Nationalities of board members in European MNCs:
The effect of psychic distance and other country differences
June 2010
University of Groningen
Evelien Aangeenbrug
Student n◦: 1451626
Email: e.aangeenbrug@student.rug.nl
Master‟s thesis International Business & Management
Supervisor: dr. K. Van Veen
Co-assessor: dr. P. Rao Sahib
University of Groningen
Faculty of Economics and Business
Nettelbosje 2
Abstract
This study empirically investigates the variety of nationalities found in boards of MNCs headquartered in fifteen European countries over the years 2005 through 2007. The nationalities of over 5.5 thousand board members per year in over 350 companies are investigated to answer the following question: “Which factors determine the nationalities of board members in European MNCs?” I argue that it is country differences and
characteristics that determine the nationalities of a firm‟s board members, and that these factors influence both the attitude of companies in their hiring decisions and the attitude of possible board members in their career decisions.
In extensive ordinary least squares and fractional logit analyses, the effects of psychic distance stimuli and other country differences and characteristics on flows of board members between countries were investigated. The results show that psychic distance plays an important role in determining representation of foreign nationalities in boards. Geographic and cultural distance between two countries have a negative influence, whereas colonial ties, a shared language and a shared border increase the percentage of board members from the one country in the other. Further, I found that the higher the expected income in the MNC‟s home country is compared to the expected income in country X, the greater the chance of finding board members with nationality X in this MNC‟s board. Next to that, the results show that a better quality of management schools in country X compared to the quality in country Y increases the flow of board members from X to Y. The internationalization level of the MNC‟s home country was found to have a significant positive effect on the presence of any foreign nationality in cross-section analyses, although panel data results did not confirm this.
Finally, this study validates the theoretical relevance of the fractional logit model. It shows that in studies where the dependent variable is a fraction or percentage, OLS analysis seems unsuitable, and fractional logit models should be used.
Keywords: TMT nationality diversity, European board diversity, Psychic distance,
Table of contents
1. INTRODUCTION ... 6
2. LITERATURE REVIEW ... 9
2.1 Theories underlying board diversity studies ... 9
2.1.1 The upper echelons theory ... 9
2.1.2 Conflict theory ... 10
2.2 Nationality diversity... 11
2.2.1 Firm internationalization related to the composition of the board ... 11
2.2.2 Advantages of TMT internationalization ... 13
2.3 Why is it that diversity in boards remains low?... 13
2.3.1 Causes of low diversity ... 13
2.3.2 Barriers to the development of an international executive labor market ... 14
2.4 Board internationalization research ... 14
2.4.1 The importance of board internationalization research ... 14
2.4.2 Previous studies on board internationalization ... 15
2.4.3 Foreignness versus heterogeneity ... 16
2.4.4 Why nationality matters ... 17
2.5 Problem statement, research question, and relevance of the study ... 18
2.5.1 Problem statement and research question ... 18
2.5.2 Relevance of the study ... 19
3. THEORETICAL FRAMEWORK ... 20
3.1 Internationalization theory and the behavioral theory of the firm ... 20
3.2 Labor migration literature ... 21
3.3 Relevance for board internationalization ... 22
4. HYPOTHESES ... 24
4.1 Psychic Distance ... 24
4.2 Hypotheses development ... 24
4.2.1 Hypothesis 1: Cultural distance ... 25
4.2.3: Hypothesis 3: Common language ... 28
4.2.4 Hypothesis 4: Institutional Distance ... 28
4.2.5 Hypothesis 5: Geographic distance ... 30
4.2.6 Hypothesis 6: Adjacency ... 30
4.2.7 Hypothesis 7: Expected income... 31
4.2.8 Hypothesis 8: Level of education ... 31
4.2.9 Hypothesis 9: Country internationalization levels ... 32
5. DATA ... 33
5.1 Sample ... 33
5.1.1 Sample design ... 33
5.1.2 How is the TMT defined? ... 34
5.2 Operationalization of the dependent and independent variables ... 34
5.2.1 Dependent variable ... 34 5.2.2 Independent variables ... 35 5.2.2.1 Cultural distance ... 35 5.2.2.2 Historical ties ... 36 5.2.2.3 Common language ... 36 5.2.2.4 Institutional distance ... 36
5.2.2.5 Expected income differential ... 37
5.2.2.6 Geographic distance ... 39 5.2.2.7 Adjacency ... 39 5.2.2.8 Educational level ... 39 5.2.2.9 Country internationalization ... 40 5.2.3 Control variables... 40 5.3 Formation of country-couples ... 42
6. METHODOLOGY ... 44
6.1 Ordinary least squares or Fractional logit regression? ... 44
6.1.1 Ordinary least squares estimation ... 44
6.1.2 Methods for fractional response variables ... 44
6.1.3 The Fractional Logit Model ... 45
6.2 Panel data and cross-sectional analysis... 47
6.2.1 Panel data regression with random effects ... 47
6.2.2 Cross-sectional regression ... 48
6.3.1 Summary statistics and paired correlations ... 49
6.3.2 Multicollinearity ... 49
6.3.3 Internationalization levels of boards per country ... 49
7. RESULTS OF THE REGRESSION ANALYSES ... 51
7.1 Panel data results ... 51
7.1.1 Cultural Distance ... 52
7.1.2 Colony ... 52
7.1.3 Common language... 52
7.1.4 Institutional distance ... 53
7.1.5 Geographic distance ... 53
7.1.6 Expected income differential ... 53
7.1.7 Difference in the quality of management schools ... 54
7.1.8 Globalization levels ... 54
7.1.9 Adjacency ... 54
7.1.10 Education Index... 55
7.1.11 Control variables... 55
7.2 Cross-sectional results ... 56
7.3 The difference between the OLS and FL models ... 57
7.4 Post-hoc analyses ... 59
7.4.1 The influence of the USA ... 59
7.4.2 The influence of the English language ... 60
8. DISCUSSION ... 62
8.1 Interpretation of results ... 62
8.1.1 The effect of classic psychic distance factors ... 62
8.1.2 The effect of expected income differential and differences in education levels ... 64
8.1.3 The effect of country internationalization levels ... 65
8.1.4 Interpretation of the post-hoc tests ... 66
8.2 Limitations and recommendations for future research ... 67
9. CONCLUSIONS ... 70
REFERENCES ... 72
“Business is going global, and so are boards – or at least they should be if they are to maximize their effectiveness”
- Mandl, 2003
1. Introduction
Over the past decades, nationality diversity in societal groups, including countries‟ workforces, has been changing tremendously. European Union member countries have even opened up their borders to each other initiating free movement of labor and trade between them and promoting diversity. In the United States, demographic changes, technological advances and global competition have forced organizations to reexamine policies and practices in order to deal with a dramatically changed workforce (Wentling and Palma-Rivas, 1997). Already in the 1980s the Hudson Institute published Workforce 2000: Work and Workers in the 21st Century (Johnston and Packer, 1987), which
claimed that US based companies would have to implement changes in the way people were managed in organizations, in order to learn to recruit, motivate, and retain high-quality workers from all demographic backgrounds to remain competitive.
Nationality diversity and its antecedents are the main theme of this study. Previous organizational diversity research has mostly focused on diversity within the entire workforce of an organization, while the upper echelon has received less attention. Recently, strategy researchers have focused on the role of the top management team (TMT) in creating sustainable competitive advantage and gaining above-normal performance (Carmeli and Tishler, 2006). Resource-based view strategists and upper-echelon theorists suggest that the firm‟s TMT is a critical resource for its success because of the significant influence it has on the firm‟s strategic decisions and how these are implemented (Castanias and Helfat, 1991, 2001; Hambrick and Mason, 1984; Wiersema and Bantel, 1992). It seems that firms and shareholders have set their hopes on CEO‟s and upper echelons in general, which makes it essential to carefully choose leaders, and identify the acts and skills associated with leadership (Bryman, 1996; Khurana, 2002). Diversity in TMTs has been the focal point of an increasing number of research papers, yet there is no consensus on the causes or implications of TMT diversity. This should be reason enough for upper echelon researchers to “open up the black box” (Lawrence, 1997) and further discover this neglected field of research.
In this study, I examine nationality diversity among board members from European multinational corporations (MNCs). Nationality diversity in particular has been studied only occasionally. However, in light of the enduring internationalization processes in the business world and the level of impact decision makers have, it seems rather odd that the national composition of the organization‟s most powerful entity has received so little attention.
characteristics determine which foreign nationalities we find in boards.
The paper is structured into nine sections. The next section provides a literature review that sheds light on the different theories and streams of research signifying the
importance of (board) diversity research, and ends with a description of the problem statement and research question. The third section provides a theoretical framework in which the hypotheses are embedded. In the fourth section the hypotheses are developed and grounded in literature. Section five describes the data and sample, section six
2. Literature review
This literature review will summarize the theories and streams of research that altogether explain the importance of studying the upper echelon and diversity in boards. Next, it gives an overview of previous research on nationality diversity within boards. The section concludes by formulating the problem statement and research question addressed in this study and its relevance embedded in literature.
2.1 Theories underlying board diversity studies
2.1.1 The upper echelons theory
Hambrick and Mason‟s (1984) upper echelons perspective initiated the development of a large body of research on top management teams (TMTs). The core of this upper
echelons theory builds on two pillars: (1) executives act on the basis of their personalized interpretations of the situations they face, and (2) these personalized understandings are a function of the executives‟ unique characteristics. This leads to the implicit assumption that managers‟ influence on organizational issues is best reflected by psychological characteristics, for instance their norms and values, cognitions and cultural perceptions (Knight et al., 1999). The difficulty lies in assessing these characteristics. Therefore, most researchers have used demographic characteristics as a proxy for these psychological characteristics.
Upper echelons research has focused mainly on the possible relationship between TMT characteristics and firm strategic profiles (e.g. Bantel and Jackson, 1989; Finkelstein and Hambrick, 1990; Smith, Grimm, Grannon and Chen, 1991), and on if and how TMT characteristics might be related to firm performance (e.g. D‟Aveni, 1990; Haleblian and Finkelstein, 1993). Moreover, recent research has focused on a clearer and more
comprehensive explanation of how managers affect firm outcomes, and has elaborated on and enhanced Hambrick and Mason‟s original upper echelon‟s model (Carpenter,
new synergy (Palmer and Varner, 2002; Liden, Wayne and Krainer, 2001). That is, diversity can create a new common culture within a team. Bell (1992) calls this a “transactional culture”. When the various cultures represented within a team are all well integrated, such a transactional culture can improve a team‟s performance.
Hambrick and Mason‟s (1984) theory has been, and still is, tremendously popular in various fields, and has been cited in hundreds of refereed journal articles (Carpenter, Geletkanycz and Sanders, 2004). However, despite its enormous impact, scholars argue that the large body of literature that builds on the upper echelons theory lacks integration, as a consequence of the extremely high number of studies combined with a lack of methodological review articles (Carpenter, Geletkanycz and Sanders, 2004).
Correspondingly, some of them suggest that empirical results have been ambiguous, and that over all, upper echelons research has raised more questions than it has answered (Priem, Lyon and Dess, 1999; West and Schwenk, 1996). Finally, extensions of the upper echelons perspective into the international(ization) context have been scarce (Heijltjes, Olie and Glunk, 2003). My study contributes to the field of upper echelons research by integrating both new insights in empirical methods and applying upper echelons research in the context of internationalization,
2.1.2 Conflict theory
strategic decision making (Eisenhardt, Kahwajy and Bourgeois, 1997b). On the other hand, conflict can limit team cohesiveness, waste energy and resources and ultimately weaken the quality of the TMTs decision making process (Amason, 1996).
These disadvantageous effects are now assumed to be a consequence of affective conflict: the type of conflict that is emotional and arises from personalized incompatibility or clashes (Amason and Sapienza, 1997; Simons and Peterson, 2000). This type of conflict is never beneficial and only limits team cohesiveness and impairs TMT decision making. The type of conflict that is most often beneficial to the TMT and the firm in general is called cognitive conflict and is task related, arising from judgment or perspectives (Amason and Sapienza, 1997; Simons and Peterson, 2000). It encourages critical
examinations of the decision making criteria and the status quo, and stimulates different perspectives. However, even cognitive conflict may hurt team and firm outcomes once it reaches very high levels (De Dreu and Weingart, 2003).
From the conflict literature we can thus conclude that team diversity may have many beneficial effects to the firm, although it might also spur affective conflict, and ultimately cause negative effects stemming from cognitive conflict. There are no guidelines that define the optimal level and type of diversity, making the topic a rather complex one. The same applies to the management of diversity and the conflict stemming from it.
2.2 Nationality diversity
2.2.1 Firm internationalization related to the composition of the board TMT nationality diversity (or TMT internationalization) has been a rather
As a result of this increased interaction with people from many different backgrounds, policies and processes should be implemented that minimize misunderstanding and exploit the potential benefits that come with diversity (Palmer and Varner, 2007). Logically, this may require consequences at the board level. For instance, the upper echelons theory predicts that in well-functioning organizations, boards should reflect the organization‟s activities, and thus mirror the dispersion of activities across countries (Hambrick and Mason, 1984). According to Ghoshal (1987), TMTs, as the central decision-making unit of the MNC, are faced with increasingly complex decisions as the firm internationalizes through an increase in foreign operations. This increase in
complexity in turn requires a higher degree of information-processing capacity among the TMT members in order to effectively manage the increasingly internationalized
organization (Sanders and Carpenter, 1998). And following Ashby‟s “principle of requisite variety” (1956), MNCs should strive to balance the demands posed by international operations with higher international capacity at their headquarters.
Perlmutter (1969) already predicted that the workforce of an MNC, including its top management, will slowly turn from „ethnocentric‟ (the orientation is on the home country, managerial positions are reserved for home country executives) into „geocentric‟ (the company has evolved from the geographic centre, and values cultural diversity), parallel to the process of globalization. That is, as the orientation of companies shifts towards international activities, management will become increasingly diverse in terms of nationality. The reasoning behind this is that firms need specific knowledge and
characteristics in order to be able to effectively manage foreign relationships. Therefore, a TMT should not only consist of local managers, but also needs managers that
“represent” foreign activities (Schmid and Daniel, 2006).
increasing use of English as a lingua franca in multinationals, and the internationalization of both capital, goods and services markets (Ruigrok and Greve, 2008)
2.2.2 Advantages of TMT internationalization
In addition to the common advantages from diversity that are grounded in and explained by conflict theory, nationality diversity in particular may be beneficial for several reasons. Here I give two examples.
International diversification of a firm‟s operations logically increases the level of
contextual complexity it faces. As a consequence, this increase in contextual complexity raises the information-processing needs of the firm, thereby changing the requirements on the TMT. In the early to mid-stages of foreign expansions, costs of “liability of
foreignness and newness” are typically rather high (Gomes and Ramaswamy, 1999), next to high initial learning costs and often insufficient economies of scale (Contractor, Kundu and Hsu, 2003). In general these “start-up costs” may initially outweigh the benefits of internationalization. However, foreign board members, with their knowledge of foreign markets, understanding of different cultures and complementary networks could help in reducing this “liability of foreignness and newness” (Ruigrok and Greve, 2008).
Moreover, by not only focusing on the local executive labor market, but instead accessing the international pool of executive labor, firms are able to screen a larger number of potential TMT members, which increases the chance of hiring the best suitable candidate (Ruigrok and Greve, 2008).
2.3 Why is it that diversity in boards remains low?
2.3.1 Causes of low diversity
Even though research has recognized the advantages associated with team diversity, companies‟ attitudes are often still rather reticent. For instance, levels of nationality diversity and female board members are still relatively low in most MNEs (Glunk,
Another explanation for the persisting homogeneous character of management teams can be found in the notion of social categorization (Turner, 1987). According to this notion, people have a natural tendency to feel attracted to persons similar to themselves, and to categorize the others. The associated categorization process is generally based on observable characteristics such as age, race or nationality, and status. This act of categorizing is found to create negative consequences for team dynamics and effectiveness.
2.3.2 Barriers to the development of an international executive labor market
However, the development of an international market for executive labor is slowed down by several emotional, structural and political barriers. Ruigrok and Greve (2008) cut it down into four categories. First, executives are not perfectly mobile. Language and cultural barriers, valuable home-country networks, family issues and pension systems keep executives from moving to the country that offers the most attractive employment opportunity (as would be the case in an efficient international market for executive labor). Second, there are often still some emotional and political links between a company and its home country, even when a firm is highly international. In this case, shareholders, stakeholders, media and the general public are often skeptical about leaving decision-making powers in the hands of foreigners. Third, in order to reduce risks companies often recruit their executives via personal networks, that are, especially in the early stages of firm internationalization, often dominated by domestic relations. And fourth, moderate levels of TMT nationality diversity are said to be harmful for team functioning, reducing team effectiveness levels both below the levels of homogeneous and highly
heterogeneous teams. This leaves many companies struggling on their way to internationalization (cf. Earley and Mosakowski, 2000; Lau and Murnighan, 1998).
2.4 Board internationalization research
2.4.1 The importance of board internationalization research
2007/2008; Van Veen and Marsman, 2008). Other types of diversity in TMTs have received more attention, such as the age of board members, previous experience and connections, and the size of the TMT (Staples, 2007). However, studies on diversity of TMT member nationalities are of importance for several reasons (Van Veen and
Marsman, 2008). First, nationality diversity within the TMT is an indicator of the actual level of internationalization of the MNC, since it shows to what extent the company is detached from its country of origin (Ruigrok and Van Tulder, 1995). Second, TMT nationality diversity may serve as an important indicator of a company‟s „transnational mindset‟ (Bartlett and Ghoshal, 1998). Lastly, greater knowledge on sources of
nationality diversity in TMTs is considered to benefit the MNC (i.e. Athannassiou and Nigh, 2000; Tan and Mahoney, 2006), in such way that higher nationality diversity is presumed to be a requirement for qualitative strategic decision making, which eventually leads to better company performance (Van Veen and Marsman, 2008).
2.4.2 Previous studies on board internationalization
As I already stated, the number of studies addressing the issue of nationality diversity in MNC boards is remarkably low. Only recently executive search firms (e.g. Heidrick & Struggles, Spencer Stuart) have presented descriptive overviews of internationalization of boards. In 1999, Gillies and Dickinson concluded that 36.3 per cent of the 80 largest companies worldwide had at least one foreign member on their boards in the year 1993. In his 2007 and 2008 papers, Staples studied the same set of companies, and found that diversity became more widespread, but was still not very extensive. He also found that cross-border acquisitions generally result in more multinational boards, and that multinational boards are concentrated in Europe. Other researchers have focused on European boards. Heijltjes, Olie and Glunk (2003) concluded that the number of
period 2000 to 2005, Swiss, Dutch and Finnish firms significantly increased the percentage of foreign board members, while Danish and Norwegian firms showed a downward trend. Also, they demonstrate that executive compensation levels and foreign ownership levels of the firm stimulate TMT nationality diversity. A study by Ruigrok, Greve and Nielsen (2009) shows that companies match their top managers‟ profiles to their strategies. They show that entry into new foreign markets and new cultural zones is associated with a higher level of international capacity within the board.
Thus, so far there is no consensus on the existence of possible sources or implications of TMT nationality diversity. In general, research has much more often addressed the consequences of board members‟ diversity than its antecedents. Finkelstein and
Hambrick (1996) argue that while most management researchers focus on the effects of board members and their actions, a better comprehension of the determinants of their characteristics is needed.
2.4.3 Foreignness versus heterogeneity
Until now, most studies on TMT nationality diversity only distinguished between national or non-national TMT members in measuring top management team
internationalization. In these studies, nationality diversity was simply calculated as the number of foreigners as a percentage of the total number of managers. The different nationalities of board members were not taken into account.
To my knowledge, there are practically no studies that have focused more in detail on board members‟ origins or citizenship. Ruigrok and Greve (2008) and Ruigrok, Greve and Nielsen (2009) have studied the rise of an international market for TMT members from MNCs in seven European countries. In their paper, they explored the antecedents of TMT nationality diversity taking into account the intensity of nationality heterogeneity as measured by Blau‟s (1977) index, instead of simply calculating the percentage of
the breadth of international executive labor market utilization1. They found an overall increase in the degree of nationality diversity, although their results suggested that an international market for executive labor is at best emerging very gradually, and that the development of such a market differs across countries and regions. Further analysis showed that strategies to internationalize TMTs seem to be strongly affected by the accessibility of culturally, institutionally and linguistically related host countries. This last finding points towards the existence of a pattern in nationality distribution across TMTs of different origins.
2.4.4 Why nationality matters
Analyzing boards by looking at the different nationalities of their members might prove interesting. After all, not all foreigners are the same, and the presence of one nationality might seem more remarkable than the presence of the other. In line with this reasoning Schmid and Daniel (2006) argue that the former may add relatively more “diversity” to a team than the latter (Schmid and Daniel, 2006). For instance, consider a Dutch firm. Adding a German member to its TMT may change less than adding for instance a Korean member. That is, the differences between German and Dutch culture (among others) are likely to be less profound than the differences between the Korean and Dutch culture. Therefore, arguments have been raised to measure TMT internationalization
incorporating for instance cultural distances. For example, studies in numerous fields have already applied Kogut and Singh´s (1988) measure of cultural distance to address this issue. Considering all this, when studying board internationalization, it seems more appropriate to take into account the different nationalities of board members instead of categorizing into „native‟ or „foreign‟.
1
The Blau index is calculated as follows: B = [ 1 – Σ(pi)2 ] , where B represents the heterogeneity measure
2.5 Problem statement, research question, and relevance of the study
2.5.1 Problem statement and research question
Previous research in the field of board internationalization has been mostly limited to finding an explanation for differences in the level of board internationalization across countries. Even though the country of origin of the MNC under study is often denoted as being rather important in determining their boards‟ levels of internationalization, the country of origin of the people that make up these boards has never been taken into consideration. However, assuming that each foreign nationality adds the same “degree” of internationalization to a team does not seem logical.
Also, previous studies have shown that there is still no real international executive labor market, yet the causes for this have not been thoroughly investigated. Such an
international executive labor market has often been associated with the
internationalization of firms‟ upper echelons. However, what has been ignored in practically all studies is the choice and options that those executives themselves have in defining their career paths. That is, the personal aspect of board internationalization has been disregarded. For example, European companies might be very interested in
appointing US natives as members of their boards, yet without the will of those
Americans themselves there will be no representation of Americans on European boards. In general, internationalization is, as the word already denotes, a multi-country
phenomenon. In case of board internationalization this relates to flows of board members between countries. None of the studies so far have focused on these flows.
Therefore, research into the antecedents of nationality diversity by its very definition seems essential. Only accounting for the organizations‟ home country seems
inappropriate, and according to the reasoning from section 2.4.4 the following line of thought is developed.
Addressing the issues mentioned in the problem statement, I will examine the nationality mix in European boards. I will investigate whether country differences and characteristics can explain flows of board members between countries. The research question that I will try to answer is:
Which country characteristics, differences or similarities determine board members’ nationalities?
2.5.2 Relevance of the study
Studying the trends and patterns in executive labor flows is beneficial for several reasons. First, it provides us with insights in the existence and/or rise of an international market for executive labor. If a company has access to all management talent from across the world, the probability of hiring the right candidate for the right job is biggest. Next, even though national heterogeneity is commonly said to improve team (and ultimately
3. Theoretical Framework
This section will provide theories and literature (other than the ones discussed in section two) on which the hypotheses developed in section four build. First, internationalization theory and the behavioral theory of the firm increase our understanding of the role of uncertainty within firms. Second, labor migration literature is reviewed. Whereas the former more or less represents the considerations of the firm in its internationalization decisions, this second stream of research gives an idea of what might drive persons (in this case board members) to participate in foreign boards.
3.1 Internationalization theory and the behavioral theory of the firm
“If organizations exist to unite diverse perspectives, capabilities, and talents in pursuit of common purposes and mutually beneficial results, why do they stifle diversity, seek sameness, discourage individuality, promote conformance, reward uniformity, and punish nonconformity? Because managing diversity is harder than managing uniformity -- managing diversity is more challenging, expensive, time consuming, demanding, stressful, and prone to fail” – Craig Hickman, 2006
These words, written by Craigh Hickman, are clear-cut; organizations generally promote university rather than diversity, even though this seems to be – in essence – an unnatural practice. The author gives several causes for this type of organizational behavior. For instance, managing diversity is said to be more challenging, expensive and time consuming. However, as explained in the previous section, diversity may have several benefits that are likely to outweigh these disadvantages. So why is it that organizations still seek uniformity? The answer lies in Hickmans last point: managing diversity is more prone to fail than managing uniformity.
instance patterns of organizational expansion across borders, and has shown that firms prefer to enter those markets that are relatively similar to the home market. Such markets are more easily understood, which in turn limits the risk to which the firm will be
exposed (Wiedersheim-Paul, 1972).
Thus, internationalization raises uncertainty, which in turn generates a risk for the firm. Reasonably, the internationalization of boards will bring the firm a comparable “risk”. That is, conflict theory predicts that high levels of team heterogeneity might lead to a weakening of social cohesion within the team, and ultimately to a decrease in team effectiveness. This is also what mister Hickman pointed out: diversifying the team comes with a higher probability of failure. Of course, when managed properly, the benefits of diversity often outweigh these drawbacks. The point is that there are no guarantees for success, whereas the risk experienced by the firm increases. Since organizations are generally uncertainty avoidant, as we learn from internationalization theory and behavioral theory, they will normally prefer to play safe. Therefore, the level of board nationality diversity is likely to be affected by the firm‟s risk-avoidant attitude. That is, the organization is likely to make a trade-off of risk-levels and the benefits from diversity, just as it does in deciding upon internationalization strategies of its activities.
3.2 Labor migration literature
The literature on labor migration flows also generates some interesting insights. In their study on international labor migration towards and within Europe, Salt, Clarke and Wanner (2004) state that the nationality profile of documented migration is mostly influenced by geographic proximity, and cultural and historical ties. Their study also shows that overall trends in labor flows show some similarities to those of total
movement of professional, managerial and technical staff across countries. They state that this migration by the highly-skilled is a rather new phenomenon, which came into play mostly after the mid-1980s, and probably is a consequence of economic
globalization and the activities of transnational corporations.
Salt, Clarke and Wanner further state that the characteristics of labor migration flows differ tremendously across European countries, and that identifying common trends and patterns is a rather complex task. From the migration data they study, they draw three main conclusions: (1) there is evidence of regional self-containment: most migration flows are destined to somewhere else in Europe, and often in the same region. However, they found that relative regional self-containment has been decreasing since the mid-1990s which makes the current flow pattern more diversified. (2) Migration flows across countries are highly different, reflecting a range of historical (e.g. post-colonial links) and geographic (in particular proximity) processes, and (3) the patterns found subsequently reinforce the diversity of migration across Europe. Next to that, during the 90s, most countries showed a decline in the proportion of foreigners coming from the European Union among the total population (Salt, Clarke and Schmidt, 2000).
Another study that focused specifically on highly-skilled labor migration shows the importance of OECD membership (Chaloff and Lemaitre, 2009). The results of this study demonstrate that within OECD countries, qualifications and work experience earned in non-OECD countries are much less appreciated and valued. Indeed, the study shows that in their hiring decisions, companies generally favor those candidates that earned their qualifications in an OECD country.
3.3 Relevance for board internationalization
between countries, flows of executive labor between these countries will be relatively large because of reduced cultural integration barriers. Vice versa this implies that the absence of such ties will thwart these flows. These notions have, up until now, never been further developed or tested empirically.
In line with firm internationalization theory, the behavioral theory of the firm, labor migration literature and the findings of Ruigrok and Greve (2008), I propose that it is country differences and characteristics that are likely to predict which nationalities we are most likely to find in boards of a specific country‟s organizations.
As I said earlier, one of the key concepts of the behavioral theory of the firm is the notion that organizations rather avoid dealing with uncertainty. Logically, the greater the
4. Hypotheses
4.1 Psychic Distance
As I pointed out in the previous section, I will investigate the influence of country differences (or similarities) and characteristics on the representation of various nationalities in boards in different countries. Probably, the measure of country
(dis)similarity that has been used most often is psychic distance (Sivakumar and Nakata, 2001). It was first coined in the 1950s by Beckermann (1956) in his empirical research on intra-European trade flows, but only gained prominence in management literature in the 1970s, when a group of researchers at the University of Uppsala started focusing on it. They defined the concept as “the sum of factors preventing the flow of information to and from the market” (Johanson and Wiedersheim-Paul, 1975, p. 308), and further developed it. Since that time, many researchers have elaborated on the construct, and how it should be operationalized. However, consensus on what constitutes Johanson‟s “sum of factors” has never truly been reached.
In this research I will include psychic distance factors that have been widely used in trade, FDI and migration literature, and are in my opinion most appropriate for this study. In the following section I develop hypotheses that will ultimately be empirically tested. Hypotheses 1 through 6 closely relate to psychic distance research and have been included in the “sum of factors” many times before. These factors are: cultural distance, historical ties, common language, institutional distance, geographic distance and
adjacency.
4.2 Hypotheses development
internationalization purposes (country globalization levels).
4.2.1 Hypothesis 1: Cultural distance
The majority of researchers that have used psychic distance in their research, relied mainly on Hofstede‟s (1980) research and measures of culture to quantify the cultural distance between countries.
National culture is of major importance in deciding upon a firm‟s strategy, as strategies are generally chosen on the basis of assumptions regarding the environment, and about relationships among people (Schneider, 1989). Cultural familiarity theory holds that an organization is less likely to invest in culturally dissimilar countries. According to Cartwright and Cooper (1993), firms better opt for a minimal exposure to unknown foreign phenomena (market presence, employees etc.) since the different psychological environment of a foreign culture will lead to conflict.
Also, Belot and Ederveen (2005) find evidence that cultural barriers play a crucial role in labor migration between OECD countries, and their explanatory power in migration patterns goes beyond that of differentials in economic variables.
Cultural distance reflects a difference in cultural values between countries. The construct of cultural distance is primarily inspired by Geert Hofstede‟s work (1980). In his 1980 study, Hofstede found four dimensions of culture that have since been used extensively in classifying countries according to their national cultures, and to determine between-country cultural distance. Kogut and Singh (1988) designed a measure that allows to calculate an index of cultural distance between countries based on these cultural dimensions. Logically, this index is a simplification of complex issues (Chapman, Gajewska-De Mattos, Clegg and Jennings Buckley, 2008), though many studies have confirmed the validity of the dimensions, proving their reliability (for an overview see Søndergaard, 1994).
From all the above, and in line with the internationalization process theory, it seems straightforward that if companies include foreigners on their TMT, they would rather include members from countries that have a relatively low cultural distance to the firm‟s home country since this limits the complexities in for instance coordination and
markets that are culturally relatively similar appeal most, since this makes working, communicating and settling in another country much easier. Therefore, my first hypothesis is as follows:
H1: The greater the cultural distance between two countries, the smaller the flow of board members between them.
4.2.2 Hypothesis 2: Perceptions and historical ties
Even though cultural distance has often been used as a synonym and proxy for psychic distance (O‟Grady et al., 1996) resulting in an interchangeable use of the concepts, the two certainly need to be distinguished. According to Dow and Karunaratna (2006), it is not necessarily incorrect to use cultural distance (for instance based on Hofstede‟s scales), though we need to be aware that this is only one of the aspects that determine psychic distance. What researchers often “forget” is that psychic distance is determined by the extent to which a foreign country is perceived as similar to the home country (Sousa and Bradley, 2006). Thus, psychic distance is not, unlike cultural distance, absolute, but cognitive, as it is a subjective perception. To be precise, managers‟
perceived understanding of foreign markets depends not only on actual knowledge, but in addition also on beliefs, assumptions and generalizations that are held to be true
(O‟Grady et al., 1996). Isn‟t it true that our decisions are based on what we think we know, instead of on what we surely know?
Even though historical ties do not directly refer to country differences, they often appear associated with reduced psychic distance. Many researchers claim that colonial or historical ties affect the perception of psychic distance (e.g. Brewer, 2007; Dow and Karunaratna, 2006). Brewer (2007) argues that this happens directly, since a historical tie results in a more detailed knowledge of the country, whereas Child, Ng and Wong (2002) expect it to be an indirect process, where historical ties lead to political or institutional and linguistic similarities between two countries. Hotho (2009) considers a historical tie a “country-level psychic distance stimulus that positively affects perceptions of familiarity with a particular country by fostering more or less collectively shared beliefs about that country” (p. 73). Also, according to postcolonial critic Edward Said (1978), colonial ties are reflected in national education systems and for instance literary histories of the colonizers. Ultimately, collective knowledge structures of both the colonized countries and the colonizers are affected by these particular representations.
A question that comes to mind is whether both “positive” and “negative” historical ties would decrease psychic distance. Naturally, it can be argued that historical ties for which the undertone is negative could have a different effect on psychic distance than those with a more positive connotation. However, following Hotho (2009), here I separate the
historical tie‟s emotional connotation from the perception of familiarity resulting from it. Thus, I assume that all historical ties, irrespective of the sentiments associated with them, negatively affect psychic distance as a result of a feeling of familiarity stemming from these ties.
In this study, I use colonial ties as a proxy for the familiarity perceptions that are part of psychic distance. I argue that one perceives a foreign country to be relatively familiar if there is a colonial history between this country and the home country. Subsequently, one thinks to posses particular knowledge about this foreign context, which reduces the uncertainty experienced towards it, regardless of the accuracy or existence of this knowledge. Accordingly, since uncertainty is reduced, the same reasoning that lead to Hypothesis 1 applies here, and I hypothesize the following:
H2: A historical tie between two countries positively affects the flow of board members between them.
4.2.3: Hypothesis 3: Common language
Another factor that is argued to influence the distance between countries, and that has received support from most psychic distance researchers is common language (e.g. Johanson and Wiedersheim-Paul, 1975; Johanson and Vahlne, 1977; Dow and Karunaratna, 2006). It seems straightforward that if two countries share a common language, this facilitates communication and reduces misunderstanding and ultimately conflict. A first analysis of European TMTs by Ruigrok and Greve (2008) shows that firm strategies to internationalize TMTs seem to be influenced by linguistically related countries. Also, international labor migration literature provides support for the
importance of language in the determination of flows, origins and destinations of (executive) labor (Salt, Clarke and Wanner, 2004; Belot and Ederveen, 2005).
H3: When two countries share a common language this positively affects the flow of board members between them.
4.2.4 Hypothesis 4: Institutional Distance
institutional distance (which is composed of regulative, normative and cognitive distances) may affect the location and entry mode strategy of MNEs. Others predict a relationship between the institutional distance and expatriate adjustment (Ramsey, 2005) and subsidiary staffing strategies (Gaur, Delios & Singh, 2007). In all, new institutional theory is based on the notion of legitimacy and transferability of particular practices and ideas: a concept that might thus also be applicable to team diversity.
Van Veen and Elbertsen (2008) argue that the accessibility of MNC boards for foreigners is to some extent determined by the institutional rules and procedures of the country in which the MNC has its roots. In their study, they find that governance regime effects on nationality diversity are strong.
In his study on the location of foreign investment, Hotho (2009) is one of the first to link institutional differences to psychic distance. Even though he does not find significant support for his prediction that institutional country differences negatively affect
locational attractiveness, this is not to say that institutions do not matter in diversification decisions.
Labor migration research gives us insights in the location decisions made by potential foreign board members. According to Shrestha (1987), the institutional framework of different countries has to be taken into account in the examination of drivers of labor migration. She argues that when institutional goals are in harmony with people‟s needs and desires, their migration behavior will generally reflect their compliance with state policies. Belot and Ederveen (2005) argue that even when there is complete free movement of labor between two countries, workers will be confronted with a series of institutional barriers hampering their movement. These institutional barriers can be consequences of for instance governmental policies, such as the lack of recognition of qualifications or non-transferability of pension rights, but they can also flow from the worker‟s personal ideology that clashes with a host country‟s institutional environment (e.g., one‟s view on the size of the informal sector, the extent of the tax system, the extent of bureaucratic red tape).
Considering all the above, I expect the following:
4.2.5 Hypothesis 5: Geographic distance
It is rather straightforward that geographic distance often determines the willingness of executives to move to a foreign country. Also, close proximity promotes the exchange of information between countries (Ghemawat, 2001). The further away the country, the less likely people will be informed about for instance job opportunities. Next to that,
geographic distance often relates to cultural distance between two countries, in that cultures of neighboring countries are often quite similar. On the other hand, this is a risky insinuation, since this assertion does not always hold. However, even if it is not true, people widely believe it is, and likewise geographic distance relates to psychic distance, which renders it indispensable in studying the predecessors of foreign TMT membership.
H5: The greater geographic distance between two countries, the less likely it is to find flows of board members between them.
4.2.6 Hypothesis 6: Adjacency
In addition to geographic distance, I will also take a look at the effect of adjacency. When geographic distance is measured as the distance from capital to capital, relevant distance is likely to be overestimated (Linders, Slangen, De Groot & Beugelsdijk, 2005).
Especially for neighboring countries this overestimation can be rather large. For instance, when measured like this, the geographic distance between The Netherlands and France is smaller than the distance between the former and Germany. Intuitively, this does not really make sense, since The Netherlands and Germany are adjacent countries. Also, irrespective of distance, a shared border might affect the psychic closeness to a country, since it might increase familiarity perceptions. More sophisticated measures of
H6: When two countries share a border, the flow of board members between them increases.
4.2.7 Hypothesis 7: Expected income
The classical model of Harris and Todaro (1970) stresses the importance of economic incentives in labor migration patterns. The model assumes that individuals base their migration decisions on the differential between the expected income in the destination country and the expected income in their home country. Ruigrok and Greve (2008) also state that compensation functions as a price mechanism in the executive labor market and plays an important role in the evolution of executive careers. That is, it is said to
influence important career-decisions made by top managers, such as going abroad. Ruigrok and Greve further say that these top managers will be at least partially motivated by a relatively high expected monetary reward and the status associated with it, not least because of the substantial switching costs for them to move to another country. In their study, they find a highly significant relation between executive compensation levels and the percentage of foreign TMT members and the level of nationality diversity.
In line with the Harris-Todaro model and the findings of Ruigrok and Greve, I predict that:
H7: The greater the (positive) differential between top managers‟ incomes between two countries, the greater the chance that we find persons from the country with the lower income in boards of the other country.
4.2.8 Hypothesis 8: Level of education
Logically, the chance of finding a person of a certain country in a board will be
considerably smaller when there are simply no suitable candidates in that country. Some countries have a relatively good educational system which breeds talent. In these
countries, people often have the opportunity to follow secondary and tertiary education after finishing primary school, and are being motivated by the government to do so. These countries are most likely to have a great pool of talent, part of which is destined to become the future leaders of large multinationals, either at home or abroad.
quality of education in country b is relatively good. Intuitively, based on educational background only, country b would rather draw managerial talent from its own pool than from country a. The reverse applies to country a, which should be interested in attracting talent from country b.
Therefore, I hypothesize the following:
H8: The greater the difference between education levels of two countries, the greater the likelihood of finding persons from the country with the higher level in boards of the country with the lower education level.
4.2.9 Hypothesis 9: Country internationalization levels
Just like one‟s nationality probably determines what will be one‟s primary language, it also affects the likelihood of knowing other languages and which specific languages will ever be readily comprehended (Church, 1982). This in turn often relates to the extent to which a country is internationalized (Klee, 2009). Foreign language acquaintance is only one factor that illustrates the relevance of internationalization levels in globalization decisions. When countries are relatively internationalized they normally have some experience in various international activities, often leaving them more tolerant and less risk-avoiding concerning such activities. This generally translates into the same attitude for its residents.
I therefore expect the following:
H9a: When a country has a relatively high internationalization level, the probability of finding people from this country in foreign boards is higher.
5. Data
5.1 Sample
5.1.1 Sample design
I collected data constituting a stratified sample that covers companies from 15 European countries2 over the years 2005, 2006 and 2007. In general, the companies included are the European „blue-chip‟ companies which in 2005 were listed in their country of origin, usually MNCs. In total, the sample covers 361 companies in 2005, for 2006 there are 354 left, and in 2007 a total of 347 are included. Some companies do not appear in all three years because of for instance acquisition, bankruptcy or merger. The 2005 sample is equal to the one used by Van Veen and Marsman (2008) in their study on nationality diversity. The companies included are listed in Table 1 of the Appendix.
For the year 2005, the total number of managers in my dataset is 5683, for 2006 I collected data on 5691 managers, and for 2007 there are 5563 managers in total. The company’s country of origin is defined as the place where its headquarters are situated. Information on these managers‟ nationalities could be collected from the companies‟ annual reports, additional reports and their websites, and if not available, other secondary sources such as Zoom Information and Forbes.com were consulted. In the rare case that a board member had two nationalities, I used the nationality of the country in which the person was raised and educated, built a career etc. Also, if someone‟s nationality was not explicitly mentioned, I derived it from information on education and career as well. In cases where none of this information was present (an extremely small proportion of the data) I derived nationality from their names, and assigned them the nationality for which their name was most typical (with use of websites like ancestry.com and
behindthename.com which give information on the background of names and their
2 The following countries are represented: Luxemburg, The Netherlands, UK, Germany, Belgium, Sweden,
distribution across the world etc.).
5.1.2 How is the TMT defined?
Defining who exactly are part of the upper echelon, or the TMT, is a rather complex task. Countries have their own corporate governance regulations, and the upper layers in companies are structured in different ways across countries. Carpenter, Geletkanycz and Sanders (2004) also report that the empirical definition of the TMT varies extremely across studies and over time. Nevertheless, most studies have used the dominant coalition as the central construct (e.g. the board of directors) (Finkelstein and Hambrick, 1990), hereby subjecting the construct to convenience sampling. The advantage of using convenience sampling here is that it is representative of the theoretical construct, and is the least complicated alternative since information on these dominant coalitions is relatively easily obtained through annual reports and other publicly accessible documents.
I choose to follow this line of reasoning, and define the upper echelon of a company as being either the board of directors in companies with a one-tier structure, and for two-tier companies I included the members of both the executive and supervisory board.
Consequently, I consider both executives and non-executives as constituting a firm‟s upper echelon.
5.2 Operationalization of the dependent and independent variables
5.2.1 Dependent variable
Since I am interested in the effect of country differences and characteristics on the
country level where I use all the available information. For instance, I could also choose to just look at the presence of a certain nationality in foreign boards, in which case the DV can only take on values of zero or one (and would thus be binary). That is, it would be one if a nationality is indeed represented in a certain country, and zero otherwise. However, in this case I would lose valuable information. For example, imagine that 30 per cent of the foreign board members in country i is represented by nationality Y, and 1 per cent by nationality X. Treating the DV as a binary variable would award both
nationalities a „one‟ (because both nationalities are in fact represented in the other country), even though there is a great difference in the level of representation. Therefore, I choose to operationalize the DV as a percentage.
5.2.2 Independent variables
5.2.2.1 Cultural distance. In succession of numerous scholars, I use Hofstede‟s (1980) cultural indices and Kogut and Singh‟s (1988) formula in measuring cultural distance. Like this, I am able to express cultural distance between two countries in one figure. This measure is based on the assumption that culture is a multidimensional phenomenon where Hofstede‟s four cultural dimensions proxy for its various fundamentals.
The dimensions are (1) power distance – the extent to which less influential members of organizations accept and expect that power is distributed unevenly, (2) individualism – the extent to which people are expected to look after themselves as opposed to integrate into cohesive groups, (3) masculinity - the degree to which “traditional” gender roles are assigned in a culture, and (4) uncertainty avoidance – the degree to which a society tolerates uncertainty and ambiguity. In a later study, Hofstede added the fifth dimension “long term orientation” which I will not include in my study since it is only available for a very limited number of countries. I collected the country values for the dimensions from the website www.geert-hofstede.nl. Values are available for 48 countries represented in my database. Consequently, I cannot calculate cultural distances for country couples in which at least one country has not been assigned Hofstede values.
The formula constructed by Kogut and Singh (1988) is as follows:
CDi,j is the cultural distance between countries i and j, and Ii,k, is country i's score on
cultural dimension k. Vk means that the squared differences between the countries are
deflated by the variance of that dimension, measured over all countries in the sample. Thus, cultural distance is the sum of the four squared differences weighted by the inverse of their variances, divided by four to yield an equally weighted average.
5.2.2.2 Historical ties. The data on historical ties are obtained from the CEPII dataset3. The data is presented in the form of a dummy which takes on the value of 0 for country couples that have never been involved in any colonial relationship with each other, and 1 for those who have.
5.2.2.3 Common language. The common language variable is also represented by a dummy. For this variable, I use a dummy that becomes 1 for country couples in which at least 9% of both populations speak a same language. Data for this variable are obtained from the CEPII dataset.
5.2.2.4 Institutional distance. I obtain information on countries‟ institutional
environments from the Global Competitiveness Report (GCR), like Hotho (2009) did in his study on the location of foreign investment. This Report is based largely on Executive Survey responses and is published annually by the World Economic Forum. Using the GCR has several advantages. First, it includes country data on a wide range of
institutional characteristics. Also, the survey questions on which country scores are based are standardized which facilitates systematic comparison between countries. A very important point in this respect is the fact that it is top executives’ opinions on the most important issues affecting their working environment on which the results of the GCR are based. And not only are the respondents business leaders, they also come from companies that represent the main sectors of a country‟s economy (including domestic companies with foreign operations and foreign companies located in the domestic market), and respondents ideally have some international experience. Since I focus on multinationals‟
3
board members, this certainly is a good match with my sample.
When it comes to a nation‟s public institutions, quantitative measurement is extremely difficult to obtain and therefore, the Survey serves as an instrument for gathering qualitative data. Even though a certain degree of perception bias is inevitable in such measurement, a comparison of these qualitative outcomes with available quantitative measures provides evidence that the Survey is, in fact, capturing an underlying reality instead of only reflecting on individuals‟ opinions. Moreover, its use is acknowledged by a range of international organizations such as the World Bank and the United States Agency for International Development (USAID).
In this study I will use the GCR 2005-2006. The institutional construct in this edition is based on 27 factors (which were all awarded equal weights in calculating the general construct scores), which can be found in Table 2 in the Appendix. Each year, the composition of the institutional construct in the GCR differs drastically, which makes comparison across years not viable. Therefore, I make the assumption that institutional environments do not change overnight, and accordingly use the same data on this variable for all three years in my dataset.
Institutional distance between two countries is operationalized as the absolute difference in their scores on the institutions index:
(2) IDij = | Insti – Instj |
IDij is the institutional distance between countries i and j, and Insti is country i‟s score on
the institutional pillar. I use the absolute difference since institutional practices are normally not simply assumed to be good or bad practice, and differences in institutional environments are not so much an indicator of quality differences as an indicator of contradictory “ways of living”. Thus, when calculating institutional distance, the sign is of no importance, only the fact that the country scores differ.
countries‟ Gini coefficients (Gini, 1912). The Gini coefficient is a measure of statistical dispersion that reflects the inequality of income distribution and is based on the Lorenz curve. It ranges from 0 to 1, with 0 corresponding to complete equality, and higher scores indicating a more unequal income distribution. When the coefficient is 1 there is
complete inequality. Thus, high Gini coefficients generally indicate that a
disproportionately high proportion of a country‟s total income is earned by the richest layer of its inhabitants, therefore the incomes for persons in higher positions (such as top managers) will be much higher than the GDP per capita of that country and lower for the ones that hold low profile jobs. Throughout the world, Gini coefficients on income range from approximately 0.247 in Denmark to 0.707 in Namibia.
Since I am only interested in the expected income for board members, I use these two constructs to create a proxy for “top job” income levels. By multiplying GDP per capita and the Gini coefficient, I obtain a fraction of the average GDP per capita that is higher when, ceteris paribus, the Gini coefficient increases (thus, when a relatively large proportion of income is intended for those holding higher positions).
Thus, my measure of expected income4 is calculated as follows:
(3) Iei = Ginii x GDP/capi
Where Iei is expected income in country Iei.
In the analysis, I study the effect of the income differential. Thus, the variable reflects the expected income differential between country i and country j, calculated as
(4) dIeij = Iei – Iej
Country i is the home country of the firm (or destination country of the board member), and country j is the manager‟s country of origin. Therefore, if dIeij is positive, chances of
finding a manager from country j in boards from firms based in country i are expected to
4 Note that this proxy does not give the real expected income, but more of an expected level. The Ie i differs
fromreal income in dollars since the Gini coefficient is always smaller than zero, and Iei will thus always
be smaller than the GDP per capita, which is most certainly not the case for top incomes. Therefore, one can only consider Iei an indicator of the level of income and in this respect these levels are comparable
be greater than in case this differential would be negative.
Data on GDP per capita (in thousands of US $) for the years 2005, 2006 and 2007 is obtained from the CIA World Factbook editions 2005, 2006 and 2007 which are published online by the US based Central Intelligence Agency5. I found the Gini coefficients for all countries in the Human Development Report 2009 published by the United Nations Development Programme (UNDP). These Gini coefficients are based on country-specific information covering the years 1992 to 2007, thus including the period under study.
5.2.2.6 Geographic distance. Distance is based on the geographic coordinates of the most important cities or agglomerations (in terms of population). I take this measure instead of simply taking the distance between the capital cities of both countries, since I believe it is most relevant for this study (since it concerns business headquarters and movement of people). In most cases, these two measures do not differ from each other. Distances are calculated in thousands of kilometers. Data come from the CEPII database.
5.2.2.7 Adjacency. For neighboring countries, I include a dummy variable that becomes 1 in case of adjacency and is 0 for country couples that do not share a border. I obtained data on contiguous countries from the CEPII database.
5.2.2.8 Educational level. For this variable, I use two different measures.
First, I again use one of the components of the GCR 2005-2006, namely the perceived quality of management schools (QMS) per country. Since I focus on the movement of top managers between countries, this type of education seems most relevant. Countries can be assigned a score from 1 through 7 on this construct, with 1 indicating a poor quality or limited availability of management schools, and 7 being an indicator of an average quality of management schools that is among the best in the world. Since I am interested in the effect of the difference in quality between countries i and j, the variable is
calculated as in formula (5):
5
(5) dQMSij = QMSi – QMSj
Second, I will use the more general Education Index from the Human Development Report 2007/2008. The education index measures a country‟s relative achievement in both adult literacy and combined primary, secondary and tertiary enrolment rates6. It is thus not so much a quality measure as a measure of the overall participation in education (where one might assume that a higher Education Index results in an increased pool of talent). The index values range between zero and one.
The data that were used to calculate this index cover the period 1995-2005, which
provides a good estimation of education levels for the period under study in this research, since these levels are rather stable and are not likely to be improved or deteriorate within short time spans (and are thus not believed to have radically changed during the period 2005-2007 either). The difference in education levels between two countries is
(6) dEDUij = EDUi – EDUj
5.2.2.9 Country internationalization. As a measure of country internationalization I use the KOF index of globalization, which rates countries on their level of
internationalization. The index measures the three main dimensions of globalization: economic, social, and political7. KOF scores have a theoretical range from zero through a hundred. Since the data on which KOF scores for a certain year are based always have a three year lag, I use the KOF scores from 2008 for my 2005 analyses, and 2009 and ‟10 scores for the 2006 and ‟07 analyses respectively.
5.2.3 Control variables Country size
A firm‟s home market size is said to be an indicator of the size of the domestic executive labor market (Ruigrok and Greve, 2008). Therefore, firms with large domestic markets
6
The education index is calculated as follows.
