Subjective Model Uncertainty, Variance Risk Premium, and Speculative Trading

Abstract

Motivated by a novel empirical finding that variance risk premium (VRP) predicts trading volume, we analyze an asset pricing model where agents are initially uncertain about their subjective models for interpreting public news announcements. Such a setting is micro-founded by ambivalence in psychology and obtains closed-form solutions. Our model explains the negative uncertainty premium in options and endogenously generates VRP. In particular, the initial uncertainty about signal precision (mean) sharply predicts that options and VRP are unspanned (spanned) and that VRP negatively (positively) predicts future trading volume.

Keywords: subjective model uncertainty, signal precision or mean, option-implied uncertainty premium, variance risk premium, trading volume

JEL Classification: G11, G12, G13