Global Rates
ECB Sep hike
92c
25bp increase on Sep 10 almost fully priced; ECB hiked in June, first rise since 2023
BoJ Sep hike
84.5c
Verified at source today, up from the 72c we carried unverified yesterday — our retired miss keeps running
Fed Sep hike
27.5c
Down from 29c. The one major central bank the market does not expect to move next month
US 10Y / 30Y
4.691% / 5.24%
Buyback relief fully erased in a session; 30Y back from 5.196% to 5.25%
Bund 10Y
3.27%
Highest since 2011; Germany selling 30-year paper at the highest yield since 2011
JGB 10Y
~2.83%
Eased after 2.945% on Aug 18 — first time above that level since September 1996
Europe and Japan are hiking and the market barely doubts it; America is the one everybody watches and its September meeting is priced at 27.5c. That divergence is the global-rates story today, and it is visible in one place: Polymarket now prices a 25bp ECB increase on Sep 10 at 92c and a 25bp Bank of Japan increase on Sep 18 at 84.5c, both verified at source in this session, against 27.5c for a Fed hike on Sep 16. Three developed-market central banks, the same oil shock, and two of them are already moving. The bond markets have noticed. Germany's 10-year Bund reached 3.27%, its highest since 2011, as Berlin prepared to sell 30-year paper at the highest yield since that same year, on fading hopes of a quick Iran resolution and $93 Brent. Japan's 10-year JGB touched 2.945% on Aug 18, its first print above 2.94% since September 1996, before easing to around 2.83%. And the American long end did something more interesting than either: it took a policy intervention head-on and won. On Wednesday the US Treasury announced it would at least double long-end liquidity-support buybacks to $4bn per operation for the 10-20y and 20-30y sectors, effective Sep 9 through Nov 4; the 10-year fell 6bp to 4.647% and the 30-year 9bp to 5.196%. By Thursday's close the entire move was gone — 10-year back to 4.70%, 30-year back to 5.25% — even with Treasury Secretary Scott Bessent saying the operations could exceed the announced $4bn per issue. This morning the US 10-year is 4.691%. This has a direct and uncomfortable consequence for this book. Yesterday's letter cut fair value on pos-017, our bet that the US 10-year touches 4.80% before 2027, from 74 to 60, on the reasoning that an official buyer of size now stood between the barrier and its target. That reasoning lasted one session. We put it back up: FV 60 to 70, because the barrier is now 10.9bp away rather than 15.3bp and because a buyback is a secondary-market liquidity operation that does not retire net supply. Two double-digit fair-value revisions in two days on one position is instability, not agility, and we are naming it. The Bank of Japan hike we flagged at 40c on Aug 3 and declined to chase at 42.5c on Aug 10 now trades at 84.5c — verified today at source, up from the 72c we carried unverified yesterday. It stays retired as a completed miss and the number keeps getting worse in public, which is the correct place for it. No global trade. Book: 5 open, $332 staked, +$93.74 unrealized, +$611.08 realized, 11/14.
Today's Market Moves
10Y Touches 4.8% (pos-017)
64.0%→66.5%+2.5pp
FV reversed 60 -> 70, one day after we cut it 74 -> 60. The cut rested on the US Treasury's doubled long-end buybacks suppressing the term premium; Thursday erased the entire announcement move, taking the 10-year to 4.70% and the 30-year to 5.25% even as Bessent said operations could exceed $4bn per issue. Barrier maths: 10.9bp over ~91 sessions at ~4bp daily vol = 77.5% raw, reduced policy haircut to 70 against a 66.5 mid (63 bid / 70 ask). Take-profit gate stays a 66c bid and will not be raised back to 72c.
BoJ September hike (retired miss)
72.0%→84.5%+12.5pp
Now verified at source rather than carried — yesterday we flagged the 72c as an unverified mark from Aug 13 and today's reading is 84.5c, with No Change at 14.5c. The miss is 44.5 points deep from the 40c where we flagged it on Aug 3. MUFG looks for 1.25% in September and DBS expects the pace to accelerate to a quarter-point every three to four months. Our published sub-30c re-entry gate is unreachable and we are not moving it.
ECB September hike (new to the page)
0.0%→92.0%+92.0pp
We have not carried this market before and we are adding it as context, not as a trade. Polymarket prices a 25bp ECB increase on Sep 10 at 92c against 8.5c for no change. The ECB hiked a quarter point in June, its first rise since 2023, and Lagarde has warned that renewed Middle East hostilities pose upside risk to the inflation outlook. At 92c there is no edge for anyone; the value here is as the cleanest available statement that the euro-area oil shock is being answered with policy.
Bund 10Y
3.2%→3.27%+0.07pp
Highest since 2011, with Germany selling 30-year paper at the highest yield since that year. The driver is the same one driving our US position: fading hopes of a quick Iran resolution, Brent near $93, and investors demanding more compensation to lend to indebted governments. Europe's long end is doing without intervention what the US long end did in spite of one.
Fed Sep hike (context)
29.0%→27.5%-1.5pp
Refused for a fourth time, and today on the cleanest ground: the CME FedWatch anchor of 34.6% is CARRIED from Aug 19 and could not be re-verified at source in this session, which under our Aug 6 rule makes it untradeable regardless of the implied 7.1pp gap. The gap is under the 10pp bar in any case and the correlation cap blocks a fifth hawkish-Fed ticket independently. Warsh's first Jackson Hole keynote as Chair is Aug 28.
Screening Table
| # | Market | Expiry | Market Price | Fair Value | Gap (pp) | Direction | Volume | Confidence |
|---|---|---|---|---|---|---|---|---|
| 1 | US 10Y Touches 4.8% Before 2027 | Dec 31 | 66.5% | 70% | +3.5pp | HOLD $25 YES — FV reversed 60 -> 70, gap under the bar | $$60K on the leg | 6/10 |
| 2 | BoJ September Hike (retired miss) | Sep 18 | 84.5% | 84.5% | 0pp | NO TRADE — retired; price verified at source today, no edge | $$114K on the leg | 3/10 |
| 3 | ECB September Hike | Sep 10 | 92.0% | 92% | 0pp | NO TRADE — fully priced; carried as context only | $$91K on the leg | 3/10 |
| 4 | Fed Sep Hike | Sep 16 | 27.5% | 34.6% | +7.1pp | NO TRADE — FV is a CARRIED number; under bar; correlation-capped | $$8.8M on the leg | 3/10 |
| 5 | Fed Rate Hike in 2026 | Dec 9 | 49.5% | 55% | +5.5pp | HOLD $100 runner — 1.5 points above the 48c floor | $$7.7M total | 6/10 |
Top 5 Opportunities
1
US 10Y Touches 4.8% Before 2027 — the intervention that lasted one day — YES
↑ BUY YES+3.5pp
Market price
66.5%
Fair value
70%
Gap: +3.5pp
The global-rates version of this correction is the more useful one, because the comparison across three markets shows why we were wrong. On Wednesday the US Treasury announced it would at least double long-end liquidity-support buybacks to at least $4bn per operation in the 10-20y and 20-30y sectors, running Sep 9 through Nov 4. The 10-year fell 6bp to 4.647%, the 30-year 9bp to 5.196%, and we responded by cutting fair value on this position from 74 to 60 — a fourteen-point write-down on the theory that an official buyer of size had taken control of the sector our barrier needs. By Thursday's close the whole move was gone: the 10-year rebounded to 4.70%, within 5bp of this week's 20-month high of 4.75%, and the 30-year returned to 5.25%. Bessent had by then said the accelerated operations could exceed $4bn per issue and pointed to a coming fiscal plan, and it did not hold the curve. Look at the other two long ends over the same week for the reason. Germany's 10-year Bund is at 3.27%, the highest since 2011, with 30-year paper being sold at the highest yield since 2011, and no one is intervening. Japan's 10-year JGB touched 2.945% on Aug 18, the highest since September 1996, and no one is intervening there either. What is repricing global duration is not a US-specific liquidity problem that buybacks can address — it is deficits, heavy issuance including AI-related corporate supply, and $93 oil with the Strait of Hormuz shut. A secondary-market buyback does not retire net supply and cannot stand against that. We should have priced the intervention as noise inside a global repricing rather than as a new actor, and we did not. The correction: barrier now 10.9bp away, roughly 91 sessions left, ~4bp daily vol, sigma 38.2bp, 2*(1-Phi(0.2856)) = 77.5% raw. We keep a reduced policy haircut because the programme is real and dated, and take fair value to 70 against a 66.5 mid.
▵ Bull case
- A dated, announced, talked-up intervention failed to hold the curve for one full session
- Bund at a 2011 high and JGB at a 1996 high — this is a global repricing, not a US liquidity event
- Barrier 10.9bp away, 91 sessions, 77.5% on raw maths
- Brent near $93 with Hormuz shut and new US sanctions on Iran being prepared
▿ Bear case
- We revised fair value by 14 points down and 10 points up in two sessions — that is instability
- The buyback programme still exists and still runs Sep 9 to Nov 4
- 63 bid / 70 ask: a 7-point spread makes the 66.5 mid notional
- The revision runs in the direction that flatters the book, which deserves extra suspicion
2
Bank of Japan September Hike (retired miss — now verified, and worse) — YES
↑ BUY YES0pp
Market price
84.5%
Fair value
84.5%
Gap: 0pp
Yesterday we published this market at 72c and labelled the number CARRIED from Aug 13, because we had not re-quoted it at source. We quoted it at source today: a 25bp September increase trades at 84.5c, with no change at 14.5c and a 50bp move at 0.85c. So the honest update is that the miss is 12.5 points deeper than the stale number we printed, and 44.5 points deep from the 40c at which we flagged the trade on Aug 3. We declined to chase at 42.5c on Aug 10, watched 72c on Aug 13, and it is 84.5c today. There is no way to write that which makes it look better, so we simply keep printing it. The fundamentals have moved with the price: MUFG expects the policy rate at 1.25% in September, 1.50% in January 2027 and 1.75% in June 2027; DBS expects the BoJ to accelerate from roughly two moves a year to one every three to four months. The 10-year JGB touched 2.945% on Aug 18, its highest since September 1996, on precisely this expectation, before easing to about 2.83%. Our published re-entry gate is sub-30c. It will not be reached, and we are not moving it, because a gate whose purpose is to stop you chasing is worthless the moment you move it to permit a chase.
▵ Bull case
- Sell-side consensus has hardened around September and a faster path after it
- JGB 10Y at a 1996 high confirms the market agrees
- Our Aug 3 fundamental analysis was right in every respect except owning it
▿ Bear case
- 84.5c leaves 15.5 points of upside — no edge and no reason to be here
- The re-entry gate of sub-30c is unreachable and stays where it is
- The failure was attendance and hesitation, not analysis, and entering now would not fix either
3
ECB September Hike (new context market) — YES
↑ BUY YES0pp
Market price
92.0%
Fair value
92%
Gap: 0pp
We are adding this to the page for the first time, and the reason is comparative rather than tradeable. Polymarket prices a 25bp ECB increase on Sep 10 at 92c against 8.5c for no change and effectively zero for a cut. The ECB delivered a quarter-point hike in June, its first increase since 2023, driven by the energy shock from the Iran war, and Lagarde has since warned that renewed Middle East hostilities pose upside risk to the euro-area inflation outlook. The ECB's own projections have headline inflation averaging 3.0% in 2026 before easing to 2.3% in 2027. Set that beside a market pricing the Fed's September meeting at 27.5c. Three central banks, the same $93 Brent, the same closed Strait of Hormuz, and the euro area is 92% priced to move, Japan 84.5%, the United States 27.5%. We are not asserting that the Fed is mispriced — the US has a different labour market, a different energy exposure and a chairman who has stopped telegraphing meetings. But we are recording the divergence, because it is the most striking feature of global rates this week and because it is the environment in which our one non-Fed position, a bet on the US long end, has to work. At 92c there is nothing to buy and we are not pretending otherwise.
▵ Bull case
- ECB already hiked in June, its first since 2023 — the direction is established
- Lagarde has explicitly flagged Middle East energy as an upside inflation risk
- Bund at a 2011 high says the bond market agrees
▿ Bear case
- 92c: no edge for anyone, and we do not trade fully priced markets
- Thin volume on the leg relative to the US Fed markets
- Adds nothing to the book except context
4
Fed September Hike (screened — refused for a fourth time) — YES
↑ BUY YES+7.1pp
Market price
27.5%
Fair value
34.6%
Gap: +7.1pp
The refusal comes with a data-quality disclosure before any analysis. Our fair value of 34.6 is CME FedWatch as of Aug 19. We attempted to re-verify it at source in this session and could not obtain a current numeric reading, so it is CARRIED. Under the rule published on Aug 6, an unverified price is not a tradeable input, and that settles the question before the 10pp bar or the correlation cap are consulted. Both of those also block it: the notional gap is 7.1pp against a 10pp bar, and this would be a fifth open position paying off on a hawkish Fed in a book of five. What deserves attention is the direction of travel. Since Aug 13 the September hike has fallen from 32.5c to 27.5c, and it has done so through a week in which the US 10-year rose to 4.70%, Brent gained more than 4% to $93, initial claims beat at 206K, and Philadelphia Fed factory growth hit its strongest since 2021 — while the ECB's September hike sits at 92c and the BoJ's at 84.5c on the same energy shock. One reading is that the equity market's 1% decline on borrowing-cost fears is doing the tightening for the Fed. Another is that a committee chaired by Kevin Warsh, who has said he will not be constrained by market pricing and has stopped pre-announcing decisions, is simply harder to forecast, which flattens the price of every individual meeting outcome. We do not know which, and we do not need to.
▵ Bull case
- $8.8M of volume on the leg — genuinely sizeable if a verified gap ever appeared
- Warsh's first Jackson Hole keynote as Chair on Aug 28 is a real pre-meeting catalyst
- Global peers at 92c and 84.5c on the same shock make 27.5c look conspicuous
▿ Bear case
- Fair value is CARRIED from Aug 19 and untradeable under our own rule
- 7.1pp is under the 10pp bar even taken at face value
- Correlation cap blocks it independently
- A market falling steadily into supportive data is usually pricing something we have not identified
5
Fed Rate Hike in 2026 (runner) — the floor is 1.5 points away — YES
↑ BUY YES+5.5pp
Market price
49.5%
Fair value
55%
Gap: +5.5pp
Included here because the global comparison sharpens the puzzle. The full-year US hike market fell half a point to 49.5c in a week when every input that should support it firmed: the 10-year rebounded to 4.70% through a Treasury intervention, Brent went to $93, claims beat at 206K, and factory activity in Philadelphia hit a five-year high. Meanwhile the euro area is 92% priced to hike in three weeks and Japan 84.5% priced to hike in four, on the same oil shock. The internal US structure is at least coherent with itself: September 27.5c, hike-by-October 40.5c, full-year 49.5c, leaving roughly 9 points of December-only hike. Fair value stays 55 — three FOMC members dissented toward a hike in July, and oil is the dominant input to the inflation the committee is fighting. The gap is 5.5pp, under the bar, so no add. What matters operationally is the floor: we published 48c on June 10, the market is 49.5c, and if it trades below 48 we sell the runner and print the result the same day. Naming it in advance is the whole point.
▵ Bull case
- Every hike input firmed this week: yields, oil, claims, factory activity
- ECB and BoJ are already moving on the same shock
- Term structure across September, October and full-year reconciles cleanly
▿ Bear case
- 1.5 points above a published floor we will honour in public
- Price falling into supportive data is a warning, not an opportunity
- US consumer is the divergence: July retail sales -0.6%, equities down 1% on rate fears