Lock at 7.03% or Wait for the Dots: What the Fed’s September Hike Did to the Buy-Versus-Wait Math

Updated September 27, 2026. Every rate, index and dollar figure below was read on that date from the issuing institution named at the end of this page. No lender quote, broker estimate or retailer price appears anywhere in it.

An open house sign outside a home for sale

Photo: Wikimedia Commons, CC BY-SA 4.0

The week this page was written, and the four weeks that undid it

At the end of August the 30-year fixed average had gone almost nowhere: Freddie Mac's Primary Mortgage Market Survey printed 6.65 percent for the week of August 20 and 6.66 percent for the week of August 27. One basis point. Writing "rates barely moved" was correct.

Then the following happened, in this order, all of it on the record:

Week ending30-year fixed15-year fixedWhat else happened that week
Aug 20, 20266.65%5.95%—
Aug 27, 20266.66%5.98%—
Sept 3, 20266.71%6.04%WTI crude rose from $87.03 (Aug 31) to $92.55 (Sept 3)
Sept 10, 20266.76%6.09%10-year Treasury 4.95% on Sept 10
Sept 17, 20266.95%6.26%FOMC raised the target range 25 bp on Sept 16, vote 12-0
Sept 24, 20267.03%6.42%10-year Treasury 5.18% on Sept 24

The federal funds target range is now 3-3/4 to 4 percent, effective September 17, 2026. That is a hike, not a cut, and the 30-year fixed has moved 37 basis points in four weeks. So the sentence this page originally led with is now a historical note, and the question worth the reader's time has changed with it: not whether affordability improved in August, but what the September hike does to the arithmetic of locking now versus waiting.

Three claims from that stretch that do not survive a check

Three companion pieces on this site covered the same fortnight. Checked against the issuing sources, each has a problem worth correcting here rather than quietly leaving in place.

  1. "Mortgage rates ticked up today" and "mortgage rates jumped again today" are not statements anyone can source. Freddie Mac's PMMS is a weekly survey: it averages rate locks from Thursday through Wednesday and publishes each Thursday at noon Eastern. There is no official daily mortgage-rate series in the United States. A daily figure can only come from a private rate tracker, which is not a primary source and was not cited.
  2. On the day one of those pieces called a "jump," the benchmark did not move. The Federal Reserve's H.15 10-year Treasury constant maturity closed at 4.75 percent on August 31, 4.79 on September 1, and 4.79 on September 2 — unchanged on the second day. A four-basis-point rise followed by a flat day is not a jump. The oil move was real: WTI went from $87.03 on August 31 to $91.48 on September 1, up 5.1 percent. It simply did not transmit into the rate benchmark that week. The rate move that mattered came two weeks later, from the FOMC.
  3. The "30 percent demand gap" is not an index. The National Association of Realtors does not publish a metric by that name. The nearest published language is its chief economist's observation, in the August pending home sales release of September 17, that contract signings are running approximately 30 percent below pre-pandemic levels nationally. That is a comparison to a baseline, stated by a person, not a tracked statistic — and it should be attributed that way.

Plan A, Plan B and Plan C, on the same house

Here is the comparison that replaces the weekly rate headline. The house is the national median existing home, $429,100, from NAR's August 2026 release of September 10. Twenty percent down is $85,820, leaving a loan of $343,280. Only principal and interest are shown, because taxes and insurance are identical across the three plans and would only pad every row equally.

PlanRate and termMonthly P&IPaid over 5 yearsPaid over 10 yearsPaid over the full term
A. Lock now7.03%, 30 years (PMMS, week of Sept 24)$2,290.77$137,446$274,893$824,678
B. Had you locked a month ago6.66%, 30 years (PMMS, week of Aug 27)$2,206.01$132,361$264,721$794,164
C. Shorter term instead6.42%, 15 years (PMMS, week of Sept 24)$2,975.26$178,516$357,031$535,547
Reference points, same loan: at 6.00% over 30 years the payment is $2,058.14; at 7.50% it is $2,400.26.

The arithmetic is mine; the rates, the price and the survey dates are the issuers'. Three results fall out of it.

The four weeks cost $84.76 a month. That is $5,086 over five years, $10,171 over ten, and $30,514 across the full thirty. Not catastrophic, and not nothing — it is roughly the difference between a five-year-old used car and no car.

Plan C costs $684.49 more every month and saves $289,131 in total. The fifteen-year loan is the single largest lever on this page, and it is the one nobody writes a weekly headline about because its rate does not move dramatically. It is also the plan most households cannot actually choose: $2,975 a month against a median household income of $87,460 (Census, released September 15, 2026) is 41 percent of gross income before taxes, insurance or anything else.

The spread between the best and worst case on this page is $123,749 over ten years — Plan C's $357,031 against Plan A's $274,893 is the wrong comparison, because Plan C is buying down principal rather than spending. The honest comparison is 6.00 percent against 7.50 percent on the same 30-year term: $246,976 versus $288,032 over ten years, a $41,056 range driven entirely by when the ink dries.

What the Committee's own projections say about waiting

"Wait for rates to come down" is a plan with a testable premise, and the Federal Reserve publishes the test. The Summary of Economic Projections released alongside the September 16, 2026 decision gives a median federal funds rate of 4.1 percent at the end of 2026 and 4.1 percent at the end of 2027, falling to 3.9 percent at end-2028. The midpoint of the range the Committee just set is 3.875 percent.

Read that twice. The FOMC's own median projection for the end of this year sits above the range it just moved to. Whatever the case for waiting, it cannot be built on the Committee's published expectations, which point to at least one more move in the same direction before the year is out. The same projections put median PCE inflation at 3.7 percent for 2026, 2.3 percent for 2027 and 2.1 percent for 2028, and median unemployment at 4.1 percent in all three years. The statement's own words on the September action: it "will support a timelier return to the Committee's 2 percent goal."

The other side of the ledger, which is rent

Waiting is not free, because the alternative to a mortgage payment is a rent payment. The Census Bureau's Housing Vacancies and Homeownership release for the second quarter of 2026, published July 28, puts the national median asking rent at $1,531 a month, the rental vacancy rate at 7.3 percent, the homeowner vacancy rate at 1.2 percent, and the homeownership rate at 65.0 percent. BLS puts rent of primary residence up 2.7 percent over twelve months in its August 2026 CPI, and owners' equivalent rent up 3.1 percent.

Escalating that median rent at the published 2.7 percent gives $96,956 over five years and $207,728 over ten. Against Plan A's $137,446 and $274,893 in principal and interest, renting looks cheaper by $40,490 over five years and $67,165 over ten — until you remember that a large share of the mortgage rows is principal, which you keep, and that the renter's column has no down payment in it either. Put the $85,820 down payment back on the owner's side and the five-year gap widens to $126,310; take out the principal portion the owner builds and it narrows again. This page will not pretend to settle that with a single number, because doing so requires an assumption about home price appreciation that nobody can source. FHFA's House Price Index rose 2.1 percent year over year in the second quarter of 2026 and 0.3 percent quarter over quarter — that is the published rate of change, not a forecast.

Where the August affordability improvement actually came from, and where it went

NAR's own Housing Affordability Index stood at 103.3 in the most recent figure available when this page was first written, against 98.3 a year earlier — a real improvement, and the basis for the original headline. Two things qualify it. That index is published monthly with roughly a two-month lag, so the reading reflected conditions in early summer, not the last week of August. And it is built from the prevailing rate, the median price and median family income; the rate input has since risen 37 basis points, which moves the index the other way.

The supply side moved in the buyer's favour over the same stretch and is still moving. NAR's August existing-home sales release of September 10 reports a 3.98 million annual pace, a median price of $429,100, and inventory at 4.9 months of supply, which NAR describes as the highest in over a decade. Census and HUD's new residential sales release of September 24 puts the median new-home price at $393,700, the average at $478,700, the pace at 684,000 and months of supply at 8.5. More inventory and a lower median for new construction than for existing homes is an unusual configuration and is the part of this market that favours a buyer right now.

Two numbers this page could not put in the table

Property tax and homeowners insurance are the two line items most likely to decide a household's actual monthly cost, and neither could be sourced cleanly. The Census American Community Survey publishes median real estate taxes paid and median homeowners insurance cost as named tables, but the figures live behind a query interface this page could not read. What is available is the bundled figure: median selected monthly owner costs for mortgaged households of $2,035, which is 21.4 percent of income at the median and includes taxes, insurance, utilities and fees together. Multiplied out, that bundle alone is $122,100 over five years and $244,200 over ten — of the same order as the principal and interest above, which is the reason no rate-versus-rate comparison should ever be presented as the whole decision.

For a buyer using FHA financing, the insurance premium is published and does belong here: an upfront premium of 1.75 percent of the base loan amount, and an annual premium that for a 30-year term on a loan at or under $726,200 runs 50 basis points at 90 percent loan-to-value or less, 50 basis points for 90 to 95 percent, and 55 basis points above 95 percent — the last two for the life of the loan rather than eleven years. The 2026 conforming loan limit for a one-unit property in the contiguous states is $832,750, with a high-cost ceiling of $1,249,125.

Where the numbers leave the decision

Reading only what the issuing institutions publish, the case for treating the next few weeks as a waiting game is weak, and it is weak for a specific reason rather than a mood. The Committee's own median dot is above its own new target range; the ten-year Treasury has gone from 4.75 to 5.18 percent in under a month; and the observable cost of having waited four weeks on this loan is already $10,171 over ten years. Against that, the buyer's genuine advantage is on the supply side, where inventory at 4.9 months is the loosest in over a decade and new-home median pricing is below existing-home median pricing. If someone were deciding today on the published figures alone, the defensible framing would be to negotiate against inventory rather than to speculate on the rate, and to price the fifteen-year option honestly before dismissing it — because the $289,131 it saves dwarfs every basis point argument on this page, and the reason to reject it should be that the payment does not fit, not that the rate did not make the news.

Five figures to write down before signing anything

  1. The Thursday number, not a daily one. Freddie Mac publishes at noon Eastern each Thursday. Any rate quoted to you between Thursdays is a lender's price, not a national average, and comparing the two is comparing different things.
  2. Your payment at plus 50 basis points. On this loan that is roughly $115 a month. If that figure breaks the budget, the lock period is the most important term in the contract, not the rate.
  3. The bundled owner cost, not just P&I. The federal median for mortgaged households is $2,035 a month including taxes, insurance and utilities. Get your own three numbers before the offer, not after.
  4. Months of supply in your market. The national figure is 4.9 months for existing homes and 8.5 for new. Above six months is conventionally a buyer's market, and it is the only variable on this list you can negotiate against.
  5. The fifteen-year payment, calculated once. Even if you reject it, knowing it costs $684 more a month and saves $289,131 changes how you think about the thirty-year loan you do sign.

Tip: Every figure in the table above is an amortisation calculation you can and should redo with your own loan amount rather than the national median, and doing it on paper once tends to stick better than a web form. A financial calculator handles the payment and the payoff schedule in one place. (These are Amazon Associate links — we may earn a small commission on qualifying purchases.)

Related on this site

Institutions these figures were read from

This page summarises published institutional figures and is not personalised financial advice; it is written by no one holding a mortgage licence, a real-estate licence or a financial-planning credential. Rates, indexes and federal limits change, and every figure here carries the date it was read.

Comments

Popular posts from this blog

USDA Food Budget for Two Adults (2026): Weekly and Monthly Costs

Lowering Monthly Housing Costs — Part 2: Escrow, Taxes, PMI and Insurance

Two 2026 Recalls, One Household Ledger: What Registering Your Products Is Worth Over Five and Ten Years