Taming the global factor zoo

Abstract

This paper examines monthly returns from over 48,120 stocks across 36 countries/regions, employing an iterative two-step LASSO methodology to identify key factors in global markets. The result is a novel global factor model that incorporates factors such as market dynamics, profit growth, quality, momentum, investment, size, and debt issuance. This model outperforms existing approaches by providing a superior explanation of global asset pricing anomalies and achieving lower average pricing errors. A distinguishing feature of our model is its efficacy in explicating anomalies in local markets, diverging from traditional models that typically confine their scope to local market dynamics. Overall, this research highlights the potential of machine learning-based frameworks in developing a more comprehensive and robust global factor model for international asset pricing.