What difference do new factor models make in portfolio allocation?
Abstract
This paper examines the economic implications of new factor models and shows that the Hou, Xue, and Zhang (HXZ, 2015a) four-factor model outperforms the Fama and French (FF5, 2015a) five-factor model for investing in anomalies in- and out-of-sample. The difference in certainty-equivalent returns between the two models can be more than 6% per year under modest model uncertainty and margin requirements. The outperformance of the HXZ model appears to come from its better ability to describe the mean rather than the covariance matrix of asset returns.
This paper has been withdrawn.