Avoiding Unpriced Risk Using a Probability-Based Approach to Security Selection

Abstract

We examine an approach to portfolio construction that that uses probability-based security analysis and find considerable evidence that the approach produces abnormal returns. High-minus-low value weighted portfolio returns are 0.50% per month and are monotonically increasing across probability quintiles. We argue that much of the information impounded in to prices is based on assumption and is unverifiable and that this, in turn, exposes prices to investor biases. Investors are therefore assuming risk of which they are unaware and for which they are not compensated. The approach we examine is able to identify this unpriced risk and thus produce abnormal returns with lower total risk.

Keywords: Required growth probability, stock selection, Fama French factors

JEL Classification: G10, G11, G14