Thesis Topic:
‘’Accounting measures of performance and Stock Price performance as predictors of CEO Compensation in different time periods and within different sized firms.’’ An Empirical Analysis
Konstantinos Theodorakopoulos
Email:
Erasmus School of Economics
Department of Economics
Supervisor: M. van Dongen
Abstract
The compensation of CEOs’ is a field of study which attracts much of attention in recent years. This study will try to find the effect of different measures of performance on total compensation, and which measure or combination of measures gives more incentives to CEOs’ in different time periods and within different sized firms. The measures of interest are Earnings per Share, Stock Price, Net Income, and the combination of Net income with Stock Price and Earnings per Share with Stock Price. The sample periodincludes 17 years of observations and is separated in three sub periods (1992 –2001, 2002 -2005, 2006 – 2009). All the measures of performance are found to have a positive correlation with the total CEO compensation. Also the combination of Net Income with Stock Price is found to affect more the level of CEO compensation thanEarnings per Share, Net Income, and Stock Price separatelyfor the time period of 1992-2001. For the time periods of 2002-2005,and 2006-2009Earnings per Share constitutes the measure which will provide proper incentives to CEOs’. In addition, Net Income is found to be significant and have decreasing pay performance sensitivity with the increase of firm size for all time periods. Stock Price is found to be a significant predictor of total compensation with an increasing coefficient the bigger the firm size according to market value categorization and with a decreasing coefficient according to revenue categorization. Furthermore EPS is insignificant in predicting total compensation in the time period before SOX legislation but after this period it is found to be significant with pay performance sensitivity which decreases the bigger the firm size according to revenue categorization and an increasing coefficient according to market value separation. Total compensation is increasing until 2007 but in 2008 and 2009 it decreases due to the financial crisis.
Key words: CEO Compensation, Measures of Performance, Earnings per Share, Net Income, Stock Price, Firm Size, SOX Legislation, Financial Crisis.
Table of Contents
Chapter 1 Introduction………………………………………………………………………………….……….4
Chapter 2 Motivation………………………………………………………………………………………….….6
Chapter 3 Theoretical Background…………………………………………………………………….……
Chapter 4 Literature Review…………………………………………………………………………………..9
4.1 Compensation, Firm performance, and Firm size………………………………....9
4.2 Sarbanes Oxley Act (SOX) legislation and CEO compensation………..……15
4.3 Financially distressed firms and CEO Compensation……………………..…….18
Chapter 5 Hypotheses Development………………………………………………………………….….20
Chapter 6 Research Methodology…………………………………………………………………………21
6.1 Research Method……………………………………………………….…………………….……………21
6.2 Variable Definition……………………………………………………………………..……………..…
6.3 Data………………………………………………………………….……………………………………..……..
Chapter 7 Results…………………………………………………………………………………………………...……27
Chapter 8 Conclusion & Limitations..………………….………………………………………..………
8.1 Conclusion………………………………………………………………………………………………….…45
8.2 Limitations...... 47
REFERENCES……………………………………………………………………………………………………………
Chapter 1 Introduction
The compensation of top executives is a broad field of study which receives great of attention in many academic literatures, with the paper of Jensen & Murphy being one of the most frequently cited work. Many literatures focus on compensation contracts and if they provide the CEOs with proper incentives. This study will focus on different measures of performance and try to find which type of measure is better in explaining the changes in CEO compensation. The measures of performance examined are accounting measures, including net income and earnings per share. Also stock price is examined as a measure of performance. After proceeding to our empirical analysis we will find which measure(s) (stock price or accounting measure(s)) is (are) better predictor(s) of CEO compensation and what the link between pay and performance is.
The research question that is formulated is:
What is the pay-performance sensitivitybetween the different measures of performance and total CEO compensation, and which is the best measure of performance by taking into consideration:
1) The pay performance sensitivity and
2) How much of the variation of the total CEO compensation is explained by each measure.
Generally it is found to be a weak relationship between pay and performance when compensation is based on stocks and options. It is difficult to compare the studies because every single study uses different measures of performance and defines CEO compensation in a different way.
The research question is relevant to management accounting and control except of corporate governance because the goal of management control is to observe whether people act in the best interest of the organization. So, by answering the research question we can see which measure(s) gives more incentives to CEO’s in orderto act for the best of the organization by increasing their performance and align their interests with those of shareholders.
In this study I will test the effect of stock price (SP) on CEO compensation as a measure of firm performance, and I mean how total CEO compensation is affected by stock price changes. In my analysis I will also test how net income (NI) affects CEO compensation, if it is selected as a measure of performance and how earnings per share (EPS) contribute to CEO compensation. In addition, I will test the overall effect that combined measures of performance have on CEO compensation. The combined measures are SP with EPS, and SP with NI. It would be very interestingto test the overall effect of the two accounting measures togetheron CEO compensation, but this is not possible due tothe high correlation between EPS and NI.
The firms will be separated into three categories (large, medium, and small) according to their 1) Market value, and 2) revenue, in order to see if and how firm size affects the results. In this process I will split the whole time period (1992-2009) into three smaller time periods (1992-2001, 2002-2005, and 2006-2009) and see whether situations (like financial crisis, currency crashes, recessions, bubbles, sovereign debts, and the introduction of SOX legislation) that occurred in these periods have an impact in the results. At the end I will be able to know the relation between each measure of performance and combination of measures with CEO compensation and which measure or combination of measures is (are) the best predictor(s) of CEO compensation in different sized firms and different time periods.
Chapter 2 Motivation
In this chapter we will discuss the motivation of this study. As it is stated in the introduction the focus of this study is set on finding out the relation between pay and performance and what is the best measure of performance that best explains the variance of the total compensation. In previous literatures no model has been developed to test these variables in the way that this study does.There is a lot of literature about the pay performance relation and is analyzed in chapter 4. The most important is that this study will take into consideration the size of the firm which, as we have seen from various studies like (Schaefer (1998) and Baker & Hall (2004)), is a very important factor that relates to pay performance sensitivity. Moreover we use a very long time period from 1992-2009. During this time period different economic factors took place. For example on March 2001- November 2001 we face the early 2000s recession, in 2002 we have the introduction of SOX legislation, and from December 2007-June 2009 we lived in the great recession. So we separate the whole time period into three subcategories (1992-2001, 2002-2005 and 2006-2009) in order to find out if and how my results differ in these time periods. No study before has split the time period in this way and categorize the firms by their size as this study does. So this kind of categorizations makes the research questions and this study unique. It is a genuine empirical analysisabout US that will draw important conclusions and reveal if the results can be generalized or not for all of the three periods and different sized firms.
Chapter 3 Theoretical Background
An important theory for the relation between pay and performance of executives is agency theory. According to agency theory a basic problem that arises is that the personal goals of managers may not be in alignment with the maximization of shareholders wealth. This comes from the fact that shareholders authorize managers to manage the firm’s assets and in that way create different interests between shareholders and managers. Agency theory supports that in order to deal with different interests between shareholders and managers, compensation plans should be designed to align the interests of risk averse executives with those of shareholders. The solution according to this theory is that executive compensation should be based on performance with some monitoring (stockholders monitor managers actions) to be taken into account. Performance is based on financial measures like return on assets, return on equity, earnings per share (EPS), and stock price changes. Sometimes it is very expensive and difficult to monitor the agents’ actions.So, agency theory suggests thatifcompensation is based on financial performance,managers act in shareholder’s interestsand maximizeboth financial performance and their own wealth,Stroh et al (1996).
On the other hand, tournament theory introduced by Lazear & Rosen (1981) suggests that the compensation of an executive should be mainly related to his rank in the organization and not to his actual performance. According to tournament theory managers should not be paid according to their marginal output because it is more probable to decrease their effort and responsibilities (Henderson & Frederickson (2001). Thus a big difference in payment between CEO’s and other managers of lower ranking, will affect the motivation of these lower ranked managersin order to exert more effort andincrease their payment by improvingthe performance of the firm. The results on the research of this theory are not clear because in some casesbig differences in compensation structure have caused decrease in productivity Bloom and Michel (2002). In contrast, other references state that compensation differences are positively related to firm performance Eriksson (1999).
Expectancy theory developed by Vroom (1964) predicts 3 factors that are of great importance in the motivation of the employees. Firstly, it states that the bigger the expectationof a strong and positive relationship between effort and performance thenthe bigger the incentive to exert bigger effort. The other factor has to do with performanceoutcome expectancy, meaning the expectation that the compensation of the executive will betied to his actual performance. So, the higher this expectation, the more motivated executives will be to exert effort. Thirdly, the greater the degree that an individual values a reward the bigger the motivation for the individual to exert bigger effort. This means that an executive increases his performance if his outcome (payment) is important to him. As we can see the reward or otherwise payment is a leading indicator of performance by increasing motivation.
Social comparisonis a theory that relates with the level of CEO compensation. According to this theory the compensation of executives is determined by the compensation committee. This committee is composed by members of directors that do not have any relation with the company in which they determine the level of compensation of a certain CEO. This means that they can have CEO positions in other companies, or be outside directors. O’Reilly, Main, and Crystal (1988) state that a comparison process takes place from the board of directors. They compare their own experience, abilities and the compensation that they receive to the experience and abilities of the CEO that they evaluate. Under these criteria theydetermine the level of compensation that a CEO should be paid.
All of these theories provide some evidence for the motivations of employees and on what measures CEO compensation should be based. This study will have in principal both agency and expectancy theory, in order to investigate how different measures of performance affect CEO compensation and which measure of performance is the best (by explaining most of the variance of the total compensation) and motivates the most CEO’s to exert more effort.
Chapter 4 Literature Review
This chapter will be separated into three subsections. In the first on it is provided an overview of earlier studies about the relation between CEO compensation and firm performance and the effect of firm size on the pay performance sensitivity.In the second one the impact of SOX legislationon CEO compensation is presented and in the third subsection how different types of firms (financial or non financial) affect the level of compensation and how the distress of a company is importantin affecting CEO compensation.
1)Compensation, Firm Performance and Firm Size.
First of all, it is very important to state that there are different definitions about CEO compensation. Every study does not measure CEO compensation in the same way. For example, Murphy (1999) measures CEO compensation with base salary, annual bonus, stock options and other forms of incentives like restricted stocks, long term incentive plans and retirement plans. The difference is embedded to the components that CEO compensation is formed. Moreover the performance measures that each study uses are of major importance because there are different ways to measure firm performance. Some studies for example use stock based measures like shareholder returns, and some other use accounting based measures like (ROE, ROA, EPS, NI).
To start with, Jensen & Murphy (1990) conducted a study in order to examine the relation between pay and firm performance (the connection between executive compensation and shareholders wealth). The results of this study were that CEO’s wealth changed by $3.25 if shareholder’s wealth changed by $1000. Also the change by $1000 of shareholders wealth the salary and bonus changes by 2.2$. Moreover every $1000 change of shareholders wealth total pay changes by 3.3 $. Although this relation is found to be statistically significant, it is suggested to be weak in order to provide CEO’s with proper incentives. The major reason that makes this relation economically insignificant is that political forces can also constrain CEO compensation.In additionAttaway (2000) and Veliyath & Bishop (1995) do notfind the relationship between firm performance and CEO compensation to be strong. On the other hand Hall & Liebman (1998) find a strong relationship between firm performance and totalCEO compensation. Moreover they have shown that the level of CEO compensation and the relationship between compensation and performance (pay – performance sensitivity) have risen dramatically since 1980 due to the increase of stock options. It is of great interest to mention that Hall & Liebman give two definitions of CEO compensation. 1) ’flow’ or ’direct’ compensation equals salary, bonus, received stocks, received options, and other compensation. 2) ‘Stock’ or ‘total’ compensation equals ‘flow’ compensation, and all options and stocks owned by the CEO. Murphy (1999) suggests that the relation between CEO compensation and the wealth of shareholders become stronger during the years. He used two models. From the first one he estimated the pay performance sensitivity (PPS) defined as the change of CEO compensation due to $1000 change of shareholder wealth. From the second one he estimated the pay performance elasticity (PPE) defined as the percent change of cash compensation due to a 1% change inshareholders wealth. In addition, Murphy (1999) stated that there are two kind of relations between CEOs’ compensation and firm performance. The first one has to do with the ownership of stocks and stock options by the CEO. This ownership is very important in aligning the goals of CEOs’ and shareholders, and determining the wealth of the CEO. This means that if the firm’s stock price increases, then from the ownership the wealth of the CEO will increase. The second relation has to do with the two major forms of CEO compensation, salary and bonuses. These two forms are affected by different accounting measures of performance and stock price performance, revealing the relation that exists with compensation and performance.
Other authors mention that some other features of compensation contracts are not taken into account by the empirical work. Boschen & Smith (1995) suggest that in order to find a link between compensation and performance we have to look in their long term relation because many executives are related with the firms for many years. In this way they find a stronger link between pay and performance than Jensen & Murphy. Aggarwal and Samwick (1999) test the relationship between compensation and performance with the variance in firm performance. For example if higher variance leads to lower pay performance sensitivity. They find that the relation of compensation and performance will increase with the variance of firm performance, and thisrelation between pay and performanceis more economically significant than what Jensen & Murphy found.
Firm performance is a determinant of executive compensation. The argument that exists is: because the CEO is the only responsible for the performance of the firm, rewards should be aligned tofirm performance. Frank (1984).On the other hand there is no clear evidence to support the relation between performance and rewards of executivesalthough it is logical to base CEO compensation on firm performance, Murphy (1985). Murphy’s explanation about this outcome was that earlier studies have not taken into consideration all the forms of CEO compensation and have focused only in salary and bonuses giving a not realistic result about the true relation of compensation and firm performance.Another study by Jensen & Murphy (1987) found some correlation between changes in compensation and performance when using as measuresof performance earnings per share and return on equity.
According to Holmstrom (1979) compensation contracts for risk averse CEO’s should not only be aligned with the increase of shareholders wealth but also other variables that will provide important information about the CEOs’ actions and choices have to be taken into account. Thesevariables include accounting measures of performance, and measures of performance based on others executives in the same industry, or market. In contrast, Gibbons & Murphy (1989) argue that basing compensation on such additional variables will not be productive because these variables do notprovide correct incentives that will not be in alignment with the maximization of shareholders wealth.This is not modeled in Holmstrom’s analysis.