An Empirical Evaluation of Hierarchical Clustering-Based Portfolio Selection Models

Abstract

This paper investigates the empirical performance of hierarchical clustering-based asset allocation models. These models use clustering techniques to inform portfolio construction without relying on traditional optimization.
We evaluate several variants of these methods and compare their out-of-sample performance with standard portfolio selection strategies. Our results suggest that hierarchical clustering models offer competitive and often superior risk-adjusted returns, demonstrating robustness across different market conditions and asset universes.