Multifactor Models and Their Consistency with the APT

Abstract

We examine the consistency of several prominent multifactor models from the empirical asset pricing literature with the arbitrage pricing theory (APT) framework. We follow the APT-related literature and estimate the common factor structure from a rich cross-section (associated with 42 major CAPM anomalies) by employing the asymptotic principal components method. Our benchmark model contains six statistical factors and clearly dominates, in both economic and statistical terms, most of the empirical multifactor models proposed in the literature by a good margin. These results represent a critical challenge to the current workhorse models in terms of explaining large-scale equity risk premiums.

Keywords: asset pricing; linear multifactor models; APT; equity risk factors; statistical factors; stock market anomalies; cross-section of stock returns; asymptotic principal components; spanning regressions

JEL Classification: G10; G12