Generic credit scores on the market are a widely used tool by lenders to mitigate risk by knowing which applicants have a high likelihood of being delinquent on their payments. Each generic risk score on the market has been developed to predict a nonspecific payment performance behavior. Auto lenders should consider if the generic score works on their specific portfolio type with their own unique characteristics that are indicative of their own payment behaviors when choosing a generic scorecard.
CRIF Achieve routinely performs bi-annual model validations on generic industry scores on behalf of auto lenders, and we see that the performance of generic scores vary widely between different lenders. Lender A may see a generic score provide a KS in the low 20s while Lender B may use the same generic score and observe a KS in the high 40s. For these examples, we use KS as the measurement of the scorecard predictive power, higher KS measurements indicate better scores for predicting and delineating riskier population segments from those that are less risky.
Generic scores are not developed to provide optimal decisions for any individual lender but to work adequately across many lenders and under many conditions. Further, generic scores are generally not built on recent economic conditions but on very old economic cycles which may not be not be aligned with how consumers are behaving in the current environment. The previously mentioned KS results in the high 40s and low 20s only serve to illustrate how a score’s effectiveness in predicting risk is heavily dependent on the population on which it is used.
Custom scorecards help lenders make more efficient decisions by providing more granularity to target applications with a lower risk profile for that specific institution. With custom risk scores, lenders can enter riskier segments but still decline the segments of the applicant population that lead to delinquency and have a high likelihood of leading to loss. Here are two scenarios where we expect the generic auto risk score to return a lower performing KS and would recommend a custom risk scorecard for decisioning:
- An auto lender with delinquency trends that deviate from the norm will benefit from the use of a custom scorecard. For example, very prime consumers will have transactional characteristics that may be very similar to each other but very different from an average shopper. A custom scorecard can be designed to integrate precise consumer profiles or loan features which better describe the financial capacity of the consumer.
- Lenders often use multiple decision matrices in addition to a generic auto risk score to drill down on less risky population segments from within the low- to mid-tier score tiers. These matrices generally blend application information with additional credit information to improve the accuracy of the decision process. Usage of these additional decision matrices can be an inefficient addition to the decision process if not used properly. A custom scorecard can use a hybrid of credit and application information to eliminate the need additional decision matrices thus adding efficiency by providing less manual review, more optimization in risk decisioning and less chance of an underwriter’s bias in the overall lending process.
CRIF Achieve can help turn acquisition and performance data into business intelligence for banks, credit unions, and finance companies. Built with proven technology and expertise, CRIF Achieve’s analytic decision management infrastructure enables us to design affordable, customized solutions of our clients that are:
- Multi-dimensional to manage every category of risk financial institutions face today
- Scalable to provide the analytics needed today and to lay the foundation for the analytics of tomorrow
- Expandable across the credit cycle and lending functions
- Agile for rapid return on decision making investments
Our success stories speak for themselves. For a detailed example of how our CRIF Achieve team of experts helped East Idaho Credit Union, please click the button below to request our case study.
Erika Ward is a manager on the Achieve Analytics team and is a product manager for the RapUP credit bureau data aggregation solution.
Photo Credit: Steve Snodgrass