Nobody's coming to save mortgage rates

Welp, the Fed hiked. 😬 And officials are signaling at least one more could be coming this year. Volatility probably keeps mortgage rates elevated through the rest of the year.

Today, Karen Postiglioni looks at where lenders can actually find growth if cheaper mortgages aren’t coming. Plus, BofA’s Jeana Curro breaks down the Fed, MBS 🌎 & why the rest of ’26 could stay pretty spicy.

By Karen Postiglioni

AI is already making mortgage production cheaper & salespeople more productive. Pennymac says its rollout of Vesta in its consumer direct channel produced roughly 50% efficiency gains for loan officers, cut end-to-end loan processing time by about 25%, & reduced operational cost to originate by 25%. Rocket says its AI tools are helping its bankers handle nearly 40% more clients while improving conversion. 

This is a good thing! But let’s be clear: if you’re still debating whether to automate fulfillment or give your salesforce better tools, those are no longer differentiators. Those are table stakes.

Now let’s run some math. My favorite!

Let me be clear: $500 in savings is not trivial. It’s more than half of today’s average pre-tax profit. Go get those savings, but do keep your eye on where the costs sit.

STRATMOR’s most recent public retail breakdown, based on ‘18-’22 benchmarking, puts about 55% of origination expense on the sales side & 22% in fulfillment.

STRATMOR uses $2,500 as a typical fulfillment cost & asks what happens if a lender cuts it by a very healthy 20%. There’s the $500. But once you automate a task, that savings is captured. The next project has a smaller cost base left to attack. You can only cut the same cost once.

So, what’s next?

Better’s interim CEO Daniel Lewis spoke plainly last month about the strategy: “Looking ahead, Better will win by leveraging that experience to manufacture mortgages efficiently, not by outspending competitors on customer acquisition.”

Totally fair! Nobody wants to win by paying the most for the same customer. But those aren’t the only two choices. The industry is already getting better at manufacturing, prospecting & conversion. Most of that still starts w/ someone who already looks like a mortgage opportunity. What about the people who don’t?

The Fed’s May ‘26 household report found that 16% of U.S. adults used BNPL in ‘25. Among those users, 26% paid late. Only 14% of BNPL users correctly answered both questions the Fed asked about how BNPL affects their credit.

Some of those consumers clearly have work to do to become more financially ready, while others may be closer than they think, or getting colder w/o even knowing it.

A ‘23 Fannie Mae study found that 90% of consumers overstate or don’t know the minimum down payment required for a typical mortgage (still). Roughly 32% could not correctly approximate the minimum credit score required by lenders. There’s still a big gap between what consumers assume & what this industry knows. And gaps are opportunities!

Customer acquisition finds someone who is ready, but demand creation helps more people become, or stay, ready. With permission, technology can tell someone where they actually stand. What’s keeping them from qualifying? What isn’t? Which programs fit? What would actually improve their position?

Then update that guidance as things change. That’s more useful than an LO calling every 60 days or a CRM wishing them happy birthday until ‘29. (That’s when AI is killing us all anyway, per Jacob Coxon.)

AI doesn’t eliminate acquisition; consumers still have to find you and give you permission to engage. But once they do, not ready today doesn’t have to mean dead lead.

Start w/ the customers you already paid to acquire. We’re very good—some more than others—at waking up when rates move, equity changes or the recapture model lights up. What about the years in between? With permission, help customers stay mortgage-ready, so the relationship remains useful, before the next transaction signals appear.

We talk a lot about recapturing customers, but we should spend more time helping to keep them recapturable.

There is another payoff for lenders: LO retention. There will always be someone willing to throw more bps at a productive originator. If your only answer is more comp, that game gets expensive fast. At some point, the better answer is more business.

What if part of the lender value proposition were simple: We help create your future pipeline. Let technology do the long-tail work until there is something worth handing to an LO. Another lender offers 10 more bps? Fine. Can they bring the LO another 10 customers? That is harder to match.

Keep automating & keep improving conversion. But don’t confuse table stakes w/ strategy. Fulfillment savings are finite b/c you can only cut the same cost once.

The next part of the strategy has to include expanding who can become a customer, keeping the customers you already have ready for what comes next, & giving LOs more business instead of endlessly bidding up comp. That’s the move from customer acquisition to demand creation. The next mortgage starts before the consumer looks like a mortgage lead.

When I got on the phone w/ BofA’s Jeana Curro Wednesday morning, the Fed hadn’t hiked yet. A few hours later, it did 🤷 .

The FOMC raised rates by 25 bps to a 3.75%-4.00% target range, the first hike since ‘23. Fed officials also signaled they expect at least one more increase before the end of the year. Four policymakers said they anticipate two more hikes in ‘26 to tamp down inflation.

That was pretty much the setup Curro, who runs agency MBS strategy at BofA, had laid out for me this morning. BofA’s economists have been calling for three hikes this year: September, October & December. Their thesis is that Kevin Warsh reverses Powell’s three cuts.

“Either they hike three times or they don’t at all,” Curro told me before the Fed decision. “You don’t want to start hiking in October prior to midterms.”

Well, hike No. 1 is officially in the books, & if you’re waiting for the Fed to ride in & rescue 🚑 mortgage rates, Curro doesn’t see much reason for optimism. The 10-year Treasury cracked 5% on Monday & was sitting around 4.96% when Curro & I spoke Wednesday morning. As I write this at 3:42 p.m. est, it’s over 5% again.

BofA’s rates team still sees the 10-year eventually settling around 4.50%. Keep something close to the roughly 200 bps gap between Treasurys & primary mortgage rates, & that leaves mortgages somewhere in the 6.50%-7% neighborhood.

“It’s really hard to get mortgage rates down if Treasury rates continue to move higher,” Curro said.

Remember the administration floating the idea of the GSEs buying $200B of MBS this year? Boy do I have an update for you…