Corporate Governance and the Business Cycle

Transcription

1 BSc Business Administration Author: Rune Hjorth Nielsen (286850) Author: (300134) Advisor: Baran Siyahhan Corporate governance and the business cycle Department of Business Studies Aarhus School of Business University of Aarhus May 2011

2 Abstract Corporate governance is important in today s business world. It can be viewed as a holistic model for how a corporation should deal with different aspects of itsenvironment. Further it has drawn attention due to the turbulent macroeconomic environment of the recent years. The aim of this thesis is to investigate the relationship between anti-takeover provisions and shareholder value. In order to create a theoretical understanding of how anti-takeover provisions fit into the world of corporate governance a literature review of agency theory is carried out. In this literature review subjects such as agent-principal relationship, separation of ownership and control, agency cost and asymmetric information are discussed. Further an introduction of the G-index and E-index is made, and discussions of the two indexes as well as a discussion of the provisions in the E-index are made in order to decide which of the two indexes to use in our analysis. In order to test the problem formulation empirical hypotheses are stated. These are tested with data obtained from the Wharton Research Data Service. The focus of the analysisis on non-dual stock class companies in the U.S. To investigate the problem formulation we have chosen to make use of the analytical approach. This is done due to the consideration of the large amount of empiric and numeric data, which is handled in the thesis. We state two hypotheses, and to empirically test these hypotheses we utilize the Fama-MacBeth method. In our analysis for the first hypothesis we do not find any support for the relationship between stock return and the anti-takeover provisions in the period 1998 to This is supported by very strong p-values. This is also the case for the second hypothesis and we cannot find any statistical significant indication for regulation being a substitute for limiting the amount of antitakeover provisions implemented. The second conclusion is somewhat uncertain due to biased estimators.

3 Contents 1. Introduction Methodology Problem statement Hypotheses formulation Hypothesis I Hypothesis II Corporategovernance Agency theory Anti-takeover provisions and the relating hypotheses Discussion of the indexes Supermajority, limitation on amending bylaws and charters Golden parachutes Staggered boards Poison pill The choice of Index The data sources The E-index The data describing firm performance Monthly movements of the stock The processing of the data set The variables Choice of statistical method Fama-MacBeth compared to other panel data methods The efficiency of the OLS estimators The efficiency of the standard errors Evidence of Hypothesis I Utilizing E-index as an Independent Variable Utilizing the democracy dummy Evidence of hypothesis II Results without industry adjustment Results with industry adjustments... 63

4 13. Discussion of the results Hypothesis I Hypothesis II Conclusion References... 69

5 1. Introduction Within the recent years one of the main developments in the research of agency problems and its relation to stockholder return arises from the development of the G-index. The index was created and introduced by Gompers, Ishii and Metric, who published it in 2003 in the article Corporate Governance and Equity Prices. This article, other than introducing the index, provides the first conclusions made based on the index. The main conclusion here is the finding of a relatively strong relationship between abnormal stock returns and certain ownership control provisions. In 2009 Bebchuk, Cohen and Ferrell published What Matters in Corporate Governance with a revision of the 24 provisions in the G-index and their relevance for shareholders value. They found that only 6 of the provisions from the G-indexwas relevant for shareholder value. These were used to construct the E-index. The articles Corporate Governance and Equity Prices and What matters in corporate governance build on corporate governance data from 1990, 1993, 1995, 1998, 2000 and 2002 and estimate the relationship in the period and , respectively. The period used by Gompers et al. (2003) represent a time with financial stability, high returns and economic progress in the society as a whole.this can pose a challenge in adopting the conclusions reached by the authors to the present time period and business environment. The time period used by Bebchuk et al.(2009) also includes the time period from 2000 to 2003, which we argue are more volatile than most of the 1990 s, however, we focus only on the volatile period and not on the stable period of the early to mid 1990 s. We choose only to focus on the period from 1998 to The reason for this is that we try to mimic the situation managers and shareholders face today, which is very different from the situation in the 1990 s where the talk about the so-called New Economy was at its highest. This period of overly optimistic projections of the future, might have altered the way some shareholders controlled the management of a company, away from what would be optimal in another part of the business cycle as the one we are experiencing today. The perception of the business environment today is that it is much more uncertain than in the 1990 s even though talks about double dip and extremely long recessions have decreased lately. Our suspicion is backed up by Cremers and Nair (2005) who extended the period of the G-index to Page 1 of 71

6 2001(capturing some financial instability) and found it caused a 1 percentage point drop in the abnormal returns between the democracy and dictatorship portfolios, but more importantly rendered the relationship insignificant. Therefore away to estimate a more current relationship between stockholder return and antitakeover provisions implemented might be to focus on a time period with a high degree of instability in the financial markets. We will attempt to do that with a dataset focused on the yearsfrom 1998 to This time period will therefore include the dot com bubble burst, 9/11 and the Asian crisis. It willthereby better mimic the way the perception of the economic future is today. We will further attempt an investigation of another much discussed business idea at the moment; regulation. Some researchers have found a relationship between the capital structure of a company and the level of regulation within that industry [Jiraporn and Gleason (2007)]. We would like to investigate if we can find any evidence that shareholders of companies within a heavily regulated financial industry have a lesser need for control over the management compared to shareholders of companies not within a heavily regulated industry. This should hopefully give us an indication if the regulation can be seen as a substitute for limiting the number of antitakeover provisions implemented. 2. Methodology In this chapter the basic methodological approach will be presented. Each methodological view has it concepts, advantages and disadvantages. For our thesis we have chosen to make use of the techniques and methodical procedures offered by the analytical approach. Our reasoning for this choice is that our thesis has a foundation on empirical data. Therefore the analytical approach seems as the right choice and will allow us to handle the vast amount of data [Arbnor and Bjerke (2009)]. The thesis will follow the suggested analytical steps offered by Arbnor and Bjerke (2009). It is as following; formulating the hypotheses, planning the study, designing methods for collecting data, coding and arranging data and finally testing the hypotheses. Using this approach will assure each step in the thesis is a development of the previous step and provide a solid structure throughout the thesis. Page 2 of 71

7 We will start with formulating our problem and drafting our hypotheses. After this a segment will contain a literature review of the papers which relates to our hypotheses. It will ensure a theoretical understanding of the problems in question. In planning our study we have chosen to follow the methodology proposed in the articles concerning the Governance index and the Entrenchment index.specifically we intend to mimic the analysis on page 141 in the section Evidence for Hypothesis III in Gompers et al. (2003). The value of this will be showed throughout the thesis. At last we will use the collected data to test our hypotheses and test for any suggested relationships which will be proposed in the form of the hypotheses. 3. Problem statement We want to investigate the relationship between the anti-takeover provisions described by the E- index and stockholder return. We further investigate if regulation can be a substitute for antitakeover provisions for heavily regulated financial industries. Both investigations are made in the time period from 1998 to To research this problem statement we will begin with the hypothesis formulation. It should be noted that the thesis is delimited to non-dual stock class U.S companies. 4. Hypotheses formulation As mentioned in the introduction the focus of this thesis will be on two different hypotheses. These will be explained below. 4.1 Hypothesis I The time period is known for its unstable economy. This differs from the early to the mid 1990 s known for its stable economy. The two articles containing the G-index and E-index are based on the year range and , respectively. Therefore one can argue the indexes are based on a dataset influenced by stable fluctuations and a healthy economy. We do realize that the unstable times from 1998 to 2003 is included in the analysis by Bebchuk et Page 3 of 71

8 al.(2009), and that they still find that they still find a relationship between stock returns and some antitakeover provisions (the ones included in the E-index). However, we have a suspicion that focusing only on a longer period of an unstable economy might change the conclusion reached by Gompers et al. and Bebchuk et al. This suspicion is raised through Cremers and Nair (2005) who extended the period of the years from 1999 to 2001(capturing some financial instability) in the G-index and found it caused a 1 percentage point drop in the abnormal returns between the democracy and dictatorship portfolios and, more importantly, made the results insignificant. Therefore our first hypothesis is that different perceptions of the current and future macroeconomic environment might affect the effect of the antitakeover provisions as the ones in the E-index. A quite likely way this could occur would be from the following way of thought. Shareholders might get very satisfied with the current management in a business friendly macroeconomic environment and therefore allow the passing of provisions which limits the shareholders power and control of the management. The same argument of course works the other way in a less beneficial economic period shareholders might seek to gain more shareholder power over the management. This is supported in the literature by Schoar and Washington (2010). They investigated whether governance changes are predicted by either bad or good firm performances. They findspecial meetings, where changes in the governance provisions are proposed, which are not beneficial to shareholders, are more often called in the current or following quarter when positive earnings surprises occurs. Furthermore the provisions have a passage rate of 96 percent at special meetings. In short their findings suggest the following: managers uses times of good firm performance to pass ballot items that hurt shareholder interest Thus the seeds of bad governance seem to be sown in good times (Schoar and Washington, 2010, p. 16). One would also expect the same relationship the other way around, so periods of economic down turn would lead to restricting the managerial power. This seems also to be the general tendency in the society today where the general public in recent years has raised critical questions about for example golden parachutes for the CEO s. In response to this the Securities and Exchange Commission(SEC) has begun to introduce rules which put obligations on public firms to hold adversary votes on compensations schemes and other provisions which limits the power of the shareholders. Page 4 of 71

9 02JAN AUG MAR NOV JUN FEB SEP APR DEC JUL MAR OCT MAY JAN AUG APR NOV JUL FEB OCT MAY DEC AUG MAR NOV JUN FEB SEP MAY DEC JUL MAR2009 Corporate Governance and the Business Cycle The hypothesis will be tested with the same setup as used in the Gompers et al. (2003) under hypothesis III, with the main difference being that we will use a different sample from a different time period. The sample we will use for our analysis is based on more financial unstable times only, compared to Bebchuk et al. (2009) and Gompers et al. (2003), which better mimics the financial situation we are facing today. Please note that this hypothesis is not intended to show the causal relationship between the economic environment (including the perception of this) and the agency theory as described by thee-index. The aim is simply to evaluate the relationship. Before discussing the model we want to investigate we show that there is increased volatility in the markets in the time period 1998 to This is seen in the discussion below. Exhibit The VIX, for the time period from 1 st of January 1990 to 1 st of October VIX index Source: Wharton Research Data Services (WRDS) Exhibit above shows the VIX. It measures the near-term volatility conveyed from 30-day options on the S&P 500 stock index. This gives us a measure of how investors perceive the volatility off the markets 30 days into the future. As these investors are the shareholders this gives us some great information about how one side of the principal-agent relationshipview the future. The key feature we use from the VIX is that the index spikes during periods of market turmoil. This Page 5 of 71

10 happens becausewhen market turmoil occurs people start hedging more to secure their portfolios. If your portfolio contains mainly stocks, a way to hedge your position would be to buy put options on the S&P 500 index. Since the S&P 500 option market is dominated by buyers who want to hedge by going short on put options, when the financial markets become perceived as unstable, this clearly shows up on the VIX. The way it shows up is by a spike. This rises from the fact that when demand for buying put options increases, prices must go up. This means, ceteris paribus, that volatility of the options must go up. This feature of the VIX has caused this index to be nicknamed the Investor Fear Gauge [Whaley (2009)]. From the VIX it is possible to see the market turmoil that has existed in the past years, so we can evaluate which time period we want to use in our analysis. As it can be seen from Exhibit there is a big difference between the periods we want to analyze in our regression than what Gompers et al. (2003) used in their analysis. Gompers et al. (2003) used the period from 1990 to When looking at this period it can be seen that the VIX index is stable at a low level and with only a limited number of spikes, at least until the year This corresponds to an investor expectation of a relatively stable S&P 500 index (seen from the low and stale value of the VIX) and with no significant market turmoil (seen from the lack of sudden spikes in the VIX). The exception from this is the beginning and end of Gompers et al. (2003) s sample period, where a few spikes occur. The spikes in the beginning of the period ( ) correspond to the Iraqi invasion into Kuwait and the following US invasion of Iraq [Whaley (2009)]. It should be remembered that the VIX shows the expected volatility within the next 30 days. Therefore we would prefer to see a rather long period of spikes showing up in the index as a long period with spikes would show a long period of market turmoil. A rather long period of market turmoil would be necessary to see changes in the corporate governance provisions for the company, since it can be a long process to alterthese. An example of this could be board structures and the bylaws of the boards. It is difficult for us to ascertain if the period of instability corresponding to the market turmoil rising from the first gulf war, was long enough to potentially change the corporate governance provisionsof a company. As mentioned above changes in the governance provisions can be very long to implement so the effects of turmoil might be delayed. We therefore need to take a look at the turmoil created before 1990 to make sure no changes were made in response to the turmoil before 1990 to fully show how stable the period from 1990 to 1998 was. Since our data source, WRDS, does not contain numbers for the VIX index before 1990, we need to find other data to see this. WRDS has a similar index Page 6 of 71

11 04JAN FEB MAR MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG AUG SEP OCT NOV DEC1989 Corporate Governance and the Business Cycle that isalmost exactly correlated to the VIX index, the difference only being that this is based on the S&P 100 instead of the S&P 500. This index is called the VXO. The VXO and the VIX are as close to being perfect substitutes as possible with a very high correlation (0,9898) on the daily return [Whaley (2009)]. Exhibit The VXO for the time period from the 1 st of January 1988 to the 31 st of December CBOE S&P100 Volatility Index Source: Wharton Research Data Services As it can be seen from Exhibit there is not a very high number of spikes in the period on the VXO. The only spikes is in the beginning of 1988 (probably related to the aftermath of the October 1987 stock market crash) and the spikes rising from the mini crash in relation to the news arising of the failed attempt for a leveraged buyout of UAL [Whaley (2009)]. These periods of spikes seem very short and it seems unlikely that these spikes have led to investors being so insecure that they felt the need to start the process of changing the corporate governance provisions in the companies they had investments in. As mentioned earlier there are a few spikes in These correspond to the Asian crisis, and the mini crisis it created arising from a big sell-off of stocks. The market turmoil arising from this is probably not within the data range used by Gompers et al. (2003) as it happened in October Page 7 of 71

12 As mentioned earlier the process of shifting the corporate governance provisions is rather long, so this is not within the 1998 data. The spikes in 1998 and 1999 are not accounted for in the analysis by Gompers et al., at least in their relation to the governance variable of their research. The reason for this is that the G-index is assumed constant from 1998 to The alteration of the G-index (which describes the corporate governance provisions in Gompers et al. (2003)) values can therefore not have been included as a response to the increased amount of market turmoil in Gompers et al. (2003), as they simply do not change their G-index data after the unstable period begins in late From the above discussion we can conclude that the data period used in Gompers et al. (2003) is a very stable period with only the period around the Iraqi invasion of Kuwait and the following war on Iraq creating some uncertainty that the time period in fact stable. This concern must be seen as minor though. If we instead look at the period from primo 1998 to ultimo 2003, in Exhibit 4.1.1, it can immediately be seen that this is a period with a much higher level of market uncertainty. The general level of the VIX index is higher, but more importantly a much larger number of spikes occurs between 1998 and 2003 than between 1990 and 1998 (remember the discussion above that for the purposes of this article the spikes from late 1997 and onwards relates to the time period of 1998 to 2003 only). This corresponds to a period with more investor insecurity in general (higher level of the VIX), as well as more periods with market turmoil (seen from the very frequent spikes). It should be noted that until primo 2003 the spikes are very frequent. This period contains the Asian Crisis and its aftermath (October 1997 mini crash and the Russian financial crisis), the dot com bubble and the terrorist attack on September 11 th. We choose to include all of 2003 in our sample as we this way hope to catch any process changing the governance provisions of the company initiated during the end of the period with high level of instability in the financial markets (ending primo 2003), given the slow speed of implementation of the changes of the governance provisions. As discussed earlier the analysis conducted by Bebchuk et al. encompasses the whole time period from 1990 to 2003, meaning that they account for some increased market volatility during parts of their sample period. However we intend to study a period only consisting of volatile financial markets, therefore we use the time period from1998 to 2003 as our sample. Page 8 of 71

13 So the sample we will use will be different from what was used by Gompers et al. (2003) and Bebchuk et al.(2009). As mentioned earlier we will use data on the governance provisions from the years of 1998, 2000, and 2002, as this will cover our sample period from 1998 to 2003 due to assumptions put forward later in this thesis. This time periodis chosen for the reasons discussed above. Although the sample will be changed in our analysis, our data generation method will be very similar to the one used by Gompers et al. (2003) in their hypothesis III on page 141. That is we will estimate a similar regression equation. This will give us a regression with this setup: Stockholde r _ return financial 1 financial... x financial x xe _ index Not that we choose to use the E-index instead of the G-index. The reasons for this are discussed in section Discussion of the Indexes. As it can be seen, the regression we want to estimate consists of 3 kinds of variables. The dependent variable will be the stockholder return, while the independent variables will consist of a group of financial variables (describing such things as company size and yield ratios) and the E-index. These different kinds of variables will be further discussed in the data section of this thesis. The hypothesis formulated her, is that the difference in economic climate might change the relationship between stockholder return and the governance index. We therefore want to test if we can find the same relationship as Bebchuk et al. (2009) did. The specific statistical hypotheses we want to test after when the regression is estimated are the following: H H o A : : x1 x1 0 0 This hypothesis test will show if the E-index can explain some of the variance in the stock prices. If we can reject H 0 we will find a similar relationship to the ones found by Bebchuk et al. (2009). Page 9 of 71

14 4.2 Hypothesis II The second hypothesis we wish to investigate is an extension of what is analyzed in hypothesis I. We want to extend the analysis from hypothesis I in the aspect of how regulation would affect the relationship between stock returns and corporate governance provisions for heavily regulated financial industries. This is inspired from the current macroeconomic environment and the general trend of more regulation of financial companies. This trend of more regulation of financial companies has been triggered by the financial crisis, as these companies behavior has been largely credited for causing the financial crisis. The increase in regulation is for example seen by the Basel III agreement, which stands as a global regulatory package on bank adequacy and liquidity. It sets up new standards, requirements and buffers for bank liquidity and leverage thereby trying to avoid a future crisis like the recent credit crunch. The extension from hypothesis I is that hypothesis II is going to relate heavily to one of the conclusions reached in the article Capital structure, shareholder rights, and corporate governance published by Jiraporn and Gleason (2007). The specific conclusion from this article is the fact that they have found strong evidence of a relation between the regulatory level of industry of an company and the leverage ratio of a company within that regulated industry. Their specific conclusion is that the debt level of the company is correlated to the amount of regulation in the way that regulation could be seen as a substitute for shareholder rights, here seen as governance provisions. By adding this to our analysis we hope to investigate if we can find a similarsubstitutional relationship, as Jiraporn and Gleason (2007), between increased regulation of financial industries and importance of the governance provisions, just with the different focus that arises from using investor return as the dependent variable and using financial companies instead of utility companies. This different approach would hopefully give a better knowledge for shareholders to evaluate the effect of regulation on a financial company in which they have some ownership through their shares. If we find a similar conclusion through our setup we would make a more direct relationship between regulation of financial companies and stock return as the main, if not sole, objective for shareholders is to maximize this. This is specifically the case in this framework as we generally focus on agency theory issues and its impact for shareholders. This subject has also been investigated by other authors than Jiraporn and Gleason. One example is the work by Stacey Kole and Kenneth Lehn (1997). Their study are conducted in the mid- to late Page 10 of 71

15 1990 s in a period were deregulation was more relevant, therefore their perspective is from the deregulation. These studies show the issues that could rise from deregulation. These issues all relates to the idea that when industries become less regulated the performance of the managers becomes more important as the managers become less limited in their actions. Kole and Lehn (1997) argues that the more deregulation there is infused into a market, the more difficult it gets to determine the individual effects of the decisions the management takes. This is because the deregulation causes, from Kole and Lehn (1997) s argumentation, instability. The way this is going to be tested is by estimating the same equation as in hypothesis I. The estimation is going to be run on two different datasets though. How these datasets are defined and the specific regressions we want to estimate, as well as the relating hypotheses we want to test, is found in the Evidence for Hypothesis II section of this thesis. 5. Corporategovernance Corporate governance is a topic relevant to every firm and organization in the economic world of today, may it be a privately or publicly based firm. The term corporate governance cannot be boiled down to one single factor. This is because it concerns such wide areas as the firms relationship to shareholders and stakeholders, the goals of the organization, the relationship between anti-takeover provisions and the firm value and the relationship to the community surrounding the organization. It can be argued that corporate governance is a holistic model containing different disciplines on how an organization can or should behave. For a preliminary definition of corporate governance we turn toward the OECD. In 2004 the OECD published the Principle of Corporate Governance. The publication offers 12 key standards for sound financial system and good corporate governance. In the OECD publication corporate governance is defined as following: Corporate governance is one key element in improving economic efficiency and growth as well as enhancing investor confidence. Corporate governance involves a set of relationships between a company s management, its board, its shareholders and other stakeholders. (OECD, 204, p. 11). The OECD adds further to the definition by stating; Page 11 of 71

16 As companies play a pivotal role in our economies and we rely increasingly on private sector institutions to manage personal savings and secure retirement incomes, good corporate governance is important to broad and growing segments of the population (OECD, 204, p. 3). OECD uses in their definition above the term good corporate governance. This raises an interesting question; what is good corporate governance? A way to approach this question is to look at a firm guilty of having performed bad corporate governance. One of the most well-known and largest failures in American corporate history was caused by Enron. Enron collapsed November 2001 which caused shareholders their money, the workers their jobs and retirements [Robert B. Thompson (2003)]. The following government subcommittee investigation, which was placed to investigate the board of Enron, found the board guilty of not acting on several red flags in Enron. The Enron Board failed to provide the prudent oversight and checks and balances that its fiduciary obligations required and a company like Enron needed the Enron Board contributed to the company s collapse and bears a share of the responsibility for it [The role of the board of directors in Enron s collapse, 2002, p. 59]. In other words, the board did not live up to their responsibility and showed bad corporate governance. An introduction to the overall theme corporate governance has now been made. It can quickly be realized that it is a very broad topic and many theses can be written about it. Therefore we need to limit our self in our further work on the thesis. We have chosen to go in depth with anti-takeover provisions and how they influence the value of the firm, as it is already described in our hypothesis formulation. 5.1 Agency theory Anti-takeover provisions are in general an issue which gives raise to conflicts between the management of a firm and the shareholders. This is due to the loss of wealth the shareholders might experience in relation to antitakeover provisions as showed in both Gompers et al. (2003) and Bebchuk et al. (2009). To gain an understanding of the theory behind anti-takeover provisions the next section of this thesis will be about agency theory. The section will include such areas as the principal-agent Page 12 of 71

17 relationship, the agency cost associated with this relationship, the corporate control mechanism and finally two hypotheses will be presented offering different views on anti-takeover provisions. The relationship between the owners and the management has existed as long as there have been corporations. This properties of this relationship is gathered under the term called agency theory. In brief terms it states that the interest of the manager is not necessarily aligned with the interests of the shareholders. This agency relationship is first acknowledged by Adam Smith in the start of the industrial era where the introduction of publicly owned firm occurred. He states the following in his work Wealth of Nations (1776). The directors of such companies, however, being the managers rather of other people s money than of their own, it cannot be well expected, that they should watch over it with the same anxious vigilance with which the partners in a private copartnery frequently watcher over their own. Like the stewards of a rich man, they are apt to consider attention to small matters as not for their master s honor, and very easily give themselves a dispensation from having it. Negligence and profusion, therefore, must prevail, more or less, in the management of such a company. 1 The agency relationship is also a witness of separation of ownership and control, a phrase coined by Berle and Means in 1932 in their work The modern corporation and private property. They argue that the owners of the corporations have been separated from the control of them. With diverge interest between the shareholders and the management, and separation of ownership and control, agency cost is bound to happen. Jensen and Meckling (1976) go into detail with this situation. They define an agency relationship as a contract under which the principal (shareholders) engage the agent (management) to have him perform some services. Thereby the principal need to give up some decision power to the agent. According to Jensen and Meckling if both parties are trying to maximize their own utility there are reasons to believe agency costs could occur. This is due to the fact that the agent will not always act in the best interest of the principal. Agency costs will likely increase when the manager s ownership claims of the companydecreases, as the company s interest becomes less important for the manager. 1 The quote is taken from the article: Theory of the firm: Managerial behavior, agency costs and ownership structures by Jensen and Meckling (1976) page 1. Page 13 of 71

18 This will increase the incentive of the manager to spend more money on non-pecuniary 2 benefits. The extra-cost of the non-pecuniary benefits will be on the bill of the firm and thereby of the principal. Agency cost is defined as following by Jensen and Meckling (1976): the monitoring expenditures 3 by the principal plus the bonding expenditures by the agent plus the residual loss 4. The principal need some devices to try and align the interest of the parties by disciplining them. These devices can for example be salary, bonuses, monitoring of the agent, or stock options for the management through specifications in their contracts. Other more corporate oriented mechanisms can be the debt leverage, block ownership or the risk of takeovers. The uncertainty between the principal and agent can be related to asymmetric information. This can be further divided into two subjects; adverse selection and moral hazard [Eisenhardt (1989)]. The adverse selection problem in agency theory might here be that the agent has claimed to have certain skills and therefore demand a higher wage in the hiring process. Adverse selection therefore arises since the principal cannot completely verify the skills of the agent or when the agent works [Eisenhardt (1989)]. Moral hazard is another form of asymmetric information. It occurs when the manager (agent) have an incentive to behave in self-interest and diverge from the agreement between the parties. This is for example seen when the manager spends time and money on non-pecuniary items. The principal has a couple of options in order to try to prevent the adverse selection and moral hazard problem. The first option is to use information systems such as budgeting or reporting systems in order to survey the agent. Another option is to use outcome-based contracts in order to align the goals of the principal and the agent. However, all these options are associated with increased costs for the principal [Eisenhardt (1989)]. If these options fail and the manager is not maximizing the shareholder value then according tothe theory of the market for corporate control will discipline the manager by the fear of either being 2 Non-pecuniary benefits can for example be more spending on better offices and means of business travelling. 3 Includes for example the cost of the effort to trying to control the agent through for example budgets, rules and other directives towards trying to regulate the behavior of the agent. 4 Residual cost is describe in Jensen and Meckling (1976) as In most agency relationships the principal and the agent will incur positive monitoring and bonding costs (non-pecuniary as well as pecuniary), and in addition there will be some divergence between the agent s decisions and those decisions which would maximize the welfare of the principal. The dollar equivalent of the reduction in welfare experienced by the principal as a result of this divergence is also a cost of the agency relationship, and we refer to this latter cost as the residual loss. Page 14 of 71

19 replaced or the firm being bought up and the agents subsequently being replaced enabling new agents to come in an run the firm more effectively with greater benefits for the shareholders [Manne (1965)]. However, the management has the opportunity to restrict and compromise the market for corporate control by for example establishing anti-takeover provisions. This can be viewed as managerial entrenchment since it entrenches the management and makes it impossible for external shareholders to take over the firm and make sure it is run in the most effective way. Anti-takeover provisions were adopted by many firms in the 1980 s during the large hostile takeover wave at that time due to corporate raiders. This caused 57% of the largest firms in USA to be either takeover targets or being restructured on their own [Mitchell and Mullherin (1996)]. Antitakeover provisions in their essence make it more difficult for outsiders to acquire a company by establishing certain methods to stall the acquirement or even render it impossible. 5.2 Anti-takeover provisions and the relating hypotheses Some argue anti-takeover provisions is only in the interest of the managers since it helps to secure their own job positions while others argue it is in the interest of the shareholders due to enabling the managers to bargain the price of the shares up in an takeover when they are sure hostile takeovers cannot occur and instead the bidders need to make a collective offer for the shares. The two hypotheses are in the literature known as the managerial entrenchment hypothesis and the shareholder interests hypothesis. In the literature there are two hypotheses concerning anti-takeover provisions; the managerial entrenchment hypothesis and the stockholders interests hypothesis. Each hypothesis represents their unique approach to anti-takeover provisions and whether they are a benefit for the shareholders or not. The managerial entrenchment hypothesis holds that anti-takeover provisions are set in place to protect and entrench the management s positions and enhance their decision making abilities. This is done on the cost of the shareholders power in the firm. It almost goes without saying, that the managerial entrenchment hypothesis has as an assumption of separation of ownership and control as proposed by Berle and Mean (1932). Page 15 of 71

20 The hypothesis recognizes the shareholders need to incur agency costs in order to ensure that the manager will not take actions which could harm the shareholders. It can be related to the asymmetry information dilemma and the agency cost proposed by Jensen and Meckling (1976) as previous mentioned in the thesis. According to the managerial entrenchment hypothesis the separation of ownership and control gives rise to the following disciplinary mechanisms from the shareholders in order to try to reduce the agency cost [DeAngelo and Rice (1981)]. 1. Management-stockholder contracting. 2. Managerial labor market forces. 3. The threat of job loss through either stockholder vote or stockholder vote. Since none of the three disciplinary mechanism is costless for the shareholders it is assumed they are imperfect at disciplining the managers and will never fully achieve to eliminate the agency problem. For example it is costly to monitor the managers and detect any non-optimal performance, and a proxy fight in order to replace the managers at a shareholder meeting is also known to be very costly [DeAngelo and Rice (1981)]. So one might argue, as long as the cost of the inefficiency of the managers is lower than the cost of having them replaced, they will be secure in their job. DeAngelo and Rice (1981) raise an interesting question. How can it be, when many of the antitakeover provisions need to be approved by shareholders, they get passed? DeAngelo and Rice offer the following three explanations: 1. The behavior of the stockholders is irrational and they have not enough interest or understanding for the subject to care. 2. Stockholders may recognize the negative effect of the provisions but still vote for them, in order to maintain a working relationship with the management. 3. The information and transactions cost required to fight an antitakeover amendment are sufficiently large to rule out effective opposition. That is, the cost for fighting the provision is greater than accepting it. To summarize, anti-takeover provisions serves as a method for the management to entrench themselves and secure their job positions. This entrenchment will bring costs for the shareholder in form of agency cost and lower firm value. The other hypothesis, the shareholder interests Page 16 of 71

21 hypothesis, has the opposite view of anti-takeover provisions and will be explained in the following segment. According to the stockholder interests hypothesis anti-takeover provisions will increase the current wealth of the stockholders. This is on the basis of two reasons. First, the gain for the shareholders according to the hypothesis anti-takeover provisions will enable the management to negotiate better deals for the shareholders seen as one group and not individuals. Consider for example the situation where shareholders have received a tender offer. A single stockholder might accept a premium which fulfills his own utility maximization, but not the rest of the shareholder group. Therefore as one united group the shareholders will be able to achieve a higher premium. However, one could also argue that instead of a higher premium the anti-takeover provisions might discourage certain groups from initiating a takeover bid, and thereby keeping potential premiums away from shareholders [DeAngelo and Rice (1981)]. Second, the anti-takeover provisions can be viewed as a long-term contract with the management. This will encourage the management diverge from short-term projects and instead engage in longterm projects which might appear as unprofitable in the short-run, but will be profitable in the long run [DeAngelo and Rice (1981)]. We have now seen how the relationship between the shareholders and the management can cause conflicts due to the separation of ownership and control. This situation can lead to agency cost which then will decrease the shareholder value. However, according to the corporate control theory managers who do not maximize the profit for the shareholders will in the end be replaced by for example a takeover. But management can protect themselves from this mechanism by the usage of anti-takeover provisions. The next section will look at the G-index and E-index which, as already mentioned, consists of antitakeover provisions. The purpose is to see if there are benefits or certain disadvantages relating to the two indexes, and if it can be argued which one will be more beneficial for us to use in our thesis. Page 17 of 71

22 6. Discussion of the indexes During our research of the existing literature in the field of corporate governance we came across two indexes, the E-index andg-index in the articles Gompers et al. (2003) and Bebchuk et al. (2009), respectively. In order to determine the relevance of the two indexes in the academic world, we looked at how many times the two indexes have been cited in other academic literature writings. According to Google Scholar Bebchuk et al. (2009) and the Gompers et al. (2003) have been cited 950 and 2350, respectively 5. It should be noted that Bebchuk et al. (2009) has only existed for two years (published in 2009) and therefore 950 is to be considered a relative high amount of citing s. It can be argued the higher amount of citing s an article has received the more well-respected it is in the academic world. At least a strong relationship must exist In order to assess which of the two indexes to use in our empirical analysis we looked at the differences between them. The main difference between the two indexes is the number of governance provisions used. The G-index uses all of the 24 anti-takeover provisions in the WRDS corporate governance database, while the E-index narrows the number of provisions down to only 6. This can be seen as a criticism of the G-index due to the fact that Gompers et al. (2003) have not made any consideration of whether or not the 24 provisions are relevant to use in connection with shareholder value. The six provisions in the E-index can be divided into two groups, limits on shareholder voting power and measures against hostile takeovers. They are as follows; staggered boards, limits to shareholder amendments of by bylaws, supermajority requirements for mergers, supermajority requirements for charter amendments, poison pills and golden parachute arrangements. To measure the amount of shareholder power both indexes assign a value of one to a firm if a provision is present 6. This is done in order to make the index as simple as possible by not taking into account the question whether or not each provision has the same influence in restricting shareholder rights. However, as noted by Gompers et al. (2003)this approach sacrifices the 5 The information about the number of citing s has been collected 04/21/ An exception of this approach is the presence of secret ballots and cumulative voting. Gompers et al. (2003) adds one point to the firms cumulative score if they do not have these provision, due to Gompers et al. (2003) view them as an increase in shareholder rights. Page 18 of 71

23 precision on the cost of simplicity in the effort of constructing an index. This is to be considered a weakness in both of the indexes. Both Gompers et al. (2003) and Bebchuk et al (2009) have decided to omit any dual stock class common stock companies therefore the two indexes only consists of single stock class companies. The omitting of the dual stock class companies (less than 10 percent of the total firms) is done due to the fact, that it is difficult to compare dual stock class and single stock class companies because of the differences in governance structures between them [Gompers et al. 2003]. The differences in governance structure arises from dual stock companies having two types of stocks; an inferior and a superior stock in relation to the amount of voting power attached to them. The superior stocks are usually owned by the managers and directors of the firm creating a powerful tool for resisting takeover attempts from outsiders. Therefore it is meaningful to omit the dual stock class firms when testing for a relationship between anti-takeover provisions and firm value. The data source for the indexes is as mentioned the IRRC defense database and this can be considered to be a weak spot. This is due to the fact that IRRC does not update every single company between each of the publications, but firms are given the opportunity to review their status before publication. This information might therefore not depict the correct governance score of the firms [Gompers et. al (2003)]. Therefore the publications should be seen as a noisy measure of corporate governance and not a complete measure. Nonetheless there should not be any reason to suspect the presence of a systematic bias [Gompers et. al. (2003)]. However, this weakness is present in both of the indexes. As mentioned earlier Bebchuk et al. (2009)narrows the provisions down to six. In order to narrow the index down to 6 provisions Bebchuk et al. (2009) used a procedure, which included interviews with six experienced M&A practitioners from six major corporate law firms together with controlling the provisions voted on in the resolutions in the Georgeson Shareholder 7 [Bebchuk et al.(2009)]. One can argue this is a strong point of the E-index, since the authors has made some consideration about which of the provisions is relevant for their research design, and not just using all of them. 7 Georgeson Shareholder is active in shareholder activism and in shareholder consulting services to corporations and shareholders groups working to influence corporate strategy. Page 19 of 71

24 After empirical testing the governance provisions they find that 18 of the 24 provisions in the G- index are insignificant and not correlated with the firm value. The 18 provisions are assigned their own index named the O-index. The authors argue the O-index create unnecessary noise when admitting them into an index, thereby the E-index have a strong advantage in contrast to the G- index. This causes them to warn against using a kitchen sink approach and creating indexes by just adding multiple provisions as done in Gompers et al. (2003). The focus should instead be on those provisions which really matter in relation to firm value [Bebchuk et al.(2009)]. Both indexes have the same weaknesses that they include not completely up to date information and the summation of the provisions without correcting for the different impact of each provision. The E-index holds an advantage from the fact that the authors have made some consideration of the relevance of the 24 provisions. Thereby they can omit the unnecessary provisions which cause noise. To progress our evaluation of the indexes we will now look at the 6 provisions in the E-index. As already mentioned and proven by the authors there are in general a negative relationship between these provisions and firm value. However, is this always the case? In order to further investigate this matter, we will take a closer look at the 6 provisions in the E- index and see if there is some literature in the academic world, which might indicate that there are actual benefits for shareholders in connection with anti-takeover provisions. A discussion of the three provisions golden parachutes, staggered boards and poison pills will be provided in the next chapter. The three other provisions (limitation on amending bylaws and charter and supermajority to approve a merger) will be discussed under one single headline, due to the fact they all limits the shareholders voting power. 6.1 Supermajority, limitation on amending bylaws and charters A supermajority requires at least two thirds of the shareholders vote (sometimes more) on issues such as mergers, takeovers or replacement of the board for a proposal to pass at special meetings and general assemblies. Due to this provision a hostile acquirer who has 51 % of a firm do not Page 20 of 71