The 10-year hit 5.29%. Inflation bets added 12 basis points of it.

The 10-year Treasury rose 85 basis points in Q3 to a 2002 high. Treasury's own TIPS curve puts 73 of them on real yields and only 12 on inflation.

The Editors · Markets · 8 min read · Oct 1, 2026

The 10-year hit 5.29%. Inflation bets added 12 basis points of it.

By the numbers

5.29%10-year Treasury yield at the September 30 close, up from 4.44% on June 30

2.93%10-year real (TIPS) yield, the highest close since November 2008

+12 bphow much of the 85 bp quarterly rise came from inflation compensation

0.89%NY Fed model term premium on September 30, near its highest since 2014

22 of 27 bpshare of the post-hike rise the NY Fed model assigns to term premium

The 10-year Treasury yield closed at 5.29% on September 30, and touched 5.342% intraday on October 1, the highest since early 2002. Most coverage blames oil, the war in the Middle East and inflation. Treasury's own numbers give inflation fears a small part of the move. Of the 85 basis points the 10-year added between June 30 and September 30, the inflation compensation priced into TIPS accounts for 12. The other 73 are a rise in the real yield, which closed at 2.93%, its highest since November 2008.

So what the market priced in the quarter was a higher Fed path and a higher fee for lending for ten years, with only a little more inflation on top. The New York Fed's term premium model splits the quarter roughly in half between those two. And since the Fed hiked on September 16, it puts 22 of the 27 basis points of the rise on the term premium alone.

That matters if you own bonds, plan to buy a house or are waiting for rates to fall. A term premium doesn't come down because the Fed stops hiking.

What the 10-year did in the third quarter

Treasury's daily par yield curve shows the move plainly. The 10-year sat at 4.44% on June 30 and closed at 5.29% on September 30. Reuters called it the biggest quarterly rise this century for the 10-year. The 2-year went from 4.14% to 4.88% over the same days. The 30-year went from 4.91% to 5.64%.

The Fed moved in the middle of it. On September 16 the FOMC raised the target range by 25 basis points to 3.75%-4.00%, its first hike since 2023, in a 12-0 vote. Sixteen of the 18 officials projected another hike before year-end, and by October 1 overnight index swaps fully priced one. We covered how the market got to that September hike in the trimmed-mean piece.

The inflation data then came in softer. The BEA's August report, out September 30, showed core PCE up 0.2% on the month and 3.0% on the year. Headline PCE was 3.4%. The 10-year still closed at its high that day.

Split one: real yield versus inflation compensation

A nominal Treasury yield is roughly two things added together: the real yield an investor demands, and the inflation they expect to lose along the way. Treasury publishes both halves. The real yield comes from TIPS, the inflation-protected bonds. Subtract it from the nominal yield and what's left is the breakeven, the market's inflation compensation.

Here is the quarter, from Treasury's two curves:

June 30: nominal 4.44%, real 2.20%, breakeven 2.24%.
September 30: nominal 5.29%, real 2.93%, breakeven 2.36%.

85 bp73 bp12 bp

View as table

What moved the 10-year Treasury yield, June 30 to September 30, 2026
label	What moved the 10-year Treasury yield, June 30 to September 30, 2026
Total rise in the 10-year yield	85 bp
Real yield (TIPS)	73 bp
Inflation compensation (breakeven)	12 bp

Source: US Treasury daily nominal and real par yield curves, September 30, 2026

The breakeven rose 12 basis points in a quarter of war and oil headlines. The real yield rose 73. A 10-year real yield of 2.93% hasn't printed on Treasury's real curve since November 24, 2008, and back then it was a crisis spike in a market nobody could trade. In the 2023 selloff the 10-year real yield peaked at 2.52%.

One caveat on this split. Breakevens carry their own distortions: TIPS trade less than nominal Treasuries, and in a selloff that liquidity gap can move the number. It doesn't stretch to explain 73 basis points.

Split two: expected Fed path versus term premium

A real yield can rise for two reasons. Investors may expect short-term rates to stay higher for longer, or they may demand extra pay for tying money up for ten years. The second part is the term premium. Nobody observes it directly, so it has to be estimated.

The New York Fed publishes the most cited estimate, the Adrian-Crump-Moench (ACM) model, daily. It works on a zero-coupon 10-year yield, so its level runs a few basis points off Treasury's par curve. The split it gives for the quarter:

June 30: model yield 4.48%, of which expected short rates 3.97% and term premium 0.51%.
September 30: model yield 5.26%, of which expected short rates 4.37% and term premium 0.89%.

That's 78 basis points in total, about 41 from the expected path and 38 from the term premium. Roughly half and half.

The timing is the part worth reading. On September 16, hike day, the model had the yield at 4.99% and the term premium at 0.67%. By September 30 the yield was 27 basis points higher, and 22 of them came from the term premium. The expected path rose about 6. Once the Fed had acted, the market barely changed its view of where rates were headed. It changed what it wanted to be paid to hold the bond.

The same model printed 0.89% on August 17 and 0.89% in May 2025, and it hasn't been above 1% since June 2014. So 0.89% sits at the top of a range that has held for a decade.

Why the term premium rose

The reasons on the record are about supply and demand for money. Bloomberg's October 1 report lists massive government borrowing, strong growth, and AI infrastructure spending adding to the demand for capital, alongside oil.

That's our read, and it comes with a limit: the ACM model says how much of the yield is term premium. It doesn't say why. Other models, such as the Fed Board's Kim-Wright, give different levels. What the data supports is narrower: in the second half of September, the extra yield came from compensation for holding duration, and only a sliver came from expected inflation.

What it means if you hold bonds or borrow

For a borrower, the pass-through is already here. Freddie Mac's survey on October 1 put the 30-year fixed mortgage at 7.28%, up from 7.03% a week earlier, the highest since November 2023. In September the spread between mortgages and Treasuries was the story, as in our MBS buyback piece. Now the base rate itself is moving.

For a saver weighing a 10-year Treasury, the real yield is the number. At 2.93%, a TIPS buyer locks in inflation plus 2.93% a year for ten years, if held to maturity. A nominal 10-year at 5.29% pays more only if inflation averages under 2.36%. Core PCE runs at 3.0% today. Both are market prices, posted daily on Treasury's site, and you can check them before you buy.

For anyone waiting on a Fed pause to bring yields down, the split is the warning. Expected policy explains about half of the quarter's rise and almost none of the last two weeks. If the term premium holds, a pause at the October 28 meeting would move the 2-year more than the 10-year.

What would prove this read wrong

If the next CPI or PCE prints hot and breakevens jump, inflation would take over the story, and the 12 basis points would look early rather than small. If the term premium falls back to the 0.5-0.6% range of June while the Fed stays on hold, the September move will read as a positioning flush that unwound. The September jobs report and the October 28 FOMC are the next two tests.