Volatility Connectedness and Network Topology in the Reinsurance Sector

This thesis analyzes systemic risk and volatility spillovers in the global reinsurance market by utilizing Bayesian Vector Autoregressive (BVAR) modeling with network analysis. The financial impact of the reinsurance sector in the overall stability remains relatively unexplored in systemic risk literature, with uncertain and discordant findings. We addressed this gap by studying 27 publicly listed global reinsurance companies over the period October 2015 to May 2025, using daily stock prices to construct Yang-Zhang volatility series. We then employed a BVAR framework with hierarchical Minnesota, SUR and SOC priors, estimated through Metropolis-Hastings MCMC sampling. The adjacency matrix is derived using Generalized Forecast Error Variance Decomposition (GFEVD) and is used as the basis for graph theoretic analysis and the pairwise volatility association of the companies. The static analysis identified a highly integrated reinsurance network, with a Total Connectedness of 74.28, low modularity, and disassortative structure, with European companies identified as the main source of variance contributors, and Asian firms as principal recipients. The dynamic analysis was based on a 250-day rolling window BVAR estimation across 349 windows, uncovering a bimodal regime structure of standard equilibrium and high connectivity. We find that asset-side shocks of financial and geopolitical origin tend to amplify the overall reinsurance market connectedness. Finally we cross-validate the network centrality rankings against independent marketbased tail-risk measures and BVAR stress-test propagation. The topological hubs identified by the network coincides with the firms flagged as systemically important by these independent measures, confirming that centrality captures an economically meaningful dimensions of systemic risk rather than an abstract topological property. At the same time, no firm exhibits a positive capital shortfall under any scenario, indicating that systemic risk in reinsurance arises from interconnection and shock propagation rather than from financial leverage or capital insufficiency.