University of Amsterdam
Amsterdam Business School
Master in International Finance
Master Thesis
Corporate Governance Challenge for Initial Public
Offering Corporations
MA Ying
Thesis supervisor: Dr. Tomislav Ladika
2
Acknowledgement
First of all, I would like to sincerely appreciate my supervisor Dr. Tomislav Ladika for his valuable guidance and advices. He helped me with the choice of thesis topic, and provided me with the Form S-1 Registration Statement. His suggestions and comments regarding my thesis details inspired me a lot. I learned how to collect data, test my hypotheses and write a professional thesis from his supervising.
In addition, I would like to thank my friends, Xie Linli and Marta Utama. Xie helped me a lot with regression commands and Marta gave me lots of suggestions on my thesis writing.
Finally, I would like to appreciate my parents and friends who supported me during the period.
This thesis is a conclusion about my master studies. I will never forget the time I spent in Amsterdam Business School.
3
Abstract
This paper examines the influence of entrenchment of managers on corporate governance choices by newly public firms at the time of initial public offerings and in the following five years. There are two main results. One suggests that longer-tenured CEOs surround themselves with fewer new directors and fewer independent directors. Another result shows that longer-tenured CEOs are less likely to have classified boards, poison pills, or golden parachutes.
4
Table of Contents
1 Introduction ... 5
2 Literature Review ... 8
3 Hypothesis and Methodology ... 15
3.1 Motivation... 15
3.2 Hypothesis ... 15
3.3 Methodology ... 16
4 Data and Statistics ... 18
4.1 Data before initial public offering ... 18
4.2 Data after initial public offering ... 20
5 Result ... 23
5.1 Part 1 ... 23
5.2 Part 2 ... 26
5.3 Part 3 ... 28
6 Omitted variable ... 30
7 Conclusion ... 31
Reference ... 32
5
1 Introduction
An initial public offering (IPO) takes place when the equity demand of firms cannot be satisfied by single or individual investor. There are two major advantages of going public, greater liquidity and better access to capital. By going public, companies are able to provide their private equity investors the ability to diversify. Most importantly, public corporations will typically have access to much larger amounts of capital through the public markets, both from the initial public offering and from subsequent offerings. However, the major advantage of undertaking an IPO is also one of the major disadvantages of IPO: When investors can choose to diversify their holdings, the equity holders of a corporation will become more widely dispersed. This lack of ownership concentration will decrease investors’ ability to monitor the company’s management, and investors may decrease the price they are willing to pay as a result of the loss of control. The agency problems are more likely to occur when a company is publicly traded, because in public firms there is a separation of ownership and control.
We have to admit that it is not an easy task to set up a good corporate governance structure for a company which just finishes the IPO process. Firms face diverse governance choices, for example whether to deploy takeover defenses and how to construct the board of directors. The choices of corporate governance are important, as they will influence the firm value.
Here, we take Facebook as an example.
The social networking company Facebook Inc. held its IPO on May 18, 2012. This IPO was one of the biggest in both technology and Internet histories, with a peak market capitalization of over $104 billion. Unfortunately, the underlying Facebook (Nasdaq: FB) stock price wasn't so charming as it dropped under $29 a share for the first time the next Thursday. That is more than 23% below its IPO price of $38 on May 18, 2012. The stock price is around $25 currently.1 It keeps declining almost the
6
whole year since going public. What is the reason behind this price declining fact? A major problem is that FB needed to expand into mobile advertising, but even expert investors did not really know the best way to do this. As a newly public firm, Facebook faced a big uncertainty and good corporate governance can support the managers to make wise decision and satisfy all the shareholders as well as themselves. Corporate governance is more important for firms facing more risk. In this thesis, the main topic is about how corporate governance evolves in IPO firms.
Based on this main topic, I found one direction to do my research. Related studies suggest that firms with entrenched or powerful CEOs tend to have poor governance. For example, Adam et al. (2005) finds that firms operated by CEOs who have strong power to influence firm decisions will suffer more variations in firm performance. Brennan and Franks (1997) claim that powerful managers take actions to ensure the continuation of their private benefits of control at the time of IPO and in following several years thereafter. Sah and Stiglitz (1991) also suggest that firms with powerful CEOs have more risk from decision errors. What’s more, for firms that performed well in the past, their managers are more likely to be entrenched (Adam et al. 2005). Reasoning from these findings and other literatures about managerial entrenchment, entrenched managers may set bad governance at IPO or shortly after IPO to secure their private benefit and power. These findings give an interesting investigate direction to me and light my research hypothesis that firms with managers entrenchment are likely to have bad corporate governance at the time of IPO and within the five years following IPO. This thesis focuses on the relationship between corporate performance and managerial entrenchment in newly public firms.
To figure out this relationship, I hand-collected financial data before IPO and governance data at the time of IPO from Form S-1 Registration Statements of 168 newly public technology firms that took IPO from 1996 to 2012 in the United States. There are two reasons that I look specifically at technology firms. One reason is that they tend to be growing fast, but also face a lot of uncertainty. Another reason is that many technology firms do IPOs just a few years after being founded. I acquired
7
governance and directors data in five years after IPO from Risk Metrics databases in Wharton Research Data Services (WRDS). I used Logit analysis to test poison pill, classified boards and golden parachute as dependent variables because these are all dummy variables, and used Ordinary least Squares methods (OLS) to test about G-index, E-index, fraction of independent directors on board, percentage of managers who worked less than three years, and fraction of Equity payment. The main proxy in my thesis to measure managerial entrenchment is CEO working years at the firm. The longer the CEO works in a firm, the more likely he will be entrenched. I also tried firm performance before the IPO, but this is a secondary proxy because the results were not significant.
There are two estimation results. One suggests that longer-tenured CEOs surround themselves with fewer new directors and fewer independent directors. The other one shows that longer-tenured CEOs are less likely to have classified boards, poison pills, or golden parachutes. The results give some supports for my hypotheses. As strong CEOs, they don’t want to be surrounded by new ideas and new managers. At the same time, the confidence level of strong CEOs is really high so they do not find it necessary to apply protections. My research contributes to the existing studies in two ways. Firstly, it extends the studies on corporate governance in newly public firms. Secondly, it tests the direct impact of managers’ entrenchment before IPO on corporate governance choices at the time of IPO.
This paper is organized as follows. Section 1 gives the introduction of my study. Section 2 discusses literature review. Section 3 describes hypotheses and methodology. Section 4 illustrates data collection. Section 5 reports the results and interpretations. Section 6 discusses potential omitted variable bias. Section 7 concludes the study.
8
2 Literature Review
Adam Smith (1776) seems to be the first economist to come up with boards of directors. 156 years later, Berle and Means (1932) put forward a similar point of view. Their main idea was that in public corporations, shareholders do not manage the company although they own it. At the same time, managers do not own a lot of capital but they manage it on a daily basis. This represents the situation that ownership and control are separated. In private firms, however, managers often own a lot of money in the firm. Both of these indicate the agency problems that have typically caught economists’ eyes for many years.
Hermalin and Weisbach (2003) mentioned a plausible hypothesis that boards are an endogenously solution, generated by the markets, to solve organizational design problem. Board is a determined institution that helps to mitigate the agency problems that puzzle large organizations. One of the major conflicts within the board is between the directors and the CEO. The CEO has incentives to “capture” the board, so that he can make his job much more secure and increase the benefits he can get from being a CEO. However, Directors have incentives to maintain their independence. They want to monitor the CEO and to replace the CEO if he has poor performance. A lot of empirical regularities have been established by researchers. First of all, board composition which is measured by the insider-outsider ratio has no correlation with firm performance. However, the number of directors on a firm’s board has negative relation with the financial performance in a firm. Secondly, they find that board characteristics have significant effects on board actions. If a firm has higher percentage of outside directors and smaller boards, it tends to make relatively better, or at least different, decisions regarding to poison pills, acquisitions, CEO replacement and executive compensation. Last but not least, boards appear to develop over time in the light of the relationship between the bargaining position of the CEO and that of the existing directors. There are many important factors affecting changes to boards such as CEO turnover, firm performance, and changes in the
9
structure of ownership.
The only way those executives can impact firm performance is that they have influence over key decision making processes. Based on this concern, Adams, Almeida, and Ferreira (2005) develop and test a hypothesis namely that companies with CEOs who have more effects on decisions should face more uncertainty in firm results. They find that stock returns are more flexible for firms with powerful CEOs. They believe that in a corporation in which the CEO makes most of the final policies, the risk caused by misjudgment is not diversified very well. In other words, the probability of occurrence of either very bad or very good decisions is higher in a firm in which the CEO has greater ability to affect decisions compare with a firm in which many executives are involved in the process of decision making. As a result, they come up with the hypothesis that uncertainty in firm performance goes up together with the degree of CEO involvements, because when CEO has more power, he will be more willing to make decisions with extreme consequences. Their findings show that the relationship between executive characteristics and organizational variables has significant effects for firm performance.
In the structure of a board, economists always focus on board composition and board size.
Numbers of papers have discussed the question that whether more outside board is better or not using several different methods. MacAvoy et al. (1983), Hermalin and Weisbach (1991), Mehran (1995), Klein (1998), and Bhagat and Black (2000) all suggest that there are no significant relationships between corporation accounting performance and the percentage of outside directors within a board. There is another approach, suggested by Morcket al. (1988), is to use Tobin’s Q to define the firm performance. This idea is that Tobin’s Q reflects the “value added” of intangible factors such as governance. Hermalin and Weisbach (1991) and Bhagat and Black (2000) use this method and find that there is no significant relationship between the proportion of outside directors and Tobin’s Q. The same results with the first accounting performance measurements. Finally, Bhagat and Black (2000) test whether there is influence of board composition on long-term stock market and firm
10
output. Same with the above two methods, they do not find any connection between the composition of the board and the performance of a firm. All in all, there is little chance to say that the composition of board has any cross sectional relationship with firm performance.
Regarding to board composition, Jensen (1993) and Lipton and Lorsch (1992) argue that smaller boards will be more effective than larger boards. The common idea is that when boards become too big, agency problems such as director free-riding will grow within the board. As a result, the board will become less meaningful and less involved in the management process. Yermack (1996) does an empirically test and finds evidence for this statement. He analyzes the relationship between board size and Tobin’s Q on a large sample of U.S. companies. At the same time, he controls other variables that will probably affect Tobin’s Q. Yermack’s results demonstrated that there is an obvious inverse relationship between board size and Tobin’s Q.
There are many other important studies related with corporate governance. Bertrand and Mullainathan (2003) assert that smaller ownership will generate moral hazard because managers bear little financial costs if they adhere to pursue their own goals rather than try to maximize shareholders’ wealth. This kind of moral hazard problem can be eliminated by several corporate governance mechanisms such as takeover threats, large shareholders and effective boards. They realize that blue-collar workers’ wages have increased a lot because of the anti takeover laws, and white-collar workers’ wages have jumped even higher. This indicated that managers prefer to pay workers, especially white-collar workers, higher wages, and this statement is consistent with stakeholder theories of the corporations. However, they notice that these kinds of higher wages did not, on net, transfer into greater operating efficiency. This can be referred to that stakeholder protection did not “pay for itself.” They also provide strong evidence of a decrease in the levels of both plant construction and destruction has little effect on overall company size. The decline, rather than increase, in plant construction develops a significant fact about managerial preferences. The majority of managers might be better characterized by
11
what they term “quiet life” models rather than by empire-building models. The average manager in their sample does not appear to try his best to increase the size of companies. Instead, it seems that he tries to avoid establishing new plants, and at the same time he avoids destroying old plants. Managers who are poorly governed probably prefer to avoid making difficult decisions and doing costly efforts related with shutting down old plants or setting up new plants. It can be noticed that the wage results can also fit into this quiet life view very well. Because providing high wages is an effective method for managers to obtain peaceful relationships with their workers.
Most likely, shareholders will accept restrictions of their rights within a firm if these restrictions can maximize their wealth, but few of them known about the ideal state which can balance the power and profit. Gompers, Ishii, and Metrick (2003) examine the relationship between firm performance and shareholder rights. Investor Responsibility Research Center (IRRC) in line with 24 governance provisions, Gompers, Ishii, and Metrick (2003) find out that when giving each provision equal percentage, an index followed these 24 provisions was inversely correlated with firm value. This firm value was measured by Tobin’s Q, as well as stockholder returns during the decade of the 1990s. A “Governance Index”, use G index for short, was created by them to stand for the level of shareholder rights. An investment strategy which can earn abnormal returns of 8.5% per year during the sample period is to buy firms in the lowest deciles of the index which means with strongest shareholder rights and sell firms in the highest deciles of the index which means with weakest shareholder rights. They claim that firms with weaker shareholder rights will have lower firm value, lower sales growth, lower profits, higher capital expenditures, and made much more corporate acquisitions. They present several corporate governance provisions such as anti-greenmail, classified board, compensation plans, golden parachutes, and poison pills.
Bebchuk, Cohen and Ferrell (2008) come up with an index to measure entrenchment, denoted by E index, which is based on 6 provisions: limits to shareholder bylaw amendments, staggered boards, golden parachutes, poison pills,
12
and super majority requirements for mergers and charter amendments. The first step they take is to examine whether these entrenching provisions have certain relationship with lower firm value measured by Tobin’s Q. The result illustrated that higher E index are significantly related with lower Tobin’s Q value. Entrenchment plays an important role to enable the firms with low Tobin’s Q to maintain their low-Q status. Low-Q firms with a relatively high entrenchment level might be prevented from being taken over or forced to make changes that would increase their Tobin’s Q. After this they begin to analyze the extent to which the 6 provisions in the E index are responsible for the correlation between the IRRC provisions and declined stockholder returns during the decade of 1990s. A strategy of selling firms with high E index scores and, simultaneously, buying firms with low E index scores will generate substantial abnormal returns. The eighteen provisions in the G index that are not in the E index represent “noise,” the main advantage of E index is that it can provide an accurate measure of corporate governance quality and it is not affected by the “noise” created by these provisions. Before Bebchuk, Cohen and Ferrell (2008), Cremers and Nair (2005) create another index based on only four of the provisions in the G index and demonstrate that it is inversely correlated with Tobin’s Q, but they do not try to present either that each of the provisions used in their index matters or that other provisions do not matter. Before this one, in another relevant report, Bebchuk and Cohen (2005) suggest that staggered boards are negatively correlated with Tobin’s Q when any other IRRC provisions are controlled. That work did not indicate which IRRC provisions other than staggered boards are inversely correlated with firm value, measured by Tobin’s Q. Bebchuk, Cohen and Ferrell’s article is the only effort to offer a full illustration of the IRRC provisions that do and do not matter although the work using the G index is quite large.
Classified board, also called staggered board, is a key provision in the twenty-four IRRC governance provisions and is legally firmed in 1995. Classified board provision groups the directors in at least two classes with each class being elected at different annual meeting and thus makes it harder for hostile takeover to
13
gain control of a firm. This provision has experienced increasing requesting declassifying from shareholders and 10 large institutional investors since the last decades. Bebchuk, and Cohen (2005) concentrate on classified board and study on around 1400 firms in years between 1995 and 2002. They find that “staggered board alone is also negatively associated with firm value”. Faleye (2007) confirms the results and points out that classified boards help to entrench management and decrease the function of directors.
Poison pill, an important governance provision, was invented in 1982 and has become a popular tool against hostile takeover among U.S. public firms after its introduction (Ryngaert 1988). It grants the existing shareholders (except the acquirer) rights to buy target company’s shares at a benefited price if a certain amount of target company’s shares are acquired by a hostile takeover (Bebchuk and Cohen 2005). Poison pill provision is similar as classified board provision in several aspects, such as making hostile takeover much harder to complete, making shareholders much harder to change bad managers, helping entrench management and decreasing firm value. Malatesta and Walkling (1988) confirm that firms with poison pill have less than average profit in the industries and the below-average performance exists at least in the year before poison pill adoption. Ryngaert (1988) conducts a similar result when researching the effect of poison pill adoption on stock price. Both of the two provisions are double-edged sword because these provisions could protect the interests of existing shareholder and managers when a hostile takeover occurs and make the firm operations stable and consistent.
Mason A. Carpenter, Timothy G. Pollock and Myleen M. Leary put agency and behavioral together and develop a theory of ‘reasoned risk-taking,’ which means the nature of risks undertaken is a result of governance mechanisms and stakeholder characteristics. They illustrate their theory by predicting when corporate governance should be related with strategic risk-seeking regardless a firm's technical core— the degree to which it has expanded internationally. They find that technology-based IPO firms are much less likely to make big global sales when they are backed by a venture capital. IPO firms with huge amount of insider ownerships are very likely global
14
risk-seekers.
The paper by Yael V. Hochberg shows the effects of venture capital backing on the corporate governance of the firm which just finish the IPO. He conducts three independent tests to examine effectively how governance differs between venture-backed and non-venture-backed firms. Firstly, he finds that non venture-backed firms have higher earnings management. Secondly, venture-backed firms will have a higher wealth effect upon the announcement of the adoption of a poison pill. Finally, venture-backed firms have more independent managers. He gives evidence that supports that these effects are not common to all pre-IPO large shareholders.
In the report by Michelle Lowry, he claims that busy directors offer advantages for many firms. While busy directors may be less effective, their experiences make them excellent advisors. Among IPO firms, which have little experience with public markets and likely rely on their directors for advising, he find busy boards to be common and to contribute positively to firm value. Moreover, these positive effects of busy boards extend to all but the most established firms.
I find that among all the researches on the governance, they look at all firms instead of focusing on firms with high growth and uncertainty (i.e. young tech firms). In my thesis, I try to fill this hole to focus on high technology firms which just finish the initial public offerings. These firms face more uncertainty, and corporate governance choices are more important for them.
15
3 Hypothesis and Methodology
3.1 Motivation
According to previous literatures, corporate governance matters more for firms which face more uncertainty about what is the right decision for a manager to make. Because if everyone knows what is the right decision is, then the manager can be fired if he undertakes any other decision. Firms that recently did IPOs are interesting because they often face a lot of uncertainty. These firms have little experience with corporate governance, but at the same time it is crucial that managers make good decisions, because the firms' business models often are not fully developed. So, governance is especially important at these firms.
3.2 Hypothesis
The main hypothesis is whether managerial entrenchment affects governance after the IPO. To test the hypothesis, firm performance is one way to measure entrenchment. One reason this could occur is that if a firm performs well initially, the board of directors will not want to stand up to management, allowing it to become entrenched. Another way also the main proxy to measure managerial entrenchment in this thesis is CEO working years at the firm. The longer the CEO works, the more likely he will be entrenched.
A company success very early, it is more likely to choose worse governance because management is entrenched. The assumption here is that early success leads to management entrenchment. The separation of ownership and control creates the possibility of management entrenchment. When a firm has a long successful history, it is highly possible that the managers and shareholders become good friends, and shareholders follow the instructions of the managers. Facing little threat of being fired and replaced, managers are free to run the firm in their own best interest. As a result, managers may make decisions that benefit themselves at
16
investors’ expense. Management entrenchment and corporate governance go hand in hand. Management entrenchment is often seen as worsening corporate governance and facilitating agency-motivated investments. So, when a firm success early, entrenched managers might be more likely to make bad governance decisions. The poor governance are measured by adopting classified board, choosing less independent board of directors and having higher average G-index value, adopting poison pill and having less average fraction of independent directors on board.
3.3 Methodology
In order to test my hypotheses, the data for newly IPO firms’ financial data before IPO, corporate governance mechanisms at the time of IPO and corporate governance data in following five years after IPO are needed. My focus is on the relation between measure of governance and measure of entrenchment at the time before, of and after IPO.
I will use Logit regression to test my hypotheses about poison pill, classified boards, and golden parachute and use Ordinary least Squares method (OLS) to test my hypotheses about G-index, E-index, fraction of equity pay, percentage of independent directors joined the firm less than three years ago, percentage of all board members that joined within past three years.
The main regression model is as follows:
CorpGovernancei,t
= β0+ β1Entrenchmenti,t+ β2log Asseti,t + β3log Liabilityi,t + β4log Agei,t+ Year fixed effectsi,t+ εi,t
CorpGovernancei,t stands for corporate governance mechanisms at the time
17
directors. (2) Percentage of independent directors joined the firm less than three years ago. (3) Percentage of all board members that joined within past three years. (4) The fraction of equity pay. (5) G-index, (6) E-index, (7) Poison pill, (8) Classified board, (9) Golden parachute.
Entrenchmenti,t measures the entrenchment level of a company by five
proxies: (1) Revenue growth, (2) Net income growth, (3) CEO working years, (4) CEO working years * Revenue growth, (5) CEO working years * Net income growth.
The model uses three standard control variables, asset, liability and average age of managers to control for firm characteristics. And I also add year fixed effects in each of these regressions to control for the change over time. There is no need to add industry fixed effects because all the companies in my sample are from the same industry, i.e. technology industry.
18
4 Data and Statistics
In the initial sample of this thesis, it examines 184 high tech firms that did IPOs in the past 20 years in the United States. Some firms were private companies for a long time before going public while others were around for just a couple years. Information and data about the newly public firms before and at the time of IPO are hand collected from IPO firm’s Form S-1 Registration Statement, for period after IPO is collected from COMPUSTAT database.
4.1 Data before initial public offering
A corporation must first file a registration statement, namely form S-1, with the Securities and Exchange Commission (SEC) before it can trade in the public market place. Form S-1 requires firms to provide comprehensive information about the plan to use the capital proceeds, explain the current enterprise model and competitor situation, as well as demonstrate a brief specimen page of the planned security, provide price making methodology and any dilutions that will take place to other listed securities. The SEC will review all information those companies provided for possible misstatements or omissions. This certification process will no doubt increase the credibility and authenticity of those almost-public firms when compared to other private firms. Each company will have a unique GVKEY code to represent its own S-1 form.
Because the time for different companies to go public is different, the data given covered the period from 1991 to 2012. The majority of firms will provide the financial data of 5 years prior their public time. In this work, all of these available data before their IPO are collected from their S-1 forms. The statistics in this thesis will be expressed in thousands.
In an S-1 form, lots of information about a company can be found. These almost-public companies will describe the initial stock offering, summarize the
19
consolidated financial data, demonstrate the risk factors they face, and reveal the management situation as well as compensation policies. Among all of the information we can get from S-1 form, in this thesis we mainly focus on following dataset: (1) firm’s founding or incorporation year and initial public offering year, (2) board composition and structures, such as who is setting on the board and their name, age and position and also some information to infer whether they are independent directors, (3) whether they have poison pill, classified board, golden parachute in anti-takeover provisions at the time of IPO, (4) revenue, net income, total assets, total liability, equity pay and total pay which are from the selected consolidated financial data section of S-1 form.
20
Table 2.Abbreviation -- Managerial Information
4.2 Data after initial public offering
Wharton Research Data Services (WRDS) is the leading data research platform and business intelligence tool for over 30,000 corporate, academic, government and nonprofit clients in 31 countries. In this thesis, all the information about those high-tech companies after their IPO is collected from WRDS. The following data was analyzed: G index, E index, Poison pill, Classified board, Golden parachutes. Based on the GVKEY code of each company, we can find out there is another code corresponding with it, namely CUSIP code. With the help of their CUSIP code, all the information after their IPO we need can be obtained in the Directors and Governance databases in Risk Metrics.
21
Table 3.Abbreviation -- After IPO Information
Together with their financial reports on line, we can determine the difference before IPO and after IPO regarding to their financial performance. We also analyze the directors and officers. By knowing who sits on the board can help us determine the characteristic of the board. Also, we try to find out whether there is management entrenchment in these companies by looking into the experiences of the boards, such as whether they have worked in competing companies before.
Data summary table 4 below presents descriptive statistics of dependent and independent variables of the sample. This table provides statistics for 168 firms took IPO in 1996-2012. Summary statistics of financial statistics are presented in
22
Panel A; Panel B presents descriptive statistics for managerial information. Of the 188 newly public firms, the mean managers’ age for all the firms is 46.9. In addition, the mean CEO working years for all the companies is 6.48 years. On average, 64.3 percentages of managers working in each firm for less than 3 years, and 67 percentages of directors working less than 3 years. Also, 38.9 percentages of board members are independent in all the companies on average.
Table 4.Data Summary Table
Panel A N Mean Standard
deviation Median 25th percentiles 75th percentiles Revenue growth(%) 160 4.822 15.985 0.659 0.212 2.303
Net income growth(%) 168 2.211 10.166 0.336 -0.427 1.397
Equity pay/Total pay 168 0.228 0.236 0.176 0.010 0.354
Asset 168 20906901.993 259722274.691 67333.000 20319.000 230496.000
Liability 167 2162073.050 23523804.897 33955.000 9283.000 131400.000
Panel B N Mean Standard
deviation Median 25th percentiles 75th percentiles G-index 80 7.413 2.151 7.000 6.000 9.000 E-index 94 1.798 1.151 2.000 1.000 2.000 Prison pill 118 0.229 0.422 0.000 0.000 0.000 Classified board 118 0.602 0.492 1.000 0.000 1.000 Golden parachute 84 0.500 0.503 0.500 0.000 1.000
Ceo working years 168 6.482 5.974 4.000 2.000 8.000
Age 168 46.905 4.855 46.500 44.000 50.000
percentage of all managers working less than 3 years
168 0.643 0.291 0.667 0.444 0.923
percentage of directors working less than 3 years
164 0.670 0.342 0.750 0.400 1.000
percentage of
23
5 Result
5.1 Part 1
The first result I get form my regression is that longer-tenured CEOs surround themselves with fewer new directors and fewer independent directors.
In order to examine the association between corporate performance and entrenchment level at the time of IPO, I run six least square regressions using the model presented in methodology section with the independent of CEO working years. In all the regressions, I include the control variables, the nature log of asset, nature log of liability and nature log of age. The main purpose to include control variables is to reduce the omitted variable bias. At the same time, all of those regressions include the year fixed effects. It’s important to include the year fixed effects because it makes these regressions comparable with each other.
Table 5 reports least square regressions on CEO working years. These regressions on CEO working years are major regression in my study. There is a limitation in my study because there are only 79 and 92 firms have the G-index and E-index, respectively. The sample size is not enough to do regressions with year fixed effects from 1996 to 2012. As the year fixed effects are linearly dependent, so the multi-co linearity assumption is violated if I use all the years from 1996 to 2012. As a result, I need to omit some of the year fixed effects when I type in the command equations in Eviews. I have to delete the year fixed effects from 2009 to 2012 because the number of firms took IPO in 2009 to 2012 is much smaller than other years.
According to Table 5, firms in which CEO has worked for long time have fewer new directors or managers, and also fewer independent directors. I found that if CEO works one year longer at firm, percentage of independent director decreases by 0.37%. This is consistent with strong CEOs do not wish to surround themselves with people who have new opinions or perspectives. When a CEO works in a firm
24
for a long period, he will be very familiar with all the directors who work for him. It is highly possible that the CEO has become good friends with these directors. Because of this, these directors are more loyalty to the CEO and will less likely to make wrong decisions just for their own benefit. This will make the CEO feel much safer. This result is consistent with my hypothesis, CEOs who work in a firm for longer time, are more likely to be entrenched. As a result, they are more likely to make bad corporate governance decisions.
25
Table 5.Result 1
This table shows the least square regression results of 6 dependent variables on past performance. Governance performance is represented by CEO working years in the firm. The control variables are nature log of assets, liability and manager’s average age. Year fixed effects are included. Firms in the sample went public during 1996-2012. Values in parentheses are robust t-statistics. Significance at the 10%, 5%, and 1% levels for a two-sided test are indicated by *, **, and *** respectively.
26
5.2 Part 2
The second result I get from my regression is that longer-tenured CEOs are less likely to have classified boards, poison pills, or golden parachutes.
In order to examine the association between CEO entrenchment level and these three provisions, namely poison pill, classified board and golden parachute, I run three Logit regressions using the model presented in methodology section with the independent variable CEO working years in firms. In all the regressions, I include the control variables, the nature log of asset, nature log of liability and nature log of age. The main purpose to include control variables is to reduce the omitted variable bias. At the same time, all of those regressions include the year fixed effects. It is important to include the year fixed effects because it makes these regressions comparable with each other.
Table 6 demonstrates the Logit regressions on CEO working years. These
dependent variables are dummy variables, equal to 1 if firms have this mechanism in their anti-takeover provisions at the time of initial public offering, otherwise equal to 0. There is one limit in my study. The information for these mechanisms in sample companies is not all available, so the number of observations is less than the sample size.
According to Table 6, firms in which CEO has worked for long time are less likely to use poison pills, classified boards, or golden parachutes. This is consistent with stronger CEOs adopting better governance. This result seems conflict with another one; however, I would explain it as stronger CEOs are confident enough to skip these protections. Classified boards and poison pills protect CEOs from hostile shareholders or outside investors. Also, golden parachutes protect CEOs who might get fired. However, powerful CEOs have more control over the board of directors, and hence may be less likely to get fired and may prevent the board from soliciting takeover offers. As a result, these mechanisms are not necessary for CEOs who already worked in the firm for long time.
Together with the result 1, we find the reason why CEO will not adopt these protection mechanisms, they already worked in one company for a long time and
27
most of directors are loyalty to the CEO. There are few independent directors or new ideas. These will make the CEO pay little attention on the protection mechanisms.
Table 6.Result 2
This table shows the Logit analysis of poison pill, classified board and golden parachute on past performance. The dependent variables are all dummy variables equal to 1 if the firm has this kind of protections at the time of IPO and equal to 0 otherwise. Governance performance is represented by CEO working years in the firm. The control variables are nature log of assets, liability and manager’s average age. Year fixed effects are included. Firms in the sample went public during 1996-2012. Values in parentheses are robust t-statistics. Significance at the 10%, 5%, and 1% levels for a two-sided test are indicated by *, **, and *** respectively.
28
5.3 Part 3
In this part, I explain other regressions which have insignificant results.
In order to examine the association between CEO entrenchment level and corporate governance, I run many other regressions using the model presented in methodology section with four other independent variables, i.e. Revenue growth, Net income growth, CEO working years* Revenue growth and CEO working years* Net income growth. In all the regressions, I include the control variables, the nature log of asset, nature log of liability and nature log of age. At the same time, all of those regressions include the year fixed effects.
Table 7 shows that there is no significant relationship between corporate performance and manager entrenchment when we use Revenue growth and Net income growth to measure the level of entrenchment. At the same time, the interaction terms with CEO working years are mainly insignificant. This means that CEOs who work at company longer are less likely to have a golden parachute or classified board, but this relationship does not vary with firm performance. In other words, it doesn't matter if the long-tenured CEO is at a firm that has performed well or performed poorly; in both cases they are less likely to have a golden parachute or classified board.
29
Table 7.Result 3
This table shows the regression results of all the dependent variables on past performance. Governance performance is represented by Revenue growth, Net income growth and two interaction terms with CEO working years in the firm. The control variables are nature log of assets, liability and manager’s average age. Year fixed effects are included. Firms in the sample went public during 1996-2012. Values in parentheses are robust t-statistics. Significance at the 10%, 5%, and 1% levels for a two-sided test are indicated by *, **, and *** respectively.
30
6 Omitted variable
The model subjects to two main potential omitted variable bias.
First of all, not all the firm characteristics that affect corporate governance could be included in the model as control variables. According to Gompers, Ishii, and Metrick (2003), G-index is not only associated with firm size but also positively associated with share price, trading volume and institutional ownership. Thus it would be better to include ownership by institutional investors at IPO, share price on the first trading day, and trading volume at the time of IPO as control variables when testing the relation between past performance and G-index. Gompers et al. (2003) find that classified board is positively related to the other IRRC corporate provisions. Based on the argument, the regressions with classified board at the time of IPO should also control for the G-index except classified board at that time. In addition, since dual class structure would affect the voting power of shares, it should be controlled for when testing on the fraction of independent directors on board.
In the second place, also the main problem is that in some specific situations, it is better for some firms to keep their CEO and other leadership for a long time. For example, CEOs with long tenure might have founded the firm, and such firms might also perform better and therefore have better governance (e.g. Google). Or, perhaps more innovative firms retain their CEO and other directors for a long time. Take Apple for example, the CEO had a significant influence to Apple’s business; the CEO himself already became part of the brand of his company. As a result, change CEO will have negative effect on the efficiency of the company as well as the stock price. So in real life, there are some cases that omitted variables will have relationship with both CEO working years and corporate governance.
31
7 Conclusion
In my study, I examine the corporate governance choices at the time of IPO by newly public firms that went public in the years from 1996 to 2012. Based on data acquired from Risk Metrics database and hand collected data from firms’ Form S-1 Registration Statement, I carry out sets of regressions using Logit or Ordinary least Squares methods to examine the different corporate governance options by newly IPO firms. There are two estimation results. One suggests that longer-tenured CEOs surround themselves with fewer new directors and fewer independent directors. The other one shows that longer-tenured CEOs are less likely to have classified boards, poison pills, or golden parachutes. My results suggest that stronger CEOs don’t choose good governance since they do not like new ideas or new directors, however, when it comes to their own ability, they choose good governance to skip protection mechanisms because they are very confident about themselves. I could not find significant relation between past performances, which are measured by revenue growth and net income growth, and governance performances.
The study has some limitations.
In the first place, since data of G-index and E-index are not applicable for every firm, the number of observations is less than the sample size. As a result, I need to omit some years fixed effect. At the same time, not all the technological firms apply poison pill, classified boards and golden parachute, the observation number is not enough. Secondly, I use G-Index and E-index as the measure of corporate governance. However, G-Index and E-index are also related to managerial entrenchment. This is a reverse causality problem --- an entrenched manager can cause firm governance to deteriorate, but poor governance can also lead to managerial entrenchment. Solving this problem is beyond the scope of my thesis.
The future study could conduct on this problem. We need to learn more and try to figure out the relationship between corporate governance and managerial entrenchment when it comes to G-index and E-index.
32
Reference
Adams, R., Almeida, H., Ferreira, D., 2005, Powerful CEOs and Their Impact on Corporate Performance, the Review of Financial Studies, Vol. 18, No. 4 (Winter, 2005), pp. 1403-1432.
Aggarwal et al. Initial Public Offerings, Journal of management, 2002, 40: 1066-1089.
Bertrand, M., Mullainathan, S., 2003, Enjoying the quiet life? Corporate governance and managerial preferences, Journal of Political Economy 111(5), 1043-1075.
Berk, J., Demarzo, P., 2011, Corporate Finance, Pearson Education.
Boulton, T., Smart, S., Zutter, C., 2010, IPO Underpricing and International Corporate Governance, Journal of International Business Studies (2010) 41, 206–222.
Bebchuk, L., Cohen, A., Ferrell, A., 2008, What Matters in Corporate Governance? The Oxford University Press.
Chazen, L. et al., 1999, Governance Issues for IPO Companies, the Corporate Board, 1999.
Gompers, P., Ishii, J., Metrik, A., 2003, Corporate Governance and Equity Prices, the Quarterly Journal of Economics, February 2003.
Garvey, G., Milbourn, T., Asymmetric benchmarking in Compensation: Executives Are Paid For Good Luck but Not Punished for Bad. Journal of Financial Economics.
33
Determined Institution: A Survey of the Economic Literature, FRBNY Economic Policy Review, 7-26.
Hermalin, B., Weisbach, M., 1998, Endogenously Chosen Boards of Directors and Their Monitoring of the CEO, The American Economic Review, 96-118.
Hochberg, Yael V. 2003, Venture Capital and Corporate Governance in the Newly Public Firm.
Jensen, M., Meckling, W., 1976, Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, Journal of Financial Economics 3, 305-360.
Kuns, L., Initial Public Offer UnderPricing: The Issuer’s View, The Journal of Finance, 1994, 27, 43-69.
Masulis, R., Wang, C., Xie, F., 2007, Corporate Governance and Acquirer Returns, Journal of Finance, 62(4), 1851-1889.
Prendergast, C., 2002, The Tenuous Trade-off between Risk and Incentives, Journal of Political Economy, 2002, vol. 110, no.5.
REITER, B., 2010, Governance After the IPO, Lexpert Magazine, 2010.
Sandler, R., Weinstein, E., 2011, Corporate Governance Practices for Initial Public Offerings in the United States, The Conference Board, 2012.
Shleifer, A., Vishny, R., 1997, A survey of Corporate Governance, The Journal of Finance, Vol. 52, No. 2 (Jun., 1997), 737-783.
