通过额外兑付改善协调困境:一项实验研究

Pregame Resource Exchange as a Mechanism in Overcoming the Coordination Dilemma:An Experimental Study

当多方参与者面临资源互补的协调博弈时,尽管各方利益一致,以往研究却发现参与者往往最终都选择不付出资源,从而整个集体陷入协调困境。本研究创新地设计了一个可以利用这些未付出的资源改善协调困境的额外兑付机制:允许参加者在博弈之前使用他们在协调博弈中不打算付出的资源,并以此为信号进行正向归纳,从而使得各方能够协调一致。通过对照实验,本文证实了该机制对改善协调困境效果显著。存在机制干预的小组协调成功的比重达到了91.7%,相比对照组提高了50%。在此基础上,本研究通过两个辅助实验进一步确认了机制的作用。本文证实了正向归纳在改善协调困境方面对人群整体有着很好的效果,同时机制中发送信号的成本基于结果而变化的特点也能较好地激发参与者发送信号的意愿。本文提出的机制有着实施简单、适用限制较少的优点,具有较好的现实应用意义。

Organizational contexts often require the efficient coordination of independent agents with strategic complementarities. However, given that hard work benefits all, team members tend to exert minimal effort. Thus, an important organizational question is how to improve the coordination within a team when it is not possible to directly incentivize employee effort. We observe a phenomenon that has not been noted in the literature. Specifically, when team members are stuck in a state of inefficient coordination, the available and remaining efforts that are not devoted to coordination can be considered a waste of resources. Although in reality, team members tend to allocate these efforts to other work, we believe that the remaining efforts could be an important tool for improving coordination. We use a novel design, by adding a pre-stage before the coordination game, to examine whether publicly revealing one's efforts in other work has a positive effect in improving coordination.We study the experimental frameworks of Van Huyck et al.(1990), which comprise seven symmetric, Pareto-ranked, pure strategy equilibria on the main diagonal with four players. We replicate the design in Van Huyck et al.(1990) as the control treatment. In the main treatment, we add a pre-stage before the coordination game. In the pre-stage, the players are allowed to directly exchange some of their six units of effort for payment, with an exchange rate of 1:1. The units exchanged can be observed by the group members. The members then play the coordination game, in which they are allowed to use any of the remaining units. We use two additional treatments to separately test the effects of the two key features in the main treatment, making resource exchange observable and an endogenously generated strategy space. In the silent treatment, we implement the mechanism but hide the information on the subjects exchange choices. In the artificial treatment, we remove the exchange stage in the main treatment and exogenously give the subjects the same strategy spaces used in the main treatment. We find that extra pre-game resource exchanges effectively improve the coordination, with 91.7% of the groups in the main treatment reaching coordination success, which is 50% higher than the control treatment. The performance in the silent treatment and artificial treatment is not significantly different from that in the control treatment. This implies that the better performance in the main treatment is attributable to the endogeneity of the strategy spaces formed by exchanging resources, and that sharing information on resource exchange makes forward induction feasible.Our paper makes three contributions to the literature. First, we use the available but remaining resources that are not used in the coordination task as a tool to improve coordination. Here, the efforts allocated to the pregame and coordination game are complementary. Hence, the resource exchange in the pregame can serve as an effective communication tool and truthfully signal an individual's intention in the coordination task. Second, we propose a novel way of signaling, in which the cost of communication dynamically changes over the periods, which serves as the signal of the strategy in the coordination task. The cost of communication converges to zero when the team members achieve efficient coordination. Numerous studies have found that costless communication can be effective in improving coordination but with demanding requirements, such as in large groups when all of the members send signals. Different from the costless and costly communication discussed above, our results show that with the dynamic costs used in our mechanism, the team members are willing to signal their intentions for coordination. Third, although forward induction has been found to be an important tool, it introduces sample selection problems. In our study, all of the players participate in both the pregame and the coordination game. As a result, forward induction functions in all of the samples, which thus resolves the selection problem, and can explain 82.2% of the improved coordination.In summary, we propose a novel design to ensure the pregame communication confirms both the goal and the strategy. Our mechanism allows all players to conduct forward induction and makes the players' commitments reliable and enforceable.