Personal Credit Improves Small-Business Default Prediction
This study using the Gies Consumer and Small Business Credit Panel finds that integrating business owners' personal credit bureau data with business-level and tradeline features raises AUROC from about 0.78 to roughly 0.83, with gradient boosting models (XGBoost, LightGBM, CatBoost) performing best. Personal credit variables—credit score, outstanding balances, recent inquiries—rank among top predictors, suggesting alternative data strengthens small-business default models and validation.
Key Points
- 1Demonstrates AUC increase from ~0.78 to ~0.83 when adding owner personal credit data
- 2Shows gradient boosting models (XGBoost, LightGBM, CatBoost) capture nonlinear interactions effectively
- 3Encourages lenders to include personal credit attributes to improve risk models and validation
Scoring Rationale
Demonstrates actionable accuracy gains and strong relevance, limited by single-study scope and external validation needs.
Sources
Public references used for this report.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Ad Tech problems
