CEO Compensation in Finland, Sweden and the U.K. 08

CEO Compensation in Finland, Sweden and

the U.K.

An empirical study of the main variables influencing CEO

compensation.

Author:

Simo Joonas Häkkinen

Supervisor:

Dr. Andreea N. Kiss

Co-assessor:

Dr. Rajiv K. Kozhikode

June, 2011

University of Groningen

Faculty of Economics and Business

Department of International Business and Management

ABSTRACT

PREFACE

This thesis is conducted as the final conclusion of my Msc International Business and Management program at the University of Groningen. During the last five months I have been studying CEO compensation and the variables influencing it.

I would like to thank a few people who have been supporting me and been giving advise. First of all, I would like to thank my supervisor Dr. Andreea N. Kiss for her valuable feedback comments and tips during the thesis project.

Last but not least I would like to thank my family and friends, especially my girlfriend and parents, for supporting and advising me during the thesis process.

TABLE OF CONTENT

2.2 Managerial Power Theory ... 10 

3. Theoretical Framework for Executive Compensation ... 12 

3.1 Firm-Level Variables ... 13 

3.2 Governance Variables ... 15 

3.3 CEO Characteristic Variables ... 18 

4. Methodology ... 21 

4.1 Method and data collection ... 21 

4.2 Data Sample ... 21  4.3 Operationalization of Variables ... 22  4.3.1 Dependent variable ... 22  4.3.2 Independent variables ... 23  4.3.3 Control Variable ... 25  4.4 Descriptive Statistics ... 25 

4.5 Assumptions of the model ... 26 

4.5.1 Comparability of the sub-samples ... 26 

4.5.2 Multicollinearity ... 26 

4.5.3 Outliers ... 27 

4.6. Additional Analysis ... 28 

5. Results ... 29 

5.1 Regression Analysis of the Whole Sample ... 29 

5.2 Descriptive Statistics of CEO Compensation Components ... 29 

5.3 Descriptive Statistics of Independent Variables ... 32 

5.4 Results of the Multiple Regression Analyses ... 34 

5.4.1 Firm-level variables ... 36 

5.4.2 Governance-level variables ... 36 

5.4.3 CEO characteristics ... 36 

5.4.4 Combined regression results ... 37 

5.5 Additional Analyses ... 37 

5.5.1 CEO salary and benefits ... 38 

5.5.2 CEO bonuses ... 39 

5.5.3 Options released to the CEO in 2009 ... 41 

5.5.4 CEO’s Option Percentage ... 41 

5.4.5 Bonus Percentage ... 42 

5.6 Discussion of the results ... 43 

5.6.1 Descriptive Statistics ... 43 

5.6.1 Firm-level variables ... 45 

5.6.2 Governance-level variables ... 48 

5.6.3 CEO characteristic variables ... 49 

5.6.4 Conclusion ... 50 

6. Conclusion ... 52 

6.1 Practical Implications ... 53 

7. Limitations and suggestions for further research ... 55 

List of Tables and Figures

List of Tables

Table 1 CEO Compensation Components ... 31

Table 2 Descriptive Statistics of the Independent Variables ... 33

Table 3 Results of the Multiple Regression Analyses ... 35

Table 4 CEO Salary and Benefits ... 39

Table 5 CEO Bonuses ... 40

Table 6 Options Released in 2009 ... 41

Table 7 CEO's Option Percentage ... 42

Table 8 Bonus Percentage ... 43

Table 9 Signs of the Hypotheses and Results ... 51

List of Appendices

Appendix 2 Regression Analysis – Whole Sample ... 62 

Appendix 3 Correlation Matrix ... 63 

Appendix 4 Signs of Additional Analyses ... 64 

1. Introduction

The purpose of this research is to develop a comparative study, which links remuneration packages in Finland, Sweden the U.K to companies’ performance and other company and governance variables and Chief Executive Officer (CEO) characteristics. A lot of literature has been published from this field, corporate governance, or at least from the remuneration programs in the U.S. However, the literature, which focuses on the comparison of executives' pay and company’s performance has been lacking consensus (Bruce, 2005). Literature on smaller European countries is limited and not much research has been done on those countries. The U.K. has an Anglo-American governance and pay system and that is why it is more closely related to the U.S. than any other European country (Bruce, 2005), including Sweden and Finland. The U.K. is also widely covered regarding executives’ compensation, and will provide a good target for comparison. Another main differences between these countries are the level of unionizing and political direction. The level of unionizing is low in the U.K., but high in Finland and Sweden, whereas the political direction is more towards socialism in Finland and Sweden than in the U.K. Bruce (2005) stated that when the unionizing is high, stakeholders are in a stronger position to influence the decision making process in a company, including remuneration packages for CEOs, which are then likely to be smaller. The number of studies done regarding the executive remuneration in Finland and Sweden is low. Conducting a study on these countries will give new insights in the field of executive remuneration in specifically smaller countries in Europe.

(Kirkpatrick, 2009).

Most of the research has been focusing on the corporate governance and remuneration systems in the U.S. (Bebchuk, 2009; Ferrarini, 2005; Erkens, 2009), while the situation in Europe, especially in small countries, has not been addressed so widely. This raises the question of whether there are differences in pay practices among these countries and whether variables, which influence the executive compensation in the U.S, have a different effect for the sample countries, namely Finland, Sweden and the U.K. in this study.

This research will not only focus on the impact of companies’ performance on CEO remuneration, but will also address other variables which might influence the remuneration of executives, namely firm size, size of a board, ownership structure, CEO tenure and CEO age. A tremendous amount of research has been conducted on the pay-performance relationship. Other variables, which have also received a lot of attention in previous literature, are firm size, size of board and ownership structure (Barkema, & Gomez-Meija, 1998). According to Zheng (2010), CEO tenure and age are the most important CEO characteristic variables. Therefore, the main purpose of this paper is to explore whether the main variables in existing literature on the U.S. influence the remuneration of company’s CEO in the U.K., Finland and Sweden. The main research question is the following:

2. Theoretical Framework

In this section, the relevant literature and main theories on executive compensation are explained. Due to the several different variables that need to be taken into account and their different theoretical bases, only one theory is not enough to explain the compensation of executives. The main theories behind the different variables influencing executive compensation are agency theory, information-processing theory, equity theory, managerial productivity theory, social comparison theory, managerial discretion theory and resource dependence theory (Barkema, & Gomez-Meija, 1998). Even though the executive compensation can be explained by many theories, however the main theories in this research will be the agency theory and the managerial power theory, which is the special case of an agency problem (Bruce, 2005). The reason for choosing these theories to explain the relationships between the independent variables and executive compensation is the variable selection and the fact that these theories have been the dominant theories explaining the executive compensation.

In previous literature, the main theory explaining the relationship between firm performance, size and executive compensation has been the agency theory and its special cases/extensions, which are usually combined with optimal contracting theory. Agency theory and its underlying assumption, optimal contracting, indicate that the executive remuneration is determined in the way that it will maximize shareholders' value and minimize agency costs, namely self-interest and opportunistic behaviour (Barkema, & Gomez-Meija, 1998; Gomez-Mejia & Wiseman, 1997; Henderson et al., 2010; Lippert & Moore, 1994). Although optimal contracting has been the dominant approach exploring executives' compensation, the objects of shareholders and managers cannot be fully settled (Boyer, 2005). Therefore, another important theory exploring executives’ compensation is the managerial power theory, which implies that executive compensation cannot be explained by optimal contracting theory alone (Bebchuk & Fried, 2002).

size and performance, are explained by the agency theory, others, such as board size and ownership structure, can be explained by the managerial power approach (Core et al., 1999; Bebchuk & Fried, 2002). To explain CEO tenure and age, both agency theory and managerial power theory are used.

In the next chapter, the theories behind the variables, namely agency theory and managerial power theory are explained, which is followed by the section where the conceptual model is presented and hypotheses are derived.

2.1 Agency Theory

The most well-known theory regarding the compensation of executives is the agency theory (Bebchuk & Fried, 2003; Barkema & Gomez-Meija, 1998; Bryan et al., 2000 Gomez-Mejia & Wiseman, 1997; Henderson et al., 2010; Lippert & Moore, 1994). Agency theory suggests that the diverging interest between shareholders and managers may occur when ownership and control is separated (Nyberg et al., 2010). Agency theory takes the perspective that directors are the shareholders’ agents, which potentially leads to diverging interests between shareholders and directors or, in other words, ownership and control, which is called an agency problem (Bryan et al., 2000). An agent compensation and equity ownership will lead to an alignment between directors and shareholders, because then executives have a higher incentive, since their own investments are at stake. In other words, the agency theory and its basic assumption, optimal contracting, suggest that remuneration packages should be structured in a way (linking pay to performance) that they favour the interest of shareholders and in a way that the alignment between shareholders and managers would be fulfilled (Bryan et al., 2000).

than would be optimal for shareholders. According to Hall (2003), companies have increasingly complied with agency theory and increased the use of the equity ownership of managers. Before 1985, the proportion of equity ownership of managers from total remuneration was zero (Hall, 2003). In 1985, companies started to use equity-based compensation, and the proportion of equity-based compensation amounted to already 66% in 2002 in the U.S. (Hall, 2003).

However, equity ownership of executives is only one solution for the agency problem. According to the agency theory, other ways to minimize the agency problems are internal and external corporate control. Internal corporate control assumes that the board, institutional investors or large shareholders, monitor the executives, whereas external corporate control comprises the market for corporate control. This refers to the active merger and acquisition markets to discipline harmful actions of executives (Nyberg et al., 2010; Jong & Ronald, 2007). Therefore, the managers do not behave only in their own interest but in the interest of shareholders, and the firm’s performance does not suffer from executives’ misbehaviour.

As a conclusion, instrumentalizing equity-based compensation and internal and external control, managers act more in the interest of shareholders and try to maximize shareholder value, which in other words should improve the performance of a firm. In Finland and Sweden, the use of performance based remuneration and long-term incentives are low compared to the U.K. In these two countries, base salary of executives is the largest proportion of the compensation package (HayGroup, 2010). In this case, agency theory might be one option to explain the compensation in the sample countries and the differences between low and high equity-based compensation countries.

2.2 Managerial Power Theory

agency problem, but also as an agency problem itself. The basic assumption of the managerial power theory is the same as the agency theory’s underlying assumption that the separation of ownership and control leads to diverging interests between shareholders and managers (Bebchuk & Fried, 2002). As the name of the theory indicates, managers have power to influence their own pay – therefore, managers with more power are paid more (Bebchuk & Fried, 2003).

3. Theoretical Framework for Executive Compensation

Based on the review of the theory and recent framework of Barkema & Gomez-Meija (1998), the following theoretical framework is developed. In the research of Barkema & Gomez-Meija (1998), the authors developed a framework of the factors influencing executive compensation, which is based on existing literature and U.S data. In their framework there are 18 variables in three categories (see Appendix 1). The variables in this study are chosen from the above-mentioned research, where the authors presented the current state of executive compensation including different variables influencing executive compensation. The authors used the most widely covered variables, namely firm size and performance as firm-level variables and board composition and ownership structure as governance variables. In this thesis, these four most widely covered variables are explored. The authors stated individual characteristics in their framework as well, which indicate the characteristics of a company’s CEO. In this research, the two most widely covered CEO characteristics, CEO tenure and age, are taken into account (Westphal & Zajac, 1994).

Figure 1 Conceptual Model: Determinants of executive compensation

Source: Author

3.1 Firm-Level Variables

There are several firm-level variables, such as firm size, firm performance, level of internationalization, growth opportunities, firm diversification, etc. However, the two firm-level variables, which are best covered and most often quoted in the literature of executive compensation, are firm performance and firm size (see Appendix 1). In general it is assumed that compensation is higher if the firm performance and size are higher. The underlying reasons for these two firm-level variables as determinants of executive compensation are discussed below.

Firm Performance

which should closely link the CEO’s pay to the firm’s performance (Barkema & Gomez-Meija, 1998). However, researchers have found only weak or even insignificant relationships between firm performance and CEO pay (Jensen & Murphy, 1990).

According to Hall & Liebman (1998), in the last two decades the use of stock options

for senior managers has increased enormously. The authors also found out that the relationship between the CEO's wealth changes and firm performance is very strong, which is due to the increased amount of stock options. Stock options can motivate managers and are a powerful incentive to link executives' pay to performance. Managers want to perform better in order to increase the share price and thus receive more pay. There are some disadvantages as well, e.g. excessive risk-taking (Ferrarini, 2005).

Even though hundreds of papers have been written on executive remuneration in the past 20 years, it is startling that there is no common trend or standard empirical specification. Reasons for this might be timing, focus on a specific industry or maybe the proportion of different kinds of remunerations paid to executives (Hallock et al., 2010). There are differences in pay practices between the U.K, Finland and Sweden. The major difference is the proportion of remunerations paid to executives. In Sweden, the percentage of base salary is high, almost 70% of the total remuneration, compared to other countries. In Finland it is approximately 55% and in the U.K. only approximately 25% (HayGroup, 2010). Agency theory predicts that pay-performance has stronger relationship if stock options are used. Equity-based compensation links the CEO’s wealth to the company – in order to receive more rewards, the CEO wants to perform better and thus increase share prices (Ferrarini, 2005). Due to the fact that the base salary is much higher in Sweden and Finland than in the U.K., and following the prediction of agency theory, the following hypothesis is derived.

Hypothesis 1.

Firm Size

Firm size is another variable, which has been examined several times. Firm size is the second most important factor, which influences the remuneration of the CEO. Generally, there has been a strong relationship between firm size and executive remuneration (Gomez-Mejia & Wiseman, 1997). When the size of a firm is larger, it needs more effort on the CEO's behalf; so correspondingly, the managerial product should increase the rewards of CEO (Henderson & Fredrickson, 1996; Finkelstein & Boyd, 1998). Even a small ownership on the side of the CEO in a large firm and a small increase in share price increases CEO’s wealth remarkably (Baker & Hall, 2004). The financial rewards of executives indeed are larger in larger firm (Jensen & Murphy, 1990; Hartzell & Starks; 2003, Carpenter, 2002). According to the managerial power theory, CEOs of larger firms have more power because the organizations are more complex, harder to control and large companies have higher growth opportunities, therefore the compensation will be higher too (Core et al., 1999). Due to the results from previous research and the theoretical arguments presented, the following hypothesis is derived.

Hypothesis 2.

Firm size is positively related to the level of CEO compensation.

3.2 Governance Variables

Ownership Structure

Ownership structure is one of the important parts of corporate governance and also has an important influence on risk taking and the executives’ remuneration packages. Corporate governance is about agency conflict (Mulbert, 2009). Based on Mulbert (2009) and Heremans (2007), there are three types of agency problems which start with a different ownership structure. Ownership structure is an important factor influencing the remunerations of executives. According to Werner et al. (2005), there is a significant difference in the compensation practices in different types of ownership structures. Their results showed that there is a strong relationship between the compensation of executives and other employees and firm performance if the firm is owner-controlled or owner-managed. Large outside shareholders and institutional shareholders have an important role when shaping the risk-appetite and compensation of executives. Institutional investors are regarded to be potentially strong corporate monitors, but the influence on the compensation packages of executives is doubtful (Werner et al.,2005).

As stated in the managerial power theory (Bebchuk & Fried, 2002) the more shares a CEO owns in a company, the more can he/she influence the election of directors or discourage hostile takeover attempts. The percentage, to which a CEO increases his/her ownership in a company, tends to decrease the ownership of outside block holders (Bebchuk & Fried, 2002). Another important factor, which might influence the compensation of the executive, is whether the ownership structure is dispersed or concentrated (Ferrarini & Moloney, 2005). In a dispersed shareholder system, boards cannot be expected to bargain at arm’s length with managers, which result in the ability to managers to substantially influence their own pay (Bebchuk & Fried, 2003). Shareholders in a firm with concentrated ownership, or firms with a block holder, can monitor the management more effectively and directly and do not need high-powered incentive contracts. In this system the manager cannot influence his/her own pay so much; instead, compensation is monitored and adjusted by block holders (Ferrarini & Moloney, 2005).

following hypothesis is derived due to the following facts: First, the percentage with which a CEO increases his/her ownership in a company tends to be linked to a decrease in the ownership of outside block holders, which means that the CEO has more power to influence his/her own pay. Secondly, in a dispersed ownership structure the executives can influence their own pay more easily – the following hypothesis is derived.

Hypothesis 3.

The level of CEO compensation is lower in companies with block holders than in companies with a dispersed ownership structure.

Board Size

According to the framework of Barkema and Gomez-Meija (1998), board size is the second most important governance factor, which influences executives’ remuneration. This is due to the fact that the board of directors is supposed to represent shareholders and act in their interest. In companies with publicly traded stocks, the tasks of the board of directors are to hire, fire and compensate the CEOs (Barkema & Gomez-Meija, 1998). If the board is relatively weak or ineffectual (large), the compensation of a CEO is higher and vice versa. Board size is significantly positive related to the compensation of CEO. When the board is large it is more difficult for directors to organize in opposition to the CEO, which leads to the situation where the CEO has more power and can influence his/her own pay (Core et al., 1999).

As stated in the managerial power theory, it is difficult for directors to organize in opposition to the CEO when the board is large and therefore the CEO will have more power to influence on his/her own pay. Moreover, in a study of Zheng (2010), the author stated that boards are less effective if they are large and large boards are associated with higher managerial power. According to the managerial power theory and the studies mentioned above, following hypothesis is derived.

Hypothesis 4.

3.3 CEO Characteristic Variables

As stated in a framework of Barkema & Gomez-Meija (1998), individual characteristics influence the CEOs pay as well (see Appendix 1). Numerous studies have explored the relationship between CEO tenure, age and CEO compensation (Westphal & Zajac, 1994). In these studies CEOs tenure has received the most attention followed by the CEOs’ age (Westphal & Zajac, 1994; Finkelstein & Boyd, 1998; Zheng, 2010; Chung & Pruitt, 1996; Henderson et al., 2010; Deckop, 1988; Baker & Hall, 2004). In this research these two main CEO characteristics, namely CEO tenure and age are taken into account. The underlying reasons for these two CEO characteristic variables as determinants of executive compensation are discussed below.

CEO Tenure

and ultimately more power to influence his/her own pay (Fredrickson et al., 1988). The compensation packages become more preferable to CEOs with longer tenure (Cyert, 2002).

Managerial power theory also argues that there is a positive relationship between CEO tenure and compensation. The more years a CEO has been working in a company, the more power he/she will have. There is a positive relationship between CEO tenure and cash compensation, but a strong negative relationship between the CEO tenure and equity-based compensation (Cyert, 2002). Finkelstein & Boyd (1998) also found a strong positive relationship between cash compensation and CEO tenure. As the agency theory suggests, the low equity-based compensation is the main reason for a CEO’s self-interest and opportunistic behaviour (Finkelstein & Boyd, 1998). Due to the facts that it is stated in previous literature that CEOs with a longer tenure will receive more compensation because they have more power and because the equity-based compensation is lower, which leads to self-interest and opportunistic behaviour of the CEOs, the following hypothesis is derived.

Hypothesis 5.

The CEO’s tenure is positively related to the level of CEO compensation.

CEO Age

A CEO’s experience influences the level of compensation and managerial power. Older CEOs have more experience and that is why it is assumed that they have higher salaries as well (Chung & Pruitt, 1996). According to the agency theory, equity-based compensation restrains the CEO’s self interest and opportunistic behaviour (Bryan et al., 2000). As stated above, CEO’s age has a negative relationship to equity-based compensation, since older CEOs prefer cash compensation. This leads possibly to opportunistic behaviour on behalf of the CEO, who is acting out of self-interest, which in turn leads to a higher compensation. According to Zheng (2010), ‘the career concern effect can also manifest differently between the early and later years of tenure, as it relates to older CEOs and usually CEOs become closer to the retirement age at the end of tenure’ (Zheng, 2010, p. 856).

In a previous chapter it was stated that the managerial power is higher when the tenure is higher and CEOs can influence their own pay to a higher degree (Cyert, 2002). Since older age is usually accompanied by a longer tenure, the managerial power effect could be related to CEO’s age as well, which in this case means that CEOs can influence their own pay more. In summary, older CEOs are more experienced, they prefer more cash compensation and they have more managerial power. Since all of these aspects are related to higher compensation of CEOs, the following hypothesis is derived.

Hypothesis 6.

4. Methodology

In this section the method and data collection is introduced. Moreover, the data sample, operationalization of the variables and statistical methods are explained. In the end of this part of the thesis, additional analyses are introduces, which is followed by the part where the result are presented.

4.1 Method and data collection

The data regarding the dependent variable, namely CEO compensation, is collected from the companies’ annual reports and web sites, meaning the use of secondary data. Moreover, all independent variables, which were not available in any databases available at the University of Groningen, were also collected from the annual reports and companies’ websites. These variables included the age and the tenure of the CEO, the number of board members and ownership structure (main shareholder). The data regarding the other independent variables and control variable, namely performance measures, firm size and age of a company, are obtained on the databases Thompson DataStream and Worldscope, which are available at the University of Groningen.

4.2 Data Sample

As mentioned earlier, the U.K., Finland and Sweden are the countries chosen to be examined in this study. From each country 39 companies are selected, which makes the sample size n = 117 companies altogether. The companies in this study are all located in the industrial sector because this is largest sector in Finland, which of the three countries has the smallest amount of companies, and thus is the limiting factor. All companies from the industrial sector in Finland are chosen. From Sweden and the U.K., companies with comparable sizes are selected randomly from the industrial sector, resulting in 39 companies from each of the countries.

Outliers were excluded. As a conclusion, the final sample size is n1 = 31 companies in

Finland, n2 = 30 companies in Sweden and n3 = 34 in the U.K., resulting in the

sample size of ntotal = 95 companies.

4.3 Operationalization of Variables

In this section, the operationalization of the variables is to be explained and argued for on the basis of previous studies. In the next sections, the dependent variable CEO compensation as well as the independent variables firm performance, firm size, ownership structure, board size, CEO tenure and CEO age will be discussed. Moreover, the control variable will be introduced.

4.3.1 Dependent variable

In this research there will be several dependent variables. However in the main analysis, the dependent variable will be total cash compensation. In additional analyses, salary and bonuses are explored separately. In addition, options released in 2009, option percentage and bonus percentage are investigated. In the following section more specific explanation is conducted for the dependent variable.

CEO Compensation

additional analysis. Moreover, options released to the CEO in 2009 and bonus percentage is covered. Bonus percentage comprises the percentage of how much the bonus can be from the CEO’s annual salary. The board decides the bonus percentage and most of the companies disclose this information in their annual reports.

4.3.2 Independent variables

In this section the operationalization of independent variables, namely firm performance, firm size, ownership structure, board size, CEO tenure and CEO age are discussed and the choices supported on the basis of previous studies.

Firm Performance

A frequently used performance measure in existing literature is the ROA (return on assets) (Werner et al., 2005). For example Core et al. (1999) used accounting return on assets, which is the ratio of earnings before interest and taxes to total assets. Bryan et al. (2000) used ROA to measure firm’s performance, as did Werner et al. (2005). Moreover Carpenter (2002) used ROA to measure firm’s performance as well. All of the above-mentioned studies investigated the relationship between executive compensation and firm performance and other firm- and governance-level variables. Some of the studies have used several performance measurements, including ROA, return on equity (ROE) and stock returns (Finkelstein & Boyd, 1998; Bryan et al. 2000). The majority of the studies related to executive compensation and corporate governance have used ROA as one performance measure, and some of the studies have used only ROA to measure performance (Carpenter, 2002; Werner et al., 2005). Additionally, profit margin (net income/total sales) is used to measure the performance.

As a conclusion ROA, ROE and profit margin are used to measure the firm’s performance in this research, due to the fact that these are frequently used performance variable in existing literature.

Firm Size

understood firm size as the market value of a company and Core et al. (1999), Lippert & Moore (1994), Harvey & Shrieves (2001) and Zheng (2010) used annual sales to measure firm size. As can be seen above, there are several ways to measure firm size. However, it seems that total sales is very popular way to measure firm size. As a conclusion, total sales of a company is used to measure the firm size in this research.

Ownership structure

Ownership structure in previous research is measured dichotomous: whether the company has a dispersed ownership structure or whether the firm has a block holder. Several studies have defined a block holder as a shareholder when it owns at least 5% of the firm’s shares: accordingly, Tosi & Gomez-Meija (1994), Core et al. (1999), Bertrand & Mullainathan (2001), Cyert (2002) and Harvey & Shrieves (2001) measured external shareholders as block holders when the ownership was at least 5%. However in the sample there are only a few companies, which have a major shareholder owning less than 5%. That might lead to weak results because almost every company would have a block holder, therefore the ownership structure in this thesis is measured as a continuous variable, by the major shareholder (percentage the major shareholder owns from the total amount of the shares).

Board Size

Board size is simply measured as the number of board members, in other words by the total number of directors on a board (Core et al., 1999).

CEO Tenure

CEO tenure is measured by the years a CEO has served as a CEO in a said firm (Lippert & Moore, 1994).

CEO Age

4.3.3 Control Variable

According to Chung & Pruitt (1995), the age of a company might influence various CEO and firm characteristics. The authors stated that firm’s growth opportunities, size and CEO age may vary among companies with different age and that is why the authors used firm age as a control variable. Therefore firm age is used as a control variable in this study as well.

4.4 Descriptive Statistics

In order to examine the influence of the independent variables on the dependent variable, multiple regression analysis is conducted. In the annual reports of the sample companies it is stated that the bonuses are paid according to the previous year performance. That is the reason why performance measures are one year prior to the dependent variable. The year of interest in this study is 2009. In Finland and Sweden the financial year starts at the 1st of January and ends at the 31st of December, whereas in the U.K. it starts at the 1st of April and ends at the 31st of March. As a conclusion, all performance measures are collected in the fiscal year of 2008, while the other independent and dependent variables are collected in the fiscal year of 2009.

The main analysis includes the regression of the dependent variable, total cash

compensation, on all of the independent variables. However, first the regression is

done in three categories, namely firm level, governance level and CEO characteristics. To test the variables in three different categories allows examining the influence of these different categories on the dependent variable. Since the independent variables include firm performance, firm size, ownership structure, board size, CEO tenure and

CEO age, the dependent variable is total cash compensation and the control variable

CEOComp = β0 + β1(FirmPerformancei)+ β2 (FirmSizei)+ β3 (OwnershipStructurei)+

β4 (BoardSizei)+ β5 (CEOTenurei) + β6 (CEOAgei)+ β7 (CompanyAgei) +εi

β0 Intercept of the line

β1, β2, β3… βn Coefficients of the predictors (for example β1 is coefficient of frim

performance)

εi The difference between predicted and observed value of the CEO

compensation

i One of the sample firms

4.5 Assumptions of the model

Before running a parametric test like the multiple regression analysis, several assumptions concerning the data have to be taken into account. In this study the comparability of the sub-samples, multicollinearity and outliers were checked, which are explained below.

4.5.1 Comparability of the sub-samples

Due to the fact that all of the listed companies (OMX Helsinki) from the industrial sector in Finland were chosen as sample companies for the sub-sample of n1, the

comparability between the country-samples is an important factor. After having randomly chosen the companies from Sweden (OMX Stockholm) and the U.K. (London Stock Exchange), it was tested if the company sizes are comparable. The comparability of the company sizes was tested with one-way ANOVA, where total sales was added as the dependent variable and country as the factor variable. The effect of country on total sales turned out not to be significant (F(2,92) = 0,41, p = 0,665). Therefore the sample consists of companies with comparable sizes in all three countries.

4.5.2 Multicollinearity

hard to determine the true influence of a specific independent on the dependent variable. It can be seen in the correlation table (see Appendix 2) that there are several variables, which are highly correlated to each other. Performance measures, namely ROA and ROE were highly correlated (r = .768, p =. 000). There was also a relatively high and significant correlation of r = .504, p = .000 between total sales and board size in the sample.

However, according to Field (2009), when the correlation coefficient is very high (more than .80) it usually leads to the exclusion of the variable. Since the board size and firm performance measures are not in the same category and the correlation coefficient is lower than .80, neither board size nor any of the firm performance measures are excluded. There were no other high correlations between the independent variables.

4.5.3 Outliers

As already mentioned, three outliers were excluded from the data set. The outliers were detected by exploring the Mahalanobis distances and Cook’s distances. Cook’s distance measures the overall influence of a single case on the whole model. It is said that a Cook’s distance value greater than 1 maybe a “cause for concern” (Field, 2009). There were no values greater than 1 in Cook’s distance in the data sample. The other test to measure outliers was Mahalanobis distance, which measures the distances of cases from the means of the independent variables. In small samples (n=100) with fewer independent variables Mahalanobis distances greater than 15 are problematic (Field, 2009).

4.6. Additional Analysis

5. Results

This section will discuss the empirical results. First it is tested if there are differences between the countries due to the fact that the purpose of this study is to compare the countries. Secondly the descriptive statistics of the compensation components and independent variables are presented, which is followed by the regression analyses and additional tests.

5.1 Regression Analysis of the Whole Sample

First, the regression analysis was conducted from the whole sample (n=95). The regression analysis was conducted with the dummy variables to see if there are differences between the countries. The U.K. was added to the reference category to see if Finland and Sweden have negative relationship to the total cash compensation compared to the U.K. as it is expected. It can be seen from the Appendix 2 that Finland and Sweden are negatively related to the total cash compensation with β = -.289, p = .010 in Finland and β = -.204, p = .074 in Sweden. The results indicate that there is negative significant relationship to the total cash compensation in Finland compared to the U.K. and negative approaching significance to the total cash compensation in Sweden compared to the U.K. In addition to dummy variables the firm size has positive significant relationship to total cash compensation in the whole sample with β = .496, p = .000. As a conclusion in Finland and Sweden the total cash compensation is lower compared to the U.K. The whole model can explain 37,5% of the variance in the level of CEO total cash compensation. Results from the regression analysis with the dummy variables can be found from Appendix 2.

5.2 Descriptive Statistics of CEO Compensation Components

the CEO. In the U.K., the mean bonus is approximately three times higher than in Finland or Sweden (bonus mean in Finland: 84752,48, in Sweden: 101807,83 and in the U.K.: 296475,56). The bonus percentage is also considerably higher in the U.K. than in Finland or Sweden (mean bonus percentage in Finland: 60,36, in Swede: 70,50 and in the U.K.: 122,92). Base salary is higher in the U.K. but the difference is not as remarkable as in the case of bonuses (base salary in Finland: 352658,77, in Sweden: 354877,57 and in the U.K.: 454257,59).

As mentioned earlier in this study, in the U.K. equity based compensation is more popular than in Finland and Sweden. It can be seen from the table below as well. In the U.K 59% of the companies released options in 2009, whereas the amount was 19% in Finland and 27% in Sweden. In the U.K. CEOs hold significantly more options compared to the two Nordic countries. The option holdings of the CEOs were remarkably higher in the U.K. as well. The mean of options hold by the CEO in the U.K. was approximately 0,86% of the total number of shares, whereas the percentage was 0,16 in Finland and 0,07 in Sweden.

As a conclusion, Finland has the lowest level of salary, bonus and total cash compensation, whereas in the U.K. all of these components are remarkably higher than in Finland and Sweden. In the U.K. the equity based compensation is remarkably higher than in Finland and Sweden as well.

5.3 Descriptive Statistics of Independent Variables

This section will introduce the descriptive statistics of the independent variables and control variable, including mean, standard deviation, minimum and maximum. The table below represents the descriptive statistics of the independent and control variables.

Table 2 Descriptive Statistics of the Independent Variables Finland Mean Std.

Deviation Minimum Maximum Total Sales 1257199,61 1690385,074 30383 6400000 ROA 7,4316 7,15792 -23,28 17,39 ROE 18,2913 17,76174 -59,78 46,03 Profit Margin 4,950581 5,1506554 -16,7490 15,6810 Major Shareholder 20,2342 13,25176 4,10 55,81 Board Size 6,45 1,767 3 11 CEO Tenure 6,3965 5,78481 1,00 22,00 CEO Age 51,58 6,546 34 65 Company Age 14,9106 5,24877 4,64 23,18 Sweden Mean Std.

5.4 Results of the Multiple Regression Analyses

Results of the Multiple Regression Analyses

5.4.1 Firm-level variables

Firm size is significantly related to the total cash compensation (salary and benefits plus bonuses) of the CEO. In Finland there is a high positive effect of β = 0.807 with

p = .000; in Sweden β = 0.510 with p = .009 and in the U.K. β = 0.587 with p = .001

the effect is still medium-sized. Moreover the results suggest that in none of the countries there is a significant relationship between the performance measures and total cash compensation of the CEO. In the case of Sweden, the control variable, company age, shows a significant positive relationship with total cash compensation with β = .346 and p = .051, indicating that CEOs in older companies get a higher total cash compensation. The amount of variance explained by the model is an adjusted R2 = .485 for Finland, R2 = .246, for Sweden and R2 = .267 for the U.K., respectively.

5.4.2 Governance-level variables

Board size is strongly and significantly related to the total cash compensation of the CEO in Finland and Sweden (Finland: β = 0.502, p = .003, Sweden: β = 0.435, p = .015). In the U.K., the relationship is only approaching significance with p = .070, β = 0.347. However, it is clear that there is a positive relationship between the board size and the CEO’s total cash compensation. In Finland and Sweden the relationship is stronger and more significant than in the U.K. The ownership structure did not show any relationship with the compensation variable. These two governance-level variables explain in Finland 30,7%, in Sweden 23,9% and in the U.K. 16,0% of the variance in the level of CEO total cash compensation.

5.4.3 CEO characteristics

5.4.4 Combined regression results

In the multiple regression analysis where all of the independent variables were included, results are slightly different. Firm size is significantly related to the total cash compensation in Finland and the U.K., and approaching significance in Sweden (Finland: β = .599 and p = .003, Sweden: β = .496 and p = .072, the U.K.: β = .548 and p = .013) suggesting that in a larger company, a CEO will receive higher total cash compensation. Board size is also significantly positively related to the total cash compensation in Finland (β = .323 and p = .061), but not in Sweden and the U.K. This might be due to the fact that firm size and board size are relatively highly correlated to each other (r = .504, p = .000). Other variables, which were significantly related in the regression analysis done separately in three categories were CEO tenure and, in Finland, also age of a CEO.

In the regression where all independent variables were included at once, these variables did not reach significance anymore. This might be due to the fact that CEO age is significantly correlated with the board size with r = .218, p = .034 and had approaching significance to the size of a company with r = .186, p = .071. There were no other significant relationships between the independent variable and total cash compensation. The multiple regression analysis indicates that all independent variables together in this study can explain 58,5% in Finland, 27,8% in Sweden and 18,9% in the U.K. of the variance in the level of CEO total cash compensation. Due to the significant correlations between some of the independent variables, the result of the multiple regression analysis seems to be biased. That is the reason why the results in the separate regression analyses are used to support or falsify the hypotheses.

5.5 Additional Analyses

the U.K. As mentioned above, there are significant correlations between the independent variables. This is the reason why the following regression analyses are done separately in three different categories, namely firm-level, governance-level and CEO characteristic categories. A table indicating the results of the separate analyses can be found after the each results chapter of the separate analyses.

5.5.1 CEO salary and benefits

The following paragraphs will investigate the relationship between the independent variables and the salary and benefits of the CEO.

Firm performance and size

After controlling for the company's age, the performance variables did not have any relationship on the salary and benefits of CEOs in Finland and Sweden. However, in the U.K., the profit margin of a company (net profit/total sales) shows a significant medium-sized relationship with CEO salary and benefits (β = .289, p = .031), meaning that a higher profit margin predicts higher CEO salary and benefits. Moreover the firm size has significant relationship with CEO salary and benefits in all of the sample countries (Finland: β = 0.687, p = .001; Sweden: β = 0.528, p = .003; the U.K.: β = 0.860, p= .000): a bigger company size predicts higher CEO salary and benefits.

Ownership structure and board size

After controlling for the effect of the company age, the results indicate a significant positive relationship between the CEO salary and benefits and the board size in all of the sample countries – a bigger board predicts higher salary and benefits (Finland: β = 0.581, p = .002; Sweden: β = 0.342, p = .052, approaching significance; the U.K.:

β = 0.574, p = .001). The ownership structure was not related to the salary and

benefits of the CEO in any of the sample countries. The control variable has a positive and significant predictive value on the CEO salary and benefits in Sweden (β = 0.381,

p = .032).

CEO tenure and age

the CEO age and tenure and CEO salary and bonuses in Sweden or in the U.K. The only other significant relationship was found between the age of a company and salary level of the CEO in Sweden (β = 0.576, p = .009). There is also a negative relationship approaching significance between the age of a company and salary level of the CEO in the U.K (β = -.320, p = .088), indicating that in the U.K. younger the company the higher salary CEO will receive, whereas in Sweden the case is an inverse.

Table 4 CEO Salary and Benefits

Finland Sweden The U.K.

Firm Performance ns ns + Firm Size + + + Ownership Structure ns ns ns Board Size + + + CEO Tenure - ns ns CEO Age + ns ns

+ significant positive relationship - significant negative relationship ns no significant results

Conclusion

It can be seen in the table above that in Finland firm size, board size, CEO tenure and age are significantly related to the salary and benefit of the CEO. In Sweden firm size and board size are significantly related to the salary and benefit of the CEO. In the U.K. the trend is the same, with the exception that in the U.K. firm performance shows also significant relationship to the salary and benefit of the CEO. However the table above indicates that in Finland more variables influence the level of CEO salary and benefit than in Sweden and the U.K.

5.5.2 CEO bonuses

This chapter will investigate the relationship between the independent variables and the bonuses of the CEO.

Firm performance and size

bonuses in Finland and the U.K., but not in Sweden. In Finland, Beta is high and positive, namely 0.786 with p = .001, and in the U.K. 0.384 with p = .045. No significant predictive relationships between company's performance and CEO bonuses were detected.

Ownership structure and board size

The board size shows a significant and strong predictive relationship with the bonuses paid in Finland (β = 0.380, p = .040) and Sweden (β = 0.427, p = .018), but not in the U.K (β = 0.189, p = .342). The ownership structure did not have any relationship with the bonuses paid to the CEO neither in Finland nor in the U.K., but in Sweden there is significant negative one (β = -.339, p = .049) indicating that the larger stake the major shareholder holds, the lower is the CEO’s bonuses.

CEO tenure and age

In the case of bonuses, the only significant relationship is the one between CEO age and the bonuses paid to the CEO in Finland. The relationship is relatively strong with

β = 0.452 and p = .018. In Sweden and the U.K., the CEO age or tenure do not have

any influence on the bonus paid to the CEO.

Table 5 CEO Bonuses

Finland Sweden The U.K.

Firm Performance ns ns ns Firm Size + ns + Ownership Structure ns - ns Board Size + + ns CEO Tenure ns ns ns CEO Age + ns ns

+ significant positive relationship - significant negative relationship ns no significant results

Conclusion

are more variables influencing CEO bonuses in Finland than in Sweden and the U.K.

5.5.3 Options released to the CEO in 2009

The first of the long-term compensation test asks the question whether the company's performance is related to the options released to the CEO in 2009. The results indicate that not even the whole model is significant in the U.K. and Sweden, meaning that it does not constitute a better predictive model than the mean. In Finland, the model is significant (p = .029), but there is only one relationship approaching significance with

p = .081 between ROA and options. The relationship is strongly negative with the

beta coefficient value of β = -1.319. However this indicates that there might be a strong negative relationship between the prior ROA and options released to the CEO. There were no other significant predictive relationships between the independent variables and the options released to the CEO in 2009.

Table 6 Options Released in 2009

Finland Sweden The U.K.

Firm Performance -* ns ns Firm Size ns ns ns Ownership Structure ns ns ns Board Size ns ns ns CEO Tenure ns ns ns CEO Age ns ns ns

* means approaching significant - means significant negative relationship ns means no significant results

5.5.4 CEO’s Option Percentage

This chapter will investigate the relationship between the independent variables and the options of the CEO.

Firm performance and size

Ownership structure and board size

Option percentage showed significant negative results with the board size in the U.K. (β = -0.389, p = .024) and Sweden (β = -0.412, p = .028), and a negative relationship approaching significance in Finland (β = -0.311, p = .097). Additionally, in Finland, there is also a negative relationship approaching significance (β = -0.311, p = .087) between the option percentage and ownership structure, indicating that the larger stake the major shareholder holds, the lower is the CEO’s option holdings.

CEO tenure and age

In the U.K. significant negative results were found between the CEO age and the option percentage (β = -0.453, p = .013), indicating that the older the CEO the less he/she holds options.

Table 7 CEO's Option Percentage

Finland Sweden The U.K.

Firm Performance ns ns ns Firm Size ns ns ns Ownership Structure - ns ns Board Size -* - - CEO Tenure ns ns ns CEO Age ns ns -

* means approaching significant - means significant negative relationship ns means no significant results

Conclusion

As it can be seen in the table above and in the previous sub-chapters, there are indeed predictive relationships between the independent and the option percentage variable. The results are all negative. In Finland there are two variables out of six, which have an influence on the option percentage of the CEO. In the U.K. there are also two variables, which are significant, whereas in Sweden there is only one. When taking all of the compensation variables into account, in Finland there are more independent variables influencing the compensation of the CEO.

5.4.5 Bonus Percentage

if companies with a higher/lower bonus percentage potential have higher/lower performance or if they are bigger/smaller. However, there is only one significant result, which occurs in Sweden. There is a significant relationship with p = .006 and β = 0.574 between total sales and the bonus percentage, indicating that larger companies in Sweden have higher bonus potential. There are no other significant relationships between the independent variables and the bonus percentage.

Table 8 Bonus Percentage

Finland Sweden The U.K.

Firm Performance ns ns ns Firm Size ns + ns Ownership Structure ns ns ns Board Size ns ns ns CEO Tenure ns ns ns CEO Age ns ns ns

+ means significant positive relationship ns means no significant results

5.6 Discussion of the results

The hypotheses, which were examined in this paper, are mainly based on the agency theory, the managerial power theory and the results of previous studies done in the field of CEO compensation, mainly originating in the U.S. To support or falsify the hypotheses, multiple regression analyses between the independent variables in three categories and the total cash compensation of the CEO are used. Furthermore the results of the regression analyses between the independent variables and additional compensation variables are discussed. The results are discussed in three categories namely firm-level, governance-level and CEO characteristics. However, first the descriptive statistics regarding the compensation measures are discussed.

5.6.1 Descriptive Statistics

percentage of base salary is high, almost 70% of the total remuneration, compared to the other countries: in Finland it is approximately 55% and in U.K. only approximately 25% (HayGroup, 2010). This means that the overall compensation of the CEO is remarkably higher in the U.K. because the base salary is already much higher than in Finland and Sweden even though it accounts for only 25% of the total remuneration in the U.K. The reason for this cannot be found from the data set or analyses because the data set is comparable, but one possible reason for this could be the high level of unionizing in Finland and Sweden and other possible country factors, such as wealth of a country and political direction.

In Finland and Sweden the level of unionizing is very high, 74% in Finland and 71% in Sweden, whereas in the U.K. the level of unionizing is only 27% (etui). According to the stakeholder theory, the executives suspend their own immediate self-interest and take company’s stakeholders into account as a whole, while making important enterprise decisions (Driver & Thompson, 2002). This is the case in high unionizing countries as well because unions are supposed to drive workers rights and that is why companies have to take stakeholders into account when making company decisions (Singh & Agarwal, 2002). Moreover Singh & Agarwal (2002) stated that unions are pressuring executives with strategies including negotiating executive pay concessions in collective bargaining and with media pressure, hoping to reduce the executives' pays.

sample countries.

Due to the comparable sample and the facts that there are no significant differences in the independent variables between the sample countries, the most logical reason for differences in CEO compensation between the countries should be found on the country level. Additionally to the level of unionizing, the wealth of the country could be one explanation but not in this case. In 2009, the GDP per capita was €31843 in Finland, €31181 in Sweden and €25117 in the U.K. (WorldBank). If country wealth would be the explanation for the results found, the GDP per capita would even suggest that in Finland and Sweden the compensation should be higher because these countries are wealthier. However political direction of a country could be one explanation for the differences in pay practices through the countries. The political direction in Finland and Sweden is more socialistic than in the U.K. (more capitalistic) and in these two Nordic countries the equality is important factor, which might influence the salaries as well. The political direction influences the taxation system as well, which might also influence the compensation practices.

There are several other variables, which might have an influence on the level of the CEO’s compensation. However, in the author's opinion, the most probable reason for the low level of compensation in Finland and Sweden is the level of unionizing and political direction of a country.

5.6.1 Firm-level variables

As already seen from the main analyses, firm performance is not related to the CEO total cash compensation in any of the sample countries. However, the profit margin has a significant positive relationship with the CEO salary and benefits in the U.K. This indicates that the first hypothesis is partly supported because it stated that the company’s performance is positively related to the CEO compensation in the U.K., but not in Finland and Sweden. The performance is not related to any of the CEO compensation measures in Finland and Sweden but is indeed related to the salary and benefit of the CEO in the U.K., indicating that CEOs in companies with better performance, measured by profit margin, have higher salaries and benefits. Any of the performance measures are not correlated with firm size or board size, so the reason for this relation in the U.K. is something else.

prior firm performance and options released in Finland. This indicates that if the firm performance is low the company will release options next year. This makes also sense according to the agency theory. When CEO owns equity in a firm, they are more likely to act in the interest of the shareholders and improve performance (Bruce, 2005). This indicates that Finnish companies try to motivate CEOs to better results but this also indicates that CEOs are rewarded with stock options due to the poor performance.

Another interesting fact is that the firm performance is not related to the bonuses paid to the CEO. This appears strange indeed, because bonuses are supposed to be paid due to the prior year performance. The reason for this might be that the CEOs of companies with already high levels of performance receive a certain level of bonus, whereas a CEO of a company which increases the performance remarkably will receive more bonuses, but still the bonus is the same as the company’s which showed already a high performance, or even smaller. Due to this reason, the influence of the company's performance on the CEO's bonus cannot be detected without a longitudinal study. However Fattorusso et al. (2007) did not find any significant relationships between the company’s performance and CEO bonuses either, even though longitudinal study was conducted.

As performance-pay relationship has received enormous amount of attention in the existing literature (Barkema, & Gomez-Meija, 1998) and it seems that there is no relationship between CEO compensation and firm performance or at least no common trend or empirical specification, one suggestion is proposed. Agency theory suggests that rewarding CEOs with long-term incentives (options) would link CEO compensation to firm performance (Bryan et al., 2000). However this has not been the case in all of the studies done in the field of CEO compensation (Bruce, 2005), and not in this thesis either. According to Seppänen (2009) the CEO compensation should be linked to the company’s dividends. The more shareholders receive dividends the more CEOs would receive compensation. This might solve the problem of the weak performance-pay relationship. When the pay would be linked to the company’s dividends it would decrease the opportunistic behaviour of the CEOs. The pay would not be related to the share price as it is if options are used. The CEOs would not only want to increase the share price but would also want to make more profit and that way the company would be able to pay more dividends to the shareholder and CEO would receive more pay. In this case the roles of the owners and managers would stay separated but the interests would be common (Seppänen, 2009). The idea of linking CEO compensation to the company’s dividends sounds very good. Seppänen (2009) did not propose any disadvantages of this idea. However there might be some disadvantages as well. If the CEO compensation would be linked to the company’s dividends the CEO of the company would make it sure that the company will pay dividends. When the profits are given to the shareholders and managers there will not be money for the investments anymore. In a long run this would mean that the company is not growing and not making so much profits because the company does not do so many investments as they would if the money would stay in a company. The idea is good but it might weaken the company’s performance and growth in a long run due to the weakened level of company investments. However this idea sounds like a best possible solution for linking the firm performance to the compensation of the CEO. If companies would link CEO compensation to the dividends, the performance-pay relationship might be easier to explain or at least there would be one.

CEO in every sample country, related to the bonuses in Finland and the U.K. and related to the bonus percentage in Sweden. This result supports the second hypothesis and indicates that larger companies have higher CEO remunerations, as suggested by managerial power theory and previous studies (e.g. Jensen & Murphy, 1990; Hartzell & Starks, 2003; Carpenter, 2002). In the case of the U.K. and Finland the results are consistent with those of Fattorusso et al. (2007), who found out, using British companies from FTSE 350, that firm size is associated with higher CEO bonuses.

5.6.2 Governance-level variables

It was stated in the hypothesis of ownership structure that companies, which have block holders, have a lower CEO compensation due to the monitoring effect of the large shareholder (Bebchuk & Fried, 2002). The regression analysis showed that the relationship is negative in every country, but it cannot be said that it is reliable because the results are not significant. So the results are consistent to those of Core et al. (1999) who stated that when block holder exists, CEO compensation is decreasing. Significant negative results were found between the bonuses paid to the CEO and ownership structure in Sweden and between option percentage and ownership structure in Finland. The U.K. did not show any significant results. In the case of Sweden this means that the larger the major shareholder, the lower is the bonus and in Finland the results indicates that the larger the major shareholder, the less the CEO holds options. As a conclusion, the hypothesis regarding the ownership structure is rejected, but it can be stated that in Finland and Sweden the size of the major shareholder matters, which is not the case in the U.K. The reason could be found in the mean size of the major shareholder. On average, the major shareholder is approximately 25,2% larger in Sweden, and 7,7% larger in Finland than in the U.K., meaning that the shareholders are indeed larger in Finland and Sweden than in the U.K., which could lead to higher control and smaller CEO compensation. Especially in the case of Sweden the major shareholder is much larger than in the U.K. and quite much larger than in Finland as well. This might be the reason for the relationship bonuses paid to the CEO and ownership structure in Sweden because the largest shareholders are larger in Sweden, they can control more and influence more on the bonuses paid to the CEO.

higher is the compensation. This supports the fourth hypothesis, which stated that the board size is positively related to the compensation of the CEO. This also indicates that the results from previous studies conducted on U.S. samples and the presumption of the managerial power theory are applicable in the sample countries.

The board size is positively related to the salary and benefits in all of the sample countries, positively related to the bonuses in Finland and Sweden, and negatively related to the CEO’s option holdings in all of the sample countries. The last one describes that the larger the board, the less the CEO holds options. This result is slightly in contradiction with the previous studies and the managerial power theory. It would be logical that CEOs would receive more options if they have more power. A possible explanation for this could be that CEOs with more power in companies with larger boards prefer more cash components because they receive them immediately, with a relatively low risk of not getting them, which is the case in long-term objectives. According to Bryan et al. (2000), CEOs closer to retirement age prefer to minimize the uncontrollable effects of the markets on their personal wealth and that is why they prefer cash compensation over the equity-based compensation. This statement would read like, if the CEOs have more power they will prefer more cash compensation than equity based compensation because cash compensation is not influenced by the uncontrollable effects of the markets. This is also true if we look at the results. Board size is positively related to the cash components but not to the long-term compensation components in this study, indicating that the board is weak when it is large, CEO has more power and he/she can influence more on his/her own pay.

5.6.3 CEO characteristic variables