Washington buys long bonds with short money while the Fed refuses to move
No buyback can outbid a war that reprices the debt it is meant to protect.
In the same week that the 30-year Treasury touched a yield last seen before the 2008 crash, the Federal Reserve sat with its cash rate frozen.
The gap between those two facts is now the whole argument of the US bond market (Fiscal Times, Aug 19). The long end sells off on a wars deficit; the front end is held in place by a central bank that will not move. Two facts that insist on being true at once are a promise that someone is about to pay.
The benchmark 10-year, just under 4% at the end of February, crossed 4.7% this week, and Fiscal Times reported on August 19 that the 30-year hit 5.34% on Tuesday, its highest reading since June 2007. The short end never budged, because the Fed kept rates steady at its July 28-29 meeting, and CoinDesk reported on August 19 that several officials even wanted a 25-basis-point increase. Steepening. The curve is steepening, which is the market's way of saying the near future looks managed while the distant one looks expensive.
Three institutions hold the rope, and each tugs against the others. Treasury Secretary Scott Bessent wants long yields down so his refinancing is affordable. Fed Chair Warsh wants the front end pinned while he watches oil-driven inflation. President Trump wants the long end to look calm, and posted an escalation against Iran with no regard for either man. Bessent's department borrows, Warsh's sets the price of that borrowing, and Trump's lights the fire that resets it.
Fiscal Times reported on August 19 that on Wednesday morning the Treasury said it would at least double its buybacks of 10-to-30-year debt, from $2 billion to at least $4 billion per operation, running September 9 through November 4 — the day after the midterm elections.
Beneath that sits the pressure: the national debt has crossed $40 trillion, the deficit is expected to top $2 trillion this year, and interest on the federal debt cost more than $1.2 trillion in fiscal 2025 and nearly as much again with a month left in fiscal 2026. Brandywine's Jack McIntyre summed the mood, calling sentiment at the long end about as bearish as he has seen in a long time.
Two wars, not one
The plural in "wars deficit" matters, because there are two wars, not one. There is the shooting war with Iran, running since late February, and there is the Red Sea, where, the New York Times reported on August 20, the Houthis have been choking the alternative route Saudi Arabia uses when Hormuz is shut. Both lift the oil price that feeds the inflation the long end fears. The 20-year auction that set off the sell-off had been lackluster the day before the Treasury moved, 247 Wall St noted on August 20.
For a few hours the lever moved the market. 247 Wall St reported on August 20 that the 30-year fell to 5.184% and the 10-year to 4.637%, and stock futures and gold rose. Then Trump posted Wednesday night that he was launching an "ECONOMIC D-DAY" against Iran, the most crushing economic operation yet, and by Thursday the 30-year had climbed back past 5.26%, cancelling Bessent's move in under a day. Brent crude, already at $92 a barrel, pushed toward $94.
Buying your own long debt to hold down a yield carries a name from recent history: yield curve control, and its last great practitioner was the Bank of Japan. Tokyo capped its 10-year for years, speculators shorted it, and the BOJ held the line until the currency broke and the operation surrendered. This time the buyer is the fiscal arm of the state rather than the central bank, at a scale no government has tried.
The counter-example bites the other way, though. A determined buyer can hold a band for years, and what undoes it is not its own money but the terms of the fight — and Bessent's fight is with a war, not with hedge funds.
Part of the relief is imaginary. As Peter Boockvar put it, this is a rearrangement of the maturity schedule rather than a debt paydown — Fiscal Times reported on August 19 that Bessent will fund the buybacks by issuing more short-term bills, shifting what the government owes toward the very front end that is holding.
Rearranging does not shrink interest expense; it can grow it, because the Treasury borrows cheap at the short end to retire the expensive long debt whose price it is propping up. Evercore ISI's Krishna Guha told Fiscal Times on August 19 that the real change is none, since the need to finance huge deficits plus a tidal wave of new hyperscaler borrowing is untouched.
Who pays for it
The long holders take the first blow — Japan above all, and the foreign central banks, pension funds and insurers who must own duration and cannot simply dump a market the Treasury is crowding. Next come the borrowers Fiscal Times listed on August 19: the hyperscalers issuing corporate debt at a rapid clip to build data centers, and every American refinancing a 30-year mortgage. The yield you suppress on Thursday is the rate they pay on Friday.
The clearest profit sits at the front end. Money-market funds, banks and the Treasury's own short-dated account absorb the fresh bills that fund Bessent's buybacks, earning a policy rate the Fed refused to move at its July meeting. Warsh is the quietest winner of them all, as CoinDesk argued on August 19: the longer the cash rate stays frozen, the longer the war's inflation is measured against someone else's lengthening curve. Brent at $94 pays the producers who got there before the headline, as 247 Wall St noted on August 20.
The read is confirmed if the Treasury keeps expanding the buybacks operation by operation while the bills' share of its debt climbs, and if the 30-year retests 5.34% before the September 9 program even starts. It breaks if a genuine paydown appears — buybacks funded by maturing principal rather than fresh bills — or if Warsh relents and cuts, which would invert today's steepener and hand the long end back to the inflation it fears (Fiscal Times, Aug 19). None of that has happened yet.
A finance minister who buys his own longest bonds while the war pumps the oil that raises the cost of the next one is choosing which end of the curve pays for it.