Closing Briefing — Friday, August 28, 2026

Source: Finviz group screener, Performance table view (g=sector&v=140&o=name), read after the close. 1-Day is the Change % column, 1-Week Perf Week, YTD Perf YTD. Finviz classification, not GICS.

YTD reconciliation, and the drift is down to one group. Compounding each group's 27 August YTD by Friday's one-day move reproduces the published YTD to within 0.02 percentage points for ten of the eleven groups — the cleanest reconciliation of the reporting window. Worked examples: technology 1.2560 × 0.9859 = 1.2383 → +23.83% against a published +23.81%, deviation 0.019 pp; communication services 0.9809 × 1.0144 = 0.9950 → -0.50% against -0.49%, deviation 0.008 pp. The single exception is real estate, where 1.0983 × 0.9959 = 1.0938 → +9.38% against a published +9.58%, a 0.200 pp gap that is now identical in sign and near-identical in size to Thursday's 0.195 pp. Two consecutive sessions of the same downward drift in the same group is a constituent or dividend adjustment inside the bucket, not a data error; it is flagged and carried. Basic materials, which drifted 0.108 pp on Thursday, reconciles to 0.101 pp this session.

Thursday inverted exactly. The two groups that led on Friday — communication services +1.44% and consumer cyclical +1.36% — were the third- and fourth-worst on Thursday at -0.76% and -0.98%, and the group that led Thursday by 288 basis points, technology at +3.09%, was the second-worst on Friday at -1.41%. That is a two-day round trip in which the index net gained 0.47% and the sector ranking completely reversed, which is what a positioning unwind looks like rather than a rotation. The composition trap matters more than usual this session: Amazon (+3.97%) sits in consumer cyclical and Alphabet (+1.74% on the A line) and Meta (+1.21%) in communication services, so the two leading Finviz groups are carried by three mega-caps a GICS reader files under technology or discretionary — and both groups remain the only negative ones on the year, at -0.49% and -2.73%.

The weekly column has turned defensive-negative and the yearly one has not moved. Over five sessions communication services +1.27% and technology +1.22% lead while healthcare -2.18%, energy -2.12% and industrials -1.49% lag, and healthcare's weekly decline arrives with Moderna -3.35% and Incyte -2.64% on the day. On the year the shape is unchanged and extreme: energy +35.89% and technology +23.81% against consumer cyclical -2.73%, utilities -0.95% and communication services -0.49%, now three negative groups after utilities crossed below zero on Friday's 1.12% decline.

Higher

Lower

Analyst and corporate actions

Sourcing, disclosed. The Earnings Whispers day pages remain behind a cookie-and-usage-agreement consent banner, which this unattended session did not accept. The rosters below are captured from the Nasdaq earnings calendar API for each date and screened name by name against an S&P 500 constituent list. Nasdaq publishes a before-open / after-close bucket rather than a clock time, so no clock times are asserted this session; confirm every time against company investor relations before trading a date.

The current week is finished. Friday 28 August was its last session and it carried no S&P 500 reporter, so no day remains outstanding; the next constituent to report is Medtronic before the open on Tuesday 1 September.

Mon 8/31. No S&P 500 reporter on either bucket. The Nasdaq capture returns fourteen names for the date, none of them constituents.

Tue 9/1. BMO: Medtronic (MDT). AMC: Palo Alto Networks (PANW), Dell Technologies (DELL).

Wed 9/2. BMO: Brown-Forman (BF.B). AMC: Broadcom (AVGO), Hewlett Packard Enterprise (HPE), NetApp (NTAP).

Thu 9/3. BMO: Campbell's (CPB), Toro (TTC). AMC: Lululemon Athletica (LULU). Timing bucket not published: Copart (CPRT) — the reviewed calendar carries no before-open or after-close designation for this date; confirm with company investor relations.

Fri 9/4. No S&P 500 reporter on either bucket.

Changes vs. the prior calendar (8/27 report):

Source: U.S. Department of the Treasury Daily Treasury Par Yield Curve Rates, Text View for August 2026, read after publication. Changes are versus the 27 August official row (1-day) and the 21 August official row (1-week).

Off-table bills, extracted and reported here because they carry the financing story. 1.5 Mo 3.83% (+4 bp on the day, +6 bp on the week), 2 Mo 3.86% (+5 bp, +6 bp), 4 Mo 3.94% (+6 bp, +4 bp), 6 Mo 4.02% (+8 bp, +7 bp). The 6-month is the fulcrum: it is the shortest tenor that fully contains the 16 September meeting, and it cheapened 8 bp on the day against the 1-month's 3 bp. See Section 9 block b.

The read: a policy-led bear flattener, and its shape is a near-perfect monotonic decay. Start at the 2-year and walk out: +14, +11, +10, +7, +6, +3, +3 basis points at 2, 3, 5, 7, 10, 20 and 30 years. A curve that cheapens most where the policy rate lives and least where the term premium lives has repriced the path and left the destination alone — and it did so on a speech with no numbers in it about rates. 2s30s flattened 11 bp to 88 and 2s10s 8 bp to 39, the largest one-day flattening Bloomberg can find since Warsh's first press conference as Chairman in June. Note the diagnostic that is missing: 3M10Y was unchanged at 83 bp, because the 3-month bill cheapened 6 bp alongside the 10-year's 6 bp. The bill matched the note, which means even the segment with almost no meeting content took the message.

The week now has one shape and it is the opposite of the level. On five sessions the 1-year is 12 bp cheaper at 4.15%, the 2-year 10 bp, the 3-year 10 bp and the 5-year 5 bp, while the 10-year is 1 bp richer, the 20-year 4 bp and the 30-year 5 bp richer at 5.22%. 2s30s has flattened 15 bp in a week and 2s10s 11 bp. Twenty-four hours ago this report described that configuration as "pricing policy without pricing term premium," on a 5 bp weekly move at the one-year point. The one-year has now moved 12 and the argument has strengthened rather than changed: the market has spent the week pulling the tightening forward and buying the long bond against it.

The Street's test, and the long end passed it. Bloomberg reported on Thursday that JPMorgan, Apollo and Morgan Stanley all expected a credibly hawkish Warsh to trigger buying of 30-year bonds, whose yields "hit the highest since 2007 last week," and Apollo's Torsten Slok warned that no framework guidance would mean "a much higher move in long rates." Both conditions were met simultaneously — hawkish message, no framework guidance — and the 30-year rose three basis points. Bloomberg's Greg Ritchie put the structural point: the three largest daily moves in the Treasury curve since Warsh became Chairman in May have all followed his own appearances, which is the precise outcome his "hall-of-mirrors" critique was written to avoid. Natixis's Christopher Hodge: "His notion from July that markets were playing the ref is clearly off-base. The Fed isn't the ref, it's a huge player and the market movements reflect that reality." MFS's Alex Mackey added the forward warning: "The chairman reaffirmed his expectation to offer less communication, underscoring the uncertainty the market will confront in anticipation of future messages. Events like today's speech are ripe for bouts of volatility." Bloomberg's live board marks the 10-year at 4.72%, +4 bp, against the official par 4.73%, +6 bp; the level gap is one basis point and the change gap two, a 4:59 p.m. live mark against a 3:30 p.m. bid-side construct, and it has now run at one basis point for four consecutive sessions. See Section 9 block c and Section 12 idea 1.

Source: Federal Reserve Bank of New York Economic Indicators Calendar for August and September 2026 (all times Eastern). Consensus figures are carried only where independently verified; where none is verified, the sensitivity note describes what the market is positioned for instead of asserting an expectation.

Current week (Aug 24-28) — still to come

Nothing remains. Friday 28 August was the week's final session and its calendar cleared with the New York Fed staff nowcast at 12:45 and the Holston-Laubach-Williams r-star estimates at 14:00.

Next week (Aug 31 - Sep 4)

Current target range: 3.50%-3.75%. Two vendors, two snapshots twelve minutes apart, and a 3.8 percentage-point gap that measures the last twelve minutes of the ZQ session.

CME FedWatch headline — 16 September 2026 meeting. Data as of 28 Aug 2026, 04:57:18 p.m. CT (5:57 p.m. ET), read from the FedWatch probability table.

Provenance of every column, stated — and one prior-session correction. CME's NOW column carries a 04:57:18 CT timestamp; the wall clock at the time of the pull was after 5:00 p.m. ET, and the ZQ contract's Central-time session close is 4:00 p.m. CT, so the reading resolves as p.m. and sits within the hour after the CT close — a near-settlement snapshot rather than an indicative intraday read, and materially better provenance than Thursday's 10:19 a.m. ET capture. 1 DAY carries the legend date 27 August and prints 35.4% for the hike, against the 33.9% this report published from CME's morning column and the 34.4% it published from Investing.com's evening card. Thursday's true close was therefore about 1.5 pp higher than the CME figure published in the previous edition; the correction is recorded here, and it does not change any conclusion drawn on Thursday. 1 WEEK (21 August) and 1 MONTH (28 July) are genuine reference dates rather than chart reads and are used in calculations. The Investing.com matrix below is timestamped 28 Aug 2026, 05:45 p.m. EDT and is the primary source for every number in parts (a), (b) and (c).

The CME-versus-Investing.com gap, quantified. CME puts the September hike at 59.7% at 5:57 p.m. ET; Investing.com at 55.9% at 5:45 p.m. ET — a 3.8 percentage-point difference across twelve minutes, against 0.5 pp across a twelve-hour gap on Thursday. The gap is large in probability and tiny in price. Investing.com publishes the September future at 96.305, down 2.0 bp from Thursday's 96.325; because the 16 September meeting sits mid-month, only about 47% of the contract's averaging period is affected by the decision, so a single basis point of ZQ price is worth roughly 10 percentage points of hike probability. The entire 3.8 pp vendor gap is therefore four-tenths of a basis point of contract price, and the entire 21.5 pp move on Investing.com's own card is two basis points. That leverage is the most important number in this section: it means a fifth of a tick reprices the meeting, and it means the 59.7% headline is far less robust than it looks.

One-day, one-week and multi-day momentum. The September hike rose 24.3 pp on CME's own columns (35.4% to 59.7%) and 21.8 pp on Investing.com's (34.1% to 55.9%) — by an order of magnitude the largest single-session move of the reporting window, which has to this point measured its repricings in ones and twos. On a one-week view it is 59.7% against 39.9% on CME and 55.9% against 39.0% on Investing.com, so essentially the whole weekly move happened in six hours. The one-month column tells a different and useful story: on 28 July CME priced 55.8% at 375-400 plus 20.2% at 400-425, a cumulative 76.0% chance of at least one hike by September. Friday's 59.7% is still 16.3 pp below where the market sat a month ago, which is the discipline to keep — this was a violent recovery of an old position, not the establishment of a new one. Further out, every horizon moved together: October's cumulative-above went to 70.1% from 53.8%, December's to 88.4% from 74.6%, January 2027's to 91.8% from 79.4%, and the peak of the strip, June through September 2027, to 95.9% from about 88%. The probability of a cut at any 2026 meeting remains 0.0%.

(a) Current-year meeting distributions

Investing.com Fed Rate Monitor, updated 28 Aug 2026 05:45 p.m. EDT. Format: current [prior day] [prior week]. Modal range in bold.

Sums are 100.0%, 100.1% and 100.0% on the published figures. Two facts in this table are new to the reporting window. First, September's modal outcome is no longer a hold — 55.9% at +25 bp against 44.1% at the current range, the first time the meeting has been modal-hike on this vendor. Second, and larger, December's modal outcome is no longer one hike but two: 4.00-4.25% at 38.9%, narrowly ahead of 3.75-4.00% at 38.5%, having been 25.2% and 44.8% twenty-four hours earlier. The hold column at December has collapsed from 25.4% to 11.6% in a session and from 28.9% in a week, and the +75 bp bucket has gone from 4.6% to 11.0%. October's hold, which this report noted falling below 48% for the first time on Thursday, is now 30.0%.

(b) Next-year meeting path

Modal range, its probability, and the cumulative probability above and below the current 3.50-3.75% range, with the contract price that draws it.

The whole 2027 strip is now modal 4.00-4.25%, and it got there in one session. On Thursday the modal range was 3.75-4.00% from January through April and 4.00-4.25% only from June to October, with December tied. On Friday every meeting from January 2027 to December 2027 is modal 4.00-4.25%, and the shape has changed as well as the level: January's cumulative-above is 91.8%, up 12.4 pp on the day, and the peak has moved forward and flattened, sitting at 95.9% across June, July and September 2027. ZQ prices fall from 95.985 in January to a joint trough of 95.770 in July and September 2027 — 15.0 bp lower than Thursday's 95.920 trough — before recovering to 95.810 by December. The first non-trivial cut probability, 0.8% at 3.25-3.50%, still appears only at December 2027, down from 1.2% on Thursday. A strip that adds a full range to its modal path while shrinking the odds of any cut is not widening its distribution; it is translating one.

(c) Year-end probability ladders

Year-end 2026 — the 9 December meeting.

Cumulative above the current range: 88.4%. Cumulative below: 0.0%. Sum: 100.0%.

Year-end 2027 — the 8 December meeting.

Cumulative above the current range: 92.6%. Cumulative below: 0.8%. Sum: 100.0%.

Rounding, transparently. Every figure is reproduced at the vendor's own one-decimal precision. Column sums of 99.9% or 100.1% are rounding artefacts of that precision, not missing probability mass; no cell has been rescaled, and cells the vendor does not publish are shown as 0.0% only where the vendor's own card omits the range entirely, which under CME methodology means a probability below the rounding floor. The October 2026 and January and March 2027 rows sum to 100.1% for exactly this reason.

(a) IG and HY credit spreads

ICE BofA option-adjusted spreads via FRED. FRED publishes with a one-business-day lag: the levels below are as of 27 August 2026, not the 28 August close. Same-day direction is cross-checked against the cash-market proxies underneath.

CDX — the full six-step ladder was worked again and is reported so the gap stays auditable. (1) Bloomberg in Chrome: the /markets and /markets/rates-bonds boards were rendered and carry the Bloomberg Fixed Income Indices and global government yields but no CDX line; the site's individual quote pages returned a bot-detection interstitial, which was not circumvented. (2) WSJ Market Data bonds page: rendered, headlined "Short-Term Treasury Yields Rise as Warsh Targets Inflation in Jackson Hole Speech," and its credit tables again did not populate a CDX row. (3) Cbonds carries dedicated CDX.NA.IG 5Y and CDX.NA.HY 5Y index pages but masks the levels behind a subscription; S&P Dow Jones Indices publishes index-family and methodology documentation, not the daily running spread. (4) FT Markets Data and Reuters credit wraps returned no dated CDX quote for 28 August. (5) TradingView and Barchart symbol searches for CDX resolve to an unrelated ETF. (6) Cash-market proxies, labelled as proxies: HYG closed $79.74, -0.16%, and LQD $106.30, -0.41%, against 27 August closes of $79.87 and $106.73. Two and a half times as much loss in the IG vehicle as the HY one, on a day the 10-year cheapened 6 bp, is duration doing the damage and credit doing none — which is the same message the FRED series carries through 27 August. No CDX level is published here, because an undated third-party digest number is not a CDX level.

The Chairman just made this table a policy variable. Warsh's economic assessment named it in terms: "Credit spreads on corporate bonds and leveraged loans are near the low ends of their historical ranges, and issuance volumes in these markets have been quite strong this year," and, on bank lending standards, "credit and loan markets are showing few signs of policy restraint." Read the rows against that sentence. IG at 79 bp is exactly where 2026 started and a basis point tighter on the day. HY at 263 bp is 18 bp tighter on the year and 12 bp tighter on the week, the largest weekly tightening of the reporting window. Only the tail dissents: CCC held 1,031 bp, unchanged on the day, which took the CCC-minus-HY differential out to 768 bp from 764 — reversing Thursday's first narrowing and returning to within a basis point of the record. So the aggregate credit market is doing precisely what the Chairman said it is doing, and the distressed cohort is not. That divergence is now the cleanest available statement of what a hike would actually bite.

(b) Money-market and funding plumbing

New York Fed reference rates, published ~8:00 a.m. ET for the prior business day. The rates below carry the 27 August 2026 effective date.

The overnight market went quiet exactly as the term market woke up. Every published rate was unchanged on the 27 August effective date — SOFR at 3.64%, a basis point below the 3.65% IORB, with the 99th percentile steady at 3.72% and volume $23bn lighter at $2,836bn. The facility drained further: overnight reverse repo take-up fell to $175m on 28 August from $456m on the 27th and $702m on the 26th, and is now below the $200m of 21 August. Reserve balances have no new print — the Federal Reserve's H.4.1 series still reads $2.9249tn for the week ended 26 August, $68bn below the 5 August peak and falling for a fourth week, and the next observation covers the week containing month-end.

The bill strip is where the message landed, and the fulcrum is the six-month. The 6-month bill cheapened 8 bp to 4.02% and the 4-month 6 bp to 3.94%, against the 1-month's 3 bp to 3.84% and the 1.5-month's 4 bp to 3.83% (Section 6, off-table tenors). The gradient is the point: the 6-month is the shortest tenor that fully spans the 16 September meeting, and it moved nearly three times as much as the 1-month, which does not span it at all. On the week the ordering is the same — 6-month +7 bp, 2-month +6 bp, 1.5-month +6 bp, 4-month +4 bp, 1-month +4 bp, 3-month +2 bp. What is notable is that the very front cheapened at all with overnight funding flat and reverse repo emptying, because it means 31 August month-end is being priced in term while the overnight market shows no strain whatsoever. Watch whether SOFR goes back through IORB on the 31st with the September operation on the 9th behind it.

(c) Rates volatility and swap spreads

The MOVE series failed for a third consecutive session and this time it did not move at all. The Investing.com page still marks 69.86, +0.42 (+0.60%) with a 27/08 timestamp, a day range of 69.44-69.86 and a stated "previous close" of 95.74 that is 26 points outside that range. Both the staleness and the internal inconsistency are disqualifying, so the level is published with its vintage and the change withheld for the third edition running. What can be said without the vendor is the cross-market fact that matters: equity volatility fell on the day the two-year moved 14 basis points and the curve delivered its largest flattening since June. If rates volatility did rise on Friday — and a 14 bp front-end day almost guarantees it did — then MOVE and VIX have decoupled hard, and the equity market is the one that has not marked. The swap-spread basis is unchanged in substance: Treasuries have outperformed equivalent-maturity swaps since the buyback announcement, compressing the 30-year spread to a six-month extreme, with Fed researchers putting hedge-fund swap-spread positions at a record $305bn last year against under $50bn in 2022. That basis has now survived its largest scheduled event, which removes a catalyst rather than a risk.

(d) Issuance, leveraged loans and private credit

Levels from the TradingEconomics currency board taken after the U.S. close, with the vendor's own date column reading Aug/28, so the published %Chg measures the completed Friday session rather than a rolled Asian one. Investing.com instrument pages are used for USD/CNH and USD/TWD, which TradingEconomics does not carry on the majors board. Quote basis: USD per unit for EUR, GBP, AUD and NZD; units per USD for JPY, CHF, CAD, KRW, TWD, CNY and CNH.

The take: the dollar finally did the thing it has failed to do all week, and it took a central banker to make it. DXY rose 0.52% to 99.673, more than the previous three sessions combined, on a day when the two-year cheapened 14 bp and the September hike probability nearly doubled. Société Générale's Kit Juckes had set the precondition earlier in the week — dollar strength "will only return when (if) domestic data turn stronger and put pressure on the Fed to tighten" — and Friday is the counterexample that clarifies the rule: the data got materially worse, with Chicago at 47.1, and the dollar rose anyway, because the pressure to tighten arrived from the Chairman rather than from the economy. That distinction matters for durability. A dollar rally underwritten by growth is self-reinforcing; one underwritten by a communication style survives exactly as long as the next data point allows, and the next data point is ISM manufacturing on Tuesday.

The yen through 160 is the position with the most convexity into the weekend. USD/JPY rose 0.42% to 159.972 on TradingEconomics and 160.09 on Bloomberg's mark, the weakest level in a month, on a day Japanese 10-year yields also rose 4 bp to 2.91%. A bond market selling off while its currency weakens is a carry trade being added to, not unwound — and 160 is the handle at which Tokyo has historically found its voice. The Nikkei rose 0.41% in the same session, so the equity market is being paid by the currency, which is precisely the configuration that makes an intervention headline expensive to be short into over a three-day gap with no U.S. futures session until Sunday 6:00 p.m. ET.

The Asian crosses inverted again, for a third consecutive session, and the mechanism is still the Bank of Korea. USD/KRW fell 0.23% to 1378.64 and USD/TWD 0.06% to 31.661 — both local currencies stronger on a day the dollar index rose half a percent, which is a genuinely unusual pairing. And they did it while the Kospi fell 1.79% and Taiwan rose 0.77%, so the currency signal is not tracking the equity signal in either market. The reading this report has carried for two sessions holds: the rate differential from the Bank of Korea's second consecutive hike to 3% is doing the work, foreign equity flow is not, and currency-first-equity-second is the ordering that usually survives. Against that, USD/CNH rose 0.19% to 6.7313, giving back Thursday's closing extreme — the one Asian currency that took the dollar move at face value, and the one whose fix is administered.

Settlement basis, stated: all rows are the Investing.com per-contract historical close for 28 August, which is this report's settle series of record. Weekly and year-to-date columns are TradingEconomics spot returns on the vendor's Aug/28 date stamp and are shown for direction and magnitude only. Day changes are computed against the same Investing.com series' 27 August closes, which differ by fractions of a cent from the quote-page figures published in the prior edition; the reconciliation is in Data Notes.

*\YTD figures marked with an asterisk are TradingEconomics spot year-to-date returns, not futures returns on the contracts quoted above. Weekly columns are TradingEconomics spot weekly changes on the same caveat. The daily settles, changes and percentage moves are on the futures basis named in each row. Mixing the two would be a basis error; they are presented in separate columns for exactly that reason.

The metals complex broke, and silver's intraday range is the whole argument. Silver printed 71.160 and settled 66.255 — a 6.89% high-to-close fade and a 4.57% loss on the day — while gold printed 4,685.45 and settled 4,504.10, a 3.87% fade and a 3.43% loss, its largest single-session decline of the reporting window. The gold-silver ratio went out to 67.98 from 67.18, ending three consecutive sessions of compression, because the higher-beta leg gave back more. Bitcoin fell 3.2% to $77,505.37 and ether 3% to $2,431.68 (Bloomberg), back below the $80,000 reclaimed on Thursday. The mechanism is a rise in the expected real policy rate: a Chairman who says underlying inflation has not meaningfully improved and that "we have work to do" is simultaneously raising the nominal path and validating the inflation the metals were hedging — and the first effect dominates for a non-yielding asset. Note what this does to Thursday's framing. Bloomberg argued on Thursday that bitcoin and the AI trade were "benefiting from the same retail appetite for risk" rather than one funding the other. Friday tested that: bitcoin fell 3.2%, silver 4.57%, gold 3.43% and SOX 3.47%, all on the same catalyst. Four assets with four different theses moving together on one speech is a single-factor tape, and the factor is the discount rate.

The crack spreads went again, for a third session, and the gasoline leg has now moved $9.49 in three days. On the same 42-gallon basis this report has used all month, computed against front-month WTI:

RBOB has now risen 2.07%, 2.01% and 2.61% on three consecutive sessions while crude fell 0.16%, rose 1.58% and fell 0.11% — up, down and sideways in the barrel, and the product went up regardless every time. Bloomberg reported that funds added bullish gasoline bets at the fastest pace in six months, which is both the confirmation and the warning: the positioning that has driven the move is now visible, and the September calendar works structurally against the long leg from here. The energy complex as a whole did nothing — Finviz energy closed +0.05% with SLB +4.22% and Halliburton among the few movers, and the group is -2.12% on the week against a barrel that is 4.16% lower. Equity is marking the crude weakness and ignoring the refining margin.

Desk-style ideas for institutional investors. Each carries an explicit expression, catalyst and invalidation. These are not personalized investment advice; verify independently and size to your own mandate before acting.

1. The rates trade — the calendar spread paid its entire thesis in one session; take half the remaining half, raise the stop hard

Mark first, honestly, because this is the best day the book has had. Long ZQU6 (September 2026) against short ZQZ6 (December 2026), DV01-matched one-for-one at $41.67 per basis point per contract, entered eight sessions ago at 96.325 / 96.160 for a spread of 16.5 bp, marked through 18.5, 20.5, 18.5 and 19.5 bp with the stop last raised to 18.5. Friday's mark: ZQU6 96.305, ZQZ6 96.035 — a spread of 27.0 bp. That is +7.5 bp on the day, worth +$312.53 per contract pair on the retained half, and it leaves the trade +10.5 bp from entry — more than triple the best previous mark, achieved on a single speech.

Why it worked, and it is the mechanism, precisely. The spread needs December to absorb more of every repricing than September. Friday delivered the largest version imaginable: ZQU6 fell 2.0 bp while ZQZ6 fell 9.5 bp, because the September meeting sits mid-month and only captures about 47% of its contract's averaging period while December captures all of it and inherits October's repricing on top. In probability terms, September's hike odds rose 21.8 pp on Investing.com and December's cumulative-above rose 13.8 pp, but December's modal outcome changed range — from one hike to two — which is worth far more in price than September's larger probability move.

The modal path, the base case and the tails. Modal path: a 25 bp hike on 16 September is now the modal outcome for the first time (Investing.com 55.9%, CME 59.7%, ease 0.0%); a hike is modal at 28 October too, at 52.1% against 70.1% cumulative above; and two hikes by 9 December is now the single most likely year-end state, at 38.9% for 4.00-4.25% against 38.5% for one and 11.6% for none. The 2027 strip is modal 4.00-4.25% at every meeting, troughing at 95.770 in July and September 2027, 15 bp below Thursday. Base case: the market has moved from pricing one hike late to pricing two, and a September-versus-December calendar spread is the cleanest expression of exactly that migration. Tail one, and it is the live one: ISM manufacturing on 1 September confirms Chicago's 47.1, the strip unwinds a 59.7% September hike, and the spread compresses violently — remember that one basis point of ZQ price is worth about ten percentage points of probability, so this can retrace 5 bp in an hour. Tail two: payrolls on 4 September come in hot, September goes above 75%, and the spread compresses from the other side as the front contract catches up. Practical implication: a trade that has made 10.5 bp in eight sessions on a thesis that just became consensus is a trade to harvest, not to defend. Take half of the remaining half — go to a quarter — and raise the stop to 23.0 bp, which locks in 6.5 bp on whatever is left.

Expression: long ZQU6 / short ZQZ6, DV01-matched one-for-one, quarter size after the trim. Catalyst: ISM manufacturing and JOLTS 9/1 10:00; ADP 9/2 08:15; payrolls 9/4 08:30; the buyback operation 9 September; the 16 September FOMC. Invalidation: the spread through 23.0 bp; or the September cumulative hike printing below 45% on either vendor; or any 2026 meeting showing a non-zero cut probability. Sizing: a quarter, at $41.67 per basis point per pair.

2. Long the 20-year against the 30-year, on the November refunding — hold the half, the event risk is behind it

Expression: long the 20-year bond against short the 30-year, DV01-neutral, half size. Mark: 20s30s at 1 bp, unchanged on the day — the 30-year at 5.22% and the 20-year at 5.21%, both 3 bp higher — and 1 bp flatter on the week, a small loss. What changed: the event this position was carried through has happened, and it resolved the good way for the long end without resolving anything for the spread. JPMorgan, Apollo and Morgan Stanley were all publicly positioned for a hawkish Warsh to buy 30-year bonds; he was hawkish, and the 30-year rose three basis points against the two-year's fourteen, richening 5 bp on the week. Alyce Andres's framing — that "the composition of any bond selloff" is what matters — was answered decisively in favour of a market adjusting to a new Fed regime rather than questioning whether it has one. That removes the tail this trade was most exposed to. The issuance thesis is untouched: Citi has pushed its forecast for larger auctions to 2028 and raised the tail risk that Treasury eliminates the 20-year. Catalyst: the 9 September buyback operation and its maturity buckets; Treasury's 4 November quarterly refunding. Invalidation, unchanged: 20s30s through -3 bp, or an explicit Treasury statement ruling out changes to long-end auction sizes. Sizing: a half, unchanged.

3. Protection on the CCC cohort funded in IG — go back to a half; the Chairman named the trade

Expression: long CCC-exposed credit protection (or short a levered-loan / CCC-heavy vehicle) against long IG cash. Mark: CCC 1,031 bp, unchanged; IG 79 bp, -1 bp on the 27 August FRED update, taking the CCC-minus-HY differential back out to 768 bp from 764 — a 4 bp gain and a full reversal of Thursday's counter-signal. The reason to add rather than hold: on Thursday this position was cut to a quarter because the tail had tightened 8 bp, the most in a session that month, and one counter-signal at that magnitude deserved respect. Friday returned the differential to within a basis point of its record and produced something the thesis never had — an explicit statement from the Federal Reserve Chairman that credit spreads "near the low ends of their historical ranges" are evidence that policy is not restrictive. That is the mechanism by which tight aggregate credit stops being a support and becomes a cause. If the strip is right that a September hike is 59.7% likely, the cohort that pays floating on leveraged loans is the one that finds out first, and it is priced at a record dispersion to the index. Action: back to a half. Catalyst: 31 August month-end; ISM manufacturing 9/1; the September IG calendar clearing; Broadcom 9/2, because the AI capex line is what the calendar funds. Invalidation: the differential back through 750 bp, or IG widening beyond 85 bp — which would mean the whole complex is repricing rather than the tail. Sizing: a half.

4. New — short the debasement complex against long the dollar, into the real-rate turn

Expression: short an equal-weighted basket of Comex gold, Comex silver and a bitcoin proxy against long the dollar index, quarter size, dollar-notional matched. Thesis: Friday is the first session of the reporting window in which all three legs of the debasement trade fell together and fell hard — gold -3.43%, silver -4.57%, bitcoin -3.2% — against a dollar index +0.52%, on a catalyst that is structurally repeatable. The catalyst is not inflation; it is the expected real policy rate, and a Chairman who has committed to "a discipline, not to a decision" while stating that PCE at 3.7% and 54% of the basket above 3% means "we have work to do" has raised it without touching the target. Note the two intraday fades, because they are the evidence that this was liquidation rather than repricing: silver made 71.160 and closed 66.255 (-6.89%), gold made 4,685.45 and closed 4,504.10 (-3.87%). Positioning that gives back that much of a day's range is positioning that was long into the event. Catalyst: ISM manufacturing 9/1, payrolls 9/4 and CPI 9/11 — note that all three cut both ways, which is why this is a quarter and not a half; Chinese physical demand on Monday's Shanghai open, which has to absorb a gap it never traded. Invalidation: gold reclaiming $4,664, its 27 August settle, which would mean the move was a one-day liquidation; or DXY back below 99.11, Thursday's close. Sizing: a quarter, and note the obvious risk — Warsh explicitly flagged that "the recent rise in overall commodity prices also bears watching," and wheat is up 51% on the year (Section 2 item 10), so the inflation the metals hedge is visibly accelerating in the one place he named.

5. Long silver against short gold — close it; the thesis was invalidated by name

Expression: long Comex silver against short Comex December gold, notional-matched, small. Mark: silver -4.57% against gold -3.43% is a 1.14-point loss, the worst session of the trade's life, and the gold-silver ratio went out to 67.98 from 67.18, ending three consecutive sessions of compression. The honest reading: the stated invalidation levels were the ratio through 71 and copper below $6.40, and neither was hit — the ratio is 67.98 and copper closed $6.5460. This is being closed anyway, on the ground that the thesis was invalidated even though the levels were not. The position was underwritten on monetary debasement, on the argument that silver is levered to both the industrial and the monetary leg. On Friday the person most able to invalidate that thesis stood up and did so, and the higher-beta leg took the larger loss — which is what a ratio trade does when the shared factor turns against both legs. Holding a position whose reason has been publicly refuted because its stop has not been hit is how a small loss becomes a large one. Action: close it, book the 1.14 points, and note that the trade still made money over its life — it gained 1.60 points on Thursday and 1.07 ratio points across the three sessions before Friday gave 0.80 of them back.

6. Long the gasoline crack against the distillate crack — take the last third off; three sessions, $9.49, done

Expression: long the RBOB crack against short the heating-oil crack, both on the standard 42-gallon basis against front-month WTI, barrel-for-barrel, currently a third of a small position. Mark: the gasoline crack rose $3.80 to $62.51 while the distillate crack rose $2.79 to $98.97 — a $1.01 move in the differential, on top of $2.33 and $3.34, for a three-session total of $6.68. Why to finish rather than ride: the differential's daily gain has now decayed from $3.34 to $2.33 to $1.01 across three sessions while the underlying gasoline crack has gone from $53.02 to $62.51, and Bloomberg reports funds adding bullish gasoline bets at the fastest pace in six months. A trend whose second derivative has turned negative, whose cumulative move is roughly a full quarterly range, and whose positioning has just become a headline is a trend to be out of. The Labor Day driving-season roll-off is next week. Action: take the last third off; flat. Catalyst for anyone re-entering later: weekly EIA product inventories; the September gasoline calendar; enforcement detail on the Iran sanctions programme.

7. Short the AI-halo basket against long the name that actually reported — hold, but the mark is bad and the reason matters

Expression: short an equal-weighted basket of Synopsys, Palo Alto Networks, ServiceNow, Fortinet and Adobe against long Nvidia, beta-adjusted, quarter size. Mark: the basket returned -4.79%, -2.94%, +4.54%, -3.92%, +0.82% for an average of -1.26%, against Nvidia -4.58% — a 3.32-point loss on day one. The honest reading: the short leg worked exactly as designed in three of five names, with Synopsys and Fortinet giving back roughly a third of Thursday's gaps. The long leg is what failed, and it failed for a reason that has nothing to do with the thesis: Nvidia fell as part of a 3.47% semiconductor decline driven by the front end, not by any reassessment of its guidance. ServiceNow +4.54% is the genuine problem — it is the one halo name that extended, and it is the single largest contributor to the loss. Invalidation check: Nvidia at $217.55 is above the $209.66 pre-print level, so that condition is intact; the basket has outperformed Nvidia by 3.32 pp cumulatively against a 6 pp threshold, so that condition is intact too, with roughly half the budget spent in one session. Action: hold the quarter, do not add, and be disciplined about the threshold. Catalyst: Palo Alto Networks reports after the close on Tuesday 1 September — the largest name in the short basket, which has now given back 2.94% of its 12.83% pre-print gain; Broadcom, HPE and NetApp Wednesday 2 September.

Prior closes, marked forward. The on-balance-sheet-versus-off-balance-sheet AI funding pair, closed on Thursday at a loss, would have lost a further 9.52 points on Friday: Marvell fell 10.26% against Broadcom's 0.74%. That is the first exit in five editions that was vindicated within a session rather than punished, and it is recorded with the same weight as the four that were not. The short-utilities-versus-S&P pair, closed a week ago, would have made 0.87 points: utilities fell 1.12% against the index's 0.25%. The equal-weight-versus-cap-weight pair would have roughly broken even on 234 advancers against 259 decliners under a 0.25% decline — a far kinder tape than Thursday's, and a reminder that the exit was timed to the wrong variable rather than the wrong direction.

The vol note. VIX closed 14.43, down 0.55%, on a day the two-year cheapened 14 basis points, 2s30s flattened 11, a regional PMI missed by 11.2 points and the September hike probability moved 24.3 points on CME's own columns. A 14.43 handle asks for roughly a 0.90% daily move. Consider what is in front of it: a three-day gap with no U.S. futures session until Sunday 6:00 p.m. ET, into an Asian market that has not traded any of Friday's repricing and a Chinese physical gold bid that has to absorb a 3.43% gap; 31 August month-end; ISM manufacturing and JOLTS on 1 September, the single release most capable of unwinding the whole of Friday's move; ADP on the 2nd; ISM services and claims on the 3rd; and August payrolls on 4 September. Rates volatility cannot be marked — the MOVE series has not updated for three sessions — but the direction is not in doubt on a 14 bp front-end day, which means the two volatility markets have almost certainly decoupled and the equity one is the one that has not marked. MFS's Alex Mackey supplied the forward-looking version: less communication means more "bouts of volatility" around each future message. Own gamma dated 31 August through 4 September, in the index rather than in single names, and prefer it to vega — the catalysts are dense, dated and clustered, which is a gamma configuration, not a term-structure one.

Crowded consensuses to stress-test, with the numbers.

The two-sided geopolitical tape. Escalation: the U.S.-Iran war is at its six-month mark with the White House confirming no negotiations are under way; Black Sea attacks have halted almost all grain shipments through ports accounting for 70% of Russian grain exports, taking wheat to a three-year high; Canada's retaliation on $20bn of U.S. goods lands 8 September; and Washington is still weighing semiconductor tariffs reaching data-centre servers. De-escalation: crude ended the week 4.16% lower and finished Friday down 0.11% at $83.44, which is a market that does not believe the escalation is supply-relevant. Domestically, President Trump's Truth Social attack on "Big Ag" as a "nasty Monopoly," with legal documents being drawn to let farmers process their own food, knocked Tyson and JBS more than 1% pre-market on the same day grain hit multi-year highs — a two-sided squeeze on the protein complex.

Structural watch items. The six-month bill cheapened 8 bp to 4.02% on meeting risk while overnight funding was unchanged and reverse repo take-up fell to $175m, into a 31 August month-end; reserve balances at $2.9249tn, $68bn below the 5 August peak with no new print until the week containing the turn; a record ~$1.4tn of 2026 IG issuance funding a capex cycle Warsh says is more than half AI; token sales at the two leading labs above $100bn annualised, up more than 500% year on year, per the Chairman's own text; Japanese 10-year yields at 2.91% with the yen through 160; a Korean won that has strengthened for three sessions while the Kospi fell 2.88% in two; and a three-day gap in which Asia and Europe must price a 14 bp U.S. front-end move, a 3.47% semiconductor decline and a 3.43% gold decline that none of them traded.

Full source links and the complete Data Notes & Conflicts appendix are in the companion file US_CrossAsset_Daily_2026-08-28_DataNotes.txt, alongside the canonical Markdown report of record.