Master thesis
Business Administration – International Management (M. Sc.)
Usage of mediator platforms to lower perceived distance:
An analysis of ownership patterns among foreign subsidiaries of
German MNEs
Name: Barbara Keller
Student number: 11086467
First Supervisor: Dr. Carsten Gelhard
Second Supervisor: Dr. Lori DiVito
Deadline: 24
thJune, 2016
2
Statement of originality
This document is written by student Barbara Keller who declares to take full
responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no
sources other than those mentioned in the text and its references have been used in
creating it.
The Faculty of Economics and Business is responsible solely for the supervision of
completion of the work, not for the contents.
3
Table of contents
1. INTRODUCTION 6 2. LITERATURE REVIEW 8 2.1. DISTANCE 8 2.1.1. CULTURAL DISTANCE 8 2.1.2. ADMINISTRATIVE DISTANCE 10 2.1.3. GEOGRAPHIC DISTANCE 11 2.1.4. ECONOMIC DISTANCE 122.2. ENTRY MODES AND OWNERSHIP LEVELS 12
2.3. MEDIATOR PLATFORMS AND RESEARCH GAP 20
3. CONCEPTUAL FRAMEWORK 22
3.1. HYPOTHESES DEVELOPMENT –CULTURAL DISTANCE 22
3.2. HYPOTHESES DEVELOPMENT –ADMINISTRATIVE DISTANCE 24
3.3. HYPOTHESES DEVELOPMENT –GEOGRAPHIC DISTANCE 25
3.4. HYPOTHESES DEVELOPMENT –ECONOMIC DISTANCE 26
4. METHODOLOGY 27
4.1. SAMPLE DESCRIPTION AND DATA 28
4.2. VARIABLES 28 4.3. ANALYTICAL APPROACH 31 5. DISCUSSION 39 6. CONCLUSION 41 ACKNOWLEDGEMENT 43 REFERENCES 44 APPENDIX 54
4
List of tables
Table 1 Descriptive Statistics ... 32 Table 2 Correlation Table ... 33 Table 3 Regression analysis ... 36
List of graphs
Graph 1 Slope diagram – Geographic Distance & OECD Membership ... 38 Graph 2 Slope diagram – Economic Distance & OECD Membership ... 38
List of abbreviations
CAGE = Cultural, administrative, geographic and economic FDI = Foreign direct investment
GDP = Gross domestic product
GLOBE = Global Leadership and Organisational Behaviour Effectiveness JV = Joint Venture
MNE = Multinational enterprise
OECD = Organisation for Economic Co-operation and Development OLI = Ownership, location and internalization
R&D = Research and development RBV = Resource-based view TCT = Transaction cost theory US = United States
WOS = Wholly owned subsidiary WTO = World Trade Organization
5
Abstract
There are various theories that try to determine the differences among countries and their effect on doing business abroad. One of these effects is the ownership structure of multinational enterprises’ (MNEs) foreign subsidiaries. In order to decrease risk and uncertainties, it is important to investigate measures to diminish the negative effects of distance. One measure could be home and host country’s membership of the same mediator platform. A mediator platform membership of both home and host country is expected to moderate the effects of distance and thus positively influences the ownership levels of MNEs’ subsidiaries in relatively distant host countries.
This thesis aims at an investigation of a possible moderating effect of mediator platform membership, here the membership of the Organisation for Economic Co-operation and Development (OECD), on the relation between cultural, administrative, geographic and economic distance, and level of subsidiary ownership. For this reason, a regression analysis is conducted by examining a data sample of 3380 MNEs originating from Germany with 39282 subsidiaries in 65 countries.
The outcomes are counter-intuitive regarding cultural, administrative and economic distance because they support the relation between greater distance among home and host country and greater MNEs’ ownership percentage of subsidiaries. Nevertheless, the expected ownership-inhibiting effect of distance arises when checking for geographic distance. When including the mediator platform into the analysis, the results support the presumption that home and host countries’ membership of the same mediator platform increases the likelihood of an MNE to possess higher ownership percentages of its foreign subsidiaries in administratively and economically distant host countries. This effect cannot be found for culturally and geographically distant host countries.
Key words: CAGE distance; Ownership pattern; Foreign Direct Investment; German
6
1. Introduction
Doing international business is based on the assumption of a firm that participates in cross-national business activities (Peng, 2012). If a firm ‘owns and [/or] controls value creating activities in two or more different countries’, it is considered an MNE (Buckley & Casson, 2009, p.1564). These value-creating activities are accomplished with foreign direct investments (FDIs) and comprise import and export of goods and services among countries but also the establishment and acquisition of assets abroad (Buckley & Casson, 2009).
Managing cross-border business is manifold and more complex than being engaged in only one country. Different currencies between trading countries, or existent political risks in the home country as well as in host countries might be complicating cross-national issues. Additionally, the institutional and cultural environment, and also the level of development and societal values vary significantly among home and host countries depending on regional settings (Aggarwal et al. 2008). These differences require a country-specific flexibility and adaptability when doing international business in order to meet the emerging strategic challenges.
A firm’s idea of crossing borders comes along with difficulties in evaluating opportunities and possible problems that could arise in future host countries. Problems a foreign firm encounters, which do not appear for domestic competitors, can be explained by the different dimensions of the companies’ origin. These differences are combined by the term Distance. Distance is a crucial factor when dealing with a multinational environment because it influences several international business related matters (Hutzschenreuter, Kleindienst & Lange, 2015). It affects decisions concerning subsidiary’s potential host country, international performance, the relation between headquarter and its affiliates, knowledge exchange within the whole organization, the choice of entry mode and thus, the amount of ownership the MNE possesses of its subsidiary abroad (Kirkman, Lowe & Gibson, 2006; Tihanyi, Griffith & Russell, 2005). The level of ownership is a very sensitive topic, since it is dependent on various internal but also external influences. The kind of business, the future strategic targets and the company’s culture are internal factors that might affect this decision. Institutional
7 aspects, economic differences and cultural issues belong to external influences. In a nutshell, the decision of ownership level is related to high risk but also great opportunities.
Since MNEs aim at entering new markets and know that handling distance can be expensive, but the incapability of overcoming these dimensions of distance will be more expensive, they focus on ways to narrow the levels of distance when going abroad. Platforms for information exchange or trade negotiations are possible moderators which could diminish the perceived distance and consequently facilitate the process of doing business abroad. More precise, these mediator platforms could influence the subsidiary ownership decision made by the MNE. This research gap is already mentioned by De Villa, Rajwani and Lawton (2015) and since it is an interesting topic to investigate possible effects, this thesis delves into the question if home and host country’s membership of the same mediator platform influences the relation between various levels of business inhibiting distance and MNEs’ pattern of subsidiary ownership.
This thesis starts with a deeper insight into the subject of distance, by highlighting Ghemawat’s (2001) four distance dimensions, namely cultural, administrative, geographic and economic distance. For each distance, the corresponding section will provide information about the conducted research. The succeeding section gives an overview of several entry modes and ownership strategies as well as their underlying theories. Subsequently, the next part informs about mediator platforms and especially about the Organisation for Economic Co-operation and Development (OECD), and concludes in an explanation of the research gap. The conceptual framework points out several hypotheses with and without the influence of the OECD on the relation between cultural, administrative, geographic and economic distance, and MNE’s subsidiary ownership percentage. The methodology part describes the conducted analysis, and contains the explanation of the outcomes of the hypothesis tests. The discussion comprises an interpretation of the results and tries to outline reasons for deviations from expected outcomes. The conclusion summarizes all findings and gives guidance for future research.
8
2. Literature review
The following section gives a brief overview of the literature of the distance dimensions, entry mode strategies, ownership patterns and the corresponding underlying theories. Thereafter, mediator platforms and their purpose will be explained, concluding in an outline of the research gap.
2.1. Distance
When a company is doing business abroad, it encounters difficulties, or to be more precise additional costs, which do not appear for an indigenous company. These extra costs which arise because of competitive disadvantages for MNEs’ subsidiaries are called Liability of foreignness (Hymer, 1976; Zaheer, 1995). Liability of foreignness emerges as a result of missing knowledge about the surroundings, prevailing differences in cultural, political and economic matters, and geographic distance, which complex communication and coordination (Zaheer, 1995).
A distance which is a composition of cultural, structural and linguistic differences is the psychic distance (Nordström & Vahlne, 1994). It can be described as aspects that have a positive or negative impact on the information exchange occurring between current or potential business partners (Nordström & Vahlne, 1994). Psychic distance consists inter alia of the differences between home and host country concerning economic development, education, business and local language, and culture (Vahlne & Wiedersheim-Paul, 1973).
Ghemawat (2001) combines in his studies the cultural, administrative, geographic and economic (CAGE) dimensions of distance in his CAGE-model.
The following part gives profound information about the distances of Ghemawat’s (2001) model.
2.1.1. Cultural Distance
A focal point in distance literature is the conducted research on cultural distance.Different languages, races, religions and also social norms, which are described by North (1990) as informal
9 rules, increase the cultural distance between two countries (Ghemawat, 2001). Cultural distance influences consumers’ product preferences, which are important for companies that work in the area of consumer goods or media. Hofstede (1983) analyzes cultural differences among countries by using four dimensions to illustrate differences in organization structures, in people’s incentives within organizations, and in people’s and organizations’ concerns within a society. Hofstede (1983) utilizes the cultural dimensions of Power Distance, Uncertainty Avoidance, Individualism and Masculinity, and adds in his later work Long-term Orientation (Hofstede, 1991; Hofstede & Bond, 1988) and Indulgence (Hofstede, Hofstede & Minkov, 2010).
The dimension of power distance describes the level of human inequality in social matters and also the level of authority distribution (Hofstede, 2011; Minkov & Hofstede, 2011). The uncertainty avoidance index accounts for people’s stress resistance when dealing with unknown or ambiguous, future-affecting events (Hofstede, 2011; Hofstede, 2016; Minkov & Hofstede, 2011). Individualism measures the importance of being an individual person compared to being part of the society (Hofstede, 2016; Minkov & Hofstede, 2011). Masculinity comprises the difference in emotional behavior and working style between genders (Hofstede, 2011, Hofstede, 2016; Minkov & Hofstede, 2011). Long-term orientation determines people’s time preference to plan over a long time range or to rather maximize utility in the present (Hofstede, 2011; Hofstede, 2016). Indulgence identifies people’s level of freedom to enjoy their life, and the level of society’s pressures to follow specific informal rules (Hofstede, 2011; Hofstede, 2016).
Several points of critique are expressed about Hofstede’s work because of his reduction of differences across nations to only cultural aspects (Ghemawat, 2001), his assumption that results of one company are representative for a country’s whole population (Lu, 2006; Smith 1996), his ecological fallacy to assume the behavior of specific managers is representative for the average population of a country (Berry, Guillén & Zhou, 2010) and his assumption of immutability of cultural distance across nations whereas researchers found evidence for its changeability (Inglehart & Baker, 2000; Shenkar, 2001; Webber, 1969).
10 Schwartz (1994) criticizes Hofstede’s (1980) cultural dimensions’ study to miss relevant questions in its underlying questionnaire, to generalize cultures with countries, as well as to not respect the varying meanings of words across countries, and consequently to ignore the differing understanding of the questions among countries. Schwartz’ (1994, 1999) conducts a separate research about cultural distance and results in seven dimensions: conservatism, intellectual and affective autonomy, hierarchy, egalitarian commitment, mastery, and harmony.
Another culture measuring research is conducted by the Global Leadership and Organisational Behaviour Effectiveness (GLOBE) project (House et al., 2002), which identifies seven cultural dimensions: Uncertainty avoidance, power distance, societal collectivism, in-group collectivism, gender egalitarianism, assertiveness, future orientation, performance orientation and humane orientation. The GLOBE dimensions are based on studies of Hofstede (1980, 1991), Kluckhohn and Strodtbeck (1961) and McClelland (1961), and although using various ways to determine the dimensions, the results are similar to prior findings and suggest a conceptual relation and an empirical correlation with the dimensions of Hofstede (1980, 1991).
2.1.2. Administrative Distance
Administrative circumstances of home and host countries are changeable over time and tremendously affect business. Administrative, political or institutional distance can be positively influenced by countries’ common political or historical unions (Ghemawat, 2001). Political ties like historical colony backgrounds, trade arrangements, political unions or common currency diminish administrative distance and hence facilitate business relations (Ghemawat, 2001; Henisz, 2000; La Porta et al., 1998; Whitley, 1992).
The opposite effect can be seen, when dealing with political hostility, institutional weakness caused by corruption or social conflicts, or inhibiting government policies such as constraints on FDIs, tariffs and trade quota (Ghemawat, 2001). Administrative issues can arise through home countries’ restrictive legislative changes as well as through target countries’ barriers to foreign
11 companies, both resulting in consequences inhibiting international business (Ghemawat, 2001). Weak institutional infrastructure is another negative influence factor for cross-border economic activities, since countries with corruption and social issues are not as aspirational as countries with strong institutional infrastructure and a proper working legislative system (Ghemawat, 2001). The regulative restrictiveness varies among countries (Ang & Michailova, 2008) as well as the structure and in-depth explanations of the regulatory process for foreigners (Chao & Kumar, 2010). Understanding the administrative environment and following its rules when doing business in a foreign country is a crucial factor to increase the possibility of a company’s success in order to maintain a player in the market (Dikova, Sahib, & van Witteloostuijn, 2010).
2.1.3. Geographic Distance
Geographic distance consists of country size, distance from border to border within a country, waterways and ocean access, topographical characteristics, and also of countries’ infrastructure for communication, information and transportation (Ghemawat, 2001). It affects communication and transportation costs and therefore, it is an important factor for high-coordination businesses or companies handling heavy goods (Ghemawat, 2001). Geographic distance is supposed to influence cross-border business activities, such as trade or FDIs (Anderson, 1979; Deardorff, 1998) and also gravity model supporters in international business and trade are convinced of its importance for MNEs’ decisions (Fratianni & Oh, 2009; Hamilton & Winters, 1992).
Due to its long recognition, there are various ways of measuring geographic distance among countries (Berry, Guillén & Zhou, 2010). For example, Krishna (2003) calculates the geographic distance by the usage of the direct line distance with the great circle method which utilizes the coordinates of the geographic centers of each country. Lòpez-Duarte and Vidal-Suárez (2013) use the flight distance between capitals to measure geographic distance among countries. Another example is Chen (2004) who decides to measure geographic distance according to the countries’ latitude as
12 well as longitude of regional capitals and concludes that a higher distance diminishes international trade among countries.
2.1.4. Economic Distance
Economic distance describes the differences of economic conditions among countries, such as consumers’ level of income and prosperity which also affects cost and quality of several resources, and the level of knowledge and information (Ghemawat, 2001). Although economic distance complex the business relations, some companies take advantage of this arbitrage by successfully shifting labor-intensive tasks to low-labor cost countries (Ghemawat, 2001).
Economic distance, similar to administrative distance, is changeable over time (Inglehart & Baker, 2000; Shenkar, 2001). It is determined by several indicators such as gross domestic product (GDP) per capita which describes the income level, current inflation rates, and cross-border trade intensity which is calculated by the ratio of the sum of exports and imports, and GDP (Berry, Guillén & Zhou, 2010).
2.2. Entry Modes and Ownership Levels
Since the above mentioned kinds of distances are important influence factors when doing business abroad, distance also affects the choice of entry mode or more general the ownership pattern MNEs follow when dealing with subsidiaries (Brouthers & Brouthers, 2003; Contractor et al., 2014; Erramilli & Rao, 1993; Hennart & Larimo, 1998).
Hennart (1988, 1989, 2000) divides the entry mode possibilities in two categories: contract-modes, where the payment is conducted ex ante, and equity-contract-modes, where partners are paid ex post from profits of the venture. Similar to Hennart (1988, 1989, 2000), Pan and Tse (2000) use non-equity and equity modes to divide the possibilities of entry choice. Non-equity modes are considered as direct as well as indirect exporting, and contractual agreements like licensing, research and development (R&D) contracts, and alliances (Pan & Tse, 2000). Equity entry modes are considered
13 as equity joint ventures (JV) such as majority equity JV, minority equity JV as well as 50% share equity JV, and wholly owned subsidiaries (WOS), which can further be divided into greenfield and acquisition (Pan & Tse, 2000).
Diverse theories are applied to find explanatory approaches for entry mode decisions. One theory is the resource-based view (RBV) which is based on the idea of the dependence of a firm’s lasting success on its possession of valuable, rare, imperfectly imitable and non-substitutable resources or capabilities (Barney, 1991). RBV is mainly concerned with experience of internationalization which is a focal point in Johanson and Vahlne’s (1977) internationalization theory. This theory, also called the Uppsala model (Johanson & Vahlne, 1977), describes the development of doing business abroad as a series of little steps which incrementally increase firms’ international involvement. It suggests to start with exports and develop towards a WOS and begin in countries which are more familiar, and develop to countries which are less familiar to the MNE (Johanson and Vahlne, 1977). Engaging their research into internationalization theory, Anderson and Gatignon (1986) as well as Erramilli and Rao (1990) set up entry mode decisions on a continuum of increasing extent of commitment, control and risk. This means that export and contractual agreements possess the lowest rates, joint ventures have intermediate rates and wholly owned subsidiaries (WOS) own the highest rates of commitment, control and risk (Brouthers & Hennart, 2007). Researchers add the field of managerial control, which needs to be intensified the riskier the entry mode (Hill, Hwang, & Kim, 1990) and the higher the resource commitment (Hennart, 1989).
Researchers use resource advantages based on experience (Aulakh & Kotabe, 1997; Cho & Padmanabhan, 2001; Delios & Henisz, 2000; in Brouthers & Hennart, 2007) such as international experience’s duration and extent (Erramilli, 1991) to measure the degree of international experience. Erramilli (1991) discovers a U-shaped relationship between the duration as well as the extent of international experience, and choice of entry mode, examining that less experience and much experience are related to modes of full control, and mean level of experience leads to modes which
14 are market-based. Aulakh and Kotabe (1997), Cho and Padmanabhan (2001), and Delios and Henisz (2000) investigate the influence of experience-related resource benefits on choosing an entry mode and discover results comparable to those of Erramilli (1991).
Another approach for entry mode decisions is the transaction cost theory (TCT) which is based on the thought, that organizations’ existence is dependent on their superior skills of hierarchical monitoring to diminish opportunistic behavior of their employees (Goshal & Moran, 1996). TCT points out that managers’ behavior and decisions are influenced by their bounded rationality which means that possible partners could act opportunistically if they had the chance to do so (Brouthers & Hennart, 2007).
As potential influences on managers’ decisions, Williamson (1985) examines three dimensions, asset specificity, uncertainty and frequency (Geyskens, Steenkamp & Kumar, 2006). Asset specificity can be measured through R&D and/or advertising intensity or investments in technology asset specificity (Palenzuela & Bobillo, 1999), physical asset specificity and human asset specificity (Brouthers & Brouthers, 2003; Erramilli & Rao, 1993). Numerous researchers support Williamson’s (1985) idea and find evidence that asset specificity is an influencing factor towards high control modes (Brouthers & Brouthers, 2003; Brouthers, Brouthers & Werner, 2003; Erramilli & Rao, 1993; Gatignon & Anderson, 1988; Hennart & Larimo, 1998) but Delios and Beamish (1999) discovered an effect in the opposite direction. Thus, it can be concluded that a significant connection between asset specificity and the choice of entry modes exists, but its effect is ambiguous (Brouthers & Brouthers, 2003; Delios & Beamish, 1999; Gatignon & Anderson, 1988; Hennart & Larimo, 1998).
Uncertainty is divided by Williamson (1985) into external-market specific and internal-behavioral uncertainty. External uncertainty impedes consideration of all possible influences before contract conclusion (Brouthers & Hennart, 2007). According to Zhao, Luo and Suh (2004) the most usual factors here are country risk and cultural distance. Delios and Beamish (1999) used the Euromoney Country Risk Index as measurement for country risk in their research. Whereas Kim and
15 Hwang (1992) and Makino and Neupert (2000) used industry growth and industry concentration ratio in their studies. The perceived political and economic stability is also used to measure country risk (K. D. Brouthers, 2002; Brouthers and Brouthers, 2003; Kim and Hwang, 1992). Internal-behavioral uncertainty complicates the performance measurement after a contract conclusion (Brouthers & Hennart, 2007) and consists of both international experience (Zhao, Luo & Suh, 2004) and also non-experience-based factors such as perceived complication in finding the right business partner and the anticipated capability of enforcement, control and supervision over it (K. D. Brouthers, 2002; Brouthers & Brouthers, 2003; Brouthers, Brouthers & Werner, 2003). International experience can be measured through the duration of international operations (Contractor & Kundu, 1998), the total amount of foreign investments (Delios & Beamish, 1999; Gatignon & Anderson, 1988; Gomes-Casseres, 1989) or the ratio of foreign to total number of investments (Contractor & Kundu, 1998).
Researchers’ findings about the influence of experience are blended. Gatignon and Anderson (1988) and Hennart (1991) find evidence that MNEs who possess international experience are less likely to establish a JV in a foreign country. Whereas Gomes-Casseres (1989) and Padmanabhan and Cho (1996) indicate the opposite findings. Zhao, Luo and Suh (2004) illustrate a tendency of experienced MNEs towards WOS. Researchers’ results about non-experience based influences are also mixed such as a relation between higher internal-behavioral uncertainty and choosing a WOS (K. D. Brouthers, 2002), a relation between lower internal-behavioral uncertainty and the choice of WOS when dealing with service firms (Brouthers & Brouthers, 2003) and with manufacturing firms (Brouthers & Brouthers, 2003), and also an evidence for no existing connection between internal-behavioral uncertainty and the ownership choice (Brouthers, Brouthers & Werner, 2003).
Williamson’s (1985) third dimension, frequency, implies firms’ decisions between using internalization or contracting partners in the market (Brouthers & Hennart, 2007). Measurements of frequency are channel volume (Klein, Frazier & Roth, 1990; in Brouthers & Hennart, 2007) and perceptual activity (Taylor, Zou & Osland, 1998; in Brouthers & Hennart, 2007). Due to the fact that the incorporation of processes carries high costs, internalization can only be warrantable if the amount
16 of conducted operations is high enough (Brouthers & Hennart, 2007). This affects the decision between contractual agreements and equity modes but it is still unclear whether it also influences the decision between WOS and JV (Brouthers & Hennart, 2007).
The institutional theory is another approach which analyzes the institutional environment that influences the entry mode choice (Brouthers & Hennart, 2007). Most researchers focus on host countries’ institutional environment or disparities between home and host country (Brouthers & Hennart, 2007) but some also concentrate on institutional aspects of the home countries (Erramilli, 1996; Pan, 2002). Institutional theory started with an emphasis on risk of host countries and perceived uncertainty but broadens the view with Scott’s (1995) approach of the new institutional theory. Scott (1995) divided the institutional framework in his studies into three country-depending ‘pillars’, namely the regulatory, cognitive and normative dimension. According to Scott (1995), the ability to influence social behavior by introducing rules, monitoring units and punitive measures belongs to the regulatory part. The cognitive dimension consists of the establishment of a common conception of social life which is culturally supported (Scott, 1995). Consequently, the normative dimension covers the establishment of moral expectations which guides social behavior by prescriptions, evaluations and obligations (Scott, 1995; Scott, 2003).
Several researchers adopt Scott’s (1995) subdivision and investigate the impact of these areas on the choice of entry mode. Yiu and Makino (2002) discover a direct effect of regulatory, cognitive and normative dimensions on the choice of entry mode. Brouthers, Brouthers and Werner (2002) determine important influences of competition, government policy, macroeconomic, material and product on the entry mode decision. The studies of Davis, Desai and Francis (2000), and Lu (2002) examine the cognitive level and the burden of isomorphism, a homogenization process (DiMaggio & Powell, 1983) caused by environmental pressures that push firms to resemble other firms in the same surroundings (Hawley, 1968). Their results vary but they discover a tendency of adaptation to isomorphic pressures when choosing the mode of entry (Brouthers & Hennart, 2007). The regulatory
17 dimension and its influencing role in the explanation of the entry mode choice is indicated by Meyer (2001), K. D. Brouthers (2002) and Meyer and Nguyen (2005). Moreover, Uhlenbruck et al. (2006) find evidence in their research for an existing significant relationship between host countries’ corruption and the selection of the entry mode.
Another important factor in the institutional theory is cultural distance and therefore its inclusion is widespread (Barkema & Vermeulen, 1997; Brouthers & Brouthers, 2001; Yiu & Makino, 2002). Among others, Yiu and Makino (2002), Hennart and Larimo (1998), and Kogut and Singh (1988) find evidence that cultural distance, which mirrors disparities of Scott’s (1995) normative dimension of different countries, is a big component when looking for an explanation of a chosen entry mode. Using Hofstede’s (1983) cultural dimensions, Barkema and Vermeulen (1997), and Shane (1993, 1994) find evidence for an influencing effect of Power Distance and Uncertainty Avoidance on the choice of entry mode but Brouthers and Brouthers (2001) point out that the influence is more likely indirect due to risk of investment and, according to Cho and Padmanabhan (2005), due to decision-specific know-how.
An additional approach to examine the choice of entry mode is Dunning’s (1993) eclectic paradigm which tries to combine TCT, RBV and institutional theory in order to discover possible interactions between these theories (Brouthers & Hennart, 2007). Dunning (1993) determines his eclectic paradigm to possess an ownership, location and internalization (OLI) framework focusing on specific OLI advantages (Brouthers & Hennart, 2007). Agarwal and Ramaswami (1992) indicate the influence of OLI dimensions on the choice of entry mode but also single dimensions are investigated. The influence of ownership advantages on the entry mode, such as decision-specific know-how (Padmanabhan & Cho, 1999), as well as its location-dependency is investigated. Hence, Erramilli, Agarwal and Kim (1997) conclude that ownership advantages have a varying effect on choosing the right entry mode in diverse countries. According to Brouthers, Brouthers and Werner (1999), MNEs
18 that include the OLI advantages in their decision process of entry mode, are more likely to display higher performance results.
In addition to the previous rational theories, Erramilli and Rao (1990), Shi, Ho and Siu (2001), Tsai and Cheng (2002), and Gil et al. (2006) suggest the influence of different entry motives on the choice of how to establish organizational structures abroad (Brouthers & Hennart, 2007). Furthermore, numerous researchers (Aulakh & Kotabe, 1997; Barkema & Vermeulen, 1997; Brouthers, Brouthers & Werner, 1999; Delios & Beamish, 1999; Erramilli, 1991; Erramilli & Rao, 1993; Erramilli, Agarwal & Kim, 1997; Hennart, 1991) analyze potential home and host country effects on entry mode choices but conclude that if country and/or institutional differences were considered, neither home nor host country characteristics make MNEs’ choices deviate from settled theories of entry modes (Brouthers & Hennart, 2007). Aulakh and Kotabe (1997), Brouthers, Brouthers and Werner (1999), Brouthers, Brouthers and Werner (2000, 2003), and K. D. Brouthers (2002) notice that firms which choose entry modes according to transaction cost and institutional theories analysis’ suggestion, result in better performance than firms applying different entry modes (Brouthers & Hennart, 2007).
The entry mode decision is the first step of the ownership pattern of an MNE. Even though a company decides for a specific entry mode, it might decrease or increase its subsidiary ownership later. Researchers examine several influence factors of the ownership behavior of MNEs, such as firm-specific assets, size of the investing firm, its experience in the specific host country, political risk of the host country, and industry concentration (Brouthers & Brouthers, 2003; Delios & Beamish, 1999; Erramilli & Rao, 1993; Hennart & Larimo, 1998).
Hennart (1991) investigates the choice of JV and WOS by using the TCT in an analysis of subsidiaries based in the United States (US) of Japanese firms. In determining variables that influence costs and benefits of starting a JV compared to a WOS, he concludes that Japanese firms are more likely to use a JV instead of a WOS when 1) diversifying into an unexperienced industry, 2) having
19 limited experience in local market, 3) needing admittance to natural resources, or 4) entering a high-growth industry (Hennart, 1991). Hennart’s (1991) suggestion is that Japanese firms’ preference taking JVs over WOSs is higher when they need to incorporate high transaction cost assets from local US firms, such as attendance to networks of distributors and natural resources, market, and industry-specific knowledge (Hennart, 1991). Gomes-Casseres (1990), Hill, Hwang and Kim (1990), Kim and Hwang (1992), and Delios and Beamish (1999) generally support Hennart’s (1991) explanations of TCT influences on the ownership mode choice but also highlight the important influence of non-transaction cost variables, that belong to the institutional theory, in terms of restrictions of host government, strategic factors, host country risk and host country uncertainty (Makino & Neupert, 2000).
Despite the supportive evidences of the TCT, Parkhe (1992) questions TCT’s assumption of managers’ propensity to opportunistic behavior and contend that their propensity of trust differs significantly depending on the nationality of their partners with whom they build a strategic alliance with (Makino & Neupert, 2000). According to Shane (1994), the nation-dependent level of trust impacts the perceived transaction costs and thus, it affects the appeal of internalization and also the entry mode choice (Makino & Neupert, 2000). Erramilli (1996) identifies cultural and economic variables representing nationality and examines their impact on subsidiary ownership preferences among multinational advertising agencies.
Hennart and Larimo (1998) build on Erramilli’s (1996) research and focus on the impact of national culture scores and its cultural score distance between home and host countries on subsidiary ownership strategies. They investigate MNEs based in Japan and Finland which establish subsidiaries in the US. They examine two theories: the “National Character” theory, and the TCT with the role of cultural distance. The prior one uses Shetty’s (1979) view of each country being specific regarding its psychological characteristics and its influence on location-specific ownership patterns of MNEs. Applying this theory by using Hofstede’s (1980) cultural dimensions, this indicates that the lower the Uncertainty Avoidance and Power Distance of MNEs’ home countries, the higher the probability that
20 the MNE will use shared-equity mode (Hennart & Larimo, 1998). Hennart and Larimo (1998) find no support for the national psychological characteristics to be an influencing factor for ownership modes but they do find support for cultural distance having a strong impact on MNEs’ ownership inclination of foreign affiliates.
2.3. Mediator Platforms and Research Gap
De Villa, Rajwani and Lawton (2015) investigate the influence of the political environment on the market entry modes in greater depth and examine future research possibilities concerning the potential influences of trade associations, that work as mediator platforms, on MNEs’ ownership decisions. Since it is crucial for firms that enter a new country to be informed about the host country’s environmental influences, but the wealth of information is hard to get and even harder to apply, mediator platforms are a potential knowledge provider. Additionally, if contributing to trade agreements, they also work as business facilitators. This additional knowledge of the new market and the potential positive influence on policies or trade agreements could affect the ownership mode of MNEs’ subsidiaries, since difficulties with a market’s cultural, administrative, geographic and economic situation are important aspects of the perceived distance between home and host country.
Mediator platforms can act as a contributor to circumvent distance and to improve the information flow which is needed to go abroad successfully and be competitive. They can be divided into national mediator platforms and international ones.
An example for a national mediator platform is the country-focused German chamber of commerce, the “Deutsche Auslandshandelskammern”. The network of German chambers of commerce represents a mediator platform which supports German companies that go abroad, or foreign companies that want to start business in the German market. Each country that is economically important for Germany has its own chamber of commerce organization. The German chambers of commerce can be found at 130 locations in 90 countries and represent German economic interests (Deutsche Auslandshandelskammern).
21 An example for an international mediator platform is the World Trade Organization (WTO) which is an organization that works on a global international level on cross-national trade regulations to facilitate conducting business for several participants of the world market (World Trade Organization (2)). It consists of 162 member states whereby some countries had already been member of the pre-WTO organization, General Agreement on Tariffs and Trade, also called GATT (World Trade Organization (1)).
In addition to that, another international mediator platform is the Commonwealth that consists of 53 countries and which’s organization inter alia supports countries with trade negotiations to improve the small business sector (Commonwealth (1); Commonwealth (2)). Furthermore, the Commonwealth Secretariat tries to improve economical and societal developments, and strengthen democratic systems, rule of law and well-working governance (Commonwealth (2)).
The Organisation for Economic Co-operation and Development is another international mediator platform which provides a platform for governments with the purpose of sharing experiences, and the mission of promoting policies that support the improvement of the prosperity of humans all over the world in economic and social matters (Organisation for Economic Co-operation and Development (1)). The OECD was founded in 1961 but originates from the former Organisation for European Economic Cooperation which was initially meant to support the Marshall Plan concerning Europe’s reconstruction after World War II (Organisation for Economic Co-operation and Development (3)). Today’s OECD consists of 34 member countries around the world, developed countries as well as emerging ones (Organisation for Economic Co-operation and Development (1)).
In order to evaluate the influence mediator platforms exert, the OECD is taken as a prime example in this analysis since it might be powerful enough to affect MNEs’ business strategies. Therefore, the influence of the OECD as a mediator platform on the CAGE distances between home and host country and its effect on ownership patterns of MNEs is an interesting field for which research needs to be done.
22
3. Conceptual Framework
The following part continues with an explanation of the conceptual framework of the conducted research and the development of hypotheses.
3.1. Hypotheses Development – Cultural Distance
By using Hofstede’s (1980) cultural dimensions, Erramilli (1996) finds support for MNEs which are based in countries that are characterized by high Power Distance and low Uncertainty Avoidance to have an intrinsic inclination for full ownership of their subsidiaries abroad. Hennart and Larimo (1998) also mention a TCT-based assertion that MNEs with higher cultural distance between home and host country are more likely to choose shared-equity modes. Hennart (1988) argues that TCT determines the degree of ownership by executing a cost-benefit analysis of shared ownership to full ownership. High transaction costs, meaning costly purchasing and contracting, and higher replication costs, compared to buying the usage right through starting a shared-equity venture, make the option of sharing equity useful if a company wants or needs to implement firm-specific assets of another company (Hennart & Larimo, 1998). High transaction costs are expected when dealing with tacit knowledge about technology and country-specificities, good reputation and distribution networks (Hennart, 1982; in Hennart & Larimo, 1998). This suggests that the higher the cultural distance between MNE’s home and host country, the higher the probability of using shared-equity modes (Hennart & Larimo, 1998).
Makino and Neupert (2000) further examine transaction cost and cultural influences on the choice of ownership levels. Makino and Neupert’s (2000) interest is driven by the question whether TCT is able to clarify that choice regardless the foreign investing firm’s home country, or whether it is abated or modified by the investing firm’s national characteristics. They find supportive evidence for Erramilli’s (1996) research, that the propensity of the choice between JV and WOS is influenced by the national culture perspective. Moreover, their results propose that the decision between JV and WOS is also affected by transaction cost factors regardless the origin of the investing MNE (Makino
23 & Neupert, 2000). Since Erramilli (1996) uses Power Distance and Uncertainty Avoidance as indicators for national culture, Makino and Neupert (2000) repeat this in their study, but still, it is questionable whether these two variables are able to detect the national culture of a whole country.
MNEs that have to deal with cultural distance are involved in handling local social routines that lead to implicit assumptions due to ambiguity and unfamiliarity, which causes the need for adaptation and adjustment (Chakrabarti, Gupta-Mukherjee & Jayaraman, 2009; Hofstede, 2001; Morosini, Shane & Singh, 1998). The higher the distance, the greater the perceived uncertainty and risk of the feasibility of the transfer of business processes and the establishment of sustainable relationships between local stakeholders (Contractor et al., 2014).
According to the above mentioned results of various researchers, subsidiary ownership is expected to be lower when dealing with a subsidiary in a culturally distant country. Hence, the first hypothesis is as follows:
Hypothesis 1: The relationship between home and host country’s cultural distance and the ownership percentage of MNEs’ subsidiaries abroad is negative.
Being a member of a mediator platform could increase information exchange about market know-how among countries and consequently, a foreign MNE would not be as dependent on external knowledge of a JV partner as without being member. Furthermore, cultural differences might be already known and solutions might already been spread among mediator platform members. Thus, this distance can be decreased by applying proven plans for overcoming it. Therefore, the initiatives of mediator platforms are expected to have an increasing effect on the ownership level. Hence, the second hypothesis is developed:
Hypothesis 2: The relationship between home and host country’s cultural distance and the ownership percentage of MNEs’ subsidiaries abroad is moderated by countries’ mediator platform membership, so that if home and host country are members of the same mediator platform, an MNE will be more likely to establish higher subsidiary ownership.
24
3.2. Hypotheses Development – Administrative Distance
Administrative distance indicates the difference between the MNE’s home and host country’s institutional environment (Dikova, Sahib & van Witteloostuijn, 2010; Peng, Wang & Jiang, 2008). It influences MNEs especially in their decision about the ownership structure of their subsidiaries (Contractor et al., 2014). The greater the distance, the less familiarity of the other country’s business systems and institutions and thus, the greater the uncertainty about a successfully feasible exchange of resources and processes from the MNE to the host country’s subsidiary (Chen & Hennart, 2004; Contractor et al., 2014; Demirbag, Glaister & Tatoglu, 2007). Moreover, the more different and complex an administrative system, the more time is needed to adapt to it, whereby the time needed is also critical for the success of the investment (Contractor et al., 2014; Dikova, Sahib & van Witteloostuijn, 2010; Peng, Wang & Jiang, 2008). Xu, Pan, and Beamish (2004) demonstrate that the bigger the regulative and normative institutional distance, the lower is the ownership percentage of foreign subsidiaries. Also a gap between corruption and transparency levels among home and host country is more likely to result in a lower ownership percentage compared to a wholly owned subsidiary (Demirbag, Glaister & Tatoglu, 2007). Hence, the third hypothesis concludes:
Hypothesis 3: The relationship between home and host country’s administrative distance and the ownership percentage of MNEs’ subsidiaries abroad is negative.
The membership of a mediator platform is expected to affect the subsidiary ownership level as a moderator. By guiding MNEs with recommendations for action and also leading trade negotiations, the potential risk is lowered and MNEs become more familiar with host countries’ systems. Furthermore, mediator platforms act as a consulting tool for policy makers which advise proven ways of advancement (Alasuutari, 2005). By using this method, and combining it with knowledge production that shows equal indicators and concepts among countries, a growing convergence is the result (Alasuutari, 2005). Thus, the fourth hypothesis is developed:
25 Hypothesis 4: The relationship between home and host country’s administrative distance and the ownership percentage of MNEs’ subsidiaries abroad is moderated by countries’ mediator platform membership, so that if home and host country are members of the same mediator platform, an MNE will be more likely to establish higher subsidiary ownership.
3.3. Hypotheses Development – Geographic Distance
Buckley and Casson (1979) as well as Mariotti and Piscitello (1995) find evidence for vicinity causing a reduction of entry barriers, of transportation issues, and of conditions among information processing (Shenkar, 2001). Vachani (1991) states the importance of individual contact among employees for a successful dissemination and exchange of know-how and other valuable resources. Proximity decreases the effort, hence it diminishes the expenses for control and coordination among the management as well as the expenses for monitoring representatives’ conduct (Shenkar, 2001). Since a geographic closer host country might embody less risk and less effort for MNEs, an MNE is expected to use higher subsidiary ownership levels with decreasing geographic distance.
Thus, the fifth hypothesis is as follows:
Hypothesis 5: The relationship between home and host country’s geographic distance and the ownership percentage of MNEs’ subsidiaries abroad is negative.
Since the membership of a mediator platform affects the trade activity of countries (Tomz, Goldstein and Rivers, 2007), a proper infrastructure within a country and between trading countries might gain more importance. By improving accessibility of the country, crucial approach roads but also flight connections might be enhanced. Therefore, a moderating effect is anticipated and the sixth hypothesis is the following:
26 Hypothesis 6: The relationship between home and host country’s geographic distance and the ownership percentage of MNEs’ subsidiaries abroad is moderated by countries’ mediator platform membership, so that if home and host country are members of the same mediator platform, an MNE will be more likely to establish higher subsidiary ownership.
3.4. Hypotheses Development – Economic Distance
Economic distance’s indicators, such as income level, current inflation rates and cross-border trade intensity, correlate with purchasing power, economic stability as well as the candidness to influences from abroad (Berry, Guillén & Zhou, 2010; Caves, 1996). These factors in turn affect market entry modes, and the survival and performance of a company (Berry, Guillén & Zhou, 2010; Caves, 1996). The effect of economic distance on an MNE’s choice of host country and its entry mode strategy is already confirmed (Yeung, 1997; Zaheer & Zaheer, 1997). Also Mitra and Golder (2002) analyze the discouraging effect of economic distance on foreign market entry. Additionally, it might be easier for MNEs to transfer their current already existing business model into a WOS in an economically closer host country, since and customers’ prosperity level will be similar. Therefore, the seventh hypothesis can be concluded:
Hypothesis 7: The relationship between home and host country’s economic distance and ownership percentage of MNEs’ subsidiaries abroad is negative.
Trade negotiations conducted by the mediator platform could lead to trade agreements which in turn result in an assimilation of economic levels. Tomz, Goldstein and Rivers (2007) show in their analysis concerning effects of GATT membership, that GATT membership of a country accounts for a tremendous increase in trade. Furthermore, Rose’s (2005) research also confirms a positive economic effect of OECD membership on trade. Growing trade affects the prosperity of societies and might lead to an adjustment of well-being. Through a home and host country’s membership of the
27 same mediator platform, the effect of economic distance might be reduced. Consequently, the eighth hypothesis is as follows:
Hypothesis 8: The relationship between home and host country’s economic distance and the ownership percentage of MNEs’ subsidiaries abroad is moderated by countries’ mediator platform membership, so that if home and host country are members of the same mediator platform, an MNE will be more likely to establish higher subsidiary ownership.
The following model provides an overview of all hypotheses.
4. Methodology
The next part describes the sample as well as the data collection and calculation of each variable. Subsequently, several tests check the fulfillment of the assumptions to finally conduct a regression analysis. Such a regression analysis is necessary to assess a possible moderating effect of mediator platform membership on the relation between the CAGE distances and the MNEs’ ownership level of the subsidiaries.
28
4.1. Sample description and data
The analysis for this study is based on a sample of MNEs that have at least one subsidiary abroad. To be able to neglect possible home country effects and to get comparable results, a fixed MNE’s home country is necessary. Therefore, the MNEs’ home country is defined as Germany. To meet the statistical requirements and ensure the availability of the data, the OECD is determined as mediator platform. Here, the data of more than 25 member countries and more than 25 non-member countries is accessible for all variables. To obtain a broad setting for this research and use as many OECD member and non-member countries as possible, there are no further restrictions made. To be included in the analysis, a country needs at least 25 subsidiaries of German MNEs in the dataset to meet the statistical requirements. The final sample consists of 3380 German MNEs with a total of 39282 subsidiaries in 65 countries, of which 32 are OECD members and 33 are OECD non-members. A list of all host countries is located in the appendix.
4.2. Variables
Contractor et al. (2014) extract their company-level data from AMADEUS, a pan-European database which is focused on public and private enterprises concerning 43 countries. Since the mediator platform used here, the OECD, is a global platform, it is reasonable to take a database with a broader range of country data. This leads to Orbis database, which delivers data of listed and unlisted enterprises worldwide, comprising detailed information about the country, revenue and number of employees for the MNE as well as the subsidiary.
The dependent variable, Subsidiary Ownership, indicates the ownership percentage an MNE possesses of its subsidiary and is extracted form Obis database. The information concerning the moderating variable, the host country’s OECD membership, is extracted online (Organisation for Economic Co-operation and Development (2)). The control variables of MNE Revenue and its MNE Employees, which are also extracted from Orbis database, are used to verify the effect of the independent variables on the dependent variable.
29 The independent variables of Cultural, Administrative, Geographic and Economic Distance originate from several sources and are emphasized in the following part.
Cultural Distance consists of Hofstede’s dimensions (Hofstede, 1983; Hofstede, 1991; Hofstede & Bond, 1988; Hofstede, Hofstede & Minkov, 2010), namely power distance, individualism, masculinity, uncertainty avoidance, long-term orientation and indulgence, and the corresponding values are based on data which is extracted from Geert Hofstede’s website (Hofstede, 2016).
Despite criticizing internal validity and way of scale rating of Hofstede’s dimensions, Kogut and Singh (1988) compile an index to calculate the cultural distance from a specific country to another by using Hofstede’s (1980) first four dimensions, indicating the higher the score, the greater the cultural distance. This procedure is used by several researchers (Barkema & Vermeulen, 1998; Brouthers & Brouthers, 2000; Cho & Padmanabhan, 1995). Another index for combining cultural dimensions is the Euclidean Distance which is used by Barkema & Vermeulen (1997), Brouthers & Brouthers (2001), and Vermeulen & Barkema (2001).
The dimensions of Hofstede (Hofstede, 1983; Hofstede, 1991; Hofstede & Bond, 1988; Hofstede, Hofstede & Minkov, 2010) and those of Schwartz (1994) are tested by Drogendijk and Slangen (2006) using both the Kogut-Singh-Index (1988) and the Euclidean Distance. Their results show an existing comparability of Hofstede’s (Hofstede, 1983; Hofstede, 1991; Hofstede & Bond, 1988; Hofstede, Hofstede & Minkov, 2010) and Schwartz’ (1994) dimensions, as well as of the Kogut-Singh-Index (1988) and the Euclidean Distance. Since the Kogut-Singh-Index (1988) accumulates the cultural differences for each dimension in square divided by each dimension’s variance and then divide it by the amount of dimensions used, it is no problem to extend this index from four to six cultural dimensions. This procedure is wide-spread among researchers (Barkema, Bell, & Pennings, 1996; Barkema & Vermeulen, 1997; Benito & Gripsrud, 1992; Morosini, Shane & Singh, 1998) and therefore it is adapted to this study.
30 Adapted Kogut-Singh (1988) formula:
𝐶𝐷
#=
&'()&'* + ,' -'./ 0Here, CDj stands for the cultural distance of host country j and Iij refers to the index of cultural dimension i of host country j. The index h refers to the home country h, which in this study is Germany. Vi is the variance of the dimension-specific index i.
The Administrative Distance is measured among 11 years through the world bank’s six government indicators: voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law, and control of corruption (World bank, 2016). The measurement range of the government score goes from -2.5 to +2.5, meaning the higher the score the more advanced the institutional setups. To sum up the six dimensions and calculate the Administrative Distance between home and host countries, the Morosini, Shane and Singh’s (1998) formula is utilized. This working method is also used by other researchers (Contractor et al., 2014; Dikova, Sahib, & van Witteloostuijn, 2010) and hence it is adopted to this study. By using Morosini, Shane and Singh’s (1998) formula, the administrative distance for each year is calculated, and to avoid discrepancies, the mean over the 11 years is taken.
Morosini, Shane and Singh’s (1998) formula:
𝐴𝐷
2#=
0(𝐼
52567
− 𝐼
5#)
:Here, ADhj refers to the Administrative Distance between home country h, here Germany, and host country j. Iih is the score of governance indicator i of the home country h, and Iij is the score of governance indicator i of host country j.
The Geographic Distance is measured through the flying distance in kilometers from home country’s capital, here Germany’s capital Berlin, to the host country’s capital, and then the logarithm is taken. This process is also used by Lòpez-Duarte and Vidal-Suárez (2013) and the data is extracted from an online database (Worldatlas, 2016).
31 The Economic Distance is measured by using the ratio of the host country’s GDP growth rate to home country’s GDP growth rate for each of the 11 years and then the median is taken. Although Lòpez-Duarte and Vidal-Suárez (2013) and Contractor et al. (2014) use the difference between home and host GDP growth rates, here it is decided to use the median of the ratio of GDP growth rates among all years in order to guarantee comparability between each country.
Some notes need to be made to follow the analysis. For many companies there is only direct subsidiary ownership given but no information about total ownership. Without further knowledge, the total ownership is not computable because unknown indirect ownership connection could be existent. The available direct ownership demonstrates the MNE’s minimum percentage of total ownership of the subsidiary. If the direct ownership is available but the total ownership is missing, the direct ownership is used as ownership indicator to maintain the quantity requirements of a proper data analysis.
4.3. Analytical Approach
Firstly, frequencies of the data are checked to spot missing values of the required variables. If data is missing, the corresponding unit is excluded from the analysis. The final data is checked for outliers but none extreme outlier can be identified. Subsequently, the dichotomous categorical variable of OECD Membership is recoded to “1 = OECD member” and “0 = OECD non-member” to run the analysis.
Secondly, the database is used to extract information on descriptive statistics. Table 1 shows the minimum, maximum, mean and standard deviation of the variables. 78% of the companies which are part of the database are located in countries that are OECD members, the rest is located in OECD non-member countries.
32
Variables N Minimum Maximum Mean Std. Deviation
MNE Revenue 39282 -447142 253890289 25526022.5 44252647.4 MNE Employees 39282 1 592,586 73562.95 119572.244 OECD Membership 39282 0 1 .78 .411 Cultural Distance 39282 .34 5.36 1.61 .848 Administrative Distance 39282 .28 41.28 5.06 7.33 Geographic Distance 39282 5.82 9.81 7.591 1.123 Economic Distance 39282 -.43 2.5 1.092 .518 Subsidiary Ownership 39282 .01 100 78.779 36.452
Table 1 Descriptive Statistics
When checking the correlations between the variables, see table 2, the strongest positive correlation is found between MNE Revenue and MNE Employees (r = .842), which is statistically significant (p < .001). This correlation is absolutely transparent due to the corollary of the relationship between number of employees and the earned revenue. Even though both variables correlate highly with each other, depending on the industry both variables have meanings which are different enough from each other to keep them both in the analysis. The second strongest correlation is between Administrative Distance and Economic Distance (r = .613) and it is also statistically significant (p < .001). The strongest negative correlation occurs between OECD Membership and Administrative Distance (r = -.780) and is also statistically significant (p < .001). The second strongest negative correlation is between OECD Membership and Economic Distance (r = -.511) which is also statistically significant (p < .001).
33
Variables Mean SD 1 2 3 4 5 6 7 8
1. MNE Revenue 25526022.5 44252647.4 Pearson -
Sig. (2-tailed) - 2. MNE Employees 73562.95 119572.244 Pearson .842 - Sig. (2-tailed) .000 - 3. OECD Membership .78 .411 Pearson -.006 -.039 - Sig. (2-tailed) .202 .000 - 4. Cultural Distance 1.61 .848 Pearson .042 .064 -.310 - Sig. (2-tailed) .000 .000 .000 - 5. Administrative Distance 5.06 7.33 Pearson -.005 .025 -.780 .251 - Sig. (2-tailed) .333 .000 .000 .000 - 6. Geographic Distance 7.591 1.123 Pearson .082 .091 -.425 .272 .312 - Sig. (2-tailed) .000 .000 .000 .000 .000 - 7. Economic Distance 1.092 .518 Pearson -.049 -.024 -.511 .118 .613 -.245 - Sig. (2-tailed) .000 .000 .000 .000 .000 .000 - 8. Subsidiary Ownership 78.779 36.452 Pearson -.012 .022 -.103 .028 .110 -.162 .251 - Sig. (2-tailed) .017 .000 .000 .000 .000 .000 .000 -
By conducting the Kolmogorov-Smirnov test, it is found that Subsidiary Ownership is statistically significant (p < .001), and skewness and kurtosis (skewness = -1.387 and kurtosis = .228) are above the value of 1.96 times standard error. Ergo, it can be concluded that Subsidiary Ownership is not normally distributed and highly left skewed with a pointy distribution. By trying to transform the distribution of Subsidiary Ownership through a logarithm function, it is determined that even with a transformation, the data will not fit a normal distribution. This is not a big problem as long as it is noticed and the test results are interpreted with caution.
34 Subsequently, an independent t-test is conducted to check both groups of OECD members and non-members on equal variances and equal means concerning their Subsidiary Ownership. The test shows a significance value of p < .001 for both the Levene’s test for equality of variances and the t-test for equality of means, which indicates that variances and means of each group statistically differ from each other.
To test the hypotheses, a moderator analysis is necessary to investigate if the moderator variable, here the OECD Membership, influences the effect of Cultural, Administrative, Geographic and Economic Distance on Subsidiary Ownership. Before this analysis can be conducted, it is necessary to check several assumptions of the regression model: Linearity, non-collinearity, independence of residuals, homoscedasticity of residuals and normality of residuals.
In this analysis, linearity can be assumed. Non-collinearity is tested and confirmed for MNE Revenue (VIF = 3.442), MNE Employees (VIF = 3.446), OECD Membership (VIF = 3.166), Cultural Distance (VIF = 1.142), Administrative Distance (VIF = 3.287), Geographic Distance (VIF = 2.096), and Economic Distance (VIF = 2.689). When testing the independence of residuals, the scatterplot of the standardized residuals and the standardized predicted values looks rectangular. So, the independence of residuals can be assumed. Although the Durbin-Watson test suggests a positive correlation of residuals (Durbin-Watson = .329), the visual test of the scatterplot provides sufficient evidence for the independence of the residuals. The test for homoscedasticity shows a distribution that despite slight deviations, basically follows the diagonal target line which indicates heteroscedasticity.
After completing the tests of assumptions, a regression analysis can be conducted to investigate the effect of OECD Membership, Cultural Distance, Administrative Distance, Geographic Distance and Economic Distance on Subsidiary Ownership, controlling for MNE Revenue and MNE Employees. For this analysis, the standardized values of Cultural, Administrative, Geographic and Economic Distance, and OECD Membership are calculated and each standardized distance is multiplied with the standardized OECD Membership to generate the moderator interaction
35 variables CD*OECD, AD*OECD, GD*OECD and ED*OECD in order to examine the interaction effect of OECD Membership. CD*OECD refers to the interaction of Cultural Distance and OECD Membership. AD*OECD reflects the interaction of Administrative Distance and OECD Membership. The interaction of Geographic Distance and OECD Membership is represented by GD*OECD. And ED*OECD measures the interaction of Economic Distance and OECD Membership. Table 3 shows the results of the conducted regression analysis.
The first model of the hierarchical multiple regression includes the predictors MNE Revenue and MNE Employees. This model is statistically significant F (2, 39279) = 71.282; p < .001; and explains 0.4% of the variance of Subsidiary Ownership (R2 = .004).
When adding OECD Membership, Cultural Distance, Administrative Distance, Geographic Distance and Economic Distance to MNE Revenue and MNE Employees for the second model, the model is statistically significant (p < .001) and explains 8.1% of the variance of Subsidiary Ownership with R2 = .081; F (7, 39274) = 491.865. Thus, OECD Membership, Cultural Distance, Administrative Distance, Geographic Distance and Economic Distance account for a statistically significant additional amount of variance above and beyond the influence of MNE Revenue and MNE Employees (R2 Change = .077; F Change (5, 39274) = 657.715; p < .001).
In the third model, the interaction variables complement the analysis and consequently this model explains 10.1% (R2 = .101) of the variance of Subsidiary Ownership F (11, 39270) = 400.748; with a statistical significance of p < .001. Ergo, OECD Membership and the interaction variables of CD*OECD, AD*OECD, GD*OECD and ED*OECD account for a statistically significant additional amount of variance above and beyond the influence of MNE Revenue, MNE Employees, OECD Membership, Cultural, Administrative, Geographic and Economic Distance (R2 Change = .020; F Change (4, 39270) = 221.924; p < .001).
36
R R2 Change R2 Change Sig. F B SE β Sig.
Model 1 .060 .004 .004 .000 MNE Revenue .000 .000 -.104 .000 MNE Employees .000 .000 .109 .000 Model 2 .284 .081 .077 .000 MNE Revenue .000 .000 -.076 .000 MNE Employees .000 .000 .100 .000 OECD Membership -8.198 .763 -.092 .000 Cultural Distance 1.135 .222 .026 .000 Administrative Distance -.174 .044 -.035 .000 Geographic Distance -5.073 .227 -.156 .000 Economic Distance 12.813 .558 .182 .000 Model 3 .318 .101 .020 .000 MNE Revenue .000 .000 -.065 .000 MNE Employees .000 .000 .094 .000 OECD Membership -5.687 .948 -.064 .000 Cultural Distance 1.466 .228 .034 .000 Administrative Distance .405 .058 .081 .000 Geographic Distance -3.689 .244 -.114 .000 Economic Distance 17.482 .621 .249 .000 CD*OECD .185 .192 .006 .335 AD*OECD 2.722 .274 .110 .000 GD*OECD -1.931 .314 -.052 .000 ED*OECD 3.357 .252 .122 .000
Table 3 Regression analysis
MNE Revenue is statistically significant (p < .001) with a standardized coefficient of β = -.065. This shows a slightly negative effect of MNE Revenue on Subsidiary Ownership, indicating that a higher MNE Revenue results in a lower Subsidiary Ownership percentage. MNE Employees is also statistically significant (p < .001) with a slightly positive standardized coefficient of β = .094, which demonstrates that a higher amount of MNE Employees positively affects higher Subsidiary Ownership.
OECD Membership is statistically significant (p < .001) with a slightly negative standardized coefficient of β = -.064. This indicates a negative influence of OECD Membership on the level of Subsidiary Ownership.
