Daily Macro US
Fed hike 2026
55.5c
+6.0 points over the weekend on the US-Canada tariff restart; Friday's flagged oddity resolved upward
10Y Treasury
4.70%
-3.8bp with oil; 30Y 5.229%, -4.7bp. Friday closed near 4.738%, a whisker off the 4.75% 20-month high
Brent crude
$92.41
-2.1% (WTI $84.80, -2.6%) as Bessent unveils the Iran sanctions package; Iran threatens Hormuz seizures
US-Canada tariffs
50% / $20bn
Talks collapsed Friday night. Energy, potash and critical minerals carved out. Carney retaliates Sep 8
Core PCE (Jul)
3.29% YoY
Cleveland Fed nowcast verified today; 0.25% MoM. The print lands Wednesday Aug 26 with Q2 GDP revision
Book
+$100.63
5 open, $332 staked; +$611.08 realized, 11/14. No trades
On Friday we flagged what we called the day's genuine oddity and explicitly declined to resolve it: the Fed-hike-in-2026 market was falling into hot data, down to 49.5c and one and a half points above the 48c floor we published in June, while claims beat, factory activity hit a five-year high and Brent sat near $93. We wrote that we held no view strong enough to trade against it. Over the weekend the market resolved the oddity itself, and it resolved upward. Fed-hike-in-2026 is 55.5c this afternoon, up six full points from Friday. The September meeting is 32.5c against 27.5c on Friday, five points higher. The trigger is not subtle. US-Canada trade talks collapsed late Friday night: Washington has imposed 50% tariffs on roughly $20bn of Canadian goods under Section 338, and Prime Minister Carney has promised dollar-for-dollar retaliation from September 8 on steel, dairy, appliances, agricultural equipment, pulp, paper and electronics. A trade war restarting three weeks before an FOMC meeting is an inflation impulse, and the hike market repriced for it. Here is the part we have to say plainly. On Friday this position was 49.5c against our published fair value of 55, a 5.5-point gap, and we did not add because 5.5 is under our 10-point bar. Today it is 55.5c. Adding on Friday would have made money. The rule cost us that, and the rule was still right: a 5.5pp gap is not an edge, it is noise plus a fair value we compute ourselves, and a book that adds on 5.5-point gaps will eventually add on 2-point gaps. We publish the foregone gain rather than quietly not mentioning it. What we will not do is chase it now. Our fair value stays 55 and the market is 55.5, so we are marked half a point above our own number and the edge on pos-010 is gone. We are not raising fair value to justify holding, and here is why we think 55 is still right rather than conveniently low: the tariff package deliberately carves out energy, potash and critical minerals, the three Canadian exports with the most direct pass-through into US prices, and 50% on $20bn is about $10bn of duties in a $30tn economy. Meanwhile the far larger inflation input went the other way today. Brent is $92.41, down 2.1%, and WTI is $84.80, down 2.6%, because Treasury Secretary Bessent is unveiling what he has called an economic D-Day sanctions package against Iran and the oil market is reading sanctions as the alternative to escalation rather than as escalation. Iran has warned of Hormuz ship seizures in response and the rial is at a record low. Long-end yields fell with oil: the 10-year is 4.70%, down 3.8bp, and the 30-year is 5.229%, down 4.7bp. So the honest description of today is that the market added six points of hike probability on a $10bn tariff and ignored a $2 fall in Brent, and we do not think that arithmetic works. We hold and we do not add. One internal inconsistency worth naming: hike-by-October is unchanged at 40.5c while hike-by-September rose to 33.5c and the full year rose to 55.5c. That implies the September-to-October increment collapsed from 13 points to 7 while November-December widened from 9 to 15. On a leg doing under $500K of volume, the likeliest explanation is a stale quote rather than a signal, and we are recording it as an observation, not an opportunity. FedWatch, third consecutive session: we could not verify a September probability at source. The CME tool is behind a registration wall in this session. A secondary source published Friday reports 36.6%. That is not a verified price and under our Aug 6 rule it is not a tradeable input, so the September ticket is refused for a fifth time before the 10pp bar or the correlation cap are even reached. On pos-017 the fair value moves 70 to 72 and we want to be clear that this is arithmetic, not opinion: the 10-year fell to 4.70%, the barrier is now 10.0bp, 93 sessions remain at about 4bp of daily vol for sigma of 38.6bp, and the same formula with the same haircut we used yesterday returns 72. After two double-digit revisions in two days, a two-point mechanical update is what a stable process is supposed to look like. Friday's close, which our Friday letter went out before: the Dow rose 517.80 points or 0.98% to 53,277.01, the S&P 500 added 0.43% to 7,674.37 and the Nasdaq 0.43% to 26,180.45, though all three still posted weekly losses. Today is mixed, with the Dow up about 0.3%, the S&P down about 0.3% and the Nasdaq down about 0.8% as Micron falls more than 6%. The week's real events are Wednesday's PCE report with the Q2 GDP second estimate, and Warsh's first Jackson Hole keynote as Chair on Friday. No trades today. Book: 5 open, $332 staked, +$100.63 unrealized, +$611.08 realized, 11/14.
Today's Market Moves
Fed Rate Hike in 2026 (pos-010)
49.5%→55.5%+6.0pp
The largest single move in the book today and the resolution of the question we flagged on Friday without answering. US-Canada trade talks collapsed late Friday: 50% tariffs on about $20bn of Canadian goods, dollar-for-dollar retaliation from Sep 8. The hike market added six points. We refused to add on Friday at 49.5c against a 55 fair value because 5.5pp is under our 10pp bar, and that refusal cost us money — published here rather than omitted. We are not chasing it at 55.5c and we are not marking fair value up to make holding look better. FV stays 55, so we are 0.5pp above our own number and the edge is gone. Energy, potash and critical minerals are carved out of the tariff, and Brent fell 2.1% the same day; $10bn of duties against a $2 move in crude is not six points of hike probability.
10Y Touches 4.8% (pos-017)
66.5%→67.0%+0.5pp
FV 70 -> 72, and the size of that move is the point. The 10-year fell 3.8bp to 4.70%, so the barrier is 10.0bp rather than 10.9bp; 93 sessions remain to Dec 31 at about 4bp daily vol, sigma 38.6bp, 2*(1-Phi(0.2592)) = 79.6% raw, same haircut as yesterday for the Sep 9-Nov 4 buyback window -> 72. Nothing here is a judgment change; the yield moved and the formula moved with it. Market 67 mid, 64 bid / 70 ask. Take-profit gate unchanged at a 66c BID and still not triggered at 64c — one point closer than Friday, and we will publish it the day it prints.
US unemployment >= 5.0% (pos-004)
9.4%→12.5%+3.1pp
The worst mark move in the book today and the one we would rather not lead with, so we are listing it third instead of last. YES rose 3.1 points, cutting our NO from +$20.80 to +$16.67. Caveat that cuts both ways: this leg trades on about $123K of lifetime volume with an 11.1 bid against a 13.9 ask, so a 2.8-point spread means the 12.5 mid is largely notional. We can find no news to explain it; July payrolls at -23,000 were known before Friday. Either a thin book moved or someone knows something about the August employment report. Review line unchanged at YES below 8c, which is now further away, not closer.
Fed Sep Hike (context)
27.5%→32.5%+5.0pp
Fifth consecutive session in the letter, fifth refusal, and this time the reason is the same one as the last two: we could not verify a CME FedWatch September probability at source, because the tool sits behind a registration wall. A secondary source published Friday reports 36.6%. Against 32.5c that would be a 4.1pp notional gap, under the bar even taken at face value, on a number our own Aug 6 rule says we may not trade. Note also that hike-by-September at 33.5c reconciles cleanly with the 32.5c September meeting leg.
Zero Fed Cuts 2026 (pos-013)
86.45%→86.05%-0.4pp
Down four tenths, which is noise on the book's largest stake. FV stays 88 for a 1.95pp gap. Wednesday's PCE report is the first genuine test in three weeks: the Cleveland Fed nowcast for July core PCE, verified today, is 3.29% YoY and 0.25% MoM. A print near that is a Fed nowhere close to cutting. The efficiency trim stays armed at 90c and was not touched.
Fed Funds End 2026 = 4.0% (pos-011)
31.05%→30.75%-0.3pp
Back to almost exactly the 30 we published on Aug 13 as the cross-market reconciliation, which is the least interesting possible outcome and therefore the honest one. Ladder today: 3.5% at 8.65, 3.75% at 44.7, 4.0% at 30.75, 4.25% at 11.8. The 3.75% bracket gained more than five points on the tariff news while the 4.25%-and-above brackets fell, which is a distribution shifting toward exactly one hike and away from two. That is consistent with a hike-in-2026 market at 55.5c and inconsistent with panic. FV 30, marked 0.75pp above it, hold to resolution, efficiency review still armed at 40c.
Screening Table
| # | Market | Expiry | Market Price | Fair Value | Gap (pp) | Direction | Volume | Confidence |
|---|---|---|---|---|---|---|---|---|
| 1 | Fed Rate Hike in 2026 | Dec 9 | 55.5% | 55% | -0.5pp | HOLD $100 runner — edge gone; we did not add at 49.5c and we say what that cost | $$7.9M total | 6/10 |
| 2 | 10Y Touches 4.8% Before 2027 | Dec 31 | 67.0% | 72% | +5.0pp | HOLD $25 YES — FV 70 -> 72 mechanically; under the bar, no add | $$60K on the leg | 6/10 |
| 3 | Fed Sep Hike | Sep 16 | 32.5% | 36.6% | +4.1pp | NO TRADE — FV unverifiable at source; under bar; correlation-capped | $$9.6M on the leg | 3/10 |
| 4 | US Unemployment >= 5.0% in 2026 | Jan 31 | 12.5% | 8% | +4.5pp | HOLD $100 NO — worst mark move today; spread 11.1/13.9 makes the mid notional | $$123K on the leg | 6/10 |
| 5 | Zero Fed Cuts in 2026 | Dec 30 | 86.05% | 88% | +1.95pp | HOLD $82 — trim armed at 90c; PCE Wednesday is the test | $$7.5M total | 7/10 |
Market vs Fundamentals
Market Price (red) vs Estimated Fair Value (green) — %
Top 5 Opportunities
1
Fed Rate Hike in 2026 — the oddity we flagged resolved, and we were not on it — YES
↑ BUY YES-0.5pp
Market price
55.5%
Fair value
55%
Gap: -0.5pp
Friday's letter ended this section by naming a puzzle and refusing to trade it. Yields were up, Brent was near $93, initial claims had beaten consensus and Philadelphia Fed factory activity was at a five-year high, and the Fed-hike-in-2026 market fell to 49.5c, one and a half points above the 48c floor we published on June 10. We offered two explanations, said we did not know which was right, and held. Over the weekend the market answered: 55.5c today, six points higher. The cause is not mysterious. US-Canada trade talks collapsed late Friday night. Washington has imposed 50% tariffs on roughly $20bn of Canadian goods under Section 338 of the Tariff Act of 1930, and Prime Minister Carney has committed to dollar-for-dollar retaliation effective September 8 covering steel, dairy, appliances, agricultural equipment, pulp, paper and electronics. Carney's account is that the US introduced late demands over Canada's other trading relationships, its auto sector and protections for Canadian culture. A trade war restarting three weeks before an FOMC meeting reads as an inflation shock, and the hike market bought it. Now the uncomfortable accounting. We could have added on Friday at 49.5c against our own published fair value of 55. We did not, because 5.5 points is under our 10-point bar. That refusal cost real money on paper, and we print it here rather than let it vanish. We also still think the rule was right. A 5.5-point gap against a fair value we compute ourselves is inside our own error bar, and a book that adds on 5.5-point gaps ends up adding on 2-point gaps and calling it conviction. The rule exists so that Friday's decision is not a mood. What we will not do is buy it now at 55.5c. Fair value stays 55, which means we are marked half a point above our own number and there is no edge left in this position in either direction. We considered raising fair value and rejected it, and the reasoning is specific rather than stubborn: the tariff package explicitly carves out energy, potash and critical minerals — precisely the Canadian exports with the shortest path into US consumer prices — and 50% on $20bn is roughly $10bn of duties in a $30tn economy. On the same day, the far larger inflation input moved the other way, with Brent down 2.1% to $92.41 and WTI down 2.6% to $84.80 as Bessent unveiled the Iran sanctions package. Six points of hike probability for $10bn of tariffs and minus $2 on crude does not reconcile. We hold the runner, we add nothing, the 48c floor stays and so does the trim above 72c.
▵ Bull case
- Tariff restart is a genuine new inflation channel three weeks before the September FOMC
- Canada retaliates dollar-for-dollar from Sep 8, so the impulse compounds rather than fades
- Three July dissents toward a hike and a broader hawkish bloc in the August minutes
- The 3.75% end-2026 bracket gained over five points today — the ladder agrees with the move
▿ Bear case
- Energy, potash and critical minerals are carved out; the direct price channel is deliberately small
- $10bn of duties in a $30tn economy is a rounding error against a $2 move in Brent
- We are marked above our own fair value — no edge remains in either direction
- The July labour data that halved September odds three weeks ago has not been revisited or revised
2
10Y Treasury Touches 4.8% Before 2027 — a two-point move, which is the point — YES
↑ BUY YES+5.0pp
Market price
67.0%
Fair value
72%
Gap: +5.0pp
Two sessions ago we cut fair value on this position from 74 to 60. One session ago we put it back to 70 and said that two double-digit revisions in two days on a single position was instability rather than agility. Today fair value goes to 72, and the reason we are writing a full section about a two-point change is that it is the first move in three sessions that came out of the formula rather than out of a narrative. The 10-year fell 3.8bp to 4.70%, tracking Brent down as the Iran sanctions package landed as a diplomatic instrument rather than an escalation. That leaves the barrier 10.0bp away instead of 10.9bp. Ninety-three business sessions remain to December 31. At roughly 4bp of daily volatility, sigma is 38.6bp, and the two-sided touch probability 2*(1-Phi(10.0/38.6)) is 79.6%. We apply the same haircut we applied yesterday for the Treasury's enlarged long-end buyback window running September 9 to November 4, and we get 72. Same formula, same haircut, different yield. That is what a process is supposed to do when the input moves a little. The market is 67 mid on a 64 bid and a 70 ask, so the seven-point spread remains wide enough that the mid is a convenience rather than a price we could transact. The position is +$76.52 on a $25 stake and remains the engine of this book. The gate is where it has been since Thursday: a take-profit at a 66c BID, which we lowered from 72c when fair value was 60 and then deliberately declined to raise back when fair value recovered. The bid is 64c today, two points away and one point closer than Friday. If it prints 66 we sell and the sale appears in the next letter. We are saying that in advance for the fourth time so that nobody has to trust us about it afterwards. The structural case is unchanged and mostly not American: the Bund is at 3.2502%, near its highest since 2011, the 10-year JGB is at 2.90% and pushing on levels last seen in 1996, and the gilt is above 5%. Long ends are repricing everywhere on deficits and inflation, and a US buyback programme that expires on November 4 does not cover December.
▵ Bull case
- Barrier is 10.0bp with 93 sessions left; raw two-sided touch probability 79.6%
- Global long ends confirm the mechanism: Bund 3.2502%, JGB 2.90%, gilt 5.05%
- The buyback window ends Nov 4 and the barrier runs to Dec 31 — December is unprotected
- The tariff restart adds an inflation channel that argues for a higher term premium
▿ Bear case
- 64 bid / 70 ask on $60K of volume — the 67 mid is not a price we could exit at
- Oil rolling over is the single fastest way to take 15bp out of the 10-year
- This is now the third fair-value number in three sessions, even if today's came from the formula
- Iran sanctions could yet end the war, and a peace print takes yields down, not up
3
Fed September Hike (screened — refused for a fifth time) — YES
↑ BUY YES+4.1pp
Market price
32.5%
Fair value
36.6%
Gap: +4.1pp
Fifth appearance, fifth refusal, and the disclosure goes first again. We could not obtain a CME FedWatch September probability at source in this session. The tool is behind a registration wall and the underlying data endpoint returned an error. This is the third consecutive session in which we have failed to verify this number, which is itself worth stating: a fair value we cannot re-source is not a fair value we should be quoting, and we are close to retiring it from the letter entirely rather than carrying it forward with an apology each day. The 36.6% figure is from a secondary source published Friday, not from CME, and under the rule we published on August 6 an unverified price is not a tradeable input. That settles it before any other consideration. For completeness: Polymarket prices the September 25bp increase at 32.5c, up five points from Friday, with no change at 66.5c. Against the secondary 36.6 the notional gap is 4.1pp, comfortably under our 10pp bar even if it were verified. The correlation cap blocks it a third time independently — four of five open positions already pay off on a hawkish Fed. There is one observation in the term structure worth recording without acting on it. Hike-by-September is 33.5c and reconciles cleanly with the 32.5c September meeting leg. But hike-by-October is unchanged at 40.5c while the full year rose to 55.5c. That compresses the September-to-October increment from about 13 points to 7 and widens November-December from about 9 to 15. On a leg turning over less than $500K, the most likely explanation is a stale quote on a thin market rather than a genuine view that the Fed will skip October and move in December. We note it because a reader could otherwise find the same numbers and conclude we had missed something. We had not; we simply do not think it is real, and we are not going to build a position on an inference about a quote that has not updated.
▵ Bull case
- Deep market at $9.6M on this leg if the gap were ever real and verified
- Warsh's first Jackson Hole keynote as Chair on Friday is a genuine repricing catalyst
- The tariff restart is a live inflation channel that the September meeting has to price
▿ Bear case
- Fair value unverifiable at source for a third straight session — not a tradeable input
- 4.1pp is under the 10pp bar even taken at face value
- Correlation cap blocks it independently: this would be a fifth hawkish-Fed ticket
- Warsh could equally use Jackson Hole to slow the market down
4
US Unemployment reaches 5.0% in 2026 — our worst mark today — NO
↑ BUY YES+4.5pp
Market price
12.5%
Fair value
8%
Gap: +4.5pp
This is the position that moved most against us today and it gets a full section for that reason rather than a line in a table. YES on unemployment reaching at least 5.0% during 2026 rose from 9.4c on Friday to 12.5c, cutting the mark on our NO from +$20.80 to +$16.67. Three points is a large move on a market that has drifted our way for four straight sessions, and the honest answer is that we cannot source a reason for it. There was no labour-market release today. July payrolls at -23,000 were public before Friday's letter, initial claims at 206K beat consensus last week and unemployment stands at 4.1%. The mechanical caveat matters here more than usual: this leg has about $123K of lifetime volume and quotes 11.1 bid against 13.9 ask. A 2.8-point spread means the 12.5 mid is an arithmetic convenience, not a level anyone traded, and a single sized order can move it. So the plausible readings are a thin book being pushed, or someone positioning ahead of the August employment report and the preliminary payrolls benchmark revision due at the end of this week. Our fair value is 8 on the YES, which is where it has been: getting from 4.1% to 5.0% by December requires a genuine recession, and nothing in the claims data supports one — continuing claims at 1.799 million say the newly unemployed take longer to re-hire, not that employers are firing. Against a 12.5 mid our NO is worth 87.5c and the gap is 4.5pp in our favour, under the bar, so we hold and add nothing. The published review line is unchanged: YES below 8c triggers a review, not an automatic sale. That line is further away today than it was on Friday, which is the plain way of saying this position got worse.
▵ Bull case
- Unemployment is 4.1% and needs +0.9pp by December to breach — that requires a recession
- Initial claims 206K beat a 210K consensus last week; employers are not firing
- Our NO is worth 87.5c against a 75c entry and has been the quiet compounder of this book
▿ Bear case
- Moved 3.1 points against us today with no identifiable cause, which is the uncomfortable kind
- July payrolls were -23,000; the labour market is visibly decelerating
- $123K of volume and a 2.8-point spread mean we cannot trust the mid in either direction
- The preliminary payrolls benchmark revision lands Friday and revisions have been running negative
5
Zero Fed Rate Cuts in 2026 — YES
↑ BUY YES+1.95pp
Market price
86.05%
Fair value
88%
Gap: +1.95pp
The book's largest stake, down four tenths of a point to 86.05c and +$8.93 on $82. There is nothing to report from today and we would rather say so than manufacture a development. What is worth setting up is Wednesday. The July PCE report arrives on August 26 alongside the second estimate of Q2 GDP, and it is the first hard inflation reading this position has faced in three weeks. We verified the Cleveland Fed nowcast at source this afternoon: July PCE is modelled at 0.15% month-over-month and 3.65% year-over-year, with core PCE at 0.25% and 3.29%. The model also has August running hotter on the headline, at 0.34% monthly and 3.73% annually, which is the oil pass-through from a Brent price that spent the month above $90. A core reading near 3.3% against a 2% target, with three FOMC members having dissented toward a hike in July and a broader hawkish bloc visible in the August minutes, is a committee with no cutting case at all. The soft spot we named on Friday has not gone away: continuing claims at 1.799 million and rising is the mechanism by which a cooling labour market eventually forces cuts, and the preliminary payrolls benchmark revision on Friday could make the labour picture look worse retrospectively. Neither operates on a 2026 timescale. Fair value stays 88, the gap is 1.95pp, and the efficiency trim stays armed at 90c: at that price the remaining ten points are not worth $82 of capital in a five-position book.
▵ Bull case
- Cleveland Fed nowcast, verified today: July core PCE 3.29% YoY against a 2% target
- August headline PCE nowcast at 3.73% YoY on oil pass-through
- Tariff restart adds a second inflation channel on top of energy
- Three July dissents were toward a hike and none toward a cut
▿ Bear case
- 86.05c leaves 13.95 points of upside for $82 — the capital efficiency is poor and getting worse
- Continuing claims at 1.799M and rising is the early shape of a cutting cycle
- Friday's payrolls benchmark revision could reframe the labour market as much weaker than reported
- A soft PCE print on Wednesday would take pressure off yields and off this trade at once