OneMain Financial
Machine-learning credit underwriting
OneMain reported underwriting model predictive power of 2.17x its bureau-score benchmark
Problem
OneMain needed to distinguish credit risk among borrowers using information beyond traditional bureau scores.
Approach
OneMain expanded its proprietary underwriting models with machine learning and additional data sources, including cash-flow data. It combined advanced analytics with human interaction in underwriting. Its 2023 investor presentation compared model predictive power against VantageScore v3 using the KS score.
Results
Takeaway
The comparison measured separation between credit-risk groups. It did not establish a corresponding reduction in loan losses.
Categories
- Industry
- Finance and insurance
- Business function
- Finance, risk and legal
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Goal
- Quality and risk
- Evidence
- Company's own claim
Sources
Source summaries
OneMain described its machine-learning underwriting infrastructure and use of additional borrower data. It reported 2023 model predictive power of 2.17x VantageScore v3, citing Experian and internal analysis.
Full details are in the original sources.