Abstract Much has been written about financial intermediaries, and there is general agreement that these firms, which account for about 8%1 of U.S. gross national product, are somehow important. Despite the volume of past studies, however, research on this topic remains at a relatively primitive stage. This is primarily so because in Arrow-Debreu economies such organizations are unneeded. Until quite recently, serious analysis of intermediaries was therefore hindered by the lack of convincing general equilibrium theories that give rise to trading frictions.