Whether it is an online marketplace lender offering to refinance the student loan debt of a recent college graduate or an online small-business lender providing an entrepreneur with a loan when no one else will, there is no doubt innovation in the online lending sector is changing how Americans gain access to credit.
This expanding market segment takes great pride in using “next-generation” underwriting and credit scoring risk models. In particular, many online lenders are incorporating noncredit information such as income, education history (i.e., type of degree and college), professional licenses and consumer-supplied information in an effort to strike the right balance between properly assessing credit risk and serving consumers typically shunned by traditional lenders because of a thin credit history.
Regulatory concerns
The exponential growth of the online lending sector has caught the attention of regulators — such as the U.S. Treasury Department, the Federal Deposit Insurance Corporation, Congress and the California Business Development Office — who are interested in learning more about how online marketplace lenders are assessing the credit risk of consumers and small businesses.
At least one official, Antonio Weiss, a counselor to the Treasury secretary, has publicly raised concerns about the use of so-called nontraditional data in the underwriting process, particularly data gleaned from social media accounts. Weiss said that “just because a credit decision is made by an algorithm, doesn’t mean it is fair,” citing the need for lenders to be aware of compliance with fair lending obligations when integrating nontraditional credit data.
Innovative and “tried and true” are not mutually exclusive
Some have suggested the only way to assuage regulatory concerns and control risk is by using tried-and-true legacy credit risk models. The fact is, however, online marketplace lenders can — and should — continue to push the envelope on innovative underwriting and business models, so long as these models properly gauge credit risk and ensure compliance with fair lending rules. It’s not a simple either-or scenario.
Lenders always must ensure their scoring analytics are based upon predictive and accurate data. That’s why lenders historically have relied on credit history, which is based upon data consumers can dispute using their rights under the Fair Credit Reporting Act. Statistically sound and validated scores protect consumers from discrimination and lenders from disparate impact claims under the Equal Credit Opportunity Act. The Office of the Comptroller of the Currency guidance on model risk management is an example of regulators’ focus on holding responsible the entities they oversee for the validation, testing and accuracy of their models.
Marketplace lenders who want to push the limit can look to credit scoring models now being used in the marketplace without negatively impacting credit quality or raising fair lending risk. For example, VantageScore® 3.0 allows for the scoring of 30 million to 35 million more people who currently are unscoreable under legacy credit score models. VantageScore does this by using a broader, deeper set of credit file data and more advanced modeling techniques. This allows the VantageScore model to capture unique consumer behaviors more accurately.
In conclusion, online marketplace lenders should continue innovating with their own “secret sauce” and custom decisioning systems that may include a mix of noncredit factors. But they also can stay ahead of the curve by relying on innovative “tried-and-true” credit score models, like VantageScore 3.0. These models incorporate the best of both worlds by leaning on innovative scoring analytics that are more inclusive, while providing marketplace lenders with assurances the decisioning is both statistically sound and compliant with fair lending laws.
VantageScore® is a registered trademark of VantageScore Solutions, LLC.
