Dynamic Time Warping Clustering for diversified portfolios: an empirical analysis

The Sharpe ratio is used to measure risk-adjusted returns for traded securities and it is quite useful to analyze long-term investments. It is one of the measures used by fund managers in the assessment of investment performance. In this guise, we assume it could be a key indicator to group securities that show similar dynamics of the Sharpe ratio over time. We use a hierarchical clustering analysis to identify groups of securities from the S&P500, which show similar dynamics of the Sharpe ratio during the last decade. Using a Dynamic Time Warping method, we identify two different clusters. To investigate how different Sharpe ratio dynamics may impact on portfolios’ diversification, we use the Most Diversified Portfolio approach (Choueifaty and Coignard, 2008) to evaluate the gain in terms of diversification and provide some insights.