Quantifying systemic risk in financial networks : empirical study based on China and South Africa

Abstract: The aftermath of the global financial crisis in 2008-09 has attracted increasing attention from both policymakers and researchers. Focusing on the intricacies of financial networks, this thesis comprises three empirical studies to deepen our comprehension in both direct and indirect risk contagion. The first study centers on the risk contagion of the Chinese interbank market using an agent-based modeling (ABM) approach. Examining a dataset encompassing 299 Chinese banks across five tiers from 2014 to 2019, the study employs the DebtRank algorithm to identify importance and vulnerabilities, thereby revealing risks among lower-tiered banks. Moreover, the results indicate a positively correlation between a bank\u2019s influence within the network with its size and financial connectivity. The second empirical investigation focuses on the South African interbank market over 168 months from 2008 to 2021, employing ABM and DebtRank alongside panel regression analysis. The findings reveal that size has an ambiguous effect on vulnerability between large and small banks, suggesting policymakers should not rely solely on size when monitoring a bank\u2019s vulnerability. In addition, the study underscores the impact of capital and interbank lending ratios on a bank\u2019s vulnerability, introducing the interbank-lending-to-equity multiple as a pivotal indicator for policymakers. The third study extends the analysis to indirect contagion, exploring systemic risk within Chinese financial stocks. Employing a two-layer network structure, the study incorporates return spillover and a bipartite network between mutual funds and stocks. Simulation results underscore the significance of liquidity and the flow-performance relationship of mutual funds when identifying systematically important financial institutions