Capital Markets Commentary Week Ending 6-19-26

MBS Demand: Who's Buying Mortgage Bonds?
Demand for agency mortgage-backed securities remains broadly supportive heading into the summer, though the buyer mix is shifting.

The GSEs (Fannie Mae and Freddie Mac) have dialed back the pace of their own MBS purchases, but they're effectively acting as a backstop. If mortgage bond spreads were to widen meaningfully, they'd likely step back in to stabilize the market. That implicit floor underneath spreads is an important dynamic for rate stability.

Banks were active buyers earlier this spring, taking advantage of wider spreads in March to add mortgage bonds opportunistically. For the rest of the year, though, most banks are expected to shift into maintenance mode, reinvesting paydowns rather than growing their MBS holdings, unless spreads widen enough to make fresh purchases attractive again.

On the positive side, mortgage REITs and foreign private investors continue to provide steady demand. And money managers have been a notable source of strength. Strong fund inflows have allowed them to keep buying mortgages even as they've trimmed their overall allocation to the sector. In other words, more money coming in the door has more than offset a modest portfolio rebalancing away from MBS.

The net picture is a market with enough demand to keep spreads relatively well-behaved, but without the aggressive buying pressure that would drive meaningful tightening on its own. Any sustained spread compression will likely require a catalyst, such as the decline in rate volatility now underway from the Iran peace deal framework.

Credit Score Optimization: What "Best Score Wins" Means for the Numbers You See
The GSEs are gradually rolling out VantageScore 4.0 (VS4) as an alternative to classic FICO for setting loan-level price adjustments (LLPAs). Rocket is already offering VS4 to borrowers, the GSE pilot program has reportedly expanded from 21 to roughly 30 lenders, and a small number of VS4-scored loans are beginning to appear in mortgage-backed securities pools. Adoption is still in its early stages, this will be a slow build, not an overnight switch, but the momentum is real.

That expansion raises an important question about the credit scores investors will actually see on these loans going forward.

Here's the issue. Under the new framework, an originator can pull both a borrower's FICO score and their VS4 score, then use whichever one produces a better LLPA outcome, meaning lower fees and a cheaper loan. The GSEs, however, only disclose the single score that was used. Investors never see the other one.

On average, VS4 scores run about 20 points higher than FICO for the same borrower. But that's just an average, there's enormous variance around it. For some borrowers, VS4 is dramatically higher than FICO; for others, it's actually lower. That spread is what creates the optimization opportunity.

When a lender pulls both scores and picks the better one, the math is straightforward: they'll use VS4 when it's unusually favorable and FICO when it's unusually favorable. The score that loses the comparison is never reported. The result is that every disclosed score, whether labeled FICO or VS4, is a cherry-picked winner, not a representative number.

Historical analysis of GSE loans originated between 2013 and 2023 illustrates the impact. In a simulated best-of-both scenario, lenders would have chosen FICO about 48% of the time and VS4 about 52% of the time. The average FICO scores in the cherry-picked group came in roughly 10 points higher than the full-population FICO average, and the average VS4 scores came in about 19 points higher. In other words, the scores investors see will systematically overstate borrower credit quality relative to what either scoring model would show on its own.

This doesn't mean the loans are bad, but it does mean that a "720 FICO" in the new system isn't quite the same "720 FICO" investors are used to seeing. As score optimization becomes more widespread, the credit scores disclosed on GSE pools will drift upward in a way that reflects lender strategy, not an actual improvement in borrower creditworthiness. It's an important nuance for anyone pricing credit risk off of disclosed scores.

Policy Watch: What's Still on the Table (and What Probably Isn't)

Streamline Refinances: A Big Idea Still Waiting on the Plumbing
One of the most impactful policy moves still potentially in play is a conventional streamline refinance program, a simplified, low-documentation refi path similar to what FHA and VA have offered for years. The idea surfaced in the White House's March executive order on housing, and recent reporting from HousingWire suggests the administration is targeting a year-end timeline. The key gatekeeper is the CFPB, which would need to revise its Qualified Mortgage rules to make a GSE streamline program possible. President Trump recently nominated Brian Johnson as the agency's permanent director, a role that has been filled only on an interim basis until now, which could provide the institutional stability needed to get it done.

If the regulatory groundwork gets laid, standing up the actual program at the GSE level should be relatively straightforward. But there's a significant catch. A streamline refi program would dramatically increase prepayment risk on existing mortgage bonds. Faster prepayments make MBS less valuable to investors, which pushes mortgage rates higher to compensate. That is exactly the opposite of the administration's stated goal of lowering borrowing costs. It would also erode the value of the GSEs' own MBS holdings from the $200 billion purchase program, though the accounting pain might not show up immediately. In short, a streamline program is a powerful affordability tool the next time rates fall meaningfully, but launching it in a rising-rate environment would be counterproductive.

VA Streamline Refis: Congress Moves to Raise the Cost
On the legislative side, the House passed H.R. 6047, which would nearly triple the funding fee on VA Interest Rate Reduction Refinance Loans (IRRRLs), the VA's streamline refi product, from 0.50% to 1.40%. The fee increase is designed to offset the cost of expanded veteran benefits elsewhere in the bill, not to address any issue with the refi program itself.

For borrowers, the impact is meaningful but not dramatic. Assuming the higher fee gets rolled into the loan balance, as most borrowers do, it works out to roughly an 8-basis-point increase in effective borrowing cost when measured in monthly payment terms. That's enough to modestly reduce the incentive to refinance through the VA program, particularly for borrowers with smaller rate savings on the table. The bill still needs to clear the Senate, so it's far from final, but it's worth watching as a potential headwind for VA refi volumes later in the year.

LLPA Revenue and the Credit Score Gaming Problem
There's a direct financial consequence for the GSEs as score optimization becomes widespread. When lenders pull both FICO and VantageScore 4.0 and use whichever produces a lower LLPA, the GSEs collect less fee revenue on the same pool of borrowers than they would have under the old single-score system. As the VS4 pilot expands to more lenders and more of these optimized loans flow into securities pools, that revenue leakage adds up. The GSEs will almost certainly need to recalibrate their LLPA grids to account for it.

That recalibration could open the door to a broader repricing conversation. FHFA Director Pulte has discussed LLPA adjustments for some time, and the score-gaming dynamic may provide the catalyst. Questions on the table include whether to steepen the pricing grid, charging more for higher-risk loans and less for lower-risk ones, and whether to revisit LLPAs on investor properties and second homes, potentially making GSE pricing more competitive to keep that volume from migrating to private channels.

MIP and G-Fee Cuts: Fading from the Conversation
Earlier in the year, particularly around the Davos meetings, there was real speculation that the administration might pursue administrative cuts to FHA mortgage insurance premiums or GSE guarantee fees as quick-win affordability measures. Those are actions that don't require legislation, just a policy decision. But the window appears to be closing. The simplest argument: if they were going to do it, they could have already, and they haven't.

There are good reasons for the hesitation. The FHA insurance fund's capital ratio of 11.47% looks healthy on paper, but roughly half of that is accounting-based net present value rather than actual cash reserves. And current MIP levels are already about as low as they can realistically go while still covering expected losses on new originations. A cut might generate a favorable headline, but it would come with real financial risk to the fund, a tradeoff the administration appears unwilling to make.