On Tuesday of this week the thirty-year US Treasury bond yielded 5.337 percent, the highest level since 2007. On Wednesday the Treasury Department announced it would at least double its buybacks of long-dated debt, and the market responded immediately: the thirty-year yield fell nine basis points to 5.196 percent, the ten-year fell 5.7 basis points to 4.647 percent. By Thursday the entire move had been given back. The thirty-year stood at 5.238 percent, the ten-year above 4.7 percent, and the Dow Jones fell 1.3 percent, roughly 700 points. On Friday the two yields sat at about 5.25 and 4.71 percent.
The relief the world’s largest borrower granted itself lasted exactly one trading day. That, rather than the announcement itself, is the news of the week — not the intervention, but its half-life.
What happened here matters well beyond nine basis points. A debtor stepped into its own bond market as a buyer in order to influence the price it pays for new debt. That is not illegitimate, it is not even new, and it is emphatically not monetary policy. But it moves a boundary that has been treated as settled for forty years: that the level of long-term interest rates is set in the market and not in the finance ministry. And it does so eleven days before a new Federal Reserve chair delivers his first substantive address at Jackson Hole.
What was actually announced on Wednesday
The details are less dramatic than the market reaction implies. Treasury is raising the maximum size of its liquidity support buybacks for nominal coupon securities in the ten-to-twenty-year and twenty-to-thirty-year maturity buckets from two billion dollars to at least four billion dollars per operation. The change takes effect on September 9 and runs through November 4. The operations are generally conducted once a week.
Treasury Secretary Scott Bessent went further on television on Thursday: We are going to make a market in these. We routinely do buybacks, and we are going to increase the size of the buyback — it could be more than four billion dollars per issue. He added: We have a big toolkit. As justification he said part of the point was signalling, to show that yields do not reflect the underlying fundamentals of the Iran conflict.
The buyback programme itself dates from May 2024. It was explicitly not introduced as a rate instrument but as liquidity support: market participants were to be given a way to sell older, thinly traded off-the-run securities back to the issuer. The purpose was the tradability of the market, not its price level. It is precisely that purpose which was quietly widened this week.
A buyback retires nothing — it swaps maturities
Here is the point most coverage leaves out. A Treasury buyback does not reduce the national debt by a single dollar. It is paid for out of cash on hand, and that cash is typically raised by issuing new short-dated bills. A thirty-year bond disappears and four-week, three-month or one-year paper takes its place. Total debt outstanding is essentially unchanged — it crossed forty trillion dollars for the first time this week.
What does change is something else: the quantity of duration that private investors are required to hold. That is the actual transmission channel. When the issuer takes long paper out of the market and puts short paper in, the aggregate interest rate risk the market must carry falls, and with it the premium demanded for carrying it. It is the same mechanism through which central bank asset purchases worked — only without a central bank, without a balance sheet expansion, and without a monetary policy mandate.
And it has a price. Bills stood at 22.0 percent of total US debt outstanding as of the end of October 2025. Treasury’s own advisory committee regards an average of roughly 20 percent as the appropriate trade-off between low debt service cost and rollover risk. Every additional buyback funded with bills pushes that share higher. You are purchasing a lower yield at the long end in exchange for greater refinancing risk at the short end. That is not an argument against the measure, but it is the offsetting entry, and it appears in no headline.
Nine basis points against fifteen: the historical yardstick
How large can the effect be at all? There is a remarkably precise historical benchmark. In 2011 the Federal Reserve Bank of San Francisco re-estimated the 1961 Operation Twist — the Kennedy administration’s attempt to lower long rates by buying long securities and selling short ones. The finding: of six potentially market-moving announcements, four had statistically significant effects, and cumulatively the entire programme lowered long-term Treasury yields by about 0.15 percentage points, or fifteen basis points. Securities with five or more years to maturity fell between six and nine basis points.
Set those side by side. A complete programme, run for months by the central bank and the Treasury together, was historically worth fifteen basis points. A single press release this week was worth nine — and gave all of them back within twenty-four hours. Krishna Guha of Evercore ISI framed it as follows: a moderately bigger buyback programme amounts to a weak form of Operation Twist, which in itself will have little enduring impact and could backfire.
The scale comparison explains why. Four billion dollars per weekly operation stands against a debt stock of forty trillion. That is one hundredth of one percent. What moved on Wednesday was therefore not supply but expectation: the market traded a declaration of intent, not purchases — not a single transaction had taken place. Signals with no volume behind them have a short shelf life, and this one was unusually short.
Japan has already finished this experiment
Anyone who wants to know how the story ends does not have to speculate. Japan ran the same experiment eight months ago. In December 2025 the Japanese finance ministry announced it would cut new issuance of super-long government bonds in the following fiscal year to around 17 trillion yen — the lowest in seventeen years and a reduction of nearly a fifth. Monthly volumes in twenty-, thirty- and forty-year paper were each to fall by 100 billion yen. The trigger was identical: a run of record yields at the long end and concern about the new government’s expansionary fiscal stance.
The market reaction was identical too. The thirty-year Japanese yield fell from a record 3.45 percent to 3.38 percent. Seven basis points. Today, eight months later, that same yield sits at roughly 4.14 percent, close to its all-time high. The ten-year Japanese government bond yields 2.945 percent, a level last seen in September 1996.
This is not proof that supply management never works. It is a very concrete indication of what it can and cannot do: it can slow the pace of a move, it can close liquidity gaps in individual securities, and it can prevent a disorderly auction. It cannot set the level when the reason for the level lies outside the bond market. In Japan that reason was fiscal policy. In the United States it is fiscal policy too.
Bessent against Bessent
The most intellectually awkward feature of this week is that the Treasury Secretary has already made the best available criticism of his own instrument — only in the opposite direction.
In June 2024, still a private investor, Bessent accused his predecessor Janet Yellen of having taken control of monetary policy through the composition of issuance, thereby easing financial conditions substantially. In a written statement in November 2024 he was blunter: Yellen had distorted Treasury markets by borrowing more than a trillion dollars in shorter-term debt relative to historical norms. His argument then was exactly the one set out above: throttling long issuance and selling bills instead removes duration from the market and depresses long-term rates — and that decision does not belong in the Treasury building.
The argument was correct. It remains correct today. Buying back long paper and funding it with short paper is mechanically the same procedure, executed through the secondary market rather than the auction calendar. Whoever advanced the distortion thesis in 2024 cannot defuse it in 2026 by calling the instrument liquidity support. Asked directly, Bessent has said the intervention is about market liquidity and not about controlling the yield curve. The distinction is real — but it lies in the intent, not in the effect, and markets trade effects.
Eleven days before Jackson Hole: who sets the long end?
The timing makes this delicate. The Jackson Hole symposium runs from August 27 to 29, with Kevin Warsh delivering his first keynote as Fed chair on the 28th. The federal funds rate stands at 3.50 to 3.75 percent, the Fed has been on hold, and Warsh has repeatedly said he prefers the market to determine rates.
That produces a contradiction which has received too little attention. If the Treasury artificially compresses the term premium at the long end, it eases financial conditions for the entire economy — mortgages, corporate bonds and capital spending decisions hang off the long end, not off the policy rate. A central bank targeting two percent inflation has to offset that easing somewhere. The most likely place is the short-term policy rate. The buyback intended to reduce the interest burden can therefore raise the rate the Treasury pays on precisely the bills with which it funds the buyback.
Then there is the institutional question. There is no sharp line between where Treasury’s remit ends and the Fed’s begins in the government bond market, and Warsh himself has argued more than once for giving Treasury greater weight in balance sheet matters. That position is hard to reconcile with a complaint that Treasury is using its weight. The August 28 speech will therefore be judged less on what it says about the rate path than on whether it draws that line at all.
Britain, and why this is not an American problem
For investors outside the United States this is not an American curiosity, because the move at the long end is global. The thirty-year gilt reached 5.82 to 5.84 percent this week, within sight of six percent, which places the United Kingdom at the sharp end of the same trade — a large deficit, a heavy long-dated issuance calendar, and a domestic pension industry that spent two decades being the price-insensitive buyer of last resort and has now largely finished buying. The thirty-year German Bund yielded near 3.73 percent, the highest since 2011; the thirty-year French bond reached 4.8558 percent, the highest since September 2008. Japanese thirty-year paper sits at 4.14 percent. Oil is above eighty-five dollars a barrel, roughly fifty percent higher than in January, after the President declined to extend the sixty-day ceasefire with Iran.
The practical consequence for a portfolio is a redistribution rather than a uniform loss. Long-duration equities — real estate investment trusts, regulated utilities carrying heavy debt, and the unprofitable end of the growth complex — take the direct hit, because their valuation is close to a pure function of the long real rate. Life insurers and annuity writers are on the other side of that trade: their reinvestment yield is the best it has been in fifteen years, and the same move that compresses a REIT’s net asset value repairs an insurer’s liability discount rate. Banks sit uncomfortably in between, since a steeper curve helps net interest margin while the mark on available-for-sale securities gets worse. Anyone holding a long Treasury fund should be clear about what the last three days demonstrated: the duration is doing the work, and the issuer is not able to protect it.
The counterargument — and what to watch now
Fairness requires the other side. First, Bessent may simply be right on the substance: if part of the yield rise genuinely stems from the Iran conflict and the oil price, then it is cyclical and not a durable repricing of creditworthiness. Second, buybacks of off-the-run securities serve a real function — they demonstrably improve tradability in the less liquid corners of the market, and a functioning market is the precondition for any orderly issuance. Third, in saying that the country will have to grow its way out of this and that there is a good chance peak deficits are already behind it, the Secretary made a testable claim. If it holds, the intervention was a bridge rather than a substitute.
The numbers give him little support so far. The fiscal 2026 deficit is estimated at roughly 1.9 trillion dollars, after 1.4 trillion through the end of June alone. Net interest outlays rise from 970 billion dollars in 2025 to more than a trillion in 2026, an increase of 69 billion or seven percent. A trillion dollars of annual interest is a sum no buyback programme of this size touches.
Three dates structure the coming weeks. On August 28 Warsh speaks at Jackson Hole and will either answer the boundary question or leave it open. On September 9 the first operation at the new size takes place — the first opportunity to measure whether actual purchases behave differently from a mere announcement. And on November 4 the arrangement expires; whether it is extended, expanded or quietly allowed to lapse is the most honest available statement of how the Treasury judges its own success.
Until then this week’s finding stands, and it is uncomfortable in its clarity. The largest borrower in the world announced it would buy its own bonds in order to lower their yield. The market gave it nine basis points and took them back the next day. The price level at the long end is currently set not by supply but by the question of whether enough price-sensitive buyers are willing to fund thirty years of fiscal policy. To that question a buyback programme is not an answer. It is a polite delay.
Try TradingView Free for 30 Days
Plus get a $15 discount on your first subscription through this link.
Read more in our topic hub: Topic Hub: Fed Policy, Rates & Bonds 2026
