Reason Codes and Refusal Policy in Contestable Algorithmic Credit Decisions: A Reference Design for Kazakhstani Law Tested on Public Credit Data — Reproducibility Package

Abstract

This repository provides the reproducibility package accompanying the manuscript "Reason Codes and Refusal Policy in Contestable Algorithmic Credit Decisions: A Reference Design for Kazakhstani Law Tested on Public Credit Data", prepared for submission to the International Journal of Financial Studies. The research examines the legal and technical requirements for explaining and contesting algorithmic credit decisions under Kazakhstani legislation. It evaluates reason codes, refusal policies, explanation robustness, simulated data corrections, and operational requirements using two publicly available benchmark datasets: South German Credit and Default of Credit Card Clients. The repository includes: - Research protocol and analysis configurations;- Python source code and automated tests;- Public benchmark datasets and processed data;- A verified register of relevant legal provisions;- Trained models and simulation outputs;- Explanation and simulated appeal records;- Statistical analyses, robustness checks, tables and figures;- Environment specifications and SHA-256 checksums. The authors designed the study, prepared the original research protocol, source code and manuscript, conducted the substantive analyses, and verified the final computational results, legal sources and reported findings. AI tools provided auxiliary assistance and a separately documented second-coding consistency check, which does not constitute independent validation by legal professionals. The analyses are based on historical public datasets and simulated decisions. The package contains no customer records from Kazakhstani financial institutions and no data collected directly from human participants. The results do not establish the legal compliance or performance of any real bank. Funding: Science Committee of the Ministry of Science and Higher Education of the Republic of Kazakhstan, Grant AP23489796, "Problems of regulating the legal regime of big data: domestic and international experience". The associated manuscript is prepared for submission to the International Journal of Financial Studies.