The Spatial Structure of Foreign Subsidiaries and MNE Expansion Strategy

Guoliang F. Jiang *Dalhousie University

Rowe School of Business6100 University Avenue,

Halifax, Nova Scotia, CanadaB3H 4R2

Tel: (902) 494-8999Fax: (902) 494-1107

Email: frank.jiang@dal.ca

Guy L. F. HolburnWestern University

Ivey Business School1255 Western Road,

London, Ontario, CanadaN6G 0N1

Tel: (519) 661-4247Fax: (519) 661-3485

Email: gholburn@ivey.uwo.ca

Paul W. BeamishWestern University

Ivey Business School1255 Western Road,

London, Ontario, CanadaN6G 0N1

Tel: (519) 661-3237Fax: (519) 661-3485

Email: pbeamish@ivey.uwo.ca

* Corresponding author

(Forthcoming at the Journal of World Business)

1

The Spatial Structure of Foreign Subsidiaries and MNE Expansion Strategy

ABSTRACT

Drawing on internalization theory and economic geography research, we examine how the

spatial structure of MNE subsidiaries in supranational regions affects subsidiary location choices.

Our analysis of foreign production investments by Japanese manufacturing firms from 1971-

2006 supports our theoretical predictions: firms were more likely to establish new production

subsidiaries in countries geographically more proximate to existing production subsidiaries, but

not to trading subsidiaries, in the same region. The proximity effect diminished for production

subsidiaries engaged in accessing natural resources or R&D. Performance of production

subsidiaries was also stronger for those closer to other production subsidiaries in the same

region.

Keywords: Spatial transaction costs; Geographic distance; Internalization; Location choice

2

1. Introduction

What geographic factors influence MNE location strategy? Research on country location of

MNE subsidiaries has largely focused on the effect of distance between an MNE’s home country

and a potential host country (Kang & Jiang, 2012; Ragozzino, 2009; Slangen, 2011), and on host

country location-specific attributes such as industry agglomeration (Head, Ries, & Swenson,

1995; Kim, Delios, & Xu, 2010; Wheeler & Mody, 1992), resource access (Kolstad & Wiig,

2012; Schotter & Beamish, 2013), and government policies towards FDI (Mudambi, 1995; Zhou,

Delios, & Yang, 2002). While there is an increasing emphasis on how the regional and global

configuration of MNE subsidiaries affects MNE investment strategy and performance

(Beugelsdijk & Mudambi, 2013; Dunning, 1998), prior studies have not typically accounted for

cross-border spatial linkages between subsidiaries, a defining attribute of MNEs’ international

operations. In this study, we focus on cross-border linkages within the firm and specifically

examine the question: how do spatial relations between subsidiaries influence MNE expansion

strategy?

MNEs are complex geographic networks of activities undertaken at interdependent

subsidiaries, linked by cross-border flows of goods, information, finance and managerial

authority. The cross-border spatial structure of an MNE’s subsidiaries reflects accumulated

country choices for investment locations over time, creating a corporate geography superimposed

on territorial geography. A subsidiary and the country market in which it is located are

simultaneously situated in both territorial geography and corporate geography. Spatial relations

among subsidiaries thus generate a distinct layer of spatial variation that can influence MNEs’

investment strategy. This type of spatial variation differs from that arising from home-host

country distance (Kang & Jiang, 2012; Slangen, 2011) or subnational industry clustering (Head

3

et al., 1995; Kim et al., 2010): the former centers on the firm and its internal relationship whereas

the latter focuses on macro-level country or cluster characteristics that are exogenous to the focal

firm. By examining an understudied yet consequential geographic aspect of the MNE, this study

provides new conceptual and empirical insights to the analysis of MNE expansion strategy.

We augment an internalization theory framework with core concepts from economic

geography research, especially the notion of spatial transaction costs, to explore the relationship

between corporate geography and FDI strategy (McCann, 2008; McCann & Shefer, 2004). The

theoretical premise of our study is that spatial transaction costs between subsidiaries, consisting

of information and transport costs, create barriers to internalization and foreign investment. We

argue that MNEs can reduce cross-border spatial transaction costs by locating new production

subsidiaries closer to existing ones within the same supranational regions, assuming subsidiaries

within the same MNE are willing to cooperate. We argue further that the benefit of proximity

depends on the functional focus (specifically, production and trading) of existing subsidiaries,

and on the strategic mandate of the new subsidiary (such as accessing local natural resources or

conducting R&D).

Our analysis of foreign production investments by Japanese public manufacturing firms

between 1971 and 2006 confirms our central proposition that the proximity of a potential host

country to an MNE’s existing production subsidiaries, but not to trading subsidiaries, in the same

supranational region increases the probability of production entry into that country. However,

this positive influence of geographic proximity diminishes when a new subsidiary is engaged in

accessing local natural resources or conducting R&D. Yet, the proximity to trading subsidiaries

increases the probability of production entries with an R&D mandate. Finally, we find consistent

proximity effects when we examine subsidiary performance, namely that subsidiaries more

4

proximate to other subsidiaries within the same region tend to perform better than more distant

ones.

Overall, this study makes three contributions. First, in response to recent calls for more

research on the organizational spatial dimension of MNE geography (Beugelsdijk, McCann, &

Mudambi, 2010; Beugelsdijk & Mudambi, 2013; Buckley, 2009), our analysis demonstrates that

cross-border spatial relations between subsidiaries constitute an important corporate geographic

characteristic that affects the relative attractiveness of a potential investment location. Second,

we provide a more fine-grained location choice analysis by distinguishing between production

and trading subsidiaries and also between strategic mandates of new entries. Our results reveal

that territorial geography and corporate geography simultaneously influence location choices and

that their relative impact is dependent on a subsidiary’s functional and strategic focuses. Finally,

building on the core concept of spatial transaction costs from the economic geography literature,

this study enriches internalization research by offering new insights into an understudied spatial

determinant of internalization effectiveness – inter-subsidiary proximity. As far as we are aware,

our analysis is among the first to systematically examine how the spatial structure of MNE

subsidiaries affects MNE expansion strategy. A direct practical implication is that intrafirm

spatial proximity at a regional level may offer a competitive advantage when MNE operations

become increasingly interdependent (Ahlstrom, 2015; Iammarino & McCann, 2013).

2. Theoretical Background

Internalization is a central concept in international business that is at the core of theories of

foreign direct investment (Buckley & Casson, 1976; Rugman, 1981). Internalization theory

argues that firms and markets are alternative institutions for organizing interdependent economic

5

activities located in different countries (Hennart, 1982). Conceptualized as a process of making a

market within a firm, internalization is a mechanism for retaining competitive advantages

derived from the firm’s intangible assets and capabilities that cannot be readily transferred across

firm boundaries using market-based mechanisms such as licencing. The MNE thus manages a

complex set of interrelated cross-border activities using a planned system of internal markets

rather than resorting to imperfect or nonexistent external markets.

Despite the broad applicability of internalization theory, the organizational costs of

internalization, especially those associated with MNE spatial structure, are often

underemphasized in previous research, and the benefits and costs of internalization are often

assumed to be invariant to distance (Buckley, 2009, p.234). Although prior research has

combined location factors and internalization in explaining the development of MNEs, it has

primarily focused on cost minimization based on transport cost savings or cost differentials in

inputs, such as labor and natural resources (Iammarino & McCann, 2013; Rugman, 1981). In

particular, Beugelsdijk et al. (2010) argued that researchers should distinguish between two

separate yet related dimensions of MNE geography – ‘place’ and ‘space’. While place is

concerned with location-specific attributes – such as natural resource endowment, institutional

environment, and home-host country distance – and their impact on MNE foreign location

choices, space emphasizes the role of the firm in geographic space based on organizational

connectivity. International business research has largely focused on the ‘place’ aspect of location

strategy (Kang & Jiang, 2012) and the variation in ‘space’ arising from industry clustering (Kim

et al., 2010). The extant literature, however, has developed fewer insights about the geographic

configuration or spatial structure of MNE subsidiaries, overlooking a form of spatial variation

that can impact spatial transaction costs within the MNE.

6

In the economic geography literature, spatial transaction costs are broadly defined as “the

costs associated with engaging in and coordinating activities across space” (McCann, 2008,

p.355). In the context of foreign trade and investment, spatial transaction costs consist of three

distinct types: information transmission costs (hereafter “information costs”), transport costs, and

tariff costs. Unlike the first two, tariff costs are institutional costs that are not geographic in their

construction. Thus, following the tradition in the field, we focus on information costs and

transport costs in this study (McCann & Shefer, 2004). Information costs are associated with

moving knowledge and information across geographic space between transacting parties,

whereas transport costs are associated with moving goods across geographic space. The

definition of these two types of spatial transactions costs are explicitly geographical and the costs

incurred depend on the distance covered (McCann, 2008). The concept of spatial transaction

costs can be readily applied to the analysis of MNE subsidiaries as they are geographically

dispersed units engaged in coordinated activities across space.

The role of information costs is crucial to internalization analysis when examining the

costs of control, monitoring and exploiting proprietary knowledge. Information asymmetry

resulting from geographic distance discourages locating internally coordinated activities in

distant locations (Ragozzino, 2009; Slangen, 2011). Research on equity investment performance

has also established that analysts and investors located proximate to a firm have an information

advantage over those in distant locations, enabling investors to earn superior returns from

geographically proximate investments (Coval & Moskowitz, 1999; Malloy, 2005).

Another aspect of information costs is concerned with knowledge transfer within the MNE.

MNEs synthesize, integrate and disseminate locally originated knowledge, whether it is

developed at the headquarters or subsidiaries, and seek to apply it more broadly across countries

7

(Ahlstrom et al., 2014; Fang et al., 2010; Kogut & Zander, 1993; Qian & Delios, 2008). Prior

studies have shown that spatial proximity between subsidiaries facilitates knowledge transfer

within the firm. Hansen and Lovas (2004) find that large geographic distances between units

reduce the tendency for staff to seek information from other units. Ambos and Ambos (2009)

show that geographic distance limits knowledge transfer between subsidiaries and that the

negative effect of distance is particularly salient when personal coordination mechanisms are

used to facilitate knowledge transfer. These findings imply that geographic proximity between

subsidiaries is advantageous to the firm; however, there is little evidence as to how inter-

subsidiary proximity may influence MNE expansion strategy.

Meanwhile, the MNE can be viewed as a differentiated network within which subsidiaries

possess different types and levels of knowledge (Nohria & Ghoshal, 1994). The development of

a subsidiary’s capabilities can alter its political relations with the headquarters and other

subsidiaries (Mudambi & Navarra, 2004; Mudambi, Pedersen, & Andersson, 2014; Mudambi,

Piscitello, & Rabbiosi, 2014). For instance, knowledge transfer is found to be more effective

between subsidiaries whose main activities are complementary rather than substitutive because a

substitutive relationship may reduce subsidiaries’ willingness to share the knowledge (Andersson

et al., 2015; Gupta & Govindarajan, 2000). In the current study, the initial choice of a new

subsidiary’s location and its activities may be politically influenced by existing subsidiaries;1

however, it is conceivable that market entry decisions largely reflect the strategic direction

designed by the headquarters (Buckley, Devinney, & Louviere, 2007). Our conceptual model

thus assumes that the relationship between the new subsidiary and existing ones is non-

competitive. Although this partial equilibrium approach restricts the conceptual scope of our

1 We thank an anonymous referee for this insight.

8

analysis, the resulting theoretical parsimony enables us to more effectively demonstrate the

effects of inter-subsidiary spatial proximity.

In the case of transport costs, while the transport costs of bulk commodities have fallen in

recent decades, the transportation and logistics components of manufactured goods have

increased (McCann & Shefer, 2004). Geographic distance also has implications for the

opportunity costs associated with delivery time. As the quantity and complexity of market

information increases, production and logistics operations at MNEs have become increasingly

complicated and fragmented as evidenced by the growth of intra-firm trade and reliance on just-

in-time delivery strategies (Schonberger, 1996). As a result, the average lead times of shipments

have fallen over the years whereas shipment frequency has increased. This increased frequency

of transactions, combined with storage, inventory-handling and other forms of logistics costs, has

been found to increase the economic costs of geographic distance (McCann, 1998, 2011). These

findings suggest that the spatial structure of the MNE’s subsidiaries is likely to influence location

choices towards creating the least-cost production system.

Both international business and economic geography research have identified a distinctive

regional configuration to MNEs’ international strategy (Delios & Beamish, 2005; Dicken, 2007;

Rugman & Verbeke, 2004). Supranational regions represent a spatial platform for MNEs to

coordinate the scale and scope of their operations. Although prior research has shown that MNEs

tend to concentrate investment activities, especially production investment, by regional

boundaries (Arregle, Beamish, & Hebert, 2009), few studies have examined the inter-subsidiary

relations within a regional framework. One exception is a recent study that examined how inter-

subsidiary learning improves subsidiary performance (Kim, Lu, & Rhee, 2012). The authors

found that Japanese foreign subsidiaries utilized formal, routinized channels and interpersonal

9

communications to create opportunities for managers and personnel to interact with each other

within the same region. These findings along with other studies on MNE regional configuration

clearly indicate that MNEs’ international strategies substantially coalesce around discernible

geographic regions; therefore, the influence of intra-firm spatiality on location choices of

overseas production is likely to be regionally dependent too.

On the other hand, subsidiaries are established by the parent firm with certain objectives,

which determine the function and strategic mandate of the subsidiary. However, the traditional

internalization framework assumes that foreign subsidiaries equally benefit from market

internalization and few studies have explicitly specified underlying subsidiary mandates, bar a

few exceptions (Goerzen, Asmussen, & Nielsen, 2013; Makino, Lau, & Yeh, 2002). Several

studies have examined different functional activities, such as production, sales and services,

which reflect, though not directly indicate, investors’ objectives (Alcacer, 2006; Castellani,

Jimenez, & Zanfei, 2013; Enright, 2009).

Furthermore, investment in the same functional activity, especially production, can be

motivated by different strategic objectives. Dunning and Lundan (2008) suggest that foreign

production can be classified as natural resource-seeking, market-seeking, efficiency-seeking and

strategic-assets-seeking. Firms exhibit differential responses to the same location attributes, such

as corruption and governance hazards, when they have different mandates (Brouthers, Gao, &

McNicol, 2008; Hakkala, Norback, & Svaleryd, 2008; Slangen & Beugelsdijk, 2010). Though

existing evidence is limited and sometimes inconsistent, it illustrates the importance of

incorporating variation in subsidiary mandates into the analysis of MNE expansion strategy.

Our review suggests that an examination of subsidiary spatial structure in supranational

regions through the lens of spatial transactions costs can yield new explanations for MNE

10

location strategy. Prior studies also indicate that distinguishing subsidiary functions and strategic

mandates can help discern the differential influences of spatial transaction costs. We thus adopt a

theoretical approach that explicitly considers the impact of subsidiary heterogeneity along

functional and strategic dimensions. Our analysis follows the tradition in the internalization and

economic geography literatures of focusing on foreign investment in production subsidiaries

(Buckley & Casson, 1976; McCann, 1998; Schonberger, 1996). However, overseas production

facilities are supported by nonproduction units such as geographically dispersed trading

operations. Production and trading are the most prevalent functions in MNE networks, but they

differ significantly in activities, tangible and intangible assets, and informational needs (Alcacer,

2006; Makino, Beamish, & Zhao, 2004). Therefore, spatial transaction costs associated with

existing production and trading subsidiaries are likely to differ when considering new production

entries, making subsidiary function a potential contingency in the analysis of subsidiary spatial

structure. Similarly, subsidiary strategic mandates also have implications for assessing intrafirm

spatial relations. Spatial transaction costs may become less of a constraint when strategic

mandates, such as accessing local natural resources and R&D, are more directly influenced by

territorial geography than corporate geography. Building on this theoretical approach we develop

a set of hypotheses in the next section that collectively examine the main effect of subsidiary

spatial structure on investment strategy, and its nuanced effects under different subsidiary

functions and mandates.

3. Hypotheses

For several reasons we argue that the proximity of a potential host country to a firm’s

existing production subsidiaries in a region, controlling for home-host country distance,

11

increases the probability of production entry into that country. First, inter-subsidiary proximity

can enhance operational efficiency and flexibility of MNE production systems. One obvious

advantage is that transport costs between subsidiaries, and associated logistics costs, will be

lower when they are located in countries closer to each other. In addition, subsidiaries within

MNE regional networks often rely on timely transmission of goods and information to link

activities along value chains. Since geographic distance between subsidiaries increases

transportation and information transmission costs, more spatially proximate subsidiaries within a

region are better positioned to coordinate operations with each other.

Second, search for investment opportunities is likely to be spatially bounded since greater

distance leads to a higher degree of information asymmetry. Foreign subsidiaries often serve

regional country markets proximate to their locations and rely on regional suppliers from these

countries. Therefore, an MNE is likely to have better knowledge of the institutional and market

conditions of the countries closer to its existing establishments in a region than more distant

ones. This localized knowledge helps lower the liability of regional foreignness and facilitates

information gathering and processing, increasing the probability of identifying investment

opportunities in the proximity of its current operations (Johanson & Vahlne, 2009; Rugman &

Verbeke, 2007).

Third, an MNE’s competitive advantage depends in part on its ability to integrate complex

knowledge, particularly tacit knowledge, residing in geographically dispersed units (Kogut &

Zander, 1993). The transmission of tacit knowledge, which is context-dependent and embodied

in personal interactions, requires shared understanding of knowledge context and mutual trust

that builds on face-to-face contact and informal coordination mechanisms (Ghoshal, Korine, &

Szulanski, 1994). However, greater geographic distance increases communication barriers such

12

as travel time across time zones, discouraging staff in one location from interacting with those in

more distant subsidiaries (Ambos & Ambos, 2009). Moreover, search for knowledge can be

constrained by managers’ cognitive biases related to existing organizational routines (Levinthal,

1997). Managers located in one country are more likely to identify with regional peers in more

proximate markets because they share common beliefs and routines that arise from greater

similarity in business environments (Kim et al., 2012). Familiarity and similarity not only enable

subsidiaries to achieve greater efficiency in knowledge transfer but they may also create bias

against knowledge from distant locations (Hansen & Lovas, 2004).

In summary, geographic proximity between regional production subsidiaries will lower

spatial transaction costs within the firm. In additional to increased efficiency in coordinating and

monitoring cross-border production, lower spatial transaction costs help the firm more

effectively deploy and protect competitive advantages derived from intangible assets and

capabilities. From an internalization perspective, spatial transaction costs can be seen as partly

determining the extent to which markets can be internalized and, in the current context, a firm is

more likely to locate a new production subsidiary in the proximity of existing production

facilities in the same supranational region. Hence, we posit:

Hypothesis 1: The greater the proximity of a country to a firm's existing production

subsidiaries in a region, the greater the probability of the firm locating a new production

subsidiary in that country.

We argue that the effect of proximity to regional subsidiaries is likely to be conditioned by

the functional activities of existing units. Trading subsidiaries create value by linking production

with demand from foreign customers. They can simultaneously engage in a variety of activities,

including market research, intelligence collection, advertising, selecting foreign distributors,

13

after-sale services, logistics, and financing. The main mandate, however, is to stimulate and

generate demand for the MNE’s products and to provide the necessary services to process and

fill customer orders (Balabanis, 2000; Goerzen & Makino, 2007). Since information collection,

due diligence and sales require frequent and close contact with customers, local businesses and

government agencies, physical presence in foreign markets is often critical for trading

subsidiaries. At the same time, the type of services offered by trading subsidiaries and

knowledge created through these activities are closely linked to the local environments they face.

Thus, unlike production subsidiaries which significantly depend on internalized transfer of

proprietary technology, trading operations draw heavily on localized market knowledge.

As is the case between production subsidiaries, geographic distance between trading and

production subsidiaries increases both information and transport costs, making spatial proximity

a favorable condition for future production investment. However, we argue that proximity of

new production facilities to trading operations is not as consequential as that to production

operations. First, the relevance of knowledge originated in the same functional activities is

higher than knowledge created through activities of different functions. Same-function

subsidiaries, thus, are likely to serve as a more important source of knowledge when the firm

invests in a region. Trading units rely on location-specific market knowledge in order to improve

effectiveness in linking supply and demand, whereas production units primarily depend on

proprietary technology and tacit operational knowledge to gain efficiency. Since knowledge

overlap is limited, the quantity and complexity of information exchange between trading and

production subsidiaries will be less significant compared to those between production units.

There is some evidence that regional meetings in MNEs are organized by functional areas to

promote knowledge sharing between subsidiaries (Kim et al., 2012). Thus, we expect that the

14

benefits of spatial proximity attributable to lower information costs will be smaller between

trading and production subsidiaries.

Second, a low level of inter-functional interdependence arises in part due to the differences

between corporate geography and territorial geography. The importance of local physical

presence dictates that the location of trading activities largely conforms to the spatial distribution

of markets (Alcacer, 2006). In contrast, the location of production subsidiaries is to a greater

extent determined by spatial transaction costs among peer production units and the distribution

of factor endowments. Therefore, spatial transaction costs associated with trading units are more

of a function of territorial geography (i.e. the spatial distribution of markets) than of corporate

geography (i.e. the spatial distribution of foreign operations). There are limits to minimizing

spatial transaction costs by locating production in the proximity of trading operations because it

is uneconomical to spread production capacity in the same way that trading operations are

structured. Thus, according to the internalization framework, the impact of spatial proximity

between production and trading subsidiaries on subsequent production investment will be limited

because, in this case, territorial geography is a more influential factor than corporate geography

in determining the extent of market internalization. Thus, we posit:

Hypothesis 2: The effect of a country’s proximity to a firm’s existing production

subsidiaries in a region on subsequent production entry into that country is stronger than that

for proximity to existing trading subsidiaries.

The impact of spatial transaction costs on investment decisions is likely to be contingent on

a subsidiary’s mandate. We argue that location choices of natural resource-seeking subsidiaries

will be less responsive to inter-subsidiary proximity than those of subsidiaries mandated to seek

market, efficiency or strategic assets. Since natural resources are relatively immobile and

15

expensive to transport, the location of a production unit with a natural resource-seeking mandate

is largely determined by the geographic distribution of resource endowments. In other words, the

influence of territorial geography strengthens in relation to corporate geography when seeking

natural resources is part of a production subsidiary’s mandate.

There is also evidence that firms develop location specific capabilities to cope with

business and political risks when seeking natural resources in foreign countries (Cuervo-Cazurra

& Genc, 2008; Holburn & Zelner, 2010; Schotter & Beamish, 2013). This location-specific

knowledge, though valuable, is likely to be less applicable to other units that do not actively seek

local natural sources. In contrast, subsidiaries of market-seeking and efficiency-seeking

mandates occupy the final and intermediate stages within a firm’s international production

network. They rely heavily on the exploitation of common proprietary knowledge and integrated

production processes to enhance their operational efficiency. Similarly, asset-seeking

investments, often through acquisition, are undertaken to strengthen a firm’s ownership-specific

advantage and to complement a firm’s existing foreign operations (Dunning & Lundan, 2008).

Therefore, internal communication and logistic coordination are particularly important for these

investments; consequently, the reduction of spatial transaction costs is likely to be more

beneficial for these production subsidiaries than for those with a natural resource-seeking

mandate. Thus, we posit:

Hypothesis 3: The effect of a country’s proximity to a firm’s existing production

subsidiaries in a region on subsequent production entry into that country will be weaker when

the new subsidiary has a natural resource-seeking mandate.

Analogous to Hypothesis 3, we expect that an R&D mandate will diminish the impact of

inter-subsidiary proximity on the location choice of production subsidiaries. Foreign subsidiaries

16

increasingly assume a significant role not just as a recipient and enabler of knowledge transfer

within the firm but also as a source of innovation (Birkinshaw & Hood, 1998; Feinberg & Gupta,

2004). These subsidiaries tend to engage in exploitative and/or explorative types of R&D. The

former focuses on extending the existing knowledge base through incremental innovation and

product development whereas, in the latter case, subsidiaries are directed toward more basic

research and the development of new technical knowledge (Frost, 2001).

The locations of exploitative R&D activities have been found to be closely related to local

industry strength (Frost, Birkinshaw, & Ensign, 2002). On the other hand, firms engaged in more

explorative research are often drawn to locations with large talent pools and strong knowledge

infrastructure, such as universities and technology centers (Demirbag & Glaister, 2010).

However, regardless of the orientation of R&D activities, external environmental conditions,

especially supply side forces, predominantly determine locational advantages of a potential host

country (Kuemmerle, 1999). The influences of these forces, which reflect territorial geography

rather than corporate geography, are likely to strengthen when R&D is part of a production

subsidiary’s mandate. Therefore, while lower spatial transaction costs resulting from geographic

proximity between production units may remain beneficial, their impact is likely to decline in the

case of an R&D mandate. Thus, we posit:

Hypothesis 4: The effect of a country’s proximity to a firm’s existing production

subsidiaries in a region on subsequent production entry into that country will be weaker when

the new subsidiary has an R&D mandate.

In contrast, while we have argued that proximity to trading subsidiaries is not as influential

for production entries as proximity to other regional production subsidiaries (Hypothesis 2), we

expect its impact to increase when a production subsidiary is mandated to conduct R&D.

17

Overseas subsidiaries create new products or skills in response to knowledge of local conditions,

such as the knowledge about consumer tastes or local regulations. Acquiring and transmitting

knowledge of local markets are considered a main source of information costs for overseas R&D

activities (Buckley & Carter, 2004). A fundamental role of trading subsidiaries is to collect and

disseminate information about local markets through a variety of activities, including market

research, due diligence and after-sale service. Location-specific market knowledge originating in

trading units, though of limited value in improving production efficiency, may prove particularly

useful for R&D activities. Therefore, by lowering spatial transaction costs, proximity to trading

subsidiaries will help address information challenges facing production subsidiaries with an

R&D mandate. Thus, we posit:

Hypothesis 5: The effect of a country’s proximity to a firm’s existing trading subsidiaries

in a region on subsequent production entry into that country will be stronger when the new

subsidiary has an R&D mandate.

4. Method

4.1 Sample

We test our hypotheses using data on the country location choices for foreign production

subsidiaries established by Japanese public manufacturing firms between 1971 and 2006. Japan

has been one of the world’s largest sources of outward FDI for three decades. The manufacturing

sector is particularly active in foreign expansion, and manufacturing firms have entered a wide

range of countries, providing significant variation in location choices during the sample period.

Subsidiary data was compiled from Kaigai Shinshutsu Kigyou Souran (Japanese Overseas

Investment), a directory of operating foreign subsidiaries of Japanese firms. The Japanese

18

Overseas Investment directory provides information on each subsidiary, including its host

country, industry classification and founding year. Data on parent firms was extracted from the

Nikkei Economic Electronic Databank System (NEEDS). NEEDS reports financial information

for firms listed on the Tokyo Stock Exchange and is available from 1971. The study period was

determined mainly by data availability.

Our sample consists of large, diversified companies whose investments span various

industries. The proximity variables and subsidiary counts with regard to production entries are

based on the grouping of subsidiaries sharing the same 2-digit Standard Industrial Classification

(SIC) code (all trading subsidiaries share the same industry code). Production and trading

account for nearly 90% of the subsidiaries reported by sample firms. Other subsidiaries are

engaged in finance, real estate, agriculture, and miscellaneous services, and are excluded from

the operationalization of variables of theoretical interest.

We use the regional classifications of the World Bank and the United Nations to identify

seven regions: Africa, Asia, Europe, Latin America & the Caribbean, Middle East & North

Africa (MENA), North America and Oceania (Arregle et al., 2009) (see Appendix 1). Since our

analysis is concerned with the impact of the spatial structure of existing subsidiaries, we exclude

firms’ first production entries into a region. The resulting pooled dataset for estimating

production entry contains 6,875 entries in 518,342 firm-year-country observations.

4.2 Measures

4.2.1 Dependent Variable

Our unit of analysis is a firm’s entry decision for a potential host country in a given year.

Accordingly, the dependent variable – Entry – is a binary variable. For each firm-year-country

19

observation, “1” indicates that the parent firm established a production subsidiary; “0” indicates

otherwise.

4.2.2 Independent Variables

We construct two proximity variables – Proximity to Regional Production Subsidiaries and

Proximity to Regional Trading Subsidiaries. These are defined as the reciprocal of the average

distance between a potential host country and a firm’s existing subsidiaries of a given function in

a region:

Proximity i=∑

jN j

∑j

DISTANCE ij∗N j

(1)

where ‘DISTANCEij’ is the distance in thousands of kilometers from the largest city (by

population) of the host country i to the largest city of a same-region country j where the firm’s

existing production (or trading) subsidiaries are located, and ‘Nj’ is the number of production (or

trading) subsidiaries in country j. The proximity variables are set to zero when there are no

regional subsidiaries. This measure reflects the theoretical underpinning that spatial transaction

costs directly depend on geographic distance covered (McCann, 2008). It is also consistent with

operationalization approaches adopted in the location choice literature (Chacar & Lieberman,

2003; Chung, 2001; Mitra & Golder, 2002; Nachum & Wymbs, 2005), and is analogous to a

widely used centrality measure in the network literature (Freeman, 1979).

4.2.3 Control Variables

We control for country level factors that existing research finds can influence location

decisions. GDP and GDP Growth Rate, adjusted for purchasing power parity, are proxies for

market size and potential size, two factors that are commonly attractive to investors. GDP

20

Growth Rate is measured as the average annual growth in real GDP in the preceding three-year

period. We include GDP per capita as a measure of a country’s prosperity and consumer

purchasing power. Trade Ratio is the ratio of a country’s total exports and imports to its GDP.

While there is some evidence that FDI complements trade, research on MNE investment

suggests that there may also be a substitution effect between foreign production and export

activities (Blonigen, 2001). We include both existing FDI Stock and FDI as a percentage of

GDP, measured as the annual dollar value of net foreign direct investment in a country

(excluding that from Japan) as a percentage of GDP, to capture the country’s attractiveness to

foreign investors. FDI data was obtained from UNCTAD except for that on Taiwan, which was

obtained from Taiwan’s Ministry of Economic Affairs. Country Resource Exports, measured as a

host country's exports of primary goods (including fuels, minerals, agricultural and food

resources) as a percentage of GDP, is included to account for the country’s natural resource

endowments (Isham et al., 2005). We also include Country Patent Grants, measured as the

logged count of yearly patent grants received by a country, as a proxy for the country’s

innovative capacity (Furman, Porter, & Stern, 2002).

We include three variables to control for a host country’s institutional environment.

Political Stability, constructed using the Henisz’s (2002) POLCONV data, measures constraints

embedded in a country’s political structures, which support credible policy commitments. Rule

of Law is drawn from the Rule of Law index in the Governance Indicators data (Kaufmann,

Kraay, & Mastruzzi, 2008). For missing data years we interpolate using the closest year for

which data is available. We use Hofstede’s (2001) cultural dimensions to construct Cultural

Distance between Japan and a potential host country (Kogut & Singh, 1988), though we exclude

the fifth dimension – long-term orientation – due to the lack of data available for the majority of

21

countries in our sample. We also include Bilateral Geographic Distance between home and host

countries measured as the distance in thousand kilometers between Tokyo and the largest city of

a host country. Finally, to control for the potential influence of spatial relations with non-regional

subsidiaries, we include Proximity to Non-regional Production Subsidiaries using the same

method as described in Equation (1).

We include Country Production Operation and Country Trading Operation measured,

respectively, as the logged count of a firm’s production and trading subsidiaries in a host country

to account for MNEs’ propensity to repeatedly invest in a country and for the differences in the

density of investment activities between firms over time (Guillén, 2003; Kogut & Chang, 1996).

A firm’s foreign investment strategy can also be influenced by the strategic choices of peer firms

(Henisz & Delios, 2001; Jiang, Holburn, & Beamish, 2014). We thus include Industry Country

Production Operation measured as the logged count of production subsidiaries with the same

industry codes established by other Japanese firms in a host country. To distinguish the spatial

element from the firm’s overall propensity to repeatedly invest in a region, we include Region

Production Operation measured as the logged count of production subsidiaries in the same

region excluding the focal host country.

Parent firm size can affect the level of foreign investment so we include Firm Size

measured as the logarithm of parent company sales adjusted by the industry mean. Research has

shown that firms of higher capital intensity tend to expand abroad. Capital Intensity is measured

as the logged ratio of total fixed assets to total sales adjusted by the industry mean (Caves, 2007).

4.3 Estimation Approach and Statistical Interpretation

22

Conditional logit is a common estimation approach in location choice research. Although a

conditional logit model could be applied to our data it has several limitations. First, the

‘independence of irrelevant alternatives’ (IIA) assumption, which means that the probability

ratios do not depend on other alternatives in the choice set, is not satisfied in many empirical

settings (Martin, Swaminathan, & Tihanyi, 2007). Second, the model does not accommodate

chooser-specific attributes that are invariant across choices as independent variables – such as

year of entry and parent firm size.

A logit model is an alternative estimation technique for this study that is not limited by the

constraints of a conditional logit model. However, the small proportion of entry events relative to

non-entry events indicates a rare events data structure. King and Zeng (2001) showed that a

logistic regression can produce biased estimation of event probability when applied to rare

events data, and developed a choice-based sampling procedure to correct for potential estimation

bias. Following recent application of this technique in management research (Cockburn &

MacGarvie, 2011; Folta & O'Brien, 2004; Hallen, 2008; Singh, 2005), we derive a sample

consisting of all entry events and randomly selected non-entry observations. For each entry

event, ten non-entry firm-year-country observations from the same year are included. We then

implement a rare event logit model for the constructed sample. We also include parent firm

industry, year, and region fixed effects using dummy variables to account for unobserved

industry, temporal, and regional heterogeneities.

5. Results

Table 1 presents summary statistics and correlations for the subsample of production

entries constructed using the choice-based sampling procedure. All the individual variables’

23

variance inflation factor values are less than three, indicating that multicollinearity is unlikely to

be a concern. Table 2 reports the rare event logit estimation results: Model 1 is the baseline

specification without proximity variables or interaction terms, which are included in Model 2 and

other alternative specifications. The coefficient estimates are highly consistent between Models 1

and 2, and we use Model 2 as the preferred specification to test Hypotheses 1 and 2.

***************************Insert Tables 1-2 about here***************************

The control variables largely behave as expected. Estimated results for GDP and GDP

Growth Rate indicate that Japanese firms are more likely to enter countries with greater market

size and economic growth rate. A country that has a lower GDP per capita is also found to be

more attractive to MNEs, perhaps to access lower-cost labor. High levels of Country Resource

Exports and Country Patent Grants appear to be attractive locational attributes to Japanese

MNEs. Consistent with expectations, FDI as percentage of GDP is positively related to market

entry. FDI Stock, however, is negatively related to entries; this may reflect the fact that a large

number of production investments occurred in developing countries which tend to have lower

levels of accumulated FDI, especially in the 1980s and 1990s. A higher value of Rule of Law is

positively related to production entries and, as expected, Cultural Distance deters entry.

Bilateral Geographic Distance and Proximity to Non-regional Production Subsidiaries are

negatively related to production entry. This latter result may imply that when a potential host

country is distant from a firm’s subsidiaries in other regions, it is costly to service this market

using those subsidiaries, thereby making direct investment in this country more attractive.

Investment decisions exhibit a strong country-specific focus as indicated by the positive

effects of Country Production Operation and Industry Country Production Operation. Country

24

Trading Operation does not have a significant impact on production entries, implying relative

low interdependence between the two functions as far as country-level aggregation is concerned.

Region Production Operation has a statistically weak yet positive effect on production entry.

Larger Firm Size is negatively associated with foreign investment. One possible explanation for

this result is that dominant firms in the domestic market tend to be less internationalized than

smaller firms (Mascarenhas, 1986). We found no significant influence of Capital Intensity.

Hypothesis 1 predicts that the spatial proximity between a potential host country and a

firm’s existing production subsidiaries in the same region enhances the probability of the firm

locating a production subsidiary in that country. Coefficient estimates in a non-linear model

cannot be directly interpreted as marginal effects as in a linear model (Hoetker, 2007). Hence, in

order to enable accurate interpretation, we adopt a simulation-based procedure to assess the

impact of independent variables on the probability of production entry (King, Tomz, &

Wittenberg, 2000). This procedure uses Monte Carlo simulation to convert the raw output of a

rare event logit model into changes in predicated probabilities. Following common practice, we

simulated the parameters 1,000 times in this study. The simulation results are most clearly

understood through graphic presentation (Zelner, 2009).

Based on the estimation results from Model 2, the upper curve in Figure 1 shows the

simulated effects of Proximity to Regional Production Subsidiaries on the probability of

production entry (measured on the y-axis). Consistent with our prediction, the upward curve

clearly indicates that inter-subsidiary proximity is positively related to production entry. The

simulation enables intuitive assessments of the magnitude of the hypothesized effect: all else

equal, the probability of production entry increases by 30% when Proximity to Regional

Production Subsidiaries increases by one standard deviation from its sample mean.2 The

2 Other variables are set at their sample mean in simulations.

25

simulated confidence interval of this positive change is statistically significant at the 5% level.

These results provide strong support for Hypothesis 1.

***************************Insert Figure 1 about here***************************

Hypothesis 2 proposes that the positive proximity effect derived from the spatial relations

with production subsidiaries is greater than that with trading subsidiaries. The lower curve in

Figure 1 depicts the simulated effects of Proximity to Regional Trading Subsidiaries on the

probability of production entry. The significant difference in the slopes of the upper and lower

curves implies that the magnitude of the impact of Proximity to Regional Production

Subsidiaries is greater than that of Proximity to Regional Trading Subsidiaries. For instance,

entry probability only increases by 1% when Proximity to Regional Trading Subsidiaries

increases by one standard deviation from its sample mean and this increase is not statistically

significant. A Wald test also confirms that the effect of Proximity to Regional Production

Subsidiaries is greater than that of Proximity to Regional Trading Subsidiaries (chi2=62.1(1),

p<0.001). Therefore, Hypothesis 2 is strongly supported.

Hypotheses 3-5 focus on whether the effects of proximities to regional production and

trading subsidiaries differ depending on the strategic mandates of new production subsidiaries.

The Japan Overseas Investment publication provides self-reported investment purposes of

foreign subsidies, including two that are pertinent to our theoretical model – “access natural

resources or materials” and “development of products and planning”. 3 However, only

approximately half of the sampled subsidiaries reported their strategic mandates. Model 3 reports

3 The Japan Overseas Investment classified FDI into fifteen commonly observed purposes. Other investment purposes are “access local market,” “alliance with customers,” “access labor force,” “reverse imports to Japan,” “development of overseas production network,” “preferential treatment of local government” “development of overseas distribution network,” “export to third countries,” “exchange rate risk measures,” “loyalty and information collection,” “expansion into new business,” “enhanced regional headquarters,” and “hedge to trade fiction.”

26

the estimation results when we only include entries whose mandates are reported, and results are

consistent with those from Model 2. We then use a subgroup analysis approach and estimate

entries of different mandates in separate regression models (Hoetker, 2006; Schotter & Beamish,

2013). We divide our sample based on whether “access natural resources or materials” or

“development of products and planning” is listed as an investment mandate.

Hypothesis 3 predicts that the positive effect of proximity to other regional production

subsidiaries on subsequent production entry will diminish when the new subsidiary has a natural

resource-seeking mandate. The results of the subgroup analysis (Models 4 and 5) show a

significant difference in coefficient estimates for Proximity to Regional Production Subsidiaries.

The coefficient is statistically insignificant in Model 4 (natural resource-seeking group) but is

statistically significant (p<0.01) in Model 5 (non-natural resource-seeking group), providing

strong support for Hypothesis 3.

Hypothesis 4 similarly predicts that an R&D mandate will lead to a smaller effect of

proximity to other regional production subsidiaries on subsequent production entry. The

coefficient estimates for Proximity to Regional Production Subsidiaries are both positive in

Models 6 and 7, but only marginally significant (p<0.1) in Model 6 (R&D group), providing

some support for Hypothesis 4. Because we use divided groups, the sample means are different;

therefore, the marginal effect is not directly comparable across the models. Further simulation

shows that the probability of production entry increases by 25% (R&D group) and 29% (non-

R&D group), respectively, when Proximity to Regional Production Subsidiaries increases by one

standard deviation from its sample mean. The lack of stronger support for the hypothesis implies

that tight internal linkages may remain critical even when firms attempt to tap into local

27

intellectual resources (Alcacer & Zhao, 2012), and that the R&D mandate, though important,

may not be the primary strategic objective of production subsidiaries.

Hypothesis 5 further proposes that the effect of proximity to regional trading subsidiaries

on subsequent production entry will be stronger when a new entry has an R&D mandate. The

results of the subgroup analysis (Models 6 and 7) confirm that the coefficient estimates for

Proximity to Regional Trading Subsidiaries are different. The coefficient is statistically

significant (p<0.01) and signed as expected in Model 6 (R&D group), and is statistically

insignificant in Model 7 (non-R&D group). Thus, Hypothesis 5 is strongly supported.

5.1 Subsidiary Performance

An underlying assumption of our conceptual model is that lower spatial transaction costs

facilitate internalization, which in turn improves organizational performance. Thus, to further

analyze the validity of our model, we conduct additional analysis to examine how the spatial

structure of foreign subsidiaries affects their performance.

Due to the absence of detailed financial performance data at the subsidiary level, we utilize

instead a categorical measure provided by the Japanese Overseas Investment data source. Each

year, subsidiary general managers are asked to assess and report their subsidiaries’ performance

in three ascending categories: loss, break-even, and gain. Thus, our dependent variable –

Subsidiary Performance – is an ordinal variable. On average, subsidiaries in our location choice

model reported performance data 5.2 times between 1986 and 2006.4 Prior research has shown

that subjective performance data is highly correlated with objective performance data, alleviating

potential concerns about reporting bias (Isobe, Makino, & Montgomery, 2000).

4 Performance data is only available between 1986 and 2006.

28

We modify other measures in order to properly model subsidiary performance. First, we

extend all time-varying variables, including the main proximity measures, from a subsidiary’s

founding year to all operating years. Second, we include several subsidiary-specific variables.

Subsidiary Size is measured as the logarithm of the number of employees. Subsidiary Age is

measured as the logarithm of the difference between the founding year and the year performance

was reported. Subsidiary Ownership is a categorical variable that classifies a parent firm’s

ownership position in a subsidiary as whole ownership, majority or equal ownership, or minority

ownership. In addition, we include Expatriate Ratio, measured as the ratio of the number of

Japanese Employees to total subsidiary employment, to account for managerial control of the

parent firm at a subsidiary (Peng & Beamish, 2014). Finally, we include Parent Firm

Performance, measured as the preceding three-year average of return on assets to control for the

impact of parent firm performance on subsidiary performance.

Given the ordinal nature of our dependent variable, we employed an ordered logit model to

examine subsidiary performance. Since multiple performance observations from the same

subsidiary are likely to be correlated, we use a clustered variance estimator to control for intra-

group correlation. Our specification also includes region, subsidiary industry, and year fixed

effects. Model 8 in Table 2 presents the estimation results of the ordered logit model. The

coefficient estimate for Proximity to Regional Production Subsidiaries is statistically significant

(p<0.05). When Proximity to Regional Production Subsidiaries increases by one standard

deviation from its sample mean, the simulated probability of a subsidiary reporting a ‘gain’

increases from 0.64 to 0.66 and the simulated probability of reporting ‘loss’ decreases from 0.16

to 0.15. These changes are statistically different from zero at the 5% level. We find no significant

relationship between Proximity to Regional Trading Subsidiaries and subsidiary performance.

29

Overall, the results of our performance analysis are consistent with the central findings from our

location choice analysis. This provides additional support for our overall theoretical prediction

that inter-subsidiary proximity is conducive to foreign production investment.

5.2 Robustness of Findings

We employ alternative estimation approaches to examine the sensitivity of our results to

the choice of specification. Two common methods in prior studies are conditional logit and event

history analysis using discrete time logit (Henisz & Delios, 2001; Kalnins, 2004). Since the

conditional logit model does not accommodate firm-specific variables, Firm Size, Capital

Intensity and year and industry fixed effects are omitted. Due to the longitudinal data structure of

the discrete time logit analysis, measures associated with production subsidiaries are no longer

based on subsidiaries in the same industry. Instead, they are now based on the grouping by the

same parent firm. Industry Country Production Operation is omitted from the model for the

same reason. We also remove all subsidiaries of less than 20 employees to ensure that our

analysis includes subsidiaries with substantial strategic significance (Beamish & Inkpen, 1998).

The results also remain consistent when we include Subsidiary Size in the production entry

models. Furthermore, we substitute the cultural distance measure with the psychic distance data

(covering 25 host countries) provided by Hakanson and Ambos (2010). The coefficient estimates

on terms of theoretical interest in these alternative specifications are consistent with the main

results reported in Model 2, indicating reliability of our findings. Finally, additional analysis

shows that overseas investment rates increased following the signing of the Plaza Accord, but

fell during the economic slump of the 1990s.5

6. Discussion

5 Exhibits are omitted due to space constraints.

30

In this study we argue that country location choices for MNEs are influenced by the cross-

country spatial structure of existing regional subsidiaries. A statistical analysis of foreign

production investments by Japanese manufacturing firms provides substantial empirical support

for our theoretical predictions. We find that the proximity of a host country to a firm’s regional

production subsidiaries, but not to trading subsidiaries, enhances the probability of production

investment in that country. This positive effect, however, diminishes when the new production

subsidiary has a mandate to seek local natural resources or to conduct R&D. In contrast, the

effect of proximity to regional trading subsidiaries strengths when R&D is part of a new

subsidiary’s mandate.

6.1 Contributions

Our study contributes to MNE research in three ways. It first contributes to the growing

literature on the strategic importance of geography by demonstrating intra-firm spatiality as a

distinct source of spatial variation and a significant determinant of MNE foreign investment

decisions. Our findings complement prior location studies that focus on spatial variations

resulting from home-host country distance and subnational industry clustering by showing that

modeling the cross-country spatial linkages between foreign subsidiaries can yield new insights

into MNE location strategies. In addition, this study echoes prior research that shows the cultural

profile of a firm’s existing location portfolio can affect the rate of subsequent internationalization

and corporate performance (Hutzschenreuter & Voll, 2008; Hutzschenreuter, Voll, & Verbeke,

2011).

Second, our analysis advances prior research on the geographic distribution of foreign

investment, which has largely omitted consideration of interdependencies between subsidiary

31

functions and mandates (Alcacer, 2006; Flores & Aguilera, 2007). We explicitly examine

interactions between subsidiary activities and corporate spatial structure embedded in territorial

geography. In particular, the significant differences between the effect of proximity to production

subsidiaries and that to trading subsidiaries imply that operational differences and knowledge

requirements associated with different functions are likely to serve as a boundary condition as to

how much internalization advantage firms can gain through spatial proximity, and how corporate

geography shapes investment strategy. The results also imply more generally that subsidiary

function may be a significant determinant of inter-subsidiary relationships. For instance,

functional differences may directly determine the degree of knowledge complementariness

between a source subsidiary and a receiving subsidiary, which in turn affects knowledge transfer

between them (Andersson et al., 2015).

On the other hand, our unique measures of subsidiary strategic mandate allow us to tease

out how the geographic configuration of MNE subsidiaries and investment motivations interact

to influence location strategies. Results on the differential influences of inter-subsidiary

proximity in the cases of natural resource-seeking and R&D mandates offer novel evidence on

the key role of managerial intention in determining the geographic pattern of foreign expansion

(Hutzschenreuter, Pedersen, & Volberda, 2007). Overall, our findings with respect to subsidiary

level attributes, such as subsidiary function and strategic mandates, provide a more sophisticated

understanding of MNE location strategy.

Finally, our study enriches internalization theory through a focused analysis of inter-

subsidiary spatial relations. Our findings suggest that economizing on spatial transaction costs,

which contributes to overall internalization costs, may enhance the firm’s competitive advantage.

A positive association between a subsidiary’s performance and its geographical proximity to

32

other regional subsidiaries, as shown in our additional analysis, further implies that lower

communication and logistic costs resulting from geographic proximity can improve operational

coordination and the efficiency of the internal market for firm knowledge. Researchers have

repeatedly noted the lack of research on spatial elements in internalization analysis (Buckley,

2009; McCann & Mudambi, 2004). As the MNE’s operation increasingly relies on integrated

subsidiary systems across countries, there is a clear need to extend spatial analysis beyond home-

host country distance and subnational agglomeration to include cross-country connectivity

between subsidiaries. Our analysis provides specifications and insights that can facilitate more

nuanced application of the internalization framework in examining the growth and performance

of the MNE.

Building on the notion of spatial transaction costs, our analysis also establishes a stronger

connection between regional strategy research and internalization theory. We develop new

insights for the literature by showing that concentration and connectivity are distinct aspects of

firms’ regional strategy. Besides the stock of existing regional operations (Arregle et al., 2009;

Rugman & Verbeke, 2004), the spatial relations between the subsidiaries constitute structural

conditions for intermediate levels of international integration, and thus influence subsequent

investment decisions.

6.2 Managerial Implications

The central findings of this study support the notion that both ‘space’ and ‘place’ are

critical dimensions of potential investment locations. While it is imperative to evaluate the

economic and institutional conditions of a location, managers should account for how the spatial

structure and connectivity of a firm’s existing investment portfolio can influence the

33

performance of overseas operations. These two distinct dimensions imply potential trade-offs in

selecting investment locations. Firms are typically attracted to country markets that have

investor-friendly business environments and/or substantial market potential. However, a location

with attractive attributes that is remote from a firm’s existing regional operations can create

significant information and logistic costs. As MNEs increasingly rely on cross-border production

systems and locally-developed expertise to improve operating efficiency, managers are advised

not to base their location decisions solely on environmental factors, such as market size and

national culture. Instead, managers should be aware of the costs of spatial isolation and carefully

weigh the benefits of inter-subsidiary proximity in relation to other environmental factors.

Another key managerial implication is that firms can gain regional competitive advantage

using ‘geographic proximity’ as a tool to reduce cross-border coordination costs in a complex

global MNE system. Intrafirm spatial tightness may prove a structural advantage when MNEs

simultaneously link, reconcile, and integrate an increasing number of issues as their operations

become more diverse and more interdependent. Managers should also recognize that the benefits

of inter-subsidiary proximity vary by function and subsidiary mandate, and thus should monitor

and adjust over time the firm’s geographic configuration to enhance internal communication and

logistic efficiency. Especially, our findings imply that it is necessary to build geographically

disparate innovation processes around multiple trading or sales units that can offer rich

information on local markets. Tight spatial linkages between innovative activities and existing

production and trading operations not only can serve as a mechanism for information sharing,

but also help firms internalize proprietary knowledge and reduce the risk of knowledge outflows

(Alcacer & Zhao, 2012).

34

Although our findings show that geographic distance between subsidiaries can pose

significant costs, there are also ways to alleviate these obstacles, especially information costs.

The more that managers are aware of distance-related costs, the more they will undertake

appropriate measures to address the challenge. Managers are advised to proactively adopt

informal and formal communication mechanisms, particularly face-to-face networking, to

enhance interactions between regional subsidiaries. In this effort, subsidiaries located afar from

others should receive extra resources since they have the most to gain from mitigating the

negative effect of geographic distance.

6.3 Limitations and Future Research

Naturally our results should be interpreted with some caution due to a variety of

limitations, which also suggest directions for future research. First, since the conceptual focus of

our analysis is cross-border spatial linkages within a firm, we did not explore industry clustering

effects at the subnational level. However, we recognize that foreign subsidiaries are often located

in particular agglomerations within a country and are embedded in localized competitive and

institutional settings (Beugelsdijk & Mudambi, 2013). Just as intra-firm spatial relationships

have strategic consequences, connections with local suppliers, competitors, and other external

stakeholders can convey benefits or liabilities. A possible spatial implication is that a strong

agglomeration effect may influence a subsidiary’s dependence on sister subsidiaries in proximate

countries. Future research could extend the scope of our study and explore potential interactions

between local agglomeration, particularly at the state, provincial, or municipal levels, and the

transnational structure of MNE operations.

35

In addition, since granular, street-level location information for many subsidiaries is

incomplete in our data, we chose to use the distances between the largest cities of respective

countries as proxies for geographic distances between subsidiaries. Although this proxy

reasonably approximates the distance between subsidiaries and is commonly used in the

literature (Ambos & Schlegelmilch, 2004; Flores & Aguilera, 2007; Kang & Kim, 2010;

Ramasamy, Yeung, & Laforet, 2012), it nonetheless is suboptimal when a host country has a

large territorial area. Due to the same data constraint, our analysis does not include measures for

distance between subsidiaries in the same country, though Country Production Operation and

Country Trading Operation partially account for firm-specific spatial variation within a country.

To understand better how intra-firm spatial structure affects investment strategy, researchers

could assemble a more complete picture of the physical distribution of foreign operations both

across and within countries for a smaller number of firms.

Second, our conceptual model does not take into account potential competitive

relationships between subsidiaries. However, cooperation and competition can exist

simultaneously between MNE subsidiaries as they vie for internal resources, corporate attention

and political power (Andersson et al., 2015; Bouquet & Birkinshaw, 2008; Luo, 2005; Mudambi,

Pedersen, et al., 2014). Thus, geographic proximity between subsidiaries may enhance

competition at the same time as it improves coordination. Future research may build on our

partial equilibrium analysis to explore whether and how competitive and political dynamics

between subsidiaries moderate the influence of corporate spatial structure. Moreover, subsidiary

mandates may change over time and, as a result, alter the benefits of spatial proximity between

subsidiaries (Birkinshaw & Hood, 1998). An interesting extension of our study would be to

36

investigate how the evolution of subsidiary mandate and inter-subsidiary spatial relations

influence each other.6

Third, the generalizability of our results is also limited by the fact that our sample consists

of firms from Japan only. Japanese MNEs historically have adopted a more centralized approach

to the management of foreign operations than have western firms, and their production networks

tend to be more tightly integrated (Bartlett & Ghoshal, 1998). In addition, while the sample firms

established a large number of plants in North America and Europe during the studied period,

Asia accounted for the largest portion of production investments. Since firms may adjust their

investment strategy depending on where they primarily operate, our results may be partially

driven by the idiosyncrasies of the Asia Pacific region. It will be particularly fruitful to utilize a

sample of MNEs from diverse home countries to provide additional tests of, and to refine, our

theoretical model. Alternatively, researchers may delve deeper into the specific institutional and

cultural characteristics of a home country to examine how attributes such as tolerance of

uncertainty interact with corporate geography in determining firms’ investment decisions.7

7. Conclusion

This paper finds that when an MNE establishes new production facilities it tends to choose

country markets more proximate to its existing regional subsidiaries – and these subsidiaries

subsequently perform better than more distant ones. Spatial proximity to existing production

subsidiaries is less consequential for natural resource-seeking investment; however, proximity to

trading operations is preferable when R&D is part of a new subsidiary’s strategic mandate. In

conclusion, we show that the spatial configuration of existing subsidiaries is a significant

6 We thank an anonymous referee for this insight.7 We thank an anonymous referee for this insight.

37

geographic determinant of MNE expansion strategy, and that location strategies built on

corporate geography can be a source of competitive advantage.

REFERENCES

Ahlstrom, D. (2015). From the editors: Publishing in the Journal of World Business. Journal of World Business, 50, 251-255.

Ahlstrom, D., Levitas, E., Hitt, M., Dacin, M. T., & Zhu, H. (2014). The three faces of China: Strategic alliance partner selection in three ethnic Chinese economies. Journal of World Business, 49, 572-585.

Alcacer, J. (2006). Location choices across the value chain: How activity and capability influence collocation. Management Science, 52, 1457-1471.

Alcacer, J., & Zhao, M. (2012). Local R&D strategies and multilocation firms: The role of internal linkages. Management Science, 58, 734-753.

Ambos, B., & Schlegelmilch, B. (2004). The use of international R&D teams: An empirical investigation of selected contingency factors. Journal of World Business, 39, 37-48.

Ambos, T. C., & Ambos, B. (2009). The impact of distance on knowledge transfer effectiveness in multinational corporations. Journal of International Management, 15, 1-14.

Andersson, U., Gaur, A., Mudambi, R., & Persson, M. (2015). Unpacking interunit knowledge transfer in multinational enterprises. Global Strategy Journal, 5, 241-255.

Arregle, J. L., Beamish, P., & Hebert, L. (2009). The regional dimension of MNEs' foreign subsidiary localization. Journal of International Business Studies, 40, 86-107.

Balabanis, G. (2000). Factors affecting export intermediaries' service offerings: The British example. Journal of International Business Studies, 31, 83-99.

Bartlett, C., & Ghoshal, S. (1998). Managing across borders: The transnational solution (2nd ed.). Boston, MA: Harvard Business School Press.

Beamish, P. W., & Inkpen, A. C. (1998). Japanese firms and the decline of the Japanese expatriate. Journal of World Business, 33, 35-50.

Beugelsdijk, S., McCann, P., & Mudambi, R. (2010). Introduction: Place, space and organization - economic geography and the multinational enterprise. Journal of Economic Geography, 10, 485-493.

Beugelsdijk, S., & Mudambi, R. (2013). MNEs as border-crossing multi-location enterprises: The role of discontinuities in geographic space. Journal of International Business Studies, 44, 413-426.

Birkinshaw, J., & Hood, N. (1998). Multinational subsidiary evolution: Capability and charter change in foreign-owned subsidiary companies. Academy of Management Review, 23, 773-795.

38

Blonigen, B. A. (2001). In search of substitution between foreign production and exports. Journal of International Economics, 53, 81-104.

Bouquet, C., & Birkinshaw, J. (2008). Weight versus voice: How foreign subsidiaries gain attention from corporate headquarters. Academy of Management Journal, 51, 577-601.

Brouthers, L. E., Gao, Y., & McNicol, J. P. (2008). Corruption and market attractiveness influences on different types of FDI. Strategic Management Journal, 29, 673-680.

Buckley, P. (2009). Internalisation thinking: From the multinational enterprise to the global factory. International Business Review, 18, 224-235.

Buckley, P., & Carter, M. (2004). A formal analysis of knowledge combination in multinational enterprises. Journal of International Business Studies, 35, 371-384.

Buckley, P., & Casson, M. (1976). The Future of the Multinational Enterprise. London: Macmillan.

Buckley, P., Devinney, T., & Louviere, J. (2007). Do managers behave the way theory suggests? A choice-theoretic examination of foreign direct investment location decision-making. Journal of International Business Studies, 38, 1069-1094.

Castellani, D., Jimenez, A., & Zanfei, A. (2013). How remote are R&D labs? Distance factors and international innovative activities. Journal of International Business Studies, 44, 649-675.

Caves, R. E. (2007). Multinational Enterprise and Economic Analysis (3rd ed.). Cambridge, U.K.: Cambridge University Press.

Chacar, A. S., & Lieberman, M. B. (2003). Organizing for technological innovation in the U.S. pharmaceutical industry. In J. Baum & O. Sorenson (Eds.), Advances in Strategic Management: Geography and Strategy (Vol. 20, pp. 299-322). Kidlington: JAI.

Chung, W. (2001). Mode, size, and location of foreign direct investments and industry markups. Journal of Economic Behavior and Organization, 45, 185-211.

Cockburn, I. M., & MacGarvie, M. J. (2011). Entry and patenting in the software industry. Management Science, 57, 915-933.

Coval, J. D., & Moskowitz, T. J. (1999). Home bias at home: Local equity preference in domestic portfolio. Journal of Finance, 54, 2045-2073.

Cuervo-Cazurra, A., & Genc, M. (2008). Transforming disadvantages into advantages: Developing-country MNEs in the least developed countries. Journal of International Business Studies, 39, 957-979.

Delios, A., & Beamish, P. W. (2005). Regional and global strategies of Japanese firms. Management International Review, 45, 19-36.

Demirbag, M., & Glaister, K. W. (2010). Factors determining offshore location choice for R&D projects: A comparative study of developed and emerging regions. Journal of Management Studies, 47, 1534-1560.

Dicken, P. (2007). Global Shift: Mapping the Changing Contours of the World Economy (5th ed.). New York: Guilford Press.

39

Dunning, J. H. (1998). Location and the multinational enterprise: A neglected factor? Journal of International Business Studies, 29, 45-66.

Dunning, J. H., & Lundan, S. (2008). Multinational enterprises and the global economy (2nd ed.). Cheltenham, UK: Edward Elgar.

Enright, M. (2009). The location of activities of manufacturing multinationals in the Asia-Pacific. Journal of International Business Studies, 40, 818-839.

Fang, Y., Jiang, G., Makino, S., & Beamish, P. (2010). Multinational firm knowledge, use of expatriates, and foreign subsidiary performance. Journal of Management Studies, 47, 27-54.

Feinberg, S. E., & Gupta, A. K. (2004). Knowledge spillovers and the assignment of R&D responsibilities to foreign subsidiaries. Strategic Management Journal, 25, 823-845.

Flores, R. G., & Aguilera, R. V. (2007). Globalization and location choice: an analysis of US multinational firms in 1980 and 2000. Journal of International Business Studies, 38, 1187-1210.

Folta, T. B., & O'Brien, J. (2004). Entry in the presence of dueling options. Strategic Management Journal, 25, 121-138.

Freeman, L. C. (1979). Centrality in social networks: Conceptual clarification. Social Networks, 1, 215-239.

Frost, T. (2001). The geographic sources of foreign subsidiaries' innovations. Strategic Management Journal, 22, 101-123.

Frost, T., Birkinshaw, J., & Ensign, P. (2002). Centers of excellence in multinational corporations. Strategic Management Journal, 23, 997-1018.

Furman, J., Porter, M., & Stern, S. (2002). The determinants of national innovative capacity. Research Policy, 31, 899-933.

Ghoshal, S., Korine, H., & Szulanski, G. (1994). Interunit communication in multinational corporations. Management Science, 40, 96-110.

Goerzen, A., Asmussen, C., & Nielsen, B. (2013). Global cities and multinational enterprise location strategy. Journal of International Business Studies, 44, 427-450.

Goerzen, A., & Makino, S. (2007). Multinational corporation internationalization in the service sector: A study of Japanese trading companies. Journal of International Business Studies, 38, 1149-1169.

Guillén, M. F. (2003). Experience, imitation, and the sequence of foreign entry: Wholly owned and joint-venture manufacturing by South Korean firms and business groups in China, 1987-1995. Journal of International Business Studies, 34, 185-198.

Gupta, A., & Govindarajan, V. (2000). Knowledge flows within multinational corporations. Strategic Management Journal, 21, 473-496.

Hakanson, L., & Ambos, B. (2010). The antecedents of psychic distance. Journal of International Management, 16, 195-210.

40

Hakkala, K. N., Norback, P., & Svaleryd, H. (2008). Asymmetric effects of corruption on FDI: Evidence from Swedish multinaitonal firms. Review of Economics and Statistics, 90, 627-642.

Hallen, B. L. (2008). The causes and consequences of the initial network positions of new organizations: From whom do entrepreneurs receive investments? Administrative Science Quarterly, 53, 685-718.

Hansen, M. T., & Lovas, B. (2004). How do multinational companies leverage technological competencies? Moving from single to interdependent explanations. Strategic Management Journal, 25, 801-822.

Head, K., Ries, J., & Swenson, D. (1995). Agglomeration benefits and location choice: Evidence from Japanese manufacturing investments in the United States. Journal of International Economics, 38, 223-247.

Henisz, W. J. (2002). The institutional environment for infrastructure investment. Industrial and Corporate Change, 11, 355-389.

Henisz, W. J., & Delios, A. (2001). Uncertainty, imitation, and plant location: Japanese multinational corporations, 1990-1996. Administrative Science Quarterly, 46, 443-475.

Hennart, J. F. (1982). A theory of multinational enterprise. Ann Arbor: University of Michigan Press.

Hoetker, G. (2006). Do modular products lead to modular organizations? Strategic Management Journal, 27, 501-518.

Hoetker, G. (2007). The use of logit and probit models in strategic management research: Critical issues. Strategic Management Journal, 28, 331-343.

Hofstede, G. (2001). Culture's Consequences: Comparing Values, Behaviors, Institutions, and Organizations Across Nations. Thousand Oaks, CA: Sage Publications.

Holburn, G., & Zelner, B. (2010). Political capabilities, policy risk, and international investment strategy: Evidence from the global electric power generation industry. Strategic Management Journal, 31, 1290-1315.

Hutzschenreuter, T., Pedersen, T., & Volberda, H. W. (2007). The role of path dependency and managerial intentionality: A perspective on international business research. Journal of International Business Studies, 38, 1055-1068.

Hutzschenreuter, T., & Voll, J. C. (2008). Performance effects of ‘‘added cultural distance’’ in the path of international expansion: The case of German multinational enterprises. Journal of International Business Studies, 39, 53-70.

Hutzschenreuter, T., Voll, J. C., & Verbeke, A. (2011). The impact of added cultural distance and cultural diversity on international expansion patterns: A Penrosean perspective. Journal of Management Studies, 48, 305-329.

Iammarino, S., & McCann, P. (2013). Multinationals and economic geography. Cheltenham, UK: Edward Elgar.

Isham, J., Woolcock, M., Pritchett, L., & Busby, G. (2005). The varieties of resource experience: Natural resource export structures and the political economy of economic growth. World

41

Bank Economic Review, 19, 141-174.

Isobe, T., Makino, S., & Montgomery, D. (2000). Resource commitment, entry timing, and market performance of foreign direct investments in emerging economies: The case of Japanese international joint ventures in China. Academy of Management Journal, 43, 468-484.

Jiang, G., Holburn, G., & Beamish, P. (2014). The impact of vicarious experience on foreign location strategy. Journal of International Management, 20, 345-358.

Johanson, J., & Vahlne, J. E. (2009). The Uppsala internationalization process model revisited: From liability of foreignness to liability of outsidership. Journal of International Business Studies, 40, 1411–1431.

Kalnins, A. (2004). Divisional multimarket contact within and between multiunit organizations. Academy of Management Journal, 47, 117-128.

Kang, J. K., & Kim, J. M. (2010). Do foreign investors exhibit a corporate governance disadvantage? An information asymmetry perspective. Journal of International Business Studies, 41, 1415-1438.

Kang, Y., & Jiang, F. (2012). China's outward foreign direct investment: Location choice and firm ownership. Journal of World Business, 45, 45-53.

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2008). Governance matters VII: Aggregate and individual governance indicators, 1996-2007. In: World Bank Policy Research Working Paper No. 4654.

Kim, T.-Y., Delios, A., & Xu, D. (2010). Organizational geography, experiential learning and subsidiary exit: Japanese foreign expansions in China, 1979-2001. Journal of Economic Geography, 10, 579-597.

Kim, Y.-C., Lu, J. W., & Rhee, M. (2012). Learning from age difference: Interorganizational learning and survival in Japanese foreign subsidiaries. Journal of International Business Studies, 43, 719-745.

King, G., Tomz, M., & Wittenberg, J. (2000). Making the most of statistical analyses: Improving interpretation and presentation. American Journal of Political Science, 44, 347-361.

King, G., & Zeng, L. (2001). Logistic regression in rare events data. Political Analysis, 9, 137-163.

Kogut, B., & Chang, S. J. (1996). Platform investments and volatile exchange rates: Direct investment in the U.S. by Japanese electronic companies. Review of Economics and Statistics, 78, 221-231.

Kogut, B., & Singh, H. (1988). The effect of national culture on the choice of entry mode. Journal of International Business Studies, 19, 411-432.

Kogut, B., & Zander, U. (1993). Knowledge of the firm and the evolutionary theory of the multinational corporation. Journal of International Business Studies, 24, 625-645.

Kolstad, I., & Wiig, A. (2012). What determines Chinese outward FDI? Journal of World Business, 47, 26-34.

42

Kuemmerle, W. (1999). The drivers of foreign direct investment into research and development: An empirical investigation. Journal of International Business Studies, 30, 1-24.

Levinthal, D. A. (1997). Adaptation on rugged landscapes. Management Science, 43, 934-950.

Luo, Y. D. (2005). Toward coopetition within a multinational enterprise: A perspective from foreign subsidiaries. Journal of World Business, 40, 71-90.

Makino, S., Beamish, P., & Zhao, N. (2004). The characteristics and performance of Japanese FDI in less developed and developed countries. Journal of World Business, 39, 377-392.

Makino, S., Lau, C.-M., & Yeh, R.-S. (2002). Asset-exploitation versus asset-seeking: Implications for location choice of foreign direct investment from newly industrialized economies. Journal of International Business Studies, 33, 403-421.

Malloy, C. J. (2005). The geography of equity analysis. Journal of Finance, 60, 719-755.

Martin, X., Swaminathan, A., & Tihanyi, L. (2007). Modeling international expansion. Research Methodology in Strategy and Management, 4, 103-119.

Mascarenhas, B. (1986). International strategies of non-dominant firms. Journal of International Business Studies, 17, 1-25.

McCann, P. (1998). The economics of industrial location: A logistics-costs approach. Berlin and New York: Springer.

McCann, P. (2008). Globalization and economic geography: The world is curved, not flat. Cambridge Journal of Regions, Economy and Society, 1, 351-370.

McCann, P. (2011). International business and economic geography: Knowledge, time and transaction costs. Journal of Economic Geography, 11, 309-317.

McCann, P., & Mudambi, R. (2004). The location behavior of the multinational enterprise: Some analytical issues. Growth and Change, 35, 491-524.

McCann, P., & Shefer, D. (2004). Location, agglomeration and infrastructure. Papers in Regional Science, 83, 177-196.

Mitra, D., & Golder, P. N. (2002). Whose culture matters? Near-market knowledge and its impact on foreign market entry timing. Journal of Marketing Research, 39, 350-365.

Mudambi, R. (1995). The MNE investment location decision: Some empirical evidence. Managerial and Decision Economics, 16, 249-257.

Mudambi, R., & Navarra, P. (2004). Is knowledge power? Knowledge flows, subsidiary power and rent-seeking within MNCs. Journal of International Business Studies, 35, 385-406.

Mudambi, R., Pedersen, T., & Andersson, U. (2014). How subsidiaries gain power in multinational corporations. Journal of World Business, 49, 101-113.

Mudambi, R., Piscitello, L., & Rabbiosi, L. (2014). Reverse knowledge transfer in MNEs: Subsidiary innovativeness and entry modes. Long range planning, 47, 49-63.

Nachum, L., & Wymbs, C. (2005). Product differentiation, external economies and MNE location choices: M&As in global cities. Journal of International Business Studies, 36, 415-434.

43

Nohria, N., & Ghoshal, S. (1994). Differentiated fit and shared values: Alternatives for managing headquarters-subsidiary relations. Strategic Management Journal, 15, 491-502.

Peng, G. Z., & Beamish, P. (2014). MNC subsidiary size and expatriate control: Resource-dependence and learning perspectives. Journal of World Business, 49, 51-62.

Qian, L., & Delios, A. (2008). Internalization and experience: Japanese banks' international expansion, 1980-1998. Journal of International Business Studies, 39, 231-248.

Ragozzino, R. (2009). The effects of geographic distance on the foreign acquisition activity of U.S. firms. Management International Review, 49, 509-535.

Ramasamy, B., Yeung, M., & Laforet, S. (2012). China's outward foreign direct investment: Location choice and firm ownership. Journal of World Business, 47, 17-25.

Rugman, A. M. (1981). Inside the Multinationals: The Economics of Internal Markets. New York: Columbia University Press.

Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35, 3-18.

Rugman, A. M., & Verbeke, A. (2007). Liabilities of regional foreignness and the use of firm-level versus country-level data: A response to Dunning et al. Journal of International Business Studies, 38, 200-205.

Schonberger, R. J. (1996). World class manufacturing: The next decade. New York: Free Press.

Schotter, A., & Beamish, P. (2013). The hassle factor: An explanation for managerial location shunning. Journal of International Business Studies, 44, 521-544.

Singh, J. (2005). Collaborative networks as determinants of knowledge diffusion patterns. Management Science, 51, 756-770.

Slangen, A. (2011). A communication-based theory of the choice between greenfield and acquisition entry. Journal of Management Studies, 48, 1699-1726.

Slangen, A., & Beugelsdijk, S. (2010). The impact of institutional hazards on foreign multinational activity: A contingency perspective. Journal of International Business Studies, 41, 980-995.

Wheeler, D., & Mody, A. (1992). International investment location decision: The case of United States firms. Journal of International Economics, 33, 57-76.

Zelner, B. A. (2009). Using simulation to interpret results from Logit, Probit, and other nonlinear models. Strategic Management Journal, 30, 1335-1348.

Zhou, C. H., Delios, A., & Yang, J. (2002). Locational determinants of Japanese foreign direct investment in China. Asia Pacific Journal of Management, 19, 63-86.

44

TABLE 1. Descriptive Statistics and Bivariate Correlations

Variable Mean S. D. Min Max 1 2 3 4 5 6 7 8 9 10 111 Entry 0.09 0.29 0 12 Proximity to Regional Production Subsidiaries 0.33 0.60 0 9.00 0.0763 Proximity to Regional Trading Subsidiaries 0.34 0.67 0 9.00 0.015 0.3704 GDP 0.51 1.27 0.00 10.9 0.444 0.024 0.0315 GDP Growth Rate 0.07 0.05 -0.17 0.44 0.209 -0.004 -0.034 0.0536 GDP per capita 9.39 8.63 0.15 61.9 -0.025 0.305 0.347 0.223 -0.1247 Trade Ratio 0.62 0.50 0.04 4.13 0.032 0.180 0.157 -0.180 0.096 0.2378 FDI as a percentage of GDP 2.22 5.23 -46.1 88.2 0.022 0.107 0.097 -0.009 -0.022 0.277 0.3879 FDI Stock 0.05 0.20 0.00 2.80 0.202 0.059 0.066 0.771 -0.028 0.399 -0.052 0.027

10 Country Resource Exports 41.0 31.3 0 100 -0.146 -0.243 -0.245 -0.231 0.040 -0.400 -0.259 -0.141 -0.16811 Country Patent Grants 4.71 2.90 0.69 11.9 0.190 0.232 0.269 0.506 -0.051 0.547 -0.083 0.042 0.396 -0.45812 Cultural Distance 3.22 1.62 0.72 9.35 0.057 0.061 0.059 -0.037 0.057 0.149 0.252 0.067 0.002 0.015 -0.06713 Political Stability 0.53 0.31 0 0.90 -0.123 0.198 0.242 -0.008 -0.181 0.531 0.176 0.161 0.157 -0.323 0.37814 Rule of Law 0.51 0.99 -1.53 2.36 0.002 0.301 0.348 0.083 0.050 0.752 0.286 0.125 0.231 -0.396 0.48715 Bilateral Geographic Distance 9.63 3.91 1.16 18.6 -0.354 -0.163 -0.121 -0.196 -0.234 -0.084 -0.302 -0.062 -0.022 0.432 -0.25016 Proximity to Non-regional Production Subsidiaries 0.10 0.04 0 0.94 -0.210 0.047 0.091 -0.114 -0.169 0.124 -0.055 0.008 -0.046 -0.056 0.07017 Country Production Operation 0.13 0.36 0 3.37 0.419 0.059 0.060 0.501 0.157 0.060 0.025 0.011 0.311 -0.153 0.24618 Country Trading Operation 0.08 0.27 0 2.8 0.198 0.143 0.186 0.364 0.043 0.227 0.068 0.010 0.305 -0.145 0.27019 Industry Country Production Operation 1.23 1.6 0 6.5 0.519 0.142 0.109 0.599 0.225 0.168 0.136 0.043 0.379 -0.295 0.40320 Region Production Operation 0.67 0.83 0 4.17 0.235 0.375 0.174 0.096 0.128 0.019 0.198 0.042 -0.014 -0.278 0.10721 Firm Size 1.39 1.29 -3.38 5.11 -0.002 0.108 0.148 -0.002 -0.015 0.003 -0.010 -0.020 -0.004 -0.001 0.00322 Capital Intensity 4.78 18.4 -54.0 72.5 -0.001 0.035 0.052 -0.001 -0.022 0.017 0.004 0.004 0.004 -0.013 0.003

Variable 12 13 14 15 16 17 18 19 20 2113 Political Stability 0.08914 Rule of Law 0.297 0.57715 Bilateral Geographic Distance -0.206 0.029 -0.14816 Proximity to Non-regional Production Subsidiaries -0.023 0.107 0.089 0.06017 Country Production Operation 0.043 -0.038 0.058 -0.282 -0.13018 Country Trading Operation 0.035 0.135 0.234 -0.115 -0.041 0.43419 Industry Country Production Operation 0.059 0.032 0.156 -0.518 -0.303 0.574 0.38620 Region Production Operation 0.097 -0.001 0.031 -0.484 -0.097 0.362 0.219 0.39121 Firm Size 0.005 0.006 0.020 0.012 0.081 0.139 0.171 0.001 0.27722 Capital Intensity 0.000 0.014 0.007 0.007 0.049 0.020 -0.002 -0.016 0.044 0.178

N=43,546; Correlation with absolute values greater than 0.008 are significant at the p<0.05

45

TABLE 2. Rare Event Logit Models of Production EntriesVariable

Proximity to Regional Production Subsidiaries 0.44 ** 0.43 ** -0.05 0.44 ** 0.34 † 0.44 ** 0.16 *(0.03) (0.04) (0.31) (0.04) (0.2) (0.04) (0.06)

Proximity to Regional Trading Subsidiaries 0.01 0.07 -0.36 0.07 0.36 * 0.02 -0.01(0.04) (0.06) (0.27) (0.06) (0.17) (0.06) (0.05)

GDP 0.39 ** 0.41 ** 0.51 ** 0.37 † 0.51 ** 0.19 0.51 ** 0.00 †(0.03) (0.03) (0.06) (0.19) (0.06) (0.19) (0.06) (0.00)

GDP Growth Rate 5.62 ** 6.05 ** 6.12 ** 3.15 6.05 ** -3.83 6.22 ** 1.91 *(0.78) (0.81) (0.83) (3.08) (0.87) (5.42) (0.84) (0.94)

GDP per capita -0.04 ** -0.05 ** -0.06 ** -0.06 † -0.06 ** -0.06 † -0.06 ** 0.00 †(0.01) (0.01) (0.01) (0.03) (0.01) (0.04) (0.01) (0.00)

Trade Ratio -0.03 -0.02 -0.15 † -0.16 -0.13 -0.03 -0.14 † 0.06(0.05) (0.06) (0.08) (0.24) (0.09) (0.33) (0.09) (0.09)

FDI as a percentage of GDP 0.04 ** 0.04 ** 0.03 ** 0.02 0.03 ** 0.02 0.03 ** 0.01(0.00) (0.00) (0.01) (0.03) (0.01) (0.03) (0.01) (0.01)

FDI Stock -0.78 ** -0.79 ** -1.36 ** -0.60 -1.27 ** -0.05 -1.4 ** 0.21 *(0.14) (0.15) (0.3) (0.95) (0.31) (0.74) (0.33) (0.09)

Country Resource Exports 0.02 ** 0.01 ** 0.01 ** 0.02 ** 0.01 ** 0.00 0.01 ** 0.00(0.00) (0.00) (0.00) (0.01) (0.00) (0.01) (0.00) (0.00)

Country Patent Grants 0.03 † 0.04 * 0.01 -0.02 0.01 -0.08 0.01 0.04(0.02) (0.02) (0.02) (0.08) (0.02) (0.09) (0.02) (0.02)

Cultural Distance -0.08 ** -0.08 ** 0.03 0.19 * 0.02 -0.29 † 0.04 † -0.02(0.03) (0.03) (0.03) (0.09) (0.03) (0.17) (0.03) (0.05)

Political Stability -0.23 * -0.19 0.17 1.28 * 0.11 -0.14 0.16 -0.30 †(0.12) (0.12) (0.14) (0.55) (0.15) (0.86) (0.15) (0.18)

Rule of Law 0.22 ** 0.16 † 0.32 ** 0.26 0.33 ** 1.08 ** 0.32 ** -0.09(0.08) (0.08) (0.06) (0.22) (0.07) (0.36) (0.06) (0.13)

Bilateral Geographic Distance -0.06 ** -0.05 * -0.18 ** -0.16 * -0.18 ** -0.17 * -0.18 ** 0.00(0.02) (0.02) (0.02) (0.08) (0.02) (0.07) (0.02) (0.00)

Proximity to Non-regional Production Subsidiaries -6.22 ** -9.11 ** -7.64 * -8.53 ** -8.4 * -8.65 ** 2.45 ** (0.84) (1.11) (3.71) (1.18) (4.22) (1.15) (0.69)

Country Production Operation 0.28 ** 0.34 ** 0.22 * 0.68 * 0.19 † 0.91 * 0.19 * 0.10(0.06) (0.06) (0.09) (0.29) (0.1) (0.4) (0.1) (0.07)

Country Trading Operation 0.02 0.00 -0.10 -0.06 -0.09 0.26 -0.11 -0.04(0.07) (0.07) (0.12) (0.43) (0.12) (0.45) (0.12) (0.07)

Industry Country Production Operation 0.63 ** 0.59 ** 0.52 ** 0.35 ** 0.51 ** 0.53 ** 0.51 ** -0.06(0.03) (0.03) (0.03) (0.13) (0.04) (0.17) (0.03) (0.04)

Region Production Operation 0.07 † 0.06 -0.11 * 0.16 -0.12 * -0.65 * -0.08 0.03(0.04) (0.04) (0.05) (0.18) (0.05) (0.3) (0.05) (0.05)

Firm Size -0.06 * -0.03 -0.11 ** -0.2 * -0.11 ** -0.10 -0.10 ** 0.04(0.02) (0.02) (0.03) (0.09) (0.03) (0.16) (0.03) (0.03)

Capital Intensity 0.00 0.00 0.00 0.01 † 0.00 0.01 † 0.00 0.00(0.00) (0.00) (0.00) (0.01) (0.00) (0.01) (0.00) (0.00)

Subsidiary Size 0.10 ** (0.03)

Subsidiary Age 0.84 ** (0.05)

Expatriate Ratio 0.05 (0.28)

Parent Firm Performance 4.10 ** (0.76)

Number of observations† p < .10; * p < .05; ** p < .01. Fixed effects (region, industry, year, and subsidiary ownership) are not reported.a. NRS = Natural Resource-Seeking

20,521 18,726 43,546 43,546 21,438 1,330 20,108 917

Reported Performance

Model 7 Model 8Mandate NRSa Non-NRS R&D Non-R&D Subsidiary

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

46

FIGURE 1: Impact of Proximity to Regional Subsidiaries on Production Entry

11.

52

2.5

Pro

babi

lity

of P

rodu

ctio

n E

ntry

(Nor

mal

ized

to 1

whe

n P

roxi

mity

is e

qual

to z

ero)

0 .5 1 1.5 2Proximity to Regional Subsidiaries

Proximity to Regional Production Subsidiaries

Proximity to Regional Trading Subsidiaries

APPENDIX 1: Regions and Countries

Region CountryAsia Bangladesh, China, Hong Kong, India, Indonesia, South Korea, Malaysia,

Pakistan, Philippines, Singapore, Taiwan, Thailand, Turkey, and Vietnam.Europe Albania, Austria, Belarus, Belgium, Bulgaria, Croatia, Czech, Denmark,

Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Macedonia, Moldova, Netherlands, Norway, Poland, Portugal, Romania, Russia, Slovakia, Slovenia, Spain, Sweden, Switzerland, Ukraine, and the United Kingdom.

North America Canada, Mexico, and the U.S.A.Latin America & the Caribbean

Argentina, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Jamaica, Panama, Peru, Suriname, Trinidad & Tobago, Uruguay, and Venezuela.

Oceania Australia and New ZealandAfrica Ghana, Mali, Nigeria, Rwanda, South Africa, Tanzania, Uganda, Zambia,

and Zimbabwe.Middle East & North Africa

Algeria, Egypt, Iran, Iraq, Israel, Jordan, Morocco, and Saudi Arabia.

47