April 2021 (Revised May 2022)

Dynamic Factor Copula Models with Estimated Cluster Assignments

Dong Hwan Oh and Andrew J. Patton


This paper proposes a dynamic multi-factor copula for use in high dimensional time series applications. A novel feature of our model is that the assignment of individual variables to groups is estimated from the data, rather than being pre-assigned using SIC industry codes, market capitalization ranks, or other ad hoc methods. We adapt the k-means clustering algorithm for use in our application and show that it has excellent finite-sample properties. Applying the new model to returns on 110 US equities, we find around 20 clusters to be optimal. In out-of-sample forecasts, we find that a model with as few as five estimated clusters significantly outperforms an otherwise identical model with 21 clusters formed using two-digit SIC codes.

Keywords: correlation, tail risk, multivariate density forecast

DOI: https://doi.org/10.17016/FEDS.2021.029r1

PDF: Full Paper

Related Materials: Accessible materials (.zip)

Original Paper: PDF Accessible materials (.zip)

Disclaimer: The economic research that is linked from this page represents the views of the authors and does not indicate concurrence either by other members of the Board's staff or by the Board of Governors. The economic research and their conclusions are often preliminary and are circulated to stimulate discussion and critical comment. The Board values having a staff that conducts research on a wide range of economic topics and that explores a diverse array of perspectives on those topics. The resulting conversations in academia, the economic policy community, and the broader public are important to sharpening our collective thinking.

Back to Top
Last Update: July 13, 2022