Daily Macro US

Aug 21, 2026·US Macro Polymarket·5 markets screened

10Y Treasury

4.691%

Buyback relief fully erased; 4.70% Thursday, 5bp from this week's 20-month high of 4.75%

30Y Treasury

5.24%

Back from 5.196% — nearly all of Wednesday's 9bp buyback drop given back in one session

Initial claims

206K

Aug 15 week, vs 210K consensus, down from a revised 212K; continuing claims 1.799M

Brent crude

~$93

+4% on the week; Hormuz shut, 8 vessels attacked this month, new US sanctions on Iran being prepared

Fed Sep hike

27.5c

Down from 29c while yields and oil rose. FedWatch 34.6% is CARRIED from Aug 19, not re-verified today

Book

+$93.74

5 open, $332 staked; +$611.08 realized, 11/14

Twenty-four hours after we cut our own fair value on pos-017 from 74 to 60, the market erased the reason we did it. On Wednesday the Treasury announced it would at least double its long-end liquidity-support buybacks to $4bn per operation, the 10-year fell about 6bp to 4.647% and the 30-year gave up 9bp to 5.196%, and we wrote that a policy actor with an unlimited balance sheet now stood between our barrier bet and its target. On Thursday the entire move was given back. The 10-year rose to 4.70%, within 5bp of this week's 20-month high of 4.75%, and the 30-year climbed back to 5.25% — this despite Treasury Secretary Scott Bessent saying the accelerated buybacks could exceed the announced $4bn per issue and pointing to a coming fiscal plan. This morning the 10-year is 4.691%, down 1.5bp on the day but a full 4.4bp above where our fair-value cut was made. The tape's verdict on the buyback is that it is a liquidity tool, not a purchase programme, and it does not change net supply. Ours was slower. We took fourteen points off a fair value on the strength of a one-day move and the one-day move did not survive one more day. So we correct in the other direction and we say why: the barrier arithmetic has improved because the yield rose — 10.9bp to travel with roughly 91 sessions left at about 4bp of daily vol gives 77.5% on the raw maths — and the policy haircut is smaller than we thought, because we have now watched it fail. Fair value on pos-017 goes 60 to 70. The market is 66.5 mid, 63 bid / 70 ask. We are back below our own fair value, the position is +$75.76 on a $25 stake, and it is again the engine of this book. What we are NOT doing is raising the take-profit gate back to a 72c bid. We lowered it to a 66c bid yesterday; a gate that ratchets upward whenever the tape turns friendly is not a gate, it is a mood. It stays at a 66c bid and the bid is 63c, so it is not triggered today and we will honour it if it prints. That is the cost of moving a published rule in a hurry, and we pay it. The rest of the tape was firm in the same direction. Initial claims for the Aug 15 week came in at 206K against a 210K consensus and down from a revised 212K, with continuing claims at 1.799 million. The Philadelphia Fed's factory index posted its strongest growth since 2021. Brent is near $93, up more than 4% on the week, with the Strait of Hormuz still shut, eight vessels attacked in transit this month and Washington preparing sweeping new sanctions on Iran. Equities did not like any of it: the Dow lost 703.84 points, or 1.32%, to 52,759.21, the S&P 500 fell 0.87% to 7,641.16 and the Nasdaq dropped 1% to 26,067.17 on the view that higher borrowing costs are a problem for the bull market. And here is the day's genuine oddity, which we flag rather than resolve: with yields up, oil up and factory activity at a five-year high, the Fed-hike-in-2026 market went DOWN, from 50.0c to 49.5c, one and a half points above the 48c floor we published in June. The September meeting slipped from 29c to 27.5c. We hold no view strong enough to trade against that, but a hike market fading into hot data is worth naming out loud. CARRIED NUMBER, flagged per our own rule: CME FedWatch's 34.6% September read is from Aug 19 and we could not re-verify it at source in this session, so it is presented as carried and nothing is decided from it. No trades today. 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

We are reversing our own cut of yesterday: FV 60 back up to 70. The reason we cut — a Treasury bid of size suppressing the long end — was tested on Thursday and the market erased the entire buyback move, taking the 10-year to 4.70% and the 30-year back to 5.25%, even after Bessent said operations could exceed $4bn per issue. Raw barrier maths improves with the yield: 10.9bp over ~91 sessions at 4bp daily vol gives 2*(1-Phi(0.2856)) = 77.5%. A reduced policy haircut takes FV to 70 against a 66.5 mid. Two 10-plus-point fair-value revisions in two days on the same position is genuine instability and we are not dressing it up.

Fed Rate Hike in 2026 (pos-010)

50.0%→49.5%-0.5pp

The day's oddity. Yields up, Brent near $93, Philadelphia Fed factory growth the strongest since 2021, claims 206K against a 210K consensus — and the hike market fell half a point to 49.5c, one and a half points above the 48c floor published on June 10. The internal term structure is still coherent: September at 27.5c, hike-by-October at 40.5c, full-year at 49.5c leaves about 9 points of December-only hike. FV stays 55, gap 5.5pp, under the bar. If it trades below 48c the runner is sold the same day and printed here.

Fed Funds End 2026 = 4.0% (pos-011)

29.6%→31.05%+1.45pp

Second consecutive session above the 30 we published as the cross-market reconciliation on Aug 13. The full ladder today: 3.5% at 7.55, 3.75% at 39.25, 4.0% at 31.05, 4.25% at 13.0 — internally coherent and centred where a split committee with an oil shock belongs. The consequence of the call landing is that we now mark 1.05pp ABOVE our own fair value of 30. Hold to resolution, no add, efficiency review still armed at 40c.

Zero Fed Cuts 2026 (pos-013)

85.3%→86.45%+1.15pp

Up a point on firm data. Claims at 206K and a five-year high in Philadelphia Fed factory activity give no cut case at all, and unemployment is 4.1%. FV stays 88, so the gap is 1.55pp. The efficiency trim stays armed at 90c and was not touched.

US unemployment >= 5.0% (pos-004)

9.9%→9.4%-0.5pp

Continues to drift our way on a labour market that is cooling without cracking. Claims fell to 206K and beat consensus; continuing claims rose to 1.799 million, which says the newly unemployed are taking longer to find work but employers are not firing. Our NO is worth 90.6c against a 75c entry, +$20.80. Review line unchanged at YES below 8c.

Fed Sep Hike (context)

29.0%→27.5%-1.5pp

Fourth consecutive session lower, and a fourth refusal. Against the CARRIED FedWatch reading of 34.6% from Aug 19 the gap would be 7.1pp — still under the 10pp bar, and built on a number we did not re-verify at source today, which under our own Aug 6 rule is not a tradeable input. The correlation cap blocks it independently: this would be a fifth ticket paying off on the same hawkish-Fed view. Warsh gives his first Jackson Hole keynote as Chair next Friday, Aug 28.

Screening Table

# ↑	Market ↕	Expiry ↕	Market Price ↕	Fair Value ↕	Gap (pp) ↕	Direction ↕	Volume ↕	Confidence ↕
1	10Y Touches 4.8% Before 2027	Dec 31	66.5%	70%	+3.5pp	HOLD $25 YES — FV reversed 60 -> 70; gap under the bar, no add	$$60K on the leg	6/10
2	Fed Rate Hike in 2026	Dec 9	49.5%	55%	+5.5pp	HOLD $100 runner — 1.5 points above the 48c floor	$$7.7M total	6/10
3	Fed Sep Hike	Sep 16	27.5%	34.6%	+7.1pp	NO TRADE — FV is a CARRIED number; under bar; correlation-capped	$$8.8M on the leg	3/10
4	Zero Fed Cuts in 2026	Dec 30	86.45%	88%	+1.55pp	HOLD $82 — trim armed at 90c	$$7.4M total	7/10
5	Fed Funds End 2026 = 4.0%	Dec 8	31.05%	30%	-1.05pp	HOLD $25 — now marked above our own FV; edge gone	$$1.37M total	5/10

Market vs Fundamentals

Market Price (red) vs Estimated Fair Value (green) — %

Top 5 Opportunities

10Y Treasury Touches 4.8% Before 2027 — we reverse yesterday's cut — YES

Dec 31, 2026·$60K on the leg·Confidence ★★★☆☆ 6/10

↑ BUY YES+3.5pp

Market price

66.5%

Fair value

70%

Gap: +3.5pp

Yesterday's letter led with a fourteen-point cut to this position's fair value, from 74 to 60, on the grounds that the Treasury's decision to at least double long-end buybacks to $4bn per operation had put an official buyer of size between our barrier and its target. The evidence was a 6bp fall in the 10-year to 4.647% on the announcement, with Brent rising at the same time — oil up, term premium down, our published mechanism failing a clean test. One session later the market unwound the entire move. On Thursday long-maturity yields rebounded, the 10-year rose to 4.70% and the 30-year to 5.25%, erasing almost all of Wednesday's decline, and it happened even as Bessent said the accelerated operations could exceed $4bn per issue and flagged a coming fiscal plan. This morning the 10-year is 4.691%. The market's reading is the one the mechanics support: a buyback is a liquidity operation in secondary markets, it does not retire net supply, and it cannot hold a curve that is repricing on deficits, AI-related corporate issuance and $93 oil. Our reading was too fast. We wrote fourteen points off a fair value on a single session's price action, and the single session did not survive. So the correction runs the other way and the arithmetic is better than it was: the barrier is now 10.9bp away rather than 15.3bp, roughly 91 sessions remain, daily vol is about 4bp giving sigma of 38.2bp, and 2*(1-Phi(10.9/38.2)) = 77.5%. We keep a policy haircut, because the buyback programme is real and runs Sep 9 to Nov 4, but a smaller one now that we have watched it fail in real time: fair value 70. The market is 66.5 mid, 63 bid, 70 ask, so we are back below our own fair value and the position is +$75.76 on a $25 stake. One thing we will not do is move the take-profit gate back up. We lowered it to a 66c bid yesterday when fair value was 60; raising it to 72c today because the number moved our way would make it a mood rather than a rule. It stays at a 66c bid, the bid is 63c, it is not triggered, and if it prints we sell and publish it.

▵ Bull case

Barrier is 10.9bp away, down from 15.3bp yesterday — raw maths 77.5%
The buyback move was fully erased within one session even with Bessent talking it up
Brent near $93 with Hormuz shut and new Iran sanctions being prepared
The programme expires Nov 4; December is unprotected and the barrier runs to Dec 31

▿ Bear case

Two fair-value revisions of 10+ points in two days on one position is real instability
The buyback programme still exists and still covers Sep 9 to Nov 4
63 bid / 70 ask — the 66.5 mid we carry is notional and the spread is 7 points
We are correcting toward the direction that flatters the book, which is exactly when to be suspicious of ourselves

Fed Rate Hike in 2026 (runner) — a hike market fading into hot data — YES

Dec 9, 2026·$7.7M total·Confidence ★★★☆☆ 6/10

↑ BUY YES+5.5pp

Market price

49.5%

Fair value

55%

Gap: +5.5pp

This is the position closest to a published exit and it deserves the plainest writing in the letter. The runner trades at 49.5c against a 48c floor we published on June 10. One and a half points. If it prints below 48c we sell it and the loss appears in the next letter on the same day, and we are saying that now, in advance, so nobody has to take our word for it afterwards. What makes today unusual is that the price fell into data that argues for a hike. Initial claims came in at 206K against a 210K consensus, down from a revised 212K. The Philadelphia Fed's factory index recorded its strongest growth since 2021. Brent is near $93 and rising, the Strait of Hormuz is still shut, and the US is preparing sweeping new sanctions on Iran. Long-end yields rebounded hard enough to erase a Treasury intervention. Every one of those is an input to a hike, and the hike market went down half a point. We can offer two explanations and we do not know which is right: either the market reads the equity selloff — Dow -1.32%, S&P -0.87%, Nasdaq -1% on borrowing-cost fears — as a financial-conditions tightening that does the Fed's work for it, or the July retail sales print at -0.6% and a consumer that is visibly slowing simply outweigh the rest. Fair value stays 55 because the underlying case has not changed: three FOMC members dissented toward a hike in July, oil is the dominant inflation input and it is going up. The internal cross-check reconciles cleanly — September 27.5c, October 40.5c, full-year 49.5c leaves about 9 points of December-only hike. At 49.5c against 55 the gap is 5.5pp, under the 10pp bar, so no add.

▵ Bull case

Claims 206K beat, Philly Fed strongest since 2021, Brent near $93 — the hike inputs are all firming
Three July dissents in favour of a hike; the committee is genuinely split
Term structure across September, October and full-year is internally coherent

▿ Bear case

One and a half points above a published floor we will honour in public
The price is falling into hot data, which means the market sees something we do not
Consumer still soft: July retail sales -0.6%, and equities sold off on borrowing costs
Warsh has said he will not be constrained by market pricing — that cuts both ways

Zero Fed Rate Cuts in 2026 — YES

Dec 30, 2026·$7.4M total·Confidence ★★★★☆ 7/10

↑ BUY YES+1.55pp

Market price

86.45%

Fair value

88%

Gap: +1.55pp

The book's largest stake and its least eventful idea, up 1.15 points to 86.45c and +$9.35 on $82. Today's data pushed it the right way and the reason is simple enough to state in one line: nothing in this week's numbers gives a committee a reason to cut. Initial claims fell to 206K, beating a 210K consensus and falling from a revised 212K. The Philadelphia Fed factory index printed its strongest reading since 2021. Unemployment is 4.1%, down from 4.2% in June. Headline inflation was 3.4% in July against a 2% target. Three FOMC members dissented toward a hike at the July meeting and none toward a cut. The one soft spot worth naming honestly is continuing claims at 1.799 million, up 18,000 and slightly above forecast — that is the signal that people who lose jobs are taking longer to find new ones, which is the mechanism by which a cooling labour market eventually becomes a cutting cycle. It is not close to that yet, and it would need to run for months before it were. Fair value stays 88 and the efficiency trim stays armed at 90c: at that price the remaining ten points of upside are not worth $82 of capital in a five-position book.

▵ Bull case

Claims 206K, unemployment 4.1%, Philly Fed at a five-year high — no cut case in the data
Inflation at 3.4% against a 2% target with Brent near $93
Three July dissents were toward a hike, not a cut

▿ Bear case

86.45c leaves 13.55 points of upside for $82 — poor capital efficiency
Continuing claims rising to 1.799M is the early shape of a labour-market problem
A Q4 growth shock would reprice this fast, and equities just sold off 1% on rate fears

Fed September Hike (screened — refused for a fourth time) — YES

Sep 16, 2026·$8.8M on the leg·Confidence ★★☆☆☆ 3/10

↑ BUY YES+7.1pp

Market price

27.5%

Fair value

34.6%

Gap: +7.1pp

Fourth consecutive letter, fourth refusal, and this time the transparency note comes before the analysis. The fair value of 34.6 is CME FedWatch's reading 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 a CARRIED number. Under the rule we published on Aug 6 — an unverified price is not a tradeable input — that alone settles it: we do not enter on a gap computed against a two-day-old anchor, no matter how attractive the arithmetic looks. For completeness, the arithmetic: Polymarket has the September hike at 27.5c, down from 29c yesterday and 32.5c a week ago, so against the carried 34.6 the notional gap is 7.1pp. Still under the 10pp bar even if it were verified. And the correlation cap blocks it a third time over: four of our five open positions already pay off if the Fed does not cut and probably hikes. The interesting question this market poses is not whether to buy it but why it keeps falling. Since Aug 13 the September hike has lost five points while the 10-year rose, Brent gained more than 4% on the week, claims beat and factory activity hit a five-year high. Either the market is treating the equity selloff as sufficient tightening, or it has concluded that a Fed chaired by Kevin Warsh — who has said he will not be constrained by market pricing and no longer telegraphs meetings — is simply less forecastable, which lowers the price of any specific meeting outcome. We do not know, and we are not paying to find out.

▵ Bull case

Deep market at $8.8M on this leg — sizeable if the gap were ever real and verified
Warsh's first Jackson Hole keynote as Chair on Aug 28 is a genuine repricing catalyst
Hike inputs firmed all week while the price fell

▿ Bear case

The fair value is a CARRIED number from Aug 19 — not a tradeable input under our own rule
7.1pp is under the 10pp bar even taken at face value
Correlation cap blocks it independently: fifth hawkish-Fed ticket
A market falling into supportive data usually knows something

Fed Funds Rate End 2026 = 4.0% — YES

Dec 8, 2026·$1.37M total·Confidence ★★☆☆☆ 5/10

↑ BUY YES-1.05pp

Market price

31.05%

Fair value

30%

Gap: -1.05pp

Eight sessions ago this was the worst position in the book at 22.65c and we published the arithmetic for why it was cheap rather than broken: hike-in-2026 near 54.5, the brackets at and above 4.25% summing to roughly 20, about 4 points belonging to a hike that later gets reversed, so exactly-one-hike-and-hold reconciles near 30. It printed 29.6 yesterday and 31.05 today. The pre-published number was reached and then passed, which means the honest report is that we are now marked 1.05 points above our own fair value — the same situation we described as uncomfortable on pos-017 yesterday, in the opposite direction and on a much smaller ticket. The bracket ladder remains internally coherent: 3.5% at 7.55, 3.75% at 39.25, 4.0% at 31.05, 4.25% at 13.0, with the sub-3% tail near zero. That is a distribution centred on one hike, which is what a split committee facing $93 oil and 3.4% inflation should produce. There is no edge left here in either direction, the market does under $250 a day, and exiting a $25 ticket into that would cost more in spread than the 1.05 points are worth. We hold to resolution, add nothing, and keep the efficiency review armed at 40c.

▵ Bull case

The pre-published reconciliation at 30 was reached and exceeded within eight sessions
Ladder is internally coherent across every bracket
$25 ticket — cheap to hold to resolution

▿ Bear case

Marked 1.05pp above our own fair value; the trade is over
Under $250/day of volume makes any exit expensive
Yet another Fed ticket in a book that is mostly Fed tickets