Systemic Risk Measurement: Network Analysis and Conditional Value-at-Risk

The financial events that have characterized the last decades have taught the importance of preserving financial stability and monitoring the systemic risk of financial systems. The aim of this project is to provide an analysis of a sample of 40 financial institutions in the U.S. financial landscape in order to evaluate its systemic risk over a five-year period of time, from the outbreak of the COVID-19 crisis to the end of 2024. Companies are divided into four sectors: insurance companies, commercial banks, brokers/dealers and regional banks. The techniques used include the study of financial networks (correlation-based, Granger causality networks and principal component analysis) and companies' systemic risk contribution (captured through the use of Delta CoVaR). The research is subdivided into three parts. The first chapter provides a qualitative overview of the concept of systemic risk, the ways in which it is quantified, and further references for the empirical studies. The second section conducts a broad analysis of the sample using network tools, specifically correlation-based and Granger causality networks. The former provides a description of the sample, while the second measures systemic risk by evaluating direct causal relationships. The chapter also contains a PCA application able to decompose the variance-covariance matrix of stock returns. Last, the third chapter aims to estimate Delta Conditional Value-at-Risk using two different multivariate GARCH-based definitions. These measures make it possible to capture the systemic risk contribution of the sample as well as to evaluate the applicability of different definitions.