Trade is both uncertain and sequential. Money surprises are not neutral because prices at the beginning of the trading process cannot depend on its end. In contrast with fixed-price models, in this paper sellers can change prices during trade. In contrast with Lucas's article, here there is no asymmetry in the information about the money supply. The price quoted by individual sellers may adjust slowly to changes in the targeted money supply, but the distribution of quoted prices adjusts perfectly to these changes and the real price distribution is independent of the anticipated rate of change in the money supply.

Trade is both uncertain and sequential. Money surprises are not neutral because prices at the beginning of the trading process cannot depend on its end. In contrast with fixed-price models, in this paper sellers can change prices during trade. In contrast with Lucas's article, here there is no asymmetry in the information about the money supply. The price quoted by individual sellers may adjust slowly to changes in the targeted money supply, but the distribution of quoted prices adjusts perfectly to these changes and the real price distribution is independent of the anticipated rate of change in the money supply.

The standard formulation of a spot market subject to uncertain excess demand uses a tatonnement process that restricts trade until the market-clearing price is found. Here I present a model in which there is no restriction on trade during the process of the resolution of uncertainty about aggregate excess demand. The idea is to enlarge the commodity space and define goods by the probability that they will sell, in addition to other characteristics. The probability of sale characteristic is the spot market analogue of the contingencies under which delivery will take place in the Arrow-Debreu model. A failure to distinguish goods by the probability of sale characteristic can lead to the rejection of the competitive paradigm even when everyone is a price taker and the allocation is efficient.

The standard formulation of a spot market subject to uncertain excess demand uses a tatonnement process that restricts trade until the market-clearing price is found. Here I present a model in which there is no restriction on trade during the process of the resolution of uncertainty about aggregate excess demand. The idea is to enlarge the commodity space and define goods by the probability that they will sell, in addition to other characteristics. The probability of sale characteristic is the spot market analogue of the contingencies under which delivery will take place in the Arrow-Debreu model. A failure to distinguish goods by the probability of sale characteristic can lead to the rejection of the competitive paradigm even when everyone is a price taker and the allocation is efficient.

The public good aspect of information is used to account for periods in which the aggregate level of trade is low. It is shown that abstention from trade may occur when the uncertainty with respect to the market-clearing price (there is no auctioneer) gets large relative to the cost of getting information about it and relative to the cost of postponing transactions. In this case, all agents are aware of bilateral Pareto-improving trading opportunities, but these opportunities are not exploited.

The public good aspect of information is used to account for periods in which the aggregate level of trade is low. It is shown that abstention from trade may occur when the uncertainty with respect to the market-clearing price (there is no auctioneer) gets large relative to the cost of getting information about it and relative to the cost of postponing transactions. In this case, all agents are aware of bilateral Pareto-improving trading opportunities, but these opportunities are not exploited.

The paper considers the choice between two finite income paths that are subject to random variations. It is shown that if one income path, X, has more cumulative variation at the outset and less variation toward the end than another income path, Y, then X dominates Y in the sense that (almost) every decision maker prefers X to Y. The intuitive reasoning is that a decision maker with income path X can, by engaging in a fair gamble, create the random variation of the cumulative income under Y and still have the advantage that more of the uncertainty is resolved at an earlier stage.

[The paper considers the choice between two finite income paths that are subject to random variations. It is shown that if one income path, X, has more cumulative variation at the outset and less variation toward the end than another income path, Y, then X dominates Y in the sense that (almost) every decision maker prefers X to Y. The intuitive reasoning is that a decision maker with income path X can, by engaging in a fair gamble, create the random variation of the cumulative income under Y and still have the advantage that more of the uncertainty is resolved at an earlier stage.]