The Four Seasons of Commodity Futures: Insights from Topological Data Analysis

Abstract

This study introduces a new technique to analyse the evolution of correlations for multiple time series. The technique is based on applying Topological Data Analysis (TDA) and we use it to gain insights about the evolution of commodity futures markets over the 1997-2017 period. Our findings complement the existing literature and provide new insights into the dynamics of commodity futures markets in the past two decades. Our analysis has both global and local aspects and could be applied to detect changes in correlation structure in a variety of time series as well as cross sectional settings.