Lorenzo Maria Stanca (Northwestern University)

“Recursive Preferences, Correlation Aversion, and the Temporal Resolution of Uncertainty”

Abstract. Models of recursive utility are commonly associated with a preference for early resolution of uncertainty, often regarded as an important economic channel in applications. This paper provides a different understanding of recursive preferences based on attitudes toward correlation, and in particular aversion to intertemporally correlated risks. I formalize and investigate such a property. I show that an increase in correlation makes a decision maker that prefers early resolution worse off, even when increasing correlation also provides noninstrumental information about future consumption. Relatedly, I show that one can separate risk aversion from intertemporal substitution by considering a domain of choice in which pure preferences for early resolution of uncertainty play no role. Finally, I apply the insights of this paper to better understand the features possessed by existing models of recursive utility. I argue that attitudes toward correlation are the key behavioral feature driving the results of consumption-based asset pricing models.