutch;Classified;Board E0.561*** E0.463*** E0.195** E0.412*** E0.253***
(0.100) (0.096) (0.098) (0.105) (0.090) Supermajority;Voting;;Requirements E0.065** E0.082*** E0.0608** E0.0290 E0.0416* E0.0582*
(0.027) (0.028) (0.028) (0.027) (0.025) (0.031) Dutch;Supermajority;Voting;Requirements; 0.260*** 0.196** 0.0684 0.182* 0.065 (0.096) (0.094) (0.091) (0.094) (0.089) Dutch;Firm E1.792*** E1.913*** E1.653*** E1.852*** E1.981*** E2.720*** E1.972*** E2.738***
(0.357) (0.369) (0.359) (0.356) (0.358) (0.325) (0.355) (0.329) Firm;Size E0.544*** E0.535*** E0.550*** E0.539*** E0.544*** E0.534*** E0.541*** E0.530*** E0.536*** E0.404*** E0.541*** E0.398***
(0.020) (0.020) (0.020) (0.020) (0.020) (0.020) (0.020) (0.020) (0.020) (0.019) (0.020) (0.019) Trading;Volume E0.036*** E0.060*** E0.031*** E0.056*** E0.034*** E0.057*** E0.034*** E0.057*** E0.052*** E0.107*** E0.054*** E0.105***
(0.011) (0.012) (0.011) (0.012) (0.011) (0.012) (0.011) (0.012) (0.012) (0.012) (0.012) (0.013) Market;Capatilization 0.588*** 0.597*** 0.578*** 0.586*** 0.589*** 0.597*** 0.590*** 0.600*** 0.589*** 0.503*** 0.597*** 0.494***
(0.012) (0.019) (0.019) (0.019) (0.019) (0.019) (0.019) (0.019) (0.019) (0.017) (0.018) (0.018) Debt;of;Capital;in;% E0.000 E0.000 E0.000 E0.000 E0.000 E0.000 E0.000 E0.000 E0.000 E0.001* E0.000 E0.001* (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) (0.000) (0.000) Total;Stock;based;compensation 0.138*** 0.148*** 0.153*** 0.164*** 0.138*** 0.147*** 0.137*** 0.144*** 0.161*** 0.211*** 0.157*** 0.218***
(0.020) (0.020) (0.020) (0.021) (0.020) (0.020) (0.020) (0.020) (0.021) (0.023) (0.021) (0.023) Board;Size E0.050*** E0.051*** E0.047*** E0.047*** E0.051*** E0.051*** E0.051*** E0.052*** E0.048*** E0.097*** E0.048*** E0.097***
(0.005) (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) Audit;committee;independence E0.019*** E0.004 E0.018*** E0.004 E0.018*** E0.004 E0.019*** E0.004 E0.004 E0.006** E0.004 E0.006**
(0.002) (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) (0.003) (0.003) (0.003) (0.003) (0.003) Constant 1.014*** 1.201*** 1.000*** 1.205*** 0.952*** 1.140*** 0.977*** 1.199*** 1.131*** 1.919*** 1.255*** 1.870***
(0.235) (0.247) (0.239) (0.240) (0.236) (0.244) (0.238) (0.240) (0.242) (0.215) (0.201) (0.218) Observations 5919 5919 5919 5919 5919 5919 5919 5919 5919 5919 5919 5919 REsquared 0.405 0.407 0.410 0.414 0.405 0.407 0.405 0.408 0.415 0.285 0.414 0.288 Robust YES YES YES YES YES YES YES YES YES YES YES YES Country;Fe NO YES NO YES NO YES NO YES YES YES YES YES Year;Fe YES YES YES YES YES YES YES YES YES NO NO YES Indusrty;Fe YES YES YES YES YES YES YES YES YES NO YES NO Robust;standard;errors;in;parentheses
***;p<0.01,;**;p<0.05,;*;p<0.1
This table presents the estimations of model (1) where Tobin’s Q is the explainable variable and the anti-takeover provisions by Dutch firms are the variables of interest. The first row contains the results when only the variable ‘poison pill and pre-emptive rights’ (PP) is included— together with the control variables and year and industry fixed effect. The second row contains the results when the variable ‘Dutch poison pill and pre-emptive rights’ (DPP) is added. The same was done within the third till the eighth row. Here GP stand for golden parachute; DGP for Dutch golden parachute; CB for classified board; DCB for Dutch classified board; SMVR for supermajority voting rights; Dutch supermajority voting rights. Row 9 contains the results when all variables of model (1) are run together. The remaining rows show the results when year and/or industry fixed effect are excluded from the model. Standard errors are reported in parentheses.
Evidence on Hypothesis 2
H2: The TARP reduces the exercise–or the threat of exercise–of golden parachutes by TARP participants in the U.S. and therefore reduces the negative effect on Tobin’s Q by golden parachutes established by these TARP participants.
Regression 3 shows that the effect of golden parachutes, established by TARP participants changes in the year that they participate the program. This is compared to firms active within the financial sector as these are functioning as the control group. Compared to the other financial institutions, golden parachutes established by TARP participants have a more negative effect of 0.108 on Tobin’s Q. However, this effect turns out to be insignificant. In the year that the firms participate in TARP the effect of the golden parachutes changes from a more negative effect to a more positive effect of 0.183 and this is significant at a level of 5 percent. This means that, compared to non-TARP participating financial institutions, golden parachutes established by TARP-participating companies have a less negative effect on Tobin’s Q.
Regression 5 also shows that the effect of golden parachutes established by TARP participants, changes in the year they participate in the program, compared to all other U.S. firms. The variable GP in year before participating TARP indicates that before the company participated in TARP their established golden parachutes had a more negative effect of 0.192 on Tobin’s Q. This finding is significant at a level of 1 percent and this effect is exactly three times the standard deviation. The variable GP in the year of participating TARP indicates that the established golden parachutes had a more positive effect of 0.295 on Tobin’s Q than those established by U.S. firms overall. This finding is insignificant, but the effect is one and a half times the standard deviation.
Comparing the regressions discussed with the remaining regressions displays the importance of year and industry fixed effects. In regression 3 the industry fixed effects are useless since the dataset only consists of financial institutions. The group and year fixed effects within regression 3 do not add much to the explained value as using the non-participating financial institutions as the control group already filters most of the effects. For regression 5 the added year and industry fixed effects give a better fit to the regression, as the variables of interest become more significant. Thereby, when adding the fixed effects, the explained part increases from 28.9 percent to 41.7 percent.
The difference between the two variables GP in year before participating in TARP and GP in year of participating in TARP could be explained by the golden parachute restriction.
These findings do not contradict hypothesis 2. It is obvious that the treatment group was not randomly chosen and that other characteristics could cause omitted variable bias, but the change in the effect by golden parachutes on Tobin’s Q for TARP participants in the year 2007 to 2008 is so clear that a relation between golden parachutes and Tobin’s Q is very likely to exist. Tobin’s Q serves as the measure for shareholder value so given the results of model 2 it could be said– with caution– that an amendment in the use of golden parachutes will affect shareholder value. Note that this claim can only hold for U.S. firms since the TARP participants and the control group both consist of U.S. firms exclusively.
