Written by: Niels Bouma, S3205541

The effect of power and institutions on CEO compensation in an

European perspective.

Theme: comparative corporate governance & CEO careers

Master Thesis International Business & Management

MSc. International Business & Management Faculty of Economics and Business University of Groningen

Written by: Niels Bouma, S3205541

1

Abstract

CEO compensation has been on the rise exponentially, but explanations for this have not yet been conclusive. Research on compensation has mainly focused on firms in the USA. This research examines the strength of the managerial power theory and the role of the institutions labor unions and shareholder protection. It does so by looking at European firms in order to evaluate the strength of these theories outside of the USA. The data consists of cross-country panel data from 2013-2017 with 803 observations from several west- and south-European countries. The results did not find evidence for direct relationships between the institutional factors and CEO compensation, but did find a small significant relationship between CEO power, measured as CEO tenure, and CEO compensation. This research contributes to the literature by providing preliminary evidence for the role of CEO power in determining CEO compensation and also using a different geographic setting to examine its cross-country generalizability.

Introduction

Whenever a company publicly announces that the CEO receives a bonus, often there will be a discussion in the media about the fairness of the bonus. For example, in May 2020 Air-France KLM awarded its CEO with a bonus of 800.000 euros despite the bad economic situation for the firm due to the lockdown caused by COVID-19 (Pieters, 2020). Months later AIR-France KLM demanded its pilots to decrease their salary in order to save the firm from bankruptcy (Le Clercq, 2020). In a well-performing economy bonusses seem to be justified most of the times. When the economy is weak however the bonus is less straightforward, especially when the other employees are requested to accept a lower salary. Average CEO annual wages in the US have increased exponentially since the 1980’s up to 937% (Mishel & Davis, 2014). The CEO-to-worker compensation ratio, the ratio between the salary of the average employee and the CEO, has increased from ‘20 to 1’ to ‘295.9 to 1’ in 2013 (Mishel & Davis, 2014). This appears as an unfair distribution of income growth. The questions then arises of how this growth has manifested itself and why there is a difference between CEOs and the average employees.

Several theories have been developed that explain the development of the CEO compensation. For example the managerial power theory and the market based theory. Market based theory argues that the rise in compensation is caused by the supply and demand in the labor market. There is an increase in demand and willingness to pay for CEOs with the relevant

2 Trade has become more globalized over the years. This has led to challenges that arise,

because of the different ways that countries have in doing business and the structure of their institutions. There are studies that look at the institutional differences between countries and its effect on CEO compensation, with for example the differences in power structures (Greckhamer, 2015) or cultural differences (Tosi & Greckhamer, 2004), Comparing how institutional differences influence the CEO compensation may provide further inside into how firms operate in different countries. This is interesting, because research has shown that different institutional configurations have led to different strategic approaches of firms (Van Essen, Engelen & Carney, 2013). Examples of institutional differences are the protection of shareholders and the presence and size of trade unions (Huang, Jiang, Lie & Que, 2017). Trade unions differ a lot between countries and also their influence on firms (EurWORK, 2019), so it would be interesting to compare them cross-country in order to understand this institutional aspect. Shareholder protection has been widely researched in its relationship with a variety of factors such as ownership concentration (Burkart & Panunzi, 2006) and CEO turnover (Defond & Hung, 2004). However, when looking at the relationship with CEO compensation, much is still not clear. Van Essen, Heugens, Otten, Heugens & Van Oosterhout (2012) used an institutional approach in their research and looked at the quality of institutions to explain CEO compensation. However, they used a holistic view of institutions and did not specify in specific aspects of institutions. Van Essen et al. (2012) recommend to look at more specific institutional factors and also to look at interactions between institutions and their effect on firm factors. Otten & Heugens (2007) also did research on the influence of institutions on CEO compensations and found evidence that institutions do matter in CEO compensation outside Europe and they also do recommend further and more specific research about how institutions influence CEO pay.

The goal of this thesis is to examine how corporate governance configurations and national institutional characteristics can explain variety in the development of CEO compensation. So the research question is: How do the national institutional characteristics and corporate

governance mechanisms influence CEO compensation in firms?

3

Literature background

In order to answer the research question, a research model has been developed that is shown at the end of this section in figure 1. This model shows the expected relationships between the independent variables and the dependent variable. The variables and their expected

relationships will be further explained in this section.

Corporate Governance and Institutions

“Institutions are the humanly devised constraints that structure political, economic and social interaction” (North, 1991, p.1). There are informal institutions like traditions and customs

and there are formal institutions like the law and property rights. Through these institutions, human interaction becomes predictable due to its role as imposing consistency and

constraining or enabling certain behaviours (Hodgson, 2016). Institutions differ in meaning between countries and are formed differently (Aguilera & Jackson, 2003). Institutions are interdependent on each other. This interconnection creates differences between countries’ ways of doing business (Greckhamer, 2015). This is called the configurational approach or actor-centered institutionalism. This theory combines the firm level interactions between stakeholders with the interactions between the institutional environment and the firm. (Aguilera & Jackson, 2003). Aguilera & Jackson argue that agency theory alone cannot sufficiently explain corporate governance developments, because of three reasons: First of all it neglects the various identities that stakeholders have and that these differences lead to different interests. Secondly, agency theory argues that corporate governance comes to stand in a bilateral principal-agent conflict. Hereby other stakeholders that could be influenced by or want to influence the new corporate governance mechanisms, are ignored. The last reason they give, is that agency theory defines the environment very narrowly, only looking at shareholders rights, meaning that a lot of aspects and institutions are ignored.

Aguilera & Jackson then argue that the institutional perspective complements the agency theory by providing the more elaborate role of the environment, although the role of the environment should not be overstated. Actor-centered institutionalism holds into account the various aspects of both theories. This theory originated from Scharpf (1997). It argues that institutions shape the ideas and interests of the actors that interact within the institutions. The important part of the theory is that the actors still have a degree of freedom within the

institutional environment. The institutions do not fully determine the ideas, perceptions and consequently the actions of agents. Actors are able to make choices or even alter the

4 The research model is based upon the paper by Aguilera and Jackson (2003). In this article they proposed that the corporate governance has three core stakeholder groups; Management, Capital and Labor. These stakeholder groups are characterised each by their own dimensions. For example Capital has three dimensions. The first one identifies the interests of the

shareholders. These interests can be identified as either with a financial orientation or with a strategic orientation. The second dimension consists of how capital is used to let shareholders express their opinion and control of the firm. This can be either liquidity-focused by selling their shares or focused on commitment by increasing their total amount of shares in the specific firm. The last dimension is whether the ownership of the firm is consisting of debt or equity. These dimensions of corporate governance can differ per context or country and ultimately the shape of these dimensions influence the stakeholder’s interests within a specific firm.

The article argues that these differences are shaped by the institutional domains that influence corporate governance. Each dimension of corporate governance is influenced by institutional characteristics. For example, by looking at the dimension of capital, Aguilera and Jackson identified three domains; Property rights, interfirm networks and financial systems. The relationship between the dimensions and the domains is that the shape of the institutional domain influences the shape of corporate governance dimension. For example, the article argues that when the property rights within an institutional environment favour blockholders, then this leads to that capital is more likely to have strategic interests and capital is used to exercise their control by commitment. If the property rights favour the minority shareholders, then capital is more likely to have financial interests and liquidity is used as the strategy to exercise the shareholders’ control of the firm. Aguilera & Jackson argue that this is because when there is a low minority shareholder protection, smaller shareholders might have less right on information disclosure and have less power to address problems in the firm. It is then harder for small shareholders to sell their shares or to protect their financial interests. In order to gain control over the firm, shareholders need to acquire a greater amount of shares,

promoting the development of blockholders.

The general idea that the institutional domains influence the dimensions applies also to the other stakeholder groups and their complementary institutional domains. It is important to add that one domain does not necessarily determine the shape of one dimension. It is the

combination of domains that determine the final form of a certain dimension. Hence it is possible that there are institutional domains in a specific country that have a different effect on the stakeholders’ dimensions of corporate governance compared to other countries. Aguilera & Jackson give the example that countries can both have high minority property rights and high internetwork complexity. High minority property rights lead to more financial interest of shareholders, whereas high network complexity leads to more strategic interest of shareholders. This is because high network complexity is related to supplier relationships, board representation and the more complex a network is, the more chance there is that the network will arrange agreements to pursue common goals. Because these relationships

5 This research will examine institutional domains that are related to the dimensions capital and labor and their effect on the decision making of the firm. Management is not included in this research, because there is quite some research on the domains of this stakeholder group and its effect on CEO compensation (Tosi & Greckhamer, 2004) (Bryan, Nash & Patel, 2014) (Haynes, 2014). These articles already provide a rich insight in the effect of the domains of national culture on CEO compensation and thus it was decided to focus more narrowly on the other two dimensions. In this research the focus is on the decision making of CEO

compensation in the firm. These expected relationships and their effect on CEO compensation will be further explained in the next section.

CEO Compensation

The early theories about compensation are based upon the principal-agent dilemma of agency theory. This theory argues that there is a misalignment of interests between the shareholders as the principal and the CEO as the agent (Jensen & Meckling, 1976). This can lead to that the CEO acts in ways that is beneficial to its own interests but thereby harming the firm. In order to reduce this risk of misalignment, a contract will be created between the CEO and the firm. However, contracts are always incomplete, so instead of eliminating agency costs with an optimal contract, they have the function of trying to minimize the agency costs associated with contracting, monitoring and the enforcement of aligning the interests of the shareholders and the CEO (Bebchuk et al, 2002). These costs are created by the assumption that the CEO can behave opportunistically due to information asymmetry (Eisenhardt, 1989), thereby reducing shareholder value (Williamson, 1985). The contracts also determine the composition of CEO compensation. Jobs where the performance is easily monitored are more likely to be salary based, whereas when performance is hard to monitor, compensation is more based on commissions or a bonus (Eisenhardt, 1988).

This perspective that focuses on contracting is called the optimal contracting theory. The contract is a result of the negotiation between shareholders of the firm and the CEO where the incentives of the CEO are aligned with the incentives of the shareholders through

performance based compensation (Weisbach, 2007). The compensation is based upon the expected skills and performance of the CEO (Zabojnik & Murphy, 2004). The rise in CEO pay is caused by the increasing amount of skills that CEOs possess that leads to a contract with higher compensation. This is to attract the CEOs with good skills and to compensate them for the risk that CEOs receive when taking the responsibility of the firm. According to Otten & Heugens (2007), this theory is based upon two critical assumptions with their own problems: The first one is that the theory assumes that CEOs will not use their power in the contracting phase itself. But it is hard to argue that CEOs will not act in their own self-interest during the negotiations whereas they are expected to do so on other occasions (Bebchuk et al. 2002). The second assumption is that board members will indubitably protect the shareholders interest. With this assumption is the problem that board members may have shared goals with the management that oppose that of the shareholders and this can cause conflict and it

6 A second theory in CEO compensation has focused on the managerial power approach,

proposed by Bebchuk et al. (2002) as a response to back then dominant theory of the optimal contract. Bebchuk et al. did not agree with the two assumptions of optimal contracting and provided a new explanation. The managerial power theory argues that compensation is one of the possibilities of the CEO to extract rent for him- or herself. This is because CEOs have a considerate amount of power in the firm that they use to appropriate these rents. This power depends on the ownership structure and the board composition. Van Essen et al. (2015) also argue that the negotiations is not directly between the shareholders and the CEO, but between the board of the firm and the CEO. The CEO can increase its influence over the contract by increasing his influence on the board. It does not mean that a CEO can extract an unlimited amount of rents. It is limited when these additional rents are perceived as negative by the public. This creates an so-called outrage and limits further rent extraction for the CEO. The difference between the two theories is that in the managerial power approach CEO

compensation is not necessarily build upon executive skills or the external market, but based upon how much influence or power the CEO has in the board and in the negotiations. Both theories focus on the interaction between the CEO and the board. According to the theories the institutional context is not relevant in this negotiation. Bebchuk et al. do mention the rights of shareholders enforced by governmental law in their critique on the optimal contracting hypothesis, but limit their discussion to only two rights; File a lawsuit when the shareholders do not agree with the compensation package and the right to vote for stock option plans. Both rights are perceived as weak by Bebchuk et al. in protecting the

shareholders’ interests in compensation contracts. Their theory seems not to take into account that financial law changes and develops over time, which means that shareholders can gain (or lose) strength in protecting their interests and influence decisions of the firm (Katelouzou & Siems, 2015). These changes then also influence the amount of power the CEO has relative to the board and thus it should not be neglected.

In order to take the institutional environment into account this research takes an actor-centered institutionalism approach. As explained in the literature background section, this approach argues that the institutions influence and shape the interests and actions of the actors, but the actors are not fully constrained by the institutional environment. Aguilera & Jackson (2003) present the firm at the center of the model and argue that the stakeholder groups interact with each other in the general processes of decision-making and the use of firm resources. This research looks at institutional domains identified by Aguilera and Jackson and their effect on the decisions made in determining CEO compensation and also includes the managerial power theory in order to take into account the degree of freedom that the actors have in determining CEO compensation.

Hypothesis Development

CEO Power

7 CEO. It is thus important for the CEO to have power over the board so it can influence his/her own pay (Elhagrasey, Harrison & Buchholz, 1999). Elhagrasey et al. also argue that the board often goes along with the decisions of the CEO, because it can be beneficial to them as well. A higher CEO compensation seems to positively influence the reputation of the firm as well as providing an anchor for future director compensation (Crystal, 1991). These aligned interests increase the power of the CEO and its influence on the board. Daily & Johnson (1998) defined four sources of executive power: First of all there is structural power, referring to the position in the hierarchy that the CEO holds. This is an important source of power for the CEO, because the CEO is at the top of the hierarchy and is able to develop the strategies and the future direction of the firm. The second source is Ownership power, which refers to the fact that CEOs with a high amount of shares are able to steer the direction of the firm more than CEOs without an equity position. It also can prevent CEOs from being dismissed by the other shareholders. The third source of power comes from prestige. CEOs who have a high social prestige can help build legitimacy for the firm and also attract other prestigious groups to collaborate with the firm. The last source of power is expert power. This refers both to the ability to expand firm operations in a broader area as well as access to a wider network of contacts. Daily & Johnson argue that normally board members are appointed in order to gain access to new resources. When the CEO has his/her own broad network, then this increases the influence of the CEO as the board member is less useful compared to the CEO. These different sources of power for the CEO also means that CEO power can be measured in a variety of ways. A common measurement used to determine the power of the CEO is by looking at CEO duality. CEO duality means that the CEO has both control in the operations and management of the firm, and the CEO is the chairman of the board resulting in that the CEO has control over the board that is supposed to control the CEO (Conyon & Peck, 1998). This position can result in that the CEO has a lot of influence in firm decision-making,

including the CEO salary. The research that used this measurement as a proxy for CEO power found mixed evidence (Conyon & Peck, 1998)(Main & Johnston, 1993). Main et al. argue that this could be explained by the institutional differences between the samples. Vemala, Nguyen, Nguyen & Kommasani (2014)found that CEO duality was positively related with CEO compensation during the financial crisis. This could mean that CEOs were able to use their power to influence their compensation despite the bad financial situation.

This raises the question if a CEO with a lot of power can actually result in better firm performance. Some research shows that firms with a powerful CEO have a worse firm performance than firms with a less powerful CEO (Gupta, Han, Nanda & Silveri, 2018). In order to prevent this power, CEO duality is illegal in several countries in Europe, such as The Netherlands and Sweden or discouraged in a corporate governance code, such as the UK. CEO duality is a more common practice in the US. Because this research focuses on firms based in Europe, this means that CEO duality is not a strong measurement to examine CEO power. An alternative that can be used to look at the power of the CEO in the firm is CEO tenure.

8 time to understand control the information systems of the firm. This can lead to that the CEO can obscure certain information in order to increase the power that CEO has in the firm (Coughlan & Schmidt, 1985). A second reason is that the longer the CEO is at the top of the firm, the more time the CEO has to influence the board composition. Although the CEO is not directly in the board, studies still show that CEOs still have considerate influence in who to nominate and who not to nominate. Graham, Kim & Leary (2020) found evidence that after a new CEO is elected, every year with the new CEO leads to diminishing independence of the board. Their research also showed that this diminishing independence opens up the

opportunity for the CEO to increase its financial compensation. Over time this means that the CEO can shape the board in a way so that the board consists of members who would act in favour of the CEO (Fredrickson, Hambrick & Baumrin, 1988). As the board members can strongly influence the CEO compensation, it can be that they will vote in favour of higher CEO compensation. Hill & Phan (1991) also found a direct positive relationship between CEO tenure and CEO pay with the argument that powerful CEOs are able to avoid the monitoring of the shareholders and also shape the composition of the compensation in their favour compared to the shareholders. Dikolli, Mayew & Nanda (2014) found evidence that shareholders monitor less as the CEO tenure increases, providing more room for the CEO to expropriate the rents of the firm for him- or herself.

There is a lot of scientific debate in the literature about which theory can best predict the trend of increasing CEO compensation. Some evidence point in the direction of the managerial power theory. Van Essen et al. (2015) do find support for their managerial power theory in a meta-analysis, but cannot exclude with certainty other possible explanations. Akram, Haq & Umrani (2019) did also find support for managerial power theory as they found a significant relationship between CEO power and CEO compensation. Otten & Heugens (2007) find proof of the managerial power theory outside the US, but point out that contextual conditions should be included in order to make strong predictions. Other research however show evidence for the optimal contracting perspective. Bugeja, Matolcsy & Spiropoulos (2017) use the CEO pay slice as a measurement for the managerial power theory. This measurement examines the share of CEO pay compared to the total compensation of the top five executives. Their article shows that once the CEO pay slice is not in balance with the other executives, then this will be corrected in the following years, meaning that the contracts are evaluated and adapted to optimally align the interests of the CEO with the rest of the firm. Song & Wan (2019) also find evidence for the optimal contracting theory. In their article they compare managerial power with managerial ability and the results show that higher CEO compensation is better explained by rewards for the managerial talent/ability than by rent extraction of the CEO. These significant results for both theories mean that the theories should not be overlooked and that the discussion about which theory gives the best explanation is far from over.

9 Labor Unionization

The second dimension of corporate governance according to Aguilera & Jackson that could influence the CEO compensation, is the labor representation based upon the power of unions. Labor representation can be defined as employees’ ability to influence corporate decision

making and to control firms' resources (Aguilera & Jackson, 2003, p.455). This ability to

influence firms is strengthened by trade unions. The trade union has the goal to maximize the welfare of the employees by helping them in negotiating their worker conditions, such as salary, working hours and working conditions (Banning & Chiles, 2007). Aguilera & Jackson argue that unions, depending on the type of union, influence the way employees communicate with their employer and trying to influence firm decision-making. This can be either through external parties or through internal systems within the firm. Banning and Chiles argued that the trade union imposes an implicit tax on the CEO compensation. This tax is caused by the loss of shareholder value that is often the case when there is an increase in employee welfare. This loss of shareholder value is then also reflected in the CEO pay. Another study also showed the negative relationship between trade union representation and CEO compensation, due to motivations of the trade union trying to control the CEO compensation in order to take these resources from the CEO to the employees (Huang, Jiang, Lie & Que, 2017). This means that signs of increases in CEO compensation may then convince unions to stat bargaining for their employees to capture some of resources of the CEO. Based on this it would make sense that CEOs in firms with high trade union representation are less able to utilize their power to protect their pay. Trade union representation also differs per country (Blanchflower, 2006), so making a cross-cultural comparison might provide more insight than only looking at a specific country. Aguilera and Jackson give the example that Japan has enterprise-based unions that consists of members of a single firm. This means that the focus is more internal

communication systems. The US has multiple forms of unions that focus on external unions that communicate with the firm or industry. This external union could mean that the

employees are more likely to look after their own goals instead of the goals of the firm. In countries with internal communication, there may be more goal alignment between the employees and the firm, which could mean that they are more likely to make concessions in for example employee salary.

Hypothesis 2: Higher trade union representation in a country has a negative influence on CEO compensation in firms.

Shareholder Protection

Shareholder protection can be defined as the amount of protection minority shareholders have against controlling majority shareholders and give minority shareholders access to

10 relationship with a widely held ownership structure, meaning that there are a lot of minority shareholders. In countries with relative poor protection the ownership structure mostly consists of large shareholders or blockholders (La Porta et al. 1998)(Denis & McConnell, 2003). This is because control has a higher value in countries with poor shareholder

protection. This control may be used by large shareholders to influence the board composition and thus to protect their own interests (La Porta, Lopez-De-Silanez, Shleifer & Vishny, 2002). La Porta et al. (1998) also make this argument and argue that in countries with poor

shareholder protection shareholders will develop their own corporate governance mechanisms to protect their interests. This idea is supported by other research that found relationships between weak shareholder protection and the existence of blockholders in a firm (Kim, Kitsabunnarat-Chatjuthamard & Nofsinger, 2007).

Further research shows that shareholder protection influences the monitoring activity of the board. An article shows that weaker shareholder protection leads to more monitoring of the shareholders but the monitoring becomes less effective (Burkart & Panunzi, 2006). They also found that more monitoring does help against rent expropriation by the manger, but also weakens the initiative of managers, which can damage the performance of the firm. Burkart & Panunzi then argue that shareholder protection can replace some of these monitoring

activities. Managers are less able to extract firm rents for themselves when the shareholders have more rights to defend their interest and managers do not have to be extensively

monitored. This line of research shows that shareholder protection influences the firm in various ways that are relevant to the CEO.

It is important to discuss that shareholder protection also has a different relationship with CEO compensation in the theme of pay for performance. According to agency theory managers’ incentives need to be aligned with maximizing the shareholder wealth (Hubbard, Palia, 1995). However, the more these interests are aligned, the higher the amount of risk that the CEO is exposed to. This can result in that risk-averse managers choose less risky policies which do not optimize firm performance (Wright et al, 1996). This is reflected in firms who operate in risky environments. CEO compensation in those firms is relatively less based on performance compared to firms that operate in a less risky environment (Beatty & Zajac, 1994). This U-shaped relationship between level of pay for performance and firm

performance was found by Mishra, McConaughy & Gobeli, (2000). They argue that perfect compensation alignment with performance and shareholder wealth is detrimental to firm performance and thus not the solution to align the interests of managers and shareholders. Shareholder protection is also used in the literature to explain the level of pay for

11 Research has also founds some links between shareholder protection and CEO compensation. A study found that there is a negative relationship between shareholder protection and CEO compensation (Jiraporn, Kim & Davidson, 2015). This means that if there is more shareholder protection, then CEO compensation tends to decrease. This is because with high shareholder protection, powerful CEOs are not able to expropriate the rents when shareholder value is increasing. This is based upon the managerial power theory proposed by Bebchuk et al. This thus provides evidence for a relationship between the two factors, but Jiraporn et al. used only firms listen in the USA. Another study found a positive relationship between investor protection and CEO compensation, using European family firms (Bozzi, Barontini &

Miroshnychenko, 2017). They argue that when there is poor shareholder protection, CEOs are less likely to be exposed or punished for using their power and the CEO can thus freely raise its compensation without punishment. This is also similar to the Managerial Power Theory. Hypothesis 3: More shareholder protection in a specific country has a negative influence on CEO compensation in firms.

12

Research Methods

Data & Sample

The sample consists of 803 observations across 167 firms that are established in countries in West- or South Europe: France, Italy, Germany, The Netherlands, Belgium, Sweden, Ireland, Finland Denmark, Spain, Switzerland, United Kingdom. These countries were based upon the premise that data was available about their trade union density and shareholder protection index. The data about CEO compensation has been retrieved from the database of BoardEX. The data that is used in this research comes from the period 2013-2017. This period was chosen because it is the most recent period where data is available for this research. The period starts at 2013, because there was a financial crisis and recession in the years before 2013, that probably significantly influence the data in its own way.

Because data for European firms is still small, the decision was made that firms should have data about CEO compensation in at least three of the five years. Firms where the total CEO compensation equalled zero were also excluded, due to this being an exceptional situation due to the alternative mechanisms that determined the compensation. Also firms where the ISIN-number is not available were excluded from the sample, because then it would not be possible within the reasonable amount of time to cross-reference the firms to other databases in order to find their data about gross turnover, number of shareholders and profit.

BoardEX provides information about all board members of a specific firm. This means that when looking at the CEO compensation data, it was noticed that the CEO often could have multiple titles. This led to the decision that functions that mention CEO as an separate function were chosen. This means that functions such as Group CEO, Division CEO, Deputy CEO were excluded from the data, because these functions act differently compared to functions where the CEO is an unique function. This means that for example the function President/CEO/CFO is included in the data list. Data about the trade union density has been retrieved from OECD.stat. In some countries there was both administrative data and survey data available which deviate a little. It was decided that for every country the administrative data was used, because the survey data was not available for every country that appears in this research. Data about CEO tenure has also been retrieved from BoardEX. When the data of CEO tenure was not available, those firms were not included in the sample. Data about the shareholder protection index has been retrieved from the Worldbank. The data about firm gross turnover, number of shareholders and the total net profits of the firm have been retrieved from ORBIS.

Dependent Variable

13 It is also important to consider the future incentives in the compensation of the CEO, because otherwise it might give a distorted view of the compensation (Frydman, Jenter, 2010).

Independent Variables

CEO tenure is defined as the number of years that the specific person has spent as CEO of the firm.

Labor union representation can be defined as the amount of employees that are a member of a trade union. The measurement is composed as the trade union density on a country level. The variable is measured by the percentage of employees that are member of a trade union

compared to the total employees in the economy of the country. Although labor union density has been criticised for not being a very strong proxy for various of firm performance related indicators (Aidt & Tzannatos 2002), it is with the current data the closest proxy for labor union representation. But Aidt and Tzannatos also argue that trade union density could be useful in research about other variables, because trade unions seem to be effective in representing labor rights of the employees. This is because economies with a high union density seem to have lower income inequality and speedier adjustment to economic shocks. As firm performance is not a variable that is examined in this research, trade union density could still be a relevant measurement.

Shareholder protection has been measured by the Strength of Investor Protection Index created by the Worldbank. This index consists of three sub-indexes; Extent of disclosure index, extent of director liability index and the ease of shareholder suits index (Worldbank, 2020). The first one measures the extent to which investors have access to financial and firm data disclosed by the firms. The second one measures the extent to which board members can be held responsible for damage caused by transactions and what kind of punishments there are. The last one measures the extent to which investors have access to information when they file a lawsuit to the specific firm. All these sub-indices are measured on a 1 to 10 scale. The Strength of Investor Protection Index is then also calculated on a 1 to 10 scale, where 1 represents a low protection for shareholders and 10 equals a high protection for shareholders. There was no other literature found that used this measurement as a proxy for shareholder protection in a specific country, but this is partly because of the rework of the index in 2020, which means that there is little time for the scientific literature to evaluate the index. This does not mean that the index has no value, because the old version of the index received some attention from the literature (Rachisan, Bota-Avram & Grosanu, 2017) (Giofré, 2017).

Control Variables

14 Total firm assets and Tobin q were also considered as measurements of size, but their

exclusion will be explained in the results section. Because CEO compensation tends to react delayed to firm characteristics, the data about these factors was collected from 2012 up until 2016. This means that the turnover of the firm in 2012 was linked with the compensation in 2013 and the same method was used to link the other years. Another control variable is the firm profits. This variable is added because of research that found that net profits influenced the CEO compensation (Randoy & Nielsen, 2002). The number of shareholders is the third control variable and it is also used as a proxy for firm size. Another factor is that countries vary between each other in a myriad of characteristics and these can lead to so called country-effects. In order to account for this variety dummy variables were used for the countries. This means that for example the dummy variable Germany has a value of 1 when the firm

observation is from a German firm and 0 when it is not. The nationality of a firm was determined by the location of the headquarters of that firm. This is consistent with all the other country dummy variables. Time dummies were also used to exclude the possibility that one of the years was extraordinary compared to the other years and that this could influence the analysis. These are also known as time effects. The dummy is defined such that, for example 2016 has a value of 1 for all observations in 2016 and 0 for the observations in the other years.

Results

This analysis is done with the help from the software program STATA 16. Table 1 shows the descriptive values of the variables. Panel data is used in this analysis with a total of 643 observations. TUD is an abbreviation for Trade Union Density and SPI is the abbreviation for Shareholder Protection Index. TimeRole is the variable that measures CEO tenure. CEO compensation has an average of 2.557.914 million euros with a standard deviation of 2.545.763 million euros. The average tenure of this sample was 5.311 years with a standard deviation of 5.133.

Table 1 Descriptive Statistics

Variable Obs Mean Std. Dev. Min Max

Compensation 803 2557.914 2545.763 4 27082 TimeRole 803 5.311 5.133 0 29.1 TUD 803 19.456 10.003 8.98 67.2 SPI 803 6.383 1.094 4.3 7.88 Turnover 803 21751653 35547923 3.857 3.542e+08 Profits 803 1048953 2207171.7 -9198000 20245567

15 coefficients. During the process two control variables were excluded from the research. The output data in appendix A shows that the variables caused for multicollinearity or simply added no value to the research. Table 1 of appendix A shows the correlation matrix and Tobins Q does not correlate significantly with the independent variable and was thus

excluded. The variable total assets was also omitted from the research, because it caused too much multicollinearity with the other variables as can be seen in figure 1 of appendix A. Because total assets showed the highest VIF-value it was picked over turnover.

Table 2 Pairwise correlations

Variables (1) (2) (3) (4) (5) (6) (1) LogComp 1.000 (2) TimeRole 0.080** 1.000 (3) TUD -0.220*** -0.133*** 1.000 (4) SPI -0.501*** -0.018 0.362*** 1.000 (5) Profits 0.306*** -0.013 -0.134*** -0.243*** 1.000 (6) LogTurnover 0.699*** 0.011 -0.336*** -0.594*** 0.364*** 1.000 *** p<0.01, ** p<0.05, * p<0.1

Table 2 shows the correlation matrix. These correlations can give an indication of the direction of the relationship. TUD has a negative significant correlation with the CEO

compensation, which is consistent with the hypothesis. SPI has the same negative correlation with CEO compensation as TUD. There was also a small positive correlation between the CEO tenure and the compensation.

Moving on to the regressions, two models were created and used in the regression. Model 1 consists of the standard situation with CEO compensation as the dependent variable and with the control variables but without the independent variables. Model 2 includes the independent variables alongside the control variables.. Table 3 contains the results of the regression

analysis.

Table 3 Regression Results

(1) (2)

VARIABLES Model 1 Model 2

LogTurnover 0.419*** 0.413***

(0.0306) (0.0315)

Profits 4.51e-08*** 4.75e-08***

16 UK 1.601*** 3.896 (0.381) (15.01) Switzerland 2.480*** 5.119 (0.379) (11.56) France 2.355*** 5.638 (0.366) (17.80) Spain 2.321*** 5.129 (0.381) (14.62) Ireland 2.464*** 4.752 (0.369) (14.54) Italy 2.608*** 4.000 (0.402) (7.241) Scandinavia 2.820*** 1.901 (0.408) (4.888) Netherlands 2.662*** 5.263 (0.374) (12.57) TimeRole 0.0261*** (0.00566) TUD 0.0762 (0.411) SPI -0.0808 (1.964) Constant 2.157*** -1.466 (0.424) (10.12) Observations 803 803 Adjusted R-squared 0.587 0.598

Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Several statistical tests were used in order to test whether the assumptions of an OLS regression were fulfilled. The output of all the statistical tests can be found in appendix B. The DFBETA-scores were calculated for each variable in order to test whether there were outliers who significantly influenced the regression on their own. Observations with a score higher than 1 or -1 are significantly different than the other observations. However, the observations in the data showed no scores above 1 or below -1, so there was no need to omit certain observations. Figure 1 of appendix B shows the leverage-versus-squared-residual plot shows the distribution of the observations with the outliers.

The next assumption tested is normality. Figure 2 of Appendix B shows the normal

probability plot of the residuals. This graph shows that the residuals do not follow a straight line but move in a s-pattern. Further analysing this pattern a Kernel density estimate was used. This graph can be seen in figure 3 also shows a substantial deviation from the normal

17 because the p-value is lower than 0.05. This means that the hypothesis that the variable is normally distributed, is rejected.

The Breusch-Pagan test was used to determine the heteroskedasticity in the sample. This tests uses hypotheses with H0 stating that there is constant variance in the sample. From the output of the test in figure 5 can be concluded that H0 cannot be rejected with p>0.05. This means that we cannot reject the hypothesis that the sample has constant variance. The residual versus fitted values model shows the variance in the sample and can be seen in figure 6.

VIF-values were used in order to look at the assumption of multicollinearity. This was

executed by determining the VIF-values of the independent variables. The VIF values need to be below a critical value. This value can either be 10, 5 or 2 depending on the

conservativeness of the researchers. The output of figure 7 shows that all the VIF values are below 2, which means that there are no substantial levels of multicollinearity in this model. Concluding from the analysis there is a problem with normality. Transformations of the variables was not a solution for this problem, because some of the variables were already transformed and transforming the other variables did not show significant differences. However, according to the Central Limit Theorem, normality can be assumed as long as the sample size is greater than 30 or more conservatively 100. The sample size of this study is larger than 100 which means that we can assume that the sample is normally distributed. This does mean that the inferences from the regression analysis are less conclusive. A robust regression analysis, shown in figure 8 was also used to compare the coefficients of the independent variables. The coefficients do not seem to differ significantly which provides further certainty that the effects are to a certain extent robust

Hypothesis Testing

An OLS-regression was to test all the hypotheses. Hypothesis 1 states that a higher CEO tenure has a positive influence on CEO compensation in firms. In the regression CEO tenure is the independent variable and CEO compensation the dependent variable. From the analysis can be concluded that there is a significant effect of CEO tenure on total compensation with b=0.0261, p=<0.01. In order to calculate the interpretable effect, the coefficient was

exponentiated, subtracted by 1 and then multiplied by 100 percent or in formula: (e0.0261 -1)*100%)). The interpretable coefficient is 2.644. This means that every 1 year increase in CEO tenure leads to a 2.644% increase in CEO compensation. Hypothesis 1 is thus

confirmed, but it is only a small effect.

18 Hypothesis 3 states that more shareholder protection in a specific country has a negative influence on CEO compensation in firms. In the third regression SPI is the independent variable and CEO compensation the dependent variable. The analysis shows there is no significant relationship between Shareholder Protection Index and CEO compensation with b=-0.0808, p>0.1. Hypothesis 3 is thus rejected. Shareholder protection has no influence on CEO compensation.

Looking at the control variables there are some significant coefficients for profits, size and number of shareholders. However, profits and number of shareholders have such a small coefficient that practically they has not a powerful effect on CEO compensation. Turnover (size) does have a significant effect on CEO compensation in both models with b=0.412, p=<0.01. Because both the independent and the dependent variable are logarithmic, another formula is used to calculate the interpretable coefficient; (1.10.412 -1)*100%))= 4,004%. This means that for every 10% increase in turnover leads to a 4.004% increase in CEO

compensation, which is not large, but not small enough to ignore. Another thing to notice is that in model 1 several countries have an effect on CEO compensation, but in model 2 this effect is not significant. Looking at the year dummies, it can also be concluded that time did not have an effect on CEO compensation.

Discussion

The goal of this research is to understand the factors that influence the CEO compensation in Europe by looking at the managerial power theory. It also tries to explain why CEO

compensation differs per country by looking at the variety in institutions. The results of this study suggest that CEO power, measured as CEO tenure, has a small influence on CEO compensation. Furthermore the results did not find a direct effect of institutions on the CEO total compensation.

Our first hypothesis stated that CEOs with a higher tenure have more power over the firm and are able to influence their compensation. The results confirmed this hypothesis, but only with a small effect. Comparing the results to the literature the evidence suggest that CEOs who are longer in the firm, are able to influence the firm and board in the favour of the CEO. This is consistent with results from other studies (Hill & Phan, 1991) (Fredrickson, Hambrick & Baumrin, 1988) and with the managerial power theory. However, because this study does not look at variables that might explain this direct effect, it does not rule out alternative

19 Other results that stand out are non-significant relationships between the institutions and CEO compensation. This is not consistent with the work of Huang et al. (2017), who did find a negative relationship between unions and CEO compensation. This could be explained by the fact that Huang et al. used a different institutional union proxy. They looked at industrial unions whereas this study used national union scores. Aguilera & Jackson (2003) also argued that there are different types of unions. The national union scores used in this study do not differentiate between the various types of unions. This could mean that certain union types are more actively involved in influencing firm decision making than others. Banning & Chiles (2007) also found a positive relationship between unionization and CEO pay, but in this study unionization was measured at firm level looking at union- versus non-union firms. This could also explain the non-significant relationship. High national unionization does not necessarily mean that firms have a high firm-level unionization. However, data about firm-level

unionization is very hard to access and because of the limited amount of time for this research, using this measurement was not possible.

The last main finding was that there was no significant relationship between SPI and CEO compensation and thus there was no proof for hypothesis 3. Bozzi et al. (2017) did found a direct negative relationship, but in this research the Anti-director rights index (ADR-index) was used as measurement for shareholder protection. This measurement is also used on a national level, so it is less obvious why no relationship was found this time. The ADR-index was originally created by La Porta et al (1998) and was later refined by Spamann (2010). The improved version does try to measure the same concepts, but the measurements of the

concepts were improved. The ADR-index has three measurements of shareholder voting; Allowing to vote by mail, not requiring shareholders to deposit their shares before meetings and shareholders are allowed to vote cumulatively (La Porta et al, 1998). The ADR-index has three measurements for minority shareholder protection. The first one is about that minorities have a mechanism to defend their interests against decisions made by the management. The second one is that the minimum required amount of shares to have the right to organize an extraordinary shareholder meeting is 10% or below. The last one is that shareholders have the first right to buy new shares. Comparing these to the Shareholder Protection Index used by the Worldbank shows that there are no or very little similarities. The SPI index focuses heavily on information disclosure the ability to hold the management responsible for . It could be argued that when the SPI score is high, there is a high requirement of information disclosure and this prevents the power of the CEO to hide key information from shareholders, proposed by Coughlan & Schmidt, (1985). However, by comparing the two indexes, the ADR-index seems to capture a wider spectrum of shareholder rights than the SPI-index, which may explain why there was no significant relationship found in this research.

20 information disclosure requirements and also not the ability of shareholders to hold directors accountable. This leads to that these two indexes measure different things. The GI has a closer link to firms than national indicators, because the index focuses on the firm and not on the nationality of the firm. However, this also means that collecting data will be quite difficult, because all the information needs to be available in order to give a score and some items such as the presence of golden parachutes are most of the times not publicly available.

Limitations

Every study has its limitations and it is important to be aware of these limitations. The biggest limitation of this study is the relative small sample size. This is caused by the relative small availability of CEO compensation data in Europe. Another problem within this lack of data was that data of CEO compensation was not consistently tracked through the years. This means that firms needed to be excluded, because data of compensation was only available in 2013 and 2014, but not after those years, although the company was still operating. This small sample size also means that the significant relationships were not really strong and

interpretations should be made cautiously. Another limitation of this study is that CEO tenure does not necessarily measure the actual use of power as a CEO. Hill & Phan (1991) also mention this limitation as it is only an indirect proxy of CEO power and it does not measure the influence of the CEO directly.

Avenues for future research could be looking at the firm variables that function between the institutional factors and CEO compensation, in order to get a more extensive view on this relationship. Another idea for future research could be about looking for more direct measurement of CEO power and influence, because the current proxies do not measure whether the CEO actually uses its influence to raise its compensation.

A final idea is the further development of data availability of compensation in European firms. This could lead to more complete and extensive datasets, which could be used for more extensive research with more complex relationships.

Conclusion

21

References

Aguilera, R.V., Jackson, G. (2003). The Cross-National Diversity of Corporate Governance: Dimensions and Determinants. The Academy of Management Review, 28(3): 447-465. Aidt, T., Tzannatos, Z. (2002). Unions and collective bargaining - economic effects in a global environment. Directions in Development. Washington, D.C. : The World Bank. Available at http://documents.worldbank.org/curated/en/831241468740150591/Unions-and-collective-bargaining-economic-effects-in-a-global-environment

Akram, F., Haq, M.A.U., Umrani, W.A. (2019). Assessing the Effect of Managerial Power on Firm Performance through the Perceptual Lens of Executive Remuneration,

Pertanika Journal of Social Sciences & Humanities, 27(1): 293-309.

Banning, K., Chiles, T. (2007). Trade-Offs in the Labor Union-CEO Compensation Relationship. Journal of Labor Research, 28: 347-357.

Beatty, R.P, Zajac, E. (1994). Managerial incentives, monitoring and risk bearing: A study of compensation, ownership and board structure in initial public offerings. Administrative

Science Quarterly, 39(2): 313-335.

Bebchuk, L.A., Fried, J.M. (2003). Executive compensation as an agency problem. Journal of

Economic Perspectives, 17(3): 71-92.

Bebchuk, L.A., Fried, J.M., Walker, D.I. (2002). Managerial power and rent extraction in the design of executive compensation. University of Chicago Law Review, 69(3): 751–861. Belloc, F. (2013). Law, finance and innovation: the dark side of shareholder protection.

Cambridge Journal of Economics, 37: 863-888.

Blanchflower, D.G. (2006). A cross-country study of union membership, IZA Discussion

Papers, No. 2016, Institute for the Study of Labor (IZA), 1-58.

Bozzi, S., Barontini, R., Miroshnychenko, M. (2017). Investor protection and CEO compensation in family firms. Corporate Ownership & Control, 14(2): 17-29.

Bryan, S., Nash, R., Patel, A. (2014). Culture and CEO compensation. Fordham University

Working Paper, available at https://extranet.sioe.org/uploads/isnie2014/bryan_nash_patel.pdf Bugeja, M., Matolcsy, Z., Spiropoulos, H. (2017). The CEO pay slice: Mangerial power or optimal contracting? Some indirect evidence. Journal of Contemporary Accounting &

Economics, 13(1): 69-87.

Burkart, M., Panunzi, F. (2006). Agency conflicts, ownership concentration, and legal shareholder protection. Journal of Financial Intermediation, 15(1): 1-31.

Coughlan, A.T., Schmidt, R.M. (1985). Executive compensation, managerial turnover, and firm performance. Journal of Accounting and Economics, 7(1-3): 43-66.

Crystal, G.S. (1991). In search of excess: The overcompensation of American executives.

22 Daily, C.M., Johnson, J.L. (1997) Sources of CEO power and firm financial performance: A longitudinal assessment. Journal of Management, 23(2): 97-117.

Defond, M.L., Hung, M. (2004). Investor protection and corporate governance: Evidence from worldwide CEO turnover. Journal of Accounting Research, 42(2): 269-312.

Denis, D.K., McConnell, J. (2003). International Corporate Governance. Journal of Financial

and Quantitative Analysis, 38(1): 1-36.

Dikolli, S.S., Mayew, W.J., Nanda, D. (2014). CEO tenure and the performance-turnover relation. Review of Accounting Studies, 19: 281-327.

Eisenhardt, K.M. (1988). Agency- and Institutional-Theory Explanations: The Case of Retail Sales Compensation. The Academy of Management Journal, 31(3): 488-511.

Eisenhardt, K.M. (1989). Agency Theory: An Assessment and Review. The Academy of

Management Review, 14(1): 57-74.

Elhagrasey, G.M., Harrison, J.R., Buchholz, R.A. (1999). Power and pay: the politics of CEO compensation. Journal of Management and Governance, 2: 309-332.

EurWORK. (2019). Trade union density. Available at:

https://www.eurofound.europa.eu/observatories/eurwork/industrial-relations-dictionary/trade-union-density

Fredrickson, J.W., Hambrick, D.C., Baumrin, S. (1988). A model of CEO dismissal. Academy

of Management Review, 13(2): 255-270.

Finkelstein, S. (1992). Power in top management teams: dimensions, measurement, and validation. Academy of Management Journal, 35(3): 505–538.

Frydman, C., Jenter, D. (2010). CEO compensation. Annual Review of Financial Economics, 2: 75-102.

Giofré, M. (2017). Financial education, investor protection and international portfolio diversification. Journal of International Money and Finance, 71(C): 111-139.

Gompers, P., Ishii, J., Metrick, A. (2003). Corporate governance and equity prices. The

Quarterly Journal of Economics, 118(1): 107-156.

Graham, J.R., Kim, H., Leary, M. (2020). CEO-board dynamics. Journal of Financial

Economics, 137(3): 612-636.

Greckhamer, T. (2015). CEO compensation in relation to worker compensation across countries: The configurational impact of country‐level institutions. Strategic Management

Journal, 37(4): 793-815.

23 Gupta, V.K., Han, S., Nanda, V., Silveri, S.D. (2018). When Crisis Knocks, Call a Powerful CEO (or Not): Investigating the Contingent Link Between CEO Power and Firm Performance During Industry Turmoil. Group & Organization Management, 43(6): 971-998.

Haynes, K.T. (2014). Something old, something new: Culture and CEO compensation revisited. American Journal of Business, 29(1): 5-25.

Hill, W.L., Phan, P. (1991). CEO tenure as a determinant of CEO pay. The Academy of

Management Journal, 34(3): 707-7017.

Hodgson, G.M. (2016). What are Institutions? Journal of Economic Issues, 40(1): 1-25. Huang, Q., Jiang, F., Lie, E., Que, T. (2017). The Effect of Labor Unions on CEO Compensation. Journal of Financial and Quantitative Analysis, 52(2): 553-582. Hubbard, R., Palia, D. (1995). Executive pay and performance evidence from the U.S. banking industry. Journal of Financial Economics, 39(1): 105-130.

Hüttenbrink, A., Oehmichen, J., Rapp, M.S., Wolff, W. Pay-for-performance – Does one size fit all? A multi-country study of Europe and the United States. International Business Review, 23(6): 1179-1192.

Jensen, M.C., Meckling, W.H. (1976). Theory of the firm: managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3: 305-360.

Jeppson, C.T., Smith, W.W., Stone, R.S. (2009). CEO compensation and firm performance: Is there any relationship? Journal of Business and Economics Research, 7(11): 81-93.

Jiraporn, P., Kim, Y.S., Davidson, W.N. (2005) CEO compensation, shareholder rights, and corporate governance: An empirical investigation. Journal of Economics and Finance, 29: 242-258.

Katelouzou, D., Siems, M. (2015). Disappearing Paradigms in Shareholder Protection: Leximetric Evidence for 30 Countries. Journal of Corporate Law Studies, 15(1): 127-160. Kim, K.A. Kitsabunnarat-Chatjuthamard, P., Nofsinger, J.R. (2007). Large shareholders, board independence, and minority shareholder rights: Evidence from Europe. Journal of

Corporate Finance, 13: 859-880.

La Porta, R., Lopez-De-Silanez, F., Shleifer, A., Vishny, R.W. (1998). Law and finance.

Journal of Political Economy, 106(6): 1113-1155.

La Porta, R., Lopez-De-Silanez, F., Shleifer, A., Vishny, R.W. (2002). Investor Protection and Corporate Valuation. The Journal of Finance, 57(3): 1147-1170.

Le Clercq, P. (2020). Wat houden KLM-piloten nog over na flinke salarisverlaging?

Retrieved from: https://www.rtlnieuws.nl/economie/life/artikel/5167915/hoeveel-verdienen-piloten-bij-klm-en-dat-meer-dan-elders

24 Mishel, L., Davis, A. (2014). CEO pay continues to rise as typical workers are paid less.

Economic Policy Institute, #380, 1-12.

Murphy, K.J., Zabojnik, J. (2004). CEO Pay and Appointments: A Market-Based Explanation for Recent Trends. New Developments in Human-Capital Theory, 94(2): 192-196.

North, D.C. (1991). Institutions. Journal of Economic Perspectives, 5(1): 97-112.

Otten, J., Heugens, P.P.M.A.R. (2007). Extending the managerial power theory of executive pay: A cross-national test. ERIM Report Series Reference No. ERS-2007-090-ORG,

Available at SSRN: https://ssrn.com/abstract=1088215

Pieters, J. (2020). Air France-KLM CEO awarded €768K Bonus; Dutch gov’t votes against. Retrieved from: https://nltimes.nl/2020/05/26/air-france-klm-ceo-awarded-eu768k-bonus-dutch-govt-votes

Rachisan, P.R., Bota-Avram, C., Grosanu, A. (2017). Investor protection and country-level governance: cross-country empirical panel data evidence. Economic Research-Ekonomska

Istraživanja, 30(1): 806-817.

Randoy, T., Nielsen, J. (2002). Company performance, corporate governance, and CEO compensation in Norway and Sweden. Journal of Management and Governance, 6: 57-81. Scharpf, F.W. (1997). Games real actors play: Actor-centered institutionalism in policy research. Boulder, CO: Westview.

Song, W.L., Wan, K.M. (2019). Does CEO compensation reflect managerial ability or

managerial power? Evidence from the compensation of powerful CEOs. Journal of Corporate

Finance, 56: 1-14.

Spamann, H. (2010). The “anti-directors rights index” revisited. The Review of Financial

Studies, 23(2): 467-486.

Stapledon, G.P. (2004). The pay for performance dilemma. Available at SSRN: https://ssrn.com/abstract=553061

Tosi, H.L., Greckhamer, T. (2004). Culture and CEO compensation. Organizational Science, 15(6): 657-670.

Tosi, H.L., Werner, S., Katz, J.P., Gomez-Mejia, L.R. (2000). How Much Does Performance Matter? A Meta-analysis of CEO Pay Studies. Journal of management, 26(2): 301-339. Van Essen, M., Engelen, P.J., Carney, M. (2013). Does “Good” Corporate Governance Help in a Crisis? The Impact of Country- and Firm-Level Governance Mechanisms in the European Financial Crisis. Corporate governance: An International Review, 21(3): 201-224.

Van Essen, M., Heugens, P.P.M.A.R, Otten, J., Van Oosterhout, J.H. (2012). An institution-based view of executive compensation: A multilevel meta-analytic test. Journal of

25 Van Essen, M., Otten, J., Carberry, E.J. (2015). Assessing Managerial Power Theory: A Meta-Analytic Approach to Understanding the Determinants of CEO Compensation. Journal

of Management, 41(1): 164-202.

Vemala, P., Nguyen, L., Nguyen, D., Kommasani, A. (2014). CEO Compensation: Does Financial Crisis Matter? International Business Research, 7(4): 125-131.

Weisbach, M.S. (2007). Optimal Executive compensation versus managerial power: A review of Lucian Bebchuk and Jesse Fried’s Pay without performance: The unfulfilled promise of executive compensation. Journal of Economic Literature, 45(2): 419-428.

Williamson, O.E. (1985). The Economic Institutions of Capitalism. New York: The Free

Press.

Worldbank. (2020). Protecting minority investors measurement. Retrieved from:

https://www.doingbusiness.org/en/methodology/protecting-minority-investors#:

26

Appendix A

Table 1 Pairwise correlations with TobinQ

Variables (1) (2) (3) (4) (5) (6) (7) (8) (1) LogComp 1.000 (2) TimeRole 0.080** 1.000 (3) TUD -0.220*** -0.133*** 1.000 (4) SPI -0.501*** -0.018 0.362*** 1.000 (5) Profits 0.306*** -0.013 -0.134*** -0.243*** 1.000 (6) Logassets 0.601*** -0.069* -0.286*** -0.536*** 0.395*** 1.000 (7) tobinQ 0.003 0.006 0.244*** 0.213*** 0.026 -0.362*** 1.000 (8) LogTurnover 0.699*** 0.011 -0.336*** -0.594*** 0.364*** 0.921*** -0.265*** 1.000 *** p<0.01, ** p<0.05, * p<0.1

Figure 1 VIF values Assets

27

Appendix B

Outliers analysis

28 Normality Analysis

29 Figure 3 Kernel density estimate of the residuals

30 Heteroskedasticity analysis

Figure 7 Multi-collinearity analysis Figure 5 Breusch-Pagan test

Figure 6 Residuals versus fitted values plot