On April 22, the Federal Housing Finance Agency and the Department of Housing and Urban Development jointly announced the next phase of mortgage credit score modernization. Fannie Mae and Freddie Mac will update their selling policies to permit the current use of VantageScore 4.0 and the future use of FICO Score 10T. HUD simultaneously announced adoption of both models for Federal Housing Administration loans. The announcements, welcomed broadly by industry trade groups, mark the most significant structural change to GSE underwriting eligibility criteria in decades. They also introduce a category of model-transition risk that most lenders’ QC programs have not yet been designed to address.
The Pricing Backdrop
The transition is unfolding against a contentious cost environment. FICO doubled its per-score mortgage pricing to $10 in early 2026, a 100% increase that followed a 16-fold rise in wholesale pricing over the preceding five years. The Community Home Lenders of America has documented a 1,567% increase in FICO-related costs over roughly three and a half years. Senator Josh Hawley sent a letter to the Federal Trade Commission in March calling the pricing structure anticompetitive and requesting an investigation.
In response to that pressure, TransUnion cut VantageScore 4.0 pricing to $0.99 per pull and is offering it bundled at no cost alongside FICO through 2026. Experian has moved similarly. The cost differential has accelerated lender interest in VantageScore adoption at precisely the moment the GSEs are opening the door to it. Credit report costs to lenders have reportedly climbed by as much as 50% industry-wide in 2026, making the case for transition more urgent.
Where the Underwriting Risk Concentrates
The FHFA’s pilot framework creates a lender-choice regime: originators may deliver either Classic FICO or VantageScore 4.0 for loans sold to the GSEs. From a loan quality standpoint, that optionality is not neutral. FOIA disclosures have revealed that both Fannie Mae and Freddie Mac raised internal reservations about VantageScore approval as far back as 2022, with both GSEs noting that single-file underwriting carried greater model risk than the tri-merge standard. Milliman published analysis in February 2026 concluding that lender choice introduces a predictable bias to default rates as lenders systematically select whichever model produces the more favorable borrower outcome.
The downstream effect is an adverse selection dynamic embedded in GSE MBS collateral pools — a concern that fixed-income investors are beginning to price. For individual lenders, the operational risk is more immediate: loan files originated under VantageScore 4.0 are entering a pipeline in which GSE loan quality review processes, investor overlays, and post-purchase QC frameworks were calibrated to Classic FICO-scored collateral. The eligibility determinations, compensating factor thresholds, and manual underwriting referral triggers used in those processes have not yet been recalibrated.
QC and Loan Salability Exposure
The transition period — during which lenders, investors, and GSE systems are operating under mixed-model conditions — creates heightened defect risk in borrower data documentation and score-to-decision traceability. A post-closing QC review that cannot confirm which scoring model governed an eligibility determination, or verify that the model selected was applied consistently across comparable borrowers, represents a material representation and warranty gap.
Pre-funding QC protocols need to incorporate explicit model identification checkpoints. Loan processing and underwriting support functions must document not only the score obtained but the model version, the delivery channel, and the basis for model selection where choice exists. MERS compliance and closing package integrity reviews are also affected: investor data fields and transfer documentation will increasingly require score-model attribution as the dual-model environment matures and investors begin differentiating collateral pricing accordingly.
Looking Ahead
Fannie Mae has indicated it will publish historical credit score data for FICO 10T and VantageScore 4.0 this summer, providing the loan performance benchmarks lenders and investors need to calibrate overlays. Until that data is available, the industry is operating on incomplete information about relative model performance across risk tiers. Lenders moving quickly to capture VantageScore’s cost advantage before pricing grids are published and performance data is available are accepting underwriting risk that their current QC infrastructure may not be equipped to detect.
