Yannis Ioannides

Paper #: 96-04-020

The paper reviews the evolution of trading structures by examining two pertinent strands in the literature on economies with interacting agents: one, works that presume a specified topology of interactions among agents, and two, works that let random mechanisms determine that topology. The paper reviews interactive discrete choice models in isotropic settings and proposes extensions within certain stylized anisotopic settings which are particularly interesting for economists. In particular, circular patterns of interaction highlight the role of money and credit; tree-type settings depict Walrasian interactions. The paper suggests that the random topology of interaction approach, which has employed random graph theory to study evolution of trading structures, may go beyond analyses of sizes of trading groups and thus exploit the full range of possible topological properties of trading structures. The paper proposes an integration of those approaches which is intended to exploit their natural complementaries. In the simplest possible version, our sythesis involves individual decisions and expectations, randomness, and nature combining to fix an initial “primordial” topology of interaction. The dynamics of interaction move the economy from then on. The evolution of trading structures depends critically upon multiplicity and stability properties of equilibrium configurations of the interaction model. The paper addresses a number of additional topics, including matching models, spatial aspects of the evolution of trading structures and issues of statistical inference.