The role of dynamic industry conditions in foreign director appointment

i

i

i

The role of dynamic industry conditions in foreign director appointment

Master Thesis, MSc Accountancy

University of Groningen, Faculty of Economics and Business

F.J.M. (Femke) Boonen

Student number: S2540894

Supervisor dr. R.C. Trapp

Co-assessor N.J.B. Mangin

Word count 11,959

June 2018

ii

ABSTRACT

The objective of this thesis is to explore whether the external environment, in this study represented by the industry dynamism dimension, is associated with the appointment of foreign directors in US firms. By studying this relationship, I aim to complement prior literature that recognised the importance of the environment on board composition but was deficient in examining the industry dynamism forces empirically. In line with the resource dependence theory, I posit that industry dynamism could be related to appointing foreign directors. Furthermore, I propose that the presence of a nomination committee may strengthen this positive relationship since its members are less likely to be affected by bounded rationality and therefore more likely to consider the dynamic industry environment. By using a logistic regression analysis of 6,990 director appointment announcements of US-headquartered firms from the period 2003-2017, I find that all three industry dynamism proxies are significantly associated with the appointment of foreign directors. However, only two proxies show the expected, positive relationship, while one proxy shows a significant, negative relationship. Furthermore, the moderation does not yield significant results. Nevertheless, these results indicate that overall, industry-level determinants can play a role in board nationality diversity and that industry dynamism in particular can help explain the appointment of foreign directors.

Keywords: Corporate governance; Board diversity; Nationality diversity; Foreign directors;

Board internationalisation; External Environment; Industry dynamism; Nomination Committee

iii

CONTENTS

ABSTRACT ... II

1. INTRODUCTION... 1

2. THEORETICAL FRAMEWORK & HYPOTHESES ... 6

2.1.BOARD COMPOSITION AND NATIONALITY DIVERSITY ... 6

2.2.DIMENSIONS OF THE ENVIRONMENT AND BOARD COMPOSITION ... 6

2.3.HYPOTHESES DEVELOPMENT ... 8

2.3.1. The role of industry dynamism ... 8

2.3.2. The moderating role of a nomination committee ... 10

3. METHODOLOGY ... 12

3.1.SAMPLE AND DATA SELECTION ... 12

3.2.MEASURES ... 14

3.2.1. Dependent variable: foreign director appointment ... 14

3.2.2. Independent variable: industry dynamism ... 14

3.2.3. Moderating variable: nomination committee presence ... 17

3.2.4. Control variables ... 17

3.3.EMPIRICAL METHOD ... 19

4. RESULTS ... 20

4.1.DESCRIPTIVE STATISTICS ... 20

4.2.HYPOTHESIS TESTING ... 21

4.3.ROBUSTNESS CHECKS ... 23

5. CONCLUSION ... 26

5.1.DISCUSSION ... 26

5.2.IMPLICATIONS ... 27

5.2.1. Theoretical implications ... 27

5.2.2. Practical implications ... 28

5.3.LIMITATIONS AND FUTURE RESEARCH DIRECTIONS ... 28

6. REFERENCES ... 31

APPENDICES ... 35

APPENDIX A ... 35

APPENDIX B ... 36

APPENDIX C ... 38

APPENDIX D ... 39

1

1. INTRODUCTION

“With a Dutch chairman, things could have gone different” stated Het Financieele Dagblad

(Couwenbergh & Verbraeken, 2017, p. 4). It represents the conclusion of the meeting between the chairs of the AEX companies, just after the dramatic takeover of ABN Amro by the Dutch government in 2008. They argued that it could have had different outcomes if the nationality of the CEO and chairman of the board both had been Dutch, as the American chairman of the ABN Amro Supervisory board did not have “the phone numbers of Prime Minister Balkenende and the Minister of Finance Bos (…), nor did he know them” (Couwenbergh & Verbraeken, 2017, p. 14).

This highlights that executives are cognizant of the fact that the nationality of directors could play a role in uncertain situations.

As indicated by the example above, nationality diversity in the boardroom has been put high on the corporate governance agenda of executives (Kaczmarek, Kimino, & Pye, 2012).

Although prior literature has already shown the importance of international experience (e.g.

Carpenter & Fredrickson, 2001; Zhang, Kong, & Wu, 2016), empirical studies on nationality diversity are still rare (Nielsen & Nielsen, 2013). Yet, there is a growing awareness in both academic literature and business press, as well as among executives, that in an increasingly globalising world, there is an increasing demand for foreign directors (Kaczmarek et al., 2012;

Miletkov, Poulsen, & Wintoki, 2016). In addition, diversity has been highly promoted by media

, shareholders and regulatory bodies

, which has created an incentive for organisations to increase diversity in order to improve, among others, organisational outcomes and their public image of inclusiveness (Anderson, Reeb, Upadhyay, & Zhao, 2011; Frijns, Dodd, & Cimerova, 2016).

Correspondingly, the profiles of the directors have become more internationalised (Greve, Biemann, & Ruigrok, 2015) and in the past two decades there has been a substantial increase in the appointment of foreign directors, yet with significant differences among European countries (Van Veen & Marsman, 2008).

However, diversity, which is often referred to as the degree to which group members differ with respect to individual characteristics

, has frequently been described as a “double-edged sword” (see, for example, Frijns et al., 2016; García-Meca, García-Sánchez, & Martínez-Ferrero, 2015; Milliken & Martins, 1996). This indicates that on the one hand, diversity encourages valuable divergent opinions and perspectives due to the differences in knowledge, expertise (Estélyi &

Nisar, 2016), preferences and experience (Miletkov et al., 2016). On the other hand, it poses

1 A Dutch financial daily newspaper.

2 To illustrate, Het Financieele Dagblad published an article in November 2017 (Knoop & Toe Laar, 2017) to highlight the image problem of the firms at the Zuidas (the corporate district in Amsterdam) regarding the inclusion of non- nationals. They note that nationality diversity of recruited staff is only slowly increasing. Therefore, they argue that bi-cultural or non-nationals demand role models at the top to be encouraged to apply at these firms.

3 For example, the United Arab Emirates (UAE) requires the chair and the majority of the board of public joint stock companies to be UAE nationals (Choueiri & Baig, 2018). However, most countries do not have legal restrictions.

Yet, numerous corporate governance codes (e.g. in Germany and France) recommend that boards should consider the principle of diversity, which includes nationality (Seibt & Kulenkamp, 2017; Djehane, 2017).

4 Individual characteristics include for example demographic factors (e.g. age, tenure, gender and nationality) and functional factors (e.g. experience and education) (Cannella et al., 2008).

2

challenges such as coordination and communication, which could result in a higher likelihood of misunderstandings and within-group conflict (Anderson et al., 2011; De Wit, Greer, & Jehn, 2012;

Francoeur, Labelle, & Sinclair-Desgagné, 2008). Its disadvantages are also reckoned in the example of ABN Amro. Given these facts, previous literature argues that this dual aspect of diversity is not simply positive or negative, but rather depends on the type of diversity (Yi, Ndofor, He, & Wei, 2017) and the context in which the team operates (Cannella, Park, & Lee, 2008).

Despite the aforementioned growing awareness and emphasis placed on foreign directors, prior board composition and diversity literature has prioritised other dimensions of diversity, such as gender (e.g. Hutchinson, Mack, & Plastow, 2015; Triana, Miller, & Trzebiatowski, 2014) and functional diversity

(e.g. Buyl, Boone, Hendriks, & Matthyssens, 2011; Cannella, Park, & Lee, 2008). Only a few studies have investigated nationality diversity, of which most of them are investigating the outcomes of appointing foreign directors, such as firms’ financial performance (Frijns et al., 2016; Peck-Ling, Nai-Chiek, & Chee-Seong, 2016), firm risk (Hutchinson et al., 2015) and cross-border acquisition performance (Masulis, Wang, & Xie, 2012). Yet, Nielsen &

Nielsen (2013) argue that first the sources of diversity should be understood before the costs and benefits can be analysed. Given that research on the sources of nationality diversity is scarce, Greve et al. (2015) point out that “the antecedents of foreign director appointment remain a blind spot that limits our understanding [of the outcomes and potential costs and benefits]” (p. 675). In addition, the few studies that have investigated the antecedents of foreign director appointment, have particularly focused on individual-level antecedents such as demographic characteristics and capabilities (Greve et al., 2015) or firm-level determinants such as firm age and firm size (Kaczmarek et al., 2012; Miletkov et al., 2016). Yet, due to the increasing complexity and uncertainty in the organisation’s environment, academic literature and practice both recognise that firms cannot solely rely on the individual capabilities such as experience and education anymore, but should also include external factors in the board composition process (Cannella et al., 2008).

This implies that besides individual-level determinants, the external environment could also play a role in appointing directors. However, as of yet, empirical research on the relationship between the external environment and board nationality diversity is scarce.

Accordingly, this study aims to fill this research gap by examining the organisation’s external environment as a determinant of nationality diversity. A firm’s environment has been largely defined in residual terms and therefore encompasses a large range of factors (Downey, Hellriegel, & Slocum, 1975). Consequently, researchers such as Aldrich (1979) and Dess & Beard (1984) have tried to conceptualise the external environment by introducing dimensions of it and thereby enabling its operationalisation. One of the environmental dimensions is industry dynamism. This dimension refers to the instability of an industry (Boyd, 1990) and indicates the unpredictable change within the industry in which the organisation operates (Dess & Beard, 1984).

Dynamic industries are characterised by a high level of uncertainty (often referred to as volatility)

5 Functional diversity refers to the differences in functional background among directors and is based on the functional area in which the board member has previously gained experience and developed specialised knowledge (Cannella et al., 2008).

3

and therefore requires organisations to adopt new strategies and tactics in order to cope with their changing environment and to minimise uncertainty (Pfeffer, 1972).

This adaptation to the environment can be explained by the resource dependence theory.

This theory posits that organisations try to minimise their dependence on contingencies in the external environment by securing resources through, for example, their board of directors (Hillman, Withers, & Collins, 2009; Pfeffer & Salancik, 1978). From this perspective, diversity can provide the firm with important resources, as it offers a variety of experiences and a broader knowledge base (García-Meca et al., 2015), which enables firms to better cope with its external environment and to diminish uncertainty (Pfeffer, 1972). Following this theory, Boyd (1990) argues that board composition tends to reflect the external environment of firms since “the need for resources and information is a function of environmental demands” (p. 420). Given that particularly firms operating in a dynamic industry environment are subject to high levels of uncertainty, industry dynamism can be an important driver of board diversity.

Building on the aforementioned growing interest of various groups in the nationality of board members, this study will particularly focus on nationality diversity. In addition, firms are operating in an increasingly globalising world where foreign interactions are increasing (Ghemawat, 2011) and an international managerial labour market is emerging (Van Veen &

Marsman, 2008). Simultaneously, globalisation, as well as technological innovations and deregulations have substantially increased the environmental uncertainty (i.e. industry dynamism) faced by organisations (Hoque, 2004). Therefore, there is a potential increasing interest in both foreign directors and industry dynamism from organisations.

Hence, the first objective of this study is to provide empirical evidence of industry dynamism forces shaping board composition, with the particular focus on foreign director appointment. Based on the resource dependence theory, organisations try to acquire resources to minimise the uncertainty in dynamic industries (Pfeffer & Salancik, 1978). According to Pfeffer (1972, p. 226), “board size and composition are not random or independent factors, but are, rather rational organisational responses to the conditions of the external environment”. One possible action could be to attract diverse, resource-rich directors who are likely to introduce new perspectives (Hillman et al., 2009). Hence, I postulate that appointing foreign directors can be related to a dynamic industry environment.

The second objective of this thesis is to evaluate the moderating role of a nomination

committee in the selection of foreign directors. Following the abovementioned statement of Pfeffer

(1972), board composition and therefore the appointment of foreign directors is likely to be a

rational response. However, according to Ruigrok, Peck, Tacheva, Greve, & Hu (2006), the

nomination of directors by current corporate directors is subject to bounded rationality, which

implies that they are likely to select candidates who resemble themselves and possess

characteristics they know well. Therefore, they argue that nomination decisions by current

corporate directors could be not entirely rationally motivated. In contrast, nomination committees

have established structured procedures regarding the selection criteria, involvement of recruitment

firms and interviewing of nominees that contribute to rational decision making (Ruigrok et al.,

2006). As a result, they are more likely to match the resources and characteristics of the directors

4

with the needs of the firm during the nomination process, whereas current directors are less likely to (Hillman et al., 2009; Pfeffer, 1972; Ruigrok et al., 2006). Given that a firm operating in a dynamic industry environment is more likely to demand an increase in board diversity as an attempt to minimise the uncertainty, and that a nomination committee is more likely to take this into consideration, a nomination committee potentially increases the probability to appoint foreign directors in a dynamic industry environment. In other words, when organisations have a nomination committee, I expect that this will strengthen the positive relationship between dynamic industry environments and foreign director appointment.

My analysis is based on 6,990 director appointment announcements of firms headquartered in the United States (US) from the period 2003-2017. To measure the association between industry dynamism and foreign director appointment, I use three proxies that capture the volatility aspect of the industry dynamism construct. Using a logistic regression model, my findings suggest that all three proxies are significantly related to the appointment of foreign directors. However, only two proxies show the expected, positive relationship, while one proxy shows a significant, negative relationship. Furthermore, the moderation effect of a nomination committee does not yield significant results. Nevertheless, these results indicate that overall, industry-level determinants can play a role in board nationality diversity and that industry dynamism in particular can help explain the appointment of foreign directors.

The theoretical contribution of my research is threefold. First of all, it contributes to the current corporate governance, diversity and board composition literature by creating a better understanding of the relationship between dynamic industry conditions and the appointment of foreign directors. As can be derived from extant literature, industry-level characteristics as determinants of foreign director appointment have remained largely unexplored. Therefore, in an attempt to bridge this gap in literature, my results provide empirical evidence on the relationship between industry dynamism and the appointment of foreign directors. This will contribute to our knowledge about foreign director appointment, as first the sources of diversity should be understood before the cost and benefits can be understood (Nielsen & Nielsen, 2013). As a result, it puts earlier research investigating the outcomes of diversity in a context and eventually allows us to better understand how foreign directors can be related to organisational outcomes and their roles in organisations (Miletkov et al., 2016).

Secondly, I have added the moderator nomination committee to my model to examine its presence on the relationship between industry dynamism and the appointment of foreign directors.

Nomination committees have often been overlooked in academic papers (Ruigrok et al., 2006) and have been particularly related to gender diversity (Hutchinson et al., 2015). This study offers new insights into the outcomes of the inferred rational decision-making process of nomination committees and the effect of their presence on the appointment of foreign directors.

Thirdly, by employing foreign directors as a dependent variable, my study responds to calls

from Top Management Team (TMT) literature to treat individual director characteristics as

dependent variables rather than independent variables in order “to open up new avenues for

thinking about organisational adaptation (…) and to sharpen our predictions of how and why

executives’ characteristics become manifested in organisational outcomes” (Hambrick, 2007, p.

5

338). In addition, by studying its relationship with industry dynamism, it responds to the specific quest of Carpenter, Geletkancz, & Sanders (2004) and Yamak et al. (2014) for examining the effect of dynamic industry conditions as antecedents on TMT composition. They find the lack of research on this surprising and describe it as a “critical omission” (Yamak et al., 2014, p. 77) that it has not been investigated before.

Moreover, I contribute to practice as I shed light on the growing debate on how organisations choose their board of directors. Studying industry dynamism as a determinant of foreign director appointment will provide us with new insights into how international directors advance into the boardroom, which can be useful for individual foreign directors aiming for a director position. In addition, this study will also provide a better understanding of under what conditions a firm is more likely to appoint foreign directors. This can be useful for institutions and regulatory bodies aiming to promote or develop governance guidelines regarding diversity within boards, as it will enhance their awareness of the antecedents of nationality diversity. Furthermore, this study will also investigate the inferred rational decision making of nomination committees and will enhance the organisations’ knowledge about the nomination committee’s role. Lastly, it may provide valuable information for the people selecting the board, as this study shows the external environment may be taken into consideration during the nomination process.

The remainder of the thesis is organised as follows. In Section 2, the theoretical framework

is developed from which the two research hypotheses are derived. Next, in Section 3, the research

methodology is introduced, which includes a description of the sample, the variables and the

empirical procedures to test the two hypotheses. Thereafter, in Section 4, the results of the analyses

are reported. Finally, in Section 5, the thesis will be concluded by discussing the results, elaborating

on the theoretical and practical implications, discussing the limitations and addressing potential

areas for future research.

6

2. THEORETICAL FRAMEWORK & HYPOTHESES

2.1. Board composition and nationality diversity

Appointing foreign directors and thereby increasing diversity within a board, can have both positive and negative implications. On the one hand, diversity is particularly associated with the introduction of different perspectives and various alternatives, which stimulates debate, reduces group-thinking (Cannella et al., 2008; Carpenter et al., 2004) and fosters creativity and innovation (Katz, 1982). In addition, firms can benefit as foreign directors bring specific knowledge and expertise that domestic directors could be lacking (Greve, Nielsen, & Ruigrok, 2009). As a consequence, it may improve the quality and comprehensiveness of decisions (Nielsen & Nielsen, 2013). On the other hand, the appointment of foreign directors could also have costs, as communication problems could arise due to language and cultural differences or because of the lack of physical proximity (Miletkov et al., 2016; Nielsen & Nielsen, 2013). In addition, heterogeneous boards are associated with slower decision making (García-Meca et al., 2015) and a higher likelihood of conflicts (Hillman, Shropshire, & Cannella, 2007). Also, they are more likely to demand a higher compensation compared to their domestic equivalents (Miletkov et al., 2016).

Researchers have not yet reached a consensus about whether the positive outcomes are outweighing the negative. For example, board nationality diversity has been both negatively (Frijns et al., 2016; Peck-Ling et al., 2016) and positively associated with performance (Estélyi & Nisar, 2016). Nielsen and Nielsen (2013) recognise this tension and argue that its relationship is not unitary (positive or negative), but rather varies with different contexts. This is supported by Miletkov et al. (2016), who argue that the costs and benefits differ not only for each director, but also vary across firms, countries and industries. This implies that the antecedents of nationality diversity can be analysed at different levels. For example, while research at an individual-level looks at demographic, professional and social characteristics such as gender, age, academic and functional background (Greve et al., 2015; Ortiz-de-Mandojana & Aragon-Correa, 2015), the organisational-level includes factors such as a firm’s growth rate, international sales, international shareholder base and size (Miletkov et al., 2016). Also the external environment (i.e. country- and industry-level factors) is seen as an important factor in the composition of boards (Boyd, 1990).

For example, Miletkov et al., (2016) examined the country-level antecedents and found that a smaller, less educated populace and underdeveloped capital markets increase the demand of firms to attract foreign directors. According to Cannella et al. (2008), the external environment is gradually receiving more attention from both academics and practitioners since it is recognised that relying solely on individual characteristics could be suboptimal. Therefore, the external environment should be considered in the nomination process and could play a role in the appointment of foreign directors as well.

2.2. Dimensions of the environment and board composition

The environment is an important contingency factor that can affect the organisation as it can pose constraints as well as opportunities (Ghosh, Willinger, & Ghosh, 2009; Hrebiniak &

Joyce, 1985). Yet, the environment has been largely defined in residual terms (Downey et al.,

7

1975). For example, Dill (1962) mentions that environments relate to all elements that do not formally belong to the organisation. Subsequently, he argues that introducing environmental constructs or dimensions allows researchers to better capture the essence of the environment’s effect on the organisation. Following this view, a frequently used approach in defining the environment is based on the typology developed by Aldrich (1979), which has been translated into three environmental dimensions by Dess & Beard (1984). These include munificence (capacity), complexity (homogeneity-heterogeneity) and dynamism (stability-instability).

Environmental munificence refers to the capacity of the environment that allows firms to experience sustained growth through the availability of resources in the environment (Aldrich, 1979; Boyd, 1990; Tushman & Anderson, 1986). In munificent environments, organisations can generate slack resources, which offers the organisation greater flexibility and more time (Nielsen

& Nielsen, 2013). This leaves room for greater diversity in opinions and perspectives as the positive consequences of diversity are likely to outweigh the negative (Finkelstein, Hambrick, & Cannella, 2009). In contrast, Wiersema & Bantel (1993) highlight that a lack of munificence leads to conditions of scarcity, which results in organisational responses such as formalising, centralisation and decreased information processing. Subsequently, they argue that there is “less decision-making discretion for management, more rigid problem solving and adherence to traditional routines” (p.

487), which leaves less room for diversity.

Complexity represents the second dimension and refers to the range of environmental factors and the variety of inputs and outputs with which the organisation should cope (Ghosh et al., 2009). As a result, companies have to interact with several constituencies, which requires higher information processing, more experience, different attitudes and perspectives (Wiersema & Bantel, 1993). Consistent with this rationale, Anderson et al. (2011) found a positive relationship between complexity and heterogeneous boards.

The last construct is industry dynamism, which refers to the degree to which the external environment components are either stable over time or subject to change, also referred to as variability or uncertainty (Ghosh et al., 2009). This dimension particularly looks at the rate of unpredictable change within the industry in which the organisation operates (Dess & Beard, 1984) and therefore shows the extent to which the industry environment can be accurately predicted or anticipated (Pfeffer & Salancik, 1978).

The abovementioned external environmental dimensions have been reviewed in the paper

of Yamak et al. (2014). They identified three kinds of models, which postulate a relationship

between TMT composition, the external environment and its outcomes such as performance,

innovation and strategic change. These comprise direct effects models (the external environment

directly shapes the composition), mediation models (composition mediates the relationship

between the environment and outcomes) and moderation models (the external environment

moderates the relationship between composition and outcomes). The authors conclude that while

the association between the environment and board composition has received increasingly more

attention in academic literature, our understanding of how environmental forces shape board

composition is still limited. This is based on the fact that previous literature has particularly adopted

moderation models and therefore was deficient in establishing a direct link between the

8

environment and board composition. For example, Nielsen & Nielsen (2013) found a positive relationship between TMT nationality diversity and corporate performance, which was positively moderated by munificent environments. This is built on the notion that time pressure does not play a dominant role in munificent environments, which gives organisations more time to overthink decisions and therefore allows slower decision-making. Given this theoretical rationale, Nielsen &

Nielsen (2013) found that munificent environments will be better suitable for foreign directors and offer the organisation the opportunity to benefit from diversity, as the consequences of diverse boards such as innovation, creativity and slower decision making are more likely to flourish in munificent environments. Moreover, also Keck (1997) adopted a moderation model and found that industry dynamism strengthened the relationship between board tenure diversity and firm performance, while it weakened the relationship between functional diversity and firm performance. This indicates that the effect the of the environment depends on the specific diversity dimension (Yamak et al., 2014) and supports the choice to focus on one specific diversity dimension (i.e. nationality) in this study.

However, Yamak et al. (2014) argue that the aforementioned “critical omission” (p. 77) of direct effect models of industry dynamism and board composition is remarkable. In particular, since the degree of environmental uncertainty has been increasing substantially over the past decades due to technological innovations, deregulation and globalisation (Hoque, 2004). Given that industry dynamism might call for foreign directors, this dimension might be insightful as well and creates an interesting research path. As can be seen in Figure 1, I will follow the recommendation of Yamak et al. (2014) and adopt a direct effect model by studying the relationship between the environment, in this study represented by the dimension of industry dynamism, and foreign director appointment.

2.3. Hypotheses development 2.3.1. The role of industry dynamism

Industry dynamism refers to the (in)stability of an industry and can affect the organisation.

Stable environments allow organisations to be managed by established routines by using formalised structures and enforcing the current system and core values (Aldrich, 1979). In contrast, unstable environments demand more cognitive processing of board members as they should change their routines and pay more attention to environmental scanning and problem solving (Wiersema

& Bantel, 1993). In addition, the unpredictability puts high pressure on the organisation and requires the organisation to constantly adapt to the current situation (Wiersema & Bantel, 1993).

As a result, organisations try to establish a buffer, which allows them to cope with the environment and absorb uncertainty.

An explanation can be found in the resource dependence theory. This theory, originally

introduced by Pfeffer & Salancik (1978), posits that the organisation is dependent on contingencies

in the external environment. This implies that external factors can affect the organisation and

increase environmental uncertainty. Hillman et al. (2009) identified the actions organisations can

take to minimise this environmental uncertainty; besides mergers, joint ventures, other inter-

organisational relationships and political actions, boards of directors could also reduce

9

environmental uncertainty. Following this perspective, Boyd (1990) states that corporate boards are “an integral component of the effective firm” (p. 419). He argues that the board is a valuable mechanism to gain access to other resources, information and knowledge which allows firms to subsequently reduce environmental uncertainty. Consequently, boards of directors have been described as a “primary method for absorbing critical elements of environmental uncertainty into the firm” (Hillman, 2005, p. 465). As a result, board composition tends to reflect the external environment of firms (Boyd, 1990). If the current board composition does not meet the requirements of the environmental context, a change in composition is often viable in order to align the board of directors with the firm’s environment (Pfeffer, 1972). Based on this rationale, dynamic industry environments are likely to “impose a greater need for TMTs to increase the quantity and range of information required to make decisions and the number of perspectives necessary in order to deal with new and changing situations” (Yamak et al., 2014, p. 77).

This can be achieved by increasing diversity, as heterogeneous groups are characterised by enhanced adaptability and are better able to respond to a dynamic environment (Katz, 1982;

Murray, 1989). Given that information available in the external environment is filtered through perceptions (Lawrence & Lorsch, 1967), one can argue that foreign directors cope differently with the external environment than domestic directors due to their different backgrounds, knowledge and experiences which influences their perceptions. Despite the greater likelihood of conflict due to different opinions and communication problems, resolving the conflict could offer new insights and allows the group to better disentangle when adaptation is necessary (Murray, 1989). In addition, they allow boards to be more cognizant of the implications of their decisions (Cannella et al., 2008). This is especially beneficial in unstable environments since diversity increases the likelihood for successful adaptation to the external environment and coping with industry dynamism (Schwenk & Clifford, 1996). In addition, Cannella et al. (2008) posit that the benefits of diversity are more likely to outweigh the costs in these unstable environments compared to stable environments. More specifically, firms can benefit from foreign directors as they respond to the need to increase the quantity and range of information. As a result, firms are more likely to seek a foreign instead of a domestic director. Therefore, I argue that the decision to appoint a foreign director could be associated with dynamic industry conditions. Consequently, it can be hypothesised that:

Hypothesis 1: There will be a positive relationship between a dynamic industry

environment and the appointment of a foreign director.

10 2.3.2. The moderating role of a nomination committee

The composition of boards is an important but highly complex process, especially when this is done by current corporate directors (Ruigrok et al., 2006). A possible explanation for the complexity of the process can be found in several theories from both psychology as well as economics and management. First of all, following from the behavioural theory of bounded rationality, the nomination process is affected by cognitive limitations of current directors (Simon, 1991). More specifically, the decision is impaired by limited knowledge of relevant and important nomination criteria, personal values, preferences, previous decisions and by not being able to consider and evaluate all alternatives (Ruigrok et al., 2006). As a result, directors are likely to suggest individuals with similar backgrounds and characteristics (Hutchinson et al., 2015; Ruigrok et al., 2006) as they are more likely to be sympathetic with their ideas (Westphal & Zajac, 1995).

Secondly, based on the similarity-attraction paradigm (Byrne, 1971), boards display in-group bias which implies that they prefer new board members who are demographically similar (Hutchinson et al., 2015). As a result, minorities are often ignored and “a small group of traditional talent [will be] perpetuating the status quo” (Groysberg & Bell, 2012, p. 2). This provides less room for diversity and stimulates homogeneity. Thirdly, from an agency theory perspective, directors are motivated by self-interest and are therefore more likely to nominate new directors that fit their preferences, rather than make a decision which is in the best interest of the shareholders (Hutchinson et al., 2015). Given the arguments above, nominations made by other current board members may not be entirely rationally motivated.

Therefore, nomination committees have repeatedly been recommended in corporate governance codes and empirical evidence has recognised its value (Kaczmarek et al., 2012). This separate committee is responsible for the director appointment process and makes recommendations with regard to directors (Kaczmarek et al., 2012). More specifically, a nomination committee is able to particularly focus on the establishment of selection procedures, selection criteria, interviewing procedures and director profiles, but also recognises when to involve the services of recruitment firms (Hutchinson et al., 2015; Ruigrok et al., 2006). These factors contribute to delivering an optimal selection process (Hutchinson et al., 2015). As a result, a nomination committee will augment the independence of the process (Hutchinson et al., 2015)

+

DIRECTORS H1

+

NOMINATION COMMITTEE H2

INDUSTRY DYNAMISM

Figure 1- Conceptual Model

11

and reduce the aforementioned limitations of current directors (i.e. bounded rationality problem, similarity-attraction and agency problem). Eventually, this could contribute to a more rational nomination process.

Based on this theoretical rationale, Kaczmarek et al. (2012) posit that a nomination committee is “the first and most important antecedent of diversity in the boardroom” (p. 475). They argue that nomination committee members are less likely to nominate directors that are similar to themselves or in their own interests, but rather make nominations based on the fit between the candidate and the needs of the firm. Firms in dynamic industry environments are more likely to demand a more diverse board, which increases the likelihood to appoint a foreign director (Hypothesis 1). As nomination committees are more likely to consider this need, I expect that the strength of the relationship between industry dynamism and foreign director appointment may vary depending on the presence of a nomination committee. More specifically, industry dynamism may have a stronger positive association with a foreign director appointment when a firm has installed a nomination committee to arrange the director nominations (as shown in Figure 1). Therefore, it can be hypothesised that:

Hypothesis 2: A nomination committee will strengthen the positive relationship between a

dynamic industry environment and foreign director appointment.

12

3. METHODOLOGY

3.1. Sample and data selection

For my empirical analysis, I primarily rely on two datasets. The first dataset was retrieved from BoardEx, a global corporate governance database which provides director profile data. The second was used as a complementary dataset and was obtained from Compustat, a global database of financial and market information. I used the BoardEx dataset for director related announcements from North American firms as initial sample and supplemented this data with financial and industry information from Compustat. Table 1 shows the sample selection process.

Table 1

Sample selection.

My initial sample consists of 78,837 announcements from 8,739 North American firms reported in the BoardEx database for the period 2003-2017. Availability of the data in BoardEx determined the choice of time period for this study. As I only consider new appointments and do not focus on changes within the board such as changes in function, I only include the announcements that proclaim that the director “will join this board”. Hence, from the initial sample, 55,166 announcements were dropped as they did not pronounce a new appointment but rather stated a promotion or departure of a director. Furthermore, as this BoardEx dataset is merged with financial information from Compustat, it is important to match the firm data retrieved from both databases correctly. Therefore, I excluded 2,716 announcements of which a company identifier such as Ticker or ISIN was missing. Moreover, as this study particularly focuses on foreign directors, nationality is an important element. Hence, I omitted 11,449 announcements without information on the director’s nationality. Next, if the director’s age was not available in BoardEx, I hand-collected this data from the Bloomberg website. Furthermore, I removed 776 announcements because of missing network or experience data.

Next, I matched the announcement data with the companies’ financial and industry data from Compustat based on the companies’ ISINs

, or, if unable to match, I matched the data based

6 In order to match data from BoardEx with Compustat, the ISINs were converted to CUSIP.

No. of announcements All announcements for North American firms available in BoardEx for the period 2003-2017 78,837

Less: Other announcements than "will join the board" (55,166)

Less: Missing ISIN or Ticker in BoardEx (2,716)

Less: Missing director’s nationality (11,449)

Less: Missing director’s network and experience (776)

Less: Unable to match with Compustat (793)

Less: Incorrect matches Ticker (290)

Less: Less ISO country code (headquarter) not “USA” (393)

Less: Missing company (financial) information (264)

Final Sample 6,990

13

on their Tickers. However, due to the limitations of conversion to CUSIP and missing company identifier data in Compustat, I was unable to match 1,028 announcements. Therefore, I attempted to match these remaining announcements based on the companies’ legal names. Since this was successful for 235 announcements, I removed the 793 announcements from the sample that I was unable to match. Subsequently, I checked whether the databases were correctly matched by comparing the company legal names for all observations. It appeared that 290 observations that were matched based on their Ticker were incorrect. Therefore, I eliminated these announcements.

Furthermore, as I define the domestic country as the US (see paragraph 3.2.1), I removed 393 announcements from companies that do not meet this criterion (i.e. having its headquarters in the US based on the ISO country code for Headquarters). Lastly, I omitted 264 announcements because of missing company (financial) information. Eventually, these criteria resulted in a sample of 6,990 announcements for 5,469 directors of 2,825 companies headquartered in the US.

Table 2 reports the distribution of the sample by year. This table shows that the years 2006- 2011 represent 71.95% of the total sample, which indicates there were relatively more new director appointments conducted in these years

. Nevertheless, this trend is in accordance with the initial sample (Appendix A), which implies that my selected sample largely reflects the distribution by year of the initial sample. The distribution of the sample by Standard Industry Classification (SIC) code can be found in Appendix B.

Table 2

Sample description.

7 I verified whether this is consistent with the initial sample (after the removal of announcements other than “will join this board”) and ensured whether this was not only caused by the elimination of specific announcements during the sample selection (Table 1). Based on this comparison (Appendix A), I conclude that my sample is slightly more concentrated at the beginning of the sample period and less at the end of the sample period, which is largely explained by the unavailability of data in Compustat from the more recent years. Still both samples show a trend of relatively more announcements in the period 2006-2011.

Sample distribution by year.

Year Percentage Observations

2003 0.16% 11

2004 3.15% 220

2005 8.31% 581

2006 11.00% 769

2007 14.23% 995

2008 14.08% 984

2009 11.03% 771

2010 10.06% 703

2011 11.55% 807

2012 5.48% 383

2013 3.39% 237

2014 2.75% 192

2015 3.26% 228

2016 1.50% 105

2017 0.06% 4

Total 100.00% 6,990

14 3.2. Measures

3.2.1. Dependent variable: foreign director appointment

In order to circumscribe the dependent variable Foreign Director Appointment (FDA), the terms “director” and “foreign” have to be specified. Regarding the term director, I will focus on the entire board and include both executive and non-executive director appointments. This is based on the rationale that US boards are characterised by one-tier boards and are therefore often seen as a single corporate body (Liem, 2018). Furthermore, given that the members of the board of directors share a collective responsibility for the organisational outcomes and that their combined personalities, knowledge and experience influence the decision-making process, the board should be considered as one group (Wiersema & Bantel, 1993).

Defining the concept “foreign” requires two demarcations. First, in order to determine whether someone is foreign, one should define its reference point- “domestic”. This is often determined in corporate governance literature (e.g. Greve et al., 2015) by the country in which the firm’s headquarter is located, as this often indicates the main country of operation. Therefore, I follow this approach and determine domestic by the country in which the firm is headquartered

. Secondly, following the research of Frijns et al. (2016) and Greve et al. (2015), “foreign” is determined in terms of nationality. Nationality is considered as a stable factor and has a strong impact on personality, identity, cognition (Triandis & Suh, 2002), but also on attitudes and behaviours (Hofstede, 1991). Therefore, differences in nationality create various perspectives.

Since this notion forms one key assumption in this study to appoint foreign directors, I have chosen to define foreign in terms of nationality

.

Given these considerations, a newly appointed director is classified as a foreigner if the nationality of the appointed director is different from the country in which the firm is headquartered. The dependent variable FDA is therefore coded as 1 if the new director is a foreigner (i.e. a non-American appointed at an American headquartered firm), and 0 otherwise. The appointment announcements and nationality of directors were retrieved from BoardEx, the headquarter country location was identified by the ISO Country Code (Headquarters) and obtained from Compustat.

3.2.2. Independent variable: industry dynamism

For the independent variable industry dynamism, I follow the approaches used by Tosi et al. (1973) and Hmieleski & Ensley (2007) to operationalise this construct. While Tosi et al. (1973)

8 Another approach is to determine “foreign” based on the country of incorporation. However, the country of incorporation is often purely chosen based on legal regime considerations (Bebchuk & Cohen, 2003). Therefore, I have chosen for the headquarter location, as this seems to be more related to the main country of operation.

9 “Foreign” can also be determined in terms of residency (see, for example, Masulis et al., 2012; Miletkov et al., 2016).

Following this approach, someone is classified as foreign when he or she has a foreign residence address. This approach, however, has several limitations. First of all, there is a risk that the person through years of study, business, or travelling could has become more acculturated to the domestic environment (Miletkov et al., 2016). As a result, this diminishes the differences between domestic and foreign individuals. Secondly, individuals could also have a domestic residence address, but have only recently moved to the country (Miletkov et al., 2016). As a result, they are considered as “domestic”, while they have many characteristics of “foreigner”. Therefore, I considered this as a less stable indicator for foreignness. Hence, I determine foreignness in terms of nationality.

15

have introduced an approach to calculate industry dynamism, Hmieleski & Ensley (2007) further extend this method and identify three proxies to measure industry dynamism.

According to Tosi et al. (1973), firms that are operating in a less dynamic industry environment are experiencing less variance in their operations, expenditures and revenues, and therefore have more stable patterns. This implies that they have, for example, relatively stable revenue patterns and are less subject to technological change, which diminishes uncertainty and therefore indicates a less dynamic industry environment. Tosi et al. (1973) note that stability is often referred to as volatility when measuring risk. Therefore, they argue that these measures of volatility such as the coefficient of variation can be used to operationalise the construct of industry dynamism. In line with this approach, this study uses the coefficient of variation to measure volatility and to capture the degree of industry dynamism consecutively. Hence, the coefficient of variation is first calculated for each firm and for each proxy variable identified by Hmieleski &

Ensley (2007) mentioned below (i.e. Employees, Revenue, R&D), using the following Equations (1-3):

𝑀𝑒𝑎𝑛 = 𝑥̅𝑗= ∑^𝑛_𝑖=0𝑥_𝑖𝑗

𝑛 (1)

𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 = 𝜎𝑗= √∑^𝑛_𝑖=0(𝑥𝑖𝑗− 𝑥̅𝑗)²

𝑛 − 1 (2)

𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑜𝑓 𝑉𝑎𝑟𝑖𝑎𝑡𝑖𝑜𝑛 𝐹𝑖𝑟𝑚 = 𝐶𝑉_𝑗=𝜎_𝑗

𝑥̅_𝑗 (3)

where j represents the firm, x the observed values for the item measured as described listed below (i.e. Employees, R&D and Revenue) for firm j and n the number of observed values for firm j for the period 2003-2017

.

Next, in order to calculate the volatility for each specific industry (i.e. the degree of industry dynamism), I took the average of the coefficients of variation of all firms comprising the specific industry, denoted by Equation (4):

𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑜𝑓 𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡𝑠 𝑜𝑓 𝑉𝑎𝑟𝑖𝑎𝑡𝑖𝑜𝑛 (𝑑𝑒𝑔𝑟𝑒𝑒 𝑜𝑓 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐷𝑦𝑛𝑎𝑚𝑖𝑠𝑚) =∑𝑗∈𝑘𝐶𝑉𝑗

|𝑘| (4)

where j represents the firm, k the two-digit SIC code and CV the coefficient of variation calculated by Equation (3). Compustat is used to retrieve the primary four-digit SIC code, after which I recoded these to a two-digit SIC code, which represents the major group and identifies the companies’ major area of activity (Compustat, 2000)). Eventually, the industry average of the coefficients of variation constitutes the three proxy variables used in this study, where lower values denote a lower volatility, hence, a lower degree of industry dynamism. In addition, using the coefficient of variation allows me to compare industries of different sizes (Tosi et al., 1973), which

10 In principle, n = 15 years (period 2003-2017), however, if the firm is incorporated or ceased its operations during this period, n will be smaller.

16

is important since my study compares dynamic industry conditions among different industries. The degree of industry dynamism (Equation (4)) was computed for the following three proxies identified by Hmieleski & Ensley (2007) to operationalise this construct:

Number of industry employees (CV_INDEMP). The volatility in the number of industry employees represents my first measure of industry dynamism. For this variable, I obtained the item Employees (EMP) from Compustat, which represents the number of company workers as reported to shareholders and includes all employees of consolidated subsidiaries (both domestic and foreign), part-time and seasonal employees, full-time equivalent employees and officers (Compustat, 2000). EMP constitutes the variable x in Equation (1) and (2), after which the final proxy variable CV_INDEMP is computed by Equation (4).

Industry research and development intensity (CV_INDRDI). The second proxy for industry dynamism is the volatility in industry R&D intensity. The variable x in Equation (1) and (2) was measured as the ratio of the sum of R&D expenditures and capital expenditures to total assets

(Christensen, 2016). In order to establish this variable, I obtained the following items for each company from Compustat: Research and Development Expense (XRD), Capital Expenditures (CAPX) and Assets – Total (AT). XRD represents all expenses that relate to the development of new products or services, such as company-sponsored R&D, purchased R&D and software development (Compustat, 2000). CAPX relate to the funds that are used for additions to property, plant and equipment, and AT represents the total assets of a company at one point in time (Compustat, 2000). All three items are measured in millions of dollars. Following Christensen (2016), XRD and CAPX were set equal to zero when missing.

Industry revenue (CV_INDREV). The last measure of industry dynamism focuses on revenues. For this variable, the earnings before interest, taxes, depreciation and amortization (EBITDA) for each firm was obtained from Compustat and used in Equations (1) and (2) to eventually calculate this proxy for industry dynamism (Equation (4)). In accordance with Hmieleski & Ensley (2007), I used EBITDA rather than net earnings or reported earnings per share, as they highlight that EBITDA minimises the differences in depreciation and changes in financial leverage between firms.

It is important to note that I recognise the limitations of measuring volatility to capture the degree of industry dynamism. First, volatility may also indicate “the ability to cope with uncertainty, rather than the lack of uncertainty” (Tosi et al., 1973, p. 31). Second, Dess & Beard (1982) highlight that companies are often able to predict change at a stable rate which could therefore diminish uncertainty. However, Tosi et al. (1973) argue that uncertainty and volatility should be highly correlated, which implies that this measure can still be adequately used to measure the degree of industry dynamism.

11 Or, in equation form: 𝑅𝑒𝑠𝑒𝑎𝑟𝑐ℎ & 𝐷𝑒𝑣𝑒𝑙𝑜𝑝𝑚𝑒𝑛𝑡 𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦𝑖𝑗=^{𝑋𝑅𝐷𝑖𝑗+𝐶𝐴𝑃𝑋𝑖𝑗}

𝐴𝑇_𝑖𝑗 , where i is the firm-year observation of firm j, XRD the research and development expense, CAPX capital expenditures and AT Assets-Total.

17

3.2.3. Moderating variable: nomination committee presence

The moderating variable Nomination Committee presence (NC) is a dichotomous variable and is coded as 1 if the firm has a specific committee committed to the nomination of new directors, and 0 otherwise. I identified all firms with a nomination committee based on the NOMINATION COMMITTEE MEMBER item as presented by BoardEx. This item shows whether the director is a member of the nomination committee of the specific firm. Based on this item, I identified the firms with a nomination committee. Subsequently, I merged this data with my final sample of announcements based on the COMPANY ID item.

3.2.4. Control variables

In this study, I introduce several variables to control for individual- and firm-level characteristics shown by prior studies to be associated with the appointment of foreign directors.

First, as I highlighted in the previous sections, the literature recognises a shift from appointment particularly based on individual-level factors, towards appointment by taking the environment into consideration. This suggests that individual-level variables, such as demographic characteristics, experience and networks often play an important role in the selection process (Cannella et al., 2008). Therefore, I control for these factors by using a dichotomous variable for gender (GENDER) (female = 1, male = 0), measured by the item GENDER from BoardEx. Based on the findings of Greve et al. (2015), males are more likely to be appointed as foreign director. Therefore, I expect a negative relationship between GENDER and FDA. In addition, a continuous variable is used for the director’s age at the day of the announcement (DIRECTORAGE), measured as the difference between the year of the announcement (ANNOUNCEMENT DATE) and year of birth (DATE OF BIRTH) as presented by BoardEx. Based on the findings of Greve et al. (2015), this variable is expected to be positively associated with foreign director appointment. They argue that older candidates are associated with more generic experience and less socialisation failure in new teams, which potentially offsets the disadvantages related to foreign directors Moreover, the years of experience on other boards (EXPERIENCE) is also retrieved from BoardEx and measured by the EXPERIENCE ON OTHER QUOTED BOARDS-item. This variable is expected to be positively related to the likelihood of appointing a foreign director since firms “will particularly seek experienced candidates (…) with skills that can be transferred across contexts” (Greve et al., 2015, p. 683). Furthermore, based on the resource dependence theory, resource-rich directors (i.e.

directors with a large network) are more likely to be appointed (Hillman et al., 2009). Hence, I also control for the network size (NETWORKSIZE) of the individual directors. This item INDIVIDUAL NETWORK SIZE is also obtained from BoardEx, which is measured as the number of overlaps with other individual directors through employment, other activities and education (BoardEx, 2016).

Secondly, I also control for firm-level variables. First of all, firm age (FIRMAGE) is

included, since adherence to deeply embedded routines is more valued in longer-established firms,

which discourages firms to increase board diversity (Kaczmarek et al., 2012). Therefore, this

control variable is expected to be negatively related to FDA. FIRMAGE is measured as the

difference between the year of the announcement (ANNOUNCEMENT DATE) and the year of

18

incorporation (INCORPORATION YEAR), retrieved from BoardEx and Compustat, respectively.

Additionally, I will include firm size (FIRMSIZE) as Miletkov et al. (2016) argue that large firms are more likely to accept the costs associated with foreign directors. In addition, they argue that foreign directors presumably prefer to serve the boards of larger and more visible firms due to its higher perceived prestige. Following Cannella et al. (2008), FIRMSIZE is measured as the natural log of total assets. The total assets item (AT) is retrieved from Compustat.

Moreover, also firm performance will be used as a control variable since it is open to criticisms of reverse causality and can both precede or follow diversity (Hutchinson et al., 2015).

To illustrate, Kaczmarek et al. (2012) argue that firm performance has a positive effect on openness to creativity, innovation and global perspectives and therefore simultaneously on the appointment of diverse directors (e.g. females and non-nationals). However, Estélyi & Nisar (2016) found a reverse relationship and note that nationality diversity positively influences firm performance. In line with corporate governance literature (see, for example, Carter, D’Souza, Simkins, & Simpson, 2010; Hutchinson et al., 2015), Tobin’s Q is often used to capture firm value as an indicator for its performance and is measured as the firm’s market-to-book asset ratio (TOBINSQ). For this variable, I follow Bebchuk, Cohen, & Ferrell (2009) and Christensen (2016) and calculate Tobin’s Q as the market value of assets divided by the book value of assets (AT), where the market value of assets is calculated as the market value of common stock (Price Close Fiscal Year (PRCC_F) * Common Shares Outstanding at year end (CSHO)), plus the book value of debt, which is computed as total assets (AT) less the sum of book value of common stock (CEQ) and balance sheet deferred taxes (TXDB). When CEQ or TXDB were missing, I set them equal to 0. In my study, I used the natural log transformation of Tobin’s Q, as Hirsch & Seaks (1993) found this is closer to the optimal form compared to linear forms. All corresponding items (PRCC_F, CSHO, AT, CEQ, TXDB) were retrieved from Compustat. In summary, this TOBINSQ is computed by the following Equation (5):

𝑇𝑜𝑏𝑖𝑛^′𝑠 𝑄 = ln (𝑀𝑎𝑟𝑘𝑒𝑡 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠

𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠 ) = ln (𝑃𝑅𝐶𝐶_𝐹 ∗ 𝐶𝑆𝐻𝑂 + (𝐴𝑇 − 𝐶𝐸𝑄 − 𝑇𝑋𝐷𝐵)

𝐴𝑇 ) (5)

Also, return on assets (ROA) is often utilised as an alternative accounting measure to

measure firm performance (Ntim, 2015). ROA particularly relates to the firm’s financial

performance and is calculated by dividing the item Income Before Extraordinary Items (IB) by

Total Assets (AT), which were both retrieved from Compustat (Christensen, 2016). Previous

literature has shown that the sign of the relationship between the two separate performance

measures (Tobin’s Q and ROA) and board diversity has not always been similar, nor consistent

(Hutchinson et al., 2015). A potential explanation is offered by Buyl et al. (2011), who argue that

the contradicting evidence can be a result of the ambiguous implications of diversity. Hence, I will

control for both and do not expect a specific sign. An overview of all variables used in this study

can be found in Appendix C.

19 3.3. Empirical method

In this study, the dependent variable Foreign Director Appointment (FDA) is a binary variable and either takes the value 1 or 0 (1 = foreign, 0 = domestic). The independent continuous variables (CV_INDEMP, CV_INDRDI and CV_INDREV) represent the potential determinants of foreign director appointment. Therefore, I aim to model which determinants cause FDA to take on the values 1 or 0 by testing Hypothesis 1 and 2, using a logistic regression model.

The predicted value of the dependent variable FDA is modelled as a function of the independent variables and denotes the probability of assigning a foreigner, contingent on the predictors. The three predictors used in this study represent proxies that capture the same underlying construct industry dynamism. However, Lubotsky & Wittenberg (2006) highlight that using multiple proxies in one regression could result in attenuation bias in the coefficient estimate and result in insignificant individual coefficients. While researchers often include a single summary measure of the proxies in their model, Lubotsky & Wittenberg (2006) propose a different method.

This method recommends to first enter the proxies separately in the regression and only enter them simultaneously in the last model, which is based on the rationale that a simple average of the proxies assumes that all measures are equally good. Since researchers often lack this knowledge, they argue that establishing an optimal summary measure is often impossible.

Given these considerations, to test whether the industry dynamism proxies can be seen as relevant predictors of foreign director appointment (H1), I estimate the following models. Model 1 is the base model and only includes the control variables. Subsequently, in Models 2-4, I enter the proxies of industry dynamism separately (CV_INDEMP, CV_INDRDI and CV_INDREV, respectively), where after I enter the three predictors simultaneously in Model 5. Each model contains a Y-intercept (𝛼), regression coefficient(s) of the predictor variable(s) reflecting industry dynamism (𝛽

), regression coefficients of the control variables (𝛽

) and eight control variables (ROA, TOBINSQ, FIRMAGE, FIRMSIZE, GENDER, DIRECTORAGE, EXPERIENCE, NETWORKSIZE). Moreover, to test the moderation effect of a nomination committee (H2), I augmented the Models 2-5 by introducing the variable NC and the interaction terms (NC

CV_INDEMP; NC

CV_INDRDI; NC

CV_INDREV) in the Models 6-9. For Hypothesis 2, I take

the same approach and enter the interaction terms separately in the Models 6-8, but simultaneously

in Model 9. A random error is not included in the models, since the dependent variable is binary

and P(FDA=1) is the probability of appointing a foreign director, given the values of the predictor

variable(s). As a result, this removes the error term (Ryan, 2009). The logistic regression equations

for the Models can be found in Appendix D.

20

4. RESULTS

4.1. Descriptive statistics

Table 3 presents two panels of descriptive statistics for my full sample (n = 6,990) of director appointment announcements made in the period 2001-2017. Panel A provides the means, standard deviations (S.D.) and the correlations among all variables used in the analyses. Panel B reports the results of the difference of means tests between firms with foreign director appointments relative to firms without foreign director appointments.

Panel A of Table 3 shows that the mean of foreign director appointment is .098, indicating that the appointment of domestic directors was dominating in the period 2003-2017. Moreover, the means of the three industry dynamism proxies show that the average of CV_INDRDI (mean = 0.686) is relatively higher than CV_INDEMP (mean = 0.285) and CV_INDREV (mean = 0.420).

This implies that firms face, on average, the most volatility in R&D intensity. Furthermore, the mean of GENDER is 0.147, which demonstrates that the majority of the newly appointed directors in my sample is male. In addition, the appointed directors are on average 57.139 years old, have 2.348 years of experience on other quoted boards and have an average network of 1760.47 (i.e.

overlaps with other individual directors).

Furthermore, a Pearson correlation was run to determine the correlations for the continuous variables and a point-biserial correlation was run to determine the correlations for the dichotomous variables

. The results are shown in Table 3 Panel A and do not show multicollinearity concerns.

Significant correlations exist among all three independent variables; CV_INDEMP is negatively related to CV_INDRDI (r = -.13, p < .001) and significantly positively related to CV_INDREV (r

=

.64, p < .001). CV_INDRDI and CV_INDREV are also significantly negatively correlated (r = - .13, p < .001).

Panel B of Table 3 presents the difference of means test for my independent variables between firms with and without foreign director appointments. To analyse whether the average of CV_INDEMP, CV_INDRDI and CV_INDREV of firms appointing foreign directors is different from firms that do not appoint foreign directors, I performed an independent samples t-test

. The test was significant for all three measures (CV_INDEMP t

= -2,265, p < 0.05; CV_INDRDI t

= 2.547, p < 0.05; CV_INDREV t

= -3.183, p < 0.01). However, the test provides mixed findings for the association between the different proxies of industry dynamism. The results indicate that firms subject to higher degrees of CV_INDEMP and CV_INDREV are more likely to appoint more foreign directors, compared to firms operating in a less dynamic industry environment in terms of employees and revenues. This is consistent with the notion that industry dynamism potentially increases the demand for diverse perspectives, which could be obtained through the appointment of foreign directors. Yet, the results also indicate that firms experiencing

12 The continuous variables are: CV_INDEMP, CV_INDRDI, CV_INDREV, ROA, TOBINSQ, FIRMAGE, FIRMSIZE, DIRECTORAGE, EXPERIENCE and NETWORKSIZE and the dichotomous variables are: FDA, GENDER and NC.

13 The test showed a significant Levene’s test for equality of variances for all three industry dynamism measures (CV_INDEMP (p = 0.01), CV_INDRDI (p = 0.001), and CV_INDREV (p = 0.001)), hence equal variances were not assumed.