Size Rotation in the U.S. Equity Market

Abstract

In this article, Miller, Ooi, Lee, and Giamouridis develop a hybrid model that relies on the nonlinear classification decision tree (DT) approach, and also on multivariate predictive regressions, to help implement a size rotation strategy in the U.S. equity markets. They derive an investment prediction with a two-stage algorithm. In the first stage, they use a decision tree to determine whether large-cap or small-cap stocks will outperform in the subsequent quarter. In the second stage, the authors use a multiple linear regression model to predict whether large-cap stocks will outperform or underperform small-cap stocks in the next quarter. A binary variable obtained from the first stage of the analysis—the DT model—is a key variable in the second-stage model. The authors find that a size rotation strategy based on the proposed hybrid model outperforms strategies based on the constituent models, as well as alternative strategies investigated in other studies.