EventsThe 2nd International Online Conference on Risk and Financial Management
Published
This submission belongs to the session S2. Machine Learning for Next-Generation Risk Modeling of the event The 2nd International Online Conference on Risk and Financial Management
Published date
28 Oct, 2026
Academic Editor
author-avatarThanasis Stengos
Citation
Malak Ouahdi, Azz El Arab Gantare, Maryem Naili, Traditional Determinants of Bank Performance and the Incremental Role of ESG in the MENA Region: Evidence from Fixed-Effects Regression and Elastic Net Machine Learning, in Proceedings of The 2nd International Online Conference on Risk and Financial Management, 2 November–4 November 2026, MDPI: Basel, Switzerland
Share
Email
Facebook
Twitter
LinkedIn

Traditional Determinants of Bank Performance and the Incremental Role of ESG in the MENA Region: Evidence from Fixed-Effects Regression and Elastic Net Machine Learning

image
1. School of Business Administration, Al Akhawayn University, Ifrane, Morocco.
Abstract

The growing integration of Environmental, Social, and Governance (ESG) considerations into financial decision-making has intensified interest in whether sustainability-related factors provide information beyond traditional determinants of bank performance. Despite this growing attention, evidence from the Middle East and North Africa (MENA) remains limited, even though the region offers a distinctive institutional setting characterized by the coexistence of Islamic and conventional banking systems, relatively low ESG disclosure and heterogeneous macroeconomic conditions. Addressing this gap, this study investigates two interconnected questions across the largest listed commercial banks in ten MENA economies: (i) identifying and ranking key drivers of bank profitability using both econometric inference and machine learning, and (ii) assessing whether the inclusion of ESG information improves the explanatory power of the fixed-effects regression model and the predictive performance of Elastic Net models. Using a panel of 51 banks, regression results indicate that funding liquidity and deposit ratio are the most economically and statistically important determinants of bank performance, while GDP growth also exerts a robust positive effect. In contrast, the effects of bank size, capital adequacy, credit risk, and inflation vary between models. Furthermore, results show that neither the aggregated ESG measure nor its individual dimensions is associated with bank profitability after controlling for bank-specific and macroeconomic factors. The Elastic Net analysis yields consistent evidence, identifying traditional financial variables as the dominant predictors of bank profitability, whereas the inclusion of ESG variables does not improve predictive accuracy. By integrating econometric inference with machine learning within a single analytical framework, this study provides one of the first comprehensive assessments of both the explanatory and predictive relevance of ESG for bank performance in the MENA region. The findings suggest that traditional financial fundamentals continue to dominate ESG metrics in explaining and predicting bank profitability, offering important implications for investors, bank managers, regulators, and policymakers.

Keywords
Sustainable Finance
Bank profitability
Financial determinants
ESG
MENA
Fixed effects
Elastic Net
Machine learning
Panel data.
Machine-Learned Scarcity Numeraires for Water–Energy Risk and Sustainable Asset Valuation
Industry Risk in Asset Pricing: A Machine Learning Approach