] Is the Executive Compensation and Firm Performance Relationship Positive?: Evidence From Swedish Listed Firms Supervisor: Dr. Jaan Grunberg Uppsala Universitet & Referent: Dr. Niels Hermes University of Groningen 2008

(1)

Author: Konstantinos P. Petrou

E-mail author: k.petrou@student.rug.nl

2008

Supervisor: Dr. Jaan Grunberg

Uppsala Universitet

&

Referent: Dr. Niels Hermes

University of Groningen

[

Is the Executive Compensation and Firm

Performance Relationship Positive?:

Evidence From Swedish Listed Firms

]

(2)

2 Konstantinos Petrou

Abstract

Departing from the agency theory it has been argued that corporate performance

should be reflected in executive compensation. It has been reasoned that the interests

of managers and shareholders can be aligned by linking manager’s compensation to

firm performance, as a way to increase shareholders protection. This research

examines the relationship of corporate performance and CEOs remuneration for a

sample of Swedish listed firms. We provide empirical evidence for a sample of 127

Swedish listed firms from all industries, by using data from the firms’ annual reports

for the period 2005-2007. We are using a three-year panel data assuming both

contamporaneous and 1-year lagged relationships, so as to examine our hypothesis

that executive pay should reflect corporate performance. Our empirical evidence

indicates a negative relatioship between executive compensarion and firm

performance, supporting those who argue that executive compensation is an

endogenous function of the firm and is determined by other factors besides

performance.

(3)

TABLE OF CONTENTS

I: Introduction……….4

II: Theoretical framework………..6

III: Literature review………..7

3.1. Main research question……….10

IV: Corporate Governance in Sweden ……….11

V: Ownership and control………13

5.1. Foreign ownership………15

5.2. Ownership and the agency problem………..16

5.3. Differentiated voting rights………...17

VI: Board of directors………..18

6.1. Board size………...20

6.2. Anglo-American board membership……….…21

VII: Data, methods, and description of variables………23

7.1. Data………...23 7.2. Methodology……….26 7.3. Variables………27 7.3.1. CEO remuneration……….27 7.3.2. Firm Performance...………...28 7.3.3. Accounting-based measures………...28 7.3.4. Market-based measures………..29 7.3.5. Control Variables………...31 7.3.6. Dummy variables………....32

VIII: Empirical analysis and results………...34

IX: Limitations and recommendation for further research……….………..43

X: Conclusion………...44 References

(4)

I: Introduction

While there has been an explosion in research on executive compensation in U.S. firms in recent years, systematic research on executive compensation outside of the U.S., is still relatively scarce mostly due to the limited data availability (Kato, Kim and Lee, 2007). The critique on executive pay has a long history, but especially over the last years, after the recent cases of corporate fraud, the topic of executive compensation has been given significant attention. After Enron’s financial fraud and bankruptcy in December 2001, chief executive officer (CEO) remuneration in traded firms has been a popular public topic in a number of countries (Makinen, 2007). Perhaps more importantly, the case of Enron was not a unique incidence. Afterwards we have witnessed, for example, the accounting and compensation scandals of Arthur Andersen, AOL Time Warner, Tyco, WorldCom and Xerox in US, but outside of the US we have also witnessed corporate scandals, such as the case of Parmalat in Italy, the Dutch-based Royal Ahold, and Swedish Skania. These series of corporate scandals have triggered a significant attention on the need for improvement of corporate governance mechanisms, in an attempt to increase transparency and information to shareholders, in an effort to prevent fraudulent activities and restore shareholders confidence.

(5)

The aim of this paper is to add to the existing literature by presenting a quantitative empirical analysis of CEO compensation in Sweden and, in particular, to examine if CEO remuneration reflects corporate performance. So far, research on this topic has been limited in Sweden. Randoy and Nielsen (2002) examined the pay-performance relationship for a sample of Norwegian and Swedish listed firms but they found no statistical relationship. This paper fills an important gap in the literature by providing new econometric evidence on the pay-performance relation solely for CEOs of Swedish listed firms.

The paper is organized as follows. In section 2 we discuss the main theoretical concept behind this research, namely the agency theory. Section 3 summarizes the key existing literature on executive compensation and firm performance relationship. In section 4 we discuss the corporate governance framework of Sweden and we focus on its most important characteristics and distinguishing features. In section 5 we discuss the ownership and control characteristics of Swedish firms, and the impact of these characteristics in corporate performance. The role of the board and the proposed relationship between Anglo-American board membership and board size with CEO remuneration is presented in section 6. The proposed relationships between the firm performance variables and CEO compensation as well as an analysis of our empirical design, is presented in section 7. The empirical results and their interpretation are presented in section 8. In section 9 we summarize the key findings of our study and in section 10 we refer to the limitations of this study, and furthermore we suggest some topics for additional research.

(6)

The main theoretical idea behind this research is the agency theory. Agency theory suggests that the performance-based pay contract, which links pay to shareholder wealth via performance indicators, such as share prices or accounting based targets, is a powerful way of attracting, retaining, and motivating managers to pursue the shareholders’ agenda (Jensen and Murphy, 1990, 2004; Hall and Liebman, 1997). Based on agency theory it has been reasoned that the interests of managers and shareholders can be aligned by linking manager’s compensation to firm performance (Murphy; 1985, 1999). Theory suggests that linking executive compensation to firm performance can mitigate agency problems by aligning manager’s incentives with those of owners (Jensen and Meckling, 1976). However, differences in corporate governance systems may influence the effectiveness of this potential alignment mechanism (Fernandes, 2008). One of the main goals of effective corporate governance is to solve the agency problem, which was first identified by Jensen and Meckling (1976). The typical principal-agent model suggests that shareholders should ensure manager’s compensation is contractually linked to the latter’s actions. As actions are not directly observable, a contract that aims to maximize shareholder’s objectives must be devised. One way to do this (align the interests) is to make manager’s compensations a function of firm performance (Fernandes, 2008). So far, extensive literature on corporate governance and executive compensation supports the view that executive compensation is, or at least should be related to corporate performance.

III: Literature Review

(7)

the US. Empirical CEO compensation studies have been conducted within economics (e.g. Becker, 1975; Hall and Liebmann, 1998; Conyon and Murphy, 2000), finance (e.g. Jensen and Murphy, 1990; Yermack,1997), accounting (e.g. Murphy, 2000), and management (e.g. Gomez-Mejia and Wiseman, 1997)1. The available research on pay-performance relationship is documented in two widely cited survey papers of Murphy (1999) and Core et al. (2003). These surveys, mainly based on US studies, deal with the relationship between CEO and shareholder wealth (Duffhues and Kabir, 2008). The most accepted evidence is that the relationship is positive and the major force behind it is equity-based incentives received by the CEO (Hall and Liebman, 1998). Hall and Liebman (1998) use 15-year panel data on the large US firms from 1980 to 1994. In this study they find that CEO compensation is highly responsive to firm performance. On the other hand though, there are studies in the US that did not find this relationship to be positive. Differences in research methodology may explain some of the inconsistent conclusions, but there is even a lack of consensus among some studies that used identical or very similar research designs (Firth et al., 1996). Core et al. (1999) investigated the role that corporate governance mechanisms play in controlling CEO pay. They find that US firms with weaker governance mechanisms have greater agency problems, as reflected in higher pay to CEOs than their firm’s performance would justify (Basu et al., 2006). Jensen and Murphy (1990) used CEO compensation data on a sample of 1295 firms from 1974 to 1986. They estimated pay for performance models in first-differences to account how change in CEO compensation is related to change in shareholders’ wealth. Their paper is widely cited both in academic and in popular press, since it raised doubts whether the US companies are managed efficiently (Makinen, 2007). While there has been an

1

(8)

explosion in research on executive compensation in US firms in recent years, systematic research on executive compensation outside of the US, is still relatively scarce mostly due to the limited data availability (Kato et al., 2007).

Countries studied in Asia include Japan (Kato, 1997; Basu et al., 2007; Kato and Kubo, 2006), Korea (Kato et al., 2007), and China (Kato and Long, 2006; Firth et al., 2006). In general these studies find a positive pay-performance relation, similar to what has been documented in the United States (Unite et al., 2007). The pay-performance relation for firms in Japan and Korea have been shown to depend upon group affiliation (Kato et al., 2006), and in China on state ownership (Firth et al., 2006). Empirical studies from some other European countries consistently document that pay-performance sensitivity is very low for UK2, and for Germany3. Duffhues and Kabir (2008), which examined the executive compensation firm performance relationship for the case of the Netherlands, reported a negative relationship between executive compensation and all the accounting and market based variables they used in their research. Additionally some studies, like the one of Fernandes (2008) on a sample of Portuguese listed firms do not find any link between executive remuneration and corporate performance.

The objective of this paper is to report the results of empirical research into the compensation of CEOs for Swedish listed firm. According to Firth et al. (1996), different social structures and different historical economic performance mean that the American research conclusions may not be generalizable to the Scandinavian environment. In their study on a sample of Norwegian listed firms, Firth et al. (1996) reported no significant association between remuneration and corporate performance

2_{See Conyon and Murphy (2000).} 3

(9)

as measured by accounting profitability and stock returns. Randoy and Nielsen (2002) examined the relationship between firm performance, corporate governance and CEO compensation for a sample of 120 Norwegian and 104 Swedish publicly traded firms. Their empirical evidence indicates a significant statistical relationship between the size of the board and CEO compensation, and foreign board membership and CEO compensation. On the contrary, however, they do not find evidence that CEO compensation and firm performance are statistically related (Makinen, 2007). A later research of Oxelheim and Randoy (2005), examines the impact of Anglo-American financial markets on CEO compensation for Swedish and Norwegian listed firms. Their results indicate a positive relationship between Anglo-American cross-listing as well as Anglo-American board membership and CEO compensation. What is of great importance is the fact that all the research conducted for Scandinavian countries over the last years indicate a significant increase in the levels of CEO compensation (e.g. see Eisenberg, 1998). This remark is also in line with Oxelheim and Randoy (2005), which suggest that as firms become internationalized, this may also cause institutional contagion from the Anglo-American market system, which in turn is likely to have a strong effect on important dimensions of corporate governance such as CEO compensation. This remark comes to question the view that CEO compensation levels in Sweden is among the lowest among E.U. countries4.

Summing up, this study is part of the growing literature on executive compensation and corporate governance. The empirical part of this paper is based on a sample of Swedish listed firms. Sweden provides an interesting case for analyzing the effect of firm performance on CEO compensation. According to Randoy and Oxelheim (2005), countries with a relatively low level of CEO compensation can expect to see higher

4

(10)

compensation levels as their firms internationalize and they get integrated into the global market for top executives. Sweden has some of the most internationally oriented companies in the world and was recently ranked #6 in terms of relative transationality (UNCTAD, 2000). Moreover, Sweden’s is considered as a country with high investor’s protection which makes it particularly interesting as the object of this study is the linkage between firm’s performance and CEO pay. Finally, due to this internationalization, there is data availability, especially over the last years (annual reports in English), which makes this study feasible. By analyzing empirically the case of Sweden we plan to fill in an important research gap, since we are going to examine the pay-performance relationship for the 2005-2007 periods, thus providing a fresh insight regarding the corporate performance-executive compensation relationship, for a region on which relevant research is rather limited.

3.1. Main Research Question

The aim of this study is to add to the existing literature by presenting a quantitative analysis of executive compensation in Sweden and, in particular, to examine the popular belief that executive pay should reflect corporate performance. In essence, we would like to answer the following research question:

Is the relationship between firm performance and CEOs remuneration for Swedish

listed firms positive?

IV: Corporate Governance in Sweden

(11)

governance according to Schleifer and Vishny (1997) is defined “as the ways in

which suppliers of finance to corporation assure themselves of getting a return on

their investment”. When the corporate governance mechanisms are weak,

shareholders may suffer a higher risk of expropriation by the managers (Mak and Kunsandi, 2002). The way in which corporations are governed has received popular attention in recent years due to a series of corporate scandals, such as for example the so-called “Skandia scandal” in Sweden, Parmalat in Italy or Enron in the US. These are cases where management has misused corporate resources and failed to serve the interests of the owners (Eklund, 2007). In essence, the corporate governance system in a country is the institutional framework that supports the suppliers of finance to corporations and enables firms to raise substantial amount of capital (Shleifer and Vishny, 1997). As it has been indicated by La Porta et al. (1997), countries with increased investor protection, measured by the character of legal rules and the quality of law enforcement have larger and more developed capital markets.

(12)

which is characterized by somewhat higher investor protection than other civil law countries, but lower than common law countries such as the UK and the US. Swedish law, drawing on Germanic and Roman, is neither as codified as in France and other countries influenced by the Napoleonic Code, nor as dependent on judicial practice and precedents as in the Unites States (US Department of State, 2008).

There exists a large literature that stresses the considerable differences in corporate governance practices (Pedersen and Thomsen, 1997; La Porta et al., 1998). A classical distinction in this literature is between the market-based corporate governance systems, typically found in the US and the UK, and the bank-or control-based systems, typically found in Continental Europe (Bebchuck and Roe, 1999). Historically, systematic differences between countries with regard to law and enforcement have accounted for substantial variations in financial development and performance (La Porta et al., 1997; Lopez-de-Silanes et al., 1999). The form of corporate governance in the Scandinavian capital markets, and in most countries in Continental Europe, is the so-called insider or control oriented system (La Porta et al. 1999; Berglof, 2000). In this corporate governance system, the emphasis is on the ability of large shareholders to monitor corporate behavior (Angblad et al. 2001), whereas the Anglo-American system puts more emphasis on monitoring by way of board independence, a market for corporate control, and institutional monitoring (e.g., SEK and Stock Exchanges) (Thomsen et al., 2008).

(13)

responsibilities at three levels: owners / the annual general meeting of shareholders (AGM), the board of directors and the executive management (Carlsson, 2007). The annual general meeting (AGM) of shareholders which is the forum of the owners to exercise their rights involves decision making regarding issues like discharging the Board for the year passed after having received the report of the auditors and having approved the accounts, deciding on profit allocation and dividends, and electing the chairman (separately) and the board as a whole of the company to take charge until the next AGM. Also, appointing the Nomination Committee or deciding on the nomination procedure for next year’s AGM, as well as the remuneration of the board and remuneration principles for the executive management. (Carlsson, 2007). The board of directors in Swedish firms, is solely focused on its function as a board, and differs from the Anglo-Saxon system where the task of the board is mixed with that of executive management. The CEO has the full operational responsibility as outlined by the instructions given by the board. The instructions should be documented in a specific document as provided by the company act. The three distinctive levels and clear division of roles and responsibilities distinguish the Swedish system from most other systems – the Anglo-Saxon as well as the Continental Models (Carlsson, 2007).

V: Ownership and control

(14)

model resembles the typical corporate control model of Continental Europe. A distinguishing feature of the Swedish model is that control is typically based on a smaller capital base than in other European countries (Henrekson and Jakobson, 2003). Faccio and Lang’s (2002) study of more than 5000 Western European firms ranks Sweden #1 in terms of dual class shares and 6# in terms of frequency of pyramid structures, indicating an extreme separation of ownership and control (Heany and Holmen, 2007). Faccio and Lang (2002) study the ownership in Europe and find that corporations are predominantly controlled by families in Continental Europe. This control is achieved without corresponding capital through using primarily three different control mechanisms: vote differentiation of shares, pyramid ownership, and cross-holdings (Eklund, 2007). In Sweden, owner control has been leveraged by all these three mechanisms which allowed owners increased voting rights compared to the level of capital they hold in the firm. This lead to the creation of some “family spheres”, which have control of most Swedish listed firms. Swedish industry and particularly many of the large Swedish multinational corporations are dominated by private owners (Carlsson, 2007). Sweden has a tradition of large business groups and large industrial firms. Close to half of the stock market capitalization in Sweden has long been controlled by the business spheres of Handelsbanken and Wallenberg (Thomsen, 2008). The most famous financial family of Sweden is the Wallenbergs, probably unparalleled in the world, by its significant role in the development of Corporate Sweden for the last 150 years. The Wallenberg Sphere, as it is usually called, together with nine other private owner spheres5, they play leading roles in companies accounting for more than 50 per cent of the total value of Stockholm Stock Exchange (Carlsson, 2007). The long-run survival of controlling families in Sweden

5

(15)

looks as follows. 4 (Wallenberg, Bonnier, Johnson, and Soderberg) of the 14 families that established control before 1920 are still exerting control, while 6 (Rausing, Kamprad, Olsson, Wallenius, Persson, and Stenbeck) of the 23 that founded firms between 1920 and 1965 are still significant and active owners today (Hogfeldt, 2004). 31 of the largest firms today were founded before 1914, only 8 in the post-war period and none after 1970 (Hogfeldt, 2004).

5.1. Foreign ownership

A feature similar to that of most European countries is that foreign ownership of the Swedish public companies, not least through foreign mutual funds and pension funds, has become significant, varying between 35 to 40 per cent (Carlsson, 2007). There has been a dramatic change in the ownership structure in Swedish firms in the last few decades. The most obvious change is a strong increase in foreign ownership of Swedish firms. In terms of total activity most of this increase concerns acquisitions of firms listed on the Stockholm stock exchange6 controlled by Swedish owners. As a result, there has been a dramatic increase in foreign ownership of listed shares since 1993 when the restrictions on foreign ownership were abolished (Hogfeldt, 2004). Naturally this is part of a general trend towards globalization, but foreign ownership has increased more in Sweden than in other developed industrial countries (Henrekson and Jakobson, 2003). As foreign ownership (especially from Anglo-American investors) increases, the demand for increased corporate control in order to protect minority shareholders also increases. Even though foreign and institutional ownership has increased recently, the industrial owners have been able to hold on to

6_{Henrekson and Jakobson (2003) document that roughly 30 listed companies were sold to foreign}

(16)

controlling stakes in the largest publicly traded firms (Agnblad et al., 2001; Hogfeldt, 2004).

5.2. Ownership and agency problem

The agency problem refers to cases where managers pursue personal goals, rather than try to increase firm value and maximize shareholders’ wealth. In general, outright expropriation of corporate assets and investor funds by managers is likely to be small in developed countries such as the Scandinavian ones. Overinvestment in pursuit of other ends than profit maximization and misallocation of assets is more likely to be a problem (Eklund, 2007). The two most common ways of dealing with the agency aspects of corporate governance are, according to Schleifer and Vishny (1997), first legal and regulatory protection of investor and minority rights, and second large and concentrated owners. Focusing on the latter, we will control for the impact of large shareholders in the pay-performance relationship for CEOs of Swedish listed firms. Past research indicated that large shareholders have the incentive to collect information and monitor the management, and have also enough voting control to put pressure on the management (Schleifer and Vishny, 1997). Existing literature has indicated a positive relationship between block holder ownership and corporate performance. Various studies7 have indeed found that the presence of external blockholders is associated with stronger links between top management compensation and firm performance (Barkema and Gomez-Mejia, 1998). Blockholder ownership, and with this the incentive to monitor manager is one way of controlling managerial behavior. In line with the above argument we include blockholder ownership as a control variable in our research.

7

(17)

5.3. Vote differentiated rights

As noted above, in Sweden firms are allowed to have vote-differentiated shares; hence voting rights are frequently separated from the amount of capital invested. Consequently there is a possibility to retain control over a company by owning a relatively small fraction of the shares (Bjuggren et al., 2004). Dual class shares have been a powerful instrument, particularly when it was possible to apply differentiation of up to 1:10008 between A and B shares (when the so-called A-share carries 1000 times more votes than so-called a B-share). This kind of extreme differentiation was introduced in the 1920s by the Swedish tycoon, Ivar Kreuger, as a creative means to get access to foreign capital in companies such as Ericsson and SKF when Swedish law restricted foreign controlling ownership of Swedish corporations (Carlsson, 2007). For new firms, however, the maximum allowed voting-right difference is 1:10. In 2002, Ericsson is the only company still having a 1:1000 differential9 (Henrekson and Jakobson, 2003). In contrast to for example the US, dual class shares are frequently used in Sweden; Bergstrom and Rydgvist (1990) report that in the late 1980s over 70% of the firms listed on the Stockholm stock exchange issued dual class shares. However this percentage decreased over the following years and in 2003 about 55 per cent of listed firms used vote differentiation (Soderstrom et al., 2003). It seems reasonable to believe that this special characteristic of the Swedish stock market will have a certain impact on the performance of Swedish firms and create a special set of incentives that needs to be considered (Bjuggren et al., 2004). Firms with differentiated voting rights are found to have a systematically worse performance as compared to firms with one single class shares. Dual class shares drive a wedge

8_{The possibility of issuing shares with a 1:1000 voting differential was abolished in 1944, but the}

companies who already enjoyed this extreme voting differential were allowed to retain the old system.

9

(18)

between cash-flow rights and control rights. Not only does this change the control structure, but it also changes the incentive structure. Firms with only one equity class are on average investing efficiently, whereas firms with dual class share structure are over investing (Eklund, 2007). Since it may have an effect in the relationship under study, thus influencing the outcome of our research, we are going to control for the impact of voting differentiated rights too.

VI: Board of directors

The board of directors is one of several internal governance mechanisms that are intended to ensure that the interests of shareholders and managers are closely aligned, and to discipline or remove ineffective management teams (Barnhar et al., 1994). The tradition of strong owners in Sweden is reflected in the supervisory boards, which are quite strong (independent) vis-à-vis managers and have long been composed of mainly non executives (Thomsen et al., 2008). The board of Swedish companies, particularly of the large and listed ones, only includes nonexecutive directors. That goes also for the chairman of the board. In some companies, the CEO is also a member of the board, but increasingly not (Carlsson, 2007). The Swedish company act clearly states that the board is totally responsible for the management of the company. It is responsible for the strategy and resource allocation of the company, for appointing the CEO of the company, for delegating the operational responsibility of the company as well as for the monitoring and evaluation of the performance of the CEO (Carlsson, 2007).

(19)

of the day-to-day business of the company, while the supervisory board monitors, hire/fires and must approve all major decisions (Thomsen et al., 2008).

Another specific characteristic of the Swedish firm boards is the mandatory employee representation. Employees are allowed a one third representation and participate in the board work on equal terms with members appointed by the general meeting with the same duties, rights and responsibilities (Thomsen, 2008). The Law on Board representation (abbreviated LSA, in Swedish) came into force in the mid-1970’s. LSA regulates direct participation in the company decision process by representation on the board of the company (Carlsson, 2007). LSA regulates direct participation in the company decision process by representation on the Board of the company. Companies with more than 25 employees are obliged to allow two ordinary employee representatives on the Board and two deputies. Companies with more than 1000 employees are obliged to allow three ordinary representative and three deputies (Carlsson, 2007). The employee representatives have the same duties and responsibilities as Directors appointed by the AGM, i.e. they should serve the best interest of the company as a whole (Carlsson, 2007).

(20)

6.1. Board size

Most research has been focused on examining the relationship of board size and corporate performance. Jensen (1993) argues that boards with more than about 7 to 8 members are unlikely to be effective. According to him, large boards result in less effective coordination, communication and decision-making, and are more likely to be controlled by the CEO. However, because of the distinct feature of employee representation in the board10, the size of Swedish boards in general tends to be larger than other countries. A clear problem in studying board size is that the number of directors might arise endogenously as a function of other variables, such as company’s size (Core et al., 1999). Another issue is that most of the studies refer to the US and its distinct corporate governance characteristics. Despite the fact that most of the research has been conducted in the US, there is still some research for other countries. Eisenberg (1998) in a relevant study on a sample of Finnish firms reported similar findings. Similar finding were also reported from Mak and Kunsandi (2002) on a study of Singapore and Malaysian firms. The outcome of these studies provides evidence that the negative relationship between board size and firm value appears to be generalizable to environments with widely different corporate governance systems (Mak and Kunsandi, 2002). Thus, following the existing literature (e.g. Eisenberg, 1998; Yermack 1996) we would expect a negative relationship between board size and firm performance.

Now, as far as the relationship of board size and executive remuneration is concerned, existing literature supports the view that board size is positively related to executive remuneration. This consistent remark comes in contrast to the above argument that

10

(21)

board size negatively affects corporate performance. More specifically, Core et al., (1999) finds a positive board size-executive compensation relationship for a study on US firms, and the same goes for a study of Randoy and Nielsen (2002) for a sample of Swedish and Norwegian listed firms. The reason why this happens is due to ineffective monitoring end evaluation of the CEO from the board, as a result of bad corporate governance. This view is also in line with Jensen (1993), who supports that

“as groups increase in size they become less effective because the coordination and

process problems overwhelm the advantages from having more people to draw on”.

Following the above arguments, we expect board size to positively affect the level of CEO compensation, thus the following hypothesis can be stated:

H1:There is positive relationship between board size and the level of CEO

compensation

6.2. Anglo-American board membership

As already noted, many Scandinavian firms have been internationalized. They are listed in the largest and – according to the general perception11- most prestigious capital markets and/or have Anglo-American board members (Oxelheim and Randoy, 2005). The Anglo-American system is commonly regarded as the most demanding corporate governance system. The “superiority (in market performance) of the Anglo-American system is widely recognized12. The strict information requirements imposed by the Securities and Exchange Commission (SEC) provide further reasons for regarding the Anglo-American system as a good proxy for a “global governance model” (Oxelheim and Randoy, 2001). Past research suggests that a firm’s value depends on the quality of the monitoring and decision-making undertaken by its board

11_{See Economist, 2001; Lucier et al., 2004.} 12

(22)

of directors (Scleifer and Vishny, 1997). On the other hand compliance with the stricter information and monitoring requirements of a more demanding corporate governance system like the Anglo-American can substantially increase a firm’s costs. Because this can be an excess cost for many small or medium-sized firms, another way to improve the monitoring opportunities is by including foreign outsider members on the board. This alternative of “importing” a more demanding corporate governance system by having one or more representatives of that system as board members signals a higher commitment to corporate monitoring and transparency (Oxelheim and Randoy, 2001). The non-Anglo-American firms can thus signal its compliance with a more demanding corporate governance system (the Anglo-American) by recruiting an outsider Anglo-American board member (Oxelheim and Randoy, 2005).

(23)

security, the likelihood of shorter tenure, and a potentially negative impact on the reputation of the incumbent in case of dismissal (Oxelheim and Randoy, 2005). Thus we suggest that any Anglo-American influence in the internationalization process of the firm will also affect compensation schemes. Countries with a relatively low level of CEO compensation can expect to face higher compensation levels as their firms become exposed to markets with higher compensation levels (Oxelheim and Randoy, 2005). Overall, we expect Anglo-American board membership to have a direct effect on CEO compensation, thus we can form the following hypothesis:

H2: There is a positive relationship between Anglo-American board membership and

CEO compensation for Swedish listed firms.

VII: Data, methods, and description of variables

7.1. Data

(24)

data. In addition to the information on CEO remuneration, we collect data on various firm performance measures and key control variables like firm size and leverage. This required collection of balance sheet and income statement data (Duffhues and Kabir, 2008). All data were compiled from annual reports of firms, plus their web-sites, which proved to be a useful source during the data-collection procedure. Data on stock price performance are first collected from Datastream, and whenever necessary, complemented with data from OMX Nordic exchange group13.

Since firm size is related to the level of executive compensation (Gomez-Mejia, 1994), we narrowed down the scope of this research only to the firms listed in the large and mid cap indices of the Stockholm Stock Exchange. This led us to an original sample of 144 firms from all industries. Subsequently, 17 firms were excluded from our research. 5 firms were excluded because they published annual reports only in Swedish, and incomplete data was the reason why 6 more firms were excluded from our sample. 5 firms were excluded because their annual reports were not available. Usually only 2007 report was available for these firms, thus they had to be excluded from our research. Finally, 1 firm (Ericsson) was excluded, because it would be a giant firm in the sample, which could drive our results. Additionally, we have incomplete data on executive compensation for two firms (there are no data for 2005)14, and also we have incomplete data on stock performance for 18 firms. 3 firms from the large cap index and 15 from the medium cap index from all industries. We miss 1 year (2004) data on stock performance for 3 firms, 2 years data for 10, and 3

13

www.omxnordicexchange.com

14

Husqvarna AB, annual report of 2005 refer only to parent company, not to the group, thus data are not comparable.

(25)

years data for 5 firms. Finally, for one firm it was not possible to find information for its stock performance.

(26)

7.2. Methodology

In order to examine the pay-performanc relationship for Swedish listed firms, we followed Duffhues and Kabir (2008) selection of variables and methods. Using it as our reference point and building upon it, we enriched their model with additional dummy variables in order to control for the impact of the distinctive Swedish corporate governance characteristics in the relationship under study. We adopt the standard empirical literature on executive compensation and we use the following regression model to test the performance between CEO remuneration and firm performance:

Pay

it

= α

0

+ α

1

Perf

it

+ α

2

Size

it

+ α

3

Lev

it

+ λ

it

+ δ

it

. (1)

The dependent variable (

Pay

it

)

is the amount of compensation paid to the CEO of

firm i in period t. The explanatory variable (

Perf

it) refers to the performance of firm i

in period t. Firm size (

Size

) is measured in two ways, as the logarithm of total assets, and the logarithm of total sales. We use ownership (

λ

_{) dummies, so as to control for}

the impact of large block holder ownership in the examined relationship, and finally we use dummies to control the impact of corporate govenance characteristics (δ) on CEO compensation. So, following Diffhues and Kabir (2008), if executive compensation is predicted to be higher in companies with higher performance, then the estimated regression coefficient

α

₁_{should be positive.}

(27)

Basu et al., 2007; Duffhues and Kabir, 2008) we perform additional regressions for each industry seperately15 so as to control for industry-specific effects on pay level.

7.3. Variables

7.3.1. CEO Compensation

The dependent variable is the amount of compensation paid to the CEO of each firm. We construct three proxy measures for CEO compensation:

1. Basic

o It includes only the basic salary and the amount of cash benefits

directly paid to the CEO during the year.

2. Variable

o It refers to the amount of variable compensation paid to CEOs as a

reward for achieving the short term goals of the firm. Variable

compensation is based in previous year’s performance, since the bonus

paid to the CEO of a company i in year t is refers to corporate goals

achieved in the year t-1.

3. Total compensation (Totalcomp)

o Is the sum of the above two variables, plus all other type of

remuneration paid to CEOs as included in companies’ annual reports.

This usually includes, extra benefits (e.g. company car, school fees for

the CEOs children etc.), other financial instruments (usually not

specified in the annual reports), pension costs, long-term variable

compensation plan (usually for all employees or the executive

members), and in a few cases perquisites.

15

(28)

By substituting

Pay

it with the three CEO compensation proxies we form the

following regressions:

Basic

it

= α

0

+ α

1

Perf

it

+ α

2

Size

it

+ α

3

Lev

it

+ λ

it

+ δ

it

. (2)

Variable

it

= α

0

+ α

1

Perf

it

+ α

2

Size

it

+ α

3

Lev

it

+ λ

it

+ δ

it

. (3)

Totalcomp

it

= α

0

+ α

1

Perf

it

+ α

2

Size

it

+ α

3

Lev

it

+ λ

it

+ δ

it

. (4)

7.3.2. Firm performance

The explanatory variable, firm performance is measured in a number of ways. We consider both accounting-based and capital market-based performance measures (Duffhues and Kabir, 2008). Investors and scholars often have difficulties assessing a company’s performance, due to the fact that there is a wide range of instruments to determine how a company is performing. These instruments are ratios, which can inform investors about the relative competitiveness of a firm and its underlying stock. A common distinction made among these instruments, is between accounting-based measures which looks at the firms’ financial results, and market-based which focus on stock performance. As already mentioned, we take both under consideration following the path of previous researchers on this topic (Kato and Kubo, 2006; Basu et al., 2007; Duffhues and Kabir, 2008).

7.3.3. Accounting-based measures

(29)

This ratio is used to evaluate a company’s operational efficiency, and indicates how much profits are created over its total sales. Several prior studies use ROA or ROS as direct estimates of profitability. In various studies, this ratio has been used as a proxy for performance (Ramamurti, 1987). Another reason why researchers prefer ROS, is because operating income is subject to less managerial discretion than net income and as dividing it by sales it is less susceptible to manipulation than total assets (Firth et al., 2006). Since theory suggests that executive compensation should be linked to firm performance (Jensen and Meckling, 1976), we would expect that firm with high ROS pay their CEOs more. Thus, we can form the following hypothesis:

H3: There is a positive relationship between ROS and CEO compensation for Swedish

listed firms.

The second variable, return on assets (ROA), gives an idea of how efficient management uses its investment base to generate income, and is defined as Net Income (before interest and taxes) over Total Assets:

(30)

CEOs will receive a reward for achieving a high ROA ratio. So the following hypothesis can be stated:

H4: There is a positive relationship between ROA and CEO compensation for Swedish

listed firms

7.3.4. Market-based measures

The variable we are using is annual stock return (RET), which measures the growth of the stock price during the year. Empirical finance literature make a extensive use of stock returns (Acker and Duck, 2006), which are calculated by looking at the daily stock price increase or decrease, measured in a percentage. Eventhough a lot of studies use monthly stock returns or sometimes even the difference of stock price at the beginning of the year and they compare it to the price at the end of the year, we decided to use daily observations so as to secure the robustness of our findings. According to Acker and Duck (2006) the choice of reference day can seriously affect the estimated statistical properties of returns, thus by using daily observations, we diminish this impact. Then the average return is calculated by taking the cumulative daily returns, and dividing it by the number of days in the observed period. In developed nations, accounting profit and stock returns are two of the major indicators of a company’s financial performance (Firth et al., 2006). The executive compensation literature16 suggests that compensation should be related to measures of stock-based-performance, not only because this is desired by shareholders, but also because high stock returns send positive signals regarding the actions taken by managers (Fernandes, 2008). This remark leads as to the formation of the following hypothesis:

16

(31)

H5: There is a positive relationship between Stock return and CEO compensation for

Swedish listed firms

The final proxy for firm performance is Tobin’s Q, or just Q (TOBINQ). Tobin’s Q is probably the most widely used valuation measure in empirical corporate finance (Drobetz, 2003), and it is a hybrid of accounting and capital market-based measure, defined as the ratio of the sum of market value of common shares and book value of debt to book value of total assets (Duffhues and Kabir, 2008):

Q is used as a proxy for firm value in several studies (e.g. Mehran, 1995). According to Drobetz (2003) a high Tobin’s Q ratio indicates that the firm has done well with its inestment decisions, thus, it can be reasoned that a high Q ratio will be positively evaluated by the board, resulting a higher reward for the CEO and following the existing literature, we can form the following hypothesis:

H6: There is a positive relationship between Tobin’s Q and CEO compensation for

Swedish listed firms

Whether Q and ROA are correct proxies for firm performance is a subject of continuing debate in the literature. In addition to well-known measurement problems associated with Q, many have argued that Q is a better proxy for the firm’s growth opportunity than its performance (Mehran, 1994). On the other hand, as these four variables have been widely used in investigating corporate performance (Duffhues and Kabir, 2008), thus we assume that they are the most appropriate for this type of research. All value variables are adjusted for inflation using CPI17 (2007=100) and are thus expressed in 2007- constant SEK prices.

17

(32)

7.3.5. Control Variables

In order to secure the robust of the findings of this research, a number of control variables have been included in our empirical design.

Firm Size

We include firm size (SIZE) as a control variable because CEO pay has been shown to be directly related to firm size (Murphy, 1999). Actually, the most consistent and enduring result from myriad studies of CEO pay is that firm size is positively and significantly associated with compensation level (Conyon et al., 1997). We use two proxy variables for firm size, the logarithm of total sales (LOGSALES), and the logarithm of total assets (LOGASSETS). Both variables have been used extensively in related research.

Leverage

Another key control Variable we are using is Leverage. It is defined as the ratio of total debt to total assets:

The reason why we use leverage as a control variable, is because debt holders may closely monitor managerial activities thereby reducing the payment of excess compensation. On the other hand, higher debts can also lead to an increase in firm risk, which in turn necessitates the payment of higher compensation (Duffhues and Kabir, 2008).

(33)

Other firm characteristics variables

We will control for the impact of strong block holders on corporate performance, by using a dummy variable equals 1 in cases where the firm has a block holder and 0 if this is not the case. We control for block ownership at 5% 10% and 20%18 level for the single largest shareholder19.

Corporate governance variables

We measure board size (BSIZE) as the number of its members including also the employee representatives. An important aspect in this study following Oxelheim and Randoy (2005), is the impact of foreign board membership in the level of CEO pay (ANGLO=1 if there is at least one Anglo-American member in the board; 0 otherwise). We use two more dummy variables, which refer to country-specific corporate governance characteristics. The first one refers to a common practice of Swedish firms, where the CEO is also a member of the board (CEOBOARD=1 if the CEO is also member of the board; 0 if not). Eventhough only one firm reported direct fees to its CEO and all other report that their CEOs receive no extra remuneration for participating in the board, still some may argue that the presence of the CEO as a member could influence other board members, thus we would like to control if this participation in the board has an effect on the CEO pay level. Our final variable controls for the effect of dual class shares stocks on firm performance (DUALCLASS=1 if a company has differentiated voting rights; 0 otherwise).

VIII: Empirical Analysis and Results

18_{Due to the separation of ownership and control through dual class shares, we control for the largest}

owner, in terms of voting control and not equity.

19

(34)

8.1. Descriptive Analysis

Table 1, presents yearly summary statistics of CEO compensation for the sample of Swedish listed firms. In order to allow a meaningful comparison over time, all

monetary values are expressed in 2007 constant SEK prices. We observe that in 2005,

Table 1. CEO compensation in Sweden

The table shows the amount of compensation paid to the CEOs of Swedish listed firms. Panel A presents data for base salary (plus benefits directly paid to CEOs), Panel B presents data for the variable part of compensation paid to CEOs, usually based on achieving short-term targets. Panel C reports those for total compensation (base salary, plus variable remuneration, plus any other type of remuneration is reported in companies’ annual reports). Compensation is expressed in thousands of SEK, in constant 2007 price after adjusting with CPI.

the firms in our sample paid its CEOs on average 7,931 thousand (t) SEK (median = 5,555). From this amount 4,073 tSEK (median = 3,274) refers to base salary and benefits and 1,686 tSEK to the variable component of total compensation paid to CEOs. In 2006 there was a sharp increase in CEOs total remuneration that amounted to 18.8%. In 2006, total compensation was 9,426 tSEK (median = 6,282), with base salary and benefits amount for 4,445 tSEK (3,501) and variable compensation 2,227 (1,172). As we observe, the high increase in average total compensation is highly attributed to a sharp increase in the variable remuneration component wich equals to

Average Median No. of obs.

Panel A: Base salary

2005 4,073 3,274 125 2006 4,445 3,501 127 2007 4,915 3,750 127 Panel B:Variable 2005 1,686 930 125 2005 2,227 1,172 127 2007 2,120 1,320 127

Panel C: Total comp.

2005 7,931 5,555 125

2006 9,426 6,282 127

(35)

32.08%. In 2007 there was a decrease in the amount of variable compensation equal to 4.80%. On the other hand, the amount of total compensation increased by 6.04% reaching 9,995 tSEK (7,071). As for the amount of base salary and benefits, we observe a relative stable but rather high increase of around 10% for both 2006 and 2007. In general, the lower values of the median reflect the fact that the sample consists of a few firms paying relatively higher amount of compensation to their CEOs.

(36)

Table 2. Summary statistics

The table shows descriptive statistics of firm performance for the sample of Swedish listed firms. The table shows the four performance metrics: return on assets (ROA), return on sales (ROS), annual stock return (RET) and Tobin’s Q (TOBINQ). Results are presented for both contamporaneous and lagged relatioships.

Other variables of interest are those who refer to firm characteristics, which are presented in Table 3. For contemporaneous relationships, the average value of total sales for the firms in our sample is 33,229,536.73 tSEK (median = 5,402,500.00). The mean for total assets is 88,974,657.24 (8,079,840.51) tSEK, and as for the market value of equity is concerned, the mean is 19,912,557.71 (2,718,170.00) tSEK. When assuming a lagged pay-performance relationship the average total sales for the firms of our sample is 28,669,427.22 tSEK (median = 4,477,510.75). The mean for the value of total assets in tSEK is 78,020,969.64 (6,198,605.00) and for the market value of equity is 17,448,963.39 (2,368,750.95). The great deviation of the mean from the median is due to the large firms of our sample which drive up the average values of our variables.

Mean Median S.D No. of obs.

(37)

Table 3. Key Firm Characteristics

The table shows descriptive statistics of key firm characteristics variables. The tale reports summary statistics for total sales, total assets and the market value of equity for our sample. Results are presented for both contamporaneous and lagged relatioships. Values are expressed in thousand SEK and in constant 2007 prices.

8.2. Regression Analysis

Table 4 presents results for the three proxies we use to measure CEOs remuneration, assuming that the pay-performance relationship is contamporaneous. The results of regressions show that CEO remuneration is significantly negative related to return on Sales. For the other performance measurements our model failed to identify any other statistically significant relation. Both proxies we use for firm size (logarithm of assets and sales) are found to be positive and highly significant at the 1% level, supporting the well-established finding in executive compensation studies, that firm size plays an important role in executive compensation. Moreover, we find a significant positive relationship between Anglo-American board membership and executive compensation, which comes in line with Oxelheim and Randoy (2005), which reported similar findings in their study on a sample of Swedish and Norwegian firms.

Mean Median No. of obs.

(38)

Table 4. Contamporaneous Relationships

T-statistics are presented in the table. Standard errors are reported in the parentheses. ***, **, * statistically significant at 1%, 5%, and 10% respectively. The dependent variable is the three proxies used to measure CEO compensation. Panel A refers to base salary plus benefits, Panel B to variable compensation, and Panel C to the amount of total compensation paid to CEOs.

Panel A Panel B Panel C

CONSTANT -8.19*** (1180.10) -3.68*** (1756.38) -7.17*** (3983.60) BLOCK5% 1.34 (396.96) 0.60 (1147.84) -0.42 (2567.26) BLOCK10% -1.04 (378.80) -0.02 (1092.96) -0.40 (2449.20) BLOCK20% -0.54 (357.06) 0.06 (1076.21) -0.64 (2421.98) ANGLO 7.71*** (240.14) 2.77*** (292.59) 6.78*** (662.67) BSIZE 3.59*** (56.20) -2.15** (68.92) 2.40** (153.14) CEOBOARD -2.84*** (243.80) 0.19 (300.61) 0.06 (673.40) DUALCLASS 4.09*** (234.80) 1.74* (287.15) 1.80* (642.78) LEV 0.767 (583.68) -1.76* (706.99) 0.77 (583.68) LOGASSETS 3.98*** (237.26) 2.66*** (290.77) 3.75*** (652.47) LOGSALES 3.34*** (254.81) 2.18** (308.42) 3.32*** (698.54) RET 0.48 (1565.05) 1.34 (5.74) 0.63 (12.93) ROA -1.20 (4.67) 0.50 (1908.65) 0.74 (4254.65) ROS -2.877434*** (63.96) -1.84* (78.41) -2.64*** (172.77) TOBINQ -0.737399 (583.91) -1.67* (714.12) 1.21 (1609.06) Adjusted R-squared 0.497988 0.130364 0.409549 F-statistic 32.41279 5.747017 22.96461

(39)

This outcome comes to support Oxelheim and Randoy’s arqument, that the presence of Anglo-American members in the board has a positive effect on the level of CEO pay. We also find a positive relationship between board size and CEO compensation, a finding consistent with previous studies, like that of Randoy and Nielsen (2002), and this finding supports the view that large board become less effective due to coordination and process problems, leading to ineffective monitoring and evaluation of the CEO from the board. An other remarkable outcome is the fact that firms which have differentiated voting rights pay their executives more, and this finding comes to support those who argue that dual-class shares and the seperation of ownership and control lead to increased agency costs. Our model does not identify any other statistical significant relationship with the other variables we used in our research. Block holder ownership does not have an effect on the pay-performance relationship for our sample. For the cases where CEO is also a member of the board we find a statistical signficant negative relationship only for one of our proxies, that of the basic salary, but is not statistical significant when we refer to the total amount of compensation paid to CEOs. The truth is though, that this came to contrast what we expected, since we expected that when CEOs are also member of the board it is easier for them to influence the board in order to receive higher compensation. As far as the other control variable we used, leverage, we didn’t find a statistical significant association, assuming that the pay-performance relationship is contamporaneous, but we find a significant negative association when assuming that the pay-performance relationship is lagged.

(40)

T-statistics values are presented. Standard errors are in parentheses. ***, **, * statistically sigificnt at the 1%, 5%, and 10% respectively. Besides the regression for the full sample, we performed regression for each industry seperately. For the sake of brevity the tables of these regressions are not presented in this section, but they are included in the appendices.

Table 5. Lagged Relationships

Panel A Panel B Panel C

CONSTANT -0.36 (1426.50) -0.16 (1782.15) -0.19 (4070.70) BLOCK5% 1.48 (378.37) 1.61 (472.70) 0.06 (2605.56) BLOCK10% -0.01 (867.99) -0.29 (1078.95) -0.14 (2494.09) BLOCK20% 0.49 (239.66) 0.01 (299.41) -0.51 (2451.37) ANGLO 8.20*** (231.93) 3.00*** (289.76) 6.94*** (664.17) BSIZE 3.95*** (56.05) -1.38 (70.03) 2.62*** (169.26) CEOBOARD -1.96* (240.83) -0.48 (300.88) 0.15 (687.18) DUALCLASS 4.48*** (228.56) 2.01** (285.55) 2.38** (652.23) LEV -4.76*** (1795.83) -2.25** (2243.56) -4.60*** (5648.92) LOGASSETS 3.74*** (235.94) 2.76*** (294.76) 3.63*** (673.43) LOGSALES 3.10*** (247.41) 1.23 (309.10) 3.88*** (669.92) RET -1.22 (4.70) 1.90* (5.87) 0.96 (13.24) ROA 1.02 (1395.13) 0.90 (1742.96) 0.60 (3981.57) ROS -1.73* (70.47) -1.74* (88.05) -1.86* (200.74) TOBINQ -5.58*** (1712.51) -2.99*** (2139.47) -4.71*** (5433.21) Adjusted R-squared 0.527873 0.145983 0.404423 F-statistic 33.68217 5.996618 20.84894

(41)

(42)

endogenous function of the firm which is influenced by other important parameters besides performance. Finally an important finding of our study has to do with the part of variable remuneration paid to the CEOs of Swedish listed firms. Usually most studies examined the pay-performance relationship using one proxy, the whole amount of compensation paid to CEOs and they don’t account for its components as we do in our research. By focusing on its components, and more specifically at the variable component, which refers to the extra compensation paid to CEOs in the form of bonuses for achieving short term corporate goals and which is directly related to previous year’s corporate performance, we managed to identify an important relationship. Our findings indicate a statistical positive relationship between stock performance and variable compensation, which means that market-based goals are more important than accounting-based targets in determining variable compensation for the CEOs of Swedish listed firms. In figure 1, a list of the hypothesized relationships and the empirical result are presented in a visual summary of our key findings.

Figure 2: Results of Hypotheses

*The hypothesized relation was statistical significant on the opposite (negative) direction.

**The hypothesized relation was statistical significant only for the variable part of CEO compensation. ***The hypothesized relation was statistical significant on the opposite (negative) direction.

Hypothesized Relation

Result

H1:There is positive relationship between board size and the level

of CEO compensation Supported

H2: There is a positive relationship between Anglo-American board

membership and CEO compensation Supported

H3: There is a positive relationship between ROS and CEO

compensation for Swedish listed firms. Not Supported* H4: There is a positive relationship between ROA and CEO

compensation for Swedish listed firms Not Supported H5: There is a positive relationship between Stock return and CEO

compensation

Partially Supported**

H6: There is a positive relationship between Tobin’s Q and CEO

(43)

In order to secure the validiy of our findings a number of robustness checks were introduced. We performed the same regression again, only this time for each performance variable separately, but this did not seem to have an effect in our findings. Additionally we performed an analysis by controlling for fixed effects, in order to control for unobserved heterogeneity in our data. In this case, controlling for fixed effects prooved to have an impact in our model. Eventhough the relationship with the performance metrics remains the same, two variables, the impact of Anlo-American board membership and Dual-Class shares loose their statistical significance when we control for fixed effects in our model.

(44)

IX: Limitations and recommendations for further research

The most important limitation of this research, is that it ignores perharps the most important component of incentive compensation, namely the value of stock options. Unfortunately the market value of stock option is not disclosed by most firms, nor was it possible for us to estimate it. Some studies have measured the value of stock options by using the Black-Scholes formula20, but because of its complexity it was not possible for us to use it. Perhaps, an other limitation is that our sample consists only from medium and large sized companies, but due to the scope and the time constraints of this research it was not possible to include small cap firms into our sample.

Drawing in the above limitations, it would be very interesting to perform an analysis including the value of the stock options, and also enlarging the sample including all listed companies on the Stockholm stock exchange so as to draw safer conclusions upon this topic. Moreover in order to control for endogeneity, which seems to be the most important constraint in executive compensation research, non-financial performance measures should be included in the empirical model, so as to draw more reliable conclusions. One other interesting topic for further research could be to examine not if executive compensation is determined by firm performance, but whether there is a convergence in the paying levels and practices for executives among European countries, as an outcome of the increased internationalization of firms and its effects on all corporate governance aspects, including executive compensation.

X: Conclusions

20_{Black, Fischer; Myron Scholes (1973). “The pricing of Options and Corporate Liabilities”. Journal of}

(45)

This paper has provided an empirical analysis on the pay-performance relationship for CEOs of Swedish listed firms. This study, contributed to the growing literature on executive compensation by analysing data from Sweden, a country for which relevant research is rather limited. Contrary to the popular belief that executive compensation should be determined by corporate performance, the most robust finding of our research indicates a negative relationship between CEOs remuneration and Tobin’s Q, as well as return on sales. Although it appears to be an intriquining finding, our study is broadly in line with the vast amount of work in the finance and management literature that fails to find a positive relationship between executive compensation and firm performance (Duffhues and Kabir, 2008). This comes to support those who argue that executive compensation is an endogenous function of the firm, thus additional non-performance related factors influence the levels of executive compensation in most firms. On the other hand, our findings indicate a positive statistical relationship between return on stock and the variable part of CEOs compensation. This finding leads to the conclusion that market-based goals are more important than acounting-based targets in determining the amount of variable reward paid to CEOs. This outcome is in line with the existing executive compensation literature, which suggests that compensation should be related to measures of stock-based-performance, not only because this is desired by shareholders, but also because high stock returns send positive signals regarding the actions taken by managers (Fernandes, 2008).

(46)

Acknowledgments

I would like to thank my supervisor Jann Grunberg for his contribution and guidance in order to complete this paper. Also my referent Niels Hermes, the person who introduced me into the field of empirical research, for his comments and assistance in order to finish my paper. The paper is dedicated to my parents.

References

Acker Daniella, and Duck Nigel W. (2006). “Reference-Day Risk and the Use of Monthly Returns Data: A Warning Note”, Working Paper.

Angblad J., Berglof E., and Svancar H. (2001). “Ownership and Control in Sweden”, The Control of Corporate Europe, Oxford University Press, Oxford.

Antle Rick, and Smith Abbie. (1986). “An Empirical Investigation of the Relative Performance Evaluation of Corporate Executives”, Journal of Accounting Research, 24: p. 1-39.

Barnhart S.W., Marr Wayne M., and Rosenstein S. (1994). “Firm Performance and Board Composition: Some New Evidence”, Managerial and Decision Economics, 15: p. 329-340.

Barkema Harry G., and Gomez-Mejia Luis R. (1998). “Managerial Compensation and Firm Performance: A General Research Framework”, Academy of Management Journal, Vol.41: No 2, p. 135-145.

Basu Sudista, Hwank Lee-Seok, Mitsudome Toshiaki, and Weintrop Joseph. (2007). “Corporate Governance, Top Executive Compensation and Firm Performance in Japan”, Pacific-Basin Finance Journal, 15: p. 56-79.