Ownership concentration and owner types : executive tenure in Europe affected by ownership structures

(1)

Master’s Thesis

Ownership concentration and owner types:

executive tenure in Europe affected by

ownership structures

Martin van Nistelrooij 10248390

(2)

Statement of originality

This document is written by Martin van Nistelrooij who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

4.3. Data sources...23 4.4. Dependent variable...24 4.5. Independent variables...25 4.6 Control variables...28 4.7 Analyses...28 5. Results...29 5.1. Descriptive statistics...31 5.2. Pre-tests...31 5.3. Tests...33 6. Discussion...36 6.1. Interpretations...36 6.2. Implications...37 6.3. Limitations...40 6.4. Future research...42 7. Conclusion...43 Reference list...45 Appendix...50

(4)

ABSTRACT Examining the impact of institutional ownership on corporate governance practices, this study conducts an empirical analysis on the effect of ownership structures on executive tenure. With the link between performance and executive tenure as an indicator of good corporate governance, the effectiveness of corporate governance is tested in the context of the agency theory. Two prevalent mechanisms that contribute to the possible effect of investor influence are examined: ownership concentration and pressure-sensitivity. A cross-sectional study is conducted and regression analyses are performed on data of 195 European firms in the year 2015. The results indicate a positive relationship between performance and executive tenure, but the effectiveness of corporate governance cannot be assessed for as there is no relationship found between ownership concentration and executive tenure. In addition, there is no moderating effect found of pressure-sensitivity. These findings are in favor of the stewardship theory and contributes to the understanding of the corporate governance mechanisms used by shareholders.

1. Introduction

In April 2016, the European hotel talent market was impacted by the announcement of the acquisition of Carlson Rezidor by the Spanish HNA group. With acquisitions argued to be in the best interest of shareholders, the consequences of this major M&A deal had an impactful echo in the shareholders’ meetings and ultimately the board room. A battle started between Oceanwood, a large British hedge fund and the acquisition-hungry Chinese conglomerate HNA Group. Oceanwood had the concern that this accelerated movement into the European leisure market would have unfavorable consequences for the performance the collectively owned NH Hotels, which had a market capitalization of €1.6B at that time (Fortado, 2016)). Both institutional investors accused each other of misalignment of their interest with those of the shareholders and HNA even wrote a letter to the shareholders stating that Oceanwood pursued a dangerously short-term oriented strategy (“HNA Sends Open Letter”, 2016). Ultimately, Oceanwood forced the dismissal of CEO Federico J. Gonzales

(5)

Tejera, as they argued that conflicting interests had emerged within the company, which was later labeled by HNA as a “coup” to retain influence (Fortado, 2016).

This example is a clear demonstration of the expression of investor influence of institutional owners impacting executive tenure in Europe. When foreign institutions own relatively more shares than domestic owners within a particular country, foreign institutions can have a major impact on local corporate governance (Gillan & Starks, 2003). It is argued that corporate governance practices from Anglo-American economies have transferred to other markets such as Europe by growing cross-border institutional investments (Aguilera & Jackson, 2010). In addition, these institutional investors are proven to take a more active stance towards shareholder activism than in other countries (Gillan & Starks, 2003). Most studies on foreign ownership examine the effect on firm outcomes, whereas the effects on organizational processes and corporate governance is still underexposed (Desender et al., 2014). A specific topic of interest is the expression of investor influence by dismissal of poor performing executives. As some studies have shown that firm performance does improve after replacing top executives in the U.S., there are also studies that find no evidence for a link between post-turnover performance in stakeholder-oriented economies (Firth et al., 2006). Findings of other studies show that large institutional investors may have business relationships with affiliated firms which results in managerial entrenchment (Gillan & Starks, 2003). This results in no significant improvement of performance either, which leaves the monitoring role of institutional investors open for debate. These divergent findings indicate an underexplored field of study of the effect of institutional ownership on executive tenure. This study aims to address this gap in existing literature as it explores the impact of investor influence on executive tenure.

In the following section investor influence is addressed in the context of globalization with respect to previous research. Two main determinants of investor influence emerge, which are ownership concentration and pressure-sensitivity. Their differences will be explained and theory about their hypothetical link with executive tenure will be discussed. The next section will decompose this link into several related constructs. In this section the elements of each construct will be explained

(6)

and described within the context of their interaction. Two hypotheses will be proposed based on theory and rhetoric arguments. Then, the methodology within this study will be discusses in detail and the results will be illustrated. This part is followed by an extensive part about the possible implications of the results, together with the limitations of this study. Afterwards, a conclusion is provided.

2. Literature review

2.1. Foreign ownership and investor influence

From the 1960s on there has been an increase in foreign investment activities and international trade flows (Lensink et al., 2007). Not surprisingly, many scholars have investigated the different influences of foreign ownership on firms, especially corporate governance. There are many different definitions of corporate governance in existing literature, but the definition of Shleifer and Vishny (1997) conforms most to ownership. They refer to corporate governance as “the ways in which suppliers of finance to corporations assure themselves of getting a return on their investment” (p. 737). Within the current state of globalization, a popular field of study is the change of corporate governance induced by foreign investment and several studies have found evidence for a central role in stimulating change in corporate governance systems internationally (Gillan & Starke, 2003). In most research addressing cross-border differences in corporate governance, the separation between shareholder-orientated and stakeholder-orientated systems is made. The former type of corporate governance is known as the Anglo-American model and is most prevalent in the U.S. and the U.K.. These markets are characterized by a large market for corporate control and dispersed ownership and tend to emphasize shareholder benefits. In contrast, the Continental European model, mainly common in Europe and Asia, is characterized by large blockholders and long-term debt markets and generally has a more stakeholder-oriented nature (Aguilera & Jackson, 2003). These two models have served

(7)

as the foundation for the majority of studies regarding corporate governance. Within this framework, a recent study of Desender et al. (2014) tested whether and to what extent foreign shareholder-oriented owners shape corporate governance in stakeholder-oriented organizations by introducing practices typically implemented within shareholder-oriented organizations. Their findings support this prediction of changing corporate governance practices and they state that future research should explore the heterogeneity in objectives of foreign investors in order to fully understand this change (Desender et al., 2014).

According to Dahlquist and Robertsson (2001), most features associated with foreign ownership are driven by the fact that foreign owners typically are mutual funds or institutional investors. The importance of these institutional investors in the global capital market has tripled since the 1990s. Institutional investors account for almost 80 percent of the daily U.S. stock exchange transactions and control almost nearly half the of the Unites States equity market (Ferreira & Matos, 2008). Furthermore, according to the International Monetary Fund (2005), institutional investors manage over US$20 trillion in equities, of which more than US$2 trillion is invested by U.S.-based institutions in overseas non-U.S. stocks (Ferreira & Matos, 2008). As these numbers indicate, there seems to be a significant urge for institutional investors to invest abroad. However, many studies have shown that investors are usually “home biased” (Coval & Moskowitz, 1999; Dahlquist & Robertsson, 2001; Kang & Stulz, 1997). This signifies that as investors face barriers in selecting and investing in shares of foreign organizations, they are likely to prefer investments within their home countries rather than engage in foreign ownership. According to Dahlquist and Robertsson (2001), these barriers consist of explicit barriers and implicit barriers. Explicit barriers are directly observable obstacles such as foreign exchange control and withholding taxes. However, as these barriers have significantly been reduced over the past twenty years in developed markets, they can therefore no longer explain the observed portfolio allocations of investors nowadays. On the other hand, implicit barriers include information asymmetries, and in order to overcome these barriers, the influence by

(8)

investors on the management of a firm is considered to be a vital instrument to do so (Dahlquist & Robertsson, 2001).

According to theory, the two prevalent mechanisms by which this influence is exerted is by “voice” and “exit”. The former can be conceptualized as direct intervention within the practices of a firm. This can be done by exercising shareholder votes during official or private meetings. Meanwhile, exit is defined as selling shares and therewith signaling a stance towards certain managerial practices (McCahery et al., 2016). This external control mechanism is increasing in importance in parallel with the rise of institutional ownership, as institutions’ direct influence can be fairly powerful because of their large stake within organizations. For instance, institutional investors can act collectively in deciding to sell shares or to purposely not invest within a particular company, which can have major consequences (Gillan & Starks, 2003). Simultaneously, expressing investor influence by voice is considered to increase with institutional ownership, and thus foreign investment. Like exit, this is due to their relatively sizable share in the capital structure of organizations, as this provides them with more voting power and accordingly investor influence. This significant investor influence might be the reason why Desender et al. (2014) refer to an “institutional investor bias” when analyzing foreign ownership, as it seems that explicitly institutional investors have the abilities to overcome the implicit barriers of foreign investment; directly through their voting power and indirectly by trading large quantities of equity (Dahlquist & Robertsson, 2001; Gillan & Starks, 2003).

2.2. Shareholder activism and top executive tenure

This exploitation of ownership in relation to the ability to control top management touches on the agency theory. The agency theory is the most dominant theory when addressing corporate governance and is concerned with finding the most efficient contract between the principal and the agent. With the shareholder being the principal and the top management team as the agent, this contract tries to solve two issues: to motivate the top management team to pursue the interest of the shareholders and to share the risk between both parties (Eisenhardt, 1989). The risk-sharing element of the agency

(9)

theory is based on the assumption that managers are more likely to be risk averse than shareholders. This is due to the opportunity for shareholders to diversify their investment portfolio and thereby spread their financial risk, whereas managers do not have the possibility to diffuse uncertainties regarding consequences of their decisions because they can only fulfill one job position at a time. In order to motivate the agent to pursue the interests of the principal and to share the risk between the two parties, the principal could link the incentives for top management to a behavior-based contract. In this case, input is used as a measure for performance. Alternatively, an outcome-based contract can be used to align the interests of shareholders and top management, which is based on the final results delivered by management (Eisenhardt, 1989).

These two issues related to the agency problem are direct consequences of the separation of ownership and control, as this provides top management with the ability to pursue their own interests rather than those of the shareholders (Gillan & Starks, 2003). By monitoring, shareholders are able to discipline managers of poor performing organizations and reduce this managerial opportunism. Differences exist between the two main models of corporate governance in how this monitoring process is implemented. As mentioned, the shareholder-oriented Anglo-American model of corporate governance is characterized by dispersed ownership. Therefore, it is argued that the incentives for shareholders to monitor corporate management are reduced, because costs and benefits associated with the monitoring process are not equally spread among all shareholders and thereby creates opportunities of free-riding. A consequence is a lack of involvement of shareholders within the firm makes, which makes it more complex to correctly assess managerial performance. As a result, the return on corporate stock is often used as an indicator to asses this company performance, as this method provides a quick and relatively effortless evaluation (Lazonick & O’Sullivan, 2000; Bezemer, 2010). In contrast, many authors argue that within the Continental European model, large blockholders and institutional owners do have significant incentives to carefully monitor and control executives and thereby reduce free-rider possibilities (Shleifer & Vishny, 1986; Admati et al., 1993; Huddart, 1993; Maug, 1998; Noe, 2002). Because of this shareholder activism, institutional investors

(10)

gain specific organizational knowledge and potentially overcome the implicit barrier of information asymmetry, which results in careful assessment of managerial performance and interference when necessary. These incentives for institutional investors to interfere with managerial decision-making might be a reason why institutional investors have been particularly successful in changing corporate governance within firms (Dalton et al., 2003).

Chi and Wang (2009) argue that an important measure of this corporate governance effectiveness can be derived from executive tenure. Tenure of top management has been under severe scrutiny lately and many of this research focused on changes of CEOs, company chairmen and other directors with respect to corporate governance. The outcomes of these studies have had practical implications in forming recommendations and guidelines for monitoring and replacing top management (Firth et al., 2006). The overarching focus of these studies is the interference by shareholders in the form of disciplining executives of poor performing companies by dismissal. Several authors have proven that the relationship between performance and tenure is affected by ownership concentration (Volpin, 2002). Firth et al. (2006) state that empirical studies from the U.S. and Europe generally confirm the link between poor corporate performance and executive tenure, but that the rate of dismissal depends on the ownership of the firm. Nevertheless, there is still little consensus in this field of study, as other scholars find contrasting results and partially offset the relationship between investor influence and executive tenure (Kaplan, 1994; Franks & Mayer, 2001). These diverging findings may result from a crucial aspect within the alignment of incentives between shareholders and managers. To a certain degree, the output-based contract of return on corporate stock as an indicator of managerial performance does not only neglect the input of managers, but also fails to assess performance over the long run (Miles, 1993). This shareholder myopia is also referred to as “short-termism” and is considered to have a negative impact on firm performance, as scholars assume that there is an optimal trade-off between the long-run and the short-term, referred to as the concept of intertemporal choice (Jackson & Petraki, 2011; Laverty, 1996; Lazonick & O’Sullivan, 2000). Basically, this concept justifies short-termism as decisions and

(11)

outcomes that pursue results that are best for the short-term but suboptimal over the long run (Laverty, 1996). In context of the agency theory, this myopic perspective on performance is passed down to top executives and by the use of effective mechanisms of corporate governance aligned with their actions. In this context, many theorists argue that managers tend to pursue short-term results and often neglect firm performance over the long run by the impact of the market of corporate control (Jackson & Petraki, 2011). By definition, this shrinkage of time horizons is accompanied by a growing element of uncertainty. According to agency theory, this rise in uncertainty has to be accounted for when compensation managers, as managers need to be offered higher outcome-based rewards in order to accept to bear the increasing risk. Examples of this are bonuses and stock options and statistics support this reasoning, as we see a significant rise in the use of these models of compensation not only in the U.S., but also in Europe (Sanders & Tuschke, 2007). However, potential misalignment of incentives, or simply failure of top management to achieve these short-term profits could result in disciplinary actions in the form of dismissal, initiated by shareholders.

As mentioned, most firms within the U.S. and the U.K. monitor executives based on this principle of short-term profits. Common sense would suggest that this focus would result in higher uncertainty and therewith a decrease in executive tenure. However, Kaplan (1994) finds no evidence for similar sensitivity of tenure to stock performance between U.S. firms and Japanese firms. A possible explanation could be a neutralizing effect of short-termism and lacking investor influence, as both effects exist alongside in the same market in the U.S.. More recently, in the context of globalization and foreign ownership, the current discussion about comparative corporate governance has included the influence of various global pressures. For that reason, it is argued that the traditional Anglo-Saxon and Continental European models of corporate governance as proposed earlier are no longer taken for granted (Aguilera & Jackson, 2010). Many studies have shown that corporate governance practices, especially those affiliated with shareholder-oriented ideologies, are exported from the U.S. to other countries and are likely to be recombined and interpreted with local practices before adopted. This may result in particular practices of corporate governance only partly

(12)

implemented and regularly leads to new or hybrid versions of these practices (Aguilera & Jackson, 2003, 2010). With the rising presence of large U.S. institutional investors in Europe, it is therefore interesting to explore this phenomenon in the context of executive tenure. Therefore, this study focusses on the effect of investor influence on executive tenure. However, with ownership concentration and as the most prevalent variable explaining investor influence and therefore managerial tenure, empirical evidence regarding the effectiveness of institutional investors’ monitoring abilities is to some extent divergent (Gillan & Starks, 2003). This implies the need for a deeper understanding of ownership to fully understand the impact of investor influence on executive tenure.

2.3. Owner types and objective functions

The diverging results of monitoring abilities of institutional investors are argued to be the consequence of a differing monitoring strength among various types of institutional investors (Bushee, 1998). Several studies have investigated the monitoring function being affected by the relationship between shareholders and directors. Results show that shareholders may engage in a business relationships that overshadows the role of objective monitoring integrated in the investor relationship (Brickley et al., 1988; Daily et al., 2003; Dalton et al., 2003; Gillan & Starks, 2003). This implies that shareholders may have differing objectives when investing. Jensen (2001) argues that the existence of multiple objectives makes it impossible for a shareholder to maximize more than one of these objectives at a time. Therefore, shareholders emphasize different objectives more than others and integrate these different levels of importance into a single objective function (Chang & Wong, 2009). In context of the agency theory and investor influence, the dissimilar objective functions associated with these divergent relationships must be accounted for. Chang and Wong (2009) explore the effect of various ownership objectives on the relationship between managerial turnover and financial performance. They do this in de context of state-ownership in China, but state that these finding are relevant for publicly listed firms too, as this market is also characterized by the presence

(13)

of multiple ownership objective functions. They suggest that exploring the trade-offs between these various objective functions and their impact on the performance-turnover relationship would provide greater insights in how managers in publicly listed firms are monitored (Chang & Wong, 2009). Complementary, Chi and Wang (2009) find that various subtypes of private ownership have differently affected the sensitivity of executive turnover to performance in China. Therefore, it is interesting to explore the potential interacting effects of objective functions on executive turnover in a European setting.

According to Lee and O’Neill (2003), the strength and impact of these combined objective functions are comprised in ownership structures. A firm’s ownership structure represents the all shareholders in that organization and therefore integrates all possible objective functions. Given the various ownership structures of firms globally, it is assumable that these different objective functions did not arise in isolation (Lee & O’Neill, 2003). Therefore, this paper explores the relationship between ownership structures and executive turnover in the context of internationalization. With growing foreign ownership, several studies have shown that corporate governance practices from the U.S. are transferred to other countries (Djelic, 1998; Fiss & Zajac, 2004). Again, this market is characterized by dispersed ownership and therefore is assumed to have a more short-term oriented nature. In combination with foreign ownership in Europe often characterized by large U.S.-based institutional owners and thereby by shareholder activism, this influence may be combined with the interest to align shareholders’ short-term incentives with those of top executives. As stated above, the impact of institutional investor influence over governance practices depends on the objective functions of these shareholders, which varies among owner types. Accordingly, institutional ownership structures conceptualize ownership concentration and the interaction of various objective functions simultaneously. Therefore, this study explores the relationship between institutional ownership structures and executive tenure in Europe. With investor influence in the context of corporate governance often analyzed within the framework of the agency theory, there is little consensus in existing literature regarding the relationship between ownership and corporate

(14)

governance practices. The agency theory provides a general starting point for analyzing ownership affecting corporate governance processes as it is widely applicable in the context of corporate governance. Nevertheless, Gillan and Starks (2003) state that “the agency problem as envisioned by Berle and Means (1932) and Roe (1990) might not be prevalent in economies where ownership structures differ” (p.14). As the agency theory solely focuses on the bilateral contract between shareholders and management, the theory is argued to often overlook important interdependencies among shareholders (Aguilera & Jackson, 2003). To capture this interaction, this study examines the effect of ownership structures affecting executive turnover in an empirical study. In particular, the moderating effect of owner types on ownership concentration affecting executive turnover is examined. Accordingly, this study tries to answer the following research question:

What is the effect of institutional ownership structures on executive tenure in Europe?

3. Theoretical framework

3.1. Firm performance and tenure

When analyzing the impact of ownership structures on executive tenure within the context of the agency theory, managerial incentives need to be illustrated in order to understand the potential impact of institutional investors. The encompassing assumption of managerial opportunism is commonly used in existing literature as a framework to clarify these incentives. Managerial opportunism is defined as the tendency for managers to solely act within self-interest with guile (Williamson, 1975, 1985). This means that when managers are not monitored, they tend to shirk because they prefer leisure to work. This assumption is exposed to much criticism, initially by Donaldson (1990). Donaldson (1990) proposes stewardship theory as a response to the agency theory which assumes that the interests of managers and shareholders could be already aligned beforehand. Despite this

(15)

valid argument, it is crucial to mention that employees can always make honest mistakes. This assumption regarding cognition is incorporated within the agency theory as bounded rationality and therefore this study takes monitoring and control of shareholders for granted in order to achieve optimal performance and overcome managerial opportunism.

According to agency theory, top management is responsible for overall firm performance (Firth et al., 2006). Performance is linked to incentives, in order to motivate managers to do their job. This way, firm performance is related to managerial opportunism, as achieving good results is now in the interest of managers. To achieve these results, management enhances their competences. Vafeas (2003) argues that this increase in commitment is linked to long-term director engagement. This organizational commitment provides managers with crucial firm-specific knowledge and fosters the alignment of interests between shareholders and managers, as managers expand their effort to achieve company goals (Buchanan, 1974). Therefore, it can be stated that organizational commitment is a direct consequence of long-term director engagement and is accompanied with experience. An extreme interpretation of this relationship is that dismissal leads to a waste of this experience and competence and an efficient market would lead to long-term survival of executives as they can best protect the interest of the shareholders (Vafeas, 2003; Vance, 1983).

However, this concept about tenure is prone to several counter arguments. Besides the notion that executives can be incompetent to lead the firm, there are also concerns that executives who serve the firm too long may get entrenched (Vafeas, 2003). This entrenchment is an elaboration of managerial opportunism, as it is serves the interests of the executive to have a stable job position. Outcomes of managerial entrenchment are that managers can dominate the board and benefit themselves at the expense of shareholders (Allgood & Farrell, 2000). These detrimental effects need to be contested and top management should be replaced when performance is poor. This disciplining of managers by dismissal is considered to be a supplementary instrument to address the agency problem. It is argued that good corporate governance results in a correct evaluation of performance and subsequently in potential executive dismissal (Firth et al., 2006). So, good corporate governance

(16)

is characterized by embracing the commitment of executives induced by long-term engagement, but simultaneously interfering when performance decreases due to inability or entrenchment of executives. This presumption constitutes a direct link between performance and executive tenure, thence the following hypothesis is proposed:

H1: Executive tenure increases with firm performance.

3.2. Concentration of ownership

The strength of this link between performance and executive tenure depends strongly on the effectiveness of corporate governance. In turn, this effectiveness of corporate governance is directly related to the presence of information asymmetries. These information asymmetries arise from the lack of ability of managers to communicate information regarding the frim and from the unwillingness of shareholders to assemble this information (Lee & O’Neill, 2003). Top executives are considered to be in possession of the complete extent of organizational information, as they have to take decisions concerning the direction of the firm based on this knowledge (Jackson & Petraki, 2011). This gap between information possessed by top management and information obtained by shareholders constitutes the information asymmetry. So, the more information the investors acquire, the smaller this gap. The reduction of this gap is crucial in order to sufficiently understand the activities of top management and therewith improve the accurate evaluation of their performance (Lee & O’Neill. 2003). This evaluation is a determining factor of the effectiveness of corporate governance because it affects the feasibility of investor influence and subsequently overcome possible entrenchment effects. Lack of information possessed by shareholders puts them in a position where they are unable to assess performance and interfere when necessary. This way, entrenched managers are not held accountable for poor performance (Firth et al., 2006).

As mentioned, institutional investors have an incentive to reduce this information gap. This incentive is a direct consequence of their considerable dependence on the performance of the

(17)

affiliated firm because of their large equity share. Therefore, large institutional investors tend to engage in shareholder activism in order to exert investor influence to improve performance. The most prevalent mechanism by which investor influence can be exerted is by voice. McCahery et al. (2016) find that most voice is exercised behind the scenes. Only when private interventions fail shareholders tend to take public measures. An effective shareholder engagement channel is voting against management on shareholder meetings (McCahery et al., 2016). By doing so, shareholders can push their interest to the top of the agenda and put severe pressure on management. In contrast, investors with a smaller stake in an organization might not have this incentive to overcome information asymmetries (Lee & O’Neill, 2003). Costs to obtain information are considered too high compared to the relatively small benefit they gain by monitoring and interfering. So, as ownership disperses, the incentive to exercise investor influence diffuses with it. This diffusion contributes to the free-rider problem, which defines minor shareholders benefitting from the monitoring of other shareholders, without sharing the costs of it. Ultimately, this results in an under-provision of investor influence, as large institutional shareholders may refuse to bear these monitoring costs, since costs and benefits become irrationally out of proportion (Bhide, 1994). This relocates power towards managers and results in “strong managers and weak owners” (Roe, 1994; Walsh & Seward, 1990). This is argued to cause a fragile relationship between performance and tenure, as executives become entrenched due to this failure of corporate governance (Hill & Phan, 1991; Morck et al., 1988). So, a large presence of minor shareholders who refuse to exert voice can result in entrenched management not held accountable for poor firm performance.

A second and more easy way shareholders can instigate managerial performance is by selling shares. By doing so, shareholders put pressure on management by taking a stance towards managerial practices (Gillan & Starks, 2003). However, a fundamental issue emerges with this mechanism of exit. Because of the large equity stakes of institutional owners, they cannot sell large holdings in the company without enduring losses due to a drop in price of stock. Theoretically, this constricts managers from exercising influence through exit (Lee & O’Neil, 2003; McCahery et al., 2016).

(18)

Nevertheless, shareholders can govern without intervening (Admati & Pfleiderer, 2009; Edmans, 2009; Edmans & Manso, 2011). The concept of exit is based on large shareholders owning significant quantities of information about the firm value and thereby have the ability to profit of this information advantage by trading (McCahery et al., 2016). Top management tends to avoid exit by large equity holders if they care about stock prices caused by equity-based pay or potential take-over pressures. Therefore, corporate governance effectiveness can be increased by the threat of exit. This instrument of corporate governance is less present in a diffused ownership structure. In a diffused ownership structure, shareholders do not have the ability to exploit such an information advantage as information asymmetries persist.

Several studies have found evidence for voice and exit to be complements. These studies show that the effectiveness of voice is improved if it is backed up by the threat of exit. This even accounts for situations where managers are not concerned about the short-term stock price (Dasgupta & Piacentino, 2015; Levit, 2014). Moreover, the additional effect of short time horizons embodied by shareholders in relation to shareholder activism remains unclear. Some argue that short-termism fosters shareholders to intervene more often as shareholders pursue short-term profits. Others agree with the argumentation that short-term relationships do not induce shareholders to engage in shareholder activism because the market for corporate control fosters low engagement (McCahery et al, 2016). Because of these diverging effects, this paper assumes both effects to compensate each other.

Thus, the mechanisms of voice and exit are better available to, and more frequently used by large holders of equity in comparison to smaller shareholders. This way, investor influence increases in parallel with ownership concentration. This investor influence is required in order to maintain effective corporate governance. Failure of corporate governance results in a misevaluation of performance and the inability to interfere by shareholders. This interference is necessary in order to overcome management entrenchment and maintain a legitimate relationship between firm performance and executive tenure. Thence the following hypothesis is constituted:

(19)

H2: Executive tenure decreases with the concentration of ownership.

3.3. Owner types

When analyzing the effectiveness of corporate governance, the evaluation of performance is an essential aspect. Chang and Wong (2009) state that different levels of importance to firm performance in shareholders’ objective functions arise with the variation of owner types. In existing literature concerning ownership composition, various objective functions have emerged and divided into those related to the investor relationship and the business relationship. According to this theory, the objectives related to the investor relationship account for all owners of equity, as the original tendency of the investor relationship is to safeguard ownership stakes. This relationship emphasizes firm performance and therefore contributes to the effectiveness of corporate governance. However, institutional investors may also engage in a business relationship with affiliated firm. This is the case when investors depend on the business of that firm (Brickley et al., 1988). In other words, a business relationship emerges if to some extent economic activity arises from the relationship between the firm and the institutional investor (Kochhar & David, 1996). As management is contingent on the actions of shareholders but also vice versa, this relationship creates a mutual dependence. Within the context of the agency theory, the business relationship and investor relationships can cause counteracting pressures on executives, as dependent relationships can endanger the objectivity of the monitoring function (David et al., 1998). To capture these pressures, Brickley et al. (1988) have formulated two opposing categories of institutional investors: pressure-resistant and pressure-sensitive investors.

Pressure-sensitive owners are more prone to managerial influence and therefore may experience more opposing pressures in the context of investor influence (Brickley et al., 1988). As they strongly depend on the actions of top management, these shareholders tend to be more supportive of management than other investors (Gillan & Starks, 2003). Consequently, this may result in an unlikeliness to express investor influence in what is perceived as active monitoring of

(20)

decision-making managers, since an activist stance towards managerial practices through voice or exit may be penalized by the firm by termination of the business relationship (Dalton et al., 2003; Kochhar & David, 1996). Banks, insurance companies, and nonbank trusts tend to fall into this owner category, as they frequently have current or potential business relations with a firm they hold equity positions in (Brickley et al., 1988; Gillan & Starks, 2003). In contrast, pressure-resistant investors have no direct business relationships with firms in which they hold equity and therefore have no conflict of interest. They therefore are assumed to exert influence by the use of voice and emphasize the objectivity function of their monitoring task. Public pension funds and mutual funds are typically assigned to this category (Brickley et al., 1988; Kochhar & David, 1996).

Besides these two main categories, Brickley et al. (1988) introduce a third category labeled pressure-indeterminate institutional investors. This outsider category includes shareholders that do have a business relationship with the affiliated firm, but do not exert any investor influence. This reluctance is mostly present in small investors, as the influence gained from their equity stake does not outweigh the dependence on the affiliated firm (David, 1996). The motivations behind this unwillingness of shareholder activism cannot always be clearly defined, but is mostly present within corporate pension funds, brokerage houses, and investment counseling firms (Brickley et al., 1988; Kochhar & David, 1996). Because this study focusses on investor influence as a consequence of shareholder activism, these owner types are perceived to be included in the category of pressure-sensitive owners.

The power acquired through the possession of large equity stakes of institutional investors may be offset by their dependence on the firm for business activity (Cook, 1997; Levine & White, 1961). The focus of pressure-sensitive owners deviates from firm performance to establishing close relationships with top management and thereby reduces investor influence. With investor influence induced by concentrated ownership mainly driven by incentives to express shareholder activism, pressure-sensitive shareholders tend to engage less in actions to control management. Therefore, this mutual dependence may have a negative impact on the effect of ownership concentration on executive

(21)

tenure. As stated, different objective functions emerge within ownership structures (Lee & O’Neil, 2003). Effective corporate governance can be negatively influenced by ownership structures in which objective functions integrated in the business relationship overshadow those of the investor relationship. Chang and Wong (2009) confirm this assumption by implying that the interactivity of these objective functions may influence the overall stance towards firm performance. This interactivity of objective functions with respect to mutual dependence has to be accounted for when addressing the impact of ownership concentration on executive tenure. A pressure-sensitivity ratio of a particular ownership structure is assumed to represents the strength of the mutual dependence within a particular firm. Subsequently, the following hypothesis is constituted:

H3: Pressure-sensitivity negatively influences the relationship between ownership concentration and executive tenure.

Figure 1 shows a visualization of all tested relationships as previously described within one conceptual model.

Firm performance

Executive tenure H1(+)

H2(-) Figure 1. Conceptual model

Ownership concentration

Pressure-sensitivity H3(-)

(22)

4. Method

4.1. Study design

This study emphasizes on examining an issue by explaining the relationship between variables. Therefore, this study follows an explanatory study design. This study follows a deductive approach wherein quantitative data is used to answer the research question. This means that the data collection occurs in systematic and structured manner (Saunders et al., 2012). This latter is mainly important to ensure the reliability of the research and this section describes all consecutive steps taken whilst carrying out this research in detail in order to secure the quality of the research. The research strategy of this study is that of archival research. In contrast to what the name may imply, archival research can refer to recent documents, which makes it a suitable research strategy for addressing currently interesting topics (Saunders et al., 2012). Secondary archival data are used to conduct the study and is extracted from two various data sources: the Amadeus database of Bureau van Dijk and BoardEx. An important reason for this type of research strategy was the availability of data, as there was only a limited time to conduct this research. In addition, archival data are reliable, free of mitigate biases, and do not depend on response rates (DesJardine & Bansal, 2014). The downside of using secondary data is that it is assembled in order to serve a different purpose than in this study. This involves that some reports may not comprise the exact information necessary in order to answer the research question.

4.2. Sample and time frame

In order to best represent the European market, purposive sampling was used as a method to determine the appropriate sample. The initial sample frame consisted of a list of the 500 largest companies of 2015 based on market capitalization, retrieved from the website of the Financial Times (https://www.ft.com) As the scope of this study encompasses the European market in general, the companies were sampled based on the size, as this characteristic puts organizations in the position

(23)

where they have the most influence on, and are most influenced by the market and are therefore best generalizable to this population.

Banks and insurance companies were excluded, as the Amadeus database did not include these firm categories into their archives. Unfortunately, access was restricted to other databases that included these data. Next to that, two dual listed companies (Unilever and Reed Elsevier) were excluded from the sample, as this would bias the sample. This potential bias would be generated by the existence of two different ownership structures induced by dual listing. The influence of separate ownership structures on a single board composition of executives could not be accounted for sufficiently, as additional effects outside the scope of this study might emerge. After eliminating firms with missing data, the sample concluded 195 companies of which sufficient data were available to construct all variables to conduct the research.

This study follows a cross-sectional study design, meaning that it investigates a certain issue in a particular point in time. This point in time is chosen to be the year 2015, as this is the most recent year of which all of the data to conduct the research are available. As this study explores a recent phenomenon, the most updated data are necessary in order to address the current gap in existing literature.

4.3. Data sources

As mentioned, the data used to conduct this study was provided by archival secondary data sources. The data source to extract data about ownership structures in Europe was Bureau van Dijk’s Amadeus database. This database includes detailed company data of around 21 million European organizations. This information contains detailed corporate structures, company financials and industry related statistics in standard formats which allows the data the be compared across countries (Amadeus, 2016). Most importantly, it holds detailed information about ownership structures, including relative percentages of shares owned by essentially all shareholders within the equity structure of the firm. These percentages are combined with the notation of the type of shareholder. This makes them easily

(24)

subdivided into particular categories with respect to their theoretical influence on management. Additional company data was also extracted from the Amadeus database. This included the financials return on assets and sales, as well as firm industry codes.

The data which contained information about executive tenure was provided by BoardEx. BoardEx is an archival data source that incorporates several statistics concerning corporate governance and boardroom processes. The information concerns individual information of executives of publicly listed companies on a global scope, including their resume information, business networks and occasionally compensation data (BoardEx, 2016). A time-consuming element of this study was the extraction of this information. This needed to be done separately for each individual firm by hand. Naturally, this data was thereafter merged with the data collected from the Amadeus database in order to generate one dataset.

4.4. Dependent variable

4.4.1. Executive tenure

Executive tenure was measured as the mean of number of years within the board of all board members of which data was available. As the data within the BoardEx database is gathered on an individual basis, occasionally information of some board members was missing. The average tenure therefore provided an acceptable measure of the individual amount of years executives served the board. Despite this resolution, missing data of individual board members slightly biased the measure. Important to mention is that the data concerns executives whom are still sitting members of the board.

An important aspect which is taken into account in this study is the board structure of the firm. The current state of corporate governance allows for two separate board structures: a one-tier board structure or a two-tier board structure. In the case of a one-tier board structure, the directors within of the supervisory board are simultaneously the executive directors and are therefore

(25)

aggregated into one board of directors. Within a two-tier board structure these functions are subdivided into two disjointed boards, namely the executive board and supervisory board. As this study focusses on the tenure of executives, data from the executive board was used as the board was structured according to a two-tier board structure. Otherwise, information from the general board of directors was utilized as an assessment of executive tenure.

4.5. Independent variables

4.5.1. Performance

To construct performance affecting the tenure of top executives, return on assets (ROA) is used as a measure for this variable (Allgood & Farrell, 2000). Although the first section of this paper identifies return on stock as an indicator of performance by shareholders, several studies have shown that accounting measures are better predictors of management behavior compared to stock returns (Allgood & Farrell, 2000). As ROA is based on operating profit, it is less susceptible to manipulation by managers. This latter is important as we want to investigate the effect of investor influence on managers by correct evaluation of performance. Therefore, a possible confounding effect of performance manipulation is not desirable.

4.5.2. Ownership structures

An ownership relation is a link between two entities: a shareholder and an affiliated company. The affiliated company is always a corporation, a shareholder could in theory be a corporation, individual, government or a collectively described entity. A link can be derived from either a direct relationship or an indirect relationship. A direct relationship is indicated when a particular entity owns a certain percentage in an organization. Total ownership is indicated when an entity holds a total stake within an organization, but the path through which ownership is held is not specified. If the percentages of the holdings within mediating organizations are known, the weighted averages of stock held are not

(26)

computed. This is because it cannot be asserted with certainty that there are no other mediating links that could potentially influence this calculation. However, when multiple minor shareholders are controlled by one entity, the total percentage of ownership is calculated by making the sum of all direct ownership percentages of the shareholders controlled by that same controlling company. Therefore, the sum of all total ownership percentages of one particular affiliated company can exceed the hundred percent (Amadeus, 2016). When no total ownership percentage could be constituted, direct ownership only served as the percentage of controlling stock of an investor. Otherwise, the total ownership percentage was used to measure the potential influence of owners.

In order to adequately measure the impact of ownership structures, several shareholders with no controlling ownership percentages were deleted from the dataset. Next to ownership concentration, this study investigates the impact of ownership sensitivity. Shareholders that did not comprise to either pressure-sensitive owners or pressure-resistant owners according to theory were deleted from the dataset. Hence, data of unnamed private and public investors, self-ownership, individuals, inside equity and public authorities were removed from the dataset. In addition, shareholders that owned less than three percent of stock were removed from the dataset. The rule of three percent is derived from a paper of Lee and O’Neill (2003), where they argue that by this the influence of large owners are captured. As pension funds commonly own less than five percent of stock, this lower limit had to be considerably below this percentage, but still had to be substantial enough to capture large institutions. Next, only shareholders with voting rights were included to capture the influence of ownership structures. Shareholders that own stock with non-voting rights are not able to engage in shareholder activism and exert this by investor influence and therefore they cannot exert any influence with respect to the research question.

4.5.3. Ownership concentration

Eventually, ownership concentration was measured as the sum of total percentage of stock held by shareholders that owned at least three per cent of an organizations outstanding stock, as performed

(27)

by Lee and O’Neill (2003). By doing so, shareholders with no controlling interest are excluded from this sum and therewith do not contribute to the total percentage of ownership concentration. This means, that the higher the sum of the percentages of shareholders included in the dataset, the higher the ownership concentration. A limitation of this method is that as if the data about a particular ownership structure is incomplete, the sum of the total ownership percentages may decrease in reliability, as there may be additional shareholders with or without controlling interest. As Bureau van Dijk claims that the dataset is as complete as possible, this method is assumed to be reliable (Amadeus, 2016; Saunders et al., 2012).

4.5.4. Ownership sensitivity

This study makes a dichotomous distinction between sensitive shareholders and pressure-resistant shareholders. According to theory, pressure-sensitive shareholders are considered to be bank and insurance companies, as they tend to usually have business relationships with the firms they invest in (Brickley et al., 1988). The overarching category of financial corporations are included into this category as well. In addition, pressure-indeterminate investors are added to this category in order to capture investor influence, as pressure-indeterminate investors do not exert investor influence. Thence, brokerage houses and investment counseling firms were incorporated into this category.

With this subdivision of owner types in two categories, the percentages of these two separate categories were accumulated per firm. So, this resulted in a percentage of pressure-sensitive and pressure-indeterminate owner stakes. In order to capture pressure-sensitivity, the following formula was used:

Pressure-sensitivity = % pressure-sensitive owner stakes / (% pressure-resistant owner stakes + % pressure-sensitive owner stakes)

(28)

4.6. Control variables

The addition of control variables to the analyses has the function to control for potential confounding effects. These effects are expected to have an impact on the dependent and independent variables tested and are therefore controlled for (Saunders et al., 2012). With respect to executive turnover, several firm-level control variables were included. The first is firm size, as it is argued that managers get more entrenched in large firms (Chang & Wong, 2009). To account for firm size, a logarithm of sales is added (Maury, 2006). The second control variable that is accounted for is the age of executives. Theory has shown that the age of executives has a strong relation to executive tenure (Kang & Shivdasani, 1995; Murphy & Zimmerman, 1993; Weisbach, 1988). With respect to the second and third hypothesis, performance is added as a control variable. As argued, performance is assumed to directly affect tenure, so it has to be kept constant when testing those hypotheses to eliminate its possible confounding effect.

4.7. Analyses

The sample data was structured according to various continues variables per firm. Given this sample and research goal, hierarchical linear regression models appeared to be the most appropriate for testing all three hypotheses. Within a hierarchical regression model, predictors are inserted at a specific order, so that potential increments in explained variance and changes within regression coefficients could be ascribed to that additional predictor (Cohen et al., 2013). In order to test the first hypothesis, the following regression was estimated:

Executive tenure = β0 + β1 x Performance + F1 + ɛ

, where β0 is the constant and β1 is the model coefficient of the regression. F1 is the vector of control

(29)

The second hypothesis is almost identical to the first hypothesis, except for the change in predictor and the inclusion of an additional control variable. Therefore, the following regression was estimated to test the second hypothesis:

Executive tenure = β0 + β1 x Ownership concentration + F1 + ɛ

Again, β0 is the constant and β1 is the model coefficient of the regression. The vector of control

variables includes performance as an extra control variable compared to the first regression estimation and, together firm size and firm performance. These controls are incorporated in F1 and ɛ is the error term.

For the third hypothesis, a moderating effect had to be tested for. First, the means of the independent variables were centered. Thereafter, the following regression was estimated:

Executive turnover = β0 + β1 x (Ownership concentration x Pressure-sensitivity) + F1 + ɛ

Again, β0 is the constant and β1 is the model coefficient of the regression. F1 is the vector of control

variables, which includes firm size, age of executives and firm performance, whereas ɛ is the error term.

5. Results

5.1. Descriptive statistics

In table 1 some descriptive statistics are provided. Minimum average tenure within a company of the sample is 0.2 years, which equals 2.4 months and the maximum is 30.5 years. Current executives are on average 7.4 years active within the board. However, a standard deviation of 5.022 indicates tenure

(30)

being relatively large. Revealed by a mean of 0.855, most of the ownership structures are dominated by pressure-sensitive owners. The average age of executives within the board is 55.9 years, with 34 years the youngest and 70 years the oldest. Remark that performance is a natural logarithm of sales and therefore cannot be interpreted by its original basis of unit.

Table 1. Descriptive statistics

Variables N Minimum Maximum Mean SD

1. Tenure 195 0.200 30.500 7.401 5.022 2. Pressure-sensitivity 195 0.000 1.000 0.855 0.209 3. Ownership concentration 195 0.030 2.061 0.481 0.332 4. Performance 195 -13.930 33.190 7.076 6.645 5. Age 195 34.000 70.000 55.898 5.910 6. Size 195 14.954 26.086 22.814 1.540 5.2. Pre-tests

In order to perform regression analyses, there are several assumptions that have to be met. The first, sample size, is assumed to be sufficient according to the rules of thumb by Green (1991). Therefore, there could be tested for the overall model and for the individual predictors within the model. Next, to check whether the data was normally distributed, the kurtosis and skewness statistics were calculated for all independent variables. H0 predicts that the data is normally distributed compared to H1 predicting the data it not normally distributed. The method to test for H0 is by calculating the lower and upper boundaries of the skewness and kurtosis statistics within the 95 percent confidence interval (C.I.), which are provided in table 2 and table 3. For the variables to be normally distributed, this tests for 0 to be within the lower and upper boundaries of the confidence interval. This is a very powerful test and will often reject H0. Therefore, empirical criteria are set at acceptable values between -1 and +1.

(31)

Table 2. Kurtosis test

Variable Lower C.I. Upper C.I.

1. Tenure 2.428 3.786 2. Pressure-sensitivity 3.458 4.816 3. Ownership concentration 2.525 3.883 4. Performance 1.766 3.125 5. Age -0.252 1.106 6. Size 4.087 5.445

Table 3. Skewness test

Variable Lower C.I. Upper C.I.

1. Tenure 1.081 1.763 2. Pressure-sensitivity -2.299 -1.617 3. Ownership concentration 0.971 1.654 4. Performance 0.647 1.329 5. Age -0.565 0.118 6. Size -1.732 -1.050

As shown in table 2, kurtosis was for all variables leptokurtic except for age. Therefore, all variables except age are transformed by using the following formula:

X* = ((X-M)(X-M)2)-0.2

, with M being the median of variable X (Van Casteren, personal communication, n.d.). Afterwards, the kurtosis was within the boundaries of the confidence interval. As the statistics show in table 3,

(32)

several variables showed positive skewness. However, as a consequence of the recalculation of the variables because of the positive kurtosis, the recomputed variables showed evidence for not rejecting H0 in the skewness test, meaning that skewness was solved for simultaneously during the transformation of the variables. Moreover, linearity assumes that the correlation between variables is linear. This assumption is checked for by looking at the Pearson correlation coefficients between the independent variables and the dependent variable, which are shown in table 4. If the correlations were significant, linearity could directly be assumed. The non-significant correlations were checked within scatter plots, which showed no non-linearity patterns. Therefore, it was concluded that the data met the linearity assumption. Non-collinearity, independence and homoscedasticity were also checked with the use of scatter plots, and subsequently was concluded that the data met these assumptions in order to test for regression as well. In addition, the VIF collinearity statistics showed no perfect multicollinearity between all of the variables.

Table 4. Means, standard deviations, correlations

Variables M SD 1 2 3 4 5 1. Tenure .251 2.278 2. Pressure-sensitivity -.099 .315 -.093 3. Ownership concentration .036 .454 .048 .035 4. Performance .289 2.632 .168* .113 -.077 5. Age 55.898 5.910 .488**_-.110 _-.031 _.044 6. Size -.079 1.119 .015 .041 -.125 -.136 .193** n=195; *p<0.05, **p<0.01

Table 4 provides summary statistics and Pearson correlation coefficients between all variables. As age is the only variable that is not transformed, the rest of the values cannot be interpreted by their original basis of unit. The statistics do provide some insights in the individual and

(33)

mutual distribution of the sample. The Pearson correlation coefficients between tenure and performance of 0.168 (p<0.05) indicates an absence of relation. Tenure and age do show a tendency to positive relation because of a correlation coefficient of 0.488 (p<0.01). Also, a negligible link between firm size and executive age is suggested by these statistics, showing a correlation coefficient of 0.193 (p<0.01). There are neither significant coefficients calculated between all variables and pressure-sensitivity, nor between ownership concentration and all variables.

Altogether, the pre-tests indicated that the data met sufficient criteria in order to conduct regression analyses. Despite the notion that the assumption of normality was violated by failing the kurtosis test, all other assumptions were met. As kurtosis is solved for by precomputing most of the variables, the sample data is assumed to be suitable to perform the regression analyses.

5.3. Tests

Hierarchical multiple regression was performed to examine the ability of firm performance to predict levels of executive tenure, after controlling for age and size and is included in model 1. Important statistics of this regression are provided in table 5 and additional statistics are provided in the appendix. In the first step of the hierarchical multiple regression, two predictor variables were inserted: age and size. This model was statistically significant (F=31.112; p<0.001) (see appendix). The value of R² of 0.245 indicates that 24.5% of the variance of executive tenure was explained by the independent variables included in this model. After entry of performance at step 2, the total variance explained by the model as a whole was 26.3% (R²=0.263; p<0.001). The inclusion of performance explained an extra 1.9% of variance in executive tenure, after controlling for age and size, indicated by the value of R² Change of 0.019 (p<0.05). In the final model, two out of three predictor variables were statistically significant, with age documenting a higher beta value (β=0.494; p<0.001) than performance (β=0.138; p<0.05). Furthermore, if age increases for one, executive tenure increases for 0.190. If performance increases for one, executive tenure increases for 0.120. Therefore, hypothesis 1 is supported.

(34)

Table 5. Model 1 summary R R² R² Change B SE β t Step 1 .495 0.245*** Age 0.194 0.025 0.504*** 7.884 Size -0.166 0.130 -0.082 -1.277 Step 2 0.513 0.263*** 0.019* Age 0.190 0.024 0.494*** 7.783 Size -0.124 0.130 -0.061 -0.951 Performance 0.120 0.054 0.138* 2.199 *p<0.05, **p<0.01, ***p<0.001

In order to test the potential effect of ownership concentration on executive tenure, and for a possible moderating effect of pressure-sensitivity, again a hierarchical regression was performed. Some of the most important statistics of this second model are shown in table 6, the rest of the outcomes can be found in the appendix. The first step of the regression included the same actions as within the second step of the first model. After adding concentration at the second step, the total variance explained by the model as a whole was 26.8% (R²=0.268; p<0.001). This model was statistically significant (F=17.384; p<0.001) (see appendix). The addition of concentration explains no additional variance in executive tenure, after controlling for age, size and performance, as the value of R² Change of 0.005 is not significant. In the second step, two out of four predictor variables were statistically significant, with age documenting a beta value (β=0.494; p<0.001) compared to performance (β=0.145; p<0.05). In this model, if age increases for one executive tenure increases for 0.190. Further, if performance increases for one, executive tenure increases for 0.125. The results do not provide any evidence to support hypothesis 2.

(35)

Lastly, the moderating effect was tested for. When the interaction variable was added to the regression, there was no significant change in variance explained by the model (R²=0.268; R² Change-0.000). Also, the moderating variable did not show significant statistics (p>0.05). Thence, the effect of concentration on executive tenure does not depend on pressure-sensitivity. Therefore, hypothesis 3 is not supported.

Table 6. Model 2 summary

R R² R² Change B SE β t Step 1 .513 0.263*** Age 0.190 0.024 0.494*** 7.783 Size -0.124 0.130 -0.061 -0.951 Performance 0.120 0.054 0.138* 2.199 Step 2 0.518 0.268*** .005 Age 0.190 0.024 0.494*** _7.787 Size -0.105 0.131 -0.051 -0.797 Performance 0.125 0.055 0.145* 2.293 Concentration 0.342 0.315 0.068 1.084 Step 3 0.518 0.268 0.000 Age 0.190 0.025 0.493*** _7.743 Size -0.104 0.132 -0.051 -0.791 Performance 0.125 0.055 0.145* 2.285 Concentration 0.333 0.319 0.066 1.043 Pressure-sensitivity* 0.030 0.143 0.013 0.209 Concentration *p<0.05, **p<0.01, ***p<0.001

(36)

6. Discussion

6.1. Interpretations

The findings of the first model show a positive relationship between performance and tenure when controlled for firm size and the age of executives. This relationship is limited, indicated by the R² Change of 0.019, meaning that a modest 1.9% of variability of earnings can be accounted for by adding performance to the model. The findings display a standardized beta value of 0.138 (p<0.05) and can be directly compared to the significant beta value of age (β=0.494; p<0.001). This means that age predicts relatively more of the variance of tenure within the model than performance does. As performance increases for one, tenure increases for 0.120 (B=0.120). Remarkably, firm size does not account for any of the variance of tenure, as its beta coefficient is not significant (p>0.05).

In the second step of the second model, there is a small increase in R², however this is not significant. Therefore, the second hypothesis is not supported, as ownership concentration does not explain some of the variance of tenure when controlled for age, size, and performance. Interestingly, there is an increase in the beta coefficient of performance. This might imply that when controlled for concentration, more variance of the model could be explained by performance. Therefore, the findings indicate that performance and ownership concentration are somehow related with respect to tenure. With the absence of ownership concentration affecting tenure in the second model, the third hypothesis is not supported either. There is no change in explained variation constituted, neither is this change significant (R² Change=0.000; p>0.05). Therefore, it can be stated that this study did not find support for the effect of ownership concentration on executive tenure, let alone a moderating effect of pressure-sensitivity on this relationship.

Ownership concentration and owner types : executive tenure in Europe affected by ownership structures