Abstract:     Enterprises within the same industry often form a patent pool to allow cross-licensing of patents among pool members. Patent pools are an effective institutional arrangement to eliminate patent licensing barriers, save transaction costs, allow mutual acquisition of complementary technologies, reduce patent disputes, and jointly strengthen technology development and trade. Patent pools also promote efficiency because members of a pool often share technical standards. To the extent that these goals are accomplished, members within a patent pool can contain outside competition and profit in global markets. Nevertheless, some members of the same patent pool end up in litigation over patent disputes for individual interests.
   The purpose of this research was to test whether alliances between two members within patent pools reduce patent litigation sued by alliance partner, and whether the risk of litigation is higher or lower depending on characteristics of the partnership and of each enterprise. We posed the counterintuitive hypothesis that repeated alliances actually increase the risk of litigation. Based on the theories of alliance learning, coopetition and social network, this paper focused on the alliance network of patent pools, and used a manual collected database of 14,454 pairs of alliance between 76 members of 12 pools managed by MPEGLA in the period of 2006-2018. Information about these pairs was gathered from public databases. Negative binomial regression model was employed to empirically analyze the effect of alliances within patent pools on patent litigation sued by alliance partner and the moderating effect of exploratory collaboration, competitive learning, network centrality and technological similarity. 
    The empirical results show that there was a positive association between alliances within patent pools and member′s patent litigation intensity sued by alliance partner. This risk of litigation was higher when the two members had a similar technology base. The risk was lower when the partners engaged in exploratory collaboration and allowed competitive learning. The risk was also lower when one partner had higher network centrality than the other. This study expands the influence path of alliances within patent pools on the interaction between partners, and is the first to document that alliance partnerships increase the risk of litigation between partners. Despite the benefits of partnership, alliances are still subject to competition to maximize individual gain. The results have clear implications for developing strategies to reduce litigation between alliance partners, and provide evidence in support of theories in the field of management. In the current study, the risk of litigation was reduced when enterprises engaged in an exploratory collaboration. In this type of collaboration, enterprises engage in more intensive interaction to share expertise, explore new technologies, and develop new products. Closer interactions can engender mutual trust and reduce opportunistic behavior, lessening the chance of litigation over patent disputes. Strong alliance learning ability is another factor than can lower the risk of litigation. Learning ability describes each firm′s access to information about the other firm′s strengths as well as weaknesses. This mutual awareness deters each partner from litigation because learning ability strengthens R&D and the chance to make breakthroughs in key technologies, thus improving each firm′s industry status and bargaining power. 

Key words:  patent pool, exploratory collaboration, competitive learning, network centrality, technological similarity, patent litigation