The determinants of Supervisory Board compensation in a two-tier system
An empirical study of Dutch listed firms
Author: G.M.N. Nijenhuis 10675825 Supervisor: Drs. C.M. Hsieh August, 2014
Key words: Supervisory Board compensation, two-tier system, Corporate Governance
University of Amsterdam Faculty of Economics and Business
MSc. Business Economics – Track Organizational Economics
Abstract: In this thesis, I adopt an empirical approach and examine the economic determinants of the level of Supervisory Board compensation in the Netherlands for which limited empirical evidence exists. Based on a hand-collected dataset consisting of Dutch listed firms for the period 2007-2013, I aim to provide insights in the ‘black box’ of Supervisory Board compensation in the Netherlands. The descriptive statistics reveal that the average per capita board compensation level increases from almost €46,000 in 2007 to more or less €58,000 in 2013, a compound annual growth rate of 3.52%. Central to this study is to establish the relations between key characteristics at the firm-, the board- and the governance level on the one hand, and the compensation of these members on the other hand. By adopting panel data analysis to allow for unobserved firm heterogeneity, I find support for the marginal productivity theory and the institutional perspective theory. Different aspects of firm complexity and risk are reflected in Supervisory Board compensation levels. Specifically, measures of firm size, the firm’s internationalization level, the time spent on the Supervisory Board tasks, and the diversity within the board are all positively and significantly determining these compensation levels.
Table of contents
List of tables and figures ... 3
List of abbreviations ... 3
1. Introduction ... 4
2. The Dutch Corporate Governance Code ... 5
3. Related literature ... 6
3.1. Conceptual approaches explaining director compensation ... 6
3.2. Empirical determinants of Supervisory Board compensation ... 8
4. Hypotheses ... 10
4.1. Firm level ... 11
4.2. Board level ... 13
4.3. Governance level ... 14
5. Data and variable construction ... 16
6. Descriptive statistics and model specifications ... 21
7. Econometric models ... 27
8. Results ... 28
9. Discussion and conclusion ... 33
10. Limitations and future research ... 36
11. References ... 37
List of tables and figures
Table 1 Overview of the hypotheses
16
Table 2 Variable definitions, descriptions and sources
20
Table 3 Descriptive statistics of the total SB compensation levels
22
Table 4 Descriptive statistics of the independent variables
23
Table 5 Correlation matrix of the full sample
26
Table 6 Correlation matrix of the subsample
26
Table 7 Regression results of the full sample
30
Table 8 Regression results of the subsample
32
Table 9 Overview of the hypothesized and empirical signs
33
Table 10 Summary of the survey on the determinants of SB compensation
39
Table 11 Company and data samples overview
41
Figure 1 Research model ... 11
Figure 2 The development of the annual per capita board compensation ... 21
List of abbreviations
AGM
Annual General Meeting
BoM
Board of Management
DCGC Dutch Corporate Governance Code
FE
Fixed Effects
IOS
Investment Opportunity Set
OLS
Ordinary Least Squares
RE
Random effects
ROIC
Return On Invested Capital
ROA
Return On Assets
1. Introduction
After the explosion of corporate scandals that, starting from 2001, burst over the financial markets in the US and in Europe, practitioners, politicians and scholars have become much more critical in evaluating Supervisory Boards and the incentives they receive. Traditionally, Supervisory Boards are seen as a group of older, wise men that had a highly successful career and that are willing to advice the management board on strategic challenges and decisions. However, in the light of recent global scandals such as Enron and WorldCom and the role that supervisors played in these, this topic has increased in relevancy. Particularly, the Supervisory Board’s role as an effective corporate governance instrument for monitoring the behavior of managers and protecting the interests of shareholders is a focal point in increasing corporate governance regulations.
Yet, the topics that are neglected in this respect and that seriously cast doubts on the effectiveness of Supervisory Boards are the compensation levels and coinciding incentives received by these members. CEOs’ pay is a much more widely discussed topic; however, very little attention is given to the compensation of Supervisory Boards. Even though companies spend large sums of money on rewarding these members, little is known about the determinants of their compensation levels. For the case of the Netherlands, the marginal attention that is given to this topic can be partly attributed to limited data availability. Data on Supervisory Board compensation is hard to come by, whereas data for executive board members is more easily attainable. Nevertheless, the introduction of corporate governance codes and disclosure rules resulted in a stream of data on company’s compensation sums, and as a consequence, slowly increased the interest and ability to study this topic empirically. Especially in the US, where the problem of lacking data is less prevalent, many studies focus on Supervisory Boards’ compensation as a remedy for dispersed interests of shareholders on the one hand and managers on the other.
The applicability of existing studies to help explain the determinants of Supervisory Board compensation levels in the Netherlands is low though, which originates in the difference between the governance- and compensation structures of countries. Whereas US studies provide evidence for the one-tier governance setting for example, the Netherlands has established a two-tier setting for which systematic evidence is hardly available. Yet, these settings differ significantly in internal governance. The fundamental difference is that the monitoring of executive board members is defined to be an additional task of the board itself in a one-tier setting, whereas these responsibilities are delegated to a Supervisory Board in the two-tier setting.1 The different division of the roles and responsibilities of board members in these two settings results in low comparability between them, as the differences in tasks are probably reflected in their compensation levels.
A second reason why the majority of the studies cannot be used to explain Supervisory Board compensation levels in the Netherlands is the fact that most countries adopt different compensation structures for their Supervisory Board members. Whereas other countries with a one-tier or two-tier setting allow for performance-based compensation for Supervisory Board members (i.e. the US, Germany, the UK), in the Netherlands the predominant Supervisory Board compensation elements are limited to cash retainers, not contingent on the firm’s performance. Given these differences, most of the literature studying Supervisory Board compensation focuses on the optimal contracting for Supervisory Board members (e.g. Andreas et al. 2012). The central paradigm in most of these studies is whether performance-based compensation can align dispersed interests of the owners (i.e.
1
In a one-tier structure, the Board of Management and the Supervisory Board form one organ. The Supervisory Board members are referred to as non-executive directors and the Board of Management members as executive directors. In this thesis, I use the term ‘non-executive directors’ when referring to studies in one-tier settings, and Supervisory Board members when referring to the two-tier setting.
the shareholders) and the (non-) executive board members. Although this leads to interesting research questions largely build on agency theoretic assumptions, the majority of these studies cannot be applied to the case of the Netherlands.
In this thesis, I adopt an empirical approach and examine the economic determinants of the level of Supervisory Board compensation in the Netherlands for which limited empirical evidence exists. Based on a hand-collected dataset consisting of Dutch listed firms for the period 2007-2013, I aim to provide insights in the ‘black box’ of Supervisory Board compensation in the Netherlands. The descriptive statistics reveal that the average per capita board compensation level increases from almost €46,000 in 2007 to more or less €58,000 in 2013, a compound annual growth rate of 3.52%. Central to this study is to establish the relations between key characteristics at the firm-, the board- and the governance level on the one hand, and the compensation of these members on the other hand. The general arguments for the hypotheses are based on less conventional theories explaining Supervisory Board compensation levels, such as the marginal productivity theory and the institutional perspective theory. Further, this thesis contributes to existing literature by adopting panel data techniques that are only recently used on Supervisory Board compensation. The measures of firm size, the firm’s internationalization level, the time spent on the Supervisory Board tasks and the diversity within the board are all positively and significantly determining these compensation levels.
The remainder of this thesis is organized as follows. First, a discussion is provided on the Dutch Corporate Governance Code to paint a brief picture of the Dutch corporate environment. Then, two literature reviews on the conceptual approaches and the empirical determinants of Supervisory Board compensation levels previous studies have identified are provided. Subsequently, I dive into the hypotheses followed by the specifics of the dataset and the variables’ construction. Then, I give a short description of the summary statistics after which the econometric models are discussed. The regression results then follow, these results are discussed and conclusions are drawn in subsequent sections. Lastly, I address the limitations of this research and suggest questions for future research.
2. The Dutch Corporate Governance Code
This section gives an insight into the Dutch Corporate Governance Code (‘DCGC’), as this set of principles and best practice provisions are considered to have important consequences for Supervisory Board (‘SB’) compensation structures. Moreover, the DCGC defines the roles and responsibilities of SB members. As I expect to find a positive relation between increases in the complexity of these roles and responsibilities on the one hand and compensation levels on the other, it deems important to shortly discuss those.
The DCGC is an attempt to create a uniform set of rules introduced as a response to recent global scandals, by which listed companies have to try to abide. The code serves as a self-regulating instrument and is aimed at influencing the behavior of the BoM, the SB and the shareholders. The code is built on the principle of ‘comply-or-explain’, originating in the desire to create a spirit of good corporate governance and not to install a mere minimum of rules. One of the code’s main pillars is effective supervision by the SB and its accountability for this. This objective should create stakeholders’ confidence in the SB.
According to the code, the SB has a couple of roles and responsibilities. In general, the SB is responsible for supervising the BoM’s policies and the general affairs of the company. In addition, it
provides advice, weighs all stakeholders’ interests, and designs compensation policies for the BoM aimed at maximizing incentive alignment between the BoM and the shareholders.
Thus, the DCGC specifically states that the SB is responsible for setting executive compensation packages, but there is no mentioning of what organ is responsible for determining SB compensation levels. Some companies state in their annual report that the shareholders determine the compensation levels, but in practice it is hard to believe that this usually dispersed group of owners collectively determine these compensation levels.2
Although the DCGC does not provide provisions about the particular organ that is responsible for determining SB compensation levels, it does state specific best practice provisions concerning the structure of SB compensation levels. Most importantly, the compensation for the SB should not depend on the results of company and thus, SB members may not be granted any shares or options as a means of compensation (DCGC best practice provisions III.7.1, III.7.2.). In practice, almost all companies adhere to this provision.3
Nevertheless, standards for determining and assessing SB compensation levels are still limited. This thesis aims to contribute to this ‘black box’ of SB compensation by investigating its determinants.
3. Related literature
Issues related to SB compensation slowly gain the interest of researchers due to two main reasons. In the light of recent global scandals, the perceived disruption of the power between the BoM and the SB, and increased attention for executive compensation, have resulted in the rise of questions regarding the effectiveness of the supervision of such firms. Second, an increase in the availability of data on the topic has enabled researchers to try to find answers to some of these questions empirically.
My aim is to paint a detailed picture of this stream of literature along two reviews. The first review is intended to give a broad overview of the conceptual approaches used to explain (SB) compensation levels. The second, more systematic review focuses on the empirical determinants previous studies have identified.
3.1. Conceptual approaches explaining director compensation
This literature review discusses the various approaches used in the literature to explain compensation levels and structures. The majority of these theories are applied extensively to explain executive remuneration. However, more recent studies argue that some of these theories also help to understand the compensation levels of SB members (Linn and Park 2005; Andreas et al. 2012). In total, I identified five theories that are applicable to the case of the Netherlands and considered appropriate given the nature of the dataset. The difficulty of applying more conventional theories such as those based on agency theoretic assumptions lies in the lack of performance related pay to SB members in this case.4 Therefore, less conventional theories are discussed here.
In general, the two theories that apply most to the Netherlands are the marginal productivity theory and the institutional perspective theory. First, the marginal productivity theory hypothesizes that employees are paid the value they add to the firm, i.e. their marginal product (Biewen and
2
See for example Fugro’s annual report 2013, p. 88 3
With regards to my study, the only exceptions are AMG and Crucell.
4For a description of these theories and the particulars of why they do not apply to my dataset can be found in the appendix. For a study that base their hypotheses on these theories and the determinants of SB compensation, I refer to Andreas et al. (2012).
Weiser 2014). According to this theory, employees’ wages reflect an equilibrium level of their marginal product and labor market supply. Linn and Park (2005) apply this theory and find evidence that non-executive directors’ marginal productivity interacts with the need for firm monitoring to ensure the maximum gain in terms of firm growth and the exploitation of opportunities.5 That is, compensation levels are set with the intention of maximizing productivity by attracting non-executive directors with a high marginal product.6 In a similar vein, Brick et al. (2006) find a positive relation between director compensation and variables reflecting the need for monitoring. They argue that this relationship might be due to unobserved firm complexity requiring higher levels of effort and skills (i.e. marginal product) by non-executive directors.
Secondly, the institutional perspective theory suggests that compensation levels and structures may be the result of macro social processes (Andreas et al. 2012). This theory argues that organizations feel pressured to conform to the expectations of different parties (in-) directly involved in the organization, in order to enhance or protect their legitimacy (Berrone and Gomez-Mejia 2009). More specifically, actions to protect legitimacy are defined within the industry in which a firm operates (and thus competes) and firms are pressured to adopt them, for example by imitating other firms (DiMaggio and Powell 1983). Hempel and Fay (1994) argue that SB compensation should be set such that it attracts new non-executive directors and the appropriate way is to benchmark with other firms.
A third pattern of thought that takes on a different, less often used approach in explaining SB compensation is the resource dependency theory (Andreas et al. 2012). Firms are always in need of resources (i.e. physical and financial) as well as information from the environment. In that respect, firms are dependent on external sources (Pfeffer and Salancik 1978). According to Andreas et al. (2012), a typical solution to reduce uncertainty and dependency on these external resources is to establish networks with other firms. The SB may serve this purpose by providing the firm with critical information through multiple board memberships and board interlocks.
Fourth of all, a different perspective commonly used to explain supervisory board behavior is the stewardship theory. This theory directly opposes standard agency theoretic assumptions of self-serving behavior and diverging interests (Andreas et al. 2012). Stewardship theory presumes a natural motivation of the individual to act in the best interest of the organization (Muth & Donaldson 1998). That said, SB members are expected to act out of non-monetary motives (Andreas et al. 2012). Coles and Hoi (2003) and Yermack (2004) extend the stewardship theory and find evidence that non-executive directors receive additional incentives through reputational concerns: if a firm performs well, the director has a higher chance of gaining additional board memberships.
Lastly, the managerial power theory is occasionally applied to explain compensation levels and structures. This theory is an extension of the conventional agency theory, which argues that the separation of ownership and control of firms leads to diverged interests (Bebchuk & Fried 2003). This managerial power theory states that compensation packages of board members are partly responsible for creating agency problems, as unequal bargaining power leads to deviations from optimal contracts (Bebchuk et al. 2002). Ryan and Wiggins (2004) extent the managerial power theory by applying it to non-executive director compensation and conclude that compensation levels
5
Firm size and its investment opportunity set serve as proxies for the need for firm monitoring.
6 They conclude this by finding a positive, statistically significant relation between total non-executive directors compensation and firm size.
are the result of a bargaining ‘game’ between the CEO and other directors. The party with relatively more power determines the structure and level of compensation for the non-executive directors.7
In conclusion, these theories help building a framework on which the arguments for researching particular determinants of SB compensation are based. When reviewing these theories, it becomes clear that some are more relevant than others for the case of the Netherlands. In particular, the marginal productivity theory is predominantly used in this thesis to analyze to what extent SB members are rewarded for the time and effort they spend on their assigned roles and responsibilities. SB members’ value added lies in monitoring and assisting the BoM. As a result, if the need for firm monitoring increases, the productivity of SB members is required to increase simultaneously and consequently, in combination with an existing SB members labor market, a different compensation level is appropriate.
Moreover, the institutional perspective theory may provide relevant insights into which determinants are used when firms benchmark their compensation levels. More specifically, most firms in my dataset state in their remuneration policy that they aim to compensate at the median level of their peer groups. These peer groups serve as reference points for their compensation levels and consist of comparable firms in size, complexity, industry and geography.8 In light of the institutional perspective theory, this benchmarking could be perceived as feeling pressured to adhere and conform to the norm. Put differently, certain compensation levels are set as signals of commitment to increase market recognition and keeping up with peer companies. A couple of reasons are important here. Firstly, firms want to retain competent SB members. Secondly, firms want to conform to market expectations to seek legitimacy and consequently attract highly skilled SB members (Andreas et al. 2012).
The last three theories discussed above are useful in developing some of the hypotheses. In light of the resource dependency theory, SB compensation levels may serve as a means to attract individuals that can provide critical resources to the firm through additional board memberships. Furthermore, the stewardship theory gives relevant insights in the potential non-monetary incentives SB members have. Finally, the managerial power theory serves as the foundation when assessing whether a distorted balance of the SB’s relative power could explain its compensation levels.
3.2. Empirical determinants of Supervisory Board compensation
This second literature review identifies the empirical determinants previous studies use to explain SB compensation. Before giving an extensive review of studies directly relevant to the subject, it is important to disclose the search process for these articles. The process that is followed consists of three parts. First of all, the search engines for papers provided by the University of Amsterdam and Google Scholar are used.9 Next, a set of keywords and relevant journals are determined.10 The reference lists of the papers that are found in this process are checked for other relevant studies (and in turn, their reference lists are checked).
Then, these papers are assessed using the following criterion. The papers need to have performed statistical analyses using some measure of SB compensation as a dependent variable.
7
The characteristics of executive and non-executive directors serve as proxies for relative power, i.e. the CEO tenure, CEO/chair duality, board size, director affiliations and whether a director serves on an inside/outside dominated board (i.e. more non-executives or executives).
8 To read more about this, see for example Ahold‘s remuneration policy in their 2013 annual report on page 69. 9
i.e. www.uba.uva.nl and www.scholar.google.nl 10
I use the following keywords: supervisory board compensation/remuneration, outside director compensation, non-executive board compensation and board of directors compensation. Journals are deemed relevant when they are refereed journals and preferably considered A journals (i.e. journal of corporate finance, journal of financial economics, etc.).
Thus, studies using SB compensation as an independent variable are excluded from the list. In the end, there are 12 papers that meet the criterion. In order to create a systematic overview of the determinants these empirical studies use to explain SB compensation, a survey is developed. The main findings are reported in Table 10 in the appendix. To ensure a structured overview, different classifications are created of commonly studied determinants. These classifications are measures of firm characteristics, corporate performance and two governance parts, namely ownership structure and board characteristics.
With regard to the results, most of these empirical studies base their hypotheses on agency theoretic assumptions and focus on whether the observed patterns in compensation levels and structures reduce agency costs and increase interests alignment. In line with this theory, the majority of the literature concentrates on the pay-performance sensitivity of the non-executive directors’ compensation packages in different scenarios. For example, Hempel and Fay (1994) cast doubt on the effectiveness of incentive-based compensation as a means to increase future firm performance. They argue that it is almost impossible to isolate the effect of the board’s actions and efforts on future firm performance as many other determinants play a role as well. Therefore, they argue, rewarding directors for their performance sounds reasonable in theory; in practice this link is not so easily measurable.
Whereas Hempel and Fay (1994) focus on the effectiveness of performance-based compensation, other studies are already convinced of its theorized purpose. Linn and Park (2005) state that the observed compensation structures for non-executive directors in the U.S. induce the appropriate incentives. That is, board compensation policies are designed such that they minimize agency problems through compensating partly in stock (options).11 While Linn and Park study agency problems in a one-tier setting, Andreas et al. (2012) extend the literature by analyzing the two-tier case of Germany and draw similar conclusions. Using panel data analysis, they state that particularly firms with weak governance mechanisms (i.e. dispersed ownership resulting in lower monitoring power of the owners) structure their directors’ compensation packages in a way that provides incentives to monitor executives.12
The particular effect that weak governance mechanisms (i.e. the dispersion of firm ownership) has on compensation structures and levels is more closely studied by Menozzi et al. (2008). During the years of interest, a liberalization process took place in Italy leading to the entrance of more private investors and less public control. They believe that this dispersion of ownership and control leads to increased agency problems. However, the evidence shows that publicly- and privately owned firms do not compensate their board members differently. Therefore, the firm’s ownership structure is not necessarily responsible for agency problems.
A stream of literature that uses a different angle in assessing agency problems looks into potential collusions between the executive board and the non-executive directors. Fernandes (2008) discusses the role non-executive board members play in the level of executive compensation packages. He argues that installing ‘independent’ board members does not solve the agency problem, since he finds that boards with zero non-executive board members have lower executive compensation levels than boards with a positive number of non-executives. Thus, the independency of these non-executive members is impeached. In a similar vein, Ryan and Wiggins (2004) question the independency of non-executive board members by using a bargaining framework. They argue
11
Bryan et al. (2000) draw similar conclusions.
12 Elston and Goldberg (2003) conclude, by examining firms in the U.S., that lower ownership concentration leads to a higher ability of non-executive board members to extract higher levels of compensation.
that boards consisting of a higher proportion of outside directors have a bargaining advantage over the CEO. As a result, compensation structures are set more in alignment with shareholders’ interests (i.e. more equity-based compensation) and vice versa for boards with a lower proportion of outside directors. A third paper that looks into potential collusions between executive and non-executive board members is provided by Brick et al. (2006). They find evidence that excess compensation of both parties is correlated and associated with firm underperformance. Hence, they draw the conclusion that the evidence shows signs of mutual back scratching or cronyism. These three studies shed serious light on potential conspiracy between executive and non-executive board members and the effect on compensation levels. However, Hempel and Fay (1994) find opposing evidence and conclude that no clear link between executive and non-executive compensation exists. Therefore, conspiracy between them to inflate compensation levels (for example as a means of reciprocity) is refuted.
In contrast to the agency theory on which most of the hypotheses of these papers are built, Yermack (2004) takes a different approach based on the reputational concerns of the directors. By studying the career paths of certain directors during a five-year period, he finds evidence that non-executive directors receive positive incentives from the likelihood to obtain additional board seats conditional of the performance of the initial firm.
Besides the paper by Yermack (2004), one other paper considered in this review bases its research on a different theory than agency theory. Linn and Park (2005) find a positive relation between a firm’s investment opportunity set (i.e. firms with a greater proportion of their value deriving from growth opportunities) and board compensation. Thus, they argue that compensation packages are designed in a way that attracts directors whose marginal productivity is the highest.
After reviewing the different studies that are directly and indirectly relevant for this thesis, an existing research gap is identified that I try to address. First of all, the majority of these studies are performed in a one-tier governance system and empirical evidence resulting from two-tier settings is limited. The papers based on data from two-tier systems point out that this is mostly due to limited data availability (Elston and Goldberg 2003; Andreas et al. 2012). Admittedly, data is hard to come by for the Netherlands as well. This is in sharp contrast with the U.S., where some prominent databases on director compensation exist.13 Moreover, the countries on which the two-tier studies in this survey are based, are still very different from the Netherlands. The most important difference is that SB members in the Netherlands should not be compensated based on performance according to the DCGC. Thus, arguments derived from agency theory apply to a lesser extent. Therefore, the contribution of this study lies in providing evidence for the case of the Netherlands as well as testing less conventional theories on compensation levels of SB members.
4. Hypotheses
The two literature reviews provide a solid background on which the development of testable hypotheses for the empirical analysis is built. Figure 1 shows the research model and the hypotheses graphically. The general arguments for these hypotheses are based on the first literature review. The second literature review helps me develop the arguments for the specific hypotheses.
With regards to the general arguments for the hypotheses, the marginal productivity theory and the institutional perspective theory play an important role. In accordance with the former theory, the marginal product of SB members lies in monitoring and assisting the BoM. Consequently, if the
demand for these tasks rises, the productivity level of SB members is required to increase. This theory is used in this thesis to analyze to what extent SB members are rewarded for the time and effort they spend on their assigned roles and responsibilities. In line with previous research, firm complexity and risk are expected to determine the need for monitoring, the difficulty of the task and the required effort by SB members (Brick et al. 2006; Andreas et al. 2012).
The institutional perspective theory states that firms feel pressured to conform to the market in order to reduce uncertainty and increase legitimacy to different (external) parties. Most firms in the dataset achieve this through compensating similarly to comparable firms in terms of size, complexity, industry and geography.14 Therefore, these firms’ policies already reveal some of the aspects firms take into account when establishing compensation levels.
Overall, the hypotheses I develop are based on the relation between firm complexity and risk on the one hand, and SB compensation levels on the other. The hypotheses are classified in three parts, i.e. firm-, board and corporate governance characteristics.
Figure 1 Research model
4.1. Firm level
The independent variables at the firm level serve as proxies for firm complexity. One of the most consistent finding in the SB compensation literature is the positive relation between firm size and pay. Generally, it is believed that firm processes become more complex with size (Cordeiro et al. 2000; Brick et al. 2006). Therefore, more effort is required to ensure effective monitoring and good oversight of all these processes. Another important reason Brick et al. (2006) point out is that larger firms have more resources at hand, resulting in increased possibilities to pay their SB members more. Moreover, following the marginal productivity theory, the monitoring task of SB members adds more value when overseeing more complex firm processes. Since workers are paid their marginal product, the compensation levels should increase.
14 In this study I do not control for industry specifications (in this case I refer to Menozzi et al. 2014). However, the other three parameters I classify under the denominator firm complexity.
Firm characteristics
- Firm size - Risk - Growth opportunities - Internationalization - Firm performanceBoard characteristics
- Time commitment - Board size - Bargaining frameworkCorporate Governance
- Board diversity- Additional board memberships
Supervisory Board
compensation level
In short, a larger firm size leads to higher firm complexity, which increases the need for monitoring and consequently raises compensation levels. This leads to the first hypothesis:
H1: The level of SB compensation is positively related to firm size
Another often used proxy for firm complexity is a firm’s risk exposure (Linn and Park 2005; Andreas et al. 2012). Higher risk rates for firms require more effort and an increased task difficulty of the company’s management. In addition, managing these risks requires extra monitoring. Consequently, it is expected that the monitoring effort and task difficulty of SB members rise as well. Moreover, since it is possible to hold SB members accountable for their actions and decisions, according to standard theories in finance, they should be compensated for the higher risk they face (Hempel and Fay 1994).15
Concluding, a higher risk rate of a firm increases its corporate complexity and thus, the need for monitoring. As a result, compensation levels will increase to reflect the increased complexity of monitoring. Therefore, the following hypothesis is developed:
H2: The level of SB compensation is positively related to a firm’s risk rate
A third proxy for firm complexity is a firm’s growth opportunities. According to the literature, firms with more growth opportunities are regarded as having a larger Investment Opportunity Set (‘IOS’) (Linn and Park 2005). This IOS can be considered as the firm’s future investment options and is therefore, valuable to a firm. In order to make good use of this IOS, specialized human capital is required. Therefore, Linn and Park (2005) argue that directors with a high marginal productivity are attracted to produce the maximal gain. Their results indicate that higher compensation levels are the result of attracting highly productive directors by firms with a large IOS.16
Cordeiro et al. (2000) point out another reason for a positive relation between a firm’s growth opportunities and compensation levels. Namely, SB members are required to put in more effort in dealing with more frequent and complex issues resulting from a firm’s rapid growth. Thus, this higher effort level should be reflected in their compensation level. Furthermore, Bryan et al. (2000) state that information asymmetry is typical for high growth firms. To ensure SB members are aware of all that is happening within the firm in order to monitor effectively, more time and effort should be devoted to their tasks.
In conclusion, firms with more growth opportunities have more complicated company processes, and therefore, require more monitoring. In line with the general arguments, the increased time and effort by SB members should be reflected in compensation levels. Thus, the following is hypothesized:
H3: The level of SB compensation is positively related to a firm’s growth opportunities
In light of the globalization, a firm’s degree of internationalization is a more recent proxy for firm complexity. Although to the best of my knowledge not applied to SB compensation yet, it has been used to determine executive compensation levels (e.g. Miller et al. 2002). Internationalizing firms
15
Each year at the company’s AGM, a voting round decides whether SB members are discharged from their liability. Thus, they actually run higher risks themselves, not just the firm.
16 Yermack (2004) finds similar results. He argues that directors’ compensation levels should have a steeper slope and more convexity when the firm has more growth opportunities.
require strong capabilities to adapt to different cultures and institutions, to process increasing information streams and manage increasingly complex firm processes (Miller et al. 2002). This leads to higher firm complexity as well as more exposure to risk. In turn, SB members are required to have a wider set of skills and experience to oversee these processes, and thus, the expectation is that they receive higher compensation levels as a result.
Hence, a higher degree of internationalization leads to a more complex business environment and hence more time and effort to monitor managers. Therefore, the following relation is expected:
H4: The level of SB compensation is positively related to a firm’s degree of internationalization
Although not a proxy for firm complexity, most papers studying the determinants of SB compensation include some measure of firm performance in their analysis. They include firm performance measures because the central issue of the agency theory and optimal contracting theory is the pay-for-performance sensitivity.17 Although performance-related compensation is
almost non-existent in the Netherlands, an indirect positive relation between firm performance and SB compensation levels might exist. Some of the studies considered in the systematic literature review in table 10 report two regressions, one on cash compensation of non-executive directors as well as one on the stock-based compensation.18 Even on the cash related compensation, which one can consider as similar to the SB compensation structure in the Netherlands, these researchers find significant and positive relations between firm performance and compensation in cash.
The reason to expect a positive relation is that better performing firms may want to reward their SB members for their share in the company’s performance through offering higher fixed fees. Moreover, these firms have more resources at hand to offer higher compensation levels. So, the following hypothesis is developed:
H5: The level of SB compensation is positively related to firm performance 4.2. Board level
Although firm-level characteristics determine the required level of monitoring through the level of firm complexity and risk, characteristics of the board itself determine the capability of delivering effective monitoring services. Boards can differ widely in their abilities and quality of services, which are assumed to be affected by a couple of characteristics. Thus, natural determinants of SB compensation levels are found in the board composition itself. Therefore, three hypotheses are developed based on board-level characteristics.
One of the most frequently used measures to proxy the board’s efforts is the time commitment of SB members. Intuitively, the amount of effort board members have to put in working on their tasks should be reflected in the compensation they receive. Companies that do not adequately compensate their workers for the time they spent have a hard time attracting and retaining SB members (Cordeiro et al. 2000). Some companies even pay attendance fees in order to create incentives for their members to be present at meetings. Furthermore, as firm complexity increases, the need for monitoring also increases, which should be reflected in time commitment. Therefore, the following is hypothesized:
17 For more information on these two theories, I refer to the appendix. 18 These studies are indicated with a 1 in table 10 in the appendix.
H6: The level of SB compensation is positively related to the time commitment of their members
A second often used proxy for measuring a board’s effectiveness is its size. Two opposing views have developed in this respect. On the one hand, boards may grow until reaching a size that causes less effective monitoring as coordination and decision-making become less efficient (Ryan and Wiggins 2004; Andreas et al. 2012). On the other hand, adding more members to the board increases the capabilities, skills and experience present within the board. This results in enhanced monitoring and more incentive alignment (Menozzi et al. 2014). Almost all papers that use board size as a determinant for SB compensation reviewed in table 10 in the appendix, report significant negative relations.19 Following the literature, I expect board size to negatively influence compensation levels because the quality of the monitoring services decreases. Thus, the following is hypothesized:
H7: The level of SB compensation is negatively related to the SB size
A third hypothesis at the board level is the bargaining framework of executive board members and SB members, and is used in the literature to assess board independence and its effect on executive compensation. The relative power either party has over the other is used as an explanatory variable of the pay setting of executive compensation (Fernandes 2008). Feng et al. (2007) and Fernandes (2008) find that a higher number of non-executives in the board leads to higher executive compensation, which serves as evidence for collusion. This contradicts with the assigned role of non-executives, namely to protect shareholders’ interests and monitor executives. The main reason these authors give is that executives are, in general, responsible for appointing new non-executive board members for approval at the AGM; therefore, these non-executives may feel indebted towards the executives. This erodes board independence. Moreover, as Yermack (2004) states, the lack of a competitive labor market for non-executive directors results in fewer incentives for these members to act on behalf of shareholders. They have a lot to lose and a little to gain.
Extending these arguments to SB compensation, it can be argued that if the SB has relative power over the executives, this may induce chasing self-interests. This statement is supported by the study of Ryan and Wiggins (2004) who find evidence that the managerial power theory also applies to non-executive director compensation. Especially in the Netherlands, where in practice it is less clear who is responsible for the pay setting process of SB, bargaining power of the SB may result in extraordinary high SB compensation levels. Therefore, the following hypothesis is tested:
H8: The level of SB compensation is positively related with its bargaining power over the management board
4.3. Governance level
The next set of hypotheses is based on characteristics at the governance level, supported by certain principles set out in the DCGC. The two pillars on which the DCGC is built are good entrepreneurship, i.e. integrity and transparency, and effective and accountable supervision. One of the goals in this respect is to ensure the independency of the SB. Two parts in the code relate to this SB independency. The first explicitly defines independency and the second part relates to the expertise and composition of the board that is expected to enhance independency (DCGC principle III.2 and III.3). With regards to the former, SB members are not allowed to hold more than five board
memberships at other Dutch listed firms to ensure proper performance of their duties (DCGC best practice provision III.3.4). Regarding the latter, the code states: “The supervisory board shall aim for
a diverse composition in terms of such factors as gender and age.” (DCGC 2008, p. 22). Therefore, it
seems relevant to take these principles of board independency into account and assess the influence on compensation levels.
The first hypothesis that relates to the SB independency deals with board diversity. A variety of papers study the effect of more diverse supervisory boards on monitoring performance (e.g. van Ees et al. 2008; Adams and Ferreira 2009). Generally, the conclusions are similar; a more diverse SB is equivalent to a board consisting of a diverse set of skills, views and opinions (van Ees et al. 2008). This increases the objectivity of the SB and improves monitoring performance (Carter et al. 2003; Adams and Ferreira 2009). An important reason for this is that tapping into a broader talent pool by increasing the board’s diversity decreases the chances of possible collusions, as most of these new members do not belong to the ‘old boys club’ (Adams and Ferreira 2009). Thus, the independency of the SB is enhanced. Especially in the Netherlands, where SB members have the reputation to be part of this sort of old boys clubs, diversity is expected to have positive effects on monitoring performance. Adams and Ferreira (2009) also find evidence that gender-diverse boards allocate more effort to monitoring.
Regarding the SB compensation levels, a couple of effects can be expected. The first one is pointed out by Ryan & Wiggins (2004), who state that independency of the board increases the bargaining power of non-executive directors that may be misused to achieve higher compensation levels. A second effect may be the result of a new equilibrium on the labor market. Since the code establishes best practice provisions aimed at increasing board diversity, it narrows the company’s profile of potentially suitable SB members. As a result, a shortage on the labor market for these directors may arise. Consequently, companies may feel the need to pay higher or more (internationally) competitive compensation levels in order to attract these directors. Therefore, a positive relation between board diversity and compensation levels is expected. The two aforementioned effects result in the following hypothesis:
H9: The level of SB compensation is positively related to board diversity
Another aspect that is assumed to relate to the level of board independence and effective monitoring is the number of additional board memberships of SB members. The literature shows that positive and negative effects on compensation levels prevail. Intuitively, those SB members performing well at their tasks are more likely to be asked for additional board memberships (Yermack 2004). Indeed, according to the stewardship theory, the possibility to gain additional board memberships serves as an incentive to being a good steward for the company (Coles and Hoi 2003). Therefore, the number of additional board memberships may serve as a proxy for well-performing SB members.
In contrast to this positive effect, it is argued that so-called ‘busy’ SB members deliver monitoring services of poorer quality. In response to this, the DCGC imposes a maximum number of five additional board memberships for SB members to ensure proper performance by board members. Both this positive and negative effect of the monitoring quality of SB members could influence their compensation levels accordingly.
Additionally, in line with the resource dependency perspective, SB members’ expertise gained through multiple board memberships serves as a contributor to the firm’s required critical resources
(Andreas et al. 2012). Consequently, the level of resource contribution by Supervisory Board members through their networks is expected to greatly influence their compensation levels (Andreas et al. 2012). Moreover, the increased expertise, experience and knowledge developed through memberships at multiple companies, results in higher performance by SB members (Boyd 1996; Ryan and Wiggins 2004).
However, applying all of these arguments to determine the effect on individual compensation levels is challenging. In practice, Dutch firms set compensation levels equal for all members, only differentiating with respect to attendance rates per individual and specific position within the SB. However, one could argue that companies set SB compensation at certain levels to attract and retain SB members.
In line with the arguments of the positive effects, the number of additional board memberships is expected to serve as a proxy for the quality of effective monitoring. Overall, it is expected that compensation levels increase as a result, either through higher performance of SB members or through the desire to attract and retain highly skilled members. This leads to the final hypothesis:
H10: The level of SB compensation is positively related to additional board memberships
An overview of the expectations of the relations between the independent variables and compensation levels are shown below in table 1.
Table 1 Overview of the hypotheses
5. Data and variable construction
In order to empirically test these hypotheses, the dataset and the operationalization of variables that measure the specific relations need further specifications. The sample consists of Dutch firms that are listed on the Euronext stock exchange and are a part of the AEX- and AMX indices. Together, these companies form the 50 largest companies on the Dutch stock exchange in terms of market capitalization. The years of interest are 2007-2013. The independent variables at the firm level are lagged one year and run from 2006 until 2012. Thus, the sample contains both cross-sectional and time-series data.
Overview of the hypotheses Independent variables
Firm level Expected sign
H1. Firm size + H2. Risk + H3. Growth opportunities + H4. Internationalization + H5. Firm performance + Board level H6. Time commitment + H7. Board size -H8. Bargaining + Governance level H9. Diversity + H10. Add. Memberships +
The data sample consists solely of public firms and the reason for this is twofold. First, the DCGC is particularly aimed at public firms. Only listed firms are required to address their (non-) compliance with the code in their annual report. Secondly, listed firms need to publicly disclose information about compensation levels and curricula vitae of their SB and BoM members. Thus, the lack of publicly available data for non-listed firms further drove the decision to include only public companies.
The reason to consider only Dutch firms lies in the corporate systems of different countries. Historically, the Netherlands has had a two-tier system in which the SB and the BoM form two organs. Moreover, other countries that have a two-tier system allow for performance-based compensation, which is in contrast with the case of the Netherlands. Therefore, the comparability of Dutch companies with companies headquartered in other countries is low. Furthermore, including other countries in the sample might result in less powerful results as country specific effects may have undesired influences and induces a bias.
Furthermore, the data sample excludes companies with a one-tier structure for two reasons.20
First, compensation levels for SB members of these companies are materially higher. Moreover, the roles and responsibilities of SB members in a one-tier setting are different resulting in a low like-for-like comparison between the compensation levels in both settings as they (assumingly) reflect these different tasks.
In the end, I use two samples in the empirical study. The first one is an unbalanced panel dataset comprising of 55 firms over a period of 7 years with a total of 303 observations. Second, a subsample is tested for robustness purposes. This subsample includes the 31 firms that were consistently listed on the Dutch stock exchange throughout our time span, resulting in 217 observations. There are several reasons for testing two samples. First, in the initial sample, I do not want to exclude firms that are delisted or (de-) merged during the time period as this might bias the sample towards more stable companies with good performance. Therefore, including all firm-year observations is expected to paint a more realistic picture of the Dutch corporate environment and its coinciding SB compensation levels. On the contrary, testing a sample of consistent constituents increases the like-for-like comparison between years and interpretations of potentially changed results do not have to take into account differences in the yearly samples. A disadvantage of the subsample is the possible overrepresentation of larger firms. Companies that recede from the AEX end up in the AMX, and are thus still included in the subsample; while if this happens to AMX companies this results in exclusion from the subsample. For an overview of company names, the particular years for which they have data and the companies that are used in both samples, I refer to table 11 in the appendix.
There are multiple sources for the data and the majority of them are hand-collected. In particular, data for the dependent variable are taken from companies’ annual reports, AGM minutes, and partly from the Focus Orange Executive Compensation database.21 The independent variables at the firm level are collected from Thomson Datastream available at the University of Amsterdam. The data on the board- and governance level are hand-collected as well from curricula vitae in annual reports.
In order to operationalize the dependent variable, specific assumptions are made. First of all, each company reports policy levels regarding the compensation levels for the SB in its annual report as well as the actual compensation received. The policy levels reflect the compensation levels that
20
The one-tier companies that are excluded are Aperam, ArcelorMittal, D.E. MasterBlenders 1753, Gemalto, Reed Elsevier, Shell, Unilever, Heineken (as of 2013) and Air-france KLM.
21 In addition, Dr. K. van Veen gave parts of the data for the year 2007. This dataset was part of a larger dataset used for a thesis written by S. van der Meer in 2009. The author expresses her gratitude for being able to make use of this data.
under normal circumstances, without any irregularities, are rewarded to SB members. Here, I only include policy levels as actual compensations levels can differ widely due to resignations or entrance during the year, missed meetings and other unplanned events. Therefore, simplicity and comparability purposes drive the decision to only include policy levels.
Furthermore, three SB compensation components are taken into account. First, fixed fees are paid for the membership in a firm’s SB and these differ with respect to the position a person holds in the board. Each board has a chairman and members, and some companies have a separate vice-chairman. In this study, the vice-chairman’s compensation components are excluded due to a limited number of companies applying such a fee. Moreover, the relatively small difference between the fees for members on the one hand, and the vice-chairman on the other, further drives this decision. The second component of SB compensation is the committee fees, which are paid according to the participation of an individual in one or more committees and their specific position within this committee (i.e. chairman or member). Lastly, some firms pay their SB members attendance fees for the board- and/or committee meetings.
Additionally, the following calculations and assumptions are developed partly based on previous research (Bryan et al. 2000; Elston and Goldberg 2003; Linn and Park 2005; Andreas et al. 2012; Menozzi et al. 2014):
• Each member serves on the board for the entire year • Each member attends all meetings
Some companies report different meeting attendance fees for different nationalities or meetings locations. No such distinction is used in this dataset
Expense reimbursements are excluded since they are not considered direct compensation Some companies denote compensation levels in other currencies. In such cases, the yearly
average exchange rates are used to convert into euros 22
With regards to the committee fees, the assumption is made that committee seats and accompanying fees are divided equally over members and that the committees’ chairmen positions are fulfilled once
Per capita board compensation is derived from the total SB compensation and used as the dependent variable
The dependent variable is operationalized as the natural logarithm of per capita SB compensation in accordance with previous studies (Ryan and Wiggins 2004; Brick et al. 2006; Andreas et al. 2012). The reason for doing so is twofold. First, it allows for a positively skewed distribution to be more normally distributed. Second, the interpretability of the results becomes easier.
In addition to the dependent variable, the operationalization of the independent variables is largely in line with previous empirical studies interested in SB as well as BoM compensation. Table 2 gives an overview of the measures of the independent variables as well as their data sources. First of all, the firm level variables are lagged one year with the dependent variable, since policy compensation levels are usually determined prior to the year of interest. Moreover, policies regarding SB compensation are rarely changed during a financial year, which reduces the usefulness of including firm characteristics of the concurrent year. Therefore, the expectation is that firm level characteristics of the prior year will play a more prominent role in setting compensation levels. In
22 This is the case for Advanced Metallurgical Group (AMG) in the years 2008 until 2013. The yearly average (ask) exchange rate is taken from www.oanda.com.
addition, using one-year lagged variables decreases potential endogeneity (Linn and Park 2005; Brick et al. 2006). The measures of the firm level independent variables are all in line with previous research.
Secondly, the independent variables at the board level are not lagged as these measures are expected to influence compensation levels simultaneously. Especially the number of board meetings will influence compensation levels directly through attendance fees. The measure of board size, i.e. the total number of SB members at year-end, is in line with previous research by Ryan and Wiggins (2004) and Brick et al. (2006). The operationalization of the bargaining framework is in line with Fernandes (2008), who uses the proportion of non-executives in the board as the measure of the relative bargaining power.
Lastly, the two independent variables at the governance level are measured along five variables. The variable reflecting diversity in the SB is operationalized by four measures in line with previous research (Carter et al. 2003; Adams and Ferreira 2009). First, a dummy variable indicates whether a firm has at least one woman in the SB or not. Then, the fraction of female SB members is taken into account as a second variable. This allows me to determine if there is any difference in compensation levels when firms have all males or a combination of males and females in the SB.23 Moreover, I can test whether there is a positive relation between the fraction of women in the board and compensation levels. Similar variables are constructed for the nationality-diversity within the board. The second variable at the governance level, i.e. additional board memberships, is operationalized by per capita board memberships, thereby conforming to previous research (Yermack 2004). Board memberships are limited to other prominent, possibly foreign companies and include SB as well as BoM memberships. Thus, board memberships at non-profit organizations, such as schools, hospitals and charities, are not taken into account.
23 I did not find SBs consisting of all females.
Table 2 Variable definitions, descriptions and sources
This table depicts the different measures for the independent variables at the firm-, board- and governance level, as well as listing previous studies that used similar measures. Variable definitions and sources
Variable Description Source Previous research
A - Firm level
Sizet-1 Total sales and the number of employees, both in natural logarithm Datastream Brick et al. 2002; Linn and Park 2005; Fernandes 2008
Riskt-1 Price volatility of the stock returns Datastream Fernandes 2008; Andreas et al. 2012
Growtht-1 Market-to-book value Datastream Bryan et al. 2000; Adams and Ferreira 2009
Internationalizationt-1 Percentage of foreign sales Datastream Miller et al. 2002
Performancet-1 Dividend yield, ROA, ROIC Datastream Andreas et al. 2012; Menozzi et al. 2014
B - Board level
Time Natural logarithm of the number of board meetings per year Annual reports Hempel and Fay 1994; Brick et al. 2002; Andreas et al. 2012 Size Natural logarithm of the number of year-end SB members Annual reports Brick et al. 2002; Ryan and Wiggins 2004
Bargaining SB size divided by the BoM size Annual reports Fernandes 2008
C - Governance level
Diversity Dummy variable for women and other nationalities in the SB Annual reports Carter et al. 2003; Adams and Ferreira 2009 Fraction of women and other nationalities in the SB Annual reports Carter et al. 2003; Adams and Ferreira 2009 Additional seats Average number of additional board memberships Annual reports Yermack 2004
6. Descriptive statistics and model specifications
The data samples that are developed based on the specifics of the previous sections are used to analyze the different variables in more detail using descriptive statistics. Furthermore, they need to be adjusted to meet the assumptions on which the regression models rely. First, the summary statistics are aimed at giving more insight into the specifics of the dependent and independent variables used in the empirical tests. Figure 2 below depicts the development of the dependent variable, i.e. the annual per capita board compensation.
Figure 2 The development of the annual per capita board compensation
Overall, an upward trend during the period of interest is detected for the median values of annual per capita board compensation, resulting in a compound annual growth rate of 3.52%.24 The only exception to this trend is the year 2010, which shows a decrease in the average compensation level. From table 3 below, which provides more specific values of the dependent variable distinguishing between the two board positions, it seems that this is caused by a decrease in the members’ median value from 2009 to 2010. Generally, for the chairmen, the overall increase in the median values equals a compound annual growth rate of 2.52%. The median compensation level of members grew over the period of interest with a compound annual growth rate of 3.81%. On average, the chairmen earn 1.4 times more than members. Over the entire period, the minimum compensation levels for the chairmen are awarded by Vastned Retail in 2008 (AMX) and SNS Reaal in 2007 (AMX) and the maximum value by Unibail-Rodamco (AEX) in 2009. ING (AEX) in 2007 and Advanced Metallurgical Group (AEX) in 2013 respectively account for the minimum and maximum amounts awarded to the members.
24 Given the large standard deviations in table 3, the median values are considered a better reflection, less influenced by extreme observations.
Table 3 Descriptive statistics of the total SB compensation levels
The descriptive statistics of the independent variables are depicted in table 4 below. Regarding the firm level variables, the mean and median values reflect the average company in the dataset. The main takeaways of these variables are the large standard deviations of the firm size measures, indicating a wide variety of firms in terms of size. In addition, the relatively high mean value of the percentage of foreign sales indicates a high degree of internationalization among the firms. The board level and governance level variables depict the average SB of the dataset. The average additional board memberships of SB members are in line with DCGC posing the maximum of five additional board memberships. Overall, most of the variables in table 4 have large standard deviations that are sometimes even larger than their mean (i.e. market-to-book value, dividend yield, ROA). The primary aim of the subsequent analyses is to research whether these fluctuations also influence SB compensation levels.
Total SB compensation levels
Position Year 2007 2008 2009 2010 2011 2012 2013 2007-2013 Obs. 44 43 42 46 45 42 41 303 Chairman Mean 60,974 61,603 64,524 64,242 67,927 75,427 76,891 67,241 Median 57,750 56,667 57,788 58,917 62,083 66,750 68,750 60,700 SD 17,573 20,821 24,891 24,208 24,653 26,697 27,170 24,361 Minimum 30,000 30,000 30,500 31,975 30,500 41,500 41,500 30,000 Maximum 112,000 119,778 142,229 131,000 130,909 130,909 141,301 142,229 Member Mean 43,267 43,875 45,784 45,954 48,738 53,849 55,001 47,977 Median 42,750 42,800 45,955 44,643 47,000 51,500 55,542 46,500 SD 12,247 12,890 14,022 14,850 14,736 15,342 15,886 14,817 Minimum 20,000 23,000 23,804 23,000 24,379 33,500 33,500 20,000 Maximum 67,000 74,778 80,741 85,333 83,857 83,857 90,829 90,829
Table 4 Descriptive statistics of the independent variables
This table shows the aggregated descriptive statistics of all variables. Note that the firm level variables are measured over the years 2006-2012, whereas the other variables' period of interest is 2007-2013. The compensation statistics and the sales variable are in EUR.
Descriptive statistics
Variable Measures Obs. Mean SD 25th percentile Median 75th percentile
Per capita total compensation 303 51,044 15,533 39,500 48,333 59,571
(LN) Per capita total compensation 303 10.795 0.304 10.584 10.786 10.995
A - Firm level
Size Sales (in mln) 303 6,580 1,140 838 2,090 7,610
Employees 301 35,451 82,892 3,498 9,547 26,156
Risk Price volatility 272 28.559 8.455 22.680 27.025 33.920
Growth Market-to-book value 284 2.494 4.889 1.055 1.730 2.695
Internationalization % Foreign sales 281 65.643 34.645 50.580 67.440 87.320
Performance Dividend yield 284 3.778 4.051 1.575 3.065 5.015
ROIC 294 13.127 20.879 4.670 10.055 14.730
ROA 295 8.253 10.784 2.860 6.360 9.410
B - Board level
Time Meetings 301 8.525 3.422 6.000 8.000 10.000
Size Number of SB members 303 6.244 2.021 5.000 6.000 7.000
Bargaining SB size relative to BoM size 301 2.000 0.896 1.400 1.750 2.333
C - Governance level
Diversity Foreigners in the SB (0,1) 300 0.683 0.466 0.000 1.000 1.000
% Foreign SB members 300 0.290 0.258 0.000 0.286 0.500
Women in the SB (0,1) 300 0.527 0.500 0.000 1.000 1.000
% Female SB members 300 0.112 0.119 0.000 0.118 0.200
