As one of the key requirements for meeting Regulation B's definition of a credit scoring system, validations are an essential part of a sound lending, pricing and score card monitoring program. More credit unions are being advised that score card validations should be performed regularly by an unbiased third party. Validations always have been a requirement of Regulation B. However, due to recent economic events there is a more concerted effort by regulatory agencies to ensure validations are performed on a regular basis.
This article discusses why validations are required, which regulations they can satisfy and what else they can do for your credit union.
The main purpose of a validation is to determine how well the score differentiates between good and bad accounts. The more predictive the score, the more confidence you can have in using the score to make approval and pricing decisions. Regular validations provide an ongoing assessment of the model's predictive power and will assist your credit union in determining if the time is right to change scores or to build or redevelop a custom model.
A predictive score card ensures compliance by reducing judgmental decisions and lender discretion and maintaining lender confidence that the score card accurately distinguishes between good and bad accounts.
Recently, the predictive ability of many score cards has been questioned. There is a perception that scores were not predictive in the time leading up to the current economic downturn. Our analysis shows that most score cards did maintain their ability to separate good and bad accounts. The increases in delinquency and loss rates resulted from a change in lending policies and strategies, not a failure of the score's predictive ability.
Many credit unions decided to approve higher-risk applicants, who were more sensitive to economic changes. In some cases, validation results show that these changes in strategies resulted in up to a tenfold increase in the loss rates at certain score intervals compared with historical rates. By performing regular validations, a credit union could have detected the changes in bad rates, loss rates and distributions. Cutoff scores and lending policies could have been adjusted to meet the organization's desired risk tolerance.
One key element of a successful lending program is score card adherence. Minimizing overrides eliminates unnecessary risk, ensures consistency and assists in the equal treatment of applicants. In turn, these results allow you to demonstrate compliance with the Equal Credit Opportunity Act. Adherence, or lack thereof, to the score card's recommendation is one of the leading factors in a credit union not achieving goals for volume and risk tolerance.
A validation measures overall score card adherence and overrides above and below the cutoff score. With significant overrides of the score card, the credit union is at risk of decreasing the score card's overall predictive ability. Too many high-side overrides also can result in lost income opportunities on lower-risk accounts. At the other extreme, excessive low-side overrides can create larger numbers of high-risk accounts that significantly increase delinquency and loss rates. These increases can result in the need to augment reserves unexpectedly and the credit union falling below minimum capital requirements.
Is your organization approving and pricing appropriately? This is another question a validation can precisely answer. A validation provides approval, booking, bad and loss rates by score intervals. These statistics allow you to establish cutoff scores and tier breaks to maximize approvals and bookings while setting rates that are appropriate for the risk level. This information also can be used to measure future populations to determine the need for adjustments to cutoff scores or tier breaks. Without adequate monitoring and validations, changes within a portfolio can result in higher risk loans and evaporation of expected profitability.
Regular validations not only will ensure regulatory compliance but also will give insight into your portfolio. As part of a robust score card monitoring program, validations provide the information needed to proactively manage your lending program.
Andrew Bieno is a strategic consultant with Experian Decision Analytics. He can be reached at 678-731-1179 or firstname.lastname@example.org