Title: Transitory and persistent networks
Authors: Jozef Barunik - UTIA AV CR vvi (Czech Republic) [presenting]
Michael Ellington - University of Liverpool (United Kingdom)
Abstract: A novel framework is proposed to measure connectedness from variance decompositions. The approach accounts for the characteristics of shocks creating network structures. Using frequency domain techniques, we measure connectedness that forms on the transitory and persistent component of shocks. We outline a procedure to test for statistical differences in transitory and persistent network connectedness. Monte Carlo evidence shows our measures reliably track connectedness and correctly identify statistical differences. We show that our connectedness measures enhance our understanding of systemic risks emerging from sectoral uncertainty networks. Therefore, they may serve as a monitoring tool for macro-prudential supervisors and investors alike.