FICO and VantageScore release competing analyses following Fannie and Freddie data drop

FICO and VantageScore used the same newly released Fannie Mae/Freddie Mac credit score datasets to tell very different stories about how their models perform and what the GSE credit score transition means for risk management and borrower access to mortgage credit. For credit and collection professionals, the split underscores how model choice, time horizons, and stress-period performance can influence risk views at a moment when FHFA, Fannie, and Freddie are opening the door to credit score competition.

Background: GSE Credit Score Modernization

Fannie Mae and Freddie Mac are in the midst of a multi‑year credit score “modernization” that replaces the long‑used Classic FICO model with two newer scores: FICO Score 10T and VantageScore 4.0. FHFA has directed the GSEs to move to a “multi‑score” framework and also shift from a tri‑merge to a bi‑merge credit report requirement, with lenders allowed on an interim basis to deliver loans using either Classic FICO or VantageScore 4.0 while new models are phased in.fhfa+2

To support that transition, Fannie Mae and Freddie Mac have begun releasing large historical datasets so market participants can run their own model comparison and validation exercises. In 2024, the GSEs first made historical VantageScore 4.0 credit scores available for single‑family loans acquired from April 2013 through March 2023, and in July 2026 they followed with more than a decade of historical FICO 10T loan‑level data covering April 2013 through September 2025.

The New Data Drop

The latest “data drop” giving rise to dueling analyses includes GSE loan‑level performance data keyed to both legacy and next‑generation credit scores. Together with earlier releases, the new files allow users to compare how Classic FICO, FICO 10T, and VantageScore 4.0 rank‑order risk over multiple vintages, including the COVID‑19 stress period, and across different distributions of borrower characteristics. Lenders, investors, and regulators can now observe default, delinquency, and prepayment outcomes at the loan level across more than a decade, using the same datasets Fannie and Freddie used in their own model validation and approval processes.

The data are available directly from Fannie Mae and Freddie Mac to approved users and are intended to support both risk management and operational planning as the Enterprises move further into a multi‑score market. Industry trade groups, including the American Bankers Association, had previously pressed FHFA to ensure that model calibration and validation would not be limited to post‑crisis samples, arguing that broader histories are needed to meet prudential standards—a concern the expanded datasets help to address.

FICO’s Analysis: Emphasizing Limits and Caution

FICO’s public response to the GSE data release stresses what it describes as limitations and caveats in the newly available credit score information. According to coverage of its analysis, FICO argues that certain aspects of the datasets—such as the time windows and portfolio composition captured—make it difficult to draw definitive conclusions about long‑run model performance or to extrapolate results to all future market environments.

FICO’s messaging also leans heavily on its sustained role as the incumbent GSE score and its broader regulatory and supervisory acceptance across banking and capital markets. In related materials about FICO 10T, the company continues to highlight what it says is superior predictive power relative to older FICO models, including an estimated 17 percent reduction in mortgage defaults when using FICO 10T compared with legacy scores, and emphasizes its incorporation of trended data and payment histories.

VantageScore’s Analysis: Expanding Access and Predictive Power

VantageScore, by contrast, quickly used the same Fannie/Freddie data environment to claim that its VantageScore 4.0 model better predicts mortgage risk and expands access to credit. In a recent op‑ed and supporting materials, VantageScore points to analyses by large lenders and independent research finding that VantageScore 4.0 captures up to 13 percent more incremental defaults than legacy FICO models when tested on millions of mortgages over a decade, including stress periods.

A new study by independent firm Prosperity Now, based on GSE loan‑level data, concluded that VantageScore 4.0 “consistently better rank‑orders mortgage borrower risk and better separates risk than FICO,” including during the COVID‑19 environment, and that loans with a high FICO score but a low VantageScore 4.0 were in fact much riskier than FICO alone suggested. VantageScore also emphasizes its broader coverage—claiming the ability to score roughly 96 percent of U.S. adults by incorporating rent and utility payments—and argues that this helps bring more thin‑file and historically underserved borrowers into the mortgage market as Fannie and Freddie transition models.

Divergent Narratives from the Same Data

Despite relying on overlapping Fannie/Freddie datasets, FICO and VantageScore are highlighting different slices of the risk story to support their respective positions. VantageScore’s narrative centers on improved rank‑ordering and broader inclusion, arguing that the new data confirm its model identifies marginally riskier loans that older scores mis‑classify, while not materially increasing risk when its scores are higher than FICO’s.

FICO, meanwhile, points to what it characterizes as shortcomings in the available data and stresses the dangers of reading too much into a limited set of vintages or stress periods, particularly when supervisory standards still rest heavily on FICO‑based infrastructures. Independent research is mixed: some studies, such as work from the Urban Institute and other think tanks, have found Classic FICO modestly outperforms VantageScore 4.0 on certain Fannie Mae vintages, while other analyses favor VantageScore depending on time period and risk metric, underscoring that performance comparisons are sensitive to sample and methodology choices.

Regulatory and Policy Context

Regulators, led by FHFA, have pushed the GSEs toward a more competitive credit score framework, explicitly validating both FICO 10T and VantageScore 4.0 after multi‑year testing. FHFA has framed the move as a way to improve risk management through modernized models that incorporate trended data and alternative payment histories and to reduce costs for borrowers by introducing competition into a market that had long relied on a single dominant score.

The Federal Housing Administration has also signaled its willingness to allow VantageScore 4.0 and FICO 10T for FHA‑insured mortgage underwriting, further broadening the potential footprint of multi‑score adoption. At the same time, prudential regulators and industry trade groups continue to stress the importance of robust model governance, back‑testing across multiple economic cycles, and clear guidance on how lenders should operationalize multiple scores for underwriting, pricing, and capital purposes.

Implications for Lenders and Servicers

For originators and servicers, the competing analyses highlight the practical challenges of operating in a multi‑score environment while maintaining compliance and consistent risk controls. Lenders must decide how to integrate FICO 10T and VantageScore 4.0 into existing underwriting engines, what score (or combination of scores) to rely on for overlays and pricing, and how to document their choices for internal model risk management and examiner review.

Servicers and investors, in turn, will need to reconcile how different credit models may affect expectations for default, prepayment, and loss severity across pools, particularly when servicing advances, MSR pricing, and credit‑risk transfer structures have historically been calibrated on FICO‑based distributions. For the collections and recovery side of the business, expanded scoring coverage—especially for thin‑file borrowers—may influence pre‑charge‑off and post‑charge‑off strategies, segmentation, and expected recovery paths on both mortgage and non‑mortgage portfolios tied to homeownership.

Consumer and Fair Lending Considerations

Consumer advocates and some policymakers have long argued that reliance on older FICO models in the GSE channel disproportionately constrained access for minority and thin‑file borrowers and did not fully reflect improvements in rent and utility payment reporting. VantageScore’s use of alternative data and its broader scoreable population are central to its claim that the credit score transition will expand homeownership opportunities without compromising safety and soundness, a claim it buttresses with the Prosperity Now and large‑bank studies.

However, fair lending and UDAAP risk considerations remain front‑of‑mind as lenders adjust cutoffs, overlays, and pricing to reflect new scores, especially if different models produce different score distributions across protected classes. Both FICO and VantageScore will likely face ongoing scrutiny from regulators, academics, and advocacy groups assessing whether model changes and multi‑score strategies reduce or exacerbate disparities in approval rates, pricing, and long‑term default outcomes across demographic groups.bankingjournal.

Key Questions for Credit and Collection Firms

For readers of Credit and Collection News, the dueling analyses raise several questions that will shape how credit and collection strategies evolve as GSE score modernization proceeds. Among them:

• How will lenders translate multi‑score model performance into concrete underwriting and pricing rules, and what will that mean for future borrower delinquency patterns feeding into collections pipelines?

• Will expanded scoring coverage and possible marginal loosening for certain segments lead to larger or more diverse servicing and recovery portfolios, and how should agencies adjust segmentation and treatment strategies in response?

• How will examiners and investors evaluate model risk when different studies reach conflicting conclusions about whether FICO or VantageScore better rank‑orders risk in specific portfolios and vintages?

As the GSEs continue to roll out additional data and implementation guidance, credit and collection professionals will need to track both the technical model debates and the operational decisions lenders make in response, since today’s score choice will help shape tomorrow’s delinquency and recovery landscape.