The Treasury stepped into the bond market. It lasted one day.

🔵 On Wednesday morning last week, the U.S. Treasury Department did something it had not been expected to do. It announced that it would at least double the size of its bond repurchase operations for long-dated government debt, raising the maximum from $2 billion to at least $4 billion per operation beginning September 9. The announcement landed the same day total public debt crossed $40 trillion for the first time. Long-term yields fell immediately: the 30-year dropped from 5.26% to as low as 5.18%, the 10-year from 4.68% to 4.63%. It was, by any measure, a successful morning. And by Friday the entire move had been given back, with the 30-year closing at 5.27% — above where it started. What happened over those three days is a genuinely useful lesson about who sets long-term interest rates, and it matters this week because Federal Reserve Chair Kevin Warsh speaks at Jackson Hole on Friday. Here is the calm read.

Key points

• Treasury doubled its long-bond buybacks to at least $4 billion per operation from September 9.
• The 30-year yield fell to 5.18%, then finished the week at 5.27% — higher than before.
• Total public debt crossed $40 trillion for the first time on the same day.

What a buyback is, and what it isn't

The mechanics are simpler than the coverage suggests.

When the Treasury conducts a buyback, it purchases its own outstanding bonds from investors in the open market. That reduces the supply of those bonds available for trading, and when supply falls while demand holds, prices rise and yields fall. Treasury has run these operations routinely since 2024, primarily to support liquidity in less actively traded issues.

What changed last week was the size and the signal. The department targeted the 10-to-20-year and 20-to-30-year portions of the market — the stretch that Reuters described as having seen a buyers' strike since late June. Bessent followed the next day by telling CNBC that the $4 billion figure was a floor rather than a ceiling, that current yields "don't reflect market fundamentals," and that his department would "make a market" in longer-dated securities.

Three points worth holding:

• The sums are modest relative to the market. Treasury said it would repurchase up to $69 billion across all maturities between August 6 and November 5, rising to a maximum of $83 billion under the expanded plan. Against roughly $40 trillion outstanding, that is a small share.
• The underlying pressures did not change. As Anshul Sharma of Savvy Wealth put it, the move "doesn't solve the underlying issues around deficits, inflation, or Treasury supply. But it buys some time."
• The market tested it within 48 hours. By Thursday the 30-year was back at 5.27%, and it stayed there through Friday.

Why the long end keeps rising

The forces behind this are worth naming plainly, because none of them is a mystery and none is quickly solved.

Fixed-income strategists attribute the selloff that began in June to three things: a federal budget deficit on course to exceed last year's, inflation persistently above the Federal Reserve's 2% target, and an unusual volume of corporate bond issuance — much of it from technology companies raising money for artificial intelligence and data centre construction — competing with Treasuries for the same buyers.

To that list, last week added a fourth item: the Strait of Hormuz standoff remains unresolved, and Treasury Secretary Bessent is expected to outline further sanctions on Iran.

For households, the practical translation is straightforward. The 10-year yield, which sat near 4.65% on Friday, anchors mortgage rates, auto loans and much consumer borrowing. The 30-year matters more for annuity pricing and long-dated bond funds. Neither is set by the Federal Reserve.

🔍 The week that will test all of it

Three events land between Wednesday and Friday, and together they represent the densest calendar since July.

• Nvidia reports Wednesday after the close. It is the closest thing the artificial intelligence trade has to a quarterly referendum, and it arrives after a week in which the Nasdaq fell roughly 2% and semiconductor funds dropped more than 4%.
• The July personal consumption expenditures index publishes Wednesday morning, alongside a revision to second-quarter GDP. Core PCE is expected to hold at 3.3% annually, with the headline figure at 3.7%.
• Chair Warsh delivers his first Jackson Hole keynote as chair on Friday, nineteen days before the September rate decision.

One detail about Warsh's speech is worth knowing in advance: the advertised topic is payments innovation, not monetary policy. He has also spent his tenure removing forward guidance and shortening statements. Investors hoping for a clear signal about September may find the speech says less than they want, and the risk of over-reading a passing remark is real.

❓ Today's question

"Can the government actually control interest rates by buying its own bonds?"

Short answer: only at the margins, and only temporarily. Buying back bonds reduces the supply available to trade, which can push prices up and yields down — but the effect depends on the size of the purchases relative to the market, and last week's operation was small against roughly $40 trillion of outstanding debt. More importantly, the forces pushing long-term yields higher are structural rather than technical: deficits, inflation above target, and heavy competing bond issuance from corporations. A buyback can smooth a disorderly stretch of trading, which is a legitimate purpose, but it cannot change what investors think about repayment over thirty years. That is why the yield returned to its starting point within two days, and why several analysts described the move as buying time rather than solving anything.

There is a particular kind of financial news that arrives with the appearance of decisiveness and turns out to be a conversation rather than a conclusion. Last Wednesday looked like an intervention that worked — a headline, a sharp move, an apparent resolution. By Friday the number was back where it started, not because anyone failed but because the thing being addressed was never a two-day problem. The bond market is, in the end, several million people forming a view about what the United States will be able to repay over the next three decades, and no single announcement changes that view for long. That is not alarming. It is simply how a very large market works, and it is worth remembering the next time a decisive-sounding morning gets reported as a turning point. 🔔

Regards,
David Ellison

Found this information useful?