INTRODUCTION

Despite a spurt of progress on the theory of incomplete security markets over the past decade or so, there is relatively little yet in the way of testable implications or practical prescriptions. This essay presents a synopsis of some of the available models and points toward some of the important issues still standing in the way of the theory's fruition. This is not a general survey of the topic.

An incomplete security market is widely taken to be one in which certain contracts are not available for trade, a definition that is perhaps too rigid. The point is that active trading is observed only in contracts with satisfactory enforcement mechanisms, sufficiently low transactions costs, and enough incentive for exchange, relative to general market opportunities. Most economists would surely agree that markets will and do open in response to the emergence of these conditions. On the other hand, there is no well-understood reason to believe that the resulting constellation of traded contracts is therefore “appropriate” in some constrained sense of efficiency, or that “markets are as complete as they ought to be.” There is also strong empirical evidence that security prices are not even approximately consistent with complete markets, at least under standard preference assumptions.

Most currently active security markets seem to have arisen in response to entrepreneurial incentives to provide markets for insurance, the need for corporations to raise capital in a convenient form, and historical accidents. The process of security market innovation, interesting in itself, also affects the pricing of securities and the allocation of contingent consumption. It is only beginning to be modeled. As yet, most of the theoretical attention of financial economists has been on complete markets models, for which wehave a reasonably clear and elegant theory without much empirical basis, or to models with an exogenously fixed set of contracts. The theory's real success story has been the application of arbitrage-free restrictions to obtain relative prices for securities that are theoretically redundant, without saying much about the allocation of consumption or the pricing of the “primitive” set of underlying securities.