The Fed Is About to Send Treasury Yields Soaring — Because the Shock Absorber Is Gone

The 10-year Treasury was already trading near 5% when the Federal Reserve sat down for its September meeting — and the most interesting thing about the rate decision is no longer the number itself, but what happens to that 5% level the moment the Fed acts. Kevin Warsh, in his first Jackson Hole address since taking the chair, turned what had been a near-certain hold into a coin flip on a hike: CME futures put the odds of a September move at 56%, with prediction markets near 50%. Two-year yields jumped about 20 basis points on the speech as markets began pricing two hikes by March.

Everyone is focusing on the wrong fight.

The Treasury is doing its best to keep a lid on the long end. Secretary Bessent built the Treasury General Account — the government's checking account at the Fed — to roughly $950 billion and doubled the size of bond buybacks, operations he calls a "Treasury Twist," buying long-dated debt with cash and short-term bills to hold yields down. First enhanced operation, September 9. The White House, reportedly, treats 5% on the 10-year as a line not to be crossed. This is the narrative most commentary grabs: a government with a $950 billion war chest defending the long bond.

The problem is that the Fed holds $4.3 trillion of Treasuries. A $4 billion buyback operation against that is a rounding error.

Here is the mechanism that actually matters, the one most commentary skips. From 2022 through late 2025, the Fed shrank its balance sheet by more than $2 trillion through quantitative tightening — shedding Treasury and mortgage holdings without breaking the Treasury market or spiking long yields. It could do that because a giant pool of idle cash sat parked at the Fed's reverse repo facility, absorbing the withdrawal. That pool peaked at a record $2.55 trillion at the end of 2022. It functioned as the shock absorber for the entire tightening cycle. As QT drained reserves, money parked in that facility drained first, topping the system back up, cushioning every dollar of duration the Fed stopped holding.

That pool is now empty. Reverse repo usage has fallen to roughly $1 billion — down about $2.5 trillion from the peak — and QT itself was paused in December 2025. The cushion is spent. There is no remaining pool of idle cash to absorb the next withdrawal from the system.

Now the Fed is about to tighten again, and with a cushion and without one are different games. During the first QT, the Fed could shrink its balance sheet and raise rates because the reverse repo pool absorbed the strain; the private market never had to swallow that much duration. Today, with the pool drained, whichever lever Warsh pulls — a hike, or the resumed balance-sheet runoff he has telegraphed he wants to shrink the Fed's footprint — the money comes straight out of real buyers of duration. Banks lend reserves into a scarcer funding market, and the term premium that compresses long yields has no reason to stay low. That is the plumbing by which the Fed sends the 10-year through 5%: not through the rate itself, but by removing the buffer that made the last tightening painless.

Read the incentives. Warsh is fighting inflation, not deflation — 12-month PCE at 3.7%, the six-month run rate 4.1%, inflation above the 2% target for more than five years, crude above $100 a barrel on Middle East escalation feeding straight into the August CPI expected at 0.4% month over month. He told Jackson Hole the Fed must ensure underlying inflation is moving to its objective "clearly and at sufficient speed," and that it is "ready to act." He is simultaneously pulling back on forward guidance — removing the Fed's implicit promise that kept long-term uncertainty low. Three FOMC members already dissented for a hike in July. Nothing in his incentives points toward defending the 5% line the way the Treasury is.

Which is the concede-then-flip: yes, the base case is still a hold. A Reuters poll of 93 economists had roughly 70% expecting no change in September and 56% expecting the Fed to sit on its hands all year — but that second share has tumbled from 80% in a matter of months, and the portion expecting at least one hike more than doubled. The consensus is moving toward Warsh, not away from him.

Hold on to the two variables that would break this reading. First, the timing of the decision relative to the data — the Fed's September call lands right after the fresh inflation print, so an upside surprise is the accelerant, a downside one the exit. Second, the balance-sheet question, which matters more than the rate: whether Warsh restarts the runoff that price action has already stopped treating as a free lunch. If inflation breaks lower and the Fed holds with the Treasury still leaning on the long end with its cash, then 5% holds and the ceiling survives. If not, the last shock absorber is gone, and the Fed — not the bond market, not the Treasury — is the one sending yields soaring.