Abstract
The increasing complexity of global financial systems has necessitated advanced methodologies for monitoring systemic risk and contagion. This study investigates the dynamic interconnectedness and spillover effects among major global financial markets using a Dynamic Bayesian Network (DBN) approach. Unlike static network models, DBNs allow for the estimation of time-varying dependencies and the identification of causal transmission channels during periods of market stress. Utilizing a comprehensive dataset of daily returns from equity, commodity, and currency markets spanning the period 2010–2023, we quantify the evolution of network topology and the intensity of contagion. Our findings reveal that financial interconnectedness is highly regime-dependent, with significant surges observed during the COVID-19 pandemic and subsequent geopolitical shifts. We identify systemically important nodes that act as primary transmitters of shocks, demonstrating that the 'ripple effect' of financial contagion is more pronounced in highly integrated markets. The results indicate that DBNs provide a superior framework for capturing the non-linear and lagged dependencies inherent in financial data compared to traditional VAR-based models. Furthermore, the analysis highlights the role of economic policy uncertainty as a driver of cross-market spillovers. These results have significant implications for macro-prudential regulation, suggesting that monitoring the dynamic structure of financial networks is crucial for early warning systems and the mitigation of systemic collapse. The study concludes that policy interventions should focus on enhancing the resilience of central network hubs to prevent the cascading failure of the global financial architecture.