An Improved Approach for Modeling Bank Loan Default in Pursuit of Sustainable Banking

Noura Metawa^*, Rania Itani

Assistant Professor of Finance, University of Sharjah, Sharjah, UAE

Emails: nmetawa@sharjah.ac.ae;  rania.itani@dmu.ac.uk

Abstract

This article presents our research effort to explore the convergence of sustainable banking practices and predictive modeling for bank loan defaults, with a primary emphasis on addressing the pressing need for resilient financial systems. To this end, an applied methodology is presented in this study to model bank loan defaults, emphasizing the incorporation of sustainability criteria into predictive analytics. Given the temporal nature of load data, our approach leverages Long Short-Term Memory (LSTM) networks as its backbone process for predictive modeling.  The empirical results of the public case study underscored the enhanced predictive accuracy completed through this approach, emphasizing the pivotal function of integrating sustainability metrics in predicting mortgage defaults inside the banking area.

Keywords: Loan Default; Sustainable Finance; Credit Risk Modeling; Banking Sustainability; Machine Learning; Financial Stability; Default Probability Estimation; Risk Management; Econometric Modeling; Banking Practices.