🔹 The 30-year just hit a 2004 high, and the Fed isn't why
Stocks set fresh records as Hegseth confirms the Iran ceasefire and oil retreats nearly 4%.
| ◆ |
The Far End of the Curve Is Doing the Talking
The 30-year reached a level last seen in 2004, and the Fed isn't the reason.
The Close
Stocks went almost nowhere on Thursday. The S&P 500 finished 1.90 points lower at 7,704.13, the Nasdaq added 3.34 points, and the Dow slipped 162 points. Bloomberg noted most of its companies fell even so. The real action sat in the bond market, where the 30-year Treasury yield touched 5.501%, a level last seen in June 2004.
Oil climbing again and a week of hot data both played a part. But the shape of the move is what caught my eye. The short end barely budged while the long end jumped, and the same thing happened in Tokyo, London and Frankfurt on the same day.
One Big Thing
Two things decide what a long bond pays. One is where investors think the Fed will set short-term rates over the life of that bond. The other is the extra compensation they want for tying up money through decades they cannot see into. Economists call the second part the term premium. Most days it sits still.
Thursday it moved, and you can see it in the spread between maturities. The 2-year note, which is essentially a wager on the Fed's next few meetings, rose about four basis points to 4.941%. The 30-year bond rose roughly ten, to 5.501%. Nobody believes the Fed will still be raising rates in 2050. Something other than monetary policy is being priced here.
That something is what it costs to lend to a heavily indebted government for thirty years while inflation is unsettled and energy prices keep jumping. Ed Al-Hussainy of Columbia Threadneedle told Bloomberg that "people are running out of superlatives" for the 30-year.
The 2-year is a bet on the Fed. The 30-year is a bet on the country.
Washington has been trying to lean against it. In August the Treasury at least doubled the size of its buyback operations in long-dated bonds, from $2 billion apiece to $4 billion or more, buying back its own debt to steady the long end. Thursday's operation went ahead. Yields rose anyway.
And this is not an American story. Japan's 10-year government bond yield reached its highest since 1996 the same day, while British gilts and German bunds hit multi-year highs. When the price of long money rises everywhere at once, no single central bank explains it.
| 2-year · 4.94% | 2023 |
| 10-year · 5.22% | June 2007 |
| 30-year · 5.50% | June 2004 |
Each bar shows how far back you'd travel to find that maturity yielding what it yielded Thursday. Three years for the 2-year. Nineteen for the 10-year. Twenty-two for the 30-year. The further out the promise, the deeper into history the price has gone.
Source: CNBC Treasury data, Sept. 24, 2026, intraday highs.
The Deeper Read
None of this stays abstract for long. The 30-year mortgage is priced off the long end of the Treasury curve, and Freddie Mac reported Thursday that the average 30-year fixed rate rose to 7.03%, up from 6.95% a week earlier and 6.30% a year ago. That is the first reading above 7% since January 2025. A bond auction in Washington becomes a mortgage quote in Ohio within days.
Here's the part I'd hold in the other hand. The same move that raised the cost of a house handed savers the richest long-dated income in twenty-two years. Locking in roughly 5.5% for thirty years is a genuine offer, and whether it turns out well depends entirely on what inflation averages between now and 2056.
Which is precisely the uncertainty the market is charging for. Buyers demanding 5.5% are not predicting disaster. They are saying they will not lend for three decades on faith alone, at a moment when the government's borrowing is heavy and energy prices keep resetting the inflation math.
Freddie Mac's chief economist framed the backdrop as a solid labor market and an economy growing at a healthy rate. This is not a panic. It is a repricing, which tends to be slower and more stubborn.
What This Means For You
If you own bond funds, this is the week to look up one number: duration. Every fund publishes it, usually on the first page of the fact sheet, and it's a rough estimate of how much the fund falls in price when yields rise by one percentage point. A duration of six means a drop of about 6%. A long-bond fund can run past fifteen.
That single number explains why two holdings both labelled "bonds" can behave nothing alike in a week like this one. Plenty of people discover their conservative sleeve was the most rate-sensitive thing they owned, and they discover it from a statement rather than a fact sheet.
Knowing it isn't a signal to sell anything. Duration cuts the other way too, and the funds falling hardest now will gain hardest whenever yields turn. The point is to be able to predict roughly how your own portfolio behaves before the market shows you.
The Long Game
A 30-year yield is the closest thing markets produce to an opinion about the next generation. It prices the years after the current Fed chair, the current president and the current war are all footnotes. Watching it move is like watching a very long shadow shift by a few inches.
Which is the useful perspective for a week of superlatives. The long bond has been wrong before and will be again. Your own thirty years are the only forecast you're required to have a view on, and that one you plan for rather than predict.
| ◆ |
— Anthony
