a dual-role typology of multinational subsidiaries

A dual-role typology of multinational subsidiaries

Jue Wang a, Xiaming Liu b,*, Xiaoying Li b

a School of International Business, Southwestern University of Finance and Economics, Chengdu, Chinab School of Management, Birkbeck College, University of London, Malet Street, Bloomsbury, London WC1E 7HX, UK

1. Introduction

Since White and Poynter’s (1984) pioneering study, there has been an increasing interest in the roles, strategies orcharacteristics of multinational subsidiaries (Bartlett, 1986; Bartlett & Ghoshal, 1989; Birkinshaw & Morrison, 1995; D’Cruz,1986; Gupta & Govindarajan, 1991; Hogenbirk & van Kranenburg, 2006; Jarillo & Martinez, 1990; Roth & Morrison, 1992;Taggart, 1998; Vereecke, Van Dierdonck, & De Meyer, 2006). This great interest has resulted mainly from the recognition thatforeign subsidiaries can contribute substantially to the vibrancy of the multinational enterprise (MNE) at the local level(Bartlett & Ghoshal, 1989), and enable the MNE to enhance and sustain competitive advantage at the corporate level (Levitt,1983; Porter, 1980; Yip, 1992; Taggart, 1998).

Ferdows (1997) argues that since subsidiaries differ in the level of creation, sharing, and absorption of innovations, theymay play different roles in the knowledge network in the company and in local environments. Although existing research hasshed considerable light on the strategic roles subsidiaries play within the MNE, one shortcoming is its excessive focus on theinternal management to achieve corporate efficiency without adequate consideration of the impact of the interactionbetween the MNE and its local business environment. White and Poynter (1984) and Hogenbirk and van Kranenburg (2006)use a subsidiary’s product, market and value added scopes to classify its roles or strategies. Bartlett (1986), Bartlett andGhoshal (1989), Roth and Morrison (1992) and Taggart (1998) apply the integration-localization/responsiveness frameworkto differentiate subsidiary roles or strategies given the international business environment. While having made significantcontributions to our understanding of subsidiary roles within MNEs, these studies may pay too much attention to theinterests and challenges facing MNEs, and not enough to how MNEs help or hurt local economies, especially developing

International Business Review 18 (2009) 578–591

A R T I C L E I N F O

Article history:

Received 21 November 2008

Received in revised form 28 June 2009

Accepted 8 July 2009

Keywords:

Dual context

Multinational enterprises

Subsidiary typology

Two-way strategic relations

A B S T R A C T

This paper argues that, since a subsidiary is embedded in a dual context of both the MNE

and the host environment, its strategic role should be assessed by its relative positions and

contributions both within the knowledge networks of the MNE and the host country.

Based on this, we develop a dual-role typology. The 369 multinational subsidiaries in our

sample from China can be classified into as many as 12 out of the 16 conceptual groups of

the typology. Our results indicate that dual activists (active both internally and externally)

account for only 12% of the total sampled multinational subsidiaries while dual loners

(inactive both internally and externally) reach 20%. The results from a larger sample by

adding 113 minority foreign share firms show that external knowledge links are positively

associated with local Chinese ownership. The central message from this paper is that a

large proportion of foreign-invested firms in China are inactive in knowledge exchange

either internally or externally or both. Managerial and policy implications are discussed.

� 2009 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: +44 0 20 7079 0895; fax: +44 0 20 7631 6769.

E-mail addresses: [email protected] (J. Wang), [email protected] (X. Liu), [email protected] (X. Li).

Contents lists available at ScienceDirect

International Business Review

journa l homepage: www.e lsev ier .com/ locate / ibusrev

0969-5931/$ – see front matter � 2009 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ibusrev.2009.07.002

mailto:[email protected]



http://www.sciencedirect.com/science/journal/09695931

http://dx.doi.org/10.1016/j.ibusrev.2009.07.002

country economies (Meyer, 2004; Ramamurti, 2004). Given that globalisation has significantly transformed the context inwhich MNEs operate, an investigation of the role of MNEs in economic development needs to be a central concern in themainstream international business research (Buckley & Ghauri, 2004; Ghauri & Yamin, 2009). Gupta and Govindarajan’s(1991) study focuses on knowledge flows within the MNE to identify different roles of subsidiaries, but does not analyseknowledge exchange between a subsidiary and its local environment.

The current paper proposes a dual-role typology of multinational subsidiaries. This typology is based on the networkapproach to MNEs (Ghoshal & Bartlett, 1988; Harzing, 1999; Hedlund, 1986; O’Donnell, 2000). While this approach hasrecently been adopted to explain knowledge creation and diffusion of MNEs (Andersson, Forsgren, & Pedersen, 2001;Andersson, Bjorkman, & Forsgren, 2005), it has not been fully applied to subsidiary typology. Following this approach, anMNE is a network of internationally dispersed units which are simultaneously embedded in two business contexts: theinternal networks of the MNE and the external (host country) environment. An MNE may not only adjust to or overcome thedifferences it encounters at the borders of its various markets, but also exploit these differences for value creationopportunities so that value chains may increasingly span multiple countries (Ghemawat, 2007). In this process of the MNE’sglobal expansion, the roles and/or strategies of its subsidiaries are both shaped by and affect the internal and externalnetworks. A subsidiary’s strategic role should be assessed by its relative positions and contributions within the knowledgenetworks of both the MNE and the host country.

This typology extends the study of Gupta and Govindarajan (1991) by considering external as well as internal knowledgeflows, and is believed to be useful not only for MNE managers but also for policy makers, especially from host countries. Therest of the paper is organized in the following way. Section 2 reviews the literature on subsidiary roles and strategies, and inSection 3 we present our dual-role typology. Section 4 explains the research methodology. In Section 5, we test the typologyon a survey data set of 369 multinational subsidiaries in China and compare the results with those when 113 firms with atleast 25% foreign ownership are added to the data set. Finally, Section 6 concludes by summarizing the paper and discussingmanagerial and policy implications and limitations of this research.

2. Literature review

The traditional view on foreign direct investment (FDI) (Buckley & Casson, 1976, 2002; Casson, 1994; Hymer, 1976)assumes a hierarchical relationship between the headquarter (HQ) of an MNE and its subsidiaries. Typically, the MNE createstechnological knowledge at the HQ and then diffuses it to subsidiaries worldwide (Almeida & Phene, 2004) while expectingthem to perform in line with the overall corporate strategy. However, as Casson, Pearce, and Singh (1991) and Buckley andGhauri (2004) note, since the early 1980s there has been a shift away from such a hierarchical relationship. During thisprocess, the favoured form of firm has become a federal structure of operating divisions drawing on a common resource ofinternational expertise, while each division belonging to the federation is able to outsource expertise from its partners(Buckley, 2004; Buckley & Ghauri, 2004). In other words, subsidiaries have seen their roles change over the past two decadesor so.

With regard to roles in the MNE, a subsidiary does not just receive and apply knowledge, but in many cases createsknowledge in its local (host country) environment and then diffuses it back to the HQ and the rest of the MNE (Almeida &Phene, 2004). This increases efficiency at both the subsidiary and corporate (MNE) levels. On the other hand, since it islocated in a host country, a subsidiary may contribute to the development of the local economy through transferringadvanced knowledge, engaging local employees and contributing to local tax. In this process, the subsidiary may also benefitfrom the local environment through learning indigenous knowledge including indigenous technology and local informationin order to increase profits, widen the global market share and enhance its reputation.

A very important contribution by White and Poynter (1984) is their recognition of different business strategies pursuedby different subsidiaries as these subsidiaries are themselves affected differently by environmental changes. If we describedall subsidiaries as branch plants, then this variation could not be captured. Hedlund’s (1986) concept of heterarchy(contrasting with hierarchy) is perfectly consistent with this idea, which implies that subsidiary roles are not necessarilyassigned by HQs. Actually, the concept of heterarchy has already proved a stimulating one in the field of multinationalstrategy and structure (Young & Tavares, 2004).

So far three main types of typology of subsidiary roles/strategies have been developed in the literature: the Product-Market-Value Added Framework, the Integration-Localisation Framework and the Knowledge Flows-Based framework. Areview of these frameworks is useful for the development of our own dual-role typology of multinational subsidiary roles orstrategies.

2.1. Typology 1: Product-market-value added framework

Based on their case studies, White and Poynter’s (1984) pioneering work divides business strategies pursued bymultinational subsidiaries into five categories, defined by the activities of subsidiaries with regard to the product, marketand value added scopes. Depending on the different combinations of these three dimensions, five types of strategies can beidentified. The first is as a miniature replica business which produces and markets some of the parent’s product lines orrelated product lines in the local country. Depending on the degree of product and marketing modifications performed by thesubsidiary, there can be three sub-strategies: adopter, adapter and innovator. The second is as a marketing satellite business

J. Wang et al. / International Business Review 18 (2009) 578–591 579

which markets products manufactured centrally into the local trading area. The third is as a rationalised manufacturer thatproduces a designated set of component parts or products for a multi-country or global market. The fourth is as a productspecialist that develops, produces, and markets a limited product line for global markets. Finally, the fifth is as a strategicindependent that has the freedom and resources to develop lines of business for either a local, multi-country or globalmarket.

Hogenbirk and van Kranenburg (2006) adopt a very similar approach to White and Poynter’s (1984), i.e. using theframework of market scope and value added scope to analyse strategic roles of subsidiaries. This typology is useful for us toidentify the particular role a subsidiary plays in the R&D, production and marketing systems within the MNE. However, itdose not discuss any interactions between the subsidiary and the local environment, although this approach suggests thatenvironmental changes affect the subsidiary.

2.2. Typology 2: Integration-localisation framework

This framework has been most widely used to classify the roles and/or strategies of subsidiaries. Porter (1986) uses aconfiguration-coordination framework at the industry level to categorise firm strategies. Bartlett (1986) develops acoordination-integration framework and group firms into three types: global (high global coordination/integration but lownational responsiveness/differentiation), transnational (high global coordination/integration and high national respon-siveness/differentiation) and multinational (low global coordination/integration but high national responsiveness/differentiation). Then, Bartlett and Ghoshal (1989) use the dimensions of degree of integration and degree of localizationto divide multinational subsidiaries into three groups. (1) An active subsidiary with a high degree of integration and a highdegree of localization. (2) A receptive subsidiary with a high degree of integration and a low degree of localization. (3) Anautonomous subsidiary with a low degree of integration but a high level of localization. In Bartlett (1986) and Bartlett andGhoshal (1989), the bottom left corner of the integration-localisation framework is empty, omitting a possible type ofsubsidiary with a low degree of both integration and localisation. Taggart (1998) extends Bartlett and Ghoshal’s Integration/localization framework by identifying a fourth type of subsidiary, i.e. an acquiescent subsidiary with a low level ofintegration and localization.

The integration/localization framework is very useful in explaining the efficiency and competitiveness of MNEs as itconsiders both the internal co-ordination and external response to local demands. However, as mentioned in the precedingsection, the framework may pay too much attention to the interests and challenges facing MNEs, and not enough to howMNEs interact with local economies, especially developing country economies. While responding to the local businessenvironment, the integration-localization framework may not take into consideration that subsidiary strategies can both beinfluenced by and affect the host country environment.

2.3. Typology 3: Knowledge flows-based framework

Gupta and Govindarajan (1991) treat an MNE as a network of capital, product and knowledge transactions among unitslocated in different countries, and believe that this perspective is consistent with the analyses of Porter (1986), Bartlett(1986), and Bartlett and Ghoshal (1989). Knowledge transactions are the focus of Gupta and Govindarajan’s (1991) studywhere knowledge flow between a subsidiary and the rest of the MNE is defined as the transfer of either expertise (such asskills and capabilities) or external market data of strategic value (such as key customers, competitors and suppliers).Subsidiaries are different in knowledge flow patterns. Based on this, Gupta and Govindarajan (1991) classify theroles/strategies of subsidiaries into 4 categories in terms of the degree of two-way knowledge flows between a subsidiaryand the MNE, namely those of global innovator (high outflow, low inflow); integrated player (high outflow, high inflow);implementor (low outflow, high inflow); and local innovator (low outflow, low inflow).

According to Gupta and Govindarajan (1991), a subsidiary plays a global innovator role when it serves as thefountainhead of knowledge for other units. An integrated player has a similar role to a global innovator, but is not self-sufficient in its knowledge needs. An implementor creates little knowledge but relies on knowledge flows from the rest of theMNE. Finally, a local innovator creates knowledge especially for the host-country market. Based on a sample of 169multinational subsidiaries Harzing and Noorderhaven (2006) test this typology and find more differences betweensubsidiaries than Gupta and Govindarajan’s (1994) own study. Using the concept of internal knowledge flows, Vereecke et al.(2006) propose a typology of plants in the international manufacturing network of multinational companies, identifying fourtypes of plants with different network roles: the isolated plants, the receivers, the hosting network players, and the activenetwork players. The typology based on a network of knowledge flows within the MNE is very useful for us to understand thecompetitiveness of the MNE in the global and host-country-specific markets. However, this particular typology onlyconsiders a two-way internal exchange of ideas between a subsidiary and the rest of the MNE, omitting possible knowledgeflows between the subsidiary and its local environment.

3. Dual-role typology of multinational subsidiaries

Over the last two decades, research on the structure and organization of multinationals has shifted its focus from one-to-one HQ-subsidiary relationships to that of managing a network of units (Kogut, 1989). Ghoshal and Bartlett (1988, p. 620)

J. Wang et al. / International Business Review 18 (2009) 578–591580

claim that the network approach ‘‘is particularly suited for the investigation of such differences in internal roles, relations,and tasks of different affiliated units and of how internal coordination mechanisms might be differentiated to match thevariety of sub-unit contexts.’’ The network approach argues that a multinational subsidiary is simultaneously embedded intwo business contexts: the internal MNE and the external (host country) environment. Thus, the roles of subsidiaries areboth shaped by and affect these two contexts. A subsidiary has a dual role to play: it interacts with the rest of the MNEinternally and with the local environment externally. These interactions allow subsidiaries to obtain access to resourcesfrom different sources and affect their internal and external business partners. In fact, differences in such interactions createdifferences in their level of competence, which in turn create differences in the roles the subsidiaries can play (Anderssonet al., 2001).

Among all resources potentially useful to a firm, knowledge may be the most important for the firm to create andmaintain competitiveness in the market. Indeed, it is widely acknowledged in strategic thinking that the ability of a firm todevelop and exploit knowledge faster than its competitors is a key component of its competitive advantage (Cooke & Beh,2007; Leonard-Barton, 1995; Nonaka & Takeuchi, 1995; Nonaka & Teece, 2001; Porter, 1980; Prahalad & Hamel, 1990; Storey& Salaman, 2005; Teece, Pisano, & Shuen, 1997). Like Gupta and Govindarajan (1991), our dual-role typology focuses onknowledge flows. In this sense, our typology is also a knowledge-based one. However, we consider not only internal, but alsoexternal two-way knowledge flows.

In this typology, different interest groups such as a multinational subsidiary, its corporate headquarters, local firms andthe host-country government are involved in the subsidiary’s internal and external knowledge linkages. They may havedifferent views on the roles and strategies of a subsidiary. A subsidiary may wish to learn more from the HQ and the localenvironment to enhance its efficiency. The MNE (HQ) may require subsidiaries to not only learn and implement know-howfrom the HQ to increase its productivity, but also create and transfer knowledge back to the HQ and the rest of the MNE inorder to enhance efficiency at the corporate level. The local firms may and the host country government may pay moreattention to how much the firm and the country can learn and benefit from the foreign subsidiary as opposed to merely theMNE’s overall efficiency. Thus, this typology allows the roles of a multinational subsidiary to be assessed not only internallyin terms of the overall strategy of the MNE, but also externally in terms of its interactions with the local environment. Thisframework can be presented in Fig. 1 below.

The top reference frame demonstrates internal interactions. Firstly, if a subsidiary actively diffuses its knowledge to, andat the same time energetically learns from, the rest of the MNE, it can be referred to as an ‘‘internal activist’’. It createsknowledge not only for its own use but also for the rest of the MNE. However, since it is impossible to create all knowledgeneeded for its operations, this subsidiary seeks knowledge from the rest of the MNE. It is characterized by high knowledgeflows in both directions. In terms of the efficiency at the corporate level, this type of subsidiary is ideal as all knowledgewithin the MNE can be effectively utilized. Secondly, if a subsidiary proactively learns from, but is less proactive to diffuse itsknowledge to, the rest of the MNE, it can be called an ‘‘internal receptor’’. It is characterized by a combination of highknowledge inflows but low knowledge outflows. Although active learning is good for this subsidiary’s own operations, therest of the MNE may be unable to substantially benefit from this subsidiary. Thirdly, if a subsidiary diffuses more knowledgeto the rest of the MNE than it learns from them, then it can be referred to as an ‘‘internal contributor’’. Compared with an

Fig. 1. Dual-role typology of multinational subsidiaries framework.


internal activist, an internal contributor diffuses relatively more knowledge than it adopts. This type of knowledge flows ismore beneficial to the rest of the MNE than to the subsidiary itself. Finally, if a subsidiary is inactive in both knowledgelearning and knowledge diffusion within the MNE, it is an ‘‘internal loner’’. Obviously, this type of subsidiary is in isolationand little benefit can be derived for the rest of the MNE and the subsidiary itself

The bottom reference frame of Fig. 1 shows the external interactions of a subsidiary. This is an extension of Gupta andGovindarajan (1991) from the internal linkages only to both internal and external linkages. Parallel to the interpretation ofthe top reference frame, firstly, if a subsidiary is proactive in both diffusing knowledge to, and learning knowledge from, itslocal partners in the host country, it can be regarded as an ‘‘external activist’’ as indicated in the up-right corner of the bottomframe. It follows that the subsidiaries in the remaining three quadrants of this frame can be referred to as an ‘‘externalreceptor’’, ‘‘external contributor’’ and ‘‘external loner’’ respectively.

The incorporation of external interactions into a multinational subsidiary typology is fully justified by the networkapproach to FDI. MNEs now regard host countries not just as their markets or sources of cheap labour, but increasingly aspotential sources of new knowledge (Dunning, 1994). Industry-specific knowledge developed in geographicallyconcentrated locations can flow to subsidiaries located there via inter-firm (and interpersonal) linkages in the region,and this can be a positive source of value creation for MNEs (Almeida & Phene, 2004).

On the other hand, the literature on FDI spillovers suggests that the most important reason why countries try to attractFDI is perhaps the prospect of acquiring modern technology, interpreted broadly to include product, process, anddistribution technology, as well as management and marketing skills (e.g. Blomstrom & Kokko, 1998; Caves, 1974). FDI is apackage of capital, technology and managerial skills, and has been viewed as an important source of both direct capitalinputs and technology and knowledge spillovers. Balasubramanyam, Salisu, and Sapsford (1996) argue that developingcountries can significantly benefit from FDI because it not only transfers production know-how and managerial skills butalso produces externalities or spillover effects. In the empirical literature, positive spillover effects are detected in thestudies of Kokko, Tansini, and Zejan (1996) on the Uruguayan manufacturing sector, Liu, Siler, Wang, and Wei (2000) on UKmanufacturing, and Li, Liu, and Parker (2001), Wei and Liu (2006), Wei, Liu, and Wang (2008) on Chinese manufacturing.However, FDI can sometimes produce insignificant or even negative spillover effects due to the lack of incentives for MNEsto diffuse their knowledge (Aitken & Harrison, 1999; Javorcik, 2004), low absorptive capabilities of local firms(Chudnovsky, Lopez, & Rossi, 2008), or MNEs’ local market stealing effect described by Aitken and Harrison (1999). Forinstance, Hu and Jefferson (2002) have detected negative horizontal spillover effects of FDI in the Chinese electronicsindustry. As the recent surveys by Blomstrom and Kokko (1998), Gorg and Strobl (2001), Gorg and Greenaway (2004) andSmeets (2008) conclude, the empirical findings are still inconclusive on the magnitude, direction, and even existence ofknowledge spillovers from FDI.

An external activist contributes substantial knowledge to, but also absorbs much knowledge from, the host country. Anexternal contributor involves high knowledge outflows to the local environment but relatively low knowledge inflows toitself. In both cases, such multinational subsidiaries help develop the local economy, and are in the best interest of the hostcountry. On the other hand, the diffusion of knowledge to local firms may negatively affect the competitive position of theMNE in this host country, and hence the MNE will have a strong incentive to prevent such knowledge flows. An externalreceptor is associated with relatively low knowledge outflows to the local environment but high knowledge inflows to thesubsidiary. In this sense, the subsidiary is sourcing knowledge from the host country, and this is beneficial to the subsidiary’sown operations and probably to the competitiveness enhancement of the MNE if the subsidiary is also an internal activist orcontributor. Finally, an external loner involves little knowledge exchange with the local environment, and plays little role inpromoting the competitiveness of the MNE and local development.

This whole framework links together the internal and external interactions (or embeddedness), and hence allows 16different combinations between these two types of interactions. For instance, a subsidiary can be very active in two-wayknowledge flows both internally and externally (a combination of internal and external activist, or dual activist), as indicatedby the linkage of the upper-right-hand quadrants of the top and bottom reference frames in Fig. 1 (combination 1), orinactive in both types of interactions (a combination of internal and external loner, or dual loner). A dual activist absorbs anddiffuses knowledge both internally and externally, and may increase overall efficiency of global knowledge flow networks,although this may have different implications for the competitiveness of the host economy and the MNE. On the other hand,a dual loner contributes little to the development of such knowledge flow networks. Similarly, a subsidiary can be activeinternally but inactive externally (a combination of internal activist/contributor with external loner/receptor), as indicatedby the linkage of the upper-right-hand quadrant of the top reference frame and the lower-left quadrant of the bottomreference frame in Fig. 1 (combination 4), or vice versa (a combination of internal loner/receptor with external activist/contributor).

The different combinations of the internal and external interactions are expected to be associated with different levels ofperformance. This is consistent with the widely accepted view in strategy and organisation research that a firm’s contextinfluences its behavior and performance (Almeida & Phene, 2004). Monteiro, Arvidsson, and Birkinshaw (2008) state thathigher frequency of knowledge flow is associated with higher subsidiary performance. In our proposed framework, anactivist energetically creates and diffuses knowledge while proactively learning from other business units while a loner is infact in a disadvantageous position within the MNE since it is unable to take advantage of the knowledge developed by otherunits. Thus, an activist should perform better than a loner. In this study, we will have a brief check of whether differentsubsidiary roles lead to different subsidiary performances.


As discussed in section III, the different combinations of the external interactions are expected to be associated withdifferent levels of the subsidiary’s impact on local firms .In our proposed framework, if an external activist vigorously seeksknowledge from the local environment while diffusing its own knowledge to indigenous firms in the host country, it shouldhave a greater positive impact on local firms than an external loner.

4. Research methodology

4.1. Sample

Probability sampling was adopted in this study. Specifically, a sample of 1223 firms was drawn from the lists of foreigninvested firms compiled by the local authorities in Beijing, Chongqing and Jiangsu Province in China using the systematicsampling method discussed in Ghauri and Gronhaug (2002, p.115). A draft research instrument was pre-tested via personalinterview with chief executives or other senior managers of 14 foreign invested firms. The questionnaire was then modifiedand finalized. This pre-test also allowed us to obtain insights into the types of multinational subsidiaries in China, andprovide an assessment of the questions’ validity and the likely reliability of the data that would be collected (Saunders, Lewis,& Thornhill, 2003).

As argued by Taggart (1998), a postal questionnaire was thought to be the appropriate method of data collection forstudying subsidiary roles or strategies because of resource constraints and generalizability of results. Following this researchstrategy, the questionnaire was sent to 1223 foreign invested firms and 493 of them responded (40.3% of total). Among therespondents, 205 (41.6%) were the founders or chief executive officers, 188 (38.1%) were chief financial officers and the rest(20.3%) were senior human resource managers. The sampled firms could be grouped into 30 industrial sectors, includingmotor, steamboat, airplane, computer as technology-intensive industries, and food, beverage, clothing and shoes as labour-intensive industries. In this whole sample, 369 firms have at least 50% foreign ownership and can be regarded asmultinational subsidiaries (as defined by Jarillo & Martinez, 1990). We first analyzed the roles or characteristics of these 369subsidiaries, and then expanded the sample size by adding 113 firms with at least 25% foreign ownership for comparison.25% is the minimum level of foreign ownership with which a firm can be regarded as a foreign invested one in China (Item 4of the Law of Chinese and Foreign Joint Ventures of the People’s Republic of China, Revised version, 2001).

Foreign invested firms located in Beijing, Chongqing and Jiangsu Province were chosen for two reasons. Firstly, because ofresource constraints, only limited locations could be included. Secondly, these three locations may represent different levelsof development in China. Between 1997 when Chongqing City was separated from Sichuan Province as a municipalitydirectly under the jurisdiction of the central government) and 2007, the average growth rates for Beijing, Jiangsu andChongqing were 12.1, 11.0 and 10.7 percent respectively. The average annual FDI inflows for these regions in the same periodwere US$ 262.2, 980.4 and 44.4 million respectively (China Statistical Yearbook and Chongqing City Statistical Yearbook,various issues). It is clear that there was a significant difference in terms of FDI inflows in these regions, although the growthdivergence was not that significant. Beijing is the capital and one of the commercial centres of China. It has attracted muchR&D FDI due to its large pool of entrepreneurial and motivated engineers and scientists, and advanced regional innovationsystem (Chen, 2008). Jiangsu Province is a highly developed industrial and commercial region in China, and it is notsurprising that it is one of the most popular destinations of inward FDI within China. Chongqing is the commercial andtransportation center of Western China, but it is less developed in terms of human capital and infrastructure, and hasattracted much less inward FDI than Beijing and Jiangsu.

4.2. Measures

As mentioned earlier, the network approach suggests that knowledge flows result from a subsidiary’s interactions withthe rest of the MNE internally and with the local environment externally. In this sense, knowledge flows may be measured bythe benefits from the interactions or embeddedness of a subsidiary within the global business networks.

To study knowledge creation and competence development, Andersson, Forsgren, and Holm (2002) and Andersson et al.(2005) use the extent to which a subsidiary’s most important local business relationships have caused its adaptation of producttechnology, production technology, standard operating procedures and business practices to measure the subsidiary’s localembeddedness. Instead, we use the answer to the question ‘‘how helpful is the interaction with your local partners for yourproduction, management, R&D and marketing practices’’ to measure knowledge inflows from local business partners to thesubsidiary. We measure external knowledge outflows and internal knowledge flows in a similar way (see Table 1). A 5-pointLikert-type scale was employed as follows: 1: Very unhelpful; 2: Unhelpful; 3: Neutral; 4: Helpful; 5: Very helpful.

To measure a subsidiary’s performance, three objective variables, including return on assets (ROA), return on equity(ROE), and return on investment (ROI) and a subjective variable were employed. In the existing literature, ROA, ROE, ROI arewidely used to measure firm performance (Tanriverd & Venkatraman, 2005). Based on Jaw, Yu, Wang, and Chen (2006), thesubjective variable was used to measure overall subsidiary performance in order to obtain a comprehensive result. In thispaper, variable OVERALL (indicating ‘‘how satisfactory the subsidiary’s overall performance is’’) was used as a subjectivevariable.

The measurement of subsidiary strategic impact on local firms was operationalized by four variables drawn on andmodified from the same set of existing studies mentioned above.


All the subjective variables on subsidiary performance and subsidiary strategic impact on local firms were measured on a5-point Likert-type scale: 1: Very disappointing; 2: Disappointing; 3: Neutral; 4: Satisfactory; 5: Highly satisfactory.

Some widely accepted rules or practices for questionnaire design are that questions need to be as easy andstraightforward as possible (Fulcher & Scott, 2003 p. 81), and that the objective of question wording is to communicate withrespondents in their own language as far as possible (Moser & Kalton, 1983, p. 320). Put another way, simple words ratherthan jargon and technical terms should be used (Oppenheim, 1996, p. 129). For these reasons, some questions posed tomanagers in our questionnaire were simplified in comparison to for example Andersson et al. (2002). Of course, too simplequestions may raise danger of providing clues to the answers being sought (Bilton et al., 2002, p. 454) and this may affect thereliability and validity of the final questionnaire. As can be seen, our questions did not contain any clues nor were they‘‘leading’’. Furthermore, as discussed below, the pilot study showed that our questionnaire was fine.

4.3. Data analysis

The statistical package used in this research is SPSS (Pallant, 2004). The survey data were analyzed using the followingsteps. Principal Components Factor (PCF) Analysis was first applied to identify the dimensions of the internal and externalinteractions. Then clustering analysis was carried out in order to test whether the sampled firms could be grouped into the 16clusters. As is well known, cluster analysis is widely used to examine the interdependent relationships among a range ofvariables. However, existing studies suggest that the use of clustering techniques in management studies has been often lessideal and that cluster analysis methods must be chosen carefully (e.g., Ketchen & Shook, 1996). Common errors include theuse of only one clustering technique, where the recommendation is to use an agglomerative technique in conjunction with adivisive technique to check the consistency of the cluster obtained. Our research uses a two-step clustering approach, whichreduces the potential biases introduced by employing a single method (Ketchen & Shook, 1996). The two-step clusteringcombines hierarchical and k-means techniques and has been used in management studies (e.g., Ferguson, Deephouse, &Ferguson, 2000; Short, Ketchen, Palmer, & Hult, 2007).

5. Results and discussions

5.1. Case studies

The interviews with the 14 multinational subsidiaries in our pre-test of questionnaire confirmed a number of differenttypes of multinational subsidiary as predicted by our dual-role typology framework. For instance, in the extreme cases, theCEO of firm CQ6 (a US motor accessory subsidiary) believed that his subsidiary had close two-way knowledge flows bothwithin the corporate and with the local Chinese partners, and hence was active both internally and externally (dual activist).On the other hand, the CFO of firm CQ30 (a South Korean motorbike accessory subsidiary) described his subsidiary as a

Table 1

Description of the variables.

Variable code Variable description

Performance variables ROA Return on assets

ROE Return on equity

ROI Return on investment

OVERALL Satisfaction of subsidiary overall performance

Strategic impact on local firms IMPACTONLP Subsidiary’s impact on local productions

IMPACTONLM Subsidiary’s impact on local management

IMPACTONLRD Subsidiary’s impact on local R&D

IMPACTONLMARKET Subsidiary’s impact on local marketing

Internal interactions A1 Production help from HQ and rest of MNE

A2 Management help from HQ and rest of MNE

A3 R&D help from HQ and rest of MNE

A4 Marketing help from HQ and rest of MNE

A5 Production help to HQ and rest of MNE

A6 Management help to HQ and rest of MNE

A7 R&D help to HQ and rest of MNE

A8 Marketing help to HQ and rest of MNE

External interactions B1 Production help from local firms

B2 Management help from local firms

B3 R&D help from local firms

B4 Marketing help from local firms

B5 Production help to local firms

B6 Management help to local firms

B7 R&D help to local firms

B8 Marketing help to local firms


dual-loner. Due to poor performance in recent years, CQ30 planned to close down soon. The interviews also revealed thatfirms CQ1, CQ2, CQ6, BJ1, BJ2, and JS1 interacted with local Chinese partners by introducing new business opportunities toeach other due to relational capital. There has been little discussion of this form of knowledge flows between foreign andlocal firms in the literature. Accordingly, this new channel was incorporated into our final questionnaire in order todetermine whether this is a prevailing practice in the interaction between multinational subsidiaries and local Chinesefirms.

A reliability test was conducted on the sample of firms interviewed. The Cronbach Alpha coefficient is 0.844: if an item isdeleted, all the values of the 16 items are larger than 0.70. Therefore, it can be concluded that the scale used for this researchis considered reliable (Pallant, 2004, p. 87). In terms of validity, all the questions are clear and easy to answer.

5.2. Postal questionnaires

In analyzing returned questionnaires, we first examined the sample of 369 subsidiaries. We expect that the 8 variables,which determine the internal interactions (A1 to A8) of a subsidiary, can be reasonably reduced to two key dimensions, i.e.internal interactions from the HQ and rest of the MNE to the subsidiary, and internal interactions from the subsidiary to theHQ and the rest of the MNE. Likewise, the 8 variables (B1 to B8) that determine the external interactions of the subsidiary canbe reasonably reduced to two key dimensions: external interactions from local firms to the subsidiary, and externalinteractions from subsidiary to local firms. The results are presented in Table 2.

Moreover, since the measures of the variables were collected using the same survey instrument, the possibility ofcommon method bias (CMB) was tested using Harman’s one factor test as advised by Podsakoff and Organ (1986). Anunrotated principal components factor analysis on the 16 measurement items yielded four factors with eigenvalues greaterthan 1.0. As several factors instead of one single factor were identified and all of them accounted for just 74% of the totalvariance, and as the first factor accounted for only 35% of the variance, a substantial amount of common method variancedoes not seem to be present (Podsakoff & Organ, 1986).

The results of PCF in Table 2 show that for internal interactions, the Kaiser-Meyer-Olkin Measure of Sampling Adequacyvalue is 0.758, which is more than 0.6, and the Sig. value of Bartlett’s Test of Sphericity is less than 0.05. Therefore, the factoranalysis is valid (Pallant, 2004). In Table 2, it can be seen that the 8 variables can be perfectly divided into 2 factors and theloadings of each of the variables on the two factors are all greater than 0.6. Factor 1 (IISTOHQ) includes A5 to A8 which showinternal interactions from the subsidiary to the HQ and the rest of the MNE. Factor 2 (IIHQTOS) combines A1 to A4, indicatinginternal interactions from the HQ and the rest of the MNE to the subsidiary. In our three-dimensional typology picture, Factor1 is represented by the top frame X axis and Factor 2 by the top frame Y axis.

The application of PCF to external interaction is also valid. Also from Table 2, the Kaiser-Meyer-Olkin Measure of SamplingAdequacy value is 0.885, and the Sig. value of Bartlett’s Test of Sphericity is less than 0.05. It can also be seen that the 8variables are divided into 2 factors and the loadings of each of the variables in the two factors are more than 0.7. To sum up,Factor 3 (EISTOL) includes B5 to B8 which indicate external interactions from the subsidiary to local firms. Factor 4 (EILTOS)combines B1 to B4 and indicates external interactions from local firms to the subsidiary. In our three-dimensional typologypicture, Factor 3 is represented by the bottom X axis, and Factor 4 is represented by the bottom Y axis.

Table 2

Factor loading of the four-dimension solution.

Variables Varimax factor loadings

Factor 1 Factor 2 Factor 3 Factor 4

A1 .876

A2 .810

A3 .787

A4 .695

A5 .905

A6 .890

A7 .839

A8 .850

B1 0.816.

B2 0.823

B3 0.864

B4 0.749

B5 0.872

B6 0.843

B7 0.792

B8 0.795

Variance .316 .382 .357 .342

Note: Loading value below 0.4 is not shown in this table. KMO (Internal) = 0.758;

KMO (External) = 0.885.


5.3. Multinational subsidiary clusters

We now proceed to conduct cluster analysis based on the four factors drawn from the above PCF examinations. Theresults are shown in Tables 3 and 4. Firstly, according to the Summary provided in Table 4, it is clear that not all of the 16conceptual groups have relevant subsidiaries to fix into. Specifically, we are unable to classify any multinational subsidiariesfrom our sample as belonging to any of the following clusters: (I) Internal activist and external contributor; (II) Internalactivist and external receptor; (III) Internal and external receptor; and (IV) Internal loner and external activist. However, all

Table 3

Results of cluster analysis.

Cluster Factor Mean Groups combine No. of cases Performance/impacts on locals Mean

1 Factor 1 1.95 Internal activist 18 Performance factor 0.84

Factor 2 0.92 Impact on local factor 0.79

Factor 3 1.41 External activist

Factor 4 0.26

2 Factor 1 �0.30 Internal receptor 17 Performance factor 0.21

Factor 2 1.15 Impact on local factor �0.20

Factor 3 0.87 External contributor

Factor 4 �1.28

3 Factor 1 0.42 Internal contributor 14 Performance factor �0.17

Factor 2 �0.84 Impact on local factor 0.28


Factor 4 �1.24

4 Factor 1 �0.44 Internal loner 27 Performance factor �0.03


Factor 3 1.01 Internal contributor

Factor 4 �0.07

5 Factor 1 �0.57 Internal receptor 11 Performance factor �0.18



Factor 4 �0.24




Factor 4 1.00




Factor 4 0.46

8 Factor 1 0.86 Internal contributor 22 Performance factor �0.14

Factor 2 �0.70 Impact on local factor �0.66

Factor 3 �0.91 External loner

Factor 4 �0.66




Factor 4 �0.51

10 Factor 1 0.71 Internal contributor 41 Performance factor 0.25



Factor 4 0.80



Factor 3 �0.19 External receptor

Factor 4 0.71




Factor 4 1.54




Factor 4 2.61


the multinational subsidiaries in our sample have been successfully classified into as many as 12 out of the 16 groups ofmultinational subsidiaries. Compared to our case studies, firm CQ6 is in Group 1 (dual activist) and firm CQ30 is in Group 14(dual loner). This indicates that the cluster grouping results are the same as the case-study results for these subsidiaries. Wecan conclude that our dual-role typology is generally appropriate in differentiating the subsidiaries’ roles, strategies orcharacteristics in both the academic and business fields.

From Tables 3 and 4 we can detect the following pattern of multinational subsidiaries’ roles or strategies in China. Firstly,the three largest groups of multinational subsidiaries are dual loners (73 cases), dual activists (45 cases), and internalcontributor/external activists (41 cases). Secondly, there are 133 internal loners with 73 dual loners, 27 internal loner/external contributors, and 33 internal loner/external receptors, and a similar number (132) of external loners (73 dual loners,16 internal activist/external loners, 22 internal contributor/external loners, and 21 internal receptor/external loners)respectively. Thirdly, there are 108 external activists (29% of the sampled multinational subsidiaries), and 61 internalactivists (17% of the total sample). Finally, there are 45 dual activists, or merely 12% of the sampled subsidiaries, while dualloners account for 20%.

From Tables 3 and 4, while some multinational subsidiaries in China are indeed actively involved in knowledgeabsorption and diffusion and help enhance the efficiency of MNEs or promote local capability development (e.g. dualactivists), slightly more than one-third of multinational subsidiaries in China do not seem to learn from and share theirknowledge with the rest of their corporations (e.g. internal loners). Hence they do not seem to contribute to the efficiency orcompetitiveness of their corporations. Furthermore, more than one-third of multinational subsidiaries in China do not seemto learn from and share their knowledge with local Chinese firms (e.g. external loners). Hence, they neither seem to gain fromthe local environment, nor do they help local development. The behaviour of this proportion of subsidiaries is inconsistentwith the main purpose of FDI policy in a typical developing country, which expects MNEs to transfer or diffuse knowledge tolocal firms. Our results are however consistent with the conclusion of a number of existing empirical studies that theexistence and magnitude of knowledge spillovers from FDI depend on a number of factors, including MNEs’ incentives to

Table 3 (Continued )

Cluster Factor Mean Groups combine No. of cases Performance/impacts on locals Mean




Factor 4 �0.87




Factor 4 �0.77




Factor 4 �0.84

Notes: For internal interaction from HQ to subsidiary, higher scores signify more internal flows from HQ to subsidiary.

For internal interaction from subsidiary to HQ, higher scores signify more internal flows from subsidiary to HQ.

For external interaction from Local to subsidiary, higher scores signify more external flows from Local to subsidiary.

For external interaction from subsidiary to local, higher scores signify more external flows from subsidiary to local.

Table 4

Summary of the subsidiary groups.

Group Combination Cluster Number Numbers of subsidiary

Dual Activist 1, 7 45

Internal activist, External contributor

Internal activist, External receptor

Internal activist, External loner 9 16

Dual Contributor 3 14

Internal contributor, External activist 10 41

Internal contributor, External receptor 12, 13 27

Internal contributor, External loner 8 22

Dual Receptor

Internal receptor, External activist 6 22

Internal receptor, External contributor 2, 5 28

Internal receptor, External loner 16 21

Dual Loner 14, 15 73

Internal loner, External activist

Internal loner, External contributor 4 27

Internal loner, External receptor 11 33


create and diffuse knowledge and the absorptive capacity of host countries/firms, as summarised in Blomstrom and Kokko(1998), Gorg and Strobl (2001), Gorg and Greenaway (2004) and Smeets (2008).

The existence of a relatively large proportion of external loners may be due to the fact that China is still an emergingeconomy, and its knowledge and skills are not seen to be important to some multinational subsidiaries, especially those fromdeveloped countries which are technology leaders. Hence they may have no incentives to learn from and share theirknowledge with local Chinese firms. The fact that there are many internal loners may be caused by subsidiary management’soverlooking of the importance of knowledge sources within their corporations.

From Table 4 it is interesting to note that no firms can be classified into the following four clusters: (I) Internal activist/external contributor; (II) Internal activist/external receptor; (III) Internal loner/external activist; and (IV) Dual receptor.These results may be caused by the particular development stage of the Chinese economy and perceptions of managers ofmultinational subsidiaries in China. As mentioned earlier, China is an emerging economy and its general technological levelis still relatively low. Given this background, internal activists may adopt two extreme strategies. Some may deemindigenous Chinese knowledge to be of little help, and conduct no knowledge exchange with local Chinese firms, leading tothe absence of (I) and (II), while other internal activists may regard indigenous Chinese knowledge to be very important andproactively learn and diffuse knowledge with local Chinese firms. As for the absence of III), it may be logical for internalloners not to be so proactive in learning and sharing knowledge externally either especially when the level of technologicalknowledge in China is widely thought to be lower than that of MNEs. Finally, the absence of IV) may be due to the fact thatinternal receptors take different external strategies. They may either proactively learn and diffuse knowledge with localChinese firms (external activists and contributors) as they feel that they can benefit from such knowledge exchange, or theyare externally inactive (external loners) as they may feel that local knowledge is not important.

As foreign partner influence may depend on the distribution of ownership by partners (Thomas, Li, & Liu, 2008), we haveexpanded our sample to include all firms with at least 25% foreign ownership for comparison. This increases the sample sizefrom 369 to 482. We have used the same methods to classify these 482 foreign invested firms (with both majority and minorityforeign ownership), and the results are presented in Table 51. It is clear that these 482 firms can again be classified into 12 out ofthe 16 conceptual groups. Similar to the subsidiary sample, there are no entries into the following three conceptual groups: (I)Internal activist/external contributor; (II) Internal activist/external receptor; and (III) Dual Receptor. There is one importantdifference between the two samples: while there is no entry of an internal loner/external activist in the subsidiary sample, wefind 38 entries in this cluster in the foreign-invested firm sample. Our explanation is that the firms in the latter sample have onaverage higher local Chinese ownership, and this helps their external knowledge links with local Chinese firms.

From Table 5, the three largest groups of foreign-invested firms are dual loners with 105 cases, followed by internalcontributor/external activists (56 cases), and internal receptor/external loners (43 cases). Secondly, there are 198 internalloners (105 dual loners, 38 internal loner/external activists, 38 internal loner/external contributors, and 17 internal loner/external receptors) and a similar number (186) of external loners (105 dual loners, 38 internal contributor/external loners,and 43 internal receptor/external loners). Finally, there are only 42 dual activists, accounting for merely 9% of the sampledforeign-invested firms, while dual loners (105) account for as high as 22%.

Compared with Tables 4 and 5, the general patterns of knowledge flows of multinational subsidiaries and foreign-invested firms in China are quite consistent. However, given that the foreign-invested firm sample produces a higherproportion of dual loners but a lower proportion of dual activists than the subsidiary sample, a firm with higher foreignownership (e.g. a subsidiary) tends to be more active both internally and externally in knowledge diffusion.

Table 5

Summary of the subsidiary groups.

Group Combination Cluster Number Number of subsidiary

Dual Activist 12, 13, 14 42

Internal activist, External contributor

Internal activist, External receptor

Internal activist, External loner

Dual Contributor 8 20

Internal contributor, External activist 11 56

Internal contributor, External receptor 4, 16 32

Internal contributor, External loner 5 38

Dual Receptor

Internal receptor, External activist 3 36

Internal receptor, External contributor 15 17

Internal receptor, External loner 9 43

Dual Loner 6, 7 105

Internal loner, External activist 2 38

Internal loner, External contributor 10 38

Internal loner, External receptor 1 17

1 The detailed results of factor loadings and cluster analysis for this sample are not reported due to space limitation, but available upon request.


5.4. Subsidiary roles and performance

After classifying multinational subsidiaries according to their roles in internal and external knowledge flows, weinvestigated the relationship between subsidiary role or strategy and subsidiary performance. Since the performancemeasurement variables in this study are ROA, ROE, ROI and OVERALL, we use factor analysis to decide whether these fourvariables can be combined into one factor to represent subsidiary performance, and use ANOVA to examine the relationshipbetween subsidiary strategy and performance.

Form the factor analysis, the Kaiser–Meyer–Olkin Measure of Sampling Adequacy value is 0.548 and the Sig. value ofBartlett’s Test of Sphericity is less than 0.05. In the Component Matrix, the loading of ROE is only�0.203, while other loadingsof ROA, ROI and OVERALL are high. Therefore, the variable ROE is dropped and the other three variables are combined intoone factor (performance factor).

The ANOVA gives a sig. value of 0.037 that is less than 0.05, indicating that there are significant differences among themean scores on the performance factor for the 16 groups. Based on our discussion, activists are expected to perform betterthan loners. The mean of the dual activist clusters (clusters 1 and 7, 0.45) is higher compared to the dual loner cluster (cluster14 and 15, �0.26). In addition, the combined mean of the four internal activist clusters is 0.34, much higher than thecombined mean of the four internal loner clusters, which is only�0.18. Furthermore, the combined mean of the four externalactivist clusters is 0.25, much higher than the combined mean of the four external loner clusters, which is only�0.18. To sumup, we have compared three different groups of performance differences between the activist clusters and loner clusters, andall the results show that activists perform better than loners.

The results on the relationship between subsidiary role and performance tend to further pinpoint the importance of ourtypology of multinational subsidiaries. FDI network theory predicts a positive relationship between a multinationalsubsidiary’s internal/external interactions and its performance. Our results help validate the dual-role typology ofmultinational subsidiaries, and offer clear links between the roles and performance of multinational subsidiaries in theworld’s largest emerging economy.

5.5. Subsidiary role and its impact on local partners

Since the impact of subsidiary role or strategy on local partners in this study is measured by IMPACTONLP, IMPACTONLM,IMPACTONLRD and IMPACTONLMARKET, we use factor analysis to decide whether these four variables can be combined intoone factor to represent its impact on local firms, and use ANOVA to test its relationship with subsidiary strategy. The resultsof factor analysis show that the Kaiser–Meyer–Olkin Measure of Sampling Adequacy value is 0.777 which is more than 0.6,and the Significant value of Bartlett’s Test of Sphericity is less than 0.05. Therefore the factor analysis is appropriate. In theComponent Matrix, all variables’ loadings are high so that all of them can be represented by one factor, i.e. the Impact factor.2

The ANOVA gives a significant value of less than 0.05 indicating that there are significant differences somewhere amongthe mean scores on the impact factor for the 16 groups. Based on our discussion in Section 3, external activists are expected tohave a more positive influence on local partners than external loners. The combined mean of the four external activistclusters is 0.54, and is positive and much higher than the combined mean of the four external loner clusters, which is only�0.58. Therefore, the statistical results confirm this expectation.

6. Conclusions

Multinational subsidiaries can play different roles in improving efficiency at both the subsidiary and corporate level andin enhancing local capability development in the host economy. This explains why there has been an increasing interestamong researchers in analyzing subsidiary typology. Following a critical assessment of the existing subsidiary typologyframeworks, we have developed a dual-role typology based on a network approach to the MNE. This typology considers boththe efficiency of multinational subsidiaries and their impact on local development and is believed to be an importantextension of Gupta and Govindarajan’s (1991) knowledge-based typology. The empirical results from an analysis of 369multinational subsidiaries in China lend strong support to our new typology as 12 out of 16 conceptual groups have relevantsubsidiaries to fix into. The results are compared with those from a larger sample of 482 firms with at least 25 percent foreignownership. The central message is that a large proportion of foreign-invested firms in China are inactive in knowledgeexchange either internally or externally or both. Our results confirm the conclusion of a number of existing empirical studiesthat the existence and magnitude of knowledge spillovers from FDI depend on a number of factors, including MNEs’incentives to create and diffuse knowledge and the absorptive capacity of host countries/firms.

Our dual-role typology may change the way we think about and study roles or strategies of multinational subsidiaries. Asa multinational subsidiary operates in business networks of the MNE and the host economy simultaneously, its role inknowledge flows needs to be assessed in a holistic way by including both its internal and external interactions rather thanthe former only. For instance, an internal activist may well enhance the competitiveness of the MNE and itself by activelydiffusing its own knowledge to and absorbing knowledge from the rest of the MNE. However, this subsidiary’s knowledge

2 The detailed test results are not tabulated but are available upon request.


diffusion ability within the MNE may be higher if it is also an external receptor or activist. To find whether this is the case, weneed to examine the subsidiary’s external knowledge exchange activities. Furthermore, an internal activist can be anexternal loner at the same time. In this case, this subsidiary plays little role in knowledge accumulation in the host economy.

There are several managerial and policy implications of this study. Firstly, managers of multinational corporations needto integrate their subsidiaries into internal networks of the corporations and encourage them to learn from local businesspartners to benefit from both sources of knowledge and enhance the competitiveness of the subsidiaries and corporations.Secondly, managers of local firms should also be encouraged to establish networks with MNEs to source knowledge from thesubsidiaries. Thirdly, host-country FDI policy may need to be selective, i.e. may need to encourage FDI which is expected todevelop extensive linkages with local firms to promote knowledge accumulation in the local economy.

The contributions of this paper to our understanding of multinational subsidiary strategies or roles need to be interpretedwith caution as there are several important limitations. One such limitation is that the multinational subsidiaries in thesample are located in only three different areas in China. Perhaps an even larger sample with the selection of subsidiarieslocated in more different areas may allow all 16 conceptual clusters to have relevant cases to fix in. Secondly, to measure asubsidiary’s impact on local partners, we used four subjective variables based on the views of subsidiary managers only. Thismay lead to a biased measurement since subsidiary managers may overestimate their subsidiaries’ positive impact on localpartners. Thus, in order to gain a fairer evaluation of subsidiaries’ impact on local partners, the views of related local firmsneed to be collected. Despite the limitations, the current study should contribute to both theoretical and empirical literatureon multinational subsidiary typology.

Acknowledgements

This research was supported by a grant from the ‘‘project 211(phase III)’’ of the Southwestern University of Finance andEconomics. The authors would like to thank Professor Pervez Ghauri and three anonymous reviewers for their helpfulcomments and suggestions on the drafts.

References

Almeida, P., & Phene, A. (2004). Subsidiaries and knowledge creation: The influence of the MMC and host country on innovation. Strategic Management Journal, 25,847–864.

Andersson, U., Bjorkman, I., & Forsgren, M. (2005). Managing subsidiary knowledge creation: The effect of control mechanisms on subsidiary local embeddedness.International Business Review, 14, 521–538.

Andersson, U., Forsgren, M., & Holm, U. (2002). The strategic impact of external networks: Subsidiary performance and competence development in themultinational corporation. Strategic Management Journal, 23, 979–996.

Andersson, U., Forsgren, M., & Pedersen, T. (2001). Subsidiary performance in multinational corporations: The importance of technology embeddedness.International Business Review, 10(1), 3–23.

Aitken, B. J., & Harrison, A. E. (1999). Do domestic firms benefit from direct foreign investment? Evidence from Venezuela. American Economic Review, 89(3), 605–618.

Balasubramanyam, V. N., Salisu, M., & Sapsford, D. (1996). Foreign direct investment and growth in EP and IS countries. Economic Journal, 106, 92–105.Bartlett, C. A. (1986). Building and managing the transnational: The new organizational challenge. In M. E. Porter (Ed.), Competition in global industries. Boston, MA:

Harvard Business School Press.Bartlett, C. A., & Ghoshal, S. (1989). Managing across borders: The transnational solution. Boston, MA: Harvard Business School Press.Bilton, T., Bonnett, K., Jones, P., Lawson, T., Skinner, D., Stanworth, M., & Webster, A. (2002). Introductory sociology. Hampshire England: Palgrave Macmillan.Birkinshaw, J., & Morrison, A. J. (1995). Configurations of strategy and structure in subsidiaries of multinational corporations. Journal of International Business

Studies, 26, 729–754.Blomstrom, M., & Kokko, A. (1998). Multinational corporations and spillovers. Journal of Economic Surveys, 12(2), 1–31.Buckley, P. (2004). The role of China in the global strategy of multinational enterprises. Journal of Chinese Economic and Business Studies, 2(1), 1–25.Buckley, P., & Casson, M. (1976). The future of the multinational enterprise. London: Macmillan.Buckley, P., & Casson, M. (2002). The future of the multinational enterprise (25th anniversary ed.). London: Palgrave Macmillan.Buckley, P., & Ghauri, P. (2004). Globalisation, economic geography and the strategy of multinational enterprises. Journal of International Business Studies, 35,

81–98.Casson, M. (1994). Why are firms hierarchical? International Journal of the Economics of Business, 1(1), 47–76.Casson, M., Pearce, R., & Singh, S. (1991). A review of recent trends. In S. Casson (Ed.), Global research strategy and international competitiveness. Oxford: Blackwell.Caves, R. E. (1974). Multinational firms, competition and productivity in host-country markets. Economica, 41, 176–193.Chen, Y. C. (2008). Why do multinational corporations locate their advanced R&D centres in Beijing? Journal of Development Studies, 44(5), 622–644.Cooke, F. L., & Beh, K. K. (2007). Continuous innovations in production technology in the manufacturing industry in Brita a networked approach? Vision, 11, 1–13.Chudnovsky, D., Lopez, A., & Rossi, G. (2008). Foreign direct investment spillovers and the absorptive capabilities of domestic firms in the Argentine manufacturing

sector (1992–2001). Journal of Development Studies, 44(5), 645–677.D’Cruz, J. R. (1986). Strategic management of subsidiaries. In H. Etemad & S. Dulude (Eds.), Managing the multinational subsidiaries. London: Croon Helm.Dunning, J. (1994). Multinational enterprises and the globalization of innovatory capacity. Research Policy, 23, 67–89.Ferdows, K. (1997). Making the most of foreign factories. Harvard Business Review, 75(2), 73–88.Ferguson, T. D., Deephouse, D. L., & Ferguson, W. L. (2000). Do strategic groups differ in reputation. Strategic Management Journal, 21, 1195–1214.Fulcher, J., & Scott, J. (2003). Sociology. Oxford: Oxford University Press.Ghauri, P., & Gronhaug, K. (2002). Research methods in business studies. London: FT Prentice Hall.Ghauri, P. N., & Yamin, M. (2009). Revisiting the impact of multinational enterprises on economic development. Journal of World Business, 44(2), 105–107.Ghemawat, P. (2007). Managing differences: The central challenge of global strategy. Harvard Business Review, 85(3), 59–68.Ghoshal, S., & Bartlett, C. A. (1988). Creation, adoption and diffusion of innovations by subsidiaries of multinational corporations. Journal of International Business

Studies, 19(3), 365–388.Gorg, H., & Greenaway, D. (2004). Much ado about nothing? Do domestic firms really benefit from foreign direct investment?. World Bank Research Observer, 19(2),

171–197.Gorg, H., & Strobl, E. (2001). Multinational companies and productivity spillovers: A meta-analysis. Economic Journal, 111(475), F723–F739.


Gupta, A. K., & Govindarajan, V. (1991). Knowledge flows and the structure of control within multinational corporations. Academy of Management Review, 16(4),768–792.

Gupta, A. K., & Govindarajan, V. (1994). Organizing for knowledge flows within MNCs. International Business Review, 43(4), 443–457.Harzing, A.-W.K. (1999). Managing the multinationals: An international study of control mechanisms. Northampton, MA: Edward Elgar.Harzing, A.-W.K., & Noorderhaven, N. (2006). Knowledge flows in MNCs: An empirical test and extension of Gupta and Govindarajan’s typology of subsidiary roles.

International Business Review, 15(3), 195–214.Hedlund, G. (1986). The hypermodern MNC—A heterarchy? Human Resource Management, 25, 9–25.Hogenbirk, A. E., & van Kranenburg, H. L. (2006). Roles of foreign owned subsidiaries in a small economy. International Business Review, 15, 53–67.Hu, A. G. Z., & Jefferson, G. H. (2002). FDI impact and spillover: Evidence from China’s electronic and textile industries. World Economy, 25, 1063–1076.Hymer, S. (1976). The international operations of national firms: A study of direct foreign investment. Cambridge, MA: MIT Press.Jarillo, J. C., & Martinez, J. I. (1990). Different roles for subsidiaries: The case of multinational corporations. Strategic Management Journal, 11, 501–512.Javorcik, B. S. (2004). Does foreign direct investment increase the productivity of domestic firms? In search of spillovers through backward linkages. American

Economic Review, 94(3), 605–627.Jaw, B. S., Yu, C., Wang, P., & Chen, Y. H. (2006). Knowledge flows and performance of multinational subsidiaries: The perspective of human capital. International

Journal of Human Resource Management, 17(2), 225–244.Ketchen, D. J., & Shook, C. L. (1996). The application of clcuster analysis in strategic management research: An analysis and critique. Strategic Management Journal,

17, 441–458.Kogut, B. (1989). Research notes and communications: A note on global strategies. Strategic Management Journal, 10, 383–389.Kokko, A., Tansini, R., & Zejan, M. C. (1996). Local technological capability and productivity spillovers from FDI in the Uruguayan manufacturing sector. Journal of

Development Studies, 32(4), 602–611.Leonard-Barton, D. (1995). Wellsprings of knowledge: Building and sustaining sources of innovation.. Boston, MA: Harvard University Press.Levitt, T. (1983). The globalization of markets. Harvard Business Review, 61(3), 92–102.Li, X., Liu, X., & Parker, D. (2001). Foreign direct investment and productivity spillovers in the Chinese manufacturing sector. Economic Systems, 25, 305–321.Liu, X., Siler, P., Wang, C., & Wei, Y. (2000). Productivity spillovers from foreign direct investment: Evidence from UK industry level panel data. Journal of

International Business Studies, 31(3), 407–425.Meyer, K. (2004). Perspectives on multinational enterprises in emerging economies. Journal of International Business Studies, 35(4), 259–276.Moser, C. A., & Kalton, G. (1983). Survey methods in social investigation. Hampshire, England: Grower House.Monteiro, L., Arvidsson, N., & Birkinshaw, J. (2008). Knowledge flows within multinational corporations: Explaining subsidiary isolation and its performance

implications. Organization Science, 19(1), 90–107.Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company. New York: Oxford University Press.Nonaka, I., & Teece, D. (Eds.). (2001). Managing industrial knowledge: creation, transfer and utilization. London: Sage Publications.O’Donnell, S. W. (2000). Managing foreign subsidiaries: agents of headquarters, or an interdependent network? Strategic Management Journal, 21, 525–548.Oppenheim, A. N. (1996). Questionnaire design, interviewing and attitude measurement. London: Printer Publishers.Pallant, J. (2004). SPSS survival manual (2nd ed.). Berkshire, England: Open University Press.Podsakoff, P. M., & Organ, D. W. (1986). Self reports in organizational research: Problems and prospects. Journal of Management, 12, 531–544.Porter, M. E. (1980). Competitive strategy. New York: Free Press.Porter, M. E. (1986). Competition in global industries: A conceptual framework. In M. E. Porter (Ed.), Competition in Global Industries (pp. 15–60). Boston, MA:

Harvard Business School Press.Prahalad, C. K. & Hamel, G. (1990, May–June). The Core competence of the corporation. Harvard Business Review. pp. 79–91.Ramamurti, R. (2004). Developing countries and MNEs, extending and enriching the research agenda. Journal of International Business Studies, 35, 277–283.Roth, K., & Morrison, A. J. (1992). Implementing global strategy: Characteristics of global subsidiary mandates. Journal of International Business Studies, 23(4),

715–735.Saunders, M., Lewis, P., & Thornhill, A. (2003). Research methods for business student (3rd ed.). London: Prentice Hall.Short, J., Ketchen, D., Palmer, T., & Hult, G. (2007). Firms, strategic group, and industry influences on performance. Strategic Management Journal, 28, 147–167.Smeets, R. (2008). Collecting the pieces of the FDI knowledge spillovers puzzle. World Bank Research Observer, 23(2), 107–138.Storey, J., & Salaman, G. (2005). Managers of innovation: Insight into making innovation happen. Oxford: Blackwell.Taggart, J. H. (1998). Strategy shifts in MNC subsidiaries. Strategic Management Journal, 19, 663–681.Tanriverd, H., & Venkatraman, N. (2005). Knowledge relatedness and the performance of multibusiness firms. Strategic Management Journal, 26(4), 97–119.Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(5), 509–533.Thomas, H., Li, X., & Liu, X. (2008). Foreign ownership structure and new product development in China. Transtional Corporations, 17(2), 17–44.Vereecke, A., Van Dierdonck, R., & De Meyer, A. (2006). A typology of plants in global manufacturing networks. Management Science, 52(11), 1737–1750.Wei, Y., & Liu, X. (2006). Productivity spillovers from R&D, exports and FDI in China’s manufacturing sector. Journal of International Business Studies, 37(4),

544–557.Wei, Y., Liu, X., & Wang, C. (2008). Mutual productivity spillovers between foreign and local firms in China. Cambridge Journal of Economics, 32, 609–631.White, R., & Poynter, T. (1984). Strategies for foreign-owned subsidiaries in Canada. Business Quarterly, 48(4), 59–69.Yip, G. S. (1992). Total global strategy: Managing for worldwide competitive advantage. Englewood Cliffs, NJ: Prentice-Hall.Young, S., & Tavares, A. T. (2004). Centralization and autonomy: back to the future. International Business Review, 13, 215–237.


a dual-role typology of multinational subsidiaries

Documents