Do top management team nationality and age composition affect the association between cultural distance and entry mode choice?

(1)

Do top management team nationality and age composition affect the

association between cultural distance and entry mode choice?

Abstract: This study researched the possible moderating effects of top management team nationality and age composition on the potential relation between cultural distance and entry mode choice. Therefore, this study executed an empirical, cross-sectional, and explanatory research with a deductive approach via quantitative secondary data. The data contained a combination of 202 wholly-owned subsidiaries and joint ventures of 51 Dutch firms in Japan, Nigeria, Brazil, China, Indonesia, India, Russia, Mexico, Turkey, and South Africa between 2010 and 2018. An SPSS logistic regression was applied for the data analysis. The results of this research indicated that firms have a higher likelihood to select joint ventures in host countries when facing contexts of greater cultural distance. The findings further suggested that top management team nationality and age composition do not impact the relationship between cultural distance and entry mode choice, while firms with nationally heterogeneous top management teams are more inclined to choose wholly-owned subsidiaries in host countries.

Keywords: multinational corporations, the Netherlands, entry mode choice, emerging markets, cultural distance, top management team nationality composition, top management team age composition, institutional theory, and upper echelons theory

Master’s Thesis, Master of Science International Business and Management Date of submission: June 14th₂₀₁₉

Author: Nyeem Khan, student number: S3521869

(2)

2

1 Introduction

Entry mode choices for host countries are imperative to firms’ internationalization processes as these choices can be accompanied by fixed and substantial resource dedications (Agarwal, & Ramaswami, 1992). And firms may then cope with additional costs through extra complexities and risks on account of cultural distance between home and host countries (Eden, & Miller, 2004). Accordingly, Kogut and Singh (1988) argued that cultural distance influences entry mode choice. Firms rely on their top management teams (TMTs), composed of executive managers, to make strategic decisions (Canella, Finkelstein, & Hambrick, 2008; Carpenter, Geletkanycz, & Sanders, 2004; Hambrick, & Mason, 1984), such as entry mode choices (Nielsen, & Nielsen, 2011). And Hambrick and Mason (1984) argued that executive managers’ features impact firms’ strategic decisions.

Theoretically, there is an abundance of international business (IB) literature on the plausible relationship between cultural distance and entry mode choice, but there still is a notable call for research on this possible relation as prior studies’ outcomes are contradictory (Beugelsdijk, Kostova, Kunst, Spadafora, & Van Essen, 2018). Numerous studies found that greater cultural distance was related to firms more likely choosing joint ventures (JVs) in host countries (Brouthers, & Brouthers, 2001; Brouthers, & Brouthers, 2003). Conversely, some studies showed that firms were more likely to be in favor of wholly-owned subsidiaries (WOS) in host countries when faced with contexts of greater cultural distance (Brouthers, & Brouthers, 2001; Chen, & Hu, 2002; Gollnhofer, & Turkina, 2015). Few studies indicated no relation (Morschett, Schramm-Klein, & Swoboda, 2010; Tihanyi, Griffith, & Russell, 2005).

(4)

4 Thus, this study applies TMT composition as a moderator on the probable association between cultural distance and entry mode choice. IB and upper echelons studies generally neglected the impact TMT nationality (Heijltjes, Olie, & Glunk, 2003; Kaczmarek, & Ruigrok, 2013; Nielsen, & Nielsen, 2013) and age (Hambrick, & Mason, 1984; Nielsen, 2010a) composition might have on firm results. This is a salient exclusion as the number of foreigners in firms’ upper echelons is increasing (Nielsen, & Nielsen, 2011). Additionally, TMTs are progressively occupied by younger executive managers (Tanikawa, Kim, & Jung, 2017). This shift in firms’ upper echelons presumably affects entry mode choices for culturally distant host countries. Particularly, since national culture may establish variances amidst nationally dissimilar executive managers concerning their interpretation of (cultural) matters and proceeding strategic decision-making (Caligiuri, Lazarova, & Zehetbauer, 2004). Additionally, Herrmann and Datta (2005) suggested that age, for instance, represents executive managers’ risk-taking tendencies and this can be related to cultural distance as cultural distance may generate settings with added risks because of the increased liability of foreignness (LOF) (Eden, & Miller, 2004).

To my knowledge, Nielsen and Nielsen (2011) is one of the few IB studies on the credible influence of TMT (nationality) composition on entry mode choice, also in relation to cultural distance. They displayed that firms with nationally heterogeneous TMTs were more likely to be in favor of JVs in host countries (Nielsen, & Nielsen, 2011). To my knowledge, the upper echelons research by Hitt and Tyler (1991) is one of the few studies related to the probable effect of TMT (age) composition on entry mode choice. Hitt and Tyler (1991) reported that firms with older executive managers had a higher likelihood to select riskier acquisitions.

(5)

5 This leads to the following research question: Do TMT nationality and age composition affect

the association between cultural distance and entry mode choice? This research is motivated

to address this gap via hypothesizing a negative association between cultural distance and entry mode choice, such that firms are more inclined to prefer JVs in host countries when experiencing contexts of greater cultural distance. This research further hypothesizes a positive moderation by TMT nationality and age composition on the above-mentioned hypothesized negative relation. Firms with nationally diverse TMTs and with TMTs with a younger average age are then more likely to be in favor of WOS in host countries when encountering situations with greater cultural distance. These hypotheses are tested via quantitative secondary data. The data contain a combination of 202 WOS and JVs of 51 Dutch firms in Mexico, Indonesia, Nigeria, Turkey (MINT), Brazil, Russia, India, China, South Africa (BRICS), and Japan between 2010 and 2018. An SPSS logistic regression is implemented for the data analysis.

The outcomes of this research imply that firms have a higher likelihood to opt for JVs in host countries when facing contexts of greater cultural distance. The results of this research also suggest that TMT nationality and age composition (in the form of average TMT age) do not impact the relationship between cultural distance and entry mode choice, although firms with nationally heterogeneous TMTs are more likely to prefer WOS in host countries. This study advances the debate in IB literature on the above-mentioned potential relation by indicating that cultural distance is a significant and negative predictor of entry mode choice. This research also enriches IB literature by arguing that TMT nationality composition is a significant and positive predictor of entry mode choice via presenting a deeper insight into firms’ upper echelons by taking entire TMTs into account rather than solely CEOs. Considering the aforementioned shortcomings of firms and the outcomes of this study, this research reinforces the urgency for firms’ executive managers to sufficiently assess cultural distance when making entry mode choices and to not see nationally heterogeneous TMTs as a potential hindrance but as a probable solution when making such choices.

(6)

6

2 Literature review and hypothesis development

2.1 Entry mode choice

The Uppsala model proposes that firms internationalize progressively by expanding resource dedications and choosing host countries based on psychic, such as cultural, distance. Firms may then initially internationalize to, for instance, culturally proximate host countries in relation to the home country and increasingly to culturally distant host countries (Johanson, & Vahlne, 1977). A possible reason for this progressive manner is firms’ absence of understanding of and local stakeholders’ prejudice and lack of local relations in host countries, also known as the LOF. Therefore, Eden and Miller (2004) suggested that the LOF causes extra complexities and risks and, ultimately, costs for firms’ operations in host countries as opposed to local firms’ operations.

Entry mode choice is defined as firms’ strategic internationalization decisions on how to enter host countries (Agarwal, & Ramaswami, 1992) as to expand their operations, such as production or sales, in such locations (Anderson, & Gatignon, 1986). IB studies on entry mode choice seemingly agree that at one end, entry mode choice is distinguished by JVs and at the other end, characterized by WOS (Beugelsdijk et al., 2018; Dikova, & Brouthers, 2015). Hence, this study focuses on JVs and WOS. JVs are defined as firms’ indirect presence in host countries through individual business units via dividing, e.g., investments and ownership with at least one other (foreign) firm. WOS are defined as firms’ direct presence in host countries through subsidiaries that are fully owned and, thus, controlled by firms (Kogut, & Singh, 1988).

(7)

7

2.2 Cultural distance and entry mode choice

Institutional theory suggests that institutions are business regulations in societies (North, 1990). This study is centered on informal, normative, and cognitive institutions. These institutional forms are tacit rules that are collective, such as societies’ norms and values (Scott, 1995). These rules vary per society and shape countries’ culture, where culture is defined as variations among groups of people regarding a shared mental programming. As such, cultures differ per country (Hofstede, 2001).

These cultural dissimilarities between home and host countries can intensify the LOF for firms. Hence, cultural distance can create situations with additional costs via extra complexities and risks. And greater cultural distance subsequently equals further increases in the LOF and the inherent costs through additional risks and complexities (Eden, & Miller, 2004). Thus, Kogut and Singh (1988) asserted that cultural distance impacts entry mode choice. Hill et al. (1990) argued that in situations with cultural distance, WOS are especially exposed as WOS are perceived as rigid due to the relatively larger resource investments. In contrast, it is suggested that JVs are less vulnerable in such situations as JVs have an increased adaptability due to the relatively smaller resource investments (Gatignon, & Anderson, 1988; Kim, & Hwang, 1992).

Numerous IB studies are available on the probable association between cultural distance and entry mode choice. Nevertheless, Kirkman et al. (2006) made an explicit statement about a notable demand for research on this possible relationship due to current studies’ conflicting conclusions. Beugelsdijk et al. (2018) also established this occurrence, showing that study advancements are still missing. Few studies delineated that cultural distance did not influence entry mode choice (Morschett et al., 2010; Tihanyi et al., 2005). Accordingly, different quantifications of institutional distance are plausibly more appropriate in association with entry mode choice, such as demographic and financial distance (Berry, Guillén, & Zhou, 2010).

(8)

8 Via WOS firms can then internalize their overseas activities for more or complete ownership as to more effectively and efficiently control such activities (Anderson, & Gatignon, 1988; Chen, & Hu, 2002; Tihanyi et al., 2005) by avoiding complexities of integrating conflicting cultures (Drogendijk, & Slangen, 2006).

As cultural distance may increase the LOF (Eden, & Miller, 2004), WOS evidently are more disruptive than JVs (Kogut, & Singh, 1988), and following multiple previous studies, this research argues that cultural distance exerts a negative influence on entry mode choice, such that greater cultural distance is associated with firms’ higher likelihood of opting for JVs in host countries (Brouthers, & Brouthers, 2001; Brouthers, & Brouthers, 2003). Firms can subsequently minimize risks (Kim, & Hwang, 1992) and, therefore, costs (Kogut, & Singh, 1988) by sharing control and resource devotions (Hill et al., 1990) through collaborating with local firms from host countries that have experience in and knowledge on their countries (Barkema et al., 1996). Hence, firms may partially avoid coping with cultural distance and decrease the LOF via JVs with such firms (Eden, & Miller, 2004) while increasing their adaptability (Gatignon, & Anderson, 1988) to modify JVs in or withdraw from culturally distant host countries if needed (Kim, & Hwang, 1992). Thus:

Hypothesis 1 (H1): Cultural distance is negatively associated with entry mode choice, such that in contexts of greater cultural distance firms have a higher likelihood to opt for JVs.

2.3 Top management team composition, cultural distance, and entry mode choice

(9)

9 Upper echelons theory suggests that executive managers’ characteristics influence firms’ results, such as strategic decisions (Hambrick, & Mason, 1984). Entry mode choices also belong to these strategic decisions (Nielsen, & Nielsen, 2011). And TMTs can face more complexities when making entry mode choices as these strategic decisions may require greater market knowledge about host countries on account of, e.g., cultural distance (Nielsen, 2010b) and the inherently increased LOF (Eden, & Miller, 2004). But, upper echelons theory also implies that executive managers are rationally bounded due to cognitive restrictions because of, for instance, the complexity and large scale of information. Therefore, executive managers may be searching for suboptimal instead of optimal solutions or decisions (Cyert, & March, 1963; Hambrick, & Mason, 1984).

Nonetheless, IB literature mainly reckoned that entry mode choices were rational (Brouthers, & Hennart, 2007). Some studies did focus on the possible impact of CEOs’ attributes on entry mode choice (Herrmann, & Datta, 2006). Still, IB literature has generally overlooked TMTs and the inherent executive managers that make such choices. Consequently, research on the probable effect of TMT composition, concerning executive managers’ attributes, on entry mode choice is scarce (Nielsen, & Nielsen, 2011).

Accordingly, following IB research should center on the plausible effect of TMT composition on entry mode choice (Canabal, & White, 2008; Dikova, & Brouthers, 2015). Previous research typically ignored the implementation of moderators on the probable relationship between cultural distance and entry mode choice (Tihanyi et al., 2005), while moderators can explain the contradicting outcomes of these studies (Kirkman et al., 2006). Brouthers and Brouthers (2001) is one of the few studies that employed a moderator and found that cultural distance positively as well as negatively affected entry mode choice. The initial negative influence became positive due to adding the risk of an investment in a particular host country as a moderator. Thus, this research topic may be reliant on the influence of moderators (Brouthers, & Brouthers, 2001). Consequently, this study employs TMT composition as a moderator on the potential relation between cultural distance and entry mode choice.

2.3.1 Top management team nationality composition

(10)

10 Previous IB studies on TMTs primarily focused on, for example, TMT tenure, experience, and education composition and their possible impact on firms’ degree of internationalization and international diversification (Nielsen, 2010a). TMT nationality composition may additionally be associated with cultural distance. As cultures vary per country, it was proposed that this variation fosters differences among managers’ feeling, thinking or cognition, and doing via their differing norms and values (Hofstede, Hofstede, & Minkov, 2010). Therefore, national culture is a national feature that can promote differences amid nationally distinct executive managers’ understanding of (cultural) affairs and ensuing strategic decision-making (Caligiuri et al., 2004). Consequently, TMT nationality composition is a focal point of this study.

Some IB studies delineated that TMT nationality composition positively influenced firms’ internationalization, such that nationally heterogeneous TMTs positively affected the degree of internationalization (Caligiuri et al., 2004; Nielsen, 2010b). Nielsen and Nielsen (2011) is one of the few studies, to my knowledge, that focused on the credible effect of TMT (nationality) composition on entry mode choice, also in relation to cultural distance. They found that TMT nationality composition was negatively associated with entry mode choice, such that firms with nationally heterogeneous TMTs were then more inclined to be in favor of JVs in host countries. Nationally diverse TMTs can better assess and comprehend cultural distance between home and host countries and the accompanying additional risks. Therefore, such TMTs may want to avoid rigid WOS in circumstances with greater cultural distance as to restrain firms’ vulnerability to risks. Moreover, JVs credibly reflect nationally heterogeneous TMTs’ own collaborative context in transnational teams (Nielsen, & Nielsen, 2011).

(11)

11 This study expects that nationally heterogeneous TMTs have more relevant resources (Nielsen, 2010b) through the cultural synergies among executive managers (Caligiuri et al., 2004; Nielsen, & Nielsen, 2013). Such TMTs can then better cope with cultural distance, not only between their firms and other (foreign) firms in JVs (Nielsen, & Nielsen, 2011) but also between home and host countries. Nationally diverse TMTs may subsequently more thoroughly process the greater market knowledge required for entry mode choices for culturally distant host countries and can, therefore, also better deal with the bounded rationality and inherent complexities accompanying such strategic decisions (Nielsen, 2010b). Nationally heterogeneous TMTs may then make more informed strategic decisions (Nielsen, & Nielsen, 2013), such that firms have a higher likelihood to afford the resources required for WOS in host countries (Gollnhofer, & Turkina, 2015; Hill et al., 1990) when facing more complex settings with greater cultural distance (Eden, & Miller, 2004). Consequently:

Hypothesis 2 (H2): TMT nationality composition positively moderates the negative relation between cultural distance and entry mode choice, such that in contexts of greater cultural distance firms are more likely to opt for WOS when possessing nationally heterogeneous TMTs.

2.3.2 Top management team age composition

TMT age composition is defined as the composition of age of executive managers who are present in firms’ TMTs (Herrmann, & Datta, 2005; Tihanyi, Ellstrand, Daily, & Dalton, 2000; Wiersema, & Bantel, 1992). Hambrick and Mason (1984) concluded that there were only a few upper echelons studies on the potential impact of executive managers’ age on firm outcomes and that no such studies existed regarding entire TMTs. More recently, Nielsen (2010a) showed that most of the research on TMT age composition focused on TMT age heterogeneity with fewer studies centering on average TMT age. Furthermore, past IB research primarily concentrated on, e.g., TMT tenure, experience, and education composition and their plausible influence on firms’ international diversification and degree of internationalization (Nielsen, 2010a). The gap in upper echelons literature surrounding TMT age composition in the form of average TMT age, especially in combination with IB literature (Nielsen, 2010a), is more striking because of the lack of recent studies.

(12)

12 A plausible rationale behind these hypotheses is that executive managers of varying ages have different experiences that form their beliefs (Bantel, & Jackson, 1989) and these managers then complement one another with their different viewpoints, which can result in more inventive strategic decisions (Wiersema, & Bantel, 1992). Nevertheless, these studies consistently concluded that the effect of TMT age heterogeneity was insignificant on firms’ international diversification (Tihanyi et al., 2000) and strategic change (Wiersema, & Bantel, 1992).

A focus of this study is on average TMT age as Herrmann and Datta (2005) insisted that average TMT age refers to executive managers’ tendency to take risks and capability to process information. The tendency to take risks can be relevant as Eden and Miller (2004) asserted that cultural distance creates settings with extra risks due to the increased LOF. Besides the potentially higher rewards (Agarwal, & Ramaswami, 1992), WOS also likely involve further risks attributable to the greater necessity for control and resources. WOS can then especially be risky in settings with cultural distance as WOS are considered irreversible of nature because of the greater resource commitment (Hill et al., 1990). The capability to process information may be relevant as entry mode choices for culturally distant host countries can require greater market knowledge (Eden, & Miller, 2004; Nielsen, 2010b).

A focus of this research also is on average TMT age as few studies found that average age of TMTs was negatively related, such that younger average age of TMTs was positively related, to firms’ international diversification (Herrmann, & Datta, 2005; Tihanyi et al., 2000) and strategic change (Wiersema, & Bantel, 1992). A probable rationale for these findings is that younger executive managers tend to make riskier strategic decisions (Child, 1974; Hambrick, & Mason, 1984) for higher rewards to develop their career (Tihanyi et al., 2000). Older executive managers may be less inclined to make such decisions, being more risk-averse and rigid as they possibly are at a phase where their financial guarantee and retirement are more salient (Carlsson, & Karlsson, 1970; Child, 1974; Wiersema, & Bantel, 1992). Moreover, Child (1974) implied that older executive managers have less mental endurance than younger executive managers and Herrmann and Datta (2005), thus, argued that older executive managers then have less capabilities to process information.

(13)

13 Hitt and Tyler (1991) expected that executive managers’ age was negatively related to riskier acquisitions, such that younger executive managers had a higher likelihood to opt for riskier acquisitions as such managers may make riskier strategic decisions (Child, 1974; Hambrick, & Mason, 1984). But, they found the opposite in that executive managers’ age was positively associated with riskier acquisitions. Firms with older executive managers were subsequently more likely to choose riskier acquisitions. A credible reason for this finding is that firms’ executive managers cannot afford to show uncertainty, regardless of the age (Hitt, & Tyler, 1991).

As younger executive managers may have more mental endurance (Child, 1974) with better capabilities to process information (Herrmann, & Datta, 2005), this study infers that TMTs with a younger average age then have the ability to better process the greater market knowledge required for entry mode choices for culturally distant host countries. Consequently, these TMTs may better cope with the bounded rationality and inherent complexities following such choices (Nielsen, 2010b) and this may increase the quality of strategic decisions (Nielsen, & Nielsen, 2013). In addition, as younger executive managers credibly make riskier strategic decisions (Child, 1974; Hambrick, & Mason, 1984) for higher rewards (Tihanyi et al., 2000), this research further expects that firms with TMTs with a younger average age subsequently are more inclined to prefer the riskier and more rewarding entry mode WOS in host countries (Agarwal, & Ramaswami, 1992) when experiencing situations with additional complexities and risks due to greater cultural distance (Eden, & Miller, 2004). Accordingly:

(14)

14

2.4 Conceptual framework

The hypotheses and accompanying variables lead to the following conceptual framework:

Figure 1 Conceptual framework

This conceptual framework depicts a negative relation between cultural distance and entry mode choice, such that firms have a higher likelihood to prefer JVs in host countries when facing circumstances with greater cultural distance. This model further delineates a positive moderation by TMT nationality and age composition on the negative association between cultural distance and entry mode choice, such that firms are more inclined to select WOS in host countries when encountering settings with greater cultural distance.

3 Research design

(15)

15

3.1 Methodology

3.1.1 Sample and data

WOS and JVs can be measured through outward foreign direct investment (FDI) (Dow, & Larimo, 2009). Consequently, this study was centered on the Netherlands as the home country as the Netherlands has the highest degree of outward FDI globally (Central Intelligence Agency, 2018). Also, this research focused on emerging markets as the host countries because IB literature commonly concentrated on developed markets (Lindsay, Rod, & Ashill, 2015). And emerging markets may progressively be of importance to IB literature as these markets are responsible for nearly 40%, and increasing, of the global gross domestic product (GDP). Emerging markets in this study were BRICS, MINT, and Japan as they account for the majority of the above-mentioned 40% and as they are frequently categorized as emerging markets (Biswas, 2018).

The research sample subsequently consisted of Dutch firms’ entry mode choices for BRICS, MINT, and Japan. The chosen time frame for these choices was between 2010 and 2018 as years prior to this period were found to positively affect Dutch firms’ entry mode choices because of the economic recession. On account of this recession, foreign firms probably were appealing acquisition targets as they were undervalued (Williams, & Martinez, 2012). This study also concentrated on this time period to mirror the latest global economic context.

(16)

16 3.1.2 Variables

Dependent variable

Entry mode choice was dichotomously operationalized, where 1 indicated WOS and where 0 represented JVs and this operationalization aligned with past IB studies on entry mode choice (Beugelsdijk et al., 2018; Dikova, & Brouthers, 2015). To differentiate between JVs and WOS, a 95% cut-off point of ownership equity over foreign entities was applied, following prior studies (Anderson, & Gatignon, 1986; Arregle, Hébert, & Beamish, 2006; Brouthers, Brouthers, & Werner, 2008). Hence, this discrimination indicated JVs when Dutch firms had less than 95% ownership equity over entities in BRICS, MINT, and Japan and indicated WOS when Dutch firms at least had 95% ownership equity over such entities. Data on Dutch firms’ entry mode choices were collected from Orbis and Orbis also shows Dutch firms’ ownership equity over foreign entities.

Independent variable

The computation of cultural distance between the Netherlands and BRICS, MINT, and Japan occurred through the Kogut and Singh formula:

𝐶𝐷_𝑗 = ∑ 4

𝑖=1

{(𝐼_𝑖𝑗 − 𝐼_𝑖𝑛) 2_/𝑉

𝑖}/4 (1)

CDj denoted the cultural distance from the home country (the Netherlands) to the jth host

country (BRICS, MINT, and Japan). With Iij and Iin representing the jth host country and the

home country, respectively, where i showed the cultural dimensions of the Hofstede framework.

Vi subsequently exhibited the variance of the ith Hofstede dimension. The 4 expressed the

(17)

17 Still, this research applied the Hofstede framework with data from the Hofstede database as most IB studies utilized this framework (Berry et al., 2010; Kirkman et al., 2006), which sustained the internal validity of this study. This study solely used the original four dimensions of the Hofstede framework, encompassing: power distance, individualism, masculinity, and uncertainty avoidance (Hofstede, 2001). The two added dimensions, long term orientation and indulgence, were excluded as they are highly correlated (Beugelsdijk et al., 2018). Another option to measure cultural distance was the Mahalanobis formula, but in the case of the Hofstede framework this formula results in a similar outcome when compared with the more widely adopted formula 1 (Ambos et al., 2018).

Moderating variables

The nationality composition of Dutch firms’ TMTs was assessed via the Blau formula:

𝐵 = 1 − ∑ 𝑃_𝑖2 (2) 𝑘

𝑖=1

Formula 2 can measure heterogeneity among team members across different groups (Blau, 1977). In this study, P represented the percentage of firms’ TMT members in a certain ith nationality, where results of B fell between 0 and 1 and where larger results of B exhibited greater heterogeneity in TMTs’ nationality composition (Nielsen, & Nielsen, 2011). Formula 2 is also widely adopted in TMT studies (Nielsen, & Nielsen, 2011), which supported the internal validity of this research. The age composition of Dutch firms’ TMTs was approximated via the average age of TMTs by adding up all TMT members’ age and dividing that outcome by the total number of TMT members (Herrmann, & Datta, 2005; Tihanyi et al., 2000; Wiersema, & Bantel, 1992). Data for both moderators originated from Orbis as Orbis possesses data on Dutch firms their TMT members’ age and nationality, for both senior and C-suite managers. Dutch firms’ annual reports and website or the TMT members’ LinkedIn pages were alternatively employed when Orbis provided insufficient data. These data corresponded with the year Dutch firms’ entry modes occurred into BRICS, MINT, and Japan between 2010 and 2018 as to reflect Dutch firms’ TMTs and their members that made entry mode choices in specific years between the aforementioned period.

Control variables

(18)

18 The size of firms was reported to positively influence entry mode choice, such that larger firms were more likely to prefer WOS (Agarwal, & Ramaswami, 1992). Larger firms may possess relatively more resources (Brouthers, & Brouthers, 2003) necessary for WOS (Gollnhofer, & Turkina, 2015; Hill et al., 1990). As Dunning and Lundan (2008) stated that firm size is mirrored by assets, the size of Dutch firms was quantified through their latest total assets in United States (U.S.) dollars and was collected from Orbis or annual reports, depending on missing data. The logarithm of Dutch firms’ total assets was used as total assets represent large figures. Logarithms transform large figures into smaller and more comprehensible figures while keeping the basic data intact (Hair, Black, Babin, & Anderson, 2009).

Firms’ international experience was also reported to positively affect entry mode choice, such that internationally more experienced firms were more likely to choose WOS (Agarwal, & Ramaswami, 1992; Dow, & Larimo, 2009). Such firms may face less risks and costs (Erramilli, 1991) due to, for instance, more effectively and efficiently accumulating and utilizing resources (Brouthers, & Brouthers, 2001; Brouthers et al., 2008). International experience was determined through firms’ share of foreign sales from total sales (Musteen, Datta, & Herrmann, 2009) as firms seemingly make strategic internationalization decisions based on prior international performance. Other measures, such as firms’ number of prior host country entries or the total years of international operations can be less fitting to capture this performance (Aulakh, & Kotabe, 1997). Dutch firms’ most recent share of foreign from total sales in U.S. dollars was gathered from Orbis or annual reports in case of lacking data.

On the TMT-level, Dutch firms’ TMT size was controlled for. Although, Nielsen and Nielsen (2011) found that TMT size had no influence on entry mode choice. Yet, TMT size can have an impact as firms with larger TMTs contain more members and, therefore, more resources (Nielsen, 2010b) required for WOS (Gollnhofer, & Turkina, 2015; Hill et al., 1990). TMT size was computed through the total number of TMT members (Nielsen, & Nielsen, 2011) with data from Dutch firms’ annual reports and website or through the Orbis database, contingent on lacking data. As with the moderating variables, data on Dutch firms’ TMT size corresponded with the year entry modes occurred as to reflect the TMTs that made entry mode choices in certain years.

(19)

19 The latest GDP per capita in U.S. dollars and infrastructure rates of BRICS, MINT, and Japan were collected from the World Bank database, where infrastructure was measured via these countries’ international Logistics Performance Index. The most recent unemployment rates of BRICS, MINT, and Japan were collected from The World Factbook database of the Central Intelligence Agency.

3.1.3 Statistical techniques

Descriptive statistics were firstly analyzed for insight into basic features of the data. Then multicollinearity was checked as this study had multiple independent (cultural distance and the control variables) and moderating (TMT nationality and age composition) variables. Pearson correlations above 0.700 or below -0.700 and variance inflation factor (VIF) values above 10 indicated multicollinearity. In SPSS the variables’ data were analyzed with a logistic regression because of the binary dependent variable (Hair et al., 2009), where 1 indicated WOS and where 0 represented JVs. The logistic regression was separated into four models.

Model 1 only contained the dependent variable entry mode choice and the control variables. Model 2 added the independent variable cultural distance to determine the impact cultural distance may have on entry mode choice. Model 3 added the moderating variables TMT nationality and age composition to evaluate their probable independent influence on entry mode choice. And model 4 added the interaction terms of both moderating variables with the independent variable as to assess the plausible moderating effects of TMT nationality and age composition on the possible association between cultural distance and entry mode choice.

(20)

20 As such, the initial logistic regression models 2, 3, and 4 were renamed to models 5, 6, and 7 (respectively) in the robustness test to clarify the substitution of the Hofstede framework by the Schwartz framework. Model 1 was not included in the robustness test as this model only includes control variables. Therefore, this model would yield identical results.

4 Results

4.1 Descriptive statistics

The data comprised a combination of 202 entry mode choices (WOS and JVs) of 51 Dutch firms for 10 host countries (BRICS, MINT, and Japan). Table 1 shows that there were differences between the 51 Dutch firms within the sample. Some firms had nationally more heterogeneous TMTs with heterogeneity scores above 0.50 with a maximum score of 0.88, whereas other firms had diversity scores of or below 0.50 with some firms having nationally completely homogeneous TMTs with a minimum score of 0.00. Some firms also had TMTs with an average age below 50 with a minimum average age of 48, while other firms had TMTs with an average age between 50 and 60 or above 60 with a maximum average age of 63.

Table 1 also exhibits variations among the 10 host countries as some of these countries were culturally more distant or proximate than others from the perspective of the Netherlands, within the context of this study. The most distant host country had a cultural distance score of 5.97 and the closest host country had a cultural distance score of 1.68 with the other host countries’ cultural distance scores falling between 1.68 and 5.97.

Table 1 Descriptive statistics

Variables Minimum Maximum Mean SD

Entry mode choice 0.00 1.00 0.47 0.50

TMT size 1.00 27.00 8.28 5.42 Firm international experience 0.00 1.00 0.73 0.28 Firm size 13.95 27.65 23.42 2.85 GDP per capita 1942.10 38428.10 8853.27 6718.61 Infrastructure rate 2.53 4.03 3.25 0.33 Unemployment rate 0.03 0.28 0.09 0.07 CD (Hofstede) 1.68 5.97 4.42 1.46 TMT nationality composition 0.00 0.88 0.48 0.28 TMT age composition 48.00 63.00 54.32 3.00

(21)

21 Of the 202 entry mode choices in the sample, 52% were JVs and 48% were WOS as represented by table 5 in appendix 1. The majority of JVs were in culturally more distant host countries and most WOS were in culturally more proximate host countries, within the context of this study. This occurrence is seen in table 5, which displays the distribution of the 202 entry mode choices across the 10 host countries in the sample and also presents these countries’ cultural distance from the Netherlands. The aforementioned results will be further clarified. Yet, a logistic regression is flexible as a logistic regression does not require assumptions regarding, for example, homoskedasticity, linearity, and data distribution. This flexibility makes the basic features of the data in descriptive statistics less relevant. That being said, multicollinearity is of importance (Hair et al., 2009).

4.2 Correlations and multicollinearity

Table 2 exhibits the Pearson correlations among all the variables in this research. As expected, cultural distance was significantly and negatively (-0.242, P = 0.001) correlated with entry mode choice and TMT nationality composition was positively and significantly (0.334, P = 0.000) correlated with entry mode choice. These findings will be further clarified. As elaborated, multicollinearity could be an issue. Pearson correlations above 0.700 or below -0.700 and VIF values above 10 indicated multicollinearity between the independent and moderating variables. The VIF values were obtained via collinearity diagnostics in a linear regression (Hair et al., 2009) of models 1 to 4. As such, the only noteworthy Pearson correlation in table 2 was that between cultural distance and unemployment rate due to being the only correlation that exceeded the threshold of -0.700 with a significant correlation of -0.868. Within the collinearity diagnostics of the linear regression of models 1 to 4, no independent and moderating variable had a VIF value that exceeded the threshold of 10. All the variables had VIF values below 4.916 and all the models had average VIF values below 2.334. Hence, there were no indications of multicollinearity between the independent and moderating variables in this research.

4.3 Logistic regression

(22)

22

Note: significance = ** P<0.01, * P<0.05. GDP = gross domestic product and TMT = top management team.

Variables 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

1. Entry mode choice

2. TMT size 0.190** 3. Firm international experience 0.179* 0.292** 4. Firm size 0.000 0.461** -0.152* 5. GDP per capita -0.037 -0.041 0.067 0.044 6. Infrastructure rate -0.069 -0.133 0.031 -0.087 0.423** 7. Unemployment rate 0.215** -0.160* -0.091 -0.186** -0.249** -0.185*

8. Cultural distance (Hofstede) -0.242** 0.142* 0.130 0.131 0.297** 0.349** -0.868** 9. TMT nationality

composition

0.334** 0.593** 0.520** 0.373** -0.063 -0.137 -0.138 0.132

10. TMT age composition -0.022 -0.098 0.182** -0.137 0.167* -0.001 0.039 -0.037 -0.037

(23)

23 Odds ratios show the direction of relations and the direct effect with the multiplicative magnitude of that effect of unit increases in independent variables and the corresponding change in the odds or likelihood that dependent variables fall into one of two categories. The logistic regression coefficients are less interpretable as they are logarithms of the odds ratios, called logits. And the logistic regression coefficients only assess the change in these logits and the direction of relations. Non-negative odds ratios below 1 indicate negative relations, odds ratios of 1 denote no relation, and odds ratios above 1 represent positive relations. Directions of odds ratios correspond with the positive and negative signs of logistic regression coefficients. In logistic regression, dependent variables that are coded 1 are represented by the odds ratios (Hair et al., 2009; Hosmer, & Lemeshow, 2000). Hence, in this study the odds ratios represented the likelihood of (Dutch) firms opting for WOS.

Table 3 depicts the logistic regression results in odds ratios divided into model 1, 2, 3, and 4. Additionally, table 6 in appendix 2 depicts the logistic regression results in the logistic regression coefficients separated into model 1, 2, 3, and 4. H1 implied that cultural distance negatively affects entry mode choice, such that firms have a higher likelihood to choose JVs in host countries when encountering circumstances with greater cultural distance. Cultural distance consistently exerts a significant and negative effect on entry mode choice in model 2 (0.497, P = 0.035), model 3 (0.435, P = 0.019), and model 4 (0.389, P = 0.011), such that firms were less inclined to opt for WOS and more inclined to opt for JVs when host countries were culturally more distant and this finding lent support to H1.

Although not separately hypothesized, the independent effect of TMT nationality composition on entry mode choice is consistently significant and positive in model 3 (3.102, P = 0.000) and in model 4 (2.992, P = 0.000). Thus, firms with nationally more heterogeneous TMTs were more likely to be in favor of WOS in host countries. Similarly, the independent effect of TMT age composition on entry mode choice is consistently negative and insignificant in model 3 (0.924, P = 0.655) and in model 4 (0.850, P = 0.411).

(24)

24 H2 suggested that TMT nationality composition positively moderates the negative association between cultural distance and entry mode choice, such that firms with nationally heterogeneous TMTs are more likely to select WOS in host countries when facing situations with greater cultural distance. The effect of the interaction between cultural distance and TMT nationality composition on entry mode choice is positive (0.112) and insignificant (P = 0.566) in model 4. Hence, H2 was not supported. Besides the insignificance, the moderating effect of H2 was as hypothesized. However, a precise or accurate interpretation of moderating effects is difficult without a graph (Aiken, & West, 1991; Hosmer, & Lemeshow, 2000). Yet, as H2 was not supported, the moderating effect of TMT nationality composition was not graphed.

H3 argued that TMT age composition positively moderates the negative relationship between cultural distance and entry mode choice, such that firms with TMTs with a younger average age are more inclined to opt for WOS in host countries when experiencing settings with greater cultural distance. The effect of the interaction between cultural distance and TMT age composition on entry mode choice is negative (-0.362) and insignificant (P = 0.055) in model 4. Thus, H3 was not supported. As with H2, the moderating effect of TMT age composition was not graphed as H3 was not supported.

4.4 Robustness test

(25)

25

Table 3 Logistic regression results with odds ratios

Variables Model 1 Model 2 Model 3 Model 4

DV = 1 for WOS and 0 for JV Entry mode choice Entry mode choice Entry mode choice Entry mode choice Control variables TMT size 1.087* 1.094* 1.039 1.046 (0.036) (0.037) (0.039) (0.041)

Firm international experience 3.145 3.536* 0.529 0.501

(0.631) (0.640) (0.840) (0.832) Firm size 0.981 0.977 0.867 0.855 (0.067) (0.069) (0.079) (0.084) GDP per capita 1.000 1.000 1.000 1.000 (0.000) (0.000) (0.000) (0.000) Infrastructure rate 0.919 1.352 1.774 1.503 (0.513) (0.551) (0.580) (0.595) Unemployment rate 5335.340* 1.361 1.300 0.249 (2.543) (4.595) (4.885) (5.145) Independent variable

Cultural distance (Hofstede) 0.497* 0.435* 0.389*

(0.332) (0.355) (0.371) Moderating variables TMT nationality composition 3.102** 2.992** (0.274) (0.274) TMT age composition 0.924 0.850 (0.178) (0.197) Interaction CD*TMT nationality composition 0.112 (0.196) CD*TMT age composition -0.362 (0.189) Model statistics Constant 0.163 0.089 3.151 7.143 Model change ꭓ2 25.440 4.488 20.823 4.442 P ꭓ2 0.000 0.034 0.000 0.109 Model ꭓ2 25.440 29.928 50.751 55.193 P ꭓ2 0.000 0.000 0.000 0.000 -2 Log likelihood 253.620 249.133 228.310 223.868 Pseudo R2 0.158 0.184 0.297 0.319

(26)

26 Cultural distance consistently exerts a significant and negative effect on entry mode choice in model 5 (0.498, P = 0.020), model 6 (0.463, P = 0.016), and model 7 (0.461, P = 0.017), in support of H1. These effects were similar to the effects in the corresponding initial models 2, 3, and 4, respectively. The largest difference was the magnitude of this effect between model 4 (0.389) and model 7 (0.461) and the significance between model 2 (P = 0.035) and model 5 (P = 0.020).

The independent effect of TMT nationality composition on entry mode choice is consistently significant and positive in model 6 (3.070, P = 0.000) and model 7 (3.067, P = 0.000). The significance of these effects was identical to the significance in their corresponding initial models 3 and 4, respectively. Still, the magnitude of these effects was comparable, with the largest difference being between the corresponding models 4 (2.992) and 7 (3.067). The independent effect of TMT age composition on entry mode choice is consistently negative and insignificant in model 6 (0.948, P = 0.764) and model 7 (0.943, P = 0.741) akin to the corresponding original models 3 and 4, respectively. The biggest distinction was in the magnitude and significance between the corresponding models 4 (0.850, P = 0.411) and 7 (0.943, P = 0.741).

As indicated, in the case of a moderation effect, the logistic regression coefficients are preferred over the odds ratios (Hosmer, & Lemeshow, 2000). Thus, for the moderating effects of TMT nationality and age composition the logistic regression coefficients are given. The effect of the interaction between cultural distance and TMT nationality composition on entry mode choice is negative (-0.028) and insignificant (P = 0.888) in model 7. Therefore, H2 was not supported as with its counterpart, model 4 (P = 0.566). Yet, model 4 displays a positive effect (0.112). The interaction effect between cultural distance and TMT age composition on entry mode choice is negative (-0.061) and insignificant (P = 0.753) in model 7. Consequently, H3 was not supported as with its corresponding original model 4 (-0.362, P = 0.055).

(27)

27

Table 4 Logistic regression results robustness test with odds ratios

Variables Model 1 Model 5 Model 6 Model 7

DV = 1 for WOS and 0 for JV Entry mode choice Entry mode choice Entry mode choice Entry mode choice Control variables TMT size 1.087* 1.090* 1.034 1.037 (0.036) (0.037) (0.039) (0.040)

Firm international experience 3.145 3.972* 0.599 0.575

(0.631) (0.648) (0.846) (0.850) Firm size 0.981 0.982 0.871 0.871 (0.067) (0.069) (0.079) (0.080) GDP per capita 1.000 1.000 1.000 1.000 (0.000) (0.000) (0.000) (0.000) Infrastructure rate 0.919 4.209 5.926* 5.964* (0.513) (0.850) (0.900) (0.901) Unemployment rate 5335.340* 176.307 527.090* 576.278* (2.543) (2.918) (3.143) (3.185) Independent variable

Cultural distance (Schwartz) 0.498* 0.463* 0.461*

(0.301) (0.320) (0.325) Moderating variables TMT nationality composition 3.070** 3.067** (0.275) (0.275) TMT age composition 0.948 0.943 (0.178) (0.179) Interaction CD*TMT nationality composition -0.028 (0.196) CD*TMT age composition -0.061 (0.192) Model statistics Constant 0.163 0.002 0.068 0.067 Model change ꭓ2 25.440 5.917 20.171 0.115 P ꭓ2 0.000 0.015 0.000 0.944 Model ꭓ2 25.440 31.357 51.528 51.643 P ꭓ2 0.000 0.000 0.000 0.000 -2 Log likelihood 253.620 247.703 227.532 227.417 Pseudo R2 0.158 0.192 0.301 0.301

(28)

28

5 Discussion and conclusion

Entry mode choices can require considerable and inflexible resource commitments, making suitable entry mode choices for host countries crucial for firms to successfully internationalize (Agarwal, & Ramaswami, 1992). Entry mode choices may be hampered by cultural distance (Kogut, & Singh, 1988) due to the likely increase in the LOF (Eden, & Miller, 2004). Although there were many IB studies on the plausible association between cultural distance and entry mode choice, the findings are indecisive (Beugelsdijk et al., 2018). Prior research mainly neglected the use of moderators on this possible relation (Tihanyi et al., 2005). And the credible influence of cultural distance on entry mode choice may rely on moderating effects (Brouthers, & Brouthers, 2001), explaining past studies’ indecisiveness (Kirkman et al., 2006).

Additionally, previous IB studies mostly neglected the possible impact of TMT composition on entry mode choice. Although TMTs, comprised of executive managers, make such strategic choices (Nielsen, & Nielsen, 2011). And it has been proposed that executive managers’ attributes affect strategic decisions (Hambrick, & Mason, 1984). Concerning research on TMTs, both IB and upper echelons studies mainly ignored TMT nationality (Kaczmarek, & Ruigrok, 2013; Nielsen, & Nielsen, 2013) and age (Nielsen, 2010a) composition. Despite the fact that foreigners are increasingly present in firms’ upper echelons (Nielsen, & Nielsen, 2011), just as younger people (Tanikawa et al., 2017). Furthermore, TMT nationality composition can be related to cultural distance. Cultures vary per country (Hofstede, 2001). Therefore, Caligiuri et al. (2004) suggested that national culture is a national characteristic that creates variations amid nationally different executive managers with respect to their comprehension of (cultural) phenomena and resulting strategic decisions. TMT age composition evidently also is related to cultural distance as age can, e.g., represent risk-taking tendencies (Herrmann, & Datta, 2005). Risk-taking may be applicable as Eden and Miller (2004) indicated that cultural distance creates situations with added risks on account of the increased LOF.

This led to the ensuing research question: Do TMT nationality and age composition affect the

association between cultural distance and entry mode choice?

(29)

29 Firms then credibly opt for a risk (Kim, & Hwang, 1992) and, thus, cost (Kogut, & Singh, 1988) reducing approach through dividing resource commitments and control (Hill et al., 1990). This approach can be realized via JVs with local firms from host countries that possess local experience in and knowledge on their respective countries (Barkema et al., 1996). Hence, firms may partly circumvent dealing with cultural distance and decrease the inherent LOF by means of JVs with such local firms (Eden, & Miller, 2004). Firms subsequently also potentially improve their versatility (Gatignon, & Anderson, 1988) to adapt JVs in or evacuate from culturally distant host countries if necessary (Kim, & Hwang, 1992).

The support of H1 is simultaneously conflicting with other prior studies as they found the opposite. These studies reported that cultural distance positively affected entry mode choice in such a way that made firms more likely to opt for WOS in host countries when they experienced contexts of greater cultural distance (Brouthers, & Brouthers, 2001; Chen, & Hu, 2002; Gollnhofer, & Turkina, 2015). The above-mentioned mixed findings possibly have multiple explanations. For example, the Hofstede, Globe, and Schwartz frameworks can assess different cultural features as the correlation between these frameworks is low. Accordingly, cultural distance scores between countries and, consequently, certain outcomes of studies may largely rely on the utilized cultural framework (Ambos et al., 2018).

Specifically, the support of H1 conflicts with Gollnhofer and Turkina (2015) and they applied the Globe framework, whereas this study used the Hofstede framework. Still, this study found similar results with both the Hofstede framework and the Schwartz framework. Regarding the support of H1, this study also conflicts with Chen and Hu (2002), while they also utilized all four original dimensions of the Hofstede framework. Cultural distance scores between countries per dimension of the Hofstede framework may then have altered over time (Kirkman et al., 2006). Nonetheless, Beugelsdijk, Maseland, and Van Hoorn (2015) implied that these scores have principally not altered.

(30)

30 This research did employ moderators and established that cultural distance then exerted a positive impact on entry mode choice as the pre-existing negative relation was positively moderated by TMT nationality composition. This finding does give merit to the idea that moderators clarify the inconclusive results regarding this topic (Kirkman et al., 2006). Yet, the moderating effects in this research of TMT nationality and age composition were insignificant. This insignificance will be further elaborated on.

Alternatively and as opposed to this study, several studies found and suggested that there is no (direct) association between cultural distance and entry mode choice (Morschett et al., 2010; Tihanyi et al., 2005). Other estimations of institutional distance may be more applicable in relation to entry mode choice, such as economic, administrative, and political distance (Berry et al., 2010). Consequently, Berry et al. (2010) proposed that these other estimations of institutional distance contribute to solving the mixed findings regarding the possible influence of, for example, cultural distance on entry mode choice.

H2 was not supported as the hypothesized moderating effect of TMT nationality composition was not significant. This effect may be insignificant as cultural distance is superfluous as an independent variable. The rationale is that entry mode choice presumably is a multilevel occurrence, where cultural distance (among other factors) can already be nested in entry mode choices for the host countries in question. And not taking into account this multilevel nature plausibly has detrimental statistical ramifications (Arregle et al., 2006). Nielsen and Nielsen (2011) realized that entry mode choice may be multilevel of nature. Therefore, they hypothesized on the independent effect of TMT nationality composition on entry mode choice while disaggregating cultural distance from entry mode choices and controlling for this distance in a multilevel approach with a hierarchical design (Nielsen, & Nielsen, 2011).

(31)

31 Moreover, presuming that entry mode choice is multilevel, where cultural distance is nested in entry mode choice (Arregle et al., 2006), it can be illogical that the hypothesized effect of H1 was significant and that H1 was supported. H1 may then not be supported when applying a multilevel method.

Still, assuming that cultural distance is nested in entry mode choice (Arregle et al., 2006): A credible reason for that firms with nationally heterogeneous TMTs were more inclined to select WOS in host countries, is that the aforementioned TMTs possess more appropriate resources (Nielsen, 2010b) necessary for WOS (Gollnhofer, & Turkina, 2015; Hill et al., 1990) in more complex situations with greater cultural distance (Eden, & Miller, 2004). These resources plausibly involve complementarities between executive managers due to their differing cultural viewpoints, values, experiences, and cognition (Caligiuri et al., 2004; Nielsen, & Nielsen, 2013). Nationally diverse TMTs may then be better equipped to deal with cultural distance, not merely between (foreign) firms in JVs (Nielsen, & Nielsen, 2011) but also between home and host countries. These TMTs then credibly more rigorously process the greater market knowledge needed for entry mode choices for culturally distant host countries and can, thus, also better cope with the bounded rationality and complexities following such strategic decisions (Nielsen, 2010b). Nationally heterogeneous TMTs may subsequently make more sophisticated strategic decisions (Nielsen, & Nielsen, 2013).

Nielsen and Nielsen (2011) reported the opposite as TMT nationality composition was negatively related to entry mode choice, such that firms with nationally diverse TMTs were more likely to select JVs in host countries. Such TMTs potentially better determine and understand cultural distance between home and host countries and the following added risks. Accordingly, nationally heterogeneous TMTs may want to circumvent inflexible WOS in contexts of greater cultural distance as to constrain firms’ exposure to risks (Nielsen, & Nielsen, 2011).

(32)

32 Table 5 in appendix 1 indicates that, in the sample of this study, WOS mostly occurred in culturally more proximate host countries, while JVs mostly occurred in culturally more distant host countries. This pattern credibly aligns with the Uppsala model as this model implies that firms internationalize progressively by selecting host countries and increasingly enlarging resource devotions in these countries contingent upon, for instance, cultural distance (Johanson, & Vahlne, 1977) due to the inherent increase in LOF (Eden, & Miller, 2004). Thus, firms can initially internationalize to culturally proximate host countries in relation to the home country and eventually to culturally distant host countries by progressively expanding resource dedications (Johanson, & Vahlne, 1977). These resource commitments can be related to the resource devotions required for entry mode choices, such as WOS and JVs (Gollnhofer, & Turkina, 2015; Hill et al., 1990). Therefore, the notion of the Uppsala model regarding firms’ progressive internationalization may still be relevant (Barkema, & Drogendijk, 2007).

No support was found for H3 as the hypothesized moderating effect of TMT age composition, in the form of average TMT age, was not significant. Because of the probable multilevel nature of entry mode choice (Arregle et al., 2006), this effect can be insignificant as with H2. Conversely, the hypothesized moderating effects of H2 and H3 can be insignificant due to measuring cultural distance in a composite manner by incorporating all four original Hofstede dimensions. Every individual Hofstede dimension may not (equally) affect particular dependent variables (Maseland, Dow, & Steel, 2018). Beugelsdijk et al. (2018) found that from the Hofstede framework, solely the dimension individualism exerted a significant impact on entry mode choice. Shenkar (2001) asserted that the Hofstede dimension uncertainty avoidance is theoretically more relevant in relation to entry mode choice than the other Hofstede dimensions as both entry mode choice and uncertainty avoidance may be related to risk. In this study, TMT age composition in the form of average TMT age can also be related to risk (Herrmann, & Datta, 2005). Maseland et al. (2018) urged that a composite method of measuring cultural distance leads to the concealment of crucial results. Hence, the interaction effects of TMT nationality and age composition with cultural distance on entry mode choice could have been significant if separate and relevant cultural dimensions were utilized.

(33)

33

5.1 Contributions

The contributions of this study are twofold. Firstly, the findings in IB literature are mixed concerning the probable impact of cultural distance on entry mode choice (Beugelsdijk et al., 2018). This study advances the discussion in IB literature on this possible relation by arguing that cultural distance is a significant and negative predictor of entry mode choice. Secondly, IB literature commonly inferred that entry mode choices were logical (Brouthers, & Hennart, 2007). However, few studies did research the potential influence of CEOs’ features on entry mode choice (Herrmann, & Datta, 2006). Still, IB literature largely omitted TMTs’ involvement in entry mode choices. While TMTs, comprised of executive managers, make such choices (Nielsen, & Nielsen, 2011). Following calls to study the potential effect of TMT composition on entry mode choice (Canabal, & White, 2008; Dikova, & Brouthers, 2015), this research aimed to fill this gap. This research furthers IB literature via proposing that TMT nationality composition is a significant and positive predictor of entry mode choice by integrating IB studies with an upper echelons perspective and considering entire TMTs instead of only CEOs.

5.2 Managerial implications

(34)

34

5.3 Limitations and future research

Every research has its caveats. The results of this research are hampered by a small sample size of 202 as Hosmer and Lemeshow (2000) stated that the recommended sample size for a logistic regression is above 400. This small sample size may also limit the generalizability of this research. The generalizability can be further restricted due to the specific sample by focusing on Dutch firms and concentrating on entry modes into BRICS, MINT, and Japan. This study did not control for industry. And entry mode choices possibly differ per industry as Dess, Ireland, and Hitt (1990) argued that varying industries handle factors, that influence such choices, in a different way. As discussed, this research did not use a multilevel approach. Entry mode choice potentially is a multilevel matter, where cultural distance may already be nested in entry mode choices. And not considering this multilevel essence can have adverse statistical implications (Arregle et al., 2006). This research measured cultural distance with formula 1 for the Schwartz framework. In hindsight, the Mahalanobis formula may be a better option for this framework as the outcome not only differs from that of formula 1, but the Mahalanobis formula also seemingly rectifies correlations among the individual dimensions (Ambos et al., 2018).

In this study, the control variables’ data came from the most recent point in time as to reflect the most recent situation. However, these data should have paralleled the year Dutch firms’ entry modes transpired into BRICS, MINT, and Japan between 2010 and 2018 as to mirror the situation when such entry mode choices were made. Data on the moderating variables concerning TMTs and their members did correspond with Dutch firms’ year of entry into the aforementioned countries. Yet, assembling these data often proved difficult on account of missing TMT data on earlier years in the period between 2010 and 2018. Hence, TMT data were frequently approximated with the data that were available and these data mostly originated from the more recent years in the above-mentioned period. A better alternative might have also been to collect the data on the control and moderating variables a year prior to Dutch firms’ entry modes as entry mode choices can be made before the actual year of entry.

(35)

35 This composite manner of measuring cultural distance and ensuing probable concealment of results can also be applicable to all the Schwartz dimensions that were used for the robustness test. Taken together, all the aforementioned limitations may make the outcomes of this research less accurate, valid, and reliable.

Considering the aforementioned results and limitations, there are fruitful theoretical implications and avenues for future studies. This study proposes a negative and significant relationship between cultural distance and entry mode choice and this discovery may make the mixed outcomes in IB literature on this probable association either less or more mixed. Previous studies largely overlooked the employment of moderators on the above-mentioned plausible relation (Tihanyi et al., 2005). Meanwhile, the credible influence of cultural distance on entry mode choice potentially depends on moderators to describe underlying mechanisms (Brouthers, & Brouthers, 2001). Moderators can then unfold the conflicting findings of prior studies on this subject (Kirkman et al., 2006). This study used moderators and reported that cultural distance (with the Hofstede framework) then positively, although insignificantly, impacted entry mode choice as the previous negative relationship was positively moderated by TMT nationality composition. Hence, more research may be required to further explain these inconsistencies through the application of moderators.

Cultural distance scores between countries and, thus, particular findings of studies can be reliant on the applied cultural framework (Ambos et al., 2018). Moreover, individual dimensions of cultural frameworks possibly do not (equally) affect specific dependent variables (Maseland et al., 2018), such as entry mode choice (Beugelsdijk et al., 2018; Shenkar, 2001). Thus, based on the research question, following research on the probable influence of cultural distance on entry mode choice should choose a cultural framework (Ambos et al., 2018) and focus on (a combination of) individual dimensions of the chosen framework. Each study then credibly has a cultural framework that is customized to the research topic (Maseland et al., 2018).

(36)

36 Still, firms can implement a broader range of entry modes. Following studies may additionally operationalize entry mode choice into, e.g., exporting, licensing, and franchising (Anderson, & Gatignon, 1986; Hill et al., 1990) or even discriminate both JVs and WOS even further into acquisition and greenfield (Beugelsdijk et al., 2018; Dikova, & Brouthers, 2015). These studies can subsequently further delineate the potential effect of cultural distance on a broader range of entry mode choices.

This research implies that TMT nationality composition is a positive and significant predictor of entry mode choice. This finding subsequently indicates that entry mode choices are not solely affected by, for example, cultural distance and CEOs’ attributes but also by executive managers in TMTs and their attributes. This result serves as a groundwork to spur additional research on the plausible effect of other TMT composition attributes on entry mode choice as to evolve IB literature with an upper echelons perspective. Especially, since past IB literature largely omitted the potential impact of TMT composition on entry mode choice (Nielsen, & Nielsen, 2011). Gender is another commonly neglected TMT composition attribute in IB and upper echelons literature, while minorities are increasingly present in firms’ upper echelons (Nielsen, 2010a). Accordingly, ensuing studies can concentrate on the plausible impact of TMT gender composition on entry mode choice or as a moderator on the possible association between cultural distance and entry mode choice.

The generalizability of this research may be limited due to a small sample size, by centering on Dutch firms, and focusing on entry modes into BRICS, MINT, and Japan. To my knowledge, only a few studies have used the perspective of entry mode choices being made by firms from emerging markets. Further studies in a similar vein can subsequently focus on firms from an emerging market serving as the home country with developed markets or other emerging markets serving as the host countries. Overall, future research on entry mode choice must consider to apply a multilevel method as Arregle et al. (2006) proposed that entry mode choice is a multilevel occurrence. As this study was faced with data that were frequently missing on Dutch firms’ TMTs, future research should also consider to collect primary data on TMTs rather than being dependent on and probably hampered by secondary data and the availability thereof.

Notes

Do top management team nationality and age composition affect the association between cultural distance and entry mode choice?