Executive compensation in public and private equity
firms
A European analysis of the effects in CEO compensation structure during the
transformation from a private to a public entity.
Name: Constantijn A. van de Rijdt Student number: 0596264
Thesis Supervisor: dr. J.E. Ligterink Document: Master thesis Date: April 2017
MSc Business Economics, specialization Finance
Statement of originality
This document is written by student Constantijn van de Rijdt who declares to take
full responsibility for the contents of this document.
I, Constantijn van de Rijdt, declare that the text and the work presented in this
document is original and that no sources other than those mentioned in the text and
its references have been used in creating it. The Faculty of Economics and Business is
responsible solely for the supervision of completion of the work, not for the contents.
Preface
This thesis deals with the effect of an IPO on the compensation for CEO's in Europe and is originated from a fascination for incentives in corporate settings. Along the process of formulating the research question and supported by a course in Corporate Finance the
intransparent, underresearched field of Private Equity designate the field of research. Little is known about the effect of managerial incentives, especially in Europe, in this particular investment asset. The aim of this study is to contribute to that gap and to enhance the analysis of providing optimal incentives. A better estimation of the direction of a
compensation element empowers organizations to prepare for the effects of the IPO and gives a better understanding of possible scenarios.
With this thesis I’ll complete the Master program Finance at the University of Amsterdam and also end my period at the University of Amsterdam. During this period, I have been privileged to experience a steep learning curve and build up strong economic knowledge. The completion of this thesis gives me the opportunity to express my gratitude to several persons in particular. First of all, I would like to thank my thesis supervisor Dr. Jeroen Ligterink for his support, clear guidance, and sharing his knowledge with me. Secondly, this thesis would not have been possible without the opportunity provided by PwC, in particular by Philip Siekman and Janet Visbeen. I have learned many new things about reward, tax and deals, and I am grateful for this opportunity. Furthermore, I would like to thank Paul de Winter, Anna Duijsings and Bastian Dorenkamper for their guidance and support. I have very much appreciated the warm welcome and support I received from the team Reward at PwC. My sincere thanks to you all!
Finally, I want to say thank you to all of my dear friends for your understanding and interest. I am deeply grateful for the warmth and support of my parents, sisters and the rest of my family. My dearest gratitude goes to my girlfriend, for her endless support, respect, love and humor.
April 2017, Amsterdam
Abstract
This paper examines the effect of an initial public offering of a Private Equity sponsored portfolio company on the compensation of the chief executive officer (CEO). Only a few prior studies have specifically examined managerial incentives at Private Equity (PE) sponsored deals and to the knowledge of the author it is the first performed for leading European countries. The aim of this study is to contribute to the gap of optimal managerial incentives in corporate settings and to make the PE industry more transparent. Knowing better what will happen changing the financial structure of an organization opens the black box of PE a bit more. The methods employed aim to present the effect of an IPO on similar companies with different ownership and financial structures by matching on size, debt and profitability. Using a sample of PE sponsored companies in the period 2005-2015 operating in European markets, evidence is provided that salary and bonuses increase after an initial public offering and the percentage of ownership and the ratio between bonus and fixed cash compensation decreases.
1. Introduction
On 28 October 2016, Financieel Dagblad published an article about the chief executive officer (CEO) of NXP Semiconductors. Qualcomm, a U.S.-based firm also operating in the semiconductor market, acquired NXP for approximately €43 billion. NXP’s CEO, Rick Clemmer, received $428 million from selling his portion of ownership stake in the company. In 2006, a consortium of private-equity (PE) firms led by Kohlberg, Kravis, Roberts, & Co (KKR). had acquired NXP for approximately €5.5 billion. In 2009, NXP experienced economic difficulties and had a high debt burden and a low level of profitability. To reverse this situation, the PE consortium asked Clemmer to lead the restructuring. After some time, NXP recovered, the share price quintupled in almost 6 years and over time Clemmer the value of his portfolio was $428 million consisting of numerous equity-based incentives.
A PE firm buys a company, enhance margins through higher prices or lower products costs, extract value, and sells it for a profit. The case of NXP is a perfect example of value creation during a PE investment cycle. The CEO, who has clearly contributed to the success, received a significant portion of the created value in the form of equity-based incentives. The incentives of the CEO and shareholders, represented by the PE firm KKR, are strongly aligned.
Private-equity firms create value by improving management and mitigating agency problems. Leveraged buyouts (LBOs), a specific form of PE, are based on three main pillars (Kaplan, 1989). The first is funding activities, and mainly involves the acquisition of companies or subsidiaries, with a high degree of leverage. Mandatory debt payments prevent managers from spending excess cash on value-destroying activities. The second pillar is the
enforcement of corporate governance instruments. Strong internal governance aligns the interests of shareholders and managers. Finally, the last pillar is the stimulation of managers to purchase or increase ownership in the firm. The increased ownership creates a stronger link between pay and performance. In other words, the principal-agent problem mitigates, because managers become owners of the firm for a larger portion. The CEO has greater motivation to serve the interests of shareholders, since he or she shares those interests. When a firm’s executives are also owners, the interests of shareholders and managers converge to a greater extent.
Executive compensation affects company performance in several ways, a proper compensation package will attract and retain the best candidates in the market. Highly motivated CEOs deliver value to their firms (Baker & Jensen, 1988). On the other hand, there are many examples of corporate scandals in which the CEO destroyed value to the detriment of shareholders. The compensation package of a CEO can act as the solution of the agency problem by aligning the interest of both parties, but it can also act as the source of the agency
problem when compensation structure is not optimal and leaves room for other behavior than acting in the best interest of the shareholders. An extreme case is that of Dennis Kozlowski, the CEO of Tyco International Ltd. Kozlowski was living in a $30 million
Manhattan apartment, with the company allegedly footing the bill. This executive sought to pursue his own gains instead of acting in the best interests of the firm’s owners, its
shareholders. Jackson (2012) explained this outcome with the managerial power hypothesis: To the detriment of shareholders, managerial power continues to drive executive
compensation in most public corporations. According to the managerial power hypothesis, public organizations have a suboptimal reward structure, since managers could potentially influence the remuneration policy. The level and structure of the CEO compensation plays an important role in incentivizing CEOs to put forth an appropriate level of effort.
Although these two contrasting examples are both very extreme, they provide a good indication how value can be created or destroyed and how the underlying mechanism will operate when the compensation structure is optimal structured or not. The one occurred in a public company and the other in a private company, financially sponsored by a PE firm. The first created value resulting in a successful exit and the second clearly destroyed value for the firm. In recent years PE firms tend to outperform other investment classes and researchers aim to assess the added value of the PE firms. Understanding the determinants of value creation strategies chosen by PE-firms is essential for financial growth and better
performance for investors. Corporate governance, the framework of rules and practices by which the board of directors ensures accountability and transparency in the relation with its stakeholders, is vital to society's creation of economic well-being and hence vital in the creation of economic value. Creating a better understanding of corporate governance mechanisms enlightens the discussion of potential marginal improvements. Furthermore it stimulate major institutional changes in places where they need to be made. This paper examines one element, the CEO compensation, of corporate governance and make the comparison with PE sponsored companies and listed companies in Europe. The main research question that it explores is: How does an initial public offering (IPO) affect CEO compensation packages at previously PE-sponsored portfolio companies in Europe?
Numerous studies have examined CEO compensation packages in listed firms, but only a few have distinguished between public and private companies. The few left all examined this topic in the U.S and non was focused on the European market. Gao, Lemmon, and Li (2012) did not specifically focus on PE-sponsored companies. Rather, they investigated private companies in general. Both Gao, Lemmon and Li (2012) and Leslie and Oyer (2008) found that the fixed cash component of compensation (i.e. the salary) is higher at public firms. This
might be indicative of the managerial power of CEOs to pursue their own interests. Another possibility is that strong corporate governance policies, such as independent boards and institutional shareholders tend to be more objective when setting pay scales. Along those lines, Leslie and Oyer (2008) and Jackson (2012) found a stronger link between pay and performance (bonuses) in PE-backed portfolio companies than in listed firms. Although Gao, Lemmon, and Li (2012) contradicted this finding about the connection between pay and performance, they did not focus on bonuses in particular, but rather see cash payments, fixed or variable as one. Furthermore, Jackson (2012) and Leslie and Oyer (2008) found a
reduction of CEO ownership after the firms went public. The latter even reports a reduction of almost 50% of the current ownership stake of the CEO. Jackson (2012) explains the retention of shares in a PE-environment by the undivided loyalty to shareholders created by one the pillars of PE i.e. the mantra of managers as owners. This claim seems to hold based on the results of Jackson (2012).
Over the past 30 years the role of PE firms changed from outside raiders to prominent advisors. Members of the PE-firm take seats the boards of portfolio companies to amplify their influence in and control over strategic decisions. This change of role is attributed to increased competition in the PE industry (Heel and Kehoe, 2005). Where the first wave of PE is defined by financial engineering strategies, such as finding alternative sources of financing and taking on leverage to change the firm’s capital structure. The second wave involves more complicated and broader strategies, including changing the corporate governance structure, operational engineering but also financial engineering. (Kaplan and Strömberg, 2009). This new acquired role of PE firms displays their oppressive influence they have on portfolio companies and during drastically reorganizations. It is essential to understand the effect of this reorganizations and research determinants of potential value creation. Due to the closed characters of the PE industry this field is under researched, mostly because missing data of sufficient quality. The main part of research on PE is on the performance of their funds, only a few focus on the determinants of value creation, such as the the corporate governance structures or the culture of an organization.
This study explores the remuneration policy this still under-researched, especially in Europe, field of private companies and PE-sponsored firms. No previous study has focused only on European PE-sponsored firms and compared fixed salaries, bonuses, and ownership percentages for CEOs of European PE-sponsored firms both before and after an IPO. Analyzing differences in executive pay is an important contribution to the field of corporate governance and more in general finance. What this papers adds to the few existing researches is a comparison of PE-sponsored companies with public companies by matching them on size
and debt rates. The novice of this study is a unique dataset containing compensation figures of previously PE-sponsored portfolio companies. The main results provided that salary and bonuses increase after an initial public offering and the percentage of ownership and the ratio between bonus and fixed cash compensation decreases. The IPO itself has no effect on the different compensation elements.
A key challenge in researching private companies is gathering reliable data. Private
companies in Europe are not required to annually disclose compensation details, and each European country sets its own disclosure rules. Nonetheless, if a private company wants to issue more capital to accelerate growth on public markets or create an exit opportunity, a possibility for block holders of shares to sell, it can possibly organize an IPO. An IPO is the first offering of a company’s shares for public trading. To sell shares on an exchange, a firm must register and submit a formal legal document providing details about the shares offered for public sale. This document, called the prospectus, needs to be approved by the domestic financial authority. The prospectus contains historical data, most of which is financial in nature. However, most prospectuses also contain details about CEO ownership and
compensation. Although not all European countries have the same disclosure standards, the registration requirement provides a method to gain more information about managerial incentives at PE-sponsored portfolio companies. Combining the contents of two commercial databases and completing the missing data by looking in prospectuses and other company proxies resulted in a unique dataset covering European PE-sponsored portfolio companies in the period 2005-2015.
This dataset forms the first sample including previously PE-sponsored portfolio firms and now listed in Europe’s leading indices and economies: Germany, France, the UK, Denmark, Sweden, and the Netherlands. Other European countries were considered but were left out due to insufficient data or their complicated disclosure rules.
The researcher also created a second sample, consisting of listed companies, by subtracting data out of CapitalIQ and both samples permit a comparison of PE-backed firms and listed companies by matching them based on size. Both samples were then transformed into panel
data. Since the research question examined the effects of a specific event (IPO), a cross-sectional research design was the most appropriate, thanks to its ability to reveal patterns in dependent variables over time.
This study makes a number of contributions to the literature on compensation. First of all, no similar study has examined this topic for leading European countries, most likely because of the difficulties of obtaining sufficient reliable data. Secondly, this study’s results provide a framework for comparing managerial incentives at public and PE-backed companies. These results are controlled for time, industry, and country effects. Moreover, performance, size,
and efficiency control variables were also included for period from 2005-2015. Lastly, the results demonstrate the individual effects of salaries, bonuses, compensation, and ownership, which permits companies considering an IPO to predict whether it would cause CEO salaries to rise or fall.
This study is structured as follows: Chapter 2 reviews the literature, discuss current theories, and presents the hypotheses. Chapter 3 explains data collection process and the methodology, defines key variables, and provides more clarifications on the hypotheses. Chapter 4 presents the results and contains a robustness check. Finally, Chapter 5 provides conclusions.
2. Literature Review
In order to understand the dynamics and general roles in corporate setting this section commences with the current literature on the agency problem and corporate governance, the mechanism to mitigate the risk on a agency problems. In the next section the literature on Private Equity is put apart and the third section looks deeper into the corporate governance. In the fourth section the main theories on executive compensation are provided and in the fifth section are corporate governance, private equity and compensation combined and discussed. The last section concludes with the hypotheses based on the preceding section.
2.1. Agency theory
Scheifler and Vishny (1997) investigate in their study international corporate governance mechanisms and find that the quality of corporate governance structure is a combination of shareholder’s protection rights and concentrated ownership. In more advanced economies, like Japan, US and Germany, the small shareholder is more protected by the legal system and the concentrated ownership, through bank finance, large shareholdings or takeovers, acts as method of control fot he managers to create value instead of destroying value (Scheifler and Vishny, 1997). A clear example of how this works is given in the study of Jarrel and Poulsen (1988a) who find that a public announcement of an anti-takeover provision, such as a poison pill, reduces the share price and so affects the shareholders wealth. Managers resist takeovers to protect their own private benefits of control by installing provisions even though this means they do not serve the shareholders interest.
A great deal of research has investigated aligning the interests of the manager and the shareholder. In his classic article, Jensen (1989) discussed managerial incentives at public organizations. As public organizations have grown in size, shareholders are more dispersed, and the ownership of management is more diluted. Furthermore, the incentives for
management have become weaker. According to his findings (Jensen, 1989), public organizations lack the internal governance mechanisms needed to operate efficiently. The often-entrenched management is prone to diverting cash flows and does not dare to take strong measures that put their position or influence at risk. Examples of the failure of management to take measure are that directors retain their seats based on their tenure period instead of their performance or that managers do not wish to terminate an
unprofitable subsidiary because size is a factor in their compensation review. Such managers do not have the appropriate incentives and deterrents to encourage them to take on the most efficient level of risk. Boards, supposedly the guardians of shareholders’ interests, carry out their duties in an ineffective manner and tend to side with the management rather than with shareholders. Managers and executives might take steps to benefit themselves at the expense of shareholders. In other words, managers tend to destroy value for the organization rather than create it. Almost all public corporations struggle with the so-called principal-agent problem, where the agent, due to an information advantage over the principal, acts to serve his own interest. Many practices have been installed to mitigate its negative, value-destroying effects, and protect shareholders from self-interested managers. The negative effect of the principal-agent problem led Jensen even to predict that public corporations will eventually cease to exist (Jensen, 1989). Although this prediction is rather strong looking, in retrospect, at the corporate world nowadays. It would be interesting however to investigate the
suggested alternative by Jensen, the Leveraged Buyout (LBO).
2.2. Private Equity
One of the first to investigate changing managerial incentives in other type of organizations was Kaplan (1989). He focused on public organizations that had experienced a management buyout (MBO) and compared them to other public corporations. He specifically focused on changes in financial performance after the MBO. The results demonstrated that operating income and net cash flows increased, while capital expenditures declined. One possible explanation for this is a reduction in the number of employees after the MBO, and this theory is called the employee wealth hypothesis. The new management decides to lay off workers, because they are inefficient and too costly. The goal of such a measure is to increase operating income by reducing wages. A second possible explanation is the information advantage or underprizing hypothesis, which concerns managers and buyout investors that have information, which is not available to bidders, about potential changes or value increases within the company at the time of the sale. The better informed investors and managers purchase the company for less than informed bidders are willing to pay. It is only due to their information advantage that managers and buyout investors can obtain post-buyout positive returns. The third, most supported hypothesis is the reduced agency cost
theory, or new incentive theory meaning that the management’s equity holdings increase because of the MBO, and new incentives lead to better financial results. Kaplan (1989) found that promoting management ownership, other words aligning the interest of both parties, improved financial performance.
While Kaplan (1989) only researched formerly public entities, an interesting next step would be to investigate managerial incentives at private firms as compared to public firms. This is in line with Jensen’s (1989) prediction regarding alternatives to public companies. He
anticipated that the most dominant corporate form would be leveraged buyouts. In a LBO, a specialized investment firm, called a PE firm or general partner (GP), acquires another portfolio company, primarily via outside debt financing. The buyer then uses equity financing to cover the rest of the purchase price – a relatively small sum. In a typical leveraged buyout transaction, the buyer gains majority control of the portfolio company, transforms it into a more efficient and more effectively managed organization and sell it to the highest bidder. During a limited investment period, generally around 5-7 years, the PE firm imposes changes to incentives and operations, they tend to create value instead of enjoying perks and other benefits. Eventually, ath the end of the investment period, the PE firms exits trough an IPO or by selling the portfolio to a corporate buyer or another PE firm. This is different in venture capital, the other main type of PE investment, the GP provides capital to accelerate the growth of a diversified group of portfolio companies. In return, the general partner receives a non-controlling portion of the equity. In such cases, the firm’s founders or angel investors using retain control.
The market for LBOs has experienced tremendous growth since the 1980s, transforming from a minor investment tool for wealthy families to a mature investment type.
The PE method of improving a firm’s efficiency is based on three main pillars (Kaplan, 1989). First, the high degree of leverage financing disciplines managers and prevents them from wasting resources. In other words, the mandatory debt payment reduces the cash flows that managers can use for activities and automatically mitigates agency costs. Moreover, it encourages managers to generate cash through alternative channels. The second pillar is the enhancement of corporate governance mechanisms. In short, it refers to corporate
governance engineering. Finance and industry professionals at the PE firm actively monitor the portfolio company. For example, the general partner (rather than the CEO) appoints board members, thus reducing agency conflicts and limiting the CEO’s managerial power. This arrangement helps to align managerial incentives and shareholders’ goals (Jackson, 2012). The third pillar is the promotion of management equity ownership. Executives as owners of the firm by investing their private wealth in is the mantra of PE philosophy. As a
result, they feel a greater sense of responsibility for ensuring a positive performance. In other words, they tend to create value instead of enjoying perks at the cost of the organization, thus aligning the interests of executives and shareholders. In addition, their compensation
structure is highly sensitive to their performance, thus further bring into line these interests (Kaplan, 1989).
Focusing on PE is an eligible choice for comparing management incentives in public and private settings. However, one could argue that comparing managerial incentives in Small or Medium Enterprises firms and publicly listed firms is like comparing apples to apples, due to their size. Leveraged buyouts target mature companies that operate globally in the same setting as publicly listed firms. Concentrating on PE-backed portfolio companies, LBOs exclusively, would provide a rare opportunity to assess the differences between public and private companies. A key challenge in research for private companies is gathering sufficient and reliable data (Harris, Jenkinson et al, 2014).
2.3 Corporate Governance
In modern corporate settings, in public and also in private organisations, there is separation between professional managers and the suppliers of finance. The latter group invest in a company with the goal of receiving a return for their investments. The former, an
entrepreneur or a manager, raises funds to finance operations or to cash out his interest in the firm. The financier need the manager's unique capabilities to generate a return on their investment and the manager need the financiers for his investment. Hence, both parties are reciprocally dependent of each other, but have different interests causing problems, so-called agency problems. The agency problem refers to the difficulties investors have in assuring that their invested money is not wasted on unattractive projects, expropriation or perquisites for the managers. An ideal solution would be to make both parties sign a complete contract, what describes what each party does in every state. However, it is impossible and too costly to create such a contract. Hence, manager and financier have to make agreements what to do at unexpected events, in other words they have to allocate residual control rights. To decide what to do every time an unexpected event occurs is not very practical, this was the main reason that the investors hires the managers in the first place. In practice this means that the manager ends up with substantial control right and the power to allocate funds as he wishes. To mitigate the risk that managers expropriate the investments of the suppliers of money corporate governance deal with this issues. Corporate governance ensures that the invested money is returned back to the original investors and nowadays in the most advanced economies the agency problems is solved rather well although not perfectly. Corporate governance limits this discretion in the form of a contract but much discretion remains at the
managers. In an ideal world would competition, in the long run, force companies to install efficient corporate governance mechanism and so minimize costs, enabling them to raise, at the lowest cost, external capital. However due to the sunk production capital this does not work in practice.
Competition is a strong force, but alone competition would not solve the problem of returning the investment to the supplier of finance. Competition will reduce the returns on capital and cut the amount the managers can expropriate, so it will cut down but not alone solve. This leaves room for improvement and there is a great deal of disagreement about the functioning of the corporate governance mechanisms. Scheifler and Vishny (1997) indicate that the combination of the presence of a large shareholder and advanced legal protection for (smaller) shareholders provide the better corporate governance structures. Legal protection for investors is important because this creates external financing for companies, what enables them to fund operations which makes them grow further or faster. If the legal protection for the investor is absent or poorly developed an investor would choose for
different investments. The presence of a large shareholders is important because this investor will have the incentive to bear the cost to monitor the management and will have the control to put pressure on the management. In a PE portfolio company, the large investor is
represented by the PE firm which has the outright control. Understanding corporate governance structures can enlighten, marginal improvements in advanced economies, but can encourage changes in places where they need to be made.
2.4 Compensation
A potential solution to further reduce the agency problem is to grant the manager with a long term incentive contract that aligns his interest with the shareholders'. The feasibility on such a contract depends on the structure, to succeed it must exceed the marginal value of the personal benefits of control of the managers. In practice this means that the contracts are substantial. To be credible it must contain a measure of performance which is verifiable in court. The optimal contract is a function of the managers’ risk aversion, the managers’ ability to control the cash flow ownership and the importance of his decisions.
Optimal incentive contracts, although expensive if the personal benefits for the managers are high, will induce the manager the act in the interest of the shareholder. A variety of elements can appear in a contract, such as fixed and variable pay, share ownership and stock options the interest.
The claim that CEOs are overpaid has been the topic of much research. Murphy (2012) explains 2 main theories, the first explanation is the managerial power theory where excessive pay is symptomatic of the agency problem and the CEO owns often a too trivial
fraction of the company’s stock to make them interested in maximizing profits. The managerial power theory, which emphasizes CEOs’ influence on the process of setting his own compensation structure. One of the results of Murphy's research indeed provides an indication of the managerial power of managers. In 2009, a crisis year there was a clear increase in the use of restricted stock unit (RSU) over stock options in the CEO compensation. The share price was low at the moment and the characteristics of the RSU, the vesting period and the valuation of fair market value, make it fair more attractive for a CEO to receive than stock options (Bebchuck and Grinstein, 2005).
The second explanation, the optimal contracting theory, views the CEO compensation contract as a method to mitigate the agency problem by aligning the interest of shareholder and manager.
Next to the main 2 theories in executive compensation Murphy (2012) found that examining the levels of pay in isolation is misleading, without examining the whole structure of
compensation package. This means that next to the cash component also the equity-based compensation components are important to investigate. Another remarkable result of his study is that the levels of CEO pay and the use of equity-based instruments are positively related to variables used to indicate the quality of corporate governance, namely board independence and institutional ownership.
Gao, Lemmon, and Li (2012) specified the share of the total compensation package composed of salary and bonuses. Further specifications regarding cash pay were not provided, however. They also found that size, cash holdings, leverage, and firm age were all positively related to CEO pay. On the other hand, accounting performance and sales growth had a negative relationship with CEO pay. The researchers controlled for size, cash holdings, CEO-specific, year, and industry effects (Gao et al. 2012).
2.5 Compensation in Private Equity and Public organisation
Three previous studies have investigated how public and private equity firms differ from each other in terms of compensation structure and policy. First, Jackson (2012) explored
differences in CEO pay at PE-owned and public firms, focusing on companies operating in the US from 2000-2004. By measuring the dollar for dollar incentive, a metric that shows the effect of an increase of the market value on the value of stock or options owned by the CEO. The results show a significant higher value for the PE owned firms compared to the public firm, in other words the increase of market value is more aligned with the compensation structure. The results demonstrate that the reward structure at PE firms has more of a link to performance than the reward structure at public firms. Jackson (2012) suggested two
market means that CEO-board negotiations yield compensation packages in the best interests of shareholders. The alternative hypothesis is the managerial power theory, which emphasizes CEOs’ influence on the process of re-electing the firm’s directors. If the CEO has power over this process, directors face strong incentives to act as the CEO desires.
Consequentially, they might not put shareholders’ interests first. In addition, directors own a small fraction of the company’s stock and do not personally bear the costs of value-destroying activities (e.g., excessive rewards for the CEO). Since directors own a small or negligible portion of the shares, they have insufficient incentives to discipline the managers who prioritize their own interests. The costs of intervening in CEO pay disputes outweighs the benefits. Hence, directors have stronger incentives to grant CEOs’ wishes than to uphold shareholders’ interests. However, CEOs do not have such managerial power at PE-owned firms, because the PE organization is usually contractually entitled to appoint the board. Therefore, the PE firm or the GP decides which directors will retain their seats and who will serve on it. Furthermore, at most PE-backed portfolio companies, the board members are also owners (Kaplan, 1989). The absence of managerial power means that board members are unequivocally motivated to promote shareholders’ interests. Another results of Jackson (2012) is his findings that the dollar-on-dollar incentives is significantly higher for private equity sponsored firms, meaning that the cost to consume perks are higher in a private equity sponsored environment. This is another indication that in public companies it is more likely that there is managerial power.
The more prominent study on this topic, however, is that of Leslie and Oyer (2008). They investigated differences in managerial incentives and pay structures in the U.S. market from 1996-2006. Just like Jackson (2012), Leslie and Oyer (2008) examined PE-backed firms that had held an IPO, and they compared these companies to public firms that had experienced a LBO, as well as to a general group of public firms. Top managers at PE-owned firms tended to hold almost twice as much equity before the IPO, earned an approximately 12% lower salary, and received a greater share of their total compensation through a variable
component than did their public counterparts (+13%). These results are controlled for size, year, and industry effects. Larger organizations (measured in total assets, number of
employees, or sales) paid higher salaries and more bonuses, and the managers of such firms held a relatively smaller portion of shares. These differences in compensation structures dissolve over a two-year period following an IPO. The percentage of management-owned shares shrank after an IPO, because of the provided possibility to managers to sell shares at the IPO. Another reason for the shrinkage is dilution of old and newly issued shares. Fixed salaries tended to increase after IPOs and variable compensation tend to be lower in public organizations. Given that executives are risk-averse and desire direct compensation over
deferred, and fixed over variable. For accepting the risk of giving up components of their salary (ownership and certain bonuses), they prefer to be compensated in fixed salaries. Leslie and Oyer’s (2008) study only included control variables for size but not for
performance and efficiency. In addition, due to their focus on the US, they did not have to control for country effects (Leslie & Oyer, 2008).
Gao, Lemmon, and Li (2012) investigated the claim of excessive CEO pay for public and private firms in the U.S. market from 1999-2008. The private firms in their sample were comparable to public firms in size and industry coverage, but they were not necessarily PE-backed portfolio companies. Their conclusions were similar to those of Leslie and Oyer (2008), and they found that CEOs at public firms had pay premiums of about 20% as compared to CEOs at private firms.
Another, related assertion is that the link between performance and pay has broken down in public firms. In contrast with this assertion and the findings of Leslie and Oyer (2008), in the research of Gao, Lemmon, and Li (2012) the pay-to-performance link was much stronger for public firms, and equity-based incentives were likewise stronger. Managers of public firms had more on-going equity incentives, while compensation packages at both types of firms were positively related to accounting performance. Dissimilar contract environments led to different CEO compensation packages. The concentration of ownership in listed firm is dominated by institutional investors where ownership in private environments is dominated by large blocks of insiders, for example family shareholders. Large, institutional owners press for a tighter link between performance and pay as a monitoring mechanism. They demand objective, performance measures to be able to control the management. On the other hand, the ‘insiders’ do not rely that much on expensive incentive pay since the CEO ownership in private firms is higher and motivation and rewarded is primarily subtracted from this percentage of ownership. The boards of private firms relied on more subjective performance evaluations to determine CEO pay than listed firms did. Hence, efficient, objective
governance and a stronger pay-to-performance link might explain the pay premium found for listed firms.
While these (Jackson, 2012; Leslie & Oyer, 2008; Gao et al., 2012), the three most prominent studies, all focused on the difference between public and private equity sponsored portfolio firms in the U.S. market, this study explored the difference between public and private firms operating in European markets. Other, previous studies, more generally focused on
compensation, have examined differences between U.S. and European firms only for listed firms. For example, Kaplan and Rauh (2010) demonstrated that CEO compensation in the US had a stronger relationship to performance. Conyon, Core, and Guay (2011) compared
U.S. and UK CEOs and found that pay was higher – and equity holdings substantially higher – in the US. For an extensive sample covering 14 countries, Fernandes et al. (2011)
documented international differences in compensation packages. They challenged the claim that U.S. pay packages are excessive and indicated that differences in pay structures seemed to be economically modest. In particular, the pay packages of U.S. CEOs were 26% higher than those of their European counterparts. More effective corporate governance policies, like independent board structures and higher levels of institutional ownership, were important determinants of this difference. Both factors encouraged a tighter link between pay and performance, resulting in higher pay and the more frequent use of equity-based pay.
2.6. Hypotheses
Compensation packages are an often-discussed and controversial topic. In an ideal world, compensation packages, or contracts, would attract talented CEOs incentivized to exert effort, avoid wasteful projects, and exploit growth opportunities. The contracts affect whether CEO seek to optimally maximize value on behalf of shareholders (Edmans & Gabaix, 2009). Bebchuck and Fried (2005) rejected this optimal contracting view for companies with a separation of control and ownership. This rejection indicates the presence of a classic
principal-agent problem in the form of the managerial power hypothesis. In short, CEOs can exert power over boards, thus influencing their own compensation structures (Bebchuck and Fried, 2005).
After an IPO, the CEO has even more influence on the board due to step back of the PE firm as large shareholder (Jackson 2012), and so he or she has more control to influence or dictate the pay-setting process. Consequently, risk-averse CEOs desire higher fixed pay. In
accordance with Leslie and Oyer (2008), this study expect that the fixed salary component of compensation would increase slightly following an IPO.
Hypothesis 1
The CEOs of Private Equity sponsored portfolio companies will have lower fixed salaries after than before the IPO.
Hypothesis 2
The effect of the IPO induces a larger change in the fixed salary component of the CEO of the PE sponsored company compared to a listed company.
Efficient executive pay contracts link compensation to firm performance. However, many empirical findings suggest that optimal contracting is not always the case. Bebchuk and Fried (2004) claimed that CEOs received compensation without needing to deliver a positive
performance. For example, during the financial crisis, CEO compensation packages in the US rose relatively faster than average wages. Hence, U.S. CEOs have been rewarded for their performances during an economic downturn. Other examples of a weak or non-existent link between CEO pay and performance include severance packages and rewarding CEOs for luck (Shaw & Zhang, 2010). In general, CEOs are not penalized for poor performances and
sometimes are even rewarded for them.
Fernandes et al. (2011) found that performance pay constituted a relatively larger share of compensation packages at U.S.-based companies, due to higher institutional ownership, more independent boards, and fewer insiders serving as shareholders. This link between pay and performance has been the topic of much discussion, although with contradicting results (Leslie & Oyer, 2008; Gao et al., 2012). Gao et al. (2012) investigated private companies in general (i.e. not specifically PE-back firms) and found a stronger link between pay and performance for public companies. However, PE has proven to provide a strong link between pay and performance and to limit managerial power in portfolio companies.
Hypothesis 3
The variable part, the bonuses, of the CEO compensation is larger before than after an IPO due to a less strong link between pay and performance in PE sponsored companies
Hypothesis 4
The effect of an IPO induces a smaller change in the variable part of the salary component of the CEO of the PE sponsored company compared to a listed company.
Fernandes et al. (2011) show that Anglo-Saxon nations have similar compensation structures, where the variable part accounts for a significant, almost 50%, part of the total compensation. In Nordic (Sweden and Norway) and continental European countries (France and the
Netherlands) the fixed salary part of the compensation package accounts for the majority share of the complete package. Germany and Italy seem to follow the Anglo-Saxon model. Looking back to the contrasting expectation of salary (hypothesis 1 & 2) and bonus
(hypothesis 3 & 4) the effect of the ratio fixed and variable compensation is unclear. Based on the results of Haubrich (1994), who found that CEO prefer risk-averse elements other riskier elements.
Hypothesis 5
Following the prior hypotheses the fixed salary will increase after the CEO and the variable pay will decrease. Hence, the ratio fixed-variable CEO compensation will decrease.
Hypothesis 6
The effect of the IPO will induce a larger chance of the fixed-variable compensation ratio for the PE sponsored firms
One of the main pillars of PE is managerial ownership in the firm. This brings the interest of shareholders and executives into alignment. PE firms oblige managers to invest their own wealth in company shares, which provides them a portion in the management incentive plan. A strong and controlling shareholder will provide better corporate governance mechanisms (Scheifler and Vishny (1997), i.e. large shareholders would stimulate a CEO to participate in the form of share based compensation and to align interest. According to Kim and Lu (2011), managerial ownership is an internal corporate governance mechanism. They examined the relationship between CEO ownership and risk-taking behaviour. Strong external pressure encouraging good corporate governance was an indicator of optimal CEO behaviour, and resulted in fewer agency problems. Such pressure holds CEOs accountable for their performance (Kim & Lu, 2011).
Overall, these two studies suggest that ownership leads to stronger incentives for CEOs. In addition, an exit, in the form of an IPO, provides an opportunity for managers to sell and decreasing their level of ownership. Another possibility is the dilution of their ownership because of newly issued shares decreasing their level of ownership (Jackson, 2012). On the other hand, equity-based pay comprises a relatively larger share of the CEO compensation package in the US than in Europe (Fernandes et al, 2011). In this study the effect of a decrease in ownership is expected to be smaller than that found by Leslie and Oyer (2008).
Hypothesis 7
The percentage management ownership in PE sponsored companies will strongly decline after an IPO.
3. Methodology
This chapter presents this study’s methodology for analysing European CEO compensation packages, including the econometric model utilized. Moreover, it provides, how data sources were identified, specifies the inclusion requirements, and describes the research methods. Further it defines the variables, and conclude with clarifying the hypotheses.
3.1. Initial public offering
This study investigates differences in managerial incentives between public and private equity sponsored firms in leading European economies. In general, obtaining data on private
companies is difficult, since they face looser regulations than listed firms. The quality and amount of disclosure of private, non-listed firms is limited since they are not required to do so. In contrast, listed firms are required to disclose financial and compensation information to the domestic financial authority, although the quality of pay disclosures significantly varies across Europe. This variety reflects, to a great extent, the national regulatory approaches (Plath, 2007). Certain countries have established legal disclosure rules, whereas others use a best-practice approach. Listed companies can choose whether they wish to comply with best practices, and in most cases, they do not have to explain their choices (Plath, 2007). For example, in the Netherlands and Germany, regulations require that firms disclose
information on management remuneration schemes, while the French rules are less strict regarding the sharing of details about compensation and ownership (Lemstra & Schütte, 2011; Zwissler, 2011; Buhart, 2011). Another difference regards whether firms must disclose their overall compensation philosophy or actual awards paid (Plath, 2007).
Pre-IPO regulations are, in general, similar in all leading European economies. A company that desires to issue public shares must register and submit a prospectus to the domestic financial authority. This prospectus should contain financial data for the three years preceding the IPO. However, prospectuses tend to use the same structure for providing compensation information as that used by already listed companies. Such prospectuses provided an opportunity, although far from perfect, to obtain data on actual paid executive compensation at PE-backed portfolio companies and CEO compensation and ownership schemes.
3.2. Econometric model
This compensation data constituted this study’s dependent variables, while financial variables functioned as controls. This study assessed the transformation from private to public via an IPO within individual companies. In contrast, Leslie and Oyer examined the difference between private equity sponsored companies that went public, public firms that became private through an LBO, and a general sample of publically listed companies. Although, the methodology was partly derived from Leslie and Oyer (2008) but modified to fit European standards. The main reasons for the adjustments were unavailable
compensation data for listed companies that became private again through a LBO and consequentially the application of matching of companies equally in size difference. This gives the following econometric model used:
𝑌"# = α + 𝛽"#𝐵𝑒𝑓𝑜𝑟𝑒𝐼𝑃𝑂"#+ 𝜃"#𝑋"#+ 𝑢"#
Here, 𝑌"# is measured managerial incentives (i.e., the dependent variable). Such incentives could include salaries, bonuses, the ratio of fixed and variable compensation, or the
percentage of CEO ownership in year t at firm i. The percentage of CEO ownership act also as a control variable. The independent variable, 𝐵𝑒𝑓𝑜𝑟𝑒𝐼𝑃𝑂"# variable is a dummy variable that equalled 1 for all periods before the IPO. This made it possible to investigate an IPO’s effects on managerial incentives. The variable 𝑋"# was a vector of financial control variables,
including Total Assets, Total Revenue, Total Debt, Cash, Number of Employees, Earnings before interest depreciation and amortization (Ebitda), and the Return on Assets (ROA).
3.3. Data collection
Since this study focused on private firms sponsored by a PE firm and public firms, there were two different samples created. As described, IPOs provide, by means of a prospectus, an opportunity to collect information about previously private companies. Hence, the primary sample consisted of PE-backed portfolio companies from 2005-2015 in Europe. The second sample is a general sample of listed firms in Europe. Table 1 (Appendix A) provides
frequencies and percentages for the various industries and the number of companies in both samples. Almost 18% (11 firms) were industrial firms in the first, PE-sample, and 8 firms (13%) were in the consumer goods sector. Again 8 firms provided consumer services, but the majority is operating in the materials industry. None PE-sponsored portfolio company in this sample operated in the energy industry. Consumer discretionary, Industrials and
Information Technology dominate the general sample, all accounting for around 20% of the sample.
For the creation of the first sample Dealogic, a financial platform offering data on
transactions, was used to identify companies that underwent an IPO during the 2005-2015 period in Europe and that received financial sponsorship from a PE firm prior to the IPO. Moreover, venture capital sponsors and LBOs that did not control more than 50% of the shares before the IPO were excluded. A total of 303 firms were at first included in the sample. For these firms, the name of the issuer, the IPO date, general industry, the International Securities Identification Number (ISIN), nationality of the issuer, and the nationality of the exchange were gathered.
The unique ISIN codes were matched in CapitalIQ by creating a watch list. Those ISIN codes unknown to CapitalIQ were added manually. In CapitalIQ, it was possible to extract the compensation data classified in salary, bonuses, total annual cash, total compensation received, and total compensation computed for all fiscal years dating back to 2006. The data is structured on executive, board and director level. CapitalIQ provided also the position,
name, and tenure period of the CEO, but to ensure accuracy, the CEO’s name and tenure period were cross-referenced in the prospectuses, other company reports, or other external sources, like Bloomberg. The same data mining approach was followed for dates prior to 2006 or missing data. The absence of strict disclosure regulations for private companies planning an IPO limited the number of companies eligible for inclusion in the sample, since data was missing on some of these companies. If data was missing for a company, it was excluded from the sample. Information on CEO ownership was collected manually using prospectuses and annual reports. For all companies, information on total assets, total cash and equivalents, total debt, ebitda, the number of employees and ROA was collected for all years going back to 2006. Data from prior years were added manually.
Cao and Lerner (2009) established that founders who serve as CEO receive different compensation and are motivated differently from 'professional' CEOs. Founders were, for these reason, excluded from the sample. The final criterion referred to the exchange
nationality of the firms. Only the leading European countries, France, the UK, Germany, Italy, the Netherlands, Sweden, and Denmark, were included. Due to insufficient data, Spain, Switzerland, Norway, and Belgium were left out of the sample. Table 3 (see Appendix A) provides the frequency and percentage of firms from different countries in both samples. Smaller countries tended to be underrepresented in the PE sample. Denmark and the Netherlands had 4 and 3 companies listed on their national exchanges, respectively, that have been sponsored by PE. However, Italy and Germany were also underrepresented in both samples. Germany is the largest economy in Europe but represented only 6.5% and 13.6% of the two samples, respectively. Furthermore, the UK was overrepresented in both samples. In the PE sample, the 26 British firms constituted 41.9% of the total, while the British share of the other sample was even higher at 45.3%.
This research study followed a panel data approach. Traditionally are panel data or cross sectional studies often used in sociology, to study the effect of a specific change on a population and in psychology, to study treatment effects across different groups of the research. In a panel data, several observations are conducted over time for the same subject, or in this case entities. The benefit of this method is that it allows researchers to detect developments or changes in the target sample at the group and individual level. Changes observed in the subjects are more accurately estimated due to the possibility to control for cultural or other external factors. In this research design, the different stages were the pre-IPO periods, the pre-IPO itself and the post-pre-IPO periods. The periods were classified in fiscal years. To study different stages, it is necessary to have data on all of the different stages, and a minimum, a time series of two consecutive observations is needed for all variables,
including control variables. While the IPO represented the change from one organizational type to another, the effect of this change is not immediately observable. That meant that minimum of three consecutive observations were needed, the pre-IPO years, the IPO years and the post-IPO years. This criterion further reduced the number of eligible firms to 62 firms in the PE sponsored firms sample.
After excluding firms with founder CEOs and missing data, 62 companies were included in the sample. Of these companies, only 45 had available CEO ownership data, although compensation data was available for all 62 companies. Table 2 (Appendix A) provides more information on all years in the sample and the number of firms that experienced an IPO. The years 2013 and 2014 were quite fertile years for organizing an IPO for PE-sponsored firms. With a combined share of more than 66%, these years were definitely overrepresented in the PE sample. The years before 2013 clearly saw less IPO activity, due to the effects of the global crisis in 2007-2009, with no IPOs in the sample for 2009 and 2011. Due to a lack of data on IPO events, gathering data for 2005 and 2006 was especially hard. For 2015, is was almost impossible to gather reliable consecutive time series of observations.
To be able to compare compensation packages at PE-backed portfolio companies to listed companies, a second sample was generated in CapitalIQ. The researcher selected public companies listed on European exchanges and, furthermore, all compensation and financial data was derived following the same procedures used for the PE sample. In addition, the last known percentage of CEO ownership was gathered. Again, this sample needed to meet certain requirements. The first requirement was a complete time series of at least three observations for salary, bonuses, and compensation in the period 2005-2015. Companies without such consecutive time series were excluded. Second, because of CEO turnover, this set contained duplicate CEO compensation data for some companies. Only current CEOs were included in the sample. This resulted in 838 unique companies with complete
compensation and financial data. Out of these 838, only 421 had ownership data available.
Both samples were combined and transferred to SPSS. In SPSS, a dummy variable was created to indicate the sample type. Moreover, both samples were transformed into a format enabling panel data research. For the PE sample, all compensation and financial data was connected to the corresponding IPO event. The fiscal years before the IPO were defined as t-1, t-2, while the fiscal years after the IPO were t+1 and t+2, and applicable for only a minor group so on. The year in which the IPO took place constituted the point of change (t=0), or the threshold. For the general sample, no threshold existed, so the most recent time series of at least three observations was utilized. This procedure was followed for all financial and
compensation data, except for the percentage of CEO ownership. Only the most recent ownership data was available in CapitalIQ, and so that information was indicative of CEO holdings at listed firms, but no time adequate time series data was available.
3.4. Background of hypotheses
One form to (partially) give away the control over a PE sponsored is through an IPO, leaving behind a less concentrated shareholder ownership. Other forms are selling the control to a corporate buyer or another PE firm, the so-called secondary buyout. In the case of an IPO it is more likely that shareholders become more dispersed and the smaller shareholders tend to create a free-rider problem (Scheifler and Vishny, 1997). A small shareholder does not want to bear the costs needed to control the management as the PE firm did before the IPO. The management will, eventually, have more room to operate more according to their wishes before the shareholders will stand against the management. The more managerial influenced internal governance comes at the expense of shareholders’ interests (Scheifler and Vishny, 1997). For the first hypothesis predicted a negative coefficient for the fixed salary component. The second hypothesis expects a higher growth in salary for the previously PE sponsored portfolio companies compared to to the control sample, indicating a size effect of the IPO.
Following this same line of thinking regarding the stronger link between pay and
performance in PE-backed portfolio companies, it was anticipated that such firms would offer higher variable cash pay. However, this link was expected to weaken after an IPO, resulting in a positive coefficient of the dummy variable BeforeIPO. Public companies will pay out lower bonuses in the post-IPO period. The third hypothesis measures the change in performance related pay, the bonuses, before and after the IPO. The expectation is a decrease of the bonus after the IPO. The fourth hypothesis expects to find a size for the effect of the IPO on the bonuses. The expected direction is in line with the third hypothesis and will decrease.
An alternative explanation might be that a CEO demands to be compensated for giving up his equity-based incentives at IPO date or post-IPO. Equity-based incentives and the variable part of cash compensation are both a function of company performance. It is likely that the CEO wishes to be compensated comparable as ex ante IPO. As a result, the coefficient of the BeforeIPO will be negative, indicating higher bonuses in the post-IPO period.
For the hypotheses on the ratio of bonus and fixed cash compensation element. The fifth hypothesis is expected to find a decreasing ratio bonus because the fixed salary component is expected to go up and the bonuses to go down. This will result in a lower variable / fixed ratio. The sixth hypothesis measures the effect size of the IPO and is also expected to go down.
The last hypothesis assumes a higher managerial ownership share before the IPO, because one foundation of PE-activity is the mantra, “managers as owners.” Hence, the direction of the BeforeIPO dummy was predicted to be positive, indicating a lower percentage of CEO ownership after an IPO. The CEOs of public companies do not need to partake in
management participation plans, and so they are less motivated to have a high degree of ownership in their organizations. It was expected that this effect would be accelerated by the possibility that shares are diluted at the IPO (Jackson 2012).
3.5. Definitions
The definitions of the variables were adopted from CapitalIQ or were derived from their usages in prospectuses. The dependent variable Salary was derived from CapitalIQ and represented the euro value of a CEO’s base cash salary in a specific fiscal year. The variable Bonus was also adopted from CapitalIQ and indicated the euro value of the CEO’s bonus in that fiscal year. It only included the cash bonus and not pensions, director fees, or other forms of variable compensation. In addition to the 2 main components of CEO compensation the underlying ratio between these 2 is also a specific variable, defined as Bonus / Salary. The last independent variable is CEO ownership defined as the percentage of ownership held in ordinary shares at the end of that fiscal year. The number of ordinary shares held by the CEO divided by total shares issued provided the percentage of ownership.
The variable CEO Ownership took into account the transformation of preference shares to ordinary shares, which usually occurs just before an IPO. It measures the percentage of ownership of total issued capital. Due to unavailable data of prices of shares before the IPO it is not possible to measure the value of the portfolio hold by the CEO. Furthermore, it did not include other types of shares, like performance shares or options, due to insufficient data from CapitalIQ, prospectuses, and other company reports. The result of leaving out the performance shares will be that the effect of lowering of ownership will be underestimated and effectively are stronger.
To improve the predictive power of the regression models, several control variables were added. All of the control variables were created using CapitalIQ data, and they belonged to the following categories: size, efficiency, and performance control variables. All financial control variables are total assets, total revenue, total debt, cash holdings, number of
employees, earnings before interest depreciation and amortization (ebitda), and the return on assets (ROA). To be able to control for size effects the ratio of debt over total assets is computed which forms the basic variable for size. Other applied variables for size are
analogue variabele such as total assets, total revenue, and the number of employees were considered. The variable Total assets was defined as the value of total assets, as provided on the firm’s balance sheet for that fiscal year. To calculate total revenue, the same method was used, and income statements were examined. The number of employees referred to the average number of employees working for the organization during that fiscal year.
Efficiency was measured by the ROA, the ratio of net income to total assets. The ROA gives an idea of managers’ efficiency in using assets to generate income. To control for the effect of performance on compensation, the variables Cash and Ebitda were created. Cash was a function of cash flows and was defined as the amount of cash on the balance sheet at the end of the fiscal year. The earnings before interest depreciation and amortization, Ebitda, was derived from income statements and defined as earnings before interest in that fiscal year. Another control variable included Total Debt, as derived from the balance sheet, at the end of the fiscal year.
The previous literature has demonstrated that size, efficiency, and performance have a relationship with compensation (Gao et al. 2012), (Murphy, 2012). The size of a company positively influences its CEO compensation packages; the larger the organization, the higher the total compensation (Jackson, 2012). Moreover, as Gao at al. (2012) demonstrated,
leverage is positively related to CEO pay, because it makes equity-based compensation riskier. On the other hand, leverage is negatively related to CEO ownership (Gao et al. 2012). Past performance is also a positive indicator of CEO pay and ownership, but cash is negatively correlated with ownership (Gao et al. 2012). The effect of efficiency on both compensation and ownership is uncertain. Table 4 (appendix B) provides the descriptive statistics of all dependent and independent variables. Estimating the dependent variable ownership delivers an insight that ownership decreases over the years, from 0.0422% to 0.0172%, in the PE-sample. Another result is that the size of public companies in the general sample is much larger compared to the private companies. But the size of private firms, measured in total revenue, appears to increase after the IPO. Moreover, the fixed part of compensation, salary, is lower in the PE-sample and increases over the years after.
In addition to controlling for size, performance and efficiency, the study also controlled for time, industry, and country effects. This meant that specific events in time or regulatory changes in a particular country could not affect the results.
4. Results
This section presents the empirical results. The first part discusses the results of goodness of fit tests and describes potential adjustments made to the variables. The second part discusses the outcome regressions of all dependent variables individually. The final part presents the robustness of the models utilized.
4.1. Testing
The possibility to interpret statistical data has basically two requirements. The first is the assumption of linearity between the independent and dependent variables. The value of the latter depends on the value of the former. In other words, a model is linear in the coefficients of the predictor with an additive random error term. The second assumption is a normally distributed sample. Looking at the descriptive statistics of table 4 (appendix B) it is striking that the standard deviations of the dependent cash variables, Salary and Bonus are very large. For the ratios Bonus/ Salary and Debt/Assets such differences are not observed. Further investigation in the variable Salary gives that the minimum cash salary paid out in the sample is €29,986 and the maximum is €2,246,719. There is clearly a huge difference between these two terms. The same huge differences appear between the lowest and highest value in the sample for the independent variables Bonus and for the control variables Total assets, Total revenue, Total debt, Ebitda and Cash. Large, absolute differences inside variables complicate the interpretation of results. A log transformation can be used to make highly skewed distributions less skewed. This can be valuable both for making patterns in the data more interpretable and for helping to meet the assumptions of inferential statistics. The assumptions of a normally distributed sample and linear regression are satisfied after
transforming the mentioned variables into logarithmic variables. Table 5 (appendix B) provides all correlations of dependent and independent variables, including the newly created variables.
Panel data studies measure different observations of the same object over time. In this study the different observations are consecutive, fiscal years and the objects are companies.
Autocorrelation or serial correlation can occur between the fiscal years. Serial correlation is the correlation of a term, here a specific observation, with the same observation at a different stage in their life cycle. In other words, it is the similarity between observations over time of the same term. Serial correlation affects the efficiency of the estimator and in case of positive serial correlation the standard errors will be smaller estimated than the actual standard errors in fact are. This will lead to estimators that predict more precise than they effectively can, which eventually will lead to biased predicted results in the form of missed turning points. The errors of observations in panel data will usually be correlated and could also have
unequal variances. This study chooses for the Toeplitz method to control for serial correlation between the terms over time. The advantage of the Toeplitz method is that it assumes
heterogeneous correlation between adjacent elements. It assumes that the pattern sequence for covariance makes steps by a unique multiple associated with that time steps instead of makings steps with a some common multiple. The result is a more efficient estimator for this sample.
4.2. Salary
The sample of Salary consists of 62 firms and 248 observations. Only firms listed in the United Kingdom, France, Netherlands, Denmark, Sweden, Germany and Italy are selected. Another additional selection is the exclusion of observations other than time period (t-1) to (t+2). To be able to measure the effect on Salary for the PE Sample, in other to check the first hypothesis a filter is installed for the PE-sponsored companies in the total sample. Table 6 (appendix C) presents the regression outcomes of all regression performed for Salary.
The first regression is the base model which reflect only the relationship between the dummy variable BeforeIPO and Salary. The intercept of 12,895 is the log of geometric mean of salary for 0-values of BeforeIPO, thus for the years after the IPO. If BeforeIPO = 0 it has a
coefficient of
0,092 measured in logarithmic scales and is significant on a 1% level. The real effect of the difference from ex ante to post IPO is the exponential of 0,092, hence a positive of 9,63%. A valid claim is that the salary ex ante IPO is 9,63% higher in more than 99 out of 100 cases. This is limited to this sample and only for previously PE-sponsored companies in the selected European countries. The second regression controls for relevant fixed effects of the country of residence, the industry and the year in which the IPO took place. For these effects The United Kingdom (UK), Real Estate and 2007 act as base references respectively. These references will be used consistently in this and the following regressions. The results show that France and Denmark, respectively negative and positive, are significantly different than the UK on a 1% level. In France, for example, the coefficient is -0,74 indicating that the salaries are lower in France. Fixed effects often capture a lot of variation in the data, and this often leads the standard errors to be larger. For the second regression this is in fact the case. Controlling for all fixed effect slightly lowers the coefficient (0.084) of BeforeIPO on the same significance level below 1%. The Salary of PE portfolio companies tend to increase when controlled for all fixed effects.
Regression 3 includes the first control variable and the relevant fixed effects, at least the ones that contributed to the significance of the model. The Industry effects are left out of these
