Toward a modelling of public goods provision : theory and experiment

在一个有公共物品的经济体中，要实现帕累托有效配置并非易事，因为公共物品的提供存在一个称为搭便车问题的激励问题。然而，格罗夫斯和莱德亚德在1977年提出了一种在纳什均衡概念下实现帕累托有效分配的机制。在以往关于公共物品的机制设计中，隐含地假设没有任何主体可以选择不参与该机制，而所有主体都必须参与。然而，在许多实际情况下，如国际条约，缔约方可能可以选择不参与所提出的机制，因此一些缔约方可以搭乘他人提供的非排他性公共产品的利益的便车。我们明确地纳入了非排他性，并研究了一个自愿参与的模型，使得每个主体可以决定她是否参与该机制。不幸的是，我们证明了一个不可能定理，即不可能设计一个每个代理都有参与动机的机制。我们在日本和美国进行了各种实验来验证这一定理。在日本的实验中，决定参与的受试者采取行动惩罚没有参与的受试者，尽管参与的受试者的收益减少。因此，没有参与的受试者了解到，不参与对他们没有好处，因此他们开始定期参与。也就是说，合作的来源似乎不是利他主义或善意，而是对其他受试者恶意行为的最大化回报反应。另一方面，美国受试者的行为符合理论预测，即美国数据支持进化稳定的均衡预测。一个悬而未决的问题是分析不同国家之间的这些显著不同的行为。

It is not easy to attain a Pareto efficient allocation in an economy with public goods, because the provision of public goods has an incentive problem called the free-rider problem. However, Groves and Ledyard proposed a mechanism that achieves a Pareto efficient allocation under the Nash equilibrium concept in 1977. Subsequently, numerous mechanisms have been proposed that satisfy additional desirable properties.In the previous mechanism design on public goods, it was implicitly assumed that no agents have a choice not to participate in the mechanism and that all agents are required to participate. In many practical circumstances such as international treaties, however, parties may have a choice not to participate in the mechanism proposed and hence some of them can free-ride on the benefit from a non-excludable public good provided by others.We explicitly incorporate non-excludability and examine a model with voluntary participation such that each agent can decide whether she participates in the mechanism or not. Unfortunately, we prove an impossibility theorem that it is impossible to design a mechanism in which each agent has an incentive to participate. We conducted various experiments in both Japan and the United State to verify this theorem. In Japanese experiments, subjects who decided to participate took actions so as to punish subjects who did not participate, although the payoffs of participating subjects decreased. Non-participating subjects thus learned that non-participation was not beneficial to them and hence they began regularly participating. That is, it seemed that the source of cooperation was not altruism or kindness but was a payoff- maximizing response to the spiteful behavior of other subjects. On the other hand, American subjects behaved as the theory predicted, i.e., the American data support the evolutionarily stable equilibrium predictions. An open question is to analyze these significantly different behaviors across countries.