The 30-Year Treasury Yield Hit a 19-Year High. Term Premium, Not Fed Bets, Is Why
7 min read
Loading...
7 min read
On 18 August 2026, the 30-year US Treasury yield touched 5.33%, its highest level since before the 2007-2008 financial crisis. That happened in the same year most forecasters started 2026 expecting the Federal Reserve to keep cutting rates. If rates were supposed to come down, why did the longest-dated government bond just pay more than it has in almost two decades?
The answer is a piece of the yield that has nothing to do with what the Fed does next meeting. It is called the term premium, and once you can see it, a lot of confusing bond-market headlines start making sense.
A long-term bond's yield is made of two separate pieces, not one:
The first piece, the expected path, is simply the market's best guess at what short-term interest rates will average out to over the life of the bond. If everyone expects the central bank to cut rates steadily for the next ten years, that piece falls.
The second piece, the term premium, is extra compensation investors demand just for the risk of locking money up that long, separate from where rates end up. Inflation could run hotter than expected. The government could need to sell far more debt than planned. A crisis could force you to sell early at a bad price. None of that is about the rate-cut-or-hike debate; it is about uncertainty over a much longer horizon, and investors want to be paid for carrying it.
This split is not a rough guess. Economists at the Federal Reserve Bank of New York (Tobias Adrian, Richard Crump and Emanuel Moench) built a model, known as the ACM model, that estimates both pieces from real Treasury market data every trading day.
On 30 April 2026, the ACM model put the 10-year Treasury yield at 4.45%, split into a 3.72% expected-rate-path component and a 0.73% term premium, according to Ferrante Capital Advisers' own reading of the New York Fed's data. Over the following months, the New York Fed's own monthly series (compiled by CEIC) showed the premium piece moving on its own: down to 0.667% in May 2026, down further to 0.513% in June, then back up sharply to about 0.80% by 13 August 2026, per MacroMicro's tracking of the same model.
Here is the timeline that makes this concrete:
| Date | What rate expectations were doing | What the 30-year Treasury yield did |
|---|---|---|
| 27 Jan 2026 | CNBC's own Fed survey found forecasters expected the Fed to cut rates two more times in 2026 | - |
| 19 May 2026 | Rate-cut bets still intact | Yield topped 5.19%, highest since before the 2007-2008 crisis |
| 12 Jun 2026 | Bloomberg reported economists pushing their rate-cut timeline out to 2027 | - |
| 31 Jul 2026 | Cut bets fading further | Yield closed at 5.27%, a level not seen since 2007 |
| 18 Aug 2026 | Market pricing for the 16 September 2026 Fed meeting had swung toward a hike, not a cut | Yield topped 5.33%, a fresh 19-year high (CNBC) |
Sources: CNBC (27 Jan, 19 May, 17-18 Aug 2026), Bloomberg (12 Jun 2026 and 17 Aug 2026 reporting on the bond selloff), fedratecalc.com's tracking of market-implied Fed odds.
By early September 2026, the market-implied probability of a 25-basis-point hike at the Fed's 16 September meeting stood at 58%, according to fedratecalc.com, a sharp reversal from the two-cuts-expected mood in January. If the 30-year yield's climb were only about the expected-rate-path piece, it should have tracked that back-and-forth in rate expectations. Instead, it climbed through both phases, the cut-expected phase and the hike-expected phase alike. That persistence is what a rising term premium looks like from the outside.
The New York Fed's own long-run data shows the term premium spent roughly 2015 through 2022 near zero or even negative, an unusually calm stretch. Since 2022 it has moved back into positive territory and stayed there. Two forces are widely cited for that shift: the end of the Fed's bond-buying programs, which had suppressed the premium by soaking up long-dated supply, and inflation that has run above the Fed's 2% target for five straight years, according to Ferrante Capital Advisers' analysis of the current regime.
On top of that, the US government's own borrowing has kept climbing. The US fiscal deficit for July 2026 was its highest monthly total since March 2021, contributing to the same worry investors have about how much long-dated debt the market needs to keep absorbing. A bigger, less certain supply of decades-long bonds is exactly the kind of risk term premium exists to price in.
Indonesia's central bank, Bank Indonesia, held its benchmark BI-Rate at 5.75% at its 18-19 August 2026 policy meeting, its second straight hold, in the first decision led by acting governor Destry Damayanti. BI has already raised that rate by a cumulative 100 basis points this year to defend the rupiah, and Kompas.id's and Bisnis.com's coverage of the same meeting noted the door remains open to a further hike, toward 6.25%, before it is open to a cut. In other words, Indonesian investors are not pricing in near-term rate relief either.
Yet Indonesia's own 10-year government bond (SUN) yield has stayed elevated all year: about 6.04% in January 2026, climbing to roughly 7.2% to 7.3% by mid-August, then easing only slightly to 7.10% on 3 September 2026, down 0.13 percentage points from 7.23% the prior session, according to TradingEconomics' and MacroMicro's tracking of the same benchmark.
Part of that is the same story as the US: Indonesia's government raised its own 2026 bond issuance target to Rp799.53 trillion to fund a wider fiscal deficit, a heavier supply load that adds its own risk compensation on top of whatever BI does with its policy rate. NetWort covered that supply-side mechanism on its own in how Indonesia's bond issuance connects to your yield. This article is about the separate, harder-to-see piece: even once you account for supply and for the BI-rate path, a portion of that 10-year yield is compensation for the sheer uncertainty of holding Indonesian government debt for a decade, not a reaction to next month's rate decision.
If you hold a bond fund, a retail SUN like ORI or SR, or you are weighing a long-tenor deposit against a shorter one, do not assume a future rate cut, from the Fed or from BI, will automatically pull long-term yields down to match. Part of what you are being paid, or what you would pay to borrow, is compensation for risks that have nothing to do with next month's rate decision. That compensation can rise even while everyone agrees rates are headed lower, and it can persist even after they arrive there.
NetWort's Macro Context page tracks the BI rate, the Fed funds rate, and benchmark bond yields side by side, so you can see for yourself whether a given yield move lines up with a rate-expectations story or looks like something else entirely.