Agency theory and America’s best paid executives

(1)

Agency theory and America’s best paid executives

Name: Gerwin Peters Student number: 5893593

E-mail: peters.gerwin@gmail.com Date: August 2015

Specialization: Economics & Finance Field: International Economics Labor Economics

Thesis coordinator:

Dr. Silvia Dominguez-Martinez Assigned supervisor:

(2)

1 Introduction

In America, executive compensation has been a popular topic in academic circuits for years, but only recently it has been in the sight of public policy makers. That is mainly caused by a growing amount of research papers, which have demonstrated the influence of compensation policies and by the fact that many policy makers and members of the public woke up in the aftermath of the financial crisis of 2008 (CNN, 2011). Due to the growing public interest, there are more intense debates about the fairness of the remuneration policies nowadays. In order to contribute to this discussion, this study will try to identify the determinants that explain these compensation amounts paid to America’s top executives. More specific, it focuses on America’s best paid executives in 2012, according to the Forbes (2012) list: ‘America’s highest paid Chief Executives’. Because the data we obtained from Forbes has not been used in academic literature so far, this thesis contributes to the discussion with analyses based on recent and factual numbers on CEOs their remunerations. However, due to time restrictions and the scope of the bachelor thesis, the sample data we used is too small to be of significant influence in the academic literature.

The research starts by analyzing the existing literature. A large amount of academic papers written on this subject is already available. According to the academic literature, two opponent views can be distinguished. The first and oldest – called the Agency Theory – suggests that variable compensation depends on firm performance. While the second – called Managerial Power Theory – disagrees to this statement and identifies the high pay-offs to top executives as a result of bad governance (Bertrand and Mullainathan, 2000; Tosi et al., 2000). This paper will test which of the two opponent views can explain the latest pay-offs made to America’s best paid executives. In order to do so, the following research questions have been proposed:

Can agency theory explain the compensation paid to America’s best paid executives?   In addition, is there a positive relationship between executive’s compensation and firm’s performance?

In order to answer these questions, the thesis will test the effect of annualized total return and years of tenure on the different variable pay-offs made. When the results show a strong relationship between annualized total return and the variable pay-offs made to the executives, the results suggest firm-performance does determine the variable pay-offs and Agency Theory best explains the compensation paid to America’s best paid executives. But if the results show a strong relationship between years of tenure and executive compensation, while showing a weak relationship between firm-performance and executive compensation, Managerial Power Theory

(4)

best explains the pay-offs. Tournament Theory does not expect total compensation to be directly based on firm performance. In Tournament Theory the pay-offs to executives are used as incentives for lower level managers to perform better, this way increasing firm-performance. So, when Agency Theory and Managerial Power Theory fail to explain the pay-offs in this thesis, Tournament Theory could be the explanation. However, if this is the case, further research is needed to prove that Tournament Theory is the explanation.

2 Literature review

This thesis will start by describing and explaining the existing literature on the matter of executive compensation. Like mentioned before, many papers are written on the topic of executive compensation. This thesis focuses on the papers which tried to define the effectiveness of the executive’s remuneration policies.

2.1 Two competing views

In the field of executive compensation theory, two competing research perspectives can be distinguished. In the Agency Theory perspective, compensation is used to solve the moral hazard problem. But in the Managerial Power Theory perspective, this compensation is a direct result of the agency problem (Bertrand and Mullainathan, 2000; Tosi et al., 2000).

During the sixties and seventies, economists as Arrow and Wilson explored risk sharing among individuals and groups. This literature described the risk-sharing problem as one that arises when cooperating parties have different attitudes toward risk (Eisenhardt, 1989).  

Ross (1973) broadened this risk-sharing literature by identifying the agency relationship. This relationship arises between two (or more) parties when one called the agent acts on behalf or as a representative for the other called the principal in a particular domain of decision problems. There are many examples of agency arrangements, but essentially all contractual arrangements between employer and employee are agency arrangements (Ross, 1973). More specific for Economic Theory, Jensen and Meckling (1976) defined an agency relationship as a contract under which the principal (shareholder) engages the agent (executive) to perform some service on his behalf which involves delegating some decision making authority to the agent. But if both parties to the relationship are utility maximizers, there is good reason to believe that the agent will not always act in the best interests of the principal. The principal can limit divergences from his interest by establishing appropriate incentives for the agent (bonding costs) and by incurring monitor costs designed to limit the aberrant activities of the agent. The bonding and monitoring costs together with the divergence in agent’s decisions and those

(5)

decisions which would maximize the welfare of the principal, are defined as the agency costs (Jensen and Meckling, 1976).

The agency problem becomes more clear in the context of the firm and separation of ownership and control. Because typically, executives own very little of the firm they control, shareholders are facing a moral hazard problem. The executives may not always maximize firm value by making their decisions (Bertrand, 2000). Shareholders use remuneration policies to reduce this moral hazard and try to align the shareholders’ and executive’s interest. According to Eisenhardt (1989), the focus of Agency Theory is to determine the most efficient contract governing the principal-agent relationship, given assumptions about people, organizations and information. If shareholders have complete information regarding the executive’s activities and the firm’s investment opportunities, the shareholders could design a contract specifying and enforcing the desired decisions to be taken. But managerial and investment opportunities are not perfectly observable by shareholders. So, in order to align executive’s and shareholders’ interest, the contract has to be designed in a way that executives become motivated to select and implement actions that increase shareholders’ wealth (Jensen and Murphy, 1990). This contract should include both short- and long-term incentive compensations that motivate the executive to put in effort in order to enhance firm performance and thus shareholders’ wealth (Jensen, 1983). However, individuals and executives tend to be risk averse. For this reason, the most suitable contract should include a fixed compensation as well (Baiman, 1990).

The competing view the Managerial Power Theory also starts with the separation of ownership and control, but argues that this same separation allows executives to gain control of the compensation setting process itself. By packing the board with their friends or any other mean of entrenchment, many executives set their own pay (Bertrand and Mullainathan, 2000). Managerial Power Theory argues that the greater the power of the executives is, the better they can extract rents from the organization. Therefore, executives are more motivated in increasing firm size than in enhancing firm performance. The result is that the executives now can decouple their compensation from firm performance (Tosi et al., 2000). The only constraints they face, are the unwillingness to draw the attention of shareholder activist groups or by the fear of becoming takeover target. However, within these constraints, they can pay themselves as much as possible (Bertrand and Mullainathan, 2000). Thus, whereas compensation in the perspective of Agency Theory is an attempt to solve moral hazard, it is in the perspective of the Managerial Power Theory the result of moral hazard.

2.2 Empirical evidence

If executive contracts are designed well, a strong relation between executive compensation and firm performance can be expected. After all, good firm performance is in shareholders’ interest,

(6)

thus when the remuneration contract is aligned, the executive should be rewarded for good firm performance.

So far, many papers have been written on the suggested relation between executive compensation and firm performance. One widely cited paper is the paper of Jensen and Murphy (1990). They based their study on 2,213 executive officers serving in 1,295 corporations listed in the Executive Compensation Surveys published in Forbes from 1974 to 1986. The purpose of Jensen’s and Murphy’s (1990) paper was to estimate the magnitude of the firm value-increasing incentives provided by salary revisions, performance based bonuses, stock options and performance based dismissal mechanisms. As a measure of firm performance, they used the change in shareholder wealth in thousands of 1986 dollar and regressed this on the different CEO wealth components. Jensen’s and Murphy’s (1990) study concluded that executive’s wealth increases by $3,25 for every $1000,00 increase in shareholder’s wealth. So, they highlighted a positive relation between executive compensation and firm performance, but the authors found that this result was of little magnitude.

Lambert et al. (1993) studied 303 large publicly traded U.S. firms in manufacturing and service markets that span many industrial sectors of the economy for the period 1982 to 1984. They tried to evaluate the ability of Tournament, Managerial Power and Agency theories to explain the observed compensation data. Their results concluded that 5,80% of the variance in executive compensation was explained by tournament model, 6,76% was explained by Managerial power theory and just 0,71% by the use of the Agency Theory. So, they found a much stronger support for the Managerial Power Theory than for the Agency Theory. However, they highlighted an alternative perspective on the Managerial Power Theory. In this perspective, shareholders allow the CEO to have long tenure or appoint board members because the shareholders believe that there are not many serious agency problems with this manager. Shareholders are willing to pay this manager a compensation premium for this trust, and managerial power does not have the types of undesirable consequences, which are typically assumed (Lambert et al., 1993).

Hall and Liebman (1998) also examined the relationship between executive compensation and firm performance. Their data set included information of 478 companies for the period 1980 to 1994. Again, most of the data was provided by the yearly Forbes’ surveys. But since the

Forbes’ data provided limited information on stock and stock option awards, they used proxy

statements to include detailed information of the amount of stock options granted during the year as well as the exercise prices and duration of the options. They are using the Black-Scholes formulae in order to value the stock and stock option grants. The results of Hall and Liebman (1998) differed from the conclusions of Jensen and Murphy (1990) and Lambert (1993). Hall and Liebman (1998) concluded a strong relationship between firm performance and executive compensation. They found that executive’s wealth often changes by millions of dollars for typical changes in firm value, a relationship generated almost entirely by changes in the value of executive holdings of stock and stock options. Hall and Liebman (1998) calculated the effect of the increase in firm value for executive compensation and found that executive’s wealth

(7)

increases with $5,29 on average for every $1000,- increase in firm value. But in order to compute the Jensen and Murphy statistics, Hall and Liebman (1998) had to incorporate the dismissal, performance sensitivity and make a few adjustments reflecting changes in salary and bonus. After making these adjustments their estimate of the Jensen and Murphy statistic rose to $6,00 for 1994. This is almost double of the $3,25, found by Jensen and Murphy. Although Hall and Liebman (1998) argue that a dramatic negative correlation exists between the Jensen and Murphy measure and the size of the firm. And since most of the firms in their sample increased in market value during the fifteen-year period, this Jensen and Murphy measure substantially underestimated the size-adjusted increase in sensitivity over time. So correcting for this error, the size-adjusted Jensen and Murphy statistic increased fourfold, rather than the almost doubling form $3,25 to $6,00 between 1980 and 1994. Hall and Liebman’s (1998) findings suggest that executive compensation is highly responsive to firm performance. But the relationship between pay and performance is almost entirely driven by changes in the value of stock and stock options. Stock and options revaluations increase median executive wealth by about $1,25 million in response to a 10 percent change in firm value. This is 53 times larger than the increase in bonus and salary solely. This means that for Hall and Liebman’s (1998) findings, stock and stock option revaluation account for about 98 percent of the relationship between executive’s pay and firm’s performance.

Himmelberg et al (1999) based their study on 600 firms randomly sampled from the Compustat data available between 1982-92 on sales, book value and the stock price. They found that managerial ownership is explained by key variables consistent with the predictions of principal-agent models. A large fraction of managerial ownership is explained by unobserved firm heterogeneity. All tough, after controlling both for observed firm characteristics and firm fixed effects, they could not conclude that changes in managerial ownership affect firm performance (Himmelberg et al, 1999)

In the paper of Core et al. (1999) the authors examined the relationship between executive compensation and firm performance as well. But they used board composition and ownership structure as measures to identify this relationship and found that these measures explain a significant amount of the variation in executive compensation. In addition, their results suggest that executives earn greater compensation when governance structures are less effective. So, their results do support the Managerial Power Theory claim. Moreover, Core et al. (1999) found that compensation arising from the characteristics of board and ownership structure was negatively related to firm performance.

Bertrand and Mullainathan (2000) empirically examined the two competing theories of executive compensation as well. In order to distinguish these theories, they examined how executive compensation responds to luck. They found substantial compensation for luck but also that better governed firms pay their executives less for luck. Bertrand and Mullainathan (2000) also examined the amount of money that executives are charged for the option contracts

(8)

they receive. Again, they find a crucial role for governance: executives in better governed firms are charged more for the options, given to them. So, their results suggest that both theories of executive compensation matter. In poorly governed firms, Managerial Power Theory fits better while in well governed firms, Agency Theory fits better.

Zhou (2000) examined executive compensation in Canada and found proof for Agency Theory. The data consisted of 755 Canadian firms over the period 1991-95. The study found evidence that CEO pay rises with firm size and compensation is tied to company performance.

One widely cited paper by Bebchuk and Fried (2006) supports the Managerial Power Theory. They claim that executives remuneration policies are decoupled from firm performance. Bebchuk and Fried (2006) explained different types of forces used to set executive remuneration policies to benefit the executive:

• Re-election incentives

The nomination of board members is always strongly influenced by the CEO. For this reason, directors in the board have a strong motivation to support the CEO with his pay arrangement. Earlier research has shown, that there is a positive correlation between CEO and director compensation. It has been proved that this is caused by corporation between the CEO and directors of the board, rather than by performance of the company.

• Loyalty and friendship 

Earlier research has shown that when the chairs of the compensation committees are appointed after the CEO has been elected, the CEO tend to receive a higher

compensation for his job.   • Collegiality and authority  

Directors tend to treat their CEO respectfully and with great deference. In most cases, it is difficult for directors to disagree and disrespect a colleague’s and leader’s decisions. • Solidarity

Most members of compensation committees have been CEO themselves or are still CEO at different companies. It is very likely that these people have formed ideas that the pay arrangements from which they have benefited, are appropriate for these functions.

• Insignificant costs of favoring the executive

Typically, directors of the board own only a small part of the firm. With this reason, the costs to these board members of approving the executive’s compensation arrangements that are overly favorable to this executive, are small.

(9)

• Ratcheting

Many boards are trying to pay their CEO more than the industry average. This has led to an increase in the average pay of executives.

Gabaix and Landier (2008) based their study on the 1000 best paid U.S. executives in the period 1992 until 2004. They constructed their total compensation variable on salary, bonus, restricted stock options granted and the Black-Scholes value of the stock options granted. Gabaix and Landier (2008) found that the executive’s equilibrium pay is dependent on the both the size of the firm and the size of the average firm in the economy. The sixfold increase in executive pay between 1980 and 2003 can be attributed to the sixfold increase in market capitalization of large U.S. companies during that period (Gabaix and Landier, 2008).

Cuñat and Guadalupe (2009) studied the extent to which changes in executive pay are driven by changes in the structure of the product market and in competition. They constructed their study on data from the five best paid executives in the banking or financial sector included in the S&P 1500 for the period 1992 till 2002. Total compensation was measured by yearly wage, bonus, stock options and other compensations. During the given period, the financial sector faced two major deregulations, leading to an increase in competition in the deregulated sectors. Cuñat and Guadalupe (2009) found that deregulations substantially changed the level and structure of the compensation: the variable components of pay increased along with performance-pay sensitivities and, at the same time, the fixed component of pay fell. The overall effect on total pay was small.

2.3 Other views

Besides the two opponent views described above, some other views are receiving more attention in the academic world. One of them is Tournament Theory. In this theory, the tournament itself is a highly competitive situation, where agents compete for a limited number of prizes. Workers are compared with each-other and ranked based on performances compared to their co-workers. Big prizes (high salaries and top-level jobs) are rewarded to the higher ranked workers. These tournaments can be an effective incentive instrument in many situations (Grund and Sliwka, 2005).

In 1999, Eriksson examined whether executive compensation levels could be explained by the use of Tournament Theory. He conducted an empirical experiment on 2600 executives in 210 Danish firms during 1992-95. Eriksson (1999) identified a stable convex relation between executive pay and job levels. The larger the number of managers considered to have significant

(10)

responsibilities in the firm, the larger the wage spread. While this bigger wage spread in the hierarchy of managers provides useful incentives to improve firm performance.

The study of Conyon et al. (2001) tested the implications of Tournament Theory based on 100 companies, listed on the UK stock index in the late 90s. Canyon et al. (2001) found a convex relationship between executive pay and organizational level and also that the gap between CEO pay and other board executives is positively related to the number of participants in the tournament. But Conyon et al (2001) also found that the variation in executive team pay has little role in determining company performance.

Also, Kale et all. (2009) examined how promotion incentives based on differential compensation between Chief Executive Officers and their potential successors affect firm performance in a tournament situation. They suggested that the alignment of executives with shareholders is an important consideration for firm performance, but the incentive structure of other top managers is important as well. Chief Executive Officers exert extra effort when their incentive is aligned to owing firm-specific equity, while lower-level managers are motivated due to two sources: how and how much they are currently paid and how much they can expect to be paid in the event of a promotion to the position of Chief Executive Officer. The last one is the prize in the tournament. Kale et all (2009) conducted this tournament experiment on the Vice Presidents and Chief Executive Officers of 1500 firms ranking in the S&P500, S&P mid-cap 400 and S&P small-mid-cap 600 indices for the period 1993 to 2004. Kale et all (2009) found that total, short-term and long-term, gaps between Chief Executive Officer and Vice President compensation affected the firm performance positively and that this positive relation between tournament incentives and firm performance remains robust to many alternate specifications. Lin’s et al. (2013) study observed the publicly traded firms in Taiwan during the period of 2002–2004. With their 1322 observations Lin’s et al (2013) explored executive pay gaps among various hierarchy levels and their influence on promotional policies and company performance. Their results confirmed the studies described above, executive pay is not based on absolute performance but on ranking. More precisely, the more influence hierarchy level has on the success of the company, the larger the pay gaps found.

Shen et al. (2009) investigated the impact of pay on CEO turnover from both the Managerial Power Theory perspective as well from the Tournament Theory perspective. Their data consisted of a sample of 313 large U.S. companies from 1988 to 1997. Shen et al (2009) found that both the level of CEO pay as its ratio over the average pay of the firm's highest paid executives have a negative impact on CEO turnover. So, according to Shen et al (2009) do their finding of the negative impact of CEO pay on CEO turnover suggest that the other executives are do not believe that there is a fair tournament at the top. Therefore, their study challenges tournament theory and the incentive effect of CEO pay proposed by it.

(11)

2.4 Long-run perspective

In 2010, Frydman and Saks published a paper that analyzed the long-run trends in executive compensation for the period 1936-2005. Their dataset consisted of top executives from 101 large companies. They found two surprising facts that go against current views of the important determinant of top executive pay. First, executive compensation was flat from the end of World War II tot the mid-1970s, even though firms grew considerably during that time. Second, the magnitude and determinants of the pay-to-performance were similar from the 1930s till the 1980s and then strengthened considerably from the mid-1980s to 2005. Their conclusion was that compensation policies often did help to align managerial incentives with those of the shareholders in the long-run, because executive wealth was sensitive to firm performance for most of their sample (Frydman and Saks, 2010).

2.5 Summary

Like Bertrand and Mullainathan (2000) and Tosi et al. (2000) already suggested, there can be identified two competing views in the field of executive compensation. Studies of Jensen and Murphy (1990), Hall and Liebman (1998), Zhou (2000) do find some kind of rationale for Agency Theory, while studies of Lambert et al (1993), Core et al. (1999) and Bebchuk and Fried (2006) find more evidence for Managerial Power Theory. So far, there is no absolute winner identified between these theories. Although Friedman and Saks’ (2010) study does suggest that in the long-term Agency Theory can explain the levels of executive compensation. Next to that, the studies of Gabaix and Landier (2008) and Cuñat and Guadalupe (2009) suggest that other factors as market capitalization and deregulation also have great explanatory power on the rise in executive compensation. The studies of Eriksson (1999), Conyon (2001), Kale et al (2009) and Lin’s et al (2009) do suggest that a different theory as Tournament Theory has explanatory power for the levels of executive compensation as well. All tough Shen et al (2009) findings do challenge tournament theory.

3 Hypotheses

This section will introduce and explain the different hypotheses. These hypotheses are formulated in order to make them testable and able to answer the different research questions. Hypothesis 1 and 2 separate direct cash bonus from other compensation paid to the executive and Hypothesis 3 combines these two variable compensation. This way the regressions offer

(12)

more insight in the special role stock options play in remuneration policies, as Hall and Liebmann (1998) and Bertrand and Mullainathan (2000) suggested. The fourth and last Hypothesis, tests whether years of tenure of the executive results in higher variable compensation paid out, as Managerial Power Theory suggests.

Hypothesis 1

▪ H1a: Executive’s bonus earned is positively influenced by Annualized total return during tenure relative to market. (Agency Theory)

▪ H1b: Executive’s bonus earned is not positively influenced by Annualized total return during tenure relative to market. (Managerial Power Theory)

The first hypothesis examines the effect of the firm performance which is translated to ‘Annualized total return during tenure relative to market’ on the ‘Executive’s bonus earned’. The ‘Executive´s bonus earned’ is the amount of USD in 2012 and is part of the total compensation that the executive earned in the fiscal year 2012. As discussed before, this part should – by the principle of Agency Theory – be completely determined by the performance of the executive and thus by the firm-performance.

The study of Lambert et all. (1993) suggested that some executives have longer tenure because the board is convinced that these executives make the firm face no serious agency costs, so we can expect that the executive´s actions increases shareholder’s wealth and thus firm performance during his tenure. Since last year’s firm performance by itself does not incorporate this effect, Lambert et al (1993) suggested that the most fair measure incorporates the total effect of the executive on firm performance. And since the market had a lot of movement by itself, our measure should be corrected for this. So this thesis measures firm performance as the ‘Annualized total return during tenure relative to market’. This variable measures the annualized return on an investment made in the stock at the time the executive has been appointed till the end of the fiscal year 2012 compared to an investment at the same time in the S&P 500. Using this method, we make sure that the total effect of the executive on the firm performance is taken into account by determining the strength of the relationship and the alternative perspective, which Lambert et all (1993) made, should not matter anymore.

Following the principle of Agency theory we can expect a positive relationship between this ‘Annualized total return during tenure relative to market’ and the executive’s bonus received in 2012.

Hypothesis 2

▪ H2a Executive’s other gains earned are positively influenced by Annualized total return during tenure relative to market. (Agency Theory)

▪ H2b: Executive’s other gains earned are not positively influenced by Annualized total return during tenure relative to market. (Managerial Power Theory)

(13)

The second hypothesis examines the effect of this ‘Annualized total return during tenure relative to market’ on ‘Executive’s other gains’. The ‘Executive’s other gains’ are the other compensations paid to the executive in the fiscal year 2012. Compensations paid, such as vested restricted stock grants, long term investment plans and the value realized by exercising stock options. Again, by the principle of Agency Theory we can expect a positive relationship between this ‘Executive’s other gains’ and ‘Annualized total return during tenure relative to market’.

Hypothesis 3

▪ H3a: Executive’s total variable compensation earned is positively influenced by Annualized total return during tenure relative to market. (Agency Theory)

▪ H3a: Executive’s total variable compensation earned is not positively influenced by Annualized total return during tenure relative to market. (Managerial Power Theory) The third hypothesis examines the effect of ‘Annualized total return during tenure relative to market’ on ‘Executive’s total variable compensation’, which is the summation of the ‘Executive’s bonus earned’ and the ‘Executive’s other gains’ from the first and second hypothesis. So, by the principle of Agency Theory we can expect a positive relationship again. Hypothesis 4

▪ H4a: Executive’s total variable compensation earned is positively influenced by years of tenure as executive in the firm. (Managerial Power Theory)

▪ H4a: Executive’s total variable compensation earned is not positively influenced by years of tenure as executive in the firm (Agency Theory)

The fourth and last hypothesis examines the effect of ‘Years of Tenure as executive in the firm’ on ‘Executive’s total variable compensation earned’. By the principle of Managerial Power Theory, this effect should be positive. The theory suggests that executives are able to extract rents from the company by selecting and influencing the board of directors. So, this effect should be positively correlated with the time that the executive served the company as Chief Executive Officer. CEOs who are leading the company for a longer time already, have had more elections and opportunities to influence the board.

(14)

The data for this thesis is collected from Forbes (2012). More precisely, the sample of this study consists of America’s 100 best paid executives according to the Forbes (2012) list: ‘America's Highest Paid CEOs’. Forbes selected these executives, based on the total amount of compensation the executives earned in 2012. This total amount was determined by the following four determinants: salary, bonus, other and stock gains. All other measures used in this study, as annualized return during tenure relative to market, years of tenure, industry and percentage of ownership are collected from the Forbes (2012) list as well.

Some of the executives in the Forbes list are also founders of the firm, so they are excluded from the data since there does not exist an agency relationship as discussed in the ‘Literature Review’. Some of the data was incomplete and has been excluded as well. Corrected for these excluded data, the sample consists of 87 observations.

5 Methodology

In the sections above, the arguments for choosing the variables and the different hypotheses have been discussed. This section will explain and introduce the technical set-up of the different variables, testing for multicollinearity and the regression models.

5.1 Definition of variables

The data and variables are collected from the Forbes (2012) list. However, not all the variables needed for this thesis were provided by Forbes (2012). The variables TOTALVARCOMP and FIRMSIZE are calculated for this thesis. These variables are still based on the values provided by Forbes (2012).

Independent variables

▪ RELtoMARKET

This is the variable for annualized total return during tenure relative to market, it is the measurement of firm performance corrected for the movement of the market itself. The annualized total return during tenure covers the tenure of the executive or from the time of the IPO or available stock history. Relative to market is the ending value of $100 invested in the stock, divided by the ending value of $1 invested in the S & P 500 (a score of 100 = the S & P 500 (Forbes, 2012).

(15)

Refers to the number of years that the executive is Chief Executive Officer of the company (Forbes, 2012).

Dependent variables

▪ BONUS

This variable refers to the direct cash bonus paid to the executive for the fiscal year 2012, given in USD (Forbes, 2012).

▪ OTHER

Refers to other compensation paid to the executive for the fiscal year 2012, such as vested restricted stock grants, long term investment plans payouts and the value realized by exercising stock options. Also the other compensation is given in USD. (Forbes; 2012).

▪ TOTALVARCOMP

Measurement, consisting of the summation of the variables BONUS and OTHER for a given executive. So, this variable refers to the total variable compensation that the executive received in 2012, given in USD.

Control variables

▪ STOCKOWNEDPER

This measurement refers to the percentage of the firm´s stock is owned by the executive (Forbes, 2012). The regressions are controlled by this control variable, because if a CEO has a larger stake in the company, the less Agency problems should be an issue.

▪ FIRMSIZE

This measurement is constructed by the variable STOCKOWNEDPER multiplied by the value of that stock owned by the executive, given in US 2012 dollars. The regressions are also controlled for this effect because the literature already proved that larger firms tend to pay their executives higher compensations (Hall and Liebman, 1998).

5.2 Testing for multicollinearity

Economic theory suggests that the independent variables for our regressions (RELtoMARKET,

TimeCEOyears, STOCKOWNEDPER & FIRMSIZE) , are possibly very correlated with each

other. Two examples :  

• A CEO’s tenure can be expected to be longer if his firm is successful, which implies a strong positive correlation between TimeCEOyears and RELtoMARKET

• For CEOs, who own a large amount of stocks of their firm, their wealth is very

dependent on the firm’s performance. So these CEOs have a stronger incentive to make their firm successful, compared to CEOs owning a smaller amount of stock.  

(16)

Based on this assumption, a strong positive correlation can be expected between

STOCKOWNEDPER and RELtoMARKET

To investigate whether we can disentangle the effects of these dependent variables, we calculated the correlation coefficients between these four variables.

5.3 Regression models

For the regression models, we preferred to use logarithmic functions of BONUS, OTHER and TOTALVARCOMP, because the intervals in the data are large and the data consists of some outliers. Outliers can result in large biases on linear variables, which is less the case when we use logarithms. For this reason, using logarithms makes the results of the regressions more accurate.

Regression 1

▪ logBONUS = αi + β1(RELtoMARKET) + β2(TimeCEOyears) + β3(STOCKOWNEDPER) + β4(FIRMSIZE) + εit.

Regression 2

▪ logOTHER = αi + β1(RELtoMARKET) + β2(TimeCEOyears) + β3(STOCKOWNEDPER) + β4(FIRMSIZE) + εit.

Regression 3

▪ logTOTALVARCOMP = αi + + β1(RELtoMARKET) + β2(TimeCEOyears) + β3(STOCKOWNEDPER) + β4(FIRMSIZE) + εit.

(17)

6 Results

This section will report the outcomes of the different regressions. The section will start with a short interpretation and explanation of the descriptive statistics. After that, we will test the independent variables for multicollinearity and the outcomes of the different OLS-regressions will be discussed and compared to the hypotheses mentioned before.

6.1 Descriptive statistics of the variables

The descriptive variables for the period 2012 are shown in Table 1. The table gives an overview of the 100 best paid executives in the United States regarding Forbes (2012). On average, their bonus is USD 4,813M, with a standard deviation of USD 4,650 M. The executives also received other variable compensation of nearly USD 8,946 M on average, with a standard deviation of USD 6,268 M. The mean of their total variable compensation is USD 13,800 M USD, with a standard deviation of USD 7,718 USD.

In return, the executives outperformed the market with 6,9% during their tenure, with a standard deviation of 8,8%. The executives have been serving the company as Chief Executive Officer with a mean of 9,09 years, with a standard deviation of 5,02 years. And the CEOs owned 0,78% of the firm they represent on average, with a standard deviation of 3,11%. 

Table 1: Descriptive statistics of the variables. The first column shows the name of the variable, the second and the third column provide the corresponding average value and standard deviation of these

variables. 

Mean Standard deviation

TimeCEOyears 9,09 5,02 SALARY in USD 1,412 M 0,0881 M BONUS in USD 4,813 M 4,650 M OTHER in USD 8,946 M 6,268 M TOTALVARCOMP in USD 13,8 M 7,718 M STOCKOWNEDPER 0,78 3,1 RELtoMARKET 1,069 0,088 FIRMSIZE in USD 36,6 B 40 B

(18)

6.2 Correlation coefficients between independent variables

To be certain that we can disentangle the effects of the independent variables, we calculated the correlation coefficients, displayed in Table 2.  

Table 2: An overview of the correlation coefficients between the dependent variables.

The strongest correlation can be found between TimeCEOyears and STOCKOWNEDPER which has the highest absolute value with 0,41. This value is not problematic when we consider multicollinearity, which is only the case if the correlation coefficient is close to 1 or -1. Based on these results, the independent variables are appropriate to use in the regressions. 

6.3 Regression results

The first regression model tested the effect of the independent variables RELtoMARKET and TimeCEOyears on the dependent variable BONUS, while controlling for STOCKOWNEDPER and FIRMSIZE. From the second column of Table 2, we can interpret that the only significant variables – by a confidence level of 90% – are TimeCEOyears and FIRMSIZE. Meaning that annualized return during tenure relative to market had no significant effect on the cash bonuses the executives received. While every extra year they already served the company as Chief Executive Officer increased their bonus with 3,9%.

The results suggest that the cash bonuses paid in 2012 to America’s best paid executives are not constructed on the basis of annualized total return relative to the market since their tenure. In other words, outperforming the market since the executive has been appointed does not influence their bonus. But the number of years they are already leading the firm as executive, do influence the value of their bonus. With these findings, we can reject Hypothesis 1.

STOCKOWNEDP

ER TimeCEOyears RELtoMARKET FIRMSIZE in USD STOCKOWNEDPER 0,410017 0,052269 -0,1038

TimeCEOyears 0,410017 0,177606 -0,204715

RELtoMARKET 0,052269 0.177606 -0,234397

(19)

Table 3:Regression results of the three regressions, described in Section 5.3. The variable coefficients are given, with the standard deviations between parentheses. *, ** and *** imply significance of the variable at levels of respectively 10%, 5% and 1%.

The second regression model tests the effect of the independent variables RELtoMARKET and TimeCEOyears on the dependent variable OTHER, while controlling for STOCKOWNEDPER and FIRMSIZE. Following from the third column of Table 3, we can conclude that RELtoMARKET is the only significant variable at a 5% significance level. But completely against Agency Theory expectations the correlation is strongly negative. For every percent the the executive outperformed the market, his other variable compensation declined by approximately 3,125%. So we also reject Hypothesis 2.

The third regression tests the effect of the independent variables RELtoMARKET and TimeCEOyears on the dependent variable OTHER, while controlling for STOCKOWNEDPER and FIRMSIZE. Form table 4 we can interpret that the only significantly important variable – at a 5% significance level – is the variable RELtoMARKET. The effect on total variable compensation is again negative, namely -2,151% for every percent extra outperforming the market since tenure. Based on these results, we reject Hypothesis 3 as well.

As described above, the number of years the executive is in the company does not significantly influence the amount of total variable compensation received by the executives. So, hypothesis 4 is rejected as well.

None of the regressions have a high R-squared value, which also suggests that the chosen independent variables do not have strong explanatory power on the dependent variables. Therefore, the results support the claim that the variable compensations paid to America’s

1 2 3

LogBONUS LogOTHER LogTOTALVARCOM_P

RELtoMARKET -1,61 -3,13** -2,15** (1,14) (1,45) (0,76) STOCKOWNEDPER -1,08 -0,79 -0,95 (3,39) (4,32) (2,28) TimeCEOyears 0,039* -0,03 -0,005 (0,02) (0,03) (0,015) FIRMSIZE 4,78 * 1,1 2,09 (2,51) (3,2 ) (1,69 ) Adjusted R-squared 0,051 0,0449 0,0985

(20)

highest earning executives is not funded on firm performance. This goes against the Agency Theory.

6.4 Limitations

After all, I want to discuss the main limitations of the study performed. First, the number of observations is a limitation of this study. The number of observations is sufficient to answer the questions regarding America’s 100 best paid executives and contributes to the literature, but the results would be more solid when the study would have contained more observations.

Second, the study focused on the two opponent views in the field but did not incorporate the third model –Tournament Theory – in the hypotheses. However, corresponding with the scope and time limits of this thesis, I had to use simplified models and reduce the amount of data. Third, the simplified model chosen couldn’t prevent the possibility of multicollinearity. The dependent variables are likely correlated. However, by calculating the correlation coefficients, I regarded this limitation and showed that the variables are used in an appropriate manner.

7 Conclusion

This thesis examined the suggested relation between firm performance and executive pay. In order to get more insight, the thesis tested which of the two competing views in the literature – Agency Theory and Managerial Power Theory – could explain the payments made to America’s 100 best paid executive officers according the Forbes (2012) list.

The results of the first regression show there is no significant proof for Agency Theory and the direct cash bonuses paid to the executives. The results of the second regression are against all expectations of Agency Theory and confirm the Managerial Power Theory, as the results of the third regression do as well. However, the fourth and last regression did not find significant proof for the Managerial Power Theory.

The findings do not support Agency Theory and are in line with the findings of Jensen and Murphy (1990) and Lambert (1993), who did not find a strong relationship between firm performance and executive pay. However, this thesis does find some significant evidence for Managerial Power Theory, supporting the findings of Core et al. (1999) and Bebchuk and Fried (2006). Interesting are the findings of Hall and Liebman’s (1998), Zhou (2000) and Frydman and Saks (2010) compared to the results of this thesis. They all found a positive relationship between pay and performance, while the results of this thesis showed a negative relationship. The findings of Bertrand and Mullainathan (2000) could be a possible explanation. They found that in poorly governed firms Managerial Power Theory fits better, while in well governed firms Agency Theory fits better. This thesis examined the 100 best paid executive officers, which are

(21)

not necessarily the 100 best governed firms of America. Next to that, the findings of Gabaix and Landier (2008) and Cuñat and Guadalupe (2009) already suggested that other factors as market capitalization and deregulation also have great explanatory power on the rise in executive compensation.

The results of this thesis suggest that there is no proof for Agency Theory based on America’s best paid officers according the Forbes (2012) list. The results suggest there is no positive relationship between firm performance and executive pay, while supporting the claim for Managerial Power Theory at least in the non-direct cash bonuses earned by the executives. As mentioned before, due to the time and scope of this thesis the number of observations and economic theories is limited. More observations would have supported a more solid claim and more economic theories would have given more explanatory power. The studies of Eriksson (1999), Conyon (2001), Kale et al (2009) and Lin’s et al (2009) suggest that Tournament Theory is the economic theory that could have explained the high pay-offs made to the executives. It could be interesting to conduct further research on this subject and dig deeper in the Tournament Theory as an explanation for the pay-offs.

(22)

References

Baiman S. (1990). Agency Research in Managerial Accounting: A Second look. Accounting,

Organizations and Society, Vol. 15, No. 4 pp. 341- 371.

Bebchuk, L.A. and Fried, J.M. (2006). Pay without performance: Overview of the issues.

Academy of management perspectives. Vol. 20, No.1, pp 5-24.

Bertrand M. and Mullainathan, S. (2000). Do CEOs set their own pay? Then ones without principals do. NBER Working Paper 7604

CNN (2011). How can we address excessive CEO pay? (http://management.fortune.cnn.com/ 2011/04/13/how-can-we-address-excessive-ceo-pay), 12 December.

Conyon, M.J., Peck, S.I. and Sadler, G.V. (2001). Corporate Tournaments and Executive Compensation: Evidence from the U.K.. Strategic Management Journal¸ Vol. 22, No.8, pp. 805-815

Core, J., Holthausen and R., Larcker, D., (1999). Corporate governance, chief executive officer, and firm performance. Journal of Financial Economics, Vol. 51, pp. 371-406.

Cuñat, V. and Guadalupe, M. (2009). Executive compensation and competition in the banking and financial sectors. Journal of Banking and Finance. Vol. 33, pp. 495-504.

Eisenhardt, K.M.. (1989). Agency Theory: An Assessment and Review. The Academy of

Management Review. Vol. 14, No. 1, pp. 57-74

Eriksson, T. (1999). Executive Compensation and Tournament Theory: Empirical Tests on Danish Data. Journal of Labor Economics, Vol. 17, No. 2, pp. 262-280

Forbes (2012). America’s Highest Paid Chief Executives (http://www.forbes.com/lists/2012 /12/ceo-compensation-12_rank.html), 12 December 2013.

Frydman C. and Saks R.E. Executive compensation: A new view from a long-term perspective, 1936-2005. The Review of Financial Studies, Vol. 23, No. 5, pp. 2099-2138

Gabaix, X. and Landier, A. (2008). Why has CEO pay increased so much. Quarterly Journal of

Economics. Vol. 123, No. 1, pp.49-100.

Grund, C. and Sliwka, D. (2005). Envy and Compassion in Tournament. Journal of Economics

and Management Strategy. Vol. 14, No.1, pp. 187-207.

Hall B.J. and Liebman J.B. (1998). Are CEOs really paid like bureaucrats? Quarterly Journal of

Economics, Vol. CXIII, No. 3, pp. 653-691

Himmelberg, C.P., Hubbard, R.G. and Palia, D. (1999). Understanding the determinants of managerial ownership and the link between ownerships and performance. Journal of

Financial Economics. Vol. 53, pp. 353-384.

Jensen M.C. and Murphy K.J. (1990). Performance Pay and Top-Management Incentives.

Journal of Political Economy, Vol. 98, No. 2, pp. 225-264

(23)

26, pp. 301-325

Jensen M.C. and Meckling, W.H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics. Vol. 3, No.4, pp. 305-360.

Lambert, R.A., Larcker, D.F., and Weigelt, K. (1993). The structure of organizational incentives. Administrative science quarterly, Vol. 38, No. 3, pp. 438-461. Lin, Y.F., Yeh, Y.M.C and Shih, Y.T. (2013). Tournament theory's perspective of

executive pay gaps, The Journal of Business Research, Vol.66, No. 5, pp 585-595. Kale, J.R., Reis, E. and Venkateswaran, A. (2009). Rank-Order Tournaments and Incentive

Alignment. The Journal of Finance, Vol.64, No. 3, pp. 1479-1512

Ross S.A. (1973). The Economic Theory of Agency: The principal’s problem. The American

Economic Review, Vol. 63, No. 2, pp 134-139

Shen W. Gentry, R.J. and Tosi Jr, H.L. (2009). The impact of pay on CEO turnover: A test of two perspectives. Journal of Business Research, Vol. 63, No. 7, pp.729-734.

Zhou, X. (2000). CEO pay, Firm Size, and Corporate Performance: Evidence from Canada, The

Agency theory and America’s best paid executives