The US Treasury’s decision to double the size of its long-end bond buyback operations bought the market exactly one trading session of relief. By the close on Aug 20, 2026, most of the yield rally that followed Treasury Secretary Scott Bessent’s announcement had already been given back — a reminder that a debt-management tool sized in single-digit billions cannot durably offset a deficit sized in single-digit trillions.
The 10-year Treasury yield touched 4.72% on Aug 17, dipped to 4.65% on Aug 19 as the upsized buyback landed, and closed back at 4.69% on Aug 20 (FRED DGS10). The 30-year followed a similar path: 5.31% → 5.19% → 5.23% over the same three sessions (FRED DGS30). More than half of the 10-year rally is gone. About a third of the 30-year rally is gone. And nothing on the short end moved at all.
Five sessions, four tenors, one round-trip
| Date | 2Y | 5Y | 10Y | 30Y | Note |
|---|---|---|---|---|---|
| Mon Aug 17 | 4.19% | 4.38% | 4.72% | 5.31% | 30Y at cycle high |
| Tue Aug 18 | 4.19% | 4.37% | 4.71% | 5.28% | Buyback upsize announced |
| Wed Aug 19 | 4.19% | 4.35% | 4.65% | 5.19% | Buyback op — yields fall |
| Thu Aug 20 | 4.19% | 4.39% | 4.69% | 5.23% | Rally starts to fade |
| Move Mon → Thu | 0 bp | +1 bp | −3 bp | −8 bp | Almost a round-trip |
What actually happened this week
Bessent’s announcement on Aug 18 — recapped in our earlier note on the doubled long-end buybacks — raised the per-operation cap for 10– to 30-year Treasuries to about $4 billion, roughly twice the prior pace. The stated purpose is to add liquidity in off-the-run long paper, but the market read it (correctly, in the short term) as a signal that Treasury would lean against the recent selloff.
The Aug 19 operation delivered. The 10-year rallied 7 basis points from Monday’s high, the 30-year rallied 12 bp, and equity markets steadied as the bond-yield discount rate stopped rising. Twenty-four hours later, most of that comfort was gone. The 30-year at 5.23% is only 8 bp below Monday’s high, and it sits again above the 5.20% threshold that had been treated all summer as the pain point where risk assets start to wobble.
Why the fix couldn’t hold
The math is unforgiving. Treasury bought roughly $4 billion of long-end paper on the upsized operation. The US has about $28 trillion of Treasury securities outstanding held by the public, and the deficit has been running at roughly $400 billion a month in FY2026. A single-op purchase equivalent to about 1% of a single month’s deficit was never going to durably reprice a $28 trillion market — it can only smooth around the edges, and that is what the Aug 19 print looked like.
The underlying drivers of the summer’s selloff have not changed. US national debt just crossed $40 trillion. Core CPI is still tracking above the Fed’s 2% target. The New York Fed’s ACM model has term premium at its highest reading since 2015 (NY Fed ACM tabs). Long-end buyers are being asked to hold duration in a fiscal regime that keeps supplying more of it, and they are pricing accordingly.
The two-year didn’t blink
The clean tell in this week’s data is the 2-year yield, which sat at 4.19% every single day from Aug 17 through Aug 20. The two-year part of the curve reflects the market’s forecast of the Fed’s policy path, not Treasury’s debt-management choices. Bessent’s buyback did not change any Fed-related fundamental, and the 2-year correctly ignored it.
That is why the buyback’s footprint on the curve is a bear-steepening pattern: long yields fell, then rose again, while short yields did nothing. The 2s10s spread ended Aug 21 at 50 basis points (FRED T10Y2Y), only slightly flatter than the 53 bp reading at the Aug 17 peak — and materially steeper than the 46 bp trough hit on Aug 19 when the buyback bid was strongest.
What to watch next
- The next buyback op. The market will look for whether Treasury sustains the $4B pace on the next scheduled operation, or trims it back once the tape stabilizes. Sustained upsizing signals commitment; a trim signals it was a one-off gesture.
- Q1 refunding announcement. The next quarterly refunding will confirm coupon-size guidance across 2Y, 5Y, 10Y, 20Y, and 30Y auctions. Any hint of tilting more issuance to bills and less to coupons would be reinforcing of the buyback story.
- Jackson Hole tone. Fed communications through the annual symposium will reset the front end of the curve. A dovish shift could rally 2s and steepen the curve further; a hawkish shift would flatten it and add to the term-premium pressure at the long end.
- Auction tails at 20Y and 30Y. The clearest live-fire test of long-end demand is the next 20-year and 30-year auction stats: bid-to-cover, tail size, and indirect (foreign) bidder share. Weak stats would confirm the buyback couldn’t restore the natural bid.
Bottom line
Treasury proved this week that it can pin long yields for a session. It also proved that a fiscal-driven selloff cannot be permanently offset by a debt-management tool an order of magnitude smaller than the flow of new supply. The 10-year is once again within 3 bp of Monday’s high; the 30-year is again above 5.20%. Anyone waiting for the long end to grind lower without a change in the fiscal or inflation story is fighting the math on the tape.
Sources
- FRED — 10-Year Treasury Constant Maturity Rate (DGS10)
- FRED — 30-Year Treasury Constant Maturity Rate (DGS30)
- FRED — 2-Year Treasury Constant Maturity Rate (DGS2)
- FRED — 5-Year Treasury Constant Maturity Rate (DGS5)
- FRED — 10Y minus 2Y Treasury yield spread (T10Y2Y)
- Federal Reserve Board — H.15 Selected Interest Rates
- US Treasury — Quarterly Refunding page
- Federal Reserve Bank of New York — ACM Term Premia estimates
Disclosure: This article is for informational purposes only and is not investment advice.