building firm-specific advantages in multinational corporations

Strategic Management Journal, Vol. 19, 221–241 (1998)

BUILDING FIRM-SPECIFIC ADVANTAGES INMULTINATIONAL CORPORATIONS: THE ROLE OFSUBSIDIARY INITIATIVE

JULIAN BIRKINSHAW1*, NEIL HOOD2 and STEFAN JONSSON1

1Institute of International Business, Stockholm School of Economics, Stockholm,Sweden2University of Strathclyde, Glasgow, Scotland

This paper investigates how subsidiary companies are able to contribute to the firm-specificadvantages of the multinational corporation (MNC). Specifically we examine the determinantsof the contributory role of the subsidiary and subsidiary initiative. The study reveals the followingsignificant relationships: (a) internal subsidiary resources in combination with initiative havea strong positive impact on the subsidiary’s contributory role; (b) subsidiary initiative isstrongly associated with the leadership and entrepreneurial culture in the subsidiary; and (c)contributory role is strongly associated with subsidiary autonomy and a low level of localcompetition. We discuss the implications of these findings and some of the theoretical issuesassociated with subsidiary initiative. Our provisional conclusion is that MNC subsidiaries cannot only contribute to firm-specific advantage creation, they can also drive the process. 1998John Wiley & Sons, Ltd.

Strat. Mgmt. J., Vol. 19, 221–241, (1998)

INTRODUCTION

A central theme of much of the recent literatureon the strategy of the multinational corporation(MNC) is the increasingly important role playedby subsidiary companies as contributors to thedevelopment of firm-specific advantages. Tra-ditional academic models that viewed subsidiariesas either ‘market access’ providers or as recipientsof the parent company’s technology transfers(Vernon, 1966) gave way in the 1980s to richerconceptualizations in which subsidiaries tappedinto leading-edge ideas, undertook importantresearch and development work, and becameactive participants in the formulation andimplementation of strategy (Bartlett and Ghoshal,1986; Gupta and Govindarajan, 1994; Hedlund,

Key words: multinational corporation; subsidiary;entrepreneurship; globalization*Correspondence to: Julian Birkinshaw, Institute ofInternational Business, Stockholm School of Econom-ics, PO Box 6501, S113-83, Stockholm, Sweden.

CCC 0143–2095/98/030221–21 $17.50 Received 30 August 1996, Revised 22 January 1997 1998 John Wiley & Sons, Ltd. Final revision received 9 May 1997

1986). The generation of firm-specific advantages,correspondingly, shifted from being the sole con-cern of the parent company to a collectiveresponsibility for the corporate network.

This paper investigateshow subsidiary com-panies are able to contribute to the firm-specificadvantages of the MNC. In one respect the paperis similar to a number of recent articles that haveexamined the different roles taken by subsidiarycompanies (Birkinshaw and Morrison, 1996; Gho-shal and Nohria, 1989; Gupta and Govindarajan,1994; Jarillo and Martinez, 1990; Roth and Mor-rison, 1992) because we are concerned withunderstanding those factors that differentiatebetween high-contributing and low-contributingsubsidiaries. However, it is also unique in twoimportant respects. First, we attempt to pry openthe ‘black box’ of the subsidiary by discussingthe various activies that occur within it, and theprocesses that link them. Second, we pick outone key activity, subsidiary initiative, and explorethe factors associated with it in detail. Our belief,which this paper provides preliminary evidence

222 J. Birkinshaw, N. Hood and S. Jonsson

for, is that the evolution of subsidiary roles canbe taken one step further than previously realized.Rather than simply seeing subsidiaries ascon-tributors to the development of firm-specificadvantages, this paper shows that they can alsodrive the process through their own initiative.

This paper is in five sections. The first sectionoffers a brief survey of the literature on firm-specific advantages and MNC subsidiaries. Thesecond section develops the theoretical conceptsof firm-specific advantage in subsidiaries and for-mally sets out the research propositions. The thirdsection describes the data collection, which wasundertaken in a sample of manufacturing subsidi-aries in Canada, Scotland and Sweden. The fourthsection describes the research findings. Finally,the fifth section offers a discussion of the resultsfrom the research and their implications for theoryand for practice.

THEORETICAL AND EMPIRICALBACKGROUND

Researchers have long recognized the centralityof ownership- or firm-specific advantages to anunderstanding of theraison d’etre of MNCs. Asfirst shown by Hymer (1976), firms engaging inoverseas production must have some form ofproprietary advantage to compensate for the natu-ral disadvantage of competing with establishedfirms in a foreign land. As stated by Dunning(1980, 1988) this firm-specific advantage can besubdivided into two distinct types of advantage:asset advantages, that stem from the exclusiveprivileged possession of income generating assets;and transaction advantages, which reflect thefirm’s ability to economize on transaction costsas a result of multinational coordination and con-trol of assets. While not asufficientcondition forforeign production, there is general agreementthat some form of firm-specific advantage isnonethelessnecessary(Dunning, 1988; Rugman,1981). Technological resources, in particular,have been the focus of many studies of firm-specific advantage (e.g., Cantwell, 1989; Rugman,1981; Teece, 1977), though research has alsoconsidered manufacturing, marketing, organi-zational, and human resources (Dunning, 1993:81).

A major problem, as pointed out by Rugmanand Verbeke (1992), is that this research has

1998 John Wiley & Sons, Ltd. Strat. Mgmt. J., Vol. 19, 221–241 (1998)

tended to assume that the MNC’s firm-specificadvantages originate in the parent company,whereas the reality is that subsidiaries can playan important part in the creation and maintenanceof such advantages. The emerging body ofresearch concerned with subsidiary roles is testa-ment to this shift in the locus of firm-specificadvantage creation (see Birkinshaw and Morrison,1996, for a review). For example, it is reportedthat subsidiaries can act as contributors to orleaders of innovation projects (Bartlett and Gho-shal, 1986); they can provide major outflows ofvalued resources to the rest of the corporation(Gupta and Govindarajan, 1994), and they cangain mandates for developing and producing cer-tain product lines on a global basis (Roth andMorrison, 1992). Terms such asspecialized con-tributor, strategic leader, and active subsidiaryhave been used to refer to those subsidiaries thatcontribute substantially to firm-specific advantage,while terms such asimplementer and branchplant are used to refer to those that do notcontribute significantly to firm-specific advan-tage.1

While there is no shortage of typologies sug-gesting that subsidiaries vary in their contributoryrole (i.e., in their contribution to firm-specificadvantage), there is no definitive evidence for thesources of such variation. A number of studieshave looked at the factors associated with differ-ences in contributory role, but they have typicallyfocused on only a subset of the potentiallyimportant factors. More specifically, three con-trasting perspectives can be discerned from theMNC subsidiary literature. The first perspectiveis one ofenvironmental determinism. Building onthe notion that the MNC operates in multipleenvironments each with is own unique character-istics, the role of each subsidiary is seen inlarge part as a function of its local environment(Ghoshal and Nohria, 1989; Westney, 1994).Where the local country is strategically important(Bartlett and Ghoshal, 1986) or where the dyna-mism of local competitors, suppliers and cus-tomers is high (Porter, 1990), the expectation isthat the subsidiary will have a correspondingly

1 We should add that we do not see an inexorable trendtowards higher value-added inall subsidiaries. Many subsidi-aries will continue to have simple market exploitation roles;others will take on the higher value-added roles describedhere.

Subsidiary Initiative in Multinational Corporations 223

important role. Industry factors, such as pressurefor local responsiveness and global integration(Jarillo and Martinez, 1990), can also be under-stood within an environmental determinism per-spective, in that they represent exogenous factorsthat the MNC has to adapt to. The second per-spective is one ofhead office assignment. Thisperspective works on the basis that head officemanagement is responsible for defining the stra-tegic imperatives of the whole company, andunderstands best how subsidiary roles can beassigned to ensure that those imperatives are met.Many studies have concentrated on facets ofstructural context (Bower, 1970) such as controland coordination mechanisms that can be used todirect the behavior of subsidiary managers, andthus to determine subsidiary role (Birkinshawand Morrison, 1996; Ghoshal, 1986; Gupta andGovindarajan, 1994; Roth and Morrison, 1992).Others have suggested that subsidiaries can beassigned roles more directly according to theirperceived importance or the growth prospects ofthe market (Bartlett and Ghoshal, 1986). Thethird perspective is ofsubsidiary choice (cf.Child, 1972), whereby the role of the subsidiaryis to a large extent open to subsidiary man-agement to define for themselves. This perspec-tive works on the assumption that subsidiary man-agement understand their local market and theirlocal capabilities better than head office, and thatthey are in the best position to decide what rolethe subsidiary should play. Rooted in the workof Canadian scholars such as White and Poynter(1984) and D’Cruz (1986), this perspectivefocuses on the specific resources and capabilitiesof the subsidiary, the aspirations of subsidiarymanagement, and the initiative and effort of sub-sidiary employees as the determinants of subsidi-ary role (Birkinshaw, 1995; Etemad and Dulude,1986; Roth and Morrison, 1992; Science Councilof Canada, 1980).

Clearly all three perspectives have considerablemerit, so for a complete understanding of thephenomenon it would be necessary to considersubsidiary, corporate, industry, and country fac-tors. However, at the same time the three perspec-tives are competing with one another for relativesalience. Is the relation of the subsidiary with isparent company the key determinant of its role?Are the attributes of the subsidiary itself moreimportant? Or is the local industrial environmentthe most important variable? While this study,


like all others, has its biases and its precon-ceptions, the relative impact of these three setsof factors will be explicitly assessed.

Management processes inside the subsidiary

Taken as a whole, the body of literature onsubsidiary management had done a far better jobof understanding aspects of subsidiary context(how the subsidiary relates to its parent, its corpo-rate network, its local environment) than ofunderstanding what actually happensinside thesubsidiary. If the subsidiary is small, focusedprimarily on the local market, and wholly depen-dent on the parent company, the inner workingsof the subsidiary are not of great consequence tothe MNC as a whole. However, subsidiary growthbrings with it an increase in resources and acorresponding reduction in parent control(Prahalad and Doz, 1981), which leads to at leastsome degree of strategic choice on the part ofsubsidiary management. At this point, how thesubsidiary is managed internally would appearto become a matter of great importance to thecorporation as a whole.

A recent study of Canadian subsidiaries byBirkinshaw (1995) provides some insight into theinternal workings of the subsidiary. Building onthe concept of induced and autonomous actionproposed by Burgelman (1983), this study trackeda series of autonomous subsidiary actions, orinitiatives, that sought to develop the internationalvalue-added scope of the subsidiary. The processmodel is illustrated in Figure 1. The developmentof specialized resources was promoted by thevision and actions of subsidiary leadership. Thesespecialized resources provided the opportunity forinitiative by subsidiary managers which led tothe development of international responsibilities(Crookell, 1986). These responsibilities rangedfrom just manufacturing (e.g., of a family ofchemicals for international sale) to product devel-opment, manufacturing, and marketing account-ability (e.g., for a new range of computermonitors). The enhanced resources and inter-national responsibilities led both to an increasein subsidiary initiative and to increased visibilityin the corporate system. Increased visibility, inturn, represented an affirmation in the ability ofsubsidiary leadership and a further stimulus forinitiative. The case of 3M Canada, in particular,showed how the process of subsidiary develop-


Figure 1. Process model of subsidiary activities and subsidiary resources

ment occurred over a long period of time andalmost exclusively through the activities of sub-sidiary management. The process represented hereechoes the work of Ghoshal and Bartlett (1994)in that initiative, resource growth, and visibilityform a virtuous circle of development that isinvigorated by the actions of top management.

It should be underlined here that initiativeappears only to be evident in a subset of thepopulation of subsidiaries. Many subsidiariesexhibit no initiative, in part because such effortswould not be positively received by head officemanagement and in part because managementdoes not have the drive or expertise to pursueinitiative. In such cases the development of spe-cialized resources and international responsi-bilities can still occur, but it rests on the activeinvolvement of parent company managers(Birkinshaw and Hood, 1997). Relatedly, the con-cept of initiative typically raises some concernsat head office because there are questions overthe motivations of the subsidiary manager: is heor she acting in the interests of the subsidiary,the corporation, or the host country? These issueswill be revisited in the discussion section ofthis paper.

THEORETICAL MODEL ANDPROPOSITIONS

We are now in a position to specify the modelthat will be examined in this study. A subsidiaryis defined as any operational unit controlled bythe MNC and situated outside the home country.In some cases there will be a single subsidiaryin the host country; in other cases there will beseveral. Consistent with the resource-based viewof the firm (Barney, 1991; Wernerfelt, 1984), asubsidiary is conceptualized as a heterogeneous


bundle of resources. Some of these resources(e.g., the salesforce) are ‘location bound’(Rugman and Verbeke, 1992), meaning that theirvalue is limited to their country or domain ofoperation. Others are not location bound, and canpotentially be leveraged by the corporation inother countries. These are the resources that offerthe potential for contributing to the MNC’s firm-specific advantage. However, there are three cri-teria that must be met before this potential is real-ized.

The first criterion relates to the value of theresources. Using a strict resource-based perspec-tive, resources need to be valuable, rare andimperfectly imitable to offer the potential of com-petitive advantage (Barney, 1991). Our preferenceis to use a less strict approach that requires thesubsidiary’s resources to bespecialized, whichwe define as superior to those available elsewherein the corporation. If the subsidiary’s specializedresources are combined with other resources else-where in the MNC, we suggest that they thenbecome part of the MNC’s firm-specific advan-tage.2

The second criterion is one of recognition bycorporate management. Recognition refers to thewidespread understanding and acceptance of thesubsidiary’s specialized resources in other partsof the MNC. The subsidiary may have expertisein process innovation, for example, but if thatexpertise remains undiscovered by other parts ofthe corporation, and focused solely on the localmarket, it can not become part of the MNC’sfirm-specific advantage. Corporate recognition can

2 Whether that firm-specific advantage also leads to a competi-tive advantage is a separate question. Firm-specific advantagesimply refers to the MNC’s ability to overcome its liabilityof foreignness; competitive advantage represents a sustainablelow-cost or differentiated position against competitors.


Figure 2. Illustration of link between subsidiary resource and MNC firm-specific advantage

probably be achieved through both top-down andbottom-up mechanisms that are akin to Burgel-man’s (1983) induced and autonomous strategicprocesses. The top-down process involves corpo-rate management identifying their leading-edgesubsidiaries through informal discussions, pro-ductivity measures, and internal benchmarkingstudies. The bottom-up process consists of entre-preneurial efforts by subsidiary management todemonstrate their expertise and willingness totake on additional responsibilities to head officemanagers (Birkinshaw and Hood, 1997).

We use the termcontributory role to refer tothe extent to which the subsidiary has specializedresources that are recognized by the corporationas a whole. This term is deliberately broad inscope because the nature of the subsidiary’s con-tribution to firm-specific advantage will varyenormously from case to case. High contributoryrole subsidiaries include the specialized contribu-tor, strategic leader, and active subsidiary typesdiscussed earlier. Also included are subsidiarieswith world product mandateand centre of excel-lence designations. The world product mandateterminology has been widely used by Canadianresearchers to refer to the subsidiary’s responsi-bility to develop, manufacture, and market a prod-uct line worldwide (Rugman and Bennett, 1982;Crookell, 1986). A centre of excellence is usuallyconceived more broadly as a unit with expertisein a primary or support activity that other partsof the corporation draw on (e.g, Forsgren, 1995).The problem with both these terms is that theysuggest a clear-cut division between those sub-sidiaries with, and those without, recognized spe-


cialized resources, whereas the reality is muchless marked. Thus, the term contributory roleis preferred because it represents a continuum.Contributory role is the dependent variable in theresearch model.3

The third criterion by which subsidiaryresources are translated into part of the MNC’sfirm-specific advantage is the effective transferand/or leverage of the resources in question. Thiscriterion is based on the observation thatresources are ‘sticky’ and often do not get trans-ferred effectively inside the firm (Szulanski,1995). Indeed, anecdotal evidence suggests thatsome centers of excellence may be establishedsolely to meet political ends, with no intentionof ever transferring the latent expertise to otherparts of the corporation, while some world prod-uct mandates exist in the mind of the subsidiarymanager without ever achieving legitimacy withother parts of the corporation. Unfortunately it isnot possible, given the data collected in thisstudy, to examine the resource transfer/leverageprocess in any detail. To some degree the processis out of the hands of the subsidiary, in thatresource transfer depends to a large degree onthe willingness of the receptor (Szulanski, 1995).However, it remains a potentially important linkthat future research should consider in moredetail.

The complete model is represented in Figure 2.

3 Though it should be clear that causality does not flow solelyin the direction of contributory role. Indeed, as Figure 1shows, the subsidiary’s existing contributory role is one ofthe major drivers of initiative.


Figure 3. Theoretical framework and path diagram for PLS analysis

The focus of the empirical part of this researchis on the factors associated with the developmentof specialized resources and contributory role. Inparticular, we apply the findings from the researchquoted earlier (Figure 1) to look at the role ofsubsidiary initiative in the process. Subsidaryinitiative is defined as the entrepreneurial pursuitof international4 market opportunities to whichthe subsidiary can apply its specialized resources.

The remainder of this section will specify theproposed relationships between the independentvariables and the two focal constructs (seeFigure 3). As suggested by the review of theliterature, these propositions represent competingideas about the drivers of subsidiary development,i.e., whether contributory role is environmentallydetermined, assigned by the parent company, ora matter of subsidiary choice. It should beobserved that the propositions specifyassociationbetween constructs rather than causation. As thediscussion so far has shown, causality in manycases is reciprocal, and with cross-sectional datait is impossible to indicate more than associationanyway. For the sake of clarity Figure 3 is, never-theless, drawn with directional arrows.

4 The international dimension should be emphasized here todistinguish initiatives that offer the potential of enhancing thesubsidiary’s contributory role from those that are fundamen-tally local in scope. We are only interested in the former here.


Subsidiary-level factors

The first set of propositions is concerned withrelationshipswithin the subsidiary unit. Consistentwith the earlier discussion and the findings ofRoth and Morrison (1992), a positive relationshipis anticipated between the subsidiary’s specializedresources and its contributory role. Subsidiaryinitiative, likewise, is proposed to positivelyinfluence the subsidiary’s contributory role. Intheoretical terms this relationship is premised onBurgelman’s (1983) analogous argument thatautonomous behavior (i.e., initiative) becomesincorporated into a concept of corporate strategy(i.e., the subsidiary’s role) through championingefforts and strategic context definition. More spe-cifically, subsidiary research by Birkinshaw(1995), Bishop and Crookell (1986), Ghoshal(1986), and Science Council of Canada (1980)all showed that aspects of subsidiary initiativehad an important influence on the role of thesubsidiary. Finally, it is also important to specifyan anticipated relationship between subsidiaryinitiative and specialized resources. As suggestedearlier, this relationship is expected to be recipro-cal in that specialized resources provide theopportunity for initiative, which in turn enhancesthe subsidiary’s resources.

Proposition 1: The level of specializedresources in the subsidiary is positively


associated with the contributory role of thesubsidiary.

Proposition 2: Subsidiary initiative is posi-tively associated with the contributory role ofthe subsidiary.

Proposition 3: The level of specializedresources in the subsidiary is positivelyassociated with subsidiary initiative.

The antecedent conditions for subsidiary initiativeand specialized subsidiary resources are antici-pated to stem from the efforts of subsidiary man-agement. Following the work of Ghoshal andBartlett (1994), top management are expected tobe instrumental in the development of a support-ive behavioral context which in turn fosters initia-tive among employees. Moreover, to the extentthat subsidiary top management has some discre-tion to directly commit resources to certain proj-ects rather than others, it is anticipated that thestrength of top management will be associatedwith the development of specialized (and poten-tially valuable) resources. This argument is anal-ogous to Hamel and Prahalad’s (1994) thesisthat firm competencies should be developed toanticipate or even drive industry evolution.Finally, the strength of top management is alsoexpected to directly influence the subsidiary’scontributory role. While specialized resources arethe underlying driver of the subsidiary’s contribu-tory role, it is typically the championing andsponsoring efforts of top management that triggerthe assignment of new international responsi-bilities or mandates to the subsidiary (Birkinshaw,1995; Bishop and Crookell, 1986; Burgelman,1983).

A related aspect of subsidiary top man-agement’s role is their responsibility for shapingthe development of an entrepreneurial culture inwhich initiative and risk-taking behavior canthrive (Kanter, 1985; Kuratko, Montagno, andHornsby, 1990; Peters and Waterman, 1982; Pin-chott, 1985). It is intuitively obvious that anentrepreneurial culture is likely to promote initia-tive.5 It is anticipated that an entrepreneurial sub-

5 Though it should be equally clear that they are not thesame thing. As defined here, initiatives are discrete cases ofentrepreneurship; entrepreneurial culture is an organizationalcontext in which certain behaviors, including initiative, arefostered.


sidiary culture will also be associated with thedevelopment of specialized resources, in muchthe same way that iniative is seen as a driver ofresource development. The implication is thateven in the absence of specific initiatives, anentrepreneurial atmosphere should still have apositive impact on the pursuit of new combi-nations of resources. In summary:

Hypotheses 4a, 4b: The actions of subsidiarymanagement (strong subsidiary leadership; anentrepreneurial subsidiary culture) are posi-tively associated with subsidiary initiative.

Hypotheses 5a, 5b: The actions of subsidiarymanagement (strong subsidiary leadership; anentrepreneurial subsidiary culture) are posi-tively associated with a high level of special-ized resources in the subsidiary.

Hypothesis 6: Strong subsidiary leadership ispositively associated with a high contributoryrole for the subsidiary.

Corporate-level factors

The traditional approach to subsidiary man-agement, as exemplified by the process school(Bartlett, 1979; Bower, 1970; Burgelman, 1983;Prahalad, 1976), conceptualized a ‘structural con-text’ for the subsidiary which consisted of thevarious facets of its relationship with the parentcompany. The subsidiary was controlled, accord-ing to this model, through the imposition (byhead office managers) of an appropriate structuralcontext that induced managers in the subsidiaryto behave in desirable ways. Aspects of contextincluded level of autonomy, formalization of acti-vites, control resources, and social control. Interms of the current study, the suggestion isthat by defining an appropriate structural context,corporate management can either promote orinhibit the development of the subsidiary’s con-tributory role.

While there have been a large number of studiesof parent–subsidiary relationships (e.g., Brandtand Hulbert, 1977; Gates and Egelhoff, 1986;Garnier, 1982; Leksell, 1984; Otterbeck 1981),Ghoshal’s (1986) dimensions of structural contextwere used as the starting point for this studybecause his study of subsidiary innovations wasclosest to our concepts of contributory role and


initiative. Ghoshal showed that the creation ofinnovation in subsidiaries was associated with highautonomy, high parent–subsidiary communicationand high normative integration. We therefore pre-dict, in an analogous manner, that decision-makingautonomy and high levels of parent–subsidiarycommunication will be associated with the subsidi-ary’s contributory role. There is a counter argu-ment to these hypotheses, namely that autonomycan indicate a lack of integration that may limitthe chances of gaining recognition for specializedresources. Nonetheless, Ghoshal’s empirical find-ings form the basis for our hypotheses. Normativeintegration, that is, the extent to which sharedvalues exist across the corporation, was not speci-fied. In our experience normative integration isvery hard to assess at a subsidiary level (Ghoshalpolled head office managers) in part because it isa corporate-wide concept.6

Hypothesis 7a, 7b: Facets of the parent–subsidiary relationship (subsidiary autonomy;parent–subsidiary communication) are posi-tively associated with a high contributory rolefor the subsidiary.

Country- and industry-level factors

The final element of the research model is theimpact of the business environment (at both acountry and industry level) on the subsidiary’scontributory role and its level of initiative. Whileit is broadly accepted that the nature of thelocal environment has a bearing on the role thesubsidiary plays in the corporation (e.g., Bartlettand Ghoshal, 1986; Ghoshal and Nohria, 1989),our interest in this study was on identifying thoseaspects of the environment that are salient to thesubsidiary’s contributory role. We focused on thelevel of competitiveness in the local market.7

Competition drives the innovation process andthe upgrading of capabilities (Porter, 1980, 1990).To the extent that the subsidiary is actively partic-

6 Instead of normative integration we attempted to measurethe related concept ofcredibility, that is, the extent to whichthe parent company has confidence that subsidiary man-agement will deliver on their objectives. However, no signifi-cant relationships with subsidiary initiative or contributoryrole were found, so we dropped the construct.7 We also looked into the quality of relationships with localsuppliers and customers (Porter, 1990) but no significantrelationships were identified.


ipating in its local marketplace, it is anticipatedthat the level of local competition will have apositive influence on the subsidiary’s own com-petitiveness, and hence on its contributory role.Porter’s (1990) diamond model, in particular,showed that the presence of clusters of firms ina single location drives the competitiveness ofthe entire cluster.

There is one caveat in order here, becausePorter’s (1990) research did not explicitly con-sider the impact of cluster development on for-eign-owned subsidiaries. Many subsidiaries havelimited decision-making autonomy, little oppor-tunity to choose their own suppliers, and limitedR&D capacity (Young, Hood, and Peters, 1994).All these constraints impede the subsidiary’sability to participate effectively in the competitiveupgrading process that Porter identified, whichsuggests that this hypothesis is tentative, giventhe current state of knowledge.

A second relevant facet of the businessenvironment is its level of globalization. Struc-tural drivers, such as the availability of economiesof scale, make certain industries more prone toglobal integration than others (Kobrin, 1991). Atone end of the spectrum are ‘pure global’ indus-tries (Porter, 1986) in which the subsidiary’sactivities are integrated with the rest of the corpo-rate network. At the other end of the spectrumare ‘multidomestic’ industries in which compe-tition in one national market is not substantiallyaffected by competition in the next. It is proposedhere that the level of subsidiary initiative isdirectly related to the level of globalization ofthe industry. Multidomestic industries do not offermuch scope for the subsidiary to influence thefirm-specific advantage of the corporation,because competition is structured on a local-for-local basis. Global industries, by contrast, requirea high level of specialization from subsidiarycompanies as each focuses on undertaking certainspecific activities on behalf of the MNC as awhole. The opportunity for initiative is thus muchgreater. Once again, it is important to emphasizethat this study is concened with internationallyoriented initiatives. It would correspondingly beexpected thatlocally focusedinitiatives are morepervasive in multidomestic industries.

Hypothesis 8: The competitiveness of thelocal market is positively associated with ahigh contributory role for the subsidiary.


Hypothesis 9: The level of industry glob-alization is positively associated with subsidi-ary initiative.

METHODOLOGY

Empirical data were collected using a mail ques-tionnaire which was completed by top managersin 229 manufacturing subsidiaries of large MNCSin Canada, Scotland, and Sweden. These countriesall have substantial populations of foreign-ownedsubsidiaries with similarities along two dimen-sions: (a) all three are relatively small countrieswith high standards of living; and (b) all three are‘peripheral’ parts of established trading blocks. Interms of generalizability, it therefore seems likelythat the findings of the study will be meaningfulto other ‘peripheral’ countries in developed areas.

Data were gathered during 1995. In each coun-try a slightly different sampling process was usedbecause of the nature of the available data bases.In Canada, the sample was drawn up from avariety of CD-ROM products and directories,including theFinancial Post 500, Report on Busi-ness 1000, and theDisclosuredata base. In Scot-land we used the data base compiled byScottishEnterprise, the inward investment agency, whichkeeps track of all foreign investors in Scotland.In Sweden we used the data bases of foreign-owned subsidiaries compiled byVeckans Affa¨rerand Compass.8 Using a standard procedure ofmailing the questionnaire to the subsidiary CEOand then mailing a remainder 4 weeks later weended up with 229 responses (34% response rate).Forty-nine subsidiaries were dropped, eitherbecause their revenues were below £15 million9

or because they had no manufacturing activity,leaving 180 usable responses. The mean annualrevenues of the sample were £203 million, witha range from £15 million through to £1.5 billion.

8 The Swedish questionnaire was translated into Swedish andback-translated to verify accuracy. Managers were sent bothEnglish and Swedish language versions, with approximatelyhalf filling in each version.9 We were unable to achieve complete consistency in subsidi-ary size across the three countries. In Canada all subsidiarieslarger than £40 million were sampled, but in Scotland andSweden we ended up polling many smaller subsidiaries aswell in order to achieve similar numbers of responses. Wefinally used a cut-off of £15 million, which meant a relativeabsence of Canadian subsidiaries in the £15–£40 millionrange.


Ninety-nine of the 180 sample subsidiariesreported having some form of internationalresponsibility. The most common parent companynationality by far was the United States (95),followed by Japan, Germany, England, Finland,and Switzerland (each with 10–20 responses).Details of response rates are listed in Table 1.

A test of nonresponse bias was conducted usingannual revenues (or number of employees in thecase of Scotland) and parent company nationalityas dependent variables, and no significant differ-ences were found. We also performed a series ofANOVAs using host country as the independentvariable, and again no major differences wereuncovered.10

The questionnaire was developed through athree-stage process. First, the draft questionnairewas reviewed by three academicians, who sug-gested improvements in wording and advice onlayout. Second, following a major revision of thequestionnaire, it was sent out to six subsidiarypresidents who were involved in an earlier study.They all filled out the questionnaire, while oneof the researchers did likewise on the basis ofhis extensive knowledge of the six companies.Responses were then compared, and where thedifferences between ‘actual’ (i.e., from the sub-sidiary president) and ‘expected’ (i.e., from theresearcher) were substantial amendments to word-ing were made. In most cases, however, responseswere very similar. At the same time, four pairsof subsidiary and head office managers were alsoasked to fill out the questionnaire, to ensure thatthe subsidiary’s answers were consistent with theperceptions in head office. No significant differ-ences were found. The interrater reliability forthese four pairs was 0.65 (using Cohen’skappa),11 an adequate but not exceptionally goodresult. Finally, once the second round of correc-tions had been made, the questionnaire was sentto a group of three managers inanothersubsidi-ary. A researcher met with these individuals to

10 The one significant difference between host countries wasthe value-adding scope of the subsidiaries. In Scotland manysubsidiaries had either manufacturing only or a predominantexport orientation, whereas in Canada and Sweden they typi-cally undertook local marketing and sales activities as well.11 This is, of course, not as high a coefficient as we wouldhave liked. Our sense from these questionnaires and fromtalking to the individuals was that head office managers wereunable to adequately answer some of the subsidiary-specificquestions (e.g., those relating to specialized resources andinitiative), which lowered the level of interrater reliability.


Table 1. Sample response rates

Canada Scotland Sweden Total

Questionnaires sent 270 182 221 673Returned blank, declined to participate 5 5 18 28Questionnaires returned complete 87 61 78 226Response rate 32% 34% 35% 34%Number used for statistical analysis 78 51 51 180(i.e., with revenues over £15 million)

discuss their responses, which resulted in a fewsmall changes.

Analytical method

The hypotheses were tested using a relativelynew multivariate analysis technique known aspartial least squares or PLS (Fornell andBookstein, 1982). PLS, like LISREL, is one ofthe so-called second-generation multivariate tech-niques that are increasingly being used to estimatecausal models with multiple independent anddependent constructs (e.g., Birkinshaw, Morrison,and Hulland, 1995; Johansson and Yip, 1994;Fornell, Lorange, and Roos, 1990). These tech-niques allow the researcher to analyze all pathsbetween constructs simultaneously, rather thanthrough a series of discrete regression models.PLS, in contrast to LISREL, has the additionaladvantage that it makes no assumptions aboutmultivariate normality in the data and it workswell with relatively small samples. Generally,PLS is preferred to LISREL in the early stagesof theory building and testing, and when theresearcher is primarily concerned with the predic-tion of the dependent variable.

PLS has one further benefit over first-generation techniques. Traditionally the researcherwould define a theoretical construct either bysumming individual items or by extracting factorscores from a factor analysis. In PLS, however,individual items are kept in their raw form asindicators of the construct,12 and their loadingson the construct then vary depending on therelationship of that construct to other constructsin the model. This is important because it makes

12 All indicators were reflective rather than formative. Thismeans that there is assumed to be an unobservable that‘causes’ the observables, rather than vice versa (Fornell,1984).


the use of reliability measures such as Cronbach’salpha redundant. Instead, the choice of whichindividual items to retain, as measures of a con-struct, becomes part of the overall model testing.Issues of reliabilty and validity can then beassessed once the model has been finalized.

Construct measurement

Construct measures were adpoted from earlierresearch where possible, most notably from pre-vious MNC subsidiary studies by Roth and Mor-rison (1992) and Ghoshal (1986). However, itproved necessary to create new measures forseveral of the key constructs as they had appar-ently not been measured before. The completewording of questions, and the correlationsbetween them, are displayed in the Appendix.

Contributory role was operationalized by ask-ing subsidiary presidents what percentage of theirrevenues (if any) were gained from ‘internationalresponsibilities’ such as world mandates or cen-ters of excellence (i.e., activities it undertook onbehalf of the corporation as a whole), so that 0percent would suggest that the subsidiary had nointernational responsibilities and 100 percentwould suggest that all their revenues were gainedfrom their international responsibilities. This mea-sure achieved our intention of recording thoseactivities that were international in scopeandrecognized by the corporation. However, itfocused on physical and technological resourceflows which meant that some aspects of the sub-sidiary’s qualitative contribution to firm-specificadvantage (e.g., sharing of ideas, knowledgeflows) were probably not picked up. Two othermeasures were also used: a simple measure ofinternational sales as a percentage of the totalrevenues, and a subjective measure of the subsidi-ary’s value-added contribution to the corporation.Both were significant correlated to the first meas-


ure (r = 0.55, 0.17 respectively), but they exhib-ited very weak loadings in the PLS analysis sothey were eventually dropped.

Specialized resourceswas operationalized usingan adapted scale from Roth and Morrison (1992).Respondents were asked to rate five differentsubsidiary capabilities (R&D, manufacturing,marketing, managing international activities, inno-vation and entrepreneurship) relative to other sub-sidiaries in the corporation. While these activitiesare all very different, we found that they allloaded strongly onto a single construct in the PLSanalysis, which we interpreted as representing thesubsidiary’s aggregate level of specializedresources.

Subsidiary initiativewas the most troublesomeconstruct to measure. Questions were wordedcarefully on the basis of previous studies(Birkinshaw, 1995; Bishop and Crookell, 1986;Science Council of Canada, 1980) to identifythe various manifestations of subsidiary initiative,from internal bidding efforts through to skunk-works-like product development. Eight questionswere crafted, which were then reworked severaltimes on the basis of discussions in the question-naire development process. Following the adviceof a reviewer, we subsequently dropped threeof these questions because of questionable facevalidity. The remaining five questions all loadedstrongly onto a single construct in the PLS analy-sis.

Following from the study quoted earlier(Birkinshaw (1985)), we also tried splitting theinitiative construct into two subconstructs: internalinitiative (the pursuit of a market opportunity thatarose inside the corporate system) and externalinitiative (the pursuit of an opportunity that aroseoutside the corporate system). While factor analy-sis suggested that these two subconstructs couldbe distinguished, a provisional PLS analysisshowed that discriminant validity between themwas poor (the path coefficient from one to theother was 0.71). We therefore chose to viewsubsidiary initiative as a single construct.

Subsidiary leadership was operationalizedusing three questions relating to the subsidiary’shistory of strong, internationally respected leaders,the credibility of the leadership with head officemanagers, and the leadership’s efforts atdeveloping middle management. Unfortunatelythese measures were only moderately correlatedwith one another (r = 0.19 to 0.59), with the


result that the PLS program put most of theweighting on the first question only. We thereforedropped the latter two from the analysis.

Entrepreneurial culture. The five highest-loading items from Kuratko et al.’s (1990)intrapreneurial assessment index were used tomeasure entrepreneurial culture. These questionswere concerned with the openness of the subsidi-ary’s working environment to entrepreneurship,risk-taking and innovation. All five items loadedstrongly on a single construct in the PLS analysis.

Subsidiary autonomy.A 7-item scale was takenfrom Roth and Morrison (1992) that asked sub-sidiary managers to identify whether certaindecisions were made in the subsidiary, divisionallevel, or head office. During the PLS analysisfour of the items were dropped because theyloaded very weakly on the construct, leavingthree items.

Communication frequency.Ghoshal’s (1986)measures of communication were used, speci-fically frequency of communication, frequency ofbusiness trips to head office, strength of workingrelations, and sharing of information. The lattertwo items were dropped in the course of the PLSanalysis, leaving the former two which reflectedthe frequency of communication between subsidi-ary and parent company.

Local competition. Beginning with the 7-itemscale developed by Woodcock (1994), weextracted two items, ‘domestic competition isintense’ and ‘competition in this country isextremely high’, which were strongly correlated(r = 0.75). Both these items loaded strongly ontothe same construct, which we interpreted as indi-cating the perceived level of local competition.

Industry globalization. The scale used by Rothand Morrison (1992) was adopted. A factoranalysis revealed one primary factor with eightitems which represented the extent to which theindustry was global. These eight items were thenreduced down to four items during the PLS analy-sis.


RESEARCH FINDINGS

PLS results are generally presented in two stages.In the first stage the ‘measurement model’ ispresented to show that the measures used asoperationalizations of the underlying constructsare reliable and valid. In the second stage, thepath coefficients between constructs can be inter-preted.

The measurement model was assessed by look-ing at the internal consistency between itemsintended to measure the same construct, and thediscriminant validity between constructs. Internalconsistency was determined using the measuresuggested by Fornell and Larcker (1981). Thismeasure is similar to Cronbach’s alpha, thoughmore appropriate because it does not assume thateach item makes an equal contribution to theconstruct. As shown in Table 2, all the constructsexceeded the level of 0.7, which is consideredgood for exploratory research (cf. Nunnally,1978). The discriminant validity of the modelwas assessed by calculating the average varianceextracted for each construct (Fornell and Larcker,1981). Table 3 shows the square root of theaverage variance along the diagonal of the corre-lation matrix. For acceptable discriminant validity,the diagonal elements should be greater than allother entries in the same row and column, as isthe case here.

Tests of hypotheses

The primary output from PLS analysis is thepath coefficients between constructs which areequivalent to standardized regression coefficients.The significance of these paths is calculated usinga jack-knifing technique (Fornell and Barlcay,1983).13 It is unfortunately not possible to testthe goodness-of-fit of a PLS model: the nearestapproximation is the percentage of varianceexplained in the endogenous constructs (i.e., theirR2 values).14 Table 4 indicates the path coef-ficients andR2 values for ‘Model 1’, which isthe model displayed in Figure 3.

13 We also calculated significance levels using the ‘bootstrap’technique offered in the PLS computer package. The jack-knife results were more conservative so they are reported here.14 While fit indices are available with PLS, they are ofquestionable validity because the objective function of PLSis to maximize the explained variance in endogenous con-structs, not to optimize the model.


Table 2. Measurement model

Construct Number of Internalitems consistency

Subsidiary 5 0.88entrepreneurshipSubsidiary leadership 1 1.00Subsidiary autonomy 3 0.81Sub-parent 2 0.76communicationLocal competition 2 0.81Global integration 4 0.73Subsidiary initiative 5 0.92Specialized resources 5 0.77Contributory role 1 1.00

Note: Internal consistency for each construct is calculated as:(Slyi)2/(Slyi)2 + SVar(ei ), wherelyi is the loading for eachitem on the construct andei is the measurement error foreach item.

The statistical analysis revealed some inter-esting relationships. As predicted there werestrong relationships between specialized resourcesand initiative and between initiative and contribu-tory role, but contrary to prediction there wasessentially no relationship between specializedresources and contributory role (path coefficient= −0.03). While there is a strongcorrelationbetween specialized resources and contributoryrole (r = 0.29: see Table 3), the inclusion of sub-sidiary initiative in the model shows that therelationship is spurious. This finding also hasstrong face validity. It suggests that specializedresources are not sufficient, in themselves, tobuild the subsidiary’s contributory role. Rather,subsidiary initiative is necessary to make thespecialized resources known to head office man-agers and thereby to gain recognition for them.

The actions of subsidiary management (strengthof subsidiary leadership, creation of an entrepre-neurial culture) had a strong positive impact onthe development of specialized resources (pathcoefficients 0.31 and 0.35 respectively), as pre-dicted. However, entrepreneurial culture had nodiscernible impact on subsidiary initiative, whilesubsidiary leadership showed small but significantrelationships with both subsidiary initiative andcontributory role.

In terms of the other factors, the level ofindustry globalization had a significant relation-ship with subsidiary initiative, indicating that,after specialized resources, the strongest predictor


Table 3. Discriminant validity

Correlations between constructs

Subsidiary entrepreneurship 0.86Subsidiary leadership 0.165 1.00Subsidiary autonomy 0.180 0.124 0.79Sub-parent communication 0.050 0.021 0.025 0.82Local competition 0.038 −0.088 −0.130 −0.032 0.87Global integration 0.049 0.188−0.005 −0.022 −0.001 0.71Subsidiary initiative 0.134 0.249 0.193 0.105−0.233 0.258 0.91Specialized resources 0.364 0.369 0.113 0.069−0.043 0.164 0.401 0.454Contributory role 0.053 0.272 0.280 0.141−0.388 0.227 0.662 0.29 1.00

Diagonals indicate the square root of the average variance extracted for the construct. Off-diagonals indicate the correlationsbetween constructs in the PLS model. Average variance extracted is calculated asSlyi

2/n where lyi is the loading for eachitem on the construct andn is the number of items.

Table 4. Summary of PLS findings for Models 1 and 2

Hypothesis Path Path coefficient Path coefficient Support forin Model 1 in Model 2 hypothesis?

1 Specialized resources—Contributory role −0.03 −0.04 No2 Subsidiary initiative—Contributory role 0.56*** 0.56*** Yes3 Specialized resources—Subsidiary initiative 0.42*** 0.36*** Yes4a Subsidiary leadership—Subsidiary initiative 0.07* 0.19* Yes4b Entrepreneurial culture—Subsidiary initiative −0.06 0.09* Some5a Subsidiary leadership—Specialized resources 0.31** 0.23* Yes5b Entrepreneurial culture—Specialized resources 0.35** 0.32* Yes6 Subsidiary leadership—Contributory role 0.11* 0.11* Yes8a Subsidiary autonomy—Contributory role 0.13* 0.14* Yes8b Parent–sub. communic.—Contributory role 0.07† 0.07† Some9 Industry globalization—Subsidiary initiative 0.18** 0.22** Yes

10 Local competition—Contributory role −0.23** −0.23** Yes

Variance explained in subsidiary initiative 0.249 0.120Variance explained in specialized resources 0.255 0.373Variance explained in contributory role 0.529 0.529

†p , 0.10; *p , 0.05; **p , 0.01; ***p , 0.001NB: There is one difference between Models 1 and 2: In Model 1 there is a path from specialized resources to subsidiaryinitiative, in Model 2 that path is reversed.

of initiative is the extent to which the subsidiaryis operating in a global industry. The threeremaining variables (subsidiary autonomy,parent–subsidiary communication, and localcompetition) had significant relationships withcontributory role. Of these, all were as predictedexcept local competition, which had a strongnegativerelationship with contributory role. Sim-ply put, this finding suggested that the lowerthe level of local competition, the greater thecontributory role of the subsidiary. We will dis-cuss possible explanations for this finding in thefollowing section.


Additional analysis

Two additional analyses were undertaken. In thefirst we simply reversed the line of causalitybetween initiative and specialized resourcesbecause, as the discussion earlier indicated, itseems likely that causality flows in both direc-tions. The results from this analysis are listedunder ‘Model 2’ in Table 4. As one would predictthis reduced the explained variance in subsidiaryinitiative (R2 from 0.249 to 0.120) and increasedit in specialized resources (R2 from 0.255 to0.373). However, it also caused small changes in


the paths throughout the model, making entrepre-neurial culture a significant predictor of subsidiaryinitiative, and reducing the significance level ofseveral other relationships.

In terms of the mechanics of PLS, the changein path direction between initiative and special-ized resources resulted in much of the variancebetween the independent constructs and contribu-tory role being channeled through subsidiaryinitiative, thus increasing the number of signifi-cant predictors of subsidiary initiative anddecreasing the significance level of the predictorsof contributory role. What this means for subsidi-ary management is harder to say, because thereis no a priori reason to prefer one model overthe other. Our preference is to concentrate on thestrong relationships, i.e., the ones that are signifi-cant in both models, and to interpret the otherswith caution. Table 4 lists the results from bothmodels and summarizes the extent to which eachhypothesis was supported.

The second additional analysis focused on therelationship between subsidiary initiative and con-tributory role. Because our interest in initiativewas restricted to internationally focused efforts, itis perhaps not surprising that a strong relationshipbetween the two constructs was obtained. Toillustrate this point, Figure 4 is a histogram ofthe relationship between the two. It shows thatthe strong correlation is driven primarily by thelow level of initiative in subsidiaries with nointernational responsibilities.

To further understand the relationships between

Figure 4. Histogram showing relationship betweensubsidiary initiative and contributory role


initiative and contributory role, we performedan additional PLS analysis using only the 99subsidiaries that had international responsibilities(i.e., where contributory role was greater than0%). This analysis resulted in a nonsignificantpath of 0.09 between initiative and contributoryrole. In other words,the role of initiative indistinguishing between medium- and high-contributory role subsidiaries is not significant.Initiative appears to have an important role toplay in generating international responsibilities inthe first place, but a questionable role in increas-ing the magnitude of those international responsi-bilities.15

DISCUSSION AND CONCLUSIONS

The findings from this research offer a numberof important insights into the process throughwhich MNC subsidiaries enhance their contribu-tory role. First, it is clear that the internal work-ings of the subsidiary matter, though perhaps notquite in the manner we had expected. Subsidiaryleadership and an entrepreneurial culture appearto promote the development of specializedresources, which in turn are strongly associatedwith the existence of subsidiary initiative. How-ever, there is no clear relationship between speci-alized resources and contributory role exceptthrough initiative. This goes somewhat againstthe findings of a number of prior studies (e.g.,Birkinshaw and Morrison, 1996; Roth and Mor-rison, 1992) that had equated subsidiary resourceswith the development of world product mandates.Our interpretation suggests that the prior studieshad not given due consideration to subsidiaryinitiative as the means by which specializedresources impact the subsidiary’s contributoryrole. However, as the additional analysis aboveindicated (Figure 4) the relationship between

15 One further insight into the relationship between initiativeand contributory role was obtained. One question on thequestionnaire asked the subsidiary presidents to assess theapproximate percentage of their international responsibilitiesthat were ‘given’ to them by the parent company vs thosethat were ‘earned’ by the subsidiary through initiative (cf.Crookell and Morrison, 1990). The mean response was 63percent earned, 37 percent given. Of course, the danger ofsocially desirable response here is very great so this findingneeds to be interpreted with care, but it suggests that subsidi-ary managers themselves believe there is an important relation-ship between initiative and contributory role.


initiative and contributory role is rather complex.It appears that initiative is an important discrimin-ator between high- and low-contributory role sub-sidiaries, but that the role of initiative may bemore critical in the early stages of contributoryrole development than in its subsequent growth.

The second key observation comes from thesurprising relationship between local competitionand both contributory role and subsidiary initia-tive. Simply stated, the survey evidence showedthat subsidiaries were more likely to have highcontributory roles and undertake initiative if do-mestic market competition was perceived to beweak. One explanation for this is a simple percep-tual bias, in that high-contribution subsidiaries arepredominantly exporters and therefore do notthink in terms of local competition, but thestrength of the relationship suggests this is notsufficient explanation. An alternative explanationis that the original hypothesis was not correctlymotivated. We grounded the hypothesis in Port-er’s (1990) thinking on national competitiveness.Specifically, we argued that the national or localbusiness environment, and in particular the levelof competition within it, would drive competitiveupgrading by participating firms. To the extentthat high-contribution subsidiaries were morecompetitive than low-contribution subsidiaries, wesuggested, one would expect to see a higherlevel of local competition in high-contributionsubsidiaries. In reality we found a strong relation-ship in the opposite direction.

Porter (1990) did not, however, give explicitconsideration to foreign ownership, other than tosuggest a process of ‘selective tapping’ by foreignsubsidiaries in leading-edge clusters. It wouldappear, then, that Porter’s study applies primarilyto cases where there are clearly defined ‘clusters’of related industries that are recognized as world-class, and which MNCs seek access to throughtheir subsidiaries. For the countries in this study,there is little evidence of such clusters: Canadaand Sweden both have leading-edge clusters inthe natural resource and heavy industry sectors,but these were underrepresented in the subsidiarysample; and Scotland has a cluster of electronicscompanies in ‘Silicon Glen’ but this is not aleading-edge cluster in terms of innovation andspin-off companies. Rather than building firm-specific advantages because of thestrengthof thelocal business environment, the subsidiaries inthis study appeared to build them on account


of the industry’s relativeweakness. The highcontributory roles appeared to be gained in suchinstances because the subsidiary was in a rela-tively protected niche. The suggestion, which can-not be tested here, is that these subsidiaries maymake relatively low-quality contributions to theMNC’s firm-specific advantage, which are lesssusceptible to upgrading through local compe-tition.16

The third important insight is that the parent–subsidiary relationship also had an important roleto play in the development of the subsidiary’scontributory role and the presence of subsidiaryinitiative. Subsidiary autonomy, in particular, hadan important influence on both initiative and con-tributory role, while parent–subsidiary communi-cation had a small positive impact on contributoryrole. When viewed in terms of the ‘competing’hypotheses described earlier, it is therefore notpossible to choose decisively between environ-mental determinism, head office assignment, andsubsidiary choice. We can clearly state that allthree perspectives are important, but because ourchoice of constructs was not comprehensive itwould be inappropriate to indicate that one per-spective is more important than the other two.

The role of subsidiary initiative

One of the key objectives of this research wasto understand the part played by subsidiary initia-tive in developing a subsidiary’s contributory role.We know from the data presented here and fromresearch interviews that initiative is absent in alarge percentage of subsidiaries. Where initiativeis present, its relationship to specialized resourcesand contributory role appears to be positivebutwith a few reservations. The data suggested someof the grounds for these reservations; this dis-cussion will consider some of the theoreticalarguments.

Most obviously, initiative is often seen by par-ent managers as subversive, that is, evidence ofsubsidiary managers acting in their own or theircountry’s interests rather than in the interests ofthe MNC as a whole. Moreover, this concern is

16 In more general terms, it is also possible that Porter’s(1990) thinking has relatively less applicability to small per-ipheral economies such as Canada, Sweden, and Scotlandthan to those economies with more dynamic clusters and moreleading-edge subsidiaries such as Japan or the United States.


not groundless. There are well-known cases ofsubsidiary managers deliberately building theirown ‘empires’, and there are more ambiguouscases where the entrepreneurial actions of subsidi-ary management could be interpreted in variousways, depending on one’s perceptions. The com-bination of bounded rationality on the part ofparent management and the decreasing depen-dence of this subsidiary on the parent (Prahaladand Doz, 1981) results in situations where parentcompany managers have to accept the actions ofsubsidiary management in good faith, or stifletheir ideas through veto.

However, if assumptions of opportunisticbehavior are temporarily suspended, subsidiaryinitiative has a potentially very powerful role toplay in the efficiency of the corporate system.Working on the basis that the MNC can bemodeled as an internal market (Ghoshal and Bart-lett, 1991), it is apparent that some of the ineffic-iences in that market arise through the stickinessof existing relationships—retaining the sameinternal component supplier, for example, justbecause it has always fulfilled a certain service.Subsidiary initiative provides a means of lubricat-ing the internal market, in that it makes otherentities within the market aware of the subsidi-ary’s distinctive capabilities and the uses to whichthey could be put. In essence, initiative enhancesthe flow of information which,ceteris paribus,improves market efficiency. To some extent thisargument is very obvious, but it is nonethelessimportant because our suspicion is that theresources and capabilities of subsidiary units arevery poorly understood by parent and sister com-pany managers around the world. If the MNC isto effectively utilize its far-flung resources(Bartlett and Ghoshal, 1986), it must first under-stand what those resources are and where theyreside. And to the extent that subsidiary managersunderstand their resources better than anyone else,it is their responsibility to proactively seek outways of utilizing those resources more effectively.

While much of this argument is speculative, itis grounded in the observation that subsidiaryinitiative is a pervasive phenomenon and tiedto the conceptual model of the MNC as aninterorganizational network. Much as Kirzner’s(1973) entrepreneurs enhanced market efficiencythrough alertness to new opportunities, thesuggestion is that subsidiary unts can enhance thedistribution of activities within the MNC through


initiative. Whether these benefits are sufficient tocounteract the dangers of opportunism and controlloss is then a separate question.

A final point that should be made here is thatwe donot see a high level of subsidiary initiativeas a driver of product diversification for theMNC. For initiatives to be accepted by the corpo-rate headquarters they must be aligned with theMNC’s existing strategic priorities, otherwise theyare likely to be viewed as self-interested behavior.We see initiatives, particularly those emanatingfrom peripheral countries like Canada, Scotland,and Sweden, as exploring opportunites at themargins of the corporation’s existing productportfolio, by building on existing technologiesand competencies rather than creating entirelynew ones. As such, this is entirely consistentwith the product diversification literature in whicha focus on core technologies and/or products istypically associated with high performance(Chatterjee and Wernerfelt, 1991; Markides,1995).

To conclude, this paper provided support forthe emerging view that subsidiaries are significantcontributors to the firm-specific advantage of theMNC, though it also raised several additionalquestions. It was not, of course, possible todemonstrate the link between contributory roleand firm-specific advantage. Contributory rolerepresented the extent to which the subsidiary hasbeen assigned responsibility for a value-addingactivity on the part of the MNC; the ability ofthe corporate system to effectively leverage thatactivity in the global market is what eventuallymakes the subsidiary’s resources part of the firm-specific advantage. Clearly there is scope forfuture research in examining this link.

This study had a number of significant limi-tations. Perhaps the most significant was thedecision to collect all data from the subsidiarygeneral manager. While this was necessary forcertain constructs, it probably created some biasin others. It is therefore recommended for futureresearch that both parent and subsidiary managersare polled where possible. Second, the focus onCanada, Scotland, and Sweden meant that gen-eralizability was limited to peripheral countriesin developed regions. We would therefore expectto see rather different relationships exhibited inother settings, such as subsidiaries in largedeveloped countries or subsidiaries in lessdeveloped regions.


Finally, the focus on subsidiary initiative pro-vides at least provisional evidence that the sub-sidiary candrive the development of firm-specificadvantage creation rather than just be a passivecontributor. This finding represents a subtle shiftin thinking on the role of the subsidiary in theMNC, because it tips the balance of responsibilityfor role development towards the subsidiary. Italso represents further evidence that the sourcesof firm-specific advantage in MNCs are increas-ingly gained outside the home country.

ACKNOWLEDGEMENT

We are grateful to comments from Gunnar Hed-lund, Jonas Ridderstra˙le, O

¨rjan Solvell, Udo

Zander, Jan Johanson, and other members of theInstitute of International Business and Universityof Uppsala. Earlier versions of this paper werepresented at the Academy of Management andAcademy of International Business meetings in1996, and a conference at the Australian GraduateSchool of Management, 1997.

REFERENCES

Barney, J. (1991). ‘Firm resources and sustained com-petitive advantage’,Journal of Management, 17(1),pp. 99–120.

Bartlett, C. A. (1979). Multinational structural evolu-tion: The changing decision environment in inter-national divisions’, unpublished doctoral dissertation,Harvard University.

Bartlett, C. A. and S. Ghoshal (1986). ‘Tap yoursubsidiaries for global reach’,Harvard BusinessReview, 64(6), pp. 87–94.

Birkinshaw, J. M. (1995). Entrepreneurship in multi-national corporations: The initiative process in Cana-dian subsidiaries’, unpublished doctoral dissertation,Western Business School.

Birkinshaw, J. M. and N. Hood (1997). ‘An empiricalstudy of development processes in foreign-ownedsubsidiaries in Canada and Scotland’,ManagementInternational Review, 4 (forthcoming).

Birkinshaw, J. M. and A. J. Morrison (1996). ‘Con-figurations of strategy and structure in multinationalsubsidiaries’,Journal of International Business Stud-ies, 26(4), pp. 729–794.

Birkinshaw, J. M., A. J. Morrison and J. Hulland(1995). ‘Structural and competitive determinants ofa global integration strategy’,Strategic ManagementJournal, 16(8), pp. 637–655.

Bishop, P. and H. H. Crookell (1986). ‘Specializationin Canadian subsidiaries’. In D. G. McFetridge (ed.),


Canadian Industry in Transition. University ofToronto Press, Toronto, pp. 305–386.

Bower, J. L. (1970).Managing the Resource AllocationProcess. Irwin, Homewood, IL.

Brandt, W. K. and J. M. Hulbert (1977). ‘Headquartersguidance in marketing strategy in the multinationalsubsidiary’, Columbia Journal of World Business,12, pp. 7–14.

Burgelman, R. A. (1983). ‘A model of the interactionof strategic behavior, corporate context and the con-cept of strategy’,Academy of Management Review,8(1), pp. 61–70.

Cantwell, J. (1989).Technological Innovation andMultinational Corporations. Basil Blackwell,Oxford.

Chatterjee S. and B. Wernerfelt (1991). ‘The linkbetween resources and type of diversification:Theory and evidence’,Strategic Management Jour-nal, 12(1), pp. 33–48.

Child, J. (1972). ‘Organization structure, environmentand performance: The role of strategic choice’,Soci-ology, 6, pp. 1–22.

Crookell, H. H. (1986). ‘Specialization and inter-national competitiveness’. In H. Etemad and L. Seg-uin Dulude (eds.),Managing the Multinational Sub-sidiary. Croom Helm, London, pp. 102–111.

Crookell, H. H. and A. J. Morrison (1990). ‘Subsidiarystrategy in a free trade environment’,Business Quar-terly, Autumn, pp. 33–39.

D’Cruz, J. R. (1986). ‘Strategic management of subsidi-aries.’ In H. Etemad and L. Seguin Dulude (eds.),Managing the Multinational Subsidiary. CroomHelm, London, pp. 75–89.

Dunning, J. H. (1980). ‘Toward an eclectic theoryof international production: Some empirical tests’,Journal of International Business Studies, 11(1),pp. 9–31.

Dunning, J. H. (1988). ‘The eclectic paradigm of inter-national production: A restatement and some pos-sible extensions’,Journal of International BusinessStudies, 19, pp. 1–31.

Dunning, J. H. (1993).Multinational Enterprises andthe Global Economy. Addison-Wesley, Wokingham.

Etemad, H. and L. Seguin Dulude (1986).Managingthe Multinational Subsidiary. Croom Helm, London.

Fornell, C. (1984). ‘A second generation of multivariateanalysis: Classification of methods and implicationsfor marketing research’, working paper, Universityof Michigan.

Fornell, C. and D. Barclay (1983). ‘Jackknifing: Asupplement to Lohmo¨ller’s LVPLS program’, Gradu-ate School of Business Administration, Universityof Michigan, Ann Arbor, MI.

Fornell, C and F. Bookstein (1982). ‘A comparativeanalysis of two structural equation models: LISRELand PLS applied to consumer exit-voice theory’,Journal of Marketing Research, XIX , pp. 440–452.

Fornell, C. and D. F. Larcker (1981). ‘Structural equ-ation models with unobservable variables andmeasurement error: Algebra and statistics’,Journalof Marketing Research, 18, pp. 382–388.

Fornell, C., P. Lorange and J. Roos (1990). ‘Thecooperative venture formation process: A latent vari-


able structural modeling approach’,ManagementScience, 36(10), pp. 1246–1255.

Forsgren, M. (1995). ‘Developing centres of excellencein multinational firms’, research proposal, Copen-hagen Business School.

Garnier, G. H. (1982). ‘Context and decision makingautonomy in foreign affiliates of U.S. multinationalcorporations’,Academy of Management Journal, 25,pp. 893–908.

Gates, S. R. and W. G. Egelhoff (1986). ‘Centralizationin headquarters–subsidiary relationships’,Journal ofInternational Business Studies, 17(2), pp. 71–92.

Ghoshal, S. (1986). ‘The innovative multinational: Adifferentiated network of organizational roles andmanagement processes’, unpublished doctoral disser-tation, Harvard Business School, Boston, MA.

Ghoshal, S. and C. A. Bartlett (1991). ‘The multi-national corporation as an interorganizational net-work’, Academy of Management Review, 15(4),pp. 603–625.

Ghoshal, S. and C. A. Bartlett (1994). ‘Linking organi-zational context and managerial action: The dimen-sions of quality of management’,Strategic Manage-ment Journal, Summer Special Issue,15, pp. 91–112.

Ghoshal, S. and H. Nohria (1989). ‘Internal differen-tiation within multinational corporations’,StrategicManagement Journal, 10(4), pp. 323–337.

Gupta, A. K. and V. Govindarajan (1994). ‘Organizingfor knowledge within MNCs’,International BusinessReview, 3(4), pp. 443–457.

Hamel, G. and C. K. Prahalad (1994).Competing forthe Future. Harvard Business School Press, Bos-ton, MA.

Hedlund, G. (1986). ‘The hypermodern MNC: A heter-archy?’ Human Resources Management, 25, pp. 9–36.

Hymer, S. H. (1976).The International Operations ofNational Firms: A Study of Direct Investment. MITPress, Cambridge, MA.

Jarillo, J.-C. and J. I. Martinez (1990). ‘Different rolesfor subsidiaries: The case of multinational corpo-rations’, Strategic Management Journal, 11(7),pp. 501–512.

Johansson, J. K. and G. S. Yip (1994). ‘Exploitingglobalization potential: U.S. and Japanese strategies’,Strategic Management Journal, 15(8), pp. 579–601.

Kanter, R. M. (1985).The Change Masters. Simon &Schuster, New York.

Kirzner, I. (1973).Competition and Entrepreneurship.University of Chicago Press, Chicago, IL.

Kobrin, S. J. (1991). ‘An empirical analysis of thedeterminants of global integration’,Strategic Man-agement Journal, Summer Special Issue,12,pp. 17–31.

Kuratko, D. F., R. V. Montagno, and J. S. Hornsby(1990). ‘Developing an intrapreneurial assessmentinstrument for an effective corporation entrepre-neurial environment’,Strategic Management Jour-nal, Summer Special Issue,11, pp. 49–58.

Leksell, L. (1984). ‘Headquarters–subsidiary relation-ships in multinational corporations’, doctoral thesis,Institute of International Business, Stockholm Schoolof Economics.


Markides, C. C. (1995).Diversification, Refocusing andEconomic Performance. MIT Press, Cambridge,MA.

Nunnally, J. C. (1978).Psychometric Theory(2nd ed.).McGraw Hill, New York.

Otterbeck, L. (1981). The Management ofHeadquarters–Subsidiary Relations in MultinationalCorporations. Gower, Hampshire, U.K.

Peters, T. J. and R. H. Waterman (1982).In Searchof Excellence. Harper & Row, New York.

Pinchott, G. III (1985). Intrapreneuring. Harper &Row, New York.

Porter, M. E. (1980).Competitive Strategy. Free Press,New York.

Porter, M. E. (1986).Competition in Global Industries.Harvard Business School Press, Boston, MA.

Porter, M. E. (1990).The Competitive Advantage ofNations. Free Press, New York.

Prahalad, C. K. (1976). ‘The strategic process in amultinational corporation’, unpublished doctoral dis-sertation, School of Business Administration, Harv-ard University.

Prahalad, C. K. and Y. L. Doz (1981). ‘An approachto strategic control in MNCs’,Sloan ManagementReview, Summer, pp. 5–13.

Roth, K. and A. J. Morrison (1992). ‘Implementingglobal strategy: Characteristics of global subsidiarymandates’,Journal of International Business Studies,23(4), pp. 715–736.

Rugman, A. (1981).Inside the Multinationals: TheEconomics of Internal Markets. Croom Helm, Lon-don.

Rugman, A. M. and J. Bennett (1982). ‘Technologytransfer and world product mandating in Canada’,Columbia Journal of World Business, 17(4),pp. 58–62.

Rugman, A. and A. Verbeke (1992). ‘A note on thetransnational solution and the transaction cost theoryof multinational strategic management’,Journal ofInternational Business Studies, 23(4), pp. 761–772.

Science Council of Canada (1980).Multinationals andIndustrial Strategy: The Role of World Product Man-dates. Science Council of Canada, Supply and Ser-vices, Ottawa, Canada.

Szulanski, G. (1995). ‘An empirical investigation ofthe barriers to transfer of best practice inside thefirm’, unpublished doctoral dissertation, INSEAD,France.

Teece, D. (1977). ‘Technology transfer by multinationalfirms: The resource cost of transferring technologicalknow-how’, Economic Journal. 87, pp. 242–261.

Vernon, R. (1966). ‘International investments and inter-national trade in the product cycle’,Quarterly Jour-nal of Economics, 80, pp. 190–207.

Wernerfelt, B. (1984). ‘A resource-based view of thefirm’, Strategic Management Journal, 5(2),pp. 171–180.

Westney, D. E. (1994). ‘Institutionalization theory andthe multinational corporation.’ In S. Ghoshal and D.E. Westney (eds.),Organization Theory and theMultinational Corporation. St Martin’s Press, NewYork, pp. 53–76.

White R. E. and T. A. Poynter (1984). ‘Strategiesfor foreign-owned subsidiaries in Canada’,BusinessQuarterly, Summer, pp. 59–69.


Woodcock, P. (1994). ‘The greenfield vs. acquisitionentry mode decision process’, unpublished doctoraldissertation, Western Business School.

Young, S., N. Hood and E. Peters (1994). ‘Multi-national enterprises and regional economic develop-ment’, Regional Studies, 28(7), pp. 657–677.

APPENDIX

Wording of questionnaire items

Contributory Role . Does your subsidiary com-pany have any international responsibilities orworld mandates (that is, does it undertake anyactivity such as manufacturing, R&D or productmanagement on behalf of the corporation as awhole?) If yes, please estimate the approximatepercentage of your subsidiary company’s revenuesthat are gained as a result of your internationalresponsibilities. (0 to 100%).

Subsidiary initiative . To what extent have thefollowing activities occurred in your subsidiaryover the past 10 years? (1) new productsdeveloped in (e.g.) Sweden and then sold inter-nationally; (2) successful bids for corporateinvestments in Sweden; (3) new internationalbusiness activities that were first started inSweden; (4) enhancements to product lines whichare already sold internationally; (5) new corporateinvestments in R&D or manufacturing attractedby Swedish management. 1= never, 5= plenti-fully.

Specialized resources. Indicate your capabilityor distinctive expertise in the following areasrelative to other subsidiaries in the corporation:(1) product or process R&D; (2) manufacturingcapability; (3) marketing capability; (4) managinginternational activities; (5) innovation andentrepreneurship. 1= far below average, 7= farabove average.

Subsidiary entrepreneurship. Indicate howcharacteristic each of the following statements isin describing your subsidiary: (1) there is topmanagement support of entrepreneurial activity;(2) top management has experience with inno-vation; (3) individual risk-takers are recognizedwhether successful or not; (4) there is encourage-ment for calculated risks; (5) risk-taker is con-


sidered a positive attribute. 1= strongly disagree,7 = strongly agree.

Subsidiary leadership. Indicate how character-istic the following statements are in describingyour subsidiary: (1) the subsidiary has a historyof strong, internationally respected leaders; (2)the credibility of subsidiary top management ishigh; (3) the subsidiary CEO or president workswith managers to focus their efforts towards thesubsidiary’s objectives. 1= strongly disagree, 7= strongly agree.

Subsidiary autonomy. Which level in your busi-ness unit has authority to make the followingdecisions? Circle the most appropriate decisionlevel based on the following (1, decision madein the subsidiary company; 2, decision made atthe sub-corporate level; 3, decision made bycorporate headquarters): (1) changes in productdesign; (2) subcontracting out large portions ofthe manufacturing instead of expanding the sub-sidiary’s own facilities; (3) switching to a newmanufacturing process.

Parent–subsidiary communication. How oftendo senior managers in your subsidiary communi-cate with their counterparts and bosses in headoffice (1 = daily, 5 = less than once a month);how often do senior and middle managers in yoursubsidiary make business trips to head office? 1= twice a month or more, 5= less than oncea year.

Local competition. Indicate how characteristiceach of the following statements is in describingyour business environment: (1) competition inthis country is extremely intense; (2) domesticcompetition is intense. 1= strongly disagree, 7= strongly agree.

Global integration. Indicate how characteristiceach of the following statements is in describingyour industry: (1) international competition isintense; (2) business activities are susceptible toglobal scale economies; (3) product awarenessexists worldwide; (4) new product introductionsoccur in all major markets simultaneously. 1=strongly disagree, 7= strongly agree.

240J.

Birkin

sha

w,

N.

Ho

od

an

dS

.Jo

nsso

n

Table A1. Person correlations between individual questions

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

1. Does your subsidiary have any 1international responsibilities ormandates?

2. New products developed in 0.56 1Sweden then soldinternationally?

3. Successful bids for corporate 0.51 0.72 1investment in Sweden?

4. New international business 0.57 0.82 0.72 1activities first started inSweden?

5. Enhancements to product lines 0.70 0.78 0.79 0.82 1which are already soldinternationally?

6. New corporate investments in 0.55 0.68 0.80 0.70 0.79 1R&D or manufacturing startedby Swedish mgmt?

7. Capability in product or process 0.32 0.43 0.36 0.47 0.42 0.43 1R&D?

8. Capability in manufacturing? 0.06 0.17 0.22 0.26 0.21 0.22 0.36 19. Capability in marketing? −0.06 0.01 0.03 0.03−0.00 0.03 0.22 0.06 1

10. Capability in managing 0.34 0.34 0.28 0.37 0.40 0.29 0.31 0.27 0.29 1international activities?

11. Capability in managing 0.04 0.16 0.18 0.20 0.10 0.23 0.30 0.09 0.37 0.16 1innovation and entrepreneurship?

12. There is top management 0.07 0.15 0.09 0.18 0.09 0.12 0.20 0.08 0.18 0.09 0.45 1support of entrepreneurialactivity

13. Top management has experience 0.05 0.13 0.15 0.14 0.08 0.17 0.28 0.23 0.09 0.07 0.44 0.67 1with innovation

14. Individual risk-takers are 0.02 0.06 0.03 0.08 0.00 0.09 0.16 0.08 0.12−0.03 0.38 0.62 0.61 1recognized whether successfulor not

1998

JohnW

iley&

Sons,

Ltd.S

trat.

Mg

mt.

J.,Vol.

19,221

–241

(1998)

Su

bsid

iary

Initia

tivein

Mu

ltina

tion

al

Co

rpo

ratio

ns

241

Table A1. (cont.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

15. There is encouragement for 0.05 0.10 0.09 0.12 0.05 0.11 0.21 0.17 0.14 0.03 0.34 0.64 0.65 0.75 1calculated risk

16. Risk-taker is considered a −0.05 0.05 0.05 0.07 0.00 0.04 0.13 0.16 0.09−0.08 0.37 0.60 0.62 0.72 0.80 1positive attribute

17. The subsidiary has a history of 0.28 0.25 0.19 0.26 0.30 0.26 0.17 0.14 0.13 0.43 0.13 0.14 0.12 0.09 0.10 0.02 1strong, internationally respectedleaders

18. Decision to change product 0.17 0.20 0.08 0.16 0.12 0.12 0.18 0.01 0.06 0.09 0.24 0.20 0.15 0.06 0.03 0.03 0.09 1design is made at what level?

19. Decision to subcontract out 0.17 0.12 0.03 0.12 0.10−0.01 0.10 0.15 0.11 0.07 0.08 0.11 0.15 0.06 0.13 0.12 0.06 0.33 1portions of manufacturing iswhat at what level?

20. Decision to switch to a new 0.32 0.19 0.18 0.21 0.21 0.16 0.25 0.02 0.06 0.17 0.15 0.11 0.18 0.13 0.08 0.09 0.14 0.34 0.38 1manufacturing process is madeat what level?

21. How often do senior managers−0.06 0.05 0.05 0.04 0.01 0.12 0.03 0.05 0.02 0.07 0.09 0.10 0.12 0.12 0.19 0.14 0.11−0.18 0.01−0.01 1communicate with others inhead office?

22. How often do managers in your 0.14 0.19 0.12 0.13 0.10 0.09 0.10 0.03 0.00 0.08 0.13 0.12−0.05 0.03 0.03 0.06−0.06−0.01 0.04−0.10 0.39 1subsidiary make business tripsto HQ?

23. Competition in this country is −0.25−1.15−0.09−0.06−0.13−0.02−0.01 0.02 0.17−0.07 0.09−0.05 0.03 0.06 0.04 0.08 0.03 0.01−0.03−0.06 0.10 0.09 1extremely intense

24. The level of domestic com- −0.40−0.28−0.15−0.22−0.27−0.19−0.14 0.02 0.12−0.12 0.11 0.06 0.05 0.05 0.05 0.11−0.12−0.06−0.09−0.11−0.05 0.02 0.57 1petition is intense

25. International competition is 0.19 0.17 0.20 0.23 0.22 0.26 0.23 0.12−0.02 0.25 0.04 0.07 0.12 0.05 0.16 0.09 0.26−0.03 0.10 0.09 0.03 0.01 0.05 0.02 1intense

26. Business activities are 0.14 0.07 0.22 0.16 0.16 0.13 0.06 0.17 0.02 0.07 0.03−0.04 0.07−0.04 0.01 0.07 0.06−0.10 0.07 0.02−0.09 0.05−0.02 0.14 0.39 1susceptible to global scaleeconomies

27. Product awareness exists 0.01−0.03 0.07 0.06 0.06 0.16 0.14 0.04 0.04−0.03 0.03 0.05 0.06 0.01 0.06 0.04 0.07−0.15−0.08−0.03 0.01 0.00 0.07 0.05 0.49 0.39 1worldwide

28. New product introductions occur 0.18 0.03 0.11 0.10 0.13 0.12 0.01−0.02−0.10 0.07−0.01−0.08−0.06−0.02−0.06−0.04 0.12−0.15−0.22−0.05 0.05−0.04−0.08−0.12 0.09 0.16 0.27 1in all major marketssimultaneously

Signifance level of correlations can be interpreted using the following critical values:p = 0.05, r = 0.15; p = 0.01, r = 0.19; p =0.001, r = 0.25.

1998

JohnW

iley&

Sons,

Ltd.S

trat.

Mg

mt.

J.,Vol.

19,221

–241

(1998)

building firm-specific advantages in multinational corporations

Documents