"I'm not doing QE," Bessent said in Texas on Tuesday, and then, describing the August bond market: a "fever," a narrative that "the United States would be unable to service its debt" — "absurd, but it became the dominant narrative."
That is the man who, at 11 a.m. Wednesday, will print a single number the entire long end is waiting on. Not a rate, not a target. A dollar size for one buyback operation.
Last week the story was directional: hike odds reportedly sat around 60/40, roughly 60%, having peaked near two-to-one earlier in the week, and the question was whether Warsh would pull the trigger. This week the market has split. Polymarket has no-change at 66% and a hike at 33%; Orrery's read of the same event reportedly puts hold at 71% (down 5 points over seven days) and the 25bp hike at 33% (up nine points over seven days).
The split is not the story. It is a symptom.
Warsh caused it. At Jackson Hole he re-centered the whole framework on inflation — 3.7% headline PCE, a 4.1% six-month annualized pace, more than 54% of components rising at an annualized rate above 3% over the past year — and then, crucially, argued that forward guidance should be "reserved for crisis conditions rather than used routinely."
Strip out the guidance and you strip out the thing that made Fed nights legible. The market can no longer read the reaction function off a dot plot; it has to author one itself, in real time, around two data prints (PPI Thursday, CPI Friday) and one Treasury number.
Hence the disagreement. When the Fed refuses to pre-commit, prediction markets fracture — 66/33 is what genuine uncertainty about an unstated reaction function looks like priced into shares.
Now the mechanism, because this is where the two catalysts of the week reveal themselves as opposing forces rather than reinforcing ones.
The funds-rate decision is a front-end event. It sets the overnight rate. But both headline catalysts of the week actually land on the long end — and they push it in opposite directions.
A hike pushes long yields up, through the inflation-credibility and term-premium channel. Warsh tightening validates the view that inflation is sticky and the Fed will have to keep leaning; that lifts the compensation investors demand to hold 30-year paper.
The buyback pushes long yields down, through supply. Bessent is retiring outstanding 10-to-20-year bonds, shrinking the float the market has to absorb.
So on the same night, one arm of Washington weighs a move that lifts the long end while another prints a number designed to press it down. They do not stack. They cancel — and the net depends entirely on how hard each one pulls.
One arm of Washington weighs a move that lifts the long end while another prints a number designed to press it down. They do not stack. They cancel.
That is why the sizing print, not the split, is the trade.
The consensus floor is $4 billion — the Treasury pledged on August 19 to "at least double" the prior $2 billion operations, so $4B is reportedly the mechanical minimum. Wrightson ICAP's Lou Crandall calls $5-6 billion the "reasonable starting range." Morgan Stanley marks the practical ceiling near $10 billion, above which the Treasury General Account's cash balance becomes a binding funding constraint.
The magnitudes are not academic. Per Wrightson's math, a $6 billion operation cuts quarterly net issuance of 20-year-plus Treasuries by about 27%; a $10 billion operation cuts it by roughly 55%. That is the size of the supply offset in play.
And here is Crandall's trap, the one that makes the number so hard to call. If the Treasury goes big, it "would amount to an admission that the Treasury had not fully thought through its hasty August 19 statement." If it comes in soft at $4 billion, disappointment feeds the exact selling the buyback was meant to cool.
Bessent broke "regular and predictable" to announce this. He acted outside the quarterly financing cycle, after the 30-year reportedly briefly hit its highest level since 2007, and then insisted it wasn't QE — reportedly reaching instead for the Operation Twist analogy, a maturity shuffle rather than a balance-sheet expansion.
That framing is doing a lot of work, and it is exactly why $4 billion is dangerous. A soft print reads as confirmation the August pledge was rushed — a fever remedy improvised under pressure and then walked back to its minimum. The narrative Bessent called "absurd" gets fresh oxygen.
So watch the right instrument. Traders expect any surprise to surface first in 30-year swap spreads, not in outright yields — because spreads are the supply-sensitive gauge, while outright price is dominated by global macro. Oil reportedly above $93 amid the Middle East escalation, the geopolitical premium, the whole inflation backdrop — that noise lives in the outright yield.
The clean tell is the spread. If the buyback beats expectations, swap spreads should move first and hardest; if it disappoints, the same. Outright 10s and 30s will be too contaminated by the macro tape to read cleanly on the night.
Which brings us to the collision. The maximally bearish long-end combination is not a hike alone, and not a soft buyback alone. It is both together — a Warsh hike stacked against a $4 billion floor.
In that scenario the term-premium channel pushes the long end up and the supply channel fails to offset it. Worse: the soft print removes the one counterweight to a hawkish long end, and it does so while the Fed is leaning the same direction as the sellers. The buyer of last resort shows up small on the same night the central bank confirms the bears' thesis.
That is the setup that repeats the round-trip. Arguably, the ingredient that kills a buyback's staying power is precisely a Fed lined up against it.
So the open thread from last week — does the buyback sequence actually move the long yield, or does the round-trip pattern repeat — is now answerable against the print. A number at or below $4 billion, especially paired with a hike, and you should expect the round-trip, only this time with the Fed pushing the same way as the sellers rather than sitting neutral.
Ignore the 66/33 argument about whether Warsh moves. That is a front-end question the whole market is loudly confused about, and the confusion is baked in.
Watch the $6 billion bar instead. Clear it, and Bessent has a credible supply offset that can hold even into a hawkish Fed. Miss it at the $4 billion floor, and he has printed an admission — and handed the long end its worst combination of the year.