Detecting Credit Risk in Egyptian Banks - Does Machine Learning Matter? - Research report

Aug 4, 2026

Upload by Sage

The study strengthens predictive modeling of credit risk in Egypt’s banking sector by distinguishing retail versus corporate credit risks and by separating banks into listed and non-listed groups. Using a comprehensive dataset covering Middle Eastern countries from 2011 to 2023, it applies machine learning—especially Random Forest—to refine credit-risk prediction. Results indicate bank-specific factors such as asset size, operating efficiency, liquidity, income diversification, and capital adequacy outperform macroeconomic indicators for both bank categories.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages17

Reading Time43 min

AccessFree to download

Introduction

Credit risk measurement and significance

Literature gaps and study contribution

Contents lists available at Vilnius University Press | Ekonomika ISSN 1392-1258 eISSN 2424-6166
2025, vol. 104(2), pp. 78–94 DOI: [https://doi.org/10.15388/Ekon.2025.104.2.5](https://doi.org/10.15388/Ekon.2025.104.2.5) | | --- | | Detecting Credit Risk in Egyptian Banks: Does Machine Learning Matter?
Doaa M. SALMAN ABDOU *
Department of Economics, Faculty of Management Sciences, October University for Modern Sciences and Arts, Cairo, Egypt Email: [dsalman@msa.edu.eg](dsalman@msa.edu.eg)
ORCID: [https://orcid.org/0000-0001-5050-6104](https://orcid.org/0000-0001-5050-6104)
Karim FARAG
Faculty of Economics and Business Administration,
Berlin School of Business and Innovation (BSBI), Germany
Email: karim.shehata@berl[insbi.com](insbi.com)
ORCID: [https://orcid.org/ 0000-0003-2661-5671](https://orcid.org/ 0000-0003-2661-5671)
Loubna ALI
Faculty of Economics and Business Administration,
Berlin School of Business and Innovation (BSBI), Germany
Email: [loubna.ali@berlinsbi.com](loubna.ali@berlinsbi.com)
ORCID: [https://orcid.org/0000-0002-6706-1890](https://orcid.org/0000-0002-6706-1890) | | Abstract. This study aims to significantly enhance the predictive modeling of credit risk within Egypt’s banking sector, particularly by differentiating between retail and corporate credit risks and categorizing banks into listed and non-listed groups. By utilizing a comprehensive dataset from Middle Eastern countries spanning 2011 to 2023, the research applies advanced machine learning techniques, including the Random Forest algorithm, to refine the predictive model.
The novelty of this research lies in its detailed exploration of credit risk determinants specific to the Egyptian banking sector, providing valuable insights into emerging economies. A distinction between various types of credit risk and bank classifications is made. The findings reveal that bank-specific factors – such asthe asset size, the operating efficiency, the liquidity, the income diversification, and the capital adequacy – are more significant predictors of credit risk than macroeconomic indicators. This trend holds for both listed and non-listed banks, thus highlighting the importance of internal metrics.
Moreover, the Random Forest algorithm demonstrates a high accuracy rate in predicting credit risk exposures, which underscores the effectiveness of machine learning in financial settings. The analysis indicates that variations in the asset size, operating efficiency, and other characteristics are crucial in influencing retail and corporate credit risks. These insights suggest that prioritizing internal bank metrics could lead to more effective credit risk management strategies than relying solely on external economic conditions.
Ultimately, this study’s predictive model is expected to enhance credit risk assessment capabilities, strengthening the financial positions of banks and fostering economic growth in the region. By bridging the gap between theoretical understanding and practical application, this research offers a novel perspective on credit risk management tailored to the unique context of the Egyptian banking sector.
Keywords: Bank-specific, retail credit risk, corporate credit risk, machine learning algorithm, Egyptian banking sector, Random Forest, institutional heterogeneity, macro-financial integration. | * Correspondent author. Received: 17/11/2024 . Revised: 04/03/2025 . Accepted: 22/03/2025 Copyright © 2025 Doaa M. Salman Abdou, Karim Farag, Loubna Ali. Published by Vilnius University Press This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Doaa M. Salman Abdou et al. Detecting Credit Risk in Egyptian Banks: Does Machine Learning Matter? Introduction Credit risk remains a pivotal determinant of banking sec

FAQs

How does the research model credit risk in Egyptian banks?

It predicts credit risk using machine learning methods, with Random Forest as a key algorithm, and it evaluates risk separately for retail and corporate exposures. Banks are also categorized into listed and non-listed groups to capture institutional differences.

What factors are found to be more important predictors of credit risk?

Bank-specific variables—asset size, operating efficiency, liquidity, income diversification, and capital adequacy—are reported as more significant predictors than macroeconomic indicators.

Why is machine learning considered useful in this context?

The Random Forest model achieves high accuracy in predicting credit risk exposures. The findings suggest internal bank metrics, identified through ML analysis, can support more effective credit risk management than relying only on external economic conditions.

Aug 4, 2026

Upload by Sage