Autonomy and Dependence_SMJ.2002

8/10/2019 Autonomy and Dependence_SMJ.2002

1/21

Strategic Management JournalStrat. Mgmt. J., 23: 881 901 (2002)

Published online 2 July 2002 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/smj.260

AUTONOMY AND DEPENDENCE OF INTERNATIONALCOOPERATIVE VENTURES: AN EXPLORATION OF

THE STRATEGIC PERFORMANCE OF U.S. VENTURES

IN MEXICO

JAMES A. ROBINS,1* STEPHEN TALLMAN2 andKARIN FLADMOE-LINDQUIST21 Department of Management, City University of Hong Kong, Hong Kong2 David Eccles School of Business, University of Utah, Salt Lake City, Utah, U.S.A.

Researchers have begun to view international cooperative ventures as complex, multipartyorganizations in which foreign and local firms and the venture itself all have distinctive roles.This approach has important implications for the venture strategies of foreign firms in emergingeconomies. This study explores relationships between the resource contributions of parent firmsand U.S. managers assessment of venture performance in a sample of established U.S.Mexicanventures. The research suggests that mature cooperative ventures are expected to achieveautonomy from parent firms in key areas at the same time that certain forms of strategicdependency also are important to success.Copyright 2002 John Wiley & Sons, Ltd.

Research on international cooperative strategieshas evolved rapidly since the Second World War.Early postwar studies focused on the multinational

corporation, treating cooperative strategies as theoutcome of investment restrictions and host coun-try partners as generic representatives of local mar-kets. Macroeconomic theories provided the basisfor many of these studies, and research typicallyemphasized patterns of corporate expansion fromhome markets (Caves, 1971; Hymer, 1960; Kindle-berger, 1969; Vernon, 1971).

A view of cooperative strategies as partner-ships in which both foreign and host country firmsplay significant roles emerged in the 1980s. Asstudies of international ventures began to givethe host country partner an independent identity,

organizational and strategic theories also took ongreater importance. Topics such as partial inter-nalization, joint venturing, strategic motives foralliance formation, and partner selection became

Key words: cooperative ventures; Mexico; internationalstrategy; resource-based strategy; emerging economy*Correspondence to: James A. Robins, Department of Man-agement, City University of Hong Kong, Tat Chee Avenue,Kowloon, Hong Kong.

central research concerns (Contractor and Lorange,1988; Geringer, 1988).

In the last decade there has been growing recog-

nition that cooperative strategies can be viewed asmore than just partnerships between a host countryorganization and a foreign firm. A view has begunto emerge of international cooperative venturesas three-party organizations in which foreign andlocal firmsand the venture itselfall are seen ashaving distinctive, interdependent roles (Lorange,1996). Although a good deal of empirical work hasbeen done on cooperative strategies (Beamish andKilling, 1997a, 1997b, 1997c), this three-party per-spective has not been used to examine strategiesfor emerging economies. Many fundamental ques-tions about the ways that interrelationships among

participants may influence the success of interna-tional cooperative strategies remain to be explored.

This study uses the concept that cooperativeventures can be treated as three-party arrange-ments to analyze cooperative strategies of U.S.firms in Mexico. The study examines the rolesthat managers in U.S. firms see for their ownfirms, their Mexican partners, and their coopera-tive ventures, and it analyzes the impact of thoseinterrelationships on strategic performance. The

Copyright 2002 John Wiley & Sons, Ltd. Received 8 July 1999

Final revision received 26 March 2002


2/21

882 J. A. Robins, S. Tallman and K. Fladmoe-Lindquist

research indicates that U.S. managers find impor-tant ongoing roles for both their own firms and

their local partners in cooperative ventures, butthat certain types of independence for ventures alsomay be vital to success.

The study makes three basic contributions tothe research on international cooperative strate-gies. It extends the three-party approach to emerg-ing economy ventures by developing new theoryand by examining empirical data on the coop-erative strategies of U.S. firms in Mexico. Theresearch also expands the scope of theories orig-inally developed for analysis of the organizationof multinational corporations and indicates how

those approaches may shed light on a very differenttopic: the performance of international cooperativeventures. Finally, the study advances understand-ing of how the success of international cooperativestrategies may be influenced by relations betweenventures and parent firms, particularly dependenceon parent firms for different types of resources.Although practicing managers might be familiarwith the idea that a degree of independence fromparent firms can be important to ICV success, thisconcept is not well established in the literature ofcooperative strategy.

The cooperative strategies examined in thisstudy involved established business ventures withsome longevity. ICVs of this type were chosenbecause cooperative strategies evolve over timeand ventures may not achieve patterns of operationthat support long-term performance until they havebeen through earlier stages of development. Crite-ria for judging performance also can be expectedto differ between established ventures and cooper-ative strategies that are still in formative stages.

The paper is organized into two major sections.The first part of the paper examines the roles of thethree parties in cooperative strategies the foreign

partner, the local partner, and the venture andoutlines a set of general premises about the linksbetween resource contributions by participant firmsand strategic performance. These ideas provide thebasis for a structural model of the relationshipsbetween resource contributions and strategic per-formance. The second part of the paper presentsan empirical investigation of this structural modelusing data on the international cooperative strate-gies of U.S. firms in Mexico. Latent variable par-tial least squares (PLS) techniques are used toderive estimates of relationships between resourcecontributions of partner firms and the evaluation

of cooperative venture strategic performance byU.S. managers. The paper concludes by discussing

some of the implications of this work for researchon international strategy and the management ofinternational cooperative ventures.

COOPERATIVE STRATEGIES: ATHREE-PARTY APPROACH

This study draws upon the OLI perspective asso-ciated with Dunning (1988), Teece (1981, 1986),and others. However, certain key assumptions ofthe OLI approach are altered for research on ICV

performance. The OLI approach was developed toexplain the economic conditions that favor onespecific form of organization the multinationalcorporationunder the assumption that observedforms of organization are efficient. This study isnot concerned with the occurrence or likelihood ofa specific form of organization or governance, andit does not assume economic efficiency.

Rather than assuming full internationalizationand economic efficiency, we are interested in fac-tors that may affect the relative performance ofa variety of ICVs. Because the research dealswith strategic performance rather than form of

organization or governance of ventures it is nec-essary to adopt a broad definition of interna-tional cooperative venture (Contractor and Lor-ange, 1988). Prior research has indicated thatboth shared-equity and nonshared-equity coop-erative ventures may operate with considerableautonomy and face similar challenges (Hagedoorn,1990; Johnson et al., 1997; Osborn and Hage-doorn, 1997; Sarkan, Cavusgil and Evirgen, 1997;Yan and Gray, 1994). This study emphasizes issuesthat may affect performance in a variety of differ-ent cooperative ventures, and shared-equity joint

ventures are not distinguished from other typesof cooperative ventures for those purposes (Florin,1997; Mjoen and Tallman, 1997).

An OLI perspective on resource dependence

The OLI perspective has argued that interna-tional business enterprises require a combination ofstrategic resources provided by foreign firms andlocally based resources in order to achieve com-petitive advantage (Dunning, 1988; Teece, 1986).The OLI work focuses on multinational corpo-rations and emphasized the conditions that make

Copyright 2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 881 901 (2002)


3/21

Autonomy and Dependence of International Cooperative Ventures 883

wholly integrated enterprise efficient. However,more recent research has begun to apply these

concepts to a broader spectrum of organizations,arguing that the creation of a complementary com-bination of resources or capabilities from part-ner firms is one of the most important objec-tives of international cooperative strategy (Con-tractor and Lorange, 1988; Chi, 1994; Tallman,Fladmoe-Lindquist, and Robins, 1997; Madhokand Tallman, 1998). As emerging economies opento global competition, firms must deal with a widerange of international competitors at the same timethat they face increased pressure for efficiency inlocal operations (Teece, 1983). The dual impera-

tives of global and host country markets demandtwo different types of resources: strategic resourcesthat support international competitive advantageand complementary resources that promote successin the local market (Tallman et al., 1997).

Implementation of cooperative strategies alsorequires a variety of resources that may not be effi-ciently acquired from either foreign or host coun-try partners. Two different types of resources mayhave this characteristic: basic inputs such as rawmaterials that can be directly sourced in compet-itive markets, and less tangible resources such asoperating routines that are most effectively devel-oped by personnel who are directly involved in thecooperative ventures. These two types of resourcesshare the common feature that it is not efficientfor an ICV to acquire them through transfer fromeither the foreign or local partner. For the purposeof this paper, this third category of resources willbe termed operating resources.

These three general types of resources stra-tegic, local, and operatingdiffer both in the func-tions they can perform in a cooperative strategyand in the degree to which they are traded incompetitive markets (Barney, 1986a). These differ-

ences in functions and markets for different typesof resources help to define the roles that partic-ipant organizations can be expected to play ininternational cooperative strategies. Table 1 pro-vides a brief summary of key characteristics ofthe resources discussed below.

Role of the foreign partner: strategic resources

As noted above, the OLI approach to multinationalenterprise suggests that the successful entry of afirm into international markets is driven by posses-sion of firm-specific strategic assets. The foreign

partner plays a critical part in contributing strategicresources that can help an ICV succeed in the face

of global competition. As companies in emergingeconomies have gained greater access to interna-tional sources of capital, goods, and services, ear-lier roles of foreign partners as conduits betweenworld markets and local ventures have diminishedin importance. In order to have an impact on ven-ture performance, multinational firms must nowbring resources to ICVs that cannot be readilyobtained through the market (Beamish et al., 2000;Chi, 1994; Teece, 1981). The resource-based viewof the firm offers insight into the types of resourcesthat may perform this role.

A resource-based view of foreign partner

contributions

The resource-based view of the firm argues thatstrategic resources have certain distinctive marketcharacteristics. Resources that are scarce, valuable,and difficult to imitate or substitute are critical tostrategic advantage (Barney, 1991; Connor, 1991;Winter, 1987). These strategic assets may be tan-gible resources such as patents, licenses or brandnames, or less tangible capabilities such as generalmanagement skills (Barney, 1986b; Mahoney and

Pandian, 1992; Winter, 1987).Contributing strategic resources to an ICV can

be important to the foreign partner for two reasons:leveraging resources to new markets represents oneof the key reasons for a foreign firm to pursueany type of international strategy, and strategicresources provide a means by which the foreignpartner can make a distinctive contribution to thecompetitive advantage of a cooperative venture.Just as an ICV must achieve competitive advan-tage over potential rivals, the foreign firm mustoffer advantages over alternative partners and non-

ICV strategies in order to be an attractive partner(Hennart, 1988; Kogut, 1988).

The stock of resources developed by a multi-national firm in prior domestic and internationaloperations is a major factor that drives entry intonew markets (Dierickx and Cool, 1989; Fladmoe-Lindquist and Tallman, 1994). These strategicresources also can provide competitive advantagein emerging economies. For example, Wal-Marthas developed world-class capabilities in organiz-ing and managing mass retailing operations in theUnited States, including supply chain management,logistics, inventory management, and information



4/21


systems. These strategic resources have provideda basis for the firm to expand operations into Latin

America and Asia through a number of coop-erative ventures. Wal-Mart profits by increasingthe scope of utilization of their strategic assets,and their cooperative ventures gain competitiveadvantage over rivals that do not have access tomanagement skills comparable to those providedby Wal-Mart. Examples of similar strategies arewidespread, including the cooperative strategiesof Loctite in many emerging economies, Star-bucks cooperative ventures with Jardine in Asia,or Volkswagenwerks cooperative strategies forsupplier development in Brazil or China.

The logic for geographic diversification throughcooperative strategy is similar to the logic for prod-uct market diversification. Just as product diversifi-cation allows firms to achieve scope economies bysharing the services of strategic assets across newproducts (Penrose, 1959; Teece, 1981; Williamson,1985), geographic diversification offers a means toachieve greater scope in the utilization of strate-gic assets across new markets (Kogut, 1988). Thisleads to a basic premise:Managers in foreign part-ner firms can be expected to see contributions of

strategic resources by their firms as having a pos-

itive effect on ICV performance.

Role of the host country partner:

complementary location-specific resources

International ventures are commonly formedby firms in response to a need for location-specific skills or resources for operation in anoverseas market (Dunning, 1988: Teece, 1983).Local (location-specific) resources may includedistribution channels, local brands, politicalinfluence, human resource management skills, or

any other capabilities that are idiosyncratic to aspecific locality. When local resources of this typeare controlled by a host country firm and notobtained equally readily through the market acooperative venture may offer an effective meansfor a foreign partner to expand into a new area(Kogut and Zander, 1993).

The capacity of a host country organization tosupply local resources to a venture is one of thebasic reasons for the existence of any form of inter-national strategy (Dunning, 1988). Without con-tributions of local resources by the host countryentity, a strategy of this type cannot be expected to

offer advantages over trade (Hennart, 1982). Tradi-tional Ricardian notions of comparative advantage

may help to explain why a foreign firm pursuesa position in the host country, but it is the abil-ity of the local partner to supply complementaryresources for the exploitation of those compara-tive advantages that makes a cooperative strategyattractive (Tallman et al., 1997).

Local resources of this type may not be freelyavailable in emerging markets for a number ofreasons. One of the defining features of an emerg-ing economy is the limited historical developmentof local industry. The development that has takenplace in many of the emerging economies of Latin

America and Asia has been dominated by a limitedset of large industrial or commercial groups suchas the Garza-Sada interests in northern Mexico orLucio Tans family of companies in the Philip-pines. A foreign firm that is seeking a capable localpartner often will have a narrow range of choicesin that type of economy.

Wal-Marts entry into Mexico again provides agood example. Few Mexican retail organizationscould have offered Wal-Mart the local resources indistribution, locations, and name recognition thatCIFRA possessed. If Wal-Mart had chosen not to

deal with one of a few established large Mexicanretail organizations, their entry into Mexico wouldhave been deeply problematic. The pure market-based alternative of building a comparable orga-nization from inputs not controlled by these largeMexican companies might not have been feasible,and it undoubtedly would have been very costlyand time-consuming.

As their cooperative strategies develop andmature, foreign firms actually may face moresevere failure of markets for local resources.Foreign and host country partners may develop

working relationships that are specialized to thecooperative strategy. These transaction-specificinvestments increase the reliance of the foreignfirm on the local partner and cement the localpartners position as a source of local resources(Kogut and Zander, 1993; Teece, 1981).

The key role that the host country partner in acooperative strategy can play in providing comple-mentary resources suggests another basic premise:Managers in foreign partner firms can be expected

to see contributions of local resources by host coun-

try partners as having a positive effect on ICV

performance.



5/21


Role of the venture: efficient resource

acquisition

Complementary contributions of local and strategicresources may be vital to alliance success, but theyare not the only types of resource transfers fromparent firms that may affect strategic performance.Certain forms of dependence on parent firms maybe costly for ICVs, either because market sourcingof resources would be cheaper or because theresources involved can be developed internallywith greater efficiency than they can be transferredto the ICV from parent organizations. Dependenceupon parent firms for either of these types ofresources may represent an inefficient solution to

the ICVs needs and reduce overall performance.For the purposes of this paper, resources that arenot efficiently acquired from either partner firm aretermed operating resources.

Transaction-cost considerations

Uncomplicated transactions involving tangiblegoods with little uncertainty about price orquality may be most efficiently carried outthrough market exchange (Williamson, 1985).Commodity-like goods such as raw materialsoften fall into this first, low transaction-costcategory. Resources of this type can be acquiredthrough market sources, and partner firms mayexpect cooperative ventures to develop their ownsourcing relationships. Continued reliance upona foreign partner for supplies of resources suchas raw materials entails unnecessary costs andcan compromise the performance of otherwiseestablished ICVs. Certain international fast foodoperations in Mexico, for example, experienceddifficulty for a number of years in developinglocal sources of supply that met reliability andquality standards set by U.S. partner organizations.

Resulting dependence upon imports from U.S.partner firms damaged strategic performance bothby inflating costs and by further inhibiting thedevelopment of more efficient sources of supplyfor the ventures.

At the other end of the transaction-cost spec-trum, it may be very costly or inefficient to attemptto acquire certain types of intangible resources byeither market sourcing or interorganizational trans-fer. Reliance upon U.S. partner firms for operat-ing routines or procedures may pose this type ofproblem for ventures. Routines and operating pro-cedures commonly have important idiosyncratic or

context-specific features that are imperfectly trans-ferable, and the implementation of effective oper-

ating procedures can involve path-dependent learn-ing that produces different results in different orga-nizations (Kogut and Zander, 1993; Nelson andWinter, 1982; Polanyi, 1975). The effort to trans-fer operating procedures from foreign firms mayrequire costly ongoing involvement of expatriatepersonnel in operational aspects of the venture andstill be relatively ineffective (Teece, 1981).

Operating problems described by a U.S. man-ager of an ICV manufacturing company in Mexicoprovide a good illustration of the costs that canaccompany this type of dependency. The addition

of new machinery or upgrade of existing produc-tion equipment in the venture typically requiredextensive involvement of U.S. personnel to assistwith basic features of process redesign such aslayout of production areas and sequencing workflow. The venture had not developed the capacityto create new operating procedures in response torelatively modest technical changes in production.Managers in the U.S. partner company viewed thisas a serious deficiency, and the regular shuttle ofpersonnel across the border significantly increasedcosts.1

For the purpose of this analysis, both of these

types of resources can be combined under thesingle rubric of operating resources. Dependenceupon the foreign parent for either type of operat-ing resources can be expected to have a negativeimpact on performance. In the case of tangibleoperating resources such as materials, the negativeeffect is a result of costly sourcing. In the caseof less tangible operating resources such as rou-tines, the negative effect is a consequence of theimperfect transferability of the resources betweenfirms. Both cases lead to a common expectation:Managers in foreign partner firms can be expected

to see dependence upon their firms for operatingresources as having a negative effect on ICV per-

formance.The three premises outlined above suggest a

straightforward view of resource dependence rela-tionships and performance for international coop-erative strategies in emerging economies. Success-ful ICVs rely on foreign parent firms for cer-tain strategic resources; they rely on host countryparents for locality-specific resources; and they

1 Personal communication from a manager who requestedanonymity.



6/21


Table 1. Resource typology

Type of resource

Strategic Local Operating(tangible)

Operating(intangible)

Importance toICV

Strategicadvantage

Effective localoperations

Cost-competitiveinputs

Efficientoperations

Efficient source Foreignpartner

Host-countrypartner

Marketprocurement

Internaldevelopment

Marketcharacteristics

Scarce,controlledby foreignparent

Semi-scarce:limited market,co-specialization

Competitivemarket

Organization-specific

Examples Patents mgt.

capabilities

Local distribution

channels

Raw materials Operating

routines

achieve independence in a variety of resourcesthat can best be sourced through the market ordeveloped internally (see Table 1). The failure todevelop and sustain these relationships can beexpected to compromise performance.

Effect of strategic resources on operating

resources

Contributions of strategic resources from the for-eign partner also may aid the development of oper-ating resources by an ICV and thus reduce theventures reliance on partner firms for operatingresources. Strategic resources include capabilitiessuch as general management skills that may raisethe level of professionalism in the venture andhelp it to achieve a degree of organizational devel-opment that reduces dependence on the foreignpartner for resources such as operating routinesand expatriate support. Ongoing difficulties expe-rienced by ventures in developing local sourcesof operating resources such as raw materials also

may signal weaknesses in management that canbe relieved by parent firm assistance in manage-ment development. This leads to the expectationthat:Contributions of strategic resources by foreignpartners will have a negative effect on contribu-

tions of operating resources.

Evaluation of performance

Although resource contributions of the types des-cribed above may all affect the success of a coop-erative strategy, their impact on performance may

be achieved in different ways. It has been com-mon to treat venture performance as a unitary

construct (Geringer and Hebert, 1989). However,the use of a single construct may suppress impor-tant information about the mechanisms by which

performance at operational levels influences strate-gic outcomes. Strategic measures of performance

such as growth, sales, or profitability are, in part,outcomes of performance in areas of operations

such as human resource management or qualitycontrol. Certain types of resource contributionsmay be linked to strategic performance primar-

ily through their effects on these operational areas,while others may have a direct impact on strategic

performance. In the former case, operational-levelmeasures of performance effectively function asintervening variables between resource contribu-

tions of partner firms and evaluation of strategicperformance.

Human resource development and quality con-trol are probably the two most important areas ofoperating performance that have the potential to

play this intervening role between inputs to a ven-ture and strategic outcomes. They are among theconcerns most widely cited by managers of emerg-

ing economy ventures, and they may be the bestresearched areas of ICV operations (e.g., Flynn,1994; Napier and Vu, 1998; Zeffane and Rugim-

bana, 1995). However, the selection of HR andquality performance as constructs for this analysis

is not meant to imply that they are the only fac-tors that might play this role for ICVs. HR andquality performance stand out among constructs

of this type, and parsimony in modeling argues for



7/21


the selection of a limited set of strong concepts(Johansson and Yip, 1994; Wunsch, 1988).2

Contributions of strategic resources by the for-eign partner would be expected to have posi-tive effects on both HR performance and qual-ity performance. Contributions of local resourcesby the host-country partner would be expected tohave positive effects on HR and quality perfor-mance. On the other hand, contributions of oper-ating resources by the foreign partner would beexpected to have negative effects on both HR andquality performance.

2 The integration of emerging economies into the global systemcreates pressures for better HR and quality performance in sev-eral ways. Technological development has been rapidly reducingadvantages associated with cheap unskilled labor in many indus-tries (Sargent and Matthews, 1997). This results in escalatingcapital intensity and growing demand for more skilled person-nel and better HR development in emerging economy industries(Minehan, 1996). Reduced trade barriers also mean that the out-put of ICVs must meet world quality standards in order to beviable (Fleury, 1995). Products face a wide range of competitorsin export and domestic markets, and quality standards must beconsiderably higher than levels that may have been acceptablefor domestic production under protectionist regimes. Failuresin many noncompetitive small and middle-sized companies inemerging economies have already been precipitated by the open-ing of domestic markets to international competition, and the

trend can be expected to continue with further global economicintegration.

These relationships suggest a variety of directand indirect effects of resource contributions on

strategic performance. Many of these effects areinterdependent, and they can best be understoodas a system of relationships represented in a struc-tural model. Figure 1 outlines a structural modelthat includes these multiple dimensions of perfor-mance, relationships among the different forms ofperformance, and potential links between resourcecontributions and strategic performance.

EMPIRICAL ANALYSIS OF RESOURCECONTRIBUTIONS AND STRATEGIC

PERFORMANCE

Data

The relationships in Figure 1 were examined usingdata collected in a survey of U.S. firms withICVs in Mexico. Information from the Ameri-can Chamber of Commerce in Mexico (AmCham)provided a basis for identifying a sample of500 U.S. firms that potentially had active ven-tures in Mexico. Detailed questionnaires were sentto managers in these firms identified as hav-ing ICV responsibilities, and 95 responses were

received. Seventy-two of the respondents sup-plied data on current ICVs. This response rate

Figure 1. Structural model, multiple dimensions of performance



8/21


is within expectations for mailed surveys of thistype (Hill, Hwang, and Kim, 1990), and it may

significantly underestimate the true proportion ofthe firms with active ICVs that responded to thissurvey.3

Items dealt with resource contributions of U.S.firms and their Mexican partners to ICVs, the per-formance of cooperative strategies, the objectiveof U.S. firms, and a variety of descriptive data onthe U.S. parent firms of respondents. Apart fromthe descriptive company data, the majority of itemswere collected using Likert-scaled items describedbelow. A general profile of the sample and descrip-tion of survey items is included in the Appendix.

Measures

Resource contributions

Data on the resource contributions made to ICVsby parent firms were derived from a set of itemsin which respondents scored the importance ofcontributions of each of several resources usingLikert scales. These items ranged from very tan-gible resources such as raw materials or patents tomuch less tangible ones such as procedures androutines. The items used in the survey were basedon Schaan and Beamish (1988). Separate ratingswere carried out for contributions by the U.S. par-ent firm and the Mexican parent. The Appendixincludes details of survey items.

3 Although the crude response rate of 19 percent is within expec-tations for mailed surveys (Hill et al., 1990), it may signifi-cantly underestimate the proportion of the firms with active ICVsthat responded to this survey. The information supplied by theAmCham that was used to define the sampling frame appears tohave overstated the number of U.S. firms that have active ICVs

in Mexico. Examination of historical AmCham data and discus-sions with managers suggest that firms without active venturesmay have constituted a significant fraction of the sample and aneven larger proportion of the nonrespondents. However, moreprecise estimation of the number of non-ICV firms inappropri-ately included in the sample was impossible without breach ofrespondent anonymity. Survey responses were required to beanonymous, and therefore it was not possible to differentiatefirms that failed to respond because they did not have activeICVs in Mexico from nonrespondents with active ICVs. An ini-tial screening question asked all respondents to return the surveyand indicate whether they had active ICVs in Mexico, but there isreason to believe that firms without active ICVs may have beenbiased toward nonresponse. The actual response raterelativeto the subgroup of firms in the sample with active ICVsislikely to have been substantially higher than the apparent ratebased on the AmCham sampling frame.

Venture performance

Respondents rated the performance of ventures ona number of different dimensions, including over-all strategic performance, growth, sales, manage-ment development, labor productivity, quality ofgoods and services, and several other key strategicand operating areas. These items were based onwork by Stopford and Wells (1972) and Geringer(1988). Following Geringer (1988), ratings of per-formance also used 5-point Likert scales (see theAppendix).

These survey measures of venture performancewere adopted for a number of reasons. Moreobjective measures of firm performance are gen-

erally based on publicly reported data on account-ing performance or securities market data. Inter-national cooperative strategies typically do notinvolve publicly traded entities, and accountingmay not be entirely disaggregated from part-ner firms. These considerations have led priorresearchers to use self-report approaches of thistype, and validated measures were therefore avail-able for this study (Geringer, 1988).

Validity of measurement

Like the great majority of survey research, thisstudy uses data collected from single respondents.This type of data inevitably raises questions aboutartificial response consistencies due to contexteffect (Harrison, Mclaughlin, and Coalter, 1996).Several aspects of the study were intended toreduce context effects such as common-methodvariance. Instrument design followed guidelinessuggested by Harrison et al. (1996), and multiple-item constructs were used in the data analysis.Context effects tend to be domain-specific andsurvey data on business performance generallyshow high validity (Venkatraman and Ramanujam,1987). Post hoc analysis was carried out usingHarmons single-factor test (Podsakoff and Organ,1986), and it showed no evidence of artificialresponse biases.4

4 Single respondent self-report data of the type employed hereare very common in organizational research, and a majority ofall social research involves self-report surveys (Harrison et al.,1996). The fact that multimethod validation of responses cannotbe carried out with data of this type has been a source of growingconcern to psychometricians (e.g., Podsakoff and Organ, 1986).However, the importance of self-report surveys as a tool ofsocial research also continues to grow, and self-report measureshave been described as virtually indispensable in many research



9/21


Methods

Specification of the structural model describedabove involved several steps, including identifica-tion of potential indicators for unobservable con-structs, refinement of construct definitions, andspecification of relationships among constructs.Potential indicators for each of the constructs inthe model were initially identified on theoreticalgrounds. Indicators then were examined for reli-ability, convergent and discriminant validity, andconstruct definitions were refined as appropriate.

A data reduction process that has become com-mon in PLS modeling was used to refine the model(Falk and Miller, 1992; Fornell, 1982; Johansson

contexts (Avolio, Yammarino, and Bass, 1991:571). Few seriouscritics argue for abandonment of survey research (Harrisonet al.,1996); instead, researchers have emphasized means of analyzingand improving validity of research based on single respondent,self-report data (Avolio et al., 1991; Harrison et al., 1996).

A variety of tactics were used to reduce context effects. Thesurvey instrument was structured with noncontiguous scales;independent and dependent measures did not have overtly similarcontent; and the auxiliary theories connecting measures toconstructs are not likely to be related to primary theories ofrelationships among constructs (Blalock, 1985). Items used in thedependent measures also were not novel in content or narrowlydefined. Narrow or novel items tend to be more problematicbecause respondents may be biased by previous scale responses

in situations where there are fewer related inputs in long-termmemory that can assist in formation of responses (Harrison et al.,1996).

Multiple-item constructs were used in the data analysis.Response biases have been shown to be more problematic at theitem level than the construct level (Avolio et al., 1991; Harrison,et al., 1996; Podsakoff and Organ, 1986). This is an area wherestructural modeling approaches such as PLS and Lisrel are usefulin avoiding problems that might be associated with item-levelanalysis.

Context effects also tend to be domain specific (Avolio et al.,1991; Podsakoff and Organ, 1986), and there is reason to believethat survey data on business performance have a relatively highlevel of validity. In a quasi-experimental design, Venkatramanand Ramanujan (1987) found that multimethod/multitrait andconfirmatory factor analysis indicated good predictive validity

for subjective measures of performance and good convergencebetween measures based on secondary data and survey measures.Dess and Robinson (1984) also have found support for thevalidity of self-report measures of business performance, andGeringer and Hebert (1989) found convergent validity betweenthe type of measures that were employed in this study andsecondary data. Self-report measures of performance have beenused in a number of other recent studies of international strategyincluding Johansson and Yip (1994) and Morrison and Roth(1992).

A post hoc analysis of potential context effects in thesedata was carried out using techniques outlined by Podsakoffand Organ (1986). The results of this analysis also suggestthat artificial response biases may not be significant. Data wereexamined using Harmons principal components approach, andindicators of independent and dependent constructs did not loadon common factors.

and Yip, 1994). The validity and reliability of indi-cators in the outer or measurement model were

examined and latent constructs were refined byeliminating unreliable indicators and creating con-structs that had greater validity and reliability. Thiswas followed by reestimation of the outer modeland reexamination of constructs for validity andreliability (Falk and Miller, 1992; Graham, Mintu,and Rodgers, 1994). Validity and reliability analy-sis for the final model are discussed below.

Indicators of latent constructs in the model wereof two types: formative and reflective. Resourcecontributions of parents were all treated as for-mative indicators in the model of ICV perfor-

mance. The underlying constructs Parent Con-tributions of Strategic Resources, Parent Contribu-tions of Local Resources, and Parent Contributionsof Operating Resourceswere conceptualized ascomposites of different specific resource contri-butions. These constructs are meaningful in thecontext of the analysis, but they would not neces-sarily be intrinsically meaningful if divorced fromthat context.

Indicators of venture performance, on the otherhand, were conceptualized as reflective indicators.Operational areas of performance such as qual-ity control and human resource development wereseen as intrinsic phenomena reflected in indicatorssuch as quality of goods or management develop-ment. Indicators of strategic outcomes were alsoseen as reflective measures: growth, sales, andmeeting strategic goals were all seen as reflectionsof the more general strategic performance of theventure.

Parent Contributions of Strategic Resources

Several items were identified as potential contribu-tions of strategic resources by U.S. parent firms.

The candidate items included management per-sonnel, technical/professional skills, brands, andpatents, copyrights and licenses. Management andtechnical/professional personnel and skills wereseen as particularly important due to the factthat they represent transfers of less tangible skillsthat may be difficult to accomplish without ICVs(Hennart, 1988; Kogut and Zander, 1993). Globalbrands and patents or copyrights are more readilytraded through markets by licensing and thereforemay be less fundamental to ICV strategy (Barney,1986a; Lieberman and Montgomery, 1988; Teece,Pisano and Shuen, 1997). Initial examination of



10/21


the measurement model suggested that brand namewas not likely to be a reliable indicator and had

little convergence with the other indicators of theconstruct. Brand was eliminated as an indicatorand the remaining indicators of strategic resourcecontributions were retained (Graham et al., 1994;Johansson and Yip, 1994).

Parent Contributions of Local Resources

Potential local resource contributions by Mexicanparents included labor, distribution channels, andlocal brands. Initial refinement of the measurementmodel suggested separation of labor from brands

and distribution channels as an independent con-struct. Two constructs for Mexican local resourceswere used in the final model: contributions ofmarketing resources (brands and distribution chan-nels), and contributions of labor.

Parent Contributions of Operating Resources

Indicators of operating resource contributions wereselected from among the resource contributions ofU.S. parent firms. Indicators included contributionsof materials and contributions of operating proce-

dures and routines. U.S. Parent Contributions ofOperating Resources was a single latent construct.

Model definition

Figure 2 illustrates the full structural model, inclu-ding construct indicators and paths among latentconstructs. Contributions of Strategic Resources byU.S. parent firms hypothetically had direct positiveeffects on strategic outcomes as well as indirecteffects through Human Resource and Quality Per-formance. Contributions of strategic resources also

were expected to have a negative link to Contribu-tions of Operating Resources, as indicated above.

Contributions of Operating Resources wereexpected to affect strategic performance bothdirectly and indirectly, through their effectson HR Performance and Quality Performance.Effects of Contributions of Operating Resourceson HR Performance and Quality Performancewere anticipated to be negative, with indirectnegative effects on strategic outcomes throughthe positive paths from HR and quality toStrategic Performance. Direct effects on StrategicPerformance also were expected to be negative.

The two different local resource constructs haddifferent logical connections to strategic perfor-

mance. Contributions of Marketing Resources wereexpected to have a direct positive effect on strate-gic outcomes, but there was no conceptual basisfor any indirect effects through Quality Perfor-mance or HR Performance. Contributions of labor,on the other hand, were expected to operate pri-marily through intermediate effects on Quality andHR Performance, although the possibility of adirect effect on Strategic Performance also wasconsidered.

Model estimation

Measurement and path models were estimatedusing latent variable partial least squares (PLS)analysis. PLS techniques are particularly useful forthis type of analysis for three reasons: they makemore modest demands upon data than covariance-matrix based structural modeling approaches suchas LISREL or EQS, they accommodate measure-ment models that include both reflective and for-mative indicators, and they are well suited to mod-eling in situations where the objective of analysisis assessing predictive relationships among vari-

ables rather than testing a general theory (Falk andMiller, 1992)The limited assumptions about underlying dis-

tributions in PLS models also preclude most para-metric forms of significance testing for path coeffi-cients (Falk and Miller, 1992; Fornell and Larcker,1981a). However, significance of path coefficientsin PLS models conventionally is estimated usingthe jackknife resampling technique (Miller, 1974;Mooney and Duval, 1993), which tends to produceconservative estimates of significance (Falk andMiller, 1992). Jackknife estimates of standard errorwere derived for path coefficients in the model and

significance was evaluated with t-statistics.

RESULTS

Outer (measurement) model

Convergent validity of the outer (measurement)model can be assessed in PLS analysis by examin-ing the construct variance extracted by indicators,with proportions greater than 0.50 indicating satis-factory convergent validity (Fornell and Larcker,1981a). Table 2 includes loadings of indicators



11/21


Figure2.

Structuralmodelforstrategicperformance,

includingindicatorsandc

onstructs



12/21


Table 2. Theta matrix of residual correlations for between-block indicator variables

1 2 3 4 5 6 7 8

(a) 1 0.024 0.011 0.000 0.011 0.065(a) 2 0.001 0.000 0.000 0.122 0.146(a) 3 0.029 0.013 0.000 0.073 0.088(b) 4 0.000 0.026 0.150(b) 5 0.000 0.012 0.071(c) 6(d) 7(d) 8

(a) Contributions of strategic resources(b) Contributions of operating resources(c) Contributions of local labor(d) Contributions of local marketing resources

Variable names1 U.S. contributions of mgt personnel/skills2 U.S. contributions of technical personnel/skills3 U.S. contributions of patents, licenses4 U.S contributions of materials5 U.S. contributions of operating routines6 Mexican contributions of labor7 Mexican contributions of brands8 Mexican contributions of channels

and proportion of variance extracted for constructs.All constructs met criteria for adequate convergentvalidity.

Discriminant validity can be evaluated in PLS

modeling by examining the correlations betweenresiduals of the indicator variables used to definedifferent latent constructs (Fornell and Larcker,1981b; Falk and Miller, 1992). If the blocksof manifest (indicator) variables are appropri-ately defined, residual correlations between indi-cators of different constructs should be minimal.Falk and Miller (1992) suggest that discriminationbetween constructs should be considered question-able if several residual interblock correlations arein excess of 0.20. No interblock residual correla-tion in this model exceeded 0.15 and most were

close to zero, suggesting good discriminant valid-ity using Falk and Millers (1992) criteria. Thetheta matrix of interblock residual correlations ispresented in Table 2.

Indicator reliability is generally evaluated interms of the loadings of individual indicators ontheir constructs. Fornell and Larcker (1981a) arguefor loadings of 70 as the criterion for satisfactoryindicator reliability, while Falk and Miller (1992)suggest a somewhat less demanding 0.55 stan-dard. All indicators met the 0.70 standard of reli-ability suggested by Fornell and Larcker (1981a)(Table 3).

Inner (path) model

As indicated above, significance of coefficients inthe inner or path model was assessed using the

jackknife, and t-statistics were calculated for pathcoefficients. Figure 3 includes path coefficients forall latent constructs. Significant path coefficients( = 0.05) are indicated by asterisks.

The soft assumptions about distributionsemployed in PLS modeling preclude summarytests of model fit such as the chi-square analysistypically employed in Lisrel modeling. The focusof PLS is on assessment of relationships amongthe variables of a theoretically identified systemrather than the creation of a comprehensivemodel of the dependent phenomenon. However,

an approximation of the strength of an overallmodel can be obtained from the average explainedvariance in endogenous constructs (Falk andMiller, 1992). The average explained variancefor the endogenous constructs of this model was39%, with no construct below the 34% level,which suggests a strong fit between model anddata. This exceeds the average variance explainedin endogenous variables in models reported byJohansson and Yip (1994), Meznar and Nigh(1995), and Falk and Miller (1992). Varianceexplained in the dependent construct, strategicperformance, was about 48%.



13/21


Table 3. Indicator loadings, residual variance, and variance extracted for constructs

Construct Indicator Loading Residual Variance extracted

Strategic Mgt. personnel 0.810 0.344 0.562resources Tech. personnel 0.714 0.490

Patents, copyrights 0.722 0.479

Local mktg Brand names 0.990 0.02 0.743resources Dist. channels 0.711 0.494

Local labor Labor 1.0 0.0 1.00

Operational Materials 0.741 0.451 0.687dependencies Routines 0.908 0.176

Quality Qual. goods 0.928 0.139 0.617performance Qual. services 0.740 0.452

HR Labor prod. 0.928 0.139 0.869development Mgt. development 0.936 0.124

Strategic Sales 0.778 0.395 0.760performance Strategic goals 0.910 0.172

Growth 0.920 0.154

Path coefficient t-statistic significant a= 0.05, jackknife estimates

Figure 3. Inner model: path coefficients

Effects of HR and Quality Performance on

Strategic Performance

HR and Quality Performance both had significantpositive effects on Strategic Performance, with the

effect of HR ( = 0.481) somewhat greater thanquality ( = 0.321). These relationships provedto be important in the analysis because certainresource contributions affected strategic perfor-mance indirectly, through their paths to HR and



14/21


Table 4. Effects of independent constructs on endogenous constructs

Effects on quality performanceStrategic resources:

direct = 0.330indirect (through operational dependence) = (0.577)(0.312) = 0.180total = 0.330 + 0.180 = 0.510

Local labor:direct = 0.568

Effects on HR development

Local labordirect = 0.548

Effects on strategic performance

Strategic resourcesindirect (through quality performance) = (0.510)(0.321) = 0.164

Local labordirect = 0.394indirect (through HR Development) = (0.548)(0.481) = 0.264indirect (through Quality Performance) = (0.568)(0.321) = 0.182total = 0.264 + 0.182 0.394 = 0.052

Local marketing resourcesdirect = 0.350

Quality Performance. Table 4 summarizes directand indirect effects on the endogenous constructsfor all independent constructs.

Contributions of Strategic Resources

The effect of Contributions of Strategic Resourceson Quality Performance proved to be positive andsignificant. However, the paths from Contribu-tions of Strategic Resources to HR Performanceand Strategic Performance were both not signif-icant. Contributions of Strategic Resources didhave positive effects on the evaluation of Strate-gic Performance of the ICVs, but all effects werethrough their influence on Quality Performance.The two-step positive effect of Contributions ofStrategic Resources through Quality Performance

was = (0.330)(0.321) = 0.106.Contributions of Strategic Resources also had

a positive effect on Strategic Performance throughtheir influence on Contributions of Operating Reso-urces. The negative effect of Contributions ofStrategic Resources on Contributions of Operat-ing Resources proved to be significant, as didthe negative effect of Contributions of Operat-ing Resources on Quality. The two negative pathsresulted in a net positive indirect effect on Qual-ity Performance = (0.577)(0.321) = 0.180,and a two-step positive effect on Strategic Perfor-mance, = (0.180)(0.321) = 0.058 (Table 4).

Contributions of Local Resources

Mexican contributions of Local Labor proved tohave significant positive paths to both HR andQuality Performance. However, Contributions of

Local Labor also had a significant negative directeffect on Strategic Performance, contrary to expec-tations. The indirect effects of Contributions ofLocal Labor on Strategic Performance were pos-itive through their paths involving HR Perfor-mance = (0.548)(0.481) = 0.264, and QualityPerformance = (0.568)(0.321) = 0.182, for atotal positive effect: = 0.182 + 0.264 = 0.446.At the same time, the direct effect was negative( = 0.394), with the result that the net effect ofContributions of Local Labor on Strategic Perfor-mance was only = 0.446 0.394 = 0.052. The

direct path from Mexican partner Contributions ofMarketing Resources to Strategic Performance waspositive and significant ( = 0.350).

Contributions of Operating Resources

As indicated above, contributions of OperatingResources had a negative effect on quality per-formance ( = 0.312) and therefore a negativeindirect effect on Strategic Performance as well( = 0.110). The direct path between Contribu-tions of Operating Resources and Strategic Per-formance was not significant, despite expectations



15/21


to the contrary. The path between Contributionsof Operating Resources and HR Performance also

proved not to be significant.

DISCUSSION

The analysis generally supported the view ofICV strategies suggested by the model developedabove, although it also included a few surprises.Contributions of strategic resources by U.S. par-ents and contributions of local resources by Mexi-can parents had the anticipated positive effects onperformance, while U.S. contributions of operating

resources had a negative indirect effect. Contribu-tions of strategic resources also appeared to dimin-ish contributions of operating resources, creatingan additional indirect positive effect on strategicperformance.

The manner in which the different types ofresource contributions achieved their effects wasmore complex. The effects of contributions ofstrategic resources were positive, but the relation-ship proved to be entirely indirect, through qual-ity performance. The path from contributions ofstrategic resources to HR performance was notsignificant, with the result that there also was no

indirect effect through HR performance.This may reflect the fact that mature ventures

operating in emerging economies often use localpersonnel at many levels from the shop floorto management. Cross-cultural differences mayrender foreign parent contributions less effectivein high cultural content activities such as HRdevelopment (Amante, 1993; Huo and VonGli-now, 1995; Napier and Vu, 1998). At the sametime, methods for quality engineering pioneeredin North America, Japan, and Western Europehave enjoyed some success in firms in emerging

economies (Lakhe and Mohanty, 1994). Standardssuch as ISO 9000 have been relatively widelyembraced, and TQM programs have been adaptedto a broad range of different settings (Farahmand,Becerra and Green, 1994). Foreign parent firmscurrently may be more capable of offering trans-ferable management and technical skills in the areaof quality management than in human resourcedevelopment.

The indirect effect of Contributions of StrategicResources on Quality Performance through reduc-tion of Contributions of Operating Resources sug-gests that involvement of foreign parents may be

more useful at strategic than operating levels ofmanagement. Foreign parents interested in improv-

ing ICV strategic performance thus may find thatemphasis on building ICV management capabil-ities is more effective than devoting resourcesto managing the venture at operational levels.Although this is not surprising, international com-panies commonly attempt to exert close control ofICVs in newly industrializing countries. The find-ing that supply of labor by the local partner appearsto have a positive effect on quality underscores thesignificance of local control of host country activi-ties and suggests that foreign partner firms need tobe able to take a strategic role while encouraging

the autonomy of host country partners in certainareas of operation.The fact that contributions of labor by the Mex-

ican parent proved to have a negative direct effecton Strategic Performance at the same time that theyhad positive indirect effects through HR Perfor-mance and Quality Performance was particularlyinteresting. The magnitude of the two positive indi-rect effects was comparable to the negative directeffect, and the relationship between Contributionsof Local Labor and Strategic Performance mighthave disappeared in an approach that treated per-formance as a unitary construct. The two positiveindirect effects are consonant with the ideas laidout above, but the negative direct effect was unex-pected. It is possible that the negative direct effectsignals underlying costs or constraints on growthassociated with reliance upon the local parent forlabor. If that is the case, labor may have someambiguity as a resource to be acquired through thelocal partner. Reliance upon the local partner mayprovide a superior pool of labor that boosts qualityand accelerates human resource development, butit also may be more costly or difficult than sourc-ing labor in the open market and may prove to

be problematic for the long-term development ofenterprises in emerging economies.

CONCLUSION

The study has a number of implications for theoret-ical and empirical research on cooperative strate-gies for emerging economies. The analysis of ICVsas complex three-party arrangements appears to bea promising approach, and the proposed adapta-tion and extension of OLI perspectives provideinsight into the relationships between ICVs and



16/21


parent firms. In combination with concepts fromresource-based and transaction-cost theories, these

ideas help to map out different areas where ICVdependence or autonomy may be important.

Theoretical implications

The modifications of the OLI perspective in thisresearch are important. When assumptions of fullinternalization and economic efficiency are sus-pended, the perspective can yield valuable insightsinto the strategic roles of parent firms in inter-national cooperative strategies. Resource-basedand transaction-cost theories supply guidelines for

understanding how different types of market fail-ure set parameters for interdependencies amongICV participants. This modified OLI analysisinwhich both a broader range of market failuresand a broader range of organizational responsesare consideredhas the potential to be a valuableaddition to the recent stream of work extending theOLI perspective. Much of the work in this vein hasbeen purely conceptual, and this study takes a use-ful step toward grounding those ideas in empiricalevidence. Further work in this area may providea more detailed view of the ways that differentmarket conditions and competitive market struc-tures help to define strategic linkages among theparticipants in cooperative ventures.

The key empirical insight to emerge from thisstudy is that a balance of ICV dependence andindependence is important to the success of strate-gies for emerging economies. Utilization of cer-tain types of strategic and local resources suppliedby parent firms appears to support venture per-formance, while other types of contributions byparents may be detrimental. These aspects of coop-erative strategies emerge only with the three-partyview of ICVs. If ICVs are viewed solely as asso-

ciations between parent firms, the liabilities asso-ciated with venture dependence on foreign parentfirms disappear.

Empirical implications

The empirical analysis is revealing in a number ofways. In addition to providing some support forthe general approach to understanding ICV strate-gies, it also offers insights into how interdepen-dencies among venture participants are structured.The ambiguous relationship between Contributionsof Labor by the emerging economy parent and ICV

success is interesting, as is the fact that Contribu-tions of Operating Resources affect Strategic Per-

formance primarily through Quality Performance.The treatment of performance as a set of interre-lated constructs rather than a single construct pro-vides valuable insights into relationships betweendifferent forms of dependence and performance.Important underlying effects of resource contribu-tions on performance would have disappeared in amodel that relied upon a unitary performance con-struct. This approach to performance is unusualin the work on ICVs, and it has the potential tomake an important contribution to future researchon cooperative strategies.

This research also has a number of empiricallimitations. It can provide insight into factors thatinfluence the evaluation of cooperative strategiesby foreign parent firms, but there is a range ofissues that it cannot address. The study focuses onperformance in established cooperative ventures,and many of the transient problems that may resultin early failure of ICVs are excluded from theanalysis. Because the study takes the vantage pointof the foreign partner, differences between partnersin the evaluation of strategic performance alsocannot be examined. This focal-firm perspective isa basic feature of strategic researchand makes it

possible to view the situation from the standpointof decision-makers but we cannot assume thatmanagers in Mexican partner firms would share thesame views as their U.S. counterparts (Gillespieand Teegan, 1995). Mexican and U.S. partners mayhave divergent strategic objectives, and operatingperformance that is satisfactory to one side maybe inadequate for the other. A separate, parallelstudy would be required to explore the strategicperspective of the host country partner.5

5 This asymmetrical, focal-firm view is an intrinsic feature of

the strategic perspective (Ansoff, 1965; Beamish, 1988; Porter,1980; Madhok and Tallman, 1998). A strategy that is highlysuccessful for a firm may simultaneously be costly to thefirms competitors, customers, suppliers, or partners (Klein,Crawford, and Alchian, 1978; Porter, 1980; Teece, 1987). Forexample, an alliance strategy that involves expropriation ofa partners technological know-how might be very successfulfrom the standpoint of the firm gaining the technologyandbarely tolerable for the alliance partner (e.g., Reich and Mankin,1986). This is an important area where strategic approaches toICVs may differ from a good deal of organizational researchon cooperative ventures. Organizational research that focuses oncriteria such as mutual satisfaction or longevity of cooperativeventures cannot be assumed to reflect the strategic objectivesof any specific firm within a cooperative venture. In extremecases, a cooperative venture that is unacceptably costly orunsatisfactory to either parent firm will fail; however, there is a



17/21


The applicability of this approach also may belimited in some cases by the role taken by gov-

ernment. ICVs that are subject to substantial statecontrol may not acquire the same level of auton-omy as ICVs in less regulated settings. This isparticularly likely to be an issue in situations wherethe local partner is closely tied to a governmentbureau or agency. Sensitivity to the political con-text is important in any effort to generalize themodel across emerging economies.

Implications for management

The research may also offer some practicalinsight into the management of ICVs in emergingeconomies. The negative relationship between con-tributions of Strategic Resources and Contributionsof Operating Resources suggests that general man-agement assistance from foreign parents may playan important role in organizational developmentfor ICVs. The fact that Contributions of Strate-gic Resources had significant effects on Strate-gic Performance only through effects on Qualitysuggests that successful long-run strategies mayrequire action by foreign parents to help build theindependent operating capabilities of ICVs.

The research provides some general support forone of the fundamental concepts that has emergedover time from the OLI perspective: the idea thata combination of strategic resources from foreignpartners and local resources from host countrypartners is vital to success. However, the researchalso contains the message that involvement ofthe foreign parent at operational levels may bedetrimental to cooperative ventures. This obser-vation may help to address the growing concernin international strategy about potential conflictsbetween global strategic demands and the influ-

ence of local institutions on ventures in emergingeconomies. Parent firms from highly developedeconomies may be most successful by providingstrategic guidance and key sources of competitiveadvantage to ICVs, relying upon local parents totake a major role in high cultural content activi-ties such as marketing or labor management, andbuilding the autonomy of ventures at operationallevels.

great deal of room for asymmetry in the outcomes that lie withinthose extremes (Lorange, 1997; Lu and Lake, 1997).

REFERENCES

Amante M. 1993. Human resource management inJapanese enterprises in the Philippines: issues andproblems. Asia Pacific Journal of Management 10:237245.

Ansoff HI. 1965. Corporate Strategy: An AnalyticApproach to Business Policy for Growth andExpansion. McGraw-Hill: New York.

Avolio B, Yammarino FJ, Bass BM. 1991. Identifyingcommon methods variance with data collected froma single source: an unresolved sticky issue. Journal of

Management 17: 571587.Barney JB. 1986a. Strategic factor markets: expectations,

luck and business strategy. Management Science 32:12311240.

Barney JB. 1986b. Organization culture: can it be a

source of sustained competitive advantage? Academyof Management Review 11: 656665.

Barney JB. 1991. Firm resources and sustained competi-tive advantage. Journal of Management 17: 99120.

Beamish PW. 1988. Multinational Joint Ventures inDeveloping Countries . Routledge: New York.

Beamish PW, Killing JP. 1997a. Cooperative Strategies:North American Perspectives. New Lexington Press:San Francisco, CA.

Beamish PW, Killing JP. 1997b. Cooperative Strategies:European Perspectives . New Lexington Press: SanFrancisco, CA.

Beamish PW, Killing JP. 1997c. Cooperative Strategies:Asian Pacific Perspectives . New Lexington Press: SanFrancisco, CA.

Beamish PW, Morrison AJ, Rosenzweig PM, Inkpen A.2000. International Management. McGraw-Hill: NewYork.

Blalock HM. 1985. Conceptualization and Measurementin the Social Sciences . Sage: Beverly Hills, CA.

Caves RE. 1971. International corporations: the industrialeconomics of foreign investment. Economica 38:127.

Chi T. 1994. Trading in strategic resources: necessaryconditions, transaction cost problems, and choice ofexchange structure. Strategic Management Journal15(4): 271290.

Choi CJ, Lee SH. 1997. A knowledge-based view ofcooperative interorganizational relationships. InCoop-erative Strategies: European Perspectives , BeamishPW, Killing JP (eds). New Lexington Press: San Fran-cisco, CA; 3358.

Connor KR. 1991. An historical comparison of resource-based theory and five schools of thought withinindustrial-organization economics: do we have a newtheory of the firm? Journal of Management 17:121154.

Contractor F, Lorange P (eds). 1988. CooperativeStrategies in International Business. Lexington Books:Lexington, MA.

Dess GG, Robinson RB Jr. 1984. Measuring organiza-tional performance in the absence of objective mea-sures: the case of the privately-held firm and con-glomerate business unit. Strategic Management Jour-nal 5(3): 265273.



18/21


Dierickx I, Cool K. 1989. Asset stock accumulation andsustainability of competitive advantage. Management

Science 35: 1504 1511.Dunning JH. 1988. The eclectic paradigm of internationalproduction: a restatement and some possible exten-sions. Journal of International Business Studies 19:132.

Falk RF, Miller NB. 1992. A Primer for Soft Modeling .University of Akron Press: Akron, OH.

Farahmand K, Becerra R, Green JR. 1994. ISO 9000certification: Johnson Controls inside story. Industrial

Engineering 26: 2223.Fladmoe-Lindquist K, Tallman S. 1994. Resource-based

strategy and competitive advantage among multina-tionals. InAdvances in Strategic Management, Vol. 10,Shrivastava P, Huff A, Dutton J (eds). JAI Press:Greenwich, CT; 4572.

Fleury A. 1995. Quality and productivity in thecompetitive strategies of Brazilian enterprises. World

Development 23: 7385.Florin J. 1997. The case for nonequity interfirm

cooperative arrangements. In Cooperative Strate-gies: North American Perspectives, Beamish PW,Killing JP (eds). New Lexington Press: San Francisco,CA; 324.

Flynn G. 1994. HR in Mexico: what you should know.Personnel Journal 73: 3444.

Fornell C (ed). 1982. A Second Generation of Multivari-ate Analysis . Praeger: New York.

Fornell C, Larcker D. 1981a. Evaluating structuralequation models with unobservable variables andmeasurement error.Journal of Marketing Research18:

3950.Fornell C, Larcker D. 1981b. Structural equation models

with unobservable variables and measurement error:algebra and statistics. Journal of Marketing Research18: 382388.

Geringer M. 1988. Joint Venture Partner Selection:Strategies for Developed Countries. Quorum: West-port, CT.

Geringer M, Hebert L. 1989. Control and performance ofinternational joint ventures. Journal of International

Business Studies 20: 235254.Gillespie K, Teegan H. 1995. Market liberalization

and international alliance formation: the Mexicanparadigm. Columbia Journal of World Business 30:5869.

Graham J, Mintu A, Rodgers W. 1994. Explorations ofnegotiation behaviors in ten foreign cultures using amodel developed in the United States. ManagementScience 40: 7295.

Hagedoorn J. 1990. Organizational modes of inter-firmco-operation and technology transfer. Technovation10: 17 30.

Harrison DA, Mclaughlin ME, Coalter TM 1996. Con-text, cognition, and common method variance: psy-chometric and verbal protocol evidence. Organiza-tional Behavior and Human Decision Processes 68:246261.

Hennart J-F. 1982. A Theory of the MultinationalEnterprise. University of Michigan Press: Ann Arbor:MI.

Hennart J-F. 1988. A transaction costs theory of equityjoint ventures. Strategic Management Journal 9(4):

361374.Hill CWL, Hwang P, Kim WC. 1990. An eclectic theory

of the choice of international entry mode. StrategicManagement Journal 11(2): 117128.

Huo YP, VonGlinow MA. 1995. On transplanting humanresource practices to China: a culture-driven approach.

International Journal of Manpower 16: 315.Hymer SH. 1960. The international operations of

multinational firms: a study of direct foreigninvestment. PhD dissertation, Massachusetts Instituteof Technology.

Johansson J, Yip G. 1994. Exploiting globalizationpotential: U.S. and Japanese strategies. Strategic

Management Journal 15(8): 579601.Johnson JL, Cullen JB, Sakano T, Takenouchi H. 1997.

Setting the stage for trust and integration inJapaneseU.S. cooperative alliances. In Coop-erative Strategies: North American Perspectives,Beamish PW, Killing JP (eds). New Lexington Press:San Francisco, CA; 227254.

Kindleberger CP. 1969. American Business Abroad: SixLectures on Direct Investment. Yale University: NewHaven, CT.

Klein B, Crawford RG, Alchian A. 1978. Verticalintegration, appropriable rents and the competitivecontracting process. Journal of Law and Economics21: 297325.

Kogut B. 1988. Joint ventures: theoretical and empiricalperspectives. Strategic Management Journal 9(4):

319332.Kogut B, Zander U. 1993. Knowledge of the firm andevolutionary theory of the multinational corpora-tion. Journal of International Business Studies 24:625645.

Lakhe RR, Mohanty RP. 1994. Total quality manage-ment: concepts, evolution, and acceptability in devel-oping economies.International Journal of Quality and

Reliability Management 11: 933.Lieberman MB, Montgomery DB. 1988. First-mover

advantages. Strategic Management Journal , SummerSpecial Issue 9: 41 58.

Lorange P. 1996. Interactive strategies: alliances andpartnerships. Long Range Planning 29: 581584.

Lorange P. 1997. Black-box protection of your core

competencies in strategic alliances. In CooperativeStrategies: European Perspectives, Beamish PW,Killing JP (eds). New Lexington Press: San Francisco,CA; 5973.

Lu Y, Lake D. 1997. Managing international jointventures: an institutional approach. In CooperativeStrategies: European Perspectives, Beamish PW,Killing JP (eds). New Lexington Press: San Francisco,CA; 7499.

Madhok A, Tallman S. 1998. Resources, transactions andrents: managing value through inter-firm collaborativerelationships.Organization Science 9: 326339.

Mahoney J, Pandian JR. 1992. The resource-based viewwithin the conversation of strategic managementStrategic Management Journal 13(5): 368380.



19/21


Meznar MB, Nigh D. 1995. Buffer or bridge? Envi-ronmental and organizational determinants of public

affairs activities in American firms. Academy of Man-agement Journal 38: 975996.

Miller RG. 1974. The jacknife: a review. Biometrika 61:115.

Minehan M. 1996. Skills shortage in Asia. HR Magazine41: 152.

Mjoen H, Tallman S. 1997. Control and performance ininternational joint ventures. Organization Science 8:257274.

Mooney CZ, Duval RD. 1993. Bootstrapping: A Non-parametric Approach to Statistical Inference. Sage:Newbury Park, CA.

Morrison AJ, Roth K. 1992. A taxonomy of business-level strategies in global industries. Strategic

Management Journal 13(6): 399418.

Napier NK, Vu VT. 1998. International human resourcemanagement in developing and transitional economycountries: a breed apart? Human Resource Manage-ment Review 8: 3977.

Nelson R, Winter C. 1982. An Evolutionary Theory ofEconomic Change. Belknap Press: Cambridge, MA.

Osborn RN, Hagedoorn J. 1997. The institutionalizationand evolutionary dynamics of interorganizationalalliances and networks. Academy of Management

Journal 40: 261278.Penrose E. 1959. The Theory of the Growth of the Firm .

Blackwell: Oxford.Podsakoff PM, Organ DW. 1986. Self-reports in organi-

zational research: problems and prospects. Journal ofManagement 12: 531544.

Polanyi M. 1975. Personal knowledge. In Meaning ,Polanyi M, Prosch H (eds). University of ChicagoPress: Chicago, IL; 2245.

Porter ME. 1980. Competitive Strategy . Free Press: NewYork.

Reich R, Mankin E. 1986. Joint ventures with Japan giveaway our future.Harvard Business Review 64: 7686.

Sargent J, Matthews L. 1997. Skill development and inte-grated manufacturing in Mexico. World Development25: 1669 1681.

Sarkan M, Cavusgil SG, Evirgen C. 1997. A commit-ment-trust mediated framework of international col-laborative venture performance. InCooperative Strate-gies: North American Perspectives, Beamish PW,Killing JP (eds). New Lexington Press: San Francisco,

CA; 255 285.Schaan JL, Beamish PW. 1988. Joint venture managers

in LDCs. In Cooperative Strategies in InternationalBusiness , Contractor F, Lorange P (eds). LexingtonBooks: Lexington, MA; 279299.

Stopford JM, Wells LT. 1972. Managing the Multina-tional Enterprise. Basic Books: New York.

Tallman S, Fladmoe-Lindquist K, Robins J. 1997. Alli-ance strategies of U.S. firms in Mexico: a managerialmodel of resources and transactional concerns foralliance success. InStrategic Discovery: Competing in

New Arenas , Thomas H, ONeal D, Alvarado R (eds).Wiley: Chichester; 155 169.

Teece DJ. 1981. The market for know-how and theefficient international transfer of technology.Annals of

the American Academy of Political and Social Science485: 8196.

Teece DJ. 1983. Multinational enterprise, internal gov-ernance, and industrial organization. American Eco-nomic Review 75: 233238.

Teece DJ. 1986. Transaction costs economics andthe multinational enterprise. Journal of Economic

Behavior and Organization 7: 2145.Teece DJ. 1987. Profiting from technological innovation:

implications for integration, collaboration, licensingand public policy. In The Competitive Challenge,Teece DJ (ed). Harper & Row: New York; 185220.

Teece DJ, Pisano G, Shuen A. 1997. Dynamic capabili-ties and strategic management. Strategic Management

Journal 18(7): 509533.Venkatraman N, Ramanujam V. 1987. Measurement of

business economic performance: an examination of

method convergence. Journal of Management 13:109122.

Vernon R. 1971. Storm Over the Multinationals. BasicBooks: New York.

Williamson OE. 1985. The Economic Institutions ofCapitalism. Free Press: New York.

Winter S. 1987. Knowledge and competence as strategicassets. In The Competitive Challenge , Teece D (ed).Harper & Row: New York; 159181.

Wunsch G. 1988. Causal Theory and Causal Modeling .Leuven University: Leuven.

Yan A, Gray B. 1994. Bargaining power, managementcontrol and performance in U.S.China joint ventures:a comparative case study. Academy of Management

Journal 37: 1478 1517.

Zeffane R, Rugimbana R. 1995. Management in theless-developed countries: a review of pertinentissues, challenges, and responses. Leadership andOrganization Development Journal 16: 2636.

APPENDIX: OVERVIEW OF SAMPLEAND SURVEY ITEMS

Sample characteristics

U.S. parent organizations typically were mediumto large-sized firms; median revenues were in

the range of $10100 million, with about a third(25) of all U.S. parent firms reporting revenuesin excess of $100 million. Employment followeda similar pattern, with median employment inthe 5001000-employee category but nearly half(34) of all U.S. parent firms greater than 10,000employees.

The ICVs were relatively long-lived, with amedian age of about 9 years and an average ageof more than 12 years. The distribution of age wasskewed by the presence of a number of older ICVs.ICVs founded before 1980 represented more thana quarter of the sample, and a number of ventures



20/21


were more than 30 years old. The youngest ICVin the sample was about 4 years old.

About 90 percent of all ICVs (65) involvedrelationships in which the formal ICV structurerequired no renegotiation to continue. More thanhalf (55.8% or 40) of the ICVs were character-ized as partnerships and shared-equity joint ven-tures. U.S. parents typically were not majorityequity holders in the shared-equity ventures; 39%(28) indicated that the U.S. parent was the major-ity equity holder, while the Mexican parent wasmajority owner in 57% (41) of cases. The remain-ing 4% (3) involved other, non-Mexican partners.

The importance of these ICVs to the U.S. parents

stands out. In about three-quarters of all cases,ICVs were characterized as being in the U.S.parents most important lines of business. U.S.parent managers also felt that it would be difficultto replace the Mexican parents. Only about 6% (4)considered it easy or relatively easy to replace theMexican parent firm, while almost 70% (50) ratedit difficult or very difficult. More than 82% (60) ofrespondents indicated that ICVs were expected tocontinue into the future for at least 5 years.

ICV managers appeared to enjoy a good deal oflatitude in decision making, regardless of the own-ership structure of the venture. A large majority ofrespondents (90.5% or 65) for both shared-equityventures and nonshared-equity ICVs indicated thatmajor decisions were undertaken without directgovernance by parent companies. The fact thatboth shared-equity and nonshared-equity ICVs hadconsiderable operating autonomy is not surprising.Research has indicated that ownership structuremay not be the primary determinant of operationalcontrol or integration for established cooperativeventures (e.g., Choi and Lee, 1997; Contractorand Lorange, 1988; Mjoen and Tallman, 1997),and a number researchers have argued for broader

conceptualizations of modes of cooperative ven-turing in which equity structure is not viewed asa central determinant of stability or performance(Hagedoorn, 1990; Johnson et al., 1997; Osbornand Hagedoorn, 1997; Sarkanet al., 1997; Yan andGray, 1994).

The most general description of ICVs examinedin this study would be to characterize them asquasi-firms. They enjoy varying degrees of legalautonomy from parents, but they are all organi-zations that have created ongoing operations thatplace them in the position to both use and developresources. The use of the term parent firm in

this paper thus does not imply the existence ofa shared-equity joint venture; the term is used

broadly to designate the U.S. and Mexican partici-pants in these cooperative strategies, regardless ofwhether the ICVs involve equity-sharing arrange-ments. This usage is employed in recognition ofthe roles that foreign and host country firms mayplay in contributing resources to ICVs and the rolethat ventures may play as recipients of resourcesand entities created for distinctive strategic pur-poses (Contractor and Lorange, 1988).

Survey items

Data for the independent and dependent measuresin the study were collected using respondent rat-ings of a variety of resource contributions to coop-erative ventures by the two partner firms. Alldata used in the analysis were based on Likert-scaled items as described below. A variety of otherdescriptive data with regard to parent firms andcooperative ventures also were collected; however,these data were collected for illustrative purposesand were not used in the structural model.

Foreign parent resource contributions

Data on foreign parent resource contributionswere collected using items based on Schaanand Beamish (1988). The general question askedrespondents: How important to this cooperativeventure areyour firms contributions of each of thefollowing types of resources? Following Schaanand Beamish (1988), these were scored on 5-point Likert scales ranging from Not importantto Very important with an additional categoryfor Not applicable. Reverse coding was used forselected items in order to improve validity. Itemsincluded: patents, licenses, copyrights; operating

procedures and routines; brand name or reputation;management personnel and skills; specializedtooling or equipment; raw materials; technicalpersonnel and skills.

Host-country parent resource contributions

These items also followed Schaan and Beamish(1988) and were scored on similar 5-point Likertscales. The general survey question was worded:How important to this cooperative venture areyour Mexican partners contributions of each ofthe following types of resources? Items included:



21/21


brand name or reputation; distribution channels;labor; management personnel; technical personnel

and skills; access to the Mexican market.

Performance

These items were based on Geringer (1988) andGeringer and Hebert (1989). The general surveyquestion was worded: Please rate the performanceof the cooperative venture on each of the following

dimensions. Following Geringer (1988), thesequestions also were scored on 5-point Likert scales

ranging from Very low to Very high with asixth category for Not applicable. Items included:profitability; quality of goods produced; quality ofservices produced; sales; growth; labor productiv-ity; management development; meeting strategicgoals.


Autonomy and Dependence_SMJ.2002

Documents

Transcript of Autonomy and Dependence_SMJ.2002