Journal of Business Research 57 (2004) 35 – 46
Foreign expansion in service industries
Separability and human capital intensity
Cyril Bouqueta,*, Louis Hébertb,1, Andrew Deliosc,2
a
b
Richard Ivey School of Business, University of Western Ontario, 1151 Richmond Street, London, Ontario, Canada N6A 3K7
Management Department, Ecole des Hautes Etudes Commerciales, 3000 Côte Ste-Catherine, Montreal, Quebec, Canada H3T 2A7
c
Department of Business Policy, National University of Singapore, 1 Business Link, Singapore 117592, Singapore
Received 2 April 2001; accepted 6 February 2002
Abstract
We investigate the effect of operating in service industries, in which separability and human capital intensity factors influence the choice
of foreign entry mode and expatriate staffing decisions. To look into this issue, we compared 14,863 instances of Japanese foreign direct
investment (FDI) into manufacturing and three service industries (wholesale trade, retail trade, and financial services). Our theoretical and
empirical analyses support the assertion that in situations where required capabilities must be developed through (1) close contacts with end
customers and (2) high levels of professional skills, specialized know-how, and customization, wholly owned subsidiaries and expatriate staff
are preferred. From our results, we draw implications for the FDI literature and offer a novel perspective on the factors influencing the
internationalization of service firms.
D 2002 Elsevier Science Inc. All rights reserved.
Keywords: Foreign expansion; Service industries; Separability; Human capital intensity
1. Introduction
When expanding internationally, a firm must determine
the appropriate mode for entering foreign markets. It must
also decide whether to staff foreign subsidiaries with local
and/or expatriate managers. Both decisions have important
consequences for a firm’s competitive advantage in new
international markets (Edstrom and Galbraith, 1977; Hill et
al., 1990). Indeed, while wholly owned subsidiaries and
expatriate staff provide foreign investors with greater control over foreign operations, they also entail substantial
resource commitments, such as capital and managerial
resources, in the host country that cannot be easily redeployed to alternative locations.
Research on the choice of entry mode and expatriate
staffing strategies has expanded considerably for some years,
with a traditional focus on manufacturing firms (Anand and
Delios, 1997; Li and Guisinger, 1992). However, the in* Corresponding author. Tel.: +1-519-661-2111x85328; fax: +1-519661-3959.
E-mail addresses: cbouquet@ivey.uwo.ca (C. Bouquet),
louis.hebert@hec.ca (L. Hébert), andrew@nus.edu.sg (A. Delios).
1
Tel.: + 1-514-340-6334; fax: + 1-514-340-5635.
2
Tel.: + 65-874-3094; fax: + 65-779-5059.
0148-2963/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved.
doi:10.1016/S0148-2963(02)00282-5
creased importance of services in developed economies and
the fast growth of foreign investment in the service sector
have fueled research on service multinationals (Aharoni and
Nachum, 2000; Boddewyn et al.,1986; Dunning, 1989). Still,
scholars have debated whether the determinants of foreign
entry decisions are the same for service and manufacturing
firms. One group suggests that theories of foreign direct
investment (FDI) apply to global service firms (Dunning,
1989; Miller and Parkhe, 1998; Yannopoulos, 1983). Another
group argues that crucial differences between goods and
services make it difficult to generalize FDI theories across
industry sectors (Boddewyn et al., 1986; Erramilli, 1992;
Gronroos, 1999).
To widen the focus of this debate and enhance our
understanding of foreign investment in the service sector,
we developed an analytical framework for examining interindustry differences in entry mode and expatriate staffing
strategies among multinational firms that compete in services
and manufacturing industries. This framework suggests that
wholly owned subsidiaries and expatriate staff are preferred
by MNCs competing in industries where required capabilities
must be developed through (1) high levels of professional
skills, specialized know-how, and customization and (2)
close interactions with end customers. Empirical tests using
36
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
a sample of 14,863 entries of Japanese multinational firms
entering the Asian, North American, and European markets
provide supporting evidence.
2. Foreign market entry decisions
Considerable attention has been given to identifying
firm- and country-level determinants of foreign market entry
decisions (Hill et al., 1990). A large portion of this literature
has used internalization theory and transaction cost analysis
to explain how companies enter foreign markets (Anderson
and Gatignon, 1986; Davis et al., 2000). Various entry
modes are available to firms, from export to licensing to
ownership-based modes such as joint ventures and wholly
owned subsidiaries. Full-control modes, such as greenfield
or acquired wholly owned subsidiaries, have been differentiated from shared-control ones, such as a greenfield or
partially acquired joint ventures. Full-control modes
increase the degree of control that an MNC can exercise
over its foreign subsidiaries, but also require greater
resource commitments compared to shared-control modes
(Anderson and Gatignon, 1986; Nitsch et al., 1996).
Research has also explored the complementary use of
systems and procedures, such as staffing or human resourcebased mechanisms (Mayrhofer and Brewster, 1996; Tung,
1982), to control foreign operations (Baliga and Jaeger,
1984). The use of expatriates deals with the limitations of
solely using ownership to protect firm-specific assets. While
a full-control mode such as a wholly owned subsidiary can
help to address opportunism problems (Sohn, 1994) shirking or hold-up by local management remain possible
(Alchian and Demsetz, 1972; Alchian and Woodward,
1988). Expatriate managers can reduce risks of opportunism
and ensure that company policies are carried out effectively
in foreign subsidiaries (Baliga and Jaeger, 1984; Edstrom
and Galbraith, 1977; Roth and Nigh, 1992).
In expanding the focus of international research from
manufacturing to service firms, scholars have explored
whether theories of multinational manufacturing firms apply
to multinational service activity (Boddewyn et al.,1986;
Dunning, 1989) and the international expansion of service
firms (Katrishen and Scordis, 1998; Li and Guisinger, 1992;
Miller and Parkhe, 1998). For the entry mode decision,
studies have highlighted the tendency of service firms to
rely on wholly owned subsidiaries (Erramilli, 1992; Erramilli and Rao, 1993). A notable difference between the
internationalization of services and manufacturing firms is
that a service firm seldom requires large-scale investments
in physical assets such as capital equipment and facilities to
establish a presence in foreign markets. The value-creating
assets of a service firm rest more on its human capital than
on its physical infrastructure (Erramilli and Rao, 1993;
Campbell and Verbeke, 1994). Accordingly, international
investments in the service sector have been found to rely
heavily on ‘‘people-transfers; that is, training programs,
visits by experts and the employment of expatriates’’
(Grosse, 1996, p. 796 emphasis added).
This stream of research has also suggested that service
MNCs face unique challenges when expanding abroad. The
distinct nature of a service firm’s assets exerts stresses on
entry mode choice decisions and the use of expatriate
managers. This idea has fueled the debate over the generalization of FDI theories across sectors and the development
of service-specific frameworks (Carman and Langeard,
1980; Erramilli, 1992; Gronroos, 1999). Many researchers
argue that there are crucial differences in the production and
delivery of services and goods. Services differ from manufactured goods along features such as the intangibility of the
offering, the separability of production and consumption,
and the perishability of inventories (Lovelock and Yip,
1996; Zeithaml et al., 1985).
Among these variables, the ‘‘separability’’ of production
and consumption has been deemed of importance for service
firms’ internationalization (Erramilli and Rao, 1993). Separability characterizes transactions by the level of interaction
required between providers and users (Hirsh, 1989). News
delivery is an example of a separable service. Like most
material goods, information may be designed, manufactured, and stocked for later delivery and consumption. Most
other services (e.g., those provided by hotels or restaurants)
necessitate the close physical proximity of buyers and
sellers (Anand and Delios, 1997; Carman and Langeard,
1980; Zeithaml et al., 1985). Separable products or services
can be transferred to overseas markets where they can be
sold to a set of foreign consumers. Inseparable ones are
location-bound. If foreign consumers are to access an
inseparable service, the consumer must come to the site at
which the service is produced. This feature of inseparable
services, in which the product of location-bound resources
must be consumed at the same time and location at which it
is produced, has important implications for entry decisions
in foreign markets (Erramilli and Rao, 1993, p. 35).
The degree of ‘‘idiosyncrasy’’ that characterizes a service
is another key factor to consider when discussing the internationalization of service firms (Erramilli and Rao, 1993;
Zeithaml et al., 1985). Because many services are laborintensive or people-centered, the marked differences between
employees in terms of skills, education, or specialized knowhow create considerable variance in performance at service
production or delivery. Firms can reduce this risk by substituting physical resources for the human element (as in the
case of automated-teller machines) or through extensive
education and training of employees. When training is
extensive and when employees have a high level of skills
when joining a firm, the degree of human capital is greater.
According to Erramilli and Rao (1993), the production/
delivery of services typically relies on a high intensity of
human capital—the skills, talent, and knowledge of a firm’s
employees that are not easily transferred to different organizational or social contexts. Below, we contend that differences in human capital intensity faced by service firms
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
influence choices about market entry mode and the deployment of expatriate staff. When considered jointly with the
degree of separability of service transactions, differences in
human capital intensity can have major implications for entry
mode choice and expatriate staffing decisions.
2.1. Entry mode choice
Firms face various costs when doing business abroad
(Hymer, 1976), some of which relate to acquiring the
experiential knowledge necessary to operate in a local
market (Johanson and Vahlne, 1977). Such a process can
take considerable time (Dierickx and Cool, 1989; Makino
and Delios, 1996). Several scholars emphasize the utility of
joint ventures to a firm when the demand for locally specific
skills is high (Anand and Delios, 1997; Carman and Langeard, 1980).
Although joint ventures can be used to acquire local
complementary assets, shared-control modes are less valuable in industries exhibiting high human capital intensity. In
a joint venture, the foreign investor may find it difficult to
produce service transactions that require important and
specialized human skills, since social assets are typically
difficult to codify and transmit across organizational boundaries (Kogut and Zander, 1993). In shared-control modes, the
skills and specialized know-how possessed by employees
face risks of appropriation and dissemination (Murray and
Kotabe, 1999), as human expertise is difficult to protect,
contractually, by patents or by copyrights (Dunning, 1989;
Grosse, 1996). In such situations, a wholly owned entry
permits a firm to address the imperfect transferability,
appropriability, and free-riding shortcomings exhibited by
joint ventures (Anderson and Gatignon, 1986; Contractor
and Lorange, 1988; Hill et al., 1990; Kobrin, 1988; Teece,
1981). Therefore, a wholly owned subsidiary would represent a more efficient mode of market entry than a joint venture
for firms operating in human capital intensive industries.
Hypothesis 1: The greater the human capital intensity in a
firm’s industry, the greater the frequency of entry by a
wholly owned subsidiary.
By forcing the ‘‘buyer into intimate contact with the
production process’’ (Carman and Langeard, 1980, p. 8), a
business that is not separable has to conduct face-to-face
transactions with its customers. This simultaneity requirement necessitates greater adaptation to local differences and
local tastes, leading firms to acquire locally specific resources
and capabilities (Anand and Delios, 1997). Yet, a close
intimacy between providers and consumers throughout the
production-delivery process creates important risks to a firm
(Carman and Langeard, 1980; Erramilli and Rao, 1993;
Zeithaml et al., 1985). In particular, where separability is
low, the performance of employees who deal directly with
customers is important for the maintenance of a firm’s quality
standards. Frequent transactions between buyers and sellers
create a need for onsite quality assurance (Zeithaml et al.,
37
1985). Quality problems may be overcome by establishing
detailed contracts/monitoring procedures with local partners.
However, the quality of a service transaction may still vary
from one local provider to the other, raising an overall
problem of consistency for the foreign investor (Carman
and Langeard, 1980, p. 8). Another option is to keep transactions within the context of the firm’s organizational boundaries by using a wholly owned subsidiary rather than a joint
venture when entering a foreign market.
Hypothesis 2: The less separable a firm’s industry, the
greater the frequency of entry by a wholly owned
subsidiary.
The risk stemming from delivering transactions that are
not separable is likely to be especially important in industries with high human capital intensity. In such industries, it
is difficult to find local partners whose employees have
acquired the skills, knowledge, and specialized know-how
that are necessary to meet a firm’s quality standards. The
expansion of Marriott in Warsaw represents a case in point
(Loveman, 1997). The reverse of this idea is that industries
with high human capital intensity and a high degree of
separability are likely to be less sensitive to the influence of
human capital intensity on mode choice. In fact, the human
capital used in the production of the firm’s product does not
need to be sited where the product is consumed. This
scenario typifies high-technology manufacturing industries
that may require a high degree of specialized knowledge
among employees, usually those involved in research and
development, and that may produce unique, sophisticated
products in the firm’s home country market.
As with a separable industry, the difficulty of independent international expansion is low, even in nonseparable
industries, when human capital intensity is low. This point is
well illustrated by the case of Seiyu, an affiliate of the
Saison group, a major retail, credit, and development
conglomerate in Japan (JETRO, 1995). Seiyu, the fifth
largest Japanese retailer, was established in 1963. It runs a
chain of supermarkets, the popular Family Mart convenience stores, and about 200 other retail chains and services.
Competition in these different retail sectors tends to be
mostly price-based in Japan. In these retail sectors, most of a
retailer’s cost-related advantages come from the development of computer systems and technologies that reduce
the costs of sourcing and purchasing. Companies in the
retail sector in Japan have been looking to reduce costs and
increase price competitiveness by increasing sourcing from
overseas. Even though cost-saving systems can be firmspecific, they are back-office functions. The service and
management provided at the front-end of retail operations,
tends to be standardized across industry competitors
(JETRO, 1995). Consequently, Seiyu has relied on a high
proportion of low-trained staff in its operations—nearly
60% of its 16,000 employees are part-time staff with little
experience or specialized know-how. Driven by the 1990s
economic slump in Japan, Seiyu has been working to
38
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
establish an overseas presence in Singapore, Thailand,
Indonesia, and Vietnam by opening a number of stores
jointly with local companies. This is a strategy followed by
other leading retailers in Japan such as Jusco, which had
joint venture stores in Shanghai, Hong Kong, and the United
States as Talbots in the state of Maryland. The risks to Seiyu
and Jusco by this strategy of high local involvement in a
separable industry are limited, because the front-end of their
retail operations is relatively standardized and involves few
individual and customized skills. Thus, the potential for
variability in service performance is low.
These examples underscore the idea that low separability
may inflict costs and risk on a service firm’s internationalization, because the production and consumption of a nonseparable service are intertwined (Carman and Langeard,
1980). The cost, and the challenges of internationalization
borne by a firm, however, are not necessarily high when the
service provided at the point of consumption involves a low
degree of human capital intensity as the risks involved in
entering by a shared-control mode, such as a joint venture, are
relatively limited (Erramilli and Rao, 1993).
Hypothesis 3: The frequency of entry by wholly owned
subsidiaries will be greatest in industries exhibiting high
human capital intensity and low separability.
2.2. Expatriate staffing decisions
Expatriates tend to be used in a foreign subsidiary for a
variety of reasons and in a variety of contexts (Black et al.,
1999; Peterson et al., 1996). We propose that human capital
intensity and the separability of production and consumption
influence staffing decisions and the propensity to use expatriate managers. As suggested earlier, it can be difficult to
identify qualified local candidates with the necessary skills to
deliver professional services in a successful and consistent
manner. Expatriates represent an effective strategy to help
foreign affiliates adhere to corporate objectives and practices
(Edstrom and Galbraith, 1977; Kobrin, 1988; Mayrhofer and
Brewster, 1996). Expatriates may have the required experience, knowledge, and socially embedded skills that can be
transferred to local managers, through socialization processes
or appropriate training programs. The deployment of expatriates is thus a key step in transferring and developing a
firm’s unique human capital in its new markets.
Hypothesis 4: The greater the human capital intensity in a
firm’s industry, the greater the proportion of expatriate staff
in foreign subsidiaries.
For a foreign entrant, low levels of separability make it
difficult to control service quality and to provide good service
consistently because the product is delivered at the site of
consumption. Carman and Langeard (1980, p. 19) noted that
low separability drove many service firms to export personnel
to overseas locations to train local workers. By helping
to maintain quality standards, expatriates contribute to
strengthen trust-based relationships established with a company’s clients. Expatriates can be used to avoid perceptions of
unfairness, reduce customer retaliation, and forestall serious
damage to a firm’s reputation (Seiders and Berry, 1998).
Hypothesis 5: The less separable a firm’s industry, the
greater the proportion of expatriate staff in foreign
subsidiaries.
We also suggest that the influence of separability will be
greatest in human-capital intensive industries. When the
production and delivery of a product/service is peoplecentered, close contacts between producers and customers
entail limited risks for the maintenance of corporate standards. By using more local employees, firms can acquire
complementary assets while reducing the costs associated
with the maintenance of expatriate staff (Kopp, 1994a).
Where human capital intensity is high, however, a firm
faces magnified risks in not using its highly trained in-house
employees to deliver its services.
Hypothesis 6: The proportion of expatriate staff in foreign
subsidiaries will be greatest in industries exhibiting high
human capital intensity and low separability.
3. Methodology
3.1. Empirical setting
To test our hypotheses, we investigated Japanese foreign
investments in the manufacturing sector and three service
sectors: retail trade, wholesale trade, and financial services.
These categories are 4 of the 10 broad industry categories
identified in the Standard Industrial Classification (SIC)
handbook (1987). Together, foreign investments in these
categories accounted for 80% of the stock of Japanese
investment worldwide in 1999. In our definition of the
categories below, we outline how these categories differ in
one of the traditional distinctions between manufacturing
and service industries, namely the level of separability.
Establishing differences between broad industry categories
along the other dimension in our analysis, namely human
capital intensity, is not as immediately clear as it is for
separability. Even though we cannot a priori distinguish
across the industries on their level of human capital intensity,
we do expect considerable variance to exist across subsectors
of the manufacturing, wholesale trade, retail trade, and
financial services industries.
3.2. Industry categories
Services and manufacturing industries have many notable differences. Yet, research has identified that analytical
frameworks developed with reference to manufacturing
firms can be applied to services firms (Boddewyn et al.,
1986; Erramilli and Rao, 1993). To promote an integration
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
of frameworks, we include the foreign entries in the manufacturing sector in our sample. Manufacturing firms can be
most readily distinguished from services firms along the
dimension of separability. However, we construct measures
and tests to account for the substantial variance that exists
within each broad industry category along the separability
and the human capital intensity dimensions.
The literature on the services industry distinguishes
manufacturing firms (SIC 20 – 29) from services along
several characteristics, most notably the separability of
production and consumption (Lovelock and Yip, 1996;
Zeithaml et al., 1985). Even though the production of
products in such industries as electronic equipment, food
products, and textiles manufacturing involve different levels
of technological sophistication and require different skill
levels among employees, each product is similarly separable. These industries’ products can be designed in Japan,
manufactured in Taiwan, and assembled in another overseas
location, which is perhaps close to major consumer markets.
In retail trade (SIC 52 – 59), consumers need to experience, try out, or inspect goods before making purchase
decisions. Not surprisingly, the importance of value-added
services at the point of product exchange (information
assistance, merchandise sampling) has often been emphasized (Anand and Delios, 1997). Retail operations are
largely undertaken at the point of consumption, in close
interaction with customers, and thus are not separable.
Wholesale trade companies (SIC 50 and 51) buy large
lots of goods usually from manufacturers, to sell them in
smaller quantities to a set of institutional customers (businesses, governments, and other wholesalers). Wholesale
trading companies provide services such as transaction
intermediation, technical, logistical, and financial advice,
as well as marketing services (Kojima and Ozawa, 1984).
Most value-added activities, such as obtaining favorable
investment terms or securing local financing, can be performed in isolation, in absence from a producer– consumer
interaction, and are more separable than the products of
retail firms (Dunning, 1989; Lovelock and Yip, 1996).
The financial services industry comprises a variety of
actors that offer several types of services to customers,
businesses, or governments. Commercial banks and thrift
institutions (SIC 60 and 61) provide a full range of depository, saving, and lending services. Security and commodity
traders (SIC 62) sell newly issued stocks, bonds, and other
financial products to individual and institutional investors.
Insurance carriers and agents (SIC 63 and 64) explain and
sell life insurance, property-casualty, and health plan policies that provide clients with protection and reimbursement
for their losses. Real estate offices (SIC 65) rent, buy, sell,
manage, and appraise estate and properties for their clients.
Holding and investment offices (SIC 67) assist customers in
raising funds for financing capital expenditures and other
forms of spending. Many financial services, particularly
those that can be computerized and automated, are separable. Other services, such as money management counsel-
39
ing, mortgage planning, risk assessment services, require
face-to-face interaction with highly skilled and trained
professionals. These services have lower levels of separability and have inspired service firms to internationalize by
following important customers abroad (Nigh et al., 1986;
Terpstra and Yu, 1988).
3.3. Sample
Our sample comprised foreign entries in these four
sectors that were a foreign subsidiary operation. We did
not include representative, branch, bureau, or liaison offices
in our definition of foreign subsidiaries. Our source
(described later) permitted us to clearly distinguish between
a foreign subsidiary and a representative, branch, bureau, or
liaison office. We focused on foreign subsidiaries to ensure
the operation was engaged actively in local sales in the local
market. Among all subsidiaries, we used three criteria to
develop our sample. The subsidiary (1) had to be in North
America, Europe, or Asia (88.3% of all cases); (2) it had to
belong to the manufacturing, wholesale trade, retail trade, or
financial services industry (80% of all cases); and (3) it had
to be established as a wholly owned subsidiary or a joint
venture, as defined in Section 3.4. Once these three criteria
were applied, the sample numbered 14,863 entries: 7475 in
the manufacturing sector, 4966 in wholesale trade, 320 in
retail trade, and 2102 in financial services.
3.4. Variables
3.4.1. Entry mode
Wholly owned subsidiaries are a greenfield or acquired
subsidiary in which one partner possessed more than 95% of
its equity. A joint venture is a greenfield or partially
acquired subsidiary in which more than one firm possessed
at least 5% of the subsidiary’s equity.
3.4.2. Expatriate ratio
This is the ratio of Japanese expatriates to total subsidiary
employment.
3.4.3. Human capital intensity and separability
Our first measure of human capital intensity was the
percentage of a two-digit SIC industry’s employees with a
bachelor degree or higher. We determined this from a
special tabulation of the Current Population Survey conducted by the U.S. Bureau of the Census for the Bureau of
Labor Statistics. The second measure of human capital
intensity used Japanese data from 1994 and 1995 on the
extent of an employee’s unique contribution to a firm, as
published in Daiwa Research Institute’s Analyst’s Guide. We
obtained three items: value added per employee, labor costs
per employee, and sales per employee. With these three
items, we constructed a factor score for each two-digit SIC
industry in our sample. Greater scores on this measure imply
higher levels of human capital intensity—greater profes-
40
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
sional skills, more specialized know-how, and customization.
We based our measure of separability on Hirsh’s (1989,
p. 66) ‘‘S-factor,’’ which measures the degree to which
products or services are produced in isolation and in
interaction with the customer. Anand and Delios (1997)
operationalized the S-factor as the percentage of employees
in downstream occupations (e.g., sales workers, sales
agents, salespersons). We developed a measure identical to
Anand and Delios (1997), but using U.S. data. We created a
second measure using Japanese data available in the Analyst’s Guide. Using this guide, we determined mean expenditures by industry for direct selling expenses plus sales,
general, and administrative expenses. Greater scores on this
measure imply greater levels of interaction between producers and customers; that is, transactions are less separable
or bounded to local resources. We tested hypotheses using
the United States- and the Japanese-based measures, and
obtained qualitatively similar results. Therefore, we only
report the results for the Japanese measures.
3.5. Control variables
For country-level control variables, we included measures of country risk and a nation’s openness to FDI (Delios
and Beamish, 1999; Kim and Hwang, 1992; Shan, 1991).
We derived the host country risk variable from indices
published in Euromoney in 1996. Data on local ownership
restrictions were obtained from the World Competitiveness
Report (1996) using the item that reported on the openness
of 47 countries to foreign equity participation. We also
included measures of cultural distance, using Kogut and
Singh’s (1988) index. Our final country-level control measured a host country’s population and its level of GDP in
1996. Both were taken as logarithms (Grubaugh, 1987;
Wheeler and Mody, 1992).
We attempted to capture features of manufacturing firms
that might influence choices about entry mode and expatriate
employment independent of our other independent variables.
One such variable is the level of technology involved in the
production of a firm’s product. Where technology is high, a
firm faces a greater risk of technological leakage when
making a market entry (Delios and Beamish, 1999). The
tacitness of a firm’s technology also creates difficulties in
contracting for technology (Oxley, 1997), thereby leading
toward a preference for wholly owned entry. This variable
was estimated with the ratio of R&D expenses to sales
revenues for the parent firm’s industry in Japan and was
obtained from the Analyst’s Guide. This R&D variable
complements our measure of human-capital intensity. It also
helped ensure that results obtained for the human capital
intensity variable are not simply due to differences in technology levels across industries, which we consider to be a
different concept than human capital intensity.
We also controlled for organizational experience (Johanson and Vahlne, 1977; Makino and Delios, 1996) measured
with the logarithm of the years of FDI experience by a
parent firm, at the time of a subsidiary’s founding. We
included measures of subsidiary size, taken as the number of
employees in a subsidiary, and subsidiary age or the years
since a subsidiary was founded (Anand and Delios, 1997;
Hennart, 1991).
4. Analysis and results
We first examined how the four basic industries varied in
their mean levels of human capital intensity and separability.
As Table 1 shows, manufacturing and retail trade involved
lower average levels of human capital intensity than wholesale trade and financial services ( p < .001). Consistent with
our earlier description of the sectors, in general, wholesale
trade and manufacturing were more separable than retail
trade and financial services ( p < .001). We next proceeded
with statistical tests based on variance in the human capital
intensity and separability measures across two-digit SIC
industries. The tests for Hypotheses 3 and 6 required an
interaction term between human capital intensity and separability. The correlation between these two variables was
.892, hence we ‘‘centered’’ the human capital intensity and
separability measures for each value by subtracting the
mean value for each. This procedure reduces the correlation
between an interaction and its composite terms, without
altering substantive interpretations of the coefficients (Aiken
and West, 1991; Erramilli and Rao, 1993). Other variables
did not exhibit collinearity problems (Table 2). Regression
diagnostics, such as variance inflation factors and a hierarchical model building process, further reduced concerns
about multicollinearity.
4.1. The choice of entry mode
Our first test of the hypotheses involves an examination
of entry mode tendencies by industry category. The data
reported in Table 3 provide support to Hypotheses 1 –3. The
preference for a wholly owned subsidiary is lowest in the
manufacturing and retail trade sectors, greater in wholesale
trade, and greatest in financial services (74.5%). A chisquare test established the significance of these differences
at the p < .001 level. The propensity to use wholly owned
subsidiaries over joint ventures appears to increase as transactions become increasingly dependent upon human capital
Table 1
Across-industry variation of separability and human capital intensity
Manuf.
Retail
Wholesale Finance Significance *
(n = 7475) (n = 320) (n = 4966) (n = 2102)
Separability
36,232 48,368
Human
0.1773 0.4232
capital intensity
28,496
2.9507
54,527
1.5879
All mean differences are significant at the 1% level.
* ( p, one-way ANOVA test).
< .001
< .001
41
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
Table 2
Correlation matrix
Variables
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1.
2.
3.
4.
5.
6.
7.
8.
Country risk
Cultural distance .173* *
FDI openness
.484* * .079* *
Population
.494* * .205* * .545* *
GDP
.568* * .409* * .114* * .043* *
Subsidiary age
.252* * .149* * .168* * .315* * .117* *
Subsidiary size .068* * .016 .064* * .032* * .004
.064* *
Host country
.100* * .038* * .008 .006
.262* * .143* * .013
experience
9. Industry R&D
.038* * .011 .017 .014 .020 *
.014
.022 * .065* *
10. Manufacturing .370* * .089* * .442* * .346* * .110* * .156* * .172* * .038* * .072* *
11. Retail
.004 .002
.003
.014 .008 .012
.002 .016 * .166* * .149* *
12. Finance
.195* * .018
.219* * .135* * .139* * .021 * .072* * .085* * .191* * .408* * .060**
13. Human
.338* * .103* * .400* * .328* * .083* * .175* * .167* * .023* * .018 * .913* * .073** .129**
capital intensity
14. Separability
.024* * .050* * .041* * .019 *
.015 .048* * .068* * .023* * .041* * .015
.100** .416** .118**
.106* * .062* * .111* * .110* * .039* * .062* * .021 *
.002 .054* * .329* * .100** .125** .418** .292 **
15. Human
capital intensity
separability
* Correlation is significant at the .05 level (two-tailed).
* * Correlation is significant at the .01 level (two-tailed).
intensity and, also, less separable, thus providing support to
Hypotheses 1 and 2. Consistent with Hypothesis 3, the
propensity to enter by wholly owned subsidiary was greatest
in financial services.
We made a second test of Hypotheses 1 –3 using logistic
regression in order to control for other influences on entry
mode behavior by industry category, and to make an
examination of our hypotheses across more precisely
defined industry categories (i.e., two-digit SIC industries).
The tests compared the propensity to expand by a wholly
owned subsidiary vs. a joint venture.
The models in Table 4 were all significant and substantive predictors of entry mode choice. Model 1 provides a
baseline estimation using control variables. In Model 2, we
added the categorical variables for industry membership in
which wholesale trade is the reference category. A chisquare test for overall model fit shows that the categorical
variables significantly improves the fit of the model over its
baseline. The signs for the coefficient estimates on the
manufacturing, retail trade, and finance dummy variables
are consistent with our hypotheses. The negative coefficients on the manufacturing and retail variables ( p < .001)
suggests that these entries are less likely to involve a wholly
Table 3
Choice of entry mode
Manuf.
Retail
Wholesale
Finance
Total
(%, n = 7475) (%, n = 320) (%, n = 4966) (%, n = 2102) (%)
Wholly
36.9
owned
Joint
63.1
venture
45.6
69.1
76.2
53.4
54.4
30.9
23.8
46.6
For differences in categories, c2 = 1762, df = 3, p < .001.
owned subsidiary compared to wholesale trade entries. The
positive coefficient on the finance variable reflects the
greater tendency of finance subsidiaries to be wholly owned
compared to wholesale trade. Thus, the frequency of entry
by wholly owned subsidiary increases when human capital
intensity and the interactions between suppliers and customers become more prominent. These findings support
Hypotheses 1 –3.
Models 3 and 4 test the effect of the human capital
intensity and separability variables on the likelihood of a
wholly owned entry. A chi-square test for overall model fit
shows that the addition of these two continuous variables
significantly improved the baseline model. In Model 3, the
human capital intensity variable was positively signed and
significant ( p < .001). In Model 4, when adding the separability variable, the coefficients on the two focal variables were
positively signed and significant. These results support
Hypotheses 1 and 2. Model 5 tests the interaction between
human capital intensity and separability. The interaction term
was positively signed, thus providing empirical support to
Hypothesis 3.
4.2. Expatriate employment
Table 5 displays employment data for the three industries.
The mean number of employees and expatriates was significantly larger in manufacturing entries ( p < .001). When measured on a proportional basis, entries in wholesale trade and
finance used significantly more expatriates than entries in
manufacturing and retail trade ( p < .001). Also, as a proportion of total employment, expatriate employment was significantly greater ( p < .001) in financial services than in
wholesale trade. These trends hold when we compare sub-
42
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
Table 4
Logistic regression model: entry mode choice
Joint venture (0) vs. wholly owned (1)
Model 1
Model 2
Model 3
Model 4
Model 5
Country-level controls
Country risk
Culture distance
FDI openness
Population
GDP
.077* * * (202.924)
.060 (3.274)
.561* * * (172.249)
.330* * (8.511)
.059* * * (9.550)
.065* * * (140.711)
.044 (1.666)
.444* * * (101.360)
.360* * (9.952)
.090* * * (20.755)
.067* * * (151.250)
.045 (1.797)
.484* * * (122.098)
.367* * * (10.359)
.087* * * (19.778)
.067* * * (150.156)
.046 (1.868)
.477* * * 118.781
.363* * (10.122)
.088* * * (20.574)
.066* * * (145.383)
.045 (1.815)
.473* * * (116.515)
.356* * (9.728)
.092* * * (22.283)
Industry-level control
R&D intensity
.183* * * (60.622)
.222* * * (80.798)
.177* * * (54.606)
.181* * * (57.284)
.185* * * (59.571)
.000 (0.003)
.003 (0.891)
.004* * * (108.901)
.009 (1.260)
.004 (1.170)
.004* * * (125.191)
.004 (0.239)
.002 (0.275)
.004* * * (121.662)
.005 (0.483)
.001 (0.137)
.004* * * (122.267)
.006 (0.588)
.002 (0.364)
.004* * * (124.360)
.375* * * (137.145)
.390* * * (150.156)
.123* * * (18.248)
.440* * * (145.207)
.159* * * (25.258)
.127* * (8.383)
5.957* * * (181.647)
5.927* * * (180.083)
5.837* * * (173.992)
Firm-level controls
Subsidiary age
Subsidiary size
Host country experience
Industry membership
Manufacturing
Retail trade
Finance
Independent variables
Human capital intensity
Separability
Human capital
intensity Separability
Constant
.825* * * (135.691)
.342y (2.718)
.530* * * (27.297)
6.718* * * (237.429)
5.460* * * (147.940)
Model indices
% correctly classified
75.8
Log-likelihood
7472.492
Model c2
2036.435
Significance of model ( p)
< .001
75.9
7223.374
2285.553
< .001
75.9
7334.64
2174.281
< .001
75.8
7316.243
2192.684
< .001
75.7
7307.794
2201.133
< .001
Wald statistics in parentheses for logistic regressions.
** p < .01, two-tailed test.
*** p < .001, two-tailed test.
y
p < .10, two-tailed test.
sidiaries of similar size. As an example, in subsidiaries in the
category of 21 –50 employees in size, manufacturing subsidiaries employed a mean of 2.31 expatriates. Retail had a
Table 5
Expatriate employment
Manufacturing Retail Wholesale Finance Significance
(n = 7475)
(n = 320) (n = 4966) (n = 2102) ( p)
5.81
Expatriate
employment
(%)
Number of
4.30
expatriates
Number of
346.54
employees
8.36
19
27.39
< .001
3.90
3.89
2.95
< .001
225.30
64.68
48.43
< .001
The expatriate employment percentage row was not computed as the ratio
of the means from the two preceding employment categories reported in
Table 6. Instead, it is a mean calculated using each subsidiary’s expatriate
employment percentage. This method gives a greater mean for expatriate
employment, but the ranking across categories is the same whichever
method is used.
mean of 2.43, while wholesale trade and financial services
were greater at 3.57 and 5.43, respectively. Thus, the proportion of expatriate staff in foreign subsidiaries is greatest in
industries characterized by high levels of human capital
intensity and low levels of separability as suggested in
Hypotheses 4– 6.
We sought additional empirical support for our hypotheses using an ordinary least squares (OLS) model, with the
dependent variable (percent expatriate employees) as a
continuous variable (Table 6). Model 1 provides a baseline
estimation. Model 2 adds the industry variables. Models 3
and 4 estimate the direct impact of the human capital
intensity and separability variables. Model 5 tests the
interaction between these two variables. The addition of
the industry dummies significantly improved model fit, as
suggested by the significant F test ( p < .001). Consistent
with Hypothesis 4, entries in the manufacturing and retail
industries used proportionally fewer expatriates (negative
and significant coefficient on manufacturing and retail) than
wholesale trade subsidiaries ( p < .001). Entries in the fin-
43
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
Table 6
OLS regression: expatriate employment levels
Percentage of expatriate staff in subsidiary
Model 1
Model 2
Model 3
Model 4
Model 5
Country-level controls
Country risk
Culture distance
FDI openness
GDP
Population
.168* * * (7.035)
.044* * (3.088)
.175* * * (11.433)
.080* * * (4.032)
.030 ( 1.606)
Firm-level controls
Subsidiary age
Subsidiary size
Host country experience
.095* * * ( 7.027) .122* * * ( 9.746) .103* * * ( 7.790) .110* * * ( 8.466) .115* * * ( 8.957)
.104* * * ( 8.608) .060* * * ( 5.323) .076* * * ( 6.422) .080* * * ( 6.840) .069* * * ( 5.932)
.068* * * ( 5.443) .079* * * ( 6.894) .079* * * ( 6.491) .081* * * ( 6.719) .088* * * ( 7.362)
Industry-level control
R&D intensity
.093* * * ( 7.720) .049* * * ( 4.251) .104* * * ( 8.883) .099* * * ( 8.522) .090* * * ( 7.806)
Industry membership
Manufacturing
Retail trade
Finance
.067* * (2.976)
.052* * * (3.957)
.087* * * (6.053)
.106* * * (5.756)
.025 ( 1.418)
.089* * * (3.802)
.051* * * (3.754)
.113* * * (7.484)
.120* * * (6.209)
.023 ( 1.265)
.072* * * (3.121)
.051* * * (3.769)
.106* * * (7.104)
.133* * * (6.953)
.025 ( 1.422)
.253* * * ( 17.939)
.065* * * ( 5.630)
.232* * * (18.363)
Independent variables
Human capital intensity
Separability
Human capital intensity Separability
Model indices
Significance of model ( P)
F value
R2
.096* * * (4.058)
.052* * * (3.755)
.121* * * (7.982)
.115* * * (5.907)
.022 ( 1.201)
< .001
101.100
.131
.243* * * (18.218)
< .001
180.492
.264
< .001
129.156
.176
.263* * * (19.765)
.135* * * (11.514)
< .001
132.020
.194
.351* * * (23.293)
.195* * * (15.444)
.166* * * (11.967)
< .001
135.798
.212
The t statistics for linear regressions are in parentheses.
* * p < .01, two-tailed test.
* * * p < .001, two-tailed test.
ance sector were the most likely to use expatriates as a
proportion of total staff (positive and significant coefficient
on the finance dummy).
Models 3, 4, and 5 test the effect of the focal variables
on the proportion of expatriates. In Model 3, the coefficient for the human capital intensity variable was positively signed and significant ( p < .001). In Model 4, the
coefficients for the human capital intensity and separability variables were both positive and significant
( p < .001). In Model 5, the addition of the interaction
term between these two variables significantly improved
model fit from 19.4% to 21.2%. Its coefficient was
positively signed and significant ( p < .001). Thus, Hypotheses 4 – 6 are supported.
Importantly, these results involve a definition of expatriate employment that emphasized the intensity of expatriate
employment use, instead of the absolute number of expatriate employees in the subsidiary. We used the proportion of
expatriate employees because we expected the number of
expatriates employed would vary positively with the number of employees in the subsidiary. Hence, in our empirical
tests, we tried to identify sources of variation in the
proportion of expatriates.
An alternative perspective of expatriate employment
would suggest that expatriate employment levels are determined independent of the number of employees in a
subsidiary. This perspective would entail that most firms
face basic requirements for expatriate employment that cut
across different sized entries made in different industry
categories. Consequently, Model 5 was reestimated with
the absolute number of expatriate employees in the subsidiary as the dependent variable in two subsamples of subsidiaries (Table 6). One subsample comprised subsidiaries
above the mean level of subsidiary size, and the other
comprised subsidiaries at and below the mean level in size.
In the below-mean group, we observe that all three coefficients for the focal variables are positively signed and
significant at the 1% level. Thus, in smaller subsidiaries,
human capital intensity and separability have separate and
interactive influences on both the intensity of expatriate
employment use and the absolute number of expatriate
employees, which support Hypotheses 4 –6. However, in
44
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
the above-mean subsidiary size sample, that is large subsidiaries, only the coefficient for the separability variable is
positively signed and significant. In short, in large subsidiaries, Hypothesis 5 was supported, but not Hypotheses 4
and 6.
5. Discussion
Our findings offer a number of insights into the determinants of foreign market entry decisions. Consistent with
our hypotheses, the propensity to use wholly owned subsidiaries over joint ventures and the reliance on expatriate
managers increased with the levels of human capital intensity and customer interaction within an industry. The proportion of wholly owned subsidiaries and expatriate
managers was greater in financial services firms than in
wholesale trade, retail trade, and manufacturing firms,
supporting the notion of a ‘‘service continuum’’ spanning
across these four industries. One endpoint in this continuum
is in financial services in which three out of four entries
were made by wholly owned subsidiaries and, on average,
expatriate managers accounted for 27% of employment. In
contrast, at the other end of the service continuum, a
majority of manufacturing entries was made by joint ventures and the entries had a mean expatriate employment
percentage of about 5%.
A crucial assumption in our analyses was that the risk of
market failure is more important for the production and
delivery of transactions that are heavily dependent upon
employee specialized know-how, and not easily separable.
Under such conditions, adverse selection, free-riding, and
appropriability concerns should lead foreign investors to
establish control modes (a wholly owned subsidiary and a
high percentage of expatriate managers) that protect the rent
potential of an expansion. Even if appropriability is not an
issue, market failure in the transfer of a firm’s assets can
arise from differences in the capabilities of firms; that is,
varying capacities to integrate via joint-ventures and the use
of local staff a firm’s skills and knowledge into the functioning of its foreign operations (Kogut and Zander, 1993).
When capabilities rely on extensive levels of professional
skills and extensive contacts with customers, firms face
substantial ambiguity when attempting to transfer firmspecific advantages across organizational boundaries (Simonin, 1999). This contingency motivates firms to use wholly
owned subsidiaries and to deploy expatriates in foreign
expansions (Madhok, 1997).
Our results challenge a few conventional notions about
the way global service firms structure foreign subsidiaries.
It is often assumed that services operations are heavily
dependent on country-specific skills for the conduct of their
international operations (Carman and Langeard, 1980),
leading them to favor joint ventures and local managers
instead of wholly owned subsidiaries (Johansson, 1990).
While this assertion may apply to services that require low
human capital intensity (e.g., the retail industry), our results
clearly indicate that not all services are the same. Industry
characteristics influence how service firms enter international markets, and how staff is used in a foreign
subsidiary.
In terms of expatriate staffing strategies, Japanese companies have often been criticized for their ‘‘rice-paper
ceiling’’ policy, an artificial barrier to advancement for local
and third-country employees (Kopp, 1994b). We questioned
this perspective by examining for a contingent relationship
between the proportion of Japanese expatriate staff and two
key sector characteristics. Rather than adopting a systematic
policy against non-Japanese employees, Japanese MNCs
appear to adopt human resource practices that balance the
need for acquiring local skills with that of protecting firmspecific assets against the risk of imperfect transferability.
This conclusion must be made with some caution, however,
as base levels of expatriate employment exhibited some
consistency across industry segments (Table 5). Further, in
the case of larger subsidiaries, significant empirical evidence was observed for the intensity of expatriate employment, not from the absolute number of expatriate
employees. Further research relying on a survey instrument
could potentially provide stronger support for this claim.
5.1. Implications
One implication of this research relates to our understanding of the distinctive features of a service firms’
internationalization process. To be operational in foreign
markets, global service firms need relatively minimal physical commitments, as few investments in capital equipment
and operating facilities are usually required. However,
certain services industries are also faced with the challenge
of transferring to foreign subsidiaries the social assets,
skills, and capabilities that have been developed through
extensive education and training of employees, as well as
close contacts with end-customers. As transferring such
human capital entails substantial risks and difficulties, many
global service firms tend to rely on wholly owned subsidiaries and expatriate management staff when expanding
internationally.
A second implication relates to the stream of research on
entry mode. Most studies have looked at the entry mode
behavior of manufacturing firms using a transaction cost or
internalization framework (Anderson and Gatignon, 1986;
Hennart, 1991). These frameworks emphasize the specific
and proprietary characteristics of a firm’s assets as drivers
of its entry mode decisions. Typical operationalization of
these characteristics involves looking at expenditure patterns on R&D or advertising, that are applicable to manufacturing firms but less so to service firms. While not
completely different, our focus is unique since it emphasizes the characteristics of the human assets in a firm—
where they are concentrated in a firm’s value chain and the
degree of training embodied in a typical employee in a
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
firm. This focus can be applied more broadly than previous
approaches to the study of the behavior of manufacturing
and service firms. Extensions of this research would help to
resolve the impasse that exists between the integration of
research in manufacturing firms and those in the services
sectors.
A third implication relates to the practice of global
strategic management. Our study suggests that certain
‘‘service’’ characteristics, such as human capital intensity
and the inseparability of the production –consumption process create problems in the control and coordination of
international activities (Fryer, 1991; Reardon et al., 1996). A
remedy to these problems is the use of wholly owned entries
and the deployment of expatriate managers. This remedy
applies not only to services, but also to several manufacturing firms, which are increasingly incorporating service
features into their products.
5.2. Conclusion
The purpose of this paper was to examine how service
characteristics such as the degree of human capital intensity
and the separability of a firm’s business influenced entry
mode and expatriate staffing decisions. Drawing from a
sample of entries in Asia, Europe, and North America by
Japanese firms, we compared and contrasted the organizational mode established upon market entry, and the employment of expatriate managers in the manufacturing, retail
trade, wholesale trade, and financial services sectors. Our
results outlined the stark differences in entry mode selection
and expatriate staffing decisions across these four industry
categories. The propensity to use wholly owned subsidiaries
over joint ventures and expatriate managers was greatest in
financial services, where service characteristics are most
important, and lowest in manufacturing. Our findings also
supported the assertion that in situations where required
capabilities involve (1) close interactions with end customers and (2) extensive levels of professional skills, specialized know-how, and customization, wholly owned
subsidiaries, and expatriate staff were preferred. To the
extent that firms from all sectors, whether in manufacturing
or services, increasingly rely on assets that are dependent
upon human capital and not easily separable for delivering
their offerings in the marketplace, our analysis points to
ways in which this reliance can be incorporated into a
framework for understanding strategic choices, such as
foreign market entry decisions.
Acknowledgements
The authors thank Paul Beamish, Tony Frost, and
Philippe Véry for their comments on an earlier version of
this paper. They gracefully acknowledge the support of the
Asian Management Institute of the Richard Ivey School of
Business for access to the data used in this study. The initial
45
draft of the paper was written while Louis Hébert visited the
Richard Ivey School of Business.
References
Aharoni Y, Nachum L. Globalization of services: some implications for
theory and practice. New York: Routledge, 2000.
Aiken LS, West SG. Multiple regression: testing and interpreting regressions. London: Sage, 1991.
Alchian A, Demsetz H. Production, information costs and economic organizations. Am Econ Rev 1972;62:777 – 95.
Alchian A, Woodward S. The firm is dead; long live the firm. J Econ Lit
1988;116:65 – 79.
Anand J, Delios A. Location specificity and the transferability of downstream assets to foreign subsidiaries. J Int Bus Stud 1997;28:579 – 603.
Anderson E, Gatignon H. Modes of foreign entry: a transaction cost analysis and propositions. J Int Bus Stud 1986;17:1 – 26.
Baliga BR, Jaeger A. Multinational corporation: control systems and delegation issues. J Int Bus Stud 1984;15:25 – 40.
Black S, Gregersen H, Mendenhall M, Stroh L. Globalizing people through
international assignments. Reading (MA): Addison-Wesley, 1999.
Boddewyn J, Halbrich MB, Perry AC. Service multinationals: conceptualization, measurement and theory. J Int Bus Stud 1986;17:41 – 57.
Campbell AJ, Verbeke A. The globalization of service multinationals. Long
Range Plan 1994;27:95 – 102.
Carman JM, Langeard E. Growth strategies of service firms. Strategic
Manage J 1980;1:7 – 22.
Contractor FJ, Lorange P. Cooperative strategies in international business.
Lexington: Lexington Books, 1988.
Davis PS, Desai AB, Francis JD. Mode of international entry: an isomorphism perspective. J Int Bus Stud 2000;31:239 – 58.
Delios A, Beamish P. Ownership strategy of Japanese firms: transactional,
institutional and experience influences. Strategic Manage J 1999;20:
915 – 33.
Dierickx I, Cool K. Asset stock accumulation and sustainability of competitive advantage. Manage Sci 1989;35:1504 – 11.
Dunning JH. Multinational enterprises and the growth of services: some
conceptual and theoretical issues. Serv Ind J 1989;9:5 – 39.
Edstrom A, Galbraith JR. Transfer of managers as a coordination and
control strategy in multinational organizations. Adm Sci Q 1977;22:
248 – 61.
Erramilli MK. Influence of some external and internal environmental factors on foreign market entry mode choice in service firms. J Bus Res
1992;25:263 – 76.
Erramilli MK, Rao CP. Service firms’ international entry-mode choice: a
modified transaction-cost analysis approach. J Mark 1993;57:19 – 38.
Fryer C. What’s different about services marketing. J Serv Mark 1991;5:
53 – 8.
Gronroos C. Internationalization strategies for services. J Serv Mark 1999;
13:290 – 7.
Grosse R. International technology transfer in services. J Int Bus Stud 1996;
27:781 – 800.
Grubaugh SG. Determinants of foreign direct investment. Rev Econ Stat
1987;69:149 – 52.
Hennart JF. The transaction costs theory of joint ventures. An empirical
study of Japanese subsidiaries in the United States. Manage Sci 1991;
37:483 – 97.
Hill CW, Hwang P, Kim WC. An eclectic theory of the choice of international entry mode. Strategic Manage J 1990;11:117 – 28.
Hirsh S. International transactions involving interactions: a conceptual
framework combining goods and services. In: Giersch H, editor.
Services in world economic growth. Tubingen: J.C.B. Mohr, 1989.
p. 63 – 84.
Hymer SH. The international operations of national firms: a study of direct
investment. Cambridge: MIT Press, 1976.
46
C. Bouquet et al. / Journal of Business Research 57 (2004) 35–46
JETRO. The changing service industries of Japan. Tokyo: Japan External
Trade Organization, 1995.
Johanson J, Vahlne JE. The internationalization process of the firm: a model
of knowledge development and increasing foreign market commitments. J Int Bus Stud 1977;8:23 – 32.
Johansson JK. Japanese service industries and their overseas potential. Serv
Ind J 1990;10:85 – 109.
Katrishen FA, Scordis NA. Economies of scale in services: a study of
multinational insurers. J Int Bus Stud 1998;29:305 – 24.
Kim WC, Hwang P. Global strategy and multinationals’ entry mode choice.
J Int Bus Stud 1992;23:29 – 53.
Kobrin SJ. Trends in ownership of American manufacturing subsidiaries in
developing countries: an inter-industry analysis. Manage Int Rev 1988;
28:73 – 84.
Kogut B, Singh H. The effect of national culture on the choice of entry
mode. J Int Bus Stud 1988;19:411 – 32.
Kogut B, Zander U. Knowledge of the firm and the evolutionary theory of
the multinational corporation. J Int Bus Stud 1993;24:625 – 45.
Kojima K, Ozawa T. Japan’s general trading companies: merchants of
economic development. London: Croon Helm, 1984.
Kopp R. International human resource policies and practices in Japanese,
European, and United States multinationals. Hum Res Manage 1994a;
33:581 – 99.
Kopp R. The rice-paper ceiling: breaking through Japanese corporate culture. Berkeley: Stone Bridge Press, 1994b.
Li JT, Guisinger S. The globalization of service multinationals in the Triad
region: Japan, Western Europe and North America. J Int Bus Stud 1992;
23:675 – 96.
Lovelock CH, Yip GS. Developing global strategies for service businesses.
Calif Manage Rev 1996;38:64 – 86.
Loveman GW. The internationalization of services. Report Case No. 9-693103. Boston: Harvard Business School, 1997.
Madhok A. Cost, value and foreign market entry: the transaction and the
firm. Strategic Manage J 1997;18:39 – 61.
Makino S, Delios A. Local knowledge transfer and performance: implications for alliance formation in Asia. J Int Bus Stud 1996;27:905 – 28.
Mayrhofer W, Brewster C. In praise of ethnocentricity: expatriate policies
of European multinationals. Int Exec 1996;38:749 – 77.
Miller SR, Parkhe A. Patterns in the expansion of U.S. banks’ foreign
operations. J Int Bus Stud 1998;29:359 – 89.
Murray JY, Kotabe M. Sourcing strategies of U.S. service companies:
a modified transaction-cost analysis. Strategic Manage J 1999;20:
791 – 809.
Nigh D, Cho KR, Krishnan S. The role of location-related factors in U.S.
banking involvement abroad: an empirical examination. J Int Bus Stud
1986;17:59 – 72.
Nitsch D, Beamish P, Makino S. Entry mode and performance of Japanese
FDI in Western Europe. Manage Int Rev 1996;36:27 – 43.
Oxley JE. Appropriability hazards and governance in strategic alliances: A
transaction cost approach. J Law Econ Organ 1997;13:387 – 409.
Peterson RB, Sargent J, Napier NK, Shim WS. Corporate expatriate HRM
policies, internationalization, and performance in the world’s largest
MNCs. Manage Int Rev 1996;36:215 – 30.
Reardon J, Erramilli MK, Dsouza D. International expansion of service
firms: problems and strategies. J Prof Serv Mark 1996;15:31 – 46.
Roth K, Nigh D. The effectiveness of headquarters – subsidiary relationships: the role of coordination, control and conflict. J Bus Res 1992;25:
277 – 301.
Seiders K, Berry LL. Service fairness: what it is and why it matters. Acad
Manage Exec 1998;12:8 – 20.
Shan W. Environmental risks and joint-ventures sharing arrangements. J Int
Bus Stud 1991;22:555 – 78.
Simonin BL. Ambiguity and the process of knowledge transfer in strategic
alliances. Strategic Manage J 1999;20:595 – 623.
Sohn D. Social knowledge as a control system: a proposition and evidence.
J Int Bus Stud 1994;25:295 – 317.
Standard Industrial Classification. US Department of Labor Statistics, 1987.
Teece DJ. The multinational enterprise: market failure and market power
considerations. Sloan Manage Rev 1981;22:3 – 17.
Terpstra V, Yu CM. Determinants of foreign investment of U.S. advertising
agencies. J Int Bus Stud 1988;19:33 – 46.
Tung R. Selection and training procedures of U.S., European and Japanese
multinationals. Calif Manage Rev 1982;25:57 – 71.
Wheeler D, Mody A. International investment location decisions: the case
of U.S. firms. J Int Econ 1992;33:57 – 76.
World Competitiveness Report. Geneva: IMD International and World Economic Forum, 1996.
Yannopoulos GN. The growth of transnational banking. In: Casson M,
editor. The growth of international business. London: Allen & Unwin,
1983. p. 236 – 57.
Zeithaml V, Parasuraman L, Berry L. Problems and strategies in services
marketing. J Mark 1985;49:33 – 46.