WHEN DOES TRUST MATTER TO ALLIANCE PERFORMANCE?

WHEN DOES TRUST MATTER TO ALLIANCE PERFORMANCE?

REKHA KRISHNANSimon Fraser University

XAVIER MARTINNIELS G. NOORDERHAVEN

Tilburg University

We examine how uncertainty moderates the trust-performance relationship in alli-ances, building on the distinction between behavioral uncertainty, which relates toanticipating and understanding partners’ actions, and externally caused environmen-tal uncertainty. We argue that trust matters more to performance under behavioraluncertainty and less under environmental uncertainty. In data from 126 internationalalliances, the positive relationship between trust and performance is stronger underhigh behavioral uncertainty and weaker under high environmental uncertainty. Weconclude that partners should concentrate on developing interorganizational trustwhere potential improvement in alliance performance justifies this effort, which inturn depends on the type of uncertainty faced.

Strategic alliances blur firm boundaries and cre-ate mutual dependence between previously inde-pendent firms (McEvily, Perrone, & Zaheer, 2003).A distinctive characteristic of strategic alliances isthat partners have to deal not only with the uncer-tainty in their environment but also with the un-certainty arising from each other’s behavior (Harri-gan, 1985). Because of partners’ dependence oneach other, previous research has emphasized theimportance of relational factors for the smoothfunctioning of strategic alliances (Powell, 1990).Although various relational mechanisms andnorms have been studied, including for instancenorms of solidarity and flexibility (Poppo & Zenger,2002: 712), none has received more attention thantrust (Gambetta, 1988; Mayer, Davis, & Schoorman,1995; McEvily et al., 2003; Sako, 1991; Zaheer,McEvily, & Perrone, 1998; Zand, 1972). Accord-ingly, a great deal of research in this tradition hasidentified interorganizational trust as a key factorcontributing to alliance success, the general viewbeing that trust has a positive effect on allianceperformance (e.g., Dyer & Chu, 2003; Mohr & Spek-man, 1994; Zaheer et al., 1998).

The existence of trust between alliance partnerscannot be taken for granted, however; partners maynot only have to cultivate trust intentionally (Parkhe,1998; Sako, 1991), but may also incur substantial realand opportunity costs in its pursuit (McEvily et al.,

2003; Poppo & Zenger, 2002). Furthermore, interor-ganizational trust need not always improve allianceperformance (McEvily et al., 2003). Indeed, research-ers are beginning to recognize that the relationshipbetween trust and alliance performance may be com-plicated and contingent on other factors. Thus, Car-son, Madhok, Varman, and John (2003) argued thatthe effect of trust on task performance in vertical R&Dcollaborations strengthens with the clients’ ability tounderstand the tasks involved. Langfred (2004) ar-gued that the effect of trust on the performance ofself-managing teams reverses when individual auton-omy is high. These studies suggest that the benefitsderived from trust may magnify under certain condi-tions and diminish under other conditions. However,contingent reasoning has yet to be applied to theeffect of uncertainty, two forms of which—behavioraland environmental—are potentially the most funda-mental strategic factors in alliances (Harrigan, 1988;Kogut, 1989; Williamson, 1985, 1991).

There is extensive support in prior research forthe overall beneficial effect of trust. Empirical stud-ies have shown that trust, by bringing about goodfaith in the intent, reliability, and fairness of part-ner behavior (Sako, 1991; Zaheer et al., 1998), al-lows for constructive interpretation of partner mo-tives (Uzzi, 1997), reduces the potential for conflict(Zaheer et al., 1998), and encourages smooth infor-mation flow between partners (Sako, 1991; Zand,1972). Trust thus mitigates uncertainty about part-ner behavior. Yet the same qualities of trust thatmitigate uncertainty about partner behavior andengender its beneficial effects may also limit thecognitive efforts of partners when they consider

The paper benefited greatly from the insights of ourassociate editor and anonymous reviewers. Any remain-ing errors are ours. This research was supported by theCentER for Economic Research at Tilburg University.

! Academy of Management Journal2006, Vol. 49, No. 5, 894–917.

894

their broader environment. Trust researchers haveargued theoretically (Webb, 1996: 292) and foundempirically (Langfred, 2004) that the perception ofreliability of information from a partner and thecognitive comfort that trust brings about also limitvariety of thought and action and attentiveness todetail. Therefore, trust may reduce the alertnessneeded when alliance partners have to respond toenvironmental uncertainty. The result may be thatpartners respond to external challenges inade-quately or not at all. Thus, as we discuss in moredetail below, trust seems to entail a trade-off be-tween the capacity to deal with behavioral uncer-tainty and environmental uncertainty.

In this article, we examine this trade-off by theoriz-ing that trust has different effects on alliance perfor-mance, depending on the levels of behavioral andenvironmental uncertainty present. We define a stra-tegic alliance as any extended cooperative agreementintended to jointly develop, manufacture, and/or dis-tribute products (Gulati, 1998; Zollo, Reuer, & Singh,2002: 701). Although trust helps alliance partners tocope with uncertainties pertaining to each other’sbehavior, trust also tends to constrain partners’ re-sponses to environmental demands, thus hinderingthem from responding appropriately to environmen-tal uncertainty. Below, we develop these argumentsin greater detail and report tests of the resulting hy-potheses in a sample of 126 international strategicalliances in India.

TRUST AND ALLIANCE PERFORMANCE

The concept of trust has received ample attentionfrom various disciplines, and although prior researchhas put forth diverse interpretations of trust, a com-mon core emerges. Building on this prior research, wedefine interorganizational trust as the expectationheld by one firm that another will not exploit itsvulnerabilities when faced with the opportunity to doso (Barney & Hansen, 1994; Mayer et al., 1995; Sako,1991). This expectation is confirmed when parties (1)demonstrate reliability by carrying out their prom-ises, (2) act fairly when dealing with each other, and(3) exhibit goodwill when unforeseen contingenciesarise. Our definition thus bases interorganizationaltrust on three related components: reliability, fair-ness, and goodwill (Dyer & Chu, 2003).

The goodwill component in the definition of trustextends beyond contractual obligations in that part-ners commit themselves and make contributions totheir relationship that go beyond what was explicitlyguaranteed (Sako, 1991: 453). Hence, trust stands tobe relevant in situations where firms make substan-tial and open commitments to a partnership. These

situations most prominently include alliances (Gu-lati, 1995; Mohr & Spekman, 1994).

One important concern with alliances is that con-flict between partners can occasion high costs or apremature breakdown of relationships (Zaheer etal., 1998). Trust helps defuse such conflict, becausetrusting partners are more likely to interpret eachother’s equivocal actions in a manner conducive tothe stability of the relationship. As Noorderhaven(2004) argued on the basis of case studies by Doz(1996), if a firm encounters unexpected actions byits partner that could be ascribed to both good andbad intentions, the presence of trust reduces thelikelihood of a negative interpretation. For in-stance, when confronted with disappointing salesof a product line, a partner might either explain theinadequate performance on the basis of an ineffec-tive promotional campaign, or view the failure assignaling a lack of commitment of the other party’sdistributors. In such equivocal situations, trust fa-cilitates mutual understanding and allows for thebenefit of the doubt. It thus reduces the costs ofinterpartner conflict as well as other transactioncosts (Dyer & Chu, 2003; Zaheer et al., 1998). Re-search also shows that such costs are negativelyrelated to alliance performance (Zaheer et al.,1998). Therefore, all else being equal, trust im-proves alliance performance. Hence:

Hypothesis 1. Ceteris paribus, trust is posi-tively related to alliance performance.

The above hypothesis does not mean that trustimproves the performance of all alliances equally.Next, we argue that the type of uncertainty thatalliance partners encounter moderates the relation-ship between trust and alliance performance. Inthis respect, we distinguish between two types ofuncertainty that are relevant to interfirm relations(Sutcliffe & Zaheer, 1998; Williamson, 1985): envi-ronmental uncertainty, which results from exoge-nous sources outside the scope of the alliance, andbehavioral uncertainty, which results from diffi-culty in anticipating and understanding the actionsof an exchange partner. Below, we develop predic-tions that the trust-performance relationship islikely to be stronger under behavioral uncertaintybut weaker under environmental uncertainty.

Trust, Behavioral Uncertainty, andAlliance Performance

Concern about partner behavior is a predominantsource of internal tension in strategic alliances(Park & Ungson, 2001; Parkhe, 1993; Sutcliffe &Zaheer, 1998). Such concerns stem in large mea-sure from behavioral uncertainty—that is, the po-

2006 895Krishnan, Martin, and Noorderhaven

tential inherent in a situation for difficulty antici-pating and understanding actions. Although thisfeature of the alliance context is never zero, itsmagnitude varies among alliances. Specifically, po-tential inability to anticipate and understand ac-tions is higher in alliances of the following twotypes (Das & Teng, 2000; Park & Ungson, 2001): (1)alliances involving high interdependence, inwhich the partners’ contributions are highly inter-twined (Nooteboom, 2002; Park & Russo, 1996;Stinchcombe, 1985) and (2) alliances in which eachpartner is able to further private interests at theexpense of collaborative interests (Khanna, Gulati,& Nohria, 1998; Park & Ungson, 2001), especiallyalliances between potential competitors (Bleeke &Ernst, 1993; Hamel, 1991; Kogut, 1988; Oxley &Sampson, 2004; Park & Russo, 1996).

Interdependence. The degree of interdependencein an alliance increases with the importance and ex-tent of the resources shared between partners andwith the resulting overlap in division of labor be-tween them (i.e., the resulting shared responsibilityfor a number of tasks) (Gulati & Singh, 1998; Kumar &Seth, 1998; Thompson, 1967). Alliances that are setup to share production facilities typically create onlyweak interdependencies (Gulati & Singh, 1998). Re-source allocations and role assignments in these part-nerships tend to be straightforward and stable, andthe division of labor is thus likely to be simple. Incontrast, alliances formed to jointly develop newtechnology or to speed up innovation (an alliance fordesigning a leading-edge microprocessor, for exam-ple) lead to high interdependence (Nickerson &Zenger, 2004: 620; Park & Russo, 1996). These alli-ances are characterized by substantial overlap be-tween the partners’ responsibilities and involve on-going mutual adjustment between partners (Gulati &Singh, 1998).

In highly interdependent activities, coordinationis difficult because the complexity associated withinterdependence discourages coordination by stan-dardization. Instead, the overlapping division oflabor calls for coordination by mutual adjustment,which precludes the use of standard rules tostreamline decision making and regularize interac-tions between interdependent alliance partners(Thompson, 1967). The variability associated withhighly interdependent alliances renders coordina-tion by mutual adjustment highly demanding anddifficult in terms of communication and decisionefforts (Gulati & Singh, 1998; Thompson, 1967).As a consequence, the higher the interdepen-dence, the more likely that any change one part-ner makes will affect the other partner in un-planned ways, and the more immediate andsevere the adverse impact of any mistake (inten-

tional or not) by a partner (Nooteboom, 2002;Thompson, 1967).

In addition, high interdependence in alliances re-quires partners to share valuable knowledge-inten-sive resources, exposing these to each other (Kumar &Seth, 1998; Nooteboom, 2002; Park & Russo, 1996;Park & Ungson, 2001). Being harder to observe, value,and protect, shared knowledge-intensive resourcesincrease the potential for misunderstandings con-cerning each partner’s intents and contributions tothe alliance (Oxley, 1999). The difficulty in discern-ing the closely intertwined partner contributions fur-ther threatens the open sharing of resources and in-formation among partners and consequentlymagnifies the coordination difficulties in high-inter-dependence alliances (Park & Russo, 1996).

Interorganizational trust stands to be especiallybeneficial in the presence of such behavioral uncer-tainty. By asserting good faith in the intent and reli-ability of partner behavior, trust allows partners toengage in constructive interpretation of each other’sactions (Zaheer et al., 1998). It also encourages part-ners to be aware of the processes and procedures thateach partner follows (Gulati & Singh, 1998). Thus,trust encourages partners to remain flexible whenmanaging their interface in the face of interdepen-dence. It also alleviates apprehensions regarding thesharing of valuable information; the resulting infor-mation exchange and socialization assist in maintain-ing effective integration and coordination. Underhigh interdependence, interorganizational trust istherefore essential for alliance performance, as it fa-cilitates mutual adjustment and allows the smoothersynchronization of critical tasks.

Conversely, we expect trust to have a weakereffect on alliance performance under conditions oflow interdependence, as the overlap in the divisionof labor between partners is lower and hence thescope for misinterpretations and tensions is likelyto be lower as well (Gulati & Singh, 1998). Hence,

Hypothesis 2. The positive relation between trustand alliance performance is stronger in allianceswith a high degree of interdependence than inalliances with low interdependence.

Interpartner competition. Interpartner competi-tion exists when a partner tries to maximize itsprivate interests at the expense of the alliance orthe other partner (Baum, Calabrese, & Silverman,2000; Park & Russo, 1996; Park & Ungson, 2001). Inalliances formed between potential competitors,concerns about opportunistic exploitation loom es-pecially large, because the partners may havestrong incentives to appropriate each other’s re-sources (Khanna et al., 1998; Oxley & Sampson,2004). Prior research has shown that alliances be-

896 OctoberAcademy of Management Journal

tween potential competitors engender greater ten-dencies of partners to engage in such “de factointernalization” (Baum et al., 2000; Hamel, 1991:84). Moreover, because potential competitors arefamiliar with the areas that their partners operatein, the potential competitors have superior capacityto absorb and reuse proprietary knowledge (Cohen& Levinthal, 1990; Park & Russo, 1996).

In situations that lend themselves to interpartnercompetition, the potential to misunderstand the mo-tives of a partner is significant, and this exacerbates afirm’s tendencies to protect its own resources, espe-cially knowledge, at the risk of hampering the alli-ance relationship (Hamel, 1991; Kale, Singh, &Perlmutter, 2000). By detracting partners from con-tributing fully to the performance of the alliance,these concerns interfere with the realization of thesynergistic benefits of the alliance (Grindley, Mow-ery, & Silverman, 1994; Madhok & Tallman, 1998).

Trust can counteract such problems by increasingeach partner’s confidence that the other will notabuse its vulnerabilities (Barney & Hansen, 1994; Bra-dach & Eccles, 1989; Mayer et al., 1995). Faith in theintentions and fairness of the other makes each part-ner more likely to respect the boundaries of the oth-er’s resources and proprietary knowledge. This mu-tual respect encourages partners to provide thesubstantive resources and accurate and timely infor-mation that enhance collaborative benefits (Sako,1991; Zand, 1972). Trust is all the more advantageouswhen the potential for interpartner competition ishigh, because it facilitates mutual understanding andcounteracts the attendant failure to cooperate.

In alliances with low potential for interpartnercompetition, conversely, the appropriation of re-sources is less likely to be of strategic concern.Hence, partners’ suspicions regarding each other’sintents within and outside the alliance are lesscrippling. Thus, the benefits of interorganizationaltrust are lower.

Hypothesis 3. The positive relation betweentrust and alliance performance is stronger inalliances in which the potential for interpart-ner competition is high than in alliances inwhich the potential for interpartner competi-tion is low.

Trust, Environmental Uncertainty, andAlliance Performance

Environmental uncertainty results from changesin the economic conditions faced by an organiza-tion that are outside its control and hard to antici-pate (Dess & Beard, 1984; Koopmans, 1957), such asinstability or unpredictability in markets (Aldrich,

1979; Dess & Beard, 1984; Wholey & Brittain, 1989).Environmental uncertainty demands speedy andresponsive decisions (Huber, Miller, & Glick, 1990:13; Mintzberg, 1978). This pressure requires organ-izations to engage in significant scanning of theirenvironment in search of accurate and reliable infor-mation that enables them to interpret and act uponthe threats and opportunities facing them (Aguilar,1967; Anderson & Paine, 1975; Hambrick, 1982).

Unpredictable changes in the environment stand toaffect the performance of an alliance (Harrigan, 1985;Kogut, 1989). To sustain performance in an uncertainenvironment, alliance partners need to monitorchanges and adjust the alliance’s strategy accordingly(Harrigan, 1985). Anderson and Paine argued that inadjusting strategy, “The critical area is not uncer-tainty per se but the processing of accurate informa-tion to deal with uncertainty” (1975: 814). This infor-mation processing may be a bottleneck because ofproblems of information overload (Mintzberg, 1978;Robertson, 1980), which are exacerbated by informa-tion unfamiliarity (Park & Sheath, 1975). In highlyuncertain environments, cognitive limitations mayintroduce considerable limitations and biases in thedecision-making process, by prompting the applica-tion of inappropriate rules-of-thumb (Barnes, 1984;Cyert & March, 1963; Schwenk, 1984).

The risk that biases will enter decision making inuncertain environments is greater in the presence oftrust, especially in alliances. Scholars are beginningto recognize the heuristic quality of trust (McEvily etal., 2003; Nooteboom, 2002; Uzzi, 1997). Like cogni-tive heuristics (Bazerman, 1998), trust enables deci-sion making under conditions of uncertainty, butmay also produce systematic biases that can result insignificant errors (Ferrin & Dirks, 2003). Specifically,when partners trust each other, their tendency toscreen the information provided by the other for ac-curacy decreases, and their inclination to accept theinformation at face value increases (McEvily et al.,2003; Szulanski, Cappetta, & Jensen, 2004; Uzzi,1997). Because trust accustoms partners to rely oneach other without doubt, each partner tends to relymore extensively on the other’s knowledge of theenvironment when scanning the opportunities andthreats faced by the alliance, while paying less atten-tion to the completeness and veracity of the informa-tion thus obtained. Trust encourages partners to min-imize redundancies in the search process byexploiting each other’s purported expertise to engagein specialized search. For instance, in internationalalliances it is common for the local partner to scan theenvironment for regulatory changes or changes inconsumer preferences in the local market, while theforeign partner monitors technological changes,global demand, or new competition from foreign


firms (Beamish & Banks, 1987). Such tendencies tendto be more pronounced in the presence of trust.

Prior research has also shown that confrontingalternative views and diverse information stimu-lates creativity and constructive criticism (Jehn,Northcraft, & Neale, 1999; Simons, Pelled, & Smith,1999). Specialized search as described above, onthe contrary, reduces variety in information andrestricts the cross-fertilization of viewpoints re-quired for crafting well-informed responses to theenvironment (Webb, 1996). In alliances, becausetrust functions as a simplifying heuristic that con-strains cognitive effort (McEvily et al., 2003), it alsostands to bias partners’ efforts to scan and makesense of their environment, and therefore it mayresult in suboptimal responses.

Trusting alliance partners may even experience“strategic blindness”—outright insensitivity to envi-ronmental changes (McEvily et al., 2003: 97). Becausepartners commit resources, effort, and time in theprocess of cultivating trust, they tend to be wary ofactions that may damage the relationship (Noot-eboom, 2002). For instance, if responding to environ-mental change would require major changes, such asbringing in a new partner or ending an alliance (Har-rigan, 1985), partners may be apprehensive about theeventuality of having to cultivate trust and adjust to anew partner all over again, or having to go it alone. Inthe presence of such apprehensions, partners tend toweigh losses resulting from responding to the envi-ronment more than the gains that might come about(Bazerman, 1998; Nooteboom, 2002). Partners mayprefer “inaction over action and status quo over anyalternatives” (Kahneman & Lovallo, 1993: 18), culmi-nating in their alliance failing to respond to demandsof its environment.

Overall, interorganizational trust stands to result ininadequate response to the challenges posed by anuncertain environment because it limits cognitive ef-forts or even causes strategic blindness. Notwith-standing good intentions, trust may thus lead part-ners into making slow and suboptimal decisions fortheir alliance, or even no decision at all. This placesthe alliance at variance with environmental de-mands. We expect that as a result the effect of trust onalliance performance will not be as positive in thepresence of high environmental uncertainty.

Under low environmental uncertainty, on thecontrary, complete and accurate environmentalscanning is less critical. Furthermore, less effort isrequired to adjust to the environment because of itsstability and predictability. Therefore, the limitingeffects of trust discussed above tend to be lessrelevant.

Scholars have consistently argued that hard-to-pre-dict changes in market environment create critical

uncertainty for organizations (Cameron, Kim, &Whetten, 1987; Delacroix & Swaminathan, 1991; Dess& Beard, 1984: 56). Many researchers have used theconcept of instability, which Child (1972) defined asthe degree of difference involved in an environmentalchange, to capture environmental uncertainty (e.g.,Bergh & Lawless, 1998; Keats & Hitt, 1988; Snyder &Glueck, 1982). Others have argued that instability isbut one dimension of uncertainty. Unpredictability—that is, the degree of irregularity in an overall patternof environmental change (Child, 1972)—also magni-fies the consequences of a changing environment(e.g., Lawrence & Lorsch, 1973; Wholey & Brittain,1989). Because environmental uncertainty is a func-tion of both instability and unpredictability (Buchko,1994), these two distinct dimensions of market vari-ation are both relevant to our investigation of the jointconsequences of trust and environmental uncer-tainty. That is, we expect instability and unpredict-ability to each reduce the trust-performancerelationship.

Hypothesis 4. The positive effect of interorgan-izational trust on alliance performance isweaker when market instability is high thanwhen it is low.

Hypothesis 5. The positive effect of interorgan-izational trust on alliance performance isweaker when market unpredictability is highthan when it is low.

METHODS

Data

Data were collected through a survey of interna-tional strategic alliances operating in India. Asstated above, we identified strategic alliances asextended cooperative agreements intended tojointly develop, manufacture, and/or distributeproducts (Gulati, 1998; Zollo et al., 2002). In thelast 15 years, India has become one of the mostattractive investment locations in the world (A.T.Kearney, 2004). Strategic alliances have been aprevalent mode of entry into the Indian market, asin the rest of the world, and the number of strategicalliances in India has also risen by more than 50percent over the past decade (Bhaumik, Beena,Bhandari, & Gokarn, 2003). Furthermore, differentindustries in India exhibit sharply different pat-terns in the variation of demand over time. Thesedifferences made strategic alliances in India a rel-evant empirical setting in which to study the role ofenvironmental uncertainty. Thus, India provided arich and suitable context in which to study theconditions for successful strategic alliances.


To obtain a target population of internationalstrategic alliances, we examined Capitaline, a sec-ondary database, and member lists of various inter-national chambers of commerce in India. We thusidentified a sample of 700 dyadic internationalstrategic alliances operating in India. FollowingParkhe (1993) and Simonin (1999), we designed theresearch to aim at respondents highly knowledge-able about their firms’ alliances. The sensitivity ofthe questions, and the fact that their top executivesdeal directly with most international alliances inIndia, demanded that managing directors or chiefexecutive officers fill in the questionnaire. Thesetarget respondents’ names were identified usingCapitaline and chamber of commerce data.

Data Collection

The questionnaires were designed and the surveyimplemented according to Dillman’s (2000) tai-lored design method, which suggests several waysto encourage response. The measurement itemswere generated through a review of prior allianceliterature. We used university faculty and doctoralstudents to assess whether the content of the itemstapped the conceptual domain of the focal con-struct (DeVellis, 1991). This assessment yielded aset of fine-tuned questionnaire items that we usedin personal interviews and early pretests with man-aging directors of Indian firms involved in interna-tional strategic alliances to verify the clarity of theitems. We modified the wording of a few itemsslightly as a result and enriched one measure, asdescribed below. This process further strengthenedcontent validity. Appendix A reports the surveyitems used in this study.

The first wave of questionnaires was sent to man-aging directors and senior executives of 700 Indianfirms with international alliances. This wave ofsurvey mailings was followed, four weeks later, bya second wave. Of the 700 managing directors andsenior executives who received questionnaires, 126responded, yielding an 18 percent response rate.This rate is comparable to those of recent surveys ofalliance managers in other emerging economies,such as 14.4 percent for China (Isobe, Makino, &Montgomery, 2000) and 19 percent for Mexico(Robins, Tallman, & Lindquist, 2002). All the re-sponses to our survey came from individuals di-rectly responsible for the alliances: 80 came fromchairpersons and managing directors of the alli-ances, 30 from presidents, vice presidents, and gen-eral managers, and 16 from full-time directors.Nearly 75 percent of the respondents had been withtheir firms for more than 5 years, and of thesealmost 25 percent had more than 20 years’ tenure.

The alliance partners of the Indian firms werespread over 21 countries. All alliances in our sam-ple were dyadic, and all belonged to industries inthe manufacturing sector, where alliances weremore prevalent than in other sectors (e.g., Parkhe,1993; Simonin, 1999). Tests of proportions showedthat the distribution of our responses according totheir two-digit manufacturing SIC code was notstatistically different from those reported in the twolandmark studies by Harrigan (1988) and Ghema-wat, Porter, and Rawlison (1986) (chi-squares were8.11 for 16 degrees of freedom and 10.04 for 17degrees of freedom, respectively). The top four in-dustries in our sample ranked in the same order asin these studies and, as others have also found (e.g.,Parkhe, 1993), were relatively high-technology in-dustries (industrial machinery and equipment,chemicals and allied products, electrical and elec-tronic equipment, and transportation equipment).

We checked the potential for nonresponse biasby comparing the characteristics of the respondentsto those of the targeted population sample. Theresults of t-tests for the sizes of the firms (p ! 0.28)and the age of the local firm (p ! 0.34) revealed nosignificant differences between respondent andnonrespondent groups. In line with Mohr and Spek-man (1994) and Poppo and Zenger (2002), we alsotested for nonresponse bias by comparing early andlate respondents. Armstrong and Overton (1977)argued that late respondents are representative ofnonrespondents. We found no significant differ-ence between early and late respondents on char-acteristics such as number of employees of the In-dian partner (p ! 0.18), alliance duration (p !0.29), and investment size (p ! 0.51).

Dependent Variable: Alliance Performance

Despite the publication of numerous studies onalliance performance (e.g., Aulakh, Kotabe, & Sa-hay, 1996; Lane, Salk, & Lyles, 2001; Mohr & Spek-man, 1994; Parkhe, 1993), no consensus exists onmeasuring this construct. The hybrid structuresand transitory nature of alliances (Buckley &Glaister, 2002; Olk, 2002) present unique chal-lenges for evaluating performance and hinder theuse of two common indicators of firm performance,financial profitability and survival. Most alliancesdo not report financial performance, which wouldin any case tend to be biased by partners’ account-ing preferences. Survival is an imperfect indicatorof success because an alliance can be successfuland discontinued—for instance, because it hasserved its purpose—or unsuccessful and not (yet)discontinued, for instance because the partners stillhope to improve the relationship (Yan & Zeng,


1999). To circumvent such hurdles, much of thealliance performance research has relied on man-agers’ evaluations of alliance success (e.g., Aulakhet al., 1996; Isobe et al., 2000; Lin & Germain, 1998;Saxton, 1997). Doing so is appropriate when re-spondents represent top management (see Olk,2002). Because the respondents in our sample werevery well informed about the alliances in whichthey were involved, we were confident that it wasproper to rely on managerial evaluations of alliancesuccess. Moreover, Geringer and Hebert (1991)found strong correlations between subjective andobjective measures of alliance performance.

We measured alliance performance using a five-item Likert scale reflecting (1) the extent to whichthe local partner is satisfied with the overall per-formance of its alliance, (2) the extent to which thelocal partner perceives the foreign partner to besatisfied with the overall performance of the alli-ance, (3) the partners’ satisfaction with respect tothe attainment of goals, (4) the extent to which thelocal partner is satisfied with the financial perfor-mance of the alliance, and (5) the extent to whichthe local partner perceives its foreign partner to besatisfied with the financial performance of the alli-ance. With a Cronbach’s alpha of .90, the perfor-mance scale demonstrated high reliability (DeVel-lis, 1991; Nunally, 1978).

Independent Variables

Trust. Although no standard scale of interorgan-izational trust exists, prior studies have typicallycovered two or more of its main dimensions: reli-ability, fairness, and goodwill (e.g., Aulakh et al.,1996; Dyer & Chu, 2003; Zaheer et al., 1998). Inkeeping with this approach, we measured trust us-ing a five-item Likert-type scale adapted fromAulakh et al. (1996) and Sako and Helper (1998)that captures fairness, reliability, and goodwillamong alliance partners.1 The measure has highstatistical reliability based on Cronbach’s alpha(.85). We also validated our trust scale using fieldinterviews based on a three-point categorization ofthe interview data, which yielded a high conver-

gent validity score of .70 (p ! .05).2 Moreover, thecorrelation between trust and performance in ourdata, .52, fitted within the range of .34 and .65reported in prior studies (e.g., Aulakh et al., 1996;Carson et al., 2004; Geyskens, Steenkamp, & Ku-mar, 1999; Zaheer et al., 1998). Altogether, thesestatistics provided evidence that our measure oftrust was reasonable and consistent.

Interdependence. Strategic rationales for form-ing an alliance, eight of which Gulati and Singh(1998) identified through an extensive review ofthe literature, captured the range of value creationmotives of the partners. On the basis of our pretestwith Indian alliance managers, we added one itemto their list: access to technology. As did Gulati andSingh (1998), we assigned the nine strategic ration-ales to one of the three classes of interdependenciesidentified by Thompson (1967): pooled, sequential,and reciprocal interdependence. We classified aspooled interdependence three strategic rationalesthat require limited coordination: sharing costs(e.g., joint materials procurement), sharing produc-tion facilities, and sharing financial resources.Three other strategic rationales were classified un-der sequential interdependence: access to financialresources, access to new markets, and access totechnology; these require intermediate levels of co-ordination. Reciprocal interdependence includedthe remaining three strategic rationales, which re-quire extensive coordination: sharing complemen-tary technology, joint development of technology,and reduction of time needed for innovation (foradditional details, see Gulati and Singh [1998:796]). Interdependence was the average compositescore of the strategic rationales present in a alliance(respondents could identify more than one strategicrationale), weighted by the type of interdependenceeach rationale represented. We assigned ordinalweights of 3, 2, and 1 for reciprocal, sequential, andpooled interdependence, respectively. Degree of in-

1 Our first, second, and third items are specific to thefairness, reliability, and goodwill aspects of trust, respec-tively. The remaining two items covered all three as-pects, indicating the overall level of trust, as opposed todistrust (Sako & Helper, 1998), in the information sharedwithin a relationship. Zaheer et al. (1998) theorized thatpredictability was a dimension of trust, alongside reli-ability and fairness, but the predictability items droppedout of their interorganizational trust scale upon factoranalysis.

2 For further detail, see the entry, “Triangulation usingfield interviews,” in Appendix C. For instance, state-ments such as the following typically received a score of3: “Our partner acquired a new brand from ‘company A’and the distribution was initially with ‘company B.’”“Our partner could have easily lifted their hands andsaid they wanted to leave the new brand with the reputed‘company B.’ But they decided to give the distributionrights of the new brand to us. Our partner always goes anextra mile for us.” Statements like the following wereassigned a score of 1: “In our alliance with a surgicalfirm, they sent us a surgical device that we later foundhad bloodstains in it. We were shocked when our partnerfailed to take responsibility for this. We did not expectthis from a partner we have known for a long time now.”


terdependence was thus a composite score, withhigher values indicating higher interdependence,calculated as follows:

CIi !

!j!1

9 wi, j*SRi, j

9 ,

where CIi was the weighted composite interdepen-dence score for alliance i; wi, j were the weightsreflecting the level of interdependence implied byeach strategic rationale j in alliance i, j ! 1, . . . , 9;and SRi, j was the presence of strategic rationale j inalliance i.

To ensure that the weighting scheme did not affectthe robustness of our findings, we reran our modelswhile coding the interdependence variable with al-ternative specifications based on Gulati and Singh(1998). First, we placed each alliance into one of thethree categories of interdependence, on the basis ofthe highest level of interdependence present in it,where reciprocal was higher than sequential, whichwas higher than pooled (Thompson, 1967: 55). Weobtained results consistent with those below whenrank-ordering the categories so that alliances withreciprocal (and possibly other) elements were as-signed a value of 2, alliances with sequential (andpossibly pooled) elements had a value of 1, and alli-ances limited to pooled interdependence had a valueof 0, as Gulati and Singh (1998: 798) suggested. Wealso replicated our results with a variable with valuesbetween 1 and 9 that were based on the rationale(s)present in an alliance.

Interpartner competition. Prior empirical re-search has classified alliances as being betweenpotential competitors when both partners operatedin the same four-digit SIC code (e.g., Mowery, Ox-ley, & Silverman, 1996; Oxley & Sampson, 2004;Park & Russo, 1996). To capture more accuratelythe extent of interpartner competition present inthe alliances, we refined the implied binary mea-sure into three categories indicating different de-grees of competitive overlap. We assigned a score of2 if an alliance operated in the same four-digit SICcode industry as both partners and neither partnerwas active in any other four-digit SIC code area.Because the partners in such an alliance are hori-zontally related and the alliance’s activities arecentral to their businesses, the concerns aboutbreeding a potential competitor are likely to be veryhigh. We gave a score of 1 if both partners operatedin the same four-digit SIC code area as that of thealliance, but one or (most often) both partners wereactive in other four-digit SIC codes as well. Thissituation indicates a horizontal relationship amongpartners, but one in which concerns about potential

competition are somewhat lower than in the situa-tion coded 2 because the overlap is less central tothe partners’ businesses. Finally, we assigned ascore of 0 if the partners did not operate in the sameindustries. This category includes pure cases ofvertical relationship of partners via alliance (forinstance, one partner supplies inputs for the alli-ance, which operates in the same business as theother partner). Though such an alliance might bestrategic for the partners, they operate in differentindustries, so they are less likely to be potentialcompetitors (Harrigan, 1985).

We performed several specification checks. First,we assigned a value of 1 if the primary operationsof both partners were in the same four-digit SICcode, and 0 otherwise. Second, we assigned a valueof 1 for the highest degree of competitive overlap—that is, if an alliance operated in the same four-digitSIC code as both partners and neither partner wasactive in any other four-digit SIC code industry,and 0 otherwise. Finally, we used the conventionalif less precise operationalization whereby partnersare considered rivals if they share an SIC code. Theresults were robust.

Environmental instability. Although environ-mental instability may have various sources, wechose to concentrate on product-market instabilityin our empirical analysis. The reason was that, withall of our alliances operating within the same coun-try, instability caused by other sources, such asregulatory regime, was likely to be consistent forthe sampled alliances. The dynamics of product-markets, on the contrary, varied between indus-tries, and were therefore a better source of environ-mental instability for testing our hypotheses. As inprevious research that has followed the same logic,our measure captures instability in the sales of eachalliance’s industry over the preceding five years(Bergh & Lawless, 1998; Keats & Hitt, 1988). Datawere obtained from Capitaline, an authoritative da-tabase of financial information about India. We re-gressed industry sales on year and divided the stan-dard error of the regression slope coefficient by themean of industry sales (Bergh & Lawless, 1998;Dess & Beard, 1984). Larger values indicated greaterenvironmental instability.

Environmental unpredictability. For the rea-sons explained above, we concentrated on the un-predictability of product-markets to gauge the ef-fect of environmental unpredictability. Ourmeasure used industry sales data spanning fiveyears, obtained from Capitaline, to capture the ex-tent to which alliance partners could predict futuretrends in product-markets from the recent past(Glick, Ogilvie, & Miller, 1990; Wholey & Brittain,1989). We measured unpredictability as the coeffi-


cient of alienation (1 – R2) of the regression ofindustry sales in the survey year on the industrysales of the five preceding years (e.g., Delacroix &Swaminathan, 1991).

Control Variables

Investment size. Because prior research sug-gested that the size of partners’ investment in analliance may affect their commitment to allianceoperations (influencing the performance of the al-liance), we controlled for investment size (Parkhe,1993). Following prior research (Luo, 2002), weoperationalized investment size as the total amountof investment by both partners measured on a five-point interval scale.

Cultural distance. Cultural distance could be re-lated to alliance performance (Luo, 2002; Pothuku-chi, Damanpour, Choi, Chen, & Park, 2002). Fol-lowing extant literature, we controlled for nationalcultural distance by using Kogut and Singh’s (1988)measure, which is based on Hofstede’s (1980) fourcultural dimensions. We used Kogut and Singh’s(1988) formula to compute the distance betweenIndia’s national culture and that of the country oforigin of each foreign partner.

Equity alliance. The governance mode within analliance (Gulati, 1995; Oxley, 1999; Oxley & Samp-son, 2004) may be indicative of the motives of thepartners and have a large impact on alliance per-formance (Osborn & Baughn, 1990; Saxton, 1997).We coded alliance governance mode by a binaryvariable, assigning 1 to alliances that involved theuse of equity and 0 for nonequity alliances (e.g.,Gulati, 1995; Saxton, 1997).

Alliance duration. Partners in long-lasting alli-ances have had enough time to develop mutualunderstanding, and thus conflicts that hamper re-lationship performance may be less likely (Lin &Germain, 1998; Martin, Swaminathan, & Mitchell,1998). Duration was measured by an item capturingthe number of years an alliance had been in exis-tence at the time of measurement (e.g., Kotabe,Martin, & Domoto, 2003; Simonin, 1999).

Quality of information exchanged. A higherquality of information exchange may influence al-liance performance and trust, irrespective of thelevel of uncertainty in an alliance (Aulakh et al.,1996). We measured the quality of the informationexchanged in an alliance with a five-item Likertscale capturing the frequency, density, and open-ness of communication, as distinguished by Guptaand Govindarajan (1991). This construct had aCronbach’s alpha of .81.

Position of respondent. We assigned a value of 1if a respondent held the CEO’s position or its equiv-

alent in a local firm participating in a sampledalliance with a foreign partner, and 0 otherwise.

Local partner size. We controlled for the size ofthe local partner by using the log of the number ofemployees (e.g., Deeds & Rothaermel, 2003).

Industry dummies. Alliances in certain indus-tries may systematically perform better than thosein other industries owing to differences in industrystructure (Steensma, Tihanyi, Lyles, & Dhanaraj,2005). To control for industry differences, we useddummy variables for the major industries in oursample, based on two-digit SIC codes.

Analysis

Measurement analysis was conducted using LIS-REL’s 8.3 maximum likelihood program (Joreskog &Sorbom, 1996). We performed confirmatory factoranalysis using LISREL to check for convergent anddiscriminant validity.

We used ordinary least squares regression analy-sis to examine alliance performance. To control forthe possible endogeneity of the choice between eq-uity or nonequity alliance (Sampson, 2004), weimplemented Heckman’s (1979) two-stage tech-nique; Appendix B discusses this technique andprovides the results of the first-stage probit model.Before calculating the interaction terms used to testHypotheses 2 through 5, we mean-centered thevariables involved (Aiken & West, 1991).

RESULTS

Reliability and Validity

All constructs displayed satisfactory levels of re-liability, as indicated by the composite reliabilitiesranging from 0.81 to 0.90 (Nunnally, 1978; see Ap-pendix A for the composite reliabilities). Conver-gent validity, the extent to which different attemptsto measure a construct agree (Campbell & Fiske,1959), can be judged by looking at the item load-ings. Each loading (!) for the multi-item constructsof trust, quality of information exchanged, and al-liance performance was significantly related to itsunderlying factor, and all standardized item load-ings were well above the cutoff of .50 (Hildebrandt,1987), supporting convergent validity.

A series of chi-square difference tests on the fac-tor correlations showed that discriminant validity,the extent to which a construct differs from others,was achieved among all the survey-based con-structs: trust, quality of information exchanged,and alliance performance (Bagozzi, 1993; Joreskog,1971). It was particularly important that discrimi-nant validity be achieved between the constructs of


trust and performance, as these were our focal la-tent constructs. We constrained the estimated cor-relation parameter between trust and performanceto 1.0 and then performed a chi-square differencetest on the values obtained for the constrained andunconstrained models. The significant differencein chi-square (!!2 " 45.59, !df " 1, p # 0.000)indicated that the two constructs were not perfectlycorrelated and that discriminant validity wasachieved (see Anderson & Gerbing, 1988). We alsocompared the comparative fit and goodness-of-fitindexes between the constrained and uncon-strained models and found that the difference wasmoderately large (!CFI " .11, !GFI " .09), againsuggesting sufficient discriminant validity (e.g., Ba-gozzi & Yi, 1990). We carried out the same proce-dures for other construct combinations, too, withsimilar results.

We collected most of our data using a singlesurvey instrument and a single informant per alli-ance. To address the potential concerns of commonmethod bias and single informant bias, we usedseveral procedural and statistical remedies. Specif-ically, we undertook the procedural remedies ofprotecting respondent anonymity, reducing itemambiguity, separating scale items for the trust andperformance measures, and obtaining data fromdifferent sources for most of the moderator vari-ables (three of four) and several control variables.Our statistical remedies included triangulation ofsurvey data with data obtained from secondarysources and from field interviews, partial correla-tion adjustment, and Harman’s (1967) one-factor

test. We also examined the results for significantinteractions, which are less likely to occur in thepresence of single informant bias (Kotabe et al.,2003). Appendix C reports each of these steps indetail. These procedures left us confident that nei-ther common method nor single informant bias wasa serious problem in our study.

Tests of Hypotheses

Table 1 reports means, standard deviations, andcorrelations for all variables. Table 2 reports theresults of the regression model predicting allianceperformance.

We tested six regression equations for the allianceperformance variable, as reported in Table 2. Afterincluding only the control variables in model 1, weintroduced the main variables in model 2. We intro-duced the interactions of trust with the behavioraland environmental uncertainty variables separatelyin models 3 and 4. In model 5, we included all theinteraction terms simultaneously. Finally, we reranmodel 5 with industry dummies to create model 6.Model 1 is significant (p # .001), and the controlvariables explain 31 percent of the variance in alli-ance performance. Alliance duration (p # .05) andthe quality of information exchanged (p # .001) arepositively related to alliance performance. Because ofthe strong relationship of the quality of informationexchanged with alliance performance and theformer’s possible overlap with trust, we replicatedour models without the quality of information ex-changed. The pattern of our findings did not change

TABLE 1Descriptive Statistics and Correlationsa

Variables Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12 13

1. Alliance performance 3.75 0.782. Position of respondent 0.58 0.49 $.043. Investment size 2.42 0.95 .17 $.064. Cultural distance 1.86 0.42 .09 $.03 .185. Local partner size 5.46 1.63 .15 $.20 .39 .236. Equity alliance 0.58 0.49 .14 .03 .17 $.02 $.067. Alliance duration 12.10 9.33 .18 $.19 .15 .02 .25 .098. Quality of information

exchanged3.12 0.74 .49 $.07 .30 .04 .19 .34 $.06

9. Interdependence 0.50 0.38 .11 $.03 .23 $.03 $.02 .14 $.12 .3210. Interpartner competition 1.28 0.63 $.24 .15 $.06 $.18 $.22 .06 .03 $.14 .0311. Environmental instability 0.04 0.03 $.03 $.01 $.07 $.01 .06 .00 .01 .03 $.03 .0112. Environmental

unpredictability0.57 0.32 .11 $.03 $.04 $.10 $.05 $.15 $.00 $.03 $.04 .11 $.19

13. Trust 3.74 0.77 .52 $.09 .15 .01 .05 .27 .17 .43 .14 $.12 $.14 .0714. "b 0.00 0.76 .13 .03 .00 $.05 $.00 .62 .11 .27 .00 .00 .00 $.00 .29

a n " 126. Correlations with absolute value greater than .17 are significant at the .05 level. Means and standard deviations reported hereare for raw scores.

b Correction for the endogeneity of the alliance governance choice (equity vs. nonequity). See Appendix B for an explanation.


TA

BLE

2R

esul

tsof

Reg

ress

ion

Ana

lysi

sfo

rA

llia

nce

Perf

orm

ance

a

Var

iabl

es

Hyp

othe

sis

Num

ber

and

Pred

icti

onM

odel

1M

odel

2M

odel

3M

odel

4M

odel

5M

odel

6

Inte

rcep

t0.

58(0

.56)

!0.

92(0

.72)

!1.

02(0

.72)

!1.

06(0

.73)

!1.

17(0

.75)

Posi

tion

ofre

spon

dent

0.07

(0.1

2)0.

15(0

.11)

0.13

(0.1

0)0.

17(0

.11)

0.15

(0.1

0)0.

14(0

.10)

Inve

stm

ent

size

0.12

(0.1

1)!

0.11

(0.1

1)!

0.12

(0.1

1)!

0.14

(0.1

1)!

0.15

(0.1

1)!

0.16

(0.1

1)C

ultu

ral

dist

ance

0.13

(0.1

5)0.

10(0

.14)

0.12

(0.1

4)0.

07(0

.13)

0.09

(0.1

3)0.

10(0

.14)

Loca

lpa

rtne

rsi

ze!

0.04

(0.0

5)0.

03(0

.05)

0.04

(0.0

5)0.

05(0

.05)

0.05

(0.0

5)0.

06(0

.05)

Equi

tyal

lian

ce!

1.06

†(0

.61)

0.78

(0.8

1)0.

91(0

.80)

0.97

(0.8

1)1.

08(0

.81)

1.19

(0.8

0)A

llia

nce

dura

tion

0.02

*(0

.01)

0.02

*(0

.01)

0.02

*(0

.01)

0.02

*(0

.01)

0.02

*(0

.01)

0.02

*(0

.01)

Qua

lity

ofin

form

atio

nex

chan

ged

0.58

***

(0.0

8)0.

29**

(0.0

9)0.

29**

(0.0

9)0.

27**

(0.0

9)0.

27**

(0.0

9)0.

28**

(0.0

9)

Mai

nef

fect

sIn

terd

epen

denc

e!

0.19

(0.3

2)!

0.24

(0.3

1)!

0.19

(0.3

1)!

0.25

(0.3

2)!

0.25

(0.3

3)In

terp

artn

erco

mpe

titi

on!

0.22

*(0

.10)

!0.

22*

(0.0

9)!

0.23

*(0

.09)

!0.

23*

(0.0

9)!

0.23

*(0

.09)

Envi

ronm

enta

lin

stab

ilit

y0.

17(1

.13)

!0.

12(0

.89)

!0.

98(1

.12)

!1.

03(0

.97)

!0.

99(1

.04)

Envi

ronm

enta

lun

pred

icta

bili

ty0.

45†

(0.2

6)0.

48†

(0.2

6)0.

49†

(0.2

6)0.

52†

(0.2

7)0.

53†

(0.2

7)

Tru

st1,

"0.

51**

*(0

.09)

0.47

***

(0.0

9)0.

51**

*(0

.09)

0.46

***

(0.0

8)0.

47**

*(0

.08)

Inte

ract

ions

Tru

st#

inte

rdep

ende

nce

2,"

0.62

*(0

.27)

0.64

*(0

.31)

0.65

*(0

.31)

Tru

st#

inte

rpar

tner

com

peti

tion

3,"

0.30

*(0

.12)

0.29

*(0

.11)

0.30

*(0

.12)

Tru

st#

envi

ronm

enta

lin

stab

ilit

y4,

!!

1.75

†(0

.90)

!1.

31†

(0.7

7)!

1.33

†(0

.78)

Tru

st#

envi

ronm

enta

lun

pred

icta

bili

ty5,

!!

0.43

†(0

.23)

!0.

49*

(0.2

1)!

0.50

*(0

.20)

Indu

stry

dum

mie

sb

!c

0.62

(0.4

0)!

0.57

(0.5

0)!

0.67

(0.5

1)!

0.71

(0.5

0)!

0.80

(0.5

1)!

0.81

(0.5

1)R

2.3

2.4

9.5

3.5

1.5

6.5

6$

R2

.17

.04

.03

.07

.00

$F

6.80

***

5.25

**3.

25*

4.25

**0.

16

an

%12

6.T

hech

ange

sin

R2

inm

odel

s3

thro

ugh

5ar

ein

com

pari

son

toth

eva

lue

for

R2

inm

odel

2.T

heco

effi

cien

tsre

port

edar

eun

stan

dard

ized

esti

mat

es,w

ith

stan

dard

erro

rsin

pare

nthe

ses.

bT

hein

dust

rydu

mm

ies

wer

eno

tsi

gnif

ican

t.c

Cor

rect

ion

for

the

endo

gene

ity

ofth

eal

lian

cego

vern

ance

choi

ce(e

quit

yvs

.non

equi

ty).

See

App

endi

xB

for

anex

plan

atio

n.†

p&

.10

*p

&.0

5**

p&

.01

***

p&

.001

appreciably. The equity alliance dummy is margin-ally and negatively related to alliance performance(p ! .10), but only in the base model.

Hypothesis 1 predicts that trust will be positivelyrelated to alliance performance. The coefficient fortrust in model 2 is positive and significant (b "0.51, p ! .001), thus supporting Hypothesis 1 andpast findings reported in the literature. Given meancentering, the trust coefficient shows the magni-tude of the relationship between trust and allianceperformance, with other variables held constant attheir mean values.

Trust, Behavioral Uncertainty, andAlliance Performance

The incremental variance accounted for by the in-teractions between the behavioral uncertainty vari-ables and trust is significant in model 3 (#R2 " 0.04,p ! .01). Hypothesis 2 predicts that alliances willbenefit more from interorganizational trust when thedegree of interdependence is higher. Hypothesis 3predicts that alliances will benefit more from interor-ganizational trust when the degree of potential inter-partner competition is higher. The coefficient of theinteraction of trust with interdependence is signifi-cant and positive (b " 0.62, p ! .05 in model 3),supporting Hypothesis 2. We also find a significantinteraction between trust and interpartner competi-tion (b " 0.30, p ! .05), thereby supporting Hypoth-

esis 3. These results show that the positive impact oftrust on alliance performance increases with interde-pendence and interpartner competition.

To further assess the implications of the regres-sion results, we plotted the relationship of trust andalliance performance over the observed range oftrust, with separate regression lines representingdifferent levels of interdependence. We created asimilar plot with interpartner competition. Theplotted lines represent the performance values ex-pected on the basis of unstandardized regressioncoefficients from the complete regression (model5). The low interdependence and low interpartnercompetition lines indicate values one standard de-viation below the mean, and the high interdepen-dence and high interpartner competition lines in-dicate values one standard deviation above themean. Figures 1 and 2 graphically support Hypoth-eses 2 and 3, respectively.

The simple slope test (Aiken & West, 1991) revealsthat the magnitude of the slope of alliance perfor-mance regressed on trust is nearly twice as large forhigh interdependence (simple slope: b " 0.57, t "5.29) as that for low interdependence (simple slope:b " 0.32, t " 3.55). The slope for high interpartnercompetition (simple slope: b " 0.66, t " 6.35) isnearly thrice as large as that for low interpartner com-petition (simple slope: b " 0.26, t " 2.41). Theseresults show that the trust-performance relationship

FIGURE 1Trust, Interdependence, and Alliance Performance


strengthens at high levels of the two variables indi-cating behavioral uncertainty.

Trust, Environmental Uncertainty, andAlliance Performance

The additional variance accounted for by the inter-actions between the environmental uncertainty vari-ables and trust is significant in model 4 (!R2 " 0.03,p # .05). Hypothesis 4 predicts that the benefits thatalliances derive from interorganizational trust dimin-ish when environmental instability is higher. Hy-pothesis 5 predicts that the benefits that alliancesderived from interorganizational trust diminish whenenvironmental unpredictability is higher. Marginallysignificant negative effects are found for the coeffi-cient of the interaction of trust with environmentalinstability in models 4 (b " –1.75, p # .10) and 5 (b "–1.31, p # .10), giving some support to Hypothesis 4.Model 4 also shows a marginally negative interactioneffect between trust and environmental unpredict-ability (b " – 0.43, p # .10), an effect that becomesstronger in the full model 5 (b " –0.49, p # .05).Overall, this pattern of findings supports Hypothesis5. These results suggest that the positive impact oftrust on alliance performance diminishes with highinstability and unpredictability in the environment ofan alliance.

To illustrate these interactions, we created plotsfor the trust–alliance performance relationship,

with separate regression lines representing differ-ent levels of environmental instability and unpre-dictability one standard deviation above and belowthe mean. Figures 3 and 4 graphically support Hy-potheses 4 and 5, respectively. The simple slopetest (Aiken & West, 1991) reveals that the magni-tude of the slopes of alliance performance regressedon trust is nearly two times smaller for high envi-ronmental instability (simple slope: b " 0.33, t "3.44) as for low environmental instability (simpleslope: b " 0.59, t " 5.06). The slope for high envi-ronmental unpredictability (simple slope: b " 0.30,t " 3.1) is more than two times smaller than forlow environmental unpredictability (simpleslope: b " 0.61, t " 5.75). Further probing of theinteractions (Aiken & West, 1991) revealed thatfor very high levels of environmental unpredict-ability, the relationship between trust and alli-ance performance becomes insignificant. Thesign of the simple slope coefficient of trust mayreverse for very high environmental unpredict-ability (b " – 0.05, t " – 0.28, at the high end ofthe observable range). These results indicate thatthe trust-performance relationship weakens andmay disappear altogether at high levels of theenvironmental uncertainty variables.

In model 6, we find that the addition of fixedeffects for the main industries in the sample doesnot substantially add to the explanatory power ofthe regression (!R2 " 0.00, n.s.). The industry dum-

FIGURE 2Trust, Interpartner Competition, and Alliance Performance


mies are not significant as a set (!F " 0.16, n.s.).Furthermore, the results reported above, includingall tests of hypotheses, do not change substantially.Thus, our results appear robust across industries.

All in all, the results consistently support ourargument that the positive relationship betweeninterorganizational trust and alliance performancestrengthens under conditions that foster behavioral

FIGURE 3Trust, Environmental Instability, and Alliance Performance

FIGURE 4Trust, Environmental Unpredictability, and Alliance Performance


uncertainty but weakens under conditions of envi-ronmental uncertainty.

DISCUSSION

By now there appears to be general support for theidea that trust is beneficial to alliances. However,recent studies have suggested that the impact of truston alliance performance may be contingent on otherfactors. Yet previous research has not yielded a gen-eral theory regarding the conditions under whichtrust facilitates or fails to facilitate alliance perfor-mance. We have presented such a theory, one basedon the distinction between behavioral and environ-mental uncertainty, and shown empirically that,apart from the positive direct relationship betweentrust and alliance performance, more subtle interac-tion effects can be distinguished. The relationshipbetween trust and alliance performance is moderatedby the type of uncertainty prevailing in a particularalliance, with behavioral uncertainty strengthening,and environmental uncertainty weakening, the rela-tionship between trust and performance.

Contributions and Implications

Our research makes several contributions. First, weextend the interorganizational trust-performance lit-erature by demonstrating that the type of uncertaintyfacing alliance partners conditions the relationshipbetween trust and alliance performance. Specifically,behavioral and environmental uncertainties have op-posite moderating effects on that relationship. Thequite distinct challenges posed by these two types ofuncertainty bring about these differing effects. Trustessentially reduces the likelihood of negative inter-pretations of partner actions by allowing for the ben-efit of the doubt. This allowance facilitates opennessin sharing knowledge and reduces fear of opportunis-tic behavior by partners. Hence, the benefits fromtrust are magnified when behavioral uncertainty ishigh. In contrast, the benefits from trust are reducedwhen environmental uncertainty is strong, becauseoverconfidence in the information provided by eachpartner restrains the vigilant environmental scanningand cross-fertilization of views that is of vital impor-tance under this condition. This contrast has impli-cations for research on trust, in that it shows that trustcan be a double-edged sword, with a performance-enhancing potential that increases under certain con-ditions but decreases under other conditions.

Second, extant research suggests that potentialcompetition and high interdependence between part-ners are likely to hamper alliances. Behavioral uncer-tainty related to partner actions, such as the potential

for misappropriation of proprietary know-how, isconsiderable in alliances that are potentially compet-itive and/or where partners are highly interdepen-dent. We contribute to this stream of research byempirically showing that trust brings about benefitsby attenuating the effects of behavioral uncertainty inalliances where potential interpartner competitionand interdependence are high. This finding under-scores the potential benefits of investing in trustwhen behavioral uncertainty is considerable and alsosuggests that trust figures among the relational mech-anisms and norms that can support alliance perfor-mance by allowing partners to realize their potentialsynergies (see Madhok & Tallman, 1998). Moreover,our findings underline the necessity of taking intoaccount both operational features (e.g., interdepen-dence) and behavioral characteristics (e.g., trust) instudying alliance success. Advising such a dual focusruns counter to the emphasis on the isolated influ-ence of either tangible alliance features or behavioralpatterns in much prior research (Contractor, 2005;Doz, 1996; Yan & Zeng, 1999). Likewise, our studyshows the benefits of examining industry factors si-multaneously with alliance- and partner-level effects.Third, the challenges environmental uncertaintyposes for firms are well documented. However, re-searchers’ knowledge about the role of environmentaluncertainty in strategic alliance performance is lim-ited. This gap is all the more relevant as environmen-tal uncertainty is commonly advanced as a leadingreason that firms form alliances in the first place (e.g.,Harrigan, 1988; Pfeffer & Nowak, 1976). Because alli-ances involve the interests of more than one firm,relational norms such as trust also shape the mannerin which partners respond to various external chal-lenges. Therefore, it is vital for alliance researchers tounderstand the role of trust in shaping responses tothe challenges posed by environmental uncertainty,and the implications of this role for performance.Prior research, for the most part, has stressed thebeneficial effects of trust. However, our study showsthat this beneficial effect decreases with environmen-tal uncertainty and suggests that, at extreme levels ofenvironmental uncertainty, trust may even have adetrimental effect on alliance performance. The cog-nitive comfort trust provides may reduce the alert-ness and cross-fertilization needed in the presence ofstrong environmental uncertainty. Hence, alliancepartners should exercise caution in relying on interor-ganizational trust and accelerate scanning and searchefforts under environmental uncertainty rather thanrelying unduly on each other to deal with exogenousuncertainty. Thus, our study adds to the researchagenda on the limits of trust (e.g., McEvily et al.,2003).

As for relevance to practice, we have established


that successful international alliances take into ac-count that trust can have different impacts underdifferent types of uncertainty. Because the inten-tional cultivation of interorganizational trust in-volves costs as well as opportunities forgone(Poppo & Zenger, 2002; Sako, 1991), such an effortshould be undertaken only when the expected pay-off is positive. Our results reveal that, at very highlevels of environmental unpredictability, the trust-performance relationship disappears and possiblyreverses. The costs of cultivating trust might thusoutweigh the expected benefits. Hence, firms oughtto expend efforts in developing interorganizationaltrust specifically when it has the potential to helpaddress behavioral uncertainty (in the presence ofhigh interdependence and/or latent competitionamong partners), and exercise caution when rely-ing on trust under conditions of strong environ-mental uncertainty (as caused by product-marketinstability or unpredictability).

Limitations and Suggestions for Future Research

This study has some limitations. First, we col-lected data concerning the perspectives of bothpartners on alliances through a survey of the Indianpartners only. Geringer and Hebert (1991: 252, 256)found a significant, positive correlation between afocal parent’s satisfaction with alliance perfor-mance and the perception by the other partner ofthis focal parent’s satisfaction. However, it wouldbe valuable to gain both partners’ perspectives oneach alliance. Yet gathering such information couldbe very challenging, especially with parent firmsoriginating from many countries, as in our sample(21 partner countries).

Second, the operationalization of our moderatingconstructs, behavioral and environmental uncer-tainty, does not preclude other sources. We do be-lieve that interdependence and interpartner com-petition are key aspects of behavioral uncertainty(Das & Teng, 2000; Park & Ungson, 2001), thoughthere may be others. Likewise, product-market in-stability and unpredictability are critical sources ofenvironmental uncertainty and are perhaps itsmost commonly used indicators in organizationalresearch (e.g. Buchko, 1994; Delacroix & Swami-nathan, 1991; Glick et al., 1990; Wholey & Brittain,1989). Nevertheless, uncertainty stemming fromregulatory and political instability may be expectedto also matter in the case of alliances located intransition economies (e.g., Delios & Henisz, 2003).

Third, single-informant bias is a potential short-coming of our research. Yet although single infor-mants provided data on both trust and alliance per-formance, we have several reasons to believe that

single-informant bias is not a serious concern in ourstudy. As outlined in Appendix C, we undertookmultiple procedural and statistical remedies, includ-ing protecting respondent anonymity, reducing itemambiguity, separating scale items for trust and perfor-mance measures, obtaining data from differentsources for most of the moderator variables (three ofthe four) and for several control variables, makingpartial correlation adjustments, and applying Har-man’s (1967) one-factor test; we also validated ourtrust and alliance performance measures, using fieldinterviews and archival data, respectively. Further-more, all our interaction effects are significant, espe-cially those that involve two survey-based measures(trust and interdependence). The significance of suchinteractions is unlikely to be an artifact of the single-informant method, as the respondents are unlikely tohave consciously theorized the moderated relation-ships when responding to the survey (Brockner, Sie-gel, Daly, Tyler, & Martin, 1997; Kotabe et al., 2003).Moreover, the interactions also make reverse causal-ity less plausible.

Our study suggests a number of interesting oppor-tunities for future research. First, our approach to thestudy of trust may be generalized beyond interfirmalliances. For instance, trust has been suggested tobenefit knowledge sharing in intraorganizational con-texts too (e.g., Makino & Inkpen, 2003; Tsai &Ghoshal, 1998). It would be interesting to explore towhat extent the moderating effects of behavioral andenvironmental uncertainty can also be found in theintraorganizational context. Our reasoning suggeststhat under conditions of strong environmental uncer-tainty, the effect of high intraorganizational trust maybe weakened (and perhaps reversed), following thesame course we found in the interorganizational con-text. Relatedly, our results on environmental uncer-tainty and trust may have implications for researchon team performance. Langfred (2004) showed thatteams that rely on trust were less likely to monitortheir team members. Future research could examinethe implications of reliance on trust for the perfor-mance of teams operating under different levels ofenvironmental uncertainty.

Second, we focused our analysis on interorganiza-tional trust, which describes the aggregate relation-ship between partner firms and is the predominantform of trust to affect performance and satisfaction atan organizational level (Zaheer et al., 1998). Trustmay also arise between the top managers of partnerfirms. These individuals generally shape initial alli-ance strategies (such as whether to use equity gover-nance, a feature that we controlled for) but not ongo-ing activities (Krishnan, 2006). As Zaheer andcolleagues (1998) found, interpersonal trust is un-likely to matter nearly as much to our outcome of


interest as interorganizational trust. Nevertheless, itwould be interesting to examine any effects of uncer-tainty on the relationships between interpersonaltrust and alliance decisions (and perhaps perfor-mance). In this respect, we would expect that indi-vidual managers, too, are prone to be blinded by trustin the presence of environmental uncertainty.

Third, uncertainty related to appropriation of pro-prietary knowledge and complete sharing of informa-tion is more salient in knowledge-intensive alliances.It has been argued that trust enables firms to cooper-ate despite such uncertainty (e.g., Dyer & Chu, 2003).Nevertheless, in knowledge-intensive contexts suchas R&D and new-product development, environmen-tal uncertainty may be high—and may occur along-side behavioral uncertainty in the case of allied firms(Harrigan, 1988; Martin & Salomon, 2003a, 2003b).Research examining such contexts could yield fur-ther insights into the conditions under which the netbenefits of trust can be sustained.

Fourth, we examined the impact of trust on alli-ance performance. Other relational mechanismsand norms may have similar contingent effects onalliance performance. For instance, Parkhe (1991)argued that tension within alliances may also resultfrom factors other than behavioral uncertainty,such as cultural differences. He suggested that rou-tines such as training—rather than trust—may go along way toward reducing cultural conflict and im-proving alliance performance. Moreover, relationalgovernance may involve “norms of flexibility, sol-idarity, bilateralism and continuance” (Poppo &Zenger, 2002: 712), norms that can operate in thepresence of uncertainty. Further research exploringsuch alternatives and complements to trust, then,would be well warranted.3 Partner reputation, es-pecially when amplified through a network of alli-ances, can have potent effects too (Gulati, 1998;Nooteboom, Berger, & Noorderhaven, 1997). Fi-nally, prior ties between partners are a potentsource of shared understanding (Zollo et al., 2002),and their role deserves attention alongside trust.

Conclusion

Our research provides significant insights intothe advantages and limitations of interorganization-al trust for strategic alliances. Specifically, thestudy underscores the need to move beyond a focuson the direct link between trust and alliance per-formance in seeking to understand the conditionsunder which trust promotes or inhibits allianceperformance. Researchers (and managers) ought totake into account the type of uncertainty facingalliance partners—that is, whether the source ofuncertainty is internal or external to their alliance.In our study, the type of uncertainty moderated therelationship between trust and alliance perfor-mance in such a way that the trust-alliance perfor-mance relationship strengthened under behavioraluncertainty and weakened under environmentaluncertainty. We hope that our study triggers futurestudies that will look in more detail at the compli-cated and contingent role of trust in inter- andintraorganizational relationships.

REFERENCES

Aguilar, F. J. 1967. Scanning the business environment.New York: Macmillan.

Aiken, L. S., & West, S. G. 1991. Multiple regression:Testing and interpreting interactions. ThousandOaks, CA: Sage.

Aldrich, H. E. 1979. Organizations and environments.Englewood Cliffs, NJ: Prentice-Hall.

Anderson, C. R., & Paine, F. T. 1975. Managerial percep-tions and strategic behavior. Academy of Manage-ment Journal, 18: 811–823.

Anderson, J. C., & Gerbing, D. W. 1988. Structural equa-tion modeling in practice: A review and recom-mended two-step approach. Psychological Bulletin,103: 411–423.

Armstrong, S. J., & Overton, T. S. 1977. Estimating non-response bias in mail surveys. Journal of MarketingResearch, 14: 396–402.

Aulakh, P., Kotabe, M., & Sahay, A. 1996. Trust andperformance in cross-border marketing partnerships:A behavioral approach. Journal of InternationalBusiness Studies, 27: 1005–1032.

Bagozzi, R. P. 1993. Assessing construct validity in person-ality research: Applications to measures of self-esteem.Journal of Research in Personality, 27: 49–87.

Bagozzi, R. P., & Yi, Y. 1990. Assessing method variancein multitrait-multimethod matrices: The case of self-reported affect and perceptions at work. Journal ofApplied Psychology, 75: 547–560.

Balasubramanyam, V. N. 2003. India: Trade policy re-view. World Economy, 26: 1357–1368.

3 It should be noted that distinguishing trust from al-ternative norms is not straightforward. In Poppo andZenger (2002: 715), three items yielded one “compositefactor with the following underlying norms and dimen-sions: open communication and sharing of information,trust, dependence, and cooperation” (! ! .78). Gibsonand Birkinshaw (2004), examining intraorganizationalfunctioning, found that nine items meant to measure“support” and “trust” (Ghoshal & Bartlett, 1994) loadedonto a single factor (! ! .93). Still, we believe that furtherscale development could yield finer differentiations tosupport the type of research suggested here.


Barnes, J. H. 1984. Cognitive biases and their impact onstrategic planning. Strategic Management Journal,5: 129–137.

Barney, J. B., & Hansen, M. H. 1994. Trustworthiness as asource of competitive advantage. Strategic Manage-ment Journal, 15: 175–190.

Baum, J. A. C., Calabrese, T., & Silverman, B. S. 2000.Don’t go it alone: Alliance networks and startups’performance in Canadian biotechnology. StrategicManagement Journal, 21: 267–294.

Bazerman, M. 1998. Judgement in managerial decisionmaking. New York: Wiley.

Beamish, P. W., & Banks, J. C. 1987. Equity joint venturesand the theory of the multinational enterprise. Jour-nal of International Business Studies, 18(2): 1–16.

Bergh, D. D., & Lawless, M. W. 1998. Portfolio restruc-turing and limits to hierarchical governance: Theeffects of environmental uncertainty and diversifica-tion strategy. Organization Science, 9: 87–102.

Bhaumik, S. K., Beena, P. L., Bhandari, L. P., & Gokarn, S.2003. Survey of FDI in India. Working paper no. 6,London Business School, Centre for New and Emerg-ing Markets.

Bleeke, J., & Ernst, D. 1993. Collaborating to compete.New York: Wiley.

Bowonder, B., & Richardson, P. K. 2000. Liberalizationand the growth of business-led R&D: The case ofIndia. R&D Management, 30: 279–288.

Bradach, J. L., & Eccles, R. G. 1989. Price, authority andtrust: From ideal types to plural forms. In W. R. Scott& J. Blake (Eds.), Annual review of sociology, vol.15: 97–118. Palo Alto, CA: Annual Reviews.

Brockner, J., Siegel, P. A., Daly, J. P., Tyler, T., & Martin,C. 1997. When trust matters: The moderating effectof outcome favorability. Administrative ScienceQuarterly, 42: 558–583.

Buchko, A. A. 1994. Conceptualization and measurementof environmental uncertainty: An assessment of theMiles and Snow perceived environmental uncer-tainty scale. Academy of Management Journal, 37:410–425.

Buckley, P. J., & Glaister, K. W. 2002. What do we knowabout international joint ventures? In F. Contractor &P. Lorange (Eds.), Cooperative strategies and alli-ances: 49–69. Amsterdam: Pergamon.

Cameron, K. S., & Kim, M. U., & Whetten, D. A. 1987.Organizational effects of decline and turbulence. Ad-ministrative Science Quarterly, 32: 222–240.

Campbell, D. T., & Fiske, D. W. 1959. Convergent anddiscriminant validation by the multitrait-multim-ethod matrix. Psychological Bulletin, 56: 81–105.

Carson, S. J., Madhok, A., Varman, R., & John, G. 2003.Information processing moderators of the effective-ness of trust-based governance in inter-firm R&D col-laboration. Organization Science, 14: 45–56.

Child, J. 1972. Organizational structure, environmentand performance: The role of strategic choice. Soci-ology, 6: 2–21.

Cohen, W. M., & Levinthal, D. A. 1990. Absorptive capacity:A new perspective on learning and innovation. Ad-ministrative Science Quarterly, 35: 128–152.

Contractor, F. J. 2005. Alliance structure and process:Will the two research streams ever meet in allianceresearch? European Management Review, 2(2):123–129.

Cyert, R. M., & March, J. G. 1963. A behavioral theory ofthe firm. Englewood Cliffs, NJ: Prentice-Hall.

Das, T. K., & Teng, B. S. 2000. Instabilities of strategicalliances: An internal tensions perspective. Organi-zation Science, 11: 77–101.

Deeds, D. L., & Rothaermel, F. T. 2003. Honeymoons andliabilities: The relationship between age and perfor-mance in research and development alliances. Journalof Product Innovation Management, 20: 468–484.

Delacroix, J., & Swaminathan, A. 1991. Cosmetic, specu-lative and adaptive organizational change in thewine industry: A longitudinal study. AdministrativeScience Quarterly, 36: 631–661.

Delios, A., & Henisz, W. J. 2003. Political hazards, expe-rience, and sequential entry strategies: The interna-tional expansion of Japanese firms, 1980–1998.Strategic Management Journal, 24: 1153–1164.

Dess, G. G., & Beard, D. W. 1984. Dimensions of organi-zational task environments. Administrative ScienceQuarterly, 29: 52–73.

DeVellis, R. F. 1991. Scale development. Newbury Park,CA: Sage.

Dhanaraj, C., Lyles, M.A., Steensma, H. K., & Tihanyi, L.,2004. Managing tacit and explicit knowledge trans-fer in IJVs: the role of relational embeddedness andthe impact on performance. Journal of InternationalBusiness Studies, 35: 428–442

Dillman, D. A. 2000. Mail and Internet surveys: Thetailored design method. New York: Wiley.

Doz, Y. L. 1996. The evolution of cooperation in strategicalliances: Initial conditions or learning processes.Strategic Management Journal, 17: 55–83.

Dyer, J. H., & Chu, W. 2003. The role of trustworthiness inreducing transaction costs and improving perfor-mance: Empirical evidence from the United States,Japan, and Korea. Organization Science, 14: 57–68.

Ferrin, D. L., & Dirks, K. T. 2003. The use of rewards toincrease and decrease trust: Mediating processes anddifferential effects. Organization Science, 14: 18–31.

Gambetta, D. 1988. Trust: The making and breaking ofcooperative relationships. Oxford, U.K.: BasilBlackwell.

Geringer, J. M., & Hebert, L. 1991. Measuring perfor-


mance of international joint ventures. Journal ofInternational Business Studies, 22: 249–264.

Geyskens, I., Steenkamp, J. B. E. M., & Kumar, N. 1999. Ameta-analysis of satisfaction in marketing channelrelationships. Journal of Marketing Research,36(2): 223–239.

Ghemawat, P., Porter, M. E., & Rawlinson, R. A. 1986.Patterns of international coalition activity. In M. E.Porter (Ed.), Competition in global industries: 345–365. Boston: Harvard Business School Press.

Ghoshal, S., & Bartlett, C. A. 1994. Linking organizationalcontext and managerial action: The dimensions ofquality of management. Strategic ManagementJournal, 15: 91–112.

Gibson, C. B., & Birkinshaw, J. 2004. The antecedents,consequences and mediating role of organizationalambidexterity. Academy of Management Journal,47: 209–226.

Glick, W. H., Ogilvie, dt, & Miller, C. C. 1990. Assessingdimensions of task environments: Intra-industryand aggregate industry measures. Paper presentedat the annual meeting of the Academy of Manage-ment, San Francisco.

Greene, W. H. 1997. Econometric analysis. Upper Sad-dle River, NJ: Prentice-Hall.

Griffith, D. A., & Myers, M. B. 2005. The performanceimplications of strategic fit of relational norm gover-nance strategies in global supply chain relation-ships. Journal of International Business Studies,36: 254–269.

Grindley, P., Mowery, D. C., & Silverman, B. S. 1994.SEMATECH and collaborative research: Lessons inthe design of high-technology consortia. Journal ofPolicy Analysis and Management, 13: 723–758.

Gulati, R. 1995. Does familiarity breed trust? The impli-cations of repeated ties for contractual choice inalliances. Academy of Management Journal, 38:85–112.

Gulati, R. 1998. Alliances and networks. Strategic Man-agement Journal, 19: 293–317.

Gulati, R., & Singh, H. 1998. The architecture of cooper-ation: managing coordination costs and appropria-tion concerns in strategic alliances. AdministrativeScience Quarterly, 43: 781–814.

Gupta, A. K., & Govindarajan, V. 1991. Knowledge flowsand the structure of control within multinationalcorporations. Academy of Management Review,16: 768–792.

Hambrick, D. C. 1982. Environmental scanning and or-ganizational strategy. Strategic Management Jour-nal, 3: 159–174.

Hamel, G. 1991. Competition for competence and inter-partner learning within international strategic alli-ances. Strategic Management Journal, 12: 83–103.

Hamilton, B. H., & Nickerson, J. A. 2003. Correcting for

endogeneity in strategic management research. Stra-tegic Organization, 1: 51–78.

Harman, H. H. 1967. Modern factor analysis. Chicago:University of Chicago Press.

Harrigan, K. R. 1985. Strategies for joint ventures. Lex-ington, MA: Lexington Books.

Harrigan. K. R. 1988. Joint ventures and competitive strat-egy. Strategic Management Journal, 9: 141–158.

Heckman, J. 1979. Sample selection bias as a specifica-tion error. Econometrica, 47: 153–161.

Hildebrandt, L. 1987. Consumer retail satisfaction in ru-ral areas: A reanalysis of survey data. Journal ofEconomic Psychology, 8: 19–42.

Hofstede, G. 1980. Culture’s consequences: Interna-tional differences in work-related values. London:Sage.

Huber, J. P., Miller, C. C., & Glick, W. H. 1990. Develop-ing more encompassing theories about organiza-tions: The centralization-effectiveness relationshipas an example. Organization Science, 1: 11–40.

Isobe, T., Makino, S., & Montgomery, D. B. 2000. Re-source commitment, entry timing and market perfor-mance of foreign direct investments in emergingeconomies: The case of Japanese international jointventures in China. Academy of Management Jour-nal, 43: 468–484.

Jehn, K. A., Northcraft, G. B., & Neale, M. A. 1999. Whydifferences make difference: A field study of diver-sity, conflict, and performance in workgroups. Ad-ministrative Science Quarterly, 44: 741–763.

Joreskog, K. G. 1971. Statistical analysis of congenerictests. Psychometrika, 36: 109–133.

Joreskog, K. G., & Sorbom, D. 1996. LISREL 8: User’sreference guide. Chicago: Scientific Software Inter-national.

Kahneman, D., & Lovallo, D. 1993. Timid choices andbold forecasts: A cognitive perspective on risk tak-ing. Management Science, 39(1): 17–31.

Kale, P., Singh, H., & Perlmutter, H. 2000. Learning andprotection of proprietary assets in strategic alliances:Building relational capital. Strategic ManagementJournal, 21: 217–237.

Kearney, A. T. 2004. FDI confidence index. Global Busi-ness Policy Council, 7: 1–44.

Keats, B. W., & Hitt, M. A. 1988. A causal model oflinkages among environmental dimensions, macroorganizational characteristics and performance.Academy of Management Journal, 31: 570–598.

Khanna, T., Gulati, R., & Nohria, N. 1998. The dynamicsof learning alliances: Competition, cooperation, andrelative scope. Strategic Management Journal, 19:193–210.

Kogut, B. 1988. Joint ventures: Theoretical and empirical


perspectives. Strategic Management Journal, 9:319–332.

Kogut, B. 1989. The stability of joint ventures: Reciproc-ity and competitive rivalry. Journal of IndustrialEconomics, 38(2): 183–198.

Kogut, B., & Singh, H. 1988. The effect of national cultureon the choice of entry mode. Journal of Interna-tional Business Studies, 19: 411–432.

Koopmans, T. C. 1957. Three essays on the state ofeconomic science. New York: McGraw-Hill.

Kotabe, M., Martin, X., & Domoto, H. 2003. Gaining fromvertical partnerships: Knowledge transfer, relation-ship duration, and supplier performance improve-ment in the U.S. and Japanese automotive industries.Strategic Management Journal, 24: 293–316.

Krishnan, R. 2006. When is trust produced and whendoes it matter? Taking a closer look at trust andalliance performance. Tilburg, the Netherlands:CentER Dissertation Series.

Kumar, S., & Seth, A. 1998. The design of coordinationand control mechanisms for managing joint ventureparent relationships. Strategic Management Jour-nal, 19: 579–599.

Lane, P. J., Salk, J. E., & Lyles, M. A. 2001. Absorptivecapacity, learning and performance in internationaljoint ventures. Strategic Management Journal, 22:1139–1161.

Langfred, C. W. 2004. Too much of a good thing? Nega-tive effects of high trust and individual autonomy inself-managing teams. Academy of ManagementJournal, 47: 385–399.

Lau, C. M., & Woodman, R. W. 1995. Understandingorganizational change: A schematic perspective.Academy of Management Journal, 38: 537–554.

Lawrence, P. R., & Lorsch, J. W. 1973. A reply to Tosi,Aldag & Storey. Administrative Science Quarterly,18: 397–398.

Lee, T. W., Mitchell, T. R., Wise, L., & Fireman, S. Anunfolding model of voluntary employee turnover.Academy of Management Journal, 39: 5–36.

Lin, X., & Germain, R. 1998. Sustaining satisfactory jointventure relationships: The role of conflict resolutionstrategy. Journal of International Business Studies,29: 179–196.

Lindell, M. K., & Whitney, D. J. 2001. Accounting forcommon method variance in cross-sectional designs.Journal of Applied Psychology, 86: 114–121.

Luo, Y. 2002. Product diversification in internationaljoint ventures: Performance implications in anemerging market. Strategic Management Journal,23: 1–20.

Madhok, A., & Tallman, S. B. 1998. Resources, transac-tions and rents: Managing value through interfirmcollaborative relationships. Organization Science,9: 326–339.

Makino, S., & Inkpen, A. C. 2003. Knowledge seeking FDIand learning across borders. In M. Easterby-Smith &M. A. Lyles (Eds.), The Blackwell handbook of or-ganizational learning and knowledge manage-ment: 233–252. Oxford, U.K.: Blackwell.

Martin, X., & Salomon, R. 2003a. Tacitness, learning, andinternational expansion: A study of foreign directinvestment in a knowledge-intensive industry. Or-ganization Science, 14: 297–311.

Martin, X., & Salomon, R. 2003b. Knowledge transfercapacity and its implications for the theory of themultinational corporation. Journal of InternationalBusiness Studies, 34: 356–373.

Martin, X., Swaminathan, A., & Mitchell, W. 1998. Orga-nizational evolution in the interorganizational envi-ronment: Incentives and constraints on internationalexpansion strategy. Administrative Science Quar-terly, 43: 566–601.

Mayer, R. C., Davis, J. H., & Schoorman, F. D. 1995. Anintegrative model of organizational trust. Academyof Management Review, 20: 709–734.

McEvily, B., Perrone, V., & Zaheer, A. 2003. Trust as anorganizing principle. Organization Science, 14: 91–103.

Mintzberg, H. 1978. Patterns in strategy formation. Man-agement Science, 24: 934–948.

Mohr, J., & Spekman, R. 1994. Characteristics of partner-ship success: Partnership attributes, communicationbehavior and conflict resolution techniques. Strate-gic Management Journal, 15: 135–152.

Mowery, D. C., Oxley, J. E., & Silverman, B. S. 1996.Strategic alliances and interfirm knowledge transfer.Strategic Management Journal, 17: 77–91.

Nickerson, J. A., & Zenger, T. R. 2004. A knowledge-based theory of the firm—The problem-solving per-spective. Organization Science, 15: 617–632.

Noorderhaven, N. G. 2004. Hermeneutic methodologyand international business research. In R. Marschan-Piekkari & C. Welch (Eds.), Handbook of qualitativeresearch methods for international business: 84–104. Cheltenham, U.K., & Northampton, MA: Ed-ward Elgar.

Nooteboom, B. 2002. Trust: Forms, foundations, func-tions, failures and figures. Cheltenham, U.K., &Northampton, MA: Edward Elgar.

Nooteboom, B, Berger, H., & Noorderhaven, N.G. 1997.Effects of trust and governance on relational risk.Academy of Management Journal, 40: 308–338.

Nunnally, J. 1978. Psychometric theory. New York:McGraw-Hill.

Olk, P. 2002. Evaluating strategic alliance performance. InF. Contractor & P. Lorange (Eds.), Cooperative strate-gies and alliances: 119–143. Amsterdam: Pergamon.

Osborn, R. N., & Baughn, C. C. 1990. Forms of interor-


ganizational governance for multinational alliances.Academy of Management Journal, 33: 503–519.

Oxley, J. E. 1999. Institutional environment and themechanisms of governance: The impact of intellec-tual property protection on the structure of inter-firm alliances. Journal of Economic Behavior andOrganization, 38: 283–309.

Oxley, J. E., & Sampson, R. C. 2004. The scope andgovernance of international R&D alliances. StrategicManagement Journal, 25: 723–750.

Park, C. V., & Sheath, J. N. 1975. Impact of prior familiarityand cognitive complexity on information processingrules. In M. L. Ray & S. Ward (Eds.), Communicatingwith consumers: The information processing ap-proach: 71–78. Beverly Hills, CA: Sage.

Park, S. H., & Russo, M. 1996. When competition eclipsescooperation: An event history analysis of alliancefailure. Management Science, 42: 875–890.

Park, S. H., & Ungson, G. R. 2001. Interfirm rivalry andmanagerial complexity: A conceptual framework ofalliance failure. Organization Science, 12: 37–53.

Parkhe, A. 1991. Interfirm diversity, organizational learn-ing, and longevity in global strategic alliances. Jour-nal of International Business Studies, 22: 579–601.

Parkhe, A. 1993. Strategic alliance structuring: A gametheoretic and transaction cost examination of inter-firm cooperation. Academy of Management Jour-nal, 36: 794–829.

Parkhe, A. 1998. Building trust in international alliances.Journal of World Business, 33: 417–437.

Pfeffer, J., & Nowak, P. 1976. Joint ventures and interor-ganizational interdependence. Administrative Sci-ence Quarterly, 21: 398–419.

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsa-koff, N. P. 2003. Common method biases in behav-ioral research: A critical review of the literature andrecommended remedies. Journal of Applied Psy-chology, 88: 879–903.

Poppo, L., & Zenger, T. 2002. Do formal contracts and rela-tional governance function as substitutes or comple-ments? Strategic Management Journal, 23: 707–726.

Pothukuchi, V., Damanpour, F., Choi, J., Chen, C. C., &Park, S. H. 2002. National and organizational culturedifferences and international joint venture perfor-mance. Journal of International Business Studies.33(2): 243–265.

Powell, W. W. 1990. Neither market nor hierarchy: Net-work forms of organizations. In B. M. Staw & L. L.Cummings (Eds.), Research in organizational be-havior, vol. 12: 295–336. Greenwich, CT: JAI Press.

Robertson, R. D. 1980. Small group decision making: Theuncertain role of information in reducing uncer-tainty. Political Behavior, 2(2): 163–188.

Robins, J. A., Tallman, S., & Lindquist, K. F. 2002. Au-tonomy and dependence of international coopera-

tive ventures: An exploration of the strategic perfor-mance of U.S. ventures in Mexico. StrategicManagement Journal, 23: 881–901.

Sako, M. 1991. The role of “trust” in Japanese buyer-supplier relationships. Ricerche Economiche, 45:449–474.

Sako, M., & Helper, S. 1998. Determinants of trust insupplier relations: Evidence from the automotive in-dustry in Japan and the United States. Journal ofEconomic Behavior and Organization, 34:387–417.

Sampson, R. C. 2004. The cost of misaligned governancein R&D alliances. Journal of Law, Economics andOrganization, 20: 484–526.

Saxton, T. 1997. The effects of partner and relationshipcharacteristics on alliance outcomes. Academy ofManagement Journal, 40: 443–460.

Schwenk, C. R. 1984. Cognitive simplification processesin strategic decision-making. Strategic Manage-ment Journal, 5: 111–128.

Shaver, J. M. 1998. Accounting for endogeneity whenassessing strategy performance: Does entry modechoice affect FDI survival? Management Science,44: 571–585.

Simonin, B. L. 1999. Ambiguity and the process ofknowledge transfer in strategic alliances. StrategicManagement Journal, 20: 595–624.

Simons, T., Pelled, L. H., & Smith, K. A. 1999. Makinguse of difference: Diversity, debate and decisioncomprehensiveness in top management teams.Academy of Management Journal, 42: 662–673.

Snyder, N. H., & Glueck, W. F. 1982. Can environmentalvolatility be measured objectively? Academy ofManagement Journal, 25: 185–192.

Steensma, H. K., Tihanyi, L., Lyles, M. A., Dhanaraj, C.2005. The evolving value of foreign partnerships intransitioning economies. Academy of ManagementJournal, 48: 213–235.

Stinchcombe, A. L. 1985. Contracts as hierarchical doc-uments. In A. Stinchcombe & C. Heimer (Eds.), Or-ganization theory and project management: 121–171. Bergen: Norwegian University Press.

Sutcliffe, K. M., & Zaheer, A. 1998. Uncertainty in thetransaction environment: An empirical test. Strate-gic Management Journal, 19: 1–23.

Szulanski, G., Cappetta, R., & Jensen, R. J. 2004. Whenand how trustworthiness matters: Knowledge trans-fer and the moderating effect of causal ambiguity.Organization Science, 15: 600–613.

Thompson, J. D. 1967. Organizations in action: Socialscience bases of administration. New York:McGraw-Hill.

Tourangeau, R., Rips, L. J., & Rasinski, K. 2000. Thepsychology of survey response. Cambridge, U.K.:Cambridge University Press.


Tsai, W., & Ghoshal, S. 1998. Social capital and the valuecreation: The role of intrafirm networks. Academyof Management Journal, 41: 464–476.

Uzzi, B. 1997. Social structure and competition in inter-firm networks: The paradox of embeddedness. Ad-ministrative Science Quarterly, 42: 35–67.

Webb, G. 1996. Trust and crises. In R. M. Kramer & T. R. Tyler(Eds.), Trust in organizations: Frontiers of theory andresearch: 288–301. Thousand Oaks, CA: Sage.

Wholey, D. R., & Brittain, J. 1989. Characterizing envi-ronmental variation. Academy of ManagementJournal, 32: 867–882.

Williamson, O. E. 1985. The economic institutions ofcapitalism. New York: Free Press.

Williamson, O. E. 1991. Comparative economic organiza-tion: The analysis of discrete structural alternatives.Administrative Science Quarterly, 36: 269–296.

Yan, A., & Zeng, Z. 1999. International joint venture insta-bility: A critique of previous research, a reconceptual-ization, and directions for future research. Journal ofInternational Business Studies, 30: 397–414.

Zaheer, A., McEvily, B., & Perrone, V. 1998. Does trustmatter? Exploring the effects of interorganizationaland interpersonal trust on performance. Organiza-tion Science, 9: 141–159.

Zand, D. E. 1972. Trust and managerial problem solving.Administrative Science Quarterly, 17: 229–239.

Zollo, M., Reuer, J. J., & Singh, H. 2002. Inter-organiza-tional routines and performance in strategic alli-ances. Organization Science, 13: 701–713.

APPENDIX ASurvey Items

Alliance Performance (! ! .90; 1 ! “strongly disagree”;5 ! “strongly agree”)

1. The objectives for which the collaboration was estab-lished are being met.

2. Our firm is satisfied with the financial performance ofthe collaboration.

3. Our foreign partner firm seems to be satisfied with thefinancial performance of the collaboration.

4. Our firm is satisfied with the overall performance ofthe collaboration.

5. Our foreign partner firm seems to be satisfied with theoverall performance of the collaboration.

Trust (! ! .85; 1 ! “strongly disagree”; 5 ! “stronglyagree”)

1. Sometimes our foreign partner changes facts slightlyin order to get what they want. (reverse-coded)

2. Our foreign partner has promised to do things withoutactually doing them later. (reverse-coded)

3. Our foreign partner has given us truthful and valuableinformation even when it did not form part of thecontract.

4. Our firm is generally doubtful of the information pro-vided to us by our foreign partner. (reverse-coded)

5. Our foreign partner firm is generally doubtful of theinformation we provide them. (reverse-coded)

Quality of Information Exchanged (! ! .80; 1 !“strongly disagree”; 5 ! “strongly agree” unless other-wise indicated)

1. Our foreign partner firm has provided relevant infor-mation whenever we asked them for it.

2. We are promptly notified by our foreign partner when-ever any major change occurs at their firm.

3. We get clear information about the plans of our foreignpartner concerning the collaboration well in advance.

4. How often do senior managers from your firm com-municate with their counterparts in the foreign part-ner firm? (1 ! “daily”; 5 ! “once a month or less”)

5. How often do senior and middle managers in your com-pany make business trips to your foreign partner firm?(1 ! “twice a month or more”; 5 ! “once a year or less”).

Alliance Duration

We asked the year in which an alliance was formedand used it to calculate duration.

Interdependence

Indicate which of the following describe the valuecreation rationales of the alliance. (The questionnaireindicated that the respondents could check more thanone rationale of the nine listed in the text.)

APPENDIX B

Controlling for the Endogeneityof the Equity/Nonequity Governance Choice

Motivation

Our sample included both equity and nonequity alli-ances. Prior research states that managers consciouslyselect from these two alliance governance modes the onethat is likely to enhance the performance of their partic-ular alliance (e.g., Gulati, 1995; Sampson, 2004). There-fore, it is likely that alliance performance depends onunobservable characteristics that determine alliance gov-ernance choices. To account for this potential endogene-ity, we used Heckman’s (1979) two-stage technique (seeHamilton & Nickerson, 2003; Shaver, 1998).

Implementation and Results

Using Heckman’s two-stage procedure, we first esti-mated a probit model of alliance governance choice (non-equity alliance vs. equity alliance) and generated the in-verse Mills ratio. We then estimated the allianceperformance model using the inverse Mills ratio from thefirst stage as a control variable. Incorporating this correctionterm into the second-stage model yields unbiased estimatesof the predictors of alliance performance (Greene, 1997). In


our first-stage model, we used two variables suggested inprior research: the tenure of respondent (e.g., Poppo &Zenger, 2002) and alliance formation before liberalization(e.g., Hamilton & Nickerson, 2003). Firstly, a respondent’sexperience within her or his firm was a proxy for corporateknowledge about interorganizational relationships (Poppo& Zenger, 2002). Respondent tenure was likely to influencethe choice of alliance governance mode, all the more so aseach of our respondents held a position of high responsi-bility. Secondly, before a major liberalization of foreigntrade relationships through legislation in 1991, the Indiangovernment exerted pressure on foreign firms to contributesome form of equity to the Indian economy (Balasubra-manyam, 2003; Bowonder & Richardson, 2000). Hence, al-liances formed before liberalization were more likely to beequity alliances, though not all were. There is no theoreticalbasis to link either of these variables directly with allianceperformance (Hamilton & Nickerson, 2003; Poppo &Zenger, 2002). Other variables from our work were alsoincluded in the first-stage model. Table B1 reports the first-stage results. Preference for an equity alliance is negativelyrelated to tenure of respondent (p ! .01) and marginallyrelated to environmental unpredictability (p ! .10), andpositively related to investment size (p ! .05) and allianceformed before liberalization (p ! .01). Because correctingfor self-selection of governance mode is important in the-ory, we included the inverse Mills ratio from this modelinto the second-stage models. The ratio’s nonsignificance inthe second stage indicated that the potential endogeneity ofgovernance mode was not adversely affecting our estimatedresults about alliance performance. The hypothesized re-

sults are similar with and without the inclusion of theinverse Mills ratio.

Rekha Krishnan ([email protected]) is an assistant professorof international business in the Faculty of Business Ad-ministration at Simon Fraser University. She receivedher Ph.D. in business from the CentER Graduate Schoolat Tilburg University. Her research interests include stra-tegic alliances, trust, and emerging markets.

Xavier Martin ([email protected]) is a professor of strategyand international business in the Faculty of Economicsand Business Administration at Tilburg University and afellow of the CentER for Economic Research. He receivedhis Ph.D. in business administration from the GraduateSchool of Business Administration at the University ofMichigan. His research interests include corporate strat-egies for international expansion, the dynamics of interor-ganizational relationships and alliances, knowledgetransfer and accumulation strategies, innovation andnew-product introduction strategies, and the perfor-mance implications of these phenomena.

Niels G. Noorderhaven ([email protected]) is aprofessor of international management in the Faculty ofEconomics and Business Administration and the head ofthe Department of Organization and Strategy at TilburgUniversity. He received his Ph.D. in business adminis-tration from Groningen University. His research focuseson interorganizational cooperation, with special atten-tion to issues related to culture and trust.

Please note: This article continues with Appendix C,on the following page.

TABLE B1Results of Probit Analysis for the First-Stage

Governance Modela

Variable Model 1

Intercept 1.86* (0.75)Tenure of respondent "0.05** (0.02)Position of respondent 0.08 (0.25)Investment size 0.39* (0.19)Cultural distance "0.13 (0.30)Local partner size "0.14 (0.09)Interdependence 1.06 (0.68)Interpartner competition 0.02 (0.21)Environmental instability 2.06 (2.42)Environmental unpredictability "0.68† (0.38)Alliance formed before liberalization 0.05** (0.02)n 126!2 160.66***

a The dependent variable was set to 0 for nonequity alliancesand 1 for equity alliances. Standard errors are in parentheses.

† p ! .10* p ! .05

** p ! .01*** p ! .001


APPENDIX CRemedies Undertaken against Common Method Bias and Single Respondent Bias

Remedy and Rationale Implementation

Procedural

Protecting respondent anonymity. This technique decreasesrespondents’ tendency to make socially desirable responses and/orbe acquiescent or lenient when crafting their responses (Podsakoff etal., 2003: 888).

Our cover letter assured respondents complete anonymity.

Reducing item ambiguity. Careful attention to the wording of itemshelps reduce item ambiguity (Tourangeau, Rips, & Rasinski, 2000).

We were careful to avoid vague concepts and double-barreled questions and to keep questions simple, all ofwhich reduce item ambiguity (Tourangeau, Rips, &Rasinski, 2000). We pretested the survey with Indianmanagers, which helped us identify and replace a fewambiguous words.

Separating scale items. Reduces the likelihood of respondents guessingthe relationship between predictor and criterion variables andconsciously matching their responses to the two measures (Parkhe,1993).

In our questionnaire, the trust and alliance performanceitems were placed far apart from each other—about 50percent of the relevant questionnaire pages apart. Itemswere not grouped by variable, and the variables were notlabeled on the basis of the reported constructs (trust, etc.).

Data from different sources. Measures based on different sources helpcontrol common method or single-informant bias (Podsakoff et al.,2003).

We obtained data on most (three-quarters) of the moderatorvariables and on several control variables from archivalsources.

StatisticalPartial correlation adjustment. If a variable can be identified that is

theoretically unrelated to at least one other variable in a study,preferably the dependent variable, then it can be used as a markervariable in controlling for common method variance (Lindell &Whitney, 2001).

We used tenure of the respondent as the marker variable, asit was theoretically unrelated to many other variables andespecially to alliance performance. The use of a variablesuch as this is consistent with existing research (e.g.,Griffith & Myers, 2005). All our significant zero-ordercorrelations remained significant after the partialcorrelation adjustment, suggesting that common methodbias was not a serious problem in our study (Lindell &Whitney, 2001).

Triangulation using archival sources. Triangulating survey data withdata from secondary sources is often used to check the convergentvalidity of a construct (Dhanaraj, Lyles, Steensma, & Tihanyi, 2004;Keats & Hitt, 1988; Parkhe, 1993).

The correlation between the subjective and secondaryalliance performance measures (i.e., return on capitalemployed) available for 35 equity alliances was highlysignificant (r ! .38, p ! 0.02). Note that our measure ofalliance performance had a general focus, but return oncapital was a focused indicator, preventing a strongercorrelation.

Triangulation using field interviews. Interview-based data can becoded to establish the reliability and validity of variables. We usedthe interview notes to validate the trust variable since our interviewfocused mainly on the quality of the relationship between partners,thus providing usable information about trust. Because theinterview notes did not provide enough useful data on how well analliance performed, we used archival data instead to validate thealliance performance construct (see “Triangulation using archivalsources”).

We used interview data available for ten alliances tovalidate the trust measure. Two independent coderscategorized the interview responses using three-pointscales to indicate the extent to which interorganizationaltrust existed in a relationship. We did not exceed three-point categorization for ease of interpretation of theinterview data (e.g., Lau & Woodman, 1995; Lee, Mitchell,Wise, & Fireman, 1996). The correlation between the trustscale obtained from the survey and the interview notescoded by the independent raters was .70 (p " .05). Nodiscrepancy was noted regarding variable content.

Harman’s one-factor test. If a substantial amount of common methodbias exists in data, a single or general factor that accounts for mostof the variance will emerge when all the variables are enteredtogether (Podsakoff et al., 2003).

An unrotated principal components factor analysis on allthe variables measured using the survey instrumentrevealed four factors with eigenvalues greater than 1.0,which together accounted for 59 percent of the totalvariance; also, the first (largest) factor did not account fora majority of the variance (19.71%).

Significance of the interaction terms. A pattern of significantinteraction terms suggests that results are unlikely to have resultedfrom single-informant bias (Kotabe, Martin, & Domoto, 2003).

All our interaction effects were significant, including theinteraction that involved two survey-based items. Supportfor interaction hypotheses is unlikely to be an artifact ofsingle-respondent bias, as it is implausible thatrespondents will consciously theorize moderatedrelationships when responding to a survey (Kotabe et al.,2003).


WHEN DOES TRUST MATTER TO ALLIANCE PERFORMANCE?

Documents