There is a conspiracy theory making the rounds on trading desks, and it is worth taking seriously precisely because it is wrong.
The theory, laid out in a Bank of America note from August 7, is that Fed Chair Kevin Warsh deliberately used his July 29 press conference to push long-term yields higher — a covert tightening of financial conditions dressed up as bad communication.
BofA's Mark Cabana and Aditya Bhave dismantle it in one move: the FOMC's own strategy statement lists "adjusting the target range for the federal funds rate" as its primary tool and does not mention long-term Treasury yields at all. Warsh has no authority over the long end. He couldn't manipulate it if he tried.
That is the tension. The market is so desperate to find a hand on the long-bond lever that it invented one for a man who doesn't have it — and the reason nobody has it is the actual story.
The market is so desperate to find a hand on the long-bond lever that it invented one for a man who doesn't have it.
This week the thread moves underneath the presser: the repricing is anchored to a documented structural change in who buys Treasuries, and that change is why no policymaker — Warsh, Bessent, the FOMC — has a working lever anymore.
Start with the two buyers.
A decade ago, a foreign central bank bought the 30-year to park reserves. It did not care what the yield was. It needed dollar assets, Treasuries were the deepest pool, and price was irrelevant to the mandate.
Today the marginal buyer of that same bond is a mutual fund or a pension. It bids only if the expected return clears a hurdle. If the yield is too low, it walks.
Barclays strategists Demi Hu and Anshul Pradan put a number on the swap: private investors now hold roughly 73% of the U.S. Treasury market, up from about 50% a decade ago. Official institutions — the Fed, foreign central banks — have fallen to the residual 27%.
Their conclusion is blunt. "The buyer base for U.S. Treasuries has changed. The same scale of debt supply may require a larger yield concession to be digested by the market."
J.P. Morgan Asset Management describes the same shift as the private sector taking "back" the bond market, reversing a two-decade trend in only four years. Central banks were, in their phrase, "non-economic buyers" who moved yields to artificially low levels. Now the price-setter demands compensation.
This is the mechanical fact under everything else. A price-insensitive buyer base let Washington issue cheaply regardless of fundamentals. A price-sensitive one does not.
And the numbers show the buyer is holding the line.
The 30-year yield stayed above 5% for 41 consecutive sessions — the longest streak since 2007. Per Bloomberg (E2), it touched roughly 5.28% on July 31, a 19-year high; Bloomberg reports that level as a print around the period following Warsh's July 29 presser.
Now watch Washington reach for the levers, and watch each one fail to touch the thing it's aimed at.
Bessent's first move was coordinated yen intervention with Japan. The logic is indirect by construction: Japan is the largest foreign holder of Treasuries at roughly $1.14 trillion, and if the yen collapses, Japan might sell Treasuries to raise dollars to defend it. Buying yen jointly, plus expanding the Fed's FIMA repo facility so Tokyo can get dollars without selling, is meant to prevent forced selling.
Note what this is not. It is not creating demand. It is preventing one specific holder from becoming a supplier. Best case, it removes a downside — it does not put a bid under the 30-year.
The second move was language. In the quarterly refunding statement, Treasury swapped "potential future increase" for "potential future changes," which the street read as room to cut long-dated issuance — a BMO survey had 61% expecting a reduction in 30-year supply.
That is real. Cutting supply is the one genuinely mechanical tool Treasury holds. But it is a confession, not a fix: you shrink the long end because you cannot fund it at a price you like, which is exactly the buyer-base problem stated in reverse.
The third move was theater. Bessent went on CNBC to endorse Warsh's communication, saying markets needed help "decoding" the Fed's remarks.
UBS's Phoebe White summarized the whole program in one line: the steps "may have limited impact," but they show "the Treasury will use every tool available to prevent a further rise in long-term Treasury yields." Every tool available — and the honest read is that the available tools do not reach.
Because the Fed's don't either. Per the Bank of America Securities report, the Fed has direct, precise control over overnight rates and almost none over the long end absent large-scale asset purchases. Everything else is indirect — policy expectations, or the term premium, which they warn is "difficult to control precisely," carries overshoot risk, and once loose "can become highly volatile."
So the picture resolves. The one entity that could buy the long end without caring about price — the central bank via QE — has spent since 2022 shrinking its book. The buyers who replaced it price on return. And the term premium, the channel through which policy might indirectly bend the long end, is the least controllable variable in the system.
Which is why the Warsh conspiracy theory is so revealing. Markets went looking for a puppeteer because they could not accept that the strings had been cut.
Now the smoking gun, and it lives in a corner of the market almost nobody watches: STRIPS.
STRIPS are zero-coupon securities carved out of Treasuries — separate the coupons and the principal, give each its own CUSIP, trade them individually. Because they pay a known amount on a known date, they are the natural instrument for anyone matching future liabilities to cash flows.
The New York Fed's first detailed look at STRIPS trading data reveals a very specific buyer. Bonds account for 98.1% of Treasuries held in stripped form. Stripping and trading concentrate in the long end.
Principal STRIPS trade with an average size of about $7.26 million — versus roughly $850 thousand for coupon STRIPS — and 88% of their volume goes through the dealer-to-client market. In the DTC segment specifically, the average principal-STRIPS trade is reportedly $10.47 million.
That is a fingerprint. Large, single-payment, long-dated, traded directly between dealers and end clients — the New York Fed attributes the high DTC share to "the value that end-users such as pension funds and insurance companies ascribe to these single-payment securities."
These are liability-driven investors. They buy the 30-year to match a pension payment or an insurance claim decades out. They are not decoding Warsh. They are not front-running a rate cut. They match cash flows, and they will only pay a price that funds the liability.
This is who is absorbing the long bond now. Not a reserve manager indifferent to yield — a matcher who demands the yield be right.
And here is the part that makes the long end genuinely unpickable. Nobody knows what "right" is, because nobody knows the anchor.
A separate New York Fed staff analysis reportedly estimates investors' 95% confidence band around r-star — the natural real rate that supposedly anchors long-run pricing — at roughly ±170 basis points. Even with full hindsight, the two-sided band stays around ±130 basis points.
Sit with that. The interval of uncertainty around the anchor is wider than most people's entire forecast range for where the 30-year should trade.
So the marginal buyer is a liability-matcher who won't move on price, the compensation he demands rests on a term premium no policymaker controls, and the real-rate anchor underneath it all is known only to within a ±170bp fog.
That is the regime. Not a lever that Bessent or Warsh is pulling badly — the absence of a lever anyone can hold.
OMFIF calls the Treasury market "living on borrowed time," a fragile foundation of buyers of whom "none of them regard Treasuries as a permanent, unconditional holding." That is directionally right but slightly dramatic. The bonds will clear. Washington can always create the currency they're paid in. What it cannot do is set the price.
And when the price of government debt is set by private hurdles rather than official mandates, the yield stops being a monetary variable and becomes a fiscal verdict. Higher for longer isn't a Fed stance anymore. It's arguably the market pricing the deficit, and it has moved the invoice from the central bank's balance sheet onto the ballot box — a 19-year-high 30-year, ahead of a midterm, that no one in Washington can actually pick up.