On the overhaul of the spillover index: connectedness amongst European financial institutions

Abstract

Quantifying systemic risk in interconnected financial markets remains a central challenge for regulators and policymakers, particularly in the aftermath of recurring financial crises. This paper proposes the Square Root Matrix (SQRTM) method as an orthogonalization technique within the Spillover Index framework, addressing the order-dependence of Cholesky decomposition and the economic inconsistency of Generalized Forecast Error Variance Decomposition (GFEVD). Mathematically, SQRTM is shown to yield a unique, order-invariant orthogonalization that preserves row-sum unity in the connectedness table – properties that neither Cholesky nor GFEVD-based Spillover Indices satisfy. Empirically, the method is applied to the volatility of 29 leading European financial institutions over the period 2007–2024. The results reveal an average connectedness of 67.8%, reaching a peak of 96.51% in the window ending March 2020, driven by overlapping European crises and the early COVID-19 market stress. A mean-reverting trend is observed throughout the sample, with no immediate signals of systemic financial distress as of 2024. The SQRTM-enhanced Spillover Index is recommended as a robust tool for systemic risk monitoring by central banks and regulatory bodies.