Time-Varying Network Connectedness of G-7 Economic Policy Uncertainties: A Locally Stationary TVP-VAR Approach

Abstract

This work analyzes the frequency-dependent network structure of Economic Policy Uncertainties (EPU) across G-7 countries between January 1998 and April 2021. We implement an approach that builds dynamic networks relying on a locally stationary Time-Varying Parameter-Vector Autoregressive model using Quasi-Bayesian Local Likelihood methods. We compute short-, medium-, and long-term network connectedness of G-7 EPUs over a period covering several economic/financial turmoils. Furthermore, we structure short-term network topologies for the Global Financial Crisis (GFC) and the COVID-19 pandemic periods. Findings of the study indicate amplified interdependencies between G-7 EPUs around well-known economic/geopolitical incidents, frequency-dependent connectedness networks among them, and stronger interdependencies than the medium-, and long-term linkages. Finally, we find that short-term spillovers are not persistent in the long-term for both turmoil periods.