EMBEDDED TIES AND THE ACQUISITION OF COMPETITIVE CAPABILITIES · Competitive capabilities Competitive capabilities are the set of organizing processes and principles a ﬁrm uses

Strategic Management JournalStrat. Mgmt. J., 26: 1033–1055 (2005)

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/smj.484

EMBEDDED TIES AND THE ACQUISITIONOF COMPETITIVE CAPABILITIES

BILL McEVILY1* and ALFRED MARCUS2

1 Tepper School of Business, Carnegie-Mellon University, Pittsburgh, Pennsylvania,U.S.A.2 Carlson School of Management, University of Minnesota, Minneapolis, Minnesota,U.S.A.

We build on previous research that explores the external acquisition of competitive capabilitiesthrough the embedded ties that firms form in networks and alliances. While information sharingand trust have been theorized to be key features of the interorganizational ties that facilitate theacquisition of competitive capabilities, we argue that these mechanisms provide an incompleteexplanation because they do not fully address the partially tacit nature of the knowledge thatunderlies competitive capabilities. Joint problem-solving arrangements play a prominent role incapability acquisition by promoting the transfer of complex and difficult-to-codify knowledge.Drawing on a set of case studies and a survey of 234 job shop manufacturers we find supportfor the role of joint problem solving with suppliers in facilitating the acquisition of competitivecapabilities. Copyright 2005 John Wiley & Sons, Ltd.

This research addresses a central question in strat-egy: how do firms acquire the capabilities theyneed to develop, sustain, and renew competitiveadvantage?1 Though searching for new sources ofcompetitive advantage is a fundamental strategicactivity (Rumelt, 1984), there is relatively littlesystematic research that explains the sources offirms’ competitive capabilities. Rather, the bulk ofresearch focuses on the performance-related out-comes of capabilities (e.g., Lieberman, Lau, andWilliams, 1990; Clark and Fujimoto, 1991; Hen-derson and Cockburn, 1994). Most of the researchthat has been done on the genesis of capabilitiesconcentrates on sources internal to the firm. For

Keywords: firm capabilities; embedded ties; networkresources*Correspondence to: Bill McEvily, Graduate School of Indus-trial Administration, Carnegie-Mellon University, Pittsburgh, PA15213-3890, U.S.A. E-mail: [email protected] We use the term ‘acquisition’ to refer to the process bywhich firms learn about, internalize, and take advantage of newcapabilities. We do not mean to indicate outsourcing or purchaseof capabilities.

instance, Penrose’s (1959) early writings suggestthat capabilities emerge as the unintended conse-quence of growth and expansion as managers findnew uses for surplus resources. Similarly, Chand-ler (1992) argues that the knowledge and skillsunderlying capabilities are ‘developed by learningthrough trial and error, feedback and evaluation’as managers solve problems such as launchingnew products and scaling up production processes.More recent work on the evolution of capabilities(Helfat, 2000) also points to internal sources offirm capabilities. Taken together, this body of workportrays firms as generating capabilities throughan incremental and path-dependent (Nelson andWinter, 1982) process of learning from their ownexperiences.

However, some strategy researchers have startedto consider how firms derive capabilities externallythrough interorganizational ties. By participatingin ongoing networks of alliances and exchange,they gain access to valuable network resourcesthat aid in the discovery of new opportunities

Copyright 2005 John Wiley & Sons, Ltd. Received 15 November 2000Final revision received 28 March 2005

1034 B. McEvily and A. Marcus

(Gulati, 1999). Network resources, unlike a firm’sinternal material resources, are akin to social capi-tal (Coleman, 1988) in that they inhere in the struc-ture of relations between firms, rather than withinthe firms themselves. Informational advantages areone form of network resources that have receivedconsiderable attention. By participating in inter-firm networks, for example, firms learn about theavailability of new alliance opportunities (Gulati,1999; Kogut, Shan, and Walker, 1992). They alsoaccess information about the capabilities and trust-worthiness of current or potential partners, therebyincreasing their capabilities for alliance formationthrough such networks (Gulati, 1999). Firms canalso augment their innovative capabilities throughinterorganizational ties by using the network topool knowledge and resources and gather andscreen relevant information (Ahuja, 2000). Main-taining a presence in innovation networks fur-ther enhances a firm’s innovative capabilities bydeveloping and maintaining its absorptive capac-ity (Cohen and Levinthal, 1990; Powell, Koput,and Smith-Doerr, 1996; Stuart, 1998; Zaheer andBell, 2005).

This paper builds on previous work by explic-itly examining the underlying mechanisms thatfacilitate the acquisition of competitive capabilitiesfrom external sources. Although information shar-ing and trust are frequently theorized to be centralto the acquisition of capabilities through interfirmties (Ahuja, 2000; Gulati, 1999; Stuart, 1998), theeffects of information sharing and trust typicallyare inferred rather than examined directly. Conse-quently, how they translate into the acquisition ofcapabilities is unclear. We maintain that joint prob-lem solving is a key intervening mechanism thatlinks information sharing and trust to the acqui-sition of capabilities. More specifically, we arguethat information sharing and trust provide only apartial explanation of the mechanisms accountingfor capability acquisition. Our claim is that sincethe knowledge underlying capabilities is partiallytacit, it is difficult to articulate and transfer, andthat joint problem-solving arrangements are a crit-ical mechanism facilitating the acquisition of capa-bilities because they promote the transfer of tacitknowledge (Uzzi, 1997). By providing a forumfor observation, experimentation, and demonstra-tion, joint problem-solving arrangements providemanagers with valuable external learning oppor-tunities to draw on during capability acquisition.While information sharing and trust are important

precursors to the acquisition of capabilities throughinterfirm ties, joint problem-solving arrangementsplay a more prominent role by promoting the trans-fer of complex and difficult-to-codify knowledge.

At the same time, we predict that not all embed-ded ties are equally influential on the acquisi-tion of capabilities. Some embedded ties are moreimportant than others owing to differences in theircontent (Burt, 1997; Gulati and Westphal, 1999;Podolny and Baron, 1997). In the case of embed-ded ties with lead customer and supplier firms, weargue that the degree of firm-specific knowledgeeach type of exchange partner possesses about afocal firm is a critical feature of the relationship.Exchange partners exerting the greatest influenceon a firm’s acquisition of capabilities are those thatare most knowledgeable of a firm’s operations andable to reduce uncertainty about how to implementa capability.

In the following section we formalize our predic-tions by synthesizing research on firm capabilitiesand integrating it with the literature on embedded-ness. We also draw on a set of case studies com-pleted at the outset of the research to inductivelyderive predictions that go beyond the existing lit-erature. To test our hypotheses we gathered surveydata from 234 job shop manufacturers in sevendifferent industries and analyzed these data withstructural equation modeling.

THEORY AND HYPOTHESES

Competitive capabilities

Competitive capabilities are the set of organizingprocesses and principles a firm uses to deploy itsresources to achieve strategic objectives (Kogutand Zander, 1992; Grant, 2002). By shaping theways in which knowledge, skill, and expertiseare coordinated and communicated within a firm,capabilities fundamentally determine what the firmcan do (Zander and Kogut, 1995). The buildingblocks of capabilities consist of theories (Guillen,1995) and frameworks (Porter, 1991), which struc-ture knowledge and organize information. Man-agerial practices and techniques also are an impor-tant component of capabilities (Eccles, Nohria, andBerkley, 1992).

The process of acquiring capabilities is an uncer-tain one because it may not be clear if the orga-nizational building blocks exist for addressing a

Copyright 2005 John Wiley & Sons, Ltd. Strat. Mgmt. J., 26: 1033–1055 (2005)

Embedded Ties 1035

firm’s particular problems or achieving its perfor-mance objectives. Moreover, a firm cannot be surethat adopting a given theory, framework, prac-tice, or technique will generate the same perfor-mance advantages that other firms appear to haveachieved (Lippman and Rumelt, 1982). The pro-cess through which capabilities emerge can bevague and difficult to reconstruct because it isoften based on idiosyncratic, trial-and-error learn-ing. Firms sometimes develop capabilities by acci-dent rather than by planning (Collins and Por-ras, 1994), with learning occurring from failureas much as from success (McGrath, MacMillan,and Venkataraman, 1995; Sitkin, 1992). Hence, theprocess through which a particular capability arisescan be difficult to identify.

Even if the practices and techniques underly-ing a capability are explicitly defined, effectiveimplementation may depend on associated know-how and skill in the firm (i.e., complementaryassets) that are not articulated. The particular mixof administrative arrangements and resources thatexist in an organization and how they have beencombined over time influence the costs and bene-fits of acquiring a capability. Consequently, resultshinge on the degree to which new capabilities canbe integrated with what the organization alreadyhas in place. Since no two organizations are thesame, the effects of acquiring theories, frame-works, practices, and techniques are hard to predictbeforehand (Dierickx and Cool, 1989) and man-agers cannot easily assess the value of a capabilitya priori.

Embedded ties and capability acquisition

Confronted with uncertainty about acquiring capa-bilities, boundedly rational managers engage inproblemistic search and make shortcuts in decisionmaking to conserve their limited time and cogni-tive resources (March and Simon, 1958). Ratherthan assessing all the costs and benefits, and acomplete set of alternatives, managers satisfice(Simon, 1982) by making decisions about capa-bilities based on information that is good enoughgiven the uncertainties they face and the oppor-tunities they discover. To overcome the uncer-tainties associated with acquiring capabilities, weargue that managers do not learn about capabil-ities in isolation. Rather, firms vicariously learnfrom the insights, experiences, or abilities previ-ously accumulated by linked organizations (Darr,

Argote, and Epple, 1995; Baum and Ingram, 1998;Kraatz, 1998; Ingram and Simons, 1999). Alongthese lines, Kraatz reports evidence indicative ofthe role that network ties play in ‘mitigating envi-ronmental uncertainty and promoting social learn-ing of adaptive responses among linked organi-zations’ (Kraatz, 1998: 622). Similarly, McEvilyand Zaheer (1999) found that diversity of con-tacts within a firm’s network, and participation inregional associations, expose the firm to new ideas,information, and opportunities leading to acquisi-tion of capabilities.

The network of interorganizational ties that firmsdraw on to learn about capabilities consists of notonly the diverse contacts managers have with theirpeers, but also the critical exchange relationshipsthey have with customer and supplier firms. Oneset of research on the role of exchange partnersin interorganizational learning has tended to focuspredominantly on knowledge sharing with leadsuppliers (e.g., Uzzi, 1997; Dyer and Nobeoka,2000). Other research has emphasized that leadcustomers can be a critical source of knowledgeleading to innovation (Argote, 1999; Von Hippel,1988). For instance, auto industry suppliers haveimproved their capabilities in quality management,just-in-time production and delivery, and productand process innovation by working closely withlead customers that had already developed exper-tise in these areas (Helper, 1991).

While the literature on interorganizational learn-ing has emphasized the notion that more deeplyembedded ties with both suppliers and customersare conducive to acquiring knowledge and capa-bilities, it has yet to systematically investigate therelative influence of embedded ties with suppli-ers vs. customers or how different elements ofembedded ties affect the acquisition of capabilities.We extend previous research on interorganizationallearning by conceptualizing embedded ties as con-sisting of fine-grained information sharing, highlevels of trust, and joint problem solving (Uzzi,1997) and by relating these three components ofembedded ties to the acquisition of capabilities.Though trust and information sharing are impor-tant factors in acquiring competitive capabilities,we suggest that joint problem solving occupiesan even more prominent role in facilitating capa-bility acquisition (see Figure 1). In our model,information sharing and trust are the preconditionsfor joint problem solving. They promote a freerexchange of ideas and a more diligent search for



Figure 1. Theoretical model of embedded ties and acquisition of competitive capabilities

solutions, which results in more rapid and effectivelearning (Jeffries and Reed, 2000). Joint problem-solving arrangements provide learning opportuni-ties by creating a forum conducive to interactionand the transfer of tacit knowledge about capabil-ities.

Transferring capabilities necessitates ‘the acqui-sition of new know-how, that is, new ways ofdoing things’ (Zander and Kogut, 1995: 78). Yet,the knowledge underlying capabilities is oftenfirm-specific and partially tacit (Polanyi, 1966;Nelson and Winter, 1982). Such knowledge isdifficult to codify and articulate, which makesit challenging to transfer (Teece, 1977; Zanderand Kogut, 1995; Szulanski, 1996). Rather thanbeing conveyed via written or physical form,idiosyncratic tacit knowledge is best transferredthrough experience, observation, or demonstration.As Hamel (1991: 99) states, ‘complex skills, basedon tacit knowledge, and arising out of a uniquecultural context may be acquirable only by up-close observation and emulation of “best in class”.’Complex skills that involve tacit knowledge andcustomization to context are promoted by jointproblem-solving arrangements that allow exchangepartners to engage in experimentation, observation,and search for solutions.

Information sharing and trust also play an impor-tant role in capability acquisition, but primarilyas precursors. More specifically, information shar-ing influences managers’ awareness of the oppor-tunities available for acquiring beneficial capa-bilities and trust influences a firm’s willingnessto accept the advice and recommendations ofan exchange partner. Taken together, informa-tion sharing and trust indirectly influence capa-bility acquisition by creating the conditions thatenable joint problem solving. Without the abil-ity to exchange sensitive and proprietary detailsabout its activities and operations, firms would

find it difficult to engage in joint problem solv-ing. Joint problem solving has a direct influenceon capability acquisition by providing an inter-active forum for learning that allows firms toobserve and experiment with capabilities in prac-tice.

In addition to affecting joint problem solving,information sharing and trust are related to eachother, as previous research has established (e.g.,Blau, 1960; Boon and Holmes, 1991; Lewickiand Bunker, 1996). While critical, the interac-tion between information sharing and trust is indi-rectly related to the acquisition of competitivecapabilities and therefore is not the focus of thispaper. Recognizing that there are reciprocal effectsbetween information sharing and trust we controlfor their interaction in our empirical models, butdo not offer a theoretical prediction for the rela-tionship.

As the intensity of joint problem solving, infor-mation sharing, and trust vary, so too does thelevel of embeddedness. Accordingly, we do notdiscretely categorize ties as being either embed-ded or arm’s length, but rather view ties as beingarrayed along a continuum anchored by these idealtypes. We elaborate the embedded ties constructsfurther and develop hypotheses that formalize pre-dictions about the effects on the acquisition ofcapabilities next.

Joint problem solving

To acquire a capability a firm must comprehend it.Joint problem solving is defined as the degree towhich the parties to an exchange ‘share the respon-sibility for maintaining the relationship itself andfor problems that arise as time goes on’ (Heideand Miner, 1992: 275). Such arrangements typi-cally involve routines for troubleshooting problemsas they arise and negotiating the mutual adapta-tions required to resolve the difficulty. Through


Embedded Ties 1037

joint problem solving, exchange partners developrelationship-specific heuristics and specialized lan-guage for conveying complex ‘chunks’ of tacitknowledge (Hansen, 1999). Such arrangementsfacilitate the transfer of situation-specific knowl-edge and, as a result, a firm will be better able tolearn about and understand a capability when it hasjoint problem-solving arrangements in place withexchange partners.

Joint problem-solving arrangements greatlyenhance the learning that occurs in exchange rela-tionships because, rather than exiting the rela-tionship when a problem arises, the parties workthrough the difficulty and receive direct feedbackabout activities and operations. This kind of inter-active relationship is particularly important for thetransfer and development of capabilities. Sinceno two firms are identical, acquiring a capabilityfrom an exchange partner requires that it be cus-tomized for the focal firm’s unique circumstancesand adapted to fit with its existing portfolio ofcapabilities. Two-way interaction is also impor-tant for transferring the tacit knowledge underlyinga capability since the recipient rarely assimilatesthe knowledge completely in a single interaction,but requires multiple interactions. By providing aforum where exchange partners can observe, expe-rience, and demonstrate the use of a capabilityin practice and receive feedback, joint problem-solving arrangements allow a firm to draw on theinsights, experience, and ability that customer andsupplier firms have with a capability. Accordingly,we hypothesize:

Hypothesis 1: Joint problem solving with leadsuppliers is positively related to acquisition ofcompetitive capabilities.

Hypothesis 2: Joint problem solving with leadcustomers is positively related to acquisition ofcompetitive capabilities.

Information sharing

For joint problem solving to occur, it is neces-sary for the exchange partners to share informationrelevant to the problem. Information sharing cap-tures ‘the degree to which each party disclosesinformation that may facilitate the other party’sactivities’ (Heide and Miner, 1992: 275). As tiesbecome more embedded, information sharing tendsto be more detailed, intricate, and proprietary than

in arm’s-length relationships in which only basicprice and quantity data are shared. Profitability,production cost data, strategic direction, and orga-nizational practices are all typical of the infor-mation exchanged in embedded ties (Uzzi, 1997).Because the information conveyed through ties thatare more highly embedded is situation-specific and‘holistic,’ in the sense that it consists of a com-posite of related details, it is both meaningful andinstructive. The information shared is also moredetailed because it pertains to the common problemthat exchange partners jointly face. Managers areable to relate the information exchanged to theirown operations and envision how it may assistthem in solving a particular problem. Moreover,since the parties to more highly embedded ties areoriented toward sustaining the relationship, theyhave an interest in seeing their exchange partnersucceed. Consequently, the parties actively provideeach other with information about potential prob-lems and opportunities that they anticipate (e.g.,market or technological trends). This informationabout potential problems and opportunities assistsfirms in narrowing the range of options to whichthey can pay attention. As exchange partners sharealternative practices and techniques for solving theproblems that arise, it enhances their awareness ofthe need for new capabilities that an organizationmight adopt. Accordingly, we predict that infor-mation sharing influences capability acquisition byfacilitating joint problem solving activities amongexchange partners.

Hypothesis 3: The relationship between infor-mation sharing with lead suppliers and acqui-sition of competitive capabilities is mediatedby joint problem solving such that informationsharing is positively related to joint problemsolving.

Hypothesis 4: The relationship between infor-mation sharing with lead customers and acqui-sition of competitive capabilities is mediatedby joint problem solving such that informationsharing is positively related to joint problemsolving.

Trust

Awareness by itself is not sufficient to cause a firmto take advantage of potentially attractive opportu-nities to upgrade its competitive capabilities. Trust



allows a firm to have confidence in the informationand advice provided by an exchange partner and tobelieve that the recommendations made are in itsown best interest (Das and Teng, 1998). For thisreason, trust and information sharing are highlyrelated (Carson et al., 2003; Ring and Van de Ven,1994). Interorganizational trust refers to a commonexpectation among the members of an organiza-tion that another organization ‘(i) makes good faithefforts to behave in accordance with any com-mitments both explicitly or implicit, (ii) is honestin whatever negotiations preceded such commit-ments, and (iii) does not take excessive advantageof another even when the opportunity is available’(Cummings and Bromiley, 1996: 303). In effect,high trust allows a firm to be confident in theveracity of the advice and recommendations of anexchange partner (Uzzi, 1997). Where trust is high,a firm expects that its exchange partner will notact in its own self-interest at the firm’s expense(Macaulay, 1963; Gulati, 1995; Zaheer and Venka-traman, 1995). Trust acts as an important filteringdevice for assessing the quality and reliability ofinformation received about opportunities. Firmsare more likely to pay attention to, weigh moreheavily, and act on information received from atrusted exchange partner (McEvily, Perrone andZaheer, 2003).

Trust also makes it possible for a firm to bemore open with its exchange partners (Dore, 1983;Ouchi, 1979). In order for lead customers and sup-pliers to provide information and guidance usefulfor solving problems and discovering new capa-bilities, it is necessary for the focal firm to revealcertain details about its operations and the chal-lenges it faces. Firms may be reluctant to revealsuch sensitive and proprietary information to acustomer out of fear that it may be used againstthe firm in the future (e.g., cost or productivitydata could be used to demand lower prices in thefuture). Firms may also not want to expose sen-sitive and proprietary information to suppliers outof concern that the information will be shared withrivals.

Similarly, when problems arise in an interfirmexchange relationship, trust facilitates the discov-ery of mutually agreeable, integrative solutions(Dore, 1983; Macneil 1980). When trust is high,exchange partners are less likely to opportunisti-cally exploit unforeseen contingencies (John, 1984)and instead view a problem as a joint one to besolved collectively. This is consistent with the idea

that firms that trust their exchange partners arecommitted to preserving their interfirm relation-ship, as opposed to merely advancing individualself gain (Uzzi, 1997). Along these lines, trustleads exchange partners to give each other thebenefit of the doubt and assume the best whenproblems emerge, rather than focus on assigningblame or questioning each other’s motives (Dyerand Chu, 2003; Zaheer, McEvily and Perrone,1998). Moreover, exchange partners that trust eachother are willing to make extra efforts beyond theletter of a contract in order to overcome difficul-ties and help each other solve problems. In sum, wepredict that trust influences the acquisition of capa-bilities by promoting joint problem solving amongexchange partners.

Hypothesis 5: The relationship between trust inlead suppliers and acquisition of competitivecapabilities is mediated by joint problem solvingsuch that trust and joint problem solving arepositively related.

Hypothesis 6: The relationship between trust inlead customers and acquisition of competitivecapabilities is mediated by joint problem solvingsuch that trust and joint problem solving arepositively related.

Content of embedded ties

We are also interested in the relative influence ofembedded ties with lead customers vs. lead sup-pliers on the acquisition of capabilities. Drawingon case studies completed at the outset of thisresearch, we suggest that there are important dif-ferences in the content of embedded ties with leadsuppliers and customers that influence the typeof knowledge exchanged and the ease of imple-menting the practices and techniques underlyinga capability. Content refers to the material (infor-mation, resources) and immaterial (social identity,authority) substance that is conveyed through atie (Burt, 1997; Podolny and Baron, 1997). Con-tent also implies the specification of ‘behavioralprocesses underlying a connection between twoactors’ (Gulati and Westphal, 1999: 473). The casestudies reveal that the content of embedded ties dif-fers in terms of the degree of firm-specific knowl-edge held by lead customers vs. lead suppliers,which holds important implications for the acqui-sition of capabilities.


Embedded Ties 1039

The case studies are based on extensive fieldinterviews with executives of four job shop man-ufacturers in the electroplating industry, their cus-tomers and suppliers, industry consultants, andsupporting regional institutions.2 In each of thefour case studies, we observed job shop manu-facturers with highly embedded ties to both theirlead customers and their lead suppliers. Relation-ships with customers tended to be long term,highly stable (with some relationships lasting upto 35 years), and few in number (e.g., 10 cus-tomers accounting for 70% of sales). Relation-ships with suppliers were also long term (e.g.,10–35 years) and few in number, with the bulkof purchases (e.g., 70–80%) coming from one ortwo suppliers. Despite the similarities between jobshops’ embedded ties with their lead customersand their lead suppliers, we also noted importantdifferences in relationships with these two kindsof exchange partners. Most importantly, the casestudies revealed that while both customers and sup-pliers convey knowledge pertinent to job shops’capabilities, the type of knowledge conveyed andhow it influences job shops’ capabilities differed.

Joint problem solving between job shops andtheir lead customers was predominantly focused onevaluating, controlling, and improving the qualitylevels of the metal coatings applied to customers’parts. The desire to achieve and maintain highlevels of quality influenced the type of exchangerelationship lead customers had with job shops.The exchange relationships became more deeplyembedded, as the following remarks indicate:

Because of the process and the sensitivity of it,we don’t move around much and as long as I’veworked here (25 years), [they are] the only onethat has plated these . . . We have a very smallsupplier base, which allows you to have a veryclose working relationship. And I go back andforth to our suppliers and they’re here. You get toknow them on a, even a personal basis because youbecome almost one . . . We’ve worked with them so

2 A total of 20 interviews were completed with 16 individualsfrom 13 different organizations. All interviews were completedon-site using a semi-structured interview format and includedplant tours. Interviews varied from 1 to 4 hours and all inter-views were tape-recorded (approximately 30 hours) and tran-scribed (roughly 300 pages). In addition to interview transcripts,the case studies are based on field notes taken during inter-views and plant tours that capture our direct observations oforganizing practices and principles. Company records, reports,and correspondence constitute a third source of data for the casestudies. Complete details of the case studies are available fromthe authors upon request.

long they can probably tell us what we need beforewe know what we need . . . They know us, whatour requirements are. We work more efficientlytogether. For me it’s a real trust type of thing. Theyget to know you and it’s just like your family, yougo to bat for your family.

Such exchange ties provide a context conducive tojoint problem solving on issues of common interestsuch as quality. In the exchange relationships weobserved, the approach taken to improving qualitycentered on identifying variations in key propertiesof metal coatings, which can be readily measuredusing objective indicators. Drawing on their analy-sis of the quality of the coatings applied, customersprovide job shops with feedback about the fre-quency and type of variation observed. This isbased on measurable and precise indicators of ajob shop’s output. Such output-based knowledgeallows job shops and their customers to narrowthe range of potential causes of variation, be theyin the job shop’s process, the way the customer hasmanufactured a particular part, or some combina-tion of the two. The CEO of one customer firm weinterviewed provided the following description ofhow they work with job shops to address qualityproblems:

There’s such an interaction between what we sup-ply to a plater. What kinds of oils do we use, areour parts clean? Our surface—what cutting toolswere used, they may have added something to thesurface of the metal. All these kinds of things arefactors. When there is a problem, I can call and sayhey, we’ve got a problem, come on over here andlook at things. Maybe we screwed up on something.

Drawing on customers’ output-based knowledge,the parties work together to pinpoint the sourcesof the problem and, when appropriate, customerssuggest specific quality practices for the job shopto consider in order to improve quality. However,customers rarely provide hands-on technical assis-tance to help job shops integrate new techniquesinto their operations.

Joint problem solving with suppliers is some-what different, due in part to differences in theability to verify the quality and performance of asupplier’s product, which in turn affects the typeof knowledge exchanged. In many ways, suppli-ers are the most important and closest technicaladvisors that job shops have. Job shops frequentlyrely on suppliers for technical advice when trou-bleshooting or developing new ways of doing



things. Because suppliers are so knowledgeable,job shops are willing to entrust critical processesin the hands of these experts. Job shops are alsovery loyal to their suppliers for the same reason.In the words of one job shop President, ‘this isa technical industry, a supplier’s knowledge andcredibility are very important. I know that my cur-rent suppliers won’t lead me astray.’

Joint problem solving with suppliers was bestexemplified in one of the job shops we studiedthat was working on a pollution prevention projectto replace cyanide with a more benign chemi-cal process. In order to facilitate the transition tothe alternative process, the job shop constructed acompletely new manufacturing line and had techni-cians from the supplier working with it on-site overa period of a couple of weeks. Initially, the twostruggled to get the new chemical process to workto the job shop’s satisfaction. Whereas the perfor-mance of the cyanide process was very consistent,the new chemical process was highly sensitive tovariations in operating conditions (e.g., time, tem-perature, concentration, pH level) that interacted inways that were not well understood. After repeatedtrial-and-error experimentation, adjustment, analy-sis, and checking, the new process finally achievedacceptable performance on a consistent basis. Asthe job shop CEO described:

They had a couple of tech people in here most ofthe time, you know holding our hand and actuallyworking with us hands on to start with. All day fora few days, every day. I mean they were here whilewe were plating and they were constantly analyz-ing and adjusting, constantly checking everythingthat was happening. And as we got a little morecomfortable, they spent less and less time [here].

Since the operating conditions, equipment, levelof expertise, and applications differ from one jobshop to the next, the way that a supplier’s productperforms in a given setting varies as well. Onesupplier representative we spoke with describedthe challenge of servicing his product to differentjob shops in the following way, ‘You may end upwith a totally divergent process, in most cases youdo because no two shops are alike.’ As a result,assessing quality and performance in exchangerelations with lead suppliers is idiosyncratic andnot readily verified with objective measures, whichis an important difference from exchange relationswith lead customers.

This underscores a key distinction between thecontent of ties with lead customers and lead sup-pliers, both in terms of the type of knowledgeexchanged and the way they work with job shopsto solve problems. Since the quality and perfor-mance of the product exchanged with suppliersis less easily verified than is the case with cus-tomers, the type of knowledge exchanged withsuppliers centers on implementation of practicesand techniques. Having the ability to observe sup-pliers demonstrate new processes in a ‘hands-on’setting is a highly effective way to solve problemsand convey knowledge that is technically com-plex and difficult to articulate. Although lead cus-tomers also exchange knowledge with job shopsin the course of solving problems, the knowl-edge is output-based due to the relative ease ofverifying product quality and performance. Conse-quently, the knowledge exchanged with lead cus-tomers during joint problem solving influences jobshops’ acquisition of capabilities by raising aware-ness of, and affecting the intention to acquire,practices and techniques. Knowledge exchangedwith lead suppliers goes one step further by alsofacilitating the implementation of practices andtechniques. Because of this difference in the con-tent of ties, we expect that embedded ties withlead suppliers will exert a greater influence on theacquisition of capabilities than embedded ties withlead customers. This leads to our final hypothe-sis:

Hypothesis 7: Joint problem solving with leadsuppliers will be more strongly related to acqui-sition of competitive capabilities than joint prob-lem solving with lead customers.

RESEARCH METHODS

To test our predictions we identified a populationof firms—job shop manufacturers—that differ intheir degree of embeddedness with lead customersand suppliers and that vary in their propensity toacquire competitive capabilities. An advantage ofthe field setting is that the job shop manufacturersthat we study are similar in size, rely on similarproduction technologies, and compete in similarmarkets. Moreover, since job shop manufacturerstend to be regionally based, we can control forgeographical differences.


Embedded Ties 1041

Research design and data collection

The sampling frame consisted of all job shop man-ufacturers operating in seven industries in twoMidwestern states: electroplating (SIC 3471), coat-ing and painting (SIC 3479), printed circuit boardmanufacturing (SIC 3672), screw machining (SIC3451), stamping (SIC 3469), sheet metal fabrica-tions (SIC 3444), and machining (SIC 3599). Weselected these industries because they were simi-lar in terms of their composition (e.g., number offirms, average age of firms, customer base).

For the purpose of our research we constructeda primary dataset using a mailed questionnaire.Officials from a regional institution located in bothstates—the Manufacturing Extension Partnership(MEP)3 —compiled and made available a list ofall firms operating in the job shop industries understudy. These lists were not limited to only thosefirms that had participated in MEP programs, butalso identified job shop manufacturers in the statesthat had not participated. These lists were alsomore comprehensive and accurate when comparedto directories available from commercial (e.g., Dunand Bradstreet) or other governmental (e.g., statedirectories of manufacturing firms) sources.

A preliminary version of the survey instrumentwas pretested among a group of 22 executivesof local job shop manufacturing companies froman industry not among those identified above.Feedback from these executives was incorporatedinto a revised version of the survey instrument,along with comments and suggestions from indus-try experts, officials from the MEPs, and severalcolleagues knowledgeable in survey design.

The final questionnaire was mailed to a ran-dom sample of 1000 chief executive officers orpresidents of job shops. We directed the surveyto CEOs/Presidents because our preliminary fieldinterviews indicated that these individuals werebest able to respond to questions about organiza-tional and strategic issues relating to their respec-tive firms. This approach is consistent with theselection of key informants knowledgeable aboutorganizational matters by virtue of their position(John and Weitz, 1988). We also implemented Dil-man’s (1978) techniques for maximizing response

3 The Manufacturing Extension Partnership (MEP) is a nationalsystem of extension centers operated by the National Instituteof Standards and Technology. The charter of the regionallybased MEPs is to promote the deployment of new manufacturingtechnology, production techniques, and business practices amongsmall manufacturers.

rate. In addition to the initial survey mailing, exten-sive follow-up communications were carried outincluding: (1) sending a reminder/thank you post-card, (2) sending a second round of mailings tononrespondents, and (3) placing a telephone callto any remaining nonrespondents.

A total of 309 executives responded to therequest for information about their company. Thisnumber is approximately 31 percent of the origi-nal 1000 firms surveyed. The follow-up telephonecalls revealed that 178 of the original 1000 firmswere not eligible to participate in the study, andwe eliminated these firms from our initial samplingframe.4 The actual response rate then equaled 38percent (i.e. 309/822) of eligible firms. Completeresponses were obtained from 234 of these firms.All of the firms from which we obtained responseshad 500 or fewer employees. Further, 75 percentof the firms had 63 or fewer employees. On aver-age, firms in our sample had sales of $2 millionand had been in operation for 30 years. Over 95percent of the firms in the sample were privatelyowned.

Testing for nonresponse bias

The overall response rate of 38 percent, whilereasonable, raises the possibility that the sampleof responding firms systematically differed fromthe remainder of the population. We addressedthe potential for nonresponse bias by comparingcertain key attributes of respondents (firm size interms of the total number of employees and annualsales) to those of a group of 50 randomly selectednonrespondents. We obtained size and sales datafor the 50 nonrespondents from one of the MEPs.t-tests revealed no significant differences betweenthe mean size (t = −1.29) and the mean sales(t = −1.83) of respondents and nonrespondents,although the near significance of the difference inthe mean sales suggests the possibility of some biasin our sample toward firms with higher sales. Tofurther confirm the representativeness of our sam-ple, we conducted a Kolmogorov–Smirnov two-sample test. For the variables of both size andsales we found no significant differences betweenour sample and the random sample of 50 nonre-spondents. p-values were, respectively, 0.225 and

4 Firms deemed ineligible to participate in our study includethose that were not job shop manufacturers, had gone out ofbusiness, or had zero employees.



0.357, suggesting that the two samples were drawnfrom the same population. While not definitive,these tests provided some assurance that the sam-ple of firms responding to the questionnaire wasclosely representative of the broader populationsurveyed (Siegel, 1956).

Operational measures

Table 1 presents the details of the measurementinstruments and scales used to operationalize ourtheoretical constructs. The Cronbach α reliabili-ties for each construct are also reported in Table 1.With the exception of the quality management con-struct, which is marginal at 0.61, constructs wereat or above the value of 0.70 (Nunnally, 1978).Descriptive statistics and zero-order correlationsamong constructs are reported in Table 2.

Acquisition of competitive capabilities

Consistent with Meyer and Goes (1988) we viewcompetitive capabilities as being acquired througha multistage, organizational process. To developa variable for statistical analysis, we created a 7-point scale that captures three primary stages asso-ciated with the acquisition of capabilities: knowl-edge–awareness, evaluation–choice, and adop-tion–implementation.

We evaluated two competitive capabilities usingthe acquisition scale: pollution prevention andquality management. Following Amit and Schoe-maker (1993), we focused on industry-specificcompetitive capabilities that were identified dur-ing the initial phase of field research. A carefulanalysis of the field interviews and subsequent con-firmation by industry experts strongly suggestedthat central to sustaining competitive advantagein this industry were: (1) effective managementof hazardous materials via pollution prevention,and (2) minimization of variations in productionprocesses via quality management. Acquisition ofcapabilities in these areas was critical because theindustry faced key challenges from environmentalrisks and regulation, heightened competition and arapidly changing market, and innovations in pro-cessing chemistry and equipment.

Pollution prevention. Pollution prevention capa-bilities have become salient with the passage oflaws and treaties governing the use, emission, andreporting of hazardous materials integral to the

production process of the industries we examined.5

According to the U.S. Environmental ProtectionAgency (1990), pollution prevention emphasizesthe judicious use of resources through sourcereduction, energy efficiency, reuse of scrap mate-rials during production, and reduced releases ofhazardous or toxic materials. We operationalizedpollution prevention with three items measuringvarious product and process changes. These itemswere adapted from a national survey of pollutionprevention conducted in one of the job shop indus-tries selected for this study (Cushnie, 1994).

Quality management. Although total quality man-agement (TQM) means different things to differ-ent people, the elimination of production defectsthrough continuous improvement of processes isa common feature of most definitions (Hackmanand Wageman, 1995). A core principle guidingthe management of quality improvement is theanalysis of variability in processes and outcomesthrough systematic data collection and statisti-cal analysis (Deming, 1986; Juran, 1974). In thisstudy, TQM was measured with three items cap-turing the use of statistical process control charts toprovide operators with feedback. These items arebased on the measurement instrument developedby Flynn, Sakakibara, and Schroeder (1995).

Embedded ties

We developed parallel instruments to measure theextent to which a firm has embedded ties withits lead customer and lead supplier in terms ofjoint problem solving, information sharing, andtrust. Our joint problem-solving construct wasmeasured using a three-item instrument based onthe scale developed and validated by Heide andMiner (1992). The instrument captures the degreeto which exchange partners share the responsibil-ity for resolving problems as they arise. Informa-tion sharing captures the degree to which partiesactively exchange information beyond the letterof the contract—information that can facilitatethe other party’s activities. We used a modified

5 Some of the most relevant examples include amendments toenvironmental regulations, such as the Superfund Amendmentsand Reauthorization Act of 1986 (SARA), the Clean WaterAct Amendments of 1987, the Global Climate Protection Actof 1987, the Clean Air Act Amendments of 1990, and thePollution Prevention Act of 1990. The Montreal Protocol signedin 1990 put additional pressure on firms to eliminate or reducesolvent use.


Embedded Ties 1043

Table 1. Measurement instruments

Measurement items Scales Internalconsistency

reliability (α)

Pollution prevention capabilities 0.721. Substitute less hazardous raw materials for more

hazardous ones1 = We know little about this

practice2. Offer new products/services because of low waste

disposal costs2 = We know about this practice,

but do not do it3. Discontinue products/services high in environmental 3 = We have considered doing this

management costs 4 = We decided not to do this

after considering itQuality management capabilities 5 = We do this from time to time1. Collect data on your company’s production process 6 = We do this most of the time 0.61

variations 7 = We do this all of the time2. Provide charts and graphs to production employees

reporting defect rates3. Conduct experiments to isolate causes of defects

Joint problem solving1. Our main [customer/supplier] works with us to 1 = Strongly disagree Customer 0.79

overcome difficulties 2 = Disagree2. We are jointly responsible with our main 3 = Slightly disagree Supplier 0.83

[customer/supplier] for getting things done 4 = Neutral3. We work with our main [customer/supplier] to help 5 = Slightly agree

solve each other’s problems 6 = Agree7 = Strongly agree

Information sharing1. Our main [customer/supplier] warns us of events that

may create problems for us

Customer 0.772. Our main [customer/supplier] shares its plans for the

future with usSupplier 0.72

3. Our main [customer/supplier] shares proprietary andsensitive information with us

Interorganizational trust1. Our main [customer/supplier] negotiates fairly with us Customer 0.872. Our main [customer/supplier] does not mislead us Supplier 0.863. Our main [customer/supplier] keeps its word

Vicarious learningTo what extent does your company: 1 = To no extent 0.821. Look at different approaches used by other 4 = To some extent

companies? 7 = To a great extent2. Get ideas about new ways of working from other

companies?3. Improve the way you work by observing what other

companies do?

Participation in regional institutions1. Obtain on-site assistance at your company from

[name of extension center]1 = We know little about this

service0.75

2. Select/install new equipment or computer systemswith [name of extension center]

2 = We know about this service,but do not do it

3. Participate in user groups or networks organized by 3 = We considered doing this[name of extension center] 4 = We decided not to do this

after considering it5 = We did this once6 = We did this a couple of times7 = We did this several times

(continued overleaf )



Table 1. (Continued )

Measurement items Scales Internal consistencyreliability (α)

Network structure (non-redundancy)1a. Please write the initials of the five most important

people not employed by your company that you relyon for advice about managing your business

(not applicable) (not applicable)

1b. Now, using the table provided indicate if thesepeople know each other. If so, circle ‘Y’ for yes

Firm sizea

1. Roughly how many full-time equivalent employeesworked for you in fiscal year 1995?

(not applicable) (not applicable)

2. Roughly how many temporary and seasonalemployees worked for you in fiscal year 1995?

a Firm size = sum of items 1 and 2 (standardized z-score).

Table 2. Descriptive statistics and zero-order correlations among constructs

Mean S.D. 1 2 3 4 5 6 7 8 9 10 11

Competitive capabilities1. Pollution prevention 3.83 1.782. Quality management 4.51 1.50 0.30∗∗

Embedded ties (supplier)3. Problem solving 5.34 1.45 0.21∗∗ 0.23∗∗

4. Information sharing 4.02 1.33 0.28∗∗ 0.13∗ 0.40∗∗

5. Trust 5.72 0.97 0.09 0.07 0.35∗∗ 0.41∗∗

Embedded ties (customer)6. Problem solving 6.50 0.77 0.09 0.12∗ 0.29∗∗ 0.10 0.027. Information sharing 5.01 1.32 0.08 0.14∗ 0.33∗∗ 0.28∗∗ 0.15∗∗ 0.20∗∗

8. Trust 5.89 1.03 0.01 0.05 0.25∗∗ 0.14∗ 0.20∗∗ 0.16∗∗ 0.57∗∗

Controls9. Vicarious learning 4.55 1.47 0.25∗∗ 0.29∗∗ 0.11 0.15∗ 0.01 0.03 0.18∗∗ 0.0510. Regional participation 2.13 1.40 0.13∗ 0.22∗∗ 0.08 0.00 −0.01 −0.00 0.13∗ 0.05 0.21∗∗

11. Network structure 0.99 0.63 0.13∗ 0.03 0.03 −0.02 −0.04 −0.01 0.03 0.04 0.03 −0.12∗

12. Firm size 0.00 1.00−0.02 0.26∗∗ 0.08 0.02 0.04 −0.10 0.01 −0.12∗ 0.03 0.03 −0.04

∗∗ p < 0.01; ∗ p < 0.05

version of the instrument developed and validatedby Heide and Miner (1992) for measuring infor-mation sharing. We operationalized interorganiza-tional trust in a lead customer or supplier with athree-item scale reflecting the degree to which theexchange partner is fair in its dealing and does nottry to take advantage of the focal firm. Interor-ganizational trust describes the extent to whichthe members of a focal firm have a collectivelyheld trust orientation toward a customer or supplierfirm (Zaheer et al., 1998). Our trust measurementinstrument is based on a shortened version of thescale developed and validated by (Cummings andBromiley, 1996).

Control variables

In order to assess the unique contribution of ourpredictions, we control for alternative explanationsfor the sources of capabilities.

Learning orientation. In addition to learning fromtheir own experiences, we have proposed that firmslearn from the experience of others, particularlylead customers and suppliers. However, it may bethe case that the overriding determinant of exter-nal sources of capabilities is a firm’s propensity todraw on its network of interorganizational ties andthat linkages to customer and supplier firms, per


Embedded Ties 1045

se, are no more significant than other ties. Withoutcontrolling for the general inclination to learn fromnetwork contacts, linkages to customer and sup-plier firms may be reflecting some of the effects offirms learning from their overall networks. To ruleout this possibility, we control for a firm’s vicari-ous learning orientation (Levitt and March, 1988).The construct was measured using an instrumentbased on the measurement scale developed byYeung and Ulrich (1994) that captures the extentto which a firm relies on the past experiences ofothers.

Network structure. Both Kraatz (1998) andMcEvily and Zaheer (1999) found that the struc-ture of a firm’s overall network influences itsacquisition of capabilities. In particular, the extentto which contacts are not directly connected (i.e.,nonredundant) affects the diversity of informa-tion accessed and the likelihood of discoveringnew information and opportunities (Burt, 1992). Tooperationalize nonredundancy, this study uses anego-centered network measure based on an instru-ment designed and developed specifically for usein the small firm context (Aldrich, Rosen, andWoodward, 1986). This instrument asks respon-dents (ego) to identify the five most importantexternal sources of advice (alters) relied upon andto report the extent to which these five sourcesknow each other. Using this matrix, a nonre-dundancy score, indicating the ratio of nonre-dundant ties per advisors, was computed as fol-lows:

Nonredundancy = (Potential Ties − Actual Ties)/

Number of Advisors

where Potential Ties = the maximum number ofties that could exist among advisors (0 to 10), orn(n − 1)/2, where n is the total number of advi-sors listed; Actual Ties = the number of ties thatdo exist among advisors (0 to 10); and Number ofAdvisors = the total number of advisors listed (0to 5).

Regional associations. Following Kraatz (1998)and McEvily and Zaheer (1999) we control forparticipation in industry or regional associationsthat provide firms with access to a greater breadthof knowledge about competitive capabilities. Thisconstruct indicated the extent to which a firm used

the services available from a regional industrialextension center. A 7-point scale similar to theacquisition scale was developed to measure partic-ipation and capture three primary decision-makingstages: knowledge–awareness, evaluation–choice,and utilization. Participation was measured with athree-item scale capturing the use of services avail-able.

Firm size. Larger firms have greater slack re-sources, such as managerial and engineering exper-tise and time, thus making experimentation withnew practices and techniques more feasible (Kel-ley and Brooks, 1991). Moreover, since the scaleeconomies associated with spreading the costs ofimplementing capabilities over a larger base ofoperations are greater in larger firms, they may findit easier to acquire capabilities than smaller firmsmay. Consequently, the final factor for which wecontrol is firm size, operationalized as the sum offull-time employees, and temporary and seasonalworkers.

Construct validity

A measure of a construct is valid to the extentthat it actually measures what it purports to mea-sure (Carmine and Zeller, 1979). The logic ofconstruct validity suggests that multiple indica-tors of the same theoretical construct should bepositively and strongly related. More specifically,the convergent validity of a construct is demon-strated by showing that each indicator loads onlyonto its associated theoretical construct. We eval-uated the convergent validity of the constructsin this study by examining the factor loadingsof each operational variable (i.e., correspondingto a survey item) on its associated latent con-struct. Factor loadings were estimated using struc-tural equation modeling. The results of this anal-ysis revealed that the factor loadings on eachconstruct were large and statistically significant(p < 0.001). This suggests that each indicator isstrongly related to its underlying construct and isevidence supporting the convergent validity of themeasures.

Discriminant validity

The degree to which two theoretical constructs dif-fer from each other indicates discriminant validity.The discriminant validity of two constructs can



be assessed by demonstrating that the correlationbetween a pair of constructs is significantly dif-ferent from unity. We test discriminant validity bycomparing a constrained structural equation model(i.e., correlation fixed to equal 1.0) with an uncon-strained model (i.e., correlation freely estimated).A significantly lower χ 2 value for the uncon-strained model supports the discriminant validitycriterion. Separate structural equation models werespecified to test the discriminant validity of thetheoretical constructs: one for the dependent vari-ables (acquisition of capabilities) and one for theindependent variables (embedded ties and controlvariables). Each model was first estimated withall correlations among latent variables left uncon-strained and then, one at a time, the correlationbetween each pair of latent variables was con-strained to unity and the models were re-estimated.In all cases, the χ 2 value for the unconstrainedmodel was significantly lower (p < 0.001) thanthe χ 2 value for the constrained model indicatingsatisfactory discriminant validity.

Assessing common method variance

As noted earlier, the bulk of the firms in our sam-ple are owner-operated small manufacturing firms.The CEO or President of these firms was typi-cally the only individual knowledgeable enoughto respond to the survey and secondary data forthe variables of interest were not available. Con-sequently, the potential for common method vari-ance may exist since a single data source wasused to measure the independent and dependentvariables. To investigate this possibility we imple-mented a variation of Harmon’s single factor testusing structural equation modeling. This techniqueassumes that ‘if a substantial amount of com-mon method variance is present . . . one “general”factor will account for the majority of covari-ance in the independent and criterion variables’(Podsakoff and Organ, 1986: 536). We tested forcommon method variance by comparing a modelloading all of the observed variables onto a sin-gle latent variable with a measurement modelthat loaded observed variables onto theoreticallyassigned latent variables. The χ 2 value for themeasurement model was significantly lower thanthe χ 2 value for the single factor model (differ-ence in χ 2 = 906.81, 2 d.f., p < 0.001), indicatinga superior fit to the data. We also evaluated the

potential for common method variance by perform-ing a principal components factor analysis on all ofthe items. The factor analysis extracted 10 factors,with the first factor accounting for 21 percent ofthe total variance. While by no means definitivetests, these results provide some indication thatthe observed relationships among constructs arenot largely accounted for by systematic varianceassociated with the measurement technique.

ANALYSIS AND RESULTS

To test the hypotheses developed earlier we spec-ified a structural equation model using LISREL 8(Joreskog and Sorbom, 1996). We used the maxi-mum likelihood procedure to estimate the model.We chose to analyze our data using structuralequation modeling because it offers the advantageof specifically modeling measurement error associ-ated with indicators, rather than assuming that con-structs are measured without error. More impor-tantly, structural equation modeling provides thecapability of simultaneously estimating a system ofstructural equations, which is important for testingthe mediation model proposed. Specifically, ourtheoretical model predicts that information shar-ing and trust influence the acquisition of capabili-ties through their effect on joint problem solving,which is hypothesized to be directly related tocapabilities acquisition (see Figure 2). Althoughnot shown, the model also includes controls forfirms’ learning orientations, network structures,participation in regional associations, and size.6

The structural equation model estimated a seriesof path coefficients reflecting the relationshipsspecified among the latent variables. The betacoefficient (β) indicates the structural path amongdependent (endogenous) latent variables, which inthis case is the link between joint problem solv-ing and acquisition of competitive capabilities.The gamma coefficient (γ ) represents the struc-tural paths between the independent (exogenous)latent variables and the dependent (endogenous)latent variables. The association between trust andjoint problem solving is an example of a gammapath coefficient. The phi coefficient (φ) represents

6 In separate analyses we have investigated the possibility ofindustry effects on the results of this model. The inclusion ofindustry controls does not alter the sign or level of statisticalsignificance of any predictor variables in the structural equationmodel.


Embedded Ties 1047

Figu

re2.

Em

bedd

edtie

s:st

ruct

ural

equa

tion

mod

el



the association among exogenous variables (e.g.,information sharing and trust). The sign and sta-tistical significance of the beta and gamma pathcoefficients serve as a test of the hypothesized rela-tionships.

The results for the structural equation modelinclude the path coefficients and t-values (shown inparentheses) corresponding to each hypothesizedrelationship, and a series of standard fit indices7

reflecting the degree of overall fit between theactual and predicted covariances among variablesin the model. The χ 2 statistic tests the correspon-dence between the model and the underlying data.

The estimated structural equation model demon-strated a χ 2 (424 d.f.)8 value of 444.36 (p = 0.24).All fit indices are 0.87 or above, indicating areasonable fit of the model with the data (seeFigure 2).

Hypothesis 1. The model provided strong supportfor the hypothesis predicting a positive relation-ship between joint problem solving with suppliersand acquisition of competitive capabilities. As pre-dicted, the joint problem solving with suppliers topollution prevent link is both positive and statisti-cally significant (β13 = 0.36, t = 3.91, p < 0.001).Similarly, the relationship between joint problemsolving with suppliers and quality managementcapabilities is positive and statistically significant(β23 = 0.20, t = 2.49, p < 0.05).

Hypothesis 2. The prediction that joint problemsolving with customers and acquisition of compet-itive capabilities would be positively related is notsupported. Neither the relationship between jointproblem solving with customers and pollution pre-vention (β14 = −0.02, t = −0.25, n.s.) nor the linkto quality management capabilities is statisticallysignificant (β24 = 0.06, t = 0.74, n.s.).

7 Four fit indices are reported for each of the models estimated.The goodness of fit index (GFI) indicates the relative amountof variance and covariance jointly explained by the model. Theadjusted goodness of fit index (AGFI) is similar to GFI, butadjusts for the number of degrees of freedom in the model. Thenormed fit index (NFI) (Bentler and Bonett, 1980) represents thepoint at which the model being evaluated falls on a scale from anull model (specifying mutual independence among indicators)to a perfect fit. The comparative fit index (CFI) (Bentler, 1990)is the same as the NFI, but corrects for small sample size bysubtracting the degrees of freedom from their corresponding chi-square values. Each index ranges from zero to 1.00, with valuescloser to 1.00 indicating a good fit. A commonly accepted ruleof thumb is that a fit index should be greater than 0.90.8 Using the guidelines developed by MacCallum, Browne, andSugawara (1996) we can conclude that the model achievesadequate statistical power given the degrees of freedom andnumber of observations.

Hypothesis 3. The hypothesis predicting a positiverelationship between information sharing and jointproblem solving with lead suppliers is supported.The relationship is positive and statistically signif-icant (γ31 = 0.57, t = 5.98, p < 0.001).

Hypothesis 4. Similarly, the hypothesized relation-ship between information sharing and joint problemsolving with lead customers is supported. As pre-dicted, the relationship is positive and statisticallysignificant (γ43 = 0.41, t = 4.96, p < 0.001).

Hypothesis 5. The predicted positive relationshipbetween trust and joint problem solving with sup-pliers is also supported by the data. The linkbetween the two variables was not only posi-tive, but also statistically significant (γ32 = 0.28,t = 3.75, p < 0.001).

Hypothesis 6. Additionally, the hypothesis relatingtrust to joint problems solving with customers issupported. As predicted, the relationship is positiveand statistically significant (γ44 = 0.50, t = 6.51,p < 0.001).

Hypothesis 7. The final prediction that joint prob-lem solving with suppliers would more stronglyinfluence the acquisition of capabilities than jointproblem solving with customers is also borne outby the data. As the model indicates, joint prob-lem solving with suppliers shows a positive effecton the acquisition of both capabilities, while jointproblem solving with customers is unrelated to thetwo capabilities.

Although not hypothesized, the structural equa-tion model also estimates the relationship betweeninformation sharing and trust. As can be seen inFigure 2, the relationship is positive and statis-tically significant for both customer (φ78 = 0.64,t = 6.89, p < 0.001) and supplier relations (φ56 =0.47, t = 5.36, p < 0.001).

Control variables

The control variables (not shown in the model) arealso significantly related to the acquisition of capa-bilities. Participation in regional institutions showsa positive and statistically significant relationshipto both pollution prevention (β15 = 0.17, t = 1.93,p < 0.055) and quality management capabilities(β25 = 0.21, t = 2.52, p < 0.05), as does vicar-ious learning (γ15 = 0.18, t = 2.34, p < 0.05 forpollution prevention and γ25 = 0.27, t = 3.54, p <

0.001 for quality management). Network structureis positively and statistically significantly related topollution prevention capabilities (γ16 = 0.24, t =


Embedded Ties 1049

2.23, p < 0.05), while the link between firm sizeand quality management capabilities is positiveand statistically significant (γ27 = 0.41, t = 5.23,p < 0.001).

Robustness of the results

The structural equation model estimated appearsto provide strong support for the hypothesizedrelationships linking embedded ties with suppli-ers to capability acquisition. However, the modelwas less supportive of the relationship betweenembedded ties with customers and the acquisi-tion of capabilities. To explore the stability ofthe results, we performed additional analyses thatmight reveal alternative formulations of the rela-tionship between embedded ties with customersand the acquisition of capabilities. In particular,we examined whether each attribute of embeddedties with customers is directly related to acquisitionof capabilities, rather than information sharing andtrust being mediated by joint problem solving. Totest this possibility we added direct paths from cus-tomer information sharing and trust to both capa-bilities. The revised model’s fit is not significantlybetter and none of the added paths are statisti-cally significant. As a further test of the possibledirect effects of customer information sharing andtrust on acquisition of capabilities, we estimated athird model that removed the paths linking infor-mation sharing and trust to joint problem solvingwith customers. Whereas the second model esti-mated direct and indirect paths, the third modelonly includes direct paths. The third model’s fitis substantially worse than the first model’s andonce again none of the direct paths from informa-tion sharing and trust to capabilities are statisticallysignificant. Taken together, the overall pattern ofresults offers consistent evidence that fails to sup-port the relationship between embedded ties withcustomers and the acquisition of capabilities.

We also repeated these analyses for the embed-ded ties with suppliers variables. Once again, nei-ther the added direct path from information shar-ing with suppliers nor the direct path from trustin suppliers reveal a statistically significant rela-tionship with either pollution prevention or qual-ity management capabilities. We then removed thepaths from information sharing and trust to jointproblem solving with suppliers to test for theirdirect effects on capability acquisition alone. Thepaths from trust in suppliers to capabilities remain

non-significant, but one of the two information-sharing paths (to pollution prevention capabilities)is positive and statistically significant (γ13 = 0.25,t = 2.42, p < 0.05). However, the fit of this modelis significantly worse than the initial model (thedifference in χ 2 = 115.57, with 2 d.f., and is sig-nificant at p < 0.001). This suggests that the initialmodel, with joint problem solving as a mediator,provides the best fit to the data.

As a final check of the robustness of the pro-posed model, we investigated alternative specifica-tions of the latent constructs to determine whethera different configuration of variables provided asuperior fit to the data. We estimated a seriesof models that explored various combinations ofvariables (e.g., combining information sharing andtrust, combining customer and supplier variables).In all cases, the fit of the alternative models wassubstantially worse than the proposed model. Anexamination of the modification indices providedby LISREL further confirmed that there are nosubstantial sources of variation that the proposedmodel has omitted. Based on these findings, weare reasonably confident that the proposed modelprovides the best fit of the data.

Discussion

The findings from this research broaden and deepenour understanding of how firms acquire competi-tive capabilities and underscore the need to recon-sider existing assumptions regarding the sources ofcapabilities. Whereas most extant work on the ori-gins of capabilities has looked within the firm, weextend recent strategy research that views capa-bilities as externally embedded in a firm’s net-work of interorganizational relationships (Ahuja,2000; Gulati, 1999; McEvily and Zaheer, 1999;Stuart, 1998). The evidence reported here is con-sistent with the notion that interfirm exchange rela-tionships represent an important external learningopportunity for discovering, evaluating, and imple-menting competitive capabilities. In addition toengaging in their own trial-and-error experimenta-tion to develop capabilities, firms learn about capa-bilities vicariously through embedded ties withcritical exchange partners. But, perhaps the moreimportant contribution of this research is to provideinsight into the underlying mechanisms that facili-tate the acquisition of capabilities through interfirmties.



Previous network research on the acquisitionof capabilities has identified information sharingand trust as key mechanisms promoting capabilityacquisition. The results reported here confirm thatthese mechanisms are in fact instrumental to theacquisition of capabilities through interfirm ties,but further indicate that they play a more sub-sidiary role than previously thought. In this studyjoint problem-solving arrangements is the moreprominent driver of capability acquisition and actsas a critical linking mechanism between embed-ded ties and the acquisition of capabilities. Jointproblem solving provides a forum for managers toimprove their comprehension of the tacit knowl-edge underlying capabilities and their understand-ing of how to customize a capability to the uniquecircumstances of their firm. Through experimen-tation, observation, and demonstration, firms areable to draw on the insights, experiences, and abil-ities that lead suppliers have with a capability.Information sharing and trust, on the other hand,exert indirect effects on the acquisition of capa-bilities by creating a foundation upon which jointproblem solving can occur. Viewed this way, infor-mation sharing and trust act as precursors to themore immediate effects of joint problem solvingon capability acquisition.

The observation that trust enables other criti-cal processes is particularly noteworthy. Recentresearch on work group performance has observeda similar finding. Rather than having a direct effecton performance, trust influences how motivationis translated into group processes and performance(Dirks, 1999; Dirks and Ferrin, 2001). More gen-erally, the accumulating evidence that trust actsas a facilitator of a variety of organizational out-comes is consistent with the notion that ‘Trust isan important social lubricant of a social system’(Arrow, 1974: 23).

Another contribution of this research is to revealthe heterogeneous influence of different types ofembedded ties on the acquisition of capabilities.The results of our empirical analysis and casestudies clearly indicate that embedded ties withlead suppliers exhibited a greater influence onthe acquisition of capabilities than did embeddedties with lead customers. The difference betweenthe effect of embedded ties with lead suppliersvs. lead customers and acquisition of capabili-ties is most evident in the extent of joint prob-lem solving that occurs in each type of relation-ship. With customers, joint problem solving is

practiced at a high level almost uniformly by allfirms and there is little variation (mean is 6.50and S.D. is 0.77). Consequently, there is littleopportunity for achieving competitive advantagethrough the acquisition of capabilities. Conversely,joint problem solving with suppliers is practicedto a lesser degree (mean = 5.34) and varies to agreater extent (S.D. = 1.45), and therefore is moreconsequential for achieving competitive advantagethrough the acquisition of capabilities. The factthat more firms engage in joint problem solvingwith customers makes it less valuable from the per-spective of achieving competitive advantage fromthe acquisition of capabilities.

The observation that lead suppliers were instru-mental in facilitating the implementation of thepractices and techniques underlying a capabilityfor competitive advantage suggests that in orderto influence capability acquisition embedded tiesmust go beyond raising awareness of opportuni-ties to enhance performance. Embedded ties mustalso reduce the uncertainty associated with imple-menting a new capability. This finding highlightsthe importance of the content of ties for the acqui-sition of capabilities. The relative ease with whichlead customers could verify the quality and per-formance of goods exchanged limited their knowl-edge of the firm’s operations. For this reason,knowledge exchanged in embedded ties with leadcustomers was output-based. Conversely, the qual-ity and performance of goods exchanged with leadsuppliers was difficult to verify, requiring lead sup-pliers to possess a greater degree of firm specificknowledge. Knowledge exchanged with lead sup-pliers tended to be more implementation-based.More generally, this set of findings indicates thatcontent, in terms of type of knowledge exchanged,varies across embedded ties and these differenceshelp explain the relative influence of interfirm tieson capability acquisition.

Taken together, the findings from this studymake important contributions to research on thesources of capabilities. Whereas previous studieshave shown how a firm’s acquisition of capabil-ities is influenced by its position in a network,or structural embeddedness, this paper illustratesthe importance of relational embeddedness (Gra-novetter, 1992; Gulati, 1998; Rowley, Behrens,and Krackhardt, 2000), or the cohesiveness of tiesand their underlying attributes. Such an empha-sis on the attributes of embedded ties is impor-tant because it provides fine-grained insights into


Embedded Ties 1051

the causal mechanisms that explain how and whyembedded ties influence firms’ abilities to acquirecompetitive capabilities.

More generally, this study holds important impli-cations for strategy research on the sources ofsustained competitive advantage (Amit and Schoe-maker, 1993; Barney, 1986). Our findings chal-lenge the assertion that differences across firmsin competitive capabilities are solely a functionof internal learning and resource endowments(Chandler, 1992; Helfat, 2000; McGrath et al.,1995). Network resources in the form of embed-ded ties with exchange partners represent anotherimportant source of variation in competitiveness.Because embedded ties tend to be idiosyncratic andpenetrate irregularly and in different degrees (Gra-novetter, 1985), they represent an important sourceof firm heterogeneity. Moreover, to the extentthat embedded ties are based on relations of trustthat evolve over time through repeated interaction,they can also be considered a relatively immo-bile resource that is difficult to acquire quickly.For these reasons, embedded ties may representa network resource that is somewhat inimitableand, therefore, a sustainable source of competitiveadvantage. At the same time, the effects of embed-ded ties on firm competitiveness do not appearto be uniformly positive. Our research points tothe heterogeneity across embedded ties with dif-ferent types of exchange partners. While someembedded ties may be beneficial for sustainingcompetitiveness, others may be inconsequential orperhaps even detrimental. Therefore, the ability todiscriminate the relative value of different types ofembedded ties is also important.

Limitations and directions for future research

Certain features of this research influence how theresults should be interpreted. Due to the cross-sectional research design employed, we were cau-tious about drawing conclusions about the causaldirection of the relationships among key con-structs. For instance, it is possible that the acqui-sition of capabilities preceded and influenced theestablishment of embedded ties with exchangepartners. However, we investigated this possibil-ity by comparing the emergence of the capabilitieswe studied in the seven industries we focused on tothe duration of relationships with lead customersand suppliers. We observed that, on average, ties

to customers and suppliers preceded the emergenceof the capabilities by 8 years.9

We also considered an alternative explanationfor the effects of exchange ties with customerson capability acquisition. Specifically, firms par-ticularly dependent on customers for businessmight be coerced into adopting such practicesand techniques. Accordingly, acquisition of capa-bilities should increase with dependence (Pfef-fer and Salancik, 1978) on a lead customer. Infact, we checked for this possibility in our dataand found the opposite—that the acquisition ofcapabilities decreased as dependence on customersincreased.10 Capability acquisition was lower whena firm was highly dependent on its lead customerfor sales. Dependence on a lead customer alsocoincided with a higher level of embeddedness.11

Together, these results might indicate the down-side of ‘overembedded’ ties. The convergence ofdependence and higher levels of embeddednesssuggests exchange relationships where the partieshave exclusive arrangements to deal primarily witheach other (e.g., sole-sourcing). Under such cir-cumstances where a firm becomes a ‘captive’ sup-plier to its lead customer the drive to innovate and

9 More specifically, we compared the average duration of rela-tionships with lead customers and suppliers to when the com-petitive capabilities we studied arrived in the seven industriesidentified. Although it is difficult to pinpoint the exact datewhen the capabilities we studied first arrived, our field inter-views confirmed that it was not until after the passage and theimplementation of the 1990 Pollution Prevention Act that pol-lution prevention practices began to emerge. We found that 84percent of the time the relationship between a firm in our studyand its lead customer was in place prior to 1990 and that 80percent of the relationships with lead suppliers were establishedprior to 1990. On average, relationships with lead customers orsuppliers were established 8 years prior to the passage of theamendments to federal legislation in 1990. Our field interviewsfurther confirmed that the emergence of quality managementpractices coincided with the rise of pollution prevention practicesin the industries we studied. These data give some indicationthat firms’ relationships with exchange partners preceded theiracquisition of competitive capabilities.10 We explored this alternative prediction by performing addi-tional analyses using data capturing the percentage of a firm’stotal annual sales coming from its lead customer as a proxy fordependence. We regressed each of the capabilities on depen-dence, controlling for other significant predictors of the capabil-ity. The results show a negative and statistically significant rela-tionship between dependence and both capabilities, contrary towhat the theory would predict. The standardized beta coefficientsfor the dependence variable in the pollution prevention and qual-ity management capabilities equations are −0.122 (t = −2.077;p < 0.05) and −0.113 (t = −2.044; p < 0.05) respectively.11 Dependence is positively correlated with two of the threeattributes of embedded ties; information sharing (ρ = 0.124;p < 0.05) and joint problem solving (ρ = 0.170; p < 0.01).



acquire new capabilities may be dulled by the lackof competitive forces. Firms may be more inclinedto upgrade and renew their capabilities when facedwith the prospect of lead customers switching toalternative sources of supply.

The focus on small manufacturing firms mayalso be a limitation of our study and restrict thegeneralizability of the findings to this population.A replication with a different sample of firmswould be a useful follow-on. In addition, it wouldbe useful for future research to obtain performancedata, which was difficult to obtain in our studysince most of the firms were privately owned.

CONCLUSION

The overall picture emerging from this study indi-cates that embedded ties clearly matter for theacquisition of competitive capabilities. Firms donot acquire capabilities in isolation. Firms dis-cover, evaluate, and learn how to implement capa-bilities in the course of interacting with keyexchange partners. Chief among these exchangerelationships are embedded ties with lead suppli-ers. In our study, well-developed joint problem-solving arrangements with key suppliers are animportant driver of the acquisition of competitivecapabilities. Hence, a firm’s capacity to compete isa function of the types of exchange partners withwhom it cultivates embedded ties and the qualityof its interfirm exchange relationships.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the thought-ful comments and suggestions provided by PaulIngram, Matt Kraatz, Susan McEvily, and AksZaheer. Mark Haveman also provided us withvaluable insights based on his extensive experi-ence working with the industry that we studied.The research leading to the development of thispaper was funded in part by the Great Lakes Pro-tection Fund and the Ewing Marion KauffmanFoundation’s Center for Entrepreneurial Leader-ship. The authors take full responsibility for theanalysis and interpretations offered. They shouldnot be construed in any way as representing theviews of either Foundation.

REFERENCES

Ahuja G. 2000. Collaboration networks, structural holes,and innovation: a longitudinal study. AdministrativeScience Quarterly 45: 425–455.

Aldrich H, Rosen B, Woodward W. 1986. Social behav-ior and entrepreneurial networks: summary. In Fron-tiers of Entrepreneurship Research , Ronstadt R, Hor-naday JA, Peterson R, Vesper K (eds). Babson Col-lege: Wellesley, MA; 239–240.

Amit R, Schoemaker P. 1993. Strategic assets andorganizational rent. Strategic Management Journal14(1): 333–346.

Argote L. 1999. Organizational Learning: Creating,Retaining and Transferring Knowledge. Kluwer:Boston, MA.

Arrow K. 1974. The Limits of Organization . Norton: NewYork.

Barney J. 1986. Strategic factor markets: expectations,luck, and business strategy. Management Science32(10): 1231–1241.

Baum JAC, Ingram P. 1998. Survival-enhancing learningin the Manhattan hotel industry, 1898–1980.Management Science 44: 996–1016.

Bentler PM. 1990. Comparative fit indexes in structuralmodels. Psychological Bulletin 107: 238–246.

Bentler PM, Bonett DG. 1980. Significance tests andgoodness of fit in the analysis of covariance structures.Psychological Bulletin 88: 588–606.

Blau PM. 1960. Exchange and Power in Social Life.Wiley: New York.

Boon SD, Holmes JG. 1991. The dynamics of interper-sonal trust: resolving uncertainty in the face of risk. InUnderstanding Personal Relationships , Duck S, Perl-man D (eds). Sage: Beverly Hills, CA; 190–211.

Burt RS. 1992. Structured Holes: The Social Structureof Competition . Harvard University Press: Cambridge,MA.

Burt RS. 1997. A note on social capital and networkcontent. Social Networks 19: 355–373.

Carmine EG, Zeller RA. 1979. Reliability and ValidityAssessment . Sage: Newbury Park, CA.

Carson SJ, Madhok A, Varman R, John G. 2003. Infor-mation processing moderators of the effectiveness oftrust-based governance in interfirm R&D collabora-tion. Organization Science 14: 45–56.

Chandler A. 1992. Organizational capabilities and theeconomic history of the industrial enterprise. Journalof Economic Perspectives 6(3): 79–100.

Clark K, Fujimoto T. 1991. Product Development in theWorld Automobile Industry . Harvard Business SchoolPress: Boston, MA.

Cohen WM, Levinthal DA. 1990. Absorptive capacity:a new perspective on learning and innovation.Administrative Science Quarterly 35: 128–152.

Coleman JS. 1988. Social capital in the creation of humancapital. American Journal of Sociology 94: S95–S120.

Collins JC, Porras JL. 1994. Built to Last: SuccessfulHabits of Visionary Companies . Harper-Business:New York.

Cummings LL, Bromiley P. 1996. The organizationaltrust inventory (OTI): development and validation. In


Embedded Ties 1053

Trust in Organizations , Kramer RM, Tyler TR (eds).Sage: Thousand Oaks, CA; 302–330.

Cushnie G. 1994. Pollution Prevention and ControlTechnology for Plating Operations . National Centerfor Manufacturing Science: Ann Arbor, MI.

Darr E, Argote L, Epple D. 1995. The acquisition,transfer and depreciation of knowledge in serviceorganizations: productivity in franchises. ManagementScience 41: 1750–1762.

Das TK, Teng BS. 1998. Between trust and control:developing confidence in partner cooperation inalliances. Academy of Management Review 23:491–512.

Deming WE. 1986. Out of the Crisis . MIT Center forAdvanced Engineering Study: Cambridge, MA.

Dierickx I, Cool K. 1989. Asset stock accumulation andsustainability of competitive advantage. ManagementScience 35(12): 1504–1513.

Dilman DA. 1978. Mail and Telephone Surveys: The TotalDesign Method . Wiley: New York.

Dirks KT. 1999. The effects of interpersonal truston work group performance. Journal of AppliedPsychology 84: 445–455.

Dirks KT, Ferrin DL. 2001. The role of trust inorganizational settings. Organization Science 12:450–467.

Dore R. 1983. Goodwill and the spirit of mar-ket capitalism. British Journal of Sociology 34:459–482.

Dyer JH, Chu W. 2003. The role of trustworthi-ness in reducing transaction costs and improvingperformance: empirical evidence from the UnitedStates, Japan, and Korea. Organization Science 14:57–68.

Dyer JH, Nobeoka K. 2000. Creating and managinga high-performance knowledge-sharing network: theToyota case. Strategic Management Journal , SpecialIssue 21(3): 345–368.

Eccles RG, Nohria N, Berkley JD. 1992. Beyond theHype: Rediscovering the Essence of Management .Harvard Business School Press: Cambridge, MA.

Flynn BB, Sakakibara S, Schroeder RG. 1995. Relation-ship between JIT and TQM: practices and per-formance. Academy of Management Journal 38(5):1325–1360.

Granovetter M. 1985. Economic action and socialstructure: the problem of embeddedness. AmericanJournal of Sociology 91(3): 481–510.

Granovetter M. 1992. Problems of explanation ineconomic sociology. In Networks and Organizations ,Nohria N, Eccles RG (eds). Harvard Business SchoolPress: Cambridge, MA; 25–56.

Grant RM. 2002. Contemporary Strategy Analysis: Con-cepts, Techniques, Applications . Blackwell: Cam-bridge, MA.

Guillen M. 1995. Models of Management: Work, Author-ity and Organization in Comparative Perspective. Uni-versity of Chicago Press: Chicago, IL.

Gulati R. 1995. Does familiarity breed trust? Theimplications of repeated ties for contractual choicein alliances. Academy of Management Journal 38:85–112.

Gulati R. 1998. Alliances and networks. StrategicManagement Journal , Special Issue 19(4): 293–317.

Gulati R. 1999. Network location and learning: theinfluence of network resources and firm capabilitieson alliance formation. Strategic Management Journal20(5): 397–420.

Gulati R, Westphal JD. 1999. Cooperative or controlling?The effects of CEO–board relations and the contentof interlocks on the formation of joint ventures.Administrative Science Quarterly 44: 473–506.

Hackman JR, Wageman R. 1995. Total quality man-agement: empirical, conceptual, and practical issues.Administrative Science Quarterly 40: 309–342.

Hamel G. 1991. Competition for competence andinter-partner learning within international strategicalliances. Strategic Management Journal , SummerSpecial Issue 12: 83–103.

Hansen M. 1999. The search-transfer problem: the roleof weak ties in sharing knowledge across organiza-tion subunits. Administrative Science Quarterly 44:82–111.

Heide JB, Miner AS. 1992. The shadow of the future:effects of anticipated interaction and frequency ofcontact on buyer–seller cooperation. Academy ofManagement Journal 35(2): 265–291.

Helfat CE. 2000. Guest editor’s introduction to the specialissue: the evolution of firm capabilities. StrategicManagement Journal , Special Issue 21(10–11):955–959.

Helper S. 1991. How much has really changedbetween U.S. automakers and their suppliers? SloanManagement Review , Summer: 15–28.

Henderson R, Cockburn I. 1994. Measuring competence?Exploring firm effects in pharmaceutical research.Strategic Management Journal , Winter Special Issue15: 63–84.

Ingram P, Simons T. 1999. The exchange of experiencein a moral economy: embedded ties and vicariouslearning in kibbutz agriculture. Academy of Manage-ment Proceedings.

Jeffries FL, Reed R. 2000. Trust and adaptation inrelational contracting. Academy of ManagementReview 25: 873–882.

John G. 1984. An empirical investigation of someantecedents of opportunism in a marketing channel.Journal of Marketing Research 21: 278–289.

John G, Weitz BA. 1988. Forward integration into dis-tribution: an empirical test of transaction cost analy-sis. Journal of Law, Economics, and Organization 4:337–355.

Joreskog KG, Sorbom D. 1996. LISREL 8: User’sReference Guide (2nd edn). SPSS: Chicago, IL.

Juran JM. 1974. The Quality Control Handbook (3rdedn). McGraw-Hill: New York.

Kelley MR, Brooks H. 1991. External learning opportu-nities and the diffusion of process innovations to smallfirms. Technological Forecasting and Social Change39: 103–125.

Kogut B, Shan W, Walker G. 1992. The make-or-cooperate decision in the context of an industrynetwork. In Networks and Organizations: Structure,



Form and Action , Nohria N, Eccles R (eds). HarvardBusiness School Press: Boston, MA; 348–365.

Kogut B, Zander U. 1992. Knowledge of the firm,combinative capabilities, and the replication oftechnology. Organizational Science 3: 383–397.

Kraatz MS. 1998. Learning by association? Interorga-nizational networks and adaptation to environmen-tal change. Academy of Management Journal 41:621–643.

Levitt B, March JG. 1988. Organizational learning.Annual Review of Sociology 14: 319–340.

Lewicki RJ, Bunker BB. 1996. Developing and maintain-ing trust in work relationships. In Trust in Organiza-tions , Kramer RM, Tyler TR (eds). Sage: ThousandOaks, CA; 114–139.

Lieberman MB, Lau LJ, Williams MD. 1990. Firm-level productivity and management influence: acomparison of U.S. and Japanese automobileproducers. Management Science 36: 1193–1215.

Lippman SA, Rumelt RP. 1982. Uncertain imitability: ananalysis of interfirm differences in efficiency undercompetition. Bell Journal of Economics 13: 418–438.

Macaulay S. 1963. Non-contractual relations in business.American Sociological Review 28: 55–67.

MacCallum RC, Browne MW, Sugawara HM. 1996.Power analysis and determination of sample size forcovariance structure modeling. Psychological Methods1: 130–149.

Macneil IR. 1980. The New Social Contract . YaleUniversity Press: New Haven, CT.

March JG, Simon HA. 1958. Organizations . Wiley: NewYork.

McEvily B, Zaheer A. 1999. Bridging ties: a sourceof firm heterogeneity in competitive capabilities.Strategic Management Journal 20(12): 1133–1156.

McEvily B, Perrone V, Zaheer A. 2003. Trust asan organizing principle. Organization Science 14:91–103.

McGrath R, MacMillan I, Venkatraman S. 1995. Defin-ing and developing capability: a strategic paradigm.Strategic Management Journal 16(4): 251–275.

Meyer AD, Goes JB. 1988. Organizational assimilationof innovations: a multilevel contextual analysis.Academy of Management Journal 4: 897–923.

Nelson R, Winter S. 1982. An Evolutionary Theoryof Economic Change. Harvard University Press:Cambridge, MA.

Nunnally J. 1978. Psychometric Theory . McGraw-Hill:New York.

Ouchi WG. 1979. A conceptual framework for the designof organizational control mechanisms. ManagementScience 25: 833–848.

Penrose ET. 1959. The Theory of the Growth of the Firm .M. E. Sharpe: White Plains, NY.

Pfeffer J, Salancik GR. 1978. The External Control ofOrganizations: A Resource Dependence Perspective.Harper & Row: New York.

Podolny JM, Baron JN. 1997. Resources and relation-ships: social networks and mobility in the workplace.American Sociological Review 62: 673–693.

Podsakoff PM, Organ DW. 1986. Self-reports in organi-zational research: problems and prospects. Journal ofManagement 12: 531–544.

Polanyi M. 1966. The Tacit Dimension . Anchor DayBooks: New York.

Porter M. 1991. Towards a dynamic theory of strategy.Strategic Management Journal , Winter Special Issue12: 95–117.

Powell WW, Koput KW, Smith-Doerr L. 1996. Interor-ganizational collaboration and the locus of innovation:networks of learning in biotechnology. AdministrativeScience Quarterly 41: 116–145.

Ring PS, Van de Ven AH. 1994. Developmental pro-cesses of cooperative interorganizational relationships.Academy of Management Review 19: 90–118.

Rowley T, Behrens D, Krackhardt D. 2000. Redundantgovernance structures: an analysis of structural andrelational embeddedness in the steel and semiconduc-tor industries. Strategic Management Journal , SpecialIssue 21: 369–386.

Rumelt RP. 1984. Toward a strategic theory of the firm.In Competitive Strategic Management , Lamb RB (ed).Prentice-Hall: Englewood Cliffs, NJ; 556–570.

Siegel S. 1956. Nonparametric Statistics for the Behav-ioral Sciences . McGraw-Hill: Tokyo.

Simon HA. 1982. Models of Bounded Rationality:Behavioral Economics and Business Organization ,Vol. 2. MIT Press: Cambridge, MA.

Sitkin S. 1992. Learning through failures: the strategy ofsmall losses. In Research in Organizational Behavior ,Staw BM, Cummings LL (eds). Vol. 14. JAI Press:Greenwich, CT; 231–266.

Stuart TE. 1998. Network positions and propensities tocollaborate: an investigation of strategic alliance for-mation in a high-technology industry. AdministrativeScience Quarterly 43: 668–698.

Szulanski G. 1996. Exploring internal stickiness: imped-iments to the transfer of best practice within the firm.Strategic Management Journal , Winter special issue17: 27–43.

Teece D. 1977. Technology transfer of multinationalfirms: the resource cost of transferring technologicalknow-how. Economic Journal 87: 242–261.

U.S. Environmental Protection Agency. 1990. Guide toPollution Prevention: The Fabricated Metal ProductsIndustry , Office of Research and Development. U.S.Government Printing Office: Washington, DC.

Uzzi B. 1997. Social structure and competition ininterfirm networks: the paradox of embeddedness.Administrative Science Quarterly 42: 35–67.

Von Hippel E. 1988. The Sources of Innovation . OxfordUniversity Press: New York.

Yeung A, Ulrich D. 1994. How organizations learn:an exploratory study on the antecedents andconsequences of organizational learning types. Paperpresented at the annual meeting of the Academy ofManagement, Dallas.

Zaheer A, Bell GG. 2005. Benefiting from networkposition: firm capabilities, structural holes andperformance. Strategic Management Journal 26(9):809–825.


Embedded Ties 1055

Zaheer A, Venkatraman N. 1995. Relational governanceas an interorganizational strategy: an empirical testof the role trust in economic exchange. StrategicManagement Journal 19(5): 373–392.

Zaheer A, McEvily B, Perrone V. 1998. Does trustmatter? Exploring the effects of interorganizational

and interpersonal trust on performance. OrganizationScience 9(2): 141–159.

Zander U, Kogut B. 1995. Knowledge and the speed oftransfer and imitation of organizational capabilities.Organization Science 6(1): 76–92.


Documents

EMBEDDED TIES AND THE ACQUISITION OF COMPETITIVE CAPABILITIES · Competitive capabilities Competitive capabilities are the set of organizing processes and principles a ﬁrm uses