The role of artificial intelligence in developing a banking risk index: an application of Adaptive Neural...

Area of Science:

• Banking and Financial Risk Management
• Quantitative Finance
• Econometrics

Background:

• Effective banking risk measurement and management are critical for financial stability.
• Existing methods may not fully capture the multidimensional nature of banking risks.
• Policymakers and managers require robust tools for risk assessment.

Purpose of the Study:

• To propose a novel risk ranking index using Mahalanobis Distance (MD).
• To determine the relative importance of various financial ratios in assessing bank risk.
• To analyze the risk profiles of 45 Gulf banks from 2016-2020.

Main Methods:

• Utilized Mahalanobis Distance (MD) to create a multidimensional risk index.
• Employed an Adaptive Neuro-Fuzzy Inference System (ANFIS) to assess ratio importance.
• Analyzed ten financial ratios across five risk areas: Capital Adequacy, Credit, Liquidity, Earning Quality, and Operational Risk.

Main Results:

• Established confidence level thresholds for MD: 4.82 (99%), 4.28 (95%), and 4.0 (90%).
• Most of the 45 Gulf banks analyzed were found to be in a sound risk position.
• Net Interest Margin was identified as the most significant risk factor, followed by Capital Adequacy Ratio, Common Equity Tier1, and Tier1 Equity.

Conclusions:

• The proposed MD index offers a valuable tool for banking risk ranking.
• Net Interest Margin is a primary driver of banking risk.
• The study provides insights for enhancing banking supervision and risk management strategies.