We report on time-varying network connectedness within three banking systems: North America, the EU, and ASEAN. The original method by Diebold and Yilmaz is improved by using exponentially weighted daily returns and ridge regularization on vector autoregression (VAR) and forecast error variance decomposition (FEVD). We compute the total network connectedness for each of the three banking systems, which quantifies regional uncertainty. Results over rolling windows of 300 days during the period between 2005 and 2015 reveal changing uncertainty patterns which are similar across regions, with common peaks associated with identifiable exogenous events. Lead-lag relationships among changes of total network connectedness of the three systems, quantified by transfer entropy, reveal that uncertainties in the three regional systems are significantly causally related, with the North American system having the largest influence on EU and ASEAN.

No abstract

No abstract

No abstract

No abstract