Information diffusion, cluster formation and entropy-based network dynamics in equity and commodity markets

Abstract

This paper investigates the dynamic causal linkages among U.S. equity and commodity futures markets via the utilization of complex network theory. We make use of rolling estimations of extended matrices and time-varying network topologies to reveal the temporal dimension of correlation and entropy relationships. A simulation analysis using randomized time series is also implemented to assess the impact of de-noising on the data dependence structure. We mainly show evidence of emphasized disparity of correlation and entropy-based centrality measurements for all markets between pre- and post-crisis periods. Our results enable the robust mapping of network influences and contagion effects while incorporating agent expectations.