Aggregation under homogeneous ambiguity: a two-fund separation result

Abstract

Contrary to the common prior model, the construction of a representative agent whose preferences follow the multiple-priors model (1989) requires strong restrictions on sets of priors and on an aggregate endowment process if we permit a large deviation among agents’ degrees of risk aversion. This paper shows that if agents’ felicity functions belong to a family of linear risk tolerance functions with the same marginal risk tolerance, the representative agent always exists at an interior equilibrium without such restrictions, and two-fund separation holds