Integrating ALM and CSR in Bayesian networks for banking risk prediction: A study of OECD banks

Abstract

• Integration of CSR and ALM in a Bayesian network framework for risk prediction. • Applies the model to liquidity, interest rate and exchange rate risks. • Uses data from OECD banks between 2015 and 2022. • Demonstrates the predictive power of Bayesian methods in finance. • First study to merge ALM and CSR in Bayesian modeling for banking risks Predicting banking risks, including liquidity, interest rate, and exchange rate risks, remains a crucial challenge for financial stability. These risks arise from the interplay of various financial and non-financial factors. This paper integrates asset and liability management (ALM) components and corporate social responsibility (CSR) scores into a Bayesian network framework to predict these risks for OECD banks. The results reveal the critical role of ALM components, especially the operational efficiency ratio and capital adequacy ratio, in predicting all three banking risks. Additionally, environmental and governance scores within CSR metrics emerge as significant contributors to financial resilience, while the social score exhibits a more limited impact. Notably, to our knowledge, this is the first research to combine ALM and CSR factors in a Bayesian model for banking risk prediction, addressing a key gap in the literature and promoting financial stability.