This is a Research Planning Grant for Women in Science and Engineering. The standard framework for competitive general equilibrium in markets relies on a centralized auctioneer to establish prices and execute trades. In contrast, the majority of actual transactions in the economy are decentralized bilateral trades between individuals. The modeling and analysis of bilateral trading mechanisms with local interactions both complements the existing general-equilibrium theoretical framework and provides an understanding of the characteristics of many existing markets. The key assumption in the project is that trade may occur only between agents who are adjacent to one another on a network. Such a fixed spatial structure of local interactions may arise from geographic obstacles or from institutional restrictions on trade such as the United States' embargo of Cuba. Other markets, such as those for over-the-counter stocks and bonds, have evolved and maintained a decentralized structure even in the absence of restrictions on trade. Understanding the properties of network-based bilateral trading may help to explain the existence and importance of these market structures. A trading mechanism based on local interactions also addresses an important criticism of existing research on decentralized trade, namely, that the number of potential trading pairs and the corresponding information requirements increase dramatically with the number of agents. With the introduction of spatial relationships, the same results can be achieved with far fewer direct interactions between agents. Although all gains from trade eventually will be exhausted, the longer it takes for the economy to converge to a common trading price, the larger the `cost` of purely local trading. Preliminary computer simulations comparing a variety of network structures confirm the importance of this issue. Not only does the speed of convergence vary greatly for different spatial structures, the process of convergence generates interesting dynamics such as the persistence of spatial correlations, or neighborhood effects, in the prices of goods over time. The assumption that agents are perfectly myopic, and hence unwilling to engage in trades which lower their current utility, plays a crucial role in existing results. Because agents do not anticipate the final equilibrium price, trades occur at prices other than those that will ultimately prevail economy-wide so long as those trades are utility improving. The project also investigates the changes in the equilibrium distribution of wealth induced by local trading and examine how the process of trading can generate path-dependent outcomes.
