Ambiguity and Information Processing in a Model off Intermediary Asset Pricing

Abstract

This paper incorporates ambiguity and information processing constraints into a model of intermediary asset pricing. Financial intermediaries (specialists) are assumed to possess greater information processing capacity. Households purchase this capacity, and then delegate their investment decisions to specialists. The delegation contract is constrained by two frictions: (1) As in He and Krishnamurthy (2012), an incentive constraint arises from a moral hazard problem, which takes the form of a minimum capital requirement, and (2) Because households can invest for themselves at any time, continued delegation is subject to a participation constraint that depends on the underlying heterogeneity in channel capacity. At the same time, both households and specialists have a preference for robustness, reflecting ambiguity about risky asset returns. Ambiguity takes the form of endogenously determined pessimistic drift distortions (Hansen and Sargent (2008)). When volatility increases, so does ambiguity, since it becomes more difficult to discriminate among models. Importantly, these endogenous drift distortions produce heterogeneous beliefs. In our model, ambiguity is scaled by the inverse of time preference, and we assume specialists are more patient. Hence, given their longer investment horizons, specialists have a stronger preference for robustness. As a result, when volatility is high specialists become relatively pessimistic, and this tightens the capital constraint and accelerates the onset of a financial crisis.