Contagion and tail risk in complex financial networks

Abstract

New contagion measures based on theories of copula, heavy-tailed distributions and networks are introduced. The measures are applied to study international stock markets contagion during the Global Financial Crisis 2008. Having declined post-crisis, the contagion risk remains above its pre-crisis level for both advanced and emerging economies. A sub-network analysis of contagion shows that the shock propagated mainly from core to periphery during the crisis. We propose an instrumental variable regression approach to deal with a potential endogeneity problem in the analysis of the contagion measures as determinants of tail risk. Endogeneity might arise as both contagion measures and tail indices are themselves estimated. The obtained results are statistically significant and suggest that more contagion-central countries tend to be less prone to tail risk.