A Simple Robust Asset Pricing Model Under Statistical Ambiguity

Abstract

We derive and test empirically a robust single-factor asset pricing market model under statistical ambiguity. The robust model can explain the cross-section of expected U.S. stock returns without additional risk factors. Further, observed asset pricing anomalies such as size and value appear as statistical characteristics of a misspecified standard CAPM because idiosyncratic ambiguity, unlike idiosyncratic risk, cannot be diversified away. Using a robust-Bayesian econometric procedure based on relative entropy, we recover the market price of statistical ambiguity, which sets a bound on stock prices that can be interpreted as investors’ “margin of safety” against the worst-case scenario.

Keywords: asset pricing, CAPM, statistical ambiguity, robust Bayesian analysis, maximum entropy

JEL Classification: G12, C58