New Credit Score Models Hold Promise for Credit Access

The need to expand access to credit is a point of broad consensus among regulators, policymakers, President Obama’s administration, academics, consumers and lenders. Over the past several years, regulators have made great strides in crafting responsible underwriting guidelines, developing servicing standards and eliminating certain products and practices to ensure long-term success for homeowners. Regulators have also been working to create clarity and certainty for lenders, although this is still a work in progress.

And yet as we work to get housing back on track, a number of reliable data sources, including the Mortgage Bankers Association and the Urban Institute, confirm that credit is too tight. This has sparked a debate as to how best to increase access to credit for qualified borrowers.

While current regulations may still be in need of some tweaks, the next opportunity is not to loosen underwriting guidelines or reintroduce risky products. It is to begin serving more qualified borrowers who are slipping through the cracks of the current credit-scoring models.

Today, lenders use credit-scoring models sanctioned by the Federal Housing Administration, Fannie Mae and Freddie Mac. But in the past ten years, technology has matured, increasingly granular data has become available, and consumer behaviors have changed. Using updated and more sophisticated models would help lenders reach significantly more consumers, many of whom would qualify for mortgages even in today’s environment of tight credit standards.

According to both Fair Isaac and VantageScore, the two largest developers of credit-scoring models, there are tens of millions of people for whom the current models cannot generate a score because they don’t meet the current key credit-scoring criteria, which is to have at least one credit account active for a minimum of six months or to have had credit activity within the past six months.

The newest model from VantageScore uses a special segmentation technique that recognizes the value of data up to and greater than 24 months old. This allows VantageScore to generate scores for 30 to 35 million of these consumers, 7.6 million of whom receive scores above 620. That is 7.6 million potential borrowers who would pass the first test of eligibility just by using an updated scoring model.

Older credit scoring models not only fail to capture all qualified borrowers, they may also eliminate them inappropriately. This is because older models do not reflect current consumer behavior. A good example is the growth in student loans. Newer models are able to extrapolate more predictive insight from those accounts than in the past. In addition, the newer models do not penalize consumers for paid debt collections because the attributes of today’s more granular data provide greater predictive strength.

Newer, more precise models may also help lower the risk premium on conventional loans because a borrower’s credit score is used to determine the premium. This would potentially lower the interest rate on many mortgages, helping consumers to better afford monthly payments.

Moreover, using newer, more accurate credit models would help to increase the number of eligible borrowers within the U.S. without weakening today’s underwriting parameters by changing debt-to-income, loan-to-value, or minimum score cut-offs.

It is encouraging to see that this issue has been gaining traction in Washington. The Federal Housing Finance Agency has asked the GSEs to consider modernizing their credit scoring requirements and has included this initiative in its 2015 scorecard. In addition, Secretary of Housing and Urban Development Julián Castro has called for expanding mortgage credit to borrowers with non-traditional credit files and promised to review ways to incorporate alternative credit-scoring models into FHA’s automated underwriting system.

This is a major undertaking, and there is certainly much work to be done. First the GSEs and FHA must validate the newer models. While certain consumers may be underscored today, we must make certain they are not over-scored by any new model.

Any change will come with implementation and technology costs, but it will also bring opportunity. Recent research from VantageScore shows that the GSEs stand to earn up to $500 million more each year by serving a larger pool of creditworthy borrowers. Lenders and others in the housing industry also stand to make additional revenue. Serving more creditworthy consumers is a true win-win for the entire market.

Updating credit-scoring models is a viable way to responsibly broaden access to credit. It is yet another actionable step towards sustaining progress in the ongoing housing recovery.

Debra W. Still is president and chief executive of Pulte Financial Services.