National Culture and CEO Compensation

(1)

National Culture and CEO Compensation

An investigation of the relation between different national cultures and

CEO compensations components

Esmee van Diepenbrugge S2536102 Accountancy & Controlling 0652690585 Moezeldreef 251, Utrecht evd1991@hotmail.com Supervisor: dr. C.P.A. Heijes Co-assessor: dr. P.G. de Heer

Subject: CEO, chief executive officer, executive, pay, compensation, corporate governance, wages, culture, Hofstede, GLOBE, remuneration, incentives, salaries, national culture, cross-nation,

(2)

Abstract

This paper examines whether variation in national culture, at country level, influences the different compensation components of CEOs. Much research has been done in factors that influence (CEO) compensation, however, little research has been done on the impact of national culture. This study looks at the cultural dimensions of Hofstede, namely power distance (PD), uncertainty avoidance (UA), individualism (INV), masculinity (MAS), and long-term orientation (LTO). There are five different remuneration components identified, namely base salary, short-term incentives, long-term incentives, pensions, and total salary. These components could be influenced by culture. Three dimensions of GLOBE are used for a robustness check, to see whether the two different models find the same results. This research is conducted by looking at the remuneration earned by CEOs in 2013. Sixteen countries are added in the sample and 143 different firms. With the dimensions of Hofstede there is a negative relationship between PD and base salary, pensions, and total salary. There is no relation found with UA and a compensation component. INV is positively related to base salary and pensions, and negatively to short-term incentives. Masculinity has a negative effect on the level of pensions. LTO has a negative effect on base salary, and a positive effect on short-term incentives. With the scores of GLOBE the findings differ. PD has a positive effect on short-term incentives, and a negative effect on pensions. UA has a positive effect on short-term incentives, long-term incentives, and pensions. LTO is positively correlated with base salary, short-term incentives, long-term incentives, pensions, and total salary.

(3)

2.2. AGENCY THEORY ... 15 2.3. CULTURE ... 16 2.3.1. Hofstede ... 18 2.3.2. GLOBE ... 20 2.3.3. Comparison ... 20 2.4. HYPOTHESES ... 21 2.4.1. Power distance ... 21 2.4.2. Uncertainty avoidances ... 22 2.4.3. Individualism ... 23 2.4.4. Masculinity ... 23 2.4.5. Long-term orientation ... 24 2.4.6. GLOBE dimensions ... 25 3. RESEARCH DESIGN ... 26 3.1. DATA ... 26 3.2. SAMPLE ... 27 3.3. INDEPENDENT VARIABLES ... 28 3.4. DEPENDENT VARIABLES ... 28 3.5. CONTROL VARIABLES ... 28 3.5.1. Age ... 28 3.5.2. Gender ... 29 3.5.3. Tenure ... 29 3.5.4. Size ... 29 3.5.5. Big four ... 30 3.5.6. Research model ... 30 3.5.7. Robustness check ... 31 4. RESULTS ... 32 4.1. DATASET ... 32 4.2. DESCRIPTIVE STATISTICS ... 32 4.3. CORRELATION ... 33 4.4. REGRESSION MODEL ... 35

(4)

4.4.1. Base salary ... 35 4.4.2. Short-term incentives ... 36 4.4.3. Long-term incentives ... 37 4.4.4. Pensions ... 38 4.4.5. Total compensation ... 39 4.4.6. Total scores ... 40 4.5. ROBUSTNESS CHECK ... 41 4.5.1. Base salary ... 41 4.5.2. Short-term incentives ... 42 4.5.3. Long-term incentives ... 43 4.5.4. Pensions ... 44 4.5.5. Total compensation ... 45 5. DISCUSSION ... 47 5.1. OVERVIEW RESULTS ... 47 5.2. DISCUSSION ON RESULTS ... 47

5.2.1. Relation control variables ... 48

5.2.2. Relation independent variables ... 48

5.3. DIFFERENCES BETWEEN HOFSTEDE AND GLOBE ... 51

6. LIMITATIONS & FURTHER RESEARCH ... 52

6.1. LIMITATIONS ... 52

6.2. FURTHER RESEARCH ... 53

7. REFERENCES ... 55

A. APPENDIX: SCORE COUNTRIES HOFSTEDE ... 66

(5)

1. Introduction

There has been a widespread discussion around the remuneration of CEOs. In this chapter we look at the public debate towards this topic. There are many factors that influence CEO compensation. Hereby a split up can be made between general factors and firm specific factors. The first will be discussed after the public debate. The latter will be discussed in paragraph 1.3. After that, the gap in the literature on the influence of national culture is presented, as well as the contribution of this research. Finally, the main research question is conceptualized in a framework.

1.1. Public debate

There has been an ongoing discussion about the compensation of CEOs (chief executive officer) for many years, which led to many research in the field (Frydman & Jenter, 2010; Murphy, 1999; Hussain, Obaid & Khan, 2014). The discussion concerning this remuneration occurred because a lot of people believe that the pay of CEOs is outrageous, leading to a widespread public attention (Mehran, 1995). Although many people find the rewards of CEOs outrageous it continues to rise. An article in the New York Times, on January 16th 2015, stated that the compensation of the Walt Disney’s CEO is still growing (Barnes, 2015). Executive pay is an international issue and is debated in different parliaments, TV shows, newspapers, and in different journals (Murphy, 1999). Another hot topic is the discussion about the remuneration of bankers. Many banks suffer losses but despite that, there is an increase in salary. On March 20 ABN AMRO decided to give top executives a pay rise of €100.000 (Kabinet: salarisverhoging ABN-AMRO, 2015), which led to a lot of commotion. Later, ABN AMRO refrained from this. The debate concerning CEO compensation is ongoing due to the continuous rise of CEO compensations. Frydman & Jenter (2010) looked at this high rise. They presented a figure that shows that CEO pay rises continuously in the U.S. (Figure 1).

(6)

More and more papers address executive compensation (examples of these papers are discussed in

paragraph 1.2.3). One reason for this is the disparity between CEO pay and the pay of “normal

employees”. The government, in some countries, tries to play a role in this by introducing laws concerning remuneration. In 1992 the Securities and Exchange Commission (SEC) adopted a compensation disclosure rule (Vafeas & Afxentiou, 1998; U.S. SEC, 1992 a, b). This rule required CEO compensation disclosure in annual reports. The commission introduced this rule to ensure that shareholders receive better information about executive compensation (U.S. SEC, 1992 c). In other countries there are also rules about disclosing executive compensation pay, like Code Tabaksblat in the Netherlands (Commission Corporate Governance, 2003). A few years ago the Dutch government introduced another regulation, which states that companies are not allowed to pay bonuses to CEOs if a company has public loans. The Dutch government also introduced the bonus ceiling. This is a new law that is introduced on the 6th of February 2015. It states that each company in the financial sector is required to implement a controlled remuneration policy (Bonusplafond, 2015). In other countries there are also regulations about the fee of CEOs. Australian Prudential Regulation Authority (APRA, 2009) developed guidance on remuneration governance, in India the Securities and Exchange Board of India (OECD, 2014) makes it mandatory for companies to get approval by a majority of minority shareholder about the remuneration packages of directors.

The remuneration of CEOs is still a trending topic and is investigated many times by scholars. Despite this, there is little known about how this remuneration is determined. Different scholars (van Essen, Otten & Carberry, 2015; Graham, Li & Qiu, 2012; Hussain, Obaid & Khan, 2014; Jouber & Fakhfakh, 2012; Kabir, 2008; Murphy, 1999) argue that despite the extensiveness of theoretical and empirical research information on how remuneration is determined is far from complete. There are still many factors valuable to explore concerning executive compensation. Therefore, it is a topic with considerable interest. Several questions arise: how is the remuneration of a CEO determined and what factors affect the remuneration? This is reviewed in the next paragraphs.

1.2. Influences on CEO compensation

In this paragraph, two topics are discussed that may have influence on the remuneration of CEOs. First the obligation of disclosure is discussed then the role of the board. These are two general factors that apply to all businesses.

(7)

1.2.1. Obligatory disclosure

Since the Securities and Exchange Commission (SEC) made legislation in 1992, there has been an obligation for firms to disclose CEO compensation in their annual reports. The mandatory disclosure of information about the remuneration of executives applies to many countries. Clarkson, van Bueren & Walker (2006) looked at remuneration disclosure in Australia. Detailed disclosure regarding executive remuneration is required since 1998 due to the Company Law Review Act. In the U.K. the Companies Act 1985 requires that the aggregate amount of directors’ emoluments are disclosed in the annual reports and accounts (Ward, 1998).

Due to the fall of Enron in 2001 more rules were developed regarding corporate governance. In The Netherlands Code Tabaksblat (Commission Corporate Governance, 2003) was introduced. This code has requirements concerning the publication of executive compensation. In Europe there are more corporate governance codes and compensation disclosure requirements. For example in the U.K. there is the Turnbull Report (Institute of Chartered Accountants in England and Wales, 1999) which is drawn up for listed companies, and also requires remuneration disclosure.

In Asia there are requirements regarding the publication of remuneration reports. Mercer (2014) conducted a research touching on executive remuneration disclosure in Asia. This study was updated in 2004 as a result of the development regarding corporate governance worldwide. It reviewed the current remuneration disclosure requirements in Asia and found that in most of the Asian countries disclosure is mandatory. It can be said many companies worldwide are obligated to disclose remuneration of executives.

The aim of mandatory disclosure is to improve corporate governance by increasing the transparency that increases shareholders value and therefore firm value. Mandating disclosure lowers agency costs and agency conflicts. Another aim of mandating disclosure is to lower the information asymmetry between shareholders and executives. This topic is further discussed in paragraph 2.2 agency theory. Craighead, Magnan & Thorne (2004) state that “regulators argue that mandated compensation disclosure improves corporate governance by permitting shareholders to enjoin boards of directors to reward executives in ways that are consistent with shareholder value creation”. (p.369)

There are also disadvantages regarding mandatory disclosure. For example the rise of disclosure costs. It is also argued that without regulations firms will only disclose information if it is beneficial for the company, so a trade-off is made between the costs and the benefits. If it is beneficial for the shareholders and therefore for the company there is no need to compel it (Lo, 2003). With this Lo says that regulation is not necessary.

(8)

Since the remuneration of CEOs became public the compensation of executives continued rising (Figure 1). Mandatory disclosure could be the cause of this. A reason for this could be the facilitation of benchmarking, due to more information about CEO pay. Pittinsky & DiPrete (2013) say that the wages of CEOs are determined by looking at benchmarking and peer groups. They found that large firms tend to benchmark towards lager firms with better paid CEOs. Faulkender & Yang (2010) and Bizjak, Lemmon & Nguyen (2011) confirm this in their research. Benchmarking could cause the rise of the CEO remuneration. Therefore this is also a disadvantage of mandatory disclosure. In

paragraph 1.2.3 previous studies benchmarking is further elaborated as a firm specific effect. 1.2.2. Independent boards

Another factor that may influence the pay of CEOs is the degree of independence of the board. As said before a lot of countries formulated rules around corporate governance, for example the Sarbanes-Oxley Act (Sox) (Institute of Chartered Accountants in England and Wales, 1999). Most of the corporate governance codes also formulate requirements concerning the independency of boards. A responsibility of the board is to appoint CEOs and to monitor them. The board of directions is invoked to represent the interests of shareholders.

After the financial crisis and accounting scandals regulators claimed that it is important to have an independent board. It is argued that before rules were developed concerning corporate governance the board was not independent enough and therefore the CEO was not monitored sufficient. A board of directors is independent if the board does not own shares of the company and does not have relationships with members of the company. Outside directors are also independent member of a board. An outside director does not work for the firm or does not have an affection and/or history with employees of that firm, and is brought in from outside the company (Rosenstein &Wyatt, 1990; Weisbach, 1988). Independent or outside independent directors do not have a bias so they can make an independent judgment about the capabilities of a CEO. This leads to a well-considered decision about CEO compensation (Chhaochharia & Grinstein, 2009). That is why most of the new corporate governance rules formulate requirements concerning this topic. Sox (Institute of Chartered Accountants in England and Wales, 1999), for example, requires that the majority of the board is independent. The idea behind these requirements is that the CEO has less power to influence the board about his own wages. According to Coughlan & Schmidt (1985) “the board approves the structure of incentives to which managers respond, including decisions about the compensation of top management” (p.44). If the board is independent it is more difficult for the CEOs to influence

(9)

them. Chhaochharia & Grinstein (2009) found that the new board requirements affected CEO compensation decisions. So corporate governance rules influences CEO pay.

1.3. Previous studies

Besides mandatory disclosure and board structure there are also firm specific determinants that could have an effect on the remuneration of CEOs. There are a lot of studies that look at factors that influence CEO compensation. The most researched determinants are size, performance, and industry & benchmarking. Previous studies concerning these topics are provided below. This section also highlights a few other determinants that are researched in previous studies.

1.3.1. Size

Some studies show that CEO compensation depends on firm size, which can be measured in different ways, like sales/revenues, assets, profitability, net income or return on assets (Ciscel and Carroll, 1980). Other factors that measure size are for instance number of employees and market value (Shalit & Sankar, 1977). Some of these measurements could also be useful for measuring performances like profitability, or return on assets. There are different scholars that link CEO compensation to one of these measurements for firm size. Cheng Wang, Venezia & Lou (2013) found that firm size results in CEOs receiving higher compensation. Tosi, Werner, Katz & Gomez-Mejia (2000) found that 40% of the variation in CEO compensation is due to firm size. There are more researchers who found a correlation between firm size and CEO compensation (Finkelstein & Hambrick, 1989; Lewellen & Huntsman, 1970; McGuire, Chui & Elbing, 1962; Roberts, 1956).

1.3.2. Performance

Performance is also a determinant that is researched a lot as an assertion of how CEO pay is determined. Researchers refer to this as pay for performance. Performance is usually measured on the basis of stock performance. Another method that can be used is for example return on equity (ROE). Banker, Darrough, Huang & Plehn-Dujowich (2012) found that the salary of a CEO positively is related to past performance. Jensen & Murphy (1990) found a small change in CEO pay if shareholders wealth increases. Abowd (1990) found that compensation is performance-based. There are more scholars who found a correlation between CEO compensation and firm performance (Coughlan and Schmidt, 1985; Finkelstein and Hambrick, 1989; Main, 1991).

(10)

1.3.3. Industry & benchmarking

Another often used factor to determine compensation is to look at competitors in an industry. As mentioned before (paragraph 1.2.1.Obligatory disclosure) scholars argue that compensation within industries are quite similar. Another way in which CEO compensation is determined is by looking at peer groups and benchmarking (Faulkender & Yang, 2010; Gravey & Milbourn, 2006).

Antle & Smith (1986) researched whether executives are rewarded as if their performances are evaluated relatively to the performance of competitors (their findings are mixed). Gibbons & Murphy (1990) found evidence that relative performance evaluation (and thus comparing with firms from the same industry) is used in compensation decisions. Usually a peer group consists of firms from the same industry (Porac, Wade & Pollock, 1999). Bizjak, Lemmon & Naveen (2008) find that the use of benchmarking is widely accepted and has a significant impact on CEO compensation. Pittinsky & Diprete (2013) investigated how CEO compensation is determined. They found that benchmarking and looking at peer groups determines the wages of CEOs. They argue that benchmarking causes the continuing rise of CEO wages. More scholars confirm this view (Bizjak, Lemmon & Nguyen, 2011; Laschever, 2013).

1.3.4. Other determinants

Next to size, performance, and industry & benchmarking there are more determinants researched by scholars that could influence the compensation of executives. Rosen (1982) found that compensation is associated with rank. Which means that wages rises with greater authority and spam of control. Other researchers looked at many different other variables. Cheng Wang, Venezia & Lou (2013) “found that the higher degree of international diversification, higher accounting earning performance, higher investment opportunities, (…) resulted in CEO's receiving higher levels of compensation” (p.27). Pathak, Hoskisson & Johnson (2014) studied the relationship between strategic changes and CEO compensation. They found that strategic changes lead to higher salary because of the uncertainties a CEO is in. There are many other factors that are used to determine CEO compensation or that influences CEO compensation. For instance the capital structure of a firm (John & John, 1993) or CEO power (Wade, O’Reilly & Pollock, 2006). As said before regulations can also influence CEO compensation (Carroll & Ciscel, 1982; Joskow, Rose & Wolfram, 1996). It can be concluded that CEO remuneration has been researched for decades but still there are some gaps to fill. The gap in the literature that this paper is trying to fill is around national culture. The effect of national culture is discussed in the next paragraph.

(11)

1.4. Culture and CEO pay

In this paragraph the effect of culture on CEO compensation is explained. First earlier research on culture and CEO compensation are summarized and reviewed. After that, the contribution of this research is set out.

As said before, scholars discussed CEO compensation in many different ways. But little research has been done into the influence of culture on CEO compensation. Culture can be “measured” in different ways. This study focuses on the dimensions of Hofstede that measures culture. This is further elaborated in section 2.3. There are cultural differences within a country but also between different countries. This study examines culture differences between countries and therefore this study highlights national culture instead of corporate culture. There already is some research done in CEO compensation and culture. Conyon & Murphy (2000) compared the differences between CEO pay in the United States and the United Kingdom. They found that CEOs in the U.S. earn 45% higher cash compensations than CEOs in the U.K. The disadvantage of their research method is that it is based on cross-national differences instead of cultural differences. There are more researches that only look at differences between two or maybe three countries. These studies primarily compare the U.S. or U.K. with other countries (Bruce, Buck & Main, 2005; Conyon & Schwalbach, 2000; Pennings, 1993).

Apart from the studies mentioned before, there are some articles published that compare more than three different countries. Fernandes, Ferreira, Matos & Murphy (2012) look at how CEO pay in the U.S. is related to fourteen other countries. Tosi & Greckhamer (2004) “examine how cultural values are related to different elements of CEO compensation in different countries” (p.657). Tosi & Greckhamer (2004) assume that “national culture plays a significant part in the nature of corporate governance systems, generally, and of pay systems, specifically” (p.658). This study was conducted in the period 1997-2001, before Enron’s fall and the financial crisis. After that, there was more regulation towards corporate governance, which could lead to different results. The study compares CEO compensation to the salaries of manufacturing employees. This is not the aim of this study. Haynes replicated this study in 2014. Grenness (2011) conducted a similar type of study but looked at less cultural variables and was also conducted before the corporate governance codes were sharpened. Bryan, Nash & Patel (2014) also looked at culture and CEO compensation. They conducted their study for a longer period, namely between 1996 and 2009, running on until after the accounting scandals. However, this study looks at a few dimensions of culture.

The question is, despite the growing regulation, does culture still influences the rewards of CEOs? It could be possible that, due to corporate governance, culture no longer influences CEO pay. There

(12)

may be more reasons for this, for example the convergence of management practices. Van der Stede (2003) found that corporate effects are larger than business-units effects in designing incentive systems. The culture set by the head office of a company plays a big role in determining the culture of the business-units, according to Van der Stede (2003). Another effect could be internationalization. Watson & Singh (2005) argue that due to globalization companies expend their operations, and compensation systems need to follow this development.

Tosi & Greckhamer (2004) say, “the mode of corporate governance in a country is one aspect of the social control of economic organization that reflects cultural values” (p.659). This view shows that culture still has an influence on CEO pay. The purpose of this study is to look at the influence of national culture on CEO compensation. According to Kabir (2008) more research concerning CEO compensation and country specific environments needs to be done to provide new insights on theoretical and empirical level. He also states that multinational studies are of more value instead of focusing on one society. There are more scholars who state that more research needs to be done concerning this topic. Pennings (1993) argues that more cross-national research is necessary, and that samples should be more divergent. Schuler & Rogovsky (1998), who investigate the use of HRM systems like pay-for performance and childcare, also call for further research. In their research they did not control for some factors like company size that may influence their findings. Haynes (2014) ends his research with the statement “that more research, relying on different sources of data and covering a wider range of countries, is needed to investigate the effects of culture on CEO compensation” (p. 23).

1.5. Conceptual model

This paper attempts to find if differences in CEO compensation arise between various countries. Differences in CEO compensation between countries may be due to the various habits in culture. Each country has different values and norms. This can affect the wages of employees in that specific country. Culture might cause the gap in CEO pay between different countries. The questions that arise are; which aspects of culture influence CEO compensation, what cultural factors cause higher or lower pay, and what is the reason for this? The main question of this research is:

(13)

Figure 2 Conceptual frameworks

To what extent does culture influence CEO compensation?

This research question is incorporated in figure 2. There are different ways to conceptualize culture. This study uses Hofstede’s framework to look into the different aspects of culture. Reasons for choosing Hofstede will be discussed in section 2.3. Hofstede has indicated several dimensions (shown in figure 2) to conceptualize culture. The different compensations components are also highlighted. Each of these components could be influenced by culture. To answer the main research question a few sub-questions are formulated.

• What does CEO remuneration consists of? • Which theories are related to this topic? • What is culture?

• How can culture be measured?

By answering these questions this study wants to examine if CEO compensation differs in dissimilar cultures and which cultural dimensions influence these distinctions in CEO pay.

1.6. Structure of the paper

The format of the paper is as follows: first different theories are discussed related to this topic (chapter 2) and hypotheses are developed. After that the research method is emphasized (chapter 3). Then the results of this research are shown (chapter 4). Next the results are discussed (chapter 5) and finally (chapter 6), the limitations of this research and opportunities for further research are explained. National Culture • Power distance • Uncertainty avoidance • Individualism • Masculinity

• Long term orientation

CEO compensation • Base salary • STI • LTI • Pensions • Total pay

(14)

2. Theoretical Framework

In this section of the paper different theories are discussed related to CEO compensation and culture. Firstly, human resource policies are explained. This is followed by the agency theory, and finally this section draws on the discussion of different definitions of culture and how culture can be measured.

2.1. Human Resource Policies

According to Schultz (1961) the definition of human capital theory is the knowledge and skills that employers acquire through education and training, which is seen as a form of capital. The human capital theory states that capacity of people to learn is also a value that is involved in the production of goods and services (Lucas, 1988, 1990). The human capital theory tries to explain the advantages of education and training as a form of investment (Aliaga, 2001). Cohn & Geske (1990) argue that investment in human capital should obtain higher wages.

Numerous studies suggest that human resource (HR) policies differ across countries (discussed below). It is important that HR policies and the corresponding compensation policies fit to the cultural values of the country where the company is established. HR practices consist of different things like planning & staging, appraising & compensating, and section & socialization (Schneider, 1988). This paper focuses on appraising & compensating. There are different ways people can be compensated. The remuneration of a CEO can consist of the following components: “basic compensation, long-term incentives, variable bonus, compulsory company contributions to social security and benefit programs, voluntary company contributions to social security and benefit programs, perquisites, and variable compensation” (Tosi & Greckhamer, 2004, p. 660). Frydman & Jenter (2010) name the following components of CEO pay: salary, annual bonus, long-term incentive plans, options, and stocks. The components this study focuses on are described at the end of this section.

Not only has the remuneration of CEOs increased, the composition of this remuneration has also changed over the years. This is clearly visible in figure 3. This figure is a representation of the 50 largest firms that are listed in the U.S. in the 1940s, 1960s, and 1990s.

(15)

Figure 3 Structure of CEO compensation (Frydman and Jenter, 2010)

According to Tosi & Greckhamer (2004) management and organizational practices are affected by culture. Each culture sees money and monetary exchange differently (Parry & Bloch, 1989). To explain the differences in CEO pay, context of cultural factors should also be considered (Conyon & Murphy, 2000; Schneider, 1988). Gomez-Meija & Welbourne (1991) argue, “companies should take culture into account when designing compensation strategies” (p. 29). They also argue that international companies should adjust their remuneration strategy based on the country in which they operate. A link must be established between the local culture and the associated norms and values of that culture. HR policies and thus rewarding systems are likely to differ across various countries. This would corroborate that the way in which CEOs are rewarded is influenced by culture. On the basis of different annual reports and available data a distinction in remuneration types is made:

Base compensation: This is salary paid in cash that consists of a monthly fixed amount. It is usually determined by benchmarking. This also includes insurance paid by the company and car allowance.

Short-term incentives (annual bonus): This remuneration is variable and depends on performance. The most common form of short-term incentives is an annual bonus. Most firms set a threshold (minimum) for performance and a cap (maximum). Under the minimum no bonuses are paid (Murphy, 1999). Bonuses are paid as a reward for achieving the performance standard.

Long-term incentives (options and shares): This form of compensation focuses on long-term payment. Examples are options and shares. An option gives the CEO the possibility to buy shares in

(16)

the future. The aim of this reward is to align the interests of the CEO with the interest of shareholders.

Pensions: This form of compensation consists of retirement plans and pensions. This type of remuneration is paid after a CEO retires and it is built up during his/her carrier.

2.2. Agency Theory

The agency theory addresses the problem that arises in an agency relationship. An agency relationship occurs when one or more persons (principal) make another person (agent) accountable, or responsible to perform services on their behalf. The agent becomes responsible for making decisions (Jensen & Meckling, 1976; Ross, 1973). There are different agency relationships, for example between employers and employees, or between debt-holders and stockholders. In this context the principals are the shareholders and the agent is the company executive (CEO).

According to Mangel & Singh (1993) the agency theory is concerned with two main problems that arise between the shareholders and the CEO, namely information asymmetry and goal incongruence. Information asymmetry refers to the problem that the agent may have more information than the principal. Goal incongruence appears when the principal and the agent do not have the same goal. Managers have few incentives to maximize the value of a company. Problems arise when an agent makes decisions that are not in the interest of the principal. Executives pursue strategic actions that could benefit them at the expense of the shareholders (Ndofor, Wesley & Priem, 2013). Mangel & Sing (1993) argue that if you want to align the interests of the agent and the principal, you could do this by structuring the compensation package.

There are two aspects of information asymmetry, namely moral hazard and adverse selection. “Moral hazard results from the inability of the principal to monitor an agent’s actions, while adverse selection corresponds to the inability of observing an agent’s private information” (Picard, 1987, p. 305). Moral hazard also refers to the lack of effort of the agent and adverse selection to the misrepresentation of the abilities of the CEO (Eisenhardt, 1989). A CEO could claim to have certain skills that the principal cannot verify. If a company has a CEO then there is a separation between decision-making and risk bearing (Fama & Jensen, 1983), because the CEO makes the decisions and the shareholders bear the risks.

To lower agency problems the principal can try to align their interest with the interest of the agent. As said before, he can do this by establishing appropriate incentives. For example by giving the CEO

(17)

shares, making both the principal and the agent want higher share value. The compensation package can ensure that the interests of the principal and the agent align. Another way to lower the agency problem is to monitor the actions of the agent. The disadvantage of this is that the monitoring costs will rise. The main idea of resolving the agency problem is to make sure that an agent behaves as if he is maximizing the principal’s welfare (Jensen & Meckling, 1976).

The agency theory focuses on the influence of performance alone, this paper also looks at the culture within a country as a possible influence on the remuneration of a CEO. Nevertheless, it is a problem all the firms in the sample deal with.

2.3. Culture

There are many different definitions of culture. Culture is something collective and covers multiple people or a group of people. According to Kluckhohn (1951) culture refers “to the distinctive way of life of a group of people, their complete design for living” (p.86). Culture is something that is learned, it is dynamic and it changes over time. Kluckhohn (1951) bases his definition of culture on other definitions. He states that:

“Culture consists in patterned ways of thinking, feeling, and reacting, acquired and transmitted mainly in symbols, constituting the distinctive achievements of human groups, including their embodiments in artifacts; the essential core of culture consists of traditional (..) ideas and especially

their attached values” (p.86).

Kroeber & Kluckhohn (1952) gave the following definition of culture:

“Culture consists of patterns, explicit and implicit, of and for behavior acquired and transmitted by symbols, constituting the distinctive achievement of human groups, including their embodiments in artifacts; the essential core of culture consists of traditional (i.e. historically derived and selected) ideas and especially their attached values; culture systems may, on the one hand, be considered as

products of action, on the other as conditioning elements of further action.” (p. 181).

There are other scholars who provide different definitions of culture, such as Eliot (2010), Hofstede (1980), House, Hanges, Javidan, Dorfman & Gupta (2004), Jahoda (2012), Keith (2011) and Wyer, Chiu & Hong (2013). Jahoda (2012) reflected on various definitions of culture. He concluded that there is not one general definition of culture that everybody agrees on. However, there are a few elements that recur in each definition. Culture is something collective, acquired and it consists of values and patterns of behavior.

(18)

As said before there is a distinction made between national culture and corporate/organizational culture. The last refers to “individual choices aggregated into critical masses of people and over time” (Wu, 2008, p.7). This involves the culture that exists within an organization. Schein (1990) uses the following definition of organizational culture:

“The pattern of basic assumptions that a given group has invented, discovered, or developed in learning to cope with its problems of external adaptation and internal integration, and that have

worked well enough to be considered valid, and, therefore, to be taught to new members as the correct way to perceive, think and feel in relation to those problems.”(p. 3)

The goal of this paper is to look at culture differences regarding CEO pay across countries, making it necessary to identify the culture of each country. Therefore, we look at national culture instead of corporate culture. National culture refers to the culture that prevails in a specific country. Hofstede (2001) refers to culture “as the collective programming of the mind that distinguishes the member of one group or category of people from another” (p. 9). The use of compensation and the value of money can be different in each country due to the national culture of that country. Parry & Bloch (1989) state that there is an enormous cultural variation in the way money is symbolized. In each culture money is viewed differently. This may explain the differences in CEO pay for different countries.

To quantify culture it is easier to look at various aspects of culture to be able to compare cultures of different countries. It is difficult to make culture measurable, however this is necessary to conduct this study. There are different scholars that made national culture classifiable. Hofstede does this by looking at dimensions. According to Soares, Farhangmehr & Shoham (2007) “Hofstede’s framework is the most widely used national cultural framework in psychology, sociology, marketing, or management studies” (p.280). Schwartz (1999) developed seven cultural dimensions that measure national culture. There are more scholars who developed a method to measure culture, such as Inglehart (1997), Maznevski & DiStefano (1995), Trompenaars & Hampden-Turner (1998), and the GLOBE team (House et al., 2004). These models (including Hofstede) are compared in the article of Taras, Rowney & Steel (2009). They state that 97,5% of all reviewed methods contain some dimensions that are similar to the dimensions of Hofstede. They also state that more than half of the reviewed models are developed through the use of students. GLOBE (besides Hofstede) is one of the few that conducted their study in firms and interviewed managers (Taras, Rowney & Steel, 2009).

(19)

There are a lot of comments on the model of Hofstede. Jones (2007) executed a study on the model of Hofstede and looked into all the scholars that did not agree on the model of Hofstede. He identified the main limitations of Hofstede according to scholars. The limitations are about the usage of surveys, cultural homogeneity, national divisions, political influences, one company approach, statistical integrity, people say it is out dated and too few dimensions are used. There are a lot of authors that express their criticism on the model of Hofstede. For example Schwartz (1999), who says that Hofstede misses data from important regions of the world. Other examples of scholars who have criticized Hofstede are Baskerville (2003), Dorfman & Howell (1988), McSweeney (2002a, 2002b), Nasif, Al-Daeaj, Ebrahimi & Thibodeaux (1991) and Smith (1998). Despite these disadvantages, the framework of Hofstede is one of the most widely used frameworks (Søndergaard, 1994). Because there is still a lot of criticism on the model of Hofstede it was decided to also use the model of GLOBE (House et al., 2004). The reason we look at GLOBE is because it is an expansion of the Hofstede model (Hofstede, 2006). Another reason is that both models measure the different cultural dimensions with scores (Shi & Wang, 2011). In the next paragraphs both models are explained.

2.3.1. Hofstede

Hofstede (1983) developed four different dimensions of culture, namely: power distance (PD), uncertainty avoidance (UA), individualism (ID) versus collectivism, and masculinity (MAS) versus femininity. In 1991 a fifth dimension was added, namely long-term versus short-term orientation (LTO). In 2010 another dimension was added, namely indulgence. This dimension is not used in this research for several reasons. There is not a lot of research conducted that uses this dimension, which makes it difficult to formulate hypotheses. Besides that, Kohun, Burcik & Skovira (2012) argue that “this dimension, as it is new, does not, as of yet have sufficient data accumulated to be as significant in conclusions as the other dimensions” (p. 990). Khan & Khan (2015) argue that indulgence is complementary to long-term orientation. Each dimension, used in this study, is explained below.

Power distance (PD)

This dimension focuses on human inequality and how societies handle this inequality. Inequality occurs in different forms like status, power and wealth (Hofstede, 2001). To measure this, the term power distance is used to look at the role of powerful and less powerful members. Every society handles power differences differently. Inequality in firms occurs for example between employers and employees. It measures the extent to which an employer can influence an employee. If a society has a high power distance index, this means that members within this society accept power inequality but

(20)

they also expect this inequality, there is no need for justification of these differences. In this type of society there are hierarchical differences, whereas in societies with a low power distance index people are more equal and if there is inequality this is warrantable. The way of dealing with power relationships is established through the values of superiors and subordinates (Hofstede, 1983).

Uncertainty avoidance (UA)

The second dimension is uncertainty avoidance. This stands for the degree to which people in a society are trying to avoid uncertainty and ambiguity (Hofstede, 2001). It gives an indication to what extent members are uncomfortable in unpredictable situations. Each society adapts to uncertainties in different ways. Societies with a high UA score tend to prefer rules, laws and regulations and prefer to operate in predictable situations (Hofstede, 1980b). In these types of societies there is need for well-defined rules and prescriptive behavior (Soares et al., 2007). Societies with a lower score are more open to change and are not seeking for lower uncertainties. Besides that, these types of societies attach less value to rules, risk avoiding, and security. It comes down to how much uncertainty a society is willing to tolerate.

Individualism versus collectivism (ID)

The third dimension stands for a preference in which individuals only take care of themselves and watch out for their own direct family (Hofstede, 2001), whereas in collective cultures (opposed) people take care of each other. A collectivist society has a greater emotional dependence on colleagues in the organization and people are more integrated in a group (Hofstede, 2001). In an individual culture people tend to only look after their own needs, and are less concerned and interested in the group in an organization (Tosi & Greckhamer, 2004). This type of culture looks at personal achievements and performance. Societies with a high score on this index are only concerned about themselves while in societies with a low score people look after each other in exchange for loyalty (Hofstede, 1984).

Masculinity versus femininity (MAS)

The fourth dimension is masculinity. Societies that are more feminine attach more value to relationships and helping each other, while masculine societies attach more value to careers, money, material rewards, and competitiveness (Hofstede, 2001). High masculine societies have more sex-differentiated occupational structures with certain jobs almost always taken by women and others mainly by men (Tosi & Greckhamer, 2004). An example could be that women are often cleaners and men often pilots. It is important to note that this dimension does not look at differences between men and women but at all the members within a culture. If a society scores high on the MAS then there is

(21)

a separation between jobs. Certain jobs are stereotyped as masculine jobs or feminine jobs (Hofstede, 2001).

Long-term versus short-term orientation (LTO)

Long-term orientation is the fifth dimension and was added in 1991. At this point only Asian countries were included (Hofstede & Minkov, 2010). In 2010 new scores of 93 countries were added. LTO measures to what extent virtuous living is a goal in a society. Societies that score high on the LTO index attach value to persistence, ordering relationships by status, observing this order, thrift and having a sense of shame. Societies that score low on this index attach value to personal steadiness and stability, protecting your “face”, respect for tradition, reciprocation of greeting, favors, and gifts (Hofstede, 2001). Long-term orientation stands for “the fostering of virtues oriented towards future rewards” (Hofstede, 2001, p.359). This dimension was added when Hofstede’s findings were replicated in Asian countries (Schneider & Barsoux, 2003). LTO cultures attach more importance to the future while STO cultures attach less value to the future.

2.3.2. GLOBE

As said before, because Hofstede is criticized a lot by scholars, this study also takes the model of GLOBE (Global Leadership and Organizational Behavior Effectiveness Research) into account. GLOBE is a project, conceived by R.J. House. He looked at the findings of Hofstede to find different aspects of culture. This project developed nine dimensions, namely: performance orientation, institutional collectivism, gender egalitarianism, uncertainty avoidance, in-group collectivism, future orientation, humane orientation, assertiveness, and power distance. GLOBE measured two sides, namely how it really is and how it should be (House, Javidan, Hanges & Dorfman, 2002). Some dimensions are the same as Hofstede, namely uncertainty avoidance, future orientation (LTO), and power distance. GLOBE split up individualism into institutional collectivism and in-group collectivism, and replaced masculinity by four other dimensions; assertiveness, performance orientation, gender egalitarianism, and humane orientation (Shi & Wang, 2011). This study only looks at the dimensions that are similar to Hofstede. This means that the dimensions power distance, uncertainty avoidance, and future orientation of GLOBE are used as a robustness check.

2.3.3. Comparison

In this paragraph both models are summarized in a table and the similarities and dissimilarities are displayed in table 1.

(22)

Hofstede Dimensions GLOBE Dimensions

Power distance Power distance

Uncertainty avoidance Uncertainty avoidance

Individualism versus collectivism Institutional collectivism In-group collectivism Human orientation

Masculine versus feminism Performance orientation

Assertiveness

Gender egalitarianism Long term versus short term orientation Future orientation

Table 1 Comparison Hofstede & GLOBE (compiled on the basis of Shi & Wang (2011)) 2.4. Hypotheses

Each cultural value can have a relation with each remuneration type. To substantiate the hypotheses, formulated in the next paragraph, there are studies used that looked at the relation between CEO compensation and the model of Hofstede and studies about compensations in other levels of a company. The different types of CEO pay this study highlights (as discussed in paragraph 2.1) are base salary, short-term incentives, long-term incentives, and pensions. All this together is the total CEO pay. Based on earlier research different hypotheses are formulated.

2.4.1. Power distance

Power distance leads to inequalities in for example wealth. Tosi & Greckhamer (2004) argue that in countries with a high power distance index the organizational structure is taller. This will result in a higher pay for each level in an organization. Power distance explains the inequalities in remuneration between hierarchical levels in organizations (Hofstede, 2001). The pay differences between these levels would be higher in high power distance societies compared to low power distance societies. Grenness (2011) argues that countries with low power distances put effort in trying to reduce income differences. It can be expected that in high power distance societies the wage gap between employees and CEO is bigger than in low power distance societies. This could lead to total higher CEO pay in high PD countries, because CEOs in high power distance countries will earn more money due to the inequality in rewards. Countries with a low power distance will try to minimize the inequalities between people (Harrison, 1993), which could lead to a lower CEO pay. Power can be expressed in authority and status (Gomez-Meija & Welbourne, 1991). A power symbol could also be remuneration. This supports the assumption that the total pay is higher in high PD societies. Gomez-Meija & Welbourne (1991) argue that the rewards are smoother and more uniform in low PD societies. This confirms the expectation that the total CEO pay is lower in low PD countries. The following hypothesis is formulated:

(23)

H1: A high degree of power distance has a positive influence on the level of total pay.

2.4.2. Uncertainty avoidances

As said before, societies with a high UA index tend to prefer rules and security. In these societies employers want to avoid risk and ambiguity (Tosi & Greckhamer, 2004). Cultures with a low score on UA are more willing to take risks. Pennings (1993) argues that high UA implies that remuneration should not be a risk, regardless of the performance. Uncertainty aversion “makes bonus and other compensation features that put some of its recipient’s income in jeopardy unsuited to its motivational climate” (p. 266). “Individuals in cultures characterized by high uncertainty avoidance may not react favorably to performance-dependent compensation since it causes them to bear more risk” (Van der Stede, 2003, p.266). It can be argued that base salary is quite riskless, so risk-averse people prefer this way of rewarding. Long-term incentives and short-term incentives are not secure because it depends on the performance of the company and is therefore less secure. Societies where uncertainties are avoided could have a higher base salary than societies where these uncertainties are not avoided.

H2: A higher degree of uncertainty avoidance has a positively influence on the level of base salary.

Other forms of remuneration that gives certainty, and does not depend on personal- or firm-performance, are pensions and retirement plans. Pensions give certainty in the future, because they are paid after you retire. By building up a pension you will have less financial uncertainties after you are done working. UA societies are probably willing to save more for their pensions, which gives them more security in the future. By doing this they avoid risks after their retirement. The hypothesis related to this is:

H3: A higher degree of uncertainty avoidance has a positive influence on the level of pensions.

It can also be argued that in nations with a high UA index people will have less performance-based incentives like long- and short-term incentives because these are not guaranteed. By making greater use of performance-based incentives an employee takes more risks because he/she does not know for sure that performance-based incentives lead to higher compensation. If an employee chooses to get more performance-based incentives he/she takes a risk. In an UA environment people will prefer to have more base salary and less performance-based incentives because they do not give any assurance. It is difficult for people to predict their performance and UA nations feel threatened by unpredictability (Johnson & Droege, 2004). “Incentive-based, achievement-oriented compensation

(24)

shifts risk to employees and increases the uncertainty of rewards (..). Given that people in cultures with high uncertainty avoidance are less tolerant of ambiguity and risk, they are likely to prefer time-based compensation, whereas those in low uncertainty avoidance cultures should have higher tolerance for uncertain, performance-based rewards” (Johnson & Droege, 2004, p.330/331). This argumentation leads to the following hypotheses:

H4: A high degree of uncertainty avoidance has a negative influence on the level of short-term incentives.

H5: A high degree of uncertainty avoidance has a negative influence on the level of long-term incentives.

2.4.3. Individualism

In an individualistic culture (as opposed to collectivism) people only look after their own interest instead of the interest of the group (Merchant, Chow, Wu, 1995). People in collective cultures see themselves as a part of a group and care for the welfare of this group. Cultures with a high individualism index want to be paid for their own performances (Schuler & Rogovsky, 1998). They want individual recognition. Hofstede (2001) argues that individualistic societies prefer equity rewarding (based on personal performance) over equally rewarding, which means that these nations prefer individual evaluation instead of group evaluation. Tosi & Greckhamer (2004) argue that in an individualistic market executives may require more performance-based compensation, rather than in a collectivist market. Management practices that fit with a group-orientated culture are inappropriate for individualistic cultures (Chow, Shields & Chan, 1991). Leaders in a collective society are more motivated by duty and commitment instead of eccentric motivation (Grenness, 2011). Individual societies prefer evaluation on the basis of their own performance, which could cause more performance-based evaluation. There are two types of performances incentives, namely on the basis of short- and long-term. It can be assumed that individual CEOs prefer more pay for performance and thus variable incentives. This leads to the following hypotheses:

H6: A high degree of individualism has a positive influence on the level of short-term incentives.

H7: A high degree of individualism has a positive influence on the level of long-term incentives.

2.4.4. Masculinity

A masculine culture (as opposed to feminine culture) attaches more value to money, accomplishment and careers, whereas a feminine environment attaches more value to relationships and other social

(25)

values (Gouveia & Ros, 2000). Van der Stede (2003) argues that members of an organization with a masculine culture are motivated by achievement and have a stronger focus on performance. In a masculine culture people are more assertive and attach value to job performance. They are also motivated by recognition of their achievements (Chiang, 2005; Tosi & Greckhamer, 2004). As said before, masculine cultures prefer recognition in the form of money. Grenness (2011) argues that money counts in a masculine society. This could result in higher base salary. Hofstede (1998) also confirms that masculine countries focus on material success and monetary incentives. Feminine countries do not prefer these types of incentives. Johnson & Droege (2004) formulated a hypothesis which states that time-based compensation, which is equal to base salary, is more desirable in feminine cultures. Thereby, it could be that high masculine countries attach less value to base salary. Based on this the following hypothesis is formulated:

H8: A high degree of masculinity has a negative influence on the level of base salary.

It could also be argued that a masculine culture prefers rewards based on performance instead of a fixed salary based on working hours. Merchant, Chow & Chan (1991) argue that countries with a high masculinity index prefer performance-based rewards whereas a high feminine index country prefers rewards based on needs. This is also argued by Newman & Nollen (1996). As argued before short-term and long-term incentives are examples of performance-based rewards. A high masculinity index could indicate that people want to be rewarded based on short- and long-term incentives. This leads to the following hypotheses:

H9: A high degree of masculinity has a positive influence on the level of short-term incentives.

H10: A high degree of masculinity has a positive influence on the level of long-term incentives.

2.4.5. Long-term orientation

The last dimension of Hofstede is long-term orientation (as opposed to short-term orientation). As said before this dimension was added later and is therefore not often used in studies concerning remuneration. People with a high long-term orientation index focus on the future. Hofstede (2001) says that these types of societies are more orientated towards future rewards. According to Newman & Nollen (1996) long-term oriented cultures prefer long-term employment rather than making quick fixes. This could also hold for the way remuneration is paid. Therefore, countries that have a future orientation prefer incentives that are in the future. These types of incentives ensure future rewards instead of rewards today. There are two types of incentives that lie in the future, namely long-term

(26)

incentives and benefits. People who are long-term orientated plan for the future (Bearden, Money & Nevins, 2006) and would probably prefer long-term incentives and pensions. Based on these arguments the following hypotheses are formulated:

H11: A high degree of long-term orientation has a positive influence on the level of long-term incentives.

H12: A high degree of long-term orientation has a positive influence on the level of pensions.

2.4.6. GLOBE dimensions

The dimensions of GLOBE are used as a robustness check, this means that there are no new hypotheses formulated. The index of GLOBE is used to corroborate the results found with Hofstede’s model. This only applies to the dimensions that correspond with the dimensions of GLOBE.

(27)

3. Research design

In this chapter the method of this research is described. First there is a list of the used data, this is followed by a description of how the sample is determined and then the different variables that are used to execute this research are explained. After that, the research model and the use of robustness check are described.

3.1. Data

There are different datasets and websites used to collect the data for this research. First of all Boardex is used to collect data. It has global information on different firms in different countries. Boardex provides information concerning CEOs, for example date of birth and tenure. Information available in WRDS (Wharton Research Data Services) is also used. WRDS has a dataset named COMPUSTAT with a sub data set named Exucomp (executive compensation). This database has information about top executive remuneration since 1992. This database focuses on countries in North America. Exucomp is therefore only used for the USA (concerning the remuneration). This information is random checked with a few annual reports. Boardex is used for the control variables age, tenure, gender, and size. All the other information (remuneration of the other countries and the big four) is gathered through annual reports. Boardex is not used for gathering information about CEO remuneration because it did not correspond with the data from the annual reports. Therefore most of the information is gathered by digging through annual reports and remuneration reports. On the basis of this it was concluded that Exucomp was reliable, and Boardex was not always accurate and reliable (concerning the CEO remuneration) the other information in Boardex was reliable.

Revenues are used as a determinate for size, the revenues of the companies were gathered through Boardex. Besides that, a list of ten biggest firms of each country was also searched on Internet to see whether it matches the information of Boardex. The differences were negligible therefore Boardex is used as a guideline to measure size and to determine the ten biggest firms of each country.

During data collection an assumption was made. If an organization did not report about a certain remuneration component I expect the value to be zero. It could be the case that an annual report does not highlight anything about bonuses, then two different assumptions can be made. Or the CEO does not receive any bonuses or it is not provided in the annual report. I assumed that the first is the case, and therefore the value is zero. This assumption was made because only big firms are included in this sample, and most countries have regulations for big firms concerning disclosure of remuneration. Therefore it is more logic that the value is zero instead of unknown.

(28)

3.2. Sample

On the basis of the dimensions of Hofstede the sample is determined. Only countries that have values on each dimension are taken into account. The two lowest and highest scoring counties per dimension are added in the sample. This leads to (4x5) 20 different countries. When there was not enough sufficient information available of a country we looked at the third country and so on. After doing this some countries overlapped which led to 13 different countries. The U.K. is added in this research because it is often used in other researches, which leads to higher comparability. Besides that, the U.K. scores high on individualism (1 point difference from Australia) and therefore it is a representative country. In the dimensions LTO only countries that were already included in the other dimensions were part of LTO dimension. Because of that, it is also chosen to add Germany that has a high score on long-term orientation, and is less influenced by the other dimensions. The dimension masculinity has two countries with a score of 70. Both countries are added in this sample, which leads to 16 different countries in the sample. In appendix A the two lowest and highest scoring countries are shown per dimension. The countries that are included in this sample are displayed in table 2 (Appendix B) and consist of 16 countries, including their Hofstede index.

Ten firms of every country are included which makes the sample size 160. These are the ten largest firms based on revenues. For some countries it was not possible to include ten firms in the sample due to the lack of sufficient information. However there is a minimum requirement of five firms for each country. This led to a sample size of 143 observations. Tosi & Greckhamer (2004) look at 23 countries. Schuler & Rogovsky (1998) only consider 11 countries. Most other researches compare only a few countries. It is difficult to determine exactly how many countries to include in the sample and how many companies because not much research is done into remuneration or management control systems, which makes it difficult to use earlier research. Kotrlik & Higgins (2001) wrote an article about sample size. In this type of research it is difficult to determine the population because the data is continuous. According to Kotrlik & Higgins (2001) if the data is continuous and the alpha is 0,05 the minimum sample size is 119. This is lower than 143, which makes the sample size correct. Ember & Ember (1998) wrote an article about cross-cultural studies. They argue that no one can examine all cultures because it is too time consuming. They say it is important to choose a sample that is representative for the research. Because this research wants to investigate the different effect of each dimension, it is appropriate to look at the highest and lowest scoring country. The sample should also be large enough (Ember & Ember, 1998). This is also the case because a lot of other studies look at fewer countries. This study looks at the remuneration of the fiscal year 2013, because not all the annual reports of 2014 are public yet. Which means that 2013 is the most recent

(29)

year to conduct this study. If the remuneration is shown in another currency than the dollars closing rate of 2013 is used.

3.3. Independent variables

Hofstede executed different studies at different times, which creates different scores for some countries. This study is going to look at the most recent scores. Hofstede has his own website www.geert-hofstede.com and http://www.geerthofstede.eu/research--vsm where all the scores are updated. These scores are used in this study.

This research performs a robustness check on the independent variables. The index of GLOBE (House et al., 2004) is used to perform this robustness check. As said earlier GLOBE identified two indexes, namely how it should be and how it really is. This research looks at the index how it really is because that is the real representation of a country.

3.4. Dependent variables

The dependent variable is the CEO compensation. There is a distinction made between base salary, short-term incentives, long-term incentives, pensions and total salary. This means there are five dependent variables that could all be influenced by the different dimensions of culture. Most of this data is gathered from annual reports.

3.5. Control variables

There are several control variables used in this research. The control variables that are used are discussed below. These variables are expected to influence the CEO compensation. Most of this data is gathered from the datasets named above.

3.5.1. Age

The age of CEOs is significantly related to compensation (Finkelstein & Hambrick, 1989; Ryan & Wiggins, 2001). A reason for this is that when a CEO becomes older his/her intellectual capabilities are enhanced due to the knowledge and experience that is gained (Mcknight, Tomkins, Weir & Hobson, 2000). There are studies conducted that look at the effect of age on compensation in other layers within a company. They found that age has a significant effect on compensation (Ostroff & Atwater, 2003). It is therefore expected that CEOs that are older earn more.

(30)

3.5.2. Gender

Another variable that can influence compensation is gender. It has been argued that if a man and a woman have the same job, the man will have a higher salary. Multiply researchers found that there is a pay gap between man and woman in the same position. For example (Tang & Wang, 2014) found that female CEOs earn less than male CEOs. Adams, Gupta, Haughton & Leeth (2007) also found differences in CEO compensation between gender. The impact of wage differences by gender is confirmed more often in other levels of an organization (e.g. Bayard, Hellerstein, Neumark & Troske, 1999; Blau & Kahn, 2000). Bugeja, Matolscy & Spiropoulos (2012) argue that gender pay gap exists in most firm levels but not at the highest level. Despite the fact that the results differ gender is still included as a control variable.

3.5.3. Tenure

Tenure stands for the time someone is working for a company. This study looks at how long someone is CEO at a certain company and not how long someone is working for that company. Hill & Phan (1991) argue that if a CEO is working for a long time for a company he/she gets more entrenched and is will get more power to peruse their own interests. Besides that the CEO is more experienced over time and will therefore be more valuable. Tenure and age are quite similar but tenure also improves firm-specific knowledge and age not per se. Johnston (2002) argues that tenure is associated with a higher CEO pay. Therefore this variable is also taken into account.

3.5.4. Size

Another determinant is firm size. As argued before, in paragraph 1.2.3 Previous studies, this can be measured in different ways, but there is no wrong or right way (Guy, 2005). This study is going to look at total revenues because this measure is often used in other studies, and also in relation with compensation (Boyd, 1994; Core, Holthausen & Larcker, 1999; Gomez-‐Mejia, 1992). Revenues is used as a determinant for performance and therefore for compensation (Stroh, Brett, Baumann & Reilly, 1996). John & Weitz (1989) also argue that sales/revenues is an objective measurement for determining compensation. Therefore this measurement is most related to this study. Another reason why revenues are used is because Boardex provides a way to range all the company on the basis of their revenues. This enables us to gather the ten biggest firms per country.

There are different reasons why a CEO of a big firm is entitled to have a high remuneration. For example Gomez-Mejia, Tosi & Hinkin (1987) state that in organizations with more hierarchical levels, the CEO earns more because of the bigger spam of control. They also argue that bigger firms

National Culture and CEO Compensation