Practical Applications of Hierarchical Clustering-Based Asset Allocation

Practical Applications Summary

In Hierarchical Clustering-Based Asset Allocation, published in the 2017 special multi-asset-class issue of The Journal of Portfolio Management, Thomas Raffinot of SILEX IP tests the performance of various asset allocation strategies, including several based on hierarchical clustering techniques. He compares the performance of portfolios constructed via hierarchical clustering methods with several defined-risk budgeting portfolios. He concludes that the portfolios constructed via hierarchical clustering outperform the others.

Hierarchical clustering techniques potentially provide a simpler approach to achieving diversification than covariance matrixes that must specify a correlation parameter for each pairing of assets.

TOPICS: Big data/machine learning, portfolio management/multi-asset allocation