Why Are Long-Term U.S. Treasuries Increasingly Difficult to Sell? The Real Problem May Not Be Inflation

Original title: Beware the Bond: Operation Twist is Back

Original author: Trader Joe

Translation: Peggy

Editor's note: This week, the 30-year U.S. Treasury yield briefly rose to around 5.34%, its highest level since 2007. Shortly after, Treasury Secretary Bessent announced an expansion of long-dated Treasury buybacks, increasing the maximum single-operation repurchase size for certain 10- to 30-year bonds from $2 billion to at least $4 billion, with the arrangements to be implemented between September 9 and November 4. Following the announcement, long-end yields pulled back, the dollar weakened, and risk assets gained some support.

On the surface, this is a relatively modest Treasury market liquidity operation. The more noteworthy question is: why has U.S. long-end rates risen to a level that requires more proactive attention from the Treasury Department? And is the market beginning to reassess the Treasury's "policy reaction function" toward long-end yields?

In "Beware the Bond," Trader Joe argues that the recent long-end selloff cannot be simply attributed to inflation. Fiscal deficits continue to generate Treasury supply, demand from traditional long-bond buyers such as Japan is shifting, and AI capital expenditures are creating substantial long-dated bond supply in the credit market—several forces are jointly raising the amount of duration that markets need to absorb.

The author further likens Bessent's expanded long-bond buybacks to a Treasury version of "Operation Twist." This analogy captures the direction of "reducing long-duration supply in the market," but the two are not equivalent: the 2011 Operation Twist was executed by the Federal Reserve, selling short-term bonds and buying long-term bonds, with the explicit goal of lowering long-term rates and easing financial conditions. The Treasury's current buyback program remains officially positioned as a tool for improving secondary market liquidity and cash management. What truly deserves attention, therefore, is not the $4 billion figure itself, but whether this tool will increasingly assume the role of managing long-end financial conditions in the future.

The following is a translation of the original article:

Earlier this week, the 30-year U.S. Treasury yield briefly rose to its highest level since 2007, U.S. stocks gave back some gains, and the dollar began to weaken.

Then Bessent stepped in.

The U.S. Treasury announced it would raise the maximum single-operation repurchase size for certain 10- to 30-year long-dated bonds from $2 billion to at least $4 billion, to be executed between September 9 and November 4. Following the announcement, the 30-year Treasury yield fell from its previous high of around 5.34% to roughly 5.2%, the dollar weakened further, and equity markets stabilized somewhat.

U.S. 30-Year Treasury Yield

The question is: is this merely a temporary patch for bond market liquidity, or does it signal a shift in the policy establishment's stance toward long-end rates?

To understand this, one must first answer another question: why have long-end yields risen to this level?

Why have long-end yields risen to this level? The issue isn't just inflation

The most intuitive explanation is inflation.

If investors worry that inflation will remain elevated over the long term, they will naturally demand higher long-term Treasury yields as compensation. But the author argues that this does not fully explain recent moves.

At least judging by consumer surveys, long-term inflation expectations have not yet shown signs of becoming unanchored. The University of Michigan's preliminary August survey shows one-year inflation expectations edging up from 4.2% to 4.3%, but five-year inflation expectations remain at 3.3%. In other words, short-term inflation concerns persist, but "long-term inflation expectations becoming unanchored" is not the only—or necessarily the most important—explanation at present.

Source: University of Michigan Consumer Survey

Long-term Treasury yields have never reflected only future short-term policy rates and inflation. Economic growth, term premium, the regulatory environment, how much debt the Treasury needs to issue, and how much long-dated U.S. debt insurers, pension funds, and overseas investors are willing to allocate all affect long-end pricing.

And what the author finds most noteworthy is that supply of long-dated bonds is steadily increasing, while traditional demand has not expanded in tandem.

The U.S. fiscal deficit means the Treasury still needs to continuously finance itself, and whether that financing is done through short-term T-bills, medium-term notes, or 30-year long bonds directly affects how much duration risk the market needs to absorb.

If the Treasury relies more on short-term T-bills for financing, it effectively reduces the supply of long-dated bonds that need to be digested by the market, putting relatively less pressure on long-end yields. Conversely, if more financing shifts toward 10-year, 20-year, and 30-year securities, the market must absorb more duration, and long-end yields may face greater upward pressure.

This is also why the debt issuance structure itself has increasingly become a macro variable.

Why is the Treasury acting now? A 5.3% long end is starting to affect financial conditions

Short-term debt can ease long-end pressure, but the liquidity buffer is thinning

The problem is that short-term debt cannot be issued indefinitely either.

Over the past few years, when the U.S. Treasury issued large amounts of T-bills, a key funding source was money market funds' balances parked in the Federal Reserve's overnight reverse repo facility (ON RRP). When short-term bond yields became more attractive, these funds could flow from RRP into T-bills, absorbing new short-dated supply without significantly draining bank reserves.

But now, this buffer is nearly depleted. Fed data shows that ON RRP usage has been near zero on most trading days. Meanwhile, as of mid-year, U.S. bank reserves stood at approximately $3.1 trillion.

In the second half of 2025, the large-scale rebuilding of the Treasury General Account (TGA) further drained banking system liquidity. Fed data shows that after the debt ceiling issue was resolved, the TGA balance increased by approximately $442 billion at one point, while reserves declined noticeably.

This is also one of the background factors behind the Fed's decision to end quantitative tightening (QT) at the end of 2025.

In October 2025, the Fed announced it would stop shrinking its balance sheet starting December 1; in December, it began purchasing short-term U.S. Treasuries through Reserve Management Purchases (RMP) to ensure bank reserves remained at an "ample" level.

These operations can easily visually resemble QE, but the policy intent is different.

QE typically involves purchasing long-term Treasuries or MBS to actively lower long-term yields and ease overall financial conditions. RMP primarily purchases short-dated securities like T-bills, with the official goal of maintaining adequate bank reserves and short-term rate control, rather than providing macroeconomic stimulus. The Fed has also explicitly emphasized that RMP does not represent a change in monetary policy stance.

The author's concern is that if the Treasury continues to increase its share of short-term financing to reduce long-dated supply, then once liquidity buffers like RRP are nearly exhausted, new short-dated debt may increasingly compete with bank reserves.

At that point, the Fed may be forced to conduct more reserve management operations to maintain system liquidity. This creates a subtle policy combination: the Treasury tries to release as little duration into the market as possible, while the Fed ensures adequate reserves at the short end.

Japan and AI are both reshaping long-dated bond supply-demand dynamics

The long end has another dimension: who buys?

Japan has long been the most important overseas investor in U.S. Treasuries. The latest TIC data from the U.S. Treasury shows that as of June 2026, Japan held approximately $1.116 trillion in U.S. Treasuries, still the largest foreign holder, but down about 2.3% from May.

At the same time, Japan's own long-term government bond yields are rising.

For Japanese insurers, banks, and pension funds, if JGBs themselves can offer increasingly attractive yields, the marginal appeal of allocating to long-dated U.S. Treasuries may naturally decline—especially after accounting for dollar hedging costs.

This does not necessarily mean Japan will persistently sell U.S. Treasuries on a large scale, but it does suggest that a structural long-bond buyer that has existed for years may no longer absorb U.S. duration as reliably as before.

The other competitor comes from AI.

AI infrastructure buildout is transitioning from an equity market story to a credit market story. Goldman Sachs Research estimates that the entire AI-related supply chain has issued nearly $500 billion in debt so far in 2026 alone; among them, hyperscale cloud providers themselves have issued approximately $194 billion. More importantly, consider the maturity structure. In this year's U.S. investment-grade credit market, roughly 40% of new issuance with maturities of 15 years or longer has already come from AI companies or AI-related financing.

This means that traditional long-duration capital pools like pension funds and insurers face more choices. They are no longer just comparing 30-year U.S. Treasuries against other sovereign bonds; they can also buy long-dated investment-grade debt from large tech companies like Amazon and Google, as well as AI-related credit assets such as data centers and infrastructure.

From the author's framework, this makes the core problem facing U.S. long bonds clearer: the Treasury needs to sell more and more debt, while the volume of other long-dated assets requiring investor absorption in global markets is also rapidly increasing.

Treasury version of Operation Twist: $4 billion is small—what really changes is the policy reaction

It is precisely against this backdrop that Bessent expanded long-dated Treasury buybacks.

The Treasury's regular buyback program began in 2024, with two officially stated purposes: improving secondary market liquidity and cash management.

Among them, liquidity support buybacks primarily purchase less liquid off-the-run Treasuries. By periodically acting as a potential buyer for these bonds, the Treasury hopes to help dealers free up inventory and improve trading conditions for off-the-run securities.

Therefore, from an institutional design perspective, this is not a QE tool established to suppress the 30-year yield.

Moreover, a minimum of $4 billion per operation is still small within a U.S. Treasury market exceeding $30 trillion. Reuters has also noted that the market generally views this scale as insufficient to address structural issues such as fiscal deficits and increased long-dated supply.

But what the author truly focuses on is not scale, but policy intent.

In the past, the Treasury could emphasize that buybacks were merely a market liquidity tool. Now, when the 30-year yield rapidly approaches two-decade highs and the Treasury immediately expands long-dated bond buybacks, the market will naturally begin to ask: if the long end continues to spiral out of control in the future, will the Treasury further adjust its buyback and issuance structure?

This is precisely why the author describes current policy as a Treasury version of "Operation Twist."

Note: Operation Twist does not involve "printing money" at its core; rather, it adjusts the maturity structure of central bank bond holdings—selling short-dated bonds and buying long-dated bonds to lower long-term rates.

The classic Operation Twist of 2011 was executed by the Federal Reserve: the Fed sold or allowed short-term Treasuries to mature while purchasing an equivalent amount of 6- to 30-year Treasuries, extending the duration of its portfolio without expanding the balance sheet, thereby reducing private-sector holdings of long-dated Treasuries and lowering long-term rates.

What is happening today is not the same operation. The Treasury has not engaged in a strict "sell short, buy long" like the Fed did back then, and the expanded buyback remains officially defined as a debt management and liquidity tool. But from the perspective of market duration supply, there is a similar direction: if the Treasury buys back more long-dated off-the-run bonds while keeping more net financing pressure at the short end, the net duration that private markets need to absorb may be relatively reduced.

This is what the author calls the "Treasury version of Operation Twist." More precisely, it is currently a market interpretation rather than an established new policy framework.

Can this approach suppress the long end? The risk may shift to the dollar and inflation

So, under what circumstances would this policy package continue to escalate? The author believes that rather than looking for an absolute "red line" for the 30-year yield, it is better to observe the speed of yield increases. Whether the 30-year yield sits at 5.2% or 5.3% may not in itself be enough to trigger policy changes; but if the market begins to see consecutive rapid jumps of around 10 basis points in single sessions, it would signal that trading order and demand are visibly deteriorating, raising the probability of further intervention by the Treasury or the Fed.

Meanwhile, long-dated bond yields are increasingly competing more directly with equities for capital. Based on data at the time of the author's article, the nominal 30-year Treasury yield was approximately 5.2%, with the real yield on long-dated TIPS near 3%. By comparison, the S&P 500 earnings yield was approximately 3.8%.

The two cannot be directly compared one-to-one—an earnings yield is not a risk-free rate, and corporate earnings will grow or decline in the future—but when the risk-free long-term real yield rises to such an elevated level, the opportunity cost equity valuations must bear is clearly increasing.

Therefore, the truly important aspect of Bessent's operation may not be pulling the 30-year yield back from above 5.3% to around 5.2% temporarily.

Rather, it is that the market has, for the first time, obtained a new observation sample: when U.S. long-end yields rise rapidly, will the Treasury respond with increasing proactiveness through buyback scale and debt maturity structure?

If the answer gradually becomes "yes," then what affects the dollar, U.S. equities, gold, and long-dated Treasuries in the future will not just be the Fed's policy reaction function, but also a new layer: the Treasury.

But this logic also has boundaries. If long-end increases are primarily driven by bond supply-demand imbalances, reducing the duration the market needs to absorb may ease pressure; if inflation expectations visibly re-accelerate, then continuing to expand buybacks and increase short-dated financing could instead make the market worry that policy is artificially suppressing financial conditions.

Therefore, what truly needs to be observed going forward is not just whether the Treasury will increase buybacks further, but whether inflation expectations, long-dated Treasury issuance structure, overseas demand, and the volatility of long-end yields are all changing simultaneously.

Only if these variables continue pointing in the same direction will the author's thesis—that "the Treasury is taking over part of long-end financial condition management"—be further validated.

Original link