Introduction - EUR



Dancing to the client’s tune: are compensation consultants conflicted?Draft Master ThesisJan-Pieter Vos287903Supervisors: Dittmann, I.; Zhang, D.Date: May 6, 2011.List of contents TOC \o "1-3" \h \z \u 1Introduction PAGEREF _Toc292487300 \h 42The complexity of executive pay PAGEREF _Toc292487301 \h 72.1Executive pay PAGEREF _Toc292487302 \h 72.1.1Components of pay PAGEREF _Toc292487303 \h 72.1.2Setting executive compensation PAGEREF _Toc292487304 \h 92.2The use of compensation consultants PAGEREF _Toc292487305 \h 92.2.1Introduction to compensation consultants PAGEREF _Toc292487306 \h 92.2.2The views on executive pay and consultant incentives PAGEREF _Toc292487307 \h 112.2.3Compensation consultant’s incentives and corporate governance PAGEREF _Toc292487308 \h 132.3Regulation on compensation consultants PAGEREF _Toc292487309 \h 152.3.1Applicability of SEC rulings PAGEREF _Toc292487310 \h 152.3.2Regulation regarding compensation consultants PAGEREF _Toc292487311 \h 162.4Conclusion PAGEREF _Toc292487312 \h 183Literature overview and research questions PAGEREF _Toc292487313 \h 203.1Quantitative research on compensation consultants PAGEREF _Toc292487314 \h 203.1.1Evidence on the effect of compensation consultants in executive compensation PAGEREF _Toc292487315 \h 203.1.2The effect of ‘conflicted’ consultants PAGEREF _Toc292487316 \h 213.2Research questions and contributions to literature PAGEREF _Toc292487317 \h 223.2.1Do consultants want to show their presence? PAGEREF _Toc292487318 \h 233.2.2Does a consultant follow the incentives of its client? PAGEREF _Toc292487319 \h 243.2.3The choice for a particular compensation consultant PAGEREF _Toc292487320 \h 263.2.4Other contributions PAGEREF _Toc292487321 \h 263.3Conclusion PAGEREF _Toc292487322 \h 274Data description and methodology PAGEREF _Toc292487323 \h 284.1Introduction PAGEREF _Toc292487324 \h 284.2Data PAGEREF _Toc292487325 \h 284.2.1Variables and sources PAGEREF _Toc292487326 \h 284.2.2Sample selection PAGEREF _Toc292487327 \h 344.2.3Descriptive statistics PAGEREF _Toc292487328 \h 344.3Methodology PAGEREF _Toc292487329 \h 374.3.1The models PAGEREF _Toc292487330 \h 374.3.2Robust covariance estimation PAGEREF _Toc292487331 \h 384.3.3Outlier influence PAGEREF _Toc292487332 \h 394.3.4Endogeneity PAGEREF _Toc292487333 \h 394.4Multicollinearity PAGEREF _Toc292487334 \h 394.5Conclusion PAGEREF _Toc292487335 \h 405Results PAGEREF _Toc292487336 \h 415.1Introduction PAGEREF _Toc292487337 \h 415.2Primary results PAGEREF _Toc292487338 \h 415.2.1Assisting versus non-assisting consultant hypothesis PAGEREF _Toc292487339 \h 415.2.2Busy board members with the same compensation consultant on multiple boards PAGEREF _Toc292487340 \h 415.2.3Busy compensation committee members with the same compensation consultant on multiple boards PAGEREF _Toc292487341 \h 425.2.4Newly assisting consultant PAGEREF _Toc292487342 \h 425.3Robustness tests PAGEREF _Toc292487343 \h 435.4Independent compensation committees PAGEREF _Toc292487344 \h 435.5Probit analysis PAGEREF _Toc292487345 \h 455.6Conclusion PAGEREF _Toc292487346 \h 466Conclusion PAGEREF _Toc292487347 \h 487Literature PAGEREF _Toc292487348 \h 508Appendices PAGEREF _Toc292487349 \h 538.1Appendix A: Legislation PAGEREF _Toc292487350 \h 538.2Appendix B: Sample selection PAGEREF _Toc292487351 \h 538.3Appendix C: List of variables PAGEREF _Toc292487352 \h 53IntroductionThis thesis discusses the role of compensation consultants. Compensation consultants are outside advisors to a firm’s Board of Directors with regard to executive compensation. Setting executive compensation is a task for the Board of Directors or its Compensation Committee. The task of setting executive compensation often is difficult and complex, because agency theory suggests that executive offers are not a priori expected to act in line with the (best) interests of a company’s shareholders, but instead engage in extracting personal benefits to the detriment of the company ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003}}{{2 Bebchuk,Lucian Arye 2003}}. A compensation package should thus align incentives of executive officers with those of the shareholders. However, research has not yet determined how the optimal compensation package should look like. In fact, contradicting evidence exists. Assigned with such a difficult task, a Board of Directors or a Compensation Committee often engages one or more so-called compensation consultants. Due to their expertise, they are able to help to solve the company’s executive compensation puzzle, for example by proposing a design for a compensation package, or by providing the Board of Directors with data on similar companies. In principle, as is suggested by the optimal contracting approach to executive compensation, the assistance of compensation consultants should lead to optimal compensation packages, which perfectly align the executives incentives with those of the shareholders.However, in recent years, compensation consultants have been under increasing scrutiny. The value of executive compensation packages have been widely criticized in the press. This criticism has extended to compensation consultants. Rather than setting executive compensation in a modest way, they have been accused of helping the executive officers to extract rents from the company by proposing excessive executive compensation ADDIN RW.CITE{{82 Icahn, C.; 2 Bebchuk,Lucian Arye 2003}}{{82 Icahn, C.; 2 Bebchuk,Lucian Arye 2003}}. The conjecture of this so-called managerial power theory is that compensation consultants have incentives to please executive officers because they are responsible for future assignments. Two incentives can be distinguished: first, the rehiring incentive, which is the incentive for a compensation consultant to act such, that he will be rehired in the next year, whereas the cross-selling incentive is the incentive to be assigned other (often more lucrative) services with the company. Both these incentives are expected to lead to consultants proposing higher executive compensation. Until now, research has established that executive compensation is higher in firms that use compensation consultants compared to firms that do not engage them{{47 Higgins, Alexandra 2007; 48 Waxman, Henry A. 2007}}. However, this is not the end of the story because these differences often can be explained by differences in economic and corporate governance characteristics of firms. Furthermore, it is not entirely clear if compensation consultants are drivers of executive pay, or that this is caused by something else (for example, factors driving the decision to use compensation consultants or not). Therefore, research has shifted to the behavior of so-called ‘conflicted’ compensation consultants. On one hand, research has focused on consultants that provide other services than executive compensation as well, but evidence has been mixed ADDIN RW.CITE{{40 Cadman,Brian 2010; 53 Murphy,Kevin J. 2010; 79 Tong,Naqiong 2011}}{{40 Cadman,Brian 2010; 53 Murphy,Kevin J. 2010; 79 Tong,Naqiong 2011}}. Research has also focused on actions in order to be rehired ADDIN RW.CITE{{53 Murphy,Kevin J. 2010; 46 Kabir,Rezaul 2010}}{{53 Murphy,Kevin J. 2010; 46 Kabir,Rezaul 2010}}; evidence also has been mixed there as well.Using a larger dataset and some new proxies to measure the effect of rehiring incentives, I find some evidence that compensation consultants act in order to please their clients. Compensation consultants seem to be associated with higher compensation if they a) have an assisting consulting role rather than a more modest role, such as providing data; and b) are the compensation consultant on another board as well, where at least one of the members of the Board of Directors or Compensation Committee serves as well (i.e. busy board members or busy compensation committee members). However, while it is expected that a newly assigned compensation consultant would like to show his presence, no difference seems to exist compared to companies which already employed a compensation consultant. These results are robust to alternative performance metrics. Being the first to perform fixed-effect analysis on the effect of compensation consultants, results suggest that the difference in executive compensation is based more on differences between firms rather than the presence of a compensation consultant.All this research has been under the assumption that a compensation consultant acts to please a firm’s executive officers. However, it is also well possible that a compensation consultant acts to please the Compensation Committee that is hiring him. The possibility of a high salary is then also conditional to the Compensation Committee’s independence from the executive officers. If a Compensation Committee is independent, a compensation consultant is expected to propose a reasonable pay rather than a high pay. To my knowledge, this assumption has not been tested extensively. Using a number of proxies which could indicate Compensation Committee independence, I find no significant (albeit economically meaningful) evidence that compensation consultants act in line with their principals’ interests.Last, I investigate possible determinants of the choice for a particular compensation consultant. I find that prior experience with the compensation consultant on another firm is a driver of the choice of a particular compensation consultant. Furthermore, as expected, I find that large firms are attracted by large compensation consultants. I don’t find evidence firms with independent compensation consultants have a higher probability to engage boutique compensation consultants (firms which do only provide executive compensation consulting services); in fact this coefficient has a negative sign.In my opinion, future research should use the ‘hard’ facts that are accessible due to improved disclosure as a result from new legislation. Furthermore, the interaction between compensation consultants and compensation committees are a field for investigation, since the clue for definitive evidence might be in there. Furthermore, since evidence suggests that prior experience with compensation consultants improves the chance to be hired and also leads to higher executive compensation, investors might be interested in that.The remainder of this thesis is as follows. In chapter REF _Ref292481263 \r \h 2, I discuss the background of this thesis. Subjects are the executive compensation problem, the role of compensation consultants, and the independence problem of them. Furthermore, I provide an overview of legislation. In chapter 3, I will then give an overview of existing quantitative research on executive compensation, after I propose my research questions. Chapter REF _Ref292485786 \r \h 4 contains a description of my dataset, some descriptive statistics, as well as my research methods. I treat the results in chapter REF _Ref292485820 \r \h 5, after which I conclude in the last chapter.The complexity of executive payExecutive pay is a subject that has been of public interest for a long time ADDIN RW.CITE{{105 Murphy,Kevin J. 1999}}{{105 Murphy,Kevin J. 1999}}. Considering the fact that CEOs wages are many times more than the average worker earns ADDIN RW.CITE{{112 Mishel, Lawrence}}{{112 Mishel, Lawrence}}, this is not strange. However, determining CEO compensation is a very demanding task, given the fact that its composition is a complex matter. Academic literature on this topic has always been present ADDIN RW.CITE{{105 Murphy,Kevin J. 1999}}{{105 Murphy,Kevin J. 1999}} and very likely will be in the future. With this chapter, I hope to shed some light on the problem of executive compensation, the process of compensation-setting and the need of compensation consultants in that process. Any topic discussed here could be subject of a whole thesis itself. Therefore the coverage will be just enough to guide the reader towards the subject of this thesis and certainly will not be in-depth. Executive payComponents of payExecutive compensation packages are typically more complex than the common-worker compensation package. Indeed, ADDIN RW.CITE{{105 Murphy,Kevin J. 1999}}{{105 Murphy,Kevin J. 1999}} pp. 2490-2493 show that executive compensation consists of multiple components, ‘normal’ salary usually not being the major share of the compensation package. More recently ADDIN RW.CITE{{113 Geiler, Philipp 2010}}{{113 Geiler, Philipp 2010}} confirms this view. Executive compensation usually consists of both short-term and long-term components of pay. Short-term components usually are a base salary, an annual bonus based on certain performance measures (e.g. revenue, or profit per share). Long-term components are stock options (which may vest at a later time), restricted stock and long-term incentive plans (i.e. plans that only pay out if a long-term goal is achieved). Furthermore, other components like retirement plans and severance payments can be part of the contract between a company and its executives ADDIN RW.CITE{{113 Geiler, Philipp 2010}}{{113 Geiler, Philipp 2010}}. Agency theory and views on executive compensationThe main challenge in executive compensation is to make sure that the CEO (of course, the same appliesADDIN RW.CITE{{114 Jensen,Michael C. 1976}} to other executives) acts in the best interest of the company’s shareholders and that compensation is set such, that the CEO’s efforts provide investors with a fair return on their investment ADDIN RW.CITE{{113 Geiler, Philipp 2010}}{{113 Geiler, Philipp 2010}}. One cannot a priori expect this, because a CEO has incentives (like personal wealth, or enjoying ‘utility’ to the detriment of shareholder’s value) not to act in the utmost interest of the shareholders. This problem can be traced back to the principal-agent relationship (in which the shareholders are the principals and the CEO is the agent) as in agency theory ADDIN RW.CITE{{114 Jensen,Michael C. 1976}}{{114 Jensen,Michael C. 1976}}. In short, there are two ways of dealing with the agency problem. First, it is possible to provide the executive with the incentives necessary to ensure that the executive acts in line with the shareholder’s interest. Secondly, good corporate governance can enhance firm performance. Later in this chapter, I briefly discuss both these ways.Setting up the optimal structure of executive pay enacts CEOs to act in line with shareholders’ interests. Setting executive compensation is then regarded as remedy to the agency problem and it is assumed that market forces shape the eventual executive compensation height and structure, which is considered to be the efficient contracting view to executive compensation ADDIN RW.CITE{{105 Murphy,Kevin J. 1999; 2 Bebchuk,Lucian Arye 2003; 113 Geiler, Philipp 2010}}{{105 Murphy,Kevin J. 1999; 2 Bebchuk,Lucian Arye 2003; 113 Geiler, Philipp 2010}}. A lot of literature has focused on the ideal structure of executive pay. In the research, it became apparent that linking executive pay to firm performance was the key to optimizing shareholder’s value ADDIN RW.CITE{{115 Bruce,Alistair 2005}}{{115 Bruce,Alistair 2005}}. Consequentially, providing executives with stocks and options was a way to let them act in the shareholder’s interest. Indeed, compensation has inter alia been traditionally linked to firm performance ADDIN RW.CITE{{106 Abowd,John M. 1999}}{{106 Abowd,John M. 1999}}. However, this also has its problems, since executives cannot control all factors affecting firm performance (e.g. an economic crisis). Also, since executives are as risk-averse as other humans, it is not strange that executives are reluctant to have compensation based on stock and options only ADDIN RW.CITE{{106 Abowd,John M. 1999}}{{106 Abowd,John M. 1999}}. In fact, using stock-based compensation also has its downsides. For example, they might provide undesirable risk-taking incentives ADDIN RW.CITE{{78 Core,John E. 2003}}{{78 Core,John E. 2003}}. {{109 DITTMANN,INGOLF 2007}} states that the optimal compensation structure (almost) has no stock incentives, but observes that this stands straight against common executive compensation practice. Indeed, ADDIN RW.CITE{{120 Edmans,Alex 2009}}{{120 Edmans,Alex 2009}} propose that that traditional efficient contracting models cannot fully explain executive compensation and that therefore additional research into new dimensions is called for. Concluding, it is no surprise that research has not yet determined how the optimal compensation structure for a specific firm given the constraints (e.g. taking into account the CEO’s risk-aversion) should look like. If anything about executive compensation is clear, it is that designing the optimal compensation package requires great expertise. Compensation packages are complex and it is therefore no surprise that they are hard to understand by the public. The ‘pessimistic’ view on executive compensation regards that executives use their power to extract rents of the company. In doing so, they benefit themselves personally to the detriment of the shareholders ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003}}{{2 Bebchuk,Lucian Arye 2003}}. This managerial power, rent extraction or skimming perspective consists of two building blocks: outrage costs and camouflage ADDIN RW.CITE{{113 Geiler, Philipp 2010}}{{113 Geiler, Philipp 2010}}. The more outrage executive compensation stirs up in the public (i.e. the press, the shareholders), the more reputational harm is done to the executives and the other directors. Furthermore, excessive compensation might lead to a loss of shareholder support. To limit outrage, managers might try to camouflage (i.e. to hide or legitimize) their extraction of rents, which could hurt optimal executive incentives and firm performance. Therefore, ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003}}{{2 Bebchuk,Lucian Arye 2003}} argue that transparency in disclosure could have a significant effect on CEO compensation. The skimming view plays an important role in companies with weak corporate governance. For example, in a variety of situations, ADDIN RW.CITE{{118 Bertrand,M. 2000}}{{118 Bertrand,M. 2000}} show that CEOs of firms with weak corporate governance are able to extract (more) rents than CEOs of firms that are well governed.Setting executive compensationTypically, executive compensation is set by a company’s Compensation Committee, consisting of independent board members. The Securities Exchange Act of 1934 delegates the definition of independence to the respective securities exchange (for example, in Section 10a-m3 for Audit Committees and in Section 10c-1 for Compensation Committees). The NYSE and NASDAC have developed formal criteria which particularly focus on existing material ties between board members, their direct family members and the company. That said, Delaware case law actually has pointed to take into account all circumstances and not only the economic ties ADDIN RW.CITE{{113 Geiler, Philipp 2010}}{{113 Geiler, Philipp 2010}}.Pursuant to Section 10c of the Securities Exchange Act of 1934, for almost any issuer of securities (i.e. firm), it is obliged to have a Compensation Committee. Its responsibility is “evaluating executive and director performance and establishing top-management compensation and benefit programs” ADDIN RW.CITE{{113 Geiler, Philipp 2010}}{{113 Geiler, Philipp 2010}} and, again pursuant to Section 10c of the Securities Exchange Act of 1934, it has the right to hire outside advisors. Please refer to paragraph 2.3.2.1 for a more comprehensive approach of legal issues. The Compensation Committee makes a proposal of the eventual compensation package, which the shareholders vote for (pursuant to Section 14a of the Securities Exchange Act of 1934). However, this concerns a non-binding vote (this is the so-called “say on pay” legislation), implying that the Board of Directors still makes the eventual decision.The use of compensation consultantsIn this chapter I elaborate on compensation consultants. First, I explain about their history, their duties and the rise of the use of compensation consultants. Then I investigate on the compensation consultant-related legislation. As the data in this thesis is from U.S. companies, I only consider U.S. legislation on the topic. I do not intend to treat the subject comprehensively. The primary objective of this chapter is to provide clarity to the reader who is not familiar with compensation consultants and to explain how compensation consultant-related legislation has developed to date. Evidence in this chapter will be primarily anecdotic, while I treat quantitative research in the next chapter. Introduction to compensation consultantsCompensation consultants have been known since the 1950s {{97 YABLON,CM 1992}}. They help a company or its Board of Directors in determining executive compensation packages ADDIN RW.CITE{{5 Armstrong,Chris S. 2008}}{{5 Armstrong,Chris S. 2008}}. Whilst setting executive pay is primarily a task for the Board of Directors of a firm or its Compensation Committee, several reasons to seek advice from compensation consultants exist. Compensation consultants typically have access to a large pool of compensation data for comparable firms ADDIN RW.CITE{{84 Bizjak,John M. 2008}}{{84 Bizjak,John M. 2008}}. Furthermore, in providing advice to many companies, compensation consultants have built up an enormous expertise in this field. They can share this expertise in various ways, for example by providing insight and advice about trends, performing analysis of managerial labor markets. Furthermore, consultants can help with various contract clauses and they can provide data ADDIN RW.CITE{{35 Conyon,Martin J. 2009}}{{35 Conyon,Martin J. 2009}}. Due to their expertise, they can provide firm-specific advice. Given that setting executive pay is a complex process, it may be efficient to engage compensation consultants rather than to design compensation packages without any outside help ADDIN RW.CITE{{35 Conyon,Martin J. 2009}}{{35 Conyon,Martin J. 2009}}. It is therefore not surprising that compensation consultants potentially have a large influence on executive compensation design.Some compensation consultants are part of large human resource consulting service companies, executive compensation consulting services being only one of the services provided by them. Consultants such as Towers Watson, Hay Group, Mercer and AON also provide compensation services for ‘lower’ personnel, actuarial services, human resource strategy consulting and strategic performance management. The services are not even necessarily limited to human resource consulting. Aon for example, is a well-known provider of various risk-assessing and risk-mitigating services. That being said, there are also boutique compensation consultants, which only provide executive compensation consulting services, such as Frederic W. Cook & Co. and Pearl Meyer & Partners.Typically, scientific literature has divided the duties of compensation consultants in two ‘roles’. The first one is to be the provider of survey data. Firms can buy surveys of executive compensation, which contain pay practices at companies comparable in terms of size and industry. ADDIN RW.CITE{{84 Bizjak,John M. 2008}}{{84 Bizjak,John M. 2008}} In the passing of time, compensation consultants have built enormous databases containing this type of information. The other role of a compensation consultant is to assist in designing executive compensation packages. A consultant can assist a client in various ways. For example, one of the largest consultants, Towers Watson, mentions the following services in its brochure: pay philosophy, incentive plan design, competitive data and market analysis, global pay leveling and design, corporate governance reviews, technical reviews, outside director compensation, compensation risk assessments, pay-for-performance analyses and executive benefits {{119 Towers Watson 2010}}. As can be seen in this list, and in the presented business cases as well, the assisting role of a consultant thus can include a wide variety of services.Strictly speaking, providing data is a compensation consultant service as well, but as this one can be separated more easily than the other services and the provision of data merely implies some sort of ‘passive’ advice to the company, I treat them as two different roles of a compensation consultant in the rest of this thesis.The views on executive pay and consultant incentivesAs explained in paragraph 1.1.1, basically two views on executive compensation exist. One can explain the use of compensation consultants in both approaches. Within the optimal contracting approach, compensation consultants are using their expertise in order to design efficient contracts which align the CEO’s interests with those of the shareholders. As became clear in paragraph 1.1.1, designing an optimal compensation package is a difficult matter that require a high degree of specialization. It therefore is very well reasonable that the Board of Directors or the Compensation Committee seeks help and advice from specialized advisors outside the firms. Seeking help may reduce the chance of setting a compensation package that does not enable the CEO to act in shareholders’ best interest and increase the chance of setting optimal compensation. In this view, compensation consultants are part of the solution to the agency problem, as they are helping a company’s Compensation Committee to create executive incentives that are perfectly in line with shareholder’s interests.In the managerial power view however, compensation consultants are part of the problem. Compensation consultants are merely a ‘camouflage technique’ used by executives in order to extract rents from the company ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003}}{{2 Bebchuk,Lucian Arye 2003}}. The compensation consultant has the reputation as an expert in the field, and he’s trusted because of his expertise. Using compensation consultants therefore suggests that the proposed executive compensation package enhances the executive to act in the shareholder’s interest, but might in fact just extract excess compensation.Why would a compensation consultant risk his reputation as an expert and be used for such purposes? The problem is that consultants have incentives to please the CEO and therefore have a conflict of interest. In essence, two kinds of incentives exist, namely the rehiring incentive and the cross-selling incentive. A consultant’s rehiring incentive comprises the incentive to act such, that he maximizes the probability of being hired by the same and other companies in the future ADDIN RW.CITE{{53 Murphy,Kevin J. 2010}}{{53 Murphy,Kevin J. 2010}}. In this light, client satisfaction is important in order to be rehired in the next year. Furthermore, if ‘busy board members’ (board members who are on the board of more than one company) are positive about the consultant’s work as well, he increases the chance to get an assignment for other firms as well. Summarizing, the rehiring incentive is to ensure future business of the executive consulting services.A compensation consultant may also have a cross-selling incentive. As I explained in paragraph 2.1, many compensation consultants are in fact consultants providing a wide variety of services in the field human resource management. These services might well generate more revenue than the executive compensation services and it is no surprise that a compensation consultant is keen to sell these services as well ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003}}{{2 Bebchuk,Lucian Arye 2003}}, hence the term cross-selling incentive. . If the person who decides on the hiring for those non-compensation services also is involved in the (hiring of) executive compensation services, satisfaction in executive compensation services might increase the chance to get assigned those other services. In fact, {{48 Waxman, Henry A. 2007}} provides examples of job advertisements which identify cross-selling as a compensation consultant’s competence.Consistent with the managerial power view, ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003}}{{2 Bebchuk,Lucian Arye 2003}} describe that consultants may (wish to) have other, more lucrative, assignments with the very same company and in order to protect these assignments, they have incentives to please the CEO by providing him with higher compensation. In this case, the CEO uses compensation consultants in order to “camouflage” his rent extraction from the firm. In this view, the consultant wants to please the executives in order to get further assignments in the future. A CEO abuses these incentives in order to receive excess compensation.Historically, there has been criticism on the role of compensation consultants. Reviewing the work of Crystal (1991), ADDIN RW.CITE{{97 YABLON,CM 1992}}{{97 YABLON,CM 1992}} describes how the CEO, assisted by his compensation consultant, can influence the board of directors in order to obtain the desired compensation package. The influence of the compensation consultants is so strong, because it is so difficult to set executive pay in a way that it is optimal from a shareholder’s point of view. In the beginning, the main critique was that compensation consultants were hired by the CEO rather than the compensation committee and that there role focused on justifying executive pay rather than designing optimal pay packages {{97 YABLON,CM 1992}}. It therefore made sense for the consultant to please the CEO, as he was the one to decide on which consultant to hire next year. The compensation consultant thus had a rehiring incentive, but of course was not insensitive to cross-selling incentives either.However, Crystal’s book appeared in 1991 and much has changed since. While executive pay is still set by the Board of Directors (or their Compensation Committee), independency of those salary-setting bodies arguably has increased ADDIN RW.CITE{{113 Geiler, Philipp 2010}}{{113 Geiler, Philipp 2010}}, p. 215. As Compensation Committees are allowed to hire their own advisors, compensation consultant’s incentives may have been diminished over the years. For example, as will be explained below, compensation committees nowadays should review consulting services in order to determine whether or not a compensation consultant can act objective or not. It is therefore not strange that not all research shares the criticism on compensation consultants. For example, consistent with the optimal contracting approach, ADDIN RW.CITE{{17 Conyon, M.J. 2006}}{{17 Conyon, M.J. 2006}} argue that the reputation held by the consultant could be of such importance that consultants will provide objective and neutral advice, although these reputation concerns may have less weight than potential benefits which they could obtain by providing CEOs with higher compensation packages. Furthermore, if a compensation committee is impressed with the consultant’s neutrality (or: independence), the consultant might be rehired next year. Rehiring incentives therefore can also play a role in this view. Contemporaneous work ADDIN RW.CITE{{79 Tong,Naqiong 2011}}{{79 Tong,Naqiong 2011}} consider a consultant’s current rehiring incentive to be such, that the consultant acts in order to be rehired by the Compensation Committee. Given its independence, it expects consultants to provide ‘proper’ compensation advices in order to maximize the chance to be rehired next year. It is well possible that the truth lies somewhere in between. Based on a number of interviews, {{50 Bender, Ruth 2008}} distinguishes between three roles of compensation consultants, i.e. providing firms with survey data, providing expert advice and providing legitimacy to the eventual compensation decision. While it may be conscious or not, compensation consultants tend to be aware of the fact that they are a important party in the decision-making process of a highly sensitive object, the CEO’s compensation. In interviews performed by {{50 Bender, Ruth 2008}}, several compensation consultants mention the need to have business, but also the need to maintain their reputation, e.g. by knowing their role, knowing who they are working for, to communicate pay decisions to major shareholders with credibility. The role of compensation consultants is therefore not necessary to the shareholders’ detriment. These mixed opinions on compensation consultants (and the mixed empirical evidence, as I will discuss in paragraph 3.1 have therefore not yet led to clarity of the influence of compensation consultants on executive pay. However, all parties seem to agree that compensation consultant independence is a prerequisite for a compensation consultant to provide objective and impartial advice. Whatever influence compensation consultants may have, at least they have to fight againstADDIN RW.CITE{{82 Icahn, C.}} the public perception that they are conflicted ADDIN RW.CITE{{74 Campos, Roel C. 2007; 82 Icahn, C.}}{{74 Campos, Roel C. 2007; 82 Icahn, C.}}Compensation consultant’s incentives and corporate governanceIn the last paragraph, it became apparent that the two streams of contracting views predict opposing effects. However, the efficient contracting and the managerial power view are not mutually exclusive, but can be used (together) to explain departures from ‘optimal’ compensation ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003}}{{2 Bebchuk,Lucian Arye 2003}}. Firm-specific circumstances may influence whether or not a company is more sensitive to managerial powers.These firm-specific circumstances mostly are of a corporate governance nature. Firms with weak corporate governance tend to be more captive for managerial power influences, whereas firms with strong corporate governance will tend more towards the optimal contracting view. It is therefore expected that executives in firms with weak corporate governance have higher executive payment than firms with strong(er) corporate governance ADDIN RW.CITE{{61 Core,John E. 1999}}{{61 Core,John E. 1999}}.However, another aspect that is important to notice, is that differences in corporate governance might affect compensation consultants’ actions. Assuming that a consultant is sensitive both the rehiring and cross-selling incentives, his actions will possibly depend on the corporate governance of the firm, most likely the strength of a compensation committee.With respect to the rehiring incentive, a compensation consultant will want to please the person responsible for his hiring. If this is the CEO, it is expected that the compensation consultant will propose a higher pay. However, if the compensation consultant is engaged by a firm’s compensation committee, he will most likely please the compensation committee. It is well possible that the actions that will please the compensation committee, are dependent of a compensation committee’s independence. If the committee is independent, a compensation consultant will most likely propose a reasonable rather than a high pay, because this is something that the compensation committee will like. However, if a firm’s compensation committee is not (entirely) independent, the committee possibly has incentives to please the CEO as well and a consultant will possibly act as if the CEO hired him personally. A further complication is that the observable criterions of compensation committee independence not necessarily imply compensation committee independence. Formally independent compensation committees may have (non-observable) incentives to please CEOs as well. {{2 Bebchuk,Lucian Arye 2003}} take the view that the Board of Directors not necessarily has a priori incentives to act in the utmost interest of the shareholders. According to them, director compensation and the social status of board seats may provide an incentive for a director to please the CEO. Moreover, getting a bad reputation likely harms the chance to be assigned as a board member at another company. Reasoning along these lines, a compensation consultant could still effectuate his incentives into high executive pay, as the directors won’t be critical as well. In my opinion, board members not necessarily will have these incentives – being ‘objective’ might pay off nowadays – but it is important to be aware of these non-observable incentives. Compensation committee independence (based on formal criteria) therefore is a proxy for good corporate governance rather than ‘real’ independence. However, formally independent compensation committees are also expected to be captive for CEO-pleasing incentives in some way and, at least for this group, some effect could be determined. Indeed, research has confirmed that differences between formal and material independence exist. ADDIN RW.CITE{{123 Hwang,Byoung-Hyoun 2009}}{{123 Hwang,Byoung-Hyoun 2009}} distinguish between ‘conventional’ (i.e. formal) independence and ‘social’ independence. They find that boards which are formally independent but socially dependent, are associated with higher pay and weaker pay-performance sensitivity than boards which are socially independent as well.More complexities exist in determining the cross-selling effects. As I will discuss in the next paragraph, a compensation committee shall determine whether or not a compensation consultant can act independently and shall, for that decision, take into account the other fees incurred. Also, some firms have the policy that a compensation committee has to approve (the fees of) all non-executive compensation services of the compensation consultant. In these cases, in order to be eligible for these more profitable services, the consultant will act in accordance with the committee’s incentives. However, if the CEO is responsible for engaging such services, a compensation consultant may calculate and come to the conclusion that it is profitable to propose a high executive salary, thereby sacrificing his chances for compensation consulting services in the future, but increasing his chances to be hired for other services.It is hard to observe how independent a compensation committee really is (although it formally might be). Consequentially, it is hard to determine which compensation contracting view will play the most important role and therefore it is hard to predict compensation consultant incentives. Rather than judging about those incentives, it makes sense to link compensation consultants to good and bad corporate governance.Regulation on compensation consultantsApplicability of SEC rulingsThe Securities Exchange Act of 1934 is the fundamental law for publicly held companies, which securities are traded in the U.S. It aims at different participants in the securities trading process {{104 Hazen, Thomas Lee 2009}}. Moreover, the Securities Exchange Act of 1934 establishes the U.S. Securities and Exchange Commission (hereinafter: the “SEC”). For example, it contains rules about trading and against manipulation of security prices. The Securities Exchange Act of 1934 imposes a number of disclosure obligations on publicly-traded companies in various places, the (in light of this thesis) most interesting being section 14, requiring a proxy statement.This paragraph mainly considers regulation regarding disclosure of interaction with compensation consultants. However, to assess the trend which these disclosure requirements fit in, I show the historical trend in Table I, for the better part based on ADDIN RW.CITE{{121 Dew-Becker,Ian 2009}}{{121 Dew-Becker,Ian 2009}}.<INSERT TABLE I>Table I suggests that the general trend has been towards more and qualitatively better disclosure on executive compensation, except for the deregulation trend in 1982. In recent years, the SEC has required firms to disclose more information about the process of setting executive compensation. The interactions with compensation consultants are part of that. Such rules were first introduced in 2006 and I will explain them in the next paragraph.Regulation regarding compensation consultantsThe right to hire a compensation consultantThe Securities Exchange Act of 1934 contains some regulation about the right to hire compensation consultants. Section 10(c) discusses a company’s Compensation Committee, which can be summarized as follows (the text being amended most recently on July 21, 2010). In short it requires inter alia U.S. publicly-traded firms to establish a Compensation Committee consisting of independent board members. Furthermore, Section 10(c) contains rules on compensation consultants and other Compensation Cpommittee advisors. Pursuant to paragraph C-1 of Section 10(c), “the Compensation Committee (…), in its capacity as a committee of the board of directors, may, in its sole discretion, retain or obtain the advice of a compensation consultant.” Further rules, which can be found in Appendix, state that the responsibility for this consultant is with the Compensation Committee and that this does not affect the duties and obligations of the Compensation Committee. The Compensation Committee should identify factors that may affect the independence of a compensation consultant and should inter alia consider (paragraph C-2) the provision of other services by the consultant, the amount of fees received as a percentage to the total revenue of the adviser, the policies and procedures of the company to prevent a conflict of interest, any business or personal relationship between the consultant and any of the Compensation Committee members, and any stock owned by the consultant. The consultant shall be paid by the company (paragraph E-1).Compensation consultant disclosureWhile regulations regarding executive compensation have existed for a very long time ADDIN RW.CITE{{121 Dew-Becker,Ian 2009}}{{121 Dew-Becker,Ian 2009}}, the compensation consultant part has only been developed recently. Since 2006, the SEC has included several rules regarding disclosure of the assistance of compensation consultants, based on the Securities Exchange Act of 1934. The SEC proposed the first rules, which compensation consultant-related rules were a part of, in the beginning of 2006. The SEC “intended to provide investors with a clearer and more complete picture of compensation to principal executive officers, principal financial officers, the other highest paid executive officers and directors” and therefore, among proposing various elements of compensation to be disclosed, proposed to require “narrative disclosure comprising both a general discussion and analysis of compensation” {{99 U.S. Securities and Exchange Commission 2006}}. Among (many) other rules, it was proposed to provide narrative disclosure “any role of compensation consultants in determining or recommending the amount or form of executive and director compensation, identifying such consultants, stating whether such consultants are engaged directly by the compensation committee (or persons performing the equivalent functions) or any other person, describing the nature and scope of their assignment, the material elements of the instructions or directions given to the consultants with respect to the performance of their duties under the engagement and identifying any executive officer within the company the consultants contacted in carrying out their assignment.” This lead to various comments, for example from compensation consultants, who objected against such compulsory disclosure. ({{100 U.S. Securities and Exchange Commission 2006}} p. 53205. In some of the compensation consultant’s letters, they stress that they merely have an advising rule, and that they are not necessarily the only ones advising. However, in the final ruling, the SEC only dropped the “required disclosure regarding contacts with executive officer” ({{100 U.S. Securities and Exchange Commission 2006}} p. 53205.The need for further disclosure was triggered by a report prepared for the U.S. House of Representatives, which concluded that executive pay was higher in firms that used consultants which also provided other services to the company and that those conflicts of interest were not reported in the company’s proxy statement {{48 Waxman, Henry A. 2007}}. Also, 21 institutional investors signed a petition for further disclosure rules requiring fees for both executive compensation services and other services provided by compensation consultants {{101 U.S. Securities and Exchange Commission: Petitions 2008}}. Potential conflicts of interest might, according to the SEC, affect the objectivity of the compensation consultant. Therefore, the SEC proposed companies to disclose the following, on top of the existing requirement {{102 U.S. Securities and Exchange Commission 2009}}: “the nature and extent of all additional services provided to the company or its affiliates during the last fiscal year by the compensation consultant and any affiliates of the consultant;the aggregate fees paid for all additional services, and the aggregate fees paid for work related to determining or recommending the amount or form of executive and director compensation;whether the decision to engage the compensation consultant or its affiliates for non-executive compensation services was made, recommended, subject to screening or reviewed by management; andwhether the board of directors or the compensation committee has approved all of these services in addition to executive compensation services.”Reactions on this proposal were positive from the investor side, but multi-service compensation consulting firms – i.e. compensation consulting firms that provide other services than executive compensation services alone – opposed the plans. According to them, the amendments were focused too much on multi-service compensation consultants and under-appreciated the role of compensation consultants on every other firm, which could lead to a distorted view about compensation consultants. Furthermore, they were anxious about their competitive position, now that sensitive and confidential information would become public {{103 U.S. Securities and Exchange Commission 2009}}. Also, the compensation consultants stressed the fact that they made recommendations regarding executive pay and did not set them. Based on the comments, the SEC maintained its point of view, but made some adjustments to the original rule. This led to a very complicated rule, which can be found in the Appendix 8.1. Its summary is that a Compensation Committee’s consultant should disclose the fees, if the fees incurred for non-executive compensation work for company management exceeds $120,000.In March 2011, the SEC proposed rules on further disclosure of compensation committee independence and their interactions with compensation consultants. Also, it proposes to remove a few disclosure exemptions, as now it is for example not obliged to disclose about compensation consultants in a non-assisting role. The consultation period ended in April 2011. Further information is not yet available {{138 U.S. Securities and Exchange Commission}}.(Non-)CompliancePursuant to Section 10c, paragraph (f) of the Securities Act of 1934, the issuer (i.e. the company) shall – except for some minor exceptions – comply with the rules and regulations in and by virtue of the Securities Act of 1934. It is hard to conclude if firms are compliant with these regulations. However, in only 20 of 3908 filed proxy statements for fiscal years 2007-2009 in my original dataset (more information follows in paragraph 4.2.2), no Compensation Discussion and Analysis was included (which was compulsory pursuant to {{100 U.S. Securities and Exchange Commission 2006}}). I did not check if these 20 exceptions are allowed or not, but it seems safe to conclude that firms generally are compliant.ConclusionIn this chapter, I introduced the problem with setting executive compensation. Executive compensation setting differs from traditional compensation setting because agency theory suggests that shareholders should set the compensation package such, that the CEO acts in line with their interests and does not extract rents from the company. However, research suggests that finding the optimal compensation package is a complicated puzzle. Therefore, it is not strange that help from outside is needed. Such is provided by compensation consultants. However, two opposing theories on executive compensation can be applied to compensation consultants. Based on the optimal contracting approach, compensation consultants help to design optimal compensation packages. Based on the managerial power or rent extraction approach however, the manager uses his power to let the consultant act in the best interest of the manager. A consultant could comply with this, because he has interests to be rehired in the next year, but also to get assigned more lucrative non-executive compensation consulting businesses. However, a compensation consultant’s incentives possibly vary with the person he is working for. Nevertheless, regulation regarding more disclosure on compensation consultant engagement is increasing.Literature overview and research questionsThe previous paragraphs have primarily provided anecdotic evidence discussing a compensation consultant’s contribution to executive compensation. In this chapter, I discuss quantitative research and my research questions, as well as my contributions to the literature.Quantitative research on compensation consultantsEvidence on the effect of compensation consultants in executive compensationGiven the lack of clarity of the role of compensation consultants and their influence on executive compensation, it is not surprising that researchers have had interest in the subject in recent times. Some of the first exploratory research was undertaken by {{47 Higgins, Alexandra 2007}}. She found that compensation consultants are associated with compensation pay levels higher than the median, while these levels were not associated with higher performance. Furthermore, compensation consultants did not appear to increase the effectiveness of incentive plans. Later that year, the US Congress called for similar research. The report concluded that conflicts of interest are actually very common for compensation consultants. In almost 50% of the cases that were investigated in the report, compensation consultants incurred fees from other services, which were almost eleven times as big as the fees incurred for compensation consulting services. Furthermore, a correlation between the extent of a compensation consultant’s conflict of interest and the level of executive pay was found {{48 Waxman, Henry A. 2007}}. While these reports were valuable as eye-opener to the potential harmful role of compensation consultants in the field of executive compensation, they can hardly be considered as proof for the role of compensation consultants, as these reports do not control for other economic variables ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}} which might influence CEO pay, such as firm size, firm risk and firm performance (as Cadman does in his own research). At the time, such research did exist for firms in the United Kingdom. ADDIN RW.CITE{{17 Conyon, M.J. 2006}}{{17 Conyon, M.J. 2006}} found that compensation consultants typically are associated with higher CEO pay and, more importantly, that this difference was greater than expected if consultants provided other services to the company.Recently, a number of studies have quantitatively examined the role of compensation consultants. ADDIN RW.CITE{{54 Armstrong,Chris S. 2008}}{{54 Armstrong,Chris S. 2008}} investigate the relation between the use of compensation consultants and CEO pay levels. After controlling for economic variables, they find that “CEO pay is generally higher in clients of most consulting firms”. However, after they control for corporate governance effects as well, the effect of compensation consultants disappear. Furthermore, their findings are not robust to propensity score-matching analysis.ADDIN RW.CITE{{35 Conyon,Martin J. 2009}}{{35 Conyon,Martin J. 2009}} also test the influence of compensation consultants in CEO pay. Analyzing data from both United States (hereinafter: “U.S.”) and United Kingdom (“U.K.”), they find that compensation consultants typically are associated with higher CEO pay in the United States but not in the United Kingdom. Furthermore, they find that the relative share of equity-based compensation in an executive’s total compensation package is larger if the firm uses compensation consultants. Since ‘pay at risk’ generally leads to higher pay, according to Conyon et al., this could be a reason for the evidence that compensation consultants typically are associated with higher CEO pay. Without performing any robustness tests they also suggest that there might be an endogeneity problem in the use of compensation consultants, since it is unknown what determines the probability for a firm to hire the consultants (and therefore, a underlying factor rather than compensation consultants could be causing potential differences in pay). In a linear regression, ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}} also find that compensation consultants are associated with higher CEO pay. However, they only do control for economic variables. ADDIN RW.CITE{{56 Voulgaris,Georgios 2010}}{{56 Voulgaris,Georgios 2010}} find weak evidence of a positive significant influence of compensation consultants, also after controlling for corporate governance effects. According to them, this mostly arises from equity-based pay compensation, while consultants have a decreasing effect on the salary part of the compensation package.The effect of ‘conflicted’ consultantsGiven the difficulty to find any tangible compensation consultant effect, it is not surprising that recent research has focused on consultant behavior rather than their presence. More specific, ‘conflicted’ compensation consultants have been under closer scrutiny, as they might have incentives to please the CEO by proposing higher compensation packages. The most important example of conflicts of interest is the cross-selling incentive for consultants providing a wide range of consulting services, most of them more lucrative than compensation consulting. By pleasing the CEO, consultants may increase their overall profitability on the client firm ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003}}{{2 Bebchuk,Lucian Arye 2003}}. Furthermore, proposing compensation packages which are too low in the CEO’s opinion, might decrease the chance to be rehired by that same firm or, indeed, by other firms ADDIN RW.CITE{{35 Conyon,Martin J. 2009}}{{35 Conyon,Martin J. 2009}}.Recent research has measured those conflicts in various ways. ADDIN RW.CITE{{54 Armstrong,Chris S. 2008}}{{54 Armstrong,Chris S. 2008}} focus on the consultants which were marked as ‘conflicted’ in the U.S. House of Representatives report {{48 Waxman, Henry A. 2007}} and does not find any significant effects for these consultants. ADDIN RW.CITE{{35 Conyon,Martin J. 2009}}{{35 Conyon,Martin J. 2009}} use a self-reporting proxy for U.K. consultants (“did the consultant provide non-compensation pay related services”) and a consultant dummy in the U.S. (“was the consultant another one than F.W. Cook or Pearl Meyer & Partners”) and find little evidence that these proxies drive executive pay rises. In ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}, conflicts of interest are the main topic of interest. They use various proxies to define conflicted consultants, namely (i) if the firms disclose that their consultants provide non-executive compensation services; (ii) if the consultant is not F.W. Cook or Pearl Meyer & Partners, which are firms providing only executive compensation services. The use of another consultant might indicate the presence of a conflict of interest; (iii) if the firms “hire their auditor for significant non-audit services, indicating a willingness to allow possible conflicts of interest among their professional service providers”; and (iv) a mixture of these three proxies. ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}} do not find evidence suggesting that these conflicts of interest are a primary driver of excessive CEO pay.ADDIN RW.CITE{{55 Kabir,Rezaul 2010}}{{55 Kabir,Rezaul 2010}} investigate the use of multiple compensation consultants, which might lead to competition (by proposing higher compensation) for those consultants in order to (i) be rehired in the future; and (ii) increase the chance to be hired for lucrative non-consulting services. Using panel data, they find evidence suggesting that equity-based CEO pay is higher when firms employ more than one compensation consultant. They also find that increases in CEO pay occur due to increases in the number of compensation consultants and that increases are positively related with CEO compensation, all “consistent with the conjecture that compensation consultants act to survive competition from other consultants.” Last but not least, they also find that compensation increases with the relative importance of the client to the compensation firm.ADDIN RW.CITE{{53 Murphy,Kevin J. 2010}}{{53 Murphy,Kevin J. 2010}} test two hypotheses, namely (i) the “repeat business” hypothesis, which tests the conflict of increasing the chances to be rehired next year (this is the same as what I mean with the rehiring incentive – JRV); and (ii) the “other services” hypothesis, which tests the conflict of being assigned with more lucrative services (the cross-selling incentive). The repeat business hypothesis is measured as a variable for who the consultant is working (exclusively for board, management or non-exclusive work), whereas the other services hypothesis is defined as several variations of the theme what “other services” were provided by the consultant. Most importantly, the name of the company’s actuary was matched to the compensation consulting firm. ADDIN RW.CITE{{53 Murphy,Kevin J. 2010}}{{53 Murphy,Kevin J. 2010}} find inconsistent evidence for the repeat business effect (i.e. executive pay is higher if the compensation consultant work exclusively for the compensation committee), but find some evidence for the other services hypothesis. Although being robust to several alternative OLS specifications, the results do not hold their significance when they perform a propensity-score matching analysis.Last, contemporaneous research ADDIN RW.CITE{{79 Tong,Naqiong 2011}}{{79 Tong,Naqiong 2011}} suggests that executive compensation is higher when non-executive compensation consulting fees paid to compensation consultants are higher and that providing other services next to executive compensation consulting services is associated with lower pay-performance sensitivity.Research questions and contributions to literatureThe two previous chapters showed the difficulty in setting executive compensation, as well as the difficulties in explaining the role of compensation consultants within this framework. In this chapter, I describe my research questions and compare them to existing literature. Each question is discussed in one paragraph. These paragraphs are structured as follows: first I explain the research question, then I discuss existing research and their problems, my contribution to existing literature. After that, I elaborate on some other problems with existing research. Do consultants want to show their presence?Until now, compensation consultant research has focused on the effects of the presence of compensation consultants ADDIN RW.CITE{{5 Armstrong,Chris S. 2008; 35 Conyon,Martin J. 2009; 56 Voulgaris,Georgios 2010}}{{5 Armstrong,Chris S. 2008; 35 Conyon,Martin J. 2009; 56 Voulgaris,Georgios 2010}} and on effects of cross-selling incentives ADDIN RW.CITE{{40 Cadman,Brian 2010; 53 Murphy,Kevin J. 2010; 79 Tong,Naqiong 2011}}{{40 Cadman,Brian 2010; 53 Murphy,Kevin J. 2010; 79 Tong,Naqiong 2011}}). However, there has not been much research into the rehiring incentives that consultants may have. In fact, only ADDIN RW.CITE{{53 Murphy,Kevin J. 2010}}{{53 Murphy,Kevin J. 2010}} has investigated this specifically. They might occur in the sense that a compensation consultant wishes to show his added value, implying that there might be a significant departure from ‘optimal’ compensation.I propose three research questions, which may serve as indications regarding rehiring incentives, which focus on active consultant behavior.Value for money?It is interesting to investigate if compensation consultants want to show their value for money. Bearing similarities with incentives that a consultant may have in case of a change of consultants, an ‘assisting’ consultant might want to show his presence, in addition merely providing data. Traditional research does not consider ‘data’ to be the same as ‘using consultants’. Data can be obtained from surveys provided by compensation consultants. ADDIN RW.CITE{{84 Bizjak,John M. 2008}}{{84 Bizjak,John M. 2008}} find that benchmarking executive pay has significant influence on executive compensation. However, they find that this influence is merely to determine the reservation wage of a CEO (thus leading to some sort of efficient contracting) rather than to extract rents from the company. Furthermore, since this has to do with a kind of rehiring incentive, it could be argued that the research of ADDIN RW.CITE{{53 Murphy,Kevin J. 2010}}{{53 Murphy,Kevin J. 2010}} also comprises this issue.However, research has not yet investigated the differences between a merely data-providing role of compensation consultants and their assistance. If compensation consultants would not perform more activities than providing data, their added value would be absent. It is therefore interesting to investigate their effects. I expect that firms that engage a compensation consultant in an assisting role are associated with higher compensation than compensation consultants with another role.Change of consultantOne can easily imagine that a newly-hired compensation consultant wishes to impress his new clients. Maybe a consultant wants to show his expertise, or impose a change. It is also possible that he has in mind that – if applicable – the previous consultant apparently has not been engaged again and that the client wishes new perspectives. ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}} find no significant contributions to executive pay if consultants are added or dropped. ADDIN RW.CITE{{46 Kabir,Rezaul 2010}}{{46 Kabir,Rezaul 2010}} find that increasing the number of compensations consultants used is associated with higher executive pay. Dropping compensation consultants yield no similar negative effect. However, there has been no research on the change from 0 to 1 consultants specifically, and also not on within-firm effects. Panel data enables me to analyze that.Do mouth-to-mouth advertising incentives exist?Until now, research has focused on rehiring and cross-selling incentives within one client. However, spin-offs to other firms have not been investigated yet. Positive experiences with compensation consultants on one firm could lead to consulting business opportunities at firms that are linked in some way, or at least share their experiences in one way. If that other firm needs a (new) compensation consultant, this might give new opportunities (and profits). As contacts in the “old boy’s network” tend to be informal by nature and the consulting business probably is one of the most mouth to mouth marketing sensitive and reputation-based businesses (see for example ADDIN RW.CITE{{134 Dawes,Philip L. 1992}}{{134 Dawes,Philip L. 1992}}), one could imagine that compensation consultants have certain incentives to please the CEO (and indirectly the board members) by proposing him a large compensation package. I am not aware of any research specifically related to compensation consultants, neither am I aware of this type of research in relation to auditors.A possible way to test this is to check compensation consultants for firms with busy board members. If the consultant is the same for these two firms, the consultant may show his grace by proposing a higher compensation package. In some ways, this considers the board of directors to have influence on the Compensation Committee’s proposal, at least by proposing some compensation consultants. However, I also test a more restricted, namely for a sample with only busy compensation committee members.Does a consultant follow the incentives of its client?So far, not much research has distinguished between compensation consultants hired by company management or compensation consultants hired by the Compensation Committee. In fact, almost all compensation consultant research previously mentioned has – apparently – assumed that conflicted compensation consultants leads to higher pay. As I explained in paragraph 2.2.1, if a compensation consultant works for a Compensation Committee, a consultant is usually expected to act in line with the Compensation Committee’s incentives. However, this feature is not included in most research. ADDIN RW.CITE{{53 Murphy,Kevin J. 2010}}{{53 Murphy,Kevin J. 2010}} test the “repeat business” effect. They assume that consultants working exclusively for the board will have a downward pressure on executive pay. Contrary to their expectations, they find that consultants working exclusively for the board actually are associated with higher executive pay. Both ADDIN RW.CITE{{56 Voulgaris,Georgios 2010}}{{56 Voulgaris,Georgios 2010}} and contemporary research assume that a consultant’s incentive is to please the compensation committee, which independence should lead to a reasonable pay rather than a high pay and align their conclusions to the efficient contracting hypothesis.The (still unresolved) question however is if compensation consultants can act objective and independent, or that they are sensitive to (informal) pressure from their clients. However, the problem is that testing the ‘real’ incentives is quite difficult. If a compensation consultant is engaged by company management, we may assume that the consultant has incentives to please the CEO and therefore proposes a high pay package. However, difficulties exist in determining compensation committee incentives.While the compensation committee, or the board of directors in general, has the task to monitor the CEO on behalf of the shareholders, one do not have to assume that the board of directors a priori have incentives to act in the interest of shareholders. In fact, directors may have incentives to please a CEO, because they attach value to having a board seat, the compensation, the social status or the contacts. On top of that, a director may wish to improve his chances of being assigned a board seat at another company. Such incentives might still lead to higher compensation.Of course, these incentives are not observable, but board independence may be a proxy for it. However, as explained in paragraph 2.3, compensation committees shall be independent and this proxy might therefore not be of much use. Furthermore, differences exist between formal independence (i.e. familial or financial ties between the director and the company) and material independence. ADDIN RW.CITE{{123 Hwang,Byoung-Hyoun 2009}}{{123 Hwang,Byoung-Hyoun 2009}} find that there is a difference between formal and material (in)dependence of boards and that material independence (which they define as two similarities between the CEO and board members on informal criteria like military history, academic background and region of origin) is associated with lower compensation than boards which are merely formally independent.I am not aware of any research that has linked material independence with compensation consultants. However, this seems very important, since the we have reasons to assume that material independence towards the CEO could well shape a compensation committee’s incentive and, therefore, compensation consultant expectations. It goes beyond the scope of this thesis to measure social independence and its effects in relation with the use of compensation consultants, but I do provide some pointers whether or not this could be a field of research.The choice for a particular compensation consultantTo question the endogeneity problem in compensation consulting research (see paragraph 3.4.2), several papers have addressed the question what determines the choice to use a compensation consultant. Estimating a Heckman selection model for the selection of a compensation consultant, ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}} find a significant negative effect for CEO ownership (percentage of shares) and a significant positive effect for the number of compensation committee meetings. They assume that powerful CEOs are less likely to engage compensation consultants. ADDIN RW.CITE{{53 Murphy,Kevin J. 2010}}{{53 Murphy,Kevin J. 2010}} perform a propensity-score matching analysis and hypothese that the choice of compensation consultants is determined by both economic variables (such as firm size, the number of compensation committee members and shareholder return) and governance variables (is the CEO the chair of the board, the percentage of directors appointed after the CEO was appointed and the percentage of non-independent directors). They find that large firms, firms with poor stock performance and manufacturing firms are more likely to use conflicted consultants.ADDIN RW.CITE{{56 Voulgaris,Georgios 2010}}{{56 Voulgaris,Georgios 2010}} run a regression with selection variables as CEO tenure and ownership, pay package complexity, fees and location of the consultant and industry competition. They find that CEO tenure and ownership do not have significant influence on executive pay, but that significant associations exist for pay complexity (measured as the number of stock, option and LTIP schemes awarded) and fees (although they measure these as audit fees and fee ratio as non-audit fees divided by total fees).However, since almost any firm uses compensation consultants nowadays (see the descriptive statistics in paragraph 4.2.3), it is more interesting to notice what determines the choice for a particular compensation consultant (i.e. a large or boutique consultant) rather than any consultant. ADDIN RW.CITE{{134 Dawes,Philip L. 1992}}{{134 Dawes,Philip L. 1992}} suggest that prior experience with the consultant might improve the probability of being hired. Furthermore, one could expect that large consultants attract large clients, and that firms with good corporate governance are eager to engage consultants that only provide executive compensation services, thus reducing the chance at any undesirable cross-selling incentives. I don’t expect firm performance to be influential on the choice for a particular compensation consultant, but I include it as control variable. This will be more of an exploratory research.Other contributionsLack of dataAs disclosure regarding compensation consultants has only been obliged since 2006, most papers to date lack data. Most papers ADDIN RW.CITE{{53 Murphy,Kevin J. 2010; 35 Conyon,Martin J. 2009; 5 Armstrong,Chris S. 2008}}{{53 Murphy,Kevin J. 2010; 35 Conyon,Martin J. 2009; 5 Armstrong,Chris S. 2008}} use only one year of data and while ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}} use two years, they treat it as cross-sectional research. Only contemporaneous research uses multiple years, such as I do. While three or four data years are still not that much, it is possible to perform some panel data analysis.EndogeneityA problem related to the lack of data is the endogeneity problem. This is a selection problem, in the sense that some unobserved dimensions may cause firms to engage a consultant or not. Not observing this problem might lead to the conclusion that compensation consultants are a driver of executive pay, whilst in reality this may be another variable. Ideally, one would use instrumental variables. However, as they are not available and unobservable (at least in my opinion), other solutions should be found. Existing literature has tried to solve this by means of propensity-score matching ADDIN RW.CITE{{5 Armstrong,Chris S. 2008; 53 Murphy,Kevin J. 2010}}{{5 Armstrong,Chris S. 2008; 53 Murphy,Kevin J. 2010}}, a Heckman selection model ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}, a switching regression model ADDIN RW.CITE{{56 Voulgaris,Georgios 2010}}{{56 Voulgaris,Georgios 2010}}, and 2SLS regressions ADDIN RW.CITE{{46 Kabir,Rezaul 2010}}{{46 Kabir,Rezaul 2010}}. I use fixed-effect panel data analysis, which mitigates the unobserved dimension problem, while I am still able to perform regression analysis with the same data. However, the risk is that between-firm differences account for differences in executive compensation rather than within-firm effects. Fixed-effect analysis would then not be of real value in itself.ConclusionIn this chapter, I discussed existing research and proposed my own research questions. Until now, evidence suggests that compensation consultant behavior rather than compensation consultant presence might have influence on executive compensation. Most notably, a compensation consultant’s conflict of interest might lead to higher executive compensation. However, this evidence is mixed. I propose some other conflicts as proxies to measure this influence. Furthermore, I suggest that compensation consultant behavior is linked to the incentives of their principal. Last, I elaborate a little on potential determinants of the choice for a particular compensation consultant and I discuss some small contributions in terms of data and a new way to treat the endogeneity problem.Data description and methodologyIntroductionIn this chapter, I first present the dataset that I use. I also provide some descriptive characteristics of my dataset. After I elaborate on my definitive sample selection, I explain the models for the research questions that I propose.DataVariables and sourcesBasically, the data that I use can be distinguished in four categories: i. data on the use of compensation consultants, ii. compensation data, iii. financial characteristics on firms, and iv. corporate governance characteristics on firms. Most variables of the last two variables use are proven to have an effect on executive compensation and are therefore used as control variables in my pensation consultant dataNo data on the use of compensation consultants is readily available. I therefore hand-collect this information from DEF-14A statements from the EDGAR database of the U.S. Securities and Exchange Commission. As discussed in paragraph 2.3, since December 2006, it is compulsory to disclose this information.Use of compensation consultantsFor each consultant, I recorded the following items.Name of the consultant. I recorded the name of consultants (and synchronized their names). In the end I merged some consultants. For example Aon and Radford Surveys & Consulting are both part of Aon and therefore I label them to Aon / Radford. Furthermore, as will become clear from the descriptive statistics, the largest consultants own a large share of the market. Other consultants are smaller. Together, I labeled them as ‘Other Consultants’. Last, some firms mentioned the use of a compensation consultant but did not provide the name of the consultant. They were marked as ‘Not named’.Appointing body. Not every consultant is engaged by the same persons within the firm. In fact, I distinguish between the following items:‘CC’: the compensation consultant was engaged by the compensation committee. I assigned this label if it was clear that the compensation committee had actively decided to engage a consultant. It was not enough just to be provided with information from a compensation consultant;‘CCS’: the CC had the sole authority to engage a consultant. Therefore, this basically is a somewhat stricter form of ‘CC’;‘M’: the compensation consultant was engaged by company management, i.e. the CEO or the Human Resources department;‘CCM’: the compensation consultant was engaged by the compensation committee in conjunction with company management;‘CCM2’: the compensation committee hired the consultant, but meanwhile the consultant was also performing work for company management;‘CCSM2’: a combination of CCS and CCM2; and‘NA’: if it did not became clear who did appoint the compensation consultant this data item was not ascertainable.Role of the consultant. As explained in paragraph 2.2.1, a compensation consultant can have different roles. I distinguished between the following:‘Assist’: the consultant provides services which cannot be described just as providing data, he really assists in designing the compensation package of the company;‘Data’: the consultant only provides compensation data (surveys) to the company;‘Customized’: the data provided by the consultant is customized, e.g. corrected for industry or firm size;‘Attend’: it is only clear that the consultant attends meetings of the compensation committee; and‘Unknown’.Other information. I also collected data on announced changes of compensation consultants and – if provided – the reasons of the change. However, due to the limited observations that I made, I decided not to include these variables in the analysis.Conflicts of compensation consultantsI create a number of dummy variables denoting potential conflicts of interest. Assisting versus other roles. This is a dummy variable which is 1 if the firm engages a compensation consultant that has an assisting role and 0 if not. Sometimes, I use this variable in a more restricted form, when it is only one in case the compensation committee engages an assisting compensation consultant. I expect a positive influence of an assisting consultant on executive compensation.Same consultant and busy board members or busy compensation committee members. To create these dummies, I use the RiskMetrics director database. I determine all busy board members (firm-year observations which do not have busy board members are labeled ‘NA’) and compare the compensation consultants that have ‘CC’ or ‘CCS’ as appointing body. If at least one same compensation consultant is engaged, the value of the variable is 1, and 0 if not. This variable is created for both busy board members and busy compensation committee members. I expect a positive influence on executive compensation.Newly assisting consultant. This is a dummy variable with value 1 if a firm did not use an assisting compensation consultant in the prior year. Its value is 0 otherwise. A related variable is the ‘year after’ dummy variable, which equals 1 if ‘Newly assisting consultant’ was 1 in the prior year.CEO / Director career history. Using BoardEx, I also tracked down whether or not CEO’s or Directors had previous (or current) working experience with a compensation consultant. If this was the case, I compared that employer with the compensation consultant. Unfortunately, to the very limited number of observations, I had to drop this variable.Groups of compensation consultantsFor my research in the determinants of the choice of particular compensation consultants, I group the consultants into a number of groups, which are not necessarily mutually exclusive. As I am interested in the choice of compensation committees, I only include firms which are engaged in an assisting role by a compensation committee and do not work with or for company management, hence the appointing body is ‘CC’ or ‘CCS’. However, the sample includes all compensation consultants that had an assisting role.Large consultants. With large consultants, I mean five largest compensation consultants, which are F.W. Cook & Co., Hewitt, Mercer, Towers Perrin and Watson Wyatt. Because the last two merged in 2009, I also included Towers Watson, which has been one of the largest consultants in this year.Small consultants. The small consultants are all consultants that are not in the group ‘Large consultants’. The groups ‘Large consultants’ and ‘Small consultants’ are mutually exclusive.Boutique consultants. This group consists of F.W. Cook & Co., and Pearl Meyer & Partners, the two largest boutique consultants, which provide executive compensation services only and have no other consulting branches whatsoever.Other large consultants. This group consists of the ‘Large consultants’ group less F.W. Cook & Co., but includes Aon/pensation dataAll data on executive compensation is derived from Execucomp. I use the following six variables, the first four of them based on ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}. As will become clear in the description of the model, all these variables are explained variables in the models.Salary. The variable salary in Execucomp. Since salary is usually set prior to the fiscal year, I use lagged control variables, again as in ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}.Bonus. The bonus is calculated as the sum of bonus (bonus) and non-equity incentives (noneq_incent) in Execucomp.Equity-based compensation. Equity-based compensation is calculated as the sum of grant date fair value of both stock and option grants (stock_awards_fv and option_awards_fv) in Execucomp.Total compensation. This is the variable total_alt1 in Execucomp. It consists of all the following parts, as well as some minor compensation aspects.Equity-based compensation share of total compensation for the CEO. This is the share of equity-based compensation within total compensation, as calculated by ‘Equity-based compensation / Total compensation’.Financial characteristicsAs financial control variables, I use a number of variables which research has confirmed to be influential on executive compensation. Firm size. As in ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}, I use a firm’s market value of assets, as can be derived from Compustat Annual Fundamentals, as a proxy for firm size. Firm size is expected to be positively associated with executive compensation ADDIN RW.CITE{{59 Smith,Clifford W. 1992; 61 Core,John E. 1999}}{{59 Smith,Clifford W. 1992; 61 Core,John E. 1999}}.Book-to-market value. The Book-to-market value gives an indication about the investment opportunities of a firm ADDIN RW.CITE{{59 Smith,Clifford W. 1992; 61 Core,John E. 1999}}{{59 Smith,Clifford W. 1992; 61 Core,John E. 1999}}ADDIN RW.CITE{{58 Core,John 1999}}. Like ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}, I calculate the book-to-market ratio as the book value of common equity divided by the market value of common equity = (at - lt - upstk) / (csho * prcc_f). I expect a negative relationship between the book-to-market ratio and executive pay, as firms with higher growth opportunities (thus a lower book-to-market ratio) tend to have greater (equity-based) compensationADDIN RW.CITE{{59 Smith,Clifford W. 1992}}{{59 Smith,Clifford W. 1992}}.Firm performance. In various economic literature, firm performance has been established as a variable which has a profound influence on executive pay ADDIN RW.CITE{{60 Murphy,Kevin J. 2000}}{{60 Murphy,Kevin J. 2000}}. As proxy for firm performance, I use a firm?s return on assets, which I calculate as a firm’s income before extraordinary items divided by total assets (ib / at). I expect a positive relationship between firm performance and executive pay. As robustness check, I use two other variables for firm performance. First, I use the buy-and hold return, calculated as BHR= (Stock pricet+ Accumulated dividend)-Stock priceBase YearStock priceBase Year, the stock price base year being 2004. Alternatively, I calculate an industry-adjusted return on assets, being the deviation from the firm’s return on assets from the weighted industry-return on assets. All data is derived from Compustat Annual Fundamentals.Inside ownership. Inside ownership, i.e. the percentage of shares of the company owned by the CEO, is associated with better corporate governance and lower executive compensation ADDIN RW.CITE{{135 Lambert,Richard A. 1993; 61 Core,John E. 1999}}{{135 Lambert,Richard A. 1993; 61 Core,John E. 1999}}. This percentage is derived from Execucomp, namely the variable shrown_excl_opts_pct.Volatility. Stock volatility is also considered to have an effect on executive compensation. I measure firm risk as stock volatility over the monthly years over the three preceding years (e.g. for fiscal year 2007, I calculate the volatility over fiscal years 2004-2006). On one hand, one could argue that, since monitoring is difficult, the increasing noise is an argument to lower incentivesADDIN RW.CITE{{77 Aggarwal,Rajesh K. 1999}}{{77 Aggarwal,Rajesh K. 1999}}, but on the contrary the line of reasoning could be that managerial risk-aversion requires higher payADDIN RW.CITE{{58 Core,John 1999}}{{58 Core,John 1999}}. As stated by ADDIN RW.CITE{{56 Voulgaris,Georgios 2010}}{{56 Voulgaris,Georgios 2010}}, it is therefore difficult to determine the effect that firm risk might have on executive compensation.Corporate governance characteristicsI also include some proxies for good (and bad) corporate governance, which should have some influence on executive compensation.Staggered board elections. This dummy variable, extracted from the RiskMetrics Governance database, has a value of 1 if staggered board elections take place and 0 if not. The presence staggered board elections makes it more difficult to quickly overtake a firm. This impedes activist shareholders to take (full) control over a company quickly ADDIN RW.CITE{{54 Armstrong,Chris S. 2008}}{{54 Armstrong,Chris S. 2008}} and therefore is associated with a weaker governance structure ADDIN RW.CITE{{63 Daines,Robert 2001}}{{63 Daines,Robert 2001}}.Dual class shares. This dummy variable, extracted from the RiskMetrics Governance database, is considered to be associated with weaker corporate governance ADDIN RW.CITE{{64 Gompers,Paul 2003}}{{64 Gompers,Paul 2003}} and, hence, higher executive compensation, dual class shares indicate weaker pensation committee independence. While compensation committee independence is not necessarily associated with lower executive compensation ADDIN RW.CITE{{111 Anderson,Ronald C. 2003; 94 BEBCHUK,LUCIAN A. 2010}}{{111 Anderson,Ronald C. 2003; 94 BEBCHUK,LUCIAN A. 2010}} itself, its interaction with the engagement of a compensation consultant might be. The RiskMetrics director database provides information on both (compensation) committee members and director independence. Combining these two provides information about compensation committee independence. I construct it in two firms, first in a dummy variable with value 1 if all members are independent and 0 if not, and also as the fraction of members that are independent.Board independence. Similar to compensation committee independence, board independence is measured as the fraction of independent compensation committee members of total compensation committee members. Board independence is considered to have a downward influence on executive pay ADDIN RW.CITE{{61 Core,John E. 1999; 123 Hwang,Byoung-Hyoun 2009; 94 BEBCHUK,LUCIAN A. 2010}}{{61 Core,John E. 1999; 123 Hwang,Byoung-Hyoun 2009; 94 BEBCHUK,LUCIAN A. 2010}}.Equity-based compensation share of total compensation for directors. I include two variables which might something about compensation committee and board independence which goes beyond the scope of formal independence. Based on the assumption that a larger share of equity-based remuneration in total compensation committee member or board remuneration leads to better alignment of director with shareholder incentives ADDIN RW.CITE{{124 Ryan,Harley E. 2004; 129 Fich,Eliezer M. 2005; 128 Ertugrul,Mine 2008}}{{124 Ryan,Harley E. 2004; 129 Fich,Eliezer M. 2005; 128 Ertugrul,Mine 2008}}. I am aware of the fact that many other factors than compensation, such as social prestige or social dependence ADDIN RW.CITE{{2 Bebchuk,Lucian Arye 2003; 123 Hwang,Byoung-Hyoun 2009}}{{2 Bebchuk,Lucian Arye 2003; 123 Hwang,Byoung-Hyoun 2009}}, but the use of this variable nevertheless might provide any pointers. The data necessary for this variable is obtained from Execucomp’s director compensation database. It is calculated following the same procedure of the CEO-equity based compensation share. However, instead of fair value, valuation under FAS123 is used, since contrary to executive compensation, fair value valuation disclosure is not obliged for director compensation. Ideally, I would calculte the share of compensation committee members as well, but unfortunately it is not possible to link the databases containing director compensation to the database containing director information. Therefore I can only calculate the weighted average ratio of equity-based compensation for directors together.Other characteristicsIndustry Control Variables In line with existing research ADDIN RW.CITE{{40 Cadman,Brian 2010; 53 Murphy,Kevin J. 2010; 35 Conyon,Martin J. 2009}}{{40 Cadman,Brian 2010; 53 Murphy,Kevin J. 2010; 35 Conyon,Martin J. 2009}} use industry control variables, in order to control for industry competition and competition for managerial talent. I measure industry effects as the two-digit code of the North-American Industry Classification System (NAICS).Sample selectionMy data consists of approximately 6000 observations, equivalent to four years of data for S&P 1500 firms. The size of my sample is dictated by the availability of proxy statements in which compensation consultant disclosure was obliged. After December 18, 2006 it became obligatory to disclose information regarding interactions with compensation consultant. I use a number of data sources, which can be divided in three parts: compensation consultant information, board member information and ‘regular’ financial information. I elaborate on them here.One important aspect requiring further elaboration is the procedure that I followed in order to link compensation consulting data to the correct year. As a rule, Compustat assigns proxy statements filed in the month January up to and including May to the previous data year. For example, a proxy filed at May 31, 2008 is assigned with data year 2007, whereas a proxy statement filed at June 1, 2008, gets data year 2008. I match my hand-collected data accordingly. Furthermore, since Execucomp compensation summations changed since 2006 and I sometimes use lagged variables (i.e. for salary), I drop the observations from data year 2006. Data year 2006 is only used as comparison for consultant change. One extra advantage is that this mitigates any ‘transition year effects’, which might occur due to (unintended) non-compliance due to unfamiliarity with the new rules {{53 Murphy,Kevin J. 2010}}.In order to balance consultant use over the year, I preferred to use a balanced panel rather than an unbalanced panel, which might provide a distorted view on the statistics. I am aware from the danger of at a selection bias in the dataset, as missing data typically is a concern with firms in distress, mergers and other irregular situations. However, tests with the original dataset saw movements in the size of samples and therefore sometimes changes in the magnitudes of the coefficients (signs and significance often remained similar). Therefore, I find it tenable to analyze with the adjusted dataset. The definitive sample consists of 741 firms * 3 years (2007-2009), leading to a total number of 2223 observations. In this data set, a number of 3143 compensation consultants were engaged (regardless of their role). A quantitative overview of the procedure described here can be found in Appendix 8.2.Descriptive statisticsBefore I turn to the methodology, I first provide some descriptive statistics for the dataset that eventually was used. Compensation consultant hiring practicesFirst, I set out a few compensation consultant hiring practices, market shares, divisions over time and industry, etcetera. The division amongst (the largest) consultants and per year can be found in Table II, Panel A.<INSERT TABLE II PANEL A>Within this dataset (the same can be said for the original dataset – not reported), five large firms exist(ed): Hewitt Associates, Towers Perrin, F.W. Cook and Mercer. After Towers Perrin and Watson Wyatt merged, they had the largest market share in 2009. As can be seen, the 14 (13) largest firms account for almost 85% of the total market. The total number of compensation consultants engaged has been approximately equal for the years 2008 and 2009. This could be an indication that around this time, most firms are fully complying with the rules.<INSERT TABLE II PANEL B>Panel B of Table II provides an overview of industries served per compensation consultant. Compensation consultant market shares per industry seem largely comparable to the figures as a whole, with few considerable differences between market shares within a industry compared to total market share.<INSERT TABLE II PANEL C>Table II Panel C provide an overview of compensation consulting roles per consultant. In most cases, the consultant has an assisting role, although in almost one third of the cases, the role is to provide data. The five main exceptions to this rule are the boutique consultants F.W. Cook & Co., Pearl Meyer & Partners, Semler Brossy and Compensia which mostly have an assisting role, and Aon/Radford, which is best known for the Radford Compensation Surveys and therefore is a data provider in most cases.<INSERT TABLE II PANEL D>Last of the consultant descriptive statistics, Panel D of Table II provides an overview about the hiring bodies (i.e. compensation committee, company management or some hybrid form) within a firm. With the caveat that a large share of the compensation consultants cannot be attributed to one specific body, in most cases the compensation committee seems to engage a compensation consultant. Compensation consultants are far less often engaged by company management. Also, the hybrid forms (CCM, CCM2, CCSM2) are not often chosen. This stands in contrast to contemporaneous research, which implies that at least CCM2 and CCSM2 would be expected more oftenADDIN RW.CITE{{79 Tong,Naqiong 2011}}{{79 Tong,Naqiong 2011}}. Possible reasons are that disclosure of other fees has only been compulsory since 2009, information about compliance is unavailable and that in this research, decisions were made on words rather than figures (in order to act consistently with the years 2007 and 2008). Furthermore, prior to that time, it was not compulsory to report about other work. Last but not least, many of these ‘unclear’ cases have been labeled as ‘Unknown’. It is therefore possible, that the table above provides a distorted view. A similar table for compensation consultants (not reported), shows that boutique compensation consultants are employed by the compensation committee more often than the other compensation consultants.Firm and compensation characteristics<INSERT TABLE III PANEL A&B>Panels A and B of Table III provide mean and median statistics for the primary explained and explanatory variables. While in absolute terms average firm size (in millions of U.S. Dollars) is greater for firms not using an (assisting) compensation consultant, the opposite is the truth for the median (which yields significant differences). This is caused by outliers in the dataset. I mitigate this outlier influence by using the natural logarithm of assets (and executive pay), which is common in executive compensation research ADDIN RW.CITE{{5 Armstrong,Chris S. 2008; 40 Cadman,Brian 2010; 35 Conyon,Martin J. 2009; 53 Murphy,Kevin J. 2010}}{{5 Armstrong,Chris S. 2008; 40 Cadman,Brian 2010; 35 Conyon,Martin J. 2009; 53 Murphy,Kevin J. 2010}}. Firm performance metrics (i.e. the book-to-market value and the return on assets) seem comparable at both mean and median levels, with the exception for the median difference for book-to-market values between consultants in an assisting role and those who don’t: the difference is significant at a 10% confidence level. That said, executive compensation mostly is significantly higher in firms using assisting compensation consultants, both in assisting and non-assisting roles. Also, the percentage of equity-based pay is significantly higher in forms that use compensation consultants. These descriptive statistics seem comparable to similar research (e.g. ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}.Table III Panel C provides similar characteristics, but now sorted out for the four conflicts which are investigated in this thesis. The difference between assisting and non-assisting in Panel B and Panel C arises from the fact that different samples were used. Panel B uses the complete sample of 2223 observations, whereas the assisting versus non-assisting role comparison in Panel C is based on the sample that compensation consultants were used (regardless of their role). The other three conflicts described in Panel C, are based on the sample that only include firms having engaged a compensation consultant in an assisting role and also then, the number of observation was reduced because, as an example, not every firm has busy board members.The upper half of Panel C indicates that executive compensation in firms that engage conflicted compensation consultants, is significantly higher than in firms that do not, both on mean as well as median level. Also, a significantly larger part of executive pay seems to be equity-based in firms using conflicted compensation consultants.On the contrary, the conflicts mentioned in the lower part of Table III Panel C yield less significant differences. For firms with busy compensation committee members employing the same compensation consultant with at least two of their firms, salary and bonus are not significantly higher, but so do the equity component of pay and (possibly therefore) total pay and the equity-based share of executive pay. However, firms that employ an assisting compensation consultant after not having done so in the previous year, seem to have no significant differences in the change of pay compared to other firms that use an assisting consultant in the same year. While the change of total pay is higher at both mean and median level, the difference is not significant.<INSERT TABLE III PANEL C>MethodologyThe modelsTo test the effect of conflicted compensation consultants, I estimate the following model:1Executive compensationit=β0+β1it×Dit+X'βit×Control Variablesit+εit ,where Executive compensation denotes (1 + the natural logarithm) of the executive compensation measure in question, β0 is a constant, D denotes the dummy variable for the proxy of conflict of interest under investigation and Control Variables represents a set both economic and corporate governance variables that are proven to be influential on executive compensation. Last, εit is the error term. Equation (1) is primarily estimated as a linear pooling model. The sample varies with the conflict that is under investigation. Investigating the assisting versus non-assisting consultant hypothesis, I estimate this model on a sample which includes all firms that used a compensation consultant in some way; hence firms that only engage compensation consultants for (customized data) or attendance are included as well. For the other estimation, the sample consists only of firms that use a compensation consultant in an assisting role.To test the assumption that compensation consultants that are engaged by compensation committees act in line with the committee’s incentives, I estimate the following model:2 Executive compensationit=β0+β1it×Committee Consultantit+β2it×Committee Independenceit+β3it×Committee Independenceit ×Committee Consultantit + X'βit×Control Variablesit+εit,where, as new variables compared to equation (1), Committee Consultant is a dummy variable with value 1 if the firm’s compensation committee engaged a compensation consultant that did not work for or with management at the same time (i.e. the appointing body was ‘CC’ or ‘CCS’) and 0 if no compensation consultant was engaged by the compensation committee. Committee independence is a proxy variable denoting the independence of a firm’s compensation committee, where 1 is considered as ‘maximum independence’ and 0 means ‘minimum independence’. Again, I estimate equation (2) as a pooling model. It is estimated on a sample which only consists of firms that use a compensation consultant in an assisting role.To test the determinants of the choice of large compensation consultants, I estimate the following probit model:3 Pr?(Compensation Consultant=1)it=β0+β1it×Same consultant on other compensation committeeit+β2it×Firm sizeit+β3it×Return on assetsit+β4it×Committee independence+ X'βit×Control Variablesit+εitAs new variables compared to equations (1) and (2), Compensation Consultant represents a dummy variable with value 1 if the compensation consultant was hired by the compensation committee and did not work with or for company management at the same time (i.e. ‘CC’ or ‘CCS’) and the consultant consists to the group of consultants that is under investigation in the estimation, for example small consultants, or boutique compensation consultants. It’s value is 0 if not all these conditions were met. Same consultant on other compensation committee is also a dummy variable with value 1 if the firm has busy compensation committee members, which engage the same compensation consultant on their other firm and 0 if the firm has busy compensation committee members, which do not engage the same compensation consultant on their other firm. Its value is NA if the firm does not have busy compensation committee members in the particular year. Firm size and return on assets are control variables in the other regressions, but are mentioned in this regression because their particular influence could be expected. In equation (3), Committee independence is only represented by a dummy variable with value 1 if all compensation committee members are independent and 0 if this is not the case.Equation (3) is estimated by a probit regression. I estimate this on a sample which consists of all firms that use a compensation consultant in an assisting role.Robust covariance estimationBecause (non-reported) tests confirm that the model is subject to both serial correlation and heteroskedasticity, I calculate adjusted standard errors using White-robust covariance estimation. This is a common method in executive compensation research (see for example ADDIN RW.CITE{{40 Cadman,Brian 2010; 35 Conyon,Martin J. 2009; 136 Loureiro,Gilberto R. 2011; 137 Conyon,M.J., He, L. 2010}}{{40 Cadman,Brian 2010; 35 Conyon,Martin J. 2009; 136 Loureiro,Gilberto R. 2011; 137 Conyon,M.J., He, L. 2010}}). More specific, I use the “white2” estimator from the ‘plm’ and ‘sandwhich’ package of R, which allows for general heteroskedasticity. Hausman tests on the models (not reported) suggest that a fixed-effects model rather than a random-effects model should be used, and therefore I estimate some of them too, once again using the “white2” estimator.Outlier influenceTo reduce the influence of outliers, I take the natural logarithm of Executive compensation and Firm size (i.e. the market value of assets) (as was explained by ADDIN RW.CITE{{5 Armstrong,Chris S. 2008}}{{5 Armstrong,Chris S. 2008}}, but is widely performed in similar research) and I winsorize Return-on-assets, Book-to-market value and Volatility at the 1% and 99% percentile (as was also done by ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}). EndogeneityCompensation consultants might not necessarily be the reason that compensation packages are larger in firms using their advice. It is well possible that they are a way for a company’s executive to extract more rents (i.e. higher compensation) from the company. If this is the case, it is wrong to conclude that compensation consultants are a driver of (excessive) CEO pay. Ideally, one would have perfect instrumental variables, controlling for the choice of a consultant and only then measure the influence of consultants. However, since such a variable is not directly available (if it exists at all), other ways are necessary of dealing with the endogeneity problem. Most studies try to mitigate this effect using a propensity-score matching approach (ADDIN RW.CITE{{54 Armstrong,Chris S. 2008; 53 Murphy,Kevin J. 2010}}{{54 Armstrong,Chris S. 2008; 53 Murphy,Kevin J. 2010}}), a Heckman selection model (ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}, or a switching regression model (ADDIN RW.CITE{{56 Voulgaris,Georgios 2010}}{{56 Voulgaris,Georgios 2010}}).Being one of the first using panel data to analyze the influence of compensation consultants, this allows me to perform fixed-effect analysis. With the help of this technique, I am able to analyse within-firm effects that occur after a firm has engaged (or dropped) a compensation consultant. It could well be that the explanatory power of the fixed-effect analysis is very low. While the analysis then may not be very worthwhile in itself, it suggests that between-firm variation has a greater influence on executive pay than within-firm variation, that for example arises due to the use of a newly-used compensation consultant.MulticollinearityTo prevent problems of multicollinearity, I run a correlation table (not reported). Judging on its values, correlation between variables are low. I therefore expect no multicollinearity problems. ConclusionIn this chapter, I explained the data and provided some descriptive statistics, suggesting the need to further analyze the effects of compensation consultants. After that, I explained the three models that I estimate in the next chapter.ResultsIntroductionThis chapter presents the results of the research of this thesis. First, I present the results for conflicted compensation consultants, based on pooling as well as fixed-effect regressions. Furthermore, I provide some robustness tests. After this primary object of analysis, I show some results of exploratory research into the influence of compensation committee independence and what determines the choice of particular compensation consultants.Primary resultsAssisting versus non-assisting consultant hypothesis<INSERT TABLE IV PANEL A>Panel A of Table IV provides evidence on the role of assisting compensation consultants. Based on pooling regressions on a sample of firms using compensation consultants, the dummy variable Assisting Role has a significant positive effect on each measure of the natural logarithm of executive compensation, as well as on the equity-based percentage of total executive pay. Even with two proxies for corporate governance effects on executive compensation included, an assisting consultants shows an significant positive effect on executive compensation. Furthermore, firm size (measured as the natural logarithm of firm assets) and the book-to-market value have a significant effect on executive compensation (as expected). The coefficients for Return on Assets are sometimes significant, but mostly in the direction that was not expected. It is only strongly positive related to bonus, which comes as no surprise since a bonus is typically linked to firm performance. However, its sign is in line with for example ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}. As can be seen, the Adjusted R2-values are typically low, but quite reasonable compared to similar research. However, caution is required with regard to the interpretation of the coefficients, as the quantities of variables differ: firm size (in millions) and executive compensation (in thousands) are measured as the natural logarithm of U.S., the dummy variables as 0 or 1, volatility as the natural logarithm of the standard deviation of monthly stock returns and the rest are ratios.Busy board members with the same compensation consultant on multiple boards<INSERT TABLE IV PANEL B>Table IV Panel B provides some evidence for the hypothesis that firms with busy board members who engage the same compensation consultant, are associated with higher executive pay than firms with busy board members, but not the same compensation consultant. This indicates that compensation consultants might be sensitive to rehiring incentives and exhibit such feelings by proposing a slight bonus on top of ‘usual’ remuneration. The difference seems to arise in the equity-based share of pay, also leading to a significant higher total executive pay, also this cannot be confirmed by the ratio of the percentage of equity-based pay in total pay. Furthermore, the models show many similarities to the models as shown in Panel A.Busy compensation committee members with the same compensation consultant on multiple boardsPanel C of Table IV provides similar analysis as panel B, but is now restricted to busy compensation committee members. One implicit assumption of this restriction is that compensation committees are the only ones to make proposals about compensation consultants. It seems hard to believe that other board members have no influence on executive compensation whatsoever, or could not propose compensation consultants. Anyway, Panel C provides very limited evidence that busy compensation committee members with the same compensation consultant on multiple boards, lead to higher executive pay. Only the equity-based compensation (and its share in total compensation) show a significant positive contribution of this potential conflict of interest.<INSERT TABLE IV PANEL C>Newly assisting consultant<INSERT TABLE IV PANEL D>Table IV Panel D provides no evidence that a newly hired consultant gets a bigger change in absolute pay than firms who did employ a compensation consultant in the previous year as well. Furthermore, given the very low Adjusted R2s, the models do not explain a lot.To be clear, this model compares the change in compensation between firms that newly hire a compensation consultant and firms that already had a compensation consultant. I also ran a similar model with the percentage of pay change as explained variable. This yields comparable results. These models does not compare what happens with executive compensation within a firm. I investigate this using fixed-effect regressions, which results can be found in Table III Panel E. This table also includes a dummy variable for the year after the compensation consultant was engaged.<INSERT TABLE IV PANEL E>The fixed-effect regressions have low Adjusted R2’s. This implies that between-firm differences have much more influence on executive compensation than within-firm differences{{57 Petersen, T. 2004}}. Combined with the absence of significant results, this suggests that the influence of compensation consultants is rather exaggerated as firms have no significant increases in executive pay after engaging compensation consultants, or that the effects of the use of compensation consultants tend to be visible after a longer time.Despite their insignificance, and doubts that may occur regarding the validity of the model, results indicate that compensation consultants could have a negative attitude towards bonuses. This might be due to the negative public sentiments about bonuses, which could cause compensation consultants advise their clients to lower bonuses (and possibly provide this compensation in another form).As robustness checks, I performed these regressions in various forms that are not reported, for example without economic and corporate governance variables, including a dummy variable indicating that a consultant was used in the year before engagement (in another sample), and with delta compensation as the explained variable. Results were more or less similar.I also run fixed-effect estimations for the conflicts presented in Table IV Panel A-D. However, results are very similar to fixed-effects in Table IV Panel E. Adjusted R2-values are 0.03 at best, and the only ‘conflict’ being significant is the Consultant in an assisting role. Robustness testsAs robustness checks as in ADDIN RW.CITE{{40 Cadman,Brian 2010}}{{40 Cadman,Brian 2010}}, I perform robustness tests where I replace the firm performance metric from Return on Assets to either industry-adjusted Return on Assets, stock performance as measured by the buy-and-hold-return, and an adjusted return-on-assets variable which only takes into account negative return-on-assets. As compensation is often linked to performance, alternative measures could well have influence.Table V provides the coefficient and t-statistic for the conflicts that have been under consideration in Table IV, panels A-D. Coefficient magnitude and significance is typically similar to the original regressions, as performed in Table IV. Also, Adjusted R2-values typically look similar to these in which only the firm performance metric is changed. It therefore seems safe to assume that the results are robust to alternative firm performance metrics.Independent compensation committeesAs explained in paragraph 3.2.2, like in similar research, an implicit assumption until now has been that conflicted compensation consultants have a tendency to advise on higher executive compensation. However, I also proposed that the consultants incentive could also be to justify a reasonable pay rather than a high pay, because a compensation consultant would want to justify his principal’s expectations rather than the CEO (who could of course be his principal). Knowing from literature that a compensation committee’s incentives could vary based on both observable and non-observable dependence, it is difficult to measure the compensation consultant incentives. I perform analysis on this, measuring compensation committee independence with four proxies: a dummy variable for compensation committee independence, being 1 if all committee members are independent and 0 if not; the fraction of independent compensation committee members; the fraction of independent board members; and the share of equity-based compensation in total director compensation. The results are presented in Table VI.<INSERT TABLE VI>The dummy variable for ‘consultant assisted committee solely’ consists of only observations where only the compensation committee and not company management was assisted by the consultant (i.e. CC or CCS). I chose for this dummy in order to abstract the possible independence effect as far as possible. For example, if CCM2 or CCSM2 were included, it would be necessary to determine two effects, namely the ‘conflicted consultant’ effect as well as the ‘committee independence’ effect.However, caution with the interpretation of these results is still required, mainly for two reasons. First, much value is attached to the dummy variable that represents the body that appointed the consultant. As was explained in the descriptive statistics section, many of the observations could not be attributed to a specific body. Moreover, as disclosure about the hiring of consultants for non-executive compensation services was not obliged until 2009, observations recorded as ‘only working for the compensation consultant’ (CC or CCS), could actual well better be labeled as ‘working with/for management too’ (CCM, CCM2 or CCSM2). It therefore could be well difficult to measure results. Secondly, results could be subject to a selection bias, because independent compensation committees could prevent their compensation consultant from working with company management, whereas less independent compensation committees might not do so.Interpreting the signs of the coefficients (not the magnitudes), two of three coefficients act according conform expectation (despite their insignificance). The Consultant assisted committee solely coefficient is positive, which is logical, since it should be interpreted under the condition that the compensation committee is not independent. Conform expectations, this leads to higher executive compensation. On the contrary, when its interaction with independence is considered (hence Consultant assisted committee solely x Compensation committee is independent), this leads to lower executive compensation, albeit not significant. The Independent compensation committee, which should be interpreted under the condition that the compensation consultant does not assist the compensation committee solely (or not at all), is significantly positive while it possibly should be expected negatively, assuming that an independent compensation committee has a downward pressure on executive compensation. However, its coefficient is also positive when removing both the consultant assisted committee solely variable and the interaction effect (which is in line with ADDIN RW.CITE{{53 Murphy,Kevin J. 2010}}{{53 Murphy,Kevin J. 2010}}).Probit analysisIn the previous paragraphs, some evidence was provided on the effects of conflicted compensation consultants on executive compensation. In this paragraph however, I investigate a different object of study, namely what determines the choice of a particular compensation consultant. Table VII presents the results.<INSERT TABLE VII>I narrow the sample to only firms that engaged a compensation consultant. Furthermore, I have interest in two categories of opposites: large consultants versus small consultants (as determined by top-5 versus all other consultants), and boutique consultants versus consultants (as defined by Pearl Meyer and F.W. Cook & Co. versus the other top-6 consultants, so one more consultant was included in order to possibly to keep the top consultants at the same number of consultants) offering a broad range of services. These were determined by using a dummy variable with value 1 if a compensation committee engaged the firm and it was working with or for the company management at the same (i.e. CC or CCS).For all regressions, it is interesting that for three of the four regressions, the engagement of the same consultant by a busy compensation committee member is a strongly positive associated with the choice of the compensation consultant. This indicates that previous experience with the consultant increases the probability of getting hired by another firm. Regarding the large versus small consultant comparison, conform expectations, larger consultants attract larger clients, as becomes clear from the coefficients for firm size, although the coefficient for other firms. Also, the choice of a large consultant seems to be associated with lower firm performance, although only the opposite coefficient is significant. Furthermore, the coefficients for staggered board elections are significant, but is positive for large consultants and negative for smaller consultants. For large consultants, increasing CEO ownership also seems to decrease the probability that a compensation committees chooses for a large compensation consultant. The comparison between boutique consultants and multi-service consultants is interesting in the sense that independent compensation committees are expected to choose for boutique consultants in order to mitigate the risk that the compensation consultant becomes conflicted (for example, in order to sell more lucrative businesses). There are a few contrasts between the categories. For example, whereas previous experience (the same compensation consultant on multiple boards) significantly increases the probability of a boutique consultant to be engaged, while this has a slight negative, but insignificant effect at the other large compensation consultants. This suggests that the positive experience is paying off for boutique consultants and small consultants, rather than large consultants. Indeed, an extra probit estimation for the ‘other large consultants’ (top-5 without Pearl Meyer) yields a similar coefficient as the ‘other large consultants’ as in the last column. Significant differences exist for Return on Assets, being significantly positive for boutique consultants, but significantly negative for the other large consultants. The coefficients are larger than the real effect is, because a one percent increase in firm performance leads to only a 1% x 1.54 = 0.0154 increase in the probability of being engaged for a boutique consultant. That said, compensation committee independence, a coefficient with a reasonably strong theoretical foundation, has a negative and insignificant contribution to the probability of a boutique consultant of to be engaged as compensation consultant. Moreover, it is positive (albeit again insignificant) for the large multi-service consultants.ConclusionIn this chapter, I provided some evidence that compensation consultants with a stronger rehiring incentive are associated with higher executive pay. Consultants with an assisting role are associated with higher executive compensation than compensation consultants who provide less far-stretching services. Furthermore, some evidence suggests that compensation consultants which serve more firms of busy board or compensation committee members, are associated with higher pay. However, the difference seems to exist between firms rather than within firms, as is suggested by fixed-effect analysis. Firms that engage a compensation consultant after having not done so in the previous year, see a decrease in executive pay. All results are robust to alternative performance metrics.Furthermore, I have investigated the usually untested assumption that a compensation consultant acts in line with his principal’s interest. While the signs are conform expectation, results are significant. However, the amount of noise in the variables and the potential selection effect that my test is subject to, could be a reason for that. Furthermore, I did not take into account social independence.Last, I did some exploratory research on the determinants of the choice of a particular consultant. Evidence suggests that prior experience with the consultant (on another board) is important for both boutique and small consultants. Some differences seem to exist between large and small firms. Large consultants attract larger firms and some difference in firm performance seems to exist. On the other hand, the choice for a boutique consultant rather than a multi-service consultant would be expected to be associated with compensation committee independence. However, this is not the case.ConclusionThe role of compensation consultants has been controversial and I foresee no change in the near future. While they are initially were considered as to help in the design of optimal compensation packages, compensation consultants have not been able to take away the common perception that they are helping executives in extracting higher compensation from the firm (and the shareholders). As was described in chapter REF _Ref292481263 \r \h 2, this has resulted in the SEC imposing stricter regulation on disclosure regarding interactions with compensation consultants.Despite these prejudices, research has not yet been able to draw a definite conclusion about the role of compensation consultants and their contribution to executive pay, not in their presence and not in their actions. In this thesis, I have not been able to do that as well. Nevertheless, I have provided some evidence that compensation consultants act in order to be rehired by their principals. Furthermore, I have suggested that knowledge and prior experience with compensation consultants at another firm might shape expectations, resulting in higher executive pay. To my knowledge, I have been the first to provide evidence for this matter. The results of my fixed-analysis effects also suggest that within-firm changes in executive pay cannot be accounted to the hiring of a compensation consultant. Due to the regulation, since 2009 it is possible to have ‘hard’ facts about compensation consultants’ cross-selling incentives. I expect future research to be able to find stronger relationships between compensation consultants and executive compensation, as is indicated by for example the research of ADDIN RW.CITE{{79 Tong,Naqiong 2011}}{{79 Tong,Naqiong 2011}}.In my opinion, an aspect that certainly requires further research is determining the incentives of the principal’s. As I have explained, in principle it should matter whether or not a Compensation Committee is really independent. The actions of a compensation consultant will depend on the wishes of his client. I tried to find some evidence suggesting that compensation consultants align their actions to their principals’ incentives, but this is insignificant. This might be attributed to noisy disclosure (which has been solved with the 2009 regulation, as now is clear whether or not a compensation consultant has worked at non-executive compensation assignments) and the lack of better ‘independence’ measures: better disclosure combined with social independence as proposed by ADDIN RW.CITE{{123 Hwang,Byoung-Hyoun 2009}}{{123 Hwang,Byoung-Hyoun 2009}} might show such effects. However, it is also possible that a selection effect exists, and that Compensation Committees who do not allow compensation consultants to work on non-executive compensation consulting assignments are really independent, whereas dependent Compensation Committees are more willing to allow this.Last but not least, I suggest that more research should be done into the determinants of the choice for a particular compensation consultant. As I have shown, the choice for a compensation consultant might not be entirely objective, since prior experience with the consultant is a pre to assign the job to him. This in turn might lead to higher executive compensation. Investors (or directors) concerned with compensation consultant independence, may appreciate such information.In any case, the jury is still out various aspects of their role and their contributions to executive compensation. Given the increasing regulation and ongoing research in the topic, I expect developments which will provide definite insights and clarity for investors.LiteratureAbowd, J. M., & Kaplan, D. S. (1999). Executive compensation: Six questions that need answering. Journal of Economic Perspectives, 13(4), 145. Aggarwal, R. K., & Samwick, A. A. (1999). The other side of the Trade‐off: The impact of risk on executive compensation. The Journal of Political Economy, 107(1), pp. 65-105. Anderson, R. C., & Bizjak, J. M. (2003). An empirical examination of the role of the CEO and the compensation committee in structuring executive pay. Journal of Banking & Finance, 27(7), 1323-1348. Armstrong, C. S., Ittner, C. D., & Larcker, D. F. (2008). Economic characteristics, corporate governance, and the influence of compensation consultants on executive pay levels SSRN.BEBCHUK, L. A., GRINSTEIN, Y., & PEYER, U. (2010). Lucky CEOs and lucky directors. The Journal of Finance, 65(6), 2363-2401. Bebchuk, L. A., & Fried, J. M. (2003). Executive compensation as an agency problem. Journal of Economic Perspectives, 17 3, 71-92. Bender, R. (2008). Paying for advice: The role of the remuneration consultant in UK listed companies. Paper presented at the Retrieved from , M. (2000). Agents with and without principals. The American Economic Review, 90(2), 203. Bizjak, J. M., Lemmon, M. L., & Naveen, L. (2008). Does the use of peer groups contribute to higher pay and less efficient compensation? Journal of Financial Economics, 90(2), 152-168. Bruce, A., Buck, T., & Main, B. G. M. (2005). Top executive remuneration: A view from europe*. Journal of Management Studies, 42(7), 1493-1506. Cadman, B. (2010). The incentives of compensation consultants and CEO pay. Journal of Accounting & Economics, 49(3), 263. Campos, R. C. (2007). Remarks before the 2007 summit on executive compensation Retrieved from ., He, L. (2010). Executive compensation in china. Conyon, M. J., Peck, S. I., & Sadler, G. V. (2009). Compensation consultants and executive pay: Evidence from the united states and the united kingdom. Academy of Management Perspectives, 23(1), 43-55. Core, J. E., Guay, W. R., & Larcker, D. F. (2003). Executive equity compensation and incentives: A survey. Economic Policy Review (19320426), 9(1), 27. Core, J. E., Holthausen, R. W., & Larcker, D. F. (1999). Corporate governance, chief executive officer compensation, and firm performance. Journal of Financial Economics, 51(3), 371-406. Core, J., & Guay, W. (1999). The use of equity grants to manage optimal equity incentive levels. Journal of Accounting and Economics, 28(2), 151-184. Daines, R., & Klausner, M. (2001). Do IPO charters maximize firm value? antitakeover protection in IPOs. Journal of Law, Economics, and Organization, 17(1), 83-120. Dawes, P. L., Dowling, G. R., & Patterson, P. G. (1992). Criteria used to select management consultants. Industrial Marketing Management, 21(3), 187-193. Dew-Becker, I. (2009). How much sunlight does it take to disinfect a boardroom? A short history of executive compensation regulation in america. CESifo Economic Studies, 55(3-4), 434-457. DITTMANN, I., & MAUG, E. (2007). Lower salaries and no options? on the optimal structure of executive pay. The Journal of Finance, 62(1), 303-343. Edmans, A., & Gabaix, X. (2009). Is CEO pay really inefficient? A survey of new optimal contracting theories. European Financial Management, 15(3), 486-496. Ertugrul, M., & Hegde, S. (2008). Board compensation practices and agency costs of debt. Journal of Corporate Finance, 14(5), 512-531. Fich, E. M., & Shivdasani, A. (2005). The impact of stock-option compensation for outside directors on firm value. Journal of Business, 78(6), 2229-2254. Geiler, P., & Renneboog, L. (2010). Executive compensation: Incentives and externalities. In H. Kent Baker, & R. C. Anderson (Eds.), Corporate governance: A synthesis of theory, research and practice (1st ed., ). Hoboken, New Jersey: John Wiley & Sons, Inc.Hazen, T. L. (2009). Securities regulation in a nutshell (10th ed.). United States of America: West.Higgins, A. (2007). The effect of compensation consultants: A study of market share and compensation policy advice. Portland, Maine: The Corporate Library. Hwang, B., & Kim, S. (2009). It pays to have friends. Journal of Financial Economics, 93(1), 138-158. Icahn, C. Compensation consultants grease the executive pay casino. Retrieved February 24, 2011, from Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305-360. Kabir, R., & Minhat, M. (2010). Multiple compensation consultants and CEO pay SSRN.Lambert, R. A., Larcker, D. F., & Weigelt, K. (1993). The structure of organizational incentives. Administrative Science Quarterly, 38(3), pp. 438-461. Loureiro, G. R., Makhija, A. K., & Zhang, D. (2011). Why do some CEOS work for a one dollar salary? SSRN eLibrary, Mishel, L. CEO-to-worker pay imbalance grows. Retrieved 3/8, 2011, from Murphy, K. J. (1999). Chapter 38 executive compensation. In Orley C. Ashenfelter and David Card (Ed.), Handbook of labor economics (pp. 2485-2563) Elsevier.Murphy, K. J. (2000). Performance standards in incentive contracts. Journal of Accounting and Economics, 30(3), 245-278. Murphy, K. J., & Sandino, T. (2010). Executive pay and “independent” compensation consultants. Journal of Accounting and Economics, 49(3), 247-262. Petersen, T. (2004). In Hardy M. A., Bryman A. (Eds.), Handbook of data analysis. London: SAGE.Ryan, H. E., & Wiggins, R. A. (2004). Who is in whose pocket? director compensation, board independence, and barriers to effective monitoring. Journal of Financial Economics, 73(3), 497-524. Smith, C. W., & Watts, R. L. (1992). The investment opportunity set and corporate financing, dividend, and compensation policies. Journal of Financial Economics, 32(3), 263-292. Tong, N., & Cen, W. (2011). Compensation consultant independence and CEO pay. SSRN eLibrary, Towers Watson. (2010). Putting clients first: World-leading executive compensation advisory services from towers watsonU.S. Securities and Exchange Commission. SEC proposes rules requiring listing standards for compensation committees and compensation consultants. Retrieved 5/6, 2011, from U.S. Securities and Exchange Commission. (2006a). 17 CFR parts 228, 229 et al. executive compensation and related person disclosure; final rule and proposed rule.17 CFR PARTS 228, 229, 239, 240, 245, 249 AND 274 RELEASE NOS. 33-8655 Proposed Rule regarding Executive Compensation and Related Party Disclosure Agency, (2006b). 17 CFR PARTS 229, 239, 240, 249, 270 and 274 [RELEASE NOS. 33-9052; 34-60280; IC-28817; File no. S7-13-09] RIN 3235-AK28 PROXY DISCLOSURE AND SOLICITATION ENHANCEMENTS AGENCY: Securities and Exchange Commission. ACTION: Proposed Rule. (2009a). 17 CFR Parts 229, 239, 240, Et Al. Proxy Disclosure Enhancements; Final Rule, (2009b). Request for Rulemaking Requiring Companies to Disclose in the Proxy Statement the Fees Associated with all Engagements for a Single Company and any Ownership Interest a Consultant Working for the Compensation Committee may have in the Parent Consulting Firm, (2008). Voulgaris, G., Stathopoulos, K., & Walker, M. (2010). Compensation consultants and CEO pay: UK evidence. Corporate Governance: An International Review, 18(6), 511-526. Waxman, H. A. (2007). Executive pay: Conflicts of interest among compensation consultantsU.S. House of Representatives. YABLON, C. (1992). IN SEARCH OF EXCESS - CRYSTAL,G. (COLUMBIA LAW REVIEW, VOL 92, PG 1868, 1992). Columbia Law Review, 92(8), R3-R3. AppendicesAppendix A: LegislationTo followAppendix B: Sample selectionTo followAppendix C: List of variablesTo follow ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download