Alliance Longevity: Examining Relational and Operational Antecedents

Long Range Planning xxx (2014) 1–17

Contents lists available at ScienceDirect

Long Range Planning journal homepage: http://www.elsevier.com/locate/lrp

Alliance Longevity: Examining Relational and Operational Antecedents Noushi Rahman, Helaine J. Korn Viewing alliance longevity as a predictor of alliance performance can lead to erroneous conclusions about interfirm collaborations. A short lifespan may be indicative of strong expeditious performance rather than weak performance. Longevity may suggest partners’ escalated commitment to a failing project rather than steady progress toward collaborative goals. Understanding how to influence alliance longevity can enrich understanding of the functioning and usefulness of strategic alliances for both scholars and practitioners. Consequently, in this study, we examine how a set of four antecedent factors that comprise operational and relational aspects of alliances contribute to alliance longevity. Results using monthly data from 167 alliances support all six hypotheses about how these factors influence alliance longevity. Ó 2012 Elsevier Ltd. All rights reserved.

Introduction The abundance of strategic alliances has prompted an interest in understanding what leads to their satisfactory performance. As performance of an alliance is often difficult to specify and measure, many researchers have turned to alliance longevity as a proxy for alliance performance (Pangarkar, 2003). Since performance is the overwhelmingly preferred dependent variable in strategy research, in general, and in strategic alliance research, in particular, we feel it is absolutely necessary to first conceptually distinguish alliance longevity from alliance performance. Alliance longevity and alliance performance are not equivalent constructs (Barkema et al., 1997; Geringer and Hebert, 1991; Glaister and Buckley, 1998; Glaister and Buckley, 1999; Hatfield et al., 1998) and the difference between these constructs is clear from their definitions. Alliance longevity refers to the time span an alliance remains in existence, i.e., the time that elapses between alliance formation and termination. Alliance performance refers to the extent to which an alliance has met the goals that it was supposed to accomplish. Ariño argues that longevity can be viewed as both a precursor and a result of performance (Ariño, 2003). An alliance may be terminated, cutting short its longevity, because it has achieved its purpose or because it is failing to achieve its purpose (Makino et al., 2007; Sadowski and Duysters, 2008). Moreover, an alliance may be terminated and not survive long enough to yield a satisfactory performance (Kanter, 1994; Park and Ungson, 2001). Alternatively, an alliance may endure, extending its longevity, because it has not yet achieved its purpose and more time is being provided to allow that to happen. Or, even when it has achieved its purpose, an alliance may endure anyway. Some alliances last well beyond their useful life because inertia sets in and no action is taken to dissolve them (Adner and Levinthal, 2004). Additionally, the anticipated effects of some alliance characteristics may differ for alliance performance and alliance longevity. For example, relatedness between alliance partners has been demonstrated to have a positive effect on shareholders’ abnormal gains (García-Casarejos et al., 2009), but we would expect that relatedness between alliance partners (Rahman and Korn, 2010) would create competition that would shorten alliance longevity. We contend, therefore, that longevity is a flawed proxy for alliance performance. Instead, it needs to be studied on its own in order to sort out the various possibilities that may explain variation in the length of time strategic alliances endure and to better afford managers with the levers they need to influence longevity to its desired state. We lay out the differences between alliance longevity and alliance performance in Table 1. Alliances with a short longevity may be associated with either weak or strong performance. On the one hand, if the financial prospects of an alliance diminish for some reason, the partners may opt to terminate the alliance early. On the other hand, if the goals of an alliance are met in an expeditious manner, then the partners may dissolve the alliance to channel their resources elsewhere. A similar argument exists for alliances with a long longevity. On the one hand, if an alliance is not performing well, the partners may opt to devote further resources to it because of their escalating commitment. On the other http://dx.doi.org/10.1016/j.lrp.2012.05.003 0024-6301/Ó 2012 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

2

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

Table 1 Contrast between alliance longevity and alliance performance Alliance Longevity

Alliance Performance

Influence of Time

As time passes, the probability of alliance dissolution increases

Effectiveness

Goal accomplishment leads to alliance dissolution

Efficiency

High efficiency leads to short alliance duration

Relational factors

Stabilises the alliance and makes partners elongate alliances in the hope for results Improves task co-ordination, which can lead to goal accomplishment, which in turn leads to shorter longevity Poorly performing alliance can cut short alliance longevity prematurely; however, poorly performing alliance may increase the escalating commitment of partnering firms and thereby extend alliance longevity

As time passes, the probability of alliance performance increases Goal accomplishment leads to alliance success High efficiency results in increased performance Relational factors do not directly impact alliance performance Operational factors facilitate alliance performance Short-living alliance may not provide sufficient time to achieve success in an alliance; however, short-living alliance may mark high efficiency in alliance, thereby denoting stellar performance.

Operational factors Impact on each other

hand, if the partners view that their alliance is making steady progress toward the eventual goal, they may favour prolonging its duration. The permutations suggest that viewing alliance longevity in terms of alliance performance would contribute to a fallacy of composition. A deeper and more nuanced understanding of alliance longevity promises to contribute to our understanding of why certain alliances dissolve quickly and others survive for years. Alliance longevity in and of itself does not represent a good or bad thing. Depending on the purpose of an alliance, more or less time may be needed to accomplish desired objectives. Sometimes, greater longevity may be desirable to allow an alliance ample time to achieve its purpose. An alliance’s lifespan allows the alliance managers to achieve various sources of power with their respective firms, increasing interfirm trust that can lower transaction costs for the partners. This is a key benefit of having an extended longevity for alliances. However, shorter alliance longevity may be preferred to free up resources no longer needed to maintain a relationship that has already achieved its purpose. So, it would be beneficial to understand what factors influence longevity to achieve whatever length may be desired in a particular circumstance. Understanding what factors contribute to alliance longevity is thus the aim of this paper. In particular, we contend that alliance longevity is affected by both relational and operational aspects of an alliance, just as others contend that alliance formation and withdrawal reflect social and task-related aspects (Greve et al., 2010; Rahman, 2007). This is because interfirm structural configurations of strategic alliances expose them to both relational and operational challenges (Das and Teng, 1996; Das and Teng, 2000; Fryxell et al., 2002). Recently, Kotabe, Martin and Domoto have observed that relational aspects facilitate the operational aspects in technology transfer alliances (Kotabe et al., 2003). In keeping with the literature, the variables in our research model are chosen on the basis of their contributions to either the relational or operational functioning of alliances. While transaction cost economics is concerned with relational hazards and associated governance mechanisms in alliances, it does not recognise the role of partnering firms’ alliance capabilities. In contrast, the dynamic capabilities literature focuses on how alliance experience and alliance units generate alliance capabilities, but the theory does not adequately address relational hazards and governance. Thus, transaction cost economics and the dynamic capabilities perspective address relational and operational aspects of alliances respectively. Since relational and operational aspects influence alliance longevity in complementary ways, highlighting determinants derived from these two theoretical perspectives would result in a richer model of alliance longevity. To that end, we draw from transaction cost and dynamic capabilities perspectives to examine four factors that determine alliance longevity. We theorise that hierarchy of alliance structure, specific alliance experience, alliance units and general alliance experience are all relevant in understanding alliance longevity. We organise this paper into five parts. First, we review theoretical arguments that help elaborate on the concept of alliance longevity. Here, we introduce the idea that both relational and operational factors need to be taken into consideration. Second, we identify four relational and operational predictors of alliance longevity, derived from transaction cost economics and dynamic capabilities perspectives, and advance six hypotheses about their relationship to alliance longevity. We describe the research methodology of the paper in the third section and then report the results of an empirical study examining these relationships. Last, we discuss theoretical and applied contributions to understanding and predicting alliance longevity. Theory and hypotheses Longevity has been considered an important concept by different disciplines. Building on various conceptualisations of longevity from an array of disciplines, we explain the significance of alliance longevity as a research construct. First, according to Taoism, longevity allows for more time to understand the world and become one with Tao. This is perhaps the most widespread reason for wanting greater longevity. Alliance longevity allows for greater time for the alliance to achieve the partners’ goals. An alliance that has been in operation for long enough will have had sufficient intermediate achievements to be considered a project moving along on the desired path. Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

3

Second, from a biological perspective, in the organisational life cycle context, alliance longevity plays a major role in determining the resource freeing-up rate, because a significant amount of resources of partnering firms are engaged in their alliance activities. Better understanding of the timespan until an alliance may terminate and free up resources would enable partnering firms to start engaging in talks for new alliances at an appropriate time. Otherwise, alliances would live out their lives until pre-planned completion dates and only then could partnering firms begin to figure out what to do next and how to redeploy resources. Third, actuarial science scholars maintain that “longevity is important, as it determines the time span during which accumulated wealth is distributed. If an individual lives longer than expected, the amount available for consumption is reduced” (Hafner and Mader, 2007, p. 29). Similarly, if an alliance has a longevity longer than initially expected and provided for, then it will exhaust the allocated resources. This notion is consistent with the gerontology argument that increased longevity would require continued devotion of resources to maintain the alliance (Yakita, 2006). Thus, better understanding of alliance longevity would allow partnering firms to allocate or budget their limited resources to their projects more efficiently. Fourth, longevity yields much needed visibility from a marketing standpoint. When a specific alliance lives long enough, the alliance project starts to become more visible in the industry. The alliance between General Motors and Toyota, New United Motor Manufacturing, Inc. (NUMMI), lasted long enough to produce several well-known models and industry experts related NUMMI’s fame to the partnering firms. So, a certain level of alliance longevity is important for the alliance’s task to attract some valuable attention within the industry. Last, from a project management perspective, it takes time to get used to the rhythms of working with another firm on any new project. As an alliance continues, it is capable of smoothing possible rough edges that may exist in collaborative work. Actual task accomplishment becomes more efficient once initial kinks are removed. Better knowledge of alliance longevity can lead partnering firms to be more patient with their alliance-specific results. Thus, there is a long and varied history of considering the contributions of longevity, meriting this as an important area of study in strategic alliances as well. Relational and operational aspects of alliance longevity The ability to influence alliance longevity may be an important instrument in managers’ arsenal of activities needed to effectively manage their resource allocations. Conditions arising from relational, operational, market, legal and political environments may influence alliance longevity. Notwithstanding, this research is strictly concerned with relational and operational conditions that either hinder or foster alliance longevity. Operational factors are task-related elements that directly affect the functioning of the alliance. Relational factors are elements of interfirm dynamics that indirectly affect the functioning of an alliance. Such a focused approach is adopted for a couple of reasons. First, partnering firms are more able to control their relational and operational conditions vis-à-vis their market, legal or political conditions. Thus, a refined understanding of the relational and operational determinants of alliance longevity would allow alliance managers to influence the longevity of their alliances by working on these conditions. Second, market, legal or political conditions explain alliance longevity at the industry level, rather than at the alliance level. Combining these conditions with relational and operational determinants will introduce problems associated with multiple levels of analysis and could perhaps be tackled in other research. Firms often form alliances for long-term strategic reasons, rather than short-term financial advantages (Glaister and Buckley, 1999). For example, learning a leaner production process, developing technological knowhow, improving their market position and gaining legitimacy within the industry are viable strategic motives that have long-term consequences. When firms steadily move forward realising their alliance-specific goals, they are likely to continue with their alliance projects (i.e., lowering the probability of unintended termination and prolonging the longevity). However, as partnering firms accomplish their alliance-specific goals, they are likely to dissolve their alliance projects (i.e., raising the probability of intended termination and limiting the longevity). Alliance task objectives have to be met through joint effort that is dependent on a good working relationship. As firms must collaborate to accomplish alliance-specific goals, the importance of a good working relationship is more salient in alliances than in market-based transactions or in independent firms. Empirical evidence suggests that alliances with stable relationships will survive longer than those with tumultuous relationships (Harrigan, 1988). Alliance members fluently cooperating with each other will enjoy such a stable relationship. Luo (2002) finds that co-operation within an alliance relationship enhances contractual adaptability, which in turn encourages more co-operation. Similarly, Poppo and Zenger (2002) advance the idea that formal contracts and relational governance function as complements, rather than substitutes for one another. And, Parkhe (2001), in one of only a few other studies explicitly considering alliance longevity, observes that different kinds of interfirm diversity influence alliance longevity differently. Interfirm differences that alliances are designed to exploit, such as differences in capabilities, may shorten the longevity of alliances as one partner achieves its goals sooner than another and decides to terminate the alliance prematurely. On the other hand, interfirm differences in culture and approach to conflict may be mitigated by learning over time, and consequently allow alliances to achieve greater longevity. Our position is consistent with recent research by Greve et al. (2010) who observe that member withdrawal from alliances has both social and task-related causes. They argue that both cohesion from social relations and friction arising during task execution can influence member withdrawal. Of the four antecedent factors that comprise our model, hierarchy of alliance structure and specific alliance experience are pertinent to the relational aspect, whereas alliance units and general alliance experience are relevant to operational aspects of alliances. These factors are chosen because of their strong relevance to Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

4

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

relational and operational aspects of alliance functioning, as predicated by transaction cost economics and dynamic capabilities perspective respectively (see Figure 1).

Hierarchy of alliance structure Transaction cost economists have studied alliances by focusing on market and hierarchy-like characteristics of alliances (Dussauge and Garrette, 1995; Gulati and Singh, 1998; Oxley, 1997; Sampson, 2004). Gulati and Singh discuss five dimensions of hierarchy: command structure and authority systems; incentive systems; standard operating procedures; dispute resolution procedures; and non-market pricing systems. They convincingly argue that different alliance structures have different degrees of hierarchy based on these dimensions. The most hierarchical alliance structure is a joint venture, which creates a separate entity through co-ownership to manage joint activities. The institutionalisation and formalisation of control and co-ordination of activities in joint ventures mimic those in independent firms, which are traditional hierarchies. Minority equity structure, involving the exchange of minority equity stakes between partners, is moderately hierarchical. “Hierarchical supervision is typically created by the investing partner joining the board of directors of the partner that received the investment. The presence of one or more individuals on the board of the partner in a minority alliance introduces a fiduciary role into the relationship and is also a vehicle for hierarchical controls” (Gulati and Singh, 1998, p. 793). Non-equity structure, based on contractual agreements, is the least hierarchical. No equity exchange or co-ownership takes place and there is no administrative structure to govern the relationship. Research on alliance structure has highlighted some specific roles that structures play. First, transaction cost economics underscores the utility of hierarchical structure in governing partner behaviour (Hennart and Oxley, 1988; Oxley, 1997). Hierarchy of alliance structure becomes critical when partner reputation is unknown and the focal firm anticipates partner opportunism due to high behavioural uncertainty and asset specificity conditions. Therefore, more hierarchical structures can deter partner opportunism and keep an alliance from terminating prematurely. Additionally, more hierarchical alliance structures are locked into a relationship by their equity investments. These projects tend to require greater levels of resource commitments, which become sunk costs when an alliance is dissolved prematurely. Studying 311 long-term contracts, Von Hirschhausen and Neumann (2008) find that “contracts linked to an assetspecific investment extend, on average, three years longer” (p. 131). Therefore, since equity investments in more hierarchical alliance structures are in effect asset-specific investments (i.e., assuming alliances as bundled assets), more hierarchical alliance structures will tend to last longer. Firms bounded by equity investments in their alliances are not going to simply write off their resource commitments to an alliance. Rather, escalation of commitment toward an alliance may make partnering firms more attached to their collaborative project, delaying a much warranted termination of the alliance (Staw, 1981; Levinthal and Mark Fichman, 1988; Adner and Levinthal, 2004). While compared with an acquisition, a joint venture is much more flexible as an investment project, compared with a non-joint venture alliance, a joint venture is much more structurally bound and requires a substantial investment by parent firms to assume ownership. This suggests that non-joint venture alliances are more flexible than joint ventures. Thus, H1: Hierarchy of alliance structure is positively related to alliance longevity.

Hierarchy of Alliance Structure

Specific Alliance Experience

H1 (+)

H3 (–) H2 (+)

Alliance Longevity

H5 (+) General Alliance Experience (asymmetry adjusted)

H6 (+) H4 (–) Number of Partners with Alliance Units - None - One firm - Both firms Figure 1. Determinants of alliance longevity

Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

5

Specific alliance experience Two kinds of alliance experience have been introduced in the literature – specific and general alliance experiences (Hoang and Rothaermel, 2005). Lessons learned from repeated alliances with a specific partner will be more applicable to future alliances with that same partner than to future alliances in general. The interfirm trust that generates between partnering firms is a unique outcome of this type of experience (Gulati, 1995). According to Ring and Van de Ven (1994), “Congruency is a cumulative product of numerous interactions; through these interactions emerge trust in the goodwill of others and an understanding of constraints on the relationship that may be imposed by a person’s organisational role” (p. 100). This interfirm trust between two partnering firms does not significantly influence the outcome of those alliances that the partnering firms are involved in separately with other firms. It must be noted that if a firm has negative experience in its alliance with a partner, it is unlikely to engage in repeated alliances with that partner. Thus, for most cases, alliance experience with a partner refers to positive experience. Partnering firms may engage in successive alliances with each other despite initial negative experiences when a negative alliance experience serves as a lesson for them, benefiting the functioning of future alliances between them. That is, notwithstanding the negative results, these partners may have a positive learning experience by identifying the problems that existed in their collaboration and the promising areas that need to be worked on. It is reasonable to assume that, under conditions of initial negative alliance experience, two firms would engage in repeated alliances only if they were able to learn from their mistakes and could avoid them in the future. Main effect The effective selection of an alliance partner is crucial to the outcome of the alliance (Holmberg and Cummings, 2009). Alliance experience makes partnering firms become familiar with each other. Familiar firms are more trusting of each other (Gulati, 1995), which is a necessary condition for co-operation. Alliances with ample inter-organisational trust and cooperation between partners are more likely to overcome intermediate obstacles and move toward accomplishing alliance goals by engaging in long-lasting relationships (Zaheer et al., 1998a). Cuypers and Martin (2008) observed that alliance experience significantly decreases the odds of dissolving the alliance. Moreover, these alliances are less likely to be plagued by partner opportunism. Thus, specific alliance experience prevents certain possibilities of premature dissolution due to a lack of trust or fear of various forms of misappropriation. Partnering firms that engage in repeated alliances may also develop routines about how to work on their collaborative projects. Zaheer et al. (1998b) found that the length of buyer-supplier alliance relationships positively affected interorganisational trust from the supplier’s point of view. Moreover, McEvily et al. (2003) note that trust can facilitate the development of interorganisational routines through structuring. They argue that “[t]he result is a network of stable and ongoing interaction patterns, both formal (e.g., routines and organisational units) and informal (e.g., cliques and coalitions)” (p. 94). Such partner-specific routines can enhance a firm’s capability of alliancing with that specific partner and encourage an ongoing relationship to continue. Therefore, H2: Partner-specific alliance experience is positively related to alliance longevity.

Moderating effect While we have argued that hierarchy of alliance structure is positively related to alliance longevity, under specific circumstances the strength of this relationship fluctuates. Though not a common phenomenon, some nonequity alliances have comparable or even greater longevity than joint ventures. For example, the Carrera-Renault nonequity dealership alliance lasted five years, as the contract had specified, whereas the Chrysler-Renault joint venture lasted less than two years (Feltrin, 2002; Levin, 1990). This deviation from expected difference in longevity between nonequity and joint venture structures suggests the presence of moderating effects. We contend that under conditions of greater specific alliance experience, influence of the hierarchy of alliance structure on alliance longevity starts to diminish. Previously, we have explained that effects of alliance experience serve as substitutes for control and co-ordination needs that are common to interorganisational relationships. This condition makes hierarchies redundant, since the control and coordination features of more hierarchical structures would remain underutilised. Greater alliance experience with a partner starts to generate interfirm trust, which in turn reduces the need to control partner behaviour (Das and Teng, 1998). Similarly, the need to co-ordinate complex tasks becomes less necessary as partnering firms develop operating routines to avert some of the complexities of subsequent alliance projects (Zollo et al., 2002). McEvily and Marcus offer several real-life vignettes about the interfirm relationship building process that lead to greater embeddedness (McEvily and Marcus, 2005). In the words of one of the managers in their field study, “We’ve worked with them so long they can probably tell us what we need before we know what we need. They know us, what our requirements are. We work more efficiently together” (p. 1039). Thus, alliances with various levels of structural hierarchy may experience similar levels of longevity as they are influenced by the effects of partnering firms’ alliance experience with each other. This argument leads to the following hypothesis: H3: Greater partner-specific alliance experience weakens the relationship between hierarchy of alliance structure and alliance longevity. Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

6

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

Alliance units Nelson (1991) describes, “Structure involves how a firm is organised and governed, and how decisions actually are made and carried out, and thus largely determines what it actually does, given the broad strategy. A firm whose strategy calls for being a technological leader that does not have a sizeable R&D operation, or whose R&D director has little input into firm decision making, clearly has a structure out of tune with its strategy” (p. 67). Along similar lines, Teece et al. (1997) note that firm capabilities are best understood in terms of structures and processes of the organisation, since they are critical in implementing value-creating strategies. Following this logic, we argue that a firm that generates a large percentage of its revenue from alliance activities ought to have an executive-level alliance role or an alliance unit spearheaded by a senior executive to manage all alliance-related activities (i.e., so that the structure is in tune with the strategy). A firm’s alliance unit is comprised of alliance managers who oversee and engage in alliance activities. It is the primary responsibility of the alliance unit to learn from prior alliance experience. The alliance unit identifies the effective and ineffective practices and refines the experience-based data to highly useful information. Thus, the alliance unit serves as a reservoir as well as a refinery of alliance management knowhow. According to Heimeriks et al. (2009), “creating a dedicated alliance function both increases external visibility and provides a firm-wide signal that alliances are deemed important” (p. 99). Recently, several large firms have established departments or units of strategic alliances to exclusively attend to crises and monitor performances of their alliances (Kale et al., 2002). This is a new characteristic in firm structure, since such units were practically nonexistent prior to the 1990s and are still relatively uncommon among firms. The existence of a department dedicated to alliance-related issues within the larger structure of a firm has significant strategic implications (Hoffmann, 2005). First, both qualitative and quantitative data get stored in the alliance unit as a firm engages in numerous alliances over time. The alliance role or unit is in a position to analyse these data, producing for the partnering firm valuable knowledge about effective and ineffective processes, practices and outcomes in managing strategic alliances. The alliance unit is often able to codify such alliance-specific knowledge (Grant, 1996; Rahman, 2004). As the literature on dynamic capabilities suggests, codification allows a partnering firm to apply lessons learned from one particular alliance to all the other alliances in which it engages (Eisenhardt and Martin, 2000; Teece et al., 1997; Zollo and Winter, 2002). Even when knowledge from alliance experience is tacit and cannot be readily codified, the alliance unit can serve as the hub of such tacit knowledge by providing a platform where managers of its different alliances can come together and share their experiences. Second, trade publications frequently report how alliance deals are initiated by top managers of two firms over some casual conversation. Hence, although partner selection is critical to effective functioning of the alliance (Geringer,1991), no systematic screening takes place to find the most competent and complementing firm as an alliance partner (Rahman, 2008). The alliance unit is responsible for such systematic search and the senior executive in charge of the alliance unit is in a position to either conduct alliance negotiations or advise the CEO of the nature of partner s/he should be looking for. By being aware of the alliancing trends and history within the industry, an alliance unit is in a unique position to advise top management (i.e., the final dealmakers) about the most appropriate candidates for a strategic alliance, leading to effective partner selection. Third, while any particular alliance manager may not know the whereabouts of different resources within a large firm, it is more likely that such information will be readily available to the alliance unit as a whole (Kale et al., 2002). When certain functions of the firm refuse to help or conform to alliance-related demands, the senior executive of the alliance unit can effectively push the issue to higher authorities (e.g., CEO or COO) and resolve the problem quickly. The various roles that alliance units play are likely to expedite the goal accomplishments of an alliance. Therefore, H4: The number of partners with formal alliance units is negatively related to alliance longevity. General alliance experience Firms build on their prior alliance experience of doing a particular task and become better in executing related tasks of later alliances. Through this process, firms acquire a dynamic capability of managing their various alliances. Dynamic capabilities of managing complex activities evolve through iterative execution of those activities. Scholars have characterised the iterative process of learning as “sequenced steps” and as “consequential” (Brown and Eisenhardt, 1997; Eisenhardt and Martin, 2000). Studying six firms in the computer industry, Brown and Eisenhardt highlight those firms’ practice of linking routines exercised in one development project to the next one. Similarly, Kim observes that Hyundai Motor Company has been benefiting from first developing relatively simpler capabilities pertaining to manufacturing process and then developing more complex capabilities of product design process (Kim, 1998). Main effect The construct, general alliance experience, refers to experience gathered from all previous interfirm collaborations (Simonin, 1997). Experience in managing alliances allows a firm to learn by doing. According to the dynamic capabilities perspective, firms learn from experience through evolution, adaptation and replication (Nelson and Winter, 1982; Teece et al., 1997). Hence, firms more experienced in managing alliances are likely to be highly skilled in alliance management, learning from prior mistakes and successes. Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

7

Simonin (1997) conjectures, “Firms with greater collaborative experience will achieve higher levels of collaborative knowhow” (p. 1158). Surveying 151 firms, he found significant support for this hypothesis. Experience is institutionalised through routines. Routines capturing alliance experience are, however, not static but dynamic in nature. That is, whereas static routines allow the mastery of a particular skill, dynamic routines accommodate the constantly changing nature of some task by offering flexibility in their applications. Simonin (p. 1158) offers a telling example: The draft of a letter of intent for a new joint venture is likely to reflect a format and content derived from previous experiences with a related, but still very different, context and situation. In that regard, the letter of intent is a manifestation of an organisational routine – the drafting process – that can be partially codified and formally updated over time. In short, general alliance experience breeds dynamic routines, which facilitate better alliance management and increases the likelihood of alliance survival. Alliance experience teaches firms the dos and don’ts of alliancing. Such firms become better at screening partners, negotiating contracts, monitoring progress toward alliance goals, managing a good working relationship and demonstrating patience at times of alliance-related crises. Moreover, Powell et al. (1996) claim that firms with greater alliance experience are more able to manage a diverse portfolio of alliances. They argue, “firms learn from exploration and experience how to recognise and structure different types of alliances” (pp. 120-121). General experience in alliance engagement lets firms learn from mistakes and successes. It can make firms more familiar with certain alliancing practices and make them aware of certain pitfalls within the alliancing process. Eisenhardt and Martin (2000) recount how Yahoo developed simple rules as its dynamic capability in managing scores of successful alliances: “Yahoo managers formed an exclusive relationship with a major credit card firm. Shortly, they recognised that this alliance restricted flexibility, especially with regard to retailers, and terminated it at great expense. The ‘no exclusive deals’ rule emerged from this mistake” (p. 1115). Zollo et al. (2002) address another key advantage of alliance experience, namely that of learning curve benefits. Learning curve benefits would make alliancing a more efficient process, precluding redundant paperwork, motions and activities. Thus, partnering firms are able to make their alliancing experience more productive and rewarding. These conditions are conducive to producing favourable intermediate goal accomplishments, keeping members content and willing to continue with the venture. The alliance management knowledge coming from general alliance experience of two partnering firms may be counterproductive if asymmetry exists in their experience bases. The longevity of alliances will further increase when the members have less asymmetry between their general alliance experience bases. Thus, we advance the following hypothesis: H5: General alliance experience (asymmetry adjusted) is positively related to alliance longevity. Moderating effect Simonin (1997) aptly states, “although knowledge about collaborations can be gleaned from external sources, much of this knowledge may not have value until it is internalised and applied to a firm’s own unique situation” (p. 1158). The primary source of alliance management knowledge for any firm is its own alliance experience. Basically, alliance experience gives a firm exposure to unique alliance-relevant situations and a hands-on experience to deal with those situations. More alliance experience means the culmination of a greater amount of data. Such experience has valuable knowledge-content that the firm needs to process and practice through some mechanism. An alliance unit generates alliance capability by accumulating, storing, integrating and diffusing alliance management knowledge (Kale et al., 2002, p. 749). The alliance unit would be able to process the rich source of data amassed through alliance experience and produce valuable alliance management knowledge for the firm. The more knowledge being processed, the higher the probability for an alliance unit to find a new alliance management best practice. If highly experienced alliance units can detect new uses that are beyond intended goals of a particular alliance, the longevity of that alliance may be extended. Overall, we expect more experienced alliance units to positively affect alliance longevity. Hence, H6: The interaction of general alliance experience and number of partners with alliance units is positively related to alliance longevity.

Research methodology Research design of this study is non-experimental, as the predictor variables are observational rather than clinically controlled. The study uses longitudinal data, which is instrumental in measuring alliance longevity. Given the limited range of the dependent variable, tobit regression is well suited to yield robust results (Greene, 1990). Data sources Considering the impressive breadth and depth of SDC alliance data, it serves as the main source for our study. We observed that alliance coverage was comprehensive from 1992. It is important to note that termination data are only limitedly available in the SDC dataset. Also, SDC does not report the minority-equity status of alliances in any systematic manner. Thus, for every Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

8

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

alliance within our sample, data on termination date and minority equity status is retrieved from the Factiva information service. Last, we use Hoover’s Online data on senior executives and their designations to determine the presence of alliance units. Sample characteristics Considering the availability of more comprehensive coverage of alliance activities from 1992 by the SDC database, we decided to study alliances that were in existence during any part of a 10-year period between January 1, 1992 and December 31, 2001. Screening for dyads with publicly-held US partnering firms, the final sampling frame for this study comprises 3,695 alliances. Since SDC includes alliances based on alliance announcements, it is important to examine carefully each one to see if it came to fruition. Given this rather time-intensive task, we first randomly sampled 1,000 alliances, of which 546 had verifiable existence. Of these 546 alliances, 241 were formed and terminated during the 10-year period of our study. We were able to code the performance for 220 of these completed alliances (91 per cent). Full data on all variables were available for 167 of these alliances, rendering this as the final sample size. A power analysis revealed that with a moderate effect size (f2 ¼ .15), a moderate significance level (a ¼ .05), and a power requirement of .95, the requisite sample size is 160 (Borenstein et al., 1997). Dependent variable: alliance longevity Consistent with prior research, alliance longevity is defined as the temporal period that an alliance remains operational (Barkema et al., 1997, p. 429). For purposes of calculation, the time span between the formation and dissolution of an alliance captures the alliance longevity. For refined assessments of alliance longevity, the time span is measured in terms of months, as opposed to years.1 Independent variables There are four distinct independent variables in the model, representing relational and operational aspects of alliances: hierarchy of alliance structure, specific alliance experience, alliance units and general alliance experience. These variables are defined and operationalised below. Hierarchy of alliance structure Hierarchy of alliance structure refers to the degree to which formal structural features exist to control and co-ordinate alliance activities (Gulati and Singh, 1998). Nonequity structure is least hierarchical, since it does not employ any significant control or co-ordination mechanism, except for the contractual agreement. Minority equity structure is moderately hierarchical because exchange of minority equity stakes usually comes with a membership in the invested firm’s board, which allows for some degree of control and co-ordination abilities to partnering firms. Joint venture is the most hierarchical because the creation of a separate autonomous company provides for various control and co-ordination features, such as incentive systems, internal pricing mechanisms, formal authority structures, operating procedures and dispute resolution procedures. From the above explanation and following Gulati and Singh’s (p. 795) approach, alliance structures are coded in terms of their degree of hierarchy: nonequity structure ¼ 1, minority equity structure ¼ 2, and joint venture ¼ 3. Specific alliance experience Alliance experience with partner has been conceived as number of prior ties in previous studies (Gulati, 1995; Gulati and Singh, 1998; Pangarkar, 2003; Zollo et al., 2002). Hence, this variable has been operationalised as the number of alliances two partners have engaged in with each other in the past. Therefore, where Nij is the number of alliances between firms i and j, Alliance Experience with Partnerij ¼ Nij. To calculate specific and general alliance experience, we included all 69,000 alliances reported by SDC database dating back to 1985. Alliance units The existence of an alliance unit or alliance position is reflected by the assignment of a senior executive in charge of all alliance activities. It is reasonable to expect that a company with an alliance unit will have in its top management team a senior executive of alliances, who is exclusively responsible for successful planning and implementation of alliances (Kale et al., 2002, p. 754). In dyadic alliances, three scenarios are possible with respect to partners having alliance units within their structures. Alliance units can be present in neither firm (N ¼ 0), any one firm (N ¼ 1), or both firms (N ¼ 2). General alliance experience Like the treatment of a specific alliance experience, scholars have operationalised this construct in terms of count measures as well (Kale et al., 2002; Zollo et al., 2002). After adding the general alliance experience of two firms, the absolute

1

Alliance longevity has been measured in months by Pangarkar (2003) and in years by Barkema et al. (1997).

Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

9

difference of their general alliance experience is subtracted to adjust for the potentially harmful effects of asymmetry in alliance experience bases. This variable was log-converted for our analyses. Control variables Considering the possibility of their influence on alliance longevity, we included partnering firm’s size differential, predetermined longevity, year (time effect), repeating firms (cluster effect), alliance type and alliance performance2 as control variables of this study. Size differential As difference in size between partner firms increases, the communicative distance may also increase. Firm size differential has been used as a control variable in studies on mergers and acquisitions and strategic alliances (Banning, 1999; Ramaswamy, 1997; Artz et al., 2002; Gulati and Singh, 1998). Size differential is operationalised as the ratio of the asset size of two alliance partners i and j: where Asseti < Assetj, Size Differentialij ¼ Asseti O Assetj Predetermined alliance longevity Even when an alliance is performing extremely well, in some instances, partner firms are obliged to dissolve the alliance at the end of the stated expiration date. Moreover, whether intended alliance duration is pre-specified or open-ended has an effect on the choice of specific contractual provisions (Reuer and Ariño, 2007). Therefore, we controlled for pre-determined longevity. Less than 5 per cent of alliances have predetermined longevity in our sample. A dummy variable is created – 1 if alliance longevity is determined during formation and 0 if alliance longevity is open-ended – to control for any effect that predetermined longevity might have on actual alliance longevity. Year (time effect) Time of existence could have an effect on the alliance longevity. To control for time effects, we computed the midpoint of each alliance’s duration during the 10-year horizon of our study and specified that in years (with two decimals). This variable had a range of 0 to 10. Repeating firms (cluster effect) Some of the partnering firms had multiple alliances in our dataset. This raised concern for a possible clustering effect. To control the influence of such clustering of firms in our sample, we introduced the variable ‘Repeating Firms’, where we distinguished between partnering firms that had multiple alliances in our sample and partnering firms that had only one alliance in our sample. Thus, in the ‘Repeating Firms’ variable, an alliance could have no firms with any other alliance in the sample (coded as 0), just one firm with no other alliance in the sample (coded as 1), or both firms with other alliances in the sample (coded as 2). Alliance type Whether an alliance is formed between competitors or between vertically-linked firms can be expected to have an influence on the longevity of the alliance. Hence, we control for alliance type in our model. Alliance types are assessed by analysing the nature of the connection between two alliancing firms. Following prior research, alliance partner’s primary SIC code is used to assess the nature of connection between two alliance partners (Mowery et al., 1996; Park and Russo, 1996; Part and Ungson, 1997). Alliance partners are horizontally linked if their primary SIC code is identical at the fourdigit level. All other remaining alliances are labelled vertical alliances. Horizontal alliances are coded as 0 and vertical alliances are coded as 1. Alliance performance Alliance performance is expected to influence alliance longevity. Thus, we controlled for alliance performance when testing our hypotheses. This variable “was coded as [1] if the alliance failed, [0] if a firm withdrew unilaterally due to a change in strategic priorities or the alliance ended with the natural expiration of the alliance contract, and [1] if the alliance terminated because firms successfully fulfilled their objectives and there was no more need to collaborate” (Reuer and Zollo, 2005, p. 108). Alliance performance was coded from the termination documents. Since alliance termination must occur for us to include alliance performance in our study, this variable could potentially cause a sample selection bias. Also, including alliances with partnering firms that have dedicated alliance units may cause a sample selection bias. We utilised the Heckman two-step approach to observe the extent of this problem (Heckman, 1979). We used ‘alliance termination’ and ‘alliances with partners that have alliance units’ as the variables that may cause sample selection. In the Heckman test, the Mill’s Lambda was not statistically significant, suggesting that sample selection is not significantly biasing our results.

2

We are grateful to an anonymous reviewer for suggesting us to control for alliance performance.

Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

10

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

Random effects tobit model We used a tobit model instead of a survival analysis model because “the use of the hazard rate masks certain details about termination, and provides little insight into why [alliances] are terminated or the critical factors that underlie termination” (Makino et al., 2007, p. 1114). Moreover, although in the same direction, results are relatively more robust when using the tobit model than when using a survival analysis model (e.g., Cox regression). Longitudinal data raises concerns for existing serial and auto correlations. Random effects models are tailored for longitudinal data and allow for unbalanced repeated entries (i.e., resulting in serial correlation) (Laird et al., 1982; Maddala, 1982). Most advanced statistical packages have random effects estimators that approximate the degree of serial correlation and compute estimates accordingly. Since we adopted the random effects tobit model for our analyses, we can expect that our results are not biased by serial correlation. Auto-correlation was not a problem in our data because 98.5 per cent of the alliances within our sample had unique partner combinations. This is not to say that firms did not have repeated alliances with each other, but rather those repeated alliances generally did not make it into our sample. The tobit model begins with the assumption of an underlying latent variable y* (i.e., the alliance’s (unobserved) propensity of experiencing a certain level of longevity). The latent variable produces the standard regression equation: 


where ui wN m; s2

0

yi ¼ b xi þ ui

y* is related to observed yi in the following way:

y i ¼ Li 



if yi  Li 

yi ¼ yi

if Li < yi < Ui

y i ¼ Ui

if yi  Ui



where Li and Ui are, respectively, the lower and upper limits of the dependent variable’s range. Tobit utilises the maximum likelihood function. The idea here is “to find that set of estimates of the parameters that, if these parameter estimates were true of the population, would have most likely generated the observed sample data” (Breen, 1996, p. 18). Significance of individual coefficients is assessed with t-tests. With respect to the above notations, the maximum likelihood function (ML) of tobit model is written as follows:

MLðb; sjyi ; xi ; Li ; Ui Þ ¼

Y yi ¼ Li

F

0

Li  b xi

s

!

Y 1 

yi ¼ yi

s

4

0

yi  b xi

s

!

Y yi ¼ Ui

" 1F

0

Ui  b xi

!#

s

0

where F is the distribution function and 4 is the density function of the standard normal evaluated at b xi =s (Maddala, 1983/ 1994, p. 152). Results Descriptive statistics We have analysed monthly data during January 1992 through December 2001 of 167 dyadic alliances formed by publiclytraded US firms. Our data contain observations on 5,759 alliance months. The average longevity was 34.5 months. In terms of alliance structure, 70 per cent had nonequity structures, 10 per cent had minority equity structures and 20 per cent had joint venture structures. In approximately 7 per cent of the alliances, the partners had prior alliance experience with each other. Almost 80 per cent of the partnering firms had some general alliance experience aside from the focal alliance. In terms of number of partners with alliance units, neither firm had a dedicated alliance role or unit in 76.5 per cent of the alliances, while 21.5 per cent had one partner with a dedicated alliance role or unit and only about 2 per cent had both partners with such a role or unit. Table 2 presents the mean, standard deviation and bivariate correlations of all the variables used in this study. Assumptions The tobit methodology assumes that the standard error of the prediction has a normal distribution and is not heteroskedastic. Histogram of the standard error of the prediction fit well with the standard normal distribution line. To assess heteroskedasticity, we ran the Breusch-Pagan test (Breusch and Pagan, 1979). The Breusch-Pagan c2 statistic was significant at p < 0.001 level, confirming a low level of heteroskedasticity in the standard error of the prediction. Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

11

Table 2 Bivariate correlations, means, and standard deviations Variables 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Mean Standard 1 Deviation

Alliance Longevity 34.49 20.61 Asset Differential 0.14 0.22 0.14 0.35 Predetermined a Longevity Mean Year 7.08 2.29 (Time Effects) 0.43 0.69 Repeating Firms (Cluster Effects)a 0.75 0.43 Alliance Typea Alliance Performancea 0.48 0.73 a Hierarchy 0.50 0.81 Specific Experience 0.43 0.94 Alliance Unitsa 0.25 0.48 General Experience 2.83 1.56 Hierarchy  Specific 2.11 1.72 Experience Alliance Units  General 5.03 3.35 Experience

2

.062 .022

3

.065

.082 .002 .032 .025 .025 .067 .036 .068 .053 .181* .220** .239** .021 .084 .089

5

6

7

8

9

10

11

12

.115

.671** .079 .004

4

.151

.026 .070

.076

.075 .124 .085 .314** .003 .074 .117 .140 .015 .049 .037 .205** .095 .035 .080 .038 .062 .034 .019 .056 .070 .048 .071 .213** .021 .100 .069 .252** .050 .060 .239** .380** .323** .136 .036 .104 .045 .206** .618** .681** .041 .147 .048

.025

.204** .010

.082

.255** .282**

.821** .78** .088

* p < .05. ** p < .01. a Spearman correlations are reported due to the ordinal nature of the variable.

Inferential statistics Drawing from transaction cost economics to represent relational aspects of alliances, we argued that hierarchy of alliance structure (H1, þ) and specific alliance experience (H2, þ) influence alliance longevity. Also, to represent operational aspects of alliances, we drew from the dynamic capabilities perspective to argue that alliance units (H4, þ) and general alliance experience (H5, þ) influence alliance longevity. We controlled for asset size difference, predetermined longevity, alliance type and alliance performance to test over and above the influence of our four independent variables and two moderating variables on alliance longevity (see Table 3). It should be noted that the control variables we used had significant influence on alliance longevity. The significance levels of the independent variables were p < 0.01 for hierarchy of alliance structure (H1), p < 0.001 for specific alliance experience (H2), p < 0.001 for alliance units (H4) and p < 0.001 for general alliance experience (H5). The significance of the moderating variables were p < 0.05 for specific alliance experience as a moderator of the relationship between hierarchy of alliance structure and alliance longevity (H3), and p < 0.05 for general alliance experience as a moderator of the relationship between alliance units and alliance longevity (H6). Therefore, as predicted, hierarchy of alliance structure, specific alliance experience, alliance units and general alliance experience were statistically significantly related to alliance longevity. Moreover, specific alliance experience weakened the positive relationship between hierarchy of alliance structure and alliance longevity. Results also supported that general alliance experience and alliance units interact in a way that extends alliance longevity. Discussion and conclusions To reach their objectives, some alliances may need ample time to gel and mature, hence alliance longevity may sometimes be a precursor for alliance performance. However, alliance longevity needs to come out of the shadow of being a proxy for alliance performance because the two concepts are not one and the same (Ariño, 2003; Makino et al., 2007). In fact, our results show that alliance performance is not related to alliance longevity at a statistically significant level. A short-lived alliance may terminate prematurely because the alliance successfully met its objectives. A long-lived alliance may linger due to inattention. These paradoxical scenarios highlight the need to study alliance longevity in its own right. Theoretical contributions Recent reviews of the alliance literature have not recognised many studies on alliance longevity (Gray and Wood, 1991; Gulati, 1998). The current research addressed this void in the alliance literature in several ways. First, there is clearly a bias on the parts of both strategy researchers as well as practitioners to focus excessively on performance. Stepping away from the mainstream inquiry about alliance performance, this study inquires what determines longevity in alliances. As a response, this research offers a model of structural determinants of alliance longevity, focusing on relational and operational aspects of alliances. Hence, the core theoretical contribution of this paper is the development of a structural model of alliance longevity. Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

12

Variables

Model 1 Coefficient

Asset Differential Predetermined Longevity Mean Year (Time Effects) Repeating Firms (Cluster Effects) Alliance Type Alliance Performance Hierarchy Specific Experience Alliance Units General Experience Hierarchy  Specific Experience Alliance Units  General Experience Constant Square root of 4 Square root of q r (variance explained) Log Likelihood L(b) Wald c2 (significance) N

10.9615*** 1.1053 2.4901*** 0.1438 0.9535 0.8731

3.3435 7.6854*** 14.8697*** 0.2108 23897.188 149.36*** df( 6) 5759

Data on 167 alliances, spanning 5759 alliance-months. * p < .05. ** p < .01. *** p < .001.

Model 2 Standard Error

Significance Level

1.7799 1.9113 0.2406 0.9424 1.5328 0.9408

0.000 0.563 0.000 0.879 0.534 0.353

2.1615 0.5130 0.1408 0.0226

0.122 0.000 0.000

Coefficient 14.4059*** 4.9833 1.9199 11.2763 4.6630 2.0635 7.0787* 7.2094*** 19.0735*** 22.1431***

43.6200*** 32.5485*** 11.7376*** 0.8849 22802.113 3005.74*** df( 10) 5759

Model 3 Standard Error

Significance Level

1.5720 7.5933 0.9287 3.7588** 6.0083 3.7312 3.2722 0.7354 5.5330 0.4621

0.000 0.512 0.039 0.003 0.438 0.580 0.031 0.000 0.001 0.000

8.4570 1.9001 0.1112 0.0121

0.000 0.000 0.000

Coefficient

Standard Error

Significance Level

14.5716*** 4.4724 1.9164* 11.3345** 4.7032 1.8083 9.4808** 10.0394*** 26.7273*** 19.7419*** 1.9741*

1.5722 7.6389 0.9341 3.7818 6.0437 3.7536 3.4612 1.4669 6.5792 1.1901 0.8134

0.000 0.558 0.040 0.003 0.436 0.630 0.006 0.000 0.000 0.000 0.015

1.8990*

0.8557

0.026

8.5348 1.9158 0.1111 0.0120

0.000 0.000 0.000

42.5276*** 32.7392*** 11.7232*** 0.8864 22796.215 3028.66*** df( 12) 5759

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

Table 3 Simultaneous tobit coefficients, standard errors, and significance levels

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

13

It should be reiterated that alliance longevity is not simply a proxy for alliance performance, but rather a separate concept in itself. Whereas longevity has a survival orientation, performance has a growth orientation. Admittedly, survival is necessary to allow growth in alliances. However, this does not mean that greater alliance longevity will ensure or even reflect better alliance performance. To tease out confounding effects of alliance performance on alliance longevity, we controlled for alliance performance when testing the effects of our independent and moderating variables. Shorter alliance longevity may even be interpreted as an indicator that an alliance is no longer needed because it has successfully accomplished its goal. Studying computer mediated groups to simulate interactions of groups that are located in different places, Wilson et al. (2006) assert that “[t]he changes in trust over time that we observed support arguments for incorporating a temporal dimension into theories of relational development.” (p. 27). Considering the great number of alliances that are formed by firms that are located in different places, a temporal dimension ought to be incorporated into theories of alliance development. This study on alliance longevity attempted to move in that direction and fill some of the extant void in the interorganisational relationship literature. Second, this study extended the application of transaction cost logic to examine the effect of hierarchy of alliance structure on alliance longevity. Conceptually we argued that hierarchy of alliance structure is positively related to alliance longevity, and empirical results supported our hypotheses. Thus, recent findings on the joint venture paradox (Stern, 2005) may not be applicable for the entire alliance population. According to the paradox, when searching for a joint venture partner, firms gravitate toward partners that have established legitimacy and capabilities (i.e., age and size); however, these partners tend to lack the flexibility needed to sustain new ventures. If this paradox were applicable, then our hypotheses pertaining to hierarchy of alliance structure and alliance longevity would not have been supported. Third, scholars have criticised transaction cost economics’ static nature (i.e., focusing on one transaction) and have called for a dynamic orientation of transaction cost economics that considers multiple transactions over time (Rindfleisch and Heide, 1997). To that end, to account for the development of relationships between partnering firms over multiple alliances, the effects of prior alliance engagements between partnering firms were considered. The results show that prior alliance engagement between alliance partners had significant independent and moderating effects on alliance longevity. Not only did specific alliance experience positively influence alliance longevity, but it also diminished the positive relationship between hierarchy of alliance structure and alliance longevity. Fourth, this study also includes concepts from the dynamic capabilities perspective to address relational aspects of alliances. Prior studies have not taken into account roles of partnering firms’ general alliance experience and alliance unit on alliance longevity simultaneously in the same model. While the dynamic capabilities perspective is a firm level theory, we transformed its firm level variables to apply them to the alliance level. For example, existence of alliance unit in firm structure (Kale et al., 2002) was transformed to number of partners with alliance units. Similarly, partnering firms’ general alliance experiences (Simonin, 1997; Zollo et al., 2002) were transformed through aggregation and adjustment for asymmetry. These transformations allowed for the inclusion of structural variables that have seldom been examined at the alliance level before. Last, findings by Kale et al. (2002) suggest that alliance units gather, interpret, codify and disseminate knowledge from general alliance experience. Building on these premises, we argued that number of partners with alliance units would enable partnering firms to manage their alliances better, thereby expediting the completion of alliance-specific tasks. For instance, by allocating necessary resources to different alliances as needs arise, an alliance unit would make the functioning of alliances more efficient. Efficient alliances would complete their tasks quicker and be more likely to dissolve sooner. We explained that since an alliance unit could quickly identify the usefulness of a particular alliance, the alliance would cease to exist as soon as the alliance unit assesses the usefulness of the alliance to be over. Alliance units are also quick to detect poorly performing alliances that do not have any hopes of becoming successful due to various contingencies. Such alliances are not allowed to linger on any longer than they deserve when partnering firms’ alliance units closely monitor alliance performance. Clearly, alliance longevity is shortened when alliances experience intended terminations due to faster goal accomplishment and unintended terminations due to early detection of failure. Managerial implications Theoretical understanding of alliance longevity has managerial implications as well. Many promising alliances fail to produce satisfactory results because of their inadequate longevity, which remains an elusive concept to alliance practitioners. Alliance managers need to recognise that greater alliance longevity may translate to more time to work on the alliance to yield satisfactory results. However, prolonging the lifespan of an alliance for escalated commitment to the collaborative project can be very counterproductive for partnering firms. A weakly performing alliance can strain the partners’ resources. Whether an alliance would require a short or long longevity is an executive decision that ought to take place prior to the formal formation of the alliance. The relational and operational determinants should be considered to ensure that the projected longevity is well aligned with the desired longevity. Partnering firms may have differing ideas of how long they want their alliance engagement to last. Das (2006) notes that “an alliance is formed with an explicit or implicit understanding among the partners that it will dissolve at some point in the future when its purpose is achieved” (p. 3). This intended timespan of the alliance or alliance horizon is negotiated by the prospective partners in light of their respective alliance horizon preference. The greater the misalignment of alliance horizon preferences, the more difficult it will be to negotiate an agreement of the alliance horizon. Our research might be particularly Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

14

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

useful in this context, because if managers know how to influence longevity, then they could bring differing expectations into synch with one another. When managing alliances, it can be tempting to take longevity as a signal of performance. If alliance managers step into this trap, they may erroneously prolong those alliances that have weak performance. Even from an agency perspective, there is incentive for alliance managers to extend the longevity of weakly performing alliances as a way of defending poor alliance management. In contrast, when alliances perform exceptionally well and accomplish goals expeditiously, dissolving them early may seem counterintuitive to practitioners. Also, there has always been concern that firms wary of alliance failure often terminate their alliances too early, preventing some of those alliances from maturing into strongly performing collaborations. Deciding a priori what kind of longevity is suitable for a firm depends not only on the complexity of the collaborative project but also the available resources for deployment. Alliance managers must take these critical issues into account when they determine how far into the future they want their collaboration to survive. Depending on this subjective projection, alliance managers can use the key levers we identified in our research to have a better control over the longevity of their alliances. To that end, this study conceptually argued that there are four structural determinants of alliance longevity, representing both relational and operational aspects of strategic alliances: hierarchy of alliance structure, specific alliance experience, alliance units and general alliance experience. All of these predictors significantly influenced alliance longevity. The results inform alliance managers about the specific structural conditions they need to focus on to achieve the desired longevity for their alliances. The relative magnitude of their effects allows managers to prioritise among structural determinants and decide which ones they want to focus on before the others. For example, our results indicated that joint ventures would have greater longevity than minority equity alliances, and minority equity alliances would have greater longevity than contractual alliances. While alliance practitioners might buy extra time for themselves by adopting a minority equity or joint venture structure, this may not be a realistic option in many instances. If alliances pose substantial relational hazards, managers should adopt a more hierarchical structure (e.g., joint venture or minority equity) to govern such collaborative ventures. However, such relational concerns dissipate when firms form repeated alliances with each other. Thus, the more hierarchical alliance structures may become a bureaucratic burden rather than a useful governance mechanism when firms with specific alliance experience with each other form strategic alliances. Also, general alliance experience can extend alliance longevity. However, if partnering firms of an alliance have different levels of general alliance experience, that difference may have potential negative effects on the focal alliance’s longevity. Therefore, alliance practitioners need to be concerned with general alliance experience of their own firms as well as that of their prospective partners. General alliance experience of the potential partner should be included in the checklist when managers engage in the partner selection process. Furthermore, alliance practitioners should note that alliance longevity is not synonymous to alliance performance. In fact, due to inertial reasons, many underperforming alliances may continue to survive beyond their useful lives. We explained earlier that an alliance unit can detect such alliances and dissolve them once they have lived out their useful lives. Thus, the negative relationship between alliance units and alliance longevity suggests that such a structural feature may help partnering firms to better manage their alliances by preventing poorly performing alliances from surviving longer than needed. Notwithstanding the merits of an alliance role or unit, managing such a role or unit is expensive. Thus, alliance practitioners planning to invest in this structural feature should take into account their firms’ expected number of alliances and their significance. The dynamic capabilities of alliancing generated by the alliance unit have to be applicable to a large pool of significant alliances in order to realise a net gain from the alliance unit investment. The greater the expected number of alliances and their significance, the higher would be the potential benefits from an alliance role or unit. Limitations and future research The current research may be extended in several paths. While we developed our model of the effects of relational and operational aspects of strategic alliances on alliance longevity by drawing from transaction cost economics and dynamic capabilities literature, there are other theoretical streams pertinent to strategic alliances that we do not consider in our model. In his lab experiment, Axelrod (1984) observed that the TIT FOR TAT game performed better than any other games. He argued that fear of future retributions (i.e., shadow of the future) encouraged co-operative behaviour rather than defection. In another study, Parkhe found that pattern of payoffs (i.e., payoffs from mutual or unilateral co-operation and mutual or unilateral defection) is significantly related to alliance performance (Parkhe, 1993). Thus, game type characterising an alliance relationship may be an influential determinant of alliance longevity (i.e., alliances characterised by the TIT FOR TAT game may last longer than those characterised by CHICKEN or STAG HUNT games). Future research may develop methodology to categorise alliances into various economic games and test whether game theoretic predictions find empirical support. Concepts from network theory have the potential to explain alliance longevity too. For example, the resources available within the network would determine the extent to which member firms value their network membership. This in turn is likely to decrease the probability of defection by any partnering firm as long as the alliance is with another firm within the same network. Also, indirect linkages between partnering firms through common suppliers or vendors may have a strong influence in fostering interfirm trust. The “strength of weak ties” may positively influence alliance longevity (Granovetter, 1973). Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

15

While behavioural uncertainty is well accounted for with transaction cost reasoning, environmental uncertainty is unaccounted for in this study. Alliances with an R&D component may face greater levels of market uncertainty. Then again, R&D alliances are generally formed between horizontal partners, and behavioural uncertainty is high in such an alliance type. Along these lines, Reddy et al. (2002) observed that joint ventures are more prevalent in medium R&D-intense industries with reciprocal knowledge flows, whereas technical agreements are more prevalent in highly R&D-intense industries. Separating the effects of behavioural and environmental uncertainties is beyond the scope of this study. Future research needs to incorporate the independent and interactive influences of different kinds of uncertainties to enrich the alliance longevity model. Acknowledging the different dynamics in dyadic and multi-partner alliances, we refrained from treating them as the same. The model we develop is specifically relevant to dyadic alliances, which comprise about 87.5 per cent of the 69,000 alliances formed during 1985 and 2001 (SDC Joint Ventures/Alliances Database). Future research may use this model as a foundation to develop a customised model of multi-partner alliance longevity. Das and Teng (2002) argued that as the number of partners increases, so too does the relational volatility and co-ordination challenges of the alliance. Examining data from 80 joint ventures, Garcia-Canal et al. (2003) found that “while relational investment significantly influences the effectiveness of dyadic JVs, formal control is pivotal in the case of multi-party JVs” (p. 743). The determinants examined in this study are expected to have a more salient role in multi-partner alliance longevity. Structural conditions explain alliance longevity at the alliance level, whereas market, legal or political conditions more likely explain alliance longevity at the industry level. Hence, an extension of the current research should explore market, legal and political conditions affecting alliance longevity. We also did not include international alliances in the empirical research, as the structural variables derived from transaction cost economics and the dynamic capabilities perspective do not explain the subtle complexities and challenges that international alliances face. Consequently, the results of this study pertain to US domestic alliances only. Nevertheless, cultural differences may be a salient determinant of international alliance longevity. We can point out one relevant variable (e.g., alliance purpose) that we were unable to account for in our model due to data limitations. Strategic alliances with different purposes, such as R&D, supply procurement, marketing, co-production and codevelopment, may have different expectations and utilities of longevity. For example, R&D alliances may be destined to higher rates of failure than are marketing agreements because of the uncertainty associated with R&D. Thus, alliance longevity of a certain level may be considered adequate for R&D alliances, but inadequate for marketing alliances that assume ongoing marketing support from the partners. The rich variation of such typology is partially captured by the ‘alliance type’ control variable of this paper. Future research should examine more closely the richness of alliance types. Lastly, alliance longevity, marked by its termination (or lack thereof), is an outcome of self-selected activities of partnering firms (Hamilton and Nickerson, 2003). We noted termination decision as a potential source of sample selection bias and tested the extent of such bias using the Heckman two-step process. The Mill’s lambda was not statistically significant, suggesting no serious threat of sample selection bias due to selecting alliances that had terminated. Also a cautionary note on generalisability needs to be addressed. Generalisability of the overall findings is limited to the extent to which longevity data were verifiable for the alliance. The results of this study do not tell us anything about the longevity of such alliances that do not have verifiable longevity data. More specifically, the findings of this study are relevant and applicable to dyadic US domestic alliances formed by publicly-traded firms, for which alliance longevity and alliance performance data are verifiable (i.e., confirmed ongoing status or termination). Acknowledgements The authors thank Giovanna Padula, Russ Coff, Bill McEvily and the two anonymous reviewers of Long Range Planning for their valuable suggestions on earlier drafts of this paper. The study benefited from the research assistance of Mark Krivis at Pace University and Zhu Zhu at Baruch College. References Adner, Roy, Levinthal, Daniel A., 2004. What is not a real option: considering boundaries for the application of real options to business strategy. Academy of Management Review 29, 74–85. Ariño, Africa, 2003. Measures of strategic alliance performance: an analysis of construct validity. Journal of International Business Studies 34, 66–79. Artz, Kendall W., Norman, Patricia M., 2002. Buyer-supplier contracting: contract choice and ex post negotiation costs. Journal of Managerial Issues 14, 399–417. Axelrod, Robert, 1984. The Evolution of Cooperation. Basic Books, New York. Banning, Kevin C., 1999. Ownership concentration and bank acquisition strategy: an empirical examination. International Journal of Organizational Analysis 7, 135–152. Barkema, Harry G., Shenkar, Oded, Vermeulen, Freek, Bell, John H.J., 1997. Working abroad, working with others: how firms learn to operate international joint ventures. Academy of Management Journal 40, 426–442. Borenstein, Michael, Cohen, Jacob, Rothstein, Hannah, 1997. Power and Precision. Erlbaum, Lawrence Associates Software and Alternative Media, Mahwah, NJ. Breen, Richard, 1996. Regression Models: Censored, Sample Selected, or Truncated Data. Sage, Thousand Oaks, CA. Breusch, Trevor S., Pagan, Adrian R., 1979. A simple test for heteroscedasticity and random coefficient variation. Econometrica 47, 1287–1294. Brown, Shona L., Eisenhardt, Kathleen M., 1997. The art of continuous change: linking complexity and time-paced evolution in relentlessly shifting organizations. Administrative Science Quarterly 42, 1–34.

Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

16

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

Cuypers, Youtha, Martin, Xavier, 2008. Tie versus tie: when do corporate development activities strengthen or disrupt buyer-supplier ties? Academy of Management Best Papers Proceedings, 30/1–30/6. Das, T.K., 2006. Strategic alliance temporalities and partner opportunism. British Journal of Management 17, 1–21. Das, T.K., Teng, Bing-Sheng, 1996. Risk types and inter-firm alliance structures. Journal of Management Studies 33, 827–843. Das, T.K., Teng, Bing-Sheng, 1998. Between trust and control: developing confidence in partner cooperation in alliances. Academy of Management Review 23, 491–512. Das, T.K., Teng, Bing-Sheng, 2000. Instabilities of strategic alliances: an internal tensions perspective. Organization Science 11, 77–101. Das, T.K., Teng, Bing-Sheng, 2002. Alliance constellations: a social exchange perspective. Academy of Management Review 27, 445–456. Levin, Doran P., 1990. Chrysler-Renault Project is Ended, New York Times, June 12, D1. Dussauge, Pierre, Garrette, Bernard, 1995. Determinants of success in international strategic alliances: evidence from the global aerospace industry. Journal of International Business Studies 26, 505–530. Eisenhardt, Kathleen M., Martin, Jeffrey A., 2000. Dynamic capabilities: what are they? Strategic Management Journal 21, 1105–1121. Feltrin, Ariverson, 2002. Carrera’s Exclusive Contract with Renault Ends, Gazeta Mercantil Online, January 8. Fryxell, Gerard E., Dooley, Robert S., Vryza, Maria, 2002. After the ink dries: the interaction of trust and control in us-based international joint ventures. Journal of Management Studies 39, 865–886. Garcia-Canal, Esteban, Valdes-Llaneza, Ana, Ariño, Africa, 2003. Effectiveness of dyadic and multi-party joint ventures. Organization Studies 24, 743–770. García-Casarejos, Nieves, Alcalde-Fradejas, Nuria, Espitia-Escuer, Manuel, 2009. Staying close to the core: lessons from studying the costs of unrelated alliances in Spanish banking. Long Range Planning 42, 194–215. Glaister, Keith W., Buckley, Peter J., 1998. Measures of performance in UK International alliances. Organization Studies 19, 89–118. Geringer, Michael J., 1991. Strategic determinants of partner selection criteria in international joint ventures. Journal of International Business Studies 22, 41–62. Geringer, Michael J., Hebert, Louis, 1991. Measuring performance of international joint ventures. Journal of International Business Studies 22, 249–263. Glaister, Keith W., Buckley, Peter J., 1999. Performance relationships in UK international alliances. Management International Review 39, 123–147. Granovetter, Mark S., 1973. The strength of weak ties. American Journal of Sociology 78, 1360–1380. Grant, Robert M., 1996. Prospering in dynamically-competitive environments: organizational capability as knowledge integration. Organization Science 7, 375–387. Gray, Barbara, Wood, Donna, 1991. Collaborative alliances: moving practice to theory. Journal of Applied Behavioral Science 27 (1), 3–22. Greene, William H., 1990. Econometric Analysis. Macmillan, New York. Greve, Henrich R., Baum, Joel A.C., Mitsuhashi, Hitoshi, Rowley, Timothy J., 2010. Built to last but falling apart: cohesion, friction, and withdrawal from interfirm alliances. Academy of Management Journal 53, 302–322. Gulati, Ranjay, 1995. Does familiarity breed trust? The implications of repeated ties in contractual choice in alliances. Academy of Management Journal 38, 85–112. Gulati, Ranjay, 1998. Alliances and networks. Strategic Management Journal 19, 293–317. Gulati, Ranjay, Singh, Harbir, 1998. The architecture of cooperation: managing coordination costs and appropriation concerns in strategic alliances. Administrative Science Quarterly 43, 781–814. Hafner, Reinhold, Mader, Wolfgang, 2007. How to deal with longevity risk: a capital market perspective, 27–30. In: Borsch, A. (Ed.), Asia-Pacific Pensions 2007. Allianz Global Investors AG, Munich, Germany. Hamilton, Barton H., Nickerson, Jack A., 2003. Correcting for endogeneity in strategic management research. Strategic Organization 1 (1), 53–80. Harrigan, Kathryn R., 1988. Strategic alliances and partner asymmetries. In: Contractor, F.J., Lorange, P. (Eds.), Cooperative Strategies in International Business. Lexington Books, Lexington, MA, pp. 205–226. Hatfield, Louise, Pearce, John A., Sleeth, Randall G., Pitts, Michael W., 1998. Toward validation of partner goal achievement as a measure of joint venture performance. Journal of Managerial Issues 10, 355–372. Heckman, James, 1979. Sample selection bias as a specification error. Econometrica 47 (1), 153–161. Heimeriks, Koen H., Klijn, Elko, Reuer, Jeffrey J., 2009. Building capabilities for alliance portfolios. Long Range Planning 42 (1), 96–114. Hennart, Jean-Francois, 1988. A Transaction costs theory of equity joint ventures. Strategic Management Journal 9, 361–374. Hoang, Ha, Rothaermel, Frank T., 2005. The effect of general and partner-specific alliance experience on joint R&D project performance. Academy of Management Journal 48, 332–345. Hoffmann, Werner H., 2005. How to manage a portfolio of alliances. Long Range Planning 38 (2), 121–143. Holmberg, Stevan R., Cummings, Jeffrey L., 2009. Building successful strategic alliance: strategic process and analytical tools for selecting partner industries and firms. Long Range Planning 42 (2), 164–193. Kale, Prashant, Dyer, Jeffrey H., Singh, Harbir, 2002. Alliance capability, stock market response, and long-term alliance success: the role of the alliance function. Strategic Management Journal 23, 747–767. Kanter, Rosabeth M., 1994. Collaborative advantage: the art of alliances. Harvard Business Review 72 (4), 96–108. Kim, Linsu, 1998. Crisis construction and organizational learning. Organization Science 9, 506–521. Kotabe, Masaaki, Martin, Xavier, Domoto, Hiroshi, 2003. Gaining from partnerships: knowledge transfer, relationship duration, and supplier performance improvement in the U.S. and Japanese automotive industries. Strategic Management Journal 24, 293–316. Laird, Nan M., Ware, James H., 1982. Random-effects models for longitudinal data. Biometrics 38, 963–974. Levinthal, Daniel A., Mark Fichman, 1988. Dynamics of interorganizational attachments: auditor-client relationships. Administrative Science Quarterly 33, 345–369. Luo, Yadong, 2002. Product diversification in international joint ventures: performance implications in an emerging market. Strategic Management Journal 23, 1–20. Maddala, G.S., 1982. Limited dependent variable models using panel data. Journal of Human Resources 22, 307–338. Maddala, G.S., 1983/1994. Limited-Dependent and Qualitative Variables in Econometrics. Cambridge University Press, Cambridge, UK. Makino, Shige, Chan, Christine M., Isobe, Takehiko, Beamish, Paul W., 2007. Intended and unintended termination of international joint ventures. Strategic Management Journal 28, 1113–1132. McEvily, Bill, Marcus, Alfred, 2005. Embedded ties and the acquisition of competitive capabilities. Strategic Management Journal 26, 1033–1055. McEvily, Bill, Perrone, Vincenzo, Zaheer, Akbar, 2003. Trust as an organizing principle. Organization Science 14, 91–103. Mowery, David C., Oxley, Joanne E., Silverman, Brian S., 1996. Strategic alliances and interfirm knowledge transfer. Strategic Management Journal 17 (Winter Special Issue), 77–91. Nelson, Richard R., 1991. Why do firms differ, and how does it matter? Strategic Management Journal 12, 61–74. Nelson, Richard R., Winter, Sidney G., 1982. An Evolutionary Theory of Economic Change. Harvard University Press, Cambridge, MA. Oxley, Joanne E., 1997. Appropriability hazards and governance in strategic alliances: a transaction cost approach. Journal of Law, Economics & Organization 13, 387–409. Pangarkar, Nitin, 2003. Determinants of alliance duration in uncertain environments: the case of the biotechnology sector. Long Range Planning 36, 269– 284. Park, Seung Ho, Russo, Michael V., 1996. When competition eclipses cooperation: an event history analysis of joint venture failure. Management Science 42, 875–890. Park, Seung Ho, Ungson, Gerardo R., 1997. The effect of national culture, organizational complementarity, and economic motivation on joint venture dissolution. Academy of Management Journal 40, 279–307.

Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003

N. Rahman, H.J. Korn / Long Range Planning xxx (2014) 1–17

17

Park, Seung Ho, Ungson, Gerardo R., 2001. Interfirm rivalry and managerial complexity: a conceptual framework of alliance failure. Organization Science 12, 37–53. Parkhe, Arvind, 1993. Strategic alliance structuring: a game theoretic and transaction cost examination of interfirm cooperation. Academy of Management Journal 36, 794–829. Parkhe, Arvind, 2001. Interfirm diversity, organizational learning, and longevity in global strategic alliances. Journal of International Business Studies 22 (4), 579–601. Poppo, Laura, Zenger, Todd, 2002. Do formal contracts and relational governance function as substitutes or complements? Strategic Management Journal 23, 707–725. Powell, Walter W., Koput, Kenneth, Smith-Doerr, Laurel, 1996. Interorganizational collaboration and the locus of innovation: networks of learning in biotechnology. Administrative Science Quarterly 41, 116–145. Rahman, Noushi, 2004. Determinants of Alliance Longevity: An Empirical Examination of Factors from Transaction Cost Economics and the Dynamic Capabilities Perspective. Unpublished doctoral dissertation. City University of New York, New York, NY. Rahman, Noushi, 2007. Duality of alliance performance. Journal of American Academy of Business 10 (2), 21–28. Rahman, Noushi, 2008. Resource and risk trade-offs in guanxi-based IJVs in China. Asia Pacific Business Review 14 (2), 233–251. Rahman, Noushi, Korn, Helaine J., 2010. Alliance structuring behavior: relative influence of alliance type and specific alliance experience. Management Decision 48 (5), 809–825. Ramaswamy, Kannan, 1997. The performance impact of strategic similarity in horizontal mergers: evidence from the U.S. banking industry. Academy of Management Journal 40, 697–715. Reddy, Sabine B., Osborn, Richard N., Hennart, Jean-Francois, 2002. The prevalence of equity and non-equity cross-border linkages: Japanese investments and alliances in the United States. Organization Studies 23, 759–780. Reuer, Jeffrey J., Ariño, Africa, 2007. Strategic alliance contracts: dimensions and determinants of contractual complexity. Strategic Management Journal 28, 313–330. Reuer, Jeffrey J., Zollo, Maurizio, 2005. Termination outcomes of research alliances. Research Policy 34, 101–115. Rindfleisch, Aric, Heide, Jan, 1997. Transaction cost analysis: past, present, and future applications. Journal of Marketing 61, 30–54. Ring, Peter S., Van de Ven, Andrew H., 1994. Developmental processes of cooperative interorganizational relationships. Academy of Management Review 19, 90–118. Sadowski, Bert, Duysters, Geert, 2008. Strategic technology alliance termination: an empirical investigation. Journal of Engineering & Technology Management 25, 305–320. Sampson, Rachelle C., 2004. Organizational choice in R&D alliances: knowledge-based and transaction cost perspectives. Managerial & Decision Economics 25, 421–436. Simonin, Bernard L., 1997. The importance of collaborative know-how: an empirical test of the learning organization. Academy of Management Journal 40, 1150–1174. Staw, Barry, 1981. The escalation of commitment to a course of action. Academy of Management Review 6, 577–587. Stern, Ithai, 2005. The Joint-Venture Paradox: Parent-Firm Characteristics, Social Cues, and Joint Venture Performance, Unpublished doctoral dissertation, University of Texas at Austin, Austin, TX. Teece, David J., Pisano, Gary, Shuen, Amy, 1997. Dynamic capabilities and strategic management. Strategic Management Journal 18, 509–533. Von Hirschhausen, Christian, Neumann, Anne, 2008. Long-term contracts and asset specificity revisited: an empirical analysis of producer-importer relations in the natural gas industry. Review of Industrial Organization 32, 131–143. Wilson, Jeanne M., Straus, Susan G., McEvily, Bill, 2006. All in due time: the development of trust in computer-mediated and face-to-face teams. Organizational Behavior and Human Decision Processes 99, 16–33. Yakita, Akira, 2006. Life expectancy, money, and growth. Journal of Population Economics 19, 579–592. Zaheer, Akbar, McEvily, Bill, Perrone, Vincenzo, 1998. Does trust matter? Exploring the effects of interorganizational and interpersonal trust on performance. Organization Science 9, 141–159. Zaheer, Akbar, McEvily, Bill, Perrone, Vincenzo, 1998. The strategic value of buyer-supplier relationships. International Journal of Purchasing and Materials Management 34, 20–26. Zollo, Maurizio, Reuer, Jeffrey J., Singh, Harbir, 2002. Inter-organizational routines and performance in strategic alliances. Organization Science 13, 701–713. Zollo, Maurizio, Winter, Sidney G., 2002. Deliberate learning and the evolution of dynamic capabilities. Organization Science 13, 339–351.

Biographies Noushi Rahman is a Professor of Management in the Lubin School of Business at Pace University. He received his PhD from the Zicklin School of Business at Baruch College/Graduate Center, City University of New York. His research interests include managing alliance outcomes, corporate social and environmental responsibility and evolution of competitive advantage. E-mail: [email protected] Helaine J. Korn is an Associate Professor of Management in the Zicklin School of Business at Baruch College, City University of New York. She received her PhD from the Stern School of Business at New York University. Her research interests include multi-market strategies, firm strategy evolution and structural factors affecting alliance outcomes. E-mail: [email protected]

Please cite this article in press as: Rahman, N., Korn, H.J., Alliance Longevity: Examining Relational and Operational Antecedents, Long Range Planning (2014), http://dx.doi.org/10.1016/j.lrp.2012.05.003