Machine Learning-Based Credit Scoring Models for Retail Lending at YES Bank

Abstract

This study, titled "Machine Learning-Based Credit Scoring Models for Retail Lending at YES Bank," evaluates retail portfolio composition, risk discrimination Gini scores, retail credit volume expansion, and financial feasibility of advanced machine learning (ML) credit scoring engines in private sector banking. Commercial banks face increasing competition in retail lending, where personal loans represent 40% and auto loans account for 28% of retail credit dispatches. A five-year project lifecycle (2021-2025) of an ML-based retail credit scoring platform at YES Bank is evaluated using capital budgeting parameters: Net Present Value (NPV), Internal Rate of Return (IRR), Payback Period (PBP), and Benefit-Cost Ratio (BCR). Quantitative analysis indicates that deep neural network and gradient boosting models achieve a Gini coefficient of 0.93 compared to 0.54 under traditional logistic regression. Scaling ML scoring deployments drives retail credit book volume to 138,000 Crores while strengthening the Provision Coverage Ratio (PCR) to 92.8%, expanding ML deployment to 94.2% and compressing retail default rates to 0.7% by 2025. The financial model yields a positive NPV of 284.5 Crores and an IRR of 38.6%, far exceeding the 10% discount hurdle rate. The study concludes that investing in ML credit scoring engines is highly viable, boosting loan portfolio quality and retail banking profitability.