Factor-Timing Model

Abstract

In this article, the authors propose a novel framework for building a factor-timing model. They introduce the concept of Akaike’s information criterion for selecting conditioning variables and evaluating various factor-timing specifications. Balancing the tradeoff between in-sample precision and complexity of the problem is the central theme of the considered methodology. In addition, the authors show how to employ information ratio quantities to track improvements in the resulting model efficacy and quantify sources of added value of factor-timing models over traditional static models.