The Macroeconomy and the Cross-Section of International Equity Index Returns: A Machine Learning Approach

Abstract

The paper evaluates the out-of-sample predictive potential of machine learning methods in the cross-section of international equity index returns using firm fundamentals and macroeconomic predictors. The relatively small number of equity indices in the cross-section compared to the multitude of predictive signals, makes this an ideal setting to examine return predictability using machine learning techniques. I find that macroeconomic signals seem to substantially improve out-of-sample performance, especially when non-linear features are incorporated via neural networks. The performance of a long-short country bet based on forecasted returns cannot be explained by standard definitions of risk.

Keywords: Asset Pricing, Equity Indices, Return Forecasting, Machine Learning, Neural Networks, Macroeconomic Predictability

JEL Classification: C21, C45, C58, C38, G12