We consider the dynamic private provision of funds to projects that generate public benefits. Participants have complete information about the environment, but imperfect information about individual actions: each period they observe only the aggregate contribution. Each player may contribute any amount in any period before the contributing horizon is reached. All Nash equilibrium outcomes are characterized. In many cases they are all also perfect Bayesian equilibrium outcomes. If the horizon is long, if the players' preferences are similar, and if they are patient or the period length is short, perfect Bayesian equilibria exist that essentially complete the project. In some of them the completion time shrinks to zero with the period length—efficiency is achieved in the limit.

We study the aggregate implications of (S, s) inventory policies in a dynamic general equilibrium model with aggregate uncertainty. Firms in the model's retail sector face idiosyncratic demand risk, and (S, s) inventory policies are optimal because of fixed order costs. The distribution of inventory holdings affects the aggregate outcome in two ways: variation in the decision to order and variation in the rate of sale through the pricing decisions of retailers. We find that both mechanisms must operate to reconcile observations that orders are more volatile than, and inventory investment is positively correlated with, sales, while remaining consistent with other salient business cycle characteristics. The model exhibits strong amplification for some shocks and persistence to a limited extent.

The paper extends Diamond's (1984) analysis of financial intermediation to allow for risk aversion of the intermediary. As in the case of risk neutrality, the agency costs of external funds provided to an intermediary are relatively small if the intermediary is financing many entrepreneurs with independent returns. Even though the intermediary is adding rather than subdividing risks, the underlying large-numbers argument is not invalidated by the presence of risk aversion. With risk aversion of entrepreneurs as well as the intermediary, financial intermediation provides insurance as well as finance. In contrast to earlier results on optimal intermediation policies under risk neutrality, the paper shows that when an intermediary is financing many entrepreneurs with independent returns, optimal intermediation policies must shift return risks away from risk averse entrepreneurs and impose them on the intermediary or on final investors.

This paper introduces a theory of market incompleteness based on the information transmission role of prices and its adverse impact on the provision of insurance in financial markets. We analyse a simple security design model in which the number and payoff of securities are endogenous. Agents have rational expectations and differ in information, endowments, and attitudes toward risk. When markets are incomplete, equilibrium prices are typically partially revealing, while full relevation is attained with complete markets. The optimality of complete or incomplete markets depends on whether the adverse selection effect (the unwillingness of agents to trade risks when they are informationally disadvantaged) is stronger or weaker than the Hirshleifer effect (the impossibility of trading risks that have already been resolved), as new securities are issued and prices reveal more information. When the Hirshleifer effect dominates, an incomplete set of securities is preferred by all agents, and generates a higher volume of trade.

Why do individuals participate in charitable gambling activities? We conduct a laboratory investigation of a model that predicts risk-neutral expected utility maximizers will participate in lotteries when they recognize that lotteries are being used to finance public goods. As predicted by the model, we find that public goods provision is higher when financed by lottery proceeds than when financed by voluntary contributions. We also find support for other comparative static predictions of the model. In particular we find that ticket purchases vary with the size of the fixed prize and with the return from the public good: lotteries with large prizes are more effective, and ticket purchases drop dramatically when the public good is not valued by subjects.