Cultural distance and the ownership mode choice : the moderating effect of International Corporate Strategy

(1)

Cultural Distance and the Ownership Mode

Choice:

The moderating effect of

International Corporate Strategy

University of Amsterdam

Faculty of Economics and Business

Msc in Business Administration

Master thesis in the field of International Strategic Management

Student Name: Azadeh Baghery Ziabari

Student Number: 10128387

Date of submission: Dec 19, 2014/ Final version

First supervisor: Lori DiVito

Second supervisor:

(2)

Abstract:

Previous studies on the relationship between cultural distance and the ownership mode choice have produced contradictory results. This has also been referred to in the literature as the national cultural distance paradox. The aim of the present study is to provide a possible explanation for this paradox, by examining the potential moderating effect of MNE’s international corporate strategy on the relationship between cultural distance and the ownership mode choice. A sample of 1044 foreign subsidiaries of 10 Dutch manufacturing

companies was used to test the two hypotheses developed in this study. Results show that the international corporate strategy of the MNE moderates the relationship between cultural distance and ownership mode choice in such way that when entering a culturally distant country for the first time, global MNEs are more likely to enter through WOSs while multi-domestic MNEs tend to establish JVs.

(3)

1

1 Introduction

This chapter provides an introduction on the topic that will be researched in this master thesis. Section 1.1 presents a general overview of the entry mode literature and its two sub-fields. Also the gap in the literature is identified in this section. Next, several shortcomings in this area of research are discussed in section 1.2. The main contributions of the present study are presented in section 1.3. Finally, section 1.4 provides an overview of the structure of the thesis.

1.1 Problem indication

For more than two decades, Foreign Direct Investment (FDI) and in particular entry mode choice has been one of the most researched topics in International Business (Harzing, 2004) and has received substantial interest from scholars in the strategy field (Mani, Antia & Rindfleisch, 2007). This is understandable, since foreign entry mode choice has been considered as one of the most crucial decisions in the

internationalization process of firms (Agarwal & Ramaswami, 1992; Harzing, 2004; Root, 1994, Morschett, Schramm-Klein & Swoboda, 2010), as it is a determinant of the likelihood that the foreign operation will be successful (Anderson & Gatignon, 1986; Erramilli & Rao, 1993; Hill, Hwang & Kim, 1990; Malhotra, Sivakumar & Zhu, 2011).

The literature suggests that there are two distinct and sequential decisions that an MNE is faced with when entering foreign markets through FDI (Brouthers & Hennart, 2007; Chang & Rosenzweig, 2001; Padmanabhan & Cho, 1996). These two decisions represent two sub-fields of the entry mode literature. The first decision concerns the choice between investing in new facilities (greenfield) or acquiring existing ones (acquisition) (Kogut & Singh, 1988), which is referred to in the literature as the establishment mode choice (Brouthers & Hennart, 2007; Padmanabhan & Cho, 1996; Slangen & van Tulder, 2009). The second decision concerns the level of ownership for the foreign subsidiary (Brouthers & Hennart, 2007). Different terms are used in the literature to refer to this decision, such as the ownership mode choice (Brouthers & Hennart, 2007), the equity structure choice (Pak & Park, 2004) and the ownership structure choice (Padmanabhan & Cho, 1996). Following Brouthers & Hennart (2007), the present study will refer to this decision as the ownership mode choice.

(7)

2

Wholly Owned Subsidiary (WOS) (Brouthers & Hennart, 2007; Chang & Rosenzweig, 2001; Slangen & van Tulder, 2009). An equity Joint Venture is a shared ownership mode (Brouthers & Hennart, 2007) and

includes partial acquisitions as well as greenfield JVs (Hennart & Larimo, 1998; Slangen & van Tulder, 2009). A WOS which is a full ownership mode (Brouthers & Hennart, 2007), includes both full acquisitions and wholly owned greenfield operations (Hennart & Larimo, 1998; Slangen & van Tulder, 2009). The focus of the present study is the latter decision, thus the choice between a JV and a WOS.

In the past decades, scholars have tried to show the impact of a range of different variables on the ownership mode choice (e.g. Agarwal & Ramaswami, 1992; Agarwal, 1994; Anderson & Gatignon, 1988; Barkema, Bell & Pennings, 1996; Brouthers & Brouthers, 2003; Erramilli, 1991; Quer, Claver & Rienda, 2007). One of the variables that has been shown to impact this choice and has received much attention is Cultural Distance (CD) (Chang, Kao, Kuo & Chiu, 2011).

Studies in this area however, have presented contradictory (Chang et al., 2011; López-Duarte & Vidal-Suarez, 2010; Wilkinson, Peng, Brouthers & Beamish, 2008) and inconclusive results (Brouthers & Brouthers, 2001; Harzing, 2004; Wang & Schaan, 2008), which is also referred to as the national cultural distance paradox (Brouthers & Brouthers, 2001; López-Duarte &Vidal-Suarez, 2010; Cho & Padmanabhan, 2005; Wang & Schaan, 2008).

Of the studies that have examined this relationship, some have found empirical evidence for a positive link between CD and shared control (e.g. Agarwal, 1994; Barkema & Vermeulen, 1997), others have found support for a negative relationship (e.g. Anand & Delios, 1997; Padmanabhan & Cho, 1996), while still others found insignificant results (e.g. Luo, 2001; Chen & Hu, 2002). Finally, there is also a recent study conducted by Wang & Schaan (2008) that suggests a curvilinear relationship. The present study attempts to resolve these inconclusive findings.

1.2 Problem statement

Scholars suggest that the mixed and inconsistent results in this area of research might be due to some common errors and empirical shortcomings in past studies (Shenkar, 2001; Harzing, 2004).

Harzing (2004) argues that one of the shortcomings in past studies is a lack of the examination of moderating effects in the relationship between CD and the ownership mode choice. It is also suggested by several other authors that the relationship between CD and the ownership mode choice should be researched

(8)

3

with the consideration of moderating variables (Brouthers & Brouthers, 2001; Cho & Padmanabhan, 2005; Malhotra, 2012). Table 1.1 suggests that in recent years some scholars have responded to this call and have incorporated moderators in their studies. However, more research still needs to be conducted in order to fully understand the role of interaction variables in the relationship between CD and the ownership mode choice. One variable that is believed to have an important moderating effect (Harzing, 2004; Brouthers & Hennart, 2007), but has not been empirically tested in previous studies is the international corporate strategy of the MNE.

Another (empirical) inadequacy of previous studies is the bias with regard to the home- and host country included in the sample (Harzing, 2004). Several scholars indicate that past research shows a US-bias (Brouthers & Brouthers, 2001; López-Duarte & Vidal-Suarez, 2010; Harzing, 2004). Indeed, of the 22 studies that have been reviewed (Table 1.1), only 7 of these studies include a country other than the U.S. or Japan as the home- or host country. There is much less known about the ownership mode choice in European MNEs however, and results of American or Japanese MNEs should not be generalized so easily to a larger number of countries (Harzing, 2004). Therefore, it is important to include home- and host countries that are underrepresented in past studies.

Finally, Slangen & van Tulder (2009) have argued more recently that when analyzing the relationship between CD and the ownership mode choice, previous studies have only concentrated on external uncertainty in their discussion of CD, while CD also reflects internal uncertainty. This might partly explain the contradicting results in this area of research.

1.3 Contributions

By addressing shortcomings in previous studies, this thesis will make several contributions to the existing entry mode research. First, it will attempt to solve the national cultural distance paradox by examining the moderating effect of a strategic variable on the relationship between CD and the ownership mode choice. Thereby, it will be the first study to examine this interaction effect empirically and answer the following research question:

‘Does the International Corporate Strategy of the MNE affect the ownership mode choice for the foreign subsidiary in a culturally distant country?’

(9)

4

knowledge on the ownership mode choice in European MNEs, which are underrepresented in previous studies.

Finally, this thesis will add to the existing literature by examining the moderating effect of MNEs’ international corporate strategy with the consideration that CD not only reflects external uncertainty, as is suggested in previous studies, but also internal uncertainty.

1.4 Thesis structure

The remainder of this thesis will be divided into four chapters. In the second chapter, a literature review is conducted and hypotheses are derived. The third chapter will discuss the sample and data collection, variable measurement and the statistical method. The results of the data analysis will be presented in chapter four. Finally, a discussion of the findings as well as the limitations of the study and recommendations for future research will be presented in chapter five.

(10)

5

Table 1.1: Empirical studies on the relationship between CD and the ownership mode choice

Study Home country Host country Empirical findings1

Gatignon & Anderson (1988) U.S. Not specified Positive

Kogut & Singh (1988) More than 11 countries including U.K., Japan, Canada & the Netherlands

U.S. Positive

Erramilli & Rao (1993) U.S. Various; including Australia, New Zealand, Japan, Canada & the market economies of Europe.

Positive

Agarwal (1994) U.S. 20 countries including including

Australia, Canada, Japan and several European countries

Positive2

Padmanabhan & Cho (1996) Japan 36 countries, not specified Negative

Anand & Delios (1997) Japan Various Negative

Barkema & Vermeulen (1997) The Netherlands 72 countries, not specified Positive Contractor & Kundu (1998) Various, not specified Various, not specified Not significant Hennart & Larimo (1998) Japan & Finland U.S. Positive

Taylor, Zhou & Ousland (1998) Japan & U.S. Various, not specified Not significant/Negative3

Brouthers & Brouthers (2001) The Netherlands, Germany, Britain & U.S.

Hungary, Poland, the Czech Republic, Russia & Romania

Positive4

Chang & Rosenzweig (2001) Japan and 10 european countries including U.K., Germany, France & Nordic countries

U.S. Positive

Luo (2001) Various including U.S., Japan, Australia, Germany, France & U.K.

China Not significant

Chen & Hu (2002) Various, not specified China Not significant Yiu & Makino (2002) Japan 23 countries including U.S., U.K.,

Germany, the Netherlands and France

Positive/ Not significant5

Pak & Park (2004) Japan 61 countries, not specified Positive Xu, Pan & Beamish (2004) Japan 44 countries, not specified Not significant Cho & Padmanabhan (2005) Japan Various, 45 countries Not significant6

Wang & Schaan (2008) Japan Various, around 50 countries Curvilinear Wilkinson et al. (2008) Japan Various, not specified Negative7

Slangen & van Tulder (2009) The Netherlands Various Not significant López-Duarte & Vidal-Suarez

(2010)

Spain Various, 33 countries Negative/Not significant8

Chang et al. (2011) Taiwan 13 countries including U.S., Japan, Germany, India, Brazil & Vietnam

Negative9

1 Positive: When CD is high, the MNE is more likely to enter the foreign country through a JV Negative: When CD is high, the MNE is more likely to enter the foreign country through a WOS 2 However, negatively moderated by multinationality (international experience).

3 Not significant for Japanese firms. Negative for U.S. firms. 4 However, negatively moderated by host-country investment risk.

5 Positive, when a 100% and a 95% equity ownership in the foreign subsidiary was chosen as the cut-off point to distinguish between a WOS and a JV. Negative, for a 80% cut-off point.

6 However, negatively moderated by general international business experience, host-country specific experience and decision-specific experience.

7 However, positively moderated by subsidiary age.

8 Negative for the non-Spanish speaking subsample. Not significant for the Spanish speaking subsample. 9 However, positively moderated by governance quality

(11)

6

2 Literature review

This chapter presents an overview of the literature on entry mode selection and the ownership mode choice. Section 2.1 provides an overview of the different entry modes by which a firm can enter a foreign country. The characteristics of these different entry modes in terms of their level of control, resource commitment, return and risk are then discussed in section 2.1.2. Transaction cost economics (TCE) is applied in section 2.2 to explain how firms choose between these modes of entry and in which cases FDI is preferred over market transactions. Section 2.3 describes how MNEs select an ownership mode for their foreign subsidiaries, once they have decided to enter through FDI. An explanation of the concept of Cultural Distance (CD) and the types of uncertainty that it represents are provided in section 2.4. The next section (2.5) discusses the relationship between CD and ownership mode choice. Finally, in section 2.6 the moderating effect of international corporate strategy on the relationship between CD and the ownership mode choice is analyzed and hypotheses are derived.

2.1 Entry modes

When entering a foreign market, firms must select a mode of entry (Erramilli & Rao, 1993). An entry mode can be viewed as an ‘institutional arrangement’ (Anderson & Gatignon, 1986; 2) that allows a firm to enter its resources, such as products, technology and human skills into a foreign country (Erramilli & Rao, 1990). Firms can choose from three entry mode alternatives; exporting, contractual modes and Foreign Direct Investment (FDI) (Andreu, Claver & Quer, 2007). These three alternatives can be placed into two categories; equity and non-equity entry modes (Harzing, 2002; Andreu et al., 2007; Taylor et al., 1998; Tihanyi, Griffith & Russell, 2005). Non-equity entry modes include exporting, and contractual arrangements (Andreu et al., 2007). The latter is also referred to by Hennart (1988) as non-equity Joint Ventures and includes

arrangements such as licensing (Harzing, 2002; Hill et al., 1990) and franchising (Erramilli, Agarwal & Chekitan, 2002).

Exporting firms sell their products manufactured in the home country to other countries (Taylor et al., 1998). This can take place either through independent intermediaries or through channels that are owned by the firm itself (Erramilli, 1991). Firms can also decide to manufacture their products overseas (Andreu et al., 2007). This occurs either through contractual modes (Erramilli, 1991), which implies the transfer of technology and management systems to a party (licensee) in the host country (Taylor et al., 1998), or through

(12)

7

FDI, involving equity ownership and management control by the firm (Chang & Rosenzweig, 2001).

Firms that enter foreign markets through FDI and own and control operations in more than one country are called Multinational Enterprises (MNEs) (Dunning & Lundan, 2008).

2.1.2 Entry modes' characteristics

The different types of entry mode are associated with different levels of control (Hill et al., 1990; Luo, 2001; Taylor et al., 1998). Higher levels of ownership in the foreign operation will give the firm more control (Agarwal & Ramaswami, 1992; Erramilli, 1991; Gaur & Lu, 2007). Thus, as firms move from contractual arrangements to FDI, the level of ownership and control in the foreign operation increases (Erramilli & Rao, 1993; Hill et al., 1990). Anderson & Gatignon (1986:3) explain that control enables firms to ''influence systems, methods and decisions'' and therefore allows the firm to enjoy a larger share of profits resulting from the foreign activities.

However, higher levels of control not only result in higher returns but also higher risks (Agarwal & Ramaswami, 1992; Pak & Park, 2004). First, a higher degree of control implies taking responsibility for decision-making in a foreign country that the firm may be unfamiliar with (Anderson & Gatignon, 1986). Second, it also means committing more resources (Agarwal & Ramaswami, 1992). Due to switching costs (Anderson & Gatignon, 1986), a higher resource commitment lowers the firm's flexibility in choosing a different entry mode in case the current mode of entry proves to be unsuccessful (Anderson & Gatignon, 1986; Hill et al., 1990).

Entry mode selection can therefore be regarded as a trade-off between returns and risks (Luo, 2001). For making this trade-off, it is crucial to understand when increased control will become more desirable, and in which cases the benefits associated with higher control will outweigh the risks of higher resource

commitment (Anderson & Gatignon, 1986). The literature presents several theories and frameworks that attempt to shed light on these questions and explain the entry mode selection of firms. Among these theories, Transaction Cost Economics (TCE) has been the most applied by scholars (Taylor et al., 1998) and will be further discussed below.

(13)

8 2.2 TCE and entry mode selection

2.2.1 Market transactions versus FDI

TCE explains in which cases firms will prefer high control entry modes over low control modes (Erramilli & Rao, 1993). High control modes are equity modes (FDI), meaning that activities are internalized and thus carried out by the MNE's own staff members (Hennart, 1989). Low control modes are contractual

agreements, which give the external partner or supplier the responsibility for carrying out the task (Erramilli & Rao, 1993).

According to TCE, the firm will choose the entry mode that minimizes transaction costs (Taylor et al., 1998; Zhao et al., 2004). Transaction costs arise when business transactions between parties take place in the market (Luo, 2001). These costs include the ex ante costs of contract negotiation (Erramilli & Rao, 1993) and the ex post costs (Hill & Kim, 1988) of monitoring the behavior and performance of the parties involved (Brouthers, 2013; Taylor et al., 1998).

TCE argues that when there are many suppliers available to perform the task at hand and thus markets can be viewed as competitive, low-control entry modes are preferred (Anderson & Gatignon, 1986; Luo, 2001; Taylor et al., 1998). The reason for this is that in a competitive market, the threat of being easily replaced prevents opportunistic behavior (Anderson & Gatignon, 1986) and puts pressure on the supplier to perform efficiently (Taylor et al., 1998). Thus the need to monitor and enforce contracts is minimized (Erramilli & Rao, 1993) and therefore transaction costs are low.

However, when the supplier becomes irreplaceable ('small numbers bargaining'), he will have the tendency to behave opportunistically (Anderson & Gatignon, 1986). Due to bounded rationality, it is difficult for firms to write complete contracts or to evaluate external partners' performance (Brouthers & Brouthers, 2003). Strict negotiation and intensive monitoring are therefore necessary in order to reduce opportunistic behavior, which however significantly increase transaction costs (Erramilli & Rao, 1993).

In such cases, when the market fails, firms will have the incentive to take greater control and to internalize activities (Anand & Delios, 1997; Luo, 2001). By allowing the firm's own employees to carry out the activities, transaction costs can be reduced, since it is easier to monitor and control the behavior of own employees than that of external partners (Erramilli & Rao, 1993).

(14)

9

Rao, 1993; Hill et al., 1990). For transactions to be internalized, the firm must invest in administrative, legal and operational infrastructures (Erramilli & Rao, 1993), resulting in bureaucratic costs (Hill et al., 1990, Luo, 2001). In case the chosen entry mode proves to be unsuccessful, this high overhead makes it difficult for firms to switch between institutional arrangements and thus lowers the firm's flexibility (Anderson & Gatignon, 1986). Therefore, the firm will only choose a high control entry mode when the reduction in transaction costs is higher than the costs of internalizing activities (Hill et al., 1990).

2.2.2 Transaction-specific assets

TCE posits that the most important determinant of small numbers bargaining and consequent market failure is asset specificity (Kogut, 1988). Transaction-specific assets can be defined as specialized physical and human investments made by suppliers (Anderson & Gatignon, 1986; Erramilli & Rao, 1993), that will lose their value in alternative uses (Yiu & Makino, 2002; Zhao et al., 2004). Anderson & Gatignon (1986) explain that since these assets develop over time and have a significant impact on performance, the partner or

supplier who possesses the asset will be difficult to replace.

One type of transaction specific assets often mentioned in the literature is proprietary know-how (Anand & Delios, 1997; Anderson & Gatignon, 1986; Hill et al., 1990). Since proprietary know-how is viewed as a firm-specific advantage, the firm will want to prevent its leakage (Hil & Kim, 1988). However, by transferring this know-how through the market, the firm runs the risk of its maladaptation or unwanted dissemination by the external supplier (Hill et al., 1990). This in turn makes its transfer through contracts costly (Anand & Delios, 1997). Also, market transfer of proprietary know-how is associated with valuation problems (Hill & Kim, 1988; Madhok, 1997), since the buyer can only know what the knowledge is worth when the seller discloses the knowledge (Hennart, 1988). However, by revealing the knowledge, the buyer will not need to pay for it anymore (Anderson & Gatignon, 1986). These problems are also unlikely to be solved through patents, since the patent system only protects explicit knowledge (Brouthers & Hennart, 2007), and proprietary know-how is difficult to codify and be patented due to its tacit nature (Chen & Hu, 2002). Because of these problems, the firm will have the incentive to transfer the know-how internally (Hill et al., 1990) through the ownership of equity (Chen & Hu, 2002).

(15)

10 2.3 TCE and the ownership mode choice

Once the firm has chosen to internalize activities and enter the foreign market through equity ownership, it is faced with yet another decision (Pak & Park, 2004), which is referred to as the ownership mode choice (Brouthers & Hennart, 2007). This means that the firm has to select a level of ownership for the foreign subsidiary; sharing equity and control with a local partner in an equity JV (Chang et al., 2011; López-Duarte & Vidal-Suarez, 2010) or taking full ownership of the foreign subsidiary by setting up a WOS (Brouthers & Hennart, 2007; Chang et al., 2011; López-Duarte & Vidal-Suarez, 2010).

According to TCE, the choice between a JV and a WOS is determined by the relative costs and benefits of the two ownership modes (Hennart, 1991; Padmanabhan & Cho, 1996).

2.3.1 The establishment of JVs

From a TCE perspective, JVs are formed when there is a bilateral need to access complementary inputs (Brouthers & Hennart, 2007; Yiu & Makino, 2002), but due to failure of both input markets, their transaction through contracts is subject to high transaction costs (Hennart, 1991). Also, their replication or acquisition is costlier than obtaining them through the establishment of a JV (Kogut, 1988; Makino & Neupert, 2000; Yiu & Makino, 2002).

2.3.2 JVs versus market transactions

Hennart (1988; 1991) explains why the establishment of JVs for accessing complementary inputs is more efficient than their acquisition through the market.

JVs can be viewed as the result of double market failure (Brouthers & Hennart, 2007). This means that the market for the MNE's input as well as the market for the local supplier's input is subject to high transaction costs (Hennart, 1988). If one of these markets were not failing, the transaction could take place through the market and the needed inputs could be licensed (Hennart, 1988; Hennart, 1991).

However, when both markets fail, it is more efficient to bundle the needed inputs through the establishment of a JV. The reason for this is that in JVs, equity is shared and thus the input suppliers become co-owners in the venture (Hennart, 1991; Hennart & Larimo, 1988). Equity owners are paid ex post from the JV profits whereas contract suppliers are paid ex ante for their contribution (Brouthers & Hennart, 2007).

(16)

11

Hennart & Larimo (1988) explain that paying suppliers from the profits of the JV, reduces their incentive to behave opportunistically, since opportunism will result in a lower JV profit and therefore, the supplier will partly bear the cost of his own opportunistic behavior. Thus, the MNE can reduce transaction costs through the formation of a JV.

2.3.3 The preference for JVs over WOSs

Since opportunism could also be reduced through full-control modes (Hennart, 1991), it is also possible to access complementary assets through replication, by setting up a greenfield operation or through full acquisition, which are both wholly-owned subsidiaries (Hennart, 1988). TCE explains in which cases both options will be more costly than the formation of a JV and therefore MNEs will prefer JVs over WOSs.

Hennart (1991) argues that JVs are formed to access complementary assets that are public goods and firm-specific. Assets that are public goods can be shared at low marginal costs, meaning that it can be used for similar or complementary products at little extra cost. This allows owners of public goods assets to charge a low price for the services resulting from these assets. Therefore, setting up a greenfield operation will be more costly for the MNE than obtaining the assets through the establishment of a JV or through full acquisition.

Because of the firm-specific character of these assets, the MNE will also choose a JV over full acquisition (Hennart, 1991). Firm-specific means that these assets cannot be separated from those that the firm does not need (Hennart, 1991) but must be purchased as a package (Hennart, 1988). Therefore, when acquiring the entire firm, the MNE must also manage unfamiliar and unrelated businesses (Kogut,1988), thereby increasing management costs (Hennart, 1988). The MNE will therefore prefer a JV since this will reduce these costs.

2.3.4 The preference for WOSs over JVs

Hennart (1991) explains that although JVs lower the incentive for opportunistic behavior by making partners co-owners in the venture, they do not completely remove this incentive. Since a JV partner receives only a part of the residual profits, he will still have the incentive to act in his own self-interest. Therefore, when conflicts between JV partners are expected to arise, MNEs will want to take full control and set up a WOS.

(17)

12

& Rosenzweig, 2001; Kogut & Singh, 1988; Padmanabhan & Cho, 1996; Yiu & Makino, 2002) for similar reasons as to why FDI is preferred over market transactions.

First, the MNE will be faced with valuation problems as discussed earlier in sub-section 2.2.2. (Hennart, 1991). Also, the transfer of proprietary assets to the subsidiary, in particular know-how, will expose the MNE to the risk of free-riding (Malhotra, 2012; Pak & Park, 2004) and the unwanted dissemination of its know-how by the joint venture partner (Malhotra, 2012; Yiu & Makino, 2002). In this case, the transaction costs associated with a JV, including the costs of partner opportunism and monitoring will be significantly higher than that of a WOS (Chang et al., 2011). For these reasons, the MNE will have the incentive to choose a WOS, thereby protecting itself against these opportunistic hazards (Chen & Hu, 2002; Pak & Park, 2004) and reducing transaction costs.

(18)

13 2.4 Cultural Distance

2.4.1 Definition of Cultural Distance

With Hofstede's operationalization of national culture through cultural dimensions, the concepts of national culture and cultural distance have been frequently applied in international business research (Cho & Padmanabhan, 2005). National culture can be defined as the norms, values and beliefs shared by a nation (Harzing, 2004). In international business research, Cultural Distance (CD) refers to the differences in these norms,values and beliefs between the MNE's home country and the country where its subsidiary is located (Chang et al., 2011; Chen & Hu, 2002). When entering a culturally distant country for the first time, MNEs will perceive considerable uncertainty in the host country environment (Wang & Schaan, 2008). It is argued that cultural distance reflects two types of uncertainty; external and internal uncertainty (Slangen & van Tulder, 2009).

2.4.2 External uncertainty

External uncertainty is defined as the uncertainty perceived by the MNE in the formal and informal institutional environment of the target country (López-Duarte & Vidal-Suarez, 2010). High CD increases uncertainty in the informal institutional environment (Chang et al., 2011; López-Duarte & Vidal-Suarez, 2010) whereas high political risk (López-Duarte & Vidal-Suarez, 2010) or low governance quality (Chang et al., 2011; Slangen & van Tulder, 2009) increases uncertainty in the formal institutional environment.

High CD increases external uncertainty because of the lack of familiarity with the norms (Tihanyi, Griffith & Russell, 2005), values (Gatignon & Anderson, 1988) and business practices (Pak & Park, 2004) of the host country. When entering a foreign country for the first time, the firm will be unfamiliar with the local conditions (Hennart, 1991). This will reduce the firm's ability to operate successfully in the host country (Tihanyi et al., 2005). It is therefore suggested that as CD increases, the need for local knowledge will also increase (Brouthers & Brouthers, 2001). Local knowledge is developed over time by local firms operating in a country (Hennart, 1988). Hennart (1988) argues that due to its tacit nature, this type of knowledge is difficult to patent and to codify. This suggests that its sale and purchase through market transactions would be difficult and subject to high transaction costs (Hennart, 1988; López-Duarte & Vidal-Suarez, 2010). Therefore, for accessing this knowledge, firms need to enter a foreign country through FDI and more

(19)

14

specifically, they will need to establish a JV. This will be discussed further in chapter 2.5.

2.4.3 Internal uncertainty

Previous research on the relationship between CD and the ownership mode choice has traditionally tried to show how external uncertainty influences this relationship and has conceptualized this type of uncertainty by using CD (López-Duarte & Vidal-Suarez, 2010).

More recently, it has been argued by Slangen & van Tulder (2009) that CD not only reflects external uncertainty as is suggested in previous research but also internal uncertainty. Internal uncertainty is experienced by firms that enter a culturally distant country for the first time through FDI (Slangen & van Tulder, 2009).

This type of uncertainty refers to uncertainty caused by unintentional conflicts and misunderstandings between employees of foreign subsidiaries and managers of the parent MNE (Slangen & van Tulder, 2009), or between the local JV partner and the parent MNE (Barkema & Vermeulen, 1997; Ding, 1997; Park & Ungson, 1997; Slangen & van Tulder, 2009). As CD increases, the potential for these unintentional conflicts and misunderstandings will also increase (Slangen & van Tulder, 2009). Since internal uncertainty is directly related to JVs and WOSs, this type of uncertainty will be discussed in more detail in section 2.5.2.

(20)

15 2.5 CD and the ownership mode choice

2.5.1 External uncertainty and ownership mode choice

Previous studies suggest that MNEs can reduce the external uncertainty associated with entry into culturally distant countries through the formation of JVs (López-Duarte & Suarez, 2010; Hennart & Larimo, 1998; Kim & Hwang, 1992; Cho & Padmanabhan, 2005; Pak & Park, 2004; Wang & Schaan, 2008; Slangen & van Tulder, 2009).

As discussed in the previous chapter, MNEs entering a culturally distant country for the first time, lack familiarity with the local norms and values and therefore also with the practices and preferences of local stakeholders (Pak & Park, 2004; Slangen & van Tulder, 2009). This will cause difficulties in transferring practices and managing the relationships with local stakeholders (Agarwal, 1994; Wang & Schaan, 2008), which in turn will increase management- and information costs (Chang et al., 2011; Wang & Schaan, 2008). It is suggested that due to their knowledge of the host country's culture and practices (Agarwal, 1994; Barkema & Vermeulen, 1997; Brouthers & Brouthers, 2001; López-Duarte & Suarez, 2010; Pak & Park, 2004; Cho & Padmanabhan , 2005) and their location-specific experience (Wang & Schaan, 2008), local firms are better able to manage the local labor force and relationships with stakeholders such as suppliers and government authorities (Hennart & Larimo, 1998; Kogut & Singh, 1988). MNEs may therefore want to enter a culturally distant country through a JV with a local partner to gain complementary knowledge and

experience in order to operate successfully in a culturally distant market (Anand & Delios, 1997).

Another way how JVs reduce external uncertainty is through their shared-equity nature (Slangen & van Tulder, 2009). Since equity is shared within a JV, the resource commitment by the MNE will be less compared to a WOS (Anderson & Gatignon, 1986; Kim & Hwang, 1992; Slangen & van Tulder, 2009; Tihanyi et al.,2005). The lower resource commitment implies lower exposure to risk (Tihanyi et al., 2005) as it reduces exit costs and thus increases the MNE's flexibility (Brouthers & Brouthers, 2001; Kim & Hwang, 1992; Slangen & van Tulder, 2009). This is advantageous for the MNE since the likelihood that entry into a culturally distant country will fail is higher than entry into countries with low cultural distance (Slangen & van Tulder, 2009).

(21)

16

2.5.2 Internal uncertainty and ownership mode choice

Entry into culturally distant countries through JVs does not only decrease external uncertainty, but also increases internal uncertainty (Slangen & van Tulder, 2009). Internal uncertainty results from

misunderstandings and conflicts between the JV partners (Barkema & Vermeulen, 1997; Ding, 1997; Park & Ungson, 1997; Slangen & van Tulder, 2009). The greater the cultural distance between the home and host country, the higher the probability of conflicts between the JV partners (Ding, 1997; Slangen & van Tulder, 2009).

These conflicts arise from difficulties in inter-partner communication (Park & Ungson, 1997) which is the result of differences between the JV partners. The higher the CD between the home and host country, the more dissimilar their approaches to communication and management (Ding, 1997; Slangen & van Tulder, 2009), administrative and organizational practices (Kogut & Singh, 1988) and response to strategic matters (Park & Ungson, 1997).

It is argued that these misunderstandings and conflicts impede the effective transfer of knowledge (Slangen & van Tulder, 2009) and cooperation between partners (Ding, 1997; Slangen & van Tulder, 2009). This may eventually lead to deterioration of performance (Ding, 1997; Slangen & van Tulder, 2009) and even the dissolution of the joint venture (Barkema & Vermeulen, 1997; Park & Ungson, 1997). The shared equity nature of JVs does not provide a remedy in this case, since in contrast to partner opportunism, these conflicts are not intentional and thus cannot be prevented (Slangen & van Tulder, 2009).

Slangen & van Tulder (2009) explain that CD may also cause internal conflicts between employees of local subsidiaries and parent MNE managers in WOSs. However, these conflicts are less likely to cause internal uncertainty, since it is easier to deal with them in WOSs. Because MNEs fully control a WOS, local employees that do not fit in the organization because of cultural differences, can be replaced relatively easy. This is more difficult in JVs however, due to the decision-making power of the local JV partner.

2.6 The moderating effect of international corporate strategy

2.6.1 Types of international corporate strategy

Firms with global operations must have a strategy for positioning themselves in the international business environment and for creating and sustaining competitive advantage (Harzing, 2002). This is referred to in the

(22)

17

literature as the international corporate strategy of the firm (Harzing, 2002). Four types of international corporate strategies are mentioned in the literature; Global, Multi-domestic, Transnational (Bartlett & Ghoshal, 1989; Martinez & Jarillo, 1991) and International strategy (Harzing, 2000). Following Harzing (2002), only the global and multi-domestic strategies will be discussed in this study. This is because the Transnational strategy is a combination of the Global and Multi-domestic strategy (Harzing, 2000).

Therefore, it shows characteristics of both these two strategies and is difficult to clearly distinguish from the Global and Multi-domestic strategies. Also, the International strategy has been considered to be the same as the Transnational strategy in many previous studies and those few studies that did make a distinction, were not able to empirically differentiate it from the other strategies (Harzing, 2000). Harzing (2002:212) therefore concludes that the Global and Multi-domestic strategies ''are the most commonly accepted and clearly defined''.

2.6.1.1 Global strategies

The improvements in transport and communication technologies in the last few decades have resulted in the homogenization of the tastes and preferences of consumers worldwide (Hill et al., 1990). Due to this, the competition in certain industries has been globalized (Harzing, 2000; Hill et al., 1990; Martinez & Jarillo, 1991) and products in these industries are highly standardized (Harzing, 2002). Examples of such industries include semiconductors, watches and aerospace (Porter, 1986).

The standardization of products enables firms operating in these industries to implement a Global strategy, which is focused on achieving cost advantages through the realization of economies of scale (Harzing, 2002; Hill et al., 1990). In order to realize economies of scale, firms integrate their value chain activities and centralize production in only a few locations (Hill et al., 1990; Martinez & Jarillo, 199; Harzing, 2000).

2.6.1.2 The role of subsidiaries in Global strategies

For producing in a cost-efficient way and realizing economies of scale (Harzing, 2002), a high level of interdependence between subsidiaries is required (Hill et al., 1990; Martinez & Jarillo, 1991).

Subsidiaries of Global MNEs usually specialize in one part of the production process only and sell and buy components and products to and from other subsidiaries (Hill et al., 1990). This is reflected in a high level of intra-firm sales and purchases (Harzing, 2000).

(23)

18

Also, in order to coordinate the different activities of subsidiaries world-wide (Slangen & Hennart, 2002), a high level of control is needed (Hill et al., 1990) and therefore decision-making will be centralized at headquarters (Harzing, 2000). Decisions about for example output levels and pricing are taken by headquarters (Hill et al., 1990) and subsidiaries are expected to carry out these decisions and strategies (Slangen & Hennart, 2008; Harzing, 2000). These subsidiaries thus have very little autonomy and are highly dependent on MNE's headquarters (Slangen & Hennart, 2008, Harzing, 2000).

2.6.1.3 Multi-domestic strategies

MNEs implementing a Multi-domestic strategy operate in industries where competition takes places on a domestic level (Harzing, 2002) and products and activities are highly differentiated (Martinez & Jarillo, 1991). This is because for some type of products, the tastes and preferences of consumers, as well as political-, legal-, and market conditions across countries differ (Hill et al., 1990, Martinez & Jarillo, 1991). Therefore, MNEs operating in industries such as consumer packaged goods and retailing (Porter, 1986), need to be locally responsive, by adapting products and processes to local conditions (Harzing, 2000).

2.6.1.4 The role of subsidiaries in multi-domestic strategies

In order to be able to successfully adapt products to local needs and conditions, foreign subsidiaries will be given considerable autonomy by the headquarters (Harzing, 2002; Slangen & Hennart, 2008; Xu & Shenkar, 2002), as they are closer to the end consumer. This means that each subsidiary will have its own production, marketing (Hill et al., 1990; Xu & Shenkar, 2002) and R&D facilities (Harzing, 2000). Also, decision-making takes place at subsidiary level and is decentralized (Harzing, 2000). The subsidiaries of these MNEs thus operate independently of other subsidiaries and headquarters (Harzing, 2000). Since the activities of the different subsidiaries do not need to be coordinated and value chain activities are not integrated, the degree of control exercised by the MNE is low (Hill et al., 1990).

(24)

19

2.6.2 CD, ownership mode choice and international strategy

As discussed in the previous sections, CD reflects both external- and internal uncertainty. The previous discussion suggests that MNEs entering a culturally distant country for the first time through WOSs will mainly experience external uncertainty while entering through a JV increases internal uncertainty. It is therefore not clear why the MNE would prefer one mode over the other. In this sub-section, it will be

explained how the level of external and internal uncertainty experienced by the MNE seems to depend on the type of international corporate strategy carried out and how these two types of uncertainties can be reduced through the firm's international strategy.

2.6.2.1 CD, JVs and Multi-domestic strategies

While all firms entering a culturally distant country for the first time will face some level of external uncertainty due to unfamiliarity with the local norms and practices, the external uncertainty experienced by firms implementing a multi-domestic strategy is expected to be significantly higher than for firms

implementing a global strategy. The reason for this is that the strategy of multi-domestic MNEs requires access to local knowledge in order to be able to adapt their products to local conditions and operate in that environment successfully (Slangen & Hennart, 2002; Brouthers & Hennart, 2007). It is therefore argued that the need for a local partner, who is familiar with the local norms and practices and who can help the firm access local knowledge, is thus especially high for firms implementing a multi-domestic strategy (Brouthers & Hennart, 2007). These MNEs can therefore substantially reduce external uncertainty through JVs.

At the same time, the international strategy of these MNEs helps them reduce internal uncertainty. It is argued that internal uncertainty resulting from conflicts and misunderstandings between JV partners is likely to occur only when there are frequent interactions between the JV partners (Brouthers & Hennart, 2007). Since subsidiaries of MNEs implementing a Multi-domestic strategy are highly autonomous and independent of the parent, interaction between the parent and the subsidiary will be low (Slangen & Tulder, 2008). This will therefore reduce the likelihood of conflict between JV partners (Slangen & Hennart, 2008) and thereby also internal uncertainty.

Thus, MNEs carrying out a Multi-domestic strategy can reduce external uncertainty by entering a foreign country through a JV while at the same time, the autonomy that subsidiaries of these MNEs are given, helps reduce the internal uncertainty that is associated with JVs. It is therefore expected that:

(25)

20

H1: When entering a culturally distant country for the first time, MNEs implementing a multi-domestic strategy are more likely to enter that country through a JV than a WOS.

2.6.2.2 CD, WOSs and Global strategies

As discussed earlier, since MNEs implementing a global strategy need to coordinate the activities of their foreign subsidiaries and these subsidiaries are expected to carry out the parent's decisions, frequent

interaction between the parent and subsidiary is necessary (Slangen & Hennart, 2002). Entering a culturally distant country through a JV would therefore substantially increase the likelihood of conflict and

misunderstandings and thus internal uncertainty. By entering through a WOS however, the likelihood of internal conflicts and thereby internal uncertainty can be reduced.

Also, since MNEs carrying out a global strategy produce standardized products that do no need to be adapted to local conditions, it is expected that unfamiliarity with the local norms and practices will pose less of a problem than for MNEs carrying out a multi-domestic strategy. Thus, the external uncertainty experienced by these MNEs will be relatively low and therefore, there is less need for a local partner.

MNEs implementing a global strategy can reduce the likelihood of internal uncertainty by entering through a WOS. At the same time, due to the standardization of their products, the external uncertainty experienced by these MNEs will be relatively low. It is therefore expected that:

H2: When entering a culturally distant country for the first time, MNEs implementing a global strategy are more likely to enter that country through a WOS than a JV.

(26)

21

Before describing the sample and method that is used to test the above hypotheses, a conceptual model will be presented below, which shows the relationship between the ownership mode choice, CD and the

international corporate strategy of the MNE:

M

X y

Figure 2.1: Conceptual model

Cultural

Distance

Ownership

mode choice

International

Corporate

Strategy

(27)

22

3 Methodology

In this chapter, it will be explained how the hypotheses that were derived from the literature review are empirically tested. In section 3.1, the sample of the study and sources for data collection will be discussed. The operationalization of the dependent, independent, moderating and control variables will be presented in section 3.2. Finally, section 3.3 discusses the statistical method that is applied in this study.

3.1 Sample and data collection

Since European countries are underrepresented in previous studies on the relationship between CD and the ownership mode choice, a European country (the Netherlands) will be taken as the home-country in this study. Although, the Netherlands is one of the smallest countries in Europe, data collected by the

International Monetary Fund shows that the Netherlands was the country with the highest outward FDI level in the world for the year 2013 (Data.imf.org, 2014).

The sample contains Dutch firms listed on Euronext Amsterdam. The AEX (Amsterdam Exchange Index) shows the performance of the 25 most actively traded Dutch companies' shares (AEX-index, n.d.).

The shares of 25 medium-sized firms are traded on the AMX (Amsterdam Midcap Index) (AMX Quote, n.d.) The reason for selecting companies that are listed on Euronext Amsterdam is that information about these companies is readily available since these are public companies and are therefore required to publish (annual) reports. Several companies are however not included in the sample. First, a study conducted by Hennart & Larimo (1998) suggests that the national origin of the MNE influences the ownership mode choice. Therefore, 10 of the 50 companies that did not have a 100% Dutch origin were excluded from the sample. Also, another 4 companies were left out as they did not have any foreign subsidiaries.

Further, since an index developed by Kobrin (1991) was used for the operationalization of the moderating variable and this index only includes manufacturing industries, the service companies listed on Euronext Amsterdam had to be excluded, which reduced the sample to 13 companies. Following Barkema & Vermeulen (1997), three companies, Akzo Nobel, Philips and Heineken, were also excluded since these companies are considerably different from the others in terms of size and international experience.

Since the focus of the present study is the ownership mode that is chosen to establish the foreign subsidiary, the level of analysis is the MNEs' foreign subsidiaries. The database Orbis was used to identify the foreign subsidiaries of the remaining ten firms in the sample. Subsidiaries in which the MNEs had a direct ownership

(28)

23

were included in the sample. 21 subsidiaries were left out due to missing data. Some subsidiaries were excluded from the sample as Hofstede’s cultural dimensions indices were not available for the countries in which these subsidiaries were located. Also, several subsidiaries had to be excluded from the sample due to missing data on government restrictions. The final sample contained 1044 subsidiaries, 548 WOSs and 496 JVs.

The following secondary sources were used for data collection: The database Orbis, the European Commission's Market Access Database, the OECD, Worldbank and Hofstede centre’s websites.

3.2 Variable Measurement 3.2.1 Dependent variable

The dependent variable in this study is the mode of ownership that is used to establish the foreign subsidiary (a JV or a WOS). Dummy variables were created where the dependent variable has the value of 1 when the chosen ownership mode is a JV and 0 when the mode of ownership is a WOS.

A review of the literature shows that scholars do not agree on what ownership percentage should be adopted for differentiating between a JV and a WOS. Some of the previous studies have used a 95% cutoff point, meaning that MNEs owning more than 95% of the equity in the foreign subsidiary are defined as WOSs (e.g. Hennart & Larimo, 1998; Padmanabhan & Cho, 1996; Wang & Schaan, 2008; López-Duarte & Vidal-Suarez, 2010). Some scholars have defined WOSs as subsidiaries in which the MNE holds 100% equity (e.g.

Gatignon & Anderson, 1988; Slangen & van Tulder, 2009; Barkema, Bell & Pennings, 1996), while still others have used an 80% cutoff point (Yiu & Makino, 2002). Since the database Orbis, defines a WOS as those subsidiaries in which the parent holds 98% or more equity, this cut-off point will be used in this study for the definition of a WOS.

Also, for the definition of JVs, different minimum ownership levels have been used in previous studies. Some believe that at least a 5% equity ownership is necessary to define the foreign subsidiary as a JV (e.g. Wang & Schaan, 2008; Makino & Neupert, 2000) otherwise the subsidiary would be considered a portfolio investment instead of FDI (López-Duarte & Vidal-Suarez, 2010).Other scholars have used a minimum ownership level of 10% (e.g. López-Duarte & Vidal-Suarez, 2010; Hennart & Larimo, 1998; Padmanabhan & Cho, 1996). There are also studies that do not use a minimum ownership level (e.g. Chang et al., 2011; Pak & Park, 2004). Following López-Duarte & Vidal-Suarez (2010) & Hennart & Larimo (1998), the 10%

(29)

24

minimum ownership level will be used in the present study. Thus, subsidiaries in which the parent holds 98% or more equity are coded as a WOS and a 10-98% equity ownership is coded as a JV.

3.2.2 Independent variable

The independent variable in this study is Cultural Distance. Following previous research (e.g. Barkema & Vermeulen, 1997; Brouthers & Brouthers, 2001; Slangen & van Tulder, 2009), the cultural distance between the Netherlands and each host-country is measured by using the Euclidean distance index. This index is derived from the Kogut and Singh index (1988). The Kogut & Singh index is based on the scores of the home-country and each host-country on the four national cultural dimensions developed by Hofstede (Slangen & van Tulder, 2009). These cultural dimensions are called uncertainty avoidance, power distance, individualism and masculinity (Chang & Rosenzweig, 2001). A fifth dimension, named long term orientation was later added by Hofstede (Barkema & Vermeulen (1997), however since scores on this dimension are only available for 23 countries (López-Duarte & Vidal-Suarez, 2010), long term orientation was not included in the present study.

The difference between the Kogut and Singh index and the Euclidean distance index is that there is an implicit assumption in the Kogut & Singh index, suggesting an equal importance of the four cultural dimensions when measuring the cultural distance between countries (Shenkar, 2012; Slangen & van Tulder). This assumption, which is not proven yet, is relaxed by the Euclidean distance index (Slangen & van Tulder, 2009). In order to calculate the CD between the Netherlands and each host-country with the Euclidean distance index, the cultural dimension scores for each country were collected from the website Hofstede centre (www.geert-hofstede.com).

Slangen & Hennart (2008) point out that the validity of Hofstede's cultural dimensions has been empirically confirmed in several studies, which makes them appropriate for determining the CD between countries. Finally, since many of the studies on the relationship between cultural distance and the ownership mode choice have measured cultural distance with this index, using the index also for this study facilitates comparability (Wang & Schaan, 2008).

(30)

25

3.2.3 Moderating variable

The moderating variable in this study is the international corporate strategy of the MNE. Several proxies for this variable can be derived from Harzing's paper (2000). First of all, as explained in section 2.6, since multi-domestic MNEs have to adapt their products to local needs and preferences, marketing and R&D activities are mainly carried out by the subsidiaries. Therefore, if the subsidiaries' marketing and R&D expenses are high relative to the MNE's HQ marketing and R&D expenses, this would indicate a multi-domestic strategy while the opposite would suggest that the MNE is carrying out a global strategy.

Also, as explained in section 2.6, subsidiaries of global MNEs usually specialize in one part of the production process only and sell and buy components to and from other subsidiaries. Therefore, a high percentage of intra-company sales relative to total sales would indicate that the MNE is carrying out a global strategy and the opposite would suggest a multi-domestic strategy.

However, since data could not be obtained for these subsidiary level variables through the secondary sources available, international corporate strategy was operationalized by using the transnational integration index (table 3.1) developed by Kobrin (1991). The index is available for 56 manufacturing industries and is measured as the proportion of international sales that are intrafirm (Kobrin, 1991). The index values are based on data from a 1982 benchmark survey on US FDI. The higher the value of the index for a certain industry, the more global that industry is. The NAICS 2012 industry classification code stated in Orbis was used to determine the industry the 10 companies in the sample are operating in. The companies were then assigned the index value of that industry. An overview of the industries and their indices are presented in table 3.2

(31)

26

Table 3.1: Index of transnational l integration for the companies in the sample

MNE’s name Industry Index

Aalberts Plumbing 0,097

Accell Other transport 0,084 ASM International Electronic components 0,385 ASML Electronic components 0,385 Corbion Other food products 0,090 DSM Other chemical products 0,196 Nutreco Grain mill products 0,077

Ten Cate Textiles 0,109

TKH Non-ferrous metals 0,228 TomTom Communications equipment 0,404

3.2.4 Control variables

Previous research has identified several variables besides cultural distance that may affect the ownership mode choice (Brouthers & Brouthers, 2001). Five control variables were therefore included in this study. Two of these variables (size and international experience) are firm-level variables and the remaining three are country-level variables.

3.2.4.1 MNE’s size

Larger firms have more financial and managerial resources available which makes it easier for them to enter a foreign market and bear the costs and risks associated with foreign entry and operation (Agarwal &

(32)

27

Ramaswami, 1992; Quer, Claver & Rienda, 2007; Wang & Schaan, 2008). This suggest that these companies are more likely to prefer high control entry modes such as wholly owned subsidiaries (Agarwal &

Ramaswami, 1992; Quer, Claver & Rienda, 2007;Wang & Schaan, 2008) which require higher resource commitment and are associated with higher risks (Erramili & Rao, 1993). Smaller companies on the other hand, often lack the required resources and therefore may need to set up a JV in order to pool resources and share the costs and risks with a local partner (Barkema & Vermeulen, 1997).

Firm size will be measured in this study by the log of the total number of employees worldwide (Erramili & Rao, 1993, Gatignon & Anderson, 1988; Brouthers & Brouthers, 2001; Wang & Schaan, 2008) for the year 2013. Data for this variable was collected from the database Orbis.

3.2.4.2 MNE's international experience

Companies with greater international experience tend to prefer entry through the establishment of WOSs since international experience reduces the risks and costs associated with entering a foreign market (Erramili, 1991, Brouthers & Brouthers, 2003). This cost- and risk reduction results from the fact that these companies have developed systems (Brouthers & Brouthers, 2001) and have gained the know-how (Barkema, Bell & Pennings, 1996) to deal with new entries. Firms with little international experience lack this know-how, which results in unfamiliarity costs and places them at a disadvantage compared to local firms (Eden & Miller, 2004). By selecting a local joint venture partner, information can be gained and the costs of unfamiliarity can be reduced (Eden & Miller, 2004).

Following Barkema, Bell & Pennings (1996) and Slangen & Hennart (2008), international experience will be operationalized by using the log of the total number of foreign countries in which subsidiaries are located. This number was obtained from Orbis.

3.2.4.3 Host countries' legal restrictions

Several scholars suggest that legal restrictions imposed by the host-country affect the ownership mode choice (Chang et al., 2011; Brouthers & Brouthers, 2003; Slangen & van Tulder, 2009; Padmanabhan & Cho, 1996). Some countries or regions within countries do not allow foreign firms to establish wholly owned subsidiaries (Anderson & Gatignon, 1988, Brouthers & Brouthers, 2003; Chang et al., 2011). MNEs can only enter these countries through the establishment of Joint Ventures. Other countries impose restrictive policies such as

(33)

28

governmental approval before the establishment of a subsidiary (Padmanabhan & Cho, 1996). Padmanabhan & Cho (1996) argue that these policies are expected to discourage full ownership modes.

This study will control for legal restrictions through a dummy variable with value 1 for entries in countries that do not allow the establishment of WOSs or impose any other restrictions on FDI, and 0 for entries into countries where there are no such restrictions. The OECD FDI regulatory restrictiveness index and The Market Access Database of the European Commission were used to control for this variable.

Previous studies have checked for restrictions at the time of entry (e.g. Brouthers & Brouthers, 2003; Padmanabhan & Cho, 1996). Although the subsidiaries’ establishment year is available for most subsidiaries in the database Orbis, the only way to obtain this information is by clicking on the individual subsidiary names. Since more than 1000 subsidiaries are included in the sample, this approach would not be feasible due to time limitations. Also, since Orbis shows the current ownership mode of the subsidiaries and the ownership mode might have changed since the establishment of the subsidiary, information on government restrictions at the time of establishment would not be very useful. Therefore, it was checked whether there were any restrictions in the host countries included in the sample for the most recent year available (2013).

3.2.4.4 Host countries' economic development

It is argued by several scholars that the economic development of the host country can be expected to be correlated with cultural distance (Barkema & Vermeulen, 1997; Harzing, 2004; Chang et al., 2011). Harzing (2004) warns that cultural distance might therefore also be a proxy for economic development if this variable is not included as a control variable (Harzing, 2004). Very few studies have included this variable as a control variable (Harzing, 2004). Indeed, of the studies reviewed (see table 1.1), economic development has been controlled for, only by three studies. These studies have taken the GNP or GDP per capita of the host country as a proxy for economic development (Barkema & Vermeulen, 1997; Chang et al., 2011).

To control for economic development in this study, the GDP per capita (year: 2013) of the host countries included in the sample was obtained from the World bank website. The natural logarithm of the GDP per capita values was calculated and used in order to correct for skewness (Chang et al., 2011).

(34)

29

3.2.4.5 Host countries’ political risk

Political risk is also expected to be correlated with cultural distance (Harzing, 2004) and was therefore controlled for in this study. Political stability is taken as a proxy for political risk.

Data for this variable was obtained through the Worldwide Governance indicators (WGI) developed by Kaufmann, Kraay and Mastruzzi (2010). The WGI consist of 6 dimensions of governance. One of these dimensions is ‘political stability and absence of Violence/Terrorism’ which is defined as:

‘Perceptions of the likelihood that the government will be destabilized or overthrown by unconstitutional or violent means, including politically-motivated violence and terrorism’ (Kaufmann, Kraay and Mastruzzi, 2010; 4).

Scores for the political stability dimension are available for 215 countries over the 1996-2012 period on the World bank website. In the present study, the average of the 1996-2012 scores was calculated for the

countries included in the sample and taken as the political stability score for these countries. The reason why the average score of all the years available was taken instead of the score for the most recent year, is that for some countries the scores showed significant changes throughout the years. The average of the 1996-2012 would therefore be a more accurate representation of the political stability of a country than the score for the most recent year.

Before presenting the statistical method in the next section, an overview of the different variables and their operationalization is presented in table 3.1.

Table 3.2: Variable description and measurement

Type of variable Variable description Operationalization and measurement

Dependent variable Ownership mode Dummy variable:

> = 98% equity ownership by MNE: WOS (0)

10% – 98% equity ownership by MNE: JV (1)

Independent variable Cultural distance Euclidean distance index Moderating variable International corporate strategy MNE's organizational structure Control variable 1 MNE’s size Total number of employees world-wide

(log)

Control variable 2 MNE's international experience # of host-countries (log)

Control variable 3 Host countries’ economic development GDP per capita (natural logarithm) Control variable 4 Host countries’ political risk Political stability score (WGI) Control variable 5 Host-countries' legal restrictions Dummy variable:

No restrictions (0) Restrictions (1)

(35)

30

3.3 Statistical Method

Since the dependent variable in the present study is dichotomous (JV or WOS), binary logistic regression was used to test the hypotheses (Gatignon & Anderson, 1988; Hennart, 1991; Erramilli & Rao, 1993; Hennart and Larimo, 1998; Slangen & Hennart, 2008).

The formal expression of logistic regression models is as follows (Slangen & Hennart, 2008):

In this expression, ‘Yi is the dependent variable, Xi is the vector of independent variables for the ith

observation, a is the intercept parameter and β is the vector of regression coefficients’ (Slangen & Hennart, 2008; 481). The dependent variable in this study is the ownership mode choice and the dummy variable coded 1 means JV entry whereas 0 means WOS entry. Therefore, a positive regression coefficient shows that the independent variable increases the probability of a JV entry while a negative regression coefficient implies the opposite relationship.

Two models were designed to examine the effect of the variables on the ownership mode choice. The first model only contains the control variables and the regression was run by using the binary logistic

regression command in SPSS. For the second model however, an add-on named Process, written by Andrew F. Hayes, was installed and used in SPSS. The process add-on allows for the estimation of direct and indirect effects in mediation and moderation models. A model must be selected before running the test. A document with 74 model templates was downloaded together with the add-on from the website www.processmacro.org. Model 1, which includes 1 moderator, was selected. An interaction variable did not have to be created since the moderating variable can be directly added in Process. Version 22 of SPSS was used to conduct statistical analysis.

Cultural distance and the ownership mode choice : the moderating effect of International Corporate Strategy