Forward-Looking Credit Scoring Matters: Refuting AEI’s Critique of VantageScore 4.0
Recently, the American Enterprise Institute (AEI) published a paper that challenges a previously published VantageScore white paper comparing Classic FICO and VantageScore 4.0 in mortgage. The AEI paper claims that “both VantageScore 4.0 and Classic FICO are effective in identifying high-risk loans, with only marginal differences between the two.” However, when taking a closer look, the true story tells itself. Their analysis rests on misunderstandings, outdated assumptions, selective omissions, and a general refusal to acknowledge what is coming next for mortgage finance.
AEI’s critique tells yesterday’s story. VantageScore 4.0 is about tomorrow.
VantageScore 4.0 Comparison Analysis Demonstrates Superior Performance Within a New Regulatory Framework
The Director of the FHFA announced on July 8, 2025, that VantageScore 4.0 or Classic FICO credit scores could be used on a tri-merge basis with immediate effect. Based on this decision, VantageScore published a “head-to-head” comparison of VantageScore 4.0 to Classic FICO, demonstrating VantageScore 4.0’s substantially superior predictive performance by over 11%.
The use of the “tri-merge” methodology by VantageScore for comparison uses the currently announced regulatory changes. Director Pulte has confirmed the “tri-merge” modernization1.
On September 5, 2025, the American Enterprise Institute published a false, outdated and inaccurate analysis using the historical credit score “aggregation method” used by FICO as the middle of three scores/lower of two scores. This method should not be relied upon by industry participants because it is inconsistent with the FHFA’s current regulatory guidance and practice.
VantageScore 4.0 Performs Better in Macro-Economic Crisis Periods
Based on a standard stress testing analysis and echoing an independent study by Bloomberg Intelligence2, the VantageScore white paper concludes that VantageScore 4.0 outperforms Classic FICO in predicting defaults entering the pandemic period by +48.5%.
The American Enterprise Institute study criticizes VantageScore for analyzing mortgages within a specific score range, claiming the approach introduces “bias.” But that criticism misses the point. Effective stress testing is about pressure-testing the segments most vulnerable to default. By concentrating on borrowers who are historically more prone to stress, analysts can see whether a model truly separates high-risk from low-risk consumers.
During the pandemic, historical default patterns made it clear which groups carried the greatest exposure. Focusing on those populations is necessary in a stress test. If a risk model can’t hold up under scrutiny where the exposures are highest, it isn’t robust. That’s exactly why the population in question was not only justified, but rather essential to proving the model’s ability to differentiate risk with precision.
Interestingly, the AEI’s white paper admits that, even after performing their own recalculation and removing what they consider to be methodology flaws, VantageScore 4.0 still demonstrates predictive improvement and captures more defaults in the riskiest decile - a 32.9% versus 32.0% for Classic FICO. But instead of highlighting this advantage, they brush this finding off as a “modest 3% lift”. However, this gain is applied to the trillion-dollar GSE portfolio, so a 3% advantage is enormous. Tens of thousands of defaults identified earlier could mean billions in avoided losses, and calling this gain “modest” is like dismissing a safety feature that prevents thousands of accidents a year.
Notably, Classic FICO was the only score used in mortgage originations in 2008, when deficiencies in stress testing led to a deterioration in underwriting standards that ultimately resulted in the collapse of the housing market (per the St. Louis Fed3).
VantageScore 4.0 Identifies More Creditworthy Borrowers with Reduced Risk
VantageScore 4.0 can responsibly expand access to credit to more than 5 million Americans who are locked out of homeownership by an old scoring system. These are veterans, immigrants, and hard-working Americans who are currently shut out of the system but can be included when using a more modern scoring technology such as VantageScore 4.0.
The American Enterprise Institute study refers to the fact that the mortgage released dataset only includes loans that FICO allowed in the market. That is true, but this “survivorship bias” (as they call it) creates an advantage to only Classic FICO.
The truth is that VantageScore 4.0 provides broader mortgage access, using smarter data and more modern technologies to correctly reveal millions of creditworthy Americans.
Multiple Areas of VantageScore and Independent Research Point to the VantageScore 4.0’s Predictive Power
Independent RBMS-based studies showcasing that VantageScore 4.0 is “better” than Classic FICO in mortgage are ignored by the AEI white paper (e.g., Bank of America, KBRA, JPMorgan Chase, and Bloomberg Intelligence).4
Additionally, even the two independent research studies (Urban Institute, Milliman)5 cited by the AEI’s white paper confirm that both models are predictive, but that VantageScore 4.0 is stronger in identifying high-risk borrowers at the lower end of the credit spectrum.
Moreover, AEI avoids engaging with the full breadth of the VantageScore’s own published white paper which provides evidence of VantageScore 4.0’s predictive superiority over the entire span of ten years of performance data, for different borrower types and timeframe slices across pre- and post-COVID periods. If AEI’s critique had merit across these broader datasets, one would expect them to address those results directly. Their silence speaks volumes.
In conclusion, forward-looking credit scoring matters. The AEI report attempts to defend the status quo by nitpicking methodology and downplaying results that show clear advantages for VantageScore 4.0. However, this is not only about model accuracy; it is about whether we keep propping up an outdated system or embrace one designed for the future.
- Tri-merge aggregation is the direction regulators have already chosen for mortgages.
- Stress-testing riskier borrowers is the right approach to evaluate predictive power, and small percentage lifts equal massive real-world impacts.
- VantageScore expands credit access to millions of creditworthy Americans.
VantageScore 4.0 is not just a comparable model; it is the future-a more inclusive, modern, forward-looking, and risk-sensitive credit scoring model that strengthens the housing finance system and expands access to responsible credit.
Footnotes
-
“Pulte says GSEs will accpt VantageScore 4.0 immediately.” HousingWire. July 8, 2025. https://www.housingwire.com/articles/pulte-says-gses-will-accept-vantagescore-4-0-immediately/ ↩
-
Erica Adelberg, Viktoria Adamova, Bloomberg Intelligence Strategy, July 22, 2025, citing “VantageScore 4.0 has 15% higher success rate at predicting defaulters.” ↩
-
The Financial Crisis Inquiry Report - Final Report Of The National Commission On The Causes Of The Financial And Economic Crisis In The United States, January 2011. https://fraser.stlouisfed.org/files/docs/historical/fct/fcic_report_final_20110127.pdf ↩
-
Pratik Gupta, Head of CLO & RMBS Research at Bank of America Securities, "*THE SCORE Podcast: Bank of America: VantageScore 4.0 is "Better" for Mortgage*", YouTube, September 11, 2024, https://www.youtube.com/watch?v=iPgilo13Gz8 citing "We conclude that VantageScore 4.0 is overall a better indicator for capturing both prepayment (elevated cashout) and delinquency."
John Sim, Ani Gelashvili, J.P. Morgan Securities LLC, “JP Morgan Chase North America Securitized Products Research”, August 2, 2024 citing “We see that VantageScore is able to capture 2-4% more ever delinquent loans than Classic FICO.”
Jack Kahan, Global Head of ABS & RMBS KBRA, July 9, 2025 citing “slight advantages in granularity observed in VS4 for lower-score borrowers.”
Erica Adelberg, Viktoria Adamova, Bloomberg Intelligence Strategy, July 22, 2025 citing “VantageScore 4.0 has 15% higher success rate at predicting defaulters.”
↩ -
Laurie Goodman, Jun Zhu. "Classic FICO versus VantageScore 4.0: A mortgage credit score comparison." Urban Institute, Dec. 2024.
Ricardo Nunez Magana. “Cracking the Tape on VantageScore 4.0.” Milliman. Sep. 2024.
↩