Bimodal Characteristic Returns and Predictability Enhancement via Machine Learning

Han, C.

Abstract

This paper documents the bimodality of momentum stocks: both high- and low-momentum stocks have nontrivial probabilities for both high and low returns. The bimodality makes the momentum strategy fundamentally risky and can cause a large loss. To alleviate the bimodality and improve return predictability, this paper develops a novel cross-sectional prediction model via machine learning. By reclassifying stocks based on their predicted financial performance, the model significantly outperforms off-the-shelf machine learning models. Tested on the US market, a value-weighted long-short portfolio earns a monthly alpha of 2.4% (t-statistic = 6.63) when regressed against the Fama-French five factors plus the momentum and short-term reversal factors.

Citation

Han, C. (2022). Bimodal Characteristic Returns and Predictability Enhancement via Machine Learning. Management Science, 68(10), 7701-7741. https://doi.org/10.1287/mnsc.2021.4189

• Journal Article Type: Article
• Acceptance Date: May 17, 2021
• Online Publication Date: Dec 13, 2021
• Publication Date: 2022-10
• Deposit Date: May 18, 2021
• Publicly Available Date: May 19, 2021
• Journal: Management Science
• Print ISSN: 0025-1909
• Electronic ISSN: 1526-5501
• Publisher: Institute for Operations Research and Management Sciences
• Peer Reviewed: Peer Reviewed
• Volume: 68
• Issue: 10
• Pages: 7701-7741
• DOI: https://doi.org/10.1287/mnsc.2021.4189
• Public URL: https://durham-repository.worktribe.com/output/1274919