Interorganizational situations – An explorative typologyLove Börjeson *Graduate School of Education, Stanford University, CERAS Building, room 448, Stanford, CA 94305-3084, USA

A R T I C L E I N F O

Article history:Received 18 May 2014Accepted 9 January 2015Available online 23 January 2015

Keywords:Inter-organizational relationsCorrespondence analysisCooperationTypology

A B S T R A C T

Firms tend to engage in interorganizational relationships (IORs) to an increasing extent, for various reasons,in different configurations and at different organizational levels. Identifying and describing types of IORsis therefore important. The research objective of this study is to synthesize earlier research on IORs andthereby develop a typology that is both coherent and nuanced. This is achieved by exploring qualitativeinterview material regarding IORs using correspondence analysis. The results show that IORs can be groupedinto four distinct situations, each of which is characterized by a combination of relationships, and thecontent of the relationships. IOR situations are further shown to be distributed in a two-dimensional spacethat depends on both the immediacy level and the establishment level of the situation.

© 2015 Elsevier Ltd. All rights reserved.

Introduction

Although the overall degree to which firms engage ininterorganizational relationships (IORs) is surprisingly difficult tooperationalize and measure, repeated studies all indicate that IORis now considered an important phenomenon that continues to grow(Hagedoorn, 2002; Hagedoorn & Narula, 1996; Morris & Hergert,1987; Narula & Hagedoorn, 1999; Schmiemann, 2006, 2007). SinceIORs were first noticed by management scholars in the late 1980s(Hagedoorn, 2002; Hladik, 1985), attempts have been made to mapthe field by building typologies, sometimes deliberately and ex-plicitly and sometimes by using auxiliary results from relatedintellectual endeavors. Previous research can be broadly divided intotwo strands: one is oriented toward comparable and coherent ty-pologies, and the other is oriented toward extended and nuanceddescriptions of IOR types.

In the first research strand, the basis for the individual types ofIORs has typically been dependencies between independent and de-pendent variables, where sets of causalities form the different types.The methods used to analyze these causalities range from deduc-tive reasoning (cf. Narula, 2001) to comprehensive empiricalapproaches, such as multinomial logistic regression analysis (cf. Teng& Das, 2008). The independent variables examined are diverse andrange from strategic goals (Hagedoorn, 1993; Osborn & Baughn,1990; Teng & Das, 2008), to earlier IOR experiences and the inter-nationalization level (Kuo, Kao, Chang, & Chiu, 2012; Teng & Das,2008), changes in the business environment (Auster, 1987), the

competency profile of the firms involved (Amaldoss & Staelin, 2010),the technological profile of the firms (Narula, 2001), the inherentrisk of the IOR (Das & Teng, 1996), and the size of the firms (Osborn& Baughn, 1990). The dependent variable is usually the configura-tion of ownership (Auster, 1987; Das & Teng, 1996; Hagedoorn, 1993;Narula, 2001; Osborn & Baughn, 1990; Teng & Das, 2008); however,the relation is occasionally reversed, with the ownership configu-ration as one of the independent variables (Amaldoss & Staelin, 2010;Ertuna & Yamak, 2011).

The consistency in the view of the ownership configuration asthe dominating dependent and discriminating IOR variable in thisfirst research strand has resulted in a largely comparable set of ty-pologies and a subsequent coherent body of research. However, thedepicted types are predisposed to repeat existing knowledge re-garding firms rather than create new knowledge regarding IORs. Thisoccurs because the ownership configuration (the commonly useddependent variable in the aforementioned studies) reflects the per-spective of the firm involved, rather than the IOR itself. PreviousIOR typologies thereby inherit knowledge regarding firms; thisknowledge may have been thoroughly tested in an intraorganiza-tional context but taken for granted in the IOR context. From theparticipating organization’s perspective, the ownership structure ofan IOR may be its most distinctive characteristic; it may simulta-neously, however, be a peripheral characteristic of the IOR for theindividuals actively involved (Guo & Ng, 2011).

In accordance with the above objections a second and alterna-tive research strand, which is qualitatively oriented and based chiefly,but not exclusively, on case studies, has extended the list of charac-teristics that may constitute an IOR. Additionally, in doing so, thealternative research strand correspondingly abandoned sets ofcausalities as the fundamental base for different types of IOR. Inthis research, ownership is only one of many possibly critical

* Stanford University, CERAS Building, room 448, Stanford, CA 94305-3084, USA.Tel.: +1 650 714 91 61; fax: +1 650 725 7395.

E-mail address: loveb@stanford.edu.

http://dx.doi.org/10.1016/j.emj.2015.01.0010263-2373/© 2015 Elsevier Ltd. All rights reserved.

European Management Journal 33 (2015) 191–200

Contents lists available at ScienceDirect

European Management Journal

journal homepage: www.elsevier.com/ locate /emj

contingencies, which also include the IOR’s goal and the characterof decision making processes (Spina, Verganti, & Zotteri, 2002); theIOR’s scope and culture (De Man, Roijakkers, & de Graauw, 2010);the relative financial magnitude of the exchanged resources, criti-cality of the IOR, availability of alternatives and switching costs, andparties’ expectations and potential (Hansen & Rasmussen, 2013); thecharacter of interpersonal cooperation (Guo & Ng, 2011); the levelof cooperation and competition on the strategic and operational levels(Kylanen & Rusko, 2011); sources of external knowledge and knowl-edge content (Arvanitis & Hollenstein, 1996, 1998); the direction ofthe relationship (vertical or horizontal) and the nature of the inter-dependency (Dowling, Roering, Carlin, & Wisnieski, 1996); the qualityof the IOR as indicated by organizational and human assets and man-agement capabilities (Lahiri & Kedia, 2011); and the different phasesof cooperation (Coltman, Bru, Perm-Ajchariyawong, Devinney, &Benito, 2009; De Man et al., 2010; Guo & Ng, 2011; Hansen &Rasmussen, 2013; Kylanen & Rusko, 2011; Lehrer, Ordanini, DeFillippi,& Miozzo, 2012). In addition to illustrating the richness of contin-gencies that may influence IORs and inculcate the dynamic natureof IORs, this alternative research strand has adopted aninterorganizational view on IORs. The extension of the list of con-tingencies is consequently both quantitative (in terms of an extendedlist of contingencies) and qualitative, representing a shift in the per-spective from the participating firms to the actual IORs. Therefore,researchers have been able to break away from the intraorganizationalbody of research and essentially produce novel insights regardinginterorganizational organizing. However, in the absence of an inte-grating theory or common discriminating IOR variable, theseadvancements run the risk of being anecdotal and limited to spe-cific cases. Similarly, the sheer number of identified contingenciesmakes them difficult to generalize and integrate in any overarchingIOR typology.

This research objective of the study presented in this paper isto develop a typology of IORs (defined as any formal or informalcooperation between firms, each of which has its own formal hi-erarchy) by combining the strength of comparable and coherenttypologies with the richness and nuances of the extended list of con-tingencies produced in case studies.

This paper begins by describing how the ownership configura-tion has been perceived as a pivotal variable to discriminate betweenIORs and the effect that this perception has had on previously de-veloped typologies. The alternative and primarily case-based meansof typologizing IORs is then described, before concluding with thestrengths and weaknesses of the respective approaches. A synthe-sized approach is subsequently suggested, shifting the focus fromthe single IOR to situations of IORs. This shift is explained, justi-fied, and related to earlier research. The combination of the densequalitative data and explorative statistics that were used to arriveat a finite set of IOR situations are then described in detail. The resultsare presented with support in the statistical model and in inter-view excerpts, before ultimately determining the developed typology.The results are then discussed in relation to earlier research andmethodological considerations and limitations. Finally, the mainresults of the study are summarized.

The (alleged) importance of ownership configuration of IORs

In previous research oriented toward comparable and coher-ent typologies, the scale along which different types of IORs aretypically located ranges from contractual arrangements to joint ven-tures (Teng & Das, 2008); however, it occasionally stretches furtherinto full ownership (Pisano, 1991). The scale is an expression of own-ership configuration. The types along the scale, typically referredto as “form” (Parmigiani & Rivera-Santos, 2011) or “structure”(Teng & Das, 2008), include joint ventures, vertical relation-ships (buyer–supplier agreements, licensing and co-branding),

franchising, project-based strategic alliances, cross-sector partner-ships, networks, trade associations, and consortia (Parmigiani &Rivera-Santos, 2011). The IOR ownership configuration is recog-nized by both practitioners (i.e., managers) and scholars as havinga decisive effect on the IOR performance (Teng & Das, 2008; Yoshino& Rangan, 1995). However, performance is not the sole aspect ofIORs that is affected by the IOR’s ownership configuration. Das andTeng (1996) extend the argument and suggest that ownership con-figuration influences nearly every aspect of IORs, including controlmechanisms, operation processes, and exit possibilities. Conse-quently, the choice of form or structure, i.e., the choice of theownership configuration, in this earlier research strand has beenconsidered one of the most important decisions that the firms par-ticipating in an IOR must make (Killing, 1983; Teece, 1992). Giventhe alleged importance of the ownership configuration, it is no sur-prise that this is the overriding variable used to discriminate betweentypes of IORs in previous research. What is surprising, given thealleged effects of the choice of the ownership configuration, is thatthe ownership configuration is repeatedly treated as a dependentvariable (Auster, 1987; Das & Teng, 1996; Hagedoorn, 1993; Kuo et al.,2012; Narula, 2001; Osborn & Baughn, 1990; Teng & Das, 2008), thuslimiting the analysis to explaining the IOR’s ownership configura-tion. However, this does not explain the effects on IORs that theownership configuration may, in turn, have. In other words, manyresearchers consider the implications of the ownership configura-tion to be important; however, these implications are rarelyempirically investigated, despite a few exceptions where the de-pendency is reversed (Amaldoss & Staelin, 2010; Ertuna & Yamak,2011). Ownership configuration thus emerges as an integrating com-ponent and, subsequently, a natural point of comparison amongdifferent typologies.

Although researchers have been significantly more inventivewhen creating independent variables, the focus on ownership con-figuration and dependencies limits and streamlines the scope ofprevious attempts at typologies. However, there is a more funda-mental criticism to be invoked against the ownership configurationas a discriminating variable for IORs. First, the ownership config-uration is not only a property but also an expression of a firmperspective, examining the IOR from the participating firms’ per-spective. However, what is important from the firm’s perspectiveis not necessarily important from the IOR’s perspective. Owner-ship configuration thus runs the risk of being more of a mirror thana peep-hole; what it truly reflects is firms’ degree of financial in-volvement in the IORs rather than IOR properties (Guo & Ng, 2011).Consequently, examining the IOR from the firm perspective is likelyto produce images of firms instead of IORs. Second (and closelyrelated), cooperation within and between firms is a destabilizingactivity that jeopardizes established norms, knowledge, roles, strat-egies and forms (Browning, Beyer, & Shetler, 1995). The one-dimensional, angular types that follow the focus on ownershipconfiguration are therefore less suitable for delineating the elusivecharacter of IORs; IORs commonly change ownership configura-tion as they evolve, and many important IORs are devoid ofownership altogether (Bidault & Cummings, 1994; Werr &Linnarsson, 2004). In fact, many IORs are unknown to manage-ment in the participating firms (Marshall, 2004). A typology thattakes ownership configuration as an a priori discriminating vari-able is consequently likely to mirror the firms involved rather thanthe IORs. Furthermore, this typology is likely to fail to discrimi-nate between the subtleties of many IORs; it may completely failto include important variants of IORs.

Finding comparability among a plethora of contingencies

If the IOR from the firm’s perspective can have different own-ership configurations and be more or less profitable, the IOR from

192 L. Börjeson/European Management Journal 33 (2015) 191–200

the IOR’s perspective is a world of its own. A shift of perspectivewill therefore result in a wider range of characteristics relevant todescribing IORs, characteristics that cannot easily be arranged in de-pendency relationships. In the research strand taking the IORperspective on IORs, sets of independent and dependent variablesare thus typically replaced by concurrent contingencies. The char-acter of the single contingencies is also different, as contingenciesmay reflect characteristics that are either invisible or irrelevant fromthe firm’s perspective, but to the fore from the perspective of theIOR. As an example, consider a “problem difficult to articulate”(Lehrer et al., 2012, p. 506); although this contingency in the endmay influence something that “counts” from the perspective of thefirm (e.g., profitability), its obvious, immediate and inevitable im-portance is found within the IOR.

The multifacetedness of the extended list of contingencies in thisalternative strand of research is an academic gain in its own right.However, it also leads to interpretative challenges directly linkedto the very same gain; the sheer number of plausible contingen-cies may very well obscure fundamental similarities underlyingseemingly different IORs (Parmigiani & Rivera-Santos, 2011). Iden-tifying these latent similarities is therefore expected to be informativein relation to both earlier and future research. As discussed above,ownership configuration is a less adequate candidate as a startingpoint for comparison between IORs because it is blind for the manynuances of IORs and for informal IORs. All IORs do, however, involveparties. This actuality is therefore a fundamental similarity of all IORs,one that also holds meaningful variation; all IORs involve parties,but these parties differ between IORs. The “with whom-question”is consequently a natural and potentially fruitful point of compar-ison between IORs. An equally fundamental actuality is that IORsrelates to some sort of content. Be it trivial or sophisticated, periph-eral or pivotal; every IOR is about something. As does the “withwhom-question”, the “about what-question” represents both sim-ilarity (all IORs are about something) and potentially fruitful variation(not all IORs are about the same thing). Although the above actu-alities are deductively derived, they do have empirical supportregarding both those with whom firms cooperate (Arvanitis &Hollenstein, 1996, 1998; Dowling et al., 1996) and what they co-operate on (Arvanitis & Hollenstein, 1996, 1998; De Man et al., 2010;Lehrer et al., 2012; Spina et al., 2002).

For each individual IOR, answering the questions “with whom?”and “about what?” is an indispensable part of describing the par-ticular IOR in question. However, previous research has offerednumerous additional suggestions for potentially important IOR char-acteristics, such as the character (Guo & Ng, 2011) or quality (Lahiri& Kedia, 2011) of cooperation or the level of commitment and trustin the relationship (Coltman et al., 2009; De Man et al., 2010). None-theless, there are no indices of any additional single characteristicthat could alone provide unequivocal comparability between IORs.Instead, the bundle of “left overs” from earlier studies can be con-sidered eventualities that may be present in a particular IOR. The“with whom?” and “regarding what?” questions are thus comple-mented with the open question “what else?”

In combination, these three questions provide the basis fordeveloping a typology of IORs, as outlined below.

A synthesized typology – from IORs to situations of IORs

Briefly, there are two basic approaches to classifying IORs in pre-vious research: producing comparable and coherent types of IORsbased chiefly on the ownership configuration as the critical vari-able or producing case-specific, multifaceted types of IORs definedby concurrent contingencies. In accordance with the first ap-proach, nuances are lost and important variants of IORs excluded.In accordance with the second approach, the results are restrictedto the specific case and comparability is lost. In this study, I avoid

these weaknesses while still capitalizing on the strengths of pre-vious research and thus create a typology that allows for bothnuances and comparability. Because IORs are unstable (Bidault &Cummings, 1994; Werr & Linnarsson, 2004), a stabile typologymeeting this double requirement entails changing the unit of anal-ysis, from the single IOR to situations of IORs. The analytical gain withthis shift is that whereas IORs change characteristics as they evolve,the situations of IORs remain the same; therefore, the latter is a moresuitable foundation for constructing a typology.

Situational understanding has been promoted in earlier studiesof certain aspects of IORs, such as trust and risk (Jarvenpaa, Shaw,& Staples, 2004; Luo, 2002; Sitkin & Pablo, 1992) and identity con-structs (Zhang & Huxham, 2009). In this study, however, thesituational approach is extended to encompass all aspects of IORs.This is accomplished by using the previously described “withwhom?”, “regarding what?”, and “what else?” questions and turningthem into three variables:

(1) Partners: partners with whom firms cooperate;(2) Content: content the IOR revolves around; and(3) Supplementary contingencies: additional eventualities char-

acterizing the relationship.

It is the various combinations of these three variables that con-stitute the situations of IORs.

The approach can be considered a continuation of the dual ty-pology developed by Arvanitis and Hollenstein (1996, 1998) in whichsources of knowledge (“with whom”) and types of knowledge (“re-garding what”) form two separate sets of clusters. These researchersfurther theorize that the “with whom” and “regarding what” areinterrelated, i.e., that the two sets of clusters somehow corre-spond. The theorized interrelatedness is explored but with no definiteresults (Arvanitis & Hollenstein, 1996, 1998).

To reach further, I use correspondence analysis in this study(Greenacre, 1984, 2007). This is a statistical method that is suit-able for exploring the interrelatedness, i.e., correspondence, ofmultinomial data without predefined dimensions that also allowsqualitatively different variables to be displayed in the same geo-metrical space (Hair, Black, Babin, & Anderson, 2010). Hence,correspondence analysis is well suited for the purposes of this studybecause it allows for a joint exploration of the three variables con-stituting the situations of IORs and because the dimensionality ofthe situations is a priori unknown. The data provided for the anal-ysis are extracted from an index based on interviews regarding IORs.Therefore, similar to the chosen analysis, the data are in accor-dance with the double requirement of a typology that is bothnuanced and comparable; interviews provide nuances, and the indexprovides comparability. The method and data are comprehen-sively described in the next section of the paper.

Research design

The study is based on interviews regarding IORs. Interviews arecoded to an index based on content. Overlaps in the index, result-ing from overlaps in the interviewees accounts regarding IORs, i.e.,when interviewees mention something in the context of some-thing else, are extracted to an intersection table. Interrelatedness,which is indicated by the overlaps in the intersection table, is ex-plored using correspondence analysis; the results are graphicallydisplayed in a perceptual map. The perceptual map and the under-lying interview excerpts are the foundation for interpretation of thevarious types of IORs. Thus, the research process can be summa-rized by the figure below: the research material is in the lower row,and the methodological activities are in the upper row (italicized)(Fig. 1).

The research process is described below.

193L. Börjeson/European Management Journal 33 (2015) 191–200

Interviews

The study presented in this paper builds on an index devel-oped from 31 interviews with managers and professionals in sevenSwedish firms. Pursuing this study’s explorative objective, the firmswere selected to present a broad sample and represent a variety ofindustries and firm sizes, ranging from large traditional manufac-turing firms to small modern software development and consultingfirms (Table A1, Appendix A). The same principle was applied to se-lecting interviewees. In each firm, three to five interviews wereconducted, and the interviewees represented general manage-ment, marketing, production and logistics, purchasing, humanresources and research and development (when applicable – notall firms had all of these functions).

The point of departure for the semi-structured interviews(Fontana & Frey, 1994) was the “with whom?”, “regarding what?”and “what else?” questions (cf. introductory section of the paper).Interviewees were first asked to identify interorganizational rela-tionships in which they participated or were affected. For the tenmost important relationships (according to the interviewees’ ownestimates), follow-up questions were asked concerning the contentaround which the relation (r)evolved and other properties or char-acteristics of the relationship that the interviewee could recall, suchas the origin of the relationship, the medium most frequently usedfor interaction, the degree of formalization of the relationship, andenablers and challenges of the respective relationship. Interviewslasted between 40 and 90 minutes and were digitally recorded andtranscribed.

Coding and index

A semi-selective coding procedure (Strauss & Corbin, 1990) wasused to code the interview transcripts with respect to variables 1–3determined in a previous section:

(1) Partners(2) Content(3) Supplementary contingencies

Thus, the coding procedure was selective in subdividing codesaccording to the predefined variables 1–3. However, the coding pro-cedure was not selective regarding the actual codes under eachvariable emergent from the interview material. Thus, exactly whichpartners, the content and any possible other characteristics foundunder each variable were based on the interviews.

Coding was initially conducted by two researchers; passages inthe interview material that presented critical and difficult delimi-tations were co-coded to establish common procedures and ensurehigh interpersonal consistency. The entire material was then recodedby a single researcher in a confirmatory coding process to furtherenhance the quality of the final index of codes.

Intersection table

A single code in the index can be considered a label attachedto the underlying content (Ryan & Bernard, 2000). However, becausecodes represent different content mentioned with unequal frequen-cy in the interviews, each code not only labels the underlyingcontent but also contains information regarding its importance, as

measured by the number of words assigned to it. Coding there-fore provides a quantitative feature to the otherwise qualitativeinterview material. In total, the index contains 158,741 words dis-tributed over 110 codes.

Because different content in the interview transcripts typicallyoverlap, the codes in the index do the same. Overlap consequentlyoccurs when interviewees mention something in the context ofsomething else. Thus, codes are not mutually exclusive, and sen-tences can, accordingly, be coded several times (but only to differentcodes). Overlaps from the index are extracted to an intersection table(Table B4, Appendix B), where the columns represent variable 1 (thepartners with whom firms cooperate) and the rows represent vari-ables 2 and 3 (the content around which the IOR revolves and thesupplementary contingencies characterizing the cooperation, re-spectively). Each cell in the intersection table shows the numberof words in that particular overlap, i.e., the number of words usedby interviewees to mention something particular in the context ofsomething else particular. Codes from the index that are not prag-matically significant, i.e., without explanatory value relative to thestudy’s purpose (Hair et al., 2010), are not included in the inter-section table. Thus, the intersection table contains 26 of the 110codes in the index (Table B4, Appendix B).

Correspondence analysis

The number of words in each cell in the intersection table pro-vides a numerical measurement of the interrelatedness between theparticular column and row that overlap (Börjeson, 2011), i.e., betweenthe variables that overlap. Therefore, the intersection table as a wholeholds information regarding the overall patterns of interrelated-ness among variables 1–3. These patterns are explored usingcorrespondence analysis (Greenacre, 1984, 2007). The analysis resultsin an a priori unknown dimensional space that is based on the linearadjustment of the chi-square distances between the observed in-tersection table and an ideal projected table (based on symmetricalmarginal probabilities) that shows how the table would appear ifthere were no interrelatedness between columns and rows.

Variables 1 and 2 are active in the statistical model, i.e., in de-fining the a priori unknown dimensional space. The supplementaryvariable 3 is then inserted without affecting the model, solely addinginterpretative information (Greenacre, 1984). The robustness andthe quality of the model are captured in several metrics (as re-ported in Appendix C): the significance of the chi-square distances;the inertia for each dimension (i.e., the contribution to the totalinertia of the model from each dimension); the contribution to theinertia of the dimensions from each variable value, and converse-ly, from each dimension to the inertia of each variable value; andthe correlation between dimensions included in the model.1 Theactual results of the analysis are reported numerically in variablevalue scores (Appendix C) and graphically in a perceptual map (Hairet al., 2010); close proximity between variable values in the maprepresents a high degree of correspondence, and distance repre-sents a low degree of correspondence. Variable value scores, i.e., thepositions of each variable value in the perceptual map, are the

1 Inertia is what similar statistical models (e.g., principal component analysis) typ-ically refer to as variance. Variance and inertia differ formally, but their interpretationis very similar.

coding extraction correspondence analysis interpretation

Interviews Index Intersection table Perceptual map Situational typology

Fig. 1. Research process.

194 L. Börjeson/European Management Journal 33 (2015) 191–200

foundation for interpreting the results. Zero scores in respective di-mensions are used as thresholds to further discriminate among thegroups’ variable values (Greenacre, 1984; Hair et al., 2010).

In this study, where the intersection table is based on an inter-view index, substantial additional interpretative support is providedby the interviews. The correspondences carved out by the statisti-cal analysis are described in the interviewees’ words, as derived fromthe transcripts that underlie each cell in the observed intersectiontable. Each variable value is accordingly described by theinterviewees, thereby supporting the interpretation.

Results

Overall result and model fit

The results from the correspondence analysis show that the in-terrelatedness between rows and columns in the intersection tableare indeed highly significant (p-value < 0.0001, cf. Table C1, Appen-dix C). Whom the firms cooperate with is thus significantlyinterrelated with the content of the relationship and with other sup-plementary contingencies of the relationship.

The first two dimensions in the statistical solution accounts for85% of the inertia (52% + 33%, respectively, cf. Table C2, Appen-dix C). The third dimension accounts for an additional 13% of theinertia. In accordance with Hair et al. (2010), this and higher di-mensions are therefore excluded from the solution. The correla-tion between the two included dimensions is also found to be

sufficiently low to provide a stable model (Table C2, Appendix C).The fit between the included codes and dimensions is generally high,with only three codes for which the two dimensions explain lessthan 50% of their inertia. Explanatory values are still sufficiently high(>40%, cf. Table C3 & C4 in Appendix C) to include the codes in themodel (Hair et al., 2010). The metrics thus indicate a good modelfit, both overall and for the individual variables (cf. Appendix C).

The results of the correspondence analysis are presented in a per-ceptual map (Fig. 2 below) and can be conceptually divided in twoparts: dimensions along which the variable values of IORs are dis-tributed and situations of IORs, consisting of groupings of variablevalues. The interpretation (including labeling) of the dimensionsdeparts from the highest and lowest scored variable values in eachdimension. The interpretation of the situations is based on the zeroscore of the dimensions, which divides the perceptual map into fourquadrants, and on the relative proximity between variable values(where high proximity indicates a high positive interrelatednessbetween variable values). The interpretation of the situations isfurther substantiated by excerpts from interviews that are codedto the relevant intersections in the interview index.

Interpreting the dimensions of IORs

The first dimension in Fig. 2 is an expression of the level ofimmediacy. Complications, successes or failures of the IOR at therightmost end of Dimension 1 (positive scores) have immediateeffects for the parties involved (typically customers, suppliers or

SUPPLIERS

BUSINESS NETWORKS

PRIVATE NETWORKS

CUSTOMERS

GROUP COMPANIES

Attenuated relations

Dependency

Betrayal

Comfortable relations

TrustReciprocity

Strategic

TacticContract

Acquire knowledge

Create relations

Formal systems

Too informal

-1.5

-1

-0.5

0

0.5

1

1.5

2

-2.8 -2.3 -1.8 -1.3 -0.8 -0.3 0.2 0.7 1.2

Dim

ensi

on 2

-L

evel

of

Est

abli

shm

ent

Dimension 1 - Level of Immediacy

VARIABLE 1: PARTNERS

Variable 2: Content

Variable 3: Supplementary contingencies

Relations

Management

Product development & Business environment

Transformation pressure

Products & production Customer needs

Product & service refinement

Business skills

Fig. 2. Dimensions and situations of IORs.

195L. Börjeson/European Management Journal 33 (2015) 191–200

business networks): frictionless cooperation between suppliers andcustomers may lower costs and increase efficiency, with immedi-ate positive effects on operative margins; daily revenues may plungeif a supplier fails to deliver input material to a production process;and investments in product development may be instantly foiledby a betrayal by someone in the business network. The immedia-cy associated with positive scores in Dimension 1 is furtherintensified by the transformation pressure that forces IOR forma-tion, in which the relationships are characterized by dependencyrather than mutual interests. To hedge against the immediacy, firmsuse contracts and trust to stabilize the IORs, a strategy that is mir-rored by the occurrences of betrayal.

At the other end of the first dimension (negative scores in Di-mension 1), IORs may be equally important, but they do not havean instantaneous impact for the parties involved. Relationships areeither taken for granted (within a business group) or do not rep-resent any considerable investment (in private networks), nullifyingboth positive and negative immediate effects of the IORs. Attenu-ated relations at the leftmost end of Dimension 1, however, indicatethat the lower level of immediacy can be mistaken by the partiesfor a lack of importance, leading to the mismanagement of rela-tionships and thus to unforeseen long-term negative effects.

The second dimension, the level of establishment of the IOR, ismore multifaceted. This dimension may be interpreted as an in-clusive heading for different expressions of the level of establishment.The level of establishment is expressed in the degree of formal reg-ulation, in which the contract and a common parent companycorrespond to positive Dimension 2 scores (upper part of Fig. 2),whereas the relationships in different networks are less formal andtend to have negative scores (lower part of Fig. 2). The level of es-tablishment is also expressed in time (tactical horizon correspondsto a high level of establishment, strategic horizon corresponds toa low level of establishment), relations (well-defined relations, e.g.,customer–supplier relations, correspond to a high level of estab-lishment, explorative relations in private networks correspond toa low level of establishment) and materialization (existing prod-ucts that are refined correspond to a high level of establishment,whereas possible future products under development correspondto a low level of establishment). The variable value that has thelowest level of establishment in Fig. 2 is the supplementary con-tingency Create relations. Creating relations is important, not inrelation to already existing products, services, customers, and sup-pliers or sister companies but in relation to future versions of thefirm.

Interpreting the situations of IORs

Using the zero scores of Dimensions 1 and 2 in Fig. 2 as thresh-olds (Greenacre, 1984; Hair et al., 2010), the variable values in theperceptual map can be divided into four situations of IORs. Each sit-uation thus corresponds to a quadrant in the perceptual map. Thesituations describe groups of variables that revolve around stronginterrelatedness between the partners with which firms cooper-ated and the content of this cooperation. Hence, Private network,for example, is strongly interrelated with content regarding Rela-tions but only modestly interrelated with content regardingManagement. In other words, in private networks, content regard-ing relationships, e.g., who is in contact with which firm or whoknows what about which things, is more commonly exchanged thanis content regarding how to manage a business. An exception to thisunambiguous pattern is Business skills (content concerning how torecognize and react to a business opportunity). Examining the per-ceptual map and the intersection table (Table B4, Appendix B),Business skills appears to be related to two types of relations: Privatenetworks and Group companies.

The supplementary contingencies in Fig. 2 display more varieddegrees of interrelatedness with the other variables. Based on thezero scores of the dimensions, all supplementary contingencies canbe attributed to the situations described below.

Tentative initiationThe Tentative initiation of IORs is characterized by a low level of im-

mediacy and a low level of establishment. The typical relationship inthis situation occurs within a private network, e.g., by coincidence, aperson meets a former colleague and discusses eventual future commonprojects. Relationships in the private network can also be more con-sciously managed and planned; however, they are linked to individualsrather than to the individual’s role in a firm (excerpt from the indexintersection Private network/Relations/Create relations):

The mingle factor. To be seen. Becoming someone is really im-portant. No one knows who [the firm] is because we’re neverout and about. Few here are amused by it, but I talk a lot aboutit. I think it’s one of the most important ways to get new cus-tomers. If you meet a person in [a bar] one night. . .. Go to [a bar],mingle, talk. It’s crucial. If you’d rather go home and makedinner . . . well, you don’t have to do it every night. (Consul-tant, Firm H)

It is difficult to note the exact content exchange outlined in thequotation above. The interpretation of this quotation and othersimilar ones is that the content of the relationship is other rela-tionships: their importance, how to combine them, and how to createand uphold them. The purpose of the relationship, i.e., the purposeof visiting a bar after work when you would prefer to go directlyhome, is to create relationships, in this case, to make the firm visible.

Without high levels of immediacy, the relationship is allowedto be explorative and strategic (excerpt from the index intersec-tion Private network/Strategic):

Being an entrepreneur is really great, but you also need to getout there, and to see: What does the market really need? Talkto people. Large networks. There’s not a general manager [in ourmarket segment] in the country that I haven’t talked to. . . I’vecontacted every single middle manager, and I basically do it everyyear. Go out and talk to them, stop by, call them, buy them lunchsometimes. . . Large networks of contacts to discuss and hear whatkind of problems they struggle with. (Business Developer andSenior Consultant, Firm G)

The exploratory element of the relationships in the above excerptoccurs because not much is immediately at stake; should one of thebusiness lunches mentioned fail to deliver, the negative effect maybe difficult to notice from the firm’s perspective. However, it is alsoclear (though more indirectly) that the formation and mainte-nance of the described relations are crucial for the long-termcompetitiveness of the firm. In the above excerpt, the intervieweeequates his networking activities with his business developmentactivities.

Deepening investmentsThe situation of Deepening investments is usually realized in busi-

ness networks and typically revolves around content related to thedevelopment of new products and how that relates to both whatis going on in one’s own industry and the more general businessenvironment. What are the products being developed by competi-tors and/or business partners, and what are the major trends in theindustry and market?2

2 Incremental refinement of already existing products and services is, in contrast,a type of content that is exchanged in relations between customers and suppliers;see below.

196 L. Börjeson/European Management Journal 33 (2015) 191–200

Although there are relative risks involved in new product de-velopment due to the required investments, the situation remainsdifficult to regulate with contracts, simply because it is difficult todetermine exactly what to regulate. This creates the combinationof a high level of immediacy and a low level of establishment thatcharacterizes the situation. For those involved, the combination isperceived as a problem of informality; much is at immediate stakewithout being secured by contracts. The informality problem ishandled with trust, but it also increases the risk of betrayal. Be-trayal, or perhaps more commonly, the “betrayal-like” use ofinformation, is a constant problem (excerpt from the index inter-sections Business network/Product development and Businessenvironment/Betrayal):

[. . . There is a risk from] a product development perspective inthese situations, namely, that you sit on a large amount of ex-clusive information. You might accidently give it away, or promotean issue so hard that your partners and potential competitorscan figure out how future products look and so on. On the otherhand, I read my partners or competitors in the same way. (Busi-ness Division Manager 2, Firm A)

Trust and betrayal occur in tandem, and “a confidence” may ac-tually rely on the fact that everyone knows each other’s agendasto gather as much information as possible and reveal as little as pos-sible in the process. The consequences of trustworthy behavior or,indeed, betrayal are immediate for the business potential of theproduct/service under development, and the interviewee mustbalance the benefit and risk of the IOR.

Commercial realizationCommercial realization is characterized by a high level of imme-

diacy and a high level of establishment. The levels of immediacyand establishment are interrelated; a high level of immediacy leadsto a desire to regulate the relationship with contracts, which, in turn,is an expression of a high level of establishment. Commercial real-ization typically occurs in relationships between suppliers andcustomers, and the content is related to existing products and cus-tomers. Relationships among suppliers, the focal firm and customersare deeply integrated, and a relationship with a supplier may revolvearound customer needs. An excerpt from the index intersectionSuppliers/Products & Production/Customer needs may be similar tothis:

The contact with the [supplier] is made when we get requestsfrom our customers that we can’t handle. Usually, we commu-nicate for a long time with [the supplier] to talk about what theactual welding should look like. . . This is not done by the en-gineer, but the welder himself, the guy in the workshop, doesthis. I just leave it all to the welder. (Business Division Manager,Firm F)

The problems in this situation are practical and have practicalsolutions, which emphasizes the high level of establishment; thedepicted discussion concerns products that are currently devel-oped and “in place” from an engineering perspective, and thesolutions under discussion are constrained by existing customerspecifications. In the above excerpt, the importance of the involve-ment of the practically oriented co-worker is consequently stressed.

High levels of immediacy in the situation mean that failure tomeet contracted requirements can lead to the immediate termi-nation of the relationship, as expressed in the index intersectionCustomer/Contract:

There are agreements with the retailers. They must meet anumber of requirements. They should have certain resources,partake in certain training, and sell a certain number of [units]and so on. It is very strictly regulated. If they don’t meet our

requirements, it’s quite easy for us to terminate the contract.(Marketing Director, Firm D)

Retailers are simply not allowed to make mistakes that woulddiminish the end-customer sales numbers; therefore, attempts aremade to formally regulate the retailers’ required competency. Shouldthe retailers fail to meet the requirements, the relationship will beterminated. Contracts thus have a double connotation. On the onehand, they stabilize the relationship by determining who does what;on the other hand, they provide exit options that can void the re-lationship altogether.

A high level of immediacy in the form of pressure to transformdue to changes in the business environment can also force anduphold relationships, even when such relationships create depen-dencies that are potentially harmful (excerpt from the indexintersection Supplier/Transformation Pressure/Dependency):

We have a set of products today that are very mechanically ori-ented, but at the same time, there’s strong growth in electro-mechanical products. So until we recruit the expertise we needin electronics and software development, we’re very depen-dent on subcontractors just to keep pace. (Business DivisionManager 1, Firm A)

Strong growth in a developing market segment (electromechani-cal products) creates ties between the firm and its suppliers. Theresulting relationship is characterized by dependency rather thanby mutual interests; in lieu of (and pending) internal product de-velopment, the interviewee’s firm buys into the latest developmentthrough its subcontractor. The core of the dependency is the factthat a company can neither know nor do everything and that it there-fore must seek external expertise and resources to solve relativelyvital tasks.

Established managementEstablished management of IORs is characterized by a low level

of immediacy and a high level of establishment. Rather than real-izing products commercially, the situation is characterized bymanagement of the relationship between currently deeply in-volved customers and suppliers or firms in the same business group;the latter represents how close two firms can become without be-coming a single firm. The content in this situation involvesManagement (excerpt from the index intersection Group companies/Management):

It starts with the HR department signaling some sort of need.It may be the result of an intense period of new recruitment orreorganization. Because we’re quite big, we sort it out within thebusiness group, management training, different kinds of edu-cation; it’s actually an ongoing activity. You bring in a larger group,run it for a couple of days focusing on basic leadership issues.(Marketing Director, Firm B)

The firms in the business group are separate from each other,but when they cooperate, they behave as a single organization. Inthe above quotation, the interviewee considers management andleadership training to be internal, though these practices are for-mally treated as a transaction between two firms (yielding, forexample, an invoice handled by the accounting department).

The combination of low immediacy and high establishmentcreates problems that mirror those in a Deepening investments sit-uation: high establishment creates comfortable relations, and lowimmediacy makes for attenuated relations (excerpt from the indexintersection Group companies/Comfortable relations/Attenuatedrelations):

Looking at the sister here in the building [sister company], it israther special because we’re in the same building and in the samebusiness group. It has both advantages and drawbacks. The

197L. Börjeson/European Management Journal 33 (2015) 191–200

advantage is closeness, familiarity, the fact that we speak the samelanguage and know each other’s businesses, etc. The drawbackis that the customer/supplier relationship becomes slightly at-tenuated because you basically deliver to yourself. You need tobe observant about that.. . . When measuring delivery prompt-ness and precision, well, then we ourselves are by far the worst.The further away the sister companies, the better the preci-sion. (Business Division Manager 2, Firm A)

Closeness makes the relationship easy, but because nothing istruly at stake if someone does not deliver on time and accordingto specifications, this simplicity is matched by a low priority andattenuated relationships.

A situational typology of IORs

The situational typology of IORs developed from the above in-terpretation of dimensions and situations of IORs is summarizedin Fig. 3.

Fig. 3 shows that situations of IORs are, similar to any actual IOR,multifaceted and rich in characteristics. However, dissimilar to any actualIOR, these situations are also stable in their establishment of charac-teristics. Therefore, a certain combination of immediacy andestablishment is consistently interrelated with certain partners and con-tents that are typical for the situation and with certain additionalcontingencies. The situations can accordingly be used to both identifyand typologize single IORs and to predict the challenges that those

involved in IORs will likely meet in various situations. In other words,managers and coworkers involved in IORs should expect to face dif-ferent but predictable difficulties, depending on the level of immediacyand establishment of the situation because a high level of immediacyadds risk to the relationships. This can be handled with formaliza-tion, resulting in a higher degree of establishment. However,formalization can also result in the conservation of harmful depen-dencies when transformation pressure is high. Trust substitutes contractswhen formalization is not an option, e.g., when it is difficult to knowwhat to formalize. However, trust induces risk of betrayal. If, converse-ly, immediacy is lower while establishment remains high, relationshipsmay be perceived as comfortable to the extent that they are ne-glected. The result is slack and mismanagement.

The only IOR situation that does not appear to be inherently prob-lematic is Tentative initiation, a situation combining a low level ofimmediacy with a low level of establishment. This is a crucial IORsituation in the longer (strategic) perspective because it is in thesesituations that the firm can create and maintain interfaces with itsenvironment. Nonetheless, this is a situation that hinges on the ini-tiative and network of the single coworker in his or her capacityas an individual (rather than as a professional). Therefore, manag-ing Tentative initiations may be difficult and potentially harmful; thetentative quality of the situation is dependent not only on the lowlevel of immediacy but also on the low level of establishment (un-derstood here as the absence of management), which allowsrelationships and exchanges in this situation to be sufficiently spec-ulative and non-committal.

Hig

h es

tabl

ishm

ent l

evel

ESTABLISHED MANAGEMENT

Involved partners: Group companies

Exchanged content: Management

Supplementary contingencies:Attenuated relations

Comfortable relations Formal systems

Tactic

COMMERCIAL REALIZATION

Involved partners: Suppliers

Customers

Exchanged content: Products & production

Product & service refinement Customer need

Supplementary contingencies:Contract

Transformation pressure Dependency

Low

est

abli

shm

ent l

evel

TENTATIVE INITIATION

Involved partners: Private network

Exchanged content: Relations

Supplementary contingencies:Create relations

Strategic perspective Reciprocity

DEEPENING INVESTMENTS

Involved partners: Business network

Exchanged content: Product development Business environment

Supplementary contingencies:Too informal

Trust Betrayal

Acquire knowledge

Low immediacy level High immediacy level

Fig. 3. A situational typology of IORs.

198 L. Börjeson/European Management Journal 33 (2015) 191–200

The situational typology may also support firms when strategizingaround a specific IOR. For example, for a firm involved in a situa-tion of deepening investments that are about to bring the result ofthe IOR to the market, the typology clearly suggests that this stephas far reaching consequences for the IOR: business partners, in-formality and trust will be replaced by suppliers, customers, contractsand dependencies. Anticipating the extent of this transformationwill allow the firm to adjust its strategy accordingly to “finish” wherethey want to in the new situation avoiding, for example, being in-terlocked in an unprofitable relationship based on the dependencyof another partner. Similarly, the typology may help firms identifyseemingly well-functioning relationships that are, in fact, stag-nated and unproductive. The “taken-for-grantedness” of such arelationship can then be challenged, and the continuation of a re-lationship can be more deliberately considered.

Discussion

Linkages to typologies of IORs in earlier research

On a general level, the legacy from the earlier two main ap-proaches to typologize IORs – by producing comparable and coherenttypes of IORs based chiefly on the ownership configuration as thecritical variable (Amaldoss & Staelin, 2010; Auster, 1987; Das & Teng,1996; Ertuna & Yamak, 2011; Hagedoorn, 1993; Narula, 2001; Osborn& Baughn, 1990; Teng & Das, 2008), or by producing case-specific,multifaceted types of IORs that are defined by concurrent contin-gencies (Arvanitis & Hollenstein, 1996, 1998; Coltman et al., 2009;De Man et al., 2010; Dowling et al., 1996; Guo & Ng, 2011; Hansen& Rasmussen, 2013; Kylanen & Rusko, 2011; Lahiri & Kedia, 2011;Lehrer et al., 2012; Spina et al., 2002) – are advanced through thesituational typology developed in this study because the identi-fied situations of IORs are nuanced, multifaceted and comparable.

On a more specific level is the core of the situations identifiedin the present typology, which are composed of the combinationof partners and relationship content. This finding is statistically strongand conspicuous in the perceptual map of the correspondence anal-ysis (Fig. 2). This finding thereby concludes the attempts by Arvanitisand Hollenstein (1996, 1998) to combine clusters of knowledge andclusters of relationships; the correspondence that they theorizedis indeed empirically verified (albeit with some conceptualalterations).

In addition, in relation to typologies based on the ownership con-figuration of IORs, an in absentia argument can be made. Similar toGuo and Ng (2011) but dissimilar to previous research in this strand(Das & Teng, 1996; Killing, 1983; Teece, 1992; Teng & Das, 2008;Yoshino & Rangan, 1995), I find no direct support in the empiricalmaterial for the importance of the IOR’s ownership configuration.This result should not be overstated because the ownership con-figuration may be implied by the type of partners engaged, e.g., arelationship within a private network, for example, is void of own-ership, in contrast to relationships within a business group. However,the ownership configuration as a primary variable to discriminatebetween IORs must be considered very limiting and, from the IOR’sperspective, less appropriate.

The results, as well as earlier research, invite reflection regard-ing the role of time in IORs. It has been suggested that time playsan important role as a variable (in the form of different stages orphases of IOR; cf. Coltman et al., 2009; Lehrer et al., 2012) or, morecommonly, as a factor for variable variation (De Man et al., 2010;Guo & Ng, 2011; Hansen & Rasmussen, 2013; Kylanen & Rusko,2011). What this study suggests is that because situations are fun-damentally different, in terms of both characteristics and the involvedpartners, purpose and content of the IOR, transitions between situ-ations are likely to involve major challenges.

Methodological considerations, limitations and future research

By applying correspondence analysis on intersections ex-tracted from indexed interviews, a priori unknown and/or poorlydefined but crucial variables and contingencies of IORs can be in-corporated into the analysis with precision and receptiveness. Thisis achieved because the delimitations of the variables and contin-gencies of IORs in this study are construed by the interviewees andbecause the “inexactness” and “wordiness” of their accounts are usedanalytically through the overlaps in the interviews, which resultsin the intersections in the index that underlies the correspon-dence analysis. The categorical fuzziness in the interviewees’ accountsthus becomes the central analytical element of the analysis.

The application of correspondence analysis to the interview datafurther paves the way for a fruitful interplay between the qualita-tive material and the quantitative analysis that strengthens thevalidity of the analysis; because the correspondences carved out bythe statistical analysis are described in the words of the intervieweesin the transcripts that underlie each cell in the observed intersec-tion table, the validity of an interpretation of a statisticalcorrespondence can be empirically verified. Hence, the immedi-ate access to the interview transcripts allows the interpretation tobe more of a listening game (reading) and less of a guessing gamethan would occur with other types of underlying material (e.g. ques-tionnaire data; c.f. Hair et al., 2010, for an elaboration on typicalinterpretative limitations related to correspondence analysis).

However, benefits of the chosen method are matched by draw-backs. Although the validity is higher in this study than in a similarstudy based on questionnaires, the reliability is lower and relies onthe rigidity of the coding process. The richness of the interview ma-terial is also made possible by the relatively small sample size, whichlimits the generalizability of the study. Although the number ofwords in the intersections in the index are more than sufficientlylarge to provide a stable statistical solution, a reasonable assump-tion remains that representations underlying the intersections inthe index are to some extent contingent, not only on different situ-ations of IORs but also on the interviewees and firms in the study.Furthermore, although sampling is aimed at producing a broadsample with respect to firm size and industry (cf. Table A1, Appen-dix A, for details regarding the firms and interviewees included inthe study), due to the small size of the sample, the remaining skewis likely to cause skew in the results. Any generalization to a largerpopulation must thus be made with care, and the statistical cer-tainty in terms of the overall fit of the statistical model applies tothe sample. However, given the broadness of the sample and thethoroughness of the coding procedure, the image of IOR situa-tions presented in Fig. 2 is a probable one, although not a likely 100%average.

The methodological limitations described above point to a naturalcontinuation of this study. Because the validity of the variables in-cluded in the analysis are relatively well-established empirically inthis and earlier studies, it is possible to leave the exploratory stageof investigations and continue with a confirmatory approach basedon more economical collection of data, e.g., questionnaires (in favorof interviews).

Conclusion

The remaining challenge from previous attempts to typologizeIORs is how to maintain comparability across types while still cap-turing the multifaceted subtleties of IORs in the same typology. Ihandle this challenge by shifting the unit of analysis from the singleIOR to situations of IORs that include the involved parties, the ex-changed content and supplementary contingencies. The analyticalgain with this shift is that whereas IORs change characteristics asthey evolve, IOR situations remain the same; therefore, the latter

199L. Börjeson/European Management Journal 33 (2015) 191–200

is a more suitable foundation for construing a typology. The dataand method are combined to achieve the same objective, in whichinterviews provide rich, nuanced material, whereas coding, indexand correspondence analysis provide comparability. Consequent-ly, the resulting typology over IOR situations is both comparable andnuanced.

The results show that IORs can be divided into four distinct situ-ations: Tentative initiation, Deepening investments, Commercialrealization and Established management. Situations of IORs are dis-tributed along two dimensions: the level of immediacy and the levelof establishment.

Tentative initiation is characterized by low levels of immediacyand establishment, and it is typically enacted in private networks,evolving around and aiming at the creation of new relations.

Deepening investments is characterized by a high level of imme-diacy and a low level of establishment. This is enacted in businessnetworks and typically involves knowledge regarding current in-dustry developments (e.g., development of new products). A typicalfeature of Deepening investments is the lack of formal regulation ofthe relationships, despite the high level of immediacy. This is com-pensated with trust, which, in turn, increases the risk of betrayal.

Commercial realization is characterized by high levels of imme-diacy and establishment. This typically involves formally regulatedrelationships between customers and suppliers, and the knowl-edge exchanged concerns products, customers’ immediate needs,and product refinement.

Finally, Established management is characterized by low immedia-cy and high establishment. It is typically enacted within a business group,and the possibilities and challenges of the situation mirror those of Deep-ening investments; relationships are typically relaxed and carry low riskbut are taken for granted and occasionally neglected.

Appendix: Supplementary material

Supplementary data to this article can be found online atdoi:10.1016/j.emj.2015.01.001.

References

Amaldoss, W., & Staelin, R. (2010) Cross-function and same-function alliances: Howdoes alliance structure affect the behavior of partnering firms? ManagementScience 56(2), 302–317.

Arvanitis, S., & Hollenstein, H. (1996) Indicators and determinants of innovativeactivity in Swiss manufacturing: An empirical analysis based on survey data. InInnovation, Patents and Technological Strategies. Organization EconomicCooperation & Development, Paris.

Arvanitis, S., & Hollenstein, H. (1998). Innovative activity and firm characteristics –a cluster analysis with firm-level data of Swiss manufacturing. Presented at theEARIE 25th Annual Conference, Swiss Institute for Business Cycle Research.

Auster, E. R. (1987) International corporate linkages – dynamic forms in changingenvironments. Columbia Journal of World Business 22(2), 3–6.

Bidault, F., & Cummings, T. (1994) Innovating through alliances – Expectations andlimitations. R & D Management 24(1), 33–45.

Börjeson, L. (2011) Förtroendets Organiseringsmetod: Samarbete, Svek och Dilemman.Mälardalens högskola, Västerås.

Browning, L. D., Beyer, J. M., & Shetler, J. C. (1995) Building cooperation in acompetitive Industry – Sematech and the semiconductor Industry. Academy ofManagement Journal 38(1), 113–151.

Coltman, T., Bru, K., Perm-Ajchariyawong, N., Devinney, T. M., & Benito, G. R. G. (2009)Supply chain contract evolution. European Management Journal 27(6), 388–401.

Das, T. K., & Teng, S. B. (1996) Risk types and inter-firm alliance structures. Journalof Management Studies 33(6), 827–843.

De Man, A. P., Roijakkers, N., & de Graauw, H. (2010) Managing dynamics throughrobust alliance governance structures: The case of KLM and Northwest Airlines.European Management Journal 28(3), 171–181.

Dowling, M., Roering, W. D., Carlin, B. A., & Wisnieski, J. (1996) Multifacetedrelationships under coopetition – Description and theory. Journal of ManagementInquiry 5(2), 155–167.

Ertuna, B., & Yamak, S. (2011) Foreign equity configurations in an emerging country:Implications for performance. European Management Journal 29(2), 117–128.

Fontana, A., & Frey, J. (1994) Interviewing: The art of science. In N. Denzin & Y. Lincoln(Eds.), Handbook of Qualitative Research. Sage, London.

Greenacre, M. (1984) Theory and Application of Correspondence Analysis. AcademicPress, London.

Greenacre, M. (2007) Correspondence Analysis in Practice (2nd ed.). Chapman &Hall/CRC, Boca Raton, FL.

Guo, L., & Ng, I. (2011) The co-production of equipment-based services:An interpersonal approach. European Management Journal 29(1), 43–50.

Hagedoorn, J. (1993) Understanding the rationale of strategic technology partnering– interorganizational modes of cooperation and sectoral differences. StrategicManagement Journal 14(5), 371–385.

Hagedoorn, J. (2002) Inter-firm R&D partnerships: An overview of major trends andpatterns since 1960. Research Policy 31(4), 477–492.

Hagedoorn, J., & Narula, R. (1996) Choosing organizational modes of strategictechnology partnering: International and sectoral differences. Journal ofInternational Business Studies 27(2), 265–284.

Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010) Multivariate Data Analysis.Pearson, Upper Saddle River.

Hansen, Z. N. L., & Rasmussen, L. B. (2013) Outsourcing relationships: Changes inpower and dependency. European Management Journal 31(6), 655–667.

Hladik, K. (1985) International Joint Ventures. Lexington Books, Lexington, MA.Jarvenpaa, S. L., Shaw, T. E., & Staples, D. S. (2004) Toward contextualized theories

of trust: The role of trust in global virtual teams. Information Systems Research15(3), 250–267.

Killing, P. (1983) Strategies for Joint Venture Success. Croom Helm, London.Kuo, A., Kao, M. S., Chang, Y. C., & Chiu, C. F. (2012) The influence of international

experience on entry mode choice: Difference between family and non-familyfirms. European Management Journal 30(3), 248–263.

Kylanen, M., & Rusko, R. (2011) Unintentional coopetition in the service industries:The case of Pyha-Luosto tourism destination in the Finnish Lapland. EuropeanManagement Journal 29(3), 193–205.

Lahiri, S., & Kedia, B. (2011) Determining quality of business-to-business relationships:A study of Indian IT-enabled service providers. European Management Journal29(1), 11–24.

Lehrer, M., Ordanini, A., DeFillippi, R., & Miozzo, M. (2012) Challenging the orthodoxyof value co-creation theory: A contingent view of co-production in design-intensive business services. European Management Journal 30(6), 499–509.

Luo, Y. D. (2002) Building trust in cross-cultural collaborations: Toward a contingencyperspective. Journal of Management 28(5), 669–694.

Marshall, C. (2004) Dating for Innovation: Recognizing and Creating Opportunities inFluid Environments through Collaborative Interorganizational Relationships(Doktorsavhandling). Handelshögskolan i Stockholm, Stockholm.

Morris, D., & Hergert, M. (1987) Trends in international collaborative agreements.Columbia Journal of World Business 22(2), 15–21.

Narula, R. (2001) Choosing between internal and non-internal R&D activities: Sometechnological and economic factors. Technology Analysis & Strategic Management13(3), 365–387.

Narula, R., & Hagedoorn, J. (1999) Innovating through strategic alliances: Movingtowards international partnerships and contractual agreements. Technovation19(5), 283–294.

Osborn, R., & Baughn, C. (1990) Forms of interorganizational governance formultinational alliances. Academy of Management Journal 33(3), 503–519.

Parmigiani, A., & Rivera-Santos, M. (2011) Clearing a path through the forest:A meta-review of interorganizational relationships. Journal of Management 37(4),1108–1136.

Pisano, G. (1991) The governance of innovation – Vertical integration and collaborativearrangements in the biotechnology industry. Research Policy 20(3), 237–249.

Ryan, G., & Bernard, H. (2000) Data management and analysis methods. In N. Denzin& Y. Lincoln (Eds.), Handbook of Qualitative Research (2nd ed.). Sage, London.

Schmiemann, M. (2006) Strategic and non-strategic inter-firm linkages in Europe.Eurostat Business Statistics.

Schmiemann, M. (2007) Inter-enterprise relations in selected economic activities.Eurostat Business Statistics.

Sitkin, S. B., & Pablo, A. L. (1992) Reconceptualizing the determinants of risk behavior.Academy of Management Review 17(1), 9–38.

Spina, G., Verganti, R., & Zotteri, G. (2002) A model of co-design relationships:Definitions and contingencies. International Journal of Technology Management23(4), 304–321.

Strauss, A., & Corbin, J. (1990) Basics of Qualitative Research: Grounded TheoryProcedures and Techniques. Sage, Newbury Park, CA.

Teece, D. J. (1992) Competition, cooperation, and innovation – Organizationalarrangements for regimes of rapid technological-progress. Journal of EconomicBehavior & Organization 18(1), 1–25.

Teng, B., & Das, T. (2008) Governance structure choice in strategic alliances – Theroles of alliance objectives, alliance management experience, and internationalpartners. Management Decision 46(5–6), 725–742.

Werr, A., & Linnarsson, H. (2004) Overcoming the innovation-alliance paradox:A case study of an explorative alliance. European Journal of Innovation Management7(1), 45–55.

Yoshino, M., & Rangan, U. (1995) Strategic Alliances: An Entrepreneurial Approach toGlobalization. Harvard Business School Press, Boston, MA.

Zhang, Y., & Huxham, C. (2009) Identity construction and trust building in developinginternational collaborations. Journal of Applied Behavioral Science 45(2), 186–211.

200 L. Börjeson/European Management Journal 33 (2015) 191–200