Taxpayers have a huge stake in ensuring that financial risks are properly and predictably measured for the government loan programs they have been forced to backstop. Credit scores are vital tools for taking those measurements, so the loan space has common reference points for creditworthiness.
But a recent evaluation from American Enterprise Institute (AEI) analysts shows that a years-long effort from the government to create “competition” out of thin air for the credit scoring market may have left taxpayers no better off—and perhaps even worse off.
AEI explored that proposition after a VantageScore White Paper claimed its VantageScore 4.0 model had a clear edge over FICO’s longstanding product by predicting “up to 49% more mortgage defaults leading into the COVID-19 pandemic period than Classic FICO.”
This would normally be great news for taxpayers, who are depending on the most accurate credit score environment possible to help avoid costly bailouts of bad loans going forward.
Unfortunately, AEI found that after adjusting for “methodological inconsistencies and selection bias,” VantageScore’s contentions may not be up to snuff:
Once these issues are corrected, the purported performance advantage of VantageScore 4.0 largely disappears. While VantageScore 4.0 has a marginal advantage over Classic FICO in capturing high risk loans within the bottom risk decile, the difference is relatively small. Across the full sample, Classic FICO performs as well as, and in some cases better than, VantageScore 4.0 on several key predictive measures.
But isn’t AEI going out on a limb in drawing its own conclusions? If so, others are sitting in the same tree with them. In late 2024, an Urban Institute study concluded that:
[b]oth credit scoring models effectively distinguish between high-risk and low-risk borrowers,” and even though “VantageScore 4.0 is marginally more effective at identifying high-risk borrowers from among those with the lowest credit scores . . . the differences are small.
The Milliman firm, which assisted with NTU’s policy paper on credit scoring in 2019, conducted its own side-by-side comparison last year and observed “there is information in using both credit scores in evaluating mortgage default risk,” specifically noting “the default rates are generally consistent between scores, but caution should be used when using Vantage directly in existing mortgage models.”
After the 2018 passage of legislation directing Fannie and Freddie’s watchdog, the Federal Housing Finance Agency (FHFA) to develop a process for evaluating credit score models, in 2022 FHFA finally approved two that the GSEs could use: VantageScore 4.0 and the newer FICO 10T. Incoming FHA Director Bill Pulte’s unexpected decision just three months ago to instead “allow” Fannie and Freddie-handled loans to use either VantageScore 4.0 or Classic FICO further roiled lending markets.
VantageScore (owned by three credit bureaus perched elsewhere in the loan ecosystem) is sure to fire back at AEI and others to defend its turf, while the AEI, Urban Institute, and Milliman studies all have important caveats.
Nonetheless, taxpayers are left to wonder whether they are on any safer ground with government mortgage liabilities now than they were when lawmakers and regulators pushed the credit score competition issue seven years ago.
Three of the four studies referenced above found varying, and relatively small differences in the predictability of the two models. And for these small differences, lenders may be incurring major overhead for converting operations to both models that could raise costs for borrowers.
Less scrupulous actors might also be tempted to play off one score against another in each loan situation to qualify as many borrowers as possible—a “gaming” scenario that defeats the purpose of the models in predicting risk, while leaving taxpayers to guarantee an even bigger government-backed lending portfolio.
Instead of more blunders that imperil the nation’s finances, taxpayers need a policy reset now from both ends of Pennsylvania Avenue. It starts with a full disclosure from FHFA, internal documents and all, about how it arrived at the 2022 credit score competition decision. And as NTU testimony has demonstrated, taxes are by far the single biggest consideration in a consumer’s mortgage closing costs.
Congress at least recognized that problem by making Private Mortgage Insurance a more attractive, tax-deductible product via the One Big Beautiful Bill Act. States and localities need to do their part by reducing transfer taxes and rationalizing the building permit process.
Meanwhile, stronger capital requirements, policies to encourage reinsurance and risk transfer away from taxpayers in the mortgage market, and a genuine plan for government divestment of the GSEs all deserve thoughtful formulation from federal leaders. These approaches will create a stronger housing market that will encourage better predictive tools to evolve organically, rather than at government’s spearpoint.
In a recent op-ed for National Mortgage News, the authors of the AEI study concluded, “given the trillions in mortgage debt, millions of borrowers, and immense taxpayer exposure, it is essential that the debate be grounded in facts rather than marketing.”
Taxpayers agree . . . and they are still wondering when public officials will recognize the need for that debate to happen.