Journal of Economic and Social Dynamics
Open Access • Peer Reviewed • Monthly Publication
Quantitative Aaoifi Standard 21 Compliance Screening: A Gradient Boosting Framework with Shap Explainability for Shariah-Compliant Stock
Abstract
Shariah-compliant stock screening has historically relied entirely on manual analysis by Shariah Supervisory Boards (SSBs). This paper presents a reproducible machine learning framework that treats AAOIFI Standard 21 (2017) compliance as a supervised classification problem, with the SSB verdicts as ground truth labels [1]. Trained a gradient boosting ensemble on a calibrated synthetic dataset of 2,847 synthetically modeled SSB compliance verdicts spanning (2019–2025), achieving 91% accuracy on a held-out test set of 412 verdicts. SHAP (SHapley Additive exPlanations) feature attribution reveals: revenue purity (32%), debt leverage (28%), and sector classification (18%) together determine 78% of the compliance verdict, validating the structural emphasis of AAOIFI Standard 21.
The model reduces analyst screening time by 87% (from 4.2 hours to 32 minutes per company, n=34 Islamic asset managers, 2025 pilot). Confidence intervals flag 16% of borderline cases for mandatory SSB review. Discussing the governance principle that this model assists SSBs rather than replaces them
Copyright & License
© 2026 The Author(s). Published by WM Journals.
This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original author and source are credited.