Daily Macro US
Brent crude
$84.90
-4.2% today, -10.0% in three sessions from $94.39 Friday. Iran-Oman framework for a temporary Hormuz corridor plus joint mine-clearing. The single most important number in this letter
10Y Treasury
4.639%
-6.5bp, Aug 25 settle. 30Y 5.174% (-5.7bp). Widens the pos-017 barrier from 9.5bp to 16.1bp and cuts our fair value by eleven points
10Y touch 4.8% (pos-017)
65.5c
-1.5, bid 64 / ask 67. FV CUT 72 -> 61 on the same model with a wider barrier. We are now marked 4.6pp ABOVE our own number
Fed hike 2026
56.5c
Flat, and that flatness is the surprise. FV cut 55 -> 53 on the oil move. Fed Sep 32.5c (-1.0), ECB Sep 96.75c (+1.25), BoJ Sep 86.0c (-1.5)
July PCE
Not yet released
08:30 ET today with Q2 GDP second estimate and durable goods. FactSet consensus +0.07% MoM headline, core +0.18%. Cleveland Fed re-stamped 08/25, unchanged, core July 0.25% MoM / 3.29% YoY
Book
+$100.82
5 open, $332 staked; +$611.08 realized, 11/14. Down $2.03 on the day. No trades
The war premium is coming out of the tape, and our book is on the wrong side of that in two places at once. Iran and Oman have agreed a framework for a temporary joint maritime corridor through the Strait of Hormuz, together with a joint mine-clearing project, with Oman's foreign minister saying he hopes to announce the interim corridor soon. Brent has gone 94.39 on Friday, 92.17 Monday, 88.58 Tuesday, and 84.90 as we write: down 10.0% in three sessions and 4.2% today alone. The 10-year settled at 4.639% last night, 6.5 basis points lower, and that single number takes eleven points off our own fair value on pos-017. This is the correlated drawdown we have been flagging in the portfolio note for weeks, arriving in the ordinary way: not as a crash, but as one piece of diplomatic news that moves oil, yields and rate-hike odds in the same direction, because all three legs of this book run off the same barrel. Second item, and it is a correction to our own work: two numbers we published yesterday do not reconcile with the settles. We gave Brent at $91.58 for Aug 25; the settle was $88.58. We gave the 10-year at 4.7050%; the settle was 4.639%. Both were intraday marks published without being labelled as such, which is our error and not a rounding difference. Today's figures are settles where a settle exists and are labelled where they are not. Third, the timing of this letter. We are writing at 04:47 ET, roughly four hours before July PCE, the Q2 GDP second estimate and durable goods all land together at 08:30. We do not have those numbers and we are not going to write as though we do. What follows prices the tape as it stands pre-release, and every fair value in it is provisional against a release we have not seen.
Today's Market Moves
Brent crude
88.58%→84.9%-3.68pp
Not a market we hold, and the most important line in the table. Three sessions, -10.0%, on a diplomatic framework rather than a barrel actually moving: Iran and Oman have agreed to establish a temporary joint maritime corridor through Hormuz and to run a joint mine-clearance project, with technical talks continuing toward a permanent arrangement. Nothing has physically transited yet. The market is pricing the removal of a risk, not the arrival of supply, which is exactly the kind of move that can reverse on one bad headline - and exactly the kind we are not allowed to assume will reverse just because it suits our book.
10Y Touches 4.8% Before 2027 (pos-017)
67.0%→65.5%-1.5pp
Bid 64 / ask 67. Our take-profit gate is a 66c BID and the bid has moved away from it, from 65 to 64. The gate is untriggered and we are not touching it. The real event here is not the 1.5 point price move but the eleven-point cut to our own fair value, which is set out in full in the top-five section.
10Y Touches 5.00% Before 2027 (watchlist)
17.5%→17.0%-0.5pp
The consistency check that failed yesterday STILL does not reproduce. On today's numbers the same model gives 31.1 against a 17.0c market, a 14.1pp gap. That is better than yesterday's 22pp and it is still a gap we do not believe. The standing rule from yesterday holds: no fair-value INCREASE on pos-017 until this check reproduces. Today's move was a cut, which the rule permits, and we note that a rule which only ever blocks good news for us is doing its job.
Zero Fed Rate Cuts in 2026 (pos-013)
86.4%→86.45%+0.05pp
Price did nothing; our opinion of it did. A 10% fall in crude over three sessions is disinflationary with a lag, and it is the first genuine argument against the no-cuts thesis we have had to write down in weeks. FV cut 88 -> 87. We are not cutting further on three sessions of oil, and we are not pretending those three sessions did not happen.
Fed Rate Hike in 2026 (pos-010)
56.5%→56.5%0.0pp
Unchanged on a day when oil fell 4.2%, which is itself information: either the hike bid is about something other than energy, or it has not repriced yet. We cut FV 55 -> 53 rather than assume the former. That leaves us marked 3.5pp above our own number, up from 1.5pp, and the floor at 48c is a long way below.
Fed Rate End 2026 = 4.0% (pos-011)
30.6%→31.05%+0.45pp
Ladder today: 3.5% 8.1 / 3.75% 43.75 / 4.0% 31.05 / 4.25% 13.0. The 4.0% bucket is exactly one hike, so lower hike odds should push it down, and it went up. FV cut 30 -> 29 for consistency with the pos-010 cut, which leaves us 2.05pp rich on a $25 ticket that costs more to exit than the gap is worth. Efficiency review still armed at 40c.
Screening Table
| # | Market | Expiry | Market Price | Fair Value | Gap (pp) | Direction | Volume | Confidence |
|---|---|---|---|---|---|---|---|---|
| 1 | 10Y Touches 4.8% Before 2027 | Dec 31 | 65.5% | 61% | -4.5pp | HOLD $25 YES - FV CUT 72 -> 61, we are now marked above our own number and the published exit gate does not cover this case | $$60K on the leg | 3/10 |
| 2 | Zero Fed Rate Cuts in 2026 | Dec 31 | 86.45% | 87% | +0.6pp | HOLD $82 YES - FV cut 88 -> 87 on the oil move; trim still armed at 90c | $$7.5M | 4/10 |
| 3 | US Unemployment 2026 >= 5.0% | Dec 31 | 12.5% | 8% | +4.5pp | HOLD $100 NO - 4.5pp on the NO, under the 10pp bar; benchmark payroll revision Friday | $$123K on the leg | 4/10 |
| 4 | Fed Rate Hike in 2026 | Dec 31 | 56.5% | 53% | -3.5pp | HOLD runner $100 YES - FV cut 55 -> 53; floor 48c, trim >72c, both far away | $$7.95M | 3/10 |
| 5 | Fed Rate End 2026 = 4.0% | Dec 31 | 31.05% | 29% | -2.1pp | HOLD $25 YES - too small to trade around; review at 40c | $$1.37M on the leg | 3/10 |
Market vs Fundamentals
Market Price (red) vs Estimated Fair Value (green) — %
Top 5 Opportunities
1
10Y Treasury Touches 4.80% Before 2027 - an eleven-point fair-value cut in one session, and a published exit rule that does not cover what happened — YES
↑ BUY YES-4.5pp
Market price
65.5%
Fair value
61%
Gap: -4.5pp
Yesterday we held this at a fair value of 72 against a 67c market and called it 5pp cheap. Today the same model, unchanged in every parameter we did not have to update, says 61 against a 65.5c market, and we are 4.6pp rich. Nothing about the method changed. One input did: the 10-year settled at 4.639% instead of the 4.7050% intraday print we wrongly published as yesterday's level. That moves the distance to 4.80% from 9.5 basis points to 16.1, and a barrier model is far more sensitive to the barrier than to anything else in it. We want to be precise about what this is and is not. It is not a change of view on the mechanism: the long end is still under global pressure and the buyback window still expires Nov 4 while the barrier runs to Dec 31. It is an admission that a position we have been marking as cheap was being marked off a number we had recorded incorrectly, and that when the number is corrected the edge inverts. The awkward part is the exit rule. Our published gate is a 66c BID, which is a take-profit written for an up-move; the bid is 64c and moving away. We have no published rule that says what to do when our own fair value falls below the market. We are not inventing one intraday, four hours before a PCE print that could move the 10-year either way - that would be exactly the improvisation the gates exist to prevent. Instead we arm a dated rule below, in advance, and we hold today.
▵ Bull case
- The mechanism is unbroken and global: the long end has been the pressure point all year, and one session of risk-premium unwind does not repair it.
- 16.1bp is still a small barrier over 91 sessions. The model has the touch as a two-in-three event even after the cut, and the market agrees at 65.5c.
- The Treasury buyback window that has capped the long end three times this month expires Nov 4, leaving roughly two months of the barrier period without that backstop.
- Today's yield move is a by-product of the oil move, not of anything structural in the fiscal or issuance picture, and oil moves reverse.
▿ Bear case
- We are marked 4.6pp above our own fair value. On any other position in this book that is the setup for a trim, and we should say plainly that the only reason we are not trimming is that our published rule does not reach this case.
- The eleven-point cut came from correcting our own data error. That is a process failure, not a market event, and it means yesterday's confident 5pp-cheap framing was wrong when we wrote it.
- If the Hormuz corridor is actually announced, the risk premium that has been doing much of the work in the long end comes out further, and 4.80% moves further away.
- The 5.00% consistency check still fails at 14pp. A model we do not fully trust is now the thing telling us we are rich, having yesterday told us we were cheap.
2
Fed Rate Hike in 2026 - the position did not move on a 4.2% oil fall, so we moved our fair value instead — YES
↑ BUY YES-3.5pp
Market price
56.5%
Fair value
53%
Gap: -3.5pp
This contract sat at 56.5c through a session in which Brent fell 4.2% and the 10-year fell 6.5 basis points. One of two things is true. Either the hike bid is genuinely about something other than energy - the tariff pass-through, the 9-3 committee split, Warsh's own preferences - or it simply has not repriced yet because the oil move happened overnight and into a pre-data morning. We do not know which, and the honest response to not knowing is to mark our own number down rather than to assume the market is telling us we were right. FV 55 -> 53. That leaves us marked 3.5pp above our own fair value on the runner, having been 1.5pp above yesterday and 0.5pp above the day before. Three sessions in a row of the market moving away from us, and now our own number moving away too. We flagged on Monday that the genuine uncertainty on this position is Warsh rather than the tariff. That is still true, and Friday is still the day it gets tested. What has changed is that the case for a hike now has to survive a crude price ten percent lower than it was on Friday.
▵ Bull case
- The tariff arithmetic is untouched by oil: 50% on roughly $20bn of Canadian goods with energy, potash and critical minerals carved out, and Carney's dollar-for-dollar retaliation starting Sep 8.
- Three FOMC members dissented in July in favour of hiking. That bloc does not disappear because crude fell for three days.
- Core inflation is the binding constraint, not headline, and the Cleveland Fed nowcast still has core PCE at 3.29% YoY for July - nowhere near target.
- A contract that refuses to fall on a large adverse move in its main macro driver is usually being held up by something, even if we cannot name it.
▿ Bear case
- A 10% fall in crude over three sessions is disinflationary with a lag and cuts directly against the energy-driven part of the hike case.
- We are now 3.5pp rich to our own fair value and have been marked above it for three consecutive sessions.
- If Warsh uses Jackson Hole to signal a preference for accelerated balance-sheet runoff over rate increases, a hawkish chair is bearish for a contract that only pays on the rate.
- The Fed September leg fell a point to 32.5c while the full-year contract did not move at all. Those two should be related, and today they were not.
3
Zero Fed Rate Cuts in 2026 - the oil move is the first real argument against this we have had to write in weeks — YES
↑ BUY YES+0.6pp
Market price
86.45%
Fair value
87%
Gap: +0.6pp
The price did nothing, moving five hundredths of a point. Our view of it moved. For most of August this position's risks were things we described as real but slow: continuing claims at 1.799M and rising, and the benchmark payroll revision on Friday. Neither operates on a 2026 timescale. A 10% fall in crude is different in kind. It feeds headline inflation within a couple of months and core with a longer lag, and it is the one thing that could plausibly give a committee that currently has no cutting case something to point at by December. We are cutting FV from 88 to 87. That is a deliberately small cut, because three sessions of oil is three sessions of oil and the thesis rests on core inflation at 3.29% YoY against a 2% target, which one barrel price does not fix. But we are not going to run a position where every new piece of information gets read as confirmation. Today's information was not confirmation, and the fair value moves accordingly. Note also that the Atlanta Fed's GDPNow, which we verified at source yesterday at 4.0% for Q3, updates today after the 08:30 releases. That is the first refresh in over a week and it will matter more to this position than the price action did.
▵ Bull case
- Cleveland Fed nowcast, re-verified at source this morning and re-stamped 08/25: July core PCE 0.25% MoM and 3.29% YoY, August core 0.27% and 3.34%. That is not a cutting profile.
- GDPNow has Q3 at 4.0%, verified at source, stamped Aug 18. Four percent annualised growth removes the growth argument for cuts entirely.
- Three FOMC dissents in July were for hiking, not cutting. The committee's centre of gravity is on the wrong side of zero for this to lose.
- Only four meetings remain in 2026 and a cut requires a majority to reverse direction from a hold that three members already think is too easy.
▿ Bear case
- Crude down 10% in three sessions is genuinely disinflationary and is the first argument against this thesis that operates on the right timescale.
- Continuing claims at 1.799M and rising, plus Friday's benchmark payroll revision, could reframe the labour market as weaker than the committee believed.
- At 86.45c the contract pays about 1.16-to-1. There is very little left to win here and the whole stake to lose.
- Our fair value of 87 is now barely above the market. On our own numbers this has become a hold-to-resolution rather than an edge.
4
US Unemployment 2026 >= 5.0% - unchanged thesis, and Friday is the day it gets its only real test — NO
↑ BUY YES+4.5pp
Market price
12.5%
Fair value
8%
Gap: +4.5pp
YES ticked up a tenth to 12.5, bid 11.4 / ask 13.6, so the mid remains arithmetic rather than a traded level. Our NO is worth 87.5c against an entry at 75c, for +$16.67. The underlying facts are unchanged and are not close. Unemployment is 4.1%. GDPNow has Q3 at 4.0%. Getting to 5.0% by year end needs roughly 1.4 million more unemployed people in four months. Friday is the only dated thing that can move this: the BLS preliminary annual benchmark revision to establishment payrolls at 10:00 ET, the same morning Warsh speaks at Jackson Hole. We said yesterday that knowing the date does not prove anyone is trading it, and that remains true - the three-point jump last week is still unexplained by anything we can point to. A benchmark revision rewrites the level of past payroll growth. It does not create nine tenths of a point of unemployment between now and December. FV on the YES stays at 8, so the NO is worth 92 against 87.5 for 4.5pp, under our 10pp bar.
▵ Bull case
- Unemployment 4.1% against a 5.0% threshold with four months left. The required move is roughly 1.4 million additional unemployed.
- GDPNow Q3 at 4.0%, verified at source. Economies growing at four percent annualised do not shed nine tenths of a point of employment.
- The bid-ask is 11.4 / 13.6 on $123K of volume. The 12.5 mid is a construct, and the bid is closer to our fair value than the mid is.
- Three FOMC members wanted to hike in July, which is not a committee looking at a labour market about to break.
▿ Bear case
- The 4.5pp gap is under our own 10pp entry bar, which means we would not put this trade on today at this price.
- The unexplained three-point jump last week has not reversed, and we still cannot account for it.
- Continuing claims at 1.799M and rising is a real signal, even if it is slow.
- Friday stacks the benchmark revision and a new chair's first Jackson Hole keynote in the same morning. That is a genuine two-sided event.
5
Fed Rate End 2026 = 4.0% - the ladder moved the wrong way and the ticket is too small to do anything about it — YES
↑ BUY YES-2.1pp
Market price
31.05%
Fair value
29%
Gap: -2.1pp
The 4.0% bucket is the one-hike outcome. On a day when the September hike leg fell a point and crude fell 4.2%, it rose 0.45 to 31.05. That is not large enough to be a signal on a leg like this, but it is the second small thing today that points the wrong way relative to the rest of the tape. We cut FV from 30 to 29 for consistency: if we are cutting the hike probability on pos-010, we cannot leave the one-hike bucket where it was. That leaves us 2.05pp rich. We are not trading it. A $25 ticket in a leg doing a few hundred dollars a day costs more in spread to exit than the entire gap is worth, and we have said so every day this position has been open. Hold to resolution.
▵ Bull case
- One hike remains the single most likely path from here if the committee moves at all, and the ladder still has 3.75% and 4.0% carrying three quarters of the distribution between them.
- The stake is $25 against a book of $332. Even a total loss is a rounding error against the realized ledger.
- The 4.25% bucket rose to 13.0 today, so the market is not uniformly reducing hike expectations.
- Resolution is mechanical at year end with no ambiguity about the settlement source.
▿ Bear case
- We are 2.05pp above our own fair value after today's cut.
- The bucket rose on a day when its main driver fell, which we cannot explain.
- Exiting costs more than the gap, which means we are locked in whether the mark is good or bad - a liquidity constraint we accepted at entry and should keep naming.
- If crude keeps falling, the no-hike bucket at 3.75% takes share from this one.