Global Rates
US Treasury buybacks
$2bn → $4bn
Long-end liquidity ops at least doubled, effective Sep 9 through Nov 4
US 30Y
5.196%
−9bp Wednesday after 5.34% Tuesday, a nineteen-year high
US 10Y
~4.64%
−6bp to 4.647% on the buyback news; 16bp from pos-017's barrier
JGB 10Y
2.895%
−4bp after touching 2.945% Aug 18, a ~30-year high, on early-hike speculation
BoJ September
Hike consensus
MUFG sees 1.25% in September; DBS sees the pace accelerating. We flagged it at 40c and never traded it
Brent crude
<$92
+5% over four sessions with Hormuz shut — and global long ends still fell
Two long ends hit multi-decade highs on Tuesday and both backed off on Wednesday, and only one of them was rescued. In Tokyo, the newly issued 10-year JGB briefly touched 2.945% on Aug 18, its highest in roughly thirty years, on speculation of an early Bank of Japan hike, then fell 4bp to 2.895% on Aug 19. In Washington the 30-year printed 5.34% on Tuesday, a nineteen-year high, and then fell 9bp to 5.196% on Wednesday — but that move had an author. The Treasury announced it is at least doubling the maximum size of its liquidity-support buyback operations in longer-dated nominal coupons, from $2bn to at least $4bn per operation across the 10-to-20-year and 20-to-30-year sectors, effective September 9 and running through November 4. The 10-year fell about 6bp to 4.647% on the news and trades near 4.64% this morning. This is the global rates story of the week and it is a governance story as much as a market one: a Treasury Secretary reaching for buybacks because the long end is at a nineteen-year high, three weeks before a Fed Chair who has publicly said he will not be constrained by market prices gives his first Jackson Hole keynote. The direct cost to this book is that pos-017, our bet that the US 10-year touches 4.80% before 2027, now has a large official buyer standing between it and its barrier. Fair value comes down from 74 to 60 and we now mark the position above our own fair value. The counter-evidence is what makes the correction necessary rather than optional: Brent traded below $92 after gaining more than 5% across four sessions, the Strait of Hormuz remains shut with eight vessels attacked in transit this month, and Trump announced fresh measures against Iran's economy on Wednesday — and US yields fell anyway. Oil up, term premium down. The mechanism we published failed a clean test. Japan is the other half of the picture and the half we are still not in. Consensus has hardened around a September BoJ hike, with MUFG looking for 1.25% in September and DBS expecting the pace to accelerate to a quarter-point every three to four months. We flagged that trade at 40c on Aug 3, declined to chase it at 42.5c on Aug 10, watched it print 72c on Aug 13, and retired it as a completed miss. We are not re-opening it and we are not pretending it away. DISCLOSURE: no letter Friday Aug 14, Monday Aug 17, Tuesday Aug 18 or Wednesday Aug 19 — four consecutive missed sessions covering both long-end records and the Treasury's response. No global trade. Book: 5 open, $332 staked, +$87.91 unrealized, +$611.08 realized, 11/14.
Today's Market Moves
10Y Touches 4.8% (pos-017)
68.0%→64.0%-4.0pp
FV cut 74 → 60. The US Treasury at least doubled long-end buyback operations to $4bn per operation for the 10-20y and 20-30y sectors, effective Sep 9 to Nov 4, and the 10-year fell 6bp to 4.647% while Brent rose. Raw barrier maths on 16bp over ~93 sessions at 4bp daily vol still gives 68%; the official-buyer haircut takes it to 60 against a 64 mid. Take-profit gate lowered from a 72c bid to a 66c bid.
US 30Y term premium
5.34%→5.196%-0.144pp
Tuesday's 5.34% was the highest since 2007 and it is what triggered the policy response. Read as a signal, the sequence says the US long end reached a level the Treasury was unwilling to tolerate — which is bullish for yields structurally and bearish for them tactically, and our barrier bet is a tactical instrument in a four-month window that is now two-thirds covered by the buyback programme.
JGB 10Y
2.945%→2.895%-0.05pp
A thirty-year high on Aug 18, then 4bp lower on Aug 19. Unlike the US move this had no official author — it was simply early-hike speculation cooling slightly. Japan's long end is now doing what ours is being prevented from doing, which is the cleanest available illustration of why the pos-017 haircut is about policy and not about the global rate cycle.
BoJ September hike (retired miss)
72.0%→72.0%0.0pp
Carried at the Aug 13 price — we did not re-quote this market at source in this session and we are flagging that rather than implying a fresh reading. The fundamentals have hardened, not softened: MUFG expects 1.25% in September and DBS expects the BoJ to accelerate to a quarter-point every three to four months from roughly twice a year. This stays retired as a completed miss; our published re-entry gate of sub-30c is unreachable.
Fed Sep hike (context)
32.5%→29.0%-3.5pp
Polymarket at 29c against CME FedWatch's 34.6% for the Sep 15-16 meeting, a 5.6pp gap, verified at both sources today. Under the 10pp bar and correlation-capped. Warsh's first Jackson Hole keynote as Chair lands Aug 28, nineteen days before the decision, and he has said his remarks will be structural rather than near-term guidance.
Screening Table
| # | Market | Expiry | Market Price | Fair Value | Gap (pp) | Direction | Volume | Confidence |
|---|---|---|---|---|---|---|---|---|
| 1 | 10Y Touches 4.8% Before 2027 | Dec 31 | 64.0% | 60% | -4pp | BUY YES | $$286K total | 6/10 |
| 2 | Fed Rate Hike in 2026 | Dec 9 | 50.0% | 55% | +5pp | HOLD $100 runner — floor at 48c is two points away | $$7.7M total | 6/10 |
| 3 | Fed Sep Hike | Sep 16 | 29.0% | 34.6% | +5.6pp | NO TRADE — under bar, correlation cap | $$869K total | 4/10 |
| 4 | BoJ September Hike (retired miss) | Sep 18 | 72.0% | 72% | 0pp | NO TRADE — retired; price carried from Aug 13, not re-verified today | $$2.8K/day | 3/10 |
| 5 | Fed Funds End 2026 = 4.0% | Dec 8 | 29.6% | 30% | +0.4pp | HOLD $25 — thesis confirmed, edge closed | $$1.37M total | 5/10 |
Top 5 Opportunities
1
US 10Y Touches 4.8% Before 2027 — the mechanism failed — YES
↑ BUY YES-4pp
Market price
64.0%
Fair value
60%
Gap: -4pp
The global-rates version of this correction is worth writing separately from the US one, because from here the failure is easier to see. Across four sessions Brent gained more than 5% to trade just below $92, the Strait of Hormuz stayed shut, eight vessels were attacked in transit during August, and Washington announced fresh measures against Iran's economy. Every one of those is a term-premium input under the thesis we have published on this position since July. The US 10-year fell. What moved it instead was the Treasury announcing that from September 9 through November 4 it will run long-end liquidity-support buybacks of at least $4bn per operation, double the prior maximum, in the 10-20y and 20-30y sectors. That is a bid of size in the sectors adjacent to the one our barrier needs. Compare Tokyo over the identical window: the 10-year JGB touched 2.945% on Aug 18, a thirty-year high, with no official counterparty leaning against it, and only cooled because hike speculation cooled. Same global cycle, different policy reaction function, opposite outcomes. That comparison is why we are haircutting for policy rather than downgrading our view of the cycle. Barrier maths: 16bp to travel, roughly 93 sessions, about 4bp of daily vol, giving 68%. Haircut to 60 for the buyback window. Market 64 mid, 58 bid, 70 ask. We hold, we do not add, we do not sell below fair value, and we lower the published take-profit gate from a 72c bid to a 66c bid.
▵ Bull case
- The buyback programme expires November 4 — December is unprotected and the barrier runs to Dec 31
- Raw barrier maths is still 68% with 16bp to go
- 4.75% on July 31 shows the level is reachable in this regime
▿ Bear case
- Oil up, yields down: the published mechanism failed a clean test this week
- An official buyer of size in adjacent sectors for two of the remaining four and a half months
- Fifth FV revision in six weeks — instability we are not going to dress up as conviction
2
Bank of Japan September Hike (retired miss — status update) — YES
↑ BUY YES0pp
Market price
72.0%
Fair value
72%
Gap: 0pp
We keep this on the page rather than letting it disappear, because the honest accounting of a miss does not end on the day you admit it. First, a transparency note that matters more than the content: the 72c here is CARRIED from Aug 13. We did not re-quote this market at source in this session, and under our own Aug 6 rule an unverified price is not a tradeable number, so it is presented as a carried mark and nothing is decided from it. What we did verify today is the fundamental direction, and it has hardened. MUFG Research expects the BoJ to take the policy rate to 1.25% in September, then 1.50% in January 2027 and 1.75% in June 2027. DBS now expects the BoJ to hike in September and to accelerate to roughly one quarter-point move every three to four months, up from about twice a year. The 10-year JGB touched 2.945% on Aug 18, a thirty-year high, on exactly this speculation. Our thesis from Aug 3 was correct in every respect except the one that pays, which is having a position on. The published re-entry gate of sub-30c will not be reached and we are not moving it, because moving a gate you set to protect yourself from chasing is just chasing with extra steps.
▵ Bull case
- Sell-side consensus has hardened toward a September hike and a faster path thereafter
- JGB 10Y at a thirty-year high confirms the market agrees
- Our Aug 3 analysis was right on the fundamentals
▿ Bear case
- Price is carried from Aug 13, not verified today — not a tradeable number
- At 72c the edge is gone; re-entry gate of sub-30c is unreachable
- The failure was attendance, not analysis, and re-entering now would not fix that
3
US 30Y Term Premium (context, not traded) — n/a
↑ BUY YES0pp
Market price
5.196%
Fair value
5.196%
Gap: 0pp
Not a position and not tradeable on our venue, but it is the variable that explains this week, so we mark it explicitly. The 30-year reached 5.34% on Tuesday, the highest since 2007, and fell 9bp to 5.196% on Wednesday when the Treasury doubled its buyback capacity. The interpretation cuts two ways and we will not pretend it only cuts one. Structurally it is a bullish signal for yields: a Treasury that has to reach for buybacks at 5.34% is telling you the market is struggling to absorb duration at current prices, which is the condition under which our 4.80% barrier eventually gets hit. Tactically it is bearish, because the intervention lands squarely inside our window, September 9 to November 4, and that window contains the seasonally heaviest issuance stretch of the year. Our barrier is a time-limited instrument, so the tactical read dominates and the fair value comes down. If we could express the structural view on a five-year horizon we would; we cannot, so we mark the position honestly instead.
▵ Bull case
- 5.34% on a 19-year high says duration absorption is genuinely strained
- Buybacks are a liquidity tool, not a purchase programme — they do not change net supply
- The programme ends Nov 4, before our barrier expires
▿ Bear case
- The intervention covers the heaviest issuance stretch of our remaining window
- It signals a policy authority willing to act against long-end yields, which is a new standing risk
- Not directly tradeable for us — we can only mark it, not hedge it
4
Fed September Hike (screened — refused for a third time) — YES
↑ BUY YES+5.6pp
Market price
29.0%
Fair value
34.6%
Gap: +5.6pp
Polymarket 29c, CME FedWatch 34.6% as of Wednesday, both verified in this session. A 5.6pp gap against a published 10pp bar is a mechanical refusal, and this is the third consecutive letter in which this exact ticket has screened and been declined. The two prior refusals were both right — the Aug 6 version, refused because we could not verify the quoted gap at source, was worth about 15 points, and the Aug 13 version at 32.5c has been worth another 3.5. We flag that because last week's letter led with a refusal that cost us $41.67 in forgone profit, and a rule that only gets discussed when it hurts is not being reported honestly. The independent blocker is concentration: four of five open positions pay off if the Fed does not cut and probably hikes. Adding a fifth would be leverage on a view we already hold, not diversification. The genuine catalyst is Warsh's Jackson Hole keynote on Aug 28, his first as Chair, nineteen days before the decision — he has signalled the speech will be structural rather than near-term guidance, and that under his chairmanship the Fed no longer telegraphs meetings in advance, which raises the variance around September rather than the level.
▵ Bull case
- 5.6pp gap verified at two independent sources today
- $869K of volume — genuinely sizeable if the gap ever widened
- Warsh's Aug 28 debut is a real pre-meeting repricing catalyst
▿ Bear case
- Under the 10pp bar; the bar does not bend for interesting setups
- Correlation cap blocks it independently — fifth hawkish Fed ticket
- A Fed that does not telegraph raises variance, which argues for smaller bets, not more
5
Fed Funds End 2026 = 4.0% (called and confirmed) — YES
↑ BUY YES+0.4pp
Market price
29.6%
Fair value
30%
Gap: +0.4pp
Seven sessions ago this bracket was the worst thing in the book at 22.65c and we published the reason we thought it was cheap rather than broken: hike-in-2026 at 54.5, the 4.25%-and-above brackets summing to roughly 20, about 4 points for a hike that later gets cut, leaving exactly-one-hike-and-hold reconciling near 30. It trades 29.6 today. We report that with the same prominence we gave the BoJ miss and last week's $41.67 refusal, because a process that only publishes its outcomes when they are bad is as dishonest as one that only publishes the good ones. The bracket structure today: 3.75% at 40.9%, 4.0% at 29.6%, 4.25% at 14.1%, 3.5% at 8.8%, and 4.5%-or-higher at 5.1% — a distribution centred exactly where a committee with three hawkish dissents and an oil shock would put it. Being right closes the trade: 0.4pp of edge is nothing, we add nothing, and we arm an efficiency review at 40c.
▵ Bull case
- Pre-published reconciliation at 30 was reached within seven sessions
- Bracket distribution is internally coherent across the whole ladder
- $25 ticket — cheap to hold to resolution
▿ Bear case
- Edge is 0.4pp; the trade is finished even if the position is not
- Under $100/day of volume makes exiting expensive
- Another Fed ticket in an over-Fed book