Long-History Principal Component Analysis in a Dynamic Factor Model with Weak Loadings

• 成果类型: Article
• 署名作者: Anderson, Robert M.; Kim, Baeho; Ryu, Dean
• 署名单位: Harbin Institute of Technology; University of California System; University of California Berkeley; Korea University; Instituto Tecnologico Autonomo de Mexico
• 刊物名称: OPERATIONS RESEARCH
• ISSN/ISSBN: 0030-364X
• DOI: 10.1287/opre.2024.1134
• 发表日期: 2026
• 关键词: variance; estimator; returns; matrix

摘要

Estimated covariance and precision matrices of asset returns significantly influence the set of portfolios compliant with risk budgets and their potential losses. Statistical risk modeling approaches often assume temporal stability for consistency with a static factor structure, typically estimated within TM-250 days of data history, resulting in finitesample estimation error when the dimension of the population exceeds the number of observations. Our study investigates the application of Principal Component Analysis (PCA) over extended data histories (e.g., TL-1,500 trading days), an approach we term Long-History PCA, to forecast the daily risk profile based on dynamic factor structures with heterogeneous factor strengths. The use of a longer data history mitigates excess dispersion bias in the estimated factor loadings, particularly in the presence of weak factors. As shown in simulations and empirical data from the United States and European stock markets, our approach substantially mitigates second-order risk bias compared with traditional methods using medium horizons (TM), both with and without the augmentation of Responsive Covariance Adjustment using a short half-life (TS) of 40 days. Open Access Statement: This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. You are free to download this work and share with others for any purpose, except commercially, and you must attribute this work as Operations Research. Copyright (c) 2026 The Author(s). https://doi.org/10.1287/opre.2024.1134, used under a Creative Commons Attribution License: https://creativecommons.org/licenses/by-nc/4.0/. Funding: R. M. Anderson gratefully acknowledges the financial support of Swiss Re through the Consortium for Data Analytics in Risk. B. Kim's research is partially supported by a Korea University Business School Research Grant. Supplemental Material: All supplemental materials, including the code, data, and files required to reproduce the results, are available at https://doi.org/10.1287/opre.2024.1134.