FICO's Julie May on how score use in mortgages is changing

March 18, 2025 6:00 AM

The push for government-related mortgage investors to adopt advanced credit metrics — originally slated for 2025 — has been put on hold indefinitely. As a result, the spotlight has shifted to the private market, where lenders are exploring new scoring models like FICO 10T and VantageScore 4.0 to expand originations while managing risk.

It's a change in the landscape that the government-sponsored enterprises, mortgage companies and providers of the modernized scores have to navigate as the current administration in Washington looks to potentially privatize the GSEs.

To better understand how private lenders are adopting these metrics, we spoke with Julie May, vice president of business-to-business scores at FICO. She shared insights on how FICO is positioning its 10T score for the nonconventional market as credit costs rise.

Below are excerpts of her remarks on the subject, edited for clarity and length.

Score modernization’s status with the GSEs

May: They've changed the date to TBD. It's just an update of the project deadline from what we understand.

We are still in contact with both of the GSEs, and working through many different processes to continue to try to make 10T data file available to the stakeholders to support the transition.

The classic FICO score has been used for a long time prior to the GSEs going into conservatorship, so irrespective of modernization we expect to remain a significant part of that ecosystem.

We were already the industry standard for lending decisions for mortgages prior to being selected by the GSEs. The GSEs really selected us because the majority of originators in the mortgage space were already using the FICO score to make lending decisions.

They used us for a long time prior to conservatorship.

The nonconforming market’s evolution

May: We are getting pretty significant adoption across multiple use cases: origination, securitization and review by investors of FICO score 10T in the nonconforming space, or loans that are not sold to the GSEs.

For example, we've had 10T used in conjunction with the government mortgage-backed securities by Cardinal Financial, which formed and traded the first government-issued mortgage backed security on loans powered by FICO10-T. Within that pool, they saw higher scores for the majority of borrowers, which enabled them to offer more favorable terms to those borrowers.

They've been adopting 10T at a very rapid rate in the nonconforming market. We have an early adopter program with over two dozen lenders who have signed up for and are actively using it. We started our early adopter program in 2023 at the very end. We're just a little over a year into that program.

Quantifying advanced score use in servicing and lending

May: It's not just originators that are using the FICO score and dual processing. When we look at our early adopter program, which allows people to look at classic FICO with 10T together, we've started to see servicers of loans use it as well.

When we look at early adopters in terms of the volumes that they cover, it's $264 billion in annualized mortgage originations and $1.43 trillion in mortgage portfolio servicing.

We see that with FICO score 10T and our simulations as it relates to the classic score that lenders could originate 5% more if they wanted to hold credit risk steady. They could see a reduction in delinquencies of up to 17% if they wanted to hold origination steady.

Securitization, technology and reseller updates

May: We're doing a lot of work with tech vendors, trading platforms and others in the ecosystem to be able to accept and process 10T so that it's easier for the lenders who want to adopt it and it's easier for investors who want to be able to understand what loans in a security looks like.

We are well integrated into resellers, making sure that people can do a process: Xactus, Factual Data, Informative Research and CoreLogic to name a few. There are more than that but those are four of the largest.

We're doing a couple of things right with what I would call the mortgage resellers, which enable lenders to pull a three-bureau report. We're integrating the ability for them to pull FICO score 10T with FICO Classic to enable dual processing for our early adopter program.

A mortgage simulator with FICO algorithm insights

May: We have a FICO score mortgage simulator. We are signing up resellers to make that available in the marketplace and the first partner that is working with us to distribute that FICO score mortgage simulator is Xactus. They were our early adopter partners.

There are other mortgage score simulators in the marketplace, but this is the only one that uses the FICO score with proprietary knowledge of the algorithm.

What that does is it allows lenders to provide guidance to consumers or potential borrowers on what is likely to happen to their FICO score if they were to take specific action. They can run different credit events scenarios looking at the applicant's data.

So for instance, if you were to pay off your car loan, what would likely happen to your FICO score? That can be very beneficial for guiding consumers either improving the terms they would get or enabling them to reach a point of qualification.

‘A myth I like to bust’

May: There's oftentimes a question about needing to have debt to have a good FICO score. It's a myth I like to bust.

We recently published information on this where we analyzed borrowers with debt and borrowers with no debt. We see that on average, borrowers with no debt score higher than the national average.

So no-debt borrowers average 737 versus a national average between 716 and 718. Then we also see that those borrowers that have no amounts owed, 80% of them have scores higher than 700.

I just point that out because it's one of the things that we do know is a big myth out there. People frequently believe that to have a high FICO score, you must have debt, or you must have large amounts of debt, and that's simply not true.

The FICO score is a lagging indicator that very consistently rank orders risk over time. What we are seeing is that one quarter of consumers are scoring 800 or better, and that that percentage was to 22% in January of 2020. More than 4 million US consumers are now scoring above 800.

Since 2021, the average score has moved between 716 and 717 in the U.S. We expect that could go down another point because of the return to student loan delinquency reporting.

Buy now, pay later’s potential

May: We looked at a half million consumers that at least had at least one Affirm BNPL loan and then we looked at that against a benchmark population of consumers which did not have Affirm BNPL loans. Then we simulated what those loans would do and how they could potentially affect the FICO score.

In those simulations, we saw that there would be higher scores or no score changes for the majority of the population for the consumers who had five or more Affirm BNPL loans. Customers with multiple BNPL loans were more likely to experience score increases.