Master Thesis International Business and Management
Institutional distance and entry mode decision:
The moderating effect of R&D capability
Author: M.M. Kuijpers Student number: S3343227 Email: m.m.kuijpers.2@student.rug.nl
Supervisor: O. Lindahl Co-assessor: L. Ge
Faculty of Economics and Business University of Groningen
Duisenberg Building, Nettelbosje 2, 9747 AE Groningen, The Netherlands P.O. Box 800, 9700 AV Groningen, The Netherlands
http://www.rug.nl/feb
Date of submission: 20 January 2020
ABSTRACT
Research on the relationship between distance and entry mode decision is well-established. Aggregated measures are often used, but therefore the distinct impact of the most relevant Worldwide Governance Indicators remains unexplored. Therefore, building upon institutional distance, this study aims to investigate the relation between institutional distance and entry mode selection. To select the most relevant indicators within this measure, this study includes a moderating perspective of R&D capability on this direct relation. Based on extensive literature, it was expected that institutional distance with regard to political stability and rule of law are positively associated to entry mode selection, preferring joint ventures over wholly owned entry modes. By analyzing 119 cross-border deals over a time frame of five years, the results of the binary logistic regression indicate that firms have a higher likelihood of selecting joint ventures in host countries facing context with greater institutional distance with regard to rule of law. Results relating to political stability did not reveal significant results. Furthermore, no significant results were found related to the moderating effect of R&D capability. Accordingly, this study challenged some existing theories regarding institutional theory and provide promising opportunities for future research.
Keywords: institutional theory, institutional distance, worldwide governance indicators, entry
TABLE OF CONTENTS
1. Introduction ... 1
2. Literature review ... 4
2.1 Theoretical background ... 4
2.1.1 Entry mode decision ... 4
2.1.2 Institutional theory ... 5 2.1.3 Institutional distance ... 6 2.1.4 R&D capability ... 9 2.2 Hypothesis development ... 10 2.2.1 Direct effects ... 10 2.2.2 Moderating effects ... 11 2.2.3 Conceptual model ... 12 3. Methodology ... 13 3.1 Data ... 13 3.2 Sample ... 13 3.3 Variables ... 15 3.3.1 Dependent variable ... 15 3.3.2 Independent variables ... 15 3.3.3 Moderating variable ... 16 3.3.4 Control variables ... 16 3.4 Statistical analysis ... 18 4. Results ... 20 4.1 Descriptive statistics ... 20
4.2 Assumptions of binomial logistic regression ... 21
4.3 Regression results ... 24
5. Discussion ... 28
5.1 Discussion of results ... 28
5.2 Theoretical implications ... 29
5.3 Practical implications ... 30
5.4 Limitations and future research ... 30
6. Conclusion ... 32
References ... 33
1
1. INTRODUCTION
“Today’s and tomorrow’s choices are shaped by the past” – D.C. North (1990: vii)
Entry mode selection is a crucial decision-making process for multinational enterprises (MNEs) when entering a foreign market, as firms have to determine to what extent it wants to control the foreign entity (Anderson & Gatignon, 1986). Despite the specific entry mode, MNEs incur costs of doing business abroad as they face liabilities of foreignness in the host country (Eden & Miller, 2004). Therefore, it is important to understand the challenges an MNE is facing when expanding abroad. Their internationalization process is based upon the Uppsala model, adhere to incremental internationalization (Johanson & Vahlne, 1977, 2009). However, the latest trends in globalization and technology have driven the world to open markets and new economies. Therefore, it is imperative to understand these influences on market entry modes. International expansion involves overcoming distance barriers, a well-known concept in international business research (Arregle, Miller, Hitt, & Beamish, 2016; Beugelsdijk, Kostova, Kunst, Spadafora, & van Essen, 2018; Hutzschenreuter, Kleindienst, & Lange, 2014; Yildiz & Fey, 2016; Zaheer, 1995). The entry mode decision of a firm varies depending on the distance between home and host country (Johanson & Vahlne, 1977; Kogut & Singh, 1988; O’Grady & Lane, 1996). The results are however ambiguous, hence Dow & Larimo (2009) called for a clarification of distance. Following this line of reasoning, this study explicitly examines the effect of institutional distance on the entry mode decision (Eden & Miller, 2004; Van Hoorn & Maseland, 2016; Xu & Shenkar, 2002).
Institutions are frequently defined as the “rules of the game” (North, 1990: 3) that provide structures to reduce uncertainty. When firms decide to internationalize, they have to adapt to the rules and structures of the host country. Hence, MNEs are facing multiple institutional environments and one of their biggest challenges is to manage these different environments while obtaining and maintaining legitimacy (Kostova & Zaheer, 1999). The institutional differences between home and host countries are often measured through measures on regulative, cognitive and normative levels (Eden & Miller, 2004; Pogrebnyakov & Maitland, 2011; Scott, 1995; Xu & Shenkar, 2002) or by means of formal and informal institutional distance (Estrin, Baghdasaryan, & Meyer, 2009; Gaur & Lu, 2007; Peng & Pleggenkuhle-Miles, 2009).
2 firms are more likely to select a greenfield investment over an acquisition or JV is the institutional distance increases. The research of Brouthers (2002) showed that companies prefer joint ventures over wholly owned entry modes when the institutional context in host countries imposes more regulations and restrictions. Yiu & Makino (2002) tested the effect of the different levels of institutional distance on entry mode and found that forces on the regulative and cognitive level more strongly influence the entry mode decision compared to the normative level.
More recent international business studies are using aggregated measures of distance as the Euclidean approach (Berry, Guillén, & Zhou, 2010; Gaur & Lu, 2007) or the Worldwide Governance Indicators (WGI) from the World Bank (Beugelsdijk, Ambos, & Nell, 2018; Cuervo-Cazurra & Genc, 2008; Hotho & Pedersen, 2012; Hutzschenreuter et al., 2014; Pogrebnyakov & Maitland, 2011; Van Hoorn & Maseland, 2016). However, these six different dimensions might not be equally important in every study. Therefore, this research tries to contribute to the existing literature by separating the WGI dimensions and testing the individual effect of the most relevant dimensions on the entry mode decision.
Furthermore, this study aims to capture the effect of the (technological) transformation of the international business environment by incorporating the firm’s research and development (R&D) efforts. On the one hand studies found that firms investing in R&D prefer a high level of ownership when expanding abroad (Belderbos, 2003) but on the other hand the institutional environment of the host country is a key determinant for firms involved in R&D (Choi & Contractor, 2016). Hence, this research hypothesizes a moderating effect of R&D capability on the relationship between institutional distance and market entry mode as R&D investments could favor a sole proprietorship. As a result, the following research question was established:
Research question: What is the relationship between institutional distance and firms’ entry mode decisions? And how is this relationship affected by a firm’s investment in R&D capability?
3 control. In addition, this research further hypothesizes a positive moderation effect of R&D capability on the previous mentioned negative relation as investments in R&D can help firms to reduce risk.
The results of this study can be used to compare the influences of the distinct institutional distance measures. Both significant and insignificant results related to the different dimensions of the Worldwide Governance Indicators are found within this study, emphasizing the importance of institutional theory. Therefore, it is recommended to other scholars to reevaluate the concept and measures of institutional distance. Furthermore, the insignificant regression results regarding the moderating effect of R&D capability demand for clarification and evaluation of the R&D objectives of the firms within the sample.
4
2. LITERATURE REVIEW
2.1 Theoretical background2.1.1 Entry mode decision
The internationalization process of firms is characterized in the Uppsala model of Johanson & Vahlne (1977, 2009). This model indicates that firms prefer to internationalize progressively by taking little, incremental steps. Market knowledge and commitment can reduce uncertainty and facilitate the extent of foreign investment, shifting from exporting to setting up a foreign subsidiary (Johanson & Vahlne, 1977). According to Anderson & Gatignon (1986), the level of control determines the entry mode decision. Here, high-control entry modes, as a wholly owned subsidiary, can enlarge the risks and returns, while low-control entry modes, as licensing, reduces risk but comes at the expense of returns. Both studies, however, recognize the impact of the unknown environment of the host country, also known as the liabilities of foreignness (Eden & Miller, 2004).
The internationalization process of firms is split in different stages, namely location choice, entry mode, degree of ownership and level of control (Beugelsdijk, Kostova, et al., 2018). While the authors affirm the importance of every single stage to successfully expand abroad, this study will focus on the second stage, the entry mode decision. More specifically, on the choice between a joint venture (JV) or wholly owned subsidiary (WOS), as used in the studies of Brouthers & Hennart (2007) and Dikova & Brouthers (2016). JVs are characterized by shared ownership, as two or more firms share control and ownership rights. WOS, differently, are fully owned and thus represent direct presence in the foreign country (Kogut & Singh, 1988). Hereby, a clear distinction is made between sole ownership and shared ownership.
The entry mode decision is based on which mode will obtain the maximum return on investment. While a WOS is favored to obtain the highest possible return, a JV is preferred when the risk of investment is higher (Brouthers, 2002). Distance between home and host country causes uncertainty and therefore influence the entry mode decision (Eden & Miller, 2004; Kogut & Singh, 1988; O’Grady & Lane, 1996; Xu & Shenkar, 2002). Hence, firms usually select an entry mode with a lower commitment level in the form of JV in distant foreign markets compared to a WOS (Anderson & Gatignon, 1986).
5 costs. By means of high control, the company can fully benefit from the economies of scale (Brouthers, 2002). Shared ownership, in contrast, entails strategic flexibility allowing firms to exit or expand depending on the success in the foreign market (Slangen, 2013). Furthermore, a JV with a local partner can help the MNE to acquire market knowledge (Tihanyi et al., 2005). The study of Brouthers (2002) includes an institutional perspective to the entry mode decision as the institutional environment influences this choice. In some host countries, the entry mode decision is limited as governments restrict or limit foreign entry. To comply with those rules and gain legitimacy in the foreign market, the entry mode should be adapted to the institutional environment (Brouthers, 2002; Yiu & Makino, 2002). In their research, Yiu & Makino (2002) state that traditional IB studies regard that the entry mode decision between a JV or WOS is solely derived on the matter of control. The involvement of institutional theory in this decision is necessary as it shapes the host environment (Brouthers, 2002; North, 1990; Yiu & Makino, 2002).
2.1.2 Institutional theory
Institutions are defined by North (1990: 3) as “the rules of the game” that provide structure to reduce uncertainty. Within institutional theory (Scott, 1995), the institutional environment is seen as major determinant of the behavior and structure of a firm. When companies expand internationally, they have to adapt to the rules of the host country. Multinational enterprises (MNEs) face multiple institutional environments, and in order to maintain legitimacy they have to comply with the different frameworks of each host country (Kostova & Zaheer, 1999). This imposes challenges for the MNE.
Besides legitimacy challenge, MNEs also face challenges regarding the transfer of practices and orientations from the parent company to the subsidiary abroad (Kostova, 1999). This implies that organizations can successfully transfer its practices to foreign subsidiaries. Ionascu, Meyer & Erstin (2004) describe this challenge as the constrained interplay between the parent company and the foreign subsidiary due to their different national environments. This is the key challenge of an MNE as a decision has to be made between local responsiveness and global integration (Xu & Shenkar, 2002). This dual pressure is one of the research streams in institutional theory that this research is based on.
6 rules and laws that organizations have to comply with (Kostova & Zaheer, 1999). This can be related to, for example, the governmental policies of a country or specific industry standards. North (1990: 55) defines this as “enforcement mechanisms” that support or restrain certain behaviors or transfer of practices.
Second, the cognitive pillar focuses on cognitive programs and schemas that people share within a specific community or country (Scott, 1995). This pillar influences how information is selected and interpreted depending on the specific environment (Kostova, 1999). National symbols, stereotypes and shared knowledge are reflected in this pillar as these are components of the social environment. By being coherent with the installed cognitive structures, organizations can be legitimate (Kostova & Zaheer, 1999).
The third dimension, the normative pillar, “goes beyond regulatory rules and cognitive structures” (Kostova & Zaheer, 1999: 69) and reflects on the norms, beliefs and values of individuals in a specific environment. This dimension determines the expected behavior (Ionascu et al., 2004), it defines what is appropriate and what is not. To this extent, legitimacy can be derived from the correspondence of the aspired values of the organization with the wider societal values (Kostova & Zaheer, 1999).
Although the three institutional environment pillars are not by definition independent, yet the overall legitimacy could be threatened if the minimum legitimacy conditions of one of the pillars is not met (Kostova & Zaheer, 1999). Furthermore, these institutional dimensions can interfere or facilitate the transfer of practices from the parent company to the subsidiary abroad (Kostova & Roth, 2002). However, the level of legitimacy and the success of the transfer of practices depend on the institutional distance between countries.
2.1.3 Institutional distance
7 Institutional distance can be split into formal and informal institutional distance (Arslan & Dikova, 2015; Estrin et al., 2009; Gaur & Lu, 2007; Peng & Pleggenkuhle-Miles, 2009) based on the three pillars of institutional theory. The regulative dimensions of institutions can be related to formal institutional distance, whereas the cognitive and normative dimensions represent informal institutional distance. This differentiation is often used to better explain the concept of each concept (Gaur & Lu, 2007). According to Gaur & Lu (2007), the formal aspects of institutions are more transparent and explicitly retrievable compared to the informal aspects. This creates less unfamiliarity risks as firms can be well informed about the formal institutional elements in the host country. However, transfer of practices can be constrained due to local laws as adoption of organizational practices is impeded (Estrin et al., 2009). With regard to informal institutions, these aspects are enclosed in the social context of the host country (Gaur & Lu, 2007). Therefore, differences in informal institutions are often related to cultural differences (Arslan & Dikova, 2015; Peng & Pleggenkuhle-Miles, 2009). Precise local knowledge is demanded to understand the implicit institutional environment and accordingly obtaining local legitimacy could be endangered.
The research of Hotho & Pedersen (2012) is based on the institutional approaches of new institutional economics, new organizational institutionalism and comparative institutionalism. These approaches differ in explaining international firm behavior and how institutions matter in international business. According to the authors, the conceptualization of institutional distance depends on the selected institutional approach (Hotho & Pedersen, 2012). As this research studies the effect of the host countries’ institutional frameworks on the entry mode decision of the MNE, the institutional approach of this study is new institutional economics. The corresponding conceptualization of institutional distance is therefore a measure that seizes the strength and quality of the institutional framework (Hotho & Pedersen, 2012). This is important as this framework can have an impact on the uncertainty an MNE is facing when entering a foreign market.
8 of this research (Xu & Shenkar, 2002). The appropriate measure of institutional distance is therefore one that captures both of these effects on the entry mode decision of the MNE. The Worldwide Governance Indicators (WGI) from the World Bank (Kaufmann, Kraay, & Zoido Lobatón, 1999) are an often-used measure in institutional research (Beugelsdijk, Ambos, et al., 2018; Cuervo-Cazurra & Genc, 2008; Hotho & Pedersen, 2012; Hutzschenreuter et al., 2014; Pogrebnyakov & Maitland, 2011; Van Hoorn & Maseland, 2016). This project consists of the following six dimensions: political instability and absence of violence, voice and accountability, government effectiveness, regulatory quality, control of corruption and rule of law. This measure thus offers the opportunity to select the most appropriate institutional distance measures. The six different dimensions will subsequently be discussed using Kaufmann et al. (1999) and the most relevant indicators within this research are selected. Political (in)stability and absence of violence, and voice and accountability are directed towards the selection, monitoring and replacement processes of governments (Kaufmann et al., 1999). More specifically, political stability is directed towards the conception on the probability of political (in)stability and violence, for example destabilization of the government or even terrorism (Kaufmann et al., 1999). To comply to the host country’s environment and obtain legitimacy, this dimension is regarded as highly relevant within this study. Voice and accountability measures aspects of the political process, civil autonomy and political rights (Kaufmann et al., 1999). Within this dimension the citizen’s participation in these processes and the freedom of media is included. Cuervo-Cazurra & Genc (2008) illustrate that firms do not determine the investment decision upon the citizen’s power to actively participate within the political process. This research accordingly excluded this dimension.
9 research. This is because the freedom of the MNE is restricted and therefore might not be able to enter the market through different entry modes.
Control of corruption and rule of law both deal with respect regarding the rules of society (Kaufmann et al., 1999). The first dimension concerns the abuse of entrusted power for private gain. This study is directed on the legitimacy challenges that MNEs face within the host market. To be able to form a JV with a (local) partner, the firm is dependent on the legitimacy of this partner and the local context. This impact the freedom of the MNE and dimension is therefore disregarded. Rule of law captures the quality of the rules and rights of society (Kaufmann et al., 1999). Intellectual property rights are included within this dimension, and as the focus of this study is on the ability of the MNE to transfer practices, including know-how, this is an important dimension. Furthermore, this dimension assesses the enforcement of contracts and the justice of (judicial) processes. In consequence, this dimension is incorporated in this research.
2.1.4 R&D capability
The (technological) transformation of the international business world requires firm to invest resources in research and development (R&D) (Cheng & Bolon, 1993). Much research has been conducted on the effect of R&D investments on firm performance. These studies find that expenditure on R&D improves the firm’s innovation process (Hsu, Lien, & Chen, 2015) and ultimately financial returns (Hsu, Chen, Chen, & Wang, 2013). This is because R&D advances the coordination within firms and supports the global integration (Cheng & Bolon, 1993), Therefore, the competitiveness of a firm is dependent on its capability to exploit R&D activities. (Hsu et al., 2013).
MNEs can transfer their knowledge and practices throughout their network, and thus increase its future performance (Hsu et al., 2013). This results, however, also in additional costs as the operations of the MNE become more complex and creates accordingly monitoring and communication challenges (Hsu et al., 2015). Furthermore, intellectual property rights expose the firm to knowledge leakages when innovations cannot effectively be protected. This shows that R&D has both advantages and disadvantages.
10 issues of control. To protect the technology and know-how and prevent knowledge leakages, solely proprietorship might be preferred by MNEs (Belderbos, 2003). However, joint ventures could lead to creating combined, distinct R&D capabilities that can be used to redeploy resources (Belderbos, 2003). Therefore, this study aims to determine whether R&D capability serve as a moderator in the relationship between institutional distance and entry mode decision.
2.2 Hypothesis development
2.2.1 Direct effects
In terms of literature that examines the entry mode decision, much research has been done on location choice, entry mode decision and degree of ownership. Furthermore, different measures of distance, as psychic distance (Hutzschenreuter et al., 2014), cultural distance (Beugelsdijk, Kostova, et al., 2018) and institutional distance (Arregle et al., 2016) are often incorporated in international business research. This study focuses on the latter, as the institutional framework induces the legitimacy and uncertainty challenges of MNEs when entering a foreign market (Hotho & Pedersen, 2012). The quality of the institutional framework affects both the performance of the local investments and the appropriateness of the selected governance modes. By separating the dimensions of the often-used WGI, this study tries to examine the distinct impact of the most relevant dimensions, selected in the above literature.
The dimension political stability relates to the quality of the government, but also to sudden changes within the government. This could lead to developments in for example the policies regarding foreign investments. As mentioned before, host countries could form limits or restrict foreign companies to enter the market. As a consequence, the MNE will have an unfavorable position in the foreign market. Furthermore, this indicator relates to the procedures how governments are selected and replaced. The MNE has to deal with the specific rules and laws of each host country, as it cannot influence this. Hereby, the transfer of organizational practices of the MNE are directly impacted. Therefore, restrictions on the regulatory level can result in selecting a subsidiary with a lower level of risk.
11 a company is able to share ownership rights with a local partner. As a result, the following hypothesis is developed:
Hypothesis 1: The greater the institutional distance with regard to political stability, the more likely an MNE will enter by JV than by WOS.
The rule of law dimension measures the level of confidence in the rules of society. Within this dimension, the degree of security, likelihood of violence and enforcement of courts are comprised as well as (intellectual) property rights and expropriation. This indicator influences the legitimacy the MNE has to establish and maintain in the host country. Establishing legitimacy can be done by being coherent with the installed structures (Kostova & Zaheer, 1999) however norms regarding security and public management can be hard to interpret. Furthermore, the MNE is already facing the liabilities of foreignness (Eden & Miller, 2004) which imposes a larger challenge on maintaining legitimacy. Therefore, the MNE is expected to enter the market by JV as it can share ownership rights with a local partner. Furthermore, property rights secure the host country’s market, while lack of these rights can reduce foreign investments (Wu, Li, & Selover, 2012) as MNEs face a higher risk of losing their resources. Expropriation could also expose a threat to the MNE as the state could take over property for public use. This means that in host countries with a low level of rule of law, the MNE has to protect its assets by means of financial resources and managerial attention. On the contrary, if the rule of law in the host country is strong, an MNE is less likely to fear expropriation and the violation of property rights. Based on the above, the following hypothesis is formulated:
Hypothesis 2: The greater the institutional distance with regard to rule of law, the more likely an MNE will enter by JV than by WOS.
2.2.2 Moderating effects
The political stability in the host country defines the investment climate as instability brings uncertainty challenges. Investments in R&D become riskier when the host country offers an unstable, uncertain market. This results in lower returns on R&D. Therefore, firms become less motivated to start long-term R&D projects within the host country. This also effects the way firms access the foreign market as they will be more likely to select an entry mode with a lower level of control. In this way, an MNE is able to cooperate with a (local) partner and share the risk. The following hypothesis is accordingly defined:
12 The institutional distance with respect to rule of law determines the quality of regulations. This includes the protection of intellectual property rights. This is an important dimension with regard to R&D capability as firms are able to protect their innovations. Entering a foreign market with an innovation that cannot be protected involves risk for the MNE. Competitors could imitate the recent developments and hereby the firm can lose its competitive advantage. Furthermore, this dimension is directed on the possible expropriation by the government. By cooperating with a local partner, the MNE is able to share the risks that can be faced when entering a host country with poor rule of law. Therefore, this study hypothesize that firms are more likely to select an entry mode with a lower level of control as risk can be shared and the R&D capability of firms cannot be jeopardized by expropriation of the government. Based on the above, the following hypothesis is formulated:
Hypothesis 4: R&D capability positively moderates the negative relation between institutional distance with regard to rule of law and the entry mode decision
2.2.3 Conceptual model
The following conceptual model is created to illustrate the relationship between the different concepts. This conceptual model details a negative relation between the different dimensions of the WGI and the entry mode decision; firms have a higher likelihood of preferring JVs when facing a host country with a greater institutional distance. Further, this model depicts a positive moderation effect of R&D capability on the negative relation between institutional distance and entry mode decision; firms are more willing to establish a WOS instead of forming a JV.
13
3. METHODOLOGY
3.1 DataQuantitative secondary data will be used in this research on firm-, industry- and country-level. This data is accessed through the databases Zephyr, World Bank and Eurostat, accessible through the University of Groningen. Zephyr is used to create a sample of cross-border deals, see section 3.2 for more details on the sample. This dataset is subsequently combined with the institutional distance variables and data on R&D capability. With respect to the independent variable, this data is collected the Worldwide Governance Indicators of the World Bank. This measure is often used in international business and institutional research as it openly accessible (Beugelsdijk, Ambos, et al., 2018; Hotho & Pedersen, 2012; Van Hoorn & Maseland, 2016). Moreover, this dataset is highly extensive as it provides over 200 country scores. The moderating variable R&D capability is in the previous literature defined and is measured by means of the proxy R&D expenses (Hsu et al., 2013). This data is accessed through Compustat, which provides market information on more than 24,000 publicly held firms. The expenses on R&D are computed here and matched to the US acquirer. Missing values are collected though the annual reports of the specific firms.
3.2 Sample
As this study is directed on how the institutional environment of host countries affects the entry mode decision, the sample of this research consist of firms from one host country (Brouthers & Hennart, 2007). Due to the availability of data, the US is selected as home country, and more especially this study is focused on US stock listed firms as their internal data is openly accessible. Foreign direct investments made by US firms (Arregle, Hébert, & Beamish, 2006; Dow & Larimo, 2009) are used to test the effect of institutional distance on the entry mode decision between WOS or JV, and how this relation is affected by investments in R&D capability.
14 In this way, acquirers from the US, and targets from all other countries except the US are selected.
The time frame of this study is based upon the most recent data to create a unique dataset, but also because the world has become significantly interconnected. Since the sample consists of firms from the US, this economy and its financial stability is considered. Hausman & Johnston (2014) indicate that the US economy was still fragile during the first months of 2012 due to the financial crisis, and started recovering more significantly in 2013. Therefore 2013 is selected as the start of this research, as from this year on firms will be more likely to invest in foreign expansion. The final year of this research is 2018, as the annual accounts for this year are closed. Based on the above criteria, 517 deals were selected.
The sample size of this study is based on the one in ten rule (Harrell, Lee, Califf, Pryor, & Rosati, 1984; Peduzzi, Concato, Kemper, Holford, & Feinsten, 1996). This general rule of thumb states that in logistics regression one parameter needs ten observations. Harrell et al. (1984) disclosed that this decreases the risk of overfitting. Overfitting happens when too many parameters are selected, as this could lead to a similar sample and hence reduce the generalizability of the study (Peduzzi et al., 1996). This study contains two independent variables, one moderating variable, one dependent variable and four control variables, so to meet this rule a sample size of 80 observations is required.
15 TABLE 3.1
Overview of host countries within sample
3.3 Variables
3.3.1 Dependent variable
The dependent variable foreign entry mode is operationalized by means of the decision between a joint venture and wholly owned subsidiary. This variable is collected through data from Zephyr and is transformed to a dichotomous variable where the value of 0 represents entry mode in the form of a JV, and a value of 1 indicates an WOS (Arregle et al., 2006; Beugelsdijk, Kostova, et al., 2018; Gaur & Lu, 2007). Within Zephyr, the percentage of the selected entry modes is registered. To make sure that a WOS is owned by just one sole owner, only deals are included of which the owner’s equity is 95% or higher (Dow & Larimo, 2009; Gaur & Lu, 2007; Yiu & Makino, 2002). Following the previous literature, a JV is undertaken jointly by two or more companies that together own 100% of the JV. Another requisite is that the US firm owns at least 30% of the shares within the firm.
3.3.2 Independent variables
The Worldwide Governance Indicators from the World Bank is a dataset consisting of six dimensions, as explained in the previous literature. Currently, these indicators are formed on almost 40 data sources, created by more than 30 organizations worldwide (Kaufmann, Kraay, & Mastruzzi, 2007). This study focusses on the two indicators political stability and rule of law. This country-level data is matched to firm-level data based on the distance between the acquirer and target country of each deal (Hotho & Pedersen, 2012; Xu & Shenkar, 2002). The country scores of two years prior to the entry of US companies are used as the decision of entering a
Host country # Host country # Host country # Host country #
Argentina 1 Denmark 2 Japan 6 New Zealand 2
Australia 2 Spain 1 Korea 2 Oman 1
Bermuda 3 France 2 Cayman Islands 3 Saudi Arabia 1
Brazil 3 United Kingdom 17 Lao 1 Sweden 5
Bahamas 1 Indonesia 1 Luxembourg 1 Singapore 1
Canada 10 Ireland 1 Mexico 3 Thailand 1
Switzerland 5 Israel 1 Nigeria 1 Taiwan 1
China 17 India 5 Netherlands 5 Vietnam 1
16 foreign country is made before the actual entry. Time-lag variables are generated to set the specific year as the time variable.
First, the home and host country scores of each dimension is derived. The estimate scores of each country is used. Composite, continuous variables are subsequently formed by calculating the distance for each of the selected WGI dimensions. This distance is determined by subtracting the home country (US) score from the host country score (Verbeke, Puck, & Van Tulder, 2017). Hence, a positive score on one of the WGI dimensions indicates that the respective dimension in the host country scores higher compared to the home country. A negative score on one of the WGI dimensions contrary represents a lower host country score compared to the home country. The WGI indicators are separately included in this study to focus on the most relevant indicators. Hereby, this study aims to contribute to the existing international business literature. The Euclidean distance measure is explicitly not used as this index aggregates all different dimensions into one measure (Berry et al., 2010; Beugelsdijk, Ambos, et al., 2018; Hotho & Pedersen, 2012).
3.3.3 Moderating variable
The moderating variable R&D capability is defined in the previous literature and is measured through R&D expenses. This study focuses on stock listed firms hence most data was openly accessible as companies are required to disclose financial statements. Some annual reports did not explicitly express values on R&D expenditure, in these cases the expenses on patents are used as patents can be seen as the products of R&D investments (Hsu et al., 2013). When firms did not report any value at all on these items, they were removed from the dataset. Providing that the aim was to collect information on 150 cross-border deals, while the sample size is based on the one in ten rule (Harrell et al., 1984; Peduzzi et al., 1996), and thus requires 100 variables, this is not causing any problems. Since this study is hypothesizing a moderating effect on the relationship between institutional distance and entry mode decision, the R&D expenses of the year prior to the cross-border entry are used. Different subgroups for these time-lags are created to lag the preceding year of each cross-country deal. The distribution of the continuous variables is first checked, as larger firms have more resources, and is later logarithmically transformed to make the data more interpretable.
3.3.4 Control variables
17 variables are incorporated in the database depending on the year of each deal. An overview of all variables can be found in table 3.2.
Firm size
The first control variable on firm-level is firm size. According to Kogut & Singh (1988), firm size is positively correlated with the likelihood of acquisitions instead of joint ventures as larger companies have more resources. Firm size can be measured by the number of employees (Brouthers, 2002; Makino, Lau, & Yeh, 2002) or by total assets (Barkema & Vermeulen, 1998; Dikova & Brouthers, 2016; Kogut & Singh, 1988; Yiu & Makino, 2002). This study incorporates the latter as resources are both financial and managerial (Kogut & Singh, 1988). The logarithm transformation of this continuous variable is used as the data is widespread and therefore harder to interpret.
Firm experience
The other often used control variable in entry mode literature is firm experience (Barkema & Vermeulen, 1998; Brouthers, 2002; Brouthers, Brouthers, & Werner, 2003; Kogut & Singh, 1988). The control variable firm experience is operationalized by the number of countries the company operates in at the year of the deal. Anderson (1993) argues that firms with more internationalization experience have more market knowledge which initiates a higher level of commitment. This is supported by Anderson & Gatignon (1986), Johanson & Vahlne (1977) and Kogut & Singh (1988). The annual reports of the selected US firms are used to collect both firm-level variables.
High-tech industry
On industry-level, this study controls for industry effects by including a control variable for type of industry (Brouthers, 2002; Gaur & Lu, 2007; Kogut & Singh, 1988), as firms in high-tech industries are more likely to acquire new firms (Dikova & Brouthers, 2016; Dikova & Van Witteloostuijn, 2007). This variable is operationalized as a dichotomous variable, coded 1 when the firm was operating in high-tech industries and coded 0 otherwise, based on the SIC code of the company. This dummy is also enclosed as robustness check of the moderating effect of R&D capability (Dikova & Van Witteloostuijn, 2007).
Host country development
18 Vermeulen, 1998) as they influence the market entry decision. Therefore, this variable is measured by the gross domestic product (GDP) per capita in the year of entry (Barkema & Vermeulen, 1998; Slangen, 2013) and is collected from the World Bank.
TABLE 3.2 Overview of variables
3.4 Statistical analysis
The method employed in this thesis is a binomial logistic regression since the dependent variable entry mode decision is a dichotomous variable, 0 is a JV and 1 is a WOS (Arregle et al., 2006; Arslan & Dikova, 2015; Arslan & Larimo, 2011; Barkema & Vermeulen, 1998; Dikova & Van Witteloostuijn, 2007). This technique is needed to understand the relationship between the independent and dependent variables, more specifically it examines the probability of selecting a WOS over a JV given the independent (different measures of institutional distance) and moderating (R&D capability) variables. The data accessed from the different databases is first collected into a spreadsheet and subsequently matched using Excel.
To study the above-mentioned relationship, this data is analysed using Stata. The relationship between the IVs and DV is tested, where an odds ratio of the institutional distance measures above 1 indicates an increase in the probability of an WOS over a JV. The interaction effect comprising R&D capability is subsequently tested. In total, five models are included in the regression analysis. The first model analyses the dependent variable entry mode decision and the control variables. Model two and three incorporate each one of the independent variables
Variables Type Data type Measurement
Entry mode decision Dependent Dichotomous JV (0) or WOS (1)
Institutional distance, with regard to political stability
Independent Ratio Worldwide Governance Indicators Institutional distance, with regard
to rule of law R&D capability Firm size Independent Moderating Control Ratio Ratio Ratio
Worldwide Governance Indicators R&D expenses, in USD, in
thousands
Total assets, in USD, in thousands Firm experience High-tech industry Control Control Ratio Dichotomous
Number of foreign countries operating in
All industries (0) but high-tech (1)
20
4. RESULTS
4.1 Descriptive statisticsTable 4.1 shows an overview of the descriptive statistics of the variables included in this binomial logistic analysis. The sample consists of 118 entry mode decisions (combination of WOS and JVs) of US firms for 35 different host countries that have been analyzed during the time frame 2013-2018. The table shows variation between the 35 host countries regarding institutional distance. Two variables, R&D capability and firm size, were logarithmically transformed in order to enhance the interpretation, as can be clearly seen in table 4.1. This redistribution ensured that the maximum was reduced, which aided in the regression analysis in terms of outliers. The variable entry mode decision and high-tech industry are both dichotomous variables, therefore there minimum and maximum value are 0 and 1 respectively.
Table 1 in the appendix shows the descriptive statistics of the variables sorted per host country. This provides some information on what country scored the largest scores on the different WGI dimensions. The largest distance between the home and host country with respect to political stability is -2.6. As indicated before, a negative represents a lower host country score compared to the home country. The total number of cross-border deals in this sample is 118, of which 74 are WOS and 44 are JVs. As a logistics regression is performed in this study, the assumptions regarding distribution of data are not required making the descriptive statistics less relevant.
TABLE 4.1 Descriptive statistics
Variables N Mean SD Min Max
Entry mode decision 118 0.6271 0.4856 0 1
Institutional distance PS 118 -0.1542 0.7779 -2.6 0.9
Institutional distance RQ 118 -0.3636 0.9163 -2.2 0.9
R&D capability 118 1367788 2596602 1820 1.31e+07
R&D capability (log) 118 12.3883 2.1296 7.5066 16.3910
Firm size
Firm size (log)
118 118 9.73e+07 16.3071 3.47e+08 2.1185 29546 10.2937 2.57e+09 21.6681 Firm experience 118 30.9322 29.6834 1 145 High-tech industry 118 0.1186 .3248 0 1
21
4.2 Assumptions of binomial logistic regression
The first assumption of binomial logistic regression is that the dependent variable should consists of only two independent groups, this to prevent a dummy variable trap (DeMaris, 1995). The entry mode decision in this research is therefore operationalized as a dichotomous variables, where the value 0 indicates a JV and the value 1 represents a WOS. Besides, the research has to include two or more independent variables at a continuous level. This assumption is fulfilled as two WGI measures with ratio values are included in this study. The next assumption is that the observations in the sample have to be independent (Peng et al., 2002). The deals included in the sample are all independent, so this assumption is met.
Another assumption is that a linear relation is needed between the independent variables and the logit transformation of the dependent variable. This can be checked using the Box-Tidwell test (Box & Tidwell, 1962). This test first computes the natural logarithm of each continuous independent variable and afterwards includes the interaction terms of each independent variable and its natural log. The results show insignificance (p > 0.05) and thus represent a linear relation between the independent and dependent variables.
Furthermore, within a logistic regression there should be no significant outliers or high leverage points present. The descriptive statistics, see table 4.1, show that there are no outliers identified after the logit transformation of R&D capability and firm size. No high leverage points, single data point with an over-proportional effect on the regression, are found using the leverage-versus-squared-residual plot.
The assumption for homoskedasticity in a logistic regression is different from linear regression, since normal distribution cannot hold for a binary outcome variable. The error variances vary for each value therefore heteroscedasticity is expected in logistic regressions. To test the contribution of each additional (independent and moderating) variable, the different variables are added one-by-one and the model is run multiple times. This increases the validity and robustness of the results as the effect of each distinct variable is examined.
22 absolute threshold for correlations is 0.7 to 0.8 (Dormann et al., 2013). Another way of investigating multicollinearity is through the Variance Inflation Factor (VIF). Here, the threshold value is >10 (Dormann et al., 2013). Likewise, tolerance (defined as 1/VIF) is another way of investigating multicollinearity. Here, the lower threshold of 0.1 should not be exceeded (Dormann et al., 2013). VIF and tolerance values can be found in table 4.2.
TABLE 4.2
VIF results of the independent variables
The results of table 4.2 and 4.3 show that collinearity can be problematic in this study. With respect to the correlations, using the threshold of 0.8 means that the two WGI dimensions break the limit, just like the correlation between the institutional distance regarding political stability and host country development. With regard to the VIF and tolerance values, the different variables to not exceed the boundaries. The high correlations between the two institutional distance variables might be expected as they measure the same underlying construct (Beugelsdijk, Ambos, et al., 2018). The variables share a considerable amount of information, this could lead to unstable coefficient estimates (Dormann et al., 2013).
This study aims to examine the distinct impact of the most relevant WGI indicators therefore the variables are not excluded during the logistic regression but included separately. However, this causes to interpret the results with caution. The same applies to the high correlation between political stability and host country development and the high correlation between R&D capability and firm size. These variables are maintained in this study as firm size and host country development are control variables in this study. Furthermore, economic development is solely measured through GDP per capita, while institutional distance with regard to political stability is measured through various variables from different sources.
Variable Variance Inflation Factor Tolerance
Institutional distance PS 4.23 0.236190 Institutional distance RL 3.05 0.328404 R&D capability 2.45 0.407742 Firm size 2.41 0.415257 Firm experience 1.43 0.700433 High-tech industry 1.15 0.871094
Host country development 3.16 0.316597
23 TABLE 4.3
Pearson correlation matrix
N=118, * p<0.05, ** p<0.01, *** p<0.001
Variables (1) (2) (3) (4) (5) (6) (7) (8)
(1) Entry mode decision 1
(2) Institutional distance PS 0.3578 1 (3) Institutional distance RL 0.4847 0.8052 1 (4) R&D capability 0.1509 0.1463 0.1118 1 (5) Firm size 0.2542 0.1018 0.0992 0.7462 1 (6) Firm experience 0.1660 -0.0017 0.0520 0.4479 0.4285 1 (7) High-tech industry 0.0119 0.0731 0.0442 0.1083 -0.0144 0.2624 1
24
4.3 Regression results
The results of the binary logistic regression are discussed based on the different models. As explained before, the dependent variable entry mode decision is a dichotomous variable with 1 representing a WOS and 0 representing a JV. The results of the regression are shown in odds ratios, derived from the odds (ratio of probabilities) for WOS and JV (Peng et al., 2002). The coefficients of logistic regressions are less interpretable as they are logarithms of the odds ratios. The dichotomous dependent variable coded 1 are represented by the odds ratios. Positive odds ratios below 1 denote negative relations, odds ratios of 1 represent no relation and odds ratios above 1 indicate positive relations. Hence, in this study the odds ratios represent the likelihood of US firms selecting a WOS. This indicates that the regression results below 1 with respect to the direct effects indicate a negative relation, preferring WOS over JVs. The results regarding the interaction effect, with a value above 1, represent a positive effect.
The categorical control variable high-tech industry is included as a dummy variable in this regression. Table 4.5 shows the results of the logistic regression in odds ratios separated into five models. The first model only includes the control variables and in model 2 and 3 the independent variables are separately included. the fourth model includes the two different institutional distance measures and here R&D capability is also included as an independent variable. Model 6 includes the interaction effect of the two independent variables and the moderator.
Model 1
25 TABLE 4.4
Binary logistic regression results for entry mode decision
Standard errors in parentheses
N=118, * p < 0.05, ** p < 0.01, ***p < 0.001
Entry mode decision Model 1 Model 2 Model 3 Model 4 Model 5 Control variables: Firm size 1.265 (0.162) 1.266 (0.163) 1.268 (0.175) 1.480* (0.275) 1.451* (0.271) Firm experience 1.006 (0.009) 1.006 (0.009) 1.006 (0.009) 1.007 (0.011) 1.009 (0.011) High-tech industry 0.955 (0.644) 0.962 (0.657) 0.870 (0.626) 1.164 (0.907) 1.113 (0.872) Host country development 1.043***
26
Model 2
Hypothesis 1 implied that institutional distance with regard to political stability negatively affects entry mode decision, so firms have a higher likelihood to choose JVs in host countries when encountering circumstances with greater institutional distance with regard to political stability. Therefore, the independent variable institutional distance regarding political stability is included in this model next to the control variables. This consistently presents an odds ratio below 1, denoting a negative relationship, however this effect is not significant in model 2 (OR = 0.966, p = 0.947). The first hypothesis can therefore not be confirmed. The different control values do not show any substantial changes, besides that the control variable host country development becomes less significant. The deviance remains the same compared to model 1.
Model 3
Hypothesis 2 suggested that institutional distance regarding government effectiveness negatively influence entry mode decision, such that firm have a higher likelihood of opting for a JV in host countries. Model 3 demonstrates an odds ratio of 0.618, implying a positive relation and more importantly, this ratio is also significant (p = 0.005). Therefore, this model shows enough support to confirm the second hypothesis at a confidence level of 95%. It is noteworthy to mention that the control variable host country development no longer shows significance (OR 1.009, p = 0.483), implying that this no longer has an effect on the entry mode decision. The deviance has a higher score compared to model 1 and 2 (chi² = 36.21, p = 0.000), implying that this model is not better fitting the data.
Model 4
27
Model 5
This model includes the interaction effects between the two different institutional distance measures and R&D capability. The odds ratio of the independent variable with regard to political stability is showing a more negative relation compared to models 2 and 4, however this odds ratio is insignificant (OR = 0.256, p = 0.668). Institutional distance measured in terms of rule of law shows an insignificant, negative relation (OR = 0.771, p = 0.835). Both the interaction terms of political stability and rule of law with R&D capability show a positive relation as the odds ratio is in both cases above 1. However, these results are not significant (OR = 1.060, p = 0.824 and OR = 1.055, p = 0.136 respectively) therefore not providing enough support for the third and fourth hypothesis. The control variables remain comparable to the previous model, keeping firm size the only significant variable (OR = 1.451, p = 0.046). The deviance of this model has the highest score compared to all the previous models (chi² = 40.10,
p = 0.000). The smaller the deviance is, the better the model fits the data. Therefore, this model
is not completely exhibiting the data. Furthermore, the significance level implies that a significant of the variance is unexplained.
4.4 Overview of results
The following table, table 4.6, provides an overview of the results. TABLE 4.6
Overview of hypotheses
Hypothesis Results
Hypothesis 1 Not confirmed
Hypothesis 2 Confirmed
28
5. DISCUSSION
5.1 Discussion of resultsThis study is designed to examine the relation between institutional distance and firm’s entry mode, and how this relationship is affected by a firm’s investments in R&D capability. Based on the literature review it was hypothesized that firms are more likely to select a JV over a WOS when the institutional distance with regard to political stability and rule of law increases. Furthermore, R&D capability was expected to have a positive moderating effect on the direct relation between institutional distance and entry mode decision.
The regression results show that hypothesis 1 is not confirmed within this study. Institutional distance with regard to political stability depicts a negative relation, implying the preference of a JV over a WOS, however this result is not found to be significant. The quality of governance in the host country can be jeopardized by abrupt changes (Cuervo-Cazurra & Genc, 2008) and these changes cause uncertainty (Hutzschenreuter et al., 2014). This study expected that this would result in higher probability of opting for a JV. However, firms can accordingly decide to first reduce this uncertainty before expanding abroad. More developed institutions hold smaller risks and induce firms to enter the host country through a WOS (Hotho & Pedersen, 2012). Furthermore, the literature contemplated that MNEs entering the foreign market have to cover the additional costs of low-quality public goods. Substandard bureaucracies make cross-border expansions less efficient and more costly (Wu, 2013). Hence, it was expected that firms would opt for a JV to share these costs. However, Eden & Miller (2004) interpret these costs as one-time costs, and once these are incurred the firm has learned how to deal with these circumstances. Therefore, in the long term, acquisition is preferred over sharing equity (Eden & Miller, 2004). Furthermore, this study included a time-lag of two years as it was expected that the entry decision is made before the actual cross-border deal. This could also have had an impact on the results of this study.
29 expropriation of the government as it is cooperating with a local partner. Therefore, the firm is more likely to trust the judicial system (Cuervo-Cazurra & Genc, 2008) and is willing to enter the market through a joint venture.
Hypothesis 3 expected to find a positive moderating effect of R&D capability on the relation between institutional distance with regard to political stability and entry mode decision. Political stability and effective governance regulations facilitate the transfer of practices and therefore the MNE is able to reduce production costs and achieve economies of scale (Hsu et al., 2015). This could encourage foreign investors to engage in more R&D investments. Firms investing in R&D capability might become more likely to invest in an entry mode with higher level of risks as the institutional environment within the host country is stable. Furthermore, Belderbos (2003) found that acquiring a local firm complements the R&D capability. Therefore, it is possible that this relation is not found to be significant as this study is not researching the specific distribution of R&D investments.
The results of the binary logistic regression show that no significant relation is found for the fourth hypothesis. The moderating effect of R&D capability on the relationship between institutional distance concerning rule of law and entry mode decision is therefore not confirmed. The results show a (small) positive moderating relation, implying that firms might be more likely to select a JV over a WOS as local partnerships help to avoid the fear of expropriation. R&D capability is however also subject to the protection of intellectual property rights. As mentioned before, this study is not researching the distribution of R&D investments. The objective of the acquirer firm might be to complement its resources and capabilities (Belderbos, 2003) and therefore a JV is the preferred entry mode. It is therefore recommended in future research to gain insight in the R&D investment objectives of the acquirer firm before measuring the moderating effect between institutional distance and entry mode decision.
5.2 Theoretical implications
30 importance of institutional theory. Therefore, this study can be seen as an extension to the existing literature where institutional distance is often measured through one single measure. With reference to the results of how this relation is affected by a firm’s investment in R&D capability, there are no significant results found. This does however not imply that this study does not yield theoretical contributions. The integration of this moderating effect shows that it is important to first evaluate the R&D objectives of firms within the sample. Furthermore, this study focused on the entry modes of existing firms, while most prior literature on this topic are concentrated on the creation of new firms with new R&D capabilities. This implies that for example the dimension rule of law within institutional distance might be less relevant, as the protection of intellectual property is no longer an issue. Finally, the results of this study with regard to the interaction effect did find the expected sign but did not find significant results. Therefore, it is recommended to test the same effect on the different dimensions of the WGI.
5.3 Practical implications
The results of this study are also relevant from a practical perspective and lead to several implications. Firms entering a foreign market can face complications as the institutional distance can be incompletely evaluated. This can result in selecting an inadequately entry mode or, even worse, in failing to enter the foreign market successfully. This research strengthens the importance of selecting the most relevant dimensions of institutional theory and, still, this selection procedure might be affected by the complexity of the institutional distance concept. This is because institutional distance is, according to the literature, approached in different ways. It is therefore recommended that managers first decide on the most relevant conceptualization of institutional distance. Thereafter, it is also of critical importance that managers select the most appropriate distance measure. In addition, with regard to the moderating effect, it is important for managers to assess the influence of firm size as this control variable shows significance when R&D capability is included in this research.
5.4 Limitations and future research
31 the time frame have an impact on the reliability of this study. Estrin et al. (2009) argue that a wider variety of entry mode decision, including greenfield investments or even exporting, can provide additional insight. This is because all entry modes have different advantages and disadvantages. In addition, the time lags included in this research have an effect on the results of this study. Hsu et al. (2013) found that expenditure on R&D have a deferred return of over five years.
This study used the most relevant WGI of Kaufmann et al. (1999), however the effect of the included dimension could be higher or lower compared to the other dimensions. It would therefore be interesting to compare the results of this study to the results of the six different dimensions. While including the six dimensions independently in this research would probably results in high correlation, or even multicollinearity, an aggregated measure as the Euclidean approach or Mahalanobis index could correct for this limitation (Berry et al., 2010; Beugelsdijk, Ambos, et al., 2018).
32
6. CONCLUSION
The aim of this research was to answer the following research questions: What is the
relationship between institutional distance and firms’ entry mode decisions? And how is this relationship affected by a firm’s investment in R&D capability?
33
REFERENCES
Andersen, O. 1993. On the internationalization process of firms: A critical analysis. Journal of International Business Studies, 24(2): 209–231.
Anderson, E., & Gatignon, H. 1986. Modes of foreign entry: A transaction cost analysis and propositions. Journal of International Business Studies, 17(3): 1–26.
Arregle, J. L., Hébert, L., & Beamish, P. W. 2006. Mode of international entry: The advantages of multilevel methods, 46(5): 597–618.
Arregle, J. L., Miller, T. L., Hitt, M. A., & Beamish, P. W. 2016. How does regional institutional complexity affect MNE internationalization? Journal of International Business Studies, 47(6): 697–722.
Arslan, A., & Dikova, D. 2015. Influences of institutional distance and MNEs’ host country experience on the ownership strategy in cross-border M&As in emerging economies. Journal of Transnational Management, 20(4): 231–256.
Arslan, A., & Larimo, J. 2011. Greenfield investments or acquisitions: Impacts of institutional distance on establishment mode choice of multinational enterprises in emerging
economies. Journal of Global Marketing, 24(4): 345–356.
Barkema, H. G., & Vermeulen, F. 1998. International expansion through start-up or acquisition: A learning perspective. Academy of Management Journal, 41(1): 7–26. Belderbos, R. 2003. Entry mode, organizational learning, and R&D in foreign affiliates:
Evidence from Japanese firms. Strategic Management Journal, 24(3): 235–259. Berry, H., Guillén, M. F., & Zhou, N. 2010. An institutional approach to cross-national
distance. Journal of International Business Studies, 41(9): 1460–1480.
Beugelsdijk, S., Ambos, B., & Nell, P. C. 2018. Conceptualizing and measuring distance in international business research: Recurring questions and best practice guidelines. Journal of International Business Studies, 49(9): 1113–1137.
34 Box, G. E. P., & Tidwell, P. W. 1962. Transformation of the independent variables.
Technometrics, 4(4): 531–550.
Brouthers, K. D. 2002. Institutional, cultural and transaction cost influeces on entry mode choice and performance. Journal of International Business Studies, 33(2): 203–221. Brouthers, K. D., Brouthers, L. E., & Werner, S. 2003. Transaction cost-enhanced entry mode
choices and firm performance. Strategic Management Journal, 24(12): 1239–1248. Brouthers, K. D., & Hennart, J. F. 2007. Boundaries of the firm: Insights from international
entry mode research. Journal of Management, 33(3): 395–425.
Cheng, J. L. C., & Bolon, D. S. 1993. The management of multinational R&D: A neglected topic in international business. Journal of International Business Studies, 24(1): 1–18. Choi, J., & Contractor, F. J. 2016. Choosing an appropriate alliance governance mode: The
role of institutional, cultural and geographical distance in international research & development (R&D) collaborations. Journal of International Business Studies, 47(2): 210–232.
Criscuolo, P. 2009. Inter-firm reverse technology transfer: The home country effect of R&D internationalization. Industrial and Corporate Change, 18(5): 869–899.
Cuervo-Cazurra, A., & Genc, M. 2008. Transforming disadvantages into advantages: developing-country MNEs in the least developed countries. Journal of International Business Studies, 39(6): 957–979.
DeMaris, A. 1995. A tutorial in logistic regression. Journal of Marriage and Family, 57(4): 956–968.
Dikova, D., & Brouthers, K. 2016. International Establishment Mode Choice: Past, Present and Future. Management International Review, 56(4): 489–530.
Dikova, D., & Van Witteloostuijn, A. 2007. Foreign direct investment mode choice: Entry and establishment modes in transition economies. Journal of International Business Studies, 38(6): 1013–1033.
35 Dow, D., & Larimo, J. 2009. Challenging the conceptualization and measurement of distance
and international experience in entry mode choice research. Journal of International Marketing, 17(2): 74–98.
Eden, L., & Miller, S. R. 2004. Distance matters: Liability of foreignness, institutional distance and ownership strategy. no. 404, Texas.
Estrin, S., Baghdasaryan, D., & Meyer, K. E. 2009. The impact of institutional and human resource distance on international entry strategies. Journal of Management Studies, 46(7): 1171–1196.
Gaur, A. S., & Lu, J. W. 2007. Ownership strategies and survival of foreign subsidiaries: Impacts of institutional distance and experience. Journal of Management, 33(1): 84– 110.
Harrell, F. E., Lee, K. L., Califf, R. M., Pryor, D. B., & Rosati, R. A. 1984. Regression modelling strategies for improved prognostic prediction. Statistics in Medicine, 3(2): 143–152.
Hausman, A., & Johnston, W. J. 2014. Timeline of a financial crisis: Introduction to the special issue. Journal of Business Research, 67(1): 2667–2670.
Hotho, J. J., & Pedersen, T. 2012. Beyond the ‘rules of the game’: Three institutional approaches and how they matter for international business. Handbook of Institutional Approaches to International Business: 236–273. Cheltenham, U.K.: Edward Elgar. Hsu, C. W., Lien, Y. C., & Chen, H. 2015. R&D internationalization and innovation
performance. International Business Review, 24(2): 187–195.
Hsu, F. J., Chen, M. Y., Chen, Y. C., & Wang, W. C. 2013. An empirical study on the relationship between R&D and financial performance. Journal of Applied Finance & Banking, 3(5): 107–119.
Hutzschenreuter, T., Kleindienst, I., & Lange, S. 2014. Added psychic distance stimuli and MNE performance: Performance effects of added cultural, governance, geographic, and economic distance in MNEs’ international expansion. Journal of International
Management, 20(1): 38–54.
