Connectedness as measures of systemic risk in the European banking system

Abstract

The recent financial crisis and sovereign debt crisis in Europe have highlighted the need for systemic risk measures for macroprudential policy purposes. This thesis im-plemented three frameworks proposed by Billio et al. (2012) and Diebold & Yilmaz (2009, 2012, 2014) to investigate the interconnectedness of the European banking sys-tem as measures of systemic risk. The frameworks are applied using market data of a sample of 28 largest banks in Europe from 2001 to 2018. The empirical results show that systemic risk measures based on market data can identify periods of financial dis-tress in the market. Besides, I also find that the European banking system in overall has become more connected, especially during the sovereign debt crisis. Different frameworks seem to depict quite different characteristics of the system connectedness. Moreover, the ranking of banks according to their contribution to the aggregate con-nectedness is not precisely consistent within itself over different periods. Nor is the banks’ ranking consistent among different frameworks. The forecast error variance de-composition framework in Diebold & Yilmaz (2009, 2012, 2014) has the best out of sample performance in terms of identifying banks with the biggest losses during the crisis period. In consistent with previous literature, I do not find a strong relationship between the connectedness measures and other measures such as MES and Delta-CoVaR. Therefore, the results call for a more systematic view of systemic risk at multi-ple aspects.