Executive compensation and firm performance
Name: Ye Chen
Student number: 10597999
Thesis supervisor: Alexandros Sikalidis Date: June 25, 2018
Word count: 12083
MSc Accountancy & Control, variant Control Amsterdam Business School
Statement of Originality
This document is written by student Ye Chen who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
The executive compensation has attracted much attention for such a long time due to its importance in stimulating executives to work hard and improving firm performance. And in this paper, an
examination of the relationship between executive compensation and firm performance has been conducted for a sample of 129 multinational companies in S&P 500 from 2002 to 2006. From the regression results, firm performance is positively related to both equity-based and cash-based compensation and, but the relationship between cash-based compensation and firm performance is much stronger than equity-based compensation and firm performance, which implies that cash-based compensation has a greater impact on firm performance than equity-based compensation. Besides, regarding the four control variables, I found that SIZE is positively and significantly related to company performance, and the rest three control variables about financial leverage are all negatively related to company performance, which is consistent with many previous findings. Then I test the correlation between different variables, and the result shows no multicollinearity between different variables.
Contents
1 Introduction………5
2 Literature review and theory………..………9
2.1 Theory background……….……….9
2.1.1 The tournament theory………9
2.1.2 The agency theory………...10
2.2 Literature review………..12
2.3 Disadvantages of equity-based compensation……….….21
3 Hypothesis development…………...………..……23
4 Research method and design………...25
4.1 Research method………...25
4.2 Data collection………...26
4.3 Design of the model………...……29
5 Empirical results………..………32
5.1 Sample characteristics………32
5.2 Regression results………...36
6 Summary and discussion…...………...43
7 Limitations and improvements…...………..…46
1.Introduction
Company shareholders have paid much attention on executive compensation since it is assumed to be closely related to firm performance. And widespread bankruptcies and bank failures during the 2008 financial crisis highlight again the importance of handling the bad times well for company executives. However, there is always asymmetric information between company executives and shareholders, and executives might choose to maximize their own interest based on it, rather than shareholder’s interest, which can be catalogued as principal-agent problem (Jensen and Meckling, 1976). To be more specific, for the sustainable development of a company, shareholders tend to focus more on company’s long-term interest, while executives might put more emphasis on short-term interest, and even try to achieve it by sacrificing company’s long-term interest, such as maximizing short-term
accounting figures. Therefore, it is worthwhile to design a suitable executive compensation structure to reach a balance between executive and shareholder’s interest. The issue of executive compensation has been studied for decades, and firm performance has generally been used as an indicator to determine the composition of executive compensation (Murphy, 1999). And the findings of previous research indicate a lot of methods to deal with this protracted struggle, such as providing incentives which are tied to firm performance closely so that executives will take the optimal actions to maximize shareholders’ returns. Based on the definition agency theory, the executive compensation package in large companies could be seen as an attempt to link executive’s interest more closely with shareholders(Jensen and
Meckling, 1976). And from previous research, I found that a conventional standard to distinguish whether executive compensation packages do indeed serve shareholders’ interests or executives themselves is the sensitivity of total executive compensation to share price performance. In general, executive compensation package is made up of several parts: base salary, short-term and long-term incentives. And long-term incentives are generally a large part of many executives’ compensation (Bloedorn & Chingos, 1994). Base salary and short-term incentives could be catalogued as cash-based compensation, while long-term incentives are normally equity-cash-based
compensation. Except for the agency theory, the tournament theory also helps explain this topic. According to Lazear and Rosen (1981), in the tournament theory,
employees are motivated by the chance to get promoted so that they can have a better salary. But for executives like CEOs, they are already in such a high position that base salary is not so attractive and they have to be motivated by extra incentives, no matter cash bonus or stock-based incentives. Besides, Jensen and Murphy (1990a) once proposed that compared to cash-based compensation, managers are more motivated to maximize shareholders’ value by equity-based compensation. However, there is no enough empirical evidence showing that when a firm put more emphasis on
executive’s equity-based compensation, it performs better.
In this paper, an examination of the relationship between different executive compensation compositions and firm performance is conducted to test whether the executives’ compensation will be correlated to the firm performance more closely when a company put more emphasis on long-term equity-based incentives. Therefore,
the research question is defined as follows: would the company have better
performance if they put more emphasis on long-term incentives? By examining pay- performance relationship, companies could better understand compensation structure, manage relevant risk related to firm’s performance so that shareholders’ interest is guaranteed. The importance of investigating executive compensation has been
stressed more and more. In general, CEOs are in top place of company’s management team. They play a role of proposing and carrying out firm strategy, which is of great importance for the development of a company. Therefore, designing executive compensation incentives could be one of the most important tasks for a company’s board of directors. In the long run, it makes the company maintain healthy and stable development, and contributes to economic prosperity.
To investigate the relationship between executive compensation structure and firm performance, firstly I choose sample companies listed in the S&P 500 because these American firms are representative and well-known. Besides, as number one economy in the world, companies in the United States are more likely to have mature executive compensation structure, stable return on assets. Therefore, these companies are representable. Then, I set up two models to test the relationship between firm
performance and equity-based and cash-based incentives respectively. Besides, expect for the main research of compensation structure, I also add some control variables which may also have an influence on firm performance. My primary findings on CEOs’ compensation structure are as follows: there is a positive relationship between
both equity-based and cash-based compensation and firm performance and the coefficient of the cash-based compensation is even larger than that of equity-based compensation, which shows a closer positive relationship between cash-based compensation and firm performance. This proves that the tournament theory also applies to the executives. Besides, by examining the relationship between firm performance and four control variables, I found that SIZE is positively and significantly related to company performance, and it can be explained by benefits brought of economies of scale. The rest three control variables could be catalogued as financial leverage. And they are all negatively and significantly related to company performance. The result of Vinasithamby (2015) also supports it. In general, profitable large companies are not willing to borrow too much money so that they could keep a low leverage, to get rid of the risk of decreasing operating efficiency resulting from financial distress cost of debt. Then I test the correlation between different variables. Return on assets is positively and significantly related to both stock-based compensation and cash-based compensation (at 1% significance level). But the correlation coefficients indicate that equity-based compensation has a closer linear relationship with firm performance than cash-based compensation, which is inconsistent with regression results. Besides, firm size has the largest positive
correlation coefficient with firm performance among all independent variables, while other three control variables (financial leverage) are negatively related to firm
2. Literature review and theory
2.1 Theory background
As mentioned in the introduction part, two theories are applicable in this paper: the tournament theory and the agency theory. Therefore, it is worthwhile to further investigate the relationship between these two theories and firm performance so that the results have a solid theory background.
2.1.1. The tournament theory
Compared to normal employees, executives are linked to high salaries more
frequently. And the definition of the tournament theory clarifies why companies are willing to pay such a high salary well: the differences on employee wages are based on relative differences between individuals, rather than based on marginal
productivity. This theory was firstly introduced by Lazear and Rosen (1981). In their experiment, participants would try to perform better when they found that the winners would be paid more rewards. Besides, Eriksson (1999) once investigated the
tournament theory as a theory of executive compensation. In his findings, a stable convex relationship between pay and job levels exists. The more responsibilities a manager takes, the larger is the wage spread. Therefore, an important implication of the tournament theory is that the difference in executive pay provides incentives for employees to work hard to improve firm performance. However, when it talks about the CEO, it seems a little bit different. CEO, usually as the leader of a company, is often difficult to be motivated. They are hardly attracted by solely higher salary. By
providing equity-based incentives, such as stock options, they themselves become a shareholder of the company. So, they would put more emphasis on the company performance and hope that the share price increases a lot in the future.
However, every coin has two sides. Kochan and Osterman (1994) once mentioned the drawbacks in their book. They proposed an extreme example: in a company where work is higher interdependent, when managers compete for higher wage, they are prevented from cooperation truthfully. Therefore, here the tournament theory plays a role of demotivating.
2.1.2. The agency theory
The primary and fundamental theory applied in this paper to help develop the hypothesis is the agency theory. The widespread use of it in a large number of literature has proved its value time and again. It has been found that agency theory has developed as the main theory in lots of researches whose topic is relevant to pay-performance relationship (e.g., Gerhart & Milkovich, 1990; Roth & O'Donnell, 1996; Stroh, Brett, Bauman, & Reilly, 1996). Besides, it is important to explain the exact impact and relation with the studied topic. According to Stroh, Brett, Baumann and Reilly (1996), the agency theory describes that there are a principal and an agent in each company. In our real life, the principal represents shareholders and the agent represents executives or managers. The basic assumption under this theory is that the agents are risk-averse, thus they are unwilling to take risky actions which may be beneficial to the company. The occurrence of agency dilemma is mainly due to the
asymmetric information, which means that the principals could not be able to monitor the agents’ behavior adequately and the agents take actions which benefits themselves by sacrificing the company’s interests. From the perspective of principals, to lower agency costs, the key point is to build and improve an effective incentive mechanism for agents, so that the behaviors of agents are confine to the interest of principals, to reach the target of ‘incentive compatibility’. According to Myerson (1979), an incentive-compatible mechanism ensures that every participant could achieve their own best outcome when they behave following their true preferences. Therefore, it is worthwhile investigating the executive compensation structure to align the principals and agents’ interests together closely. And the findings of Eisenhardt (1989) are in line with it: the agency theory could be used to examine the compensation scheme of managers. In general, there are two types of incentives: monetary and non-monetary (Hanley, Becker & Martinsen, 2006). Monetary incentives are usually money-based reward, include cash bonus and stock options. The types of non-monetary incentives are varied. It could be the promotion opportunity, flexible working arrangements, extra days off (outside of annual leave) and so on. In most cases the companies would choose equity-based incentives, such as stock options. The findings of Mehran (1995) proved it. The firm performance and the percentage of executive-owned equity and executive’s equity-based compensation are positively correlated. And equity-based compensation is widespread in firms which have much more outside directors; for firms whose share are primarily held by insiders, they are less willing to use equity-based compensation. Hall and Liebman (1998) also found a strong relationship
between CEO compensation and firm performance, with a large sample which has a time interval of 15 years. This relationship exists mainly due to the changes of CEO’s ownership of stocks or stock options. Moreover, one thing deserves to mention is that they did not find any correlation between cash-based compensation and firm
performance, while adding equity-based compensation in the compensation structure, the correlation showed up and increased significantly, which means that executives indeed put equity-based compensation in the first place. Being a shareholder of the company gives them more motivation to perform better to increase firm performance.
2.2 Literature review
Jensen and Murphy (1990) once investigated the relevant topic on CEO
compensation. In their paper, excessive pay on CEO compensation is always not the most serious problem, and public attention has been diverted to how CEOs are paid. By collecting salary and bonus data for 2505 CEOs in 1400 publicly held firms from 1974 to 1988, they reached conclusions which are not consistent with the most accepted views on CEO compensation at that time. Firstly, they found that although top executives are the headlines frequently, they did not receive salaries and bonuses as much as they were supposed to have. The specific example is that the average salary and bonus for CEOs of companies which are on the New York Stock Exchange in 1938 was even higher than that in 1988. Besides, one thing needs to be mentioned is that they did not found significant relationship between changes in company performance and changes in executive annual compensation. Taking all compositions
of CEO compensation such as salary, bonus, stock options and shares into consideration, a $1000 change in firm value only relates to $2.59 change in CEO compensation. Furthermore, Jensen and Murphy found that performance-driven pay for CEOs has been used much less than in the past. More specifically, CEO stock ownership in their sample was even 10 times greater in the 1930s than in the 1980s. Last but not least, Jensen and Murphy support the application of more aggressive pay-for-performance systems. With these systems, less talented executives would be paid much less than those who have managed companies better, and these less talented executives would be replaced soon by more talented ones. And this also lead to better company performance, and further increase in shareholders’ wealth.
Bloom and Milkovich (1998) examined the relationship between risk in the executive compensation structure and firm performance based on the agency theory. Previous research on agency-based executive compensation are mostly in line with agency theory’s classic definition, which assumes that an optimal compensation structure is trying to find a balance between manager’s effort and risk-aversion (Eisenhardt, 1990; Jensen, 1983; Levinthal, 1988). However, Bloom and Milkovich extended it on the basis of classic definition. Firstly, they examine the influence of degree of risk organization faces on the relationship between firm performance and incentive compensation so that they could study the joint effects of risk and incentive compensation on firm performance. Secondly, they choose to investigate the compensation of a variety of managers instead of only CEOs. They believe that it
contributes to the generalization of their findings. Thirdly, they examine the balance between stock-based incentive compensation and cash-based salary, and then study the relationship between risk and them, not just focus on stock-based incentive compensation, ignoring pay level related to cash. The executive compensation data for their analysis was drawn from a consulting company. Altogether 740 firms and 75 randomly managers of each firm were select, from 1981 to 1988. In their model, they use firm ownership as the control variable. And their results show that firms with higher risk would put more emphasis on long-term incentive pay, rather than short-term incentive pay. And risks play a more important role than incentive compensation regarding firm performance. And risk is also able to have an effect on both
compensation policies and firm performance. Besides, they partly agree with the interpretation of the agency theory. More specifically, although company’s
shareholders are willing to use incentive compensation to align their interests together with the interest of executives, the effect of this incentive pay is hard to predict since it is not as simple as assumed.
Sanders (1999) once studied both stock ownership and stock option pay since they play a role of financial incentives. Stock ownership and stock options could be seen as substitutes which can both used as equity-based compensation. The author chose 250 companies randomly from S&P 500, during the period from 1994 to 1996, a 3-year time interval. In this paper, firm performance was measure by total shareholder returns, which is the sum of stock price return and dividend return. And there are
three methods to measure firm risk: beta for stock return risk, standard deviation of ROA for income stream risk, and debt to equity ratio for firm’s strategic risk. Two control variables were added: the prior firm performance and firm size (the log of firm sales). Sanders has three principle findings. Firstly, there is indeed a positive and significant relationship between firm performance and both executive stock
ownership and stock option pay. Secondly, the positive relationship between firm performance and executive stock ownership becomes stronger regarding to companies with high level risk. Lastly, the positive relationship between firm performance and executive stock option pay becomes weaker regarding to companies with high level risk. He found that although stock ownership and stock option pay can both truly encourage executives to put more emphasis on stakeholders’ interests, stock option pay seems to have no downside effect on executives because when stock price drops after the granted date of the option. Nevertheless, in Sanders’ findings, stock option pay will not necessarily lead to severe problems if the firm stays in a relevant stable environment. But for fluctuant environment, stock option pay behaves not as well as stock ownership.
Tosi, Werner, katz and Luis (2000) examined the possible relationships between firm size, firm performance, and CEO compensation by a meta-analytic review of the existing studies on the determinants of CEO compensation packages. Most of existing studies have put much emphasis on the issue of control, which is the extent to what CEOs are held accountable to company shareholders. In general, CEOs are
professional managers, receive compensation packages and act on behalf of firm shareholders. In their study, the agency theory was also applied. With a sample of 137 articles or manuscripts which have investigated the relationship between CEO
compensation and firm performance or firm size, they designed a model using CEO compensation as dependent variable, the other two as independent variables. And from the regression results, both firm size and firm performance have a significant influence on CEO compensation. To be more specific, changes in firm size and firm performance account for 5% and 4% of the variance in CEO compensation,
respectively. This finding is consistent with much previous literature. Therefore, a large proportion of unexplained variance in CEO compensation needs to be examined in the future for the development of empirical theory.
When most scholars choose to treat CEO and other executives as a whole when examining the relationship between executive compensation structure and firm performance, Carpenter and Sanders (2002) investigated the top management team (expect CEO) separately. In their opinion, top management team has received little empirical attention due to the assumption that CEO and the top management team have the isomorphic compensation structure. They argued that there is enough evidence supporting the different compensation structure among executives. They chose 250 firms randomly from S&P 500 at the beginning, from 1991 to 1995. And the sample size has been decreased to 199 companies due to incomplete data for executive compensation. In their model, firm size, geographic diversification and
CEO pay structure were introduced as control variables. Their results prove the existence of interplay between CEO compensation and the top management team compensation. When CEO compensation structure is positively related to firm performance, this relationship could be influenced by the top management team compensation. The findings of this paper highlight the importance of senior executive pay (except CEO), and contribute to the generalization of relationship between executive compensation and firm performance.
Bebchuk and Fried (2003) also examined executive compensation structure. Their focus is based on companies which have separate ownership and management, where managers may have substantial power to affect the design of their compensation structure. They found that executive compensation packages were often influenced by public perception. And more specifically, their results show a decline of $2.7 million in executive compensation after critical shareholder resolutions and negative
perceptions. From the perspective of the greater shareholder environment, public perception has been one of the most significant factors in deciding the executive compensation, especially the cash-related part. However, the growing number of public criticism over both compensation levels and sensitivity to the corporate performance, suggests that conflicts of interests and agency problems are to some extent inherent to executive compensation schemes. For example, base salary itself may result in risk. From the findings of Bernoulli (1954), there is an inversely relationship between executive’s reactions to risk and their current level of wealth.
When executives have a relatively higher base salary, they are more willing to take risky activities since they are less sensitive to risk. In Bebchuk and Fried’s
conclusion, both managerial power and rent extraction have a significant influence on executive compensation structure, and they are related to firm governance. And the extent of the influence depends on the extent to which company shareholders could be able to recognized such a problem. Therefore, the analysis from financial economists on the degree of deviation of current executive compensation structure from the optimal contracting indeed could help solve this problem.
Burns and Kedia (2006) analyzed the impact of executive compensation structure on misreporting. To be more specific, they examined how managers use accounting practices to manipulate financial statements so that they could achieve compensation incentives. Firstly, they found 266 companies listed in S&P 1500 from 1995 to 2002 which were reported to have manipulated accounting figures for a nice-looking financial statement, and then compare them with those which do not have such behaviors. They found a positive and strong relationship between CEO's option sensitivity and the propensity to misreporting, indicating that when executives’ compensation is stock options, they are more willing to take misreporting actions since the characters of stock options make the executives get rid of downside risk. When they were found have illegal behaviors, the share price dropped, they could just give up exercising the options. The only loss for them is the value of options. Their paper again clarifies the conflict of shareholders and executive’s interests, and proves
that the use of stock options indeed encourages executives to conduct misreporting, and increases firm risk.
Coles, Daniel and Naveen (2006) also demonstrated the relationship between
managerial compensation and firm risk. In their paper, the sensitivity of CEO wealth to stock return fluctuations, which can be described by Vega, is the main
characteristic of executive compensation. They found that higher Vega will lead to higher firm risk because of riskier policy choices, such as investing more in R&D, borrowing more money from banks and so on. And delta is defined as the sensitivity of CEO wealth to stock price. Delta seems to align the interests of managers and shareholders closely together. The sample companies were also selected from S&P500, S&P Midcap 400 and 600, from 1992 to 2002. And they use data of base salary, cash bonus, and total compensation for the top five executives. Their findings show that there is a positive and significant relationship between stock price volatility and investments in R&D, and a negative and significant relationship between stock price volatility and capital expenditure. Moreover, their hypothesis was supported: when executive compensation structure shows higher sensitivity to stock price volatility, executives are more willing to perform riskier actions, such as investing in riskier property and using high financial leverage, which result in higher firm risk.
Cornett, Marcus and Tehranian (2008) once examined how the firm performance was influenced by governance structure and incentive-based compensation while the
performance has been adjusted because of earnings management. They referred to both accounting and finance literature documents, which indicate that governance and compensation structure have an influence on both earnings management and financial performance. They choose 100 largest companies listed in S&P 500, from 1994 to 2003. The control variables that may affect firm performance include CEO’s age, gender, tenure, board size and so on. And their results show that earnings
management indeed influences the managed performance. More specifically, the influence of governance structure on firm performance will be magnified, while the influence of incentive-based compensation becomes smaller. On the other side, earnings management is related to agency problem while executives and shareholders have asymmetric information. When incentive-based compensation is applied,
executives are motived to use discretionary accruals in earnings management.
Devers, McNamara, Wiseman and Arrfelt (2008) also investigated the relation between CEO equity-based compensation and firm’s risk-taking behavior. In their paper, Behavioral Agency Model, Agency Theory, and Prospect Theory are applied. They collected data about all publicly traded manufacturing firms in the Execucomp database from 1992 to 2005. Therefore, their sample companies were limited to a specific industry: manufacturing. They argued that firms in manufacturing industry are able to provide a lot of information what contains firms’ strategic risk. Except for all other executive compensation, they only use CEO compensation that it is
interesting and remarkable findings. Firstly, different elements in stock-based compensation play different incentive roles regarding to incentive properties and when the value of elements changes, incentive properties also make a change. Secondly, there is a big difference between the incentive properties of restricted stocks and stock options. The restricted stock compensation is negatively related to firm’s strategic risk, while compensation of stock options has a positive relationship. Thirdly, a surprising finding which is contrary to previous research is that different elements of equity-based pay have different influence on CEO’s attitudes towards risk. Overall, they concluded that firm’s strategic risk could be affected by CEO equity-based compensation significantly, but the influence is more nuanced and complex than previous finding on executive compensation.
Overall, most of the pieces of literature above indeed indicate a strong relationship between executive compensation structure and firm performance. And both cash-based compensation and stock-cash-based compensation imply firm’s risk, which in turn, affects firm performance. Besides, most literature choose to use CEO compensation to represent the whole top management team due to data collection reason or isomorphic compensation structure reason. By examining pay-performance relationship,
companies could better understand compensation structure, manage relevant risk related to firm’s performance so that shareholders’ interest is guaranteed.
2.3 Disadvantages of long-term incentive pay (equity-based compensation)
Before moving into the main research of this paper---the long-term incentive pay, it is worthwhile to talk about the drawbacks of this equity-based compensation so that all factors could be taking into consideration when analysing the regression results.
Stock options are the most common type of equity-based compensation. Sanders (1999) suggested that the main disadvantage of using stock options as long-term incentives is that executives would not bear the risk when stock price drops since they could choose not to exercise the option. The only risk they have to take is the value of the option. In such a situation, they may take riskier measures which are detrimental to firms.
Besides, Hall (2003) proposed that when executives are reward with equity-based compensation, they may choose to cash out their holdings at the expense of company shareholders. Specifically, executives are able to adjust their firm’s accounting
methods to maximize short-term accounting figures so that the share price reaches the top in a short time and then they cash out their holdings. This is achieved by
sacrificing the firm’s long-term interest for individual short-term gains.
Furthermore, in the paper of Bryan, Hwang and Lilien (2000), another type of equity-based compensation, the restricted stock, was examined. They found that the
restricted stock failed to make risk-averse managers accept risky, but profitable investment opportunities, because of the linear payoff of the restricted stock.
Overall, there must be drawbacks accomplished by using stock-based long-term incentives. When the share price deviates significantly from the expected values, a great psychological let-down appears among the executives, which cannot be ignored by shareholders of the company.
3. Hypothesis development
From the literature, we can see that there is a high possibility that the correlation between executive compensation and firm performance exist. And according to the agency theory, asymmetric information gives the executives opportunity to maximize their own interests in a way of sacrificing company’s interests. Therefore, a pattern of rewards which can align the interests of executives and shareholders together closely would be a good solution. And the equity-based compensation seems to a suitable choice. When executives themselves become the stakeholder of the company, they care more about the firm performance which can affect the stock price significantly. The first hypothesis would be formed as follows:
H1: long-term incentive pay (equity-based compensation) in CEO’s
Furthermore, except for long-term incentive pay, executive compensation package also has base salary and short-term incentives (bonus). These two are case-based compensation and according to the tournament theory, executives are already in such a high position that the chance of been promoted for them is much less than low level employees. And case-based compensation seems to be not as attractive as equity-based compensation for them. The results of Ozkan (2011) indicate that 10% increase in shareholders’ return corresponds to only 0.75% increase in UK CEOs’ case-based compensation, which means that CEOs are poorly sensitive to the increase in case-based compensation and firm performance is almost irrelevant to the change in CEOs’ case-based compensation. Besides, when executives are given a high salary, they may even become more afraid of taking risky measures, such as investing more in R&D, so that they would not be blamed by shareholders if those investments end in failure and have to give up such a high salary job. However, those risky activities, which may cost a lot in current year, maybe generate a lot of revenues for the company in the future. Therefore, the second hypothesis would be defined as follows:
H2: Base salary and bonus (cash-based compensation) in CEO’s compensation package will be negatively related to company performance.
4. Research method and design
4.1 Research method
The main research method in this paper is archival research. It is a type of primary research by referring to a lot of original archival records to gain experience of similar topics, such as how to construct the general research procedure, design the hypothesis and model. Besides, learning the outcomes of other archival records is also helpful.
In this paper, the dependent variable is firm performance (which is related to firm risk). And profitability could be seen as the measure of performance (Daft, Sormunen & Parks, 1988). By referring to a lot of previous research, there are a lot of methods which can be used to measure the profitability, such as return on assets (ROA), Tobin’s Q or stock price return. All of them have advantages and disadvantages. Fisher & McGowan (1983) and Benston (1985) once proposed that ROA is unable to provide detailed information about economic rates of return, and Tobin’s Q can better forecast firm’s growth opportunity, rather than its performance. Besides, methods like stock price return, could be easily affected by other external factors, such as overall economic environment and spread gossip. Despite of its potential flaws, ROA has been adopted by quite a lot of literature. Therefore, in this paper, ROA was chose since it is the most consistent measure of profitability. And ROA is calculated as follows: ROA = Net Income after taxes/total assets.
The independent variables in this paper include executive cash-based compensation, equity-based compensation, firm size, leverage ratio, short-term debt ratio and long-term debt ratio. Among them the equity-based compensation is the most complex part. In general, the equity-based compensation is composed of stock options, the restricted stocks, preferred stocks and performance-based shares. Besides, Buck, Bruce, Main and Udueni (2003) once introduced a new type of long-term incentive plans (LTIPs). This long-term incentive plans is also composed of stocks and options. However, these incentives would only be granted when executives achieve their setting targets. This is quite different since in most companies, executives would be given long-term incentives before they achieve targets. The value of options could be calculated directly by the Black-Scholes formula.
4.2 Data collection
The data used in this paper could be drawn from Wharton Research Data Services (WRDS). It is a well-known and professional database, providing a variety of
business-related data. For example, executive compensation data, especially the CEOs compensation data, could be found from the Compustat’s ExecuComp Database. In this paper, CEOs compensation data would be used rather than each member of executives. The reasons are as follows: CEOs compensation are representable since they normally have a complete and detailed compensation structure than other executives, and CEOs are more likely to be motivated by stock-based incentive
income, total assets, market value, long-term and short-term debts, would be drawn from the Compustat Global Database. Moreover, one thing needs to mention is that all CEO’s compensation data would be represented in thousands of dollars, while all other performance-indicator data would be represented in millions of dollars.
When selecting sample companies, one thing needs to make sure is that they are comparable. All relevant factors are necessary to be taken into consideration, such as scale of company, market position, industrial characteristics and so on. This study utilizes sample time series compensation data on companies listed in the Standard & Poor’s 500 for several reasons. Firstly, companies in the S&P 500 are representative and well-known, they are large and have common stock listed on the NYSE and NASDAQ. Besides, as number one economy in the world, companies in the United States are more likely to have mature executive compensation structure, stable return on assets. Therefore, it is reasonable to choose companies from the S&P 500. Besides, companies are selected from all kinds of industry, including financials, information technology, health care, energy and materials. The reason for that is because when limiting the research object in one specific industry, there are some problems with the generalization of findings. Base on previous literature, while a sample is too large, there may be a bias due to time-varying factor loadings; and while a sample is too small, it may result in insufficient statistical power or efficiency. Therefore, originally the estimated specific sample size was 200 public firms in this paper, accounting for nearly half of S&P 500 companies; however, during the procedure of collecting
relevant data, those companies which could not provide complete data of CEO compensation for one or more years, or firm performance such as net income, market value, short-term debt and long-term debt information, would be excluded from the analysis. And in the last, 129 companies are left. The sample size of 129 companies seems to be reasonable since it could provide sufficient statistics power. For the time interval of data, the 2008 financial crisis would not be included in this paper to examine the company performance since the crisis may have an significant effect on the company performance. Specifically, during the crisis, a lot of companies went bankrupt, and this may influence the accounts receivable of those selected companies’ return in this paper. Besides, it seems to be more difficult for companies to find profitable projects during the financial crisis. And rolling regressions with a window of 5 years, from 2002 to 2006, is estimated due to problems with the time-variation in factor loadings. The time interval is set before the 2008 financial crisis When the time interval for the research is too long, company’s policies for executive compensation may have changed a lot, which will lead to a bias of our test results. And when the time interval for the research is too short, there will not have enough observations for the research. Therefore, a 5-year time interval seems to reasonable and acceptable.
More importantly, it is worthwhile to test whether this study is feasible. From previous literature, executive compensation data could be found from Wharton Research Data Services directly or annual reports in their official website. Moreover,
company financial data could also be found from the WRDS directly. Therefore, it seems that there is no need to worry about the feasibility of this study.
4.3 Design of the model
H1: long-term incentive pay (equity-based compensation) in CEO’s compensation package will be positively related to company performance.
𝑅𝑂𝐴$= 𝛽(+ 𝛽*∗ 𝐸𝑞𝑢𝑖𝑡𝑦 − 𝑏𝑎𝑠𝑒𝑑 𝑐𝑜𝑚𝑝𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛$=*+ 𝛽>∗ 𝑆𝑖𝑧𝑒$+ 𝛽A∗ 𝐿𝑅$+ 𝛽C∗ 𝑆𝑇𝐷$+ 𝛽F∗ 𝐿𝑇𝐷$+ 𝜀
H2: Base salary and bonus (cash-based compensation) in CEO’s compensation package will be negatively related to company performance.
𝑅𝑂𝐴$= 𝛼(+ 𝛼*∗ 𝐶𝑎𝑠ℎ − 𝑏𝑎𝑠𝑒𝑑 𝑐𝑜𝑚𝑝𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛$=*+ 𝛼>∗ 𝑆𝑖𝑧𝑒$+ 𝛼A∗ 𝐿𝑅$+ 𝛼C∗ 𝑆𝑇𝐷$+ 𝛼F∗ 𝐿𝑇𝐷$+ 𝜀
where:
ROA is the return on assets (used to measure company performance);
Equity-based compensation is the ratio of long-term incentive pay divided by the total compensation;
Cash-based compensation is the ratio of base salary and bonus divided by the total compensation;
Size= logarithm of total market value, LR = debt/equity (leverage ratio); STD = short-term debts to total assets, LTD = long-term debts to total assets; (Size, LR, STD, LTD are control variables)
and ε represents error term, which is used due to correlation of different variables and heteroscedasticity.
Since all other variables are expressed as ratios and to make it comparable, I define Size as the logarithm of total market value. The three ratios of total debt to total asset (LR), short-term debts to total assets (STD), long-term debts to total asset (LTD) could be categorized as financial leverage. In the paper of Abor (2005), Ebaid (2009), Salim and Yadav (2012), they also use these three ratios as control variables when the dependent variable is return on asset. Financial leverage could be seen as an important criterion to decide the return on asset. The higher the financial leverage, the more debt a company should have, which means that the firm may have to pay a large amount of interest every year, resulting in a lower return on asset. Therefore, financial leverage is supposed to be negatively related to return on asset. Moreover, previous research suggest that firm’s size may also influence firm performance. Larger firms may have a lot of capacities, and benefit from the economies of scale (Ramaswamy, 2011; Frank and Goyal, 2013; Jermias, 2008). Therefore, the larger a firm, the better a firm may perform. In this paper, by including the size as a control variable, different firms’ operating environment could be controlled as well (Salim and Yadav, 2012). A firm’s size could be determined by its total market value. In essence, the control variables themselves are not the main research objects for experimenters and they are held the same throughout the experiment, although they may have an effect on the
experimental results. The control variables mainly help test the relationship between two other variables and explain the possible results.
The most important part in my model is the ratio between equity-based incentive pay and total compensation, which is a method followed from Mehran (1995). In his paper, the equity-based incentive pay could be defined as the sum of the value of grants of new stock options, restricted stocks, preferred stocks and performance-based shares. And due to the lagged effect, CEO’s compensation structure in year t-1
corresponds to the company performance (return on assets) in year t. The null hypothesis is 𝛽* > 0, which means that equity-based compensation is positively related to the company performance. And to make it comparable, a second hypothesis is made and I estimate that 𝛼* < 0 since the risk-averse managers who receive high salaries are not willing to invest in those projects which seem to be risky but may generate a lot of revenues in the future.
5.Emperical results
5.1 Sample characteristics
Explanation dependent variable: company performance
Table 1 Company performance statistics
Panel A: descriptive characteristics
Variables MIN MAX MEAN SD P25 median P75
Assets - Total 883.650 1459737.000 44502.359 131719.858 5710.009 14182.500 31546.250 Net Income (Loss) -17462.200 21133.000 1494.026 2614.142 338.561 738.500 1627.375
ROA -0.257 0.247 0.068 0.054 0.029 0.067 0.103
Market Value - Total - Fiscal 1038.296 310218.720 27173.748 42952.596 6566.794 12965.584 25912.850
In this paper, return on assets was used to measure the company performance. The table 1 shows the summary descriptive company performance statistics of 645 observations of selected American companies, with a time interval from 2002 to 2006. It includes variables of total assets, net income or loss, return on assets and total market value. From the data, we can see that the average assets are $44502.359 million, average net income is $1494.026 million, average ROA is 6.8%, and average market value is $27173.748 million. There is a huge difference between the minimum and maximum for different variables of different companies, such as total assets and total market value. Furthermore, except for the median, there are also 25th percentile and 75th percentile, which indicate the value below which the 25% and 75% of the observations could be found. According to 75th percentile, we can infer most selected
companies’ performance in this paper. Specifically, their total assets are about $44502.359 million, net income are about $1494.026 million, ROA are nearly 6.8%, and total market value are about $27173.748 million.
Explanation independent variable: CEOs’ compensation structure
Table 2 Compensation structure statistics
Panel B: descriptive characteristics
Variables MIN MAX MEAN SD P25 median P75
Salary ($) 0.000 5806.651 905.324 514.887 604.466 903.401 1056.875 Bonus ($) 0.000 16500.000 1503.180 1959.363 408.902 945.789 1906.080 salary+bonus 0.000 21473.073 2408.505 2282.038 1087.446 1891.667 2988.401 All Other Compensation 0.000 21698.983 414.764 1440.280 31.234 122.470 295.927 equity-based incentive pay 0.000 59916.936 3384.963 4972.454 605.274 1938.009 4238.477
Total Compensation 0.000 62621.340 7956.052 7596.378 3507.589 5495.190 9866.966
This table presents the summary descriptive compensation structure statistics of 645 observations of selected American companies, with a time interval from 2002 to 2006. It includes the CEOs’ detailed compensation structure: base salary, short-term cash-based bonus, long-term stock-based incentive pay and all other compensation.
From the data, we can see that average total compensation is $7956052, which is composed of average salary $905324, average bonus $1503180, average equity-based incentive pay $3384963, and average all other compensation $414764. The average
equity-based incentive pay is much higher than the average sum of salary and bonus, which means that equity-based incentive pay accounts for a large proportion of total compensation. Besides, from the 75th percentile, we can infer most selected
companies’ executive compensation structure in this paper. Specifically, their base salary is about $1,056,875, cash bonus is about $1,906,080, all other compensation is nearly $295,927, equity-based compensation is about $4238477, and total
compensation is about $27173.748 million. Therefore, equity-based incentive pay of 75% selected companies in this paper is much higher than the sum of salary and bonus, which is consistent with the average results.
Table 3 All variables statistics in the regression model
Variables MIN MAX MEAN SD P25 median P75
ROA -0.257 0.247 0.068 0.054 0.029 0.067 0.103 Equity-based compensation 0.000 1.000 0.370 0.274 0.143 0.361 0.584 Cash-based compensation 0.000 1.000 0.400 0.244 0.239 0.361 0.501 Size 3.016 5.492 4.147 0.492 3.814 4.114 4.416 LR 0.000 19.173 0.815 1.404 0.211 0.448 0.913 STD 0.000 0.388 0.036 0.049 0.002 0.021 0.050 LTD 0.000 0.583 0.152 0.114 0.062 0.136 0.215
This table presents the summary descriptive statistics of all variables in the regression model of 645 observations of selected American companies, with a time interval from 2002 to 2006. The year chosen for factors like return on the assets, firm size, total debt to total asset, short-term debts to total assets, long-term debts to total asset is one year after which long-term and short-term compensation are awarded due to the lagged effect of compensation on firm performance. For example, return on assets of year 2003 corresponds to the compensation of year 2002. This table includes relevant statistics for the four control variables: size, leverage ratio, short-term debt ratio, and long-term debt ratio. Size is the logarithm of total market value, and from the table, we can see that the maximum leverage ratio is 19.173, far exceeds the value of mean 0.815 and 75th percentile 0.913, which means that in my sample, some companies with extreme large leverage ratio were selected. Besides, the average long-term debt ratio is much higher than short-term debt ratio, which means that most companies are more willing to choose long-term debt. Glazer (1994) found that when companies issue long-term debt, rather than short-term debt or no debt, their product market behavior would be strongly affected by firm’s strategic considerations, which in turn, influence firm performance. Therefore, it is reasonable to use financial leverage as control variables and it is worthwhile to examine the possible effects that financial leverage may have on the firm performance.
5.2 Regression results and correlation analysis
Table 5 Regression of company performance on CEO compensation
Independent variables model 1 model 2
Equity-based compensation 0.0284*** t (3.42) p(t) (0.001) Cash-based compensation 0.0432*** t (4.90) p(t) (0.000) SIZE 0.0266*** 0.0280*** t (5.73) (6.15) p(t) (0.000) (0.000) LR -0.00618*** -0.00569*** t (-3.14) (-2.93) p(t) (0.002) (0.004) STD -0.120** -0.135*** t (-2.53) (-2.90) p(t) (0.012) (0.004) LTD -0.0177 -0.0369 t (-0.73) (-1.57) p(t) (0.265) (0.117) Adjusted-R2 0.157 0.176 α -0.0410** -0.0506** F 20.62*** 23.50*** N 645 645 t statistics in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01
This table shows the ordinary least squares estimates of regressing company performance on CEO different compensation structure. The sample consists of 645 observations of selected American companies, with a time interval from 2002 to 2006. T-statistics is the ratio of the difference between the estimated value of a parameter and its hypothesized value to its standard error. And the adjusted R2 is a tool used to measure the accuracy of the model to see if the observations fit in the model, and fluctuates between 0 and 1. The closer the R2 to 1, the more accurate the model fitting the observed numbers. It is generally believed that when the adjusted R2 in the economic model is greater than 0.1, the results of the model can be considered acceptable. From the table, we can see that the adjusted R2 for the two estimated models are 0.157 and 0.176 separately, indicated that the observations fit in the model well. Besides, the F values are 20.62 and 23.50 respectively, both of which are
significant at the 1% level, indicating that the model is applicable.
More specifically, in model 1, the p-value of the independent variable equity-based compensation is 0.001, which is lower than the 1% significance level, and the coefficient of this variable is 0.0284, indicating that equity-based incentive pay in CEO’s compensation package is positively and significantly related to company performance. And 1% increase in the percentage of equity-based incentive pay in CEO’s compensation package will lead to 0.0284% increase in return on assets. The first hypothesis is supported. In model 2, the p-value of the independent variable cash-based compensation is 0.000, which is also lower than the 1% significance level, and the coefficient of this variable is 0.0432, indicating that cash-based pay in CEO’s
compensation package is positively and significantly related to company performance. And 1% increase in the percentage of cash-based pay in CEO’s compensation package will lead to 0.0432% increase in return on assets. Therefore, the second hypothesis is rejected. And it is more surprising that the coefficient of the cash-based compensation is even larger than that of equity-based compensation, which means that when there is equal increase in both cash-based compensation and equity-based compensation, firm performs better with the increase in cash-based compensation. This result is contrary to my estimates based on the tournament theory. In the second part of this paper, I once proposed that executives such as CEO, are hardly attracted by solely higher salary since they are already in such a high position that companies should put more emphasis on compensation which can align
executives’ interest with shareholders’ interest together closely. However, from the regression results it is obvious to see that CEOs are motivated by cash-based
compensation much more and the company performs better. There may be different explanations for it. Firstly, the tournament theory also applies to executives such as CEOs, not just employees at the ground level. Secondly, most of the CEOs observed here are risk-averse. Although they are provided stock-based incentives, they are still not willing to invest in risky but profitable projects. Last but not least, some of the CEOs choose to cash out their holdings at the expense of firm’s interest in the future. Moreover, regarding control variables, the p-value of SIZE is 0.000 in both models, which are lower than the 5% significance level, and the coefficients of this variable are 0.0266 and 0.0280 respectively, indicating that SIZE is positively and
significantly related to company performance. The higher the total market value of a company, the better the company performs. And this is consistent with my estimates, which explain it by benefits brought by economies of scale, which is no available to relatively smaller firms. This result is also consistent with the findings of
Vinasithamby (2015). In his paper, except for economies of scale, there existing other reasons helping explain this phenomenon. For example, larger firms may have more borrowing capacity than smaller firms and therefore, their bankruptcy costs will decrease, which also improves their return on assets. The rest three control variables are described as financial leverage. From the table, we can see that all coefficients of these three ratios for both models are negative. This result is consistent with the previously set hypothesis. And the result of Vinasithamby (2015) also proves it. In his result, the coefficient of total debt ratio and firm’s profitability is -0.1105, which means these two variables have 11.05% significant negative relation. And the perking order theory helps explain it. The theory clarifies that in general profitable companies are not willing to borrow too much money so that they could keep a low leverage, compared to less profitable or even unprofitable companies. By doing it they could be rid of the risk of decreasing operating efficiency resulting from financial distress cost of debt. Besides, borrowing too much can be seen as a signal that the company is in a dilemma and decrease people’s expectation of their future operating profitability, which is detrimental to a company’s share price or return on assets.
Table 4 Correlation between different variables
ROA Equity-based Compensation Cash-based Compensation Size LR STD LTD
ROA 1 .225** .202** .275** -.251** -.164** -.235** Equity-based compensation .225** 1 -.407** .158** -.114** -.146** -.243** Cash-based compensation .202** -.407** 1 .009 -.006 -.043 .016 Size .275** .158** .009 1 -.077 .091* -.251** LR -.251** -.114** -.006 -.077 1 .321** .558** STD -.164** -.146** -.043 .091* .321** 1 .115** LTD -.235** -.243** .016 -.251** .558** .115** 1
*significance level =5% **significance level =1%
The correlation coefficient R reflects the statistics of the linear correlation intensity and direction between two random variables. It is helpful to provide a predictive relationship between different variables that could be exploited in practice. The value of R ranges from -1 to +1. The positive and negative values of R represent the
direction of linear correlation between two random variables. When R is larger than 0, these two random variables are positively related; and when R is smaller than 0, these two random variables are negatively related; when R equals to zero, then there is no correlation. The absolute value of R indicates the level of intimacy of linear
correlation between the two random variables. The closer the absolute value of R is to 1, the higher the level of intimacy. The closer the absolute value of R is to 0, the lower the level of intimacy.
From the table, we can see that the correlation coefficient between return on assets and equity-based compensation is 0.225, and thus these two variables are positively
and significantly related (at 1% significance level). And the correlation coefficient between return on assets and cash-based compensation is 0.202, and thus these two variables are also positively and significantly related (at 1% significance level). The previous correlation coefficient is larger than the later one, which means that stock-based compensation has a closer linear relationship with firm performance than based compensation. This result is opposite to the regression results, in which cash-based compensation has a greater effect on firm performance than equity-cash-based compensation. Besides, firm size has the largest positive correlation coefficient with firm performance among all independent variables, which means that firm size affect the firm performance the most. This result is in line with a lot of findings which also examine the relationship between firm size and firm performance. Besides, all other three control variables (financial leverage) are negatively related to firm performance at 1% significance level, which is consistent with regression results from table 3 and the original estimates. Moreover, regarding independent variable, there are also correlations. For instance, the equity-based compensation and cash-based
compensation are negatively related at 1% significance level, and the coefficient is -0.407, which means that 1% increase in the equity-based compensation will lead 0.407% decrease in the cash-based compensation. It seems reasonable since when the board of directors decide to increase equity-based executive compensation, they would choose to lower cash-based executive compensation. Besides, the absolute values of correlation coefficient between equity-based compensation and the three control variables are much larger than cash-based compensation and the three control
variables, which means that equity-based compensation has a closer linear
relationship with control variables than cash-based compensation at 1% significance level. Furthermore, the three control variables are also significantly related. For example, firm size is positively related to short-term debt ratio, negatively related to total leverage ratio and long-term debt ratio. This result could also be explained by the perking order theory. Larger companies are usually not willing to borrowing too much, which leads to a lower total leverage ratio and long-term debt ratio. And thus, larger companies prefer to choose short-term debt ratio when they encounter with much investment in profitable projects, or research and development. Besides, the short-term and long-term leverage are both positively related to the total leverage ratio. It is because that the use of no matter short-term debt or long-term debt ratio would lead to an increase in total leverage ratio. The correlation between control variables could be described as multicollinearity. Therefore, correlation analysis could help detect whether there exists collinearity between different variables; and if the multicollinearity problems exist, the effect of multiple linear regression would be influenced seriously, and even make the regression results not statistically significant. But in general, only when the correlation coefficient of explanatory variables (control variables) is above 0.8, then there may be multicollinearity. The result above shows no multicollinearity. Nevertheless, one thing needs to be mentioned is that correlation analysis is only a single factor analysis, and the final correlation results need to be determined by regression analysis.
6. Summary and Discussion
Although the level of equity-based compensation always gives rise to controversy among shareholders of the company, it is universally accepted that equity-based executive compensation could play a role of mitigating the interest conflict between managers and shareholders (Guay, Core & Larcker, 2001). And Finkelstein and Hambrick (1988) emphasized once again that the importance of compensation in incentivizing managers has been recognized by theorists for a long time. Despite of it, there is not enough research on investigating how to reward managers, especially executives like CEOs after examining the relationship between executive
compensation and firm performance. The most recent theory is focusing on how to reward employees. Employees are in ground level of a company; therefore, it is of great difference on how to reward top managers and bottom employees. This helps explain the necessity of investigating the incentive schemes for executives.
In this paper, I investigate the relationship between executive compensation structure and firm performance. There is a positive relationship between equity-based
compensation and firm performance, which supports my first hypothesis. Besides, there is also a positive relationship between cash-based compensation and firm performance, which is inconsistent with my second hypothesis. And it is more surprising that the coefficient of the cash-based compensation is even larger than that of equity-based compensation, which means that cash-based compensation has a greater influence on CEOs to perform better than equity-based compensation. The possible reason could be that most of the CEOs chosen in my sample are risk-averse,
they do not invest in risky projects which may be profitable in the future, resulting in a relatively low return on assets in the next year. Besides, when executives are aligned with equity-based compensation, they maximize short-term accounting figures to raise up the share price by sacrificing the interest of shareholders in the future and then cash out their equity holdings soon. This kind of action could be prevented by restricted stocks, which limit the time of cashing out equity holdings (maybe five years later). However, Bryan, Hwang and Lilien (2000) found that the restricted stock failed to make risk-averse managers accept risky, but profitable investment
opportunities, because of the linear payoff of the restricted stock. In the next, I also examine the relationship between firm performance and four control variables: firm size, total debt to total asset (LR), short-term debts to total assets (STD), long-term debts to total asset (LTD). SIZE is positively and significantly related to company performance, which indicates that the higher the total market value of a company, the better the company performs. And this is consistent with my estimates, which explain it by benefits brought by economies of scale, which is no available to relatively smaller firms. The rest three control variables could be catalogued as financial
leverage. And total debt to total asset, short-term debts to total assets, long-term debts to total asset are all negatively and significantly related to company performance. This is also consistent with the previously set hypothesis. And the result of Vinasithamby (2015) also proves it. In general, profitable companies are not willing to borrow too much money so that they could keep a low leverage, compared to less profitable or even unprofitable companies. By doing it they could be rid of the risk of decreasing
operating efficiency resulting from financial distress cost of debt. Besides, borrowing too much can be seen as a signal that the company is in a dilemma and decrease people’s expectation of their future operating profitability, which is detrimental to a company’s share price or return on assets. Then I test the correlation between different variables. Return on assets is positively and significantly related to both stock-based compensation and cash-based compensation (at 1% significance level). But the coefficients indicate that stock-based compensation has a closer linear
relationship with firm performance than cash-based compensation. Besides, firm size has the largest positive correlation coefficient with firm performance among all independent variables, while other three control variables (financial leverage) are negatively related to firm performance at 1% significance level, which is consistent with regression results from table 3. Moreover, regarding independent variable, there is also correlations. And equity-based compensation has a closer linear relationship with control variables than cash-based compensation at 1% significance level. Besides, the three control variables are also related. The correlation between control variables (which are explanatory variables) could be described as multicollinearity. The result in table 4 shows no multicollinearity between explanatory variables. In fact, the correlation analysis is only a single factor analysis, and the final correlation results need to be determined by regression analysis.
7. Limitations and improvements
When talking about the generalizability about the findings, it is necessary to take the limitations of a paper into consideration. And there is no doubt that there is still room for improvement of this paper.
Firstly, the equity-based compensation is measured as a whole in this paper. However, the equity-based compensation is composed of several parts: the stock options, the restricted stock and performance shares. If the measurement is based on relationship between firm performance and these separate components, then there might be a different result. Finkelstein and Hambrick (1990) found that it was still popular for practitioners and scholars to use stock options as equity-based pay for a long time. However, Hall and Murphy (2002) proposed that there is a growing voice of using restricted stock as equity-based pay instead of stock options by practitioners and scholars due to an increasing number of company scandals and the implementation of the SOX Act. Therefore, for future research, I suggest that it is worthwhile to examine the relationship between firm performance and the restricted stock to see if there is a different result when the use of restricted stock keeps growing.
Secondly, randomly chose companies may cause bias. If sample companies are limited to one or several specific industries, such as service industry or information technology industry, the findings may also differ. For instance, when examining the relation between CEO equity-based compensation and firm’s risk-taking behavior, Devers, McNamara, Wiseman and Arrfelt (2008) decided to use firms from
manufacturing industry as the research object. Although they acknowledged that focusing on one specific industry may impose restrictions on generalizing their findings to firms in other industries, they insisted that manufacturing firms are those firms which take risky investments frequently, when focusing on them, their related measurements are more reliable and the corresponding results are also more
persuasive. Therefore, for future research of relationship between firm performance (when taking firm’s risk-taking behavior into consideration) and executive
compensation, I suggest to choose firms of one or two specific industries where firm performance could be measured more accurately and timely, such as manufacturing firms and real estate firms. And then compare it with results for firms coming from a variety of industries. Furthermore, one interesting phenomenon found by Barkema and Gomez-Mejia (1998) is that the data of almost all empirical research on executive pay is drawn from American companies, and then generalize their results to the rest of the world. It is contradictory since not all countries have the same economic system and market regulations. However, when taking firms from other countries as sample, it is often difficult to find enough observations. Therefore, it is worthwhile to find a balanced solution for this dilemma.
Thirdly, in this paper financial leverage is used as control variables. In some others’ papers, CEO’s age, gender, or the percentage of shares owned by CEO could also be used as control variables (Mehran, 1995). Besides, from the regression results, we can see that there is indeed relationship between financial leverage and CEO stock-based
