Liquidity as an Insurance Problem

Risk-averse individuals wish that assets concentrate their payoffs in states of high marginal value (that is, highly likely or low endowment states). An asset or portfolio may fail to do so, by having payoffs uncorrelated to its owner needs or, even worse, by having them inversely related. The latter, which we call tier 1 illiquidity, is shown to occur in non-Walrasian markets (where a trade involves bargaining) and in incomplete Walrasian markets where optimal trading strategies are non trivial. In both cases, the high valuation of the trader biases the equilibrium price against him. The former, which we call tier 2 illiquidity, is shown to arise when individual shocks are privately observed, because moral hazard prevents contracting on them. Diamond and Dybvig (1983) and Holmström and Tirole (1998) present prominent examples of tier 2 illiquidity. However, a self-insurance model is offered to argue that the importance of this type of illiquidity is limited from a welfare perspective, provided individuals are patient enough and can trade in a perfectly competitive, complete—except for individual-level uncertainty— set of asset markets. This article characterizes an asset's liquidity as the degree of insurance it provides, thereby identifying the basic economic problem behind liquidity as one of the familiar risk-sharing kind. It also shows, by means of examples, that the problem arises when asset markets are imperfectly competitive, incomplete, or both.