174 X Jiang et al. /Journal of International Management 14(2008)173-189 dimensions into account. Hennart(2006: 1623)suggests that the main task in alliance research is to"keep it simple Hence, this study focuses on the stability dimension in particular Stability is vital for alliance survival, development and evolution, and it provides a necessary condition and a good proxy for performance gains and alliance success(Dussauge and Garrette, 1995; Beamish and Inkpen, 1995 Surprisingly however, few research efforts have been devoted to creating a comprehensive understanding of the stability issue in the strategic alliance field. With respect to the question of whether or not alliances are by nature unstable and transitional, we still have only limited knowledge. Some well-known alliances-for example, the IBM Sony-Toshiba R&D alliance and the Disney-McDonald's-Coca-Cola marketing alliance- have operated stably for many years and are considered successful. Many others, however, have been terminated shortly after they were formed. Alliance stability remains one of the least understood aspects of alliances and a big challenge for alliance researchers en In this paper, we argue that at the center of this academic gap might be the lack of a rigorous conceptualization of alliance stability. The gap may also result from limited knowledge about antecedents, elements and consequences of alliance stability. To fill this gap, we will first stress the need for stability research and discuss its relationship with alliance outcomes, and then we propose a precise conceptualization of alliance stability. We feel this conceptualization will be helpful in understanding the evolutionary nature of strategic alliances Furthermore, while studies examining the developmental process of alliances are increasing (e. g, Ring and Van de en, 1994; Doz, 1996; Das and Teng, 2002; De Rond and Bouchikhi, 2004), they have failed to incorporate the stability issue into this process. To date, little is known about which variables and factors may have impacts on stability in each specific stage of alliance development. For the purposes of this paper, the research initiatives appear to be fragmented and incomplete: while some researchers have focused on initial alliance conditions, others have covered factors at later stages. We contend that any one of these issues alone is unlikely to provide sufficient guidance to alliance managers, and the research stream is therefore in need of a comprehensive synthesis that organizes the existing literature. To fill this managerial relevance gap, we propose a process model in which the main antecedents of alliance stability will be examined. We argue that an alliance' s evolutionary dynamics depend on these factors and variables that the partners must assess and manage over its developmental stages We begin with an in-depth review and critique of prior research on alliance instability and stability. After that, we plain the important role of stability as a strategic variable, analyze its relationship with alliance outcomes, and offer a clear definition for it. A subsequent section focuses on the four commonly identified stages of alliance development and the key factors that actually determine the dynamics underlying alliance stability across the stages. The final section presents implications, limitations and directions for future research 2. Alliance instability and stability: a critical review Increasing academic attention to the complex evolving phenomenon of strategic alliances has led to research examining the specific dynamics within alliances. Table 1 lists a number of previous important and representative studies and their findings related to alliance instability and stability. As shown in Table 1, scholars have examined this topic in such areas as strategic alliances in general, special alliances(e. g, Bidault and Salgados multi-point alliances), equity joint ventures(hereafter: JVs), and international joint ventures(hereafter: IJVs) 2. 1. Previous research on alliance instability 2. .1. IV instability research Since the initial work undertaken by Franko(1971), the instability issue has long been a subject of research in strategic alliance literature. Notably, much of this prior research has addressed the issue in the context of [Vs. One possible reason is that international alliances that involve two or more firms across national and cultural boundaries have more fragile structures, more inter-partner conflicts, and a higher level of relational risks than other types of alliances. The inter-cultural and inter-organizational hybrid nature also suggests that the collaborative motives for IJVs and the factors underlying their instability are different from those associated with other types of alliances. As a result, UV instability has received much attention in the literature Reuer et al.(2002)argue that the root of alliance evolution research lies at the early work on DV instability. Past DV stability research has gone through a methodological process consisting of statistical observation, theoretical analysis, and empirical investigation. Early research(e.g, Franko, 1971; Killing, 1983; Gomes-Casseres, 1987)adopted a surveydimensions into account. Hennart (2006: 1623) suggests that the main task in alliance research is to “keep it simple”. Hence, this study focuses on the stability dimension in particular. Stability is vital for alliance survival, development and evolution, and it provides a necessary condition and a good proxy for performance gains and alliance success (Dussauge and Garrette, 1995; Beamish and Inkpen, 1995). Surprisingly however, few research efforts have been devoted to creating a comprehensive understanding of the stability issue in the strategic alliance field. With respect to the question of whether or not alliances are by nature unstable and transitional, we still have only limited knowledge. Some well-known alliances — for example, the IBM– Sony–Toshiba R&D alliance and the Disney–McDonald's–Coca-Cola marketing alliance — have operated stably for many years and are considered successful. Many others, however, have been terminated shortly after they were formed. Alliance stability remains one of the least understood aspects of alliances and a big challenge for alliance researchers. In this paper, we argue that at the center of this academic gap might be the lack of a rigorous conceptualization of alliance stability. The gap may also result from limited knowledge about antecedents, elements and consequences of alliance stability. To fill this gap, we will first stress the need for stability research and discuss its relationship with alliance outcomes, and then we propose a precise conceptualization of alliance stability. We feel this conceptualization will be helpful in understanding the evolutionary nature of strategic alliances. Furthermore, while studies examining the developmental process of alliances are increasing (e.g., Ring and Van de Ven, 1994; Doz, 1996; Das and Teng, 2002; De Rond and Bouchikhi, 2004), they have failed to incorporate the stability issue into this process. To date, little is known about which variables and factors may have impacts on stability in each specific stage of alliance development. For the purposes of this paper, the research initiatives appear to be fragmented and incomplete: while some researchers have focused on initial alliance conditions, others have covered factors at later stages. We contend that any one of these issues alone is unlikely to provide sufficient guidance to alliance managers, and the research stream is therefore in need of a comprehensive synthesis that organizes the existing literature. To fill this managerial relevance gap, we propose a process model in which the main antecedents of alliance stability will be examined. We argue that an alliance's evolutionary dynamics depend on these factors and variables that the partners must assess and manage over its developmental stages. We begin with an in-depth review and critique of prior research on alliance instability and stability. After that, we explain the important role of stability as a strategic variable, analyze its relationship with alliance outcomes, and offer a clear definition for it. A subsequent section focuses on the four commonly identified stages of alliance development and the key factors that actually determine the dynamics underlying alliance stability across the stages. The final section presents implications, limitations and directions for future research. 2. Alliance instability and stability: a critical review Increasing academic attention to the complex evolving phenomenon of strategic alliances has led to research examining the specific dynamics within alliances. Table 1 lists a number of previous important and representative studies and their findings related to alliance instability and stability. As shown in Table 1, scholars have examined this topic in such areas as strategic alliances in general, special alliances (e.g., Bidault and Salgado's multi-point alliances), equity joint ventures (hereafter: JVs), and international joint ventures (hereafter: IJVs). 2.1. Previous research on alliance instability 2.1.1. IJV instability research Since the initial work undertaken by Franko (1971), the instability issue has long been a subject of research in strategic alliance literature. Notably, much of this prior research has addressed the issue in the context of IJVs. One possible reason is that international alliances that involve two or more firms across national and cultural boundaries have more fragile structures, more inter-partner conflicts, and a higher level of relational risks than other types of alliances. The inter-cultural and inter-organizational hybrid nature also suggests that the collaborative motives for IJVs and the factors underlying their instability are different from those associated with other types of alliances. As a result, IJV instability has received much attention in the literature. Reuer et al. (2002) argue that the root of alliance evolution research lies at the early work on IJV instability. Past IJV instability research has gone through a methodological process consisting of statistical observation, theoretical analysis, and empirical investigation. Early research (e.g., Franko, 1971; Killing, 1983; Gomes-Casseres, 1987) adopted a survey- 174 X. Jiang et al. / Journal of International Management 14 (2008) 173–189