Any review of Kenneth Arrow's contributions to the economics of uncertainty is likely to focus on two topics: Arrow securities and measures of risk aversion, the first relating to allocation of risk in a general equilibrium setting and the second to characterizing behavior toward risk at the individual level. Arrow recognized early on that uncertainty could be incorporated in a general equilibrium setting through the introduction of contingent commodity markets (markets for trading goods contingent on the state of the world); the presence of such markets, together with other standard assumptions of the competitive model, would assure the first best allocation of risk for the same reasons that competitive equilibrium is Pareto efficient in the certainty model.

Since the set of contingent commodity markets would be absurdly large, Arrow explored the question: What is the minimal set of financial assets that, together with spot markets, would generate a first best allocation? The minimal set he identified is appropriately termed Arrow securities. The Arrow security model became the benchmark against which other market structures could be compared. It led naturally to analysis of the properties and performance of models in which the asset markets were incomplete (in that they provided fewer risk sharing opportunities than Arrow securities).

One line of research sought to explain why markets were in fact incomplete. Although some pointed to setup costs in operating the contingent markets, others focused on the presence of private information concerning the correct state of the world. The presence of such monopoly elements naturally entails a breakdown of perfectly competitive processes and leads to the consideration of bargaining games.