Credit Scoring And Its Applications By L C Thomas Hot Online

Credit Scoring and Its Applications

by L. C. Thomas Hot

4. Reject Inference

How do you score rejected applicants? Thomas formalized reject inference—methods to infer how rejected applicants would have performed if accepted. This is critical for building unbiased models. credit scoring and its applications by l c thomas hot

: The authors address real-world issues including scorecard monitoring, when to update models, and the impact of legislation like equal opportunity and privacy laws Blackwell's Broad Applications Credit Scoring and Its Applications by L

1.3 The Statistical Toolkit

L.C. Thomas is known for rigorously comparing and refining statistical methods. The key techniques he discusses include: Classification (Risk Scoring): The traditional model

Step 1: Separate Application, Behavioral, and Collection Scores

Do not use one model for everything. Thomas demonstrated that dynamic programming over the customer lifecycle increases risk-adjusted return by 15–25%.

Beyond traditional bank loans, the book discusses how these scoring models are applied across diverse fields:

  1. Classification (Risk Scoring): The traditional model. Separating applicants into ‘good’ (will repay) and ‘bad’ (will default). Thomas refined this by introducing survival analysis—acknowledging that when a borrower defaults matters as much as if they default.
  2. Reject Inference: This is Thomas’s most cited “hot” problem. Most models are trained only on accepted applicants. But what about the rejected ones? We never observe their performance. Thomas provided the mathematical rigor for inferring the risk of the rejected population, preventing “sample bias” that leads to overly optimistic models.
  3. Profit Scoring (Behavioral Scoring): Thomas pushed the industry beyond risk mitigation toward profit optimization. A low-risk customer who never uses interest (transactor) is less profitable than a medium-risk customer who revolvs a balance. His work on Markov decision processes allowed lenders to score not just risk, but expected monetary value.