Lies, Damn Lies & Mortgage Studies

A new Bankrate report — co-authored by its CEO, naturally 😜 — claims that 87% of American mortgage borrowers are likely overpaying on their home loans as of '25, to the tune of $3,343 a year for the typical borrower. The "study" even made the pages of the WSJ.

Luckily, there’s a fix available for consumers: Bankrate! "It's not because better rates were unavailable — it's because most borrowers never found them," the company said in a press release. The $65B borrowers overpay each year "could have been saved had borrowers used Bankrate's mortgage auction, where lenders compete in real time on price alone. That's not a market problem. It's an access problem."

I hope you had a good chuckle after reading that. In today's edition, we explain why that scary headline number is pretty much baked in from the start. Plus: a new bill would direct the GSEs to buy construction loans to ease the affordability crunch, & we dig into what that'd mean for the industry. We've also got news on where VC-backed tech bros might be juuuuust a little naive, who's got really sharp DSCR pricing, where a pair of $100M-plus producers are headed & more.

So let's get into why that 87% figure is pretty shaky. Start w/ the benchmark. "Overpayment" is defined as how much a borrower's actual rate exceeded "competitive market offers available for their exact borrower profile." And those offers come from Bankrate's own marketplace. So the finding boils down to: people who didn't use Bankrate got worse rates than Bankrate quotes. Genius.

The 87% is pretty close to tautological IMO, rigged by the construction of the study. Besides, anytime you measure a population against the best available offer, the overwhelming majority is going to land above it. Like, duh. Did I get the absolute best deal on toilet paper? Almost certainly not, but I also didn't spend two hours driving to four stores to find out.

Like every HMDA-based study, this one starts w/ a dataset that has no FICO scores in it, so Bankrate estimated them "based on typical borrower profiles." Credit is one of the largest determinants of a mortgage rate; the same profile with a 60-point spread can land in a completely different pricing grid (👋 VantageScore lol). So when Bankrate claims to compare each borrower against offers for their "exact borrower profile," they're basically guesstimating the single most important input to that profile.

A few more problems here. It relies on borrower-submitted data. A quote based on self-reported numbers is not a loan that survived underwriting, appraisal, & documentation. "$65 billion could have been saved" also assumes every borrower could simultaneously capture the marginal best rate, which obviously isn’t possible & in a real-world scenario would drive up pricing. And the study ignores consumers paying points, the certainty of closing, servicing, the overall experience & closing costs. Pretty meaningful stuff, right?

If mortgage lending were purely about getting the absolute best rate, Provident Funding would be a top-5 lender in America. Or Spirit would rule the skies. 

The study does raise some good points. Consumers do absolutely benefit by shopping around. And I don’t think we should discount that consumer marketplaces like Bankrate provide educational value to unsophisticated shoppers at minimum. How much do they ‘save’ borrowers? I don’t know. But let’s be honest about what this is: a lead-gen company w/ a huge profit motive trying to portray its business model as deeply consumer friendly while simultaneously ignoring the complexities of the industry it’s also making money off.

Fannie & Freddie may get a new line of business if one Wisconsin Republican gets his way. Rep. Scott Fitzgerald (no relation to the author🥚) dropped three GSE reform bills Thursday, the headliner being the "Working Families Home Construction Act of 2026," which would direct Fannie & Freddie to buy homebuilder construction loans & securitize them. Were it to happen (a long ☄️ shot), this would put the twins’ balance sheets behind the vertical-construction phase for the first time.

Let’s break this one down b/c it’s a pretty fun one.

The pitch is rather straightforward: If the GSEs take on more of the construction risk, lenders get way more comfortable funding builders, borrowing costs come down, & more homes are built. Homes financed this way would have to sell to buyers earning 90–130% of AMI, per Fitzgerald’s bill, so there’s some supply side, trickle-down ideology at work on the affordability component. 

"If construction financing becomes easier, more homes can be started, completed, & converted into permanent mortgages," Baird's Kirill Krylov told Bloomberg.

Fitzgerald's other two bills are the more familiar fare: one would end conservatorship (paired w/ mandates for more private MI, more CRT, & capped retained portfolios), the other would trim post-’08 disclosure rules. It’s uh, interesting timing. The bills landed a day after Trump pulled the signing on the ROAD to Housing Act 🙂‍↔️that had already cleared both chambers.