Machine learning as arbitrage: Can economics help explain AI?

Author

Huahao LU, Tsinghua University
Matthew SPIEGEL, Yale University
Hong ZHANG, Singapore Management UniversityFollow

Publication Type

Report

Version

publishedVersion

Publication Date

9-2024

Abstract

Machine learning algorithms have shown to be remarkably successful tools for predicting asset returns. However, the underlying economic mechanisms behind their performance remain unclear. This paper proposes a model-based dynamic arbitrage trading strategy that combines economic and statistical nonstationarity to demystify this black box. In predicting stock returns based on 153 firm characteristics (anomalies), our strategy ranks anomalies similarly to neural networks in the cross-section. Overall, it accounts for approximately 87.9 bps monthly alphas of the high-minus-low portfolios selected by neural networks in the time series. When unpublished anomalies and microcap stocks are excluded from trading, this strategy can fully explain the performance of neural networks. Our results reveal three economic sources of neural-network performance: a time varying strategy analogous to dynamic arbitrage, a tendency to weight portfolios on unpublished anomalies, and exposure to microcaps.

Keywords

Machine Learning, Dynamic Trading, Anomalies, Interpretable AI

Discipline

Finance | Finance and Financial Management | Growth and Development

Research Areas

Finance

Areas of Excellence

Digital transformation

First Page

1

Last Page

85

Publisher

Sim Kee Boon Institute for Financial Economics

City or Country

Singapore

Embargo Period

1-20-2025

Citation

LU, Huahao; SPIEGEL, Matthew; and ZHANG, Hong. Machine learning as arbitrage: Can economics help explain AI?. (2024). 1-85.
Available at: https://ink.library.smu.edu.sg/skbi/47

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.