The evolution of R&D collaboration in inter-organizational project networks: Effects of reference points for competitive preference

This study examines the evolution of R&D collaboration in inter-organizational project networks by simulating the play of an appropriate game model. We assume that players display reference-dependent competitive preferences, in that each player’s utility is the sum of a material payoff and a subjective utility that depends on the difference between her material payoff and those of her neighbors. Some players, called max-players, are aggressively competitive, comparing their payoffs to the maximum payoff of their neighbors; the others, called min-players, are prudent, comparing their payoffs to their neighbors’ minimum. In each period, a player decides whether to share a high level or a low level of knowledge with her neighbors. We find that the mean collaboration level across the network decreases as a parameter representing the ability to make use of a partner’s knowledge increases. When the majority of decision-makers are min-players, a higher competitive preference intensity induces greater average collaboration, but the level of collaboration decreases as preference heterogeneity increases. In contrast, a high intensity of competitive preference prevents the diffusion of collaboration when most decision-makers are max-players, and increased competitive preference heterogeneity leads to greater collaboration. Furthermore, increasing the fraction of max-players leads to a decrease in the total payoff of the entire network. Collaboration in an inter-organizational network is more robust when most decision-makers are max-players. Nonetheless, max-players and min-players are equally likely to collaborate when either type is in the majority.