Does combining the position of CEO and chairman mean reduced returns? : CEO duality in S&P500 companies between 2007 and 2014

1

Does combining the position of CEO and chairman mean reduced

returns?

CEO duality in S&P500 companies between 2007 and 2014

Name: Chantal Rietvink Student number: 10610731

Programme: Economics and Business Specialization: Finance and Organization Supervisor: E. Giambona

Date: 15-06-2016

Abstract

This thesis is about the effect of having the same individual holding both the role of CEO and chairman, also called CEO duality, on the firm’s profitability, in terms of return on equity and return on assets. The effect of changing from a dual structure to an independent structure is also

researched. Both these effects are studied in S&P500 companies during the period of 2007 until 2014. After doing a multivariate cross-sectional analysis, a positive influence of CEO duality on the return on equity and a negative influence on the return on assets was found. Switching from a dual structure to an independent structure has a negative effect on the return on equity and a positive effect on the return on assets.

2

Statement of originality

This document is written by Chantal Rietvink who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

3

1. Introduction

On December the second in 2001 Enron, America’s seventh largest company at that time, declared bankruptcy after being caught up in several accounting scandals (The Economist, 2002). It was a combination of overstating the revenue, hiding debt and creating special purpose entities that led to an enormous decrease in the share price (The Economist, 2002). After the shareholders filed a lawsuit, the U.S. Securities and Exchange Commission began their investigation and another company tried to buy to Enron (The Economist, 2002). Unfortunately the deal failed and this led to one of the biggest corporate bankruptcies in the U.S (The Economist, 2002). On the bright side, this enormous scandal has been a wakeup call for the Securities and Exchange Commission to tighten regulations, introduce the Sarbanes-Oxley Act and give more attention to the phenomenon called corporate governance (The Economist, 2002).

Corporate governance is the way in which a company is directed and controlled (Thomsen & Conyon, 2012). Especially in the US, corporate governance is mainly about the board of directors, what the role is of the board and how the board is composed (Thomsen & Conyon, 2012). The board of directors is implemented to serve as an intermediary between the shareholders of the company and the management of the company (Thomsen & Conyon, 2012). The duties of the board include making decisions about the companies assets, hiring and firing executives, supervise and monitor the executives and decide on special issues like compensation, nomination of new board members and auditing (Thomsen & Conyon, 2012). According to the Global Governance Principels, the board members have to act in the best interest of the company, shareholders and all other stakeholders. For an optimal supervising and monitoring process, it is extremely important that the members of the board are independent, in terms of no family, economic, financial or business related ties (Thomsen & Conyon, 2012).

Even though independent board members, and especially independent leadership of the board, is highly recommended by the Global Governance Principles, 71% of the S&P500 companies combined the position of CEO and chairman of the board in 2005, which means that the CEO of the company also holds to tile of the chairman (Spencer Stuart Board Index). Fortunately, this number has decreased to 52% in 2015 and also the share of independent directors within the board grew from 80% to 84% in 2015 (Spencer Stuart Board Index). Although this is a positive change, the number of companies who combine the position of CEO and chairman is still quite high. The

combination of having the same CEO as chairman is also called CEO duality and there is a great deal of literature about whether this is good for the company or not.

Following the Global Governance Principles, the supervising role of the chairman would be compromised if the chairman is also CEO. Fama and Jensen (1983) also argue this using the agency

4

theory. The agency theory recognizes that in large companies there is a separation between ownership, the shareholders, and control, the board of executives (Fama & Jensen, 1983). The executives have different incentives than the shareholders and they rather act in their own interest than in the best interest of the company (Fama & Jensen, 1983). For this particular reason large companies have to implement a board of directors (Fama & Jensen, 1983). The board has to align the interests of the executives with the interests of the shareholders, supervise the executives and intervene when something is not in line with the best interests of the company, the shareholders or other stakeholders (Fama & Jensen, 1983). This means that when the chairman of the board is also the CEO of the company, the CEO can act more in his own interest without having a board

commenting on this, leading to a poorer performance of the company.

In contrast to the agency theory, the stewardship theory argues that the executives of the company do not always act in their own interest, but they also have an intrinsic motivation to do a good job (Donaldson, 1990). This means that the executives will automatically have the same interests as the shareholders and that there is no need for a board of directors to align those interests (Donaldson, 1990). Using this argument, the company could also benefit from CEO duality. For example, the decision process would be a lot faster if the CEO is also on the board and there is no confusion whatsoever for the stakeholders and the employees of the company who the leader and the spokesman of the company is (Brickley et al., 1997; Baliga et al., 1996).

Being more convinced by the agency theory and the arguments against CEO duality, I argue that CEO duality can have a negative effect on the performance of a company. Not only the

supervising role of the board is compromised, the internal control mechanism within the company also fails and this can lead to keeping the same CEO longer than necessary (Jensen, 1993). Jensen (1993) gives some examples that this can lead to great losses and huge drops in the share price. Consequently, I am going to research to following question: “Does CEO duality have a negative effect on firm performance, in terms of return on equity and return on assets, in the S&P500 companies between the period of 2007 and 2014? And does changing from a dual structure to an independent structure within this period have a positive effect on return on equity and return on assets?”

To answer this question, I did a multivariate cross-sectional analysis. Using 440 companies who were part of the S&P500 during 2007 until 2014 and doing a multivariate OLS-regression, the results showed that having the same individual for the position as CEO as well as chairman has a positive effect on the return on equity, but a negative effect on the return on assets. To see what the effect of changing from a dual structure to an independent structure is, I used 155 companies also from the S&P500 during 2007 until 2014 and found that this change has a negative influence on the return on equity, but a positive influence on the return on assets.

5

stakeholder theory in greater detail and show what the costs and benefits of CEO duality are. After that, I will present the results of previous studies in relation to CEO duality. Then, I will explain the model and the variables I will be using to run the OLS-regression. After running the regression, I will present the results to show how I got the answers to my research question. I will finish with some implications, remarks for further research and a conclusion.

6

2. Literature review

Agency theory

Fama and Jensen (1983) explain that the process of making in a decision in large companies in divided in four steps. The first step is initiation, which is the proposal of different solutions for resource allocation, the second step is ratification, which is the choice of the best solution, the third step is implementation, which is the execution of the choice and the fourth step is monitoring, which is the measurement of the performance (Fama & Jensen, 1983). The initiation and implementation part are called decision management and the ratification and monitoring part decision control (Fama & Jensen, 1983).

In large complex companies Fama and Jensen (1983) recognize that there is a seperation between the residual claimants and the agents who initiate and implement decisions. This means that those agents do not have a financial stake in the decision they make, which could lead to actions that are not in the interests of the residual claimants (Fama & Jensen, 1983). A solution to this problem could be to write a contract between the residual and the agent (Fama & Jensen, 1983). However, Fama and Jensen (1983) elaborate that a contract is costly and difficult, because there are many aspects to take into account. The costs of structuring, monitoring and bonding a set of

contracts are called agency costs (Fama & Jensen, 1983).

A way to reduce these agency costs is to separate decision management and decision control (Fama & Jensen, 1983). According to Fama and Jensen (1983), it is best to leave the decision

management up to the agents in the company, because they have the relevant knowledge for this. The decision control could be done by the residual claimants, however Fama and Jensen (1983) explain that this can better be done by a board of directors, which is formed by the residual

claimants. In this way, the decision management and decision control will be separated, because the board will be able to fire, hire, and compensate the top executives and monitor, and control

important decisions (Fama & Jensen, 1983).

Next to the board, there are also two other mechanisms within the market for common stock which reduce the agency costs (Fama & Jensen, 1983). First of all, Fama and Jensen (1983) explain that the stock market is an excellent external monitoring device, because all the decisions which are made, will be reflected in the stock price. This means that if any of the executives makes a bad decision, the stock price will immediately drop. Secondly, Fama and Jensen (1983) provide information about the market for takeovers. This market has two mechanisms for external

monitoring (Fama & Jensen, 1983). First the stockholders have to right to replace the management if there are not happy with the current management, which is also called a proxy fight (Fama & Jensen,

7

1983). And secondly, an acquirer has to ability to make a tender offer, purchase all the shares and replace the current management (Fama & Jensen, 1983).

Costs of CEO duality

On the one hand Fama and Jensen (1983) argue that it is necessary to have a few top managers of the company on the board, because in this way it is possible for the board to have all relevant information about what is happening within the company to make the best decision. However, Fama and Jensen (1983) also argue that is important to have enough outside members on the board to secure the separation between management and control. If there are not enough outside directors on the board, there is also a chance that the managers will collude to expropriate the shareholders (Fama & Jensen, 1983). This means that the outside members are on the board to prevent the serious agency problems between the executives and the shareholders (Fama & Jensen, 1983).

If the CEO also holds the position of the chairman in the company, the checks and balances of the decision process will be disturbed and the agency costs will be higher (Fama & Jensen, 1983). If there is no separation between the decision management and decision control, the residual claimant is more exposed to opportunistic actions of the decision agents (Fama & Jensen, 1983). The CEO of the company could for example buy a corporate jet, because it is convenient for him to fly to meeting, while it would be cheaper to use a car. This is therefore not in line with the interests of the

company and the shareholders.

Next to that, Jensen (1993) argues that CEO duality is ineffective governance, because the board “sets the rules of the game for the CEO” (p. 862). Ineffective governance is part of the failure of the internal control mechanism (Jensen, 1993). The internal control mechanism is important, because Jensen (1993) gives some examples of companies in which the board failed to fire the CEO on time and this led to great losses and reductions in share prices. Jensen (1993) also explains that having the CEO on the board leads to ineffective control, because the CEO has the power over the board members and this will lead to easier approval and an avoidance of conflicts. The monitoring process can be comprised due to this.

Stewardship theory

The agency theory assumes that the executive who has the control over the company, is a self-interested person whose only goal is to maximizing his own personal economic gain (Donaldson & Davis 1991). This person is individualistic and wants to attain, mainly financial, rewards and avoid punishment (Donaldson & Davis, 1991). These characteristics are retrieved from Theory X from McGregor (Robbins & Judge, 2015). However, the theory of McClelland recognizes that the individual also has needs for achievement, power and affiliation (Robbins & Judge, 2015). On top of that, Donaldson (1990) explains that the behaviour of the individual is not always conscious, which means

8

that the individual also behaves through “habit, emotion, taken-for-granted custom, conditioned reflex, unconditioned reflex, posthypnotic suggestion, and unconscious desires” (p. 372).

This alternative view of motivations by which an individual is driven, is called stewardship theory (Donaldson & Davis, 1991). Donaldson and Davis (1991) explain that the executive also has non-financial motivators and he has an intrinsic motivation to do a good job, if he identifies himself with the corporation. Moreover, the executive wants to be a good steward of the corporate assets and this means that there is no problem of executive motivation (Donaldson & Davis, 1991).

However, Donaldson and Davis (1991) argue that the organizational structure plays an important role in deciding whether the executive will perform in line with the corporation. The best organizational structure according to Donaldson and Davis (1991) is where the CEO has complete authority over the corporation and his role is unambiguous and unchallenged.

Benefits of CEO duality

Donaldson and Davis (1991) argue that it is in the best interest of the company to have the same individual for the role of the CEO and the chairman, because then it is clear who all the authority and responsibility has. This will not only lead to an empowering role for the CEO, but also a clear

leadership for all the managers as well as the board, which will lead to superior returns for the shareholders of the company (Donaldson & Davis, 1991). The CEO can use all the possible

information in the company, his broad knowledge and expertise of the company and commitment to the company to achieve those superior returns (Muth & Donaldson, 1998). Muth and Donaldson (1998) also argue that if the CEO is also chairman, the CEO will be more committed to the company and shareholders’ return because of long-term employment.

Having a CEO who is also the chairman does not only give all the authority and responsibility to one person, there will also be no confusion about who has the absolute leadership of the company and this will lead to no friction whatsoever in the top layer of the company (Baliga et al., 1996). It is clear as well for the employees as for other stakeholders who the representative of the company is and moreover, the CEO will feel more empowered, innovative and entrepreneurial because there is no one above him in the board who will doubt his judgements (Baliga et al., 1996).

Next to that, Brickley et al. (1997) explain that having a combined position of the CEO and chairman reduces information costs. The CEO has specialized, valuable information about the company which could be beneficial for the role of the chairman (Brickley et al., 1997). When the role of the CEO and chairman is splitted, the chairman could miss out on some important information which can be critical in his monitoring role (Brickley et al., 1997). Jensen (1993) also recognizes this; he argues that there are information problems within in the board, because the CEO decides about what to put on the agenda for board meetings and what information he gives to the board. If the

9

CEO is not very co-operative, the board will not have enough information to make the proper decisions. If the CEO is chairman of the board, the board will have all the necessary information needed. Moreover, the board determines the corporate strategy and operating policy, but some boards lack the relevant knowledge to determine these things (Jensen, 1993). Having the CEO as chairman can also solve these kind of problems.

Lastly, having two leaders of a company could make it difficult to appoint who has the blame for any problems occurring in the company (Brickley et al., 1997). Faleye (2007) also gives this argument, in the sense that having both a CEO and a chairman could lead to finger pointing when the company performs poorly. There are other two other small remarks about having an independent structure. Having a CEO as well as a chairman could increase the costs for a company because they both have to pay the salary for the CEO as well as the chairman (Brickey et al., 1997). CEO duality could thus mean cost reducing within in the company. And lastly, Faleye (2007) gives the argument that having a separate CEO and chairman eliminates the possibility of using the incentive to reward the CEO with the position of the chairman, which could demotivate the CEO to do a good job.

Existing research

Donaldson and Davis (1991) did a cross-sectional analysis with 321 companies in the year 1988. They found a significant difference of 3.26% between the return on equity of companies with CEO duality and companies without CEO duality, with the dual-structured companies outperforming the

independent-structured companies. Controlling for industry effects made the percentage drop to 2.38%, but this result was still significant. The shareholders’ wealth was also higher for companies with CEO duality, but this finding was not significant.

Rechner and Dalton (1991) did a longitudinal analysis of 250 randomly selected Fortune 500 companies between the period of 1978 and 1983. Using a multivariate analysis of variance, they found significant results that the companies with an independent structure outperformed the companies with a dual structure on return of investment, return on equity and profit margin for the entire period. The results were however not consistent over time, indicating that in some periods the difference between independent and dual structure was smaller. They also did not control for industry effects.

Baliga et al. (1996) compared Fortune 500 companies who switched from a single position to a splitted position. They did not found a significant announcement effect associated with changing the structure. They also compared several accounting measures, but again they did not find a significant effect in a two-year period. Comparing companies with CEO duality to companies without CEO duality for a longer time-period, still didn’t give a significant effect of CEO duality on firm performance.

10

Brickley et al. (1997) used 661 U.S. firms in their research, in which 80% of the companies CEO duality is present in 1989. First of all, they find that almost no firm has a completely

independent chairman. Second, they find that firms who split the position, have a chairman who has a lot of knowledge and a relatively high stock ownership. They also find that firms with a dual structure have a significant higher return on capital in 1988 than independent-structured firms. On stock return the dual companies also significantly outperformed the independent companies. However, when controlling for industry effects, all the results were insignificant.

Dayha et al. (2009) investigated the relationship between CEO duality and corporate performance after introducing the Cadbury Report in the United Kingdom, which encourages

companies to have an independent structure. When comparing the ROA of firms who always splitting the position to firms who always combined the position, they do not see any difference who

performs better. Next to that, they do not find significant improvement in operating or stock price performance after the recommendation of splitting the position of the CEO and chairman.

All these researches are relatively dated and none of them give a clear answer to whether CEO duality is good for a company or not. My research is more up-to-date and hopefully it can give a clear answer whether it is good for a company’s performance to have the same individual holding both the role of the CEO and the chairman.

Hypothesis

Both the agency theory and stewardship theory give good arguments whether to combine the position of the CEO and chairman or not. The agency theory argues that it is bad for a company to combine the position of the CEO and the chairman, because the monitoring role of the chairman would be compromised and there is a disturbance in the checks and balances (Fama & Jensen, 1983). These agency costs can have a negative influence on the financial performance of the company. Next to that, having a dual structure can also lead to opportunistic behaviour of the CEO and it

compromises the internal control system of the company (Jensen, 1993). This means that combining the position of the CEO and the chairman can lead to lower firm performance, and therefore the first hypothesis is as follows:

H1: CEO duality has a negative influence on return on equity and return on assets.

On the other hand, the stewardship theory tells us that the CEO not always acts only in his best interest, but also has intrinsic motivation to do a good job (Donaldson & Davis, 1991). This means that there are no agency costs when the CEO is also the chairman, because there is no need to monitor the CEO. In fact, there are also a lot of benefits when the position of the chairman and CEO is combined. There could be a reduction in information costs (Brickley et al., 1997; Jensen, 1993), the

11

CEO can feel more empowered and have more incentives to perform well (Donaldson & Davis, 1991; Muth & Donaldson, 1998), and there is no confusion about who is leader of the company is and who the responsibility over the company has (Baliga et al., 1996; Brickley et al., 1997; Faleye, 2007). All these possible aspects of CEO duality can lead to a higher firm performance, which makes the alternative hypothesis:

H2: CEO duality has a positive influence on return on equity and return on assets.

On top of those two hypotheses, following the study of Dahya (2009) and Baliga et al. (1996) a change from a combined position of the CEO and chairman to a splitted position can have a positive effect on the firm’s profitability, because this change can reduce the agency costs and improve the internal control system (Fama & Jensen, 1983; Jensen, 1993). The third hypothesis therefore is: H3: A change from a dual structure to an independent structure has a positive influence on the return on equity and return on assets.

However, this change can also have a negative effect on the firm’s profitability, due to the

information costs and uncertainty about the leadership of the company (Baliga et al., 1996; Brickley et al., 1997; Faleye, 2007). Consequently, I define the fourth hypothesis as follows:

H4: A change from a dual structure to an independent structure has a negative influence on the return on equity and return on assets.

12

3. Data description and model set-up

Data and variables

The data consists of companies who were part of the S&P500 between the period of 2007 until 2014. The data for the title(s) of the executive is retrieved from the Execucomp database and the data for net income, total assets, total debt and dividends paid out is retrieved from the Compustat database. The two datasets are merged using Stata and the observations with missing values are dropped in the regression. After dropping the missing values and deleting some outliers, I was left with 3,328 observations in total for 440 companies. Not all the companies are fully presented in the period of 2007 until 2014, because not all the companies were part of the S&P500 during this period and sometimes not all the data was available.

The dependent variables are return on equity (ROE), following the studies of Rechner and Dalton (1991), and Donaldson and Davis (1991) and return on assets (ROA), following the studies of Muth and Donaldson (1998) and Dayha et al. (2009). Return on equity is used, because this is a measure for the shareholder return (Donaldson & Davis, 1991). Both return on assets and return on equity will be used as dependent variables to see what the effect of CEO duality is on the firm’s profitability. This means that I will use two models with different dependent variables. Return on equity is defined as net income divided by shareholders’ equity. The dependent variable return on assets is calculated as net income divided by total assets.

The independent and most important variable is CEO duality. This is a dummy variable, which is equal to 1 is the title of the CEO is also chairman of the board. The variable is equal to 0 if the CEO is not chairman of the board. Titles defined as co-chairman or executive chairman are also

considered as chairman of the board, so this means that if the CEO is for example co-chairman, the variable for CEO duality is defined as 1. In the sample of S&P500 companies, the mean is 57.38%, which means that more than half of the observations in the sample size combined the position of CEO and chairman.

To see what effect the change from a combined position to a splitted position has, I also use another model with a switch as independent variable. This switch variable is also a dummy, which equals 0 in the period before the switch to the independent structure and 1 during the dual structure. All the companies who always or never had a dual structure were excluded from this sample and that left me with 151 companies and 680 observations. All the years before the switch are included and all the years after the switch, so for some companies there is for example a time period of six years and for some companies only two years.

As control variables firm size, total debt to assets ratio and total dividend pay-out ratio will be used. I will also control for industry performance and time differences. Firm size is defined as the

13

natural logarithm of the total assets and is included to control for firm size effects (Coles et al., 2001). Baliga et al. (1996) argue that larger firms are likely to be near the stage of maturity and as a

consequence they have fewer positive net present value investment opportunities than smaller firms. This can be an influence on the market of value of the company, so this can be reflected in the equity of a company (Baliga et al., 1996).

Next to that, Palmon and Wald (2002) argue that in a small firm the combination of the two positions is preferred because of the communication problems, but in a large firm it is better for the checks and balances to split the positions with regard to agency costs. The findings of the research of Palmon and Wald (2002) support their hypothesis, an announcement for the splitting of the position results in a negative 3-day abnormal stock return for small firms, but a positive one for large firms.

The second control variable is total debt to assets ratio, computed as long term debt plus short term debt divided by total assets (Baliga et al., 1996). This ratio is included to control for the disciplining effect of capital structure (Baliga et al., 1996). Jensen and Meckling (1976) explain that if the company has a lot of debt, the managers of the company will engage in more risky projects, because if this investment turns out positive, the shareholders will benefit from the gains, but if the investment turns out negative, the debt holders will bear the costs. This means that it is preferable for the shareholders to have some debt. However, Jensen and Meckling (1976) also recognize that there are bankruptcy costs attached to having too much debt.

The total dividend pay-out ratio is also included as a control variable, because Easterbrook (1984) argues that dividends have an effect on the agency costs. Easterbrook (1984) explains that the agency costs also consist of the risk aversion of managers. Usually, managers are risk-averse, because if they make a risky investment and the company performs poor or goes bankrupt, the managers will lose their jobs (Easterbrook, 1984). This means that managers usually go for a safe investment, but this is the one with a lower expected return (Easterbrook, 1984). However, Easterbrook (1984) argues that shareholders prefer a more risky investment with a higher return, because the

debtholders will pay the failure and the shareholders will get the gains. A way to give incentives to the managers to take on more risky investments is the pay-out of dividends (Easterbrook, 1984). A constant dividend pay-out ratio forces the manager to seek out ways to raise money to be able to pay-out the dividends (Easterbrook, 1984). This means that the dividend pay-out ratio of a company can lower the agency costs.

Lastly, I will account for both the industry effects and time effects. Industry effects are important to take into account, because Schmalensee (1985) shows in his research that industry effects account for more than 75% of the variance on the industry average rates of returns. This means that the industry, in which a company operates has a great influence on the return on assets and return on equity of the company. I will use the Standard Industrial Classification (SIC) codes as

14

dummies to control for industry effects. I will also control for time effects, because part of the period of 2007 until 2014 I choose is during the financial crisis. At the end of 2008 there was a huge drop in the S&P500 (Yahoo Finance) and next to that, the S&P500 is always volatile so it’s important to take the time effects into account.

Taking all the variables into account means that I will use the following four models:

ROEit= β0+ β1∗ CEOit+ β2∗ SIZEit+ β3∗ TDit TAit

+ β4∗ DPit+ β5∗ INDUSTRY + β6 ∗ TIME + εti

ROAit= β0+ β1∗ CEOit+ β2∗ SIZEit+ β3∗ TDit TAit

+ β4∗ DPit+ β5∗ INDUSTRY + β6 ∗ TIME + εti

ROEit= β0+ β1∗ SWITCHit+ β2∗ SIZEit+ β3∗ TDit TAit

+ β4∗ DPit+ β5∗ INDUSTRY + β6 ∗ TIME + εti

ROAit= β0+ β1∗ SWITCHit+ β2∗ SIZEit+ β3∗ TDit

TAit+ β4∗ DPit+ β5∗ INDUSTRY + β6 ∗ TIME + εti

15

4. Results

Empirical findings

First looking at the correlation matrix in table two, it is remarkable that the correlation between the ROE and ROA is relatively small, around 12.1%, while Muth and Donaldson (1998) found a correlation in the data of 1994 a correlation of 59% and in data of 1993 a correlation of 83%. ROE and ROA are both measures of the firm’s profitability, so I was expecting a higher correlation between those two variables. Fortunately, the correlation between the independent variable and the control variables is low, so there will not be a problem with multicollinearity.

Moving on to the regressions, I did a stepwise regression using OLS adding variables to improve the quality of the regression. I started with a simple OLS-regression and followed with a multivariate-cross sectional OLS-regression. I used robust standard errors, making sure that the estimates of the coefficients will not be biased. For the first five regressions without taking time effects into account, I used the cluster option because the 3,328 observations are for 440 companies and consequently those observations are not independent. For the last two regressions controlling for time effects, I did a panel regression using fixed effects.

Using return on equity as a dependent variable, the effect of CEO duality is in all the regressions positive, so this means that combining the position of the CEO and the chairman does have a positive effect on the return on equity. However, only in the last two regressions the

coefficient is significant, when controlling for time effects at the 5% level and when controlling both for industry and time effects the 10% level. This is not in line with both the study of Donaldson and Davis (1991) and Rechner and Dalton (1991), who both found a negative effect of CEO duality.

Firm size is positively related to the return on equity, which means that a bigger firm has a higher return on equity in comparison with a small firm. This is contrary to the argument of Baliga et al. (1996), which stated that a larger firm has fewer investment opportunities, leading to a lower return on equity. Following this argument, I expected a negative relation between firm size and return on equity. However, the coefficient is relatively small and only in two regressions significant at the 10% level.

Another interesting outcome is the negative influence of the debt to assets ratio on the return on equity. This means that having more debt has a negative influence on the return on equity. This is also not in line with the theory of Jensen and Meckling (1973), who stated that having more debt makes the CEO take on more risky investments, which can lead to a higher return on equity. However, Jensen and Meckling (1973) also explained that having more debt does impose more risk of bankruptcy, so probably these costs have a greater influence on the return on equity.

pay-16

out ratio reduces agency costs, the effect on return on equity is really small and only in one

regression significant at the 10% level. The effect of the dividend pay-out ratio however is positive, so this in line with the explanation of Easterbrook (1984).

When using the return on assets as the dependent variable, the results are slightly mixed. In the first four regressions, CEO duality has a positive effect on the return on assets. However, when taking time and industry effects into account, combining the position of CEO and chairman has a negative influence on return on assets. Unfortunately, only when controlling for time effects, the coefficient is significantly different form zero at the 5% level.

Interestingly, firm size has a negative effect on return on assets. This is in line with Baliga et al. (1996), but firm size had a positive effect on return on equity, so that is what makes this result quite interesting. The debt to assets ratio has a negative influence on return on assets, the same result as on return on equity and this effect is in all the regressions significant at the 5% level. The dividend pay-out ratio is in all but one the regressions positive and significant, in line with the regression on return on equity.

To give an answer to the research question whether CEO duality has a positive or negative influence on the return on equity and return on assets, both part of the first and second hypotheses are true. Combining the position of the chairman and the CEO has a positive effect on the return on equity, but a negative effect on the return on assets.

The results of the regression with the switch from a dual structure to an independent structure as the independent variable give almost the same results. Using return on equity as the dependent variable, switching from a combined position to an independent structure has a negative influence on the return on equity. The results are however not significant. This is comparable to the results of the study of Dahya et al. (2009). These results are logical, because I used part of the same sample as I used in the previous regressions, which gave a positive results to a dual structure, so switching to an independent structure should have a negative influence. However, the effect of switching to an independent structure is larger than the effect of CEO duality.

Firm size is again positively related to the return on equity and the dividend pay-out ratio as well. Surprisingly, the debt to assets ratio has a positive influence on the return on equity, which is different than in the previous regressions using CEO duality. However, only the effect of firm size is significant at the 10% level, the other coefficients are not significantly different from zero.

The last regression using return on assets as a dependent variable also gives comparable results. It has a positive influence on the return on assets when a company switches from a dual structure to an independent structure. The effect is relatively small, however the results are

significant at the 5% level. Firm size has a significant negative effect on return on assets in all but one regression. This is the same result as the previous regression using CEO duality. The debt to assets

17

ratio also has a negative influence on the return on assets and this is also significant. The dividend pay-out is in all but one regression negative, but not significant.

The answer to the last part of the research question whether switching from a dual structure to an independent structure has a negative or positive influence is also mixed. Switching from a combined position to a splitted position has a negative effect on the return on equity, but it has a positive effect on the return on assets. This means that both of the third and fourth hypotheses are partly true.

18

4. Discussion and conclusion

Discussion

Several studies also found mixed results whether CEO duality has a positive influence on the firm’s performance. In my research, it is however quite strange that the effect of CEO duality is positive on the return on equity, but negative on the return on assets. Both return on equity and return on assets are indicators of the firm’s profitability, so I would have expected similar results. A possible explanation for this could be that I used the wrong data for the shareholders’ equity in Compustat, because there were several measures to choose from.

Another shortcoming in my research is that I only looked at the employment title of the CEO and when the CEO also had the chairman title, I indicated this as CEO duality. I assumed that if the CEO did not also have chairman as title, the chairman would be, for the most part, independent. However, it is also possible that the chairman of the company is also the president or another top executive, so in that case the chairman is not completely independent. Donaldson and Davis (1991) did take this into account in their research.

Although I did take time effects into account, the financial crisis was during the time period I researched and this financial crisis could have had a greater impact than I accounted for. It could be possible that due to the crisis, more companies switched from a dual structure to an independent structure because of the wide attention given to scandals with the board, but maybe it would have been better for the company to have stayed with the dual structure.

Lastly, due to the mixed results and some results that were not significant, it is hard to say what the impact of CEO duality is on the firm’s performance. For further research, it might be interesting to look at other aspects where CEO duality has an impact on. Maybe if the CEO also holds the position of the chairman, the tenure of the CEO is longer, because as chairman he has impact on the decision to fire him or not. This holds the same as for the compensation for the CEO, due to CEO duality it might be the case that the compensation of the CEO is higher than it would be with an independent structure. It would be very interesting to also look into those kinds of aspects and not only to the effect on the firm’s profitability.

Conclusion

During the period of 2005 and 2015 a lot of companies changed their managerial structure from having the same CEO as chairman to two splitted positions held by different individuals. Having CEO duality in a company can induce agency costs and it can be better for the shareholders to have an independent chairman. However, combining the position of the CEO and chairman also has

advantages. The board will have access to more information if the CEO is on the board, there is more clarity within the company about the leadership and it can empower the CEO to also have the role of

19 the chairman.

Due to these different arguments for as well as against CEO duality, I was interested in looking at what the effect of CEO duality is on the firm’s profitability in S&P500 companies during the period of 2007 until 2014. I looked at the firm’s profitability in terms of return on assets and return on equity. Surprisingly, combining the position of the CEO and the chairman has a positive effect on the return on equity, but it has a negative effect on the return on assets. This could mean that it would be good for the shareholders of the company to combine the two positions, but it would have a negative effect on the company itself. Next to that, if the company changes from a dual structure to an independent structure, the return on equity is positively influenced, but again the return on assets is negatively influenced.

The results are slightly mixed and quite interesting. The related literature also gives mixed results about whether CEO duality is good for the performance of the company or not.

Unfortunately, this makes it impossible to give a compelling conclusion and recommendation for S&P500 companies to keep the CEO also as chairman or to change this into an independent structure.

20

5. Appendix

Table 1: Summary of variables CEO duality

Variable Observations Mean Standard dev Min Max

ROA 3,328 0.066 0.081 -0.853 0.903

ROE 3,328 0.152 1.645 -36.500 37.515

CEO duality 3,328 0.557 0.497 0 1

Firm size 3,328 9.514 1.334 5.641 14.592

Debt to assets ratio 3,328 0.241 0.186 0 1.562

Dividend pay-out ratio 3,328 0.323 1.077 -17.4 23.876

Table 2: Correlation matrix

ROA ROE CEO duality Firm size Debt to assets Div pay-out

ROA 1.000 ROE 0.121 1.000 CEO duality 0.003 0.006 1.000 Firm size -0.200 0.013 0.147 1.000 Debt to assets -0.197 -0.018 -0.018 -0.036 1.000 Div pay-out -0.007 0.006 0.027 0.021 0.132 1.000

Table 3: Summary of variables switch from duality to indepedent structure

Variable Observations Mean Standard dev Min Max

ROA 680 0.059 0.087 -0.853 0.409

ROE 680 0.090 1.969 -34.327 30.992

Switch 680 0.410 0. 492 0 1

Firm size 680 9.397 1.149 6.702 13.874

Debt to assets ratio 680 0.257 0.175 0 1.183

21 Table 4: OLS regressions CEO duality

Depedent variable: ROE 1) 2) 3) 4) 5) 6) 7) Constant 0.141* (0.043) 0.002 (0.250) 0.048 (0.213) 0.049 (0.212) -0.134 (0.229) -0.028 (0.764) -0.160 (0.244) CEO duality 0.021 (0.066) 0.015 (0.065) 0.014 (0.066) 0.014 (0.065) 0.069 (0.056) 0.211* (0.155) 0.069** (0.045) Firm size 0.015 (0.024) 0.014 (0.023) 0.014 (0.023) 0.033** (0.024) 0.024 (0.086) 0.036** (0.024) Debt to assets -0.159 (0.279) -0.167 (0.284) -0.543** (0.368) -0.736* (0.593) -0.516 (0.443) Div pay-out 0.012 (0.012) 0.044 (0.034) 0.047** (0.035) 0.434 (0.037) Industry dummies No No No No Yes No Yes

Time dummies No No No No No Yes Yes

Number of obs

3,328 3,328 3,328 3,328 3,328 3,328 3,328

R-squared 0.0000 0.0002 0.0005 0.0006 0.0845 0.0004 0.0845 Between brackets robust std. err.

22 Table 5: OLS regressions CEO duality

Depedent variable: ROA Constant 0.065* (0.004) 0.180* (0.022) 0.206* (0.022) 0.206* (0.022) 0.121* (0.039) 0.118* (0.058) 0.138* (0.020) CEO duality 0.0006 (0.004) 0.005 (0.004) 0.005 (0.004) 0.005 (0.004) -0.005 (0.004) -0.011* (0.004) -0.004 (0.004) Firm size -0.012* (0.023) -0.013* (0.002) -0.013* (0.002) -0.0002 (0.004) -0.0001 (0.006) -0.002 (0.002) Debt to assets -0.089* (0.015) -0.090* (0.015) -0.086* (0.023) -0.187* (0.037) -0.088* (0.197) Div pay-out 0.0018* (0.001) 0.001 (0.001) -0.006 (0.001) 0.001 (0.001) Industry dummies No No No No Yes No Yes

Time dummies No No No No No Yes Yes

Number of obs 3,328 3,328 3,328 3,328 3,328 3,328 3,328 R-squared 0.0000 0.0411 0.0828 0.0834 0.2790 0.0381 0.2787 Between brackets robust std. err.

23 Table 6: OLS regressions dual switch to independent Depedent variable: ROE Constant 0.159* (0.084) -0.934 (0.769) -0.945** (0.715) -0.936** (0.712) -4.593** (3.547) -3.756** (2.585) -4.679** (3.539) Switch -0.169 (0.189) -0.153 (0.183) -0.153 (0.183) -0.156 (0.186) -0.329 (0.004) -0.341 (0.305) -0.410 (0.352) Firm size 0.116** (0.079) 0.116** (0.077) 0.115** (0.077) 0.443** (0.331) 0.401** (0.281) 0.436** (0.329) Debt to assets 0.027 (0.411) 0.010 (0.423) 0.333 (1.004) 0.765 (1.056) 0.334 (1.056) Div pay-out 0.014 (0.017) 0.012 (0.023) 0.058 (0.053) 0.006 (0.025) Industry dummies No No No No Yes No Yes

Time dummies No No No No No Yes Yes

Number of obs 680 680 680 680 680 680 680

R-squared 0.0018 0.0063 0.0063 0.0064 0.0983 0.0058 0.1050 Between brackets robust std. err.

24 Table 7: OLS regressions dual switch to independent Dependent variable: ROA Constant 0.053* (0.007) 0.140* (0.047) 0.167* (0.047) 0.167* (0.007) 0.087 (0.088) -0.066 (0.171) 0.116** (0.087) Switch 0.013* (0.006) 0.012* (0.007) 0.012* (0.007) 0.012* (0.007) 0.020* (0.007) 0.018* (0.009) 0.017* (0.009) Firm size -0.009* (0.079) -0.010* (0.005) -0.010* (0.005) -0.002 (0.007) 0.021 (0.019) -0.003 (0.007) Debt to assets -0.065* (0.028) -0.066* (0.028) -0.153* (0.063) -0.327* (0.117) -0.151* (0.060) Div pay-out 0.0004 (0.002) -0.0003 (0.001) -0.001 (0.002) -0.0002 (0.002) Industry dummies No No No No Yes No Yes

Time dummies No No No No No Yes Yes

Number of obs 680 680 680 680 680 680 680

R-squared 0.0055 0.0205 0.0376 0.0851 0.3481 0.0702 0.3761 Between brackets robust std. err.

25

6. Reference list

Baliga, B. R., Moyer, R. C., & Rao, R. S. (1996). CEO duality and firm performance: what's the fuss?.Strategic Management Journal,17(1), 41-53.

Brickley, J. A., Coles, J. L., & Jarrell, G. (1997). Leadership structure: Separating the CEO and chairman of the board. Journal of corporate Finance, 3(3), 189-220.

Coles, J. W., McWilliams, V. B., & Sen, N. (2001). An examination of the relationship of governance mechanisms to performance. Journal of management, 27(1), 23-50.

Dahya, J., Garcia, L. G., & Van Bommel, J. (2009). One man two hats: what's all the commotion!.

Financial Review, 44(2), 179-212.

Donaldson, L. (1990). The ethereal hand: Organizational economics and management theory. Academy of management Review, 15(3), 369-381.

Donaldson, L., & Davis, J. H. (1991). Stewardship theory or agency theory: CEO governance and shareholder returns. Australian Journal of management,16(1), 49-64.

Easterbrook, F. H. (1984). Two agency-cost explanations of dividends. The American Economic

Review, 74(4), 650-659.

Faleye, O. (2007). Does one hat fit all? The case of corporate leadership structure. Journal of

Management & Governance, 11(3), 239-259.

Fama, E.F., and Jensen, M.C. (1983). “Separation of Ownership and Control”. The Journal of Law &

Economics, 26(2), 301–325.

Global Governance Principles (2014), International Corporate Governance Network. Retrieved from https://www.icgn.org/policy.

Jensen, M. C. (1993). The modern industrial revolution, exit, and the failure of internal control systems. The Journal of Finance, 48(3), 831-880.

Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305-360.

Muth, M., & Donaldson, L. (1998). Stewardship theory and board structure: A contingency approach. Corporate Governance: An International Review, 6(1), 5-28.

Palmon, O., & Wald, J. K. (2002). Are two heads better than one? The impact of changes in

management structure on performance by firm size. Journal of Corporate Finance, 8(3), 213-226.

Robbins S.P., & Judge T.A. (2015). Organizational Behavior. Harlow: Pearson Eduction Limited. Schmalensee, R. (1985). Do markets differ much?. The American economic review, 75(3), 341-351. Spencer Stuart U.S. Board Index 2015, Spencer Stuart, 30th edition. Retrieved from

26

S&P500 Index Historical Chart. Yahoo Finance. Retrieved from

http://finance.yahoo.com/echarts?s=%5EGSPC+Interactive#{"customRangeStart":116760600 0,"customRangeEnd":1419980400,"range":"custom","allowChartStacking":true}

The real scandal, (2002, January 17th). The Economist. Retrieved from http://www.economist.com/node/940091.

Thomsen S., & Conyon M. (2012). Corporate governance: Mechanisms and systems. Berkshire: McGraw-Hill.