Credit Scoring: Is it Right for Your Bank?

The paper aims:

• To demonstrate how banks can decide whether credit scoring is right for them;
• To provide a "road map" of the steps involved in designing, implementing and monitoring a credit scoring model.

The paper argues that while credit scoring can:

• Predict the likelihood or probability of a "bad" outcome as defined by the bank;
• Focus underwriting time on borderline cases, while automatically identifying very good and very bad applicants and reducing time spent reviewing them;
• Increase the profitability of small business lending by reducing time spent on collections and workout.

But cannot:

• Predict individual loan loss;
• Approve or reject a loan application;
• Increase approval rates.

The paper presents a six-step "road map" to designing, implementing and monitoring a custom credit scoring model, including:

• Presenting the concept;
• Understanding what kind of system would work in the organisation;
• Putting together a "steering committee" to discuss strategic and technical issues;
• Designing and testing the model;
• Presenting the model and providing introductory training;
• Monitoring the models and providing follow-up training.