Contagion and tail risk in complex financial networks

Abstract

The thesis develops a unified approach to the analysis of the propagation of crises as well as economic policy throughout complex financial networks. The approach builds on new copula-based contagion measures introduced in the thesis. The developed contagion measures incorporate dependence and heavy-tailedness structure of financial and economic variables dealt with, as well as the network effect. The thesis consists of three chapters. The first chapter applies the contagion measures to study international stock market contagion during the Global Financial Crisis 2008. The findings indicate that contagion in the lower tail is stronger than in the upper one. Contagion risk has come down post-crisis, however, it still remains above its pre-crisis level. Furthermore, the chapter sheds light on the analysis of determinants of tail risk and heavy-tailedness and proposes an instrumental variable regression approach to resolve the potential endogeneity problem that arises in this kind of analysis. The second chapter extends the contagion measures to a multi-layer network setting and applies them to analyse the transmission of monetary policy in the European countries' network. This is the first study, to our knowledge, that addresses policy transmission from a network perspective. The main finding of this analysis is that the policy transmits most efficiently during severe bearish contagion and is least efficient during intense bullish contagion. This finding could be attributed to the level of attention that markets pay to policy announcements during turmoil and calm periods. The third chapter, which is a joint work with Paul Nahai-Williamson, examines solvency contagion in the interbank networks. The main insight of this chapter is that network structure matters for contagion risk and can impact the total systemic losses resulting from common shocks or individual bank defaults. This chapter further highlights the importance of capturing the network effect in the first two chapters.