What’s changed: Updated October 15, 2026: the 30-year held at 7.40%, and September inflation rose to 3.7%. Earlier, October 10: the 30-year fixed reached 7.40%, the highest since November 2023. Full update history.
The average US 30-year fixed mortgage rate stood at 7.40% in the week ending October 15, 2026, according to Freddie Mac, unchanged from the week before. That is the highest level since November 2023, and the survey first crossed 7% only three weeks earlier, for the first time since January 2025. The Federal Reserve raised its policy rate on September 16, its first move of 2026, so it is tempting to connect the two. The link is not the one most people assume. Mortgage rates had been climbing for weeks before the decision, and on the day itself the benchmark they follow barely moved. Fixed mortgage rates do not follow the Fed’s rate at all. They follow the 10-year Treasury yield, and that yield responds to what investors expect the Fed and inflation to do next, not to what the Fed did last. This guide explains what actually sets your mortgage rate, why it has risen, and what the move does and does not mean if you are buying or refinancing.
Educational, not financial advice. Savvy Investor Guide is published by The Savvy Investor Ltd. We are not financial advisers, a mortgage broker, or a registered investment adviser, and nothing here is personalized advice. Mortgage and refinancing decisions depend on your own circumstances. For advice on your situation, speak to a licensed loan officer or a qualified, fee-only financial adviser.
What this article covers: why the US 30-year fixed mortgage rate crossed 7% in September 2026, why fixed mortgage rates track the 10-year Treasury yield rather than the Federal Reserve’s policy rate, the role of the “spread” between the two, what is pushing rates up now, why it kept rising in October, what the Fed’s September rate increase does and does not change, and what all of that means for buyers and refinancers.
What it does not cover: whether you personally should buy, wait, lock, or refinance. That depends on your finances and your timeline, and a licensed professional is the right person for it.
In short
- The US 30-year fixed rate held at 7.40% in the week ending October 15, 2026 (Freddie Mac), unchanged on the week and the highest print since November 2023. It crossed 7% three weeks earlier, for the first time since January 2025. The 15-year was 6.71%.
- The Fed raised its rate by a quarter point on September 16, its first move of 2026. By then the 30-year had already risen about a full point from 5.98% in February, almost all of it while the Fed held. Fixed mortgages track the 10-year Treasury yield plus a spread, not the Fed’s overnight rate.
- That spread is about 2.1 percentage points today, down from a post-2022 peak near 2.4 points but still a little under half a point above its long-run norm. It has moved little all year, which is why the 10-year is doing nearly all the work.
- The reason rates rose through the summer was an oil shock: renewed conflict around the Strait of Hormuz pushed crude back above $84 (and briefly through $90 by late July) and lifted inflation expectations, which pushed Treasury yields, and mortgage rates, up. Since August the driver has been US jobs and inflation data, and since late September the Fed’s signals about a second hike. The 10-year closed at 5.18% on September 24, its highest since July 2007, and at 5.31% on October 5.
- At 7.40% a $400,000 loan costs about $2,770 a month in principal and interest, unchanged from the week before and about $302 more than a year ago at 6.27%, or about $3,600 a year. That year-over-year comparison turned in early September; through the summer the going rate was still cheaper than a year earlier.
- Waiting for the Fed to change course may not lower your mortgage rate. The 10-year Treasury and the inflation outlook matter far more.
The headline number, in perspective
Freddie Mac’s weekly survey put the average 30-year fixed rate at 7.40% for the week ending October 15, 2026, unchanged from the week before, when it had risen from 7.28%. The 15-year fixed averaged 6.71%, down from 6.73%. Two things about that number are worth stating precisely, and both come from Freddie Mac’s own published history rather than from coverage of it. It is the highest print since November 16, 2023, when the survey stood at 7.44%. And it is almost a point and a half above the year’s low of 5.98% in February. Most of that came in the last six weeks. The survey read 6.71% in the week ending September 3, then 6.76%, 6.95%, 7.03%, 7.28%, 7.40% and 7.40% again, about seven tenths of a point after a summer spent moving in hundredths. It crossed 7% in the week ending September 24, the first reading above that line since January 2025.
One quirk worth knowing before you panic or celebrate at a different number: several “current” mortgage rates circulate for the same week and they do not match. In mid-September, for example, Freddie Mac’s weekly survey said 6.95% for the week ending September 17. The Mortgage Bankers Association’s separate weekly survey, which measures the contract rate on actual applications rather than a lender-submitted average, had the 30-year at 7.12% for the week ending September 18, up from 6.97%, as CNBC reported. That is almost the same week and 17 basis points higher. Neither is wrong; they use different samples, different fee assumptions and slightly different weeks. Freddie Mac’s PMMS is the figure most news coverage cites, so it is the one we lead with here. If your lender quotes something 10 to 20 basis points higher, that gap is methodology, not a mistake.
It also helps to size the move in dollars rather than headlines. On a $400,000 loan over 30 years, principal and interest work out roughly as follows:
| Rate | Monthly P&I on $400,000 | vs 7.40% |
|---|---|---|
| 6.00% | about $2,398 | $372 a month less |
| 6.27% (a year ago) | about $2,468 | $302 a month less |
| 7.03% (week ending September 24, 2026) | about $2,669 | $101 a month less |
| 7.28% (week ending October 1, 2026) | about $2,737 | $33 a month less |
| 7.40% (weeks ending October 8 and 15, 2026) | about $2,770 | baseline |
Principal and interest only; excludes property taxes, insurance, PMI and HOA dues. Illustration, not a quote.
One line in this article flipped in early September, and the gap has widened every week since. Through the summer this section told you the headline was real but that the going rate was still a little cheaper than a year before. That stopped being true in the week ending September 3, when the gap was about $55 a month. Three weeks later it was about $193. The year-ago comparison for the week ending October 15, 2026 is Freddie Mac’s print of a year earlier, 6.27%. At 7.40% a $400,000 loan now costs about $302 a month more than the same loan a year ago, or roughly $3,600 over a year.
Affordability had been improving for most of the year, as incomes rose while rates drifted sideways, and six weeks do not undo all of that. But the direction of travel matters for anyone who has been waiting: the wait has cost money rather than saved it, and since February’s low of 5.98% that cost runs to almost a point and a half.
Why a Fed decision barely moves your mortgage rate
Start with the most common misunderstanding: that the Federal Reserve “sets” mortgage rates. It does not. The Fed sets the federal funds rate, the overnight rate at which banks lend reserves to each other. That rate drives short-term, floating borrowing: credit cards, home equity lines, and the initial pricing of adjustable-rate mortgages. It does not set the rate on a 30-year fixed loan.
A 30-year fixed mortgage is a long-term loan, so its rate is built on a long-term benchmark: the 10-year Treasury yield. Why the 10-year? Because homeowners typically move or refinance within about a decade, a 10-year bond has roughly the same effective life as the average mortgage. That Treasury yield is the bond market’s best guess at three things: where the Fed’s rate is likely to average over the coming years, how much inflation investors expect over that horizon, and a “term premium” they demand for tying up money for a decade.
This is why a single Fed meeting usually does little to mortgage rates. By the time the Fed acts, the bond market has already priced in what it expected. Lenders move mortgage pricing ahead of the announcement, based on the expectation, not the event. September 2026 showed it cleanly. The Fed raised its target range by a quarter point to 3.75% to 4% on September 16, its first move of the year. On the Treasury’s daily curve the 10-year closed at 5.00% the day before, 5.01% on decision day and 4.94% the day after, lower than before the increase. The mortgage survey had already risen in the weeks leading up to it. What moves rates is a change in expectations, especially about inflation and the next decision. Over the year as a whole the point is starker: the 30-year mortgage climbed from 5.98% in February to 7.40% in early October, against a single quarter-point increase from the Fed. For the Fed’s own stance and what it signals next, see our guide to why a 2026 Fed rate hike became the base case.
The 10-year Treasury, the spread, and where 7.40% comes from
If mortgages track the 10-year Treasury, why is the mortgage rate so much higher than the Treasury yield? Freddie Mac’s survey averages the rates lenders offered from the previous Thursday through Wednesday. For the survey of October 8, whose window ran from October 1 to 7, the 10-year averaged about 5.28% on the Treasury’s own daily par yield curve. The gap between that and 7.40%, roughly 2.1 percentage points, is called the spread, and it holds most of the answer.
Fannie Mae’s research breaks that spread into two parts. The first is the primary-secondary spread: the origination and servicing costs and profit that lenders add on top of the price of the mortgage bond. The second, and the one that swings the most, is the secondary spread: the extra yield investors demand to hold a mortgage-backed security (MBS) instead of a risk-free Treasury of the same maturity. From 1995 to 2005 those two pieces added up to about 1.67 percentage points. After 2022 the total ballooned to about 2.41 points on average, and at the peak of the 2023 banking stress it reached roughly 3.0 points.
Why do investors demand that extra yield on a mortgage bond? Two reasons. One is prepayment risk. Unlike a Treasury, a mortgage can be paid off early, without penalty, whenever the borrower refinances or sells. That is bad for the investor: when rates fall, borrowers refinance and hand the money back exactly when it can only be reinvested at lower yields; when rates rise, borrowers stay put and the investor is stuck holding a below-market loan. Investors charge for that one-sided deal, and they charge more when interest rates are volatile, because volatility makes the timing even less predictable. The other reason today is the Federal Reserve’s balance sheet. During the pandemic the Fed was a huge buyer of mortgage bonds, which held the spread down. It has since been shrinking that portfolio, so private investors have to absorb more mortgage supply, and they demand a higher yield to do it.
The encouraging part is that at about 2.1 points, today’s spread is narrower than the 2.4-point post-pandemic average and far below the 3-point crisis peak. It has been slowly normalizing as rate volatility eased and the market digested the Fed’s pullback. The less encouraging part is that it is still a little under half a point wider than the long-run norm, which is worth roughly $120 a month on a $400,000 loan. If that spread ever returns to its historical average, mortgage rates could fall even if Treasury yields do not move at all. That is one of the few realistic paths to lower rates that does not depend on the economy weakening.
Why mortgage rates are rising now: the data and the Fed’s next move
Through the summer of 2026 the answer was oil. A ceasefire reopened the Strait of Hormuz in mid-June, the chokepoint through which about a fifth of the world’s crude passes, and prices fell sharply. Then the ceasefire collapsed, Brent went from under $70 a barrel to above $84 in a little over a week and briefly through $90 by late July, and Treasury yields rose with the inflation expectations that came with it. That episode is the mechanism in its plainest form: something with no connection to the Federal Reserve moved the 10-year Treasury, and the mortgage rate followed.
By September the input has changed and the mechanism has not. The driver now is US economic data and what it implies for Fed policy. On September 4 the Bureau of Labor Statistics reported that nonfarm payrolls rose 162,000 in August against a consensus near 53,000, with unemployment steady at 4.1% and both June and July revised upward. A labor market running three times as hot as forecast is the kind of print that argues against rate cuts and for higher yields.
Then compare what that report did to the Fed odds with what it did to the 10-year. It moved the market’s view of the Fed a long way: futures-implied odds of a September hike went from roughly 49% the day before to about 58% after it, on CME’s FedWatch tool as reported by CNBC. It barely moved the 10-year Treasury. On the Treasury’s own daily yield curve the 10-year closed at 4.77% on September 3 and 4.78% on September 4, one basis point across the release, while the 30-year went the other way, 5.25% to 5.24%.
Watching only the Fed headlines that week, you would have expected mortgage rates to jump. They rose five basis points on the survey, and the reason sits in the slow drift of the 10-year up from 4.75% at the end of August rather than in anything that happened on the morning of the jobs report.
Then the Fed acted, and the pattern repeated. On September 16 the committee decided “to raise the target range for the federal funds rate by 1/4 percentage point to 3-3/4 to 4 percent”, and its median projection put the rate at 4.1% at the end of 2026, which implies one more increase this year. The 10-year closed lower the next day. Markets had expected the decision for weeks.
What moved the 10-year was the week that followed. On September 23 Governor Michael Barr, in a speech on housing costs, said that “In my base case, further policy adjustments are likely to be needed to ensure inflation comes down to target in a timely fashion.” The same day an auction of five-year Treasury notes met weak demand, and the 10-year rose from 4.96% to 5.11%, CNBC reported. On September 24 New York Fed President John Williams said it was “reasonable” to expect another hike by the end of the year, and the 10-year closed at 5.18%, its highest close since July 2007 on the Treasury’s daily curve. Futures-implied odds of an October hike rose from about 55% on September 22 to 77.5% on September 24 on CME’s FedWatch tool, as reported by CNBC.
The Fed’s rate did not change that week. Expectations of the next decision did the work, and that is the main channel through which the Fed reaches a fixed mortgage at all.
The data since then has pointed the same way. On September 30 the Bureau of Economic Analysis reported that prices on the PCE measure, the one the Fed targets, rose 3.4% in the year to August, or 3.0% excluding food and energy. On October 2 the Bureau of Labor Statistics reported that payrolls rose 29,000 in September with unemployment at 4.2%, and it revised July and August down by a combined 60,000. A slower month for hiring did not pull yields down. The 10-year closed at 5.28% that day, four basis points higher than the day before.
Then the Fed spoke again. The minutes of the September meeting, published on October 7, recorded that “most participants assessed that another increase in the target range for the federal funds rate would likely be appropriate by year end.” The next day Governor Christopher Waller said: “If the economic data continue to come in as expected, I anticipate additional hikes to support a timelier return of inflation to our 2 percent goal.” He put the futures market’s odds of at least one more increase by the December meeting at 85%. Neither moved the 10-year much. It closed at 5.31% on October 5, before either, and at 5.22% on the day Waller spoke. The mortgage survey, which averages a week of lender quotes, went from 7.03% to 7.28% to 7.40% over those two weeks.
September consumer prices, published on October 14, came in higher. The Bureau of Labor Statistics reported prices up 3.7% over the year, from 3.4% in August, and 2.6% excluding food and energy, from 2.4%. The 10-year did not jump on the news: it closed at 5.19% that day, two basis points lower than the day before, on the Treasury’s daily yield curve. The committee decides next on October 28.
What it means if you are buying or refinancing
None of the above tells you what to do; it tells you what you are working with. A few practical points follow from the mechanics.
- Waiting for the Fed to change course may not help. Because mortgage rates track the 10-year Treasury and already price in expected Fed moves, a decision that markets have anticipated can pass without moving mortgage rates at all, as the September increase did. Timing a home purchase to a Fed meeting is timing the wrong thing.
- Rate locks and float-downs. A rate lock fixes your quoted rate for a set window while you close. A float-down is an optional add-on, usually costing between 0.25% and 1% of the loan, that lets you take a lower rate once if the market falls before closing. Whether the fee is worth it depends on how far away closing is and how volatile rates are.
- Points and buydowns. Paying “discount points” (each is 1% of the loan) typically buys the rate down by around a quarter point, permanently. Builders and sellers are also using buydowns heavily: a March 2026 survey found 64% of homebuilders were offering incentives such as rate buydowns and closing-cost credits, which can matter more than the list price.
- Adjustable-rate mortgages. An ARM usually starts below a 30-year fixed, which is tempting when fixed rates are high, but that gap has narrowed through 2026, and the share of buyers choosing ARMs has fallen from around 21% early in the year to roughly 8% by mid-year. A smaller discount means less reward for taking on the risk of a future rate reset.
- The “date the rate, marry the house” idea, with a caveat. The industry slogan is that you commit to the home and refinance the rate later if it falls. It is a fair point that a house you love outlasts any rate, but refinancing is not free. Most 2025 to 2026 buyers need at least a 0.75-point drop before a refinance recovers its closing costs in a reasonable time. Today’s rate is about 113 basis points above where it stood a year ago, so most existing borrowers are further from that break-even than they were, and for most of them a refinance does not pay right now. The strategy works best when the payment already fits your budget today and any future refinance is a bonus, not the plan.
For the refinancing math specifically, the rule of thumb is simple: divide your closing costs by your monthly savings to get the break-even in months. Save $100 a month against $3,000 of costs and you break even in 30 months, so it only pays if you will keep the loan longer than that. For the broader housing picture, existing-home sales ran at a 3.98 million annual pace in August, before September’s rise, the first reading below 4 million since June 2025. Supply stood at 4.9 months and the median price was $429,100, up 1.6% from a year earlier. As the National Association of Realtors’ chief economist Lawrence Yun put it, “Mortgage rates and home sales move in opposite directions, so it’s not surprising to see a mild dip in home buying activity due to high mortgage rates.”
What forecasters expect
Forecasts are not facts, and the people who make them disagree, but they help calibrate expectations. The more useful thing to watch is the direction of revision rather than the level, and through 2026 the revisions have all gone one way: up.
Fannie Mae’s Economic and Strategic Research group publishes a monthly housing forecast. Its September 2026 edition puts the 30-year fixed at 6.7% for the third quarter of 2026 and 6.8% for the fourth, then 6.7% in every quarter of 2027. On an annual average basis that is 6.5% for 2026 and 6.7% for 2027. The forecast notes that its interest rate projections are based on rates as of August 31, 2026, so it was written before the September jump. At 7.40% in mid-October, the survey is already 0.6 of a point above the forecast for the fourth quarter.
That is a marked change from where the same forecaster stood a year ago. Its late-2025 editions had rates ending 2026 near 5.9%, and as recently as July 2026 the projection for the rest of the year was closer to 6.4%. The forecasts have moved up all year, and the climb since September has passed even the higher ones.
For anyone waiting, the industry’s central case is no longer that rates drift down to something beginning with a 5. It is that they sit in the high 6s into 2027, and the market is currently above that. Forecasts have been wrong in both directions before and will be again, but a reader who postponed a decision in 2025 on the strength of one has since watched it move away from them for four consecutive quarters.
Common questions
- If the Fed cuts rates, will my mortgage rate fall? Not necessarily. Fixed mortgage rates track the 10-year Treasury, which moves on inflation and growth expectations. If a Fed cut is already expected, yields may already reflect it, and mortgage rates may not fall when it happens. A cut prompted by a surprise economic slowdown could pull yields and mortgage rates down; a cut the market saw coming may not. The same works in reverse: the 10-year closed lower the day after the September 2026 increase.
- What is the “spread” and why does it matter? It is the gap between the 10-year Treasury yield and the mortgage rate, about 2.1 points in early October 2026. It covers lender costs and the extra yield investors want for holding mortgage bonds, which carry prepayment risk. When the spread is wide, as it still is versus history, mortgage rates are higher than Treasury yields alone would imply.
- Why does my lender quote a different rate than 7.40%? The 7.40% is Freddie Mac’s weekly survey average. Daily trackers and individual lender quotes vary with timing, your credit score, loan-to-value, points, and property type. A quote 10 to 20 basis points away from the survey average is normal. If a lender or servicer treats you unfairly, our guide to filing a CFPB complaint explains the federal route.
- Should I take an ARM to get a lower rate? That is a personal trade-off, not a recommendation we can make. An ARM starts lower but can reset higher later. The starting discount over a fixed loan has shrunk in 2026, so there is less reward for the reset risk than there was; whether it suits you depends on how long you plan to keep the loan.
- Are rates going to keep rising? No one knows. In the near term they track the 10-year Treasury, and the 10-year is responding to US inflation data and the Fed’s signals about a second increase. September consumer prices, published October 14, rose 3.7% over the year, up from 3.4%, ahead of the Fed’s next decision on October 28, and the minutes of its September meeting say most officials expect another increase by year end. The industry’s own forecasts have been revised upward through 2026, not downward, and the survey rate is now above them, which matters if you have been waiting for a fall.
Savvy Investor’s take
Keep one idea if nothing else: for a fixed mortgage, the Federal Reserve is mostly a distraction. The number to watch is the 10-year Treasury yield, and behind it the inflation outlook and the mortgage spread. September showed it twice. The August jobs report moved the market’s odds of a September hike about nine points in a day, and the 10-year one basis point. Then the Fed actually raised rates, and the 10-year closed lower the next day. If the Fed were setting your mortgage rate, neither of those would look the way it did.
Two things have changed since July, and both cut against the patient reader. The headline rate is 7.40%, the highest since November 2023, and more expensive than a year ago by about $302 a month on a $400,000 loan. And forecasters, who have revised upward all year, are now below the market rather than above it: their central case is the high 6s through 2027, and the survey has already passed it. Neither is a reason to hurry a decision that should not be hurried. Both undercut the idea that a lower rate arrives if you simply wait.
For anyone buying, that still argues for focusing on a payment that works at today’s rate, treating any future refinance as a bonus rather than the plan, and ignoring the temptation to time the Fed. For anyone refinancing, the math is unchanged and, for most people right now, it does not add up: the gap to break-even is wider than it was a year ago.
Sources
- Freddie Mac Primary Mortgage Market Survey (weekly 30-year and 15-year fixed rate averages)
- Fannie Mae: What Determines the Rate on a 30-Year Mortgage? (the Treasury benchmark and the mortgage spread)
- National Association of Realtors: Existing-Home Sales, August 2026 (sales, prices, inventory)
- CNBC on the September 23 Treasury selloff and the MBA weekly survey (September 23, 2026)
- US Bureau of Labor Statistics: Consumer Price Index
- US Bureau of Labor Statistics: The Employment Situation (the September 2026 report, published October 2)
- US Bureau of Economic Analysis: Personal Income and Outlays, August 2026 (the PCE price index)
- Bankrate: How the Federal Reserve affects mortgage rates
- Freddie Mac: Primary Mortgage Market Survey, full weekly history (the primary series behind every rate figure in this article)
- US Department of the Treasury: Daily Treasury Par Yield Curve Rates (the 10-year and 30-year readings)
- Fannie Mae: Economic and Strategic Research housing forecast (the quarterly 30-year projections)
- Federal Reserve: FOMC statement, September 16, 2026, Governor Barr, September 23, 2026, minutes of the September meeting (published October 7, 2026), Governor Waller, October 8, 2026 and FOMC calendars (the October 27 to 28 meeting)
- CNBC: New York Fed’s Williams on another hike by year-end (September 24, 2026, including CME FedWatch October odds)
- Savvy Investor Guide: why a 2026 Fed rate hike became the base case, and what Warsh’s first FOMC means for savers and borrowers.
- For the UK comparison, where fixed rates rose the same week for related but separate reasons: seven UK lenders raised mortgage rates over one weekend.
This article is for general information only and is not personalized financial advice. Mortgage rates cited are as of October 15, 2026 and Treasury yields as of October 14, 2026; both change daily. Savvy Investor Guide and The Savvy Investor Ltd are not financial advisers or a registered investment adviser. Mortgage and refinancing decisions depend on your own circumstances; consider speaking to a licensed loan officer or a qualified financial adviser. Fact-checked and rates refreshed October 10, 2026; rates, inflation and yields updated October 15, 2026.
Update history: October 15 back to July 30, 2026
Update, October 15, 2026: Freddie Mac’s survey of October 15 held the 30-year at 7.40%, unchanged on the week, with the 15-year at 6.71%; a year earlier the 30-year was 6.27%, so the same $400,000 loan now costs about $302 a month more than a year ago. September consumer prices rose 3.7% over the year, up from 3.4% in August, and the article now says so. The 10-year Treasury closed at 5.19% that day, two basis points lower. The notes below are kept as they were published.
Update, October 10, 2026: The 30-year fixed averaged 7.40% in the week ending October 8, the highest since November 2023, and the article has been refreshed again: the headline figures, the payment table, the spread, and the inflation, jobs and Fed news since late September. The notes below are kept as they were published.
Update, September 25, 2026: The 30-year fixed crossed 7% in the week ending September 24, and the article has been refreshed throughout: the headline figures, the payment table, the spread, the Fed’s September 16 increase and the reaction to it, and the latest forecasts. The two notes below used to sit at the top of the article, and their facts are now in the main text. The older boxes that follow them are kept as they were published.
Updated September 18, 2026. The Fed raised its target range by a quarter point to 3.75% to 4% on September 16, and its median projection now puts the rate at 4.1% at the end of 2026, which implies one more increase this year. Mortgage rates moved first. Freddie Mac’s 30-year fixed averaged 6.95% in the week to September 17, up from 6.76% a week earlier, with the 15-year at 6.26%. The 10-year Treasury, the benchmark this article tracks, closed at 5.01% on decision day and 4.94% the next day. Sources: Federal Reserve statement, projections, Freddie Mac PMMS, Treasury daily yield curve.
Updated September 8, 2026. The 30-year fixed reached 6.71% in the week ending September 3, the highest print of 2026, and for the first time since this article was written it costs more than a year ago rather than less. The Fed meets September 15 and 16 with a hike in play, so the original title, which assumed a Fed on hold, has been changed. Figures and forecasts are refreshed throughout.
Update, 21 August 2026: the long end hit a nineteen-year high and gave it back within two days. The Federal Reserve H.15 series has the 30-year Treasury at 5.31% on 17 August, then 5.28% on 18 August and 5.19% on 19 August. On this article’s own check of the Treasury curve year by year, no year from 2008 to 2025 exceeded 5.11% and 2007 peaked at 5.35%, so the 17 August reading was the highest in nineteen years and four basis points under the 2007 peak. It did not hold. By 19 August the 30-year sat below the 5.20% on 29 July that the folded note below flagged as a 2007-era high.
Mortgage rates moved the other way, which is the point of this article. Freddie Mac’s survey for the week ending 20 August puts the 30-year fixed at 6.65%, down from 6.67% the week before, with the 15-year at 5.95%. So the 30-year Treasury round-tripped 12 basis points in three sessions while the headline mortgage rate drifted down two. The 10-year, which is the maturity that actually prices a fixed mortgage, went 4.72% on 17 August to 4.71% and then 4.65% on 19 August. Against the Freddie print that is a spread of almost exactly 2.0 percentage points, close to the spread this article described in July.
What to take from it. A nineteen-year high in the 30-year Treasury made headlines on 17 August, had unwound by the 19th, and never reached the mortgage market at all. The drivers cited for the spike were a July federal deficit of $432.3 billion, inflation still above target, and firmer oil. The July FOMC minutes, released 19 August, record that “Many participants assessed that policy tightening would likely be necessary if inflation did not decline”, which is consistent with a long end pricing higher-for-longer rather than cuts. The mechanics this article sets out are unchanged: watch the 10-year rather than the 30-year, and wait for several consecutive prints before reading a move as a trend.
Update, 17 August 2026: the climb has flattened. Freddie Mac’s survey has now printed the 30-year fixed at 6.66%, 6.69% and 6.67% across the weeks ending 30 July, 6 August and 13 August, with the 15-year at 6.04%, 6.01% and 5.96%. Three weeks inside four basis points is not a trend in either direction, and after the run that carried the headline up from 6.55% in mid-July, the flattening is the thing that has changed. The mechanics this article describes are unaffected: what has paused is the climb, not the link between the 10-year Treasury and your mortgage rate.
The box below asked what the July CPI release would do. It landed on 12 August, and the surveys either side of it moved the 30-year by two basis points. This page has read a short move as a trend once already, so the honest reading of one small print is that it tells you very little on its own. What would be informative is several consecutive prints leaving the 6.6% band, in either direction.
Update, August 8, 2026: the labor market cracked, and it changes the question this article is asking. The July employment report, released August 7, showed nonfarm payrolls fell by 23,000 against a Dow Jones consensus of roughly +83,000. The revisions were larger than the miss: May was cut by 66,000 and June by 37,000, so that “employment in May and June combined is 103,000 lower than previously reported”. The unemployment rate edged down to 4.1% from 4.2%, but on a shrinking labor force rather than on hiring, and average hourly earnings growth slowed to 3.2% year over year, the weakest since May 2021. Government payrolls fell 53,000. Source: CNBC.
What it did to rates. Treasury yields fell across the curve on the print, the two year to 4.20% from 4.22% and the ten year to 4.65% from 4.67%. Because the 30 year mortgage tracks the ten year Treasury rather than the federal funds rate, a sustained move of that kind is the actual mechanism by which mortgage rates would come down. One print is not a sustained move, and the Freddie Mac survey rate quoted above was set before this data landed.
One deliberate omission. You will see precise probabilities quoted for what the Fed does in September, and this article is not going to repeat one, because the sources do not agree with each other. Some report the odds of a rate hike falling; at least one wire report describes the odds of a rate cut falling on the same data, which does not follow from a weak jobs report. The direction is not in doubt and it is away from a near term hike. The number is not reliable enough to print. The next real test is the July CPI release on August 12.
Update, August 7, 2026: the survey caught up, and it moved far less than the bond market suggested it would. Freddie Mac’s survey for the week ending August 6 puts the 30-year fixed at 6.69%, up from 6.66%, with the 15-year down to 6.01% from 6.04%. The box below predicted this release would carry the late-July jump in long yields; it carried three basis points of it, because the long end went back the other way in between. On the Treasury’s daily curve the 30-year ran 5.23%, 5.18%, 5.17% and 5.22% across August 3 to 6, so the survey week closed roughly where it opened. The long end has now round-tripped twice in two weeks, which is a caution against reading any two-day move as a trend.
The Fed side got louder and the mortgage side still did not care. On August 5 Governor Lisa Cook, who voted to hold in July, said in Anchorage that “I am prepared to act by raising rates, if necessary”, with the Fed’s preferred inflation gauge running at 3.7 percent over the twelve months to June. A member who voted to hold moving toward tightening is a broader signal than three dissenters, and the 30-year mortgage answered it with three basis points. One note on the numbers: the Mortgage Bankers Association’s own survey put its 30-year at 6.81% for the week ending July 31, per HousingWire, twelve basis points above Freddie Mac’s figure for a week ending six days later. Both are right; they sample different lenders on different days, as the section on competing trackers below explains.
Update, July 31, 2026: rates rose again, and this week the data argued the other way. Freddie Mac’s survey for the week ending July 30 puts the 30-year fixed at 6.66%, up from 6.58% the week before, with the 15-year at 6.04% from 5.96%.
What makes this the sharper illustration is what the economy was doing while that happened. The Bureau of Economic Analysis reported that “Real gross domestic product (GDP) increased at an annual rate of 1.5 percent in the second quarter of 2026”, down from 2.1 percent in the first quarter. The Fed’s preferred inflation gauge cooled in the same week: on the BEA’s personal income and outlays release, “From the preceding month, the PCE price index for June decreased 0.1 percent”, with the annual rate at 3.7 percent and the core rate at 3.3 percent.
Slower growth and a monthly fall in prices is the combination that is supposed to pull long rates down. It did not. On the Treasury’s daily curve the 30-year went from 5.20% on July 29 to 5.21% on July 30, and the 10-year from 4.67% to 4.68%, while the 2-year sat at 4.23%. The steepening described below carried on rather than unwound.
It kept going the next day. On July 31 the same curve put the 30-year at 5.27% and the 10-year at 4.75%, six and seven basis points higher again, with the 2-year at 4.28% and the 20-year at 5.28%. So the two sessions after the soft GDP and PCE prints added eleven basis points to the 30-year rather than taking any off. That matters for the week ahead as much as the week behind: Freddie Mac’s survey lags the bond market by several days, so the 6.66% quoted above reflects the curve as it stood before this move, and the next release on August 6 is the one that will carry it.
That is this article’s argument stated as plainly as it can be put. The Fed held on July 29, three of its officials wanted to tighten, the following week’s data pointed the other way, and the 30-year mortgage still went up eight basis points. None of those things set the rate a buyer was quoted this week. The long end of the Treasury curve did, and that is the leg the Fed does not control.
Update, July 30, 2026: the Fed held, and the 30-year Treasury went to its highest level since 2007 anyway. This is the argument of this article compressed into a single day. The FOMC left its target range at 3-1/2 to 3-3/4 percent on July 29, with three officials dissenting in favor of a quarter-point rise. Long rates went up regardless.
On the US Treasury’s own published daily yield curve, comparing July 28 with July 29:
- 2-year: 4.26% to 4.22%, down 4 basis points
- 10-year: 4.61% to 4.67%, up 6 basis points
- 30-year: 5.09% to 5.20%, up 11 basis points
That 5.20% is the highest the 30-year has been on the Treasury’s daily curve since 2007. No year between 2008 and 2025 recorded a reading above 5.11%; the last year it stood above today’s level was 2007, when it peaked at 5.35% on June 12.
The shape of the move matters more than any one number, and it is the cleanest illustration of this article’s point yet. The short end fell while the long end jumped. That is the curve steepening, and it says two things at once: traders trimmed their expectations of a near-term hike, and at the same moment demanded more to lend the government money for thirty years. Since your mortgage rate is priced off the 10-year Treasury rather than the Fed’s policy rate, the leg that moved against a homebuyer on Wednesday was the leg the Fed did not touch. Freddie Mac’s latest survey, for the week ending July 23, has the 30-year fixed at 6.58% and the 15-year at 5.96%, up from the 6.55% quoted below.