SVB Shock and Risk Repricing Among Selected Major Chinese Financial Institutions: Parameter Stability, Event Evidence, and Spillover Reconfiguration

This study investigates whether the March 2023 failure of Silicon Valley Bank (SVB) coincided with changes in market-risk exposure, abnormal returns, and return connectedness among five major Chinese financial institutions. Using daily data from January 2022 to December 2023, we apply CAPM and market-model regressions, structural-break tests, event-study methods, and generalized forecast-error variance decompositions. The revised design distinguishes 10 March from the first subsequent Chinese trading day, 13 March, and adds symmetric event windows, placebo tests, an alternative risk-free rate, return-series audits, significance tests, and VAR diagnostics. Full-sample Chow tests identify breaks for ICBC and CITIC Securities, but only ICBC remains significant within the ±60-day window, and none are significant within ±30 days. Only ICBC records a significant positive CAR over [−3, +3]. Overall connectedness falls from 55.53% to 47.98%, while CITIC Securities becomes the largest post-event net transmitter. The evidence therefore indicates selective repricing and reconfigured linkages rather than a system-wide effect, and does not support unique causal attribution to SVB.