CEO PAY: THE EFFECT OF USING COMPENSATION CONSULTANTS FOR COMPENSATION BENCHMARKING
Master thesis, MSc International Business and Management University of Groningen, Faculty of Economics and Business
August 30, 2011
OLIVER KOTKAS Student number: 1940406
Trummi 38A 12617, Tallinn, Estonia Telephone: +372 55 66 8880 e-mail: o.k.kotkas@student.rug.nl
Supervisor Dr. K. van Veen
Acknowledgment: I would like to gratefully acknowledge dr. Kees van Veen for his guidance and helpful comments throughout the process of writing my thesis.
Additionally I would like to express my gratitude to my family who have enabled me to follow the MSc studies in the University of Groningen.
ABSTRACT
This research examines the effect of compensation benchmarking in conjunction with the use of compensation consultants on CEO pay using a sample of 101 companies from the S&P 500 for fiscal years 2007 and 2010. The use of compensation consultants and the process of compensation benchmarking are two controversial topics among scholars, as critics often contend that both practices inflate executive pay. With this study I demonstrate that compensation consultants design the compensation peer groups in a way that enables CEOs to receive greater pay. The results indicate that using compensation consultants leads to higher CEO pay;
moreover, companies who switched from big to small compensation consulting agency between 2007 and 2010 can potentially inflate pay by benchmarking a higher target pay. Increased disclosure and transparency in the remuneration setting process has made significant changes to CEO pay, as the allocated executive pay has become more sensitive to company performance.
Keywords: CEO remuneration; compensation benchmarking; compensation consultants; peer group characteristics.
1. INTRODUCTION
There are a few issues in the history of modern corporation that have attracted the attention garnered by the executive compensation in the U.S. companies. Executive pay is a controversial and complex subject that has become an international issue debated among financial economists, as well as in the academic field, business press and the media. Most large multinational corporations (MNCs) rely on the expertise of executive compensation consultants that make recommendations on the pay levels to implement short-term and long-term incentive arrangements. Compensation consultants opine on the existing pay arrangements, as well as on accounting, tax and regulatory issues that pertain to executive remuneration design (Murphy & Sandino, 2010). Consultancy companies possess extensive knowledge about recent developments in international pay practices, which makes them capable of providing expert guidance to the compensation committees by tailoring the compensation schemes for the organization’s executives (Brancato, 2002). Compensation consultants have access to proprietary information and to more detailed industry compensation level tendencies than is generally publicly disclosed (Cadman, Carter
& Hillegeist, 2008). This allows the consultants to provide compensation committees with in depth analysis of the peer group’s compensation practices and therewith aid the firm in setting competitive pay levels. More generally, with the expertise and knowledge of the compensation consultants the firm has the potential to maximize shareholder value by designing the compensation package that aligns the interests of the managers with the interests of the shareholders (Gillan, 2001).
Top executives’ remuneration is a complex and controversial subject, as high levels of pay awarded to U.S. chief executive officers (CEOs) and to the top management have raised the question whether their compensation packages are
consistent with the shareholders’ interests (Jensen & Murphy, 1990). Critics have increasingly accused the consultants of being complicit in the alleged excess in compensation that include too many perks, hidden benefits, golden parachutes, lucrative pension deals and non-demanding performance criteria (Conyon, 2006). The accusations focus typically on the two primary sources of conflicts, which are called
“repeat businesses” and “other services” that could both lead the compensation consultants to favour the incumbent executives, when designing the remuneration packages (Murphy & Sandino, 2010). Other critics have argued that the flaws in CEOs’ pay arrangements and deviations from the interests of the shareholders are substantial and comprehensive, as the CEOs have a significant influence over their own pay resulting in a greater possibility to extract rents and to determine one’s own pay (Bebchuk & Fried, 2003).
The academic literature on executive compensation is proliferating, but scholars have not reached a consensus on why the increasing executives’ pay has outpaced the company performance over the past couple of decades (Faulkender & Yang, 2010).
There is some agreement upon the fact that the flaws in CEO pay arrangements are widespread and considerable, which is regarded as a sign of corporate governance failure and top executives’ abuse of power (Canyon, 2008; Bebchuk & Fried, 2003).
Others argue however that the level of executive compensation reflects the market equilibrium in which the board of directors optimally structures chief executives’ pay in order to motivate and retain the CEOs (Murphy & Zabojnik, 2004; Edmans, Gabaix & Landier, 2009; Kaplan & Rauh, 2009). More recently it has been found that the complexity and dynamism of compensation comparison groups gives rise to the overall upward movement of executive compensation, as the CEOs have been able to regularly “leapfrog” their compensation benchmarks by moving to the right tail of the
benchmark distribution and get larger than normative compensation increases (DiPrete, Eirich & Pittinsky, 2010). The critics of the peer group benchmarking argue that more powerful CEOs in cooperation with the compensation consultants have the ability to opportunistically choose peer firms in a manner that significantly inflates chief executives’ pay (Bebchuk & Fried, 2003). Moreover, the scholars contend that given the prevalence of benchmarking and the pay raises that are independent of the CEO or firm performance, has led to a continuous upward ratcheting of executive remuneration.
The use of compensation consultants has risen rapidly over the past few years, as they are frequently hired by the management board or by the compensation committee to assist them in designing the remuneration of chief executives (Higgins, 2007). Despite their widespread use at public firms, the empirical evidence about their effects on the composition of peer groups and as a result on the executives’ pay levels is nonexistent. Until recently there has been little public information available because U.S. firms were not required to disclose the use of compensation consultants, but as the concerns about executive pay levels have increased, the Securities and Exchange Commission (SEC) now requires all companies to provide a Compensation Disclosure and Analysis (CD&A) section in their annual proxy statement. The CD&A requires numerous disclosures by the compensation committee, including the use of compensation consultants, the list of companies in the peer group and the performance and pay targets that will be used in determining executives’ level of pay.
These new disclosure requirements enable provision of empirical evidence on the incentives of compensation consultants and their effect on the level of pay, as well as the sensitivity of CEO pay to company performance.
Previously the academic research on the topic of executive compensation has developed into two directions. The first field of investigation has examined the effect of the employment of compensation consultants on CEO compensation (Bebchuk &
Fried, 2003; Cadman et al., 2008; Conyon, 2006) and the second field has investigated the effect of compensation benchmarking in determining chief executives’ pay (Bizjak, Lemmon & Nguyen, 2011, Faulkender & Yang, 2010, Diprete et al., 2011). Despite the frequent criticism of compensation consultants’
practices (Bebchuk & Fried, 2006) and the widespread use of remuneration benchmarking (Bizjak, Lemmon & Naveen, 2008), it is surprising that the interaction of these two phenomena that are most often used for determining CEO pay, have never been examined together. When examining the occurrence of these phenomena together and investigating to what extent the change in disclosure requirements have affected CEO pay setting, this research will provide new evidence to the field of executive compensation by bringing clarity to the composition of CEO pay. More importantly, the underlying study potentially explains whether these two practices in conjunction inflate CEO remuneration or enable to create a compensation scheme that aligns CEO pay with the company’s performance.
This study contributes to the literature examining executive compensation in general and more specifically to the emerging literature on the role of compensation consultants in affecting compensation benchmarking. The setting of this research examines the role of compensation consultants in designing contracting schemes and the effect of compensation benchmarking in determining CEO pay. As the prior disclosure requirements about the use of compensation consultants have been very limited, the current research is among the first to examine their role in determining executive pay. This study also provides additional academic evidence on the
company and consultancy agency level and thereby seeks to clarify the general debate surrounding the executive compensation practices and the objectivity of compensation consultants’ practices. Consequently, the focus of this research is to bring more clarity into the topic of compensation benchmarking, more specifically by addressing the following main research question: how does compensation benchmarking affect CEO compensation among Standard & Poor’s 500 companies over the period of 2007-2010? Moreover, answers will be found to the following sub- questions: how does the composition of compensation peer groups affect the executives’ level of pay, and what is the effect of compensation consultants on compensation benchmarking? An overview of the hypothesized relationships between the variables can be seen from the conceptual model presented below (Figure 1). Hopefully this research offers new insights into the topic of compensation benchmarking and the use of compensation consultants and beneficial practical and theoretical implications can be drawn.
The remainder of the research paper is structured as follows. In section 2, I discuss the theoretical background and Section 3 provides the hypotheses. Section 4 provides the research design and the data sources. Section 5 presents the results of the empirical analysis, and in Section 6 I discuss the results and conclude with final remarks.
--- Insert Figure 1 about here ---
2. THEORETICAL BACKGROUND
2.1 Agency Theory
Before examining the empirical evidence presented in this research it is important to consider the pay-setting process and the underlying mechanisms that determine the level of executive pay. The standard economic theory of executive compensation is the principal-agent (shareholder - CEO/manager) model, which argues that firms seek to design the most efficient remuneration packages possible to attract, retain and motivate the executive officers (Laffont & Martimort, 2002). In particular, when the company has dispersed shareholders, and the ownership and management are separated, it may give rise to managers having a substantial power.
This recognition goes back to the theory of managerialism formulated by Berle and Means (1932), who argued that the directors holding office have almost the complete discretion in management and act in their own interest at the expense of the interests of the shareholders. Among the scholars this phenomenon is in modern economics recognized as an “agency problem”, first defined in the seminal work by Jensen and Meckling (1976). They define the agency relationship as a “contract under which one ore more persons (the principal(s)) engage another person (the agent) to perform some service on their behalf which involves delegating some decision making authority to the agent” (p. 308). More specifically, in the agency model the shareholders of the firm determine the executive remuneration. In practice however this is usually done by the compensation committee of the board that is working in conjunction with the compensation consultants and sets the pay on behalf of the shareholders.
2.2 Mechanisms Shaping Executive Compensation
Appropriately designed management compensation contracts are increasingly complex elements of the business environment (Lynch & Perry, 2003). Executive compensation arrangements are an outcome of arm’s-length contracting between executives trying to get the best possible deal for themselves, and boards that are seeking to ensure the best possible deal for the shareholder (Bebchuk & Fried, 2006).
The most noticeable aspect in the executive compensation in the U.S. is not the variation of pay across companies, but rather the upward ratcheting of managerial wages, as in the 1980s and 1990s the executive pay rose at a faster pace than the average wages and corporate earnings. Following Useem (1993), the upward ratcheting of executive remuneration coincided with corporations providing proportionately larger equity-based incentive schemes that had been justified as a closer alignment of the interests of managers with the interests of the shareholders.
Within the field of executive compensation there are alternative approaches studying the efficiency of contracting arrangements between the firms and their executives. In general the theories fall into two broad categories. The first emphasizes the changes in the demand and supply of managerial talent that have a profound effect on the executive pay. Following Himmelberg and Hubbard (2000) the supply of highly skilled CEOs who are capable of running large and complex firms is relatively inelastic, as an increase in the demand for skilled executives will increase overall compensation levels. Murphy and Zabojnik (2003) argue that general managerial skills have become more important in the modern firm, which is driving up executive pay. The second category sees the ability of managers in different occupations extracting rents, as Sørensen (2000) records that executives use their control of productive assets to manipulate corporate governance processes that could
protect the interests of the shareholders. In line with this argumentation, the managerial power approach argues that executives have substantial influence over their own pay. Moreover, the compensation arrangements have deviated from arm’s length contracting because the directors are influenced by the management or they are insufficiently motivated to insist shareholder-serving compensation (Bebchuk &
Fried, 2006).
Another key issue of debate over the mechanisms that shape the executive compensation is the question where does the governance come from and how well does it work. The proponent scholars have argued that it is the market and its discipline that will create the optimal shareholder value (DiPrete et al., 2010).
Meanwhile Bebchuk and Fried (2003) provide their critique of the optimal contracting theory by stating that there can’t exist a contract that would perfectly align the interest of the managers with the interests of the shareholders. Also there are scholars that are proponents of the pay for performance model who argue in opposition to managerial power theory that the variations in pay and changes in the remuneration over time is a market driven consequence of “superstar” labour markets (Jensen & Murphy, 1990; Hall & Liebman, 1998; Murphy & Zabojnik, 2004).
Academics have argued that chief executive talent is worth more in larger firms and this presumption will create incentives to hire and pay in a way that the managerial talent and firm size become positively correlated (Gabaix & Landier, 2008).
2.3 Compensation Consultants
US public companies typically employ outside consultants to provide input about the executives’ compensation (Bizjak, Lemmon & Naveen, 2000). In addition to advising on compensation, the consultants that tailor executives’ remuneration schemes also provide other non-executive consulting services to the firm such as
advice on pension plans, employee benefit plans and other actuary services. These non-executive compensation arrangements create an economic dependence on revenues controlled by executive officers and therewith form the premise for the critics in claiming consultants’ actions to be camouflaging managerial rents and instead of optimizing their pay the consultants are assumed to be rather justifying executives’ excessive pay (Bebchuk & Fried, 2003). These cross-selling interests give rise to providing biased advice in order to secure future revenues from other services provided to the organization (Bebchuk & Fried, 2003; Bebchuk & Fried, 2006; Waxman, 2007). According to this view compensation consultants recommend higher pay without requiring greater performance and due to these conflicting interests enable the executives to extract wealth from the shareholders through higher remuneration or through lower pay-performance sensitivity.
Critics allege that compensation consultants can favour the executive officers by providing the market information that is most suitable for justifying a high level of pay. According to the managerial power companies fail to index stock options that executives would not be rewarded for general market price increases. Thus the executives can benefit even when their own performance is mediocre relative to their industry peers. Following Bizjak et al., (2000) as most of the 100 largest U.S.
companies use peer groups in establishing managerial pay and set the compensation at or above the fiftieth percentile of the peer group, it will result in a steady increase of executive salaries. Critics argue that after the compensation committee has approved the compensation scheme, firms will use the industry-wide peer information provided by the compensation consultants to justify the compensation to the shareholders.
2.4 Compensation Benchmarking
Already in the early 1950s when the rates of external hiring of executives were low, scholars started to debate whether the executive officers were paid on the basis of comparisons to the external labour market or the internal labour market (Roberts, 1956). It was argued that the remuneration of executives is a combination of both and latest by the 1970s the setting of managerial pay often included institutionalized comparisons between the practices of company’s own executives with the practices of a list of companies that are comparable to the compensation director’s organization (Cook, 1981). In the last decades the recruiting of external executives has become more prevalent and as a result there has developed an explicit system for the use of benchmarking and labour market analysis for setting executive compensation (Khurana, 2002).
Contemporary benchmarking is regarded to have created a “standard model” for the process of compensation consultants determining executive pay. As a first step the compensation committees are to assess the current pay of the executive officers (the actual pay is not always obvious given the complex structure of the compensation package). The second step is to conduct a thorough labour market analysis, as the compensation committee has the task to specify whether the level of compensation of company’s management is reasonable and adequate. The reason for doing this is twofold. First, it is for ensuring that the level of pay would be competitive and not be “below market” in order to retain the “talent executives” and to prevent other companies from “stealing” their managers. Second, the market analysis is conducted to ensure that the remuneration wouldn’t be too high or too disproportionate by creating a balance between the salary, annual bonuses, incentives etc. The third step is to call in the compensation consultants that are to identify the
peer group. The compensation of executives in the peer group will provide the basis for determining the pay of the executives of the company in question (DiPrete et al., 2010). Widespread evidence of the use of compensation peer groups has been found, as about 96% of firms determine their top managers’ pay in this manner (Bizjak et al, 2008).
The majority of publicly traded companies generally target the various parts of compensation at the median pay level of the peer group, although in recent times it has become more prevalent that the biggest and most profitable corporations target their executives’ pay above the median (e.g. at the 75th percentile) (Bizjak et al, 2011). As a basis for analyzing the compensation practices within the firm, the compensation committee that is working in conjunction with compensation consultants use the information on remuneration practices at peer companies, which are generally firms with similar size that originate from the same industry. The composition of the peer companies however varies significantly among firms and industries. Due to this some scholars argue that managers may find ways to inflate their pay, resulting in that not being in accordance with the company’s performance (Bizjak et al., 2011; Faulkender & Yang, 2010).
While scholars provide evidence on the selection of compensation peers based on the labour market considerations, the ultimate composition of the peer group is based on joint discussions between the management board, the executives and the compensation consultants. Thus far scientific literature has provided contradicting evidence on the use of compensation peer groups. On the one hand the defenders of executive remuneration levels have argued that managers are not extracting rent because they are being compensated at their reservation pay, which is set by the market for corporate control and a functioning market for executive talent
(Holmstrom & Kaplan, 2003). The economic sociology provides an alternative basis for suggesting that rent capture can occur even in the firm-level competition for executive talent because the rent capture by one executive creates a higher benchmark that other executive officers can refer to as “market price”. These academics argue that the upward movement of pay supports rather the labour market pressure than governance as an explanation for executive salaries. As a result this influence will be magnified if the overcompensated executives will be chosen as peers because their wages are higher and would produce even higher compensation for firms that choose them to benchmark the level of compensation (DiPrete et al., 2010).
2.5 The Effect of Benchmarking on Executive Compensation
Seveal academics have recently documented a widespread use of peer groups and compensation benchmarking (Bizjak et al., 2008; DiPrete et al., 2010; Faulkender
& Yang, 2010; Holmstrom & Kaplan, 2003), but interpretation of these findings varies widely. Bizjak et al., (2008) found that the practice of benchmarking executive pay against peers is used to retain the executive officers by determining their reservation wage. A similar conclusion was made by Holmstrom and Kaplan (2003) who argued that the wages of executive officers are ultimately set by supply and demand, and that benchmarking practices are nothing more than looking at the market prices. There exists an extreme skew in the awards offered to the managers, as they went on to argue that “to deal with this problem, we need more effective benchmarking, not less of it” (Holmstrom & Kaplan, 2003, p. 19). In contrast Faulkender and Yang (2010) suggest that firms appear to select highly paid peers to justify their executives’ pay and will forgo lower paid executives in the same industry in favour of higher-paid executives outside the industry. These kinds of pay practises
are found to be more prevalent in companies with smaller peer groups, more busier directors, with the CEO serving as the chairman of the board and CEO having longer tenure in the board. Also Bizjak et al., (2011) have argued that companies can potentially use compensation peer groups to inflate managerial remuneration by choosing peers that are larger in size, choosing a high target pay percentile or by just choosing peer companies that employ higher compensation practices. More importantly they suggest that the structure of the peer groups seems to be constructed in a manner that biases executive compensation upward. In light of these opposing views, it can be assumed that competitive benchmarking and the use of peers can be used opportunistically to bias managerial compensation upward.
3. HYPOTHESES DEVELOPMENT
Following Murphy (1999) and Bizjak et al., (2011) firms generally tend to favour peer companies that are of similar size, active in the same industry and those that exhibit similar accounting performance and similar market-to-book values. Peer companies that are chosen from outside the company’s operating industry tend to be selected from industries that have higher stock return correlations with the firm’s own industry, have similar credit ratings and similar geographic or product diversity. More generally firms tend be drawn into the peer group that originate from the same industry which either supply or hire managerial talent from the company’s own industry. Companies that employ compensation consultants to determine executives’
remuneration however appear to select a greater ratio of highly paid peer companies to justify their executives’ compensation (Bizjak et al., 2008) than companies that determine their pay internally. Scholars have argued that compensation consultants have the tendency to please the management and therefore may seek ways to increase
CEOs’ overall pay levels. As a consequence this will inflate CEOs’ remuneration upward that is not equivalent to the increase in company’s performance. Based on the line of argumentation the following hypothesis is suggested:
Hypothesis 1: Firms that determine their CEO pay through practices of compensation benchmarking, for which they employ compensation consultants, have higher executive remuneration than companies that do not use compensation consultants’ services.
Scholars believe that there is a positive relationship between peers originating from other industries and the inflation of executive pay, but thus far they have been unable to explain the antecedents of this phenomenon, and have left it unresolved (e.g.
DiPrete et al., 2010; Bizjak et al., 2011; Faulkender & Yang, 2010). Although prior studies shed light on the conflicts surrounding the use of compensation consultants and its effect on executives’ pay, it is surprising that little has been done to find explicit evidence on the influence of the consultants on companies’ peer group.
Compensation consultants usually have the discretion to decide the list of firms that constitute the peer group. Several academics have however argued that due to the potential conflicts, a difference of interest among the board, the firm’s executives and the compensation consultants gives rise to the opportunity of composing the peer group based on economic consideration, but especially in a manner that will bias the compensation upward (Bizjak et al., 2008; Bizjak et al., 2011; Faulkender & Yang, 2010). Therefore compensation consultants, having the sole authority to determine the list of comparative firms, include in the peer group companies that originate from other industries. This in turn has a significant effect on inflating executive pay.
Moreover, based on the same economic consideration, it has been suggested that firms employing compensation consultants have started to target pay levels at higher
percentiles of the peer group compensation distribution with benchmarking the pay against peers that have higher wages (Bizjak et al., 2008). In addition it has been suggested that firms have the possibility to choose peer firms that are larger and have a significantly better performance and thus also higher compensation levels (Bizjak et al., 2011). Consequently, the following hypotheses are suggested:
Hypothesis 2a: Companies that use compensation consultants have a higher percentage of peers originating from other industries, than companies that don’t use compensation consultants.
Hypothesis 2b: Companies that use compensation consultants have a higher target pay level, than companies that don’t use compensation consultants.
I will also propose, that compensation consultants’ influence on peer group characteristics changes according to who hires the compensation consultant. Properly structured peer groups provide useful information to the board of directors for determining the compensation levels that are necessary for retaining and motivating top executives (Bizjak et al., 2011). However the compensation consultants that are giving their recommendations about executive pay are often retained by the company’s management board rather than by the compensation committee. By working for and with the head of the human resources, CFO or the CEO, it creates obvious conflicts of interests known as “repeat business”, since they make pay recommendations for the managers that hire them (and could hire them for future businesses; Murphy & Sandino, 2010). Moreover, it gives rise to composing the peer group in a manner that would bias the compensation upward that could be done by camouflaging pay, such as including companies from other industries with greater sales, as well as aiming higher pay levels at compensation peer companies (Bebchuk
& Fried, 2003; Faulkender & Yang, 2010). Compensation consultants seek to ensure
the continuity of remuneration consulting businesses with the existing client firms.
Moreover, since they recommend the compensation to the individuals that have hired them, it could be argued that under those circumstances they design the peer group in a manner that would inflate CEO pay. The theoretical argumentation discussed above leads to suggest the following hypothesis:
Hypothesis 3: The positive effect of the use of compensation consultants on the peer group characteristics is stronger when the consultant is hired by the management board as opposed to being hired by the compensation committee.
Furthermore, compensation consultants provide non-compensation related businesses like consultancy on human resource management, internal control, insurance, financial and risk management (Kabir & Minhat, 2010). These fees paid to consultants for “other services” are often significantly larger than the fees for designing the compensation package. This causes another conflict of interests, because the decision to provide company-wide consulting services are made or influenced by the same managers, whose compensation they have previously determined (Murphy & Sandino, 2010). In order to secure these “other services” for the consultancy agency, the consultants are suggested to find ways to bias firm’s remuneration benchmarking processes that would grant the firm’s CEO a greater wage. By including companies from other industries that are greater in size and have a better profitability, as well as setting the compensation benchmark higher, the consultants can inflate CEO pay irrespective of the company’s performance. More specifically, employing biased compensation consultants that provide other services inside the organization and who are not independent of the board’s actions, will lead the firm to employ compensation benchmarking practices whereby consultants include companies from other industries and target a higher compensation level at
company’s peers. Based on the theoretical argumentation discussed above, the following hypothesis is proposed:
Hypothesis 4: The positive effect of the use of compensation consultants on the peer group characteristics is stronger when the consultant is providing “other services” to the company as opposed to when the consultant is providing only remuneration services.
Academics have recently started to investigate the relation between the size of the consultancy agency and its effect on the chief executives’ level of compensation.
Tong and Cen (2011) argue that firms employing bigger consultancy agencies receive lower equity payments and as a result lower total compensation levels compared to that of companies employing small consultants. Consequently the authors argue that big consultancy agencies tend to design more optimal remuneration schemes to reduce excessive pay than smaller consultancy agencies, but thus far the academics have not managed to provide a clear reasoning that would explain this phenomenon.
Recent studies examining the relationship between the use of compensation consultants and executive remuneration have documented that consultancy agencies are increasingly aiming to secure additional businesses with the client firm that are not related to remuneration services (Cadman et al., 2010; Murphy & Sandino, 2010).
This has increasingly become the situation for the smaller consulting firms, as they have fewer clients than the big consulting agencies and as each client contributes proportionately more to their total revenues, it becomes of greater significance to obtain and please each client. The practitioners have identified 5 consulting firms as the U.S. largest compensation consultancies, which together occupy over 60% of the market share of the industry. These are Frederick Cook and Co., Towers Watson, Hewitt Associates, Mercer Human Resource Consultants and Pearl Meyer (Tong &
Cen, 2011). The other 40% of the market is covered by the smaller consultancy firms, who are perceived differently by the clients, as they generally don’t provide other consulting services. Therefore the smaller companies are under less pressure from the management board to design the optimal contract and have greater incentives to design compensation schemes that result in greater pay without increased company performance. Following Kabir and Minhat (2010) especially the smaller compensation consultants try to secure their business interests with the client firms by advising towards higher pay and thereby seek to survive from increasing competition.
Additionally, the small consulting agencies are more likely to include companies from other industries with greater sales to boost CEO pay. Crystal (1991) argues that if a compensation consultant doesn’t make pay recommendations that would increase chief executives’ pay, they will risk losing the future businesses with the client.
Based on the theoretical argumentation discussed above, the following hypotheses are proposed:
Hypothesis 5a: The positive effect of the use of compensation consultants on target pay level is stronger when the peer group is composed by a small consultancy agency and weaker when the peer group is composed by a large consultancy agency.
Hypothesis 5b: The positive effect of the use of compensation consultants on percentage of other industry peers is stronger when the peer group is composed by a small consultancy agency and weaker when the peer group is composed by a large consultancy agency.
Recently the possible effects of a change of the consultancy firm as well as issues concerning the selection of the number of compensation consultants employed by a company and its effect on CEO pay have come under greater scrutiny by the academic community (e.g. Tong & Cen, 2011; Kabir & Minhat, 2010). Tong and Cen
(2011) find that when a company switches the firm that is providing executive compensation services, the managers will have a significantly greater pay level. More specifically they argue that when a firm switches its compensation consultant from a small to a big consulting agency it will result in CEO receiving lower bonuses, lower equity and an overall lower level of compensation. However when the opposite happens (the firm switches its consultant from a big consultant to a small consultant), its chief executives will receive higher bonuses, higher salaries and an overall higher level of compensation. As the prospect of lucrative business interests incentivizes the compensation consultants to compete intensely with each other, the smaller consultancy agencies seek to expand their market share by creating a reputation of designing remuneration schemes that would result in higher managerial pay levels (Kabir & Minhat, 2010). Furthermore, it could be argued in the increasingly competitive remuneration services market the small consultancy agencies have a greater motivation to include companies from other industries with greater sales, as well as to aim a higher target percentile to inflate CEO remuneration. The theoretical reasoning discussed above lead to the suggestion of the following hypotheses:
Hypothesis 6a: The switch from a big to small compensation consulting company has a positive influence on target pay level, whereas switch from a small to big consulting company has a negative influence on target pay level.
Hypothesis 6b: The switch from a big to small compensation consulting company has a positive influence on percentage of other industry peers, whereas switch from a small to big consulting company has a negative influence on percentage of other industry peers.
It will be further argued that if multiple consultants provide remuneration services to a company, they are assumed to have stronger incentives to advocate higher CEO
compensation than acting alone. In more detail it is being suggested that if different compensation consultants specialize in different aspects of pay, firms have the possibility to benefit from employing multiple consultants to inflate executive pay (Bender, 2008).
Kabir and Minhat (2010) have argued that as the number of pay consultants employed by the company is increasing, so does the CEOs’ equity-based pay. In particular they find that an increase in the number of compensation consultants is associated with an increase in CEO compensation. There is however no corresponding decline in CEO compensation when firms reduce the number of compensation consultants. This empirical evidence lends to support the conjecture that firms employing multiple compensation consultants provide a justification of increased level of executive remuneration. Given the fact that the U.S. consulting industry is highly aggregated with the five major compensation consultancy agencies, the small consulting firms that are working in conjunction with bigger consultancy agencies are suggested to be under greater pressure to justify the level of executive pay. As the employing firm is assessing the consultants’ input based on the advice provided to the company, the smaller consulting firms have the tendency to provide the level of pay that would secure the existing business relations or expand the extent of services provided to the company. As a result it could be argued that if a firm is employing multiple compensation consultants of whom one is a small consultancy agency, there is an increased pressure for consultants to advise firms to set a significantly higher target percentile for determining CEO remuneration. Based on the line of reasoning stated above, the following hypotheses are suggested:
7a: Firms employing multiple consulting agencies target a higher target pay level than companies employing only one compensation consulting agency.
7b. Firms employing simultaneously big and small consulting agencies have a higher target pay level than companies employing either a big or a small compensation consultancy agency.
4. METHOD
4.1 Sample and Procedure
In order to study the compensation benchmarking processes in the executive boards of the S&P 500, a quantitative data collection method has been applied. First, these 500 companies traded on the New York Stock Exchange and the NASDAQ with the largest market capitalization were identified. Thereafter, based on random sampling procedure 101 companies of the S&P 500 were chosen to represent the whole population. The total sample for this research consists of 9090 hand-collected firm-year observations from fiscal years 2007 and 2010.
As the compliance deadline for the peer group disclosure was December 15th, 2006, the research is confined to begin from fiscal year 2007. This study is additionally restricted to fiscal year 2010, as this is the last fiscal year that all the firms in this sample have provided the details about their peer group as well as the fiscal year end reports. The members of the compensation peer group and the use of compensation consultants is stated in the CD&A section of companies’ annual proxy statements, which are typically stated in the firm SEC DEF-14A filings that are accessible through EDGAR. Companies’ annual financial measures are adapted from the Orbis database, Google Finance, Wikinvest and from companies’ annual reports.
4.2 Independent Variables
The CD&A section of the annual proxy statements is used for collecting the information about the CEO’s and other four top executives’ yearly remuneration, as well as for collecting the information about the peer group and the use of compensation consultants. More specifically, the following values will be taken from the annual proxy statements: yearly cash compensation, yearly salary and bonus, total yearly compensation, name of the CEO, whether compensation consultants are used, name of the compensation consultant, who retains the compensation consultant, does the compensation consultant provide other services to the company besides remuneration related services, the fees paid for other (non-remuneration related) services, whether the company has changed the compensation consultant from year 2007 to 2010, number of peer group firms, and the target pay level of the company.
Additionally, information from the proxy statements and annual reports was used to calculate subsequent variables.
Leverage ratio. The leverage ratio of a company is calculated by dividing the
company’s total debt with company’s total reported assets.
Market value of assets. The market value of a company’s assets is determined by
multiplying the share price with the number of shares outstanding and adding the company’s total debt value.
Percentage of other industry peers. The percentage of peer companies that
originate from another industries than the company in question, is determined in the same manner as other scholars have done, it when examining benchmarking procedures (e.g. Faulkender & Yang, 2010; Bizjak et al., 2011). Firstly, based on the four-digit code that is assigned to each industry, the industry code of the company in question will be determined. The industry codes were determined from the S&P
webpage. Thereafter the same procedure was replicated to all of the companies constituting the peer group. Based on these figures the company is regarded to originate from the same industry if the first two digits of the four-digit code match with the company’s first two digits. If the first two digits of the peer company differ from the company’s own industry code, the peer is regarded to be an outside industry peer. Based on this, the number of outside industry peers will be divided with the total number of company’s peers, resulting in the percentage of outside industry peers.
Peer group change. A change in a company’s peer group is regarded to have
taken place, when there has been a change by at least one company in the identified peer group between fiscal years 2007 and 2010.
Small or big consultants. The big 5 compensation consultancy agencies on the
U.S. market are Frederick Cook and Co., Towers Watson, Hewitt Associates, Mercer Human Resource Consultants and Pearl Meyer and therefore these companies will be regarded as big consultancy agencies. If the company uses consultancy agencies that do not belong to the list of the aforementioned five companies, they will be regarded as small consultancy agencies.
Change in the size of the consultancy agency. Based on the size of the
consultancy agency (classified in the previous variable) a change in the size of consultancy agency is considered to have taken place, when in 2010 the size of the consulting agency providing services to the company is different from the fiscal year 2007.
Multiple compensation consultants. A company is regarded to employ more than one compensation consulting agency, if the company reports in their CD&A section of the annual proxy statement the use of more than one consulting agency.
Small and big consulting agencies simultaneously. A company is regarded to use
the services of small and big and small consulting agencies’ simultaneously, if the company reports in their CD&A section of the annual proxy statement that they use more than one consulting company of which at least one is small and the other is big in its size (based on the classification described above).
4.3 Control Variables
In addition to the hand-collected data that is obtained from the CD&A section of the annual proxy statements, the data is supplemented with the following measures:
whether the CEO is the chairman of the board, CEO tenure, firm’s return on assets (ROA), equity, sales, price-earnings ratio (price / yearly dividend). This data is collected for determining the interdependencies that exist between the CEOs’ and other board members’ remuneration and firm’s own financial performance.
4.4 Dependent Variables
As the aim of this paper was to investigate how different benchmarking practices influence CEO’s compensation, I will investigate the effect that the abovementioned independent variables have on CEO total compensation as well as on the board’s total compensation. For obtaining the CEOs’ and top executives’ yearly remuneration I will use the CD&A section of the annual proxy statements.
CEO total compensation. CEO total compensation is calculated as a sum of U.S.
dollar yearly cash compensation, yearly incentives and yearly bonus.
Board total compensation. Board’s total compensation is calculated as a sum of U.S. dollar yearly cash compensation, yearly incentives and yearly bonus of the top five highly paid executive officers of the firm.
5. RESULTS
The correlational relationships between the main study variables were examined by conducting parametric correlations tests. Table 1 and Table 2 present the relationships that exist between executive compensation, firm characteristics and the use of compensation consultants for years 2007 and 2010 respectively. As the results reveal that the correlations with the variable total board compensation are weaker than CEO total compensation for the year 2007 as well as for year 2010, the former is excluded from further statistical analysis.
--- Insert Table 1 and Table 2 about here
---
The first objective of this research is to study the effect of compensation benchmarking on the observed level of CEO pay. In order to study this I begin with a baseline estimation of the level of compensation for CEOs among the S&P 500 companies for which I have gathered peer group and company information. In line with the other academic literature (e.g. Faulkender & Yang, (2010); Core, Holthausen
& Larcker, (1999)) I measure the regression of CEO compensation on firm and CEO characteristics to explain the observed variation in received remuneration using the following formula:
CEO Compensationi,t = a + β1(Salesi,t)+ β2 ROAi,t + β3 P/E ratioi,t+ β4 Leveragei.t, + β5 MVAi,t + β6 (CEO tenure) + β7 Dummy(CEO chair) + εi,t.
To mitigate the skew in the data the CEO total compensation is winsorized at the 2nd and 98th percentile. Table 3 presents the regression between the CEO compensation and firm’s characteristics.
--- Insert Table 3 about here ---
As shown by the 2007 results CEOs have higher pay in larger firms and in 2010 CEOs’ income is higher in larger firms, in firms that are more profitable, and have CEOs serving as the chairman of the board, consistent with results shown previously in the literature (e.g. Faulkender & Yang, 2010).
Table 4 reports the cell means, standard deviations and modes of the study variables for 2007 and 2010 respectively. Furthermore, Table 5 presents the frequencies of the study’s nominal variables for years 2007 and 2010.
--- Insert Table 4 and Table 5 about here
---
In order to formally test Hypothesis 1 the multiple regression analysis was conducted. Hypothesis 1 predicted that MNCs employing compensation consultants have higher remuneration than companies that are not using the services of a compensation consultant. Table 6 presents that for 2007 the hypothesis is rejected (b
= 3831306.50, p = .17) with Model 2 producing insignificant results (∆R² = .02, ∆F = .17, p = n.s.), whereas for 2010 the hypothesis is corroborated (b = 4393061.11, p <
.1), despite the change in Model 2 being insignificant (∆R² = .02, ∆F = .10, p = n.s.).
--- Insert Table 6 about here ---
The hypotheses 2a and 2b were proposing that the proportion of peers originating from other industries and target pay level of the CEO, are higher for firms that
employ compensation consultants. The regression analysis reveals that neither for 2007 (b = 8.48, p = .47) nor for year 2010 (b = 13.79, p = .26) the percentage of outside peers reached statistical significance. This means that Hypothesis 2a is rejected for 2007 as well as for 2010. In 2007 companies employing compensation consultants have a statistically significant higher compensation level (b = 9.94, p = .04), which means the Hypothesis 2b is accepted for 2007. On the contrary for 2010 the compensation level doesn’t reach statistical significance (b = 2.20, p = .79) and therefore Hypothesis 2b is rejected for that year.
--- Insert Table 7 about here ---
The third hypothesis was suggesting that the positive relationship between the use of compensation consultants and peer group characteristics is more pronounced, when the consultancy agency is being hired by the management board, as opposed to being employed by the compensation committee. The results from Table 9 reveal that neither for year 2007 (F = 1.08, p = .37) nor for year 2010 (F = .82, p = .49) the statistical significance was reached and therefore Hypothesis 3 is rejected.
The fourth hypothesis proposed that the positive relationship between the use of compensation consultants and peer group characteristics is more pronounced when the consultant provides “other services” to the company as opposed to providing only remuneration services. Results in Table 9 present that this relationship is statistically insignificant for year 2007 (F =.58, p = .56), as well as for 2010 (F = .01, p = .91), which means that Hypothesis 4 is rejected for both years.
To formally test Hypotheses 5a and 5b that are suggesting that the effect of using compensation consultants on peer group characteristics is moderated by the size of
the consulting agency, I used a two-way analysis of variance (ANOVA). This particular statistical model was used because it explains the observed variance in the peer group characteristics that is divided into different components that are attributable to the variation in consultancy agency’s size and the use of compensation consultants. For Hypothesis 5a, results in Table 8 indicate that in 2007 the statistical relationship is insignificant (F = .00, p = .28; F = 1.71, p = .20). On the contrary the results for 2010 suggest that smaller compensation agencies target a higher compensation level than big consultancy agencies (F = 3.47, p = .07) and therefore Hypothesis 5b is corroborated for the year 2010; for 2007 Hypothesis 5b is rejected (F = 1.70, p = .17).
--- Insert Table 8 about here ---
The sixth hypothesis proposed that a switch from a big to a small consulting agency will lead to an increase in target pay level and in the percentage of other industry peers, whereas the switch from a small to a big consulting agency will have the opposite effect. The hypothesis was tested using a linear regression analysis.
Results presented in Table 9 reveal that the switch in the size of the consulting agency does not predict the percentage of other industry peers (b = -7.26, p = .59), nor does it predict the target pay level (b = -4.62, p = .59). Therefore both Hypotheses 6a and 6b are rejected.
--- Insert Table 9 about here ---
In order to formally test Hypotheses 7a and 7b, I used linear regression analysis.
Hypothesis 7a proposed that firms who employ multiple consulting agencies have a higher target pay level than MNCs employing only one compensation consulting agency. This hypothesis did not reach statistically significant levels (Table 10) as the statistical significance for predicting target pay level was (b = 2.33, p = .54) and (b = .14, p = .97) for the years 2007 and 2010 respectively. Hypothesis 7b proposed that firms who employ simultaneously big and small compensation consulting agencies have a higher target pay level. This relationship too failed to reach statistical significance, both for 2007 (b = 1.95, p = .71) and for 2010 (b = 6.46, p = .19).
Therefore both Hypothesis 7a and 7b were rejected. In order to provide an overview of the study’s results, a summary table is presented subsequently (Table 11).
--- Insert Table 10 and Table 11 about here
---
6. DISCUSSION
6.1 Findings
The aim of this study was to explore how the use of compensation consultants has affected CEO compensation over the time period of 2007-2010 and in particular how the compensation consultants affect the compensation benchmarking through the design of the peer group characteristics. In comparison to scholars that have previously focused solely on the topic of compensation benchmarking (Bizjak et al., 2008; DiPrete et al., 2010), or the use of compensation consultants (e.g Murphy &
Sandino, 2010; Conyon, 2006), I tried to bring clarity into the topic of executive
compensation by examining the joint effect of the two methods for determining CEO pay. More importantly this research is the first study in this field investigating the moderating effect of the compensation consultants on the compensation benchmarking processes over the three years that the increased disclosure requirements have been in place. Additionally, investigating the switch in the size and the number of consultancy agencies affecting the compensation benchmarking practices were incorporated to this study to extend the scale of the moderating effect.
Results of the study indicated that CEOs’ pay is higher in companies that are bigger, more profitable, and have CEOs serving as the chairman of the board. More importantly it was proven that MNCs employing compensation consultants in 2010 have a significantly higher executive pay than companies that don’t use compensation consultants. Compensation consultants do not inflate executive pay by adding other industry peers to the peer group, but generally aim a higher compensation level than companies that don’t employ compensation consultants. On the contrary to expectations when the consultant is hired by the management or provides “other services” to the MNC, executive pay remains unaffected, whereas the size of the consulting agency is a significant factor predicting CEO compensation level. More importantly, it was noted that in conjunction with a threefold increase of employing small consultancy agencies, MNCs using small consultants have a significantly higher pay level than those using the services of big consultancy agencies.
Furthermore, there was no evidence to support the hypothesized relationship between the switch in size of the compensation consulting agency and its effect on percentage of other industry peers as well as on the level of compensation at peer firms. In addition, contrary to expectations multinational corporations that are using the services of multiple consultants and consultants of various size don’t seem to
