E FFECTIVENESS : E VIDENCE FROM THE

M ASTER T HESIS :

B USINESS A DMINISTRATION

S PECIALIZATION : F INANCE

S UPERVISORY B OARD C HARACTERISTICS AND

E FFECTIVENESS : E VIDENCE FROM THE

N ETHERLANDS

U

NIVERSITY OF

G

RONINGEN

F

ACULTY OF

E

CONOMICS

& B

USINESS

A

UTHOR

: M

AARTEN

S

PIT

(1274201) F

IRST SUPERVISOR

: H. G

ONENC

S

ECOND SUPERVISOR

: J.H.

VON

E

IJE

JEL

CODES

: G34, M19.

K

EY WORDS

:

SUPERVISORY BOARD CHARACTERISTICS

,

SUPERVISORY BOARD

EFFECTIVENESS

,

FIRM PERFORMANCE

, CEO

COMPENSATION

,

TWO

-

TIER STRUCTURE

.

A

BSTRACT

This study examines the relationships between supervisory board characteristics and supervisory board effectiveness. Previous literature provides ambiguous results and primarily concerns one-tier boards. We examine whether independence, busyness, size and ownership of Dutch (two-tier) supervisory boards influence the supervisory board’s effectiveness. We find no evidence that supervisory board ownership influences supervisory board effectiveness. We find some evidence that larger supervisory boards are less effective. Firms with larger supervisory boards have lower operating performance and pay higher CEO compensation. We find consistent evidence that more independent supervisory boards are more effective. Higher supervisory board independence is associated with higher firm perfomance. We find consistent and convincing evidence that ‘busy’ supervisory boards are more effective. A higher ratio of supervisory board members holding multiple board positions is associated with higher firm performance. Our results indicate that further research is needed before additional legislation on composition of (two-tier) boards can be justified.

T

ABLE OF

C

ONTENTS

I

NTRODUCTION

...3

1. C

ORPORATE

G

OVERNANCE

...6

1.1 B

OARD OF DIRECTORS

...8

1.1.1 T

HE

S

UPERVISORY

B

OARD IN THE

N

ETHERLANDS

...9

1.1.2 S

UMMARY

...11

2. B

OARD

C

HARACTERISTICS

& B

OARD

E

FFECTIVENESS

...12

2.1 B

OARD

I

NDEPENDENCE

...12

2.2 B

OARD

B

USYNESS

/ M

ULTIPLE

B

OARD

P

OSITIONS

...14

2.3 B

OARD

S

IZE

...16

2.4 B

OARD

O

WNERSHIP

...18

3. B

OARD

E

FFECTIVENESS

, F

IRM

P

ERFORMANCE

& CEO C

OMPENSATION

...19

3.1 F

IRM

P

ERFORMANCE

...19

3.2 CEO C

OMPENSATION

...19

4. H

YPOTHESES

...21

5. M

ETHODS

...22

5.1 S

AMPLE

...23

5.2 S

UPERVISORY

B

OARD

C

HARACTERISTICS

...24

5.3 F

IRM PERFORMANCE

...26

5.4 CEO C

OMPENSATION

...29

6. R

ESULTS

...34

6.1 S

UMMARY

...44

7. C

ONCLUSIONS

...45

8. D

ISCUSSION

...47

R

EFERENCES

...48

A

PPENDIX

I C

ORPORATE

G

OVERNANCE IN THE

N

ETHERLANDS

...54

A

PPENDIX

II R

ELEVANT

A

RTICLES FROM THE

DCGC ...58

A

PPENDIX

III I

NDUSTRY

I

NFORMATION

...60

A

PPENDIX

IV S

TATISTICAL

T

ESTS

...64

I

NTRODUCTION

The recent financial and economic crisis has increased the discussion on two important aspects of corporate governance; the functioning of the board of directors in supervising management and the compensation practices used in motivating management. The excessive amounts of compensation that managers, and particularly Chief Executive Officers (CEOs), received despite poor firm performance or even corporate failure, has once again spurred the discussion on the functioning of boards of directors and their role in setting CEO compensation. Although boards of directors are supposed to act in the best interests of investors, board members have often been accused of neglecting their supervisory role.

Several authors have argued that boards of directors fail to fulfil their role as critical supervisor (Mace, 1971, Lipton & Lorsch, 1992, Jensen, 1993, Bebchuck & Fried, 2004). The criticism is usually directed at certain board characteristics that are argued to lead to board ineffectiveness. Boards of directors are often too large, lack incentives, lack independence of management and lack enough time to perform their supervisory role, according to these critics. These aspects of boards of directors will lead to board members being either unable or unwilling to adequately perform their supervisory role. Inadequate supervision by the board of directors will in turn lead to lower firm performance and the extraction of additional compensation by management (Bebchuck & Fried, 2004).

This criticism has led to an abundant amount of studies that examine the desirability of different board characteristics. Frequently studied board characteristics are board independence (Weisbach, 1988, Rosenstein & Wyatt, 1990, Boyd, 1995, Dailey, Johnson, Ellstrand & Dalton, 1995, Millstein & MacAvoy, 1998, Bhagat & Black, 2000, Klein, 2002), board busyness (Core, Holthausen & Larcker 1999, Ferris, Jagannathan & Pritchard, 2003, Harris & Shimizu, 2004, Jiraporn, Davidson, Dadalt & Ning, 2009, Fich & Shivdasani, 2006), board size (Goodstein, Gautam & Boeker, 1994, Yermack, 1996, Eisenberg, Sundgren &

Wells, 1998, Mak & Kusnadi, 2005, ) and board ownership (Morck, Shleifer & Vishny, 1988, Mehran, 1995, Harvey & Shrieves, 2001). Board independence concerns the relation that non- executive board members have with the firm and its management. Independence from management is a prerequisite for non-executive board members to be critical of management according to Fama and Jensen (1983) and Jensen (1993). However, Boot and Macey (2003) argue that when the distance between management and non-executive board members becomes too large this will lower the board’s ability to make informed decisions. Board busyness reflects the activities that non-executive board members employ outside their board position. The busyness hypothesis suggests that when non-executive board members have too many additional duties (other executive or non-executive positions) this will hurt their ability to adequately perform their supervisory duties. However, non-executive board members that

hold multiple other board positions may also be a valuable resource to a firm’s board (Fama and Jensen, 1983). Board size is simply the number of board members that sit on a board.

Several authors (Lipton & Lorsch, 1992, Jensen, 1993) have argued that boards are generally too large to function effectively. Finally, board ownership concerns the equity stakes that board members have in the firm they have to supervise. The absence of significant ownership by (non-)executive board members may make them indifferent to whether or not investors’

interests are pursued.

The results of studies that examine these characteristics are mixed and have not provided evidence that led to consensus concerning the desirability of different board characteristics. In addition, the studies on the relation between board characteristics and board effectiveness were pre-dominantly performed on firms with a one-tier structure. In a one-tier structure the board of directors is comprised of both executive and non-executive board members. In a two-tier structure the board of directors is split up into two different bodies; the management board, consisting solely of executive board members, and the supervisory board, consisting solely of non-executive board members. In countries in which a two-tier structure is dominant, such as the Netherlands, the research on the influence of board characteristics on board effectiveness is scarce. Van Ees, Postma and Sterken (2003) study the influence of size, ownership, remuneration and independence of Dutch supervisory boards on corporate performance. They find a negative association between supervisory board size and firm performance, which is in line with the argument that boards are often too large. However, they find no evidence that supervisory board remuneration or supervisory board ownership influences firm performance. In addition, they find that supervisory board independence is negatively related to firm performance, which contradicts the criticism that board independence is generally too low.

Although the evidence concerning the relations between previously mentioned board characteristics and board effectiveness is ambiguous, some of the criticism on board characteristics has grown into conventional wisdom. The criticism has often been used by regulators and legislators to try to improve the functioning of the board of directors. In the Netherlands, ‘The Dutch Corporate Governance Code’¹ (2008) (hereinafter referred to as ‘the DCGC’) provides specific best practice provisions on several supervisory board characteristics. Among other things, the DCGC recommends limiting the number of

1 The Dutch Corporate Governance Code provides principles of good corporate governance and best practice provisions. It was drawn up in 2003 at request of, among others, the Netherlands Centre of Executive and Supervisory Board members (NCD), the Foundation for Corporate Governance Research for Pension Funds (SCGOP), the Association of Stockholders (VEB), the Association of Securities-Issuing Companies (VEUO) and the Confederation of Netherlands Industry and Employers (VNO-NCW) at the invitation of the Minister of Finance and the Minister for Economic Affairs. In 2008 the DCGC was updated to account for changes in concepts of corporate governance. None of the articles or citations presented in this paper was changed during the 2008 update.

dependent board members to one dependent board member per supervisory board (2008, III.2.1), which, according to findings by Van Ees et al. (2003), may actually hurt board effectiveness. The DCGC also recommends limiting the number of supervisory board positions per supervisory board member to a maximum of five, to assure “the proper performance of his duties” (2008, III.3.4). It thereby implicitly assumes the busyness hypothesis to be correct and rejects the notion that supervisory board members holding more than five board positions may be a valuable resource. Although the DCGC works according a comply-or-explain principle, some politicians have increased pressure to make compliance mandatory. Although mandatory compliance to the DCGC has not yet been realized, this political pressure led to the partial approval of a bill by Dutch parliament in December 2009 that will legally restrict the number of supervisory board positions to a maximum of five (two for non-retired board members) per board member². These actions clearly indicate the conviction of legislators that (some of) the previously mentioned board characteristics are related to board effectiveness and that legislation on these characteristics will increase board effectiveness. We question whether these interventions in board composition are justified.

We extend research on the influence of supervisory board characteristics on supervisory board effectiveness in a two-tier structure. The main hypothesis of this paper is dual. The first part of our main hypothesis is that supervisory board independence, busyness, size and ownership influence supervisory board effectiveness. The second part of our main hypothesis is that supervisory board effectiveness is positively related to firm performance and negatively related to the amount of CEO compensation. Several authors have argued and found that board ineffectiveness will lead to lower firm performance and higher CEO compensation. We examine the relationships between independence, busyness, size and ownership of Dutch supervisory boards and three measures of firm performance; Tobins’s Q, as a measure of firm valuation, industry-adjusted ROA, as a measure of firm profitability and market-adjusted stock returns, as a measure of overall firm performance. In addition, we examine the relationship between those supervisory board characteristics and the amount of CEO compensation. Our sample consists of 208 observations over a three-year period (2004 – 2006) for 79 Dutch listed firms (AEX, AMX, and AscX) that apply a two-tier structure.

We find no evidence that supervisory board ownership is related to firm performance or CEO compensation. Our findings do not support the argument that ownership stakes by non-executive board members increase their propensity to perform their supervisory duties effectively. We find some evidence that higher supervisory board independence raises supervisory board effectiveness. Higher supervisory board independence is associated with higher firm valuation, as measured by Tobin’s Q. In addition, although not statistically

2 Upon finishing this paper the bill had already been approved by part of Dutch parliament (Tweede Kamer) but still had to be approved by the Senate (Eerste Kamer).

significant, supervisory board independence is positively related to several measures of subsequent firm performance. Our findings seem to suggest that board independence, even for higher levels of board independence, is still weakly positively related to board effectiveness.

Our findings support regulation by the DCGC that limits dependent board members to a maximum of one per supervisory board. We find some, but not very robust, evidence that larger supervisory boards are less effective. We find that firms with larger supervisory boards have significantly lower subsequent operating performance and pay their CEOs significantly higher compensation. Finally, concerning supervisory board members holding multiple board positions, we find consistent evidence opposing the busyness hypothesis. We find that firms with supervisory boards with a higher ratio of board members holding multiple board positions have higher firm valuations, higher subsequent operating performance and higher subsequent stock performance. Our findings suggest that board members holding multiple board positions may be a valuable resource, which is in line with findings by Ferris et al.

(2003) and Mak and Kusnadi (2004). Our findings do not support Dutch legislation that limits the number of supervisory board positions a board member is allowed to have to a maximum of five.

This study contributes to literature on the desirability of different board characteristics and therefore on the desirability of legislation on board composition. It extends this research to two-tier boards. In addition, it provides insights into the determinants of CEO compensation in a two-tier structure. The next section reviews the practice of corporate governance and elaborates on the board of directors as corporate governance mechanism.

Section 2 discusses the rationale behind the supposed relations between supervisory board characteristics and supervisory board effectiveness. Section 3 describes why supervisory board effectiveness is hypothesized to be positively related to firm performance and negatively related to the amount of CEO compensation. Section 4 combines sections two and three into testable hypotheses for this study. Section 5 describes the methodology and data used in this study. Sections 6 and 7 present the results and the conclusions following from these results. The final section addresses any possible limitations of this study in the form of recommendations for future research on this topic.

1. C

ORPORATE

G

OVERNANCE

In corporations agency problems are caused by the separation of ownership and control. One of the first books to describe the possible problems caused by the separation of ownership and control was “The modern corporation and private property” by Berle and Means (1932). They predicted that firms would grow bigger and bigger and would become dependent on public markets for capital, thus creating a separation of ownership (shareholders) and control (managers). They also predicted that this separation would result in conflicting interests

between managers and shareholders. Since shareholders are often dispersed and have diversified portfolios, the gains that shareholders receive from ownership are primarily financial. The manager, however, derives his utility from financial gains and from various non-financial aspects of his managerial activities (Jensen & Meckling, 1976). These non- financial gains can take lots of different forms, for example, leisure, workload, the prestige associated with firm size or a specific project, the challenge, power, the working environment, personal relations with employees, a larger than necessary computer to play with, job retention, etc. The actions that managers want to take to maximize their own utility are therefore not likely to coincide with actions that maximize firm value³. These conflicting interests, combined with the fact that shareholders are unable to completely observe the manager’s behaviour, give rise to a moral hazard problem.

Tirole (2006) broadly divides moral hazard problems or ‘shirking’ into four different categories; insufficient effort, extravagant investments, entrenchment strategies, and self- dealing. Insufficient effort primarily refers to the inefficient allocation of work time to various tasks. Managers may allocate their time to the activities of their liking instead of to the activities that most require their attention. Extravagant investments are investments that give the manager a certain prestige but do not maximize firm value. Managers may reject to distribute excess cash when the firm does not have profitable investment opportunities (Jensen, 1986) in order to engage in empire building (Jensen & Meckling, 1976).

Entrenchment strategies may be used by managers to make dismissal by shareholders in case of poor performance more difficult. For example, managers may invest in certain activities for which their expertise is indispensable (Shleifer & Vischny, 1989). Finally, self-dealing concerns the consumption of private benefits or perks. Investors now have to deal with how to make sure that managers will take those actions that maximize firm value instead of those actions that maximize managerial utility.

The practice that deals with the ways in which suppliers of finance to corporations assure themselves of getting a return on their investment is called corporate governance (Shleifer & Vischny, 1997). This definition comprises both equity and debt holders as suppliers of finance. The primary concern of corporate governance is to protect suppliers of finance from exploitation or expropriation by a firm’s management. Two common approaches to corporate governance both rely on giving investors some power (Shleifer & Vischny, 1997). The first approach is legal protection to investors from expropriation by managers. For example, investors can be given the right to decide on important aspects such as, shares issues, acquisitions, strategy changes, etc. In addition, the legal system may allow investors to

3 Throughout this paper, ‘firm value’ will be meant to capture various possible measures of firm success, for example, return to investors, profitability, successful execution of the firm’s strategy, survival.

sue management when actions were deliberately taken to benefit management at the expense of investors. The second approach is concentrated ownership by large investors, combining significant control rights with significant cash flow rights. A large investor has higher incentives to monitor management and can more easily discipline management, making it less likely for management to act against investors’ wishes. Becht, Bolton and Röell (2002) identify three more corporate governance mechanisms; hostile takeovers and proxy voting contests, alignment of managers’ and investors’ interests through managerial compensation, and delegation of control to the board of directors. Hostile takeovers and proxy fights (temporarily) concentrate ownership or voting power to intervene in management’s actions.

This mechanism is most likely to be used when managerial performance is low and there is no controlling shareholder to discipline the manager. By tying managerial compensation to firm performance, managers can be motivated to maximize firm value. Finally, when investors delegate control to the board of directors, the non-executive board members are supposed to monitor and assist management in decision-making, while serving investors’ best interests.

1.1BOARD OF DIRECTORS

The board of directors is often described as the most important corporate governance mechanism. Jensen (1993, p.862) states that the board of directors is “…at the apex of internal control mechanisms”. In the last decades, the pursuit of diversification by investors has led to an increased dispersion of ownership. This has led to a free-rider problem concerning the collection of information about managerial performance by investors, which has increased the importance of the board of directors (Hermalin & Weisbach, 2003). Although boards of directors are required by law in most developed countries nowadays, they predate regulation and are part of the market solution to the contracting problems inside most organizations (Hermalin & Weisbach, 2003).

The duties of the board of directors can be largely divided into two different roles; the monitoring role and the service role (Van Ees & Postma, 2004). The board of directors is responsible for monitoring managers’ actions to make sure they do not benefit themselves at the expense of investors. By appointing a board of directors the investors thereby reduce some of the aspects leading to agency problems. By monitoring the manager, the board of directors is able to judge whether the manager shirks or not. This will increase the managers’

propensity to exhibit good behaviour, reducing the conflict of interests. The board of directors is also responsible for designing managerial compensation to further align managers’ and investors’ interests, for which purpose it often appoints a separate remuneration committee⁴ (RC). Because the board of directors gains superior information about the firm and its

4 Also often referred to as compensation committee.

manager, they will be better at designing a compensation contract that will further reduce the conflicts of interests than usually dispersed shareholders. This role of the board of directors is often referred to as the monitoring or governance role. The service role of the board of directors refers to the role of board members in supporting management. Through their own knowledge, experience, connections or other resources, the board members can assist management in decision-making and managing the company.

Most of the empirical work this study builds on was done in Anglo-Saxon countries in which a one-tier structure prevails. Because this study extends this research to a country with a two-tier structure, the Netherlands, a brief evaluation of board structure and duties is required. A more general description of corporate governance in the Netherlands can be found in Appendix I.

1.1.1THE SUPERVISORY BOARD IN THE NETHERLANDS

In the Netherlands the two-tier structure is the prevalent structure for the board of directors.

Although listed firms can choose between a one- and two-tier structure, practically all listed firms apply a two-tier structure⁵. A two-tier structure implies that the board of directors is split up into two different bodies; the management board (raad van bestuur), consisting solely of executive board members, and the supervisory board (raad van commissarissen), consisting solely of non-executive board members. In a one-tier structure, the board of directors consists of both executive and non-executive board members. Jensen (1993) suggests that a one-tier structure creates a board culture that decreases the propensity of non-executive board members to be critical of executive board members. In addition, a one-tier structure creates the somewhat perverse situation in which executive board members are partially responsible for duties of the entire board, such as monitoring and evaluation of their own performance or designing executive compensation. Several authors (Mace, 1971, Fama & Jensen, 1983, Jensen, 1993) have argued that an arm’s length relationship is essential for non-executive board members to be critical of executive board members. A two-tier structure may create or strengthen this arm’s length relationship. According to Cadbury (1995, p.66), the two-tier structure allows the supervisory board “…to take an entirely independent view of the reactions of management, since there is no overlap of membership between the two boards”.

However, Maassen and Van den Bosch (1999) investigate board practices in the Netherlands and argue that, although formal board structure suggests independence between the management board and the supervisory board, board practices are largely similar to board practices in a one-tier structure. Finally, a one-tier structure allows for the CEO and the chairman of the board to be the same person (CEO duality). The chairman of the board

5 In our initial sample of 143 Dutch listed firms only 4 applied a one-tier structure.

largely determines the subjects discussed in boardrooms and the information provided to board members (Jensen, 1993, Lipton & Lorsch, 1992). When the CEO is the chairman of the board he may abuse this position to maintain his information advantage over non-executive board members and influence decision making in boardrooms.

The supervisory board has historically been a powerful institution in the Netherlands.

When firms reach a certain size in the Netherlands, they are required to assume a certain corporate form, namely the structured regime. Prior to 2004, the adoption of the structured regime required the supervisory board to take over several powers from shareholders, such as, establishment and approval of the annual accounts, the election of the management board and of the supervisory board itself (called co-optation). Van Ees and Postma (2004) argue that the practice of co-optation in Dutch supervisory boards reduces the formal independence of board members. The supervisory board also acquired authority over major decisions made by the management board. In 2004 a legislative change was made (Wet Wijziging Structuurregeling) that largely granted these rights back to shareholders. However, the supervisory board has kept a powerful position. All major managerial decisions, such as, large acquisitions, selling large parts of the company, or entering into or breaking large contracts, still require the consent of the supervisory board. The supervisory board can also approve the use of defense mechanisms by management⁶. Only when a large part of shareholders disagree with actions that are approved by the supervisory board are they able to influence those actions⁷.

The supervisory board has the task to monitor the policy of management and the general course of business of the firm and the supervisory board acts in the best interest of the firm, while performing its task (section 2, art. 140, Book 2, Dutch Civil Law). The DCGC presents a similar description of the role of the supervisory board. It states that the management board and the supervisory board have the responsibility to weigh the interests of different stakeholders, generally with a view to ensuring the continuity of the enterprise. In doing so, the company endeavours to create long-term shareholder value (DCGC, 2008, p.6).

The supervisory board is also responsible for the design of managerial compensation and often appoints a separate RC for this task. According to the DCGC and Dutch Civil Law, the RC is responsible for:

a) drafting a proposal to the supervisory board for the remuneration policy to be pursued; b) drafting a proposal for the remuneration of the individual members of the management board; such proposal shall, in any event, deal with: (i) the remuneration structure and (ii) the amount of the fixed remuneration, the shares and/or options to

6 A description of these defense mechanisms can be found in Appendix I.

7 A firm’s articles of incorporation may require up to a two-third majority of votes of the shareholders meeting when deciding on important matters.

be granted and/or other variable remuneration components, pension rights, redundancy pay and other forms of compensation to be awarded, as well as the performance criteria and their application; c) preparing the remuneration report (DCGC, art III.5.10)⁸.

In the absence of a separate RC, the entire supervisory board is responsible for the previously mentioned duties (DCGC, III.5, principle). Although shareholders have to approve the design of executive compensation contracts, this is considered to be a formality and the supervisory board (or RC) determines the conditions for executive compensation⁹. Although legislation is primarily directed at the monitoring role of the supervisory board in the Netherlands, Van Ees and Postma (2004, p.92) analyse the role of boards in the Netherlands and find that boards

“…tend to articulate the … service role over the monitoring role”. Both the monitoring role and the service role therefore are important for the supervisory board in the Netherlands.

1.1.2SUMMARY

The board of directors has the duty to monitor the firm and its manager to make sure investors’ interests are pursued. The board of directors is responsible for optimally designing managerial compensation to further align managers’ and investors’ interests. In addition, the board of directors can assist management in decision-making through their service role. In the Netherlands practically all firms apply a two-tier structure, in which the board of directors is split up into two different bodies; the managing board, consisting solely of executive board members, and the supervisory board, consisting solely of non-executive or supervisory board members. Although this separation may create an arm’s length relationship that is essential

8 The adoption of the best practice provisions of the DCGC is not mandatory since the code works according to a comply-or-explain principle. However, some of the best practice provisions are also incorporated in Dutch Civil Law. The adoption of best practice provision III.5.10 of the DCGC is therefore mandatory, since it is enforced by Dutch Civil Law (Book 2, art. 135, 138 c,d,e).

9 In 2008, the shareholders of Dutch based electronics company Royal Philips rejected new pay packages for its top management, proposed by the supervisory board. The proposed compensation was rejected because it allowed for managers to be awarded with stock options in case of poor performance as well. Although it may seem obvious to reject such a proposal, up until then, similar proposals had always been adopted. This was one of the first instances in which shareholders requested and accomplished a proposal change for CEO compensation.

http://www.nrc.nl/economie/article1874407.ece/Beloning_van_top_Philips_afgewezen

In the beginning of 2009, Dutch based Oil Company Royal Dutch Shell (RDS) awarded its CEO, Jeroen van der Veer, a bonus of approximately €2 million over 2008 for which conditions were not met. RDS had to finish third or higher in a peer group of five companies (including RDS) for the CEO to receive his bonus. Although RDS finished fourth, Van der Veer still received 85% of his bonus.

RDS’ remuneration committee (RC) thought that Van der Veer’s performance was good enough for him to receive most of his bonus. Although criticised by RDS’ shareholders, the RC used its right to intervene in the ex post determination of compensation and to deviate from the ex ante made agreements. This example is representative for the role of the supervisory board and RC in setting CEO compensation in the Netherlands.

http://www.veb.net/content/HoofdMenu/Beurs/Aandeelhoudersvergaderingen/Avaartikelen/ava2009/av ashell2009.aspx

for non-executive board members to be critical of executive board members, several authors have argued that board practices and culture in a two-tier structure hardly differ from those in a one-tier structure. The supervisory board has a powerful position in the Netherlands. This powerful position is exacerbated by the supervisory board’s control over managerial compensation. Our evaluation suggests that the role and power of the supervisory board in monitoring and assisting management and setting managerial compensation is similar to that of boards of directors in Anglo-Saxon countries.

2. B

OARD

C

HARACTERISTICS

& B

OARD

E

FFECTIVENESS

Critics of board functioning have often argued that boards are not independent of management, that board members are often too busy and lack incentives to perform their duties adequately, and that boards are too large to function effectively. These characteristics are argued to lead to board ineffectiveness, making boards unable or unwilling to adequately perform their supervisory role¹⁰. Although the criticism has often been presumed to be correct, some authors have argued against this criticism, reasoning that the relation between a certain board characteristic and board effectiveness does not exist or may sometimes be opposite. This section discusses the rationale behind the supposed relationships between board characteristics and board effectiveness and provides the evidence on these relationships.

In addition it presents the Dutch regulation or legislation on the different board characteristics.

2.1BOARD INDEPENDENCE

Board independence refers to the independence by non-executive or supervisory board members from management. Several authors have argued that board independence is positively related to board effectiveness. They argue that independence is a prerequisite for non-executive board members to be critical of management and to perform their supervisory role well. A non-executive board member is usually considered to be independent when he or she meets several formal criteria, such as, not being related to any of the executive board members, not having a business relationship with the firm (other than the board position), not having been employed by the firm him- or herself in the recent past, etc. Executive board members, obviously, are never considered to be independent from management, since they are part of management. Fama and Jensen (1983) suggest that independent non-executive board members are more active in monitoring. Independent non-executive board members have incentives not to collude with managers and to make good decisions in order to preserve

10 ‘Supervisory role’ is used in this study as a comprehensive term that comprises both the monitoring role and the service role.

their value as supervisors in other contexts. Jensen (1993) argues that an arm’s length relationship is essential for non-executive board members to be critical of management.

Van Ees and Postma (2004, p.95) argue, based on Boot and Macey (2003), that an effective board is “…both well informed and capable of making painful decisions with respect to management”. Although the latter may increase with board independence, the first may require a certain proximity to management, according to them. Thus a trade-off between being well-informed and being objective may exist with respect to board independence. The presence of some dependent non-executive board members on the board may therefore be preferable to a board on which all non-executive board members are independent.

Several studies have examined the relationship between board independence and board effectiveness but found mixed evidence. Weisbach (1988), studying one-tier boards, finds that CEOs that perform poorly are more likely to be replaced in firms with a majority of non-executive board members, implying that non-executive board members are important in disciplining the CEO¹¹. Core et al (1999), studying 205 listed US-firms, find that firms with boards with more dependent non-executive board members pay higher levels of CEO compensation, which in turn leads to poor firm performance. Newman and Mozes (1999) find that CEOs of firms with dependent non-executive board members on the remuneration committee (RC) receive preferential conditions concerning their compensation.

CEOs of firms with dependent non-executive board members on the RC do not earn significantly greater compensation but the relation between CEO compensation and performance is biased in the CEO’s favour at shareholders’ expense. These findings suggest that independent board members are better supervisors than dependent board members.

Klein (2002) studies the influence of board independence on the degree of earnings management for 692 firm years (1992 – 1993) of large US listed firms with one-tier boards.

She finds that firms with a majority of independent board members¹² exhibit significantly lower degrees of earnings management, which also suggests that independent board members are more effective supervisors. However, Klein (2002) finds that when boards increase their independence from a majority of independent board members towards a board on which all non-executive board members are independent, the degree of earnings management does not decrease further. Her findings suggest that a majority of independent board members is preferable to a minority of independent board members, but that a board with solely independent non-executive board members is not preferable to a board with a majority of independent board members and some dependent non-executive board members. Van Ees et

11 Please note that Weisbach (1988) does not differentiate between independent and dependent non- executive board members.

12 Please note that, since Klein (2002) studies one-tier boards, a majority of independent board members refers to the situation where there are more independent non-executive board members than there are dependent non-executive board members and executive board members combined.

al. (2003) study the influence of supervisory board independence on firm performance for 94 Dutch listed firms. They find a negative relation between supervisory board independence and firm performance. Their findings seem to confirm findings by Klein (2002) that increasing board independence when a majority of the board is already independent¹³ does not increase, or may even hurt, board effectiveness.

Dailey et al. (1998) and Anderson and Bizjak (2003) study compensation committee independence and find no evidence that more dependent compensation committee members leads to higher levels of, or changes in, CEO compensation. Bhagat and Black (2000) perform a longitudinal study on the relation between board independence and long-term firm performance for 934 of the largest US firms (1985 – 1995). They find no evidence that higher board independence is associated with better long-term firm performance. They do find that firms suffering from low profitability increase board independence. However, increasing board independence does not lead to increased profitability.

The findings on the association between board independence and board effectiveness are mixed. In general, findings seem to suggest that a majority of independent board members is preferable to a minority of independent board members. Beyond a majority of independent board members, the relation between board independence and board effectiveness becomes unclear. This is line with a trade-off between proximity to management and objectivity towards management (Boot & Macey, 2003). However, the authors of the DCGC primarily stress the importance of objectivity towards management. It states that “…the composition of the supervisory board shall be such that the members are able to act critically and independently … of the management board…”,(DCGC III.2). It provides a best practice provision on the independence of board members that states that all supervisory board members, with the exception of not more than one person, shall be independent¹⁴ (DCGC III.2.1). The authors of the DCGC thereby assume that in practice more independent supervisory board members will always be preferable to less independent board members.

2.2BOARD BUSYNESS /MULTIPLE BOARD POSITIONS

There have been several studies linking board busyness to board effectiveness. Board busyness concerns the activities that board members employ outside their board position.

Whether a board member is considered to be busy usually depends on the number of board positions (also referred to as directorships) he holds and on whether the board member is retired. The busyness hypothesis suggests that non-executive board members that hold multiple board positions will not be able to produce the required effort in monitoring the firm

13 Mean and median percentage of independent board members in the sample of van Ees et al. (2004) are 84.3 and 100, respectively.

14 The criteria that a board member has to meet to be considered independent are presented in the DCGC, art III.2.2, which is presented in Appendix II.

and its management. Since board members have limited time to allocate to their supervisory roles, holding multiple board positions will imply that less time is available for each board position. The busyness hypothesis suggests this will reduce the ability of non-executive board members to adequately perform those supervisory roles. In addition, Lipton and Lorsch (1992) argue, when board members hold multiple board positions, the importance to them of adequately fulfilling one particular board position may decrease.

Core et al (1999), consider non-executive board members to be busy when they hold three or more (six or more for retired board members) board positions. They find that busy non-executive board members pay higher levels of CEO compensation, which in turn leads to poor firm performance, implying that ‘busy’ board members are less effective in performing their supervisory role. Jiraporn et al. (2009) find that non-executive board members holding a higher number of board positions exhibit a higher tendency to be absent from board meetings, which is in line with the busyness hypothesis. Fich and Shivdasani (2006), studying the monitoring effectiveness of busy boards (a board in which the majority of non-executive board members hold three or more board positions), find that busy boards are associated with weak corporate governance. They find that firms with busy boards exhibit lower market-to- book ratios, weaker profitability, and a lower sensitivity of CEO turnover to firm performance.

However, Fama and Jensen (1983) suggest that multiple board positions can signal higher board member quality. The appointment to multiple board positions might be the result of superior performance of the firms on which boards the board members previously served.

They argue that board members holding multiple board positions will work harder to fulfill their supervisory role. In addition, board members holding multiple board positions are likely to have larger social networks, more experience with board practices and superior access to valuable information, which can be used in assisting management (Harris & Shimizu, 2004).

Ferris et al. (2003), investigating the likelihood of securities fraud litigation, find no evidence that busy non-executive board members harm subsequent firm performance. They consider board members to be busy when they hold three or more board positions¹⁵. In addition they find that busy non-executive board members serve on more board committees and attend more committee meetings, which is inconsistent with the assumption that busy board members neglect their supervisory role. Harris and Shimizu (2004) study the influence of board busyness on the outcomes of 143 acquisitions and find evidence contradicting the busyness hypothesis. They find that board busyness, using the same measure as Core et al.

(1999), increases acquisition performance.

15 Ferris et al. (2003) do no differentiate between retired and non-retired board members.

The mixed results might suggest that with respect to multiple board positions both theories are true to some extent. When board members are appointed to more board positions this may increase their social network, their knowledge, or their experience in board practices.

However, when a board member holds too much board positions, the costs of not having enough time to adequately supervise all organizations may outweigh the added value.

Although results are ambiguous, the DCGC (2008, art.III.3.4) limits the number of non- executive board positions per supervisory board member to a maximum of five to assure “the proper performance of his duties”, for which purpose the chairmanship of a supervisory board counts double. The bill that was recently passed by the Dutch parliament is largely similar to art.III.3.4 of the DCGC. It will legally limit non-executive board positions¹⁶ to a maximum of five (two for board members that also hold an executive position in another firm). It also counts the chairmanship of a board position double. Dutch legislators thereby suggest that when board members hold more than five board positions they will not be able to adequately perform their supervisory role anymore.

2.3BOARD SIZE

The ideal size of a group of people performing a joint task is a thoroughly researched subject in all social sciences. In general, there exists, once again, a trade-off between the gains and the costs of adding one person to a group. The gains of an additional person are the additional expertise and complementary viewpoints that person brings to a group. The costs of an additional person are the costs of additional communication, coordination and process problems, caused by having a larger group. When the additional costs of adding another person to the group exceed the gains of that person’s additional expertise, the ideal group size has been reached.

Lipton and Lorsch (1992) criticize the presence of oversized boards in the US and recommend limiting board size to eight or nine board members. They argue that when boards are larger, it becomes difficult for all board members to express their ideas and opinions¹⁷. Jensen (1993) argues that board size should not exceed seven or eight board members. Larger boards are less likely to “…function effectively and are easier for the CEO to control”

(Jensen, 1993, p. 865). In addition, when boards become too large a free-rider problem may emerge. Non-executive board members may neglect to contribute to the board’s supervisory role, assuming that the contribution of the numerous other board members will suffice.

However, resource dependence theorists, according to Goodstein et al. (1994), emphasize the

16 Which board positions will be considered is still subject of debate. Most likely board positions of listed firms, large non-profit organizations and perhaps large private firms.

17 The fact that board size still often exceeds ten board members, strengthens the argumentation that board members are not even supposed to “…voice their opinions freely and frequently”, according to Lipton and Lorsch (1992, p.65).

institutional function of board size. They argue that by increasing in size, boards help to link organizations to its external environment and secure critical resources.

Goodstein et al. (1994) study the influence of board size on the ability of the board to implement strategic change in times of environmental turbulence. They investigate strategic change in 334 hospitals (with an average board size of 10 board members) over the period 1980-85, which was characterized by environmental changes in the health care industry. They find some evidence that larger hospital boards had reduced ability to implement strategic change. Yermack (1996) studies the influence of board size on firm value for large listed US firms with large one-tier boards (mean and median board size of 12) and finds a significant inverse relationship between board size and firm value, measured by Tobin’s Q. Core et al.

(1999) also study large US listed firms with large one-tier boards (mean and median board size of 13) and find that firms with larger boards pay higher levels of CEO compensation, which in turn leads to poor firm performance, indicating that larger boards are less effective in performing their supervisory role. These findings are intuitively appealing since most boards in the samples of both Yermack (1996) and Core et al (1999) are likely to have surpassed the ideal board size. However, for firms with smaller boards (the mean and median supervisory board size in our sample are 4.8 and 5) this relationship may not be as straightforward.

However, Eisenberg et al (1998) and Mak and Kusnadi (2005) extend Yermack’s research to countries¹⁸ with smaller firms, smaller boards, and other corporate governance structures and find similar results. Eisenberg et al. (1998) study the influence of board size on profitability for Finnish firms (mean and median board size of 3.7 and 3). They find a negative association between board size and industry-adjusted return on assets. There findings suggest that increasing board size, even when board size is low (2 – 5 board members) decreases board effectiveness, which seems counter-intuitive. Eisenberg et al (1998) do mention the possibility that there results are not explained by additional communication, coordination and process problems caused by larger board size, but by other factors that are correlated to board size. Mak and Kusnadi (2005) study board size of Malaysian and Singaporean firms (mean and median board size of 7.3 and 7) and find a negative association between board size and firm value (Tobin’s Q). However, they find a positive association between board size and firm value for board sizes up to and including a board size of 5. Van Ees et al. (2003) study supervisory board size for Dutch listed firms (mean and median board size of 5) and find a significant negative association between supervisory board size and firm performance.

Although most previous literature on board size suggests that board effectiveness decreases with board size, there is no consensus on what the ideal board size is (perhaps no board at all). The DCGC does not provide any best practice provision on board size. There is,

18 Eisenberg et al.’s sample concerns Finnish firms and Mak and Kusnadi’s sample concerns Malaysian and Singaporean firms

however, a legal requirement on board size. Dutch Civil Law requires that supervisory boards are comprised of at least three board members.

2.4BOARD OWNERSHIP

Although boards of directors are assumed or supposed to act in the best interests of investors, several authors have questioned if they do. Hendry (2006) argues that it is contradictory to assume that non-executive board members will selflessly pursue investors’ wishes, whereas managers are assumed to act solely (or primarily) out of self-interest. Management needs to be monitored and awarded with incentive compensation to make sure it takes firm-value maximizing decisions, but the non-executive board members are supposed to act in the firm’s best interests “…without any instruction, monitoring or significant incentivization” (Hendry, 2006, p.56). In general, the non-executive board members will prefer reappointment. Apart from the associated salary, the board position is likely to give them prestige and valuable connections (Bebchuck & Fried, 2003). According to Fama and Jensen (1983), board members’desire to be appointed to other (more prestigious) positions provides them with incentives to pursue investors’ interests. However, the power of the CEO in (re-)appointments is likely to be substantial as well. In addition the CEO may provide non-executive board members with other perks and private benefits, to make them favour him (Bebchuck & Fried, 2004). The implicit incentives board members face, may therefore not suffice to make them pursue investor’s wishes. However, if board members hold shares in the firms they are supposed to monitor, this provides them with an explicit incentive to make sure the CEO maximizes firm value.

Several studies have examined the influence of board ownership on board effectiveness. McConnel and Servaes (1990), Hermalin and Weisbach (1991) and Yermack (1996) find a significant positive but non-monotonic relationship between board ownership and firm value, measured by Tobin’s Q. However these studies were performed on firms with a one-tier structure and did not differentiate between ownership by executive and non- executive board members. Morck et al. (1988) study the influence of board ownership on firm value (Tobin’s Q) and make a distinction between ownership by executive board members and by non-executive board members. Studying 371 large US listed firms, they find that the percentage of ownership by non-executive board members (between 0 – 5 %) is positively related to firm value. Van Ees et al. (2003) study supervisory board ownership of Dutch listed firms. They find no evidence that supervisory board ownership is associated with firm performance. Core et al (1999) study the effect of ownership of non-executive board members on the amount of CEO compensation, but find no significant association between the two.

The only best practice provision in the DCGC on board ownership (art.III.7.2) states that ‘any shares held by a supervisory board member in the company on whose board he sits

are long-term investments’. Although the desirability of board ownership is not mentioned in the DCGC, art.III.7.2 does imply that board ownership creates a certain (long-term) commitment to the company.

3. B

OARD

E

FFECTIVENESS

, F

IRM

P

ERFORMANCE

& CEO C

OMPENSATION Several authors have argued and found that board effectiveness is positively related to firm performance and negatively related to the amount of CEO compensation. This section discusses the rationale behind these relationships.

3.1FIRM PERFORMANCE

Effectiveness, in general, refers to the degree to which something realizes its intended goal or purpose. Section 2 explained the supervisory role of the board of directors. The intended purpose of the board is to prevent management from shirking and to assist management in making value-maximizing decisions. Not assisting management in decision-making is likely to influence firm performance, but especially the consequences of allowing management to shirk can be severe. Insufficient effort, extravagant investments and entrenchment strategies by management are some of the forms of shirking that significantly hurt firm performance. In 1776, Adam Smith already argued that you cannot expect managers of other people’s money to watch over it with the same anxious vigilance with which they watch over their own. Since then authors from Berle and Means (1932) to Jensen and Meckling (1976) and Fama and Jensen (1983) have argued that when decision-making is in the hands of managers that do not bear a substantial share of the wealth effects of their decisions, alternative control mechanisms are needed to prevent managers from exploiting investors. Since the board of directors is one of the most important mechanisms to prevent management from shirking nowadays, we expect reduced or ineffective supervision by the board of directors to be directly reflected in firm performance. Many authors (McConnel & Servaes, 1990, Hermalin

& Weisbach, 1991, Yermack, 1996, Eisenberg et al., 1998, Vafeas, 1999, Mak & Kusnadi, 2005, Fich & Shivdasani, 2006) have found that lower board effectiveness is associated with lower firm performance.

3.2CEOCOMPENSATION

CEO compensation practices have received considerable attention ever since information about the, sometimes mind-boggling, amounts of CEO compensation has become public.

Recently, the financial and economic crisis increased the discussion on excessive CEO compensation. Although firm performance world-wide dropped dramatically, CEO compensation seemed to continue its everlasting steady increase. In addition, some have

argued that compensation practices actually caused or at least contributed to the crisis, by the excessive risk-taking it evoked in the years preceding the crisis.

From an agency theory perspective, managerial compensation can be seen as the contract between the principal (shareholders) and the agent (manager). Agency theory focuses on designing an optimal contract that aligns the principal’s and the agent’s interests at the lowest cost (Jensen & Meckling 1976, Eisenhardt, 1989). The board of directors, as an extension of shareholders’ interests, therefore has the duty to design managerial compensation at the lowest possible cost. According to Core et al (1999), designing an optimal compensation contract should be a relatively straightforward decision for an effective board, given the information available to them. Even if we look at managerial compensation from a stakeholder point of view (as opposed to the shareholder point of view that is associated with agency theory), it seems unlikely that the board of directors serves any stakeholders’ interests (other than the manager’s interests) by awarding the manager with excessive amounts of compensation. Designing managerial compensation at the lowest possible cost is thus one of the boards’ goals.

Several authors have argued that less effective supervision will lead to higher compensation. Two explanations are given for this relationship in the literature. The first concerns the power that management will be able to exert over the board (Jensen, 1993, Bebchuck & Fried 2003, 2004). They argue that management will be able to exert more power over an ineffective board and that this will lead to the consumption of private benefits and the extraction of additional compensation by management. The second explanation concerns the trade-off between board monitoring and incentive compensation (Holmström, 1979, Lippert & Moore, 1994, Harvey & Shrieves, 2001). As explained in section 1, both managerial compensation and monitoring by the board provide the manager with incentives to pursue investors’ interests. When monitoring by the board is low, the board will have to compensate for decreased monitoring by awarding the manager higher incentive compensation to make sure he will pursue investors’ interests. Higher incentive compensation will in turn lead to higher, on average, total compensation.

This makes CEO compensation practices appropriate for assessing board effectiveness. Kerr and Bettis (1987, p. 645) state that critics of the functioning of boards of directors “…have viewed the compensation decisions of boards as a surrogate test of the corporate governance process”. Core et al. (1999, p.376) state that CEO compensation “…is a frequent and observable decision and has been the subject of much of the debate regarding the effectiveness of boards of directors”. According to Hermalin & Weisbach (2003) there are several advantages of linking board characteristics to a board decision, such as CEO compensation, over linking them to some measure of firm performance. Firstly, there are many unobservable factors affecting firm performance, which may contaminate the statistical

relationship. Secondly, when examining particular tasks of the board, “…it is less likely that the endogeneity of board composition will affect the results” (2003, p.14).

Several authors (Finkelstein & Hambrick, 1989, Boyd, 1994, Mehran, 1995, Core et al, 1999, Harvey and Shrieves, 2001) have found that factors associated with lower board effectiveness lead to higher CEO compensation.

4. H

YPOTHESES

The main hypothesis of this paper is dual, which is the case in most studies that examine the desirability of board characteristics¹⁹. Our main hypothesis is thus comprised of two parts.

The first part of our main hypothesis is that supervisory board independence, busyness, size and ownership influence supervisory board effectiveness. The second part of our main hypothesis is that supervisory board effectiveness is positively related to firm performance and negatively related to the amount of CEO compensation. The conclusions we wish to draw primarily concern the first part of our main hypothesis. The conclusions we want to draw from our results are whether or not supervisory board size, busyness, independence or ownership influence the supervisory board’s ability to perform its supervisory role effectively.

If we find that a supervisory board characteristic is positively related to firm performance and negatively related to the amount of CEO compensation, we will conclude that this characteristic is positively related to supervisory board effectiveness.

Supervisory Board Independence

Several authors have argued that board member independence is a prerequisite for non- executive board members to be critical of management and to perform their supervisory role well. However, some authors have argued that increasing board independence when the board is already comprised of a majority of independent board members may not increase or may even hurt board effectiveness. Therefore, we hypothesize that a relationship exists between supervisory board independence and firm performance and between supervisory board independence and the amount of CEO compensation, but we have no ex ante prediction for the direction of these relationships.

Hypothesis 1

There is a relation between supervisory board independence and firm performance / the amount of CEO compensation.

19 Most studies do not explicitly mention the duality of their main hypothesis. Yermack (1996, p.189), for example, does. He states that “the main hypothesis of this paper is that firm value depends on the quality of monitoring and decision-making by the board of directors, and that the board’s size represents an important determinant of the board’s performance”.

Supervisory Board Busyness / Multiple Board positions

The busyness hypothesis suggests that non-executive board members holding multiple board positions will be too busy to adequately perform their supervisory role. However, others have argued that non-executive board members holding multiple board positions do not shirk their supervisory role and may bring valuable resources, such as experience, knowledge and social networks to a firm. Therefore, we predict that a relationship between supervisory board busyness and firm performance and between supervisory board busyness and the amount of CEO compensation exists, but we have no ex ante prediction for the direction of these relationships.

Hypothesis 2

There is a relation between supervisory board busyness and firm performance / the amount of CEO compensation.

Supervisory Board Size

Based on evidence by previous literature, we hypothesize that larger supervisory boards will be less effective. Larger supervisory boards will have more communication, coordination and process problems. Therefore, larger supervisory boards will pay higher amounts of CEO compensation and will be more likely to allow the CEO to shirk.

Hypothesis 3

Supervisory board size is negatively related to firm performance and positively related to the amount of CEO compensation.

Supervisory Board Ownership

Based on previous literature, we hypothesize that supervisory board members with ownership in the organization they have to supervise will have higher incentives to perform their supervisory role effectively. Therefore, supervisory boards with higher ownership will pay lower amounts of CEO compensation and will be less likely to allow the CEO to shirk.

Hypothesis 4

Supervisory board ownership is positively related to firm performance and negatively related to the amount of CEO compensation.

5. M

ETHODS

This study extends research on the relation between supervisory board characteristics and supervisory board effectiveness in a two-tier structure. Previous literature on these

relationships is ambiguous and primarily concerns firms with a one-tier structure. This study examines whether independence, busyness, size or ownership of Dutch supervisory boards are related to firm performance or the amount of CEO compensation. We estimate several straightforward models of the relation between firm performance and supervisory board characteristics. We regress three different measures of firm performance against a set of explanatory variables, consisting of our measures of supervisory board independence, busyness, size, ownership and several control variables. We estimate a similar model of the relation between the amount of CEO compensation and supervisory board characteristics. We include control variables for the economic determinants of CEO compensation and for the firm’s ownership structure. When a separate RC is appointed, variables for RC independence, busyness, size and ownership are included as well.

For our regressions we use ordinary least squares (OLS) regressions. Panel data analysis (fixed effects) on our data was performed but provided consistent non-normality of residuals and highly anomalous coefficient estimates²⁰. We therefore only report and interpret results from our pooled data estimations. We use White’s (1980) heteroscedasticity consistent standard error estimates.

5.1SAMPLE

This study examines all firms with a listing on the Amsterdam stock exchange (AEX, AMX and AScX) in the period of 2004-2006. Our initial sample consists of 143 firms. We only consider firms that had a listing throughout the entire three-year period (117 firms) and had the same CEO for at least two consecutive years (108 firms). From the remaining 108 firms, the one-tier firms are excluded (1 firm), consistent with other studies the financial firms (5 firms) are excluded, and finally 23 firms are excluded because of unavailability of data. This leads to a final sample consisting of 79 firms and 208 firm year observations.

The data concerning CEO compensation is obtained from annual reports and the website of the Vereniging Effectenbezitters (VEB) (www.veb.net). The VEB is a well-known Dutch association of private and institutional investors that defends shareholder rights in the Netherlands. The data concerning supervisory board/RC characteristics (independence, busyness, size and ownership) was taken from annual reports and REACH. REACH is a large database containing particulars of supervisory board members over time. Firm characteristics were also taken from REACH. Stock returns are taken from Datastream. Data concerning the

20 The most likely explanation for the unsuitability of panel data analysis for our data is the low variability of the dependent variables for each firm. This is caused by the nature of our dependent variables (for example, return on assets usually differs very little between consecutive years) and the low number of observations per firm (2 or 3, depending on the regression). In the extreme, where the dependent variable for a specific firm is the same for all observations, the behaviour of the dependent variable (for that firm) will be entirely explained by fixed effects. In general, this will lead to high (adjusted-) R², highly non-normal distributions of residuals and strange coefficient estimates, which was also the case in our fixed effects estimations.

presence of large shareholders was taken from the archives of the AFM, the Dutch financial markets supervisory institution (www.afm.nl/registers/).

5.2SUPERVISORY BOARD CHARACTERISTICS

We measure supervisory board independence (BOARDIND) as the ratio of supervisory board members that is independent, which is the most common approach to measuring board independence. To determine whether supervisory board members are independent or dependent (also sometimes referred to as ‘affiliated’ or ‘gray’) we use the criteria of the DCGC, which can be found in Appendix II. We use two measures for supervisory board busyness (BOARDBUSY1 & BOARDBUSY2). BOARDBUSY1 is the ratio of supervisory board members that hold three or more supervisory board positions. For this measure we only count supervisory board positions in listed firms. Our choice to consider supervisory board members with three or more board positions in listed firms to be busy is in line with Core et al (1999), Ferris et al. (2003), and Fich and Shivdasani (2006). BOARDBUSY2 reflects the requirements of the DCGC and the bill that was (partially) passed by the Dutch parliament. It is calculated as the ratio of supervisory board members that hold six or more supervisory board positions (three or more for non-retired supervisory board members). For this measure we also count board positions in private firms, and non-profit organizations, which is in line with the recently passed bill. For both measures we count chairmanship of a supervisory board double, which is in line with the DCGC. We will regress BOARDBUSY1 and BOARDBUSY2 separately to avoid multi-collinearity. We measure supervisory board size (BOARDSIZE) by the number of supervisory board members. This is in line with the methodology and findings of Yermack (1996), Eisenberg (1999), Core et al. (1999), Van Ees et al, (2003) and Mak and Kusnadi (2005). We measure supervisory board ownership (BOARDOWN) by the number of ordinary shares owned by supervisory board members divided by the total number of ordinary shares outstanding. In 109 of 208 firm year observations a separate RC was appointed. We calculate RC characteristics (RCIND, RCBUSY1, RCBUSY2, RCSIZE & RCOWN) in the same way as the supervisory board characteristics.

Table 1 presents descriptive statistics of our supervisory board and RC characteristics.

Our descriptive statistics for independence, size and ownership of Dutch supervisory boards are similar to those of Van Ees et al (2003). Supervisory board and RC independence (BOARDIND & RCIND) are both left-skewed. More than half of the supervisory boards in our sample consists entirely of independent board members (BOARDIND of 1) and more than two-thirds of the RCs consists entirely of independent RC members (RCIND of 1).

Supervisory board and RC ownership (BOARDOWN & RCOWN) are both right-skewed.

More than two-third of the supervisory boards in our sample owns less than 0.1 percent of