Upload
johnalis22
View
214
Download
0
Embed Size (px)
DESCRIPTION
strategy
Citation preview
A Review of Antecedents of InternationalStrategic Alliance Performance:
Synthesized Evidence and New Directionsfor Core Constructsijmr_335 66..85
Jeppe ChristoffersenDepartment of Accounting and Auditing, Copenhagen Business School, Solbjerg Plads 3,
DK-2000 Frederiksberg, DenmarkEmail: [email protected]
This paper provides a systematic review of 165 empirical studies on the antecedents ofperformance in international strategic alliances. It provides the most detailed displayof definitions, rationales, measures and findings currently available. Hence, this state-of-the art literature review creates an accessible pool of knowledge that is highly relevantfor future research on international strategic alliances. Further, it draws on this know-ledge pool to build a model which highlights the quite different rationales advancedby researchers to explain associations between the antecedents and performance. Themodel makes the different rationales explicit and will aid researchers in identifying teststhat can be performed to examine the links between antecedents and performance aswell as the mechanisms through which such associations operate.Finally, the synthesizedevidence is used to suggest that researchers should give increased attention to achievingcongruence between measures of antecedents and performance.
Introduction
As strategic alliances in general – and internationalstrategic alliances (ISAs) in particular – have prolif-erated over the past half century, so research hasincreased knowledge about causes and consequencesof this organizational design. The earliest studiesinclude Friedmann and Kalmanoff (1961), Tomlin-son (1970) and Franko (1971), and research in alli-ances and particularly ISAs has subsequentlyexpanded dramatically, initially focusing primarilyon broad aspects (such as motivation for collaborat-ing), on joint ventures (JVs) and on their popularityas a mode of entry into new countries (e.g. Janger1980; Killing 1983). Sparked to an extent by studiesby Harrigan (1988) and by Kogut (1988b) in aManagement International Review special issue onco-operative ventures, later studies increasingly
shifted their focus to include non-equity alliances,and towards performance and its antecedents.
The large number of empirical articles has beenmatched by a fair number of literature reviews,the earlier of which – including Beamish (1985),Geringer and Hébert (1989), Kogut (1988a) andParkhe (1993a) – reflected the content of the earlystudies, considering primarily the extent of andmotives for JV formation. Where performance andits antecedents were considered, the performancemeasures generally used were stability and manage-rial satisfaction, and the antecedents suggested weremainly numbers of partners (two vs. more than two),JV location (developed vs. developing countries) andcontrol (dominant vs. shared). As Parkhe (1993a,p. 230) noted, the focus was on constructs reflectedby data made available through ‘ “hard” data sources[which] are unlikely to capture the soft core
bs_bs_banner
International Journal of Management Reviews, Vol. 15, 66–85 (2013)DOI: 10.1111/j.1468-2370.2012.00335.x
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.Published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA02148, USA
concepts’ – such as trust, forbearance, reciprocityand opportunism – which Parkhe argued were centralto understanding inter-firm co-operation. As thenumber of articles studying alliance performancegrew through the 1990s, so did the focus of moreliterature reviews, so that Robson et al. (2002) wereable to provide a very informative vote-count reviewof 91 studies focusing on factors influencing interna-tional joint venture (IJV) performance. Their mainconclusion was that there were confused findings formost hypothesized performance drivers, but behav-ioural factors such as trust, commitment and conflicthad significant effects. These findings were sup-ported by Reus and Ritchie (2004), whose compre-hensive review of 194 empirical and conceptual IJVarticles focused both on performance and on IJVresearch in general. In 2006, Robson et al. (2006)provided support for Robson et al.’s (2002) claimsabout the key role played by behavioural factors, intheir vote-count analysis of 41 ISA articles focusingon five behavioural constructs – trust, commitment,co-operation, communication and conflict/conflictresolution – and finding general support for hypoth-eses relating them to ISA performance.
More recently, five very different but complemen-tary reviews have appeared. Culpan (2009), writingin the inaugural issue of the International Journal ofStrategic Business Alliances, presents a narrativereview which focuses on a wide range of researchtopics and paradigms. Nippa et al. (2007) present anarrative review based on 41 articles that synthesizesexisting concepts into a framework of IJV successfactors under five broad headings: parent attributes,parent fit, relationship management, IJV governanceand external environment. Beamish and Lupton(2009) provide a narrative review of the 74 mostcited JV performance articles published in the previ-ous 25 years, which focuses on identifying promi-nent academic discussions and outlining practicalimplications for managing JVs. Finally, while theother recent reviews tend to be relatively general anddiscuss groups of constructs under broad headings,the reviews by Reus and Rottig (2009) and Ren et al.(2009) focus on individual constructs. provide ameta-analytical review based on 61 IJV performancearticles and provide a path model in which conflictmediates the association between cultural distance,hierarchical control and commitment on the onehand and performance on the other. While they findoverall support for this model, they note ‘only 19percent of the variance in IJV performance isaccounted for, indicating a need to explore other
underexplored variables’ (Reus and Rottig 2009,p. 607). That is exactly what Ren et al. (2009) do,by exploring a larger number of potential perfor-mance antecedents: bargaining power, commitment,control, trust, conflict, co-operation, cultural dis-tance, justice, goal compatibility and conflict resolu-tion mechanisms. However, given the greater numberof constructs spread out over a relatively modestsample size (54 articles), the empirical backing foreach construct is considerably lower than is the casein those considered by Reus and Rottig (2009).
This paper, which provides a review of empiricalresearch pertaining to antecedents of ISA perform-ance, builds on these recent reviews and particularlythe two last-quoted: specifically, it also reviews par-ticular constructs (rather than broader concepts andissues) and finally presents them in a model thatproposes links between individual constructs, ratherthan a framework suggesting relations between thebroader issues, thus combining the fairly broad scopeof constructs found in Ren et al. 2009 with the sub-stantial empirical backing found in Reus and Rottig(2009). I draw a sample of 165 cross-sectional studies– not only of JVs but also of non-equity alliances – toidentify the antecedents of ISA performance.
I do not present more antecedents than Renet al. (2009), but whereas their model includes (forinstance) conflict resolution mechanisms and bar-gaining power, which are only considered (respec-tively) in one and three of their sample articles, theantecedents suggested in my model are all consid-ered in five or more articles, which I believe allowsme to claim that they are core constructs in the lit-erature, and to identify which rationales linking con-structs to performance can be considered dominantand which alternative. Finally, this greater sampleprovides a more extensive pool of findings, whichgives me greater security in assessing empiricalsupport for theoretical arguments. Compared withReus and Rottig (2009), I do not present moreempirical backing for each antecedent included inmy model. However, while they focus on four ante-cedents for which they are able to find large quanti-ties of empirical evidence – which is necessary fortheir meta-analytical approach – I suggest manymore antecedents, responding to their suggestion thatfuture models should include more antecedents.Considering these differences, the first contributionof this review is to provide a comprehensive refer-ence list of core constructs in the ISA literature,supported by definitions, the dominant rationalesused in linking them to ISA performance, the ways in
Antecedents of ISA Performance 67
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
which they have been measured and, finally, the find-ings: this level of detail is unprecedented in previousreviews.
The detailed review of dominant rationales linkingconstructs to performance reveals that, in most cases,researchers suggest several mechanisms throughwhich they affect performance, even in some casesshowing how mechanisms may produce opposingeffects (e.g. arguing that national cultural distanceaffects performance positively (via complementaryresources) against the more common argument ofa negative effect associated with less trust andco-operation and greater conflict). My second con-tribution is to bring attention to the different roleswhich researchers suggest constructs play, elaborat-ing on individual constructs as they are reviewed andthen integrating them into a model. This model isthus the first to demonstrate explicitly that the roleswhich researchers presume core constructs playare often very different: using it may enable futureresearch to enhance knowledge about whether agiven construct can predict performance, and also viawhich mechanism(s).
Sample selection
The first step in preparing this paper consisted ofidentifying the relevant literature. As IJMR publishesauthoritative literature reviews in the field of ‘busi-ness and management’ (Armstrong and Wilkinson2007, p. 81), I first identified journals from the ‘busi-ness’ and ‘management’ subject categories of theSocial Sciences Citation Index (SSCI) (also used byArmstrong and Wilkinson 2007). For the 25 journalswith the highest one-year impact factor, I calculatedthe proportion of journal issues in the last 20 yearsfor which there were bibliographic records in theEBSCO article database and found the percentagewas 98%, and 95% over the last 30 years: this stronglevel of coverage of the relevant domains allowed meto choose it as the primary database for my literaturesearch. I made searches of the titles, abstracts, key-words and subject terms of all articles that appearedin all peer-reviewed scholarly journals for whichEBSCO had bibliographic records, identifying allthose which included at least one ISA keyword andone performance keyword.1 I identified relevant top
journals through SSCI listings and the work of Pisani(2009) and DuBois and Reeb (2000), and for thesetop journals supplemented my EBSCO searches withsearches in other databases for journal volumes notcovered by EBSCO. After sorting the articles thusidentified based first on titles and abstracts and sub-sequently on their full text, I arrived at a total of 156articles that met all inclusion criteria (see below):while almost all were identified through EBSCO,many had to be obtained via other databases or fromthe relevant journal’s hard copies.
To assess whether my search process missed asubstantial number or sector of relevant articles, Icross-checked with earlier reviews and with thecontent of the selected articles. I found only nineadditional articles to be included, and so consideredthat, while a few relevant studies may have beenmissed, my search process had been sufficiently com-prehensive to provide a fairly representative sampleof the literature. (Appendix S1 (available in elec-tronic version only) lists all articles in the sample.)
All 165 articles in the final sample met the follow-ing inclusion criteria: each was empirical and hadbeen published in a peer-reviewed journal; eachclearly hypothesized and statistically tested relation-ships between at least one antecedent variable and atleast one performance measure as dependent vari-able. Those hypotheses (and the samples used to testthem) only considered ISAs. This means that Iexcluded articles which only tested hypotheses onmixed samples (e.g. of wholly owned subsidiaries aswell as ISAs), which made it impossible to determinerelationships for the sub-sample of ISAs. However,this criterion also meant that I did include articleswhich, at some point, tested relationships for the ISAsub-sample of a sample containing a mix of ISAsand wholly owned subsidiaries.
In a recent review, Culpan (2009) defines alliancesas ‘[c]ollaboration[s] of multiple firms (at least two)
1Relevant keywords were identified by going through everyissue of 20 years of research in Academy of ManagementJournal, Journal of International Business Studies and
Strategic Management Journal, where (as expected) manyrelevant articles were found. The keywords used for alliancesearches were: joint venture, joint ventures, JV, JVs, JV’s,IJV, IJVs, IJV’s, alliance, alliances, SA, SAs, SA’s, ISA,ISAs, ISA’s, and those for performance were performance,perform, performs, outcome, outcomes, efficiency, efficient,productivity, productive, profitability, profitable, profits,profit, competitiveness, competitive, termination, terminate,dissolution, dissolves, mortality, mortal, longevity, stability,stable, duration, endure, endures, result, result, value,satisfaction, satisfied, fulfill, fulfills, fulfillment, achieve,achieves, achievement, goal, goals, objective, objectives,payoff, pay off, payoffs, pays off, benefit, benefits, success,succeed, fail, failure, effect, effects, CAR, return.
68 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
involving their long-term commitment of resourcesto relationships that would serve strategic goalsof partners’ (Culpan 2009, p. 7). Where alliancepartners are based in different locations, or wheretheir alliance operates in a different country fromtheir headquarters, the alliance can be consideredinternational. This paper therefore defines ISAs as‘collaborations involving the long-term commit-ment of resources to relationships that serve thestrategic goals of two or more partners based indifferent countries, or whose collaborative opera-tions take place in a different country to theirheadquarters’.
In terms of relevance to the inclusion criteria, themost challenging consideration was how to deter-mine which measures were measures of allianceperformance and which were not. To guide this deter-mination, I defined alliance performance as the netpresent value that partners obtain from participatingin the alliance. While no measure exactly reflects thisconstruct, some are clearly better than others – andsome can be specifically ruled out. For instance, fol-lowing studies that have discussed measuring strate-gic alliance performance (e.g. Ariño 2003; Geringerand Hébert 1991; Hatfield et al. 1998), I focusedonly on measures that are seldom used as antecedentvariables, as such variables are presumably far fromreflecting all benefits and costs. I therefore excludedarticles that only used measures reflecting relationalaspects – such as the extent of co-operation betweenpartners (e.g. Robson and Katsikeas 2005) – orrelated to knowledge acquisition or transfer (e.g.Kim and Inkpen 2005; Simonin 2004). Furthermore,articles were only included which used general per-formance measures applicable to virtually all indus-tries: thus, those using only specific measures such aspatent performance were excluded (e.g. Nooteboomet al. 2007).
Findings: ISA performance measuresand antecedents
This section first presents five types of performancemeasures used by researchers, and then brings inconstructs that five or more studies have suggested asbeing direct antecedents of alliance performance (i.e.not moderated or moderating). Definitions, research-ers’ rationales for linking the constructs to perform-ance, the ways in which the constructs have beenmeasured and, finally, the findings of the samplestudies are presented. As the review proceeds, these
rationales and findings are used to develop a modelthat will organize the constructs that it considers.
Knowledge acquisition is often used as a measureof performance (even to the extent that Ren et al.(2009) treat it as such in their review), so I make onedeviation from the rule that all constructs suggestedas antecedents in five or more studies are included,and leave knowledge acquisition in a ‘grey zone’ – inthat it neither clearly constitutes performance norclearly constitutes an antecedent to performance –which reflects its position in the literature. (For anexcellent review of research on knowledge in thecontext of strategic alliances, see Meier (2011).)
Performance measures
This sub-section presents five types of performancemeasures used in the alliance literature, defines them,explains how they are obtained and what they reflect,and finally assesses their appropriateness under dif-ferent circumstances.
Subjective measures. The sample contains 102studies employing subjective performance measures,obtained by soliciting assessments from alliance orpartner firm managers. In their work on differencesand similarities between performance measures,Ariño (2003) and Geringer and Hébert (1991)divide subjective measures into measures of overallsatisfaction and of achievement of partner goals/objectives. However, a review of the subjective per-formance measures in this sample reveals that, whilethese categories predominate, many subjective per-formance measures cannot be assigned to them.Some measures contain several items and thuscombine items from both categories (e.g. Krishnanet al. 2006); some reflect satisfaction not only withoverall performance, but also with very specificaspects such as sales, market share, technologydevelopment, customer service and reputation (e.g.Luo et al. 2001). Thus, since subjective measurescannot all be assigned to one of the two categories(overall satisfaction and goal achievement), I do notmake that distinction at all, but assign all such per-formance measures to the same category, labelled‘subjective’.
Subjective measures have been widely discussedin the literature, with their proponents arguingthat they are appropriate because they reflect allthe advantages that firms seek when partnering(Anderson 1990). But, of course, the results aresensitive to the questions asked – for example Ariño
Antecedents of ISA Performance 69
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
(2003) argue that goal fulfilment only measuresperformance compared with initial goals, whereassatisfaction with overall performance also takesemergent goals into consideration. Subjective per-formance measures also vary with the individualsassessing them, and this can be particularly problem-atic if antecedents are also measured via subjectivemeasures, since antecedent and performance meas-ures may vary concurrently merely as a consequenceof same-source variance, which may potentiallyaffect findings (Podsakoff and Organ 1986).
Stability. Thirty-one studies employ stabilitymeasures based on event history analyses of suchevents as exits from given markets, changes in equitydistribution between JV parents, contract changes,parent takeover, third-party takeover and alliance dis-solutions – some reflecting alliance mortality andothers less drastic changes. Many researchers sawsuch events (or short alliance duration before theiroccurrence) as indicators of poor performance, onthe assumption that curtailing/closing down an alli-ance’s operation indicated that parents/partners nolonger saw it as the most appropriate organizationmode (Lu and Beamish 2006).
Measures of this type (labelled ‘stability’) reflectdecisions taken by the alliance partners and thus theirassessment of how the alliance is performing. Aswith assessments gained via subjective measures,these implicit assessments (based on managers’actions) may encompass not just financial goals, butany goals that the partner companies might have.Compared with the subjective measures – whichreflect performance in specific aspects suggested bythe researcher – stability measures encompass thepartners’ own goals, rather than those assumed by theresearcher. And, unlike subjective measures, stabilitymeasures never involve concerns about same-sourcevariance.
Many researchers doubt, however, whether stabil-ity measures really reflect poor performance. Forinstance, partners selling an alliance do not necessar-ily do so because they think it is performing poorly –quite the opposite may be true, as a well-performingalliance may be an attractive acquisition target.In fact, alliances may have been intended from theoutset as temporary organizations providing futureoptions, so their sale does not imply they can be seenas failures (Bowman and Hurry 1993; Kogut 1991).
Accounting measures. Twenty-one of the sample’sstudies employ accounting measures, that is, per-
formance measures based on financial data which areused to compute growth percentages and financialratios so as to gauge performance in the financialdomain (e.g. Luo 2002a). Measures of this type(which I label ‘Accounting’) have been criticized asbeing unable to reflect performance in non-financialdomains (e.g. Gong et al. 2005; Pangarkar and Klein2004; Pothukuchi et al. 2002). This concern datesback to the early days of alliance research, and ledAnderson (1990) to suggest subjective measures asmore appropriate, since partner goals are not alwaysfinancial. However, accounting measures have theclear advantage over subjective measures of notgiving rise to concerns about same-source variance,and they can also be interpreted with more confi-dence than stability measures (Beamish and Lupton2009). For example, comparatively high profitabilitycan always be considered an indication of compara-tively good performance, while the sale of a IJV canreflect either good or poor performance, dependingon the circumstances.
Cumulative abnormal return (labelled CAR).Eighteen studies employ cumulative abnormalreturns (CAR) as performance measures, whichreflect how shareholders and potential shareholdersin the participating partners react to allianceformation (e.g. Merchant and Schendel 2000). Theassumption is that stock markets react at the time ofannouncement in a way that reflects shareholders’assessment of the extent to which the alliance willcreate or destroy shareholder value (and thus ofits contribution to overall performance) for a givenpartner. Koh and Venkatraman (1991) have examinedthe association between CAR and managementassessments of achievement of five key objectivesand find ‘support for the construct validity and, espe-cially, the convergent validity of the market-basedmeasures’ (p. 881). However, Pearce and Hatfield(2002) argue that strategy research is hampered bythe fact that the CAR measure actually representsan assessment of partners’ strategy choice ratherthan of actual performance, which relates to thehypothesis – supported by Madhavan and Prescott’s(1995) empirical findings – that investors cannotalways predict performance impacts.
External evaluation. Finally, three studies usedevaluations made by parties outside the alliances. Liet al. (2008) use a non-profit organization’s objectiveranking of the import achievements of various alli-ances, and Griffith et al. (1998) and Hu and Chen
70 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
(1996) use whether JVs have been honoured bythe China Association of Enterprises with ForeignInvestment. I label such measures ‘External’2 – asthey have only very limited use, I do not discuss themfurther in this paper.
Antecedents of performance
This sub-section presents definitions, researchers’rationales for linking constructs to performance, theways in which the constructs are measured and,finally, the findings for each construct. Table 1presents a summary of these findings3 and Figure 1 asynthesizing model, which integrates the theoreticalarguments as well the empirical findings of the rel-evant articles, and which is developed as the studyproceeds, as the constructs are presented.4
Behavioural attributes. The first building block ofthe model is the group of behavioural constructs –trust, commitment, co-operation and conflict, whichare central in this model for two reasons. First, thedevelopment of hypotheses involving other con-structs often also invokes behavioural arguments,suggesting that researchers (at least implicitly) con-sider these behavioural constructs as mediating theassociation with performance. Second, the empiricalevidence associated with behavioural constructs sup-ports their relevance, making them candidates for asolid core around which a model can be built, align-ing with earlier reviews (particularly by Robsonet al. (2002)) which found that behavioural con-structs showed great promise and warranted attentionin future model-building efforts. Their centrality tomy model is illustrated by their grey shading in themodel depiction in Figure 1.
Commitment. Eleven articles in the sample addressthe performance implications of commitment, gener-ally referring to the partners’ intention to stay in the
relationship. Dimensions of commitment include cal-culative commitment, involving reasoning about therelative costs and benefits of the alliance, and affectivecommitment, involving partners’ identification withthe alliance and the other partner (Geyskens et al.1996; Voss et al. 2006). While (in this sample) onlyVoss et al. (2006) specifically account for this distinc-tion, most authors implicitly draw on one or bothdimension when developing their hypotheses. Whencalculative commitment exists, it is argued thatpartners weigh the costs and benefits of leaving, andwhen the latter are more than outweighed, partnersbecome more determined to work to make the alliancework. Where affective commitment is at play, whilethe effort exerted to make the alliance work isexpected to give rise to a positive effect, the originalmotivation for exerting that effort is seen as arisingfrom feelings of affection towards the partner, ratherthan from a cost/benefit assessment (Voss et al. 2006).The motivation to make the alliance work also entailssacrificing individual opportunities in pursuit ofthe common goals that can be achieved throughco-operation (Kwon 2008).
The variation in measures used is extensive. Themajority of researchers using the calculative commit-ment argument use measures consisting of partners’assessments of the extent of relation-specific invest-ments that they have made, which indicate highercosts of switching and thus of the intention to stay inthe relationship (Parkhe 1993b). Where the argu-ments are based on affective commitment, the meas-ures reflect the partners’ assessment of their sense ofbelonging and emotional attachment to both allianceand partner(s) (Voss et al. 2006). No matter whichmeasures are used, the results are fairly consistent insuggesting a positive relationship, and this seems tobe true regardless of whether commitment is meas-ured from the point of view of one or both partners.
Trust. Sixteen articles address the performanceimplications of trust, which has been defined as
the willingness of a party to be vulnerable to theactions of another party based on the expectationthat the other will perform a particular actionimportant to the trustor, irrespective of the ability tomonitor or control that other party. (Mayer et al.1995, p. 712)
Different dimensions of trust have been proposed,but are discussed explicitly in only one paper –Muthusamy et al. (2007) – where the authors inves-tigate the effect of ability-based, benevolence-based
2The number of studies employing the different typesof performance measures total 175, as a few of the 165employed more than one type. Examples are Lu and Xu(2006), who employ both sales growth (accounting) andsurvival (stability), and Luo (2001), who uses managers’assessment of satisfaction on ten different domains (subjec-tive) in combination with return on investment (accounting).3See Appendix S2 (available in electronic version only) for adetailed presentation of findings.4The size of the boxes varies, allowing the use of onlyvertical and horizontal lines, but reducing the visual com-plexity: as a result their relative sizes do not necessarilyreflect the relative importance of the constructs.
Antecedents of ISA Performance 71
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
Tabl
e1.
Sum
mar
yof
findi
ngs
Con
stru
ctN
o.of
arti
cles
Mea
sure
men
tof
ante
cede
nta
Mea
sure
men
tof
depe
nden
tva
riab
leH
ypot
hese
sR
esul
ts
+0
-+
0-
Beh
avio
ural
attr
ibut
esC
omm
itm
ent
11S
ubje
ctiv
e(1
9),
Obj
ectiv
e(2
)S
ubje
ctiv
e(1
4),A
ccou
ntin
g(5
),E
xter
nal
(2)
210
016
50
Tru
st16
Sub
ject
ive
(19)
Sub
ject
ive
(18)
,Acc
ount
ing
(1)
190
014
50
Co-
oper
atio
n5
Sub
ject
ive
(7)
Sub
ject
ive
(6),
Acc
ount
ing
(1)
70
05
20
Con
flic
t12
Sub
ject
ive
(12)
Sub
ject
ive
(10)
,S
tabi
lity
(2)
00
120
48
Dis
sim
ilar
itie
sS
ize
diss
imil
arit
iesb
9O
bjec
tive
(10)
Sub
ject
ive
(3),
Sta
bili
ty(5
),C
AR
(2)
30
90
100
Nat
iona
lcu
ltur
aldi
stan
ceIn
dex
ofdi
ffer
ence
son
cult
ural
dim
ensi
ons
12O
bjec
tive
(13)
Sub
ject
ive
(5),
Sta
bili
ty(5
),C
AR
(2),
Ext
erna
l(1
)1
012
45
4D
iffe
renc
eson
indi
vidu
alcu
ltur
aldi
men
sion
s4
Obj
ectiv
e(1
6)S
ubje
ctiv
e(6
),S
tabi
lity
(10)
03
130
133
Cou
ntry
/reg
ion
com
pari
sons
6O
bjec
tive
(17)
Sub
ject
ive
(7),
Sta
bili
ty(2
),A
ccou
ntin
g(1
),C
AR
(7)
00
174
310
Org
aniz
atio
nal
cult
ural
dist
ance
11S
ubje
ctiv
e(1
3)S
ubje
ctiv
e(1
1),
Sta
bili
ty(1
),A
ccou
ntin
g(1
)0
013
08
5R
elat
edne
ss8
Sub
ject
ive
(2),
Obj
ectiv
e(1
1)S
ubje
ctiv
e(2
),S
tabi
lity
(5),
Acc
ount
ing
(3),
CA
R(3
)13
00
85
0E
xper
ienc
eA
llia
nce
expe
rien
ce9
Obj
ectiv
e(1
3)S
ubje
ctiv
e(3
),S
tabi
lity
(7),
Acc
ount
ing
(1),
CA
R(2
)13
00
85
0In
tern
atio
nal
expe
rien
ce9
Obj
ectiv
e(1
4)S
ubje
ctiv
e(3
),S
tabi
lity
(8),
Acc
ount
ing
(1),
CA
R(2
)11
03
48
2In
tern
atio
nal
alli
ance
expe
rien
ce10
Sub
ject
ive
(1),
Obj
ectiv
e(1
2)S
ubje
ctiv
e(3
),S
tabi
lity
(3),
Acc
ount
ing
(1),
CA
R(6
)13
00
310
0P
rior
rela
tion
ship
s8
Sub
ject
ive
(1),
Obj
ectiv
e(8
)S
ubje
ctiv
e(3
),S
tabi
lity
(4),
Acc
ount
ing
(1),
CA
R(1
)9
00
35
1C
ontr
olD
omin
ant
part
ner
cont
rol
6S
ubje
ctiv
e(2
),O
bjec
tive
(5)
Sub
ject
ive
(3),
Sta
bili
ty(2
),C
AR
(1),
Ext
erna
l(1
)2
32
22
3D
omin
ant
cont
rol
byfo
reig
npa
rtne
rc7
Sub
ject
ive
(4),
Obj
ectiv
e(5
)S
ubje
ctiv
e(5
),S
tabi
lity
(1),
Acc
ount
ing
(2),
Ext
erna
l(1
)8
01
35
0D
omin
ant
cont
rol
byfo
cal
part
ner
8S
ubje
ctiv
e(1
0),
Obj
ectiv
e(4
)S
ubje
ctiv
e(1
2),
CA
R(2
)13
10
66
2
The
tabl
epr
esen
tsth
eco
nstr
ucts
ofin
tere
st,
the
num
ber
ofar
ticl
eshy
poth
esiz
ing
and
test
ing
adi
rect
line
arre
lati
onsh
ipw
ith
perf
orm
ance
,ho
wth
ean
tece
dent
has
been
mea
sure
d,ho
wth
ede
pend
ent
vari
able
has
been
mea
sure
d,th
ehy
poth
esiz
edin
flue
nce,
and
the
empi
rica
lfi
ndin
gs.
Man
yar
ticl
espr
esen
tre
sult
sfo
rnu
mer
ous
test
sof
the
sam
ehy
poth
eses
,ow
ing
toe.
g.di
ffer
ent
spec
ifica
tion
sin
clud
ing
diff
eren
tco
ntro
lva
riab
les.
The
refo
re,
one
unit
incr
ease
inth
enu
mbe
rsin
all
colu
mns
but
the
‘No.
ofst
udie
s’co
lum
nre
pres
ents
asu
mm
ary
ofco
effi
cien
tsfr
omte
sts
usin
gex
actly
iden
tica
lan
tece
dent
mea
sure
sin
com
bina
tion
wit
hde
pend
ent
vari
able
sfa
llin
gw
ithi
nth
esa
me
ofth
efi
vety
pes
for
perf
orm
ance
mea
sure
spr
esen
ted
abov
e.T
hus
ther
eis
anin
crea
sefo
rev
enm
inor
mea
sure
men
tdi
ffer
ence
sin
the
ante
cede
nts,
whe
reas
diff
eren
ces
inth
epe
rfor
man
cem
easu
res
only
impl
yin
crea
ses
inth
ese
num
bers
ifth
edi
ffer
ence
isla
rge
enou
ghfo
rth
em
easu
res
tobe
long
todi
ffer
ent
type
s.If
all
test
sw
ithi
na
sum
mar
yar
epo
sitiv
e(n
egat
ive)
and
sign
ifica
nt,
a10
%si
gnifi
canc
ele
vel
one
unit
isad
ded
toth
eco
lum
nin
dica
ted
by‘+
’(‘
-’).
Ifno
t,on
eun
itis
adde
dto
the
colu
mn
indi
cate
dby
‘0’.
a Sub
ject
ive
mea
sure
sof
ante
cede
nts
are
thos
esu
bjec
tivel
yas
sess
edby
are
pres
enta
tive
ofth
eal
lian
ceor
part
ner
com
pani
es,
whe
reas
obje
ctiv
em
easu
res
ofan
tece
dent
sar
eth
ose
base
don
arch
ival
info
rmat
ion,
but
also
ante
cede
nts
base
don
surv
eyin
form
atio
nin
thos
eca
ses
whe
reth
esu
rvey
info
rmat
ion
does
not
cons
ist
ofsu
bjec
tive
asse
ssm
ents
,bu
tra
ther
fair
lyob
ject
ive
ones
.b T
henu
mbe
rof
hypo
thes
esex
ceed
sth
enu
mbe
rof
resu
lts
beca
use
Bea
mis
han
dJu
ng(2
005)
use
the
sam
ere
sult
sto
test
hypo
thes
esw
ith
oppo
site
sign
s.c T
henu
mbe
rof
hypo
thes
esex
ceed
sth
enu
mbe
rof
resu
lts
beca
use
Hu
and
Che
n(1
996)
use
the
sam
ere
sult
sto
test
hypo
thes
esw
ith
oppo
site
sign
s.
72 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
and integrity-based trust, which refer, respectively,to each partners’ trust that the other has the requiredabilities, will act benevolently towards the trustor,and adheres to the moral and principles that thetrustor finds acceptable (Mayer et al. 1995; Muth-usamy et al. 2007). All the researchers who investi-gate the direct impact of inter-partner trust onperformance suggest a positive relationship, mostrelying on the arguments that trust reduces the risk ofopportunistic behaviour and supplements or replacesrigid control mechanisms, so reducing transactionscosts (Robson et al. 2008; Sarkar et al. 2001). Spe-cifically, these positive effects are the consequencesof the behaviours that trust engenders, which arecharacterized by little conflict (Krishnan et al. 2006)and much co-operation (Nielsen 2007). The overallresults are convincing (since almost all studies showa positive relationship), and it is worth noting thatthese findings are based on a range of different sub-jective measures of trust, from direct questions aboutthe extent of inter-partner trust (Sarkar et al. 2001) tomulti-item measures designed to uncover each of thedimensions separately (Muthusamy et al. 2007).
Co-operation. Five articles address the perform-ance implications of co-operation – that is, the actingtogether of people or groups in a co-ordinatedmanner to pursue shared or complementary goals(Argyle 1991). Heide and Miner (1992) suggestthat the dimensions of this construct are flexibility,information exchange, shared problem-solving and
restraint in the use of power. All articles in thesample posit positive effects, as co-operating reduceshazards of opportunism and lessens managerial andagency problems via more open information sharing,which reduces information asymmetry (Luo andPark 2004). This intuitively attractive notion receivesconsiderable support. Luo (2002b) shows that thepositive effects of co-operation are unabated even athigh values. In contrast to those applied to commit-ment and trust, the measures used for co-operationare well aligned with each other and with the overalldefinition of co-operation, and reflect alliance andpartner representatives’ assessments of the degree towhich partners make joint decisions on differentaspects of the JV, including strategic objectives andgoals (Luo 2002b).
Conflict. Twelve articles address issues of conflict,which the team conflict literature defines as ‘anawareness of the parties involved of discrepancies,incompatible wishes, or irreconcilable desires’ (Jehnand Mannix 2001, p. 238). While this definitionrefers to intra-organizational conflict, it aligns withunderstandings about partner conflicts in the articlesincluded in my sample. Jehn and Mannix (2001)further distinguish between task conflict, processconflict and relationship conflict, which refer, respec-tively, to disagreements over which activities andtasks should be performed, over who should performgiven tasks, and tension and emotional conflictbetween leading representatives. While none of the
Commitment
Performance
Conflict
Trust
Co-operation
National cultural distance
Organizational cultural distance
Relatedness
Alliance experience
International experience
Prior relationships
Dominant control
Dominant control by foreign partner
LEGEND:Suggested positive impact:Suggested negative impact:
Size dissimilarities
Figure 1. Model of ISA performance
Antecedents of ISA Performance 73
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
studies distinguishes explicitly between these differ-ent dimensions, they all implicitly refer to task andprocess conflicts by referring to disagreements overfunctional areas (Ding 1997), and to relationshipconflict by referring to negative feedback loopsbetween conflict and trust (Steensma and Lyles2000). Authors generally argue that conflicts affectperformance negatively, as they are likely to lead tomisunderstandings, distrust and anxiety and so resultin less than efficient integration of activities, whilethe effort required to unravel them may take up timeneeded for important decisions (Steensma and Lyles2000). The measures used to reflect conflict includeJV and partner representatives’ assessments, andgenerally fall into two types, one mainly related totask and process conflict – through items reflectingthe frequency of disagreement in various functionalareas – and the other to relationship conflicts,through items reflecting the extent of cultural misun-derstandings, inter-partner distrust, conflicting goalsand personality conflicts. Regardless of the measuresused, the results consistently suggest that conflict hasa negative influence on performance.
Implications for the model. In line with research-ers’ dominant rationales reflected above, one ‘posi-tive impact’ arrow in Figure 1 points directly fromcommitment to performance, which represents theincreased effort the parties are willing to make toensure that the alliance works (Voss et al. 2006). Apositive arrow points to co-operation, reflecting thepursuit of common rather than individual benefits(Kwon 2008), and (for the same reason) anotherpositive arrow points from trust to co-operation(Nielsen 2007), which is supplemented by a ‘nega-tive impact’ arrow, suggesting (as proposed above)that trust reduces conflict (Krishnan et al. 2006).The notion illustrated – that co-operation and conflictmediate (in part) the effects of trust and commitment– is in line both with the arguments presented above,and also with Robson et al.’s (2002) review of thestrategic alliance literature on the behavioural ante-cedents of performance, which argued that trustand commitment are the fundamental ties that bindpartners, and that co-operation and conflict relateto behaviours that may vary as a consequence ofvarying degrees of trust and commitment.
Given that the evidence for the behavioural con-structs is fairly conclusive, the main opportunitiesfor enhancing knowledge through future research lienot in testing whether and why – on average – theseconstructs affect performance, but rather in testing
whether these effects are independent of other con-ditions. Researchers of trust who have started suchwork have found that (for example) the effect oftrust on performance is moderated positively by riskcommensuration, commitment (Luo 2002a), inter-dependence (Krishnan et al. 2006; Luo 2002a) andeconomic integration, but is moderated negativelyby alliance age, market uncertainty (Luo 2002a),alliance size (Robson et al. 2008) and legalism as aconflict resolution strategy (Lin and Wang 2008).While such moderation of trust’s effects has receivedattention from a fairly wide range of researchers,only Luo and Park have tested the moderation ofco-operation’s effects, finding that the effect ofco-operation on performance is moderated positivelyby contingency adaptability, term specificity (Luo2002b) market uncertainty (Luo and Park 2004) andequity in the alliance (Luo 2008). None of the studiesconsider moderation of the effects of commitmentand conflict, which is somewhat surprising, at least inthe case of conflict, given that the team conflict lite-rature has long demonstrated that the effects of con-flict may be less clearly negative than suggested – forinstance, Jehn (1995) finds that, while task conflictmay be detrimental to groups that perform routinetasks, it may in fact be beneficial to those performingnon-routine tasks. Following up these findings – andnumerous others – from the team conflict literaturecan inform future alliance conflict research, whichhas so far focused mainly on providing depth byresearching the efficacy of different conflict resolu-tion strategies (e.g. Lin and Germain 1998; Lin andWang 2008; Lu 2006).
Dissimilarities. Many researchers have focusedon identifying the characteristics of a good ‘fit’between partners, some suggesting that dissimilari-ties between partners – and others that similarities –may play important roles. While the literature hassuggested many other examples of both – includingdissimilarities in alliance experience, age, strategicscope, strategic content, goals, ownership type andreputation (Lowen and Pope 2008; McCutchen et al.2008; Park and Ungson 1997; Saxton 1997; Yeheskelet al. 2001) – only size and cultural dissimilarities (atboth national and organizational levels) and related-ness have been suggested in a sufficient number ofthe sample studies to be included in the review.
Size dissimilarities. A construct considered par-ticularly in early studies is size dissimilarity, wheresize relates to the complexity, structures and styles of
74 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
organizations (Park and Ungson 1997). The con-struct is considered in nine articles: most researcherssuggest that size dissimilarities between partnersdamage alliance performance, arguing that they giverise to a host of other dissimilarities which also harmperformance. For instance, Park and Ungson (1997,p. 287) quote Doz’s (1988) finding that ‘JVs betweenlarge and small firms experience difficulties becauseof idiosyncratic incentives for partners, cultural vari-ation, and asymmetric policies in information distri-bution’, while Merchant and Schendel (2000, p. 727)argue that partners’ relative firm size influences theirJVs’ performance through the ensuing dissimilaritiesin their administrative protocols (systems and proce-dures). Beamish and Jung (2005) argue that thisleads to difficulties in partners’ mutual understand-ing, increasing the risk of opportunistic behaviourreplacing co-operative behaviour. However, theypresent two hypotheses with opposite signs, as theyalso suggest the possibility of a positive relationshipwhere dissimilarities give rise to complementaryabilities which can fuel better performance: Yeheskelet al. (2001) advance the same argument in onlyfinding a positive effect of dissimilar size on per-formance. Different studies use numerous measuresfor size dissimilarities, including differences, ratiosand standard deviations in sales, assets, employeenumbers and equity market values. In spite of this –and of using different measures for performance(including subjective measures, CAR and stability) –their results are fairly consistent in only beinginsignificant.
National cultural distance. One construct for whichthe findings diverge more than for any others isnational cultural distance – and at the same time it isone that receives the most attention, being addressedin 22 articles. While there are different definitionsof cultural distance, they generally refer to dissimila-rities in beliefs, values, practices and behavioursshared by members of a nation (Hofstede 1980). Thegeneral agreement is that cultural distance is nega-tively related to performance, most often supportedby the argument that it increases the risks of mistrust,misunderstandings, miscommunication and manage-rial conflicts, so hampering efficient alliancemanagement (Glaister and Buckley 1999; Kim andParkhe 2009; Makino et al. 2007). Only Beamishand Kachra (2004) hypothesize that there may be apositive relationship, following Yeheskel et al.’s(2001) reasoning in suggesting a positive effect ofsize dissimilarities, that dissimilar cultures may also
give rise to complementarities and thus synergies andinnovation.
While most researchers (with this notable excep-tion) agree with the general hypothesis, there is lessagreement with respect to measurement: in factnational cultural distance is measured in threedifferent ways – which even require separate expla-nations – as is reflected by the three separate sub-entries (in italics) in Table 1. Most studies use anindex of several dimensions of cultural distancebetween the local and foreign partners’ home coun-tries. In all but one study (which used Globe’s ninedimensions (see House et al. 2004), the index usedwas based on Hofstede’s four (later updated to five)dimensions of national cultural distance: but theirresults for these indices were almost as mixed asthey could be, nearly equally distributed betweenpositive, negative and insignificant results. A smallgroup of researchers examined the possibility thatsome cultural distance dimensions may be harmful,while others are not. While three studies whichfollow this line do not directly suggest differenteffects of individual dimensions, Barkema and Ver-meulen (1997) argue (and find some support for theidea) that only differences in ‘uncertainty avoid-ance’ and in ‘long-term orientation’ should have anegative impact on performance. However, acrossthe studies in general, there is no overwhelmingsupport for a difference between the effects ofindividual dimensions, as almost all the remainingresults are insignificant. Finally, a third group ofresearchers use dummy variables or other compari-sons of the performance of groups of alliances thatrepresent supposedly varying extents of cultural dis-tance based on partner’s nationalities or regionallocations. Chiao et al. (2009), Hanvanich et al.(2003) and Makino and Beamish (1998) offer themost refined approach, which distinguishes betweenthe cultural distance between partners and thatbetween the country of location and the partnerswhen none is local. In this way, they divide theconcept of cultural distance into two categories –which could be termed ‘location cultural distance’and ‘partner cultural distance’ – which are in effectintertwined in traditional settings where one of thepartners is local (Hanvanich et al. 2003). While thestudies that apply the simplest country or regiondummies mainly provide insignificant results, thislatter approach provides results which support theassumption that cultural distance both between part-ners and between partners and local actors may havenegative effects.
Antecedents of ISA Performance 75
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
Organizational cultural distance. A constructclosely related to national cultural distance is organ-izational cultural distance (addressed in 11 articles).Researchers explicitly framing this construct as dis-tinct from national cultural distance suggest that,whereas the latter can only exist between firms fromdifferent countries and typically relates to values,the organizational construct can exist between firmsfrom the same country and typically relates to prac-tices (Pothukuchi et al. 2002). As only few studiesexplicitly distinguish between the two, I categorizethem pragmatically, based on whether researchersask partners about any perceived cultural distancebetween them. Rationales for suggesting a negativeinfluence on performance remain largely the same asthose for national cultural distance: organizationalcultural distance can also make it difficult to estab-lish common expectations and impedes co-operation(Kim and Parkhe 2009; Lin and Germain 1998).The measures employed are subjective and based onstakeholders’ assessments of dissimilarities betweenthe partners on different aspects, ranging from mani-festations of different national cultures (Park et al.2008) to variations between same-country partners(Pothukuchi et al. 2002). The results are less diver-gent than in the case of national cultural distance:generally negative relationships with performanceare found, as well as some that are insignificant. Nopositive relationships are found.
Relatedness. The final construct in this group(which is addressed in eight articles) refers rather tosimilarities – of the business activities, products andindustries of the organizations involved (Lu and Xu2006; Luo 1997; Merchant and Schendel 2000).Most authors have focused on the performanceimpact of relatedness between a JV and one or moreof its partners, typically invoking one or both of thefollowing arguments for JV/partner relatednesshaving a positive influence on performance: thepotential for economies of scale, and for facilitatingthe transfer of knowledge and resources (Lu and Xu2006; Merchant and Schendel 2000). When partnerrelatedness is the focus, such arguments are supple-mented by the notion that related partners will bebetter able to detect and thus deter opportunisticbehaviour (Luo 1997; Merchant and Schendel 2000).In most cases, relatedness is assessed through match-ing SIC codes, but researcher or stakeholder assess-ments have also been applied. The proposed positiverelationship is supported – fairly consistently –across a number of different relatedness and per-
formance measures, including subjective, stability,CAR and accounting measures.
Implications for the model. In line with theresearchers’ rationales reflected above, Figure 1shows that size dissimilarities reduce co-operation,as suggested indirectly by most researchers andexplicitly by Beamish and Jung (2005). Figure 1 alsoshows a positive arrow from size dissimilaritiestowards performance, which directly reflects thearguments by Beamish and Jung (2005) and Yeheskelet al. (2001) that size dissimilarities may give riseto complementarity of abilities to the benefit ofperformance. These arguments suggest opposingeffects, which may explain the fact that all resultswere insignificant. The opposing arguments are inline with Parkhe’s (1991) line that, while dissimilari-ties provide reciprocal strengths, and complementaryresources are the very reason for the formation ofalliances, they may also inhibit co-operation andincrease transaction costs.
Similarly, Figure 1 illustrates the opposing effectsof national cultural distance: on the one hand, itshows a positive arrow towards performance(directly reflecting Beamish and Kachra’s (2004)argument that dissimilarities create complementa-rity); while on the other, a negative arrow reflectsa negative effect on performance through trust,co-operation and conflict, in line with Glaister andBuckley (1999), Kim and Parkhe (2009) and Makinoet al. (2007). In fact, suggesting opposing effectsfor national cultural distance aligns both with theliterature’s theoretical arguments and with the almostequal distribution between positive, negative andinsignificant results found in the sample.
While Figure 1 also reflects the negative effects oforganizational cultural distance on co-operation assuggested by the researchers who considered thisconstruct (Kim and Parkhe 2009; Lin and Germain1998), it depicts no links to trust or to conflict,reflecting the fact that they are rarely suggested byresearchers. Their absence is in line with the notionthat organizational cultural distance concerns differ-ences in practices and behaviours – potentiallycausing practical co-ordination and co-operationproblems – rather than distrust and conflict arisingfrom differences in the deeply rooted values andbeliefs more typical of national culture. Again (asresearchers have not suggested one) Figure 1 showsno positive arrow towards performance: this may bebecause operational differences (the typical manifes-tations of organizational cultural distance) do not
76 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
create synergies in the same manner as national cul-tural distances, and so cannot contribute differentviewpoints (Kim and Park 2002). So the figure onlysuggests that organizational cultural distance hasnegative effects on performance, which is in linewith the sample studies’ empirical findings.
Figure 1 also reflects the arguments and findingssupporting the positive effect of relatedness on per-formance, an effect that is suggested to operate notthrough behavioural constructs, but via the increasedpotential for economies of scale and the relativelyeasier transfer of knowledge and resources (Lu andXu 2006; Merchant and Schendel 2000).
Experience. Another group of constructs relates tothe pre-alliance experience partners have of relevantbusiness aspects: four constructs – alliance experi-ence, international experience, international allianceexperience and prior partner relationships – weresuggested in a sufficient number of articles to beincluded in this review.
Alliance experience. Experience relates to the part-ners’ learning opportunities provided by earlierevents and processes similar to those in question –one type is alliance experience, which was addressedin nine articles. Researchers agree in suggesting thatalliance experience affects performance positively,motivated mainly by arguments based on organiza-tional learning concepts: relevant alliance experiencehelps partners to reduce mistakes and make the rightdecisions, both initially and in responding to chal-lenges that may arise (McCutchen et al. 2008;Merchant and Schendel 2000). Most studies measureexperience as the number of prior alliances a partnerhas engaged in, and most present results support-ing the hypothesis of a positive association withperformance.
International experience. A similar learning oppor-tunity provided by earlier experiences concerns thelevel of partners’ previous participation in inter-national business (here ‘international experience’),which was addressed in nine articles. Internationalexperience – as well as more specific experience inthe focal host country or region – is also generallysuggested as being positively related to performance,by enabling partners to develop skills and capabili-ties that allow them to tune into foreign conditionsmore successfully (Barkema et al. 1997). Interna-tional experience can also facilitate communicationand reduce the risk of misunderstandings (Luo
1997), and thus the liability of foreignness (Luand Beamish 2006). While these arguments suggesta positive path from experience to performance,Lowen and Pope (2008) and Lu and Beamish (2006)suggest that a negative path is also possible if per-formance is measured in terms of stability. Theyargue that the more a foreign partner has learnedabout the business setting, the more likely it is that itwill opt for a wholly owned subsidiary instead of analliance. There is great diversity in the measurementof international and host country experience, butmost researchers measure it as the number of years ofsuch involvement, the number of prior or currentwholly owned subsidiaries, or as sales percentages ininternational settings or in a given country/region.Results are also mixed: with stability as the perform-ance measure, some find a positive influence, othersfind a negative influence, while the majority find noinfluence; in studies using other performance meas-ures, results are either positive or insignificant.
International alliance experience. Some research-ers have investigated the double learning opportunityarising not just from having participated in alliancesor doing business in an international context, but infact doing both in one setting – thus gaining ‘inter-national alliance experience’ (addressed in ten arti-cles). Arguments in the articles studied suggest thebeneficial effects of alliance experience and of inter-national experience are combined, and so propose apositive relationship between international allianceexperience and performance (Barkema et al. 1997;Meschi 2004). While the measures of experience areagain fairly mixed and similar to those mentionedabove, the findings are fairly consistent in suggestingthat relationships are either positive or insignificant.
Prior relationships. A final experience constructqualifying for this review is prior relationships,which refers to the learning opportunity provided bypartners’ prior business relationships with eachother. This construct was addressed in eight articles:two looked specifically at prior alliances between thepartners, while the remaining eight took a broaderoutlook to also encompass other types of prior rela-tionships between the partners. Two main argumentsare advanced to support a positive effect on perform-ance – one relates to the knowledge that partnershave gained about each other, which is expected to beuseful for understanding, for instance, resources andcapabilities (Nielsen 2007; Saxton 1997); the otherrelates to prior relationships creating social relations
Antecedents of ISA Performance 77
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
which provide the basis for trust and mutual under-standing in a later JV (Glaister and Buckley 1999;Greve et al. 2010). Again, the measures used toreflect experience and performance are diverse, asare the results – there are a few positive and negativeresults, but most are insignificant.
Implications for the model. In line with the findingsand the rationales presented above, Figure 1 showsthat alliance experience directly affects performancepositively, reflecting the argument that allianceexperience enables partners to avoid mistakes andincreases the likelihood of them making the rightdecisions (McCutchen et al. 2008; Merchant andSchendel 2000). Similarly, skills and capabilitiesaccumulated through international experience allowfirms to tune into foreign conditions, as suggestedabove and reflected by the direct positive arrow frominternational experience to performance (Barkemaet al. 1997). The figure also reflects the argument thatinternational experience moderates the damagingeffects of cultural distance by reducing the liabilityof foreignness (Lu and Beamish 2006). Although thefindings show negative associations between inter-national experience and stability – in line with thetheoretical arguments that firms with more inter-national experience would be more inclined to go italone (Lowen and Pope 2008; Lu and Beamish 2006)– there is no negative arrow pointing towards per-formance, as the effect on stability does not reflect aneffect on performance – rather it reflects an improve-ment in that alternative. To avoid making Figure 1overly complex, I do not illustrate international alli-ance experience explicitly – the suggested effects ofthat construct only combine those of alliance expe-rience and international experience – so the model’sonly remaining experience construct is prior relation-ships. Two arrows emanate from this construct, bothpositive: one points directly towards performance –which reflects the argument that partners with priorrelationships understand each others’ resources andcapabilities and thus are better able to exploit themeffectively and efficiently (Nielsen 2007; Saxton1997) – and the other towards trust, reflecting theargument that prior relationships create social rela-tions which enhance trust and mutual understanding(Glaister and Buckley 1999; Greve et al. 2010).
Control. In common with previous recent reviews(e.g. Nippa et al. 2007; Ren et al. 2009), I find thatcontrol issues have received consistent attentionfrom the earliest literature up to now. A number of
different issues have been considered, including forexample the autonomy of a JV (Newburry and Zeira1999) or the effects of different types of control(output, process or social (Aulakh et al. 1996) andstrategic, operational or structural aspects (Yan andGray 2001)); however, this review only considersthree control issues – dominant (vs. shared) control,dominant control by the foreign partner and domi-nant control by the focal partner – as the only con-structs addressed in five or more sample articles.
Dominant partner control. Control refers to thedecision-making power exercised by the partners(Killing 1983), and much research has consideredwhether evenly distributed control or dominance byone partner is the better structure for an alliancerelationship. Dominant partner control (which wasaddressed in six articles) represents the situationwhere one partner manages the alliance virtually as ifthey were a wholly owned subsidiary (Killing 1983,p. 16), and is generally viewed as being beneficial inreducing managerial complexity and alleviatingchallenges of co-ordination and thus the potentialfor conflicts between partners (Merchant 2002).However, a completely opposite argument is alsovalid – that such dominance can lead the partnerinvolved to exploit the power of its position, frustrat-ing the other partner and thus potentially leadingto conflict (Hébert 1996; Pangarkar and Lee 2001).Using equity distributions as their main measures ofcontrol, researchers report mixed results, supportingpositive, negative and insignificant relationshipsalmost equally, and thus providing little evidence thatremoves conceptual ambiguity for this construct.
Dominant control by foreign partner. When alli-ances involve partners from developed and develop-ing countries, foreign partners worried about therisks involved in doing business in local countriesoften seek to exercise more decision-making powerthan the local partner, thus creating an alliance char-acterized by dominant control by the foreign partner(Calantone and Zhao 2001; Ding 1997), a constructaddressed in seven articles. In addition to the generalargument suggesting the positive effects of dominantcontrol – reductions in managerial complexity,co-ordination challenges and inter-partner conflict –some researchers have argued that foreign partnersfrom developed countries will have more advancedmanagerial skills and technology than the localdeveloping country partner, so that foreign partnerdominance will be positively related to performance
78 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
(Calantone and Zhao 2001; Ding 1997). Given theseadditional arguments for a positive relationship,one would expect that results would provide moreconclusive support for a positive connection betweendominant control and performance. Empirical testsusing both equity distribution and stakeholderassessment measurements of dominant controlsupport that expectation to some extent, as no nega-tive results appear – but a fairly large proportion ofinsignificant results are still evident.
Dominant control by focal partner. Some research-ers suggest that asymmetrical control may haveasymmetrical effects on performance as perceived bythe partners, necessitating distinguishing betweenfocal and non-focal partners. They examine theimportance of what I define as dominant control bythe focal partner (a construct addressed in eight arti-cles) and most commonly suggest that the positiveconsequences of dominant control will be greaterfor the partner exercising that control. The mostcommon arguments advanced concern the dominantpartner’s ability to control strategic resources (Mjoenand Tallman 1997), to curb the other partner’s oppor-tunistic behaviour and thus to gain the lion’s share ofbenefits from the partnership (Yan and Gray 2001). Anumber of these studies bring extra detail to theiranalysis by investigating the effects of differenttypes of control – specific vs. overall (Luo et al.2001; Mjoen and Tallman 1997) or strategic vs.operational vs. structural (Yan and Gray 2001).However, only Luo et al. (2001) suggest these differ-ent types of control produce different effects, findingthat Chinese partners may not perceive having domi-nant overall control as beneficial, but may gainadvantages from dominant specific control. Usingmeasures mainly originating from JV or partner rep-resentatives’ assessments of relative control, mostempirical tests support the positive influence of focalpartner dominant control; however, insignificantfindings are also numerous, and two studies presentnegative results.
Implications for the model. As the rationalesrelated to dominant control by the focal partner differfrom those of dominant control only in terms of thesuggested effects on the distribution of performanceimprovements rather than their creation, Figure 1only depicts the dominant control and dominantcontrol by foreign partner constructs. The negativearrow from dominant control towards conflict repre-sents the most common assertion – that dominant
control by one partner will reduce managerialcomplexity and thus also the potential for conflicts(Merchant 2002). The less common and opposingargument is that control by one partner may frustratethe other and thus increase conflict (Hébert 1996;Pangarkar and Lee 2001): recognizing this alterna-tive (and the diverging empirical findings), I alsoinclude a positive arrow towards conflict: which ofthem prevails is likely to depend on conditioningfactors (such as, for instance, goal congruence orculture). Turning to dominant control by the foreignpartner, the model includes the same two arrowstowards conflict, but also a positive direct arrowtowards performance, which reflects the argumentthat dominant foreign partners from developed coun-tries will possess advanced managerial skills andtechnology, so that their dominance should enhanceperformance (Calantone and Zhao 2001; Ding 1997).
A note on measurement
While the model I have built up considers compo-nents at the construct level, it is clear that measure-ment issues may influence empirical researchfindings. In particular, the review of the performancemeasures used indicated that they should be chosenwith care. One concern was the use of stability meas-ures as performance measures, particularly whenconsidering antecedents that could be imagined toaffect stability through other mechanisms than byaffecting performance negatively. Another concernwas the simultaneous use of subjective measures ofboth an antecedent construct and of performance,which may give rise to same-source variance thatcould bias the findings. Such concerns suggestgreater attention needs to be paid to the combinationof measures of performance and measures of ante-cedents in empirical research in this area.
Table 2 brings support to this suggestion by sum-marizing the test coefficients on which Table 1 isbased. The numbers in bold indicate how manystatistical tests employed subjective vs. objectivemeasures of antecedents in combination with each ofthe five types of performance measures presented.5
Most importantly, these numbers show that, in 200
5Subjective measures of antecedents are those subjectivelyassessed by a representative of the alliance or partner com-panies, whereas objective measures of antecedents are thosebased on archival information, as well as independent vari-ables based on survey information, where such survey infor-mation itself does not consist of subjective assessments, butrather of (fairly) objective information (e.g. age of alliance).
Antecedents of ISA Performance 79
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
out of the 217 tests in which an antecedent wasmeasured by a subjective measure, the performancemeasure was also subjective – in most cases (but notalways) the measures were obtained from the samerespondent. Turning to the numbers not in bold, wesee that, in 155 of those 200 tests, the coefficientssupported the hypothesis, 44 were insignificant, andone directly opposed what had been hypothesized.This relatively high ratio of supportive findings is, ofcourse, likely to be a consequence of associationsbetween the constructs, but in the absence of testsinvolving objective measures of performance, itcannot be ruled out that at least some of these find-ings are the consequences of such measurement phe-nomena as same-source variance. I suggest futureresearch considering such antecedents as trust andco-operation – which are typically measured subjec-tively – could benefit if such measures were com-bined with objective measures of performance to agreater extent than has been the case hitherto.
Another observation is that, for antecedents meas-ured by objective measures, there tends to be poorsupport for the hypotheses tested (36% vs. 77% forsubjective antecedent measures). Obviously, this mayreflect the true association (or lack thereof) betweenconstructs, but again it cannot be ruled out that meas-urement plays a role, as the lack of support may be aconsequence of the objective measures providingpoor reflections of the constructs they are intended toreflect. For instance, it is quite possible that an objec-tive measure such as dominant equity ownership maynot reflect actual dominant control as accurately assubjective measures might have. So future researchconsidering such antecedents as dominant controlor different types of experience (which are typicallymeasured by objective means) might benefit if sub-jective construct measures were combined withobjective measures of performance.
Conclusion
There is no doubt that research enhancing under-standing of antecedents of ISA performance hasadvanced dramatically, especially over the pastcouple of decades. But this review shows that thereis still much that we could understand better, evenabout those antecedents of performance that havebeen considered most frequently. Robson et al.(2002) found the literature on IJVs ‘remains chaoticand to a large extent ambiguous’, and I echo thisassessment for the broader ISA literature. This paperTa
ble
2.M
easu
rem
ent
Tota
lS
ubje
ctiv
eS
tabi
lity
CA
RA
ccou
ntin
gE
xter
nal
All
Sup
.0
Opp
.A
llS
up.
0O
pp.
All
Sup
.0
Opp
.A
llS
up.
0O
pp.
All
Sup
.0
Opp
.A
llS
up.
0O
pp.
Sub
ject
ive
ante
cede
nt21
716
848
120
015
544
13
30
00
00
014
104
00
00
0O
bjec
tive
ante
cede
nt25
491
125
3881
1649
1691
4146
448
1326
927
172
87
42
1To
tal
471
259
173
3928
117
193
1794
4446
448
1326
941
276
87
42
1
80 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
contributes to the literature by identifying two poten-tial sources of ambiguity in previous research onantecedents of ISA performance, as well as by sug-gesting how they might be addressed.
One potential source of this ambiguity is the lackof attention to the measurement of central constructs.The five types of performance measures I list (sub-jective, stability, accounting measures, CAR andexternal evaluation) are shown to be very different,and may reflect different constructs, but the subse-quent detailed review of individual antecedents indi-cates that hypothesis development rarely takes suchdifferences into account. A notable exception here isLu and Beamish (2006), who suggest that inter-national experience will be positively related to asubjective performance measure, but negativelyrelated to stability, because a foreign partner who haslearned more about the business setting is morelikely to opt for a wholly owned subsidiary instead ofan alliance. The fact that they also find support forthis suggested difference demonstrates the utmostimportance of ensuring congruence between the per-formance construct assumed in the hypothesis devel-opment and that reflected by the chosen performancemeasure: the preceding section demonstrated thatthese efforts can be strengthened considerably. Bydiscussing the different performance measures usedin studies, and laying out in detail the theoreticalarguments presented in previous research, this paperprovides useful aid in the process of strengtheningthese efforts.
The lack of attention to measurement is borne outin my final note on the issue, which indicates howrare it is to see combinations of subjective and objec-tive measurements, suggesting that the choice ofmeasures too often depends more on conveniencethan on methodological considerations such as thoserelated to construct validity or same-source variance.This suspicion is amplified when one considers thedetailed review of the individual antecedents (sum-marized in Table 1, detailed in Appendix S2 (avail-able in the electronic version only)), which showsthat (for certain types of constructs) the choice ofperformance measure often seems to follow thechoice of the antecedent measure – so that (forinstance) objective experience measures are oftencombined with stability measures, and subjectivebehavioural measures with subjective performancemeasures. With the notable exceptions of Sim and Ali(2000) and Steensma and Lyles (2000), none of the32 studies involving behavioural constructs uses sta-bility as a performance measure, despite its general
popularity for this purpose. This paper identifies thetypical combinations of antecedent and performancemeasures, and thus also points to the possibilities ofchallenging those combinations to assess whethermeasurement issues impact findings. The study’sreview of the individual antecedents presents bothmeasurement choices and results of prior studies(summarized in Table 1, detailed in Appendix S2(available in electronic version only)), thus providingan easily accessible source for comparing thealternative outcomes of different combinations ofchoices.
The second potential source of ambiguity identi-fied in this review is that different explanations havebeen advanced about the mechanisms via which anumber of well-known constructs might affect per-formance, but without these different explanationsbeing tested empirically. Thus, while numerous theo-retical explanations about such causal mechanismsare available, most of them have little empiricalsupport, which gives rise to ambiguity in variousareas. For constructs for which research has pro-duced divergent results, ambiguity about the causalmechanisms for some constructs is quite obvious.One example is national cultural distance, whereresults are almost equally divided between insignifi-cant, negative and positive. This paper suggests thatthe ambiguity about constructs which yield divergentresults can be reduced considerably by testing thecausal mechanisms already suggested in the litera-ture. In the case of the example at hand, nationalcultural distance, this would entail testing the causalmechanisms with different impacts on performancethat have been suggested in prior research; a positiveeffect arising because of complementary resources(Beamish and Kachra 2004), and a negative effectvia decreased trust and co-operation and increasedconflict (Glaister and Buckley 1999; Kim and Parkhe2009; Makino et al. 2007). For other constructs –where the results all point in the same direction – theambiguity about causal mechanisms is less obvious.In these cases, previous research may indicate a rela-tionship with performance, but empirical results shedlittle light on why it exists. One such example con-cerns prior relationships between partners. Do theyinfluence performance positively because the part-ners know each other’s resources well, which enablesefficient integration of resources (Nielsen 2007;Saxton 1997), or because partners have alreadydeveloped mutual trust (Glaister and Buckley 1999;Greve et al. 2010)? Only if we know the answers tothese questions will we be able to convincingly
Antecedents of ISA Performance 81
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
suggest to managers whether (for instance) problemsof distrust can be avoided by choosing partners withwhom the company has prior relationships.
So this paper claims that there is still much ambi-guity about the causal mechanisms suggested in theliterature, both for constructs with diverging andwith converging results, and that much of it could beresolved via empirical testing which deepens under-standing of their operation. The model illustrated inFigure 1 identifies which theoretical explanationsshould first be tested by synthesizing the results ofprior research with theoretical explanations of thecasual mechanisms posited by researchers in theirhypothesis development, so my suggestions forfuture research are anchored in the collective under-standing of our field. But the model does not justpresent suggestions to be tested to increase know-ledge regarding the influence of individual con-structs, but also highlights how the constructs which,empirically, are often researched separately, are inte-grated in the theoretical rationales that researchersposit, even though the theoretical foundations ofthe literature vary (Culpan 2009). This paper makesexplicit how, in spite of their varied theoretical foun-dations, these rationales are often related. In particu-lar, the model highlights how the rationales related tobehavioural constructs are central to the rationalesbehind most other constructs, and thus argues thatthis central role should be reflected in empirical testsof the mediating effects of these constructs.
ReferencesAnderson, E. (1990). Two firms, one frontier: on assessing
joint venture performance. Sloan Management Review,31, pp. 19–30.
Argyle, M. (1991). Cooperation. The Basis of Sociability.London: Routledge.
Ariño, A. (2003). Measures of strategic alliance perform-ance: an analysis of construct validity. Journal of Inter-national Business Studies, 34, pp. 66–79.
Armstrong, S. and Wilkinson, A. (2007). Processes, proce-dures and journal development: past, present and future.International Journal of Management Reviews, 9, pp.81–93.
Aulakh, P.S., Kotabe, M. and Sahay, A. (1996). Trust andperformance in cross-border marketing partnerships: abehavioral approach. Journal of International BusinessStudies, 27, pp. 1005–1032.
Barkema, H.G. and Vermeulen, F. (1997). What differencesin the cultural backgrounds of partners are detrimentalfor international joint ventures? Journal of InternationalBusiness Studies, 28, pp. 845–864.
Barkema, H.G., Shenkar, O., Vermeulen, F. and Bell, J.H.J.(1997). Working abroad, working with others: how firmslearn to operate international joint ventures. Academy ofManagement Journal, 40, pp. 426–442.
Beamish, P.W. (1985). The characteristics of joint venturesin developed and developing countries. Columbia Journalof World Business, 20, pp. 13–19.
Beamish, P.W. and Jung, J.C. (2005). The performance andsurvival of joint ventures with parents of asymmetric size.Management International, 10, pp. 19–30.
Beamish, P.W. and Kachra, A. (2004). Number of partnersand JV performance. Journal of World Business, 39,pp. 107–120.
Beamish, P.W. and Lupton, N.C. (2009). Managing jointventures. Academy of Management Perspectives, 23,pp. 75–94.
Bowman, E.H. and Hurry, D. (1993). Strategy through theoption lens: an integrated view of resource investmentsand the incremental-choice process. Academy of Manage-ment Review, 18, pp. 760–782.
Calantone, R.J. and Zhao, Y.S. (2001). Joint ventures inChina: a comparative study of Japanese, Korean, and U.S.partners. Journal of International Marketing, 9, pp. 1–23.
Chiao, Y.C., Yu, C.M.J. and Peng, J.T.A. (2009). Partnernationality, market-focus and IJV performance: a contin-gent approach. Journal of World Business, 44, pp. 238–249.
Culpan, R. (2009). A fresh look at strategic alliances:research issues and future directions. InternationalJournal of Strategic Business Alliances, 1, pp. 4–23.
Ding, D.Z. (1997). Control, conflict, and performance: astudy of U.S.–Chinese joint ventures. Journal of Inter-national Marketing, 5, pp. 31–45.
Doz, Y.L. (1988). Technology partnerships between largerand smaller firms: some critical issues. In Contractor, F.J.and Lorange, P. (eds), Cooperative Strategies in Inter-national Business. Lexington, KY: Lexington Books,pp. 317–328.
DuBois, F.L. and Reeb, D. (2000). Ranking the internationalbusiness journals. Journal of International BusinessStudies, 31, pp. 689–704.
Franko, L.G. (1971). Joint Venture Survival in MultinationalCorporations. New York, NY: Praeger.
Friedmann, W.G. and Kalmanoff, G. (1961). Joint Inter-national Business Venture. New York, NY: ColumbiaUniversity Press.
Geringer, J.M. and Hébert, L. (1989). Control and perform-ance of international joint ventures. Journal of Inter-national Business Studies, 20, pp. 235–254.
Geringer, J.M. and Hébert, L. (1991). Measuring perform-ance of international joint ventures. Journal of Inter-national Business Studies, 22, pp. 249–264.
Geyskens, I., Steenkamp, J.B., Scheer, L.K. and Kumar, N.(1996). The effects of trust and interdependence on rela-tionship commitment: a Trans-Atlantic study. Interna-tional Journal of Research in Marketing, 13, pp. 303–317.
82 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
Glaister, K.W. and Buckley, P.J. (1999). Performance rela-tionships in UK international alliances. ManagementInternational Review (MIR), 39, pp. 123–147.
Gong, Y., Shenkar, O., Luo, Y. and Nyaw, M.-K. (2005).Human resources and international joint venture perform-ance: a system perspective. Journal of International Busi-ness Studies, 36, pp. 505–518.
Greve, H.R., Baum, J.A.C., Mitsuhashi, H. and Rowley, T.J.(2010). Built to last but falling apart: cohesion, friction,and withdrawal from interfirm alliances. Academy ofManagement Journal, 53, pp. 302–322.
Griffith, D.A., Hu, M.Y. and White, D.S. (1998). Formationand performance of multi-partner joint ventures: a Sino-foreign illustration. International Marketing Review, 15,pp. 171–187.
Hanvanich, S., Miller, S.R., Richards, M. and Cavusgil, S.T.(2003). An event study of the effects of partner and loca-tion cultural differences in joint ventures. InternationalBusiness Review, 12, pp. 1–16.
Harrigan, K.R. (1988). Strategic alliances and partner asym-metries. Management International Review (MIR), 28,pp. 53–72.
Hatfield, L., Pearce, J.A. II, Sleeth, R.G. and Pitts, M.W.(1998). Toward validation of partner goal achievementas a measure of joint venture performance. Journal ofManagerial Issues, 10, pp. 355–372.
Hébert, L. (1996). Does control matter? A path model of thecontrol–performance relationship in international jointventures. Management International, 1, pp. 27–39.
Heide, J.B. and Miner, A.S. (1992). The shadow of thefuture: effects of anticipated interaction and frequency ofcontact on buy–seller cooperation. Academy of Manage-ment Journal, 35, pp. 265–291.
Hofstede, G. (1980). Culture’s Consequences: InternationalDifferences in Work-Related Values. Beverly Hills, CA:Sage.
House, R.J., Hanges, P.J., Javidan, M., Dorfman, P.W. andGupta, V. (2004). Culture, Leadership and Organizations:The GLOBE Study of 62 Societies. Thousand Oaks, CA:Sage Publications.
Hu, M.V. and Chen, H. (1996). An empirical analysis offactors explaining foreign joint venture performance inChina. Journal of Business Research, 35, pp. 165–173.
Janger, A.R. (1980). Organization of International JointVentures. New York, NY: Conference Board.
Jehn, K.A. (1995). A multimethod examination of the ben-efits and detriments of intragroup conflict. AdministrativeScience Quarterly, 40, pp. 256–282.
Jehn, K.A. and Mannix, E.A. (2001). The dynamic nature ofconflict: a longitudinal study of intragroup conflict andgroup performance. Academy of Management Journal,44, pp. 238–251.
Killing, J.P. (1983). Strategies for Joint Venture Success.New York, NY: Praeger.
Kim, C.S. and Inkpen, A.C. (2005). Cross-border R&Dalliances, absorptive capacity and technology learning.
Journal of International Management, 11, pp. 313–329.
Kim, J. and Parkhe, A. (2009). Competing and cooperatingsimilarity in global strategic alliances: an exploratoryexamination. British Journal of Management, 20, pp.363–376.
Kim, K. and Park, J.-H. (2002). The determinants of valuecreation for partner firms in the global alliance context.Management International Review (MIR), 42, pp. 361–384.
Kogut, B. (1988a). Joint ventures: theoretical and empiricalperspectives. Strategic Management Journal, 9, pp. 319–332.
Kogut, B. (1988b). A study of the life cycle of joint ventures.Management International Review (MIR), 28, pp. 39–52.
Kogut, B. (1991). Joint ventures and the option to expandand acquire. Management Science, 37, pp. 19–33.
Koh, J. and Venkatraman, N. (1991). Joint venture forma-tions and stock market reactions: an assessment in theinformation technology sector. Academy of ManagementJournal, 34, pp. 869–892.
Krishnan, R., Martin, X. and Noorderhaven, N.G. (2006).When does trust matter to alliance performance?Academy of Management Journal, 49, pp. 894–917.
Kwon, Y.C. (2008). Antecedents and consequences ofinternational joint venture partnerships: a social exchangeperspective. International Business Review, 17, pp. 559–573.
Li, M., Zhang, Y. and Jing, R. (2008). Does ownership andculture matter to joint venture success? InternationalManagement Review, 4, pp. 90–102.
Lin, X. and Germain, R. (1998). Sustaining satisfactoryjoint venture relationships: the role of conflict resolutionstrategy. Journal of International Business Studies, 29,pp. 179–196.
Lin, X. and Wang, C.L. (2008). Enforcement and perform-ance: the role of ownership, legalism and trust in interna-tional joint ventures. Journal of World Business, 43, pp.340–351.
Lowen, A. and Pope, J. (2008). Survival analysis of interna-tional joint venture relationships. Journal of Business &Economic Studies, 14, pp. 62–80.
Lu, J.W. and Beamish, P.W. (2006). Partnering strategiesand performance of SMEs’ international joint ventures.Journal of Business Venturing, 21, pp. 461–486.
Lu, J.W. and Xu, D. (2006). Growth and survival of inter-national joint ventures: an external–internal legitimacyperspective. Journal of Management, 32, pp. 426–448.
Lu, L.-T. (2006). Conflict resolution strategy betweenforeign and local partners in joint ventures in China.Journal of American Academy of Business, Cambridge,8, pp. 236–240.
Luo, Y. (1997). Partner selection and venturing success: thecase of joint ventures with firms in the people’s republicof China. Organization Science, 8, pp. 648–662.
Antecedents of ISA Performance 83
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
Luo, Y. (2001). Antecedents and consequences of personalattachment in cross-cultural cooperative ventures.Administrative Science Quarterly, 46, pp. 177–201.
Luo, Y. (2002a). Building trust in cross-cultural collabora-tions: toward a contingency perspective. Journal ofManagement, 28, pp. 669–694.
Luo, Y. (2002b). Contract, cooperation, and performance ininternational joint ventures. Strategic ManagementJournal, 23, pp. 903–919.
Luo, Y. (2008). Procedural fairness and interfirm coopera-tion in strategic alliances. Strategic Management Journal,29, pp. 27–46.
Luo, Y. and Park, S.H. (2004). Multiparty cooperationand performance in international equity joint ventures.Journal of International Business Studies, 35, pp. 142–160.
Luo, Y., Shenkar, O. and Nyaw, M.K. (2001). A dual parentperspective on control and performance in internationaljoint ventures: lessons from a developing economy.Journal of International Business Studies, 32, pp. 41–58.
Madhavan, R. and Prescott, J.E. (1995). Market value impactof joint ventures: the effect of industry information-processing load. Academy of Management Journal, 38,pp. 900–915.
Makino, S. and Beamish, P.W. (1998). Performance andsurvival of joint ventures with non-conventional owner-ship structures. Journal of International Business Studies,29, pp. 797–818.
Makino, S., Chan, C.M., Isobe, T. and Beamish, P.W. (2007).Intended and unintended termination of internationaljoint ventures. Strategic Management Journal, 28, pp.1113–1132.
Mayer, R.C., Davis, J.H. and Schoorman, F.D. (1995). Anintegrative model of organizational trust. Academy ofManagement Review, 20, pp. 709–734.
McCutchen, W.W., Swamidass, P.M. and Teng, B.S. (2008).Strategic alliance termination and performance: the roleof task complexity, nationality, and experience. Journalof High Technology Management Research, 18, pp. 191–202.
Meier, M. (2011). Knowledge management in strategic alli-ances: a review of empirical evidence. InternationalJournal of Management Reviews, 13, pp. 1–23.
Merchant, H. (2002). Shareholder value creation via inter-national joint ventures: some additional explanations.Management International Review (MIR), 42, pp. 49–69.
Merchant, H. and Schendel, D. (2000). How do internationaljoint ventures create shareholder value? StrategicManagement Journal, 21, pp. 723–737.
Meschi, P.-X. (2004). Valuation effect of international jointventures: does experience matter? International BusinessReview, 13, pp. 595–612.
Mjoen, H. and Tallman, S. (1997). Control and performancein international joint ventures. Organization Science, 8,pp. 257–274.
Muthusamy, S.K., White, M.A. and Carr, A. (2007). Anempirical examination of the role of social exchanges inalliance performance. Journal of Managerial Issues, 19,pp. 53–75.
Newburry, W. and Zeira, Y. (1999). Autonomy and effec-tiveness of equity international joint ventures (EIJV’s): ananalysis based on EIJV’s in Hungary and Britain. Journalof Management Studies, 36, pp. 263–285.
Nielsen, B.B. (2007). Determining international strategicalliance performance: a multidimensional approach.International Business Review, 16, pp. 337–361.
Nippa, M., Beechler, S. and Klossek, A. (2007). Successfactors for managing international joint ventures: a reviewand an integrative framework. Management & Organiza-tion Review, 3, pp. 277–310.
Nooteboom, B., Van Haverbeke, W., Duysters, G., Gilsing,V. and Van Den Oord, A. (2007). Optimal cognitive dis-tance and absorptive capacity. Research Policy, 36,pp. 1016–1034.
Pangarkar, N. and Klein, S. (2004). The impact of control oninternational joint venture performance: a contingencyapproach. Journal of International Marketing, 12, pp.86–107.
Pangarkar, N. and Lee, H. (2001). Joint venture strategiesand success: an empirical study of Singapore firms.Journal of Asian Business, 17, pp. 1–13.
Park, B.I., Giroud, A., Mirza, H. and Whitelock, J. (2008).Knowledge acquisition and performance: the role offoreign parents in Korean IJVs. Asian Business &Management, 7, pp. 11–32.
Park, S.H. and Ungson, G.R. (1997). The effect of nationalculture, organizational complementarity, and economicmotivation on joint venture dissolution. Academy ofManagement Journal, 40, pp. 279–307.
Parkhe, A. (1991). Interfirm diversity, organizational learn-ing, and longevity in global strategic alliances. Journal ofInternational Business Studies, 22, pp. 579–601.
Parkhe, A. (1993a). ‘Messy’ research, methodologicalpredispositions, and theory development in internationaljoint ventures. Academy of Management Review, 18,pp. 227–268.
Parkhe, A. (1993b). Strategic alliance structuring: a gametheoretic and transaction cost examination of interfirmcooperation. Academy of Management Journal, 36, pp.794–829.
Pearce, J.A. II and Hatfield, L. (2002). Performance effectsof alternative joint venture resource responsibility struc-tures. Journal of Business Venturing, 17, pp. 343–364.
Pisani, N. (2009). International management research:investigating its recent diffusion in top managementjournals. Journal of Management, 35, pp. 199–218.
Podsakoff, P.M. and Organ, D.W. (1986). Self-reports inorganizational research: problems and prospects. Journalof Management, 12, pp. 531–544.
Pothukuchi, V., Damanpour, F., Choi, J., Chen, C.C. andPark, S.H. (2002). National and organizational culture
84 J. Christoffersen
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.
differences and international joint venture performance.Journal of International Business Studies, 33, pp. 243–265.
Ren, H., Gray, B. and Kim, K. (2009). Performance ofinternational joint ventures: what factors really makea difference and how? Journal of Management, 35, pp.805–832.
Reus, T.H. and Rottig, D. (2009). Meta-analyses of interna-tional joint venture performance determinants. Manage-ment International Review (MIR), 49, pp. 607–640.
Reus, T.H. and Ritchie, W.J. III (2004). Interpartner, parent,and environmental factors influencing the operation ofinternational joint ventures: 15 years of research. Man-agement International Review (MIR), 44, pp. 369–395.
Robson, M.J. and Katsikeas, C.S. (2005). International stra-tegic alliance relationships within the foreign investmentdecision process. International Marketing Review, 22,pp. 399–419.
Robson, M.J., Katsikeas, C.S. and Bello, D.C. (2008).Drivers and performance outcomes of trust in inter-national strategic alliances: the role of organizationalcomplexity. Organization Science, 19, pp. 647–665.
Robson, M.J., Leonidou, L.C. and Katsikeas, C.S. (2002).Factors influencing international joint venture perform-ance: theoretical perspectives, assessment, and futuredirections. Management International Review (MIR), 42,pp. 385–418.
Robson, M.J., Skarmeas, D. and Spyropoulou, S. (2006).Behavioral attributes and performance in internationalstrategic alliances. International Marketing Review, 23,pp. 585–609.
Sarkar, M.B., Echambadi, R., Cavusgil, S.T. and Aulakh,P.S. (2001). The influence of complementarity, compat-ibility, and relationship capital on alliance performance.Journal of the Academy of Marketing Science, 29, pp.358–373.
Saxton, T. (1997). The effects of partner and relationshipcharacteristics on alliance outcomes. Academy ofManagement Journal, 40, pp. 443–461.
Sim, A.B. and Ali, Y. (2000). Determinants of stability ininternational joint ventures: evidence from a developingcountry. Asia Pacific Journal of Management, 17, pp.373–397.
Simonin, B.L. (2004). An empirical investigation of theprocess of knowledge transfer in international strategicalliances. Journal of International Business Studies, 35,pp. 407–427.
Steensma, H.K. and Lyles, M.A. (2000). Explaining IJVsurvival in a transitional economy through socialexchange and knowledge-based perspectives. StrategicManagement Journal, 21, pp. 831–851.
Tomlinson, J.W.C. (1970). The Joint Venture Process inInternational Business: India and Pakistan. Cambridge,MA: MIT Press.
Voss, K.E., Johnson, J.L., Cullen, J.B., Sakano, T. andTakenouchi, H. (2006). Relational exchange inUS–Japanese marketing strategic alliances. InternationalMarketing Review, 23, pp. 610–635.
Yan, A. and Gray, B. (2001). Antecedents and effects ofparent control in international joint ventures. Journal ofManagement Studies, 38, pp. 393–416.
Yeheskel, O., Zeira, Y., Shenkar, O. and Newburry, W.(2001). Parent company dissimilarity and equity interna-tional joint venture effectiveness. Journal of InternationalManagement, 7, pp. 81–104.
Supporting informationAdditional supporting information may be found in theonline version of this article at the publisher’s web site:
Appendix S1. Sample (Word document)Appendix S2. Detailed findings (Word document)
Please note: Wiley-Blackwell are not responsible for thecontent or functionality of any supporting materials suppliedby the authors. Any queries (other than missing material)should be directed to the corresponding author for thearticle.
Antecedents of ISA Performance 85
© 2012 The AuthorInternational Journal of Management Reviews © 2012 British Academy of Management and Blackwell Publishing Ltd.