Corporate Governance in the Netherlands: Compliance with the
Frijns Code and the Effect on Firm Performance and Firm Value
H.H. Zeilstra
December 2011
University of Groningen
Faculty of Economics & Business MSc Business Administration
3
Corporate Governance in the Netherlands: Compliance with the
Frijns Code and the Effect on Firm Performance and Firm Value
AbstractCorporate Governance is gaining more attention in recent years as a result of corporate scandals and the financial crisis. Also views on the role of corporate governance codes, and therefore corporate governance codes themselves, change. This thesis provides an extensive research into the effects of the changes in the Dutch corporate governance code in the years 2009 and 2010. In order to do so, first the compliance levels with the new provisions are tested. This compliance analysis shows that the compliance level with the new provisions is high in most fields, but that it is relatively low in the field of remuneration. The compliance analysis also shows that compliance levels significantly differ between firms from different industries and that compliance levels are positively related with firm size. Afterwards the effects of compliance on firm accounting performance and firm value are investigated. No significant relations are found for the complete Dutch corporate governance code, but evidence is found for a significant positive effect of compliance with the new provisions in this code on firm value. When the new provisions are divided into different fields, significantly positive relations are found for compliance in the field of remuneration with both accounting performance and firm value. In the field of communication with shareholders only a significantly positive relation is found with firm value. In the fields of risk management, diversity and CSR, no evidence is found for an effect of compliance on firm performance and firm value. However, a significantly positive relation is found between compliance in the field of diversity and actual gender and age diversity within supervisory boards.
Keywords: corporate governance, compliance, accounting performance, firm value, the Netherlands.
Supervisor: Dr. S. Tillema Second Assessor: Dr. C.A. Huijgen
5 0
Preface
In front is my master thesis which I wrote to finish my master program ‘Organizational and Management Control’ at the Faculty of Economics and Business, University of Groningen. The reason for writing it about corporate governance in the Netherlands is that I participated twice in the investigation by the Corporate Governance Insights Centre (CGIC) into corporate governance compliance among Dutch listed companies. The CGIC is the corporate governance research centre of the University of Groningen. While carrying out my student-assistantship at the CGIC I wondered what the effects of compliance with the new provisions were. Investigating this turned out to be challenging but also a very interesting.
I would like to thank my supervisor dr. S. Tillema for her supervision and advice during the process of writing my thesis. I would also like to thank dr. C.A. Huijgen for reviewing this thesis.
I would like to thank the CGIC, and in particular dr. D.H.M Akkermans, for enabling me to use their compliance data in performing the research for this thesis.
Furthermore, I would like to thank my family, my friends and my girlfriend for their support and their patience during my study and especially during the writing of this thesis.
Hendrik Zeilstra,
6 1
Table of Contents
0Preface ... 4 75 11. Introduction ... 4 88 22. History of Corporate Governance and Corporate Governance in the Netherlands ... 4 911 32.1 Aim and History of Corporate Governance ... 5 011 42.2 Corporate Governance Codes in the Netherlands ... 5 113 52.3 Changes with the introduction of the Frijns code ... 5 215 63. Theory and Hypotheses on the Relationship between Complying and the Performance and Value of the Firm ... 5 317
7 2 44.2 Tests ... 7 132 2 54.2.1 Compliance Analysis ... 7 232 2 64.2.2 Analysis of the Relationship between Compliance and Performance and Value . 7 333 2 75. Results ... 7 435 2 85.1 Compliance Analysis ... 7 535 2 95.1.1 Overall Compliance levels ... 7 635 3 05.1.2 Compliance and Indexes ... 7 736 3 15.1.3 Compliance Differences between Years ... 7 838 3 25.2 Relation between Compliance and Performance and Value ... 7 939 3 35.2.1 Descriptive Statistics ... 8 039 3 45.2.2 Multicollinearity, Autocorrelation and Heteroscedasticity ... 8 141 3 55.2.3 Results for the Complete Code and All New Provisions ... 8 241 3 65.2.4 Results for the New Risk Management Provisions ... 8 343 3 75.2.5 Results for the New Remuneration Provisions ... 8 444 3 85.2.6 Results for the CSR Provisions ... 8 546 3 95.2.7 Results for the Diversity Provisions ... 8 647 4 05.2.8 Results for the New Communication Provisions ... 8 750 4 16. Conclusions and Discussion ... 8 852 4 27. References ... 8 956
8
1.
2Introduction
The general objective of corporate governance codes is to enhance the quality and transparency of corporate management, which should result in improving company performance and restoring investors’ confidence (Werder et al., 2005; Brown and Caylor, 2008). Restoring this confidence was necessary after large corporate scandals like Enron in the US, but also Royal Ahold in the Netherlands. As a reaction to these and other scandals, new corporate governance codes were introduced worldwide at the beginning of the 21st century. Their main goal was to prevent such scandals to happen again or at least to decrease the chances of such scandals to happen to a minimum. The codes had to reduce the agency problem between investors and managers of companies. One of these new codes was the Dutch Tabaksblat code. In December 2009, as a reaction to further developments in the corporate governance field and to increase the level of corporate governance in the Netherlands, the Frijns code was introduced as the successor of the Tabaksblat code.
The Tabaksblat code was the first mandatory code in the Netherlands and it was introduced in December 2003. The Frijns code is based on the Tabaksblat code. The overall structure of the code remained the same, but some changes and additions to the old code have been made. The main changes are in the following categories: risk management, remuneration, corporate social responsibility (CSR), diversity within the supervisory board and communication with shareholders. In other categories only some small changes have been made.
Similar to the Tabaksblat code, with the Frijns code it is legally mandated for publicly listed companies to report their compliance with the code in their annual reports. However, it is not mandated to comply with all best practice provisions, as the code works with the ‘comply or explain’ principle. By using this rule the Dutch corporate governance monitoring committee tries to let the firms act in the spirit of the code instead of stimulating them to use it as a list on which they have to tick the boxes. As previous research into the compliance of publicly listed companies with the code has shown that this compliance is rather high, the Dutch corporate governance monitoring committee argues that there is broad support for the code among the publicly listed companies (Monitoring Commissie Corporate Governance Code, 2010).
9 Netherlands by adding new provisions, compliance with this code, and its new provisions, can be expected to lead to higher firm value and better firm performance. Examples of changes which are, according to the agency theory, expected to have a positive influence on firm performance and firm value are the changes in the field of remuneration and risk management. In the remuneration provisions the focus has moved more to longer term share price, which better aligns the long term interests of managers and shareholders. Related to risk management, the provisions of the Frijns code demand more transparency on, for example strategic risks. According to agency theory this increase in transparency enables investors to better estimate the risks the firm faces. Also an overall increase in transparency will help outsiders to better monitor insiders. This forces insiders to invest in projects with a positive net present value and to reduce perks and waste, so that more of the benefits flow back to outside investors (Shleifer and Vishny, 1997). However, stakeholder theory argues that corporate governance codes do not only have to aim at shareholders’ interests, but also at interests of other stakeholders to be in the best interest of society as a whole. Besides, an increase in transparency might also benefit competitors and therefore could lead to worse firm performance and a lower firm value. The aim of this thesis is to find out whether compliance with the Frijns code and the new elements of this code is positively related with firm performance and/or firm value.
4 8Objective, key research question and sub-questions
This thesis explores the relationship between compliance with the Frijns code, and in particular the new provisions within this code, and firm performance and firm value. Performance is measured by accounting performance. The objective of this thesis is to find out whether compliance with the Frijns code helps to reduce the agency problem and leads to better firm performance and to higher firm value or whether compliance might be valuable to other stakeholders but not to shareholders.
The main question of this thesis is:
To what extent do the Dutch listed companies comply with the Frijns code (and the new elements within this code), and what is the relationship between the level of compliance on firm performance and firm value?
To answer this question, the following sub-questions will be used:
10
4. How did corporate governance regulation influence firm performance and firm value according to previous empirical research?
5. What is the actual level of compliance with the code and the new provisions in the code?
6. How does compliance with the Frijns code and the new provisions within this code affect firm performance and firm value?
7. How can these empirical findings be explained?
11
2.
3History of Corporate Governance and Corporate Governance in the
Netherlands
This chapter describes the history of corporate governance and corporate governance codes, both in the Netherlands and worldwide. The first section elaborates on the aim and the history of corporate governance worldwide. In section 2.2 an overview of the history of corporate governance in the Netherlands is given. In addition, this section gives a comprehensive description of the changes in the Dutch corporate governance code.
2.1
1 3Aim and History of Corporate Governance
Before discussing corporate governance codes, it is important to know what corporate governance exactly is. In literature, different definitions of corporate governance are used. Shleifer and Vishny (1997) give a narrow definition. They state that corporate governance deals with the agency problem: the separation of management and finance. According to them the fundamental question of corporate governance is how to assure financiers that they get a return on their financial investment. Mallin (2007) gives a broader definition of corporate governance. He states that corporate governance is concerned with both internal aspects of the company, such as internal control, as well as external aspects, such as an organization’s relationship with its shareholders and other stakeholders. According to Mallin (2007), corporate governance is also seen as an essential mechanism helping the company to attain its corporate objectives and monitoring performance is a key element in achieving these objectives. Important elements of corporate governance are the board system and members, the structure and the level of the remuneration of the board(s), internal control mechanisms, the relationship between a company’s management, the board(s), shareholders and the other stakeholders, and transparency and accountability towards shareholders. Both definitions are used in corporate governance literature, but the broad definition of Mallin (2007) better describes the way corporate governance is treated in Europe and the Netherlands as the emphasis is not only on shareholders here, but also on other stakeholders.
12 The most important aim of corporate governance codes is to enhance the quality and transparency of corporate management (Werder et al., 2005). These codes are sets of business best practices, which have proliferated around the globe mainly since the publication of the Cadbury Report in 1992 (Akkermans et al., 2007). Although the first corporate governance code was already introduced at the end of the 1970s in the US, it was only after the Cadbury Report was published in the UK that worldwide corporate governance codes were initiated (Aguilera and Cuervo-Cazurra, 2004). The Cadbury code was the first mandatory code and was using the ‘comply or explain’ rule. This means that companies were obliged to comply with the provisions in the code and that when they did not comply with a provision they had to explain in their annual report why they do not comply. The introduction of the Cadbury Report led to significant changes in board structure and management characteristics in the UK: the average board size increased, as did the number and fraction of independent board members. This change in board structure led to managers becoming more responsible for their results (Dahya et al., 2002; Dedman, 2000). Complying with the Cadbury recommendations also led to better firm performance and higher firm valuation (Dahya and McConnell, 2007; Weir and Laing, 2000).
Worldwide codes similar to the Cadbury Report were adopted. Aguilera and Cuervo-Cazurra (2004) explain this diffusion of corporate governance with a combination of endogenous as well as exogenous pressures to solve deficiencies in corporate governance systems. A corporate governance system is the combination of legislation and codes in a country. Although corporate governance codes are quite similar across countries, the level of protection might differ, as the law systems differ. The common law system, which is used in the US and the UK, provides the best shareholder protection. The civil law system provides less protection as it focusses more on protecting other stakeholders, such as employees. In the Netherlands of the civil law system is used, so the level of shareholder protection is rather low in the Netherlands (La Porta et al., 1998).
In such countries with weaker shareholder rights in regulation, company ownership is less widely dispersed. There are more controlling shareholders. In some situations the controlling shareholder is just a large shareholder, at family held corporations it even happens quite often that one or more managers are part of the controlling family. When a controlling shareholder exists, there is not an agency problem between the managers and the investors, but between the controlling shareholders and the other shareholders. The other shareholders mostly have hardly any influence (La Porta et al., 1999). In these situations, more protection of outside investors is needed. Protection can be offered by courts, government agencies or market participants themselves. A way to offer this protection is a good legally obliged corporate governance code.
13 Netherlands (Bauer et al., 2004). This led to stricter regulation with new codes and to a boom in the literature in the corporate governance field.
More recently, in a report commissioned by the European Commission, De Larosière et al. (2009) argue that corporate governance is one of the most important failures of the credit crisis. According to the report, most of the incentives encouraged financial institutions to focus on short term profits, which led to high risk taking by managers. With respect to remuneration, the report states that bonuses should be better aligned with long term shareholder interest. They should reflect actual performance and should not be ‘guaranteed’ in advance. The report also argues that risk monitoring and management practices within financial institutions have dramatically failed in the crisis. According to De Larosière et al. (2009) the risk management functions within financial institutions must be made independent and responsible for effective, independent stress testing and senior risk officers should hold a very high rank in the company’s hierarchy. Furthermore they state that internal risk assessment and proper due diligence must not be neglected by overreliance on external ratings. According to De Larosière et al. (2009) there should however be no illusion that regulation alone can solve all these problems and transform the mindset that presided over the functioning of the system.
2.2
1 4Corporate Governance Codes in the Netherlands
In the past fifteen years three corporate governance codes have been used in the Netherlands. The first attempt to initiate a corporate governance code was the Peters Report, published in 1997. The report was based on self-regulation and there was no legal obligation for firms to comply with this code, as the Peters report was an initiative of the private sector. The main goal of the report was to create more rights for shareholders in the Netherlands. At the time the report was published there were a lot of protective anti-takeover measures in Dutch listed firms, such as priority shares, certificates of shares and protective preference shares. The Peters Report was aimed at making the management and the supervisory board more accountable (De Jong et al., 2005). De Jong et al (2005) argue that the Peters report was not very successful, based on their findings that no real changes in shareholder rights evolved from it and that no relation existed between a firm’s corporate governance characteristics and its value. They conclude that a corporate governance code without the obligation to comply with the code by a government or an exchange probably will not be successful.
14 provisions was still allowed, as long as the company could give a clear explanation for this. With this rule, the Dutch corporate governance monitoring committee tried to stimulate the firms to act in the spirit of the code instead of using it as a list on which they have to tick the boxes (Monitoring Commissie Corporate Governance Code, 2010). The main objectives of the Tabaksblat code were to increase transparency and accountability of Dutch corporate governance, and to improve the quality and integrity of the management and supervisory boards of companies (Akkermans et al., 2007). Compliance with the Tabaksblat code among listed firms was high. The percentage of provisions which was complied with or about which non-compliance was explained, was 88% over the year 2004 and increased to 95% over the year 2006. After 2006 this percentage stabilized at 95% (Monitoring Commissie Corporate Governance Code 2005 - 2009). In a research into compliance over 2004, Akkermans et al. (2007) find that compliance was not only high, but also positively related to company size. Their results also show that non-compliance is especially high for provisions related to the remuneration of board members, the independence of the supervisory board and requirements with respect to internal control/risk management systems. The findings of Akkermans et al. (2007) are similar to the findings of Werder et al. (2005) who investigated the German situation. Related to studies into compliance in the Netherlands it should be kept in mind that when no proof of non-compliance with a provision is found, it is assumed that this provision is complied with, since the code only demands explanations about non-compliance. As a result of that assumption, the level of compliance will probably be overstated. Also Akkermans et al. (2007) find that explanations are quite similar across companies. This may indicate that these explanations are rather symbolic and that firms are not really trying to comply.
In December 2008, the Frijns code is introduced0
1. The main goal of the Frijns code is to
stimulate proper behavior of management board members, supervisory board members and shareholders. In line with the recommendations of De Larosière et al. (2009), the Frijns code pays more attention to the way people behave than to the way they account for this afterwards (Monitoring Commissie Corporate Governance Code, 2009). With the introduction of this code almost all previous best practice provisions as well as the comply-or-explain rule are retained. Some best practices changed and some new provisions are added. The main changes are in the following categories: risk Management, remuneration, Corporate Social Responsibility (CSR), diversity within the supervisory board and communication with shareholders. These changes will be further described in the next section.
1
15
2.3
1 5Changes with the introduction of the Frijns code
In the field of risk management the Frijns code is more explicit than the Tabaksblat code. The Tabaksblat code only stated that the management board had to declare in the annual report that the internal risk management and control systems were adequate and effective. The Frijns code demands that the management board states that the internal risk management and control systems provide a reasonable assurance that the financial reporting does not contain any errors of material importance and that the risk management and control systems worked properly in the year under review. Besides that, since the introduction of the Frijns code it has become mandatory to describe the risks related to the strategy of the company. Previously, this was not mentioned in the code. Mainly the financial risks were described. Especially the last mentioned change should provide new information to shareholders and other stakeholders.
Related to remuneration the focus of the code became more on long term actual firm performance. This is a result of the introduction of the claw-back clause and the discretionary power of the supervisory board to adjust conditionally awarded remuneration components if in their opinion this remuneration component produces an unfair result due to extraordinary circumstances. The Frijns code also demands scenario-analyses to be performed to determine the level and the structure of the remuneration, which have to take into account the share price and non-financial indicators relevant to the long-term objectives of the company. Related to transparency, it is added that the remuneration report has to explain how the chosen remuneration policy contributes to the achievement of the long-term objectivesof the companyin keeping with the risk profile. It is added because according to the Monitoring Commissie Corporate Governance (2009) this relationship is discussed within the company. The committee believes that stakeholders, especially shareholders, have the right to know the outcome of the discussion as this helps them to monitor the firm and the managers. Also the Frijns code demands more transparency about, for example, conditional shares and options and about the costs of the ‘other remuneration components’, such as lease cars and houses. Furthermore the Frijns code states that if the remuneration committee of the supervisory board makes use of a remuneration-consultant, it has to verify that this consultant does not provide advice to the members of the management board.
16 The monitoring committee thinks that a diverse composition of the supervisory board improves the quality of the supervision, as it expects that more diverse composition leads to a more independent supervisory board. The monitoring committee believes that this independency is a crucial feature of a supervisory board (Monitoring Commissie Corporate Governance Code, 2007). Therefore diversity is added to the code. This is done by demanding that the profile of the supervisory board mentions the specific objective on the diversity of the composition of the supervisory board. The profile has to be place on the website of the company. If, at a certain moment, the actual diversity does not meet the desired diversity, the supervisory board has to account for this in its report. The supervisory board also has to indicate how and within what period the objective will be achieved. Age diversity and gender diversity are considered to be the most important (Monitoring Commissie Corporate Governance Code, 2009). The fact that nationality diversity is considered to be less important might stem from the fact that many Dutch listed firms mainly operate in the Netherlands. Other forms of diversity which are also mentioned by the monitoring committee are diversity in terms of expertise and background.
17
3.
4Theory and Hypotheses on the Relationship between Complying and the
Performance and Value of the Firm
In this chapter hypotheses about the relationship between complying with the new provisions and the performance and value of the firm are derived. The first section describes the main theories related to corporate governance. In section 3.2 hypotheses are derived from the theories presented in section 3.1 and from some other theories related to the effect of compliance with the Frijns code on firm performance and on firm value in specific fields.
3.1
1 6Theories related to Corporate Governance
The main theories that have affected the development of corporate governance regulation and have informed corporate governance research are agency theory, transaction cost economics, stakeholder theory and stewardship theory (Mallin, 2007). Transaction costs economics argues that corporate governance regulation should protect the stakeholders who have the highest transaction costs when ending the relationship with the company (Williamson, 1984). Stewardship theory states that empowering managers to take autonomous executive action leads to higher returns to shareholders (Donaldson and Davis, 1991). Related to corporate governance this theory argues that the CEO should be the chairman of the board. Both these theories are not really related to the changes in the Dutch corporate governance code. Therefore they are not further described in this section.
3.1.1 2 2Agency Theory
The most influential theory in the corporate governance literature is the agency theory. The agency theory views a firm as a nexus of contracts (Mallin, 2007). Jensen and Meckling (1976) define an agency relationship as a contract under which one or more persons (the principal(s)) engage another person (the agent) to perform some service on their behalf which involves delegating some decision making authority to the agent. The core of the agency problem, described by the agency theory, is that the principal wants the agent to make the decisions which are in the best interest of the principal, while the agent makes the decisions which are in his own best interest. To solve this problem, interests of principals and agents have to be aligned. According to the agency theory, the agent will still make the decisions which are in his own best interest, but these decisions will then also be in the best interest of the principal.
18 can for example be done by stealing money, by selling assets at below market prices to another firm which is owned by the insider or by overpaying executives (La Porta et al., 2000).
Corporate governance rules can be used to decrease the agency problem between shareholders and managers, but also between controlling shareholders and other shareholders. Remuneration policies can align the managers’ interests and the interests of the shareholders by using performance based incentives. Also transparency is important. Increased transparency will decrease the information asymmetry and it will help outsiders to better monitor insiders. This forces insiders to invest in projects with a positive net present value and to reduce perks and waste, so that more of the benefits will flow back to outside investors (Shleifer and Vishny, 1997). The role of the supervisory board is to resolve agency problems between managers and shareholders by setting compensation and replacing managers that do not create value for the shareholders (Carter et al., 2003). This does however require an independent supervisory board.
3.1.2
2 3Stakeholder TheoryThe other main theory related to the changes in the corporate governance code is stakeholder theory. This theory does not only take the shareholders, the managers and the relationship between them into account, but it states that managers, when making decisions, should take account of the interests of all stakeholders of a firm. Examples of other stakeholders are employees, customers, suppliers, communities, governmental officials and the environment, but also competitors (Jensen, 2002). When taking other stakeholders into account, the focus on shareholders becomes less self-evident. However, still a lot of companies try to maximize shareholder value whilst at the same time trying to take the interests of other stakeholders into account (Mallin, 2007). This is in line with the strategic stakeholder theory, a stream within stakeholder theory which states that devoting sufficient resources and managerial attention to stakeholder relations will tend to lead to profitable outcomes for the corporation and hence for shareholders (Health and Norman, 2004).
19 management performance measurement allows managers and directors to invest in their favorite projects, even if they destroy firm value or social value.
When looking at corporate governance systems, the Anglo-American system pays mostly attention to shareholder value, which is in line with the agency theory and Jensen’s enlightened stakeholder theory. The European system pays more attention to other stakeholders than the Anglo-American system. This is in line with traditional stakeholder theory. According this theory, the focus of corporate governance should not only be on creating shareholder value, but also on creating value for society (Mallin, 2007). In the remainder of this thesis, when there is referred to stakeholder theory, traditional stakeholder theory is meant.
3.2
1 7Theories Related to Specific Elements of the Frijns Code and Hypotheses
In this section the hypotheses tested in the empirical research are derived. The hypotheses are based on theories and previous empirical research. Beside the main theories related to corporate governance, there are some other theories related to specific elements of the Frijns code which might influence the relationship between compliance with the code in these fields and the firm value and firm performance. Those theories are discussed in this section. The section starts with the hypothesis related to compliance with the code as a whole and continues with hypotheses on the specific changes.3.2.1 2 4The Frijns Code as a Whole
Based on the agency theory, it can be expected that higher levels of compliance with the Frijns code will reduce the agency problem and hence lead to a better financial firm performance (hereafter firm performance) and a higher firm value compared to non-complying with the code. The expected decrease of the agency problem is the result of more transparency to shareholders and a better alignment of the interests of managers and shareholders. As the information asymmetry between the managers and the shareholders decreases, also possibilities of expropriation by insiders decrease. This should lead to better firm performance and as a result higher firm value. Better alignment of the interests of shareholders and managers, for example by aligning remuneration with share value, is also expected to lead to higher firm value. Managers will still make decisions which are in their own best interest, but these decisions will favor shareholders more.
20 theory argues that an increase in transparency in some fields might not only be valuable to shareholders, but also to competitors. Examples of such provisions are provisions related to transparency on risk management and on performance criteria. Compliance with these provisions might give competitors information about the strategic goals of the company, which might result in lower firm performance instead of better performance. As most provisions are aimed at shareholders and not at other stakeholders, still a positive relationship is expected between corporate governance compliance and firm value and firm performance. This leads to the following hypothesis:
H1: The level of compliance with the Frijns code, and with the new provisions within this code, is
positively related with firm performance and firm value.
Previous empirical research into the relationship between corporate governance and firm performance and firm value is in line with this hypothesis. Most of such research is performed in the US. For example, Core et al. (1999) find that firms with weaker corporate governance systems have greater agency problems. They also find that these firms perform worse than firms with less agency problems, while in the firms with weaker governance systems the CEOs receive a greater compensation. Core et al. (1999) measure agency problems by board structure and ownership structure. Gompers et al. (2003) and Brown and Caylor (2008) also find positive relations between corporate governance and firm value and firm performance. Though, Brown and Caylor (2008) find this relationship only for six of the provisions which they tested. These provisions are: no former CEO serves on the board; the company is not authorized to issue blank check preferred stock; non-employees do not participate in company pension plans; the CEO serves on no more than two boards of other public companies; the auditors were ratified at the most recent annual meeting; and a simple majority (not a supermajority) vote is required to approve a merger.
Also in the rest of the world, positive relationships between the level of corporate governance and firm performance and firm value are found. For example, Bauer et al. (2004) find this positive relationship among companies included in the FTSE Eurotop 300. Their findings indicate a difference between countries from the European Monetary Union (EMU) and the UK. Bauer et al. (2004) constructed share portfolios of well-governed firms and share portfolios of poorly governed firms. In the UK their share portfolio of well-governed firms outperforms their share portfolio of poorly governed firms. In the EMU countries, instead of a strong relationship between corporate governance and stock returns, a strong relationship was found between corporate governance and firm value. This might imply that in the EMU, to a large extent, corporate governance standards have already been incorporated in stock prices.
21 valued lower in countries with weaker shareholder protection, the relationship between the level of corporate governance and firm value and firm performance is stronger in these countries (Black, 2001; Black et al., 2006; Durnev and Kim, 2005; Klapper and Love, 2002; La Porta et al., 2002; Renders et al., 2010). The main reason that the relationship is stronger in these countries is that the inter-firm variation in the level of corporate governance is higher. In these countries good governance at firm level seems to be more important for shareholders as the law provides less protection. La Porta et al. (2002) define the Netherlands as one of these countries.
3.2.2 2 5Risk Management
Transparency in the field of risk management will enable the shareholders to better estimate the firm’s risks and to better monitor the way the firm deals with these risks. Knowing what the risks are prevents shareholders from being surprised by a sudden downfall in firm performance as a result of one of these risks. Monitoring how the firm deals with the risks enables shareholders to inform the firm when they believe that the risks are too high or when they believe that the risks are not well dealt with and to potentially put pressure on the firm when nothing changes afterwards. As a result, the risks for the shareholders might decrease and the value of the firm might increase.
On the other hand, complying with the provisions related to transparency about risk management and control systems might also be valuable for other stakeholders, such as creditors and competitors. Competitors might not only use information about the way the firm deals with risks in designing their own risk management system, but information about strategic risks might also reveal competitor sensitive information about the strategy of the firm to competitors. This might harm the performance and as a result the value of the firm. Therefore, the direction of the relationship between compliance with the new provisions in the field of risk management and firm performance and firm value is unclear. This leads to the following hypothesis:
H2: The level of compliance with the new provisions in the field of risk management is related with
firm performance and firm value.
3.2.3 2 6Remuneration
22 managers to take risks which increase the short term results, but might harm the long term results of the company. Based on agency theory, it is expected that compliance with these provisions leads to better firm performance and as a result to higher firm value. The findings of Core et al. (1999) are in line with this expectation. This leads to the following hypothesis:
H3a: The level of compliance with the new non-transparency related provisions in the field of
remuneration is positively related with firm performance and firm value.
In agency theory, compliance with the transparency related remuneration provisions is positively related to firm performance and firm value. When investors believe that the current remuneration policy is a good way of stimulating the management to strive for optimal performance, they will expect high future performance and valuate the firm higher. It is possible, however, that investors dislike the remuneration policy and its relationship with the objectives of the company, and that they believe that it is not the right way to gain higher future firm performance. In this situation, knowing what the policy is enables investors to reveal their anxiety to the supervisory board at for instance the annual shareholder meeting which might result in a better fitting remuneration package. The argument of Shleifer and Vishny (1997) that increased transparency will help outsiders to better monitor insiders is also relevant here.
Based on stakeholder theory, it can be argued that compliance with the transparency related provisions is not only valuable for shareholders, but that it might also be valuable for other stakeholder. For example competitors might derive valuable information about the firm’s strategy from it. Therefore transparency about the performance criteria and about the relationship between remuneration policy and the long term objectives of the firm might also lead to lower performance. Furthermore transparency about bonuses and severance payments being granted might result in commotion among customers, which can lead to lower performance. Transparency in the field of remuneration can have both a positive as well as a negative effect on firm performance. It is unclear which effect will be stronger. The direction of the relationship between compliance with the transparency related remuneration provisions and firm performance and firm value is unclear. A relationship is however expected to exist. This leads to the following hypothesis:
H3b: The level of compliance with the new transparency related provisions in the field of
remuneration is related with firm performance and firm value.
3.2.4 2 7Corporate Social Responsibility
23 cost differences are costs of pollution and wages paid by rich-country firms in poor countries. He states that CSR involves taking actions which reduce externalized costs or avoid distributional conflicts. Heal (2005) comes up with six benefits related to CSR investments, which can result in better firm performance. Among these benefits are lower risks of having to pay the social costs of pollution in the future and lower risks of conflicts with environmental organizations, improved relations with regulators which might be valuable in negotiations with them, improved employee productivity as employees like to work for ‘good’ companies, increased brand equity and lower costs of capital. Generating brand equity is important, because investments in CSR influence customer satisfaction, which in turn is positively related to market value (Luo and Bhattacharya, 2006).
Where Durnev and Kim (2005) do not find any evidence of a positive relationship between CSR investments and firm performance, other research indicates that CSR is profitable or at least does not lead to lower profits. Dasgupta et al. (2001, 2006) find in event studies in Argentina, Chile, Korea, Mexico and the Philippines that the market value of firms increased after positive firm news related to CSR, while it declined after negative CSR news. Dowell et al. (2000) found that firms adopting a stringent global environment standard have higher market values, measured by Tobin’s Q.
The main goal of introducing CSR in the Frijns code is that more attention will be paid to CSR. Higher investments in CSR will lead to a decrease in externalized costs and are therefore expected to be good for society. So according to stakeholder theory, CSR is a good addition to the code. According to the stakeholder theory of branding and corporate culture, paying attention to CSR might also be good for firm performance and firm value as it creates goodwill among shareholders, employees and customers. Other theories, for example Jensen’s enlightened stakeholder theory, argue that investments in CSR do not necessarily have to be good for companies. Too high investments in CSR might even lead to lower firm performance. Research into socially responsible investment funds has shown that these funds underperform compared to other investment funds, but still every year more money is invested in the socially responsible funds (Renneboog et al., 2008). So it seems like investors accept lower returns when investments are socially responsible. This leads to higher values of socially responsible firms compared to other firms with the same performance.
24 mentioned in the annual report. Since it does not mention how comprehensive this CSR section has to be, compliance differs from short sections of about 10 lines to very comprehensive CSR reports of dozens of pages. Since no data is available about the investments in CSR and CSR rankings, it is not possible to test whether high levels of compliance with CSR provisions are related with high CSR investments and high CSR rankings or whether compliance in this field is more a form of ticking the boxes behavior.1
2 Also no research into this relationship has been performed yet. Therefore it is not
possible to formulate hypotheses in this field. It is however still interesting to test whether a relationship between the level of compliance with the CSR provisions and firm performance and firm value exists.
3.2.5 2 8Diversity within the Supervisory Board
Different types of diversity are relevant for Dutch listed companies. Nationality diversity is relevant as most listed companies do not only operate in the Netherlands, while also their shareholders are mostly not only Dutch. Van Veen and Elbertsen (2008) find that as a result of governance differences between countries, the percentage of non-national directors among multinational companies differs. In the Netherlands they did however find a higher level of diversity than they expected based on the governance system. They believe that this might be explained by the integration of non-national directors in the so called ‘old boys network’. Among the large multinational companies in the Netherlands they find that the percentage of non-national directors was 44% in 2005. They find comparable percentages for both the management board and the supervisory board.
Gender diversity is relevant since women are still underrepresented in top-management and board functions, while empirical evidence suggests that adding women to boards leads to better performance. Adams and Ferreira (2004) argue that it leads to more cooperative boards and to boards with fewer attendance problems, which suggests that these boards can be more effective than homogeneous boards. Francoeur et al. (2008) support this. They find that a higher proportion of women officers leads to positive and significant abnormal returns for firms operating in complex environments. Besides, firms which appointed female directors to their board were more likely to have experienced bad performance in the period before the appointment than firms which appointed male directors. So to make their companies perform equally well, the boards with new female directors have to outperform the boards who appointed new male directors (Ryan and Haslam, 2005). The empirical research in this thesis studies only current firm performance and not current firm performance compared to performance in previous years. The fact that female directors
2
25 are more likely to be appointed in firms which experienced recent bad performance might weaken the relationship between gender diversity and firm performance found in this thesis.
Also age diversity might influence the performance and the value of the firm. Aratat et al. (2010) state that age diversity can influence the performance of the firm since different generations experience different social, political, and economic environments and events. They also state that, according to Vroom and Pahl (1971), some cognitive abilities diminish with aging and so does the willingness to take risks. According to Aratat et al. (2010), a diversified representation of generations may prevent group thinking and may lead to better performance by balancing the risk taking, associated with younger managers, and the cautiousness and risk averseness, as well as depth of experience, associated with older managers.
Some argue that the members of a diverse top management team might disagree more with each other (Eisenhardt et al., 1997). However, others suggest that when conflicts are managed well, they can lead to higher quality recommendations than consensus (Schweiger et al., 1986). It looks like mostly these conflicts are managed well, as empirical research provides evidence of heterogeneous boards which outperform homogeneous boards. Dallas (2002) argues that diverse boards, which in her paper are boards with more women, minorities and non-national directors, perform better than homogeneous board. She argues that advantages related to knowledge, perspective, creativity and judgment brought forward by a heterogeneous group, may be superior to the advantages related to smoother communication and coordination within more homogeneous groups. The relationship between board diversity and firm value and firm performance is examined by e.g. Carter et al. (2003) and more recently by Aratat et al. (2010). They find significant positive relationships between board diversity, monitoring intensity and firm performance and firm value among fortune 1000 (US) and ISE-100 (Turkey) firms. Carter et al. (2003) used the percentages of women, African Americans, Asians and Hispanics on the board of directors to measure diversity. Aratat et al. (2010) used gender and generation differences as observable attributes and directors’ educational backgrounds and nationality as proxies of non-observable attributes of values, beliefs, skills and competencies.
The agency theory also expects a positive relation between the level of compliance with the diversity related provisions and firm performance and firm value. The changes are introduced because they are expected to lead to a higher degree of independency of the supervisory board (Monitoring Commissie Corporate Governance, 2007). Independency is a crucial characteristic of a supervisory board according to agency theory. A more independent supervisory board is more likely to act in the best interest of the shareholders and thus be useful in contesting the agency problem.
26 H5: The level of compliance with the new provisions in the field of diversity is positively related to
firm performance and firm value.
Hypothesis 5 can be divided in the following two sub-hypotheses:
H5a: The level of compliance with the new provisions in the field of diversity is positively related to
the actual level of diversity within the supervisory board.
H5b: The level of diversity within the supervisory board is positively related to firm performance and
firm value.
3.2.6 2 9Communication with Shareholders
According to the monitoring committee, the management board and the supervisory board should serve the interests of the company and all its stakeholders, while the shareholders mainly serve their own interests (Monitoring Commissie Corporate Governance, 2009). These interests might differ, which can possibly lead to tensions between shareholders and the boards. For example, when a point for the agenda of the shareholders meeting is raised by a shareholder, this point is not necessarily in the long term best interest of the company and its other stakeholders. During the response time, which is introduced in the Frijns code, the boards can trade-off the interests of the different stakeholders to come to a sound reaction to the proposal of the shareholder. When the maximum response time and conditions for this response time are predetermined, this can improve the communication and decrease the tension between the boards and the shareholders.
The policy on bilateral contacts with shareholders is introduced in the Frijns code to decrease the tension between stimulating the dialogue between the company and its shareholders on the one hand, and on the other hand the regulatory requirements concerning market abuse. Having such a policy is expected to lead to a decrease in information asymmetry among shareholders and between minority shareholders and managers. This might lead to a better monitoring ability among shareholders and thus might force insiders to invest in projects with a positive net present value and to reduce perks and waste. So it can lead to better performance and as a result higher firm value. Findings of La Porta et al. (2002) also indicate that better protection of minority shareholders leads to higher firm value.
Overall, compliance with the new provisions related to communication with shareholders will probably lead to less tension between shareholders and managers and to a better monitoring ability for (minority) shareholders. Based on the agency theory and the findings of La Porta et al. (2002), the following hypothesis can therefore be formulated:
H6: The level of compliance with the new provisions in the field of communication with
27
4.
5Methodology and Sample Description
The analysis consists of two parts. First, a compliance analysis is performed. The aim of this analysis is to find out what the level of compliance is and how this level differs between companies of different sizes, between companies from different industries and between 2009 and 2010. Afterwards an analysis into the relationship between the level of compliance with the code and some of its elements on the one hand and firm performance and firm value on the other hand is performed. The remainder of this chapter first elaborates on the data which is used for the tests in section 4.1. In section 4.2 the performed tests are discussed.
4.1
1 8Data Description
4.1.1 3 0Data and Measures of Compliance
The data on the level of compliance with the Frijns code for both years in which this code was in use, 2009 and 2010, is retrieved from the Corporate Governance Insights Centre (CGIC). This research centre is part of the University of Groningen. It performs yearly research into the compliance with the Dutch corporate governance code among publicly listed firms. The CGIC collects its data by investigating the public documents of the companies, such as annual reports, remuneration reports, regulations for the supervisory board and the management board, and notifications on the corporate websites. In 2009, the CGIC investigated compliance for almost all provisions of the Frijns code of which compliance is observable. In 2010, some provisions, which all had a compliance score of at least 90 per cent in 2009, were not investigated by the CGIC. One of these provisions is a new provision in the field of communication with shareholders. To make sure that this does not influence the relationship between compliance and firm performance and firm value found in this thesis, compliance with this provision in 2010 is checked. Furthermore, the CGIC data does not include the element of the CSR provision which states that the main CSR issues have to be mentioned in the annual report. Therefore also data on this element is collected for the purpose of this thesis.
When no evidence is found on non-compliance, but also no explicit information about compliance is found, the provision is assumed to be complied with by the CGIC and a 1,0 score is given. This assumption can only be made for provisions on which compliance does not necessarily have to be observable. The reason that this assumption is made is that companies are obliged to explain their non-compliance in their annual reports. Table 1 gives an overview of all possible scores and their meaning.
28 1,0) are used as proxies for compliance with a provision. Besides, the combination of complied provisions (scores 1,1 and 1,0) and explained provisions (score 0,1) is used to measure whether a company fulfils its legal obligation to comply or explain.
In calculating the compliance levels with provisions in the specific fields, the sub-provisions as checked by the CGIC are used. In calculating the levels of compliance for the complete code and for all new provisions, the sub-provisions are merged into complete provisions to prevent overweighing fields in which provisions are divided into many sub-provisions, such as remuneration. Following the method of the CGIC, when not all sub-provisions of a provision are complied with or explained, a 0,0 score is given. When some sub-provisions are complied with and some are explained, a 0,1 score is given. When all sub-provisions are complied with, but not all explicit, a 1,0 score is given. Only when all sub-provisions are explicitly complied with, a 1,1 score is given.
4.1.2 3 1Data and Measures of Firm Performance and Firm Value
The data on firm performance and firm value is retrieved from Thomson’s DataStream. Return on assets and return on equity are used as proxies for firm performance and Tobin’s Q is used as a proxy for firm value. These proxies are widely used in corporate governance research (e.g. Aratat et al., 2010; Bauer et al., 2004; Black et al., 2006; Brown and Caylor, 2008; Carter et al., 2003; Core et al., 1999; Dahya and McConnell, 2007; De Jong et al., 2005; Durnev and Kim, 2005; Gompers et al., 2003; Klapper and Love, 2002; La Porta et al., 2002). Return on assets is calculated by dividing earnings before interest and taxes by the book value of total assets. Return on equity is calculated by dividing earnings before interest and taxes by the book value of total shareholders’ equity. The formulas can be represented as:
Where: i
RoA
= Return on assets in year i.Table 1
Possible Scores in the CGIC Data and Their Meaning Score Meaning of the score
-1 Provision is not applicable.
0,0 Provision is not complied with, there is also no explanation. 0,1 Provision is not complied with, but there is an explanation.
1,0 Provision is probably complied with, since there is no evidence found of non-compliance with the principle. This is interpreted as implicit compliance with the code.
1,1 Explicit compliance with provision.
29 i
RoE
= Return on equity in year i.i
EBIT
= Earnings before interests and taxes in year i.i
TA
= Average book value of total assets in year i.i
TE
= Average book value of total shareholders’ equity in year i.Tobin’s Q is the ratio of the market value of the assets of a company and the replacement values of these assets. Since these values are not available, Tobin’s Q is approximated. Just as in other corporate governance research, Tobin’s Q is approximated by taking the sum of the total assets and the market value of equity minus the book value of equity and divide this by the book value of total assets (La Porta et al., 2002; Durnev and Kim, 2005). The formula can be represented as:
Where:
j
Q
= Tobin’s Q at moment j.j
TA
= Book value of total assets at moment j.j
MVE
= Market value of equity at moment j.j
TE
= Book value of total shareholders’ equity at moment j.As almost all annual reports are published in the second quarter of a year, Tobin’s Q is calculated based on the market value of equity and the accounting values of assets and equity at the end of this quarter.
4.1.3 3 2Diversity Data
Diversity data of Dutch listed firms is not available in the Datastream database. This data is therefore collected from the annual reports and the corporate websites of the listed firms. The forms of diversity used in this thesis are gender diversity, age diversity and nationality diversity. As a proxy for gender diversity the percentage of female directors is used. This percentage is multiplied with 2. So when there are as many male directors as female directors, the percentage is 100%. As proxy for age diversity the standard deviation of the ages of the directors is calculated. This standard deviation is transcoded in a percentage with the highest measured standard deviation coded as 100% and a standard deviation of zero as 0%. As proxy for the nationality diversity the number of nationalities in a supervisory board minus one is divided by the number of members of this board minus 1. To calculate the combined level of diversity, the average of the three diversity measures is taken.
30 4.1.4 3 3Control Variables
The choice of control variables is based on previous empirical corporate governance research (e.g. Bhagat and Black, 2002; Durnev and Kim, 2005; Klapper and Love, 2004; La Porta et al., 2002; Renders et al., 2010) and on the theory presented in the previous chapter. Table 2 gives an overview of the control variables and their measures. Industry data (NACE Rev. 2 main section classification), as well as incorporation data, is retrieved from the Bureau van Dijk’s Orbis database. Data on the other control variables is retrieved from Thomson’s Datastream. Descriptive statistics of the control variables can be found in appendix 2. As a rule of thumb, the number of observations should be at least five times as large as the number of independent variables (Pelsemacker and Kenhove, 2006). The total number of independent variables is 17 if all these control variables are used. Therefore the required number of observations is 85. When the sets of provisions which are tested in both years are used, this prerequisite is easily met.
Firm size is included as a control variable because large firms may have greater agency problems than small firms as they are harder to monitor for shareholders. Therefore, they need stricter governance mechanisms to compensate (Klapper and Love, 2002). Besides, small firms may be more sensitive for the credit crisis than large firms. Firm Age is included because older firms tend to have higher book to market ratios (Durnev and Kim, 2005). This influences Tobin’s Q negatively and might influence returns on assets and equity negatively. Growth Opportunities are included since firms with high sales growth opportunities might have relatively low returns on assets and equity, while on the other hand, as the market might expect them to grow, their Tobin’s Q might be relatively high. Leverage is included because at firms where leverage is high, shareholders’ interests might be considered to be less important compared to interests of creditors. For these firms it is more important to satisfy creditors in order to keep interest levels low. Interests of creditors can differ
Table 2 Control Variables Control Variable Description
( ) Firm size. Measured by the natural log of the book value of total assets. ( ) Natural log of company’s age in years.
Growth opportunities, proxied by the average growth of sales over the 3 previous years, winsorized at 2% and 98%.
Leverage. Measured by book value of debt divided by book value of total assets. Ownership concentration of employees.
Ownership concentration of other major shareholders (at least 5% of shares).
Negative net income indicator
Industry indicator. Only for industries with at least 10 firm year observations. NACE Rev. 2 main section classifications are used.
Indicator of AEX membership
31 from interests of shareholders. For example, large creditors are very likely to receive information directly from firms. They are less dependent on annual reports and shareholder meetings. This might influence the relationship between compliance and firm performance. On the other hand, shareholders might consider compliance to be more important when leverage is high. High leverage mostly reduces the power of the shareholders within firms. Compliance can compensate for this. It can increase the monitoring ability and thereby decrease the risks of shareholders. This might influence the relation between compliance and Tobin’s Q. Similar to other corporate governance research leverage is measured by book value of debt divided by book value of total assets (e.g. Black et al, 2006; Dahya and McConnell, 2007). Ownership concentration of employees (including managers) is included because firms with high levels of employee ownership face lower agency problems between shareholders and managers. The interests of managers which are also shareholders are better aligned with the interests of other shareholders. Ownership concentration of other major shareholders is included because this type of shareholding is mostly strategic. Large strategic shareholders might retrieve information from the company in other ways than via corporate websites and official shareholder meetings. This can reduce the positive influence of transparency and corporate governance compliance on firm value, as the information is not new for major shareholders. Industry indicators are included because in different industries asset structure and accounting practices differ (Durnev and Kim, 2005). This leads to other levels of returns on assets and equity. Besides, performance might differ across industries due to different market situations. The AEX indicator is included because the demand for shares included in large indexes tends to be higher than the demand for other shares. This leads to higher firm values without better performance. Morck and Yang (2001) prove this for S&P 500 inclusion. The year indicator is included because the market situations differed in both years, which can lead to different market values and returns on assets and equity. The year indicator decreases the autocorrelation which otherwise could appear between the 2009 and the 2010 observations, respectively.
4.1.5 3 4Sample
32 The DataStream database does not contain performance data on all of these firm years. Firm years on which performance data is not available and firms which use other accounting principles than IFRS are excluded from the sample. Some of the remaining firms have different book years than the majority of firms. They are therefore excluded from the Tobin’s Q sample, as their firm values can only be measured at different times, which might lead to incomparable firm values. Also firm years of firms which are taken over in the next year are excluded from the Tobin’s Q sample, as these firms either do not have a market value being published anymore at the end of the second quarter or this market value is highly influenced by the rumours of a take-over. On some of the remaining firm years no EBIT data is available. These firm years are excluded from the RoA and RoE sample. One firm year has a negative book value of equity. This firm year is also excluded from the RoE and the Tobin’s Q sample. The final Tobin’s Q sample contains 177 firm years. The RoA and the RoE sample respectively contain 184 and 183 firm years.
4.2
1 9Tests
4.2.1 3 5Compliance Analysis
In the compliance analysis, the level of compliance with the Frijns code and different elements of this code is investigated. This analysis is performed as a background for the analysis of the relationship between on the one hand the level of compliance with the code and on the other hand firm performance and firm value. The sets of provisions for which the level of compliance is tested are presented in table 3. For these sets of provisions it will not only be tested what the level of compliance is, but also if this level of compliance differs between companies from different indexes, between companies from different industries and between 2009 and 2010. To analyze the differences between companies from different industries and from different indexes, independent samples t-tests are used. To analyze the difference between 2009 and 2010 paired samples t-tests are used. For this analysis only the firms of which compliance data from both years is available can be used. Only industries for which at least 10 firm years are available are included in this analysis.
Table 3
Compliance Levels Analyzed in the Compliance Analysis Field of Provisions Sets of Provisions to be Analyzed
Complete Code All provisions tested in both years and new provisions
Risk Management All provisions and new provisions
Remuneration All provisions and new provisions
Diversity New provisions
Corporate Social Responsibility New provisions
33 4.2.2 3 6Analysis of the Relationship between Compliance and Performance and Value After the level of compliance with the code is analyzed, the relationship between this level of compliance and firm performance and firm value is analyzed. To analyze this relationship, first correlation analyses are performed to explore whether the level of compliance and the performance and value of companies are related. Pearson correlation tests are performed for the sets of provisions which fulfill the requirement of a normal deviation or which fulfill this requirement after the exclusion of outliers. Pearson tests are performed because they have the highest explanatory power as they calculate the correlation between actual observations. To test whether variables are normally distributed Kolmogorov-Smirnov tests are used. For all other sets of provisions only Spearman correlation tests are performed. The Spearman test first ranks the observations and then calculates the correlation between the rankings. As a result this test does not require normal distributions and can be calculated for all sets of provisions, without excluding any observations.
To deeper investigate the relationship between the level of compliance and firm performance and firm value, regression analyses are performed afterwards. Firm performance and firm value are the dependent variables in the regression analyses. The independent variable is the level of compliance. The control variables are the variables mentioned in table 2. In both the correlation and the regression analyses, almost the same sets of provisions are used. These sets of provisions are presented in table 4. Because of the low numbers of observations, the sets containing provisions of which only data for 2009 is available are not used in the regression analyses. In the results section, the emphasis is on the sets containing only new provisions. The other sets of provisions are mainly tested as background tests and are presented in the appendices.
The regressions are controlled and adjusted for multicollinearity, heteroscedasticity and autocorrelation. Multicollinearity is tested by building correlation matrixes of the independent variables and by calculating variance inflation factors (VIF). There is no scientific consensus on a maximum level of correlation between independent variables in a linear regression. According to De Pelsemacker and Van Kenhove (2006), it is often stated as a rule of thumb that none of the correlation coefficients may exceed the value of 0.50 to prevent multicollinearity problems to occur.
Table 4
Sets of Provisions Analyzed in the Correlation and Regression Analyses Field of Provisions Sets of Provisions to be Analyzed
Complete Code All provisions tested in both years and new provisions
Risk Management New provisions
Remuneration All new provisions, new transparency related provisions and new non-transparency related provisions
Diversity New provisions
34 However, when sample size and R² increase, multicollinearity problems decrease. VIFs take this into account. Therefore VIFs are calculated for variables with correlation coefficients higher than 0.50. When VIFs are below 10, no control variables have to be adjusted or removed (Chen et al., 2011; Huizingh, 2010). To correct for potential heteroscedasticity, t-statistics are calculated based on White’s standard errors (White, 1980). Autocorrelation is tested by the Durbin-Watson statistic. Values close to two indicate that autocorrelation does not occur. When values below 1 or higher than 3 are found, corrections have to be made for autocorrelation (Huizingh, 2010).
35
5.
6Results
In this chapter the results from the empirical analyses are presented. The chapter begins with the results from the compliance analysis. Afterwards the results from the analysis of the relationship between the level of compliance and firm performance and firm value are presented. Because of the large number of tests which are performed, not all tables are presented in this chapter. Most attention is paid to the new elements in the code. Tables which are not presented in this chapter are presented in appendices.
5.1
2 0Compliance Analysis
This section presents the results of the compliance analysis. First, the overall compliance levels are given for each tested set of provisions. Afterwards, it is analyzed how the compliance levels differ between companies from different indices and between the different years, 2009 and 2010. For the sake of brevity the results related to the differences in compliance levels between industries are not presented in this section. In short, these results indicate that in the construction industry and in the financial industry, the compliance levels are higher than in the other industries. The lowest compliance levels are found in the information and communication industry and in the manufacturing industry. The results support the inclusion of industry dummies in the regression analyses, as significant differences in compliance levels between industries are found. Appendix 1 shows the compliance levels for all tested sets of provisions, per index, per industry and per year. It also shows whether these compliance levels differ significantly from the other firms. In this section only the most important results are presented.
5.1.1 3 7Overall Compliance levels
The overall levels of compliance with the Frijns code and the tested sets of provisions are presented in table 5. The first thing that stands out is that on average 14.19% (i.e. 100% - 85.81%) of all provisions which are tested in both years are not completely complied with and not explained. Only one firm fulfills the legal obligation to give explanations for all the provisions which they do not comply with in both years. One other firm fulfills this obligation only in one year. For the new provisions the level of compliance or explanation is slightly lower. Of these provisions on average 16.02% (i.e. 100% - 83.98%) of the provisions are not completely complied with or explained. Also only one firm complies with the legal obligation in both years for the new provisions. Six other firms fulfill this obligation in one of the years.
