R ATIOS AND CEO C OMPENSATION F AIR V ALUE , F IRM P ERFORMANCE

(1)

M

ASTER

T

HESIS

F

AIR

V

ALUE

,

F

IRM

P

ERFORMANCE

R

ATIOS AND

CEO

C

OMPENSATION

An Investigation of the Association between Use of Fair Value and Firm

Performance Ratios and its effect over CEO Compensation, in Sweden

Author: Öznur Uyanık

Supervisor: Fredrik Karlsson

Examiner: Andreas Jansson & Fredrik Karlsson

Term: ST’17

(2)

Abstract

The purpose of this study is to test the relation, if any, between use of fair value measurement and reported performance ratios of the firms and finally its effect on CEO compensation in Sweden. This research aims to contribute knowledge of decision makers about the performance-pay link in Sweden by comparing the changes of CEO compensations across the years, before and after the use of fair value method. In this sense, firms divided in two models of ownership structure: firms with family concentration and firms with dispersed ownership structure. This paper intent to contribute the explanations of existing researches from the USA and China about exponential increase in CEO compensation after the use of fair value method, with the data of Swedish listed companies. The data set of this study was highly dependent of the accessibility of information. In this sense, this research can be contributed with different data set, with more detailed scrutinise of data and in a longer research period.

Acknowledgements

Having completed this master thesis in Linnaeus University in 2017, I would like to thank several individuals. I would like to start by thanking Andreas Jansson, Associate Professor of Accounting and Logistics at Linnaeus University, who inspired me about the subject and my supervisor Fredrik Karlsson, Senior Lecturer of Accounting and Logistics at Linnaeus University, for his sincere assistance, guidance and immensely valuable criticisms.

I owe special thanks to Ozan Uyanık for proofreading my paper, for his never-ending support and his whole year patience and endurance at times of stress. Another special thanks to my friends Ayşegül Girgin Ring and Şeniz Yılmaz for their valuable advice in econometrics and statistics.

Key words

(3)

1. Introduction

This part begins with a brief background about the amendment in measurement methods from historical cost to fair value method in Sweden because of its revolutionary approach to income recognition regardless of being realized or unrealized. This is followed by a problem discussion about the importance of CEO compensation by proceeding from agency theory and the reasons why studying this subject is interesting, which lead the reader to trace to the purpose of this research.

1.1 Background

(5)

According to Brown (2011), the benefits of this recognition and implementation within EU listed companies would be:

“. . .to prepare their consolidated accounts in accordance with IAS from 2005 onwards. . . will help eliminate barriers to cross-border trading in securities by ensuring that company accounts throughout the EU are more reliable and transparent and that

they can be more easily compared. This will in turn increase

market efficiency and reduce the cost of raising capital for companies, ultimately improving competitiveness and helping boost growth.” (EU statement issued in Brussels on 7 June 2002 - Brown (2011) p.271)

However, use of historical cost method was an obstacle in front of creating comparable financial information. According to conceptual framework of IFRS (2010), historical cost method based on the cost which occurred at the time of acquisition and provides monetary information about resources, claims and changes of them with using information about past transactions. In this sense, the value on the financial reports reflects the entity specific measurement. Here the cost, includes all costs necessary to bring the asset on working conditions for its intended use such as original purchase price, costs of site preparation, delivery and handling, installation and others related to entity specific costs.

On the other side, according to IFRS13, Fair value is the price which is supposed to be received when an asset is on sale or paid to transfer a liability at the measurement date. Fair value method does not give a firm specific information, rather market based. It provides opportunity to make comparison between firms in the market as well as between different time periods of one certain firm.

(6)

Since 1975 fair value have become one of the most debated benchmark in terms of revaluation with its consistency, reliability, comparability and its role in financial crises (Pozen, 2009). In this sense, one of the reform in accounting world with the implementation of IFRS in EU is, the allowance of use of fair value measurement method (Peterson&Plenborg, 2012).

In his speech at IASB in 2015, Hoogervorst highlights some points from the eyes of fair value method’s fans:

“The fans of fair value like it for the very reason that it does require a full update of all inputs at each reporting date. They believe this gives the most meaningful picture of the financial position and performance of an entity. They recognise that fair value may lead to volatility in the income statement, but they believe that to be an accurate reflection of economic reality.” (Hoogervorst, 2015, p2-3).

The obligation of adopting IFRS imposed by EU commission, for consolidated financial statements of listed companies (Guggiola, 2010). In Sweden, this adoption has been implemented partially; listed companies consolidated accounts had been required to be in full compliance with IFRS while individual accounts have not been required (ibid). Guggiola (2010), explains these limitations of use in some countries as a result of rigidity in financial reporting and taxation such as in Sweden and in Germany. Additionally, the enforcement responsibility left in each countries’ realm of authority so the law (Lag (2007:528) om värdepappersmarknaden) was issued in Sweden and according to this “Finansinspektionen” (FI), appointed as the responsible enforcement organ (Frejinger&Johansson, 2009).

(7)

the more comprehensive disclosures of fair value, the less certainty of information will be achieved. Those challenges of financial statement analysis are not new, however, the use of “managerial judgements” provided by fair value method may have a large impact on firm earnings and financial ratios (Peterson&Plenborg, 2012). Since the changes in measurement methods may affect the calculation of many financial performance ratios, and those ratios may affect the managerial contracts, it plays a crucial role for managerial choice of accounting.

The managerial tendency to use profit increasing accounting policies in order to increase their compensations is also discussed by many other researchers (RW Holthausen, 1990, Watts&Zimmerman, 1990, Godfrey et al, 7th ed., 2010, Peterson&Plenborg, 2012). So, following section discusses the underlying logic of managerial accounting policy preference and links this with compensation contracts.

1.2 Problem Discussion

According to Godfrey et al, (7th ed., 2010) accounting history consists of three periods; 1800-1955 is labelled as “Pragmatic Accounting” era where the explanations of accounting mainly based on the practice of accounting. Between 1956 and 1970 is labelled as “Normative Period” where accounting theorists attempt to develop best accounting practice and focused on what accountants should report. Finally, the “Positive Era”, which we are living in, have started at the early beginning of 1970s. The most attractive part of this era is, its link to agency theory. This link mainly explains the managerial logic of producing financial information and its affiliation with managerial remuneration. The choice of accounting policies in order to report maximum profit is also linked to managerial intentions to increase their compensations (ibid). In this sense, converting from historical cost method to fair value may play a crucial role.

(8)

According to fair value method entities recognise the unrealized gains and losses in their profit or loss in the period that they occurred (ibid). Penman (2007) utilizes the income definition of Hicksian which considers any wealth increase as income. In this sense, Penman (2007), asserts that the fair value measurement brings solution to income measurement since it books any increase or decrease as income or expense.

However, this approach is criticized by cost model proponents since it violates the conservatism principle (Godfrey et al, 7th ed., 2010). Additionally, physical capital proponents argue that the firms purchase their assets to be used in operation so the liquidation of them is unrealistic and the price changes of them are irrelevant for profits (ibid). At this point, one may claim that the managers aim to increase profit and may benefit this aspect of fair value method in order to cover underlying economic reality of the firm and increase their compensation.

On the other side, some discussants claim that under fair value method, the amount of depreciation is greater so the profit and dividend do not become fictive (Godfrey et al, 7th ed., 2010). However, according to conceptual framework of FASB (Financial Accounting Standard Board), fair value poses a superior consistency to historical cost except when the issue comes reliability (Hermann et al. 2006). At this point, agency problem rises since the financial information produced by managers and fair value method provides a gap to managers to be opportunistic.

The agency theory is developed by considering the firms where the security ownership is separated from its control and management. Here the risk is that the agent may not work for the best interest of the principals so there is always a risk that the agent may display opportunistic behaviours to increase his/her own wealth (Jensen & Meckling 1976).

(9)

et al, 7th ed., 2010). In this sense, agents tend to have negotiations to align their interests with principals (ibid).

In order to decrease the severity of agency problem it is suggested to tie managerial remuneration to the share prices (Godfrey et al, 7th ed., 2010). However, this obviously favours the principals’ interests (ibid). Furthermore, the sectoral situations or other impacts, which are not in the realm of managerial competence, may affect the firm value so managers will likely to refuse this remuneration plan (ibid). In this sense, it is suggested that the managerial remuneration may be linked to the firm profit which mainly use the accounting profit (Watts&Zimmerman, 1990). This kind of remuneration consists of two parts; fixed salary and bonus, which depend on a particular profit target and being over the target (Healy, 1985). As a result of this, the calculation of profit and the selection of accounting policies strongly attract the agent’s attention.

In this sense, use of fair value method has been discussed frequently due to its implementation with estimations of fair values (Zeff 2005, Pozen, 2009, Godfrey et al, 7th ed., 2010, Peterson&Plenborg, 2012). According to IFRS 13, implementation of fair value leaves a gap for managerial subjective revaluation with its ‘hierarchy' and recognizes unrealized gains, which affect the profit of the year and the CEO compensation (ibid). The combination of; existence of the subjectivity within fair value method and the reality of that the future managerial contracts will be decided by depending on the financial statement which were prepared with use of fair value method, attracts attention whether fair value method affected the managerial compensation increase or not.

(10)

of China International Marine Containers tenfold in 2008 (ibid). In this sense, it gains importance to detect how much change have been occurred at those performance ratios and CEO compensations after the use of fair value in Sweden.

On the other side, existing researches from the USA reflecting the typical Anglo-American dispersed and well-protected ownership structure under common-law tradition (Aguilera & Jackson, 2010, La Porta et al., 2008). Additionally, researches from China reflect its own complex ownership structure with civil-law origin (Liu, 2015). In Chinese listed firms, state holds considerable number of shares and plays the role of controlling shareholder especially in crucial industries (ibid). Furthermore, in those firms, state appoints CEOs and uses remuneration contracts based on financial and non-financial measures unlike the Anglo-American remuneration plans which based on stock performance (ibid).

Sweden differs from those countries with its prevailing market feature of concentrated ownership structure under civil-law tradition (La Porta et al. 1998). One of the important differences between them are; the lower minority shareholder protection in Sweden and secondly, the big owners may play a significant role at the corporate governance in the Swedish listed firms such as being CEO or being a member of the board of directors (Högfeldt 2005, La Porta 1998). In this sense, this research will be implemented by considering two models of ownership structure: firms with family concentration and firms with dispersed ownership structure. By doing this, prevailing ownership features of Swedish market will have been specified.

1.3 Purpose

(11)

This paper also intends to contribute explanations of existing researches about exponential increase in CEO compensation in the USA and China with the data and results of Swedish listed companies.

1.4 Disposition

In this section, the chapters of this study will be presented with a brief explanation of each chapter's content along the main points.

Introduction

:

This part begins with a brief background about the changes in measurement methods from historical cost to fair value method in Sweden. This is followed by a problem discussion about the importance of CEO compensation by proceeding from agency theory and the reasons why to study this subject is interesting, which lead the reader to trace to the purpose of this research.

Method:

This chapter will introduce the research approach of this study to achieve the purpose. The section consists of research approach and the presentations of the statistical analysis which will be implemented.

The theoretical framework:

This chapter begins with a discussion about agency theory and followed by presentations of fair value method, CEO compensation plans and performance ratios. This section finishes with the Literature Review.

Hypotheses:

This section begins with the presentations of hypotheses under two different ownership structures; Firms with Family Concentration and Firms with Dispersed Ownership Structure. Hypotheses are derived on the basis of the existing research and theories which will then form the basis for the quantitative research.

(12)

Examination of Hypotheses:

In this section, the hypothesis of this study will be analysed and the findings will be discussed in accordance with the arguments which were used to develop those hypotheses.

Concluding Discussion:

The chapter begins with a brief summary of the study and continues with its conclusion which answers the purpose. The limitations of this study as well as suggestions for further research will be discussed.

2. Method

This chapter will introduce the research approach of this study to achieve the purpose. The section consists of research approach and the presentations of the statistical analysis which will be implemented. The hypotheses of this research are developed in further section.

2.1 Research Approach and Empirical Method

The following chapters were written to make a presentation and reasoning of the research design of this study.

2.1.1 Deductive Approach

According to Saunders et al. (2009), deductive research approach gives opportunity for falsifying. It is also for testing interactions while inductive research approach is for exploration. They explain this point by depending on the explanation of Yin (2003):

“…based on theoretical propositions drawn from appropriate literature you specify a number of related outcomes (dependent variables) that you expect to find as a result of the implementation of a particular change management programme (independent variable) in an organisation where you intend to undertake research.” (Saunders, Lewis and Thornhill, 2009, p.500)

(13)

2.1.2 Quantitative Method

Saunders et al (2009), assert that the use of quantitative analysing techniques such as graphs, charts and statistics, is useful to examine relations and trends. In this sense, quantitative approach has been chosen for this research.

Quantitative method is appropriate for the subject of this research. However, while quantitative method is giving more objective and generalizable results, it is omitting the individual appraisals as an underlying effect of the result (Bryman &Bell 2005). This study aims to extract generalizable results so individual appraisals of board of directors will not be scrutinized.

Alternative of quantitative method would be qualitative method which is more related to interpretation of individuals’ behaviour by depending on individuals’ answers in a query (Bryman &Bell 2005). This study will be written by avoiding subjective interpretations. However, reader should be aware of that 100% objectivity is not possible.

2.1.3 Longitudinal Study

This research aims to study changes in performance ratios and CEO compensations over a time period of 2000-2010. According to Saunders et al. (2009), data can be collected by aiming a particular time’s results or changes over a period. Bryman & Bell (2005) say that longitudinal study can be used for the researches which will scrutinized data over years.

2.1.4 Trend Analysis

Saunders et al. (2009) recommend drawing a line or graph to display and obtain the trend, with examination of longitudinal data. Not only they are highlighting three common uses of this analysis, but also trend analysis can be used to make other supplementary statistical analysis (ibid).

(14)

Furthermore, Saunders et al. (2009) also draw a path to make this analysis meaningful when it is applied to multiple variables which are aggregated from different units and displaying different values. According to this:

“Longitudinal data may necessitate selecting different statistical techniques such as:

– index numbers to establish a trend or to compare trends between two or more variables

measured in different units or at different magnitudes;

– moving averages and regression analysis to determine the trend and forecast.” (Saunders et al., 2009, p. 467)

In this sense, since the years which this research will study (2000-2010) overlap with one economic recession in 2002 and one financial crisis in 2008, taking an index year has not been found reasonable. Instead of index year, trend analysis of this study will be implemented with averages and regression analysis.

2.1.5 Variable Selection

The purpose of this study is to test the relation, if any, between use of fair value measurement and reported performance ratios of the firm and finally its effect on CEO compensation in Sweden. Since the firms were allowed to use fair value method after 2005, firm ratios and CEO compensations will be scrutinized in two different periods over ten years between 2000-2010. The changes of firm ratios and CEO compensations between 2000-2004, which is the period of before fair value, will be compared with the changes in between 2005-2010, which is the period of after fair value.

The alternative period after use of fair value could be further years such as 2010-2015 where the managers have more experience in use of fair value method. It would give more recent results. However, this would risk the data continuity and decrease the value of trend analysis. Yet, this period can be scrutinized in future researches.

(15)

is stated before the cash pay of the compensation plans will be scrutinized because cash pay is related to accounting based performance ratios.

While doing these things, the differences between ownership structures will be taken under consideration because different owners have different preferences, and different level severity of agency problem (Hedlund et al. 1985, Oreland, 2008). In this sense, two different ownership structures will be compared. Those structures will be the firms with family concentration where the equity or voting rights concentration is also significant and the firms with dispersed ownership structure. Basu (2001) asserts that the small firms tend to use more conservative measurement methods while big firms not. At this point, control variables of this research will be firm size and CEO age since Bryan et al (2000) found that the older CEOs had been paid with more stock based compensation.

The data will be collected both from annual reports and web based data site of holding.se. Sample firms will be selected with consideration of information availability. Furthermore, samples will be aggregated from 20 Swedish companies which are listed on the Stockholm Stock Exchange before 2000. This data set can be enlarged and detailed by another researcher in a longer research period.

(16)

3. Theoretical Framework

This chapter begins with a discussion about agency theory and followed by presentations of fair value method, CEO compensation plans and performance ratios. This section finishes with the Literature Review.

3.1 Agency Theory

According to Alchian & Demsetz (1972) the output of a team is bigger than sum of outputs of individual workers which leads the resource owners to establish cooperative economic organizations that called firms. The main target of the resource owners in establishing the firm is to seize the plus value of the team. However, “team members may shirk” and in order to prevent shirking Alchian & Demsetz (1972) suggest monitoring team and rewarding them. Nonetheless, it creates a kind of trust problem and requires monitoring the monitor as well. Even if monitoring team has widely accepted by all parties of the market, and has been used as a tool of efficient governance, it is insufficient on its own.

Furthermore, agency theory is developed by considering the firms where the company owner is separated from its management, the risk is here that the agent may not work for the best interest of the principal (Jensen & Meckling 1976, Fama & Jensen, 1983). This risk creates agency costs; monitoring, bonding and residual loss (ibid).

Monitoring costs are the cost of monitoring agents’ activities such as audit costs. If monitoring costs increase, compensation of the managers decreases which is named as price protection (ibid). As a result of efficient price protection agents bear the cost of monitoring in their future contracts. In this sense, agents endeavour to show that they are working for the interest of principals by providing financial statements which named as bonding cost (Jensen & Meckling 1976). If managerial actions are less beneficial than expected by the principals then there is an ex-post settling-up possibility to revise manager’s contract for the next term (Fama, 1980, ibid).

(17)

will be a managerial labour market and finally managers will be paid well to align their incentives with the principals (Fama 1980, Fama & Jensen, 1983, van der Zwan 2014).

However, agency theory reflects the Anglo-American dispersed and well-protected ownership structure in contrast with the prevailing market characteristics of Sweden (Aguilera & Jackson, 2010, La Porta et al. 1998), where ownership concentration and the effectiveness of big owners in monitoring are highlighted by many different researchers (Högfeldt 2005, Oreland (2007).

Sweden is a civil-law country with its concentrated ownership structure where the firms are governed by a few big owners and families (La Porta, 1998, Högfeldt 2005). In this study, these kinds of firms will reflect ‘typical Swedish firms’. However, in order to increase contribution of this paper. Two different groups of ownership structure will be used; firms with family concentration where equity or voting rights concentration is significant with 20% (Faccio and Lang, 2002) and firms with dispersed ownership where the highest equity or voting rights are less than 10% (Kollegiet for Swedish Corporate Governance, 2016).

3.2 Fair Value Method

According to IFRS13, Fair value is the price that supposed to be received when an asset is on sale or paid to transfer a liability at the measurement date. Fair value method does not give a firm specific measurement, rather market based. It provides opportunity to make comparison between firms in the market as well as between different time periods of one certain firm. According to this measurement method, entities recognise the unrealized gains/losses in their profit or loss in the period that they occurred. Penman (2007) utilizes the income definition of Hicksian which considers any wealth increases as income. In this sense, Penman asserts that the fair value measurement brings solution to income measurement since it books any increase or decrease as an income or expense.

(18)

*The developed inputs by using market data, there are some criteria to distinguish whether an input is observable or not such as supported by market transaction, accessible information by interested parties and so on (Godfrey et al, 7th ed., 2010).

In this sense, the hierarchy categorises the inputs used in valuation techniques into three levels (IFRS 13, p.72-89 & Godfrey et al, 7th ed., 2010). First level observable inputs should be valued by the current market price at the measurement date such as securities (ibid). The first level of the hierarchy is not leaving any initiative to managers make some subjective decisions (ibid). However, if those prices are unavailable, special indexes should be used. Level 2 inputs, are the inputs other than quoted market prices included within Level 1 which are observable for the asset or liability, either directly or indirectly (ibid). In this sense, the similar or identical assets’ prices in active or inactive markets can be used. According to Level 3 inputs, the inputs are unobservable for the asset or liability so the best available information can be used to make valuation which allows the company’s internal information (IFRS 13, p.72-89 & Godfrey et al, 7th ed., 2010).

The fair value measurement, first of all, has been criticized for involving inputs that obtained by management in absence of active market (Zeff 2005). On the other hand, many scholars defend the fair value measurement concept. Barth and Clinch (1995) stressed that the revaluation under fair value method is highly relevant for investors at decision making. Findings of Barron et al. (2016) suggest that the use of third level measurements give more useful information about the firms’ future earnings. However, according to Godfrey et al, (7th ed., 2010), under fair value method, the agents (CEO, CFO), have subjective judgement possibilities embedded within this method and they are aware of that the future managerial contracts will be decided by depending on the financial statement information. Furthermore, Godfrey et al (2010) highlights accusations of fair value method applied to assets and liabilities with its exacerbating role in financial crisis. It is asserted that the firms allowed to write up assets in boom conditions then it has been required to write down them when the issue comes to liquidation.

(19)

of losses were the subject of litigation from debt providers or shareholders but recognition of gains improves the ability of managements to pay out more compensation to themselves (ibid).

In this sense, since the conservatism restricts opportunistic behaviour of managers, Basu (1997) argues that the demand for conditional conservatism has increased over the years and this can be taken as a result of higher demand for compensation based contracts. Furthermore, according to Ball et al (2003) conservative accounting principles reduces the reported income which convey the reduced managerial opportunism. At this point, it becomes important to test any increase in firm performance ratios and CEO compensation after the implementation of fair value method.

Two following subtitles will discuss the CEO compensation plans and the performance measures which are commonly used in those plans.

3.3 CEO Compensation Plans

Jensen and Murphy (1990), assert that CEOs make private cost-benefit assessment before in order to fulfil shareholder expectations. According to Bender (2004), CEO pay structure’s approach of HR is to attract and retain executives with a potential of large earnings. In this sense, Palmberg (2012) explains two approaches for setting compensation contracts; optimal contracting where the managers don’t involve in contracting process, and the managerial power approach where the managers influence contracts.

(20)

asserts that concentrated ownership and family firms weaken CEOs bargaining power. On the other side, Oreland (2007) underlines that concentrated ownership and family firms tend to exert control over CEOs and implement more efficient monitoring which decreases the level of compensation in Sweden.

At this point, Gabaix and Landier (2008) findings from different countries indicate that there is no certain effect of the family or concentrated ownership over CEO compensation. However, the findings of Gomez-Mejia et al. (2003) from the USA, indicate a negative effect while the findings from China indicate positive effect (Cheung et al. 2005). On the other side, findings of Mishel, Bernstein & Allegretto, (2007) highlights that CEOs of Swedish listed firms were earning 44 times less than their peers in the USA in 2005 although, the CEO compensations had increased by over 300 percent between 1998-2005 (Palmberg, 2012). Additionally, Oxelheim and Randøy (2005) found that firms with dispersed ownership structure indicates higher compensation in Norway and Sweden. It is also worth to pay attention that Davila&Penalva (2006) found the more CEO influence in governance decisions, the more use of cash pay and accounting based measures of performance, specifically use of ROA ratio. On the other side, Matolcsy (2000) asserts that the cash pay compensation level is induced by certain annual financial targets.

In this case, since there are both firms with dispersed and concentrated ownership structures existence at Stockholm Stock Exchange, to make this research in Sweden becomes convenient. This will provide opportunity to compare compensation changes among different kinds of firms after the implementation of fair value measurement.

3.4 Performance Ratios

(21)

Since this study will test the relation, if any, between use of fair value method and increases in CEO compensation, and since fair value method is about accounting, the accounting based contracts will be scrutinized. In this sense, the financial performance measurements will be used.

While the research result of Lambert and Larcker (1988) indicates that the firms place less weight on accounting performance, Peterson and Plenborg, (2012) underlines the importance of using profitability performance measures to prepare compensation plans. The research result of Jensen and Murphy (1990) indicates a very small link between CEO compensation and performance of the firms. However, Bender (2004) claims that the performance-pay relation legitimizes the CEO compensation contracts in the eyes of shareholders.

On the one side, Cybinski & Windsor (2013) assert that there is an increasing public anger to those exponential rises in CEO compensations in Australia. On the other side, The UK Cadbury Committee (1992) suggests rewarding executives by increases in shareholder wealth and firm performance. In this sense, the legitimation of the CEO pay with firm performance may be needed by all other actors of the economies. Additionally, according to Shaw & Zhang (2010) the ideal efficient compensation contract is the one which links the CEO compensation to firm performance in order to align facets interests. They also state that CEOs have not been penalized for lower performance of the firm rather rewarded yearly. However, Leone et al. (2006) assert that CEO cash compensation have been punished for poor performance. In this sense, studying cash compensation, gains importance with its relation to the accounting performance.

(22)

3.5 Literature Review

Literatures in this subject which is focusing the relationship between use of fair value method and firm performance ratios as well as CEO compensation are very much limited. However, the subject has been studied in China more elaborately than this paper conducted. They studied the subject with the details of the profits and losses in the financial statements which induced by use of fair value method (Shao&Chen&Mao; 2012, Hou and Jin, 2010, Xu and Zeng, 2010). In this paper, details above are not scrutinized due to time restrictions and it is suggested for future studies.

According to those studies, Hou and Jin (2010) documented that CEO compensations have been positively affected by both profits and losses from changes created by use of fair value method. Furthermore, Xu and Zeng (2010) found that CEO compensations have been positively affected by profits created by use of fair value method (PCFV) and have not been affected by losses created by use of fair value method (LCFV).

On the one side, while Taussings&Baker (1925) from the USA assert that there is no relation between CEO compensation and firm performance ratios, research results of Conyon (2006) indicates that the firm performance and CEO compensation link have gained significance over years. Additionally, Shao&Chen&Mao, (2012) assert that it has always been significant in China. However, they emphasized that the effects on CEO compensation diverges in the situations of decrease/increase in performance ratios (ibid). In other words, CEO compensations are sticky.

(23)

On the other side, there is no research which studied the relation between use of fair value method and firm performance ratios as well as CEO compensation in Sweden. In this sense, the researches of Oreland (2007, 2008) were found most related to subject of this research. He studied the link of firm performance ratios and ownership structures between 1985 and 2000. According to research of Oreland (2007, 2008) from Sweden, firms with family concentration, exhibited lower performance than any kinds of firms with dispersed ownership structure (Oreland, 2008). However, no significant link between ownership structure and CEO compensation have been detected (Oreland, 2007). Furthermore, it is stated that the firms with family ownership use less performance related compensations compared to firms with dispersed ownership structure (ibid).

However, his study scrutinized the data between 1985 and 2000 while this paper intends to scrutinize data between 2000-2010. This study will contribute the explanation of CEO compensation and firm performance association with the results of the years where comparison of measurement method changes happened and will give more actual information. Furthermore, this study will give more clear comparison with its ten-years data between dispersed and family concentration firms in Sweden with the focus of measurement method.

4. Hypotheses

This section begins with the presentations of hypotheses under two different ownership structures; Firms with Family Concentration and Firms with Dispersed Ownership Structure. Hypotheses are derived on the basis of the existing research and theories which will then form basis for the quantitative research.

(24)

La Porta et al. (1998) emphasized that the countries which have the less investor protection have the smaller debt and equity markets. Even if von Koch et al (2014) asserts that the shareholder protection of Sweden has evolved through the years it is still low and these causes a small market as well as a small-world network among governance actors such as agents, board members and owners (Sinani et al 2008). This is one of the prevailing market features in Sweden, which facilitate diffusion of information among firms (ibid). In this sense, reputation and credibility may create more powerful mechanisms to reduce severity of agency problem and may reduce the CEO compensations in any kind of ownership structure in Sweden.

In the light of this information, this study will distinguish two different ownership structures so that hypotheses will be developed by proceeding this distinction.

4.1 Firms with Family Concentration

Sweden is a civil-law country with its concentrated ownership structure where the firms are governed by a few big owners (La Porta, 1998, Högfeldt 2005). In this study, these kinds of firms will be reflected as ‘typical Swedish firms’.

According to Kollegiet for Swedish Corporate Governance (2016), big owner is the one who holds 10% of the shares or voting rights of the firm. In Sweden, this power with voting rights concentration is protected by issuing dual class shares and pyramid ownership in large firms (Högfeldt 2005). This concentration created 15 elite families in Sweden which were conducting 38% of the Swedish GDP in 2000 (ibid). Oreland (2007) describes the firms with family control as the firms where the firm is owned by one or more individuals from the same family and they control the largest equity or voting rights. In this sense, this research will operationalize the concentrated ownership with family concentration in Sweden.

However, powerful owners are perceived risky by other investors in the USA that some firms where founders control the significant vote-rights, experienced a significant price fall during the initial public offering such as Facebook Inc with its 52.5% of price fall and Groupon Inc with 79.3% (Lukomnik, 2012).

(25)

expand and sustain their existence and growth, firms need to reach new resources which is possible with new investors. In order to attract those frightened investors to the firm’s equity, profitability and shareholder value becomes key points. This makes CEOs extra important because the CEO has talent to run the firm in this path. Furthermore, Mehran (1995) asserts that CEO intentions are affected by compensation to get a measurable impact on corporate efficiency. Therefore, the incentive of CEOs needs to be induced with his/her compensation (van der Zwan, 2014, Sapp, 2008). According to Oreland (2007, 2008), the firms with family concentration tend to exert more monitoring. In this sense, the CEO compensation may be used to increase the CEO incentive to make firm’s equity attractive for foreign investors, rather than solving the agent problem.

On the other side, it is stated before according to IFRS 13, implementation of fair value provides managerial subjective revaluation opportunities with its ‘hierarchy' and recognizes unrealized gains, which affect profit of the year (ibid) and presumably CEO compensation. Furthermore, Paul (1992) asserts that CEOs may have incentives to report higher income to affect ROA as well as their compensation. In this sense, it is predicted that CEOs of the firms with family concentrated ownership structure tend to use fair value to increase performance ratios.

However, as Oreland (2008) stated that the firms with family concentration exert more monitoring, so it is predicted that the CEOs of those firms will be paid lower than the CEOs of the firms with dispersed ownership structure.

HO.1: Use of fair value method positively affects performance ratios in firms with family concentration.

HO.3. CEO compensations in firms with family

concentration increased more than CEO compensations in firms with dispersed ownership structure.

(26)

4.2 Firms with Dispersed Ownership

Oreland (2007) describes the firms with dispersed ownership as the firms where there is no controlling shareholder existence. This paper will also proceed from this description. Dispersed ownership structure is more common in the countries of common law (La Porta, 1998). In those countries, interests of minority shareholder are highly protected with law and the firms are controlled by professional managers instead of less skilled family heirs (La Porta, 1998, Oreland, 2007). The markets dominated with these features are, such as the US or the UK markets, bigger than the ones in the civil law countries (La Porta, 1998).

However, as it is stated before, agency theory is developed with consideration of the firms where the security ownership is separated from its control and management (Jensen & Meckling 1976). This separation creates agency problem (ibid) which is more severe in firms with dispersed ownership structure due to absence of controlling group to exert more intensive monitoring. Additionally, the big corporate scandals -such as Lehman Brothers and Enron- are more frequent in common law countries. As Watts&Zimmerman (1990) stated before the managerial choice of accounting is depending on the idea of increasing managerial utility. In other words, managers intend to increase their compensation. The findings of Dechow et al. (2010) are consistent with this point. He signified that the gains (realized/unrealized) reported by CEOs were taken as basis to compensation plan.

Oreland (2008) stated that firms with dispersed ownership structure display higher performance. In this case, this research predicted that the managers of the firms with dispersed ownership structure tend to benefit of the subjectivity of fair value method to increase performance ratios and their compensation.

HO.4: Use of fair value method positively affects performance ratios in firms with dispersed ownership structure.

(27)

However, as Oreland (2008) stated that the family concentrated firms exert more monitoring, which reduces the level of CEO compensation. Since the agency problem is more severe in firms with dispersed ownership structure CEOs of the firms with dispersed ownership structure will be paid more than CEOs of the firms with family concentration. Furthermore, Oxelheim and Randøy (2005) found that the firms with dispersed ownership structure indicates higher compensation in Norway and Sweden.

5. Data and Empirical Findings

This section was written to make a presentation of data as well as to make an operationalization of the variables. Furthermore, a presentation of the data processing was done.

5.1 Data

Data have been aggregated both from annual reports and web based data site of holding.se. 20 Swedish firms have been taken from large, medium and small capital lists of Stockholm Stock Exchange.

Sectors of the firms are healthcare, information technics, service, finance, trade and industry. Sample firms were selected by considering the information availability over studied 10 years of 2000-2010. Furthermore, samples have been aggregated by dividing 20 Swedish firms in two; firms with family concentration and firms with dispersed ownership.

In this sense, firms with family concentration have been operationalized with the firms where the family owners hold at least 20% of the voting rights, by depending on the definition made by Faccio and Lang (2002) that a family-owned company is the one where the largest owner is a family and has at least 20% of the votes. The firms with dispersed ownership have been operationalized with the owners who don’t hold more than 10% of voting rights. The voting right is taken as a point of departure for the

HO.6. CEO compensations in firms with dispersed

ownership structure increased more than CEO

(28)

sample aggregate in this research since it represents the owner’s capacity to influence the firm.

10 firms of dispersed ownership were divided in three capital size with 6 firms from large-cap, 3 firms from mid-cap and 1 firm from small cap. Same mix of firm size have been selected for the firms with family concentration. One of the difficulties was to find small cap firms with 10 years of stock market history. This may be due to capital improvement of the firms over years and changing the capital list.

Furthermore, ownership structures of the dispersed firms changed over years so this has been taken under consideration to aggregate data from the firms with dispersed ownership structure during the studied years of 2000-2010. One exception to this is one firm with dispersed ownership structure from the small-cap list. This firm had an investing company as its biggest shareholder with 10,04% capital and 10.7% share rights. The underlying intention of aggregating data from different capital-size lists was to reach more comprehensive results.

As it was mentioned before that CEO compensation is needed in Sweden in order to attract investors. Since this research examines the effect of fair value method over performance ratios and CEO compensation, accounting based compensation of cash pay has been chosen. Data collected by considering this and cash compensation operationalized with fixed salaries and bonuses.

In this sense, since profitability of a firm may lock investors, the profitability ratios of Return on equity (ROE) and Return on assets (ROA) were taken as the independent variables of this study. While ROE and ROA had been used as the independent variables, CEO age and the employee numbers had been used as the independent control variables of this study. Data of those variables have been collected from the annual reports of the firm.

(29)

5.2 Reliability and Validity of Data

Data have been collected from annual reports. However, one difficulty should be highlighted here. Firms were not publishing detailed information about CEO compensation before 2005 which is the year that Swedish Code of Conduct was issued. Total payments were given but many of the annual reports had not reported the cash, bonus, pension payments, other benefits and stock payments separately.

On the other hand, since use of fair value method was not compulsory in 2005 rather arbitrary. Data collected by considering the use of fair value method in 2005. However, in which inputs/items valued at their fair value have not been detected. Furthermore, whether firms had implemented fair value in every year’s financial reports after 2005 or not have not been detected neither. If there is/are missing years when the fair value method had not been implemented may harm the data continuity.

In this sense, reader should take it under consideration that there is always a risk that the data source is subject to certain faults. Furthermore, reader should also consider that this point may affect the accuracy of the results.

5.3 Data Processing of the Firms with Family Concentration

The data processing of firms with family concentration had been done by depending upon the data which were aggregated from 10 Swedish listed firms at the Stockholm Stock Exchange’s large-cap, mid-cap and small-cap lists.

Firstly, the data were analysed with correlation test and if a correlation detected, t-tests were implemented to find out any significance and finally the regression analysis had been implemented. In order to test hypotheses, trend analysis had been implemented with average function as well as had been displayed with figures and charts. The results of multiple regression analysis have been showed in the following part of this research. Reader should look at the appendices A, for each and every variable’s regression results.

(30)

t-Test: ROE-CEO Compensation t-Test: Firm Size-CEO Compensation

P(T<=t) one-tail 0.000 P(T<=t) one-tail 0.0000 t Critical one-tail 1.659 t Critical one-tail 1.6590 P(T<=t) two-tail 0.000 P(T<=t) two-tail 0.0000 t Critical two-tail 1.982 t Critical two-tail 1.9820

t-Test: ROA-CEO Compensation t-Test: CEO age-CEO Compensation

P(T<=t) one-tail 0.0000 P(T<=t) one-tail 0.0000 t Critical one-tail 1.6590 t Critical one-tail 1.6590 P(T<=t) two-tail 0.0000 P(T<=t) two-tail 0.0000 t Critical two-tail 1.9820 t Critical two-tail 1.9820

Figure 1.A: Correlation table of the firms with family concentration

1.B: T-Test Results of between CEO compensation and independent variables

Detected alpha significance in this research (α<0,05)

According to this chart, while performance ratios of ROE (+0,07) and ROA (+0,19) display very little correlation with CEO compensation, employee numbers (+0,69) which represent the variable of firm size and CEO age (+0,26) display more correlation in comparison with performance ratios. According to these correlation results, while performance ratios are increasing, CEO compensation is also increasing however the effect is negligibly low. Additionally, increase in firm size and CEO age affect the increase of CEO compensation.

(31)

On the other side, the results of T-Test between performance ratios and CEO compensation indicates lower significance as well as between control variables and CEO compensation. Despite the low significance between independent variables and CEO compensation, regression analysis has been done and the result was given with figure 3.

(32)

Figure 3: Regression analysis table of the firms with family concentration

According to multiple regression analysis results, ROE has insignificant P-value. This result is consistent with the findings of Oreland (2007) -whose research country also was Sweden and Mehran (1995) whose research country was the USA- that they asserted ROA captures and indicates the accounting performance of the firm and determination of CEO compensation is highly related with ROA. However, R square value displayed very low significance (~54,5%). When this is combined with the high value of standard error (~3,5 million) model has not been created.

As a result of these analyses, it can be asserted that CEOs of firms with family concentration have not been paid in accordance with performance ratios of ROA and ROE. This result is consistent with the findings of Taussings&Baker (1925) from the USA who asserted that there is no relation between CEO compensation and firm performance ratios.

However, the relation between employee numbers and CEO compensation indicates that the firm size plays role for the CEO remuneration in the firms with family concentration.

5.4 Data Processing of the Firms with Dispersed Ownership

The data processing of firms with dispersed ownership had been done by depending upon the data which were aggregated from 10 Swedish listed firms at the Stockholm

Regression Statistics Multiple R 0.73815316 R Square 0.54487009 Adjusted R Square 0.52753181 Standard Error 3534330.49 Observations 110 ANOVA df SS MS F Significance F

Regression 4 1.57022E+15 3.92556E+14 31.4258407 3.3380E-17 Residual 105 1.31161E+15 1.24915E+13

Total 109 2.88183E+15

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%

(33)

Stock Exchange’s large-cap, mid-cap and small-cap lists. Furthermore, data analyses were done by following the same path with data analyses of the firms with family concentration. In this sense, correlation test, t-tests and regression analysis had been implemented. The results are displayed with following figures.

Figure 4.A Correlation table of the firms with dispersed ownership structure

Figure 4.B T-Test Results of between CEO compensation and independent variables

Detected alpha significance in this research (α<0,05)

According to this chart, while performance ratios of ROE (+0,45) and ROA (+0,18) display very little positive correlation with CEO compensation, employee numbers (-0,08), which represent the variable of firm size displays negative and negligible low correlation, and CEO age (-0,045) displays negative and low correlation. According to these correlation results, while performance ratios are increasing, CEO compensation is also increasing however the effect is negligibly low. Additionally, decrease in firm size and CEO age affect the increase of CEO compensation.

On the other side, the results of T-Test between performance ratios and CEO compensation indicates lower significance as well as between control variables and

Correlation chart of firms with

dispersed ownership

ROE ROA Employees CEO age CEO compensation ROE 1 ROA 0.74 1 Employees 0.24 0.29 1 CEO age -0.14 0.00 0.42 1 CEO compensation 0.45 0.18 -0.08 -0.045 1

t-Test: ROE-CEO Compensation t-Test: ROA-CEO Compensation

P(T<=t) one-tail 0.000 P(T<=t) one-tail 0.000

t Critical one-tail 1.659 t Critical one-tail 1.659

P(T<=t) two-tail 0.000 P(T<=t) two-tail 0.000

t Critical two-tail 1.982 t Critical two-tail 1.982

t-Test: Firm Size-CEO Compensation t-Test: CEO age-CEO Compensation

P(T<=t) one-tail 0.000 P(T<=t) one-tail 0.000

t Critical one-tail 1.659 t Critical one-tail 1.659

P(T<=t) two-tail 0.000 P(T<=t) two-tail 0.000

(34)

CEO compensation. Despite the low significance between independent variables and CEO compensation, regression analysis has been done and the result was given with figure 6.

(35)

Figure 6: Regression analysis table of the firms with dispersed ownership structure

According to multiple regression analysis results of the firms with dispersed ownership structure, ROE has more significant P-value than ROA. Despite the low significances, this is consistent with the findings of Lambert and Larcker (1988) that asserted the affiliation between CEO cash compensation and ROE has strong positive time-series relation between them. However, it is not enough to explain CEO compensation in terms of performance ratios. This result is consistent with the findings of the research result of Jensen and Murphy (1990) that they found a very small link between CEO compensation and performance of the firms.

On the one side, according to results of analysis, firm size has no effect over CEO compensation. On the other side, cash remunerations have been negatively affected by the age increase of CEOs. This is consistent with the finding of Bryan et al (2000) who asserted the older CEOs had been paid with more stock based compensation. They emphasized that; since CEOs approaching retirement become more “short run” oriented, their incentives should be strictly aligned with the principals. In this sense, CEOs approaching retirement have been paid much more with stock options than cash salaries(ibid).

To sum up, it can be asserted that the CEOs of the firms with dispersed ownership structure have not been paid in accordance with performance ratios or there is very weak link between them. This result is consistent with the findings of Taussings&Baker

Regression Statistics Multiple R 0.55326971 R Square 0.30610737 Adjusted R Square 0.27967337 Standard Error 2634524.39756097 Observations 110.00000000 ANOVA df SS MS F Significance F

Regression 4 3.21496E+14 8.04E+13 11.58006 7.945002E-08 Residual 105 7.28775E+14 6.94E+12

Total 109 1.05027E+15

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%

(36)

(1925) from the USA who asserted that there is no relation -or negligible weak- between CEO compensation and firm performance ratios.

As a result, since R square value displayed very low significance (~30%) and its combination with high value of standard error, CEO compensation can be explained neither in terms of performance ratios, nor firms size and CEO age. In this sense, model has not been created.

6. Examination of Hypotheses

In this section, the hypothesis of this study will be analysed and the findings will be discussed in accordance with the arguments which were used to develop those hypotheses.

6.1 Hypotheses of the Firms with Family Concentration

This section focused to analyse the hypotheses of the family concentration category. Hypothesis 1: Use of fair value method positively affects performance ratios in firms with family concentration.

(37)

Chart 1.A: ROA-ROE Changes in Firms with Family Concentration 2000-2004

Chart 1.B ROA-ROE Changes in Firms with Family Concentration 2005-2010

(38)

In the light of given information, the graph of studied ten years (2000-2010) -as it is seen in chart 2-, ROA and ROE slopes are positive. However, it is arguable whether these positive slopes are due to use of fair value or not. Yet, it is worth to highlight that the fall of performance ratios in 2008 was not as sharp as it was in 2002 and this can be due to use of fair value.

Chart 2: ROA-ROE Changes in between 2000 and 2010

However, it is difficult to assert that there is a significant positive relation between use of fair value and performance ratios. In this sense, hypothesis 1 has not been supported by the findings.

Hypothesis 2: There is significant positive correlation between CEO compensation and use of fair value.

(39)

Chart 3: Changes in CEO Cash Compensation between 2000-2010

Chart 4A: Changes in CEO Cash Compensation between 2000-2004

(40)

Chart 3 shows the overall changes in CEO cash compensation. According to this the breaking points had negatively affected the increase of CEO cash compensation. However, as it is seen above the negative effects of the 2008’s crisis lasted longer than the effects of breaking point 2002.

As it is seen in chart 3A, CEO cash compensation had already had a positive trend and higher slope value in between 2000-2004 in comparison to slope value in between 2005-2010, which drives the hypothesis 2 to fail.

Furthermore, even if it is slightly decrease, there had been a decrease in CEO cash compensations just after the use of fair value in 2005. This can be related to findings of Matolcsy (2000) from Australia that cash pay compensation level is induced by certain annual financial targets. Furthermore, this can be in line with the findings of Leone et al. (2006) that CEO cash compensation have been punished for poor performance. In order to make this clear, trend analysis had been implemented by using the data of CEO cash compensation and ROA-ROE ratios which were lagged by one year.

Chart 5: CEO Compensation Changes in Accordance with ROA and ROE

(41)

accordance with mostly ROA and comparably less with ROE ratios in between 2003 and 2006 which is in line with assertions of both Matolcsy (2000) and Leone et al. (2006) in those years. However, the overall result is consistent with Oreland (2007) that he asserted the firms with family ownership use less performance related compensations compared to firms with dispersed ownership structure (ibid).

To sum up, these are not proofs as positive effects of use of fair value. The decrease in CEO cash compensation after use of fair value makes it difficult to assert that there is a positive relation between use of fair value and cash pay. In this sense, hypothesis 2 have not been supported by the findings.

Hypothesis 3: CEO compensations in firms with family concentration increased more than CEO compensations in firms with dispersed ownership structure

Hypothesis 6: CEO compensations in firms with dispersed ownership structure increased more than CEO compensations in firms with family concentration.

These hypotheses have been developed by depending on both agent theory and the assertion of Oreland (2008) that the firms with family concentration exert more monitoring and it lowers the remuneration. Additionally, agency problem is more severe in the firms with dispersed ownership structure, and Oreland (2008) states that the family ownership reduces the severity of agency problem. In this sense, it is predicted that the CEOs of the firms with dispersed ownership structure would be paid higher than the CEOs of the firms with family concentration and trend analysis had been implemented. Result is showed in chart 6 below.

(42)

According to this chart, CEOs of the firms with family concentration had been paid more than the CEOs of the firms with dispersed ownership. This result is inconsistent with the findings of Oxelheim and Randøy (2005) that asserted the firms with dispersed ownership structure indicates higher compensation in Norway and Sweden. Additionally, it is disagreeing with the findings of Oreland (2007) that detected no significant link between ownership structure and CEO compensation. In the light of the results of trend analysis, hypothesis 3 and 6 have not been supported by the findings.

6.2 Hypotheses of the Firms with Dispersed Ownership

This section focused to analyse the hypotheses of the dispersed ownership category. Hypothesis 4: Use of fair value method positively affects performance ratios in firms with dispersed ownership structure.

This hypothesis has been developed by depending on both agent theory and the assertion of Watts&Zimmerman (1990) that the managerial choice of accounting is depending on the idea of increasing managerial utility. This assertion is also in line with agent problem that the agent intends to work for his/her own interests. In this sense, it is predicted that the managers would take the advantage of subjectivity in fair value method and would increase performance ratios in order to increase their compensation. Findings of Dechow et al. (2010) supported this prediction that the realized/unrealized gains reported by CEOs were taken as basis to compensation plan (ibid). According to this hypothesis, the relation between use of fair value method and performance ratios in firms with family control was tested. Expected result was getting higher increase in performance ratios after the use of fair value method in between 2005-2010, in comparison with the ratios of 2000-2004.

However, the trend analysis result does not seem to fully support this hypothesis. First of all, according to correlation and regression analysis’ results there is no significant relation between ROA-ROE and CEO compensation. If CEOs are knowledgeable about this information, they would not intend to take the benefits of using fair value subjectivity to increase performance ratios.

(43)

financial situations that performance decrease was greater. Furthermore, since, the performance ratios display negative values before 2004, this result is inconsistent with the findings of Oreland (2007, 2008) which asserted that the firms with family concentration, exhibited lower performance than any kinds of firms with dispersed ownership structure (Oreland, 2008).

On the one side, first five years (2000-2004) performance ratios’ path, both ROE and ROA had already had an increasing trend with positive slope. As it is seen at the chart 7.A. On the other side, according to analysis result of second five years (2005-2010), both ROE and ROA had negative slopes.

Chart 7.A: ROA-ROE Changes in Firms with Dispersed Ownership 2000-2004

Chart 7.B: ROA-ROE Changes in Firms with Dispersed Ownership 2005-2010

(44)

According to chart 8, which is the graph of studied ten years (2000-2010), both ROA’s and ROE’s slopes are positive. However, after the use of fair value in 2005, ratios had decreased. Increases after 2006 did not last longer and while ROA started to decrease in 2007 with small values, ROE started to decrease in 2008. The magnitude of the decrease of ROE was greater than the decrease in years where historical cost method was employed.

Chart 8: ROA-ROE Changes of Firms with Dispersed Ownership in between 2000-2010

In the light of the results of trend analysis, hypothesis 4 is not been supported with the findings. In other words, there is no significant positive relation between use of fair value method and performance ratios of the firms with dispersed ownership structure. Hypothesis 5: CEO compensation positively affected by use of fair value in firms with dispersed ownership structure.

(45)

of accounting is depending on the idea of increasing managerial utility. Dechow et al. (2010) highlighted that gains (realized/unrealized) reported by CEOs were taken as basis for compensation plan (ibid). In the light of those arguments and agent theory it is predicted and expected that CEOs of the firms with dispersed ownership would benefit managerial subjectivity within fair value and they would increase both performance ratios and their compensation.

In order to test the hypothesis, trend analysis has been implemented and the results are displayed in chart 9.A, 9.B and 10.

Chart 9.A: Changes in CEO compensation between 2000-2004