Ulric B. and Evelyn L. Bray Social Sciences Seminar

Abstract: We show that an unintended consequence of the independence and recursivity axioms in dynamic settings is that preferences for risk and intertemporal substitution fully determine the preference for temporal resolution of risk. We relax these axioms and characterize a utility representation that combines Kreps and Porteus (1978) with a function capturing what the decision maker learns about the future from present risk resolution. We analyze its properties and show how to apply it to dynamic optimization problems. As an example, we revisit the equity premium puzzle, showing that our model helps resolve it, while maintaining a reasonably low timing premium.