Management of the Cutoff for Granting Consumer Credit

Abstract

Credit issued directly to consumers is processed in volume through credit-scoring computer models that analyze a large number of potential clients in order to eliminate those that have poor payment histories. This avoids financial losses that arise from payment default. However, by refusing consumers who might default based on purely financial criteria, these models also refuse a large number of consumers with potentially good payment behavior, which can also reduce retailers' potential gains. The main objective of this paper is to verify whether corporate earnings can improve by considering the operating margin when defining the cutoff for direct lending to consumers. We used a statistical measure known as the Receiver Operating Characteristic (ROC) curve and the approach offered by Stein (2005) to build simulations around real market values, which allowed us to confirm that by being flexible and taking into account the operating margin, companies may improve net earnings.

Key words:
consumer credit; credit scoring; cutoff; credit analysis; error type I and type II