Construction and empirical study of stochastic differential equations for banking systemic risk

Abstract

The banking system is composed of a large number of financial institutions closely connected through various transactions and relationships, forming a complex network that is prone to risk contagion. At the same time, financial markets are influenced by factors such as market volatility, economic cycles, and unpredictable external shocks, further increasing the complexity of systematic risk modeling and prediction. Therefore, to improve the accuracy and reduce the complexity of systemic risk in banks, this study proposes an assessment method for systemic risk in banks based on stochastic differential equations. The results showed that compared with other methods, the model fitting and prediction accuracy of assessing systemic risk in banks based on stochastic differential equations were the highest. When the number of experiments was 1,000, the model fitting degree and prediction accuracy were 0.991 and 96.3%, respectively. The extreme event capture rate and risk premium correlation of evaluating experimental data samples from 30 listed banks using a stochastic differential equation model have been improved. When the number of experiments was 1,000, the extreme event capture rate and risk premium correlation using the SDE model were 93.3% and 0.85. The proposed bank systematic risk assessment method based on stochastic differential equations has superior performance. In real-world banking scenarios, this method captures the dynamic behavior and randomness of the banking system, enabling real-time monitoring and early warning of systemic risks, improving risk management efficiency, and ensuring financial market stability.