Early Warning Systems for Financial Contagion: A Network-Based Framework for Financial Supervisory Authorities

Analysis by the President of the Financial Supervisory Authority (ASF), Alexandru Petrescu, together with ASF specialists Marius Acatrinei and Silviu Nistor

Financial supervisory authorities increasingly recognise geopolitical instability as a structural driver of systemic risk. Yet existing surveillance frameworks lack real-time tools to detect contagion dynamics as they emerge. This article presents a practical early warning framework that combines nonlinear Granger-causal network analysis with spectral methods to monitor financial interconnectedness across European capital markets. Using kernel-based causality testing, we construct time-varying directed networks that capture predictive relationships between 26 European stock indices. The spectral radius, the dominant eigenvalue of the network adjacency matrix, serves as a scalar indicator of systemic fragility: values near unity indicate normal market segmentation, while sharp increases signal reg...