Abstract:

Against the backdrop of global instability and intensifying economic deglobalization， the increasing frequency of extreme risk events—including financial crises， public health emergencies， and geopolitical conflicts—has heightened abnormal fluctuations and cross-border risk transmission in global foreign exchange markets. Understanding the heterogeneous contagion effects of different types of major shocks is thus critical for safeguarding exchange rate stability.Risk contagion in the global forex market is systematically examined through a dual-network framework， focusing on two dimensions： risk linkage （comovement） and risk spillover. Methodologically， two advanced approaches are employed. First， the R-Vine Copula model is used to construct risk linkage networks， capturing complex nonlinear and tail-dependent structures among currency pairs. Second， an Elastic Net-VAR model is applied to build directed and weighted risk spillover networks. This high-dimensional framework incorporates volatility estimates from a Skew-t-GARCH model and utilizes elastic net shrinkage techniques to effectively measure the direction and intensity of volatility spillovers. Topological analysis is further conducted by grouping currencies according to geographic region and capital openness.The empirical analysis draws on a comprehensive dataset of daily nominal broad effective exchange rate indices for 27 major economies from July 2005 to July 2023， comprising 14 developed and 13 emerging market currencies to ensure representativeness. The network structures are compared across four periods： a baseline tranquil period and three extreme event episodes—the Subprime Crisis， the Major Public Health Security Event （COVID-19）， and the Russia-Ukraine Conflict.Key findings reveal that the US dollar and Saudi riyal consistently serve as pivotal nodes in both linkage and spillover networks， acting as significant net transmitters of risk across all periods. Moreover， the evolution of risk networks is strongly event-driven： extreme events disrupt regionally clustered patterns observed during tranquil times and facilitate cross-regional contagion. A notable structural shift occurred after the Russia-Ukraine conflict， as the global exchange rate landscape exhibited signs of bloc formation. This is marked by weakened Europe-Asia cross-regional linkages and spillovers， alongside intensified risk spillovers from the US dollar to other regions. Cross-regional contagion became more concentrated among cooperating， economically similar， or geographically proximate regions. Topological analysis further shows that cross-regional risk spillovers are most severe during the Major Public Health Security Event， followed by the Subprime Crisis， and least pronounced during the Russia-Ukraine Conflict.It contributes to the existing literature by integrating the two complementary perspectives of linkage and spillover into a unified analytical framework， offering a more holistic view of risk contagion in this study. The application of R-Vine Copula and Elastic Net-VAR methods provides robust tools for capturing the high-dimensional and heterogeneous nature of global forex market interconnections. By systematically comparing the network evolution across multiple diverse extreme events， it provides nuanced insights into how different shocks reshape the global risk landscape， which is vital for formulating targeted risk monitoring and prevention policies in an era characterized by deglobalization pressures and recurrent extreme events.

Key words: exchange rate, extreme events, risk transmission, complex networks, R-Vine copula model