Beyond the Bureau: A Systematic Review and Proposed Segment-Aware Credit Risk Framework

This review of 55 credit risk studies from 2020 to 2025 establishes that the Indian personal lending literature has produced strong evidence for each of five components — borrower segmentation, UPI-based alternative data, segment-level explainability, algorithmic fairness evaluation, and regulatory compliance architecture — but has not yet assembled them within a single operational framework. The five gaps documented in Section 4 are not academic constructs. India's Economic Survey 2025-26 [6] documents that income volatility is actively preventing gig workers from obtaining formal credit. EY India [8] specifies UPI transaction data as the tool lenders need to assess this population correctly. The NBER [40] shows UPI adoption has already enabled credit access for thin-file borrowers in practice. MicroSave Consulting [7] puts the unmet credit demand at INR 80 trillion. And the RBI's Model Risk Management Circular [13] makes explainability and demographic bias testing legal obligations, not optional enhancements. The problem is real, the data infrastructure to address it exists, and the regulatory expectation is now binding. The SACRF proposed here addresses all five gaps within a design that is both regulation-compliant and operationally feasible. Section 6 demonstrates how it extends the closest existing framework [2] — Idbenjra et al.'s segmentation approach — with specific, motivated additions for the Indian context. The four testable hypotheses in Section 5E give researchers a concrete empirical agenda. Three directions stand out for follow-on work: validating the SACRF on real Account Aggregator data through the RBI sandbox programme; extending the segment definitions to cover MSME and agricultural lending portfolios [68], where the thin-file problem is equally severe; and examining privacy-preserving training mechanisms [65] for multi-lender model development that would comply with the DPDP Act 2023. The Account Aggregator framework and the expansion of India's UPI infrastructure are together creating the data conditions in which this work can be done. The research agenda the SACRF outlines is therefore timely rather than premature.

1.   FACE: FinTech Personal Loan Market Report, April 2018 to December 2024. Fintech Association for Consumer Empowerment, India (2025) 2.     Idbenjra, K., Coussement, K., De Caigny, A.: Investigating the beneficial impact of segmentation-based modelling for credit scoring. Decis. Support Syst. 179, 114170 (2024) 3.  Bakoben, M., Bellotti, T., Adams, N.: Identification of credit risk based on cluster analysis of account behaviours. J. Oper. Res. Soc. 71(5), 775–783 (2020) 4.  Reserve Bank of India: Financial Stability Report, June 2025. RBI, Mumbai (2025) 5.  TransUnion CIBIL: Quarterly Insights Report, Q4 2024. TransUnion CIBIL, Mumbai (2024) 6.    Government of India: Economic Survey 2025–26. Ministry of Finance, New Delhi (2026) 7.  MicroSave Consulting: From UPI to ULI: India’s Next Digital Infrastructure Imperative. MSC Report (2025) 8.  EY India: How Recent Changes to UPI Are Helping Fill India’s Credit Gap. Ernst and Young LLP (2025) 9.    Dastile, X., Celik, T., Potsane, M.: Statistical and machine learning models in credit scoring: A systematic literature survey. Appl. Soft Comput. 91, 106263 (2020) 10.   Shi, S., Tse, R., Luo, W., D’Addona, S., Pau, G.: Machine learning-driven credit risk: A systemic review. Neural Comput. Appl. 34(17), 14327–14339 (2022)