Modeling bank systemic risk of emerging markets under geopolitical shocks: Empirical evidence from BRICS countries

In this work, we proposed the Bank Risk Interlinkage with Dynamic Graph and Event Simulations (BRIDGES) to analyze the banking sector in BRICS nations. We implemented a Dynamic Time Warping (DTW) method to construct a dynamic network of 551 BRICS banks from 2008 to 2024 using annual balance sheets. A Temporal Graph Neural Network (TGNN) identified anomalous structural changes within the bank network over time. Agent-Based Model (ABM) simulations assess the impact of anomalies on network stability and evaluate the resilience of banking systems to internal financial distress and exogenous geopolitical shocks at the country and BRICS levels. Our simulation outcomes present key insights. The failure of major BRICS banks creates greater systemic risks than the collapse of insolvent or anomalous banks, due to induced panic. Also, relative to the crisis involving the largest banks, a nationwide correlated shock yields greater systemic disruption, leading to near-total systemic collapse. Panic over the failure of major BRICS banks and a countrywide, correlated geopolitical shock seems to pose the biggest threats to the financial stability of BRICS nations, largely absent from typical bank risk analysis models. BRIDGES detects fragility (e.g., for Russia) where standard static metrics (such as z-scores) identify stability, elevating the behavioral dimension relative to the structural network aspect, as extensively documented in studies of contagion. BRIDGES can thus quantify hidden aspects, such as the vulnerability of state-owned banks, the behavioral premium of large banks, and the extreme, nonlinear nature of damage from correlated geopolitical events compared to standard economic crises.