Despite the growing importance of strategic alliances, high termination rates of inter-firm collaborative agreements continue to puzzle academics and practitioners alike (e.g., Makino et al., 2007). As a result, increasing attention has been paid to further understanding factors leading to alliance instability and driving the dissolution of inter-firm partnerships. In this regard, prior work has specifically focused on destabilizing factors related to the alliance itself or the involved partners. Far less attention has been dedicated to investigating the impact of contextual factors, even though their relevance has been emphasized by various authors studying alliance evolution (e.g. Ariño & De La Torre, 1998). Only recently Greve et al. (2013) found evidence supporting the importance of outside options as a cause for instability in inter-firm partnerships and emphasize the need to distinguish between failure- and option-driven alliance termination. The latter refers to cases where firms withdraw even from well-performing alliances, due to the availability of alternative partners possessing a higher match quality with the focal firm.
In our study, we build upon this nascent work on option-driven alliance termination and examine, which characteristics of (existing and alternative) alliance partners induce decision makers to withdraw from established partnerships and enter into collaborative agreements with outside options. In this regard, we introduce the term “partner switch”, which we define as the substitution of an existing alliance partner with an alternative one that provides higher relative match quality along a set of alliance- and partner-specific characteristics. To examine factors that influence partner switching behavior, we develop a conjoint experiment consisting of six attributes derived from prior research on alliance formation and termination. These parameters encompass both economic and social drivers and barriers of partner switch as well as value creation and value appropriation considerations. We report our results based on a dataset of 2,440 experimental ratings, conducted by 122 alliance managers.