The Multiplex Dependency Structure of Financial Markets

We propose here a multiplex network approach to investigatesimultaneously different types of dependency in complex datasets. Inparticular, we consider multiplex networks made of four layerscorresponding, respectively, to linear, nonlinear, tail, and partialcorrelations among a set of financial time series. We construct thesparse graph on each layer using a standard network filtering procedure,and we then analyse the structural properties of the obtained multiplexnetworks. The study of the time evolution of the multiplex constructedfrom financial data uncovers important changes in intrinsicallymultiplex properties of the network, and such changes are associatedwith periods of financial stress. We observe that some features areunique to the multiplex structure and would not be visible otherwise bythe separate analysis of the single-layer networks corresponding to eachdependency measure.