Please, fake news, don’t give insurance regulators any bright ideas

A relatively minor bureaucratic change proposed by the Federal Housing Finance Agency sparked a viral storm in right-leaning news media recently, with stores that Washington Times, New York Post, National review and Fox News all report some variation of the sentiment expressed in Times headline: “Biden to raise payments for homebuyers with good credit to subsidize high-risk loans.”

The underlying issue concerns the FHFA’s recent decision—as conservator of the government-sponsored enterprises (GSEs)—to revise the loan-level price adjustments (LLPAs) charged by Fannie Mae and Freddie Mac, which together account for roughly 60% of U.S. housing mortgage loan. The LLPAs charged by GSEs are primarily determined by loan type, loan-to-value ratio and a borrower̵

What’s largely consistent in the coverage is that the changes — first announced in January, affecting loans delivered to the GSEs on or after May 1, and therefore already implemented by lenders for months — overall tend to reduce costs for those with lower credit scores and increase costs for those with higher credit scores. In fact, as part of a broader rate change announced last year, the FHFA eliminated conventional loan fees altogether for about 20% of homebuyers, funded by increased down payment fees for vacation homes, high-balance loans and cash-out refinances.

Unfortunately, the way this story has been spun in the wake of the changes would leave many news consumers with the impression that borrowers with higher credit scores will pay more directly in fees than borrowers with lower credit scores. This is certainly not the case. Comparing apples to apples, at all levels of the grid, a borrower with a higher credit score would continue to have a lower LLPA (or, in many LTV categories, none).

Writing in his Substack newsletter, Kevin Erdmann of the Mercatus Center responded to a Fox News graphic that declared, under the new rules, a “620 FICO score gets a 1.75% fee discount” while a “740 FICO score pays a fee of 1%”:

I’m pretty sure what they’ve done here is pick the low credit score that had the biggest fee reduction. Then they reported the total fee for a higher credit score. So a low down payment of 620 points has a fee that went from about 6.75% to 5% (when mortgage insurance is included). And also the fee for a 740 point went from 0.25% to 1%. (plus a 0.25% mortgage insurance fee). Why didn’t they just say the fees for 740 points increased by 0.75%? It would still get their partisan point across. It would still be strange, as it would describe mortgages with two different down payments. And that would hide the fact that the 620 points still have a fee that is more than 3% higher than the 740 points. But at least it wouldn’t be mixing levels with changes.

Ultimately, whether these changes are good or bad for the GSEs is an actuarial question. As Erdmann goes on to state, there is good reason to believe that the fees on borrowers with lower credit have been too high for an extended period.

But there are other reasons to worry about what the event could mean for insurance markets. The concern here is that government regulators — or, at worst, Congress — might think charging those with high credit scores more to subsidize those with low credit scores might actually be an idea worth emulating.

Apparently, insurers’ use of credit information in underwriting and rate setting has been the subject of public debate for four decades. At this time, while a handful of states ban this practice outright, most have passed legislation allowing it, with some caveats.

The FHFA precedent — allowed because Fannie and Freddie have been in the agency’s conservatorship for nearly 15 years — is particularly troubling given recent cases of state insurance regulators moving to restrict or prohibit the use of credit information without any express direction from state lawmakers to do that. Whether courts choose to uphold such unilateral decisions depends on the specifics of state law.

Last year, Washington State Insurance Commissioner Mike Kreidler moved to adopt a permanent rule imposing a three-year ban on the use of credit-based insurance scores, after a predecessor emergency rule to do the same was invalidated in September 2021 by Thurston County Superior Judge Indu Thomas. An August 2022 final order by Thomas found that Kreidler exceeded his authority in applying the rule when there was a specific state statute that allowed insurers to use credit scoring.

More recently, the Nevada Supreme Court ruled in February to uphold a temporary ban on the use of credit information in insurance quotes originally issued by the Nevada Division of Insurance in December 2020. The rule, which is scheduled to expire on May 20, 2024, was unsuccessfully challenged by Mutual National Association of Insurance Companies.

The rise of credit-based insurance scoring has revolutionized the industry, allowing for much greater segmentation and better matching of risk to rate. Where government residual car insurance entities once insured as much as half or more of all private passenger car risks, they now represent less than 1% of the market nationwide. It would be unfortunate if a few misleading headlines inspired ill-conceived regulation to reverse that trend.