Extrapolation Bias and the Predictability of Stock Returns by Price-Scaled Variables

Abstract

Using survey data on expectations of future stock returns, we recursively estimate the degree of extrapolative weighting in investors' beliefs (DOX). In an extrapolation framework, DOX determines the relative weight investors place on recent-versus-distant past returns. DOX varies considerably over time, and the ability of price-scaled variables to predict the year-ahead equity premium is contingent on DOX. High price-scaled variables are followed by lower returns only when the DOX is high. Our findings support extrapolation-based theories of the aggregate stock market and the interpretation of price-scaled variables as proxies for mispricing. Our results help answer a critical question: when will an overvalued asset experience a correction?

JEL Classification: G02, G12, G14