NextFin News - Treasury Secretary Scott Bessent's latest attempt to cool what he called a "fever" in the U.S. bond market backfired this week: after the Treasury Department announced a $6 billion buyback of longer-dated debt on Sept. 9, the benchmark 10-year yield jumped to 4.8568%, its highest level since November 2023, as investors judged the intervention too small to matter.
The episode is a test case for the Trump administration's new willingness to manage the bond market directly — and the early verdict is that the market will not be managed. Bessent can smooth liquidity at the margins, but the forces driving the 10-year yield above 4.8% are fiscal, geopolitical, and monetary, and none of them responds to a repurchase operation.
The sequence: a working intervention, then a failed one
The timing matters, because it shows how quickly the Treasury's leverage evaporated. On Aug. 19, the department said it would "at least double" the size of its liquidity-support buybacks for longer-dated nominal coupon securities — the 10-to-20-year and 20-to-30-year sectors — lifting the per-operation ceiling from $2 billion to "at least $4 billion," effective Sept. 9 and running through Nov. 4. The market liked the surprise: the 10-year note yield fell 5.7 basis points to 4.647% and the 30-year bond dropped 9 basis points to 5.196%.
Three weeks later, the follow-through disappointed. At 11 a.m. ET on Sept. 9, the Bureau of the Fiscal Service said it would purchase up to $6 billion in 10-year notes and 20-year bonds in the next day's operation, with maturities ranging from February 2037 to August 2046. That is triple the size of the last long-dated buyback — but short of what some investors had hoped for. The market's verdict was swift and one-directional. The 10-year yield rose 3.06 basis points to 4.835% after touching 4.8568%; the 30-year yield climbed 2.06 basis points to 5.2846%, having earlier reached its highest level since 2007; and the 2-year note yield, the most sensitive part of the curve to Federal Reserve policy, added 2.52 basis points to 4.423%, a level not seen since July 2024. The spread between 2-year and 10-year notes sat at a positive 41 basis points.
The next day, yields held near those highs. The 10-year note yield touched 4.9% intraday before settling around 4.84%, and the 30-year bond yield sat at 5.2974%. Only a strong $39 billion auction of 10-year notes on Sept. 9 — which drew demand equal to 2.71 times the amount offered, the highest ratio since 2019, and sold at a high yield of 4.834%, more than a basis point below where it traded ahead of the sale — kept the selloff from going further.
The political framing ran ahead of the market reality. On Sept. 8, speaking at a policy event in Washington, Bessent said his move was aimed at quelling a "fever" that had been building in the bond market. He suggested that yields had risen largely because the financial press had latched onto a negative narrative.
"There was, like, this fever that was building," Bessent said. "They get a hold of a narrative, and I wanted things to become more fact-based."
Then he defined the limits of his own power.
"My job is to try to push things back towards equilibrium," he said. "I don't believe that I can change the equilibrium price. But nothing's ever in equilibrium."
A day later, the market proved him half right. He could nudge prices, but the equilibrium — set by inflation, oil, deficits, and the Federal Reserve — moved the other way. Tom di Galoma, a managing director at Mischler Financial, put the trader's read plainly.
"People thought the buyback was going to be larger than $6 billion," he said. "It was a disappointment so bond prices sank and yields rose."
What a buyback can do — and what it cannot
The buyback is a debt-management tool, not monetary stimulus. The Treasury swaps newly issued or on-the-run securities for older, less liquid longer-dated bonds. It is funded by borrowing, not by money creation, so it is not quantitative easing. The department's August statement described the goal in bureaucratic language: the increase "reflects Treasury's desire to provide greater liquidity support in longer-dated nominal sectors where there is consistent strong sponsorship from market participants." Treasury's own borrowing estimates note that buybacks are not expected to significantly affect net marketable borrowing, because new issuance replaces the securities that are bought back.
Translated, the Treasury was acknowledging a buyers' strike in the long end of the curve. Since late June, dealers and investors have been reluctant to hold 20- and 30-year paper, pushing the 30-year yield to a 19-year high of 5.3% in August. By standing ready to absorb up to $6 billion in a single operation, the Treasury hoped to remind the market that it has a buyer of last resort for the maturities investors are shunning.
But the mechanism has hard limits, and they are arithmetic. A buyback removes a fixed stock of bonds from the market; it does not remove the deficit that creates them. Treasury's own borrowing estimates put total net marketable borrowing at roughly $1.99 trillion for fiscal 2026, and the 12-month rolling deficit reached $1.8 trillion in August, according to Congressional Budget Office estimates. Every bond the buyback retires is, in effect, replaced elsewhere on the balance sheet. That is why Bessent's own words — "I don't believe that I can change the equilibrium price" — are the most honest thing anyone in Washington has said about the episode. A liquidity tool cannot reprice fiscal risk.
The contrast between Aug. 19 and Sept. 9 reveals the channel through which the Treasury actually has influence: expectations. The August move worked because it was a surprise that signaled a willingness to do more. The September move failed because it was the known follow-through, and it revealed the outer edge of that willingness. Once the size is on the table, the market prices it — and then looks past it to the roughly $2 trillion of net borrowing still coming.
The fever is fiscal, geopolitical, and monetary
What drove yields higher this week was not a shortage of buyers for old 20-year bonds. It was a stack of macro shocks that no buyback can touch. Brent crude pushed above $100 a barrel for the first time since late July as hostilities between the United States and Iran intensified, with West Texas Intermediate near $96. Oil at those levels feeds directly into inflation expectations, which is why producer-price data expected to show a 5.4% year-over-year August increase — up from 4.7% — mattered more than the size of a repurchase operation.
Then there is the Fed. After an August employment report that added far more positions than expected, traders lifted the implied odds of a rate hike at the September 15-16 meeting to roughly 60%. A hike, or even a hawkish hold, raises the entire curve. Lawrence Gillum, chief fixed-income strategist at LPL Financial, framed the transmission.
"The inflationary story is getting a little stickier," he said, "and that could cause the Fed to hike rates."
Beneath both sits the fiscal arithmetic. Gross federal debt has crossed $40 trillion, and debt held by the public has reached 100% of GDP. The Committee for a Responsible Federal Budget calculates that if the 10-year yield stays about 60 basis points above the Congressional Budget Office's February forecast, it will add $2.3 trillion to the debt over the next decade — pushing debt to 125% of GDP by 2036 instead of 120%. Higher debt can also slow growth, which would boost debt further. This is the structural core of the story: the "fever" Bessent described is the market repricing the risk of holding long-duration U.S. government debt in a world of persistent deficits, sticky inflation, and geopolitical supply shocks. A buyback treats the symptom — thin liquidity in the longest maturities — while the disease — the fiscal and inflation trajectory — keeps progressing.
There is also a crowding-out channel that a buyback does nothing to address. Corporate borrowers chasing artificial-intelligence infrastructure issued about $159 billion of bonds by mid-2026, already above the roughly $121 billion the five largest hyperscalers issued in all of 2025, against a 2020-to-2024 average of just $28 billion a year. That private demand competes with Treasury supply for the same pool of long-duration capital, and it bids up the term premium that the Treasury must pay.
The counter-thesis: the market is functioning just fine
The strongest case against that read is that the bond market is not broken at all. Bessent himself has argued that the U.S. Treasury market has been the best-performing developed bond market this year, with total returns of 6% — its best year since 2020 — and that the 10-year term premium is "basically unchanged." On that view, the selloff is a relative-value rotation, not a crisis of confidence: investors are rotating out of long bonds because they expect higher short rates and higher inflation, exactly the behavior a healthy, forward-looking market should display.
He has also argued that if investors were truly worried about the United States, they would be selling Treasuries and buying German or Japanese bonds instead — and U.S. bonds have outperformed. The Sept. 9 auction supports that view. A 2.71-times bid-to-cover ratio is not the signature of a market in distress; it is the signature of a market that cleared at the right price. And U.S. borrowing costs, while elevated, remain far below the 15.82% peak the 10-year yield reached in September 1981.
There is force in that argument — but it misses the asymmetry. Yields can rise on their own; they cannot be pushed back down by a $6 billion repurchase. The Treasury can intervene to smooth liquidity, but it cannot force investors to accept a lower term premium when the fiscal outlook is deteriorating. The market's reaction to the buyback announcement — selling, not buying — shows that investors now treat debt-management fine-tuning as a signal of anxiety rather than as reassurance. When the caretaker of the market has to advertise his own interventions, the market stops taking comfort from them.
The counter-thesis also leans on a comparison that no longer holds. Bessent's claim that U.S. bonds are "the best performing in the world" may be true on a year-to-date total-return basis, but it is a backward-looking statistic. What matters for the next decade of borrowing costs is not how Treasuries performed while inflation was falling; it is how they perform while the deficit stays wide, oil stays above $100, and the Fed's next move is a hike.
Second-order risk: the line between debt management and monetary policy
Beyond the mechanics, the episode carries an institutional cost. Strategists at Barclays noted the Treasury could leave the maximum buyback size open-ended — "at least $4 billion per operation" — which preserves flexibility but leaves markets without a clear roadmap. Morgan Stanley has calculated $10 billion as a practical cap on such operations. Every step up that ladder deepens the expectation that the Treasury is managing yields, not just liquidity.
That expectation blurs the boundary between the Treasury's debt-management mandate and the Federal Reserve's monetary-policy mandate. Bessent has already crossed that line rhetorically: at a university event earlier this week he declared, "I am the house now," warning currency traders not to bet against his interventions after the Treasury helped stabilize the Japanese yen in early August. If investors come to believe that long-term rates are being steered by the fiscal authority, the term premium they demand for holding that risk rises — the opposite of the intended effect.
Traders have already noted that any surprise in buyback size is more likely to show up in 30-year swap spreads than in outright yields, because spreads reflect supply shifts while outright yields remain driven by broader macro forces. The market is telling the Treasury, in its own language, that it cannot be fine-tuned. The more the Treasury signals its desire to push yields lower, the more the market prices in the anxiety behind the signal.
Outlook: three horizons, three verdicts
The verdict on Bessent's "fever" fight is that the cyclical and structural legs of the story point in opposite directions — and the structural leg is winning. In the short term, yields will swing with the inflation prints and the Fed's September decision; a buyback can shave a basis point or two off a disorderly move, but it cannot change direction. Over the medium term, the refunding calendar and the deficit path dominate: the Treasury has said it will address future buyback sizes at the November 4 Quarterly Refunding, and every operation until then is a test of whether debt management can substitute for fiscal credibility. Over the long term, the question is whether the 10-year yield has entered a higher regime, one set by a 100%-of-GDP debt load and an inflation floor that oil shocks keep lifting.
Base case: the 10-year yield grinds between 4.8% and 5.2% as sticky inflation and steady supply offset the calming effect of buybacks. Upside case for bonds: a string of soft inflation prints and a growth scare pull the 10-year back toward 4.2%-4.4%. Downside case: oil stays above $100, the deficit widens, and the 10-year tests 5.2%-5.5%, with the 30-year pushing toward and above 5.4%, the April 2007 peak.
The signal that would prove the structural-repricing view wrong is specific and observable: if core inflation prints below 0.2% month over month for two consecutive months and the 10-year yield falls back below 4.3%, then this was a cyclical liquidity scare, not a regime shift. The signal that would prove it right is equally concrete: if the 30-year yield breaks and holds above 5.4% — the April 2007 peak — on an auction that tails with a bid-to-cover below 2.3 times, the market has moved from repricing to refusing.
Bessent is right that nothing is ever in equilibrium. But this week's lesson is that a Treasury Secretary can push prices toward equilibrium only when the market agrees on where equilibrium is — and right now, the market's equilibrium is set by the deficit, not the buyback.
Explore more exclusive insights at nextfin.ai.
