Odd versus even: comparative statics in multicommunity models
研究多社区模型中，当个体类型数超过社区数时，政策干预的福利效应如何取决于社区数量的奇偶性：偶数时帕累托改进的政策在奇数时可能使所有人更糟。
This paper examines the comparative statics properties of multicommunity models in which perfect sorting is not obtained because the number of types of individuals is greater than the number of communities. In general, equilibrium is inefficient and policy interventions can be Pareto improving. The paper shows that the welfare effects of a given policy depend critically on whether the number of communities is odd or even. The same policy intervention that is Pareto improving for an even number of communities can make all worse off if the number of communities is odd.