A Theory of Monetary Exchange
用计算复杂性刻画不同交换机制的成本，证明引入货币可将计算复杂度降至nH，并指出分散信息的问题。
The transactions cost for alternative exchange mechanisms for the household exchange problem can be characterized by the computational complexity of the exchange process. The computational complexity for any exchange mechanism is at least nH, where n is the number of goods and H is the number of households. Imposing the conditions of conservation, nonnegativity and quid pro quo results in a command exchange mechanism whose computational complexity is nH. Multiparty barter exchange, formalized using graph theory, has computational complexity equal to the minimum of (nH2, n2H). Introducing an auxiliary good, money, reduces the computational complexity to nH. A problem with decentralized information is demonstrated.