Managerial compensation, CEO characteristics and
corporate performance
A study on America’s highest paid Chief Executives
Elisabeth Maria Veldwiesch! Master thesis Finance: EBM866B20
University of Groningen Supervisor: Dr. V. Angelini Date: June 26th, 2014
Abstract
This paper investigates the effect of corporate performance and several chief executive officers’ (CEOs) characteristics on managerial compensation. The agency theory hypothesizes that CEOs are given certain incentives to select and implement actions that increase shareholder wealth. Therefore a positive relation between pay and performance is expected. The estimates indicate that managerial compensation (including salary, bonus and market value of shares owned) changes by $95 for every $1,000,000 change in shareholder wealth (which is used as a measure of corporate performance). Compensation generated by share ownership is large in relation to salary and bonus payments, although most CEOs hold small fractions of their companies’ shares.
JEL classification: G3; G32
Keywords: Chief executive officer compensation; Managerial incentives; Corporate performance; Chief executive officer characteristics; Agency theory
2
1. Introduction
In recent years the levels of compensation given to managers and executives have become a subject of great interest in several countries. The level of bonuses is of special interest. One point under discussion is whether the high levels of compensation of executives can be justified by the economic performance of the firms under their direction. See Jensen and Murphy (1990b) and Crystal (1992) for two opposite views on this. This study uses ordinary least squares to investigate the effects of corporate performance on managerial compensation as measured by shareholder wealth, CEO’s age, educational background, gender, tenure and threat of being replaced.
The agency theory hypothesizes that executives need to be given certain incentives to select and implement actions that increase shareholder wealth. In consistence with the agency theory, managerial compensation based on corporate performance seems natural, since an executive has more direct impact on corporate profits than their subordinates do. Typically an executive is an important decision maker and manager. Hence, corporate performance can be influenced by the CEO’s actions and is especially important for shareholders, since the shareholder’s objective is to maximize shareholder wealth. However, shareholders are frequently unable to access all relevant information about these actions. An executive possibly also undertake actions which would not always increase the shareholders’ wealth. CEOs could be indifferent to the selection and implementation of certain actions, as they receive a compensation which is not closely linked to corporate performance. For instance, salary is less dependent on corporate performance than the value of an executive’s shareholdings. The agency theory suggests that in these situations a compensation policy needs to be designed to make sure that executives select and implement actions that increase shareholder wealth. This policy is more effective when company information available to the CEO is equal to the information shareholders can access. However, especially publicly held firms often deal with a conflict of interests in information availability for shareholders and executives. This common example of the principal–agent problem, or agency dilemma, concerns the difficulties in motivating the agent (the CEO), to act in the best interest of the principal (the shareholders) rather than in his or her own best interest. This problem thus arises when the two parties have different interests and asymmetric information exists, i.e. usually where less information is available to the shareholders. When a CEO’s compensation is less dependent on the corporate performance, he or she will have less incentive to act in the best interest of the shareholders. Therefore most researchers in this field hypothesize that executive compensation has to depend on the firm’s performance.
3
individual characteristics of CEOs as a determinant for compensation. In the current research CEO’s age, tenure, gender, educational degree and the threat of being replaced are considered in addition to corporate performance. Increases in age and tenure and expected to increase managerial compensation. In accordance with Jensen and Meckley (1976), who suggest that with tenure, knowledge and skill increase and job performance will improve. According to many studies, differences in pay between male and female executives still exists. Consistent with Elkinawy and Stater (2011) it is expected that women’s salary remains below those of men. A higher educational degree is expected to increase managerial compensation, because among all, Jalbert et al. (2004) found that skills earned through an advanced graduate degree receive higher managerial compensation. The threat of being replaced is expected to lower managerial compensation. For instance Jensen and Murphy (1990) found an inverse relation between the threat of being replaced and managerial compensation.
This paper has one principal aim. That is to examine whether the compensation of executives is systematically related to corporate performance and to age, gender, threat of being replaced, educational degree and tenure of executives. Therefore the following two research questions are developed:
Are corporate performance and CEO’s age, education and tenure positively related to managerial
compensation? (i)
Do a CEO’s threat of being replaced and being female have a negative effect on managerial compensation? (ii)
The individual characteristics thus include age, gender, threat of being replaced, educational degree and tenure of executives. This study focuses on a dataset of the highest paid CEOs in America in 2006 to 2012. Information on more than 400 CEOs in more than 300 corporations is used. In addition, the data includes the global financial crisis with its market collapse in 2008. Therefore it can be observed how managerial compensation is related to corporate performance in years of distress.
Ordinary least squares regressions will be used to estimate the effect corporate performance and age, gender, threat of being replaced, educational degree and tenure of executives on managerial compensation. The study will emphasize the effect of all variables with panel data featuring entity fixed effects.
4
2. Discussion on literature, hypotheses development and contributions 2.1 Prior literature
Managerial compensation has received much theoretical attention since the 1990s. Most of the theoretical work has addressed the justification of the high levels of compensation by corporate performance. See for instance papers by Diamond and Verrechia (1982) and Murphy (1999). Corporate performance usually depends on unobservable executive effort. Therefore most research has been carried out within the framework of the agency theory. For instance Jensen and Meckling (1976) and Jensen and Murphy (1990a) use the agency theory in their research. As explained before, the agency theory concerns the difficulties in motivating the agent to act in the best interest of the principal rather than in his or her own best interest. The theory sets out the need to design a compensation policy. Typically this involves tying the compensation to a measure of corporate performance. This theory is also considered in the current paper.
Most prior studies provide evidence of a positive relationship between corporate performance and managerial compensation. Most of these studies were carried out on managerial compensation in large companies in the US. See Rosen (1992) and Gomex-Mejia (1994) – using stock market indicators as a measure of corporate performance- for two comprehensive reviews. Studies using accounting measures, for instance Leonard (1990) and Sloan (1993) provide the same evidence. There are few studies carried out outside the US which support a positive pay-performance relationship. Examples are Meyerson (1994) for Sweden and Kaplan (1994a) and (1994b) for Germany and Japan respectively.
5
Much of the existing empirical work on managerial compensation is in addition to managerial incentives related to risk taking. These studies examine whether managerial incentives, depending on the level of managerial risk aversion, have observable operational and policy implications. See Hall and Liebman (1998), Murphy (1999), Perry and Zenner (2000), Coles and Naveen (2006) and many others. A common conclusion is that convex payoffs should be given to executives to diminish the effect of risk aversion and provide the executives with incentives to take on value increasing, risky projects. For example see also Guay (1999) and Core and Guay (1999).
Among these incentives several other factors might influence managerial compensation. As illustrated in Murphy (1985), stock options should be included to calculate a more complete measure of a CEO’s compensation. If the value of the stock option grant is based on performance, this should provide an incentive to make decisions which increase shareholder wealth. The role of options in managerial compensation is also analysed by Ju et al. (2002), who demonstrate that a call option can either lead to a higher or a lower compensation, depending on managerial risk aversion and investment technology.
The pay-performance relationship is only a partial analysis within the agency theory. The analysis is more complete when also other factors, influencing executive’s decisions, are considered. Jensen and Murphy (1990a) and Weisbach (1988) already considered the threat of being replaced in their study. They found an inverse relation between the threat of being replaced and managerial compensation. Therefore executives are disciplined by this threat. The threat of being replaced is also considered in the current paper. Besides, some other potentially important factors governing managerial compensation are considered. These are age, gender, educational degree and tenure of executives. Some studies have accounted for a part of these factors.
Jensen and Meckley (1976) address the Human capital theory in their study, this theory suggests that with tenure, knowledge and skill increase and job performance will improve. Based on this theory, Berger et. al (1997) argues that executives with longer tenures are more likely to protect human capital. Therefore greater tenure is expected to increase managerial compensation. Brookman and Thistle (2009), and McKnight and Tomkins (2004) also state that performance and compensation increases with tenure for CEOs.
The classic “taste for discrimination” model developed by Becker (1971)suggests that certain groups which are underrepresented will be paid a lower wage than equally skilled workers in the better represented group. Elkinawy and Stater (2011) found that large firms have more male-dominated boards. They conclude that in these firms the gender differences in salary are larger. Although the number of women in top managerial positions has increased, they remain underrepresented and accordingly, their salaries remain below those of men.
Differences in educational degree that lead to differences in compensation are found by Jalbert et al. (2004) who use Forbes CEO compensation list for the period 1987-1999. They found that those CEOs who have a doctoral degree earn more than those with other degrees, and those with no graduate degree (for instance a bachelor’s degree) or a law degree earn less than others. This evidence thus suggests that skills earned through an advanced graduate degree receive higher managerial compensation.
6
found that younger executives will adopt risk averse strategies which, on average, adversely influence future firm performance. When the pay-performance relation in these firms is strong, this averse risk strategies will lower firm performance and therefore lower managerial compensation. However, since studies on a direct impact between age and managerial compensation are rare, this provides an opportunity in the current study to examine this relationship.
2.2 Hypotheses
While the literature provides considerable evidence of a relation between corporate performance and managerial compensation, there are hardly any cases where the research takes other factors, such as individual characteristics of the CEOs, into account. For example, does a higher level of corporate performance lead to higher managerial compensation, or is managerial compensation driven by some underlying and omitted factor? For instance are young men better managers and do they therefore receive a higher compensation? Does higher education positively influence managerial compensation? Does the threat of being replaced influence compensation? And last, do managers with more experience in the firm earn more salary and bonuses?
In estimating these relationships, endogeneity could possible affect the outcomes. They are likely to be biased and unlikely to quantify the magnitude of the economic effects. To address such difficulties, this paper uses data that focus on individual executives over time and between corporations. Optimal corporate performance should depend on changes in shareholder wealth, as these are affected by decisions made by CEOs, actions of other board members and employees, and demand and supply conditions. In addition several individual characteristics of the CEOs are estimated that potentially influence the level of wealth for executives. For instance, individuals with higher educational degrees are typically viewed as having more knowledge. If a master’s degree or a doctorate positively influences the CEO’s ability to make better decisions for his or her corporation, one way to increase CEO wealth is to earn a doctorate. Another way through which the corporation could increase the level of wealth of its executives is by letting them serve in this position for a longer period. Finally, if a CEO is expected to leave his or her position when he or she is close to retirement, a higher compensation is received.
In summary, the primary hypotheses are that higher levels of shareholder wealth should lead to higher salary and bonuses, and higher value of inside shareholdings. Higher education leads to higher pay, males earn more, increases in age and tenure increase compensation and the threat of being replaced lowers the compensation given to executives. The aim of the paper is that corporate performance (as measured by shareholder wealth) and CEOs’ individual characteristics are all likely to influence executives’ salary, bonus and value of shareholdings. Accordingly, ordinary least square regressions, lagged variables, and entity fixed effects are used to estimate the effects of all variables on the level of managerial compensation.
7
characteristics can either change or remain constant, and so will shareholder wealth and certain individual characteristics of the executives. Table 1 provides a summary of all expectations of the independent and dependent variables. These expectations are based on prior literature.
Table 1
Summary expectations of relationships between the dependent and independent variables.
Managerial compensation consists of the sum of salary and bonus and the market value of inside stock ownership.
Number Expectation
1 Shareholder wealth is positively related to managerial compensation 2 Age is positively related to managerial compensation
3 Gender: being male is positively related to managerial compensation Gender: being female is negatively related to managerial compensation 4 Replacement is negatively related to managerial compensation
5 Educational degree is positively related to managerial compensation 6 Tenure is positively related to managerial compensation
2.3 Additional contributions to the literature
Some contributions are outlined in the introduction and in the above. By including shareholder wealth, age, gender, educational background, tenure and threat of being replaced, the effect on CEOs’ performance-based salary, bonus and value of shareholdings are considered.
First, with the exception of Jensen and Murphy (1990a), no prior studies have been found that consider shareholder wealth as a measure of corporate performance in relation to managerial compensation. The authors obtained results that are inconsistent with the implications of the agency theory, which concludes that if shareholders had complete information regarding all the opportunities of a firm, a contract specifying and enforcing the executives to implement actions that increase shareholder wealth could be designed. The current paper also includes the effect of shareholder wealth to consider the effect on managerial compensation. Second, almost none of the prior empirical studies on this topic use individual CEO characteristics in the regressions to measure the effect of each of these on the level of CEO compensation. Such variables include educational degree, gender, tenure and threat of being replaced. By including them, a more precise measure of the effects on compensation can be estimated.
8
3. Data and methodology 3.1 Data description
In this section the sample selection will be discussed. Furthermore the different variables used in this study will be explained and summary statistics will be reported.
3.1.1 Sample selection
The dataset used in this paper is estimated by following CEOs published in Forbes’ surveys on America’s
highest paid Chief Executives. Every year Forbes ranks the 500 highest paid CEOs. Data on all 1,029 CEOs published in Forbes from 2006 to 2012 is reported. This period is chosen because of data availability constraints. Data on executives are not available in Forbes before 2006, and performance data for most companies is yet not available for 2013. Furthermore, 2006 to 2012 is considered to be a reasonable time period because it includes information on more than 700 CEOs in more than 1,000 corporations. Additionally, the data includes the global financial crisis with its market collapse in 2008, thus enabling observation of how managerial compensation related to corporate performance during the years of distress.
The sample for this study includes executives working in 737 corporations. Forbes provides information for each executive relating to age, one-year salary and bonuses, five-year salary and bonuses and on the market value of inside share ownership. These data on compensation were matched with corporate performance data obtained via Orbis. This data source reports corporate data, both on fiscal year performance of the respective company and individual characteristics for the CEO. These characteristics include gender, appointment date and educational degree.
For the purpose of this study the sample needs to contain information on all of these characteristics. In some cases Orbis fails to provide information on CEOs who are not currently serving as the executive for the company. For these executives information is obtained via Bloomberg Businessweek. By combing the CEOs data and information on their compensation from Forbes, data on fiscal year performance and individual characteristics reported by Orbis, and completing missing data from Bloomberg Businessweek, a unique data set is created for this study. After eliminating observations for which information on some variables is still missing, the final sample includes 413 executives working in 317 companies.
3.1.2 CEOs’ individual characteristics
9
and therefore also their pay, this variable is also included in the current research. Also based on prior literature, increases in age and tenure are expected to increase salary. Therefore these variables are also included. For instance, Berger et. al (1997) argue that executives with higher cash compensation and longer tenures are more likely to be entrenched and therefore to do their best to increase shareholder wealth. Jalbert et al. (2004) found that those CEOs who have a doctorate earn more than those with other degrees, and those with no graduate degree (for instance a bachelor’s degree) or a law degree earn less than others. This evidence thus suggests that skills earned through an advanced graduate degree receive higher managerial compensation Not all obtained variables of this study can be applied directly. Gender is naturally reported as either male or female, and can therefore be used directly to measure the differences in managerial compensation. On the other hand, appointment date does not have much influence on the findings. It is more important to use tenure as a variable. Therefore, tenure is measured as the number of months an executive worked for the same company. This is measured as the number of months between the appointment date and the 31st of December in the year in which the compensation is earned. Educational degree is classified into four groups according to the achieved degrees. These are bachelor’s degree, master’s degree, law degree and doctoral degree. The data are classified in these groups since the degrees mentioned in Orbis and Bloomberg business week are not reported in a systematic way. Sometimes all information on education is reported for a CEO, even including the educational institution that awarded the degree. However, in most cases only one of the four classes is reported. For this study, higher pay is expected for those who earned a master’s or doctoral degree and lower pay is expected for lower levels of education and for specific degrees which are not really related to management, such as a law degree.
10
Table 2Summary statistics
Data on executives from 2006 to 2012 are from Forbes, Orbis and Bloomberg Businessweek. Data is on number of CEOs and companies. The CEOs average age is in years and average tenure is in months. The percentages of CEOs who have either a Master’s, Bachelor’s, Doctorate or Law degree are reported. Also the percentage of male and female executives in this study is shown.
Variable Number/percentage Number of CEOs 413
Number of companies 317 Average age in years 56 Average tenure in months 79 Percentage of CEOs with a master's degree 66% Percentage of CEOs with a bachelor’s degree 17% Percentage of CEOs with a doctorate degree 6% Percentage of CEOs with a law degree 11% Percentage of male CEOs 98% Percentage of female CEOs 2%
As can be seen in the above table, this study examined more than 400 CEOs. More than 300 companies were included. Since some CEOs served in companies for several years, data for over 1,300 estimates are used. The average age of the executives is 56, which is considered to be a reasonable age to have gained knowledge and experience in the position as a CEO. This age is also considerably remote from the normal retirement age. The average tenure is 79 months, which is about 6.5 years. As previously explained, compensation is expected to increase with tenure. As to the educational background, most executives have a master’s degree, followed by a bachelor’s degree, even though the latter number is much smaller, only 17%. There are only a few executives who have a doctorate (6%) and only slightly more have a law degree. As stated previously, a law degree is considered as a specific degree not related to management, and executives with this type of degree are therefore expected to earn less. Almost all CEOs examined in this study are male (98%).The difference in percentages between males and females is probably due to the fact that in the USA still most managerial positions are held by men. Nevertheless, a difference in pay between the genders can be observed, although the difference is probably not of economic relevance considering the small sample of women relative to men.
3.1.3 Managerial compensation
11
Forbes uses the end of the year market value of shares owned and measures this as the beginning-of-period
value of the shares multiplied by the inflation adjusted rate of return on common stock.
Graph 1 displays an overview of average managerial compensation for CEOs used in this study. The bars consist of two components. The first is the average of the sum of salary and bonus; and the second is the average market value of shares owned. The bars represent annual averages of the total managerial compensation.
Graph 1
Compensation amounts
The graph displays the annual average amounts of the sum of salary and bonus and the market value of shares owned. All amounts are obtained via Forbes and are in millions of dollars. The combined amounts form the total managerial compensation in the respective year.
As can be seen in the graph, the average total managerial compensation varies on an annual basis. During the sample period; the minimum amount of average total compensation is less than $150 million in 2009 and 2010 and the maximum is over $250 million in 2012. The compensation in salary and bonus is very small compared to the value of shares owned. This value is between $9 million and $15 million, and this is between 4% and 9% of the total compensation of each year. These small percentages are consistent with the findings of Murphy (1985) who reports that the value of shareholding often exceeds levels of salary and bonus compensation by orders of magnitude. As also displayed in the graph, the value of salary and bonuses does not change significantly when the value of shares increase or decrease. This component of compensation is therefore said to be relatively stable compared to the value of shares owned. Therefore, overall, the result of the differences in total compensation is mainly a result of fluctuations in the value of shares owned.
In addition to graph 1, descriptive statistics are displayed in table 3. This table reports information on the mean, standard deviation, minimum, median and maximum value of the sum of salary and bonus, the value of
0 50 100 150 200 250 300 2007 2008 2009 2010 2011 2012
12
inside share holdings and shareholder wealth. In addition the value of shares owned are reported as percentages of the value of all shares outstanding and are separately reported for small and large firms.
Table 3
Descriptive statistics on managerial compensation
The mean, standard deviation, minimum, median and maximum value of the sum of salary and bonus, the market value of shares owned and shareholder wealth are reported. The values are in millions of dollars. The value of shares owned by CEOs are reported as a percentage of the value of total shares outstanding, by firm size. Small firms are defined as having a market value below the sample median ($8.9 billion). Large firms have a market value above the sample median. The value of shares owned is in millions of dollars. The value of shareholder wealth is reported in billions of dollars.
Mean Standard deviation 25th percentile 50th percentile 75 percentile Pay ($mil) 11.673 17.958 3.963 6.690 13.390 Value shares owned
($mil) 182.227 1175.969 6.700 13.450 30.575
Value of shares owned as percentage of value of shares outstanding
All firms 0.331% 3.186% 0.041% 0.136% 0.423% Small firms 0.449% 3.864% 0.043% 0.158% 0.484% Large firms 0.233% 2.489% 0.041% 0.122% 0.362%
13
revenue (annual turnover) below the sample median ($8.9 billion). Large firms have an annual turnover above the sample median.
Table 3 also presents an overview of the value of shares owned by executives, by firm size and also as a percentage of the value of total shares outstanding, by firm size. The executives of all firms hold an average value of 0.333% of the value of all shares outstanding. Once again, the distribution is skewed; the median executive holds only 0.136% of the firm’s stock. Twenty-five percent of the executives hold only 0.041%. CEOs of large firms tend to hold on average a smaller fraction of total shares. The value of shares owned is presented in dollar terms. Executives hold an average value of shares of $182.2 million. The median value is only $13.5 million; therefore once again the distribution is skewed. The average value of shares owned in small firms is much lower than for large firms, $63.6 million compared to $280.2 million. In consistence with Jensen and Murphy (1990a), the executives in large firms tend to have larger dollar investments in their firms’ shares while owning a smaller fraction of their firms’ shares.
3.1.4 Corporate performance
Corporate performance is measured by shareholder wealth. This measure of performance is chosen to examine the relationship between decisions made, actions undertaken by executives, and objectives of shareholders. Usually the main objective of shareholders is to increase shareholder wealth. To achieve this, an executive should undertake actions and make decisions that increase this measure of corporate performance. Therefore a positive relation is expected between shareholder wealth and managerial compensation.
14
3.1.5 Threat of being replaced
An executive is seen as replaced when another executive takes over his or her position in the company in
Forbes’ next year report on executive compensation. When management performance is poor, the threat of
management replacement increases. As stated before, Jensen and Murphy (1990a) found an inverse relation between the threat of being replaced and managerial compensation. Therefore executives are disciplined by this threat. Besides, executives who are close to retirement are more likely to turnover. However, executives are more likely to be fired when they are young. The threat of being replaced is included in this study, since previous studies found an inverse relation between the threat of being replaced and pay. This means that when an executive is expected to leave the firm in the next fiscal year, this will lower his or her payment. However, these previous studies were not entirely sure of whether the executive had actually left in the following year; this study replicates the effect of the threat of being replaced for executives who are replaced in the Forbes survey of the next year and therefore without doubt were replaced. A CEO can leave the firm, be fired, retire or die. In all these cases, the CEO is replaced by another CEO. Therefore the term being replaced is used. There is only a small fraction of executives who were replaced in the consecutive year (7%). Therefore 93% are reported in Forbes’ next year list on executive compensation for that same company. Table 4 reports replacements of CEOs, grouped according to age.
Table 4
Number of CEOs and replacements
CEOs are grouped according to age. Replaced is when another executive takes over his or her position in the company in Forbes’ next year report on executive compensation.
Full sample Less than 50 Between 50 and 55 Between 55 and 60 Between 60 and 64 64 years or older Sample size 1326 171 258 363 217 61 Replaced 97 4 12 29 24 28
The full sample of CEOs contains 1,326 men and women. Most of these executives are between 50 and 64 years of age. When a CEO is older than 55 it is more likely that he or she will have retired in a consecutive fiscal year. Young executives are more likely to be replaced than executives of 55 years and older. However, table 4 provides evidence that only about one sixth of all CEOs who had left the company in the space between two Forbes surveys were younger than 55.
3.2 Methodology
This section discusses the methodology of the study. The models that are used will be explained and the ordinary least squares regressions will be described.
15
3.2.1 Research design
As discussed in previous chapters, managerial compensation consists of the sum of salary and bonus and the market value of inside share ownership; corporate performance is measured by shareholder wealth. In this study the effect on managerial compensation is measured by corporate performance, age, gender, tenure, the threat of being replaced and educational degree. The research is divided into two main models; namely one for the sum of salary and bonus compensation and one for compensation which consists of the sum of salary and bonus and the market value of inside share ownership. Therefore both components of managerial compensation can be considered apart from one another. An advantage of estimating these two models is that both components of compensation are expected to be differently related to the principle-agent problem and their magnitudes can now be considered apart from one another. Next to the two main models, two additional models are estimated by panel data techniques. These models show entity fixed effects to correct for omitted variables that are constant over time but can fluctuate between entities. This technique is applied since the sample used contains executives who are reported several times in the period between 2006 and 2012. For each year the information on corporate performance, managerial compensation, age, tenure and threat of being replaced differs. However, naturally, the information on education and gender stay the same.
3.2.2 Specification of sum of salary and bonus model
In order to test the effect of shareholder wealth, age, gender, tenure, the threat of being replaced and educational degree an ordinary least squares regression equation is estimated. As previously elaborated, shareholder wealth in year t-1 is used to measure the previous year’s stock price and corporate performances effect on the sum of salary and bonus. The least squares regression becomes:
CEO (salary + bonus)i,t = α + β1(shareholder wealth)t-1 + β2(age)i,t+ β3(gender)i + β4(replacement)i,t+
β5(educational degree)i + β6(tenure)i,t, + ei (1)
in which the subscript t denotes time and i denotes the observation for the executive. A is the constant and β are parameters to be estimated. Salary and bonus is the sum of both components since Forbes only reports the sum. The dependent variable is measured in millions of dollars. Based on Jensen and Murphy (1990a) shareholder wealth is defined as rtVt-1 , where rt is the for inflation adjusted rate of return on common stock in
fiscal year t. Vt-1 isthe firm value at the end of the previous year. Shareholder wealth is measured in billions of
16
3.2.2 Specification of sum of salary and bonus and the value of share ownership model
The effect on total compensation (sum of salary and bonus and the market value of inside share ownership) is measured in the same way as the previous ordinary least squares regression, with the exception that now one more component of managerial compensation is added. This ordinary least squares regression thus becomes: CEO (salary+ bonus)i,t + value of shares ownedt = α + β1(shareholder wealth)t-1 + β2(age)i,t+ β3(gender)i +
β4(replacement)i,t+ β5(educational degree)i + β6(tenure)i,t, + ei (2)
in which the explanations for the variables in the previous least squares equation are the same. Only value of shares ownedt is different. These values are obtained from Forbes, which uses the end of the year market
value of shares owned and measures this as the beginning-of-period value of the shares multiplied by the inflation adjusted rate of return on common stock.
4. Results
In this section the outcomes of the ordinary least squares regressions will be discussed. The effects on the sum of salary and bonus will be discussed first, followed by the effects on total compensation. Finally, the study will emphasize the effect of all variables with panel data featuring entity fixed effects. Here, gender and educational background are no longer considered in the estimation of the equation.
4.1 Results of sum of salary and bonus model
17
Table 5Estimates of the sum of CEO salary and bonus and several personal characteristics.
Coefficients of ordinary least squares regression of the sum of CEO salary and bonus, shareholder wealth in year t-1 in billions of dollars, educational degree, age, gender, tenure and replacement. The independent variable for replaced is equal to one if the executive is serving in his last full fiscal year and zero otherwise. The independent variable of gender is equal to one if the executive is male and zero if she’s female. For education more dummies are created for the four different types of education that are used. These are; bachelor’s, master’s, doctoral and law degrees. Tenure is defined as the number of months a CEO this position in its company. ***, **, and * indicate statistical significance at respectively the 1%, 5%, and 10% levels.
Dependent variable ($mil) Sum of CEO salary and bonus
Independent variable Coefficient t-Statistic Prob.
Intercept -7.670 -1.366 0.172
Shareholder wealth in t-1 ($bill) 0.002 3.840 0.000*** Bachelor's degree -2.325 -1.259 0.208 Master's degree -1.381 -0.900 0.368 Doctoral degree 4.746 1.977 0.048** Male -0.593 -0.192 0.848 Age in years 0.244 2.696 0.007*** Tenure in months 0.081 8.558 0.000*** Replaced 1.110 0.598 0.550 R-squared 0.097 F-statistic 17.631 Prob(F-statistic) 0.000***
The first least squares regression is estimated in table 5. The coefficient for shareholder wealth in year t-1 is statistically significant at the 1% level, meaning that this variable influences the sum of salary and bonus. When last year’s shareholder wealth increases by $1 million, an executive receives an average pay increase of $2.00. This increase is lower per $1 million increase in shareholder wealth than Jensen and Murphy (1990a) found for their sample; which is an increase of $3250 in managerial compensation for the period 1974-1986. It seems that the pay-performance relationship has declined in the current sample.1 Jensen and Murphy considered that an increase of $3250 in compensation for every $1 million increase in shareholder wealth was too weak to serve as an incentive for executives to act in the shareholders’ interest. For educational degrees dummy variables are created for having bachelor’s, master’s, doctoral and law degrees. The last dummy is excluded in the equation. No significant evidence is found for the coefficients of a bachelor’s degree or a master’s degree. However, a doctorate significantly increases a CEO’s salary and bonus compensation by $4.7 million per year compared to having a law degree. This finding is consistent with prior literature, which suggests that executives with a higher degree are expected to earn more Therefore, expectation number five is confirmed, and educational degree is positively related to total managerial compensation. There is no evidence for a significant
1
18
effect of gender and the threat of being replaced. That means that being male does not significantly decrease managerial compensation and the threat of being replaced does not significantly increase managerial compensation. These relationships are the opposite of what is expected. Age is significantly and positively related to managerial compensation. As an executive grows older, he or she receives on average $0.2 million more in salary and bonus per year. This is consistent with expectation two; age and compensation is positively correlated. Tenure also significantly increases managerial compensation, on average by $0.1 million for each month he or she works for the company. This is also consistent with prior literature. The coefficient for the intercept is not statistically significant and is therefore no evidence that managerial compensation is negative when the other variables are zero.
4.2 Results of sum of salary and bonus and the value of share ownership model
19
Table 6Estimates of the sum of salary and bonus, and share ownership on shareholder wealth and several Individual characteristics.
Coefficients of ordinary least squares regression of total managerial compensation (CEO salary, bonus and the market value of inside stock ownership), shareholder wealth in year t-1 in billions of dollars, educational degree, age, gender, tenure and threat of replacement. The independent variable for replacement is equal to one if the executive is serving in his last full fiscal year and zero otherwise. The independent variable ofgender is equal to one if the executive is male and zero if she’s female. For education more dummies are created for the four different types of education that are used. These are bachelor’s, master’s, doctoral and law degrees. Tenure is defined as the number of months a CEO worked for the company. ***, **, and * indicate statistical significance at respectively the 1%, 5%, and 10% levels.
Dependent variable ($mil) Total managerial compensation
Independent variable Coefficient t-Statistic Prob. Intercept 1448.483 3.832 0.000*** Shareholder wealth in t-1 ($bill) 0.095 3.602 0.000*** Bachelor's degree -20.633 -0.166 0.868 Master's degree 126.619 1.226 0.221 Doctoral degree 197.246 1.220 0.223 Male 218.731 1.051 0.294 Age in years -33.218 -5.457 0.000*** Tenure in months 3.351 5.241 0.000*** Replaced 30.417 0.243 0.808 R-squared 0.044 F-statistic 7.620 Prob(F-statistic) 0.000***
20
one month, total managerial compensation is expected to increase with $3.4 million. The tenure’s coefficient is statistically significant. This finding is consistent with Brookman and Thistle (2009), and McKnight and Tomkins (2004) who state that performance and compensation increases with tenure for CEOs.
4.3 Results with panel data fixed effects
In this study data on executives are reported several times in the period of 2006 to 2012. For each year the information on corporate performance, managerial compensation, age, tenure and threat of being replaced differs. However, naturally, the information on education and gender stay the same. Panel data techniques with entity fixed effects are used for this study to correct for omitted variables that are constant over time but are allowed to fluctuate between entities. The two ordinary least squares regressions with panel data entity fixed effects are different from the previous least squares regressions in the independent variables. Gender and educational background are both fixed effects for the executives over time. Therefore, these variables are not present in the fixed effects estimations. Naturally, age and tenure increase every year, and the threat of being replaced can change every year and these variables are therefore still present in the estimations with fixed effects. Table 7 reports the coefficients of the least squares regressions one and two with fixed effects.
Table 7
Coefficients of ordinary least squares regressions with panel data entity fixed effects.
Estimated coefficients on the sum of CEO salary and bonus. Independent variables are shareholder wealth in year t-1, age, tenure and threat of being replaced. Tenure is defined as the number of months a CEO worked for the company. The dependent variable for replaced is equal to one if the executive is serving in his last full fiscal year and zero otherwise. ***, **, and * indicate statistical significance at respectively the 1%, 5%, and 10% levels.
Dependent variable ($mil) Sum of CEO salary and bonus
Independent variable Coefficient t-Statistic Prob.
Intercept 5.789 0.073 0.942
Shareholder wealth in t-1 ($bill) 0.002 2.259 0.024**
Age in years 0.075 0.047 0.963 Tenure in months 0.011 0.083 0.934 Replaced -2.423 -1.065 0.287 R-squared 0.624 F-statistic 3.624 Prob(F-statistic) 0.000***
21
value as for the first ordinary least squares regression. Again, the relationship is too weak to provide incentives for executives to act in the shareholder’s best interest. This estimation positively influences the R-squared, which indicates that this regression with entity fixed effects better fits the data than the previous estimations.
Table 8
Coefficients of ordinary least squares regressions with panel data entity fixed effects.
Estimated coefficients on the sum of CEO salary and bonus, and the market value of inside shares owned. Independent variables are shareholder wealth in year t-1, age, tenure and threat of being replaced. Tenure is defined as the number of months a CEO worked for the company. The dependent variable for replaced is equal to one if the executive is serving in his last full fiscal year and zero otherwise. ***, **, and * indicate statistical significance at respectively the 1%, 5%, and 10% levels.
Dependent variable ($mil) Total managerial compensation
Independent variable Coefficient t-Statistic Prob.
Intercept 52.040 0.023 0.982
Shareholder wealth in t-1 ($bill) -0.017 -0.673 0.501
Age in years 1.113 0.024 0.981 Tenure in months 1.138 0.295 0.768 Replaced -33.140 -0.508 0.612 R-squared 0.928 F-statistic 28.063 Prob(F-statistic) 0.000***
Table 8 provides the coefficients of the second least squares regression, this time with panel data with entity fixed effects applied to the regression. None of the coefficients are statistically significant. Therefore none of the variables can explain average increases or decreases in total managerial compensation. The relationships can however be discussed. According to this estimation, the relationship between shareholder wealth in year
t-1 and total compensation is negative. This is in contradiction to the previous results. As expected, the
relationships between age and tenure, and total managerial compensation are positive, and the relationship between the threat of being replaced and compensation is negative. The high R-squared indicates that this regression almost perfectly fits the data.
5. Conclusion
22
The analysis of performance pay and management characteristics for over 1,000 CEOs appearing in Forbes lists in 2006 to 2012 is used to answer the research questions: Are corporate performance and CEO’s age,
education and tenure positively related to managerial compensation? (i), and do a CEO’s threat of being replaced and being female have a negative effect on managerial compensation? (ii).
The results indicate that the relation between managerial compensation and shareholder wealth is small. On average each $1 million change in last year’s shareholder wealth corresponds to an increase in this year’s salary and bonus of about $2.00. The executive’s total pay, defined as the sum of salary and bonus and share ownership, changes by about $95.00 per $1 million increase in last year’s shareholder wealth. The compensation generated by the value of inside shareholdings is therefore large relative to the value of salary and bonus. Consistent with findings by Jensen and Murphy (1990a) this both increases in compensation for increases in shareholder wealth do not provide executives with incentives to act in the best interest of shareholders. Having a doctoral degree significantly increases a CEO’s average salary and bonus compensation by $4.7 million per year compared to having a law degree. Age is significantly positively related to managerial compensation. When an executive is one year older, he or she receives on average $0.2 million more salary and bonus per year. Tenure also significantly increases managerial compensation, on average by $0.1 million for each month he or she works for the company. Total managerial compensation is expected to decrease on average by $33.2 million when an executive’s age increases by one. When tenure increases by one month, total managerial compensation is expected to increase with $3.4 million.
The largest increase in managerial compensation is connected to the difference in the value of shares owned by executives, but such holdings are small for the majority of CEOs, especially in large firms. Median inside shareholdings for reported CEOs are less than 0.42 percent, and 75 percent of these executives hold 9.78 percent of their firm’s shares. Median ownership for executives of small firms is 0.48 percent and for large firms it is 0.36 percent.
Overall, the hypotheses that higher shareholder wealth should lead to higher salary and bonuses and higher value of inside shareholdings, that higher education leads to higher pay, that males earn more, and that the longer one holds the position as executive the more one gets paid are confirmed. The aim of this research is that managerial incentives and characteristics are both likely to influence the executives’ salary, bonus and value of shareholdings.
To reduce the principle-agent problem the executive should be content with the increase in compensation he or she receives for an increase in shareholder wealth. Naturally, an executive would certainly be content when receiving 100% of the increase in shareholder wealth (i.e. b=1). A suggestion for further research is to investigate at what percentage a CEO is on average content with the extra compensation received from an increase in shareholder wealth.
23
something that executives regularly do in order to increase or decrease the effect of the value of shares on their compensation. This buying and selling of shares might even be an indication of good or bad management decisions. In future research other measures of an executive’s performance can be included. This could for instance apply to accounting measures of corporate performance and measures of relative performance based on a benchmark.
6. References
Abowd, J., & Bognanno, M., 1995. International differences in executive and managerial compensation. In F. Richard, & L. Katz, Differences and changes in wage structures. Chicago: University of Chicago Press. Barkema, H., & Pennings, J., 1996. Top management pay: Impact of power and influence. Unpublished
manuscript.
Becker, G., 1971. The economics of discrimination. Chicago: University of Chicago Press.
Berger, P., Ofek, E., Yermack, D., 1997. Managerial entrenchment and capital structure decisions. Journal of
Finance 52, 1411-1438.
Brookman, J., Thistle, P., 2009. CEO tenure, the risk of termination and firm value. Journal of Corporate ` Finance 15, 331-344.
Brunello, G., Graziano, C., & Parigi, B., 1996. Executive compensation and firms performance in Italy. Working paper 4-96, University of Udine.
Coles, J., Daniel, N., Naveen, L,. 2006. Managerial incentives and risk-taking. Journal of Financial Economics 79, 431-468.
Conyon, M., Gregg, P., & Machin, S., 1995. Taking care of business: Executive compensation in the UK. Economic Journal, 105, 704-714.
Core, J., Guay, W., 1999. The use of equity grants to manage optimal equity incentive levels. Journal of Accounting and Economics 28, 151-184.
Crystal, G. S., 1992. In search of excess. New York: W. W. Norton.
Diamond, D., Verrechia, R., 1982. Optimal managerial incentives and equilibrium security prices. Journal of Finance 37, 275-287.
Elkinawy, S., Stater, M., 2011. Gender differences in executive compensation: Variation with board gender composition and time. Journal of Economics and Business 63, 23-45.
Gomez-Mejia, L., 1994. Executive compensation: A reassessment and a future research agenda. In Research in Personnel and Human Resources Management, vol. 12. Greenwich: JAI Press.
Guay, W., 1999. The sensitivity of CEO wealth to equity risk: an analysis of the magnitude and determinants. Journal of Financial Economics 53, 43-71.
Hall, B., Liebman, J., 1998. Are CEOs really paid like bureaucrats? Quarterly Journal of Economics 113, 653-691.
24
Jensen, M., & Meckling, W., 1976. Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3, 305-360.
Jensen, M., Murphy, K., 1990a. Performance pay and top-management incentives. Journal of Political Economy 98, 225-264.
Jensen, M., & Murphy, K. J., 1990b. CEO incentives - it's not how much you pay but how. Harvard Business Review, 68, 138-149.
Ju, N., Leland, H., Senbet, L., 2002. Options, option repricing and severance packages in managerial compensation: their effects on corporate risk. Working paper. University of Maryland.
Kaplan, S., 1994a. Top executives, turnover and corporate performance in Germany. Journal of Law, Economics and Organization, 10, 142-159.
Kaplan, S., 1994b. Top executive rewards and corporate performance: A comparison of Japan and the United States. Journal of Political Economy, 102, 510-546.
Kato, T., Long, C., 2006. CEO turnover, corporate performance, and enterprise reform in China: Evidence from micro data. Journal of Comparative Economics 2006, 796-817.
Leonard, J., 1990. Executive pay and firm performance. Industrial and Labor Relations Review, 43, 13-29. Low, A., 2009. Managerial risk-taking behavior and equity-based compensation. Journal of Financial
Economics 92, 470-490.
McConnell, J., Servaes, H., 1990. Additional evidence on equity ownership and corporate value. Journal of Financial Economics 27, 595-612.
McKnight, P., Tomkins, C., 2004. The implications of firm and individual characteristics on CEO pay. European management journal 22, 27-40.
McClelland, P., Barker, V., Oh, W., 2012. CEO career horizon and tenure: future performance implications under different contingencies. Journal of Business Research 65,1387–1393
Meyerson, E., 1994. Human capital, social capital and compensation: The relative contribution of social contacts to managers' incomes. Acta Sociologica, 37, 383-399.
Morck, R., Shleifer, A., Vishny, R., 1988. Management ownership and market valuation: an empirical analysis. Journal of Financial Economics 20, 293-315.
Murphy, K., 1985. Corporate performance and managerial remuneration. Journal of Accounting and Economics 7, 11-42.
Murphy, K., 1999. Executive compensation. In: Ashenfelter, O., Card, D. (Eds.), Handbook of Labor Economics, vol. 3b (Chapter 38). Elsevier Science, North Holland, pp. 2485-2563.
O'Reilly, C., Main, B., & Crystal, G. (1988). CEO compensation as tournament and social comparison: A tale of two theories. Administrative Science Quarterly, 33, 257-274.
Perry, T., Zenner, M., 2000. CEO compensation in the 1990s: shareholder alignment or shareholder expropriation? Wake Forest Law Review 35, 123-152.
Rosen, S., 1992. Contracts and the market for executives. In L. Werin,& H. Wijkander, Contract economics. Oxford: Basil Blackwell.
25
Weisbach, M., 1988. Outside directors and CEO turnover. Journal of Financial Economics 20, 431-460. Appendix A - Pay-performance sensitivity in different times periods
Table A.1
Coefficients of the pay-performance sensitivity in 1934-38 versus 1974-86 and 2006-12.
Included variables are the change in salary and bonus and the change in shareholder wealth at time t and the change in shareholder wealth at time t-1.Sample size in 1934-38 is 427 CEOs, in 1974-86 427 and in 2006-12 413.***, **, and * indicate statistical significance at respectively the 1%, 5%, and 10% levels. Results of 1934-38 and 1974-86 are obtained from Jensen and Murphy (1990a).
Dependent variable ($mil)
Change of sum CEO salary and bonus Independent variable 1934-38 1974-86 2006-12
Intercept 6.300 22.300 0.626
Change in shareholder wealth ($mil) 0.114 0.012 0.000 t-Statistic (5.6)*** (7.0)*** (1.3) Change in shareholder wealth in year t-1 ($mil) 0.061 0.007 0.000 t-Statistic (2.8)** (4.4)*** (2.0)**
R-squared 0.070 0.017 0.004
