1
The effect of international complexity on MNE’s financial
performance, and the moderating role of cultural diversity within
the company board
By Kristie Lijster
S2750309
k.n.lijster@student.rug.nl
M.Sc International Business and Management Master thesis
July 2018
University of Groningen Faculty of Economics and Business
Duisenberg Building, Nettelbosje 2, 9747AE Groningen, The Netherlands P.O. Box 800, 9700 AV Groningen, The Netherlands
http://www.rug.nl/feb
Supervisor: O. Lindahl Co-assessor: C.H. Slager
2
The effect of international complexity on MNE’s financial performance, and the moderating role of cultural diversity within the company board.
Table of content
Abstract ... 5 1. INTRODUCTION ... 6 2. THEORY ... 9 2.1. International complexity ... 9 2.1.1. Geographic scope ... 92.1.2. Cultural distance to MNE’s operations ... 10
2.1.2.1. Hofstede's dimensions...12
2.2. Financial performance ... 13
2.3. Cultural diversity within the company board ... 14
2.4. Hypothesis development ... 16
2.4.1. Geographic scope and MNE financial performance ... 16
2.4.2. Individualism-collectivism distance and MNE financial performance... 16
2.4.3. Power distance and MNE financial performance ... 17
2.4.4. Masculinity-femininity distance and MNE financial performance... 18
2.4.5. Uncertainty avoidance distance and MNE financial performance... 19
2.4.6. The influence of cultural diversity within the company board on the ‘geographic scope-financial performance’ relationship ... 20
2.4.7. The influence of cultural diversity within the company board on the ‘cultural distance to MNE’s operations-financial performance’ relationships ... 21
2.5. Conceptual model ... 23
3. METHODOLOGY ... 23
3.1. Target population ... 24
3.2. Sample ... 24
3.3. Measures ... 25
3.3.1. Dependent variable: MNE’s financial performance ... 25
3.3.2. Independent variable: Geographic scope ... 25
3.3.3. Independent variables: Cultural distance to an MNE’s operations in terms of IDV, PD, MAS and UA ... 26
3.3.4. Moderator: Cultural diversity among board members ... 27
3.4. Control variables ... 28
3
3.6. Data analysis ... 29
3.7. Missing data ... 30
3.8. Robustness checks ... 30
4. DATA ... 31
4.1. Testing basic assumptions ... 31
4 Overview of figures
Figure 1 Conceptual Model Figure A.1 Histogram residuals Figure A.2 Industry type
Figure A.3 Amount of culturally diverse boards in sample
Overview of tables
Table 1 Statistics variables Table 2 Correlations
Table 3 Baseline results regression analyses Table A.1 Location headquarters
Overview of abbreviations
CD Cultural diversity within the company board LOF Liability of foreignness
IBM International business management IDV Individualism-collectivism
MAS Masculinity-femininity
MMR Moderated multiple regression MNE Multinational enterprise PD Power distance
ROA Return on assets ROE Return on equity ROS Return on sales SD Standard deviation
SPSS Statistical package for the social sciences UA Uncertainty avoidance
5
Abstract
Nowadays more and more firms are expanding their domains. However, whether this internationalization is beneficial for financial performance remains unclear; the relationship has been subject to debate for half a century. This thesis takes into account researchers’ recommendations by acknowledging the multi-dimensional nature of internationalization ánd taking on a contingency perspective. Consequently, two internationalization dimensions mostly neglected by other studies were examined, labeled together as the MNE’s international complexity, and the potential moderating effect of cultural diversity among board members. Annual reports of 100 European MNEs were consulted and data were subjected to analyses in SPSS. The results showed a positive effect for the first internationalization aspect on financial performance; geographic scope. Results support a negative effect of one of the variables pertaining to the second internationalization aspect; the masculinity distance to foreign operations. Nevertheless, cultural diversity among board members is no moderator. As a result, European executives know entering multiple countries is beneficial, but caution must be paid to the masculinity distance when choosing which countries to enter. A culturally diverse board is not capable of strengthening or mitigating the consequences. Besides, this thesis makes a scientific contribution as the international complexity-financial performance relationship was unclear and knowledge about moderating variables, and more specifically cultural diversity within the company board, was in a developing phase.
6
1. INTRODUCTION
Both globalization and technological advances have speeded up since the 1950s (Meyer, Mudambi and Narula, 2011). According to Cavusgil and Knight (2015), these are the processes that facilitated firms to internationalize their businesses. Nowadays, the number of companies expanding their domains and entering foreign markets is still increasing (Hojnik, Ruzzier and Manolova, 2018). In comparison with the end of the 1960s, for instance, the number of firms with operations in foreign markets has increased from 7.000 (UNCTAD, 1994) to almost 80.000 in 2006 (UNCTAD, 2007). Motivations for this behavior are, amongst other things, to find opportunities and to increase MNE competitiveness (Hojnik et al., 2018).
This raises the question what the impact of this phenomena is on financial performance. The importance of this question in international business research has been extremely high (Berry and Kaul, 2016; de Jong and van Houten, 2014). However, previous findings are inconsistent and mostly contra dictionary. Researchers found positive (Delios and Beamish, 1999), negative (Denis, Denis and Yost, 2002), sigmoid (Lu and Beamish, 2004), U-shaped (Geringer, Beamish and daCosta, 1989) and even non-significant effects (Morck and Yeung, 1991). Consequently, for the past 40 years, there is an ongoing debate about the relationship between internationalization and financial performance.
Several researchers emphasize that internationalization is multidimensional (Goerzen and Beamish, 2003; Miller, Lavie and Delios, 2016; Sullivan, 1994). According to Miller et al. (2016) prior mixed results are partly due to the different conceptualizations. As a result, in order to make the relationship between internationalization and financial performance clearer, a distinction between different internationalization dimensions is required.
7 dimension. As a result, five different independent variables are examined. These will be labeled together in this thesis as the MNE’s international complexity.
Literature about the relationship between international complexity and MNE financial performance is incomplete and somewhat underdeveloped. With regard to geographic scope and financial performance, prior empirical results are mixed and contra dictionary (Goerzen and Beamish, 2003). Hence, this research domain requires further research. With regard to average cultural distance, it is mostly assumed firms first enter countries that are less distant before entering countries that are more distant. However, little research has been done on the actual performance effects of cultural distance (de Jong and van Houten, 2014; Miller et al., 2016). This is important, as distance is critical to the understanding of internationalization (Sullivan, 1994). Studies that have addressed the relationship between an MNE’s operational cultural distance and financial performance are Gomez-Mejia and Palich (1997), Lavie and Miller (2008) and Miller et al. (2016). However, Gomez-Mejia and Palich (1997) only focused on US firms, Lavie and Miller (2008) only focused on alliances and Miller et al. (2016) only focused on Japanese firms. Hence it can be concluded that the relationship between international complexity and MNE financial performance should be studied further.
In addition, a recent stream of literature takes a contingency perspective on the relationship, arguing the relationship between aspects of internationalization and financial performance varies with different moderating variables (Lu and Beamish, 2004). A couple of potential moderators have already been examined. Lu and Beamish (2004), for instance, studied how the possession of intangible assets could positively moderate the relationship between geographic scope and financial performance. Likewise, Jung, Noh and Chung (2016) studied the influence of labor flexibility on the relationship between the intensity aspect of internationalization and financial performance. Researchers (Bausch and Krist, 2007; Hennart, 2012; Ruigrok, Amann and Wagner, 2007) emphasize the need for studying more factors that might modify the relationship.
8 effects of cultural diversity (Nederveen Pieterse, van Knippenberg and van Diererdonck, 2013). This research stream argues cultural diversity within groups contributes to several advantages, such as more and better knowledge and more perspectives, thereby enhancing the group’s effectiveness. This thesis will examine if the theories concerning cultural diversity within groups are applicable to this context with company boards. Little is known about the moderating influence of cultural diversity within company boards on the relationship between international complexity and financial performance. As a result of cultural diversity among board members, the board’s effectiveness might improve and familiarity with cultural differences might increase. Consequently, it is likely that the effect of international complexity on financial performance is different for MNEs with highly culturally diverse boards and MNEs with few cultural diversity within the board. The research question of this thesis is therefore;
“What is the effect of the MNE’s international complexity on financial performance, and how is this relationship moderated by cultural diversity among board members?”
This thesis makes several contributions to existing literature. This thesis contributes to internationalization literature by examining the unclear relationship between internationalization and financial performance, while distinguishing between different aspects of internationalization and studying a potential moderator. In addition, this thesis contributes to cultural distance literature, as this thesis is one of the first to distinguish between different dimensions of cultural distance with respect to the MNE’s operations. Furthermore, this thesis contributes to diversity literature and, more specifically, cultural diversity literature by studying the effects of cultural diversity among board members. Besides, this thesis makes a practical contribution. Financial performance is seen as the main goal of firms (Fama and Jensen, 1983). Hence, knowledge about how financial performance is influenced is highly convenient for executives. This thesis contributes to that knowledge with respect to how the internationalization portfolio affects financial performance, what to change to achieve higher financial performance and how the board composition affects this.
9
2. THEORY
In this chapter, literature will be discussed and hypotheses will be drawn. First, literature about international complexity, MNE financial performance and cultural diversity among board members will be reviewed. Then, the direct relationships between international complexity and MNE financial performance will be discussed. After that, the moderating influence of cultural diversity among board members will be argued. This chapter will end with a conceptual model where all the relationships are shown.
2.1. International complexity
2.1.1. Geographic scope
Internationalization refers to the expansion of a firm’s sales across national borders into diverse markets or locations (Hitt, Hoskisson and Hicheon, 1997). This thesis will examine two aspects of internationalization. The first aspect is the MNE’s geographic scope. This refers to the size of the MNE’s international presence (Delios and Beamish, 1999), which can be the number of foreign subsidiaries or the number of countries the firm has operations in. In this study, geographic scope means the number of countries the MNE operates in.
There are several motivations for firms to enter a foreign country and hence increase the geographic scope. The motivations can be classified into efficiency seeking, resource seeking and market seeking motivations.
According to the first category of motivations, firms enter foreign markets for efficiency reasons. The desire for improved efficiency could be driven by investors, as they usually long for higher firm performance and growth (Hitt, Bierman, Uhlenbruck and Shimizu, 2006). An argument that is included in the first category is that firms enter foreign markets in order to secure a competitive advantage (Porter, 1990). Caves (1996) adds to this that when a firm enters many foreign countries, it could achieve economies of scale and scope. Moreover, it is said the firm could spread investment risks over multiple countries (Kim, Hwang and Burgers, 1993) and thereby improve its efficiency.
10 The last category of motivations consists of a market seeking motivation. Some firms experience a maturing market in their home country. Consequently, these firms enter foreign countries as they want to enter new markets for their products (Curci, Ling-Yee and Mackoy, 2014).
With regard to the pace of geographic scope expansions, the logic of the Uppsala model (Johanson and Vahlne, 1977) has been highly impactful. The Uppsala model highlights the role risk aversion plays: managers perceive a geographic scope expansion as risky and therefore expand the geographic scope slowly. These perceived risks result from unfamiliarity with the economy, culture and politics of the foreign country (Shrader, Oviatt and McDougall, 2000). Only after experience in foreign environments is gained, and thereby risk is declined, firms will enter new foreign countries (Håkanson and Kappen, 2017; Hashai, 2011; Johanson and Vahlne, 1977). Nevertheless, this model is based on the scope patterns of MNEs in the 1970s. Doz, Santos and Williamson (2001) argue things have changed and that lately, some firms are racing to enter foreign countries in order to stay ahead of competitors.
Regarding the effect of geographic scope on financial performance, prior studies have revealed mixed results. Similar to the influence of internationalization in general, researchers have found all sorts of effects (Goerzen and Beamish, 2003). Researchers finding non-significant effects and negative effects highlight the increased complexity associated with a big geographic scope. As the geographic scope increases, monitoring subsidiaries will get more challenging. Kostova and Zaheer (1999) complement that the risk of not gaining legitimacy in one country due to actions in another country increases. A big geographic scope namely implicates the firm must deal with multiple, diverging regulations and expectations (Miller et al., 2016). In addition, transferring information and coordinating the MNE is more difficult when the geographic scope is large, leading to additional costs (Hitt et al., 1997). This is especially true for tacit knowledge, which can only be transferred by doing (Dyer and Singh, 1998). Acknowledging that a big geographic scope leads to increased complexity, this thesis will nevertheless argue that the advantages of a big geographic scope outweigh the additional complexity.
2.1.2. Cultural distance to MNE’s operations
The second aspect of internationalization that will be examined, is the average cultural distance to the MNE’s operations. Culture has been defined by numerous researchers. One of them is Hofstede (2011: 3) who defines culture as follows: “Culture is the collective programming of
11 Another researcher attempting to define culture is Schwartz (2006), who regards culture as a set of values that are emphasized in a society, and affects how individuals behave. Important overlaps in these definitions are that culture exists at the group-level and influences behaviour. Cultural distance refers to how different cultures are. The average cultural distance to MNE’s operations is the average degree of difference in cultural characteristics between the MNE’s country of origin and the countries where it operates (Gomez-Mejia and Palich, 1997).
There are two main perspectives in measuring these cultural differences. The first perspective is ethnocentrism. According to ethnocentrism, cultures cannot be compared as they are completely unique (Maseland and van Hoorn, 2017). This means that there are no dimensions and every culture should be studied individually. The second perspective assumes that there is some universality within cultures. According to this perspective, there are concepts or problems that every society faces. To compare cultures, idiosyncrasy needs to be identified in how different societies deal with these universal concepts or problems (Maseland and van Hoorn, 2017). Within the second perspective, Hofstede’s (1980) framework is definitely the framework that is mostly used in research in the field of cultural distance. This framework looks at cultural differences at the country-level. However, Hofstede’s framework has been subject to substantial criticism. The main critique is that Hofstede’s framework is outdated (Vaiman and Brewster, 2015). Nevertheless, Beugelsdijk, Maseland and van Hoorn (2015) showed this is not the case as cultural values are relatively stable over time. In addition, researchers argue this framework is the most complete and comprehensive one (Erramilli, 1996; Shane, 1995). Therefore, Hofstede’s framework can still be used to approach distances.
12 Literature about the cultural distance to MNE’s operations usually combines different cultural dimensions into one variable. It is being argued that the total cultural distance of operations will decrease MNE financial performance (Gomez-Mejia and Palich, 1997; Miller et al., 2016; Palich and Gomez-Meija, 1999). The reason for this is that there are some additional costs due to the cultural distance in general. Gomez-Mejia and Palich (1997) argue that cultural distance of operations drive diverging consumer preferences. Customization is required in order to survive abroad and this increases complexity and operating costs. Besides, as the cultural distance to the countries increases, it is more challenging to control and monitor the subsidiaries (Gomez-Mejia and Palich, 1997). The reason for this is that the major differences add complexity. Accordingly, agency costs will rise. Furthermore, an increase in the cultural distance of an MNEs operations drives an increase in Liability of Foreignness (LOF). LOF refers to the unfamiliarity with local cultures and environments (Zaheer and Mosakowski, 1997). LOF negatively influences performance (Barkema and Drogendijk, 2007). However, this study does not combine the dimensions into one variable. This study is one of the first to differentiate between the dimensions when studying the cultural distance to an MNE’s operations. The reason for this is the belief that different dimensions have different causes for influencing performance and therefore influence performance differently. This is in line with Shenkar’s (2001) illusion of equivalence, which implies that researches should not assign every dimension equal weight (Dow, 2017). Similarly, studies performed by Barkema and Vermeulen (1998) and Lim et al. (2016) found that the effects of the different dimensions vary across contexts. Hence, a distinction between dimensions will be made. The individual dimensions will be described next.
2.1.2.1. Hofstede’s dimensions
Hofstede’s first cultural dimension, individualism-collectivism, refers to the importance of groups within a society (Hofstede, 2011). In individualistic cultures, people should only take care of themselves and their closest family. Important aspects are that people are allowed to have a personal opinion and that tasks are prioritized above relationships. In collectivistic cultures, on the contrary, people are highly integrated into groups, which protect them. Personal opinions are not allowed and relationships are more important than tasks (Hofstede, 2011).
13 Consultation with subordinates is expected, instead of commanding what subordinates need to do (Hofstede, 2011).
Hofstede’s third dimension, masculinity-femininity, contains opposites, just as individualism and collectivism (Hofstede, 2011). According to this source, the dimension refers to if women and men’s roles are different or the same; the inequality between male and female. There is a role differentiation between males and females, both emotionally and socially, in masculine societies. Males should fight and decide on family size and politics, whereas women should not fight but cry, and deal with feelings. That role differentiation is minimum in feminine societies. Values as ambition and assertiveness override in masculine societies, whereas values as quality of life, empathy and compassion predominate in feminine societies (Hofstede, 2001). Hofstede’s fourth dimension, uncertainty avoidance, refers to the forbearance of uncertainty and ambiguity (Hofstede, 2011). Societies with high scores on uncertainty avoidance feel uncomfortable in unknown situations and therefore try to minimize such situations by implementing specific rules and behavioral codes. Societies with low scores on uncertainty avoidance, on the contrary, do not feel threatened by unknown situations.
2.2. Financial performance
The importance of financial performance is highlighted by the many studies interested in factors influencing financial performance (e.g. Akisik and Gal, 2017; Post and Byron, 2015; Weichieh and Tsang, 2015). This is in accordance with the mainstream perspective on firm performance, which views firms as rational agents that are principally occupied with finance (Fama and Jensen, 1983). Although firms pursue other goals as well, such as for instance reputation improvement (Goerzen and Beamish, 2003), it must be acknowledged that financial performance is critical for firms.
Post and Byron (2015) mention financial performance is multi-dimensional. One aspect is named accounting returns. This dimension can be defined as follows: “Accounting returns
refers to how well a firm utilizes its assets and investments to generate earnings and represents past or short-term financial performance.” (Post and Byron, 2015: 1547). Market performance
is another aspect of financial performance. This can be defined as “The behavior of a security
or asset in the marketplace, reflecting external perceptions and expectations of a firm’s future or long-term value.” (Post and Byron, 2015: 1547). This study will use the first dimension. The
definition of Meyer, Mudambi and Narula (2011: 241) will be employed to define an MNE:
14
across national borders”. Hence, this thesis is about how well firms, that operate across
national borders, utilize their assets to generate earnings.
There are several manners to measure financial performance. What can be seen in the conceptual definition of market performance-based measures, is that these measures are future-oriented (Meyer, 1994). Previous studies have used Tobin’s Q as a measure for market-based performance (e.g. Lu and Beamish, 2004). An advantage of this measure is that it adjusts for market risk (Farjoun, 1998). Accounting performance-based measures, on the contrary, are based on past information (Meyer, 1994). Return on assets (ROA) is undoubtedly used the most in internationalization literature to measure accounting performance (Kirca et al., 2011). Measures based on accounting performance have received criticism for their short-term view (Aaker and Jacobson, 1987). However, multiple researchers (e.g. Lu and Beamish, 2004) found similar results for Tobin’s Q and return on assets, implying these measures are both appropriate measures for financial performance. Less popular measures in internationalization literature for measuring financial performance are return on equity (ROE) and return on sales (ROS). The measure that is most appropriate for this study will be discussed in section 3.3.1.
2.3. Cultural diversity within the company board
A board is “a group of diverse individuals who have different biases and prejudices and whose
behavior is affected by social constraints and power relations” (Ferreira, 2010: 225). As
mentioned before, company boards are an extremely important governance mechanism. Company boards have several responsibilities, including determining the organization’s strategy, setting rules, setting direction and deciding on the budget (Nueno, 2016). In other words, the board is responsible for the company as a whole. It is the company board who decides if the MNE will internationalize, where to and how. It makes, therefore, sense that certain company board characteristics could influence the degree to which international complexity affects financial performance. This line of reasoning is supported by a study performed by Hsu, Chen and Cheng (2013) that showed characteristics of decision makers, such as age, experience and educational level, influence the internationalization-performance relationship. Since research (e.g. Ferreira, 2010) has shown that cultural diversity within teams affects the effectiveness of team processes and subsequently team outputs, the board’s effectiveness may as well be influenced by its degree of cultural diversity.
15 This source subsequently refers to cultural diversity as the presence of multiple diverse cultural identities. Similarly to the independent variables about the cultural distance to the MNE’s operations, this study will differentiate between cultures based on nations. This implies that the cultural diversity of the company board will reflect the existence of diverse nationalities.
Estélyi and Nisar (2016) examined motivations for a culturally diverse company board. This source found shareholder variety as one of the main motivations. Shareholders prefer board members that share their interests. Heterogeneity among shareholders results in heterogeneity in interests, which can only be covered by several individuals with different backgrounds. Another main motivation for a culturally diverse board is product market variety. In case of major product market variety, a firm might benefit from a culturally diverse board as this implies connections to numerous people and networks.
Many researchers have studied the effects of cultural diversity. Ferreira (2010) for instance found that cultural diversity leads to creativity. The reason for this is that diverse people have different mental models, which is viewed as an input for creativity (Stahl, Maznevski, Voigt and Jonsen, 2010). Bjornskov (2008) argues cultural diversity within teams constrains the building of trust. As people trust within-group members and distrust outside-group members, trust is low within high culturally diverse teams
16
2.4. Hypothesis development
2.4.1. Geographic scope and MNE financial performance
This thesis argues that as the MNE’s geographic scope increases, and hence the MNE enters more countries, financial performance would increase. This relationship works in two ways, namely by lowering operating costs and increasing revenues.
First, the more countries the MNE enters, the more opportunities the MNE has to lower its operating costs. The MNE could, for instance, achieve economies of scale and scope. In addition, the MNE is better able to assign value-adding activities to the locations where that activity can be done the cheapest (Ruigrok, Amann and Wagner, 2007). Moreover, the MNE would have more opportunities to acquire the latest technology (Luo and Tung, 2007), which can increase efficiency and effectiveness. Finally, the bigger the MNE’s geographic scope, the more power the MNE will have (Kogut, 1985). Consequently, the MNE can lower the input prices of its suppliers.
Second, the MNE’s geographic scope can increase financial performance as it drives higher revenues. As the MNE enters more countries, the market for its products is broadened (Buckley, 1988; Jung, Noh and Chung, 2016). Further, according to Mitchell and Coles (2003), a higher geographic scope will enhance innovation, which will ultimately lead to higher revenues.
For these reasons, I hypothesize:
Hypothesis 1: A bigger geographic scope leads to higher MNE financial performance. 2.4.2. Individualism-collectivism distance and MNE financial performance
In addition, this thesis argues that the cultural distance to the MNE’s operations negatively influences financial performance.
17 Cohen and Levinthal (1990), people can only effectively communicate and learn from each other if there is a partially overlapping knowledge base. Consequently, if the ways of working are completely different people do not understand each other. Knowledge cannot be transferred efficiently and communication will be inefficient as well. Conflicts, miscommunication and inefficient knowledge transfers increase coordination costs (de Jong and van Houten, 2014).
Furthermore, differences in customer needs and preferences are huge as the individualism-collectivism dimension is a crucial determinant for innovation adoption (Flight, Allaway, Kim and Souza, 2011). This is supported by the large research stream studying the effects of individualism-collectivism on purchasing behavior (Faqih and Jaradat, 2015). Customers from countries with high scores on individualism adopt innovation more easily as these customers want to express their uniqueness (Kwok and Uncles, 2005). This implies individualistic customers seek constantly new and different products, whereas collectivistic customers seek traditional products. As the individualism-collectivism distance increases, complexity rises as the MNE is likely to be unfamiliar with the different customer preferences. According to Agency Theory, this will provide local managers with an information advantage. Agency problems arise and headquarters will exert more control over subsidiaries (O’Donnell, 2000). Consequently, agency costs will rise.
In sum, as the individualism-collectivism distance to an MNE’s operations increases, coordination costs and agency costs will rise and complexity will enlarge due to different ways of working and customer’s speed of adoption. Consequently, financial performance will decline. For these reasons, I hypothesize:
Hypothesis 2: Greater distance to the MNE’s foreign operations in terms of individualism-collectivism leads to lower MNE financial performance.
2.4.3. Power distance and MNE financial performance
18 inefficient knowledge transfer is likely due to mismatched leadership styles. Conflicts and inefficient knowledge transfer increase coordination costs (de Jong and van Houten, 2014).
In addition, there will be differences with regard to the decision making process, as high power distance countries have centralized decision making processes and low power distance countries are rather participatory. Brockner et al. (2001) found that employees in high power distance countries accept that they have little voice in decision making. In low power distance countries, employees expect a bigger voice. As a result, decision-making procedures differ. This could result in higher decision wait times, which increases coordination costs (de Jong and van Houten, 2014).
Moreover, different value propositions are preferred by customers in high and low power distant countries (Zhang, Winterich and Mittal, 2010). Customers in high power distance countries are hierarchical and buy products that match their social status (Thompson and Chmura, 2015). In addition, these customers are likely to repeat previous value proposition choices (Zhang et al., 2010) and are brand loyal (Palumbo and Herbig, 2000). For customers in low power distant countries, the opposite is true. As a result, an MNE that operates in countries that are highly distant in terms of power distance experiences complexity and agency costs, as the MNE’s headquarters is not familiar with the locally preferred value propositions.
From the section above it can be concluded that MNEs operating in highly distant countries with regard to power distance experiences complexity, coordination costs and agency costs due to diverging leadership styles, decision-making processes and value propositions. This will lower financial performance. Therefore, I hypothesize:
Hypothesis 3: Greater distance to the MNE’s foreign operations in terms of power distance leads to lower MNE financial performance.
2.4.4. Masculinity-femininity distance and MNE financial performance
19 In addition, customer preferences differ, leading to additional complexity and agency costs. Customers in masculine countries want products with which they can brag. Besides, these customers are more materialistic (Steenkamp, ter Hofstede and Wedel, 1999). Consequently, these customers buy expensive products from brands with exclusive reputations. Customers in feminine countries are less materialistic (Steenkamp et al., 1999). As a result, these customers are more likely to buy primary products, which they actually need. Consequently, if the masculinity-femininity distance between an MNE and the countries it operates in is high, this mandates adaptations to multiple, unfamiliar, local conditions, leading to complexity and agency costs.
In a nutshell, MNEs operating in countries that are distant in terms of masculinity and femininity experience more conflict and coordinating costs due to different working environment preferences. Besides, customer preferences diverge, creating additional complexity and giving rise to agency costs. Therefore I hypothesize:
Hypothesis 4: Greater distance to the MNE’s foreign operations in terms of masculinity-femininity leads to lower MNE financial performance.
2.4.5. Uncertainty avoidance distance and MNE financial performance
An MNE that operates in culturally distant countries in terms of uncertainty avoidance will experience diverging decision-making processes. Employees in countries that are highly uncertainty avoidant will spend time gathering all relevant information and considering every alternative properly, as they seek to minimize risk (Steenkamp et al., 1999). Decisions will be made centralized and decisions that provide certainty are likely to be chosen. Countries with low scores on uncertainty avoidance will behave in the opposite way. These diverging processes might result in higher decision wait times, which increases coordination costs.
Moreover, the way of working is likely to differ as well, as strong uncertainty avoidance countries have strict rules and procedures, and work in a formal way (Ollo-López and Goni-Legaz, 2015). Again, for countries that score low on uncertainty avoidance, the opposite is true. This could lead to conflicts, miscommunication and inefficient knowledge transfers since there is no common knowledge base. These events contribute to coordination costs.
20 risk-taking. They will easily switch brands (Roth, 1995) and hence their choice to buy a product will depend on other things, such as for instance price. Consequently, when the uncertainty avoidance distance to an MNE’s operations increases, the MNE will have to deal with diverging customers. This leads to complexity, and agency costs as the MNE is unfamiliar with local customers.
In sum, uncertainty avoidance distance to an MNE’s operations results in complexity, coordination costs and agency costs, due to diverging decision-making processes, diverging ways of working and variety in risk aversion of customers. Consequently, as the distance increases, financial performance declines. For these reasons, I hypothesize:
Hypothesis 5: Greater distance to the MNE’s foreign operations in terms of uncertainty avoidance leads to lower MNE financial performance.
2.4.6. The influence of cultural diversity within the company board on the ‘geographic scope-financial performance’ relationship
Shenkar (2001) is one of the first researchers recognizing cultural differences might be advantageous. Shenkar (2001) labels this the illusion of discordance, which denotes cultural differences are not permanently a negative matter. Conversely, cultural diversity within the company board may have a positive influence on the relationship between geographic scope and MNE’s financial performance, as it improves board effectiveness.
21 From the section above, it can be concluded that board member cultural diversity may have a positive moderating influence on the relationship between the geographic scope of the MNE and its financial performance. Therefore, I hypothesize:
Hypothesis 6: More cultural diversity within the company board increases the positive effect of the MNE’s geographic scope on MNE financial performance.
2.4.7. The influence of cultural diversity within the company board on the ‘cultural distance to MNE’s operations-financial performance’ relationships
In addition, cultural diversity among board members might weaken the negative relationships between the cultural distances to an MNE’s operations and financial performance. The reason for this is that the presence of multiple cultural identities in the company board enhances familiarity with foreign cultures and environments, and provides experience with handling cultural differences. Ferreira (2010) state that when the MNE enters a country and has a board member originating from that country, the MNE can make use of the board member’s familiarity and connections in the country to ensure the company runs smoothly. A study performed by Davidson (1980) provides empirical support for the claim that familiarity affects the cost of internationalization. In addition, Lim et al. (2016) confirm familiarity reduces LOF, and hence decreases LOF costs. Besides, when the company board is familiar with local customer preferences, the information advantage of the local subsidiary diminishes. Consequently, agency costs will reduce (O’Donnell, 2000). Furthermore, with knowledge about many different countries, cultures and processes, the common knowledge base increases. As a result, communication and knowledge transfers will be more effective (Cohen and Levinthal, 1990). Moreover, this knowledge makes managing and coordinating the MNE less complex (Anderson, Reeb, Upadhyay and Zhao, 2011). As a result, coordination costs decrease. Finally, experience with handling cultural differences will enhance the board’s capability of coping with cultural differences and therefore conflicts will appear less.
22
Hypothesis 7: More cultural diversity within the company board decreases the negative effect of the operational cultural distance in terms of individualism-collectivism on MNE financial performance.
For the second dimension, power distance, this means that cultural diversity within the company board ensures familiarity with other cultures, leading to knowledge about which leadership style is effective. Consequently, conflict will appear less and knowledge transfers are more efficient if the leadership styles are matched with culture. Furthermore, cultural diversity among board members ensures familiarity with customer preferences. Agency costs and complexity are reduced. For these reasons, I hypothesize:
Hypothesis 8: More cultural diversity within the company board decreases the negative effect of the operational cultural distance in terms of power distance on MNE financial performance.
For the third dimension, masculinity-femininity, this implies cultural diversity within the company board provides familiarity with the preferred working environments and experience with cultural differences, thereby reducing the potential for conflict. Again, it also ensures familiarity with customer preferences, reducing agency costs and complexity. Therefore, I hypothesize:
Hypothesis 9: More cultural diversity within the company board decreases the negative effect of the operational cultural distance in terms of masculinity-femininity on MNE financial performance.
For the fourth dimension, uncertainty avoidance, this implies cultural diversity among board members ensures familiarity with the diverging ways of working and an increased common knowledge base, reducing miscommunication and inefficient knowledge transfers. Besides, familiarity with customer’s risk aversion reduces agency costs and complexity. Therefore, I hypothesize:
23
2.5. Conceptual model
Figure 1 shows a conceptual model to make the relations more visible.
FIGURE 1 Conceptual Model
3. METHODOLOGY
24
3.1. Target population
A study performed by Palmer and Varner (2007) showed that the boards of European MNEs have a higher share of international directors than for instance the boards of MNEs from the US. Europe is, therefore, an attractive area to test the hypotheses. In addition, studying European MNEs fulfills the need to study the internationalization-financial performance relationship in Europe (Buckley and Ghauri, 2004). Accordingly, the target population of this study includes all listed European MNEs. The MNEs needed to be listed, as annual reports were to be analyzed.
Orbis database was used to seize the target population. In Orbis database, there has been filtered on ‘Eastern Europe and Western Europe’, ‘Very large companies’, ‘Publicly listed companies’ and ‘With foreign subsidiary’. Orbis classifies firms as very large when firm revenue exceeds €100 million (de Jong and van Houten, 2014). The resulting target population amounts 4.843 MNEs.
3.2. Sample
This study made use of probability sampling. Probability sampling implicates each MNE in the population has a chance to be selected (Blumberg, Cooper, Schindler, 2014). For this method was chosen as it improves the precision of the estimates. Within probability sampling, the method used to select the sample was random sampling. Following this method, the target population is assigned numbers and random numbers are picked from the target population (Van der Pligt and Blankers, 2013). This implies each MNE has an equal chance to be selected (Blumberg et al., 2014). This method was chosen, as the sample resulting from this sampling method is highly representable for the population.
25
3.3. Measures
3.3.1. Dependent variable: MNE’s financial performance
Return on assets is commonly used to measure financial performance in internationalization literature (e.g. Cuervo-Cazurra et al., 2018; Lu and Beamish, 2004). Return on assets is therefore an attractive measure, as it facilitates comparisons with earlier studies. This measure is highly appropriate, contrary to for instance Tobin’s Q, which is not only affected by internationalization, but also by future prospects and earnings surprises (Ball and Kothari, 1991). Return on assets is also more suitable than return on equity, as this measure is easily influenced by a firm’s capital structure (Hitt et al., 1997). Return on sales is occasionally used as well in internationalization literature to measure financial performance. However, Chao and Kumar (2010) claim return on assets and return on sales correlate strongly and therefore it makes no difference which one of the two is employed. Hence, the operational definition for MNE’s financial performance in this study is: Return on assets in the year 2016. Return on assets is calculated by dividing the MNE’s net income by the total value of the MNE’s end of period assets (Berry and Kaul, 2016) (see formula 1) and indicates how capable a firm is in employing its assets for the generation of incomes. The outcome is a percentage, which can take any value, both positive and negative. Financial performance is therefore a continuous variable. Concerning measurement scale, this variable is a ratio variable as distance between values is meaningful (Blumberg et al., 2014).
Formula (1) ROA = Net income / End of period assets
3.3.2. Independent variable: Geographic scope
In order to measure the geographic scope of the MNE’s operations, the number of countries the MNE operates in was counted. As mentioned before, previous research measured geographic scope either as the number of foreign subsidiaries (e.g. Sambharya, 1995) or the number of countries the MNE operates in (e.g. Lu and Beamish, 2004). Tallman and Li (1996) argue that a country count is more accurate to address an MNE’s geographic scope, as most argumentations of geographic scope leading to an advantage address country matters such as economic arbitrage and tax. Hence, geographic scope will be measured as: Number of countries
the firm operates in, in the year 2016. This variable is a discrete variable that can take any
26 foreign subsidiary. The maximum geographic scope was 196 as this was the number of countries in the world in 2016. Concerning measurement scale, this variable is a ratio variable.
3.3.3. Independent variables: Cultural distance to an MNE’s operations in terms of IDV, PD, MAS and UA
The cultural distance to an MNE’s operations was measured based on national cultures. The presumption culture exists at the country level is mainly based on two reasons. First, a shared identity is built by state borders. Second, states behave as vigours that homogenize citizens due to for instance regulations (Maseland and van Hoorn, 2017). Moreover, prevalent approaches for measuring cultural differences use culture at the country level as well (e.g. Hofstede, 1980; Schwartz, 2006).
Several manners for measuring the average cultural distance to an MNE’s operations were evaluated. The evaluations can be found in Appendix B. Eventually, a formula similar to Gomez-Mejia and Palich (1997) and Lavie and Miller (2008) was used.
The formula was slightly adapted. This was required to be able to measure the distance of Hofstede’s individual dimensions, instead of measuring cultural distance in general. In addition, the part about variance was removed since information about the variance of dimensions was not available. The formula that was used is formula 2.
Formula (2) CDd = ∑𝑐=𝑁𝑐=1[Sc − So] /N
Whereas:
CDd: cultural distance of dimension d
Sc: score on dimension d of country c
So: score on dimension d of MNE’s country of origin
N: number of countries the MNE operates in
This formula implied that for every country the MNE operates in in 2016, the absolute difference was calculated between the MNEs headquarters’ country score on the dimension and that particular country’s score on the dimension. All these individual distances were added up and divided by the total number of countries the MNE operates in.
27 Average distance between the MNE’s country of origin and the countries it operates in
in 2016, in terms of individualism-collectivism
Average distance between the MNE’s country of origin and the countries it operates in
in 2016, in terms of power distance
Average distance between the MNE’s country of origin and the countries it operates in
in 2016, in terms of masculinity-femininity
Average distance between the MNE’s country of origin and the countries it operates in
in 2016, in terms of uncertainty avoidance
Hofstede’s scores range between zero and one hundred. Consequently, the outcomes of this formula could range between zero and one hundred as well. Hence, this variable was measured as a continuous variable that could take any value between zero and one hundred. A score close to zero implies the cultural distance to operations for the respective dimension is small, whereas scores closer to one hundred imply large cultural differences between the MNE’s headquarters and its operations, with respect to that particular dimension. The variables are ratio variables.
3.3.4. Moderator: Cultural diversity among board members
Similar to the independent variables above, cultures are measured at the country level as well for the moderator. Blau’s (1977) index is commonly used to measure variety (Konrad, Prasad and Pringle, 2006). Richard, Barnett, Dwyer and Chadwick (2004) used it for instance to measure race and gender variety. Consequently, Blau’s index was employed to measure cultural diversity among board members. The index is presented in formula 3.
Formula (3) 𝐵 = 1 − ∑ki=1Pi2
Whereas:
B: cultural diversity within company board P: proportion of individuals in a category i: number of categories
28 values could take any value between zero and one, this variable is measured as a continuous variable. With regard to measurement scale, this variable is a ratio variable as well.
Consequently, the measure used was: Blau’s index, reflecting the degree of cultural
diversity within the 2016 board. Noticeable is that this index captures the extent to which
diverse nationalities exist on the company board, and not how different the members are. This matches the conceptual definition of cultural diversity.
3.4. Control variables
Finally, some control variables were added to the model. These are explained in this section. Firm size might influence MNE financial performance as larger firms can attain economies of scale and thereby enhance their financial performance. Furthermore, large firms have better access to capital (Park, Li and Tse, 2006). In order to draw correct inferences about the other relationships, it is important to take this control variable’s effect into account. The same operational definition as Goerzen and Beamish (2005) and Geringer et al. (2000) used was employed to measure firm size: Number of employees in the year 2016. This measure could take any integer value bigger than zero. Hence this variable is a discrete variable. Besides, this variable is a ratio variable.
Furthermore, the MNE’s age might influence financial performance. As time passes by, knowledge and experience accumulate and this will feasibly influence MNE performance (Murphy, Trailer and Hill, 1996). Following previous research (e.g. Cuervo-Cazurra et al., 2018; Jung, Noh and Chung, 2016) MNE age was measured as follows: Difference in years
between the MNE’s year of establishment and 2016. This variable was measured as a discrete
variable that could take any integer value from zero. Besides, this variable is a ratio variable. Finally, since prior research showed that the type of industry the firm operates in affects firm performance (Schmalensee, 1985), this variable was also included in the analysis as a control variable. The same industry sectors were used as Cuervo-Cazurra et al. (2018) identified (see Appendix A). As a result, industry sector was measured as: Categories reflecting the type
of industry the MNE operates in. This implied MNEs occupied with for instance wind energy
29 categories. Since the numbers simply represent the categories, the measurement scale of this variable is called nominal (Blumberg et al., 2014).
3.5. Data collection
This study made use of secondary data, namely annual reports of the year 2016 and Hofstede’s country scores on the cultural dimensions. The use of secondary data provides several advantages. It requires, for instance, fewer resources, it can provide new insights, it is convenient and data is permanent and available (Saunders, Lewis, Thornhill, Booij and Verckens, 2011). A disadvantage can be that the researcher has no actual control over the data and eliminating non-relevant data is time-consuming. However, usage of secondary data is for this study most appropriate as getting access to primary data is difficult. Approaching managers is rather challenging.
For information about geographic scope, financial performance, cultural diversity among board members, firm size, firm age and industry type, annual reports of the year 2016 were analyzed. Consulting annual reports is regularly done for variables such as financial performance (Hoisl, Gruber and Conti, 2017).
For the variables about the cultural distance to an MNE’s operations, annual reports of the year 2016 were analyzed to find out every country the MNE operates in. Subsequently, Hofstede’s country scores on the dimensions were used to measure distances. Hofstede’s country scores were on average available for 95,57% of the countries an MNE operates in. Moreover, for 71% of the MNEs in the sample, Hofstede’s scores covered all the countries the respective MNE operates in. Due to the lack of information available for the other 4,43% of the countries an MNE operates in, these countries were not included in the calculation. Hofstede’s country scores can be found at his website:
https://www.hofstede-insights.com/product/compare-countries/.
3.6. Data analysis
30 Irwin and McClelland, 2001). Moreover, this method is commonly used in similar studies (e.g. Frijns et al., 2016; Hoisl et al., 2017; Rafiq, Salim and Smyth, 2016). Hence, MMR was highly appropriate for this study. MMR requires the respective variables to be mean centered (Aiken and West, 1991). Consequently, new mean centered variables were computed for the five independent variables and moderator. In addition, MMR mandates the formation of interaction variables (Aiken and West, 1991). As a result, five interaction variables consisting of the product of the mean centered cultural diversity among board members and each of the mean centered independent variables were computed. After that, regressions were done with the centered independent variables, the centered moderator and the interaction variables, on the dependent variable financial performance.
Since regressions can only be run with variables measured on ratio level and dummy variables, the control variable industry type required the creation of dummy variables. As a result, dummy variables were made for the first three categories where a value of one was adopted if the MNE operates in that industry. Otherwise, a value of zero was adopted.
3.7. Missing data
In case data about one of the variables was missing, the MNE was still included in the analysis of other hypotheses, a method known as pairwise deletion (Fife, Mendoza and Berry, 2017). For this method is chosen as valuable data is maintained (Enders, 2010).
3.8. Robustness checks
31 sample produces the same results. This implies the same regression analyses are conducted, however with only a random portion of the sample. Consequently, the sensitivity of the findings is revealed.
4. DATA
This chapter starts with describing the basic tests that were performed. This was done to check if the assumptions for doing regression analyses were met. Thereafter, outliers are discussed. The chapter ends with a description of the sample and the collected data by using descriptive statistics.
4.1. Testing basic assumptions
Basic assumptions for regressions were tested before conducting regressions. The first assumption that was tested was the assumption of linearity and homoscedasticity. A scatterplot was made with the standardized residuals on the Y-axis and the standardized expected values on the X-axis. The scatterplot showed randomly distributed dots. As recognition of any pattern in the dots implies the relationship is not linear (Siero, Huisman and Kiers, 2009), the assumption of linearity and homoscedasticity is met.
The second basic assumption that was tested is the assumption of normality. This assumption can be tested by the creation of a histogram with the standardized residuals (Siero et al., 2009) where the assumption is met if distribution appears normal. Subsequently, a histogram was made with the standardized residuals. The histogram showed an approximately normal distribution, as can be seen in Appendix C figure A.1. Hence, the assumption of normality is met.
The third assumption that must be met for doing regressions is the assumption of independent errors. However, since this study does not make use of time series, it is unlikely that the errors are dependent. Hence, the assumption of independent errors is met.
32
4.2. Outliers
An outlier was noticed during data analysis. More specifically, one element had an extraordinary low score on age and high score on the number of employees. Possibly this MNE recently merged and is therefore so young. This would mean that this element’s age is not representable for the knowledge it possesses and hence does not correspond with the logic of the argument made in the methodology section. Therefore, this element’s age is not included in the analysis to ensure this value does not distort outcomes.
4.3. Descriptive Statistics
The sample consisted of 100 MNEs, which is sufficient as the minimum sample size for reliable results in this study amounts to 90 MNEs. Most of the MNEs in the sample operate in the manufacturing industry (38%), followed by the service industry (24%), “other” industries (22%), and finally the natural resource-based industry (16%). This distribution is shown in Appendix C figure A.2. The MNEs in the sample have their headquarters in nineteen different European countries. These are displayed in Appendix C table A.1, which shows that most MNEs are headquartered in Germany (14%) and the United Kingdom (14%).
41% of the MNEs that were included in the sample had a company board with no cultural diversity (see Appendix C figure A.3). This is about equal to what other researchers found (Frijns et al., 2016). The average cultural diversity index among board members is 0,2323 (SD=0,23434). That number is rather low, since the index could range between zero and one. The number has especially sunk by the 41% of MNEs with no cultural diversity within the board and hence had a score of zero on this measure. When the MNEs with no cultural diversity within the company board were excluded from the computation, the average cultural diversity index within boards increased to an amount of 0,3995 (SD=0,16481), which is an increase of 71,98%. In other words, only taking the boards with cultural diversity into account, the average cultural diversity index amounts 0,3995.
33 countries, with a major standard deviation of 12,512 which is due to the large variety. These and other statistics can be found in table 1.
TABLE 1 Statistics variables Mean Standard Deviation Minimum Maximum Firm size 9.160,01 34.959,163 20 302.562 Firm age 53,19 49,649 6 270 ROA 0,596 2,47 -47,69 53,19 Cultural diversity board 0,2323 0,23434 0 0,75 Geographic scope 11,41 12,512 2 89 IDV 20,695 10,2695 2 57,3 PD 20,057 8,7402 2 49,7 MAS 21,913 13,3086 2 61 UA 22,434 9,4965 0 59
Finally, a correlation analysis was done. The results are displayed in table 2. Noticeable is that some of the average operational cultural distances correlate with each other (namely average operational power distance with average operational individualism distance, r=0,263, p=0,009; average operational power distance with average operational uncertainty avoidance distance, r=0,318, p=0,001; and average operational individualism distance with average operational masculinity distance, r=0,218, p=0,031). With solely positive r-values, this implies high scores on one cultural distance dimension goes hand in hand with high scores on another cultural distance dimension.
35
5. RESULTS
This chapter presents the results. First, the results of the analyses are described. Then, the performed additional tests are described.
5.1. Baseline results
The first five hypotheses are direct effects with ratio variables. Therefore, a simple linear regression was used to analyze the hypotheses. The hypotheses were confirmed if the p-value was below 0.05. The formula that represents a regression equation is formula 4.
Formula (4) Y = BX + c
Where B is the coefficient of the X-variable and c is a constant value (Montgomery et al., 2012). This implies that if X is zero, Y is the value of c.
Hypotheses 6 - 10 include a moderating variable. As mentioned in the methodology chapter, these hypotheses will be analyzed with a multiple moderated regression. Testing a moderation effect requires two steps. In the first step, a regression model is made, in block 1, predicting the dependent variable based on the moderator and independent variable. For a moderation effect to occur, both the model as well as the two direct effects need to be significant (Aiken and West, 1991). After that, the interaction variable is added to the model, in block 2. If the interaction variable is significant this means that moderation occurs (Baron and Kenny, 1986; Hayes, 2018). The formula for the moderation effect is formula 5.
Formula (5) Y = B0 + B1X + B2Z + B3XZ
Where B1 represents the effect of the independent variable on the dependent variable, B2
represents the direct effect of the moderating variable on the dependent variable, and B3
represents the effect of the interaction variable on the dependent variable (Aiken and West, 1991). The intercept in the equation is represented by B0.
36 A positive significant effect is found (B=0,218, t=1,815, p=0,043). Hence, hypothesis 1 is confirmed. A greater geographic scope drives higher financial performance. The formula that represents this relationship is formula 6.
Formula (6) Y = 0,218X -4,118
In model (2) – (5) the other direct effects are considered case by case. Model (2) shows a positive relationship between the average individualism-collectivism distance to MNE’s operations and financial performance. However, this result was not significant. Hence,
Hypothesis 2 cannot be confirmed. Bigger distance to MNE’s operations in terms of
individualism-collectivism does not decrease financial performance.
In model (3), the influence of the average power distance to an MNE’s operations is considered. Again, a positive and non-significant relationship was found, leading to a rejection of hypothesis 3. Bigger distance to MNE’s operations in terms of power distance does not decrease financial performance.
Hypothesis 4 is regarded in model (4). This model shows a negative relationship between the average masculinity distance to an MNE’s operations and financial performance. This relationship is however weakly significant (B=-0,175, t=-1,604, p=0,10). Currently, geographic scope is a strong predictor of MNE financial performance and the average masculinity distance to foreign operations a negative predictor. This raises the question if possibly these two variables behave together. Hence, in model (5), return on assets was regressed on both the average masculinity-femininity distance to an MNE’s operations and the number of countries the MNE operates in. Results show a significant negative relationship between the average masculinity-femininity distance to an MNE’s operations and return on assets (B=-0,184, t=-1,714, p=0,048). In addition, the positive relationship between geographic scope and return on assets is found (B=0,227, t=1,912, p=0,035). This implies that a smaller masculinity-femininity distance between headquarters and subsidiaries and a bigger geographic scope lead to higher financial performance. Consequently, hypothesis 4 is supported. The formula that represents this relationship is formula 7.
37
TABLE 4
Baseline regression results
* p<0,10 ** p<0,05 *** p<0,01
Standard errors are presented by the number in parentheses. In model 7-11, only the results of the regression block that includes the interaction variable are shown.
38 Model (6) considers the direct effect of the average uncertainty avoidance distance to an MNE’s operations and financial performance. This effect is negative, however, not statistically significant. Hypothesis 5 must, therefore, be rejected. A bigger uncertainty avoidance distance to MNE’s operations does not decrease financial performance.
In model (7), the first moderating influence is considered. First, the direct effect of board member cultural diversity was added to the variables included in model (1). Subsequently, the interaction term regarding board member cultural diversity and geographic scope was added. These results show geographic scope remains a factor influencing financial performance (B=0,293, t=1,811, p=0,074), nevertheless is now weakly significant. However, neither a direct effect of board member cultural diversity on financial performance nor a significant model was found. Moreover, a positive moderating influence of board member cultural diversity was not found. Instead, the interaction term’s effect appeared negative and non-significant. Therefore, it can be concluded board member cultural diversity does not positively moderate the relationship between geographic scope and financial performance. Hypothesis 6 cannot be confirmed.
Models (8), (9), (10) and (11) consider the other relationships and adjunct interaction variables case by case. None of the direct and interaction effects is statistically significant. The interaction variables’ values are both positive (e.g. the average individualism-collectivism distance) and negative (e.g. the average power distance). Hence, no conclusions can be made regarding the direction of the influence of cultural diversity within company boards. Since none of the interaction variables’ effect is significant, hypothesis 7, hypothesis 8, hypothesis 9 and
hypothesis 10 must be rejected. Cultural diversity among board members does not moderate the
relationship between average cultural distance to MNE’s operations and financial performance.
5.2. Additional tests
39 p<0,05, whereas the second model was significant at p<0,01. The R2 change was significant
(∆R2=0,031, F(2,95)=4,652, p=0,038). Therefore, it can be concluded that the simple model is
correct, nevertheless, the more complex model fits the data better and is statistically significantly better than the first one.
Besides, a couple of tests were done to test the robustness of the results. Since a disproportionate share of MNEs in the sample operate in the manufacturing industry, regression analyses were run with only MNEs that do not operate in the manufacturing industry. The conclusions were nearly similar to earlier conclusions. In general, B-values became greater and p-values became a little higher. The most radical change was that geographic scope was, individually, no longer a predictor of financial performance. When regression analyses were conducted with only the MNEs that operate in the manufacturing industry, no original predictor remained significant. Still, no moderating influence of cultural diversity among board members is found. Nevertheless, it must be acknowledged that the number of observations in the latter regressions is rather small. Hence, these results might not be reliable.
After that, predictive validity was tested. This was done by running regressions with solely 75% of the MNEs. A major change was that the effect of the average masculinity distance to MNE’s operations was no longer significant, It can be concluded that the results are to some extent robust.
6. DISCUSSION
This thesis’ aim was to contribute to the debate about the relationship between internationalization and financial performance, by examining the influence of international complexity on financial performance. Two aspects were studied; the firm’s geographic scope and average cultural distance to MNE’s operations. The latter was subdivided into four variables each consisting of a cultural dimension. In addition, researchers called for research studying moderators on this relationship (Bausch and Krist, 2007; Hennart, 2012; Ruigrok, Amann and Wagner, 2007). Consequently, the moderating influence of cultural diversity among board members was examined. After conducting several analyses in SPSS it can be concluded that the results of the analyses to some extent fit the expectations.
