Doctoral Dissertation Research DRMS: Experimental Evidence and Measurement of Inter-Temporal Strategic Inefficiency

本研究从重复战略互动的角度来考察效率。也就是说，这项研究考虑了几种类型的游戏，它们是对更复杂的现实世界情况的简化表示，在这些情况下，互动发生在任意长的时间范围内的多个阶段。研究人员将研究这些游戏中回报结构的增强如何为代理人提供激励，使其默契地串通，获得比重复博弈更有效的结果。这里提出的问题不仅从机制设计的角度来看是重要的，而且在重复的寡头垄断相互作用中也有应用。首先，我们关心的是相对回报的大小如何使一个特定的结果比其他结果更具吸引力来协调，这可能是社会上次要的。对于每种类型的战略情况，我们评估了实验对象对的相对和绝对效率，根据每一阶段的平均回报，看看在博弈的限制下，代理是否能够有效地协调以获得尽可能高的回报。关于如何提高效率的知识对决策者和受限于一组特定战略参数的代理人都很重要。决策者可以决定博弈的战略参数。

This study considers efficiency from the perspective of repeated strategic interaction. Namely, the study considers several types of games that are simplified representations of more complex real world situations in which interaction occurs at multiple periods over an arbitrarily long time horizon. The researcher will examine how augmentation of the payoff structure in these games can provide incentives for agents to tacitly collude at a more efficient outcome over repeated play. The issues raised herein are important not only from a mechanism design standpoint, but also have applications in repeated oligopoly interaction. Primarily, we are concerned with how the relative magnitudes of payoffs may make a particular outcome more attractive for agents to coordinate upon in comparison to other outcomes, which may be socially inferior.For each type of strategic situation studied, we assess the relative and absolute efficiencies of experimental subject pairs in terms of average per period payoffs to see whether or not agents can effectively coordinate to extract as high a payoff as possible given the constraints of the game. Knowledge of how to induce such efficiencies is important to both policy makers, who may determine the strategic parameters of a game, and to agents confined to a particular set of strategic parameters.