FICO® Score 10T: The Most Predictive Credit Score for Mortgage Lending — Now Confirmed by Independent Research
What the Milliman analysis means for lenders navigating credit score modernization
The most critical question in mortgage lending today is also the most consequential: which credit score is most predictive — and therefore best positioned to serve lenders, investors, servicers, and the borrowers who depend on their decisions? The FHFA's approval of two new credit scores carries the promise of bringing more qualified buyers into the market and creating a more inclusive mortgage lending environment. Realizing that promise depends on choosing the right score — not just the first one available.
Global actuarial firm Milliman released the first independent, head-to-head analysis of its kind, comparing the predictive performance of FICO® Score 10T to VantageScore 4.0 across several major loan types.
The result is clear: FICO® Score 10T outperformed VantageScore 4.0 across every mortgage category tested, including GSE and FHA loans. For lenders, investors, servicers, and the borrowers at the center of every decision, the choice of credit score is not a procedural matter — it is a risk management decision with real implications for portfolio performance, mortgage pricing accuracy, and the long-term health of the housing finance system. The facts are now in.
The research: what Milliman found
To assess which score better separates mortgage borrowers who default from those who do not, Milliman analyzed four established statistical measures: the Kolmogorov-Smirnov (K-S) statistic, bottom decile lift, area under the curve (AUC), and Gini coefficient. The choice of measures is significant in its own right — the K-S, bottom decile lift, and Gini coefficient statistics are the criteria specified in the Joint Enterprise Credit Score Solicitation as the measures for accuracy and reliability in evaluating credit score models established by Freddie Mac and Fannie Mae.
That means these findings are evaluated using the same established criteria set forth by the GSEs — not just analytically sound but directly correlated to the ongoing industry discussions. Milliman applied these measures across nearly 20 million mortgage tradelines spanning six historical snapshots from 2011 to 2023, with performance windows capturing actual borrower outcomes over 18 to 24 months.
Summary of key findings:
FICO® Score 10T outperformed across all predictive measures tested. Milliman found that FICO Score 10T consistently predicted mortgage defaults more accurately than VantageScore 4.0 across all mortgage types, as well as individually, for GSE and FHA loans.
The performance advantage was strongest in FHA lending. FICO® Score 10T outperformed VantageScore 4.0 by more than 8% in FHA lending, the segment that serves first-time homebuyers and underserved borrowers. In fact, a companion Milliman analysis focused specifically on first-time homebuyers confirms this performance advantage widens for first time homebuyer mortgages, supporting sustainable access to homeownership.
FICO Score 10T’s performance advantage over VantageScore 4.0 has tripled since 2018. In the 2023 GSE loan vintage, FICO Score 10T outperformed by 8.1% - triple the already impressive 2.7% improvement on the 2018 vintage.
Milliman conducted this analysis as an independent consulting engagement. Their methodology and conclusions are their own.
Why the FICO® Score 10T predictive advantage matters for lenders
Better default prediction compounds across every origination decision. For lenders managing large volumes, a consistent improvement in predictive accuracy doesn't register in a single loan — it compounds across thousands of origination decisions, in approvals extended with confidence and in loans not made to borrowers who cannot sustain them. The bottom decile lift improvement demonstrates that FICO® Score 10T identifies a greater proportion of the highest-risk borrowers at or below the score threshold where key industry cutoffs are oftentraditionally set.
Accurate assessment of risk is critical for FHA loans — for people, not just portfolios. FHA loans serve first-time homebuyers and borrowers from underserved communities. First-time homebuyers carry meaningfully higher default risk — Milliman's analysis found they defaulted at a rate 2.25 times higher than all mortgage borrowers across all archive periods studied, and nearly 60% of could be priced differently depending on which credit score was utilized. A score that identifies that risk more accurately in precisely this population supports expanded access and responsible lending at the same time. Those are not competing objectives with FICO® Score 10T.
The performance gap is growing — and that trajectory is the most important signal in the data. The 2023 vintage, which recorded the highest default rates in the dataset, shows the largest gap in performance between FICO® Score 10T and VantageScore 4.0 of any cohort studied. A score that widens its lead precisely when conditions are hardest is demonstrating something about how it was built: the time-tested FICO Score blueprint used for decades helps to ensure that our scores hold up through economic cycles.
The data sources are the same; the results are not. Both FICO® Score 10T and VantageScore 4.0 draw from the same credit bureau sources and incorporate trended credit data. The performance gap is not explained by data access. It reflects differences in design philosophy and model architecture — the same differences FICO's own prior analysis identified, and which the Milliman data now confirms on a fully independent basis.
Why this independent analysis holds up from any angle
Independent research stands on its methodology. Milliman's approach is straightforward and traceable— the more closely you examine how this analysis was structured, the more robust the findings appear to be.
Scale, consistency, and time all point in the same direction. Milliman evaluated nearly 20 million mortgages across six archive periods spanning 2011 to 2023, covering all major loan types — GSE, FHA, VA, and non-agency. Four independent statistical measures were applied, and FICO® Score 10T led on every one of them. Consistency across measures, segments, and time periods is the definition of a robust result.
The measures used are the same ones selected by regulators for credit score modernization. When FICO® Score 10T's advantage shows up on metrics specified by the regulatory evaluation process, the findings don't just hold up analytically — they speak directly to the standard the industry has already agreed matters.
And that result is not surprising, given that documents obtained through the Freedom of Information Act revealed that FICO® Score 10T was the sole credit score model recommended by both Fannie Mae and Freddie Mac following their own lengthy, competitive, and comprehensive evaluation process.
Lender choice: a structural risk the market has not fully priced
Milliman's analysis surfaces an important finding that goes beyond score comparison at the aggregate performance level. Because the two scores could assign nearly half of all borrowers to different LLPA pricing bands, a loan-by-loan lender choice policy introduces systematic adverse selection risk.
Prior research from Milliman demonstrated that this dynamic —without offsetting controls —will bias default rates, creating greater uncertainty for loan servicers and MBS investors, and ultimately placing upward pressure on mortgage rates for borrowers. The FHFA's current policy, reaffirmed in its April 22, 2026, announcement, allows lenders to choose one score on an individual loan-by-loan basis for delivery to Freddie Mac and Fannie Mae. The questions of how an LLPA matrix will be calibrated to avoid penalizing or favoring a given score, and how the bias of lender choice will be accounted for in pricing, remain outstanding.
The more predictive score is not just the better analytical choice. It is the structural foundation that keeps the system honest and stable for lenders, for investors, and for the borrowers whose rates ultimately reflect these decisions.
The FICO® Score advantage
The Milliman analysis is the latest in a converging body of evidence pointing in the same direction. The GSEs' internal evaluation, FICO’s own prior analysis, and now this independent head-to-head study all reach the same conclusion: FICO® Score 10T delivers the strongest predictive performance available for mortgage lending today.
That performance is built on something durable. FICO® Score 10T incorporates trended credit data and rental payment history — and does so without using the mortgage-specific variables that VantageScore 4.0 includes — variables that penalize first-time homebuyers for not having already owned a home.
The pricing implications are sharpest for this population: Milliman's analysis of first-time homebuyer mortgages found that only 40.4% could fall in the same LLPA pricing band under both scores —meaning nearly 60% of first-time homebuyers could be priced differently depending on which score was used. For a population where affordability is already the primary barrier to homeownership, that degree of pricing divergence is not a technical footnote. It is a direct consequence of which score a lender chooses. FICO® Score 10T was deliberately built without those mortgage-specific variables — and it still outperforms. That is the point.
FICO's independence from the credit bureaus is a structural advantage that matters. FICO's sole interest is the performance of the score. For every stakeholder who depends on an objective, reliable measure of risk, that independence is not a marketing point. It is a structural safeguard that has underpinned the integrity of the FICO® Score for over 30 years.
The trajectory of the Milliman findings says the rest. FICO® Score 10T's lead is widening, not narrowing, and a score that proves more powerful when economic conditions are deteriorating is one that the mortgage market can depend on for the decades ahead.
What to do next
The Milliman analysis provides a firm factual foundation for the credit score decision. Read the full score performance white paper atmilliman.com/en/insight/fico-score-10t-vantagescore-4-analysis-credit-mortgage and the companion analysis focused specifically on first-time homebuyers at https://www.milliman.com/en/insight/fico-score-10t-vantagescore-4-analysis-first-time-homebuyers.
When the GSEs publish the approximately 50 million historical FICO® Score 10T scores associated with their Single Family Loan-Level datasets— expected in Summer 2026 — we encourage lenders and industry stakeholders to conduct their own analysis using the same standard metrics applied by Milliman and specified in the Joint Enterprise Credit Score Solicitation. The Milliman findings, GSE internal evaluations, and FICO prior analysis all point in the same direction. We are confident this historical GSE conforming loan data will as well.
For lenders, servicers, and investors ready to act on the opportunity these analyses confirm, three paths are available today:
The FICO® Score 10T Free Access Program provides FICO® Score 10T at no cost alongside Classic FICO, enabling side-by-side performance testing without requiring lenders to pay for an additional score.
The FICO® Archive Score Program offers mortgage lenders up to two million free archive FICO® Scores, calculated using depersonalized historical consumer credit data for retrospective testing and validation across existing portfolios. Archive FICO Scores are available through resellers and the credit bureaus.
Lenders interested in moving to FICO® Score 10T only can purchase FICO Score 10T through the FICO® Mortgage Direct License Program with a simplified, transparent pricing model of $0.99 per borrower per score with a $65 funding fee — providing a lower-cost path to production deployment with full pricing transparency.
To learn more or discuss how FICO® Score 10T can strengthen your risk management framework, contact FICO’s Mortgage and Capital Markets team directly.
The performance findings speak for themselves. The path to acting on them is open today.
The Milliman analysis was performed independently by Milliman as part of a consulting engagement with FICO. The analysis reflects Milliman's own methodology and conclusions. Read the full white paper at milliman.com.
Frequently Asked Questions
Both scores incorporate credit data from the same bureau sources, and both use trended credit data — evaluating borrower behavior over a 24-month window rather than a single point in time. The performance gap is not explained by data access. It reflects differences in model design and architecture: what variables are included, how each score weighs and interprets borrower behavior, and the design philosophy applied during model development.
Milliman is an independent actuarial and consulting firm founded in 1947 with no stake in the outcome of this analysis. The dataset was sourced directly from a major U.S. credit bureau, entirely outside of FICO's control, and Milliman's methodology and conclusions are their own.
That separation of data, methodology, and conclusions is what makes the findings credible to lenders, investors, and regulators who need to trust the analysis, not just the party commissioning it. It also stands in notable contrast to competing analyses where the organization being evaluated controls both the data selection and the analytical framework — and where non-standard, self-selected evaluation metrics are seemingly applied until a favorable result is found.
In their reports, Milliman applied four widely used, industry-standard measures — the K-S statistic, bottom decile lift, AUC, and Gini coefficient — the same measures the FHFA and GSEs themselves specified in the Joint Enterprise Credit Score Solicitation for evaluating credit score model applicants. The result was consistent across all measures, all mortgage types, and all time periods studied.
FHA loans serve first-time homebuyers and borrowers from underserved communities — the segment where credit risk sensitivity is highest and where an inaccurate assessment carries the most acute consequences. FICO® Score 10T outperformed VantageScore 4.0 by more than 8% in FHA lending on the bottom decile lift measure — the largest segment-level gap in the analysis. This matters for two reasons.
First, the bottom decile lift measures how effectively a score concentrates the highest-risk borrowers at the bottom of the distribution, exactly where lenders set cutoffs and make approval decisions. A more than 8% improvement in that measure means FICO® Score 10T identifies a meaningfully greater proportion of the borrowers most likely to default in a population where defaults carry the most severe consequences for borrowers, lenders, and the broader market.
Second, the size of the FHA gap matters in the context of the design choices each score makes. FICO® Score 10T achieves this advantage without using mortgage-specific variables that would disadvantage first-time homebuyers, delivering its greatest performance improvement precisely in the population its design was built to serve fairly.
The widening gap (bottom decile lift from 2.7% in the 2018 GSE loan vintage to 8.1% in 2023) is one of the most consequential findings in the Milliman analysis for lenders thinking about adoption timing. It suggests that the conditions under which the two scores diverge most are precisely the conditions that matter most: recent originations and elevated default environments. The 2023 vintage, which recorded the highest default rates in the dataset, produced the largest performance gap of any cohort studied.
Yes. Under the FHFA’s Validation and Approval of Credit Score Models final rule, both Fannie Mae and Freddie Mac conducted extensive testing and review of FICO® Score 10T and VantageScore 4.0 for accuracy and reliability. FICO Score 10T was the sole credit score model recommended for use by both GSEs following that evaluation process. Documents obtained through Freedom of Information Act (FOIA) requests revealed that experts within the GSEs recommended FICO Score 10T and recommended against VantageScore 4.0 for GSE use.
Both — and the distinction matters. FICO® Score 10T is a more precise score, not simply a tighter filter. Because it more accurately identifies which borrowers are genuinely at risk of default, it also more accurately identifies which borrowers are not — including some who would have been declined or priced conservatively under a less precise score.
FICO® Score 10T can expand mortgage approval rates by up to 5% in mortgage lending without adding incremental risk relative to the current score in use. It can also reduce mortgage delinquency rates by up to 17% for new originations. These are not competing outcomes. A score that is better at identifying high-risk borrowers is, by definition, also better at identifying low-risk borrowers — and that precision translates simultaneously into fewer defaults and more confident approvals. For lenders, that means both stronger portfolio performance and a larger addressable borrower population served responsibly. For investors, that means stability and predictability, which translates into more reliable cash flows.
VantageScore is jointly owned by the three major U.S. credit bureaus — Equifax, Experian, and TransUnion. The score, the data, and the distribution channel are all controlled by the same organizations.
FICO is independent from the credit bureaus. Its sole interest is the performance of the score.
That independence is a structural safeguard, not a marketing point. It is what has allowed the FICO® Score to earn and maintain the trust of lenders, investors, and market stakeholders for over 30 years — through multiple economic cycles, regulatory reviews, and significant shifts in the credit landscape.
Popular Posts
Has the Reporting of Rental Data to the Credit Reporting Agencies (CRAs) Increased?
FICO Score 10T includes rental data, but consumers can only experience the benefit of this to the extent that their rental data is reported to the CRAs
Read more
Average U.S. FICO® Score at 716, Indicating Improvement in Consumer Credit Behaviors Despite Pandemic
The FICO Score is a broad-based, independent standard measure of credit risk
Read more
FICO Statement on FHFA and FHA Updates to Credit Score Modernization
FICO supports FHFA’s announcement that the long-anticipated historical data for FICO® Score 10T will be released to the mortgage market.
Read moreTake the next step
Connect with FICO for answers to all your product and solution questions. Interested in becoming a business partner? Contact us to learn more. We look forward to hearing from you.