181bp, Narrowest Since February 2022 (September 2026): Why the Fed's Hike Doesn't Set the 30-Year Mortgage Rate — the 10-Year Yield Does

The Federal Open Market Committee meets on 15–16 September 2026 with the federal funds target range at 3-1/2 to 3-3/4 percent and three members who already dissented in July in favour of a hike. Freddie Mac's survey put the 30-year fixed mortgage at 6.76% in the week ending 10 September, while the 10-year Treasury closed 11 September at 4.96%. The gap between those two numbers — 181 basis points — is the narrowest since February 2022, and it is the part of a mortgage rate the Fed does not control and has already spent.

What the Fed actually decides on Wednesday

The Committee sets a target range for the federal funds rate, the rate at which banks lend reserves to each other overnight. That is the entirety of the instrument. Everything else — the two-year note, the ten-year note, a corporate bond, a mortgage — is priced by markets that take the current and expected path of that overnight rate as one input among several.

The statement from the July 2026 meeting is worth reading precisely because nothing happened at it. The Committee held the range at 3-1/2 to 3-3/4 percent by a 9–3 vote, with Beth Hammack, Neel Kashkari and Lorie Logan dissenting in favour of raising by a quarter point, and it described inflation as elevated relative to the 2 percent goal "in part reflecting supply shocks." September is a projection meeting, so a Summary of Economic Projections arrives alongside the decision.

The front end has already moved. Against an effective rate of 3.63%, the Treasury par yield curve for 11 September shows the 1-month bill at 3.93%, the 1-year at 4.35% and the 2-year at 4.63%. That is a market that has already priced tightening, not one waiting to hear about it. Which is the first reason Wednesday's 25 basis points, in isolation, does very little to a mortgage rate: the mortgage market has been trading the same expectations for weeks.

Where 6.76% comes from

A 30-year fixed mortgage rate is built in three layers.

The base is the 10-year Treasury yield. A 30-year loan is not held for 30 years — it is refinanced, or the house is sold — so its cash flows behave like a shorter bond, and the 10-year is the market's chosen proxy. The second layer is the spread investors demand to hold mortgage credit rather than government credit, which mostly compensates for one specific feature: the borrower owns a free option to prepay. When rates fall, the loan disappears into a refinancing and the investor gets their money back at the worst possible moment. When rates rise, the borrower keeps the cheap loan and the investor is stuck holding it. That asymmetry has a price, and the price rises with interest-rate volatility. The third layer is the originator's own margin — the difference between what a lender charges a borrower and what the loan fetches when it is sold on.

181 basis points: the spread has already done its work

Setting Freddie Mac's Primary Mortgage Market Survey against the Treasury's own daily 10-year yield for the same week produces a single series that explains most of the last four years of American housing.

Computed from Freddie Mac PMMS weekly 30-year fixed rates and US Treasury daily par yield curve 10-year rates, matched to the survey week.

Read the last column. The spread has compressed in every single year since 2023, and the 181bp printed in the week ending 10 September 2026 is the narrowest weekly reading since 10 February 2022 — before the Fed's tightening cycle had even begun.

That compression is why mortgage rates fell while Treasury yields did not. Take the two dates directly. On 2 November 2023 the 10-year was 4.67% and the mortgage 7.76%, a 309bp spread. On 10 September 2026 the 10-year was 4.95% and the mortgage 6.76%, a 181bp spread. The Treasury yield was higher on the later date. The borrower's rate was a full percentage point lower. Every basis point of that improvement came from the spread.

Why the cushion matters more than the level

A spread at 181bp against a 2014–2021 norm of 175bp is, roughly, normal. That is the point, and it cuts in an uncomfortable direction.

For three years the mortgage market had a built-in shock absorber. Treasury yields could rise and the borrower's rate would rise by less, because the spread was abnormally wide and grinding tighter as rate volatility fell and as buyers returned to mortgage-backed securities. The most recent week shows the absorber running out of travel: the 10-year rose 18 basis points between the survey weeks ending 3 and 10 September, from 4.77% to 4.95%, while the survey mortgage rate rose five, from 6.71% to 6.76%. The spread took the other thirteen. It cannot keep taking them from 181.

So the mechanism to watch on Wednesday is not the decision. It is whether the projections and the press conference move the 10-year — and, separately, whether they move implied interest-rate volatility, because a surprise that widens the expected range of future rates raises the value of the prepayment option, and can widen the spread and lift the 10-year at the same time. That combination is the one that actually reprices a mortgage.

What Thursday's housing data will and won't tell you

The Census Bureau publishes New Residential Construction for August at 8:30 a.m. Eastern on 17 September. The July report contained the split that matters. Building permits were at a seasonally adjusted annual rate of 1,443,000, up 5.0% on the month and 3.1% on the year. Housing starts were at 1,239,000, down 12.4% on the month and 13.5% on the year, with single-family starts at 808,000. Completions fell 9.1% to 1,212,000.

A permit costs relatively little and commits a builder to nothing. A start commits capital and labour to a house that will be finished and priced in a market nobody can see yet. Builders taking more of the first and fewer of the second is a statement about expected demand in 2027, and it is a cleaner signal than either line alone.

The rest of the picture supports it. New single-family home sales ran at 607,000 in July, down 10.5% on the month, against 488,000 homes for sale — 9.6 months of supply at the current sales rate, where six months is the long-standing rule of thumb for balance. The NAHB/Wells Fargo Housing Market Index sat at 35 in August, a level at which most builders call conditions poor, with 35% of them cutting prices by an average of 6% and 63% deploying sales incentives. Many of those incentives are mortgage-rate buydowns, which means the builder is paying cash to compress the very spread this piece is about, one buyer at a time.

What the release will not tell you is where rates go. Housing starts are volatile, heavily revised, and carry confidence intervals wide enough that Census flags some of its own headline moves as inconclusive — the 12.4% fall in total starts came with a ±9.5 percentage-point interval and clears the bar, but the 9.9% fall in single-family starts came with ±10.4 points and does not. Treat the monthly number as texture, not as a signal.

The honest read-across to markets

This is a rates and housing story before it is anything else, and it should be read that way. The instruments it genuinely touches are Treasuries, mortgage-backed securities, and the homebuilding and building-products companies whose gross margins absorb the buydowns described above — which is a matter of arithmetic in an income statement, not a view on any security.

The connection to the currency board is slower and runs one way only. Shelter is the heaviest single component of US consumer price inflation and is measured with a lag long enough that today's cooling shows up in a print some quarters from now. Lower measured inflation eventually lowers the expected policy path, and the expected policy path relative to other countries is what the interest-rate factor on the meter scores for the US dollar. That is a chain with four links in it. Anyone claiming Thursday's housing starts will move the dollar has skipped three of them.

What would change the picture

Three things, in descending order of how quickly they would show up.

A move in the 10-year is the fastest. With the spread back at its long-run norm, a repricing of the term premium at the long end — the mechanism examined in the August long-end selloff, with the 30-year Treasury at 5.35% on 11 September — now passes through to borrowers with very little damping. A sustained inflation surprise is next, because it moves both the expected path and the volatility around it, and August's core CPI at 0.3% against 0.2% expected is the kind of print that does it. Slowest is supply: completions running below starts eventually tightens the resale market, but 9.6 months of new-home inventory says that constraint is not the binding one today.

What will not change the picture is the 25 basis points itself. The Committee will move an overnight rate. The borrower's rate is three layers away, and the layer that did all the work since 2023 has nothing left to give.

More on how this site approaches transmission mechanisms is on the about page, and the underlying reason long yields and currencies move together is covered in bond yields and currencies.

Educational macro context only — not investment advice.