| | The tape in one paragraph. Thursday was the mirror image of Wednesday in every respect except the one that mattered: the bond market moved, and it moved because a Fed governor said the quiet part out loud. Christopher Waller told a Reuters event that "if there is continued progress toward our 2% goal, then I am willing to support holding the policy rate at its current level" — adding that his decision would be "heavily influenced by what we learn about August inflation," and that "if inflation comes in hot, I would consider a rate hike." The strip repriced a coin flip inside an afternoon: CME put the 16 September hike at 50.2% against its own settled 63.2% for Wednesday, a 13.0-point collapse, and Investing.com at 50.4% from 60.1%. Yields fell hardest where policy timing lives — the official par 1-year and 2-year each fell 5 bp, to 4.11% and 4.34%, against 2 bp at the 10-year to 4.77% and 2 bp at the 30-year to 5.25% — a front-end-led bull steepener that took 2s10s to 43 bp from 40. Equities took it straight: the S&P 500 rose 81.11 points, or 1.06%, to 7,747.71, its best session in a month, with the Dow +624.16 (+1.18%) and the Nasdaq Composite +1.40%. Breadth widened to 331 advancers against 161 decliners on Investing.com's 494-name board, and VIX fell 5.79% to 14.32. The Very-high window behind us was empty — no release rated Very high landed in the past twelve hours — but the two High ones both beat: initial claims printed 206,000 against a 205,000 WSJ consensus, and ISM services 55.4 against 54.2, a beat that would ordinarily lift hike odds and instead was buried by Waller. Ahead, the only Very high release in the next twenty-four hours is the August Employment Situation, Friday 08:30 ET, consensus +55,000 payrolls with unemployment held at 4.1%. Now the fades, and there are three. The dollar broke — Bloomberg's gauge hit its lowest since May and USD/JPY fell 1.73% to 155.985, the yen's best session of the window and roughly double Wednesday's — while gold settled $4,539.90, up $125.30 or 2.84%, a move nearly five times the dollar's. Second, the semiconductors did not come: SOX rose 0.11% against the Nasdaq 100's 1.16%, and Broadcom fell 2.74% to $357.16 on the day Macquarie upgraded it to outperform with a $490 target. Third, the leadership was pure reversal — Robinhood +16.57%, Palantir +7.71%, ServiceNow +6.49% and CrowdStrike +5.68% were, almost to the name, Wednesday's worst performers. After the bell Lululemon fell as much as 20% on a second consecutive full-year guidance cut. The index rose because the front end fell; almost nothing else about the session was a re-rating. |
| | Index | Close | Chg | %Chg | Note | | S&P 500 | 7,747.71 | +81.11 | +1.06% | Range 7,686.71-7,756.76. 331 advancers vs 161 decliners on Investing.com's 494-name board. 0.66% below the 13 August record close of 7,798.99 | | Nasdaq Composite | 26,584.06 | +366.23 | +1.40% | Range 26,325.06-26,644.57. The best of the five headline indices | | Dow Jones Industrials | 53,686.11 | +624.16 | +1.18% | Range 53,286.15-53,746.50 | | Nasdaq 100 | 29,482.32 | +338.99 | +1.16% | Range 29,160.96-29,538.76 | | Russell 2000 | 2,968.27 | +12.32 | +0.42% | Range 2,951.29-2,976.62. The day's laggard, exactly inverting Wednesday. Change computed against the prior close; see Data Notes on the vendor column | | VIX | 14.32 | -0.88 | -5.79% | Range 14.23-15.44. A second consecutive decline, and the lowest close of the reporting window | | PHLX Semiconductor (SOX) | 11,352.1 | +12.9 | +0.11% | Range 11,056.4-11,371.3. A tenth of the Nasdaq 100's move | | UST 1Y (official par) | 4.11% | -5 bp | — | +7 bp on the week. Joint-largest daily fall on the curve | | UST 2Y (official par) | 4.34% | -5 bp | — | +14 bp on the week. WSJ marks it 4.351% overnight | | UST 10Y (official par) | 4.77% | -2 bp | — | +10 bp on the week. WSJ 4.771%, Bloomberg 4.77% | | UST 30Y (official par) | 5.25% | -2 bp | — | +6 bp on the week. 20s30s still exactly flat | | Gold (Comex Dec) | $4,539.90 | +$125.30 | +2.84% | Settle basis. The largest move on any board today | | Silver (Comex Dec) | $67.704 | +$2.241 | +3.42% | Gold-silver ratio 67.05 from 67.44 | | WTI (Oct, NYMEX) | $91.30 | +$0.29 | +0.32% | Settle basis; see Data Notes on the prior-edition reconciliation | | Brent (Nov, ICE) | $95.52 | -$0.11 | -0.12% | The only energy contract lower, on a weaker dollar | | Copper (Comex Dec) | $6.6710 | +$0.0780 | +1.18% | Joined the metals bid for once | | Natural gas (Oct, NYMEX) | $2.928 | -$0.028 | -0.95% | Gave back Wednesday's $3 handle | | DXY | 98.989 | -0.59% | — | Computed against the prior edition's vendor level. Bloomberg's dollar gauge at its lowest since May |
| | 2 · Market Hot Spots (ranked by tradability) |
| | 1. | One Fed governor moved a meeting thirteen points, and everything else on the tape is downstream of it. Christopher Waller told a Reuters event he is "willing to support holding the policy rate at its current level" if disinflation continues, called the current 3.50%-3.75% range "only slightly restricting aggregate demand," and said he does "not see elevated energy prices and tariffs now as a significant source of ongoing inflation pressure." CME's September hike fell to 50.2% from a settled 63.2%; Investing.com's to 50.4% from 60.1%. Evercore's Krishna Guha: "We repeat our call that the Fed is more likely to hold than hike in September, though we think it is close and will indeed turn on the next set of inflation data." Forward catalyst: payrolls Friday 08:30, then CPI on 11 September, the last inflation print before the 15-16 September meeting. |
| | 2. | The curve bull-steepened from the front, which is the correct signature for a policy-timing repricing rather than a term-premium one. The 1-year and 2-year each fell 5 bp, the 3-year 4 bp and the 7-year 3 bp, against 2 bp at the 10-year, 20-year and 30-year — so the belly and front absorbed two-and-a-half times the long end's move. 2s10s widened to 43 bp from 40, 2s30s to 91 from 88, and 3M10Y to 88 from 87. On the week the same curve is 14 bp cheaper at the 2-year and 6 bp at the 30-year, so Thursday retraced the shape of the selloff without retracing its level. Forward catalyst: the 9 September buyback operation, and Friday's payroll. |
| | 3. | The rally's leadership was a reversal list, not a conviction list, and the overlap with Wednesday's losers is close to exact. Robinhood +16.57% to $124.72 led the index; behind it Palantir +7.71%, ServiceNow +6.49%, CrowdStrike +5.68%, Oracle +5.69%, S&P Global +4.37%, PG&E +4.73% and Omnicom +3.63% — every one of which fell on Wednesday, five of them by more than 4%. Against that, Wednesday's leaders reversed: Charter -4.77%, Celanese -3.13%, FMC -2.63%, Eastman -1.93%, Steel Dynamics -0.65%. A one-day rate-relief move that pays the highest-beta losers and takes back from the prior day's winners is a short-covering tape. Forward catalyst: whether the reversal factor survives a payroll print. |
| | 4. | Semiconductors would not participate, and that is the cleanest non-confirmation on the board. SOX rose 0.11% against the Nasdaq 100's 1.16% and the Nasdaq Composite's 1.40% — a tenth of the index move on a day long-duration equity was supposed to be the whole trade. Broadcom fell 2.74% to $357.16 even as Macquarie upgraded it to outperform with a $490 target, about 37% above the close, and Bloomberg's framing was that its AI chip forecast "failed to impress." Micron managed +0.22%, Applied Materials -0.58%, Texas Instruments -0.38%, Qualcomm -0.82%, KLA +0.39%. Forward catalyst: Oracle on 8 September as the next AI capital-expenditure read. |
| | 5. | The dollar broke and gold went with it, at five times the leverage. Bloomberg's dollar gauge fell to its lowest since May; DXY closed 98.989, down 0.59% on the prior edition's vendor level, and USD/JPY fell 1.73% to 155.985 with an intraday extreme of 155.30. Gold settled $4,539.90, up 2.84% and silver $67.704, up 3.42% — moves four to six times the dollar's, which means the metals were pricing the real rate rather than the currency. Newmont rose 4.21% as the only major-miner participant of size. Forward catalyst: the Bank of Japan on 18 September, with Nomura telling Bloomberg three consecutive hikes are possible in an extreme case. |
| | 6. | Packaged food broke as a bloc, and a dividend cut was the trigger. Campbell's fell 6.98% to $22.13 after cutting its dividend 36% and announcing a $500m savings plan alongside a sales miss; the read-across was indiscriminate — Tyson Foods -7.26%, Lamb Weston -3.76%, Conagra -3.46%, General Mills -3.25%, Kraft Heinz -3.20%, McCormick -2.27%, JM Smucker -2.23%, Bunge -2.23%. Nine names down an average of 3.6% on a session the index rose 1.06% is a 4.6-point relative move in one sub-industry. Forward catalyst: whether a second staples name cuts a payout, which would turn a read-across into a re-rating. |
| | 7. | Energy was the only red sector on a day crude settled higher, and the split was in the products. Energy fell 0.67% on Finviz's board while WTI settled $91.30, up 0.32% — because heating oil fell 1.78% to $4.5987 while RBOB rose 1.00%, exactly inverting Wednesday. Exxon -1.18%, ConocoPhillips -1.08%, EOG -2.01%, Halliburton -0.90%, Occidental -0.49%. Bloomberg reported Iran claiming fresh strikes on US bases against Saudi Arabia's smaller-than-expected increase in its flagship crude prices. Forward catalyst: any Hormuz headline into a three-day weekend. |
| | 8. | The distillate premium collapsed, and it collapsed on the second consecutive session in which crude rose. On the October basis against October WTI the distillate crack fell $3.79 to $101.85 while the gasoline crack rose $1.02 to $40.37, narrowing the differential $4.81 to $61.48. That is the invalidation condition this report wrote down twenty-four hours ago — a second straight session in which the distillate leg loses money outright while crude rises — and it has now fired. Forward catalyst: the post-Labor-Day gasoline roll-off, and the first winter distillate builds. |
| | 9. | Volatility fell in both markets at once for the first time in the window, which removes the divergence this report has been flagging. VIX fell 5.79% to 14.32, the lowest close of the window, and the MOVE index fell 6.31% to 74.68 on its own day range of 74.68-79.71 — so rate volatility and equity volatility declined together, after a fortnight in which MOVE rose while VIX did not. The MOVE/VIX ratio is 5.22. Forward catalyst: a payroll landing into a three-day weekend against a 14.32 handle that asks for roughly a 0.90% daily move. |
| | 10. | The credit tail widened for a fourth consecutive update while the aggregate did nothing, and the cash market again declined to confirm. CCC widened 4 bp to 1,053 on the 2 September FRED stamp, after 7 and 16 in the two prior updates, taking the CCC-minus-HY differential to 787 bp from 784 — a fresh window extreme. Against that, IG was unchanged at 81 and HY widened 1 to 266, while HYG closed +0.13% and LQD +0.14%. Forward catalyst: the September IG calendar, which Bloomberg sizes near $215bn, pricing into a 4.77% ten-year. |
| | 3 · Sector Performance — September 3, 2026 |
| | Sector | 1-Day | 1-Week | YTD | | Industrials | +1.87% | -1.28% | +10.95% | | Financial | +1.69% | +1.67% | +9.96% | | Consumer Cyclical | +1.40% | +0.48% | -3.59% | | Communication Services | +1.37% | +1.89% | -0.08% | | Technology | +1.25% | -0.84% | +24.93% | | Real Estate | +0.93% | -1.02% | +8.90% | | Utilities | +0.78% | -0.44% | -0.27% | | Basic Materials | +0.53% | -1.89% | +20.89% | | Healthcare | +0.26% | +0.55% | +11.29% | | Consumer Defensive | +0.17% | +0.88% | +7.29% | | Energy | -0.67% | +3.10% | +40.02% |
| | 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 closed. Compounding each group's 2 September YTD by Thursday's one-day move reproduces the published YTD to within 0.01 percentage points for ten of eleven groups. Technology, the group this report flagged at 0.345 pp of drift on Wednesday, now reconciles exactly: 1.2339 × 1.0125 = 1.24932, or +24.93% against a published +24.93%, deviation 0.000 pp. The single exception is real estate at 0.09 pp — 1.0781 × 1.0093 implies +8.81% against a published +8.90% — which is the second consecutive session that group has carried the widest deviation on the board. Clean examples: energy 1.4096 × 0.9933 = 1.40015 → +40.02% against +40.02%, deviation 0.000; consumer cyclical 0.9508 × 1.0140 = 0.96411 → -3.59% against -3.59%. A one-day correction of 0.345 pp in technology means Wednesday's gap was a vendor timing artefact rather than a compounding error, and it should not be carried forward. | | The rotation is cyclical, and it is the exact inverse of Wednesday's leadership. Ten of eleven groups closed green with industrials, financials, consumer cyclicals and communication services in front — three of which were at or near the back of Wednesday's board, industrials most starkly, at exactly 0.00% twenty-four hours earlier and +1.87% now. Energy is the only red group at -0.67%, having led Wednesday's board on the week and still leading the year at +40.02%. Basic materials, Wednesday's best group at +1.83%, fell to eighth at +0.53%, and its week is now the worst on the board at -1.89%. This is a rate-relief rotation: the groups that pay for a lower discount rate led, and the group that pays for a higher oil price lagged. | | The cross-check disagrees, and the disagreement is worth stating rather than smoothing. Investrade's GICS tally for the same session marks consumer discretionary +1.29%, financials +0.98% and technology +0.75% as the leaders — consistent with Finviz on all three — but puts materials -0.21%, healthcare -0.28% and consumer staples -0.30% at the bottom, and counts eight of eleven sectors higher against Finviz's ten-up-one-down. The three contested groups all sit within half a percentage point of zero on both boards, and the difference is a classification-and-weighting artefact: Finviz's groups are broader and less cap-concentrated than the GICS sector indices. Neither board is wrong; the leaders agree and the near-zero tail does not, so the leaders are the tradeable statement. | | The composition traps to name. Technology at +1.25% badly overstates the semiconductor complex — SOX rose 0.11% and Broadcom fell 2.74%, so the group is carried by software and hardware: ServiceNow +6.49%, Oracle +5.69%, CrowdStrike +5.68%, Dell +4.91%, Hewlett Packard Enterprise +5.04%, Gartner +4.66% and CDW +3.53%, against Microsoft +2.68% and Nvidia +1.80%. Consumer defensive at +0.17% is the group most damaged by its own internals: the nine packaged-food names in Section 2 fell an average of 3.6%, so the group's small positive is Walmart at +2.20% and Costco at -0.33% cancelling a food-and-beverage wreck. Financial at +1.69% is genuinely broad — Goldman +3.34%, Citigroup +2.83%, Morgan Stanley +2.52%, Schwab +1.98%, BlackRock +1.72%, JPMorgan +1.64%, Nasdaq Inc +3.12% — and is the one group whose day, week and year all point the same way. | | 4 · Movers & Single-Name Catalysts |
| | Higher | | • | Robinhood Markets (HOOD) +16.57% to $124.72 — the best S&P 500 performer, on prediction-market research rather than on anything the company said. Deutsche Bank's Brian Bedell argued contracts on companies' key performance indicators could become "the largest event-contract category within prediction markets," reaching $1tn of U.S. volume by 2028; Piper Sandler's Patrick Moley set a $145 target, about 16% above the close, on NFL and NCAA season volumes after World Cup-driven gains. |
| | • | Palantir (PLTR) +7.71% to $182.53 — a complete reversal of Wednesday's 5.81% decline, on 38m-plus shares. |
| | • | ServiceNow (NOW) +6.49% to $145.59 and CrowdStrike (CRWD) +5.68% to $214.97 — the two most heavily sold software names of the prior two sessions, both recovering more than their Wednesday losses. |
| | • | Principal Financial (PFG) +6.47%, Axon Enterprise (AXON) +6.15% to $538.15, Gartner (IT) +4.66% to $195.46, CBRE (CBRE) +4.70% to $148.74. |
| | • | Oracle (ORCL) +5.69% to $154.04 — a second consecutive advance, now +8.99% across two sessions, five days before it reports. |
| | • | Tesla (TSLA) +5.42% to $376.37 — the best of the megacaps by a factor of two. |
| | • | Hewlett Packard Enterprise (HPE) +5.04% to $54.44 and Dell Technologies (DELL) +4.91% to $516.39 — the after-hours rejection of HPE's beat reversed in the regular session, and Fox Advisors upgraded Dell to outperform with a $625 target, about 21% above the close. |
| | • | PG&E (PCG) +4.73% to $13.96 — up on the day Truist cut it to hold with a $17 target, which is still 22% above the close. |
| | • | Copart (CPRT) +4.42% to $33.58 — JPMorgan upgraded to overweight, target to $40 from $32, about 19% of upside. |
| | • | S&P Global (SPGI) +4.37% to $450.58, Interactive Brokers (IBKR) +4.30%, ICE (ICE) +4.27% to $164.59, Nasdaq Inc (NDAQ) +3.12%, Cboe Global (CBOE) +1.49% — the exchange and data complex bought back after Wednesday's volatility-collapse selling. |
| | • | Newmont (NEM) +4.21% to $130.43 — the gold move's only large-cap equity expression of size. |
| | • | PayPal (PYPL) +3.93% to $56.82, Goldman Sachs (GS) +3.34% to $1,037.93, KKR (KKR) +3.13% to $109.76, Citigroup (C) +2.83% to $138.14, Morgan Stanley (MS) +2.52% to $217.15. |
| | • | Omnicom (OMC) +3.63% to $84.73 — recovering most of Wednesday's uncatalysed 5.03% decline. |
| | • | TE Connectivity (TEL) +3.56% to $209.41, CDW (CDW) +3.53% to $153.92, Aptiv (APTV) +3.33% to $46.53, Salesforce (CRM) +2.92% to $264.43, Meta Platforms (META) +3.01% to $610.68, Microsoft (MSFT) +2.68% to $510.12, NetApp (NTAP) +2.55% to $185.38, FedEx (FDX) +2.45% to $324.31, Walmart (WMT) +2.20% to $108.42, Adobe (ADBE) +2.13% to $285.75. |
| | • | Nvidia (NVDA) +1.80% to $228.45 — Bloomberg reported the agreed acquisition of Hugging Face at about $13bn, firming up Wednesday's "nearing $14bn." Alphabet A +1.59%, Amazon +1.54%, Apple +1.00% to $328.21, Intel +1.80% to $91.67. |
| | Lower | | • | Tyson Foods (TSN) -7.26% to $51.76 and Campbell's (CPB) -6.98% to $22.13 — the two worst S&P 500 performers. Campbell's cut its dividend 36%, announced $500m of savings and missed on sales with a 37% earnings decline; Tyson carried no separate dated catalyst in the reviewed sources and moved with the group. |
| | • | Charter Communications (CHTR) -4.77% to $151.38 — a full round trip of Wednesday's 8.74% chief-financial-officer relief rally. |
| | • | Albemarle (ALB) -4.06% to $132.16, Lamb Weston (LW) -3.76% to $50.44, Conagra (CAG) -3.46% to $15.62, General Mills (GIS) -3.25% to $39.26, Kraft Heinz (KHC) -3.20% to $25.42. |
| | • | Celanese (CE) -3.13% to $45.20 and Dow (DOW) -2.94% to $30.36 — chemicals gave back a two-day rally in one session. |
| | • | Boston Scientific (BSX) -2.94% to $46.95, Broadcom (AVGO) -2.74% to $357.16, LyondellBasell (LYB) -2.72% to $64.76, FMC (FMC) -2.63% to $12.98. |
| | • | Chipotle (CMG) -2.47% to $37.57, McCormick (MKC) -2.27% to $52.58, Booking (BKNG) -2.24%, Bunge (BG) -2.23% to $120.12, JM Smucker (SJM) -2.23% to $128.51. |
| | • | EOG Resources (EOG) -2.01% to $145.96, Eastman Chemical (EMN) -1.93% to $70.77, DoorDash (DASH) -1.87% to $222.00, Freeport-McMoRan (FCX) -1.85% to $72.56 — the copper equity fell on a day the metal rose 1.18%, which is the day's neatest single-name contradiction. |
| | • | IDEXX (IDXX) -1.80%, Autodesk (ADSK) -1.74% to $237.52, Constellation Energy (CEG) -1.72% to $285.05, Air Products (APD) -1.69%, DuPont (DD) -1.35% to $131.13, Moderna (MRNA) -1.29% to $148.87, Exxon Mobil (XOM) -1.18% to $162.21, ConocoPhillips (COP) -1.08% to $135.72. |
| | • | Deere (DE) -0.57% to $694.41 — down on the day Evercore ISI upgraded it to outperform with an $813 target, about 17% above the close. |
| | Analyst and corporate actions | | • | Macquarie upgraded Broadcom to outperform from neutral with a $490 target, about 37% above the close — and the stock fell 2.74%. That is the single largest published upside on the day's call sheet attached to the day's most conspicuous decliner. |
| | • | Evercore ISI upgraded Deere to outperform with an $813 target; the stock fell 0.57%. Fox Advisors upgraded Dell to outperform at $625; the stock rose 4.91%. JPMorgan upgraded Copart to overweight, target to $40 from $32; the stock rose 4.42%. JPMorgan upgraded Envista to overweight, target to $32 from $29. |
| | • | Downgrades: Rothschild & Co Redburn cut Moderna to sell with an $81 target, roughly 46% below the close, and the stock fell only 1.29%; Truist cut PG&E to hold with a $17 target from $21, and the stock rose 4.73%; Bank of America cut Sonoco to neutral at $60 from $69; Evercore ISI cut Ultragenyx to in line at $16 from $34; JPMorgan cut ZTO Express to neutral at $22 from $29. |
| | • | Initiations: Seaport Research started three energy names at buy — Devon Energy ($65), Permian Resources ($45) and Viper Energy ($45); CLSA started Thermo Fisher at outperform ($748); RBC started RingCentral at outperform ($85). |
| | • | Nvidia agreed to acquire Hugging Face in a transaction valued at about $13bn (Bloomberg), adding the software platform to its AI stack. |
| | • | Adobe named Anil Chakravarthy — president of its customer experience business — as chief executive; the stock rose 2.13%. |
| | • | OpenAI released GPT-6, positioning it as a milestone toward artificial general intelligence (Bloomberg). |
| | • | Lululemon (LULU) closed +1.42% at $121.77 and fell as much as 20% after hours, cutting full-year revenue guidance to $10.35-10.5bn, a 5-7% decline, and adjusted earnings to $9.48-9.73, on second-quarter revenue of $2.4bn, down 4%, with comparable sales down 9%. Interim co-chief executive and finance chief Meghan Frank: "While we continue to navigate some challenging dynamics, we are taking a prudent approach with our revised full-year outlook." It is the second consecutive quarter of a lowered full-year outlook, ahead of incoming chief executive Heidi O'Neill. |
| | • | Non-constituent prints worth the tape: Snowflake +21% on adjusted EPS $0.62 against $0.45 and revenue $1.55bn against $1.48bn; ChargePoint +45% on tariff refunds lifting margins; Five Below +5% on 14.1% comparable growth; MicroStrategy +17.56% to $144.82 on the bitcoin move; against them Ultragenyx -46% after apazunersen missed its primary endpoint in Angelman syndrome, Aptevo -16% and Polestar -14% on cut delivery forecasts. Costco reported August comparable sales up 8.4% overall and 9.0% in the U.S., and the stock still fell 0.33%. |
| | 5 · S&P 500 Earnings Calendar — Current & Next Week (S&P 500 components only) |
| | 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. | | Current week (Aug 31 - Sep 4) — remaining sessions |
| | Fri 9/4. No S&P 500 reporter on either bucket. This is the fourth consecutive capture returning an empty Friday. | | Next week (Sep 7 - Sep 11) |
| | Mon 9/7. U.S. equity markets are closed for Labor Day. The Nasdaq capture returns no S&P 500 reporter for the date. | | Tue 9/8. Timing bucket not published: Oracle (ORCL) — the reviewed calendar carries no before-open or after-close designation; confirm with company investor relations. | | Wed 9/9. AMC: Cooper Companies (COO). | | Thu 9/10. AMC: Adobe (ADBE), Copart (CPRT). | | Fri 9/11. BMO: Kroger (KR). | | Changes vs. the prior calendar (9/2 report): | | • | No additions and no removals. The rosters for 9/8 through 9/11 reproduce the prior capture name for name and bucket for bucket, which is the first fully unchanged capture of the reporting window. |
| | • | Copart (CPRT) is confirmed on 10 September in the after-close bucket for a second consecutive capture. The prior edition flagged the 3 September to 10 September re-dating as provisional; two consecutive agreeing captures resolve it, and the provisional marker is retired. |
| | • | Oracle's missing timing bucket on 8 September persists for a fourth consecutive capture and is a settled publisher gap rather than scheduling news. |
| | • | Thursday's three reporters — Campbell's, Toro and Lululemon — have dropped out of the current-week block under the forward-only rule; their results and reactions are in Section 2 and Section 4. |
| | • | Dual listings deduped; no dual-listed constituent appears on the covered dates this capture. |
| | • | Non-members on the same dates, listed so nobody mistakes their absence for an omission: KT Corp (KT), KNOT Offshore (KNOP), Children's Place (PLCE), Hurco (HURC) and Virco (VIRC) on 9/4; Grifols (GRFS), Apartment Investment (AIV) and Dynagas LNG (DLNG) on 9/7; Casey's (CASY), ServiceTitan (TTAN), GameStop (GME), Braze (BRZE), ABM, United Natural Foods (UNFI) and Mission Produce (AVO) on 9/8; Chewy (CHWY), SailPoint (SAIL), Core & Main (CNM), AeroVironment (AVAV), Korn Ferry (KFY), Signet (SIG), American Eagle (AEO) and Academy Sports (ASO) on 9/9; Descartes (DSGX), Macy's (M), RH, National Beverage (FIZZ), Zumiez (ZUMZ), Designer Brands (DBI) and Lovesac (LOVE) on 9/10; Hooker Furnishings (HOFT) and Rent the Runway (RENT) on 9/11. Borderline membership cases are listed in Data Notes and conservatively excluded. |
| | • | What the forward calendar hands the desk. Nothing tomorrow, nothing Monday, and then a single unbucketed name on Tuesday that happens to be the most important AI capital-expenditure disclosure of the month. Oracle on 8 September is the first reporter after a session in which its own stock rose 5.69% and 8.99% across two days, into a complex whose semiconductor leg refused a 1.16% Nasdaq 100 rally and whose largest recent beat was sold. The reaction function to carry forward is the one Section 4 documents twice this week in opposite directions: guidance was rejected on Wednesday evening and rewarded on Thursday morning, with the same names moving both ways inside eighteen hours. Then Cooper Companies Wednesday, Adobe and Copart Thursday — Adobe reporting five days after naming a new chief executive, and Copart eight days after a JPMorgan upgrade that lifted its target 25%. Size the reaction, not the number. |
| | 6 · U.S. Treasury Yields — Official Par Curve |
| | Source: U.S. Department of the Treasury Daily Treasury Par Yield Curve Rates, Text View for September 2026, read after publication. Changes are versus the 2 September official row (1-day) and the 27 August official row (1-week). | | Tenor | 3 Sep | 1-Day | 1-Week | | 1 Mo | 3.83% | 0 bp | +2 bp | | 3 Mo | 3.89% | -3 bp | +5 bp | | 1 Yr | 4.11% | -5 bp | +7 bp | | 2 Yr | 4.34% | -5 bp | +14 bp | | 3 Yr | 4.41% | -4 bp | +11 bp | | 5 Yr | 4.52% | -2 bp | +14 bp | | 7 Yr | 4.63% | -3 bp | +11 bp | | 10 Yr | 4.77% | -2 bp | +10 bp | | 20 Yr | 5.25% | -2 bp | +7 bp | | 30 Yr | 5.25% | -2 bp | +6 bp |
| | Spread | 3 Sep | 1-Day | 1-Week | | 2s10s | +43 bp | +3 bp | -4 bp | | 3M10Y | +88 bp | +1 bp | +5 bp | | 2s30s | +91 bp | +3 bp | -8 bp | | 20s30s | 0 bp | 0 bp | -1 bp |
| | The read: a front-end-led bull steepener, and the diagnostic is policy timing rather than term premium. Every tenor fell, but the fall was concentrated where the September and October meetings are priced — the 1-year and 2-year each -5 bp and the 3-year -4 bp, against -2 bp at the 10-year, the 20-year and the 30-year. That is a 2.5-to-1 ratio of front-end to long-end movement on a day whose only catalyst was a governor discussing the next meeting, which is exactly the signature you would expect and the opposite of what the past fortnight delivered. Set it against the week: on a one-week view the 2-year is 14 bp cheaper and the 30-year 6 bp, so Thursday changed the shape of the selloff without changing its level, and 2s10s at 43 bp is still 4 bp flatter than a week ago even after steepening 3 bp on the day. 20s30s printed exactly 0 bp for a seventh consecutive session — the long end has now spent a week refusing to differentiate between twenty and thirty years, which remains the cleanest evidence that the buyback programme has put a floor under the very long end. | | Vendor gaps, explained as timing rather than as level disputes. WSJ's overnight quote board at 10:49 p.m. ET marks the 10-year at 4.771%, the 2-year at 4.351% and the 30-year at 5.246%, against official par closes of 4.77%, 4.34% and 5.25%; Bloomberg's live board shows the 10-year at 4.76%, down 1 bp. All three sit within a basis point of the Treasury's 3:30 p.m. bid-side snapshot, and the differences are the four to seven hours of trading between the official fixing and the reads. | | The bill curve is doing something separate, and it is a financing signal rather than a policy one. Below one year the moves do not line up with the coupon curve at all: the 1-month was unchanged at 3.83%, the 2-month rose 2 bp to 3.91% — the only tenor on the entire fourteen-point curve to rise — while the 1.5-month fell 5 bp to 3.82% and the 6-month fell 5 bp to 3.95%. All four are off the published table under the trimmed tenor set and are cited here because they carry the story. A 2-month bill cheapening 2 bp while the 1.5-month richens 5 leaves a 9 bp inversion inside a fortnight of maturity, which no policy expectation can generate; it is settlement and supply, and it belongs with the funding data in Section 9 block b. Note also that the 4-month bill this report flagged on Wednesday at 4.02% fell 3 bp to 3.99% while the 6-month fell 5 to 3.95%, so the 4-month now sits 4 bp above the 6-month against 2 bp on Wednesday. The anomaly did not close; it widened, and it widened on a day the reverse repo facility took in $702m against $525m. | | 7 · U.S. Macroeconomic Calendar |
| | Source: Federal Reserve Bank of New York Economic Indicators Calendar for September 2026 (all times Eastern), cross-checked against the Wall Street Journal and Investing.com economic calendars. Consensus figures are carried 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 31 - Sep 4) — still to come | | Date | Time ET | Release | Period | Consensus | Sensitivity | | Fri 9/4 | 08:30 | Employment Situation | Aug | +55,000 payrolls (Investing.com, Bloomberg survey); unemployment 4.1%; private payrolls +50,000; average hourly earnings +0.3% m/m, +3.0% y/y | Very high | | Fri 9/4 | 10:00 | Global Supply Chain Pressure Index | Aug | — | Low | | Fri 9/4 | 12:45 | New York Fed Staff Nowcast | — | — | Low |
| | Next week (Sep 7 - Sep 11) | | Date | Time ET | Release | Period | Consensus | Sensitivity | | Mon 9/7 | — | Labor Day — U.S. markets closed. No release on the calendar | — | — | — | | Tue 9/8 | 11:00 | Survey of Consumer Expectations | Aug | — | Medium | | Thu 9/10 | 08:30 | Initial Jobless Claims | wk ended 9/5 | No verified consensus published in the reviewed sources | High | | Thu 9/10 | 08:30 | Producer Price Index (PPI) | Aug | No verified consensus published in the reviewed sources | High | | Thu 9/10 | 10:00 | NAR Existing Home Sales | Aug | — | Medium | | Thu 9/10 | 10:00 | Wholesale Trade | Jul | — | Low | | Thu 9/10 | 11:30 | Weekly Economic Index | wk ended 9/5 | — | Low | | Fri 9/11 | 08:30 | Consumer Price Index | Aug | No verified consensus published in the reviewed sources | Very high | | Fri 9/11 | 10:00 | Michigan Consumer Survey (Preliminary) | Sep | — | Medium | | Fri 9/11 | 12:45 | New York Fed Staff Nowcast | — | — | Low |
| | The look-ahead: the data beat and the market priced a hold anyway, which tells you the reaction function has changed hands. ISM services printed 55.4 against a 54.2 consensus — up 1.3 points on July and 1.7 points above its twelve-month average — and initial claims 206,000 against 205,000, with the four-week average at 207,250. Both are the kind of prints that had been adding to hike odds all week. Instead the September meeting lost 13 points on CME, because Waller reframed the decision as an inflation decision rather than an activity decision: "my decision on the appropriate stance of policy will be heavily influenced by what we learn about August inflation." That single sentence moves the operative catalyst from Friday's payroll to CPI on 11 September, and it is why the asymmetry has inverted. Two days ago a soft payroll was the only path to a hold; now a hold is the base case and a hot CPI is the only clean path back to a hike. The supporting data are mixed rather than weak: Q2 productivity was revised to +1.4%, exactly in line, with unit labour costs at +1.2% against a +1.3% consensus and a +1.3% prior — a genuine downside surprise on the wage-cost side that supports Waller's disinflation case; the July trade deficit came in at -$88.58bn against -$90.0bn; and Challenger counted 52,881 August job cuts, up 58% on the month but down 38% on the year, with Andy Challenger noting companies "are making plans to hire more workers than last year" while "it doesn't appear those positions are being filled quickly." The hooks, in the order they can move the Fed card. (1) The August Employment Situation, Friday 08:30, the last payroll before the meeting: consensus +55,000 after a July print that fell 23,000, with unemployment held at 4.1% and average hourly earnings decelerating to +3.0% year on year from +3.2%. The Sevens Report's Tom Essaye frames the two tails cleanly — "a Goldilocks report will help reduce rate hike concerns, which should lower yields," while "a 'too hot' report will only further reinforce fears of more rate hikes" — and a 22V Research survey has 48% of investors expecting a mixed or negligible reaction against 33% risk-on and 19% risk-off. (2) CPI on 11 September, which Waller has now made the deciding print, and which will contain a barrel that has risen 9.9% in a week. (3) PPI on 10 September, the first read on whether the energy move is reaching the pipeline. (4) The 15-16 September FOMC itself, after five consecutive holds this year. The trade the calendar sets up is the weekend: a payroll on Friday morning, then three days in which nobody can act on it. |
| | 8 · Fed Funds Futures & Rate Path |
| | Current target range: 3.50%-3.75%. One governor's remarks took the September meeting from odds-on to a coin flip, and the entire 2027 strip richened a near-uniform four and a half basis points with it. | | CME FedWatch headline — 16 September 2026 meeting. Data as of 3 Sep 2026, 09:07:50 p.m. CT (10:07 p.m. ET), read from the FedWatch probability table. Contract ZQU6, mid price 96.3125, prior volume 50,046, prior open interest 234,140. | | Target rate (bps) | NOW | 1 DAY (2 SEP 2026) | 1 WEEK (27 AUG 2026) | 1 MONTH (3 AUG 2026) | | 350-375 (current) | 49.8% | 36.8% | 64.6% | 32.8% | | 375-400 | 50.2% | 63.2% | 35.4% | 67.2% |
| | Provenance of every column, stated — and Wednesday's live read corrected by 0.9 of a point. The footer timestamp reads 09:07:50 CT with no meridian; the wall clock at capture was 9:15 p.m. CT, roughly five hours after the 4:00 p.m. CT ZQ session close, and the printed probabilities agree to within two-tenths of a point with an Investing.com card stamped 09:55 p.m. EDT, so it resolves as p.m. and is an indicative post-close read rather than a settlement snapshot. The 1 DAY column carries the legend date 2 September and prints 63.2%, against the 62.3% this report published from CME's live column on Wednesday evening — a +0.9 pp correction, recorded here rather than left standing, and the third-tightest reproduction of the reporting window. 1 WEEK (27 August) and 1 MONTH (3 August) carry genuine reference dates and are used in the calculations below. The Investing.com matrix underneath is timestamped 3 Sep 2026, 09:55 p.m. EDT and is the primary source for parts (a), (b) and (c). | | The CME-versus-Investing.com gap, quantified, and it has essentially vanished. CME puts the September hike at 50.2% at 10:07 p.m. ET; Investing.com at 50.4% at 9:55 p.m. ET — a 0.2 percentage-point difference across twelve minutes, against 1.9 pp on Wednesday, 1.3 pp on Tuesday and 2.0 pp on Monday. The convergence is arithmetic rather than editorial: Investing.com publishes the September future at 96.313 against CME's 96.3125, a five-hundredths-of-a-basis-point difference in price. Because the 16 September meeting sits mid-month, only about 47% of the contract's averaging period is affected, so one basis point of ZQ price is worth roughly ten percentage points of hike probability. That leverage is the most important number in this section and it cuts both ways today: the whole 13-point move is about 1.3 bp of contract price, and the September contract richened only 1.3 bp from Wednesday's 96.300 while the December 2026 contract richened 4.5 bp and every 2027 contract 4.0 to 6.0 bp. | | One-day, one-week and multi-day momentum, and the vendor trap disclosed again. The September hike fell 13.0 pp on CME's own columns (63.2% to 50.2%) and 10.0 pp against the 60.4% this report published from Investing.com on Wednesday. Investing.com's own "previous day" column reads 60.1%, making its internal move -9.7 pp; that column is a fixed daily snapshot rather than the prior evening's live value, and the 0.3 pp gap to what this report published is the smallest of the reporting window — a marked improvement on Wednesday's 2.3 pp, and the discrepancy is carried in Data Notes regardless. On a one-week view the direction flips: 50.2% against 35.4% on CME and 50.4% against 34.4% on Investing.com, so the meeting is still 14.8 and 16.0 points more hawkish than a week ago even after Thursday's collapse. The one-month column is the discipline: on 3 August CME priced 67.2%, so Thursday's 50.2% is 17.0 points below where the market sat a month ago, the widest month-on-month gap of the reporting window. Further out, every 2026 horizon retreated hard: October's cumulative-above fell to 64.1% from 72.1%, December's to 82.7% from 88.2%, and the peak of the strip is 94.8% at September 2027 against 96.7% at July 2027 on Wednesday. The probability of a cut at any 2026 meeting remains 0.0%. | | (a) Current-year meeting distributions | | Investing.com Fed Rate Monitor, updated 3 Sep 2026 09:55 p.m. EDT. Format: current [prior day] [prior week]. Modal range in bold. The prior-day column is the vendor's own snapshot; see the disclosure above. | | Meeting | 3.50-3.75 (hold) | 3.75-4.00 (+25) | 4.00-4.25 (+50) | 4.25-4.50 (+75) | Cumulative above | Cumulative below | | Sep 16 | 49.6% [39.9] [65.6] | 50.4% [60.1] [34.4] | 0.0% | 0.0% | 50.4% | 0.0% | | Oct 28 | 35.9% [28.0] [47.5] | 50.2% [54.1] [43.0] | 13.9% [18.0] [9.5] | 0.0% | 64.1% | 0.0% | | Dec 9 | 17.2% [11.8] [26.3] | 42.7% [39.0] [45.0] | 32.8% [38.8] [24.4] | 7.2% [10.4] [4.2] | 82.7% | 0.0% |
| | Sums are 100.0% at all three meetings. Three observations. First, September is now the closest thing to a genuine coin flip the strip has produced: 50.4% against 49.6%, eight-tenths of a point apart, after a week that began at 34.4% and a Wednesday at 60.1%. Second, the December mode has flipped decisively, and it flipped the other way from Wednesday: one hike now leads at 42.7% against 32.8% for two hikes, a 9.9-point gap, where twenty-four hours ago the two were separated by three-tenths of a point in favour of two hikes and had been led by two hikes for three sessions. That is a genuine resolution of the contest this section flagged as the most consequential line on the page. Third, the deep tail was sold hardest: October's +50 bucket fell to 13.9% from 18.0% and December's +75 bucket to 7.2% from 10.4%, so 7.3 points of tail mass left the 2026 distribution in a single session — more than the 4.9 points that left it on Wednesday, and in the same direction, which makes this a two-day, 12-point unwind of the oil-shock tail rather than a one-day reversal. | | (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. | | Meeting | Future price | Modal range | Prob. | Cumulative above | Cumulative below | | Jan 27, 2027 | 96.035 | 3.75-4.00 | 36.6% | 86.8% | 0.0% | | Mar 17, 2027 | 95.935 | 4.00-4.25 | 35.8% | 92.0% | 0.0% | | Apr 28, 2027 | 95.885 | 4.00-4.25 | 34.5% | 93.1% | 0.0% | | Jun 9, 2027 | 95.820 | 4.00-4.25 | 32.6% | 94.6% | 0.0% | | Jul 28, 2027 | 95.800 | 4.00-4.25 | 32.0% | 94.7% | 0.0% | | Sep 15, 2027 | 95.795 | 4.00-4.25 | 31.9% | 94.8% | 0.0% | | Oct 27, 2027 | 95.790 | 4.00-4.25 | 31.7% | 93.8% | 0.2% | | Dec 8, 2027 | 95.820 | 4.00-4.25 | 30.9% | 91.2% | 0.9% |
| | The 2027 strip richened almost uniformly, and the uniformity is the information. The eight contracts print 96.035, 95.935, 95.885, 95.820, 95.800, 95.795, 95.790, 95.820 against Wednesday's 95.990, 95.890, 95.840, 95.775, 95.760, 95.735, 95.750, 95.775 — 4.0 to 6.0 basis points richer at every point, with no meaningful curvature. Compare that to the front: September richened 1.3 bp and December 2026 4.5 bp. A parallel shift in the back with a pinned front is what a timing repricing looks like when the terminal rate is unchanged — the market moved the hikes later, it did not remove them. The clearest evidence is January 2027, whose mode has flipped down a bucket to 3.75-4.00% at 36.6% from 4.00-4.25% at 39.3%; every meeting from March onward still modes at 4.00-4.25%. Cumulative-above fell at every point — December 2027 to 91.2% from 93.3%, June to 94.6% from 96.1% — and the first non-trivial cut probability rose again, to 0.9% at 3.25-3.50% in December 2027 from 0.8%, its highest of the window. | | (c) Year-end probability ladders | | Year-end 2026 — the 9 December meeting. | | Outcome | Range | Probability | | -75 bp | 2.75-3.00 | 0.0% | | -50 bp | 3.00-3.25 | 0.0% | | -25 bp | 3.25-3.50 | 0.0% | | Hold | 3.50-3.75 | 17.2% | | +25 bp | 3.75-4.00 | 42.7% | | +50 bp | 4.00-4.25 | 32.8% | | +75 bp | 4.25-4.50 | 7.2% | | +100 bp and beyond | 4.50 and higher | 0.0% |
| | Cumulative above the current range: 82.7%. Cumulative below: 0.0%. Sum: 99.9%. | | Year-end 2027 — the 8 December meeting. | | Outcome | Range | Probability | | -75 bp | 2.75-3.00 | 0.0% | | -50 bp | 3.00-3.25 | 0.0% | | -25 bp | 3.25-3.50 | 0.9% | | Hold | 3.50-3.75 | 7.7% | | +25 bp | 3.75-4.00 | 22.1% | | +50 bp | 4.00-4.25 | 30.9% | | +75 bp | 4.25-4.50 | 24.0% | | +100 bp | 4.50-4.75 | 10.9% | | +125 bp | 4.75-5.00 | 2.9% | | +150 bp | 5.00-5.25 | 0.4% | | +175 bp and beyond | 5.25 and higher | 0.0% |
| | Cumulative above the current range: 91.2%. Cumulative below: 0.9%. Sum: 99.8%. | | 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 shown as 0.0% are ranges the vendor's own card either omits entirely or publishes as zero, which under CME methodology means a probability below the rounding floor. The 2026 year-end ladder sums to 99.9% and the 2027 ladder to 99.8% for exactly this reason, as do the January, April, July and October 2027 rows at 99.9% and the March and June 2027 rows at 100.1%; only the September 2027 row and the three 2026 meeting rows sum to exactly 100.0%. | | | (a) IG and HY credit spreads | | ICE BofA option-adjusted spreads via FRED. FRED publishes with a one-business-day lag: the levels below carry the 2 September 2026 as-of date, not the 3 September close. Same-day direction is cross-checked against the cash-market proxies underneath and against Bloomberg and WSJ credit coverage. | | Series | FRED code | 2 Sep | 1-Day | 1-Week | YTD (from 2 Jan 2026) | | IG credit spread (ICE BofA US Corporate OAS) | BAMLC0A0CM | 81 bp | 0 bp | +1 bp | +2 bp (from 79) | | HY credit spread (ICE BofA US High Yield OAS) | BAMLH0A0HYM2 | 266 bp | +1 bp | -1 bp | -17 bp (from 283) | | CCC & lower credit spread | BAMLH0A3HYC | 1,053 bp | +4 bp | +22 bp | +165 bp (from 888) | | CDX IG 5y | — | Not retrievable this session | — | — | — | | CDX HY 5y | — | Not retrievable this session | — | — | — |
| | CDX — the full six-step ladder was worked again, this time with the local Chrome browser available, and is reported so the gap stays auditable. (1) Bloomberg in Chrome: /markets, /markets/rates-bonds, the Thursday and Friday Markets Wrap articles and a direct quote-page attempt were all rendered; a full-text scan of each returns zero occurrences of the string, and the quote page answers that the security "may only be available on the terminal." (2) WSJ Market Data bonds page in Chrome: rendered with its Treasury, consumer-rate and government-bond tables populated, zero occurrences in a full-text scan. (3) Cbonds carries dedicated CDX.NA.IG 5Y and CDX.NA.HY 5Y index pages but masks every level behind a subscription wall, printing a row of masking characters in place of the level with a 02/09/2026 subgroup date; S&P Dow Jones Indices and ICE publish index-family, options, tranche and methodology documentation rather than the daily running spread. (4) FT Markets Data and Reuters credit wraps returned no dated CDX quote for 3 September. (5) TradingView and Barchart symbol searches for CDX resolve to unrelated instruments, and the CME CDS index product pages carry specifications rather than levels. (6) Cash-market proxies, labelled as proxies: HYG closed $79.21, +0.13%, and LQD $105.50, +0.14%, against Wednesday's $79.11 and $105.35 — both reconciling to the cent against the prior edition. No CDX level is published here, because an undated third-party digest number is not a CDX level. | | The tail widened for a fourth consecutive update while the aggregate did nothing, and the cash market again declined to corroborate. CCC widened 4 bp to 1,053 on the 2 September stamp — after 16 and 7 in the two prior updates, so 27 bp across three sessions — taking the CCC-minus-HY differential to 787 bp from 784, a fresh extreme for the reporting window and 165 bp wider than where 2026 opened. Against that, IG was unchanged at 81 bp and HY widened 1 bp to 266, so the aggregate sits 2 bp wide of the January IG level and 17 bp through the January HY level. The corroboration question this report posed on Wednesday now has a second data point and the answer is the uncomfortable one: HYG rose 0.13% and LQD 0.14% on Thursday, a near-exact repeat of Wednesday's +0.01% and +0.12%, so two consecutive sessions of a widening lagged tail have been met by two consecutive sessions of a cash market registering essentially nothing. That is no longer an index catching up; it is a genuine divergence between where the tail is marked and where the tradeable proxies are priced. | | (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 2 September 2026 effective date. Reverse repo take-up is the same-day 3 September operation. | | Rate | 2 Sep | Change vs 1 Sep | 1st pct | 99th pct | Volume | | SOFR | 3.65% | -1 bp | 3.60% | 3.73% | $2,882bn | | EFFR | 3.63% | 0 bp | 3.60% | 3.65% | $114bn | | OBFR | 3.63% | 0 bp | 3.55% | 3.69% | $216bn | | TGCR | 3.63% | 0 bp | 3.56% | 3.67% | $1,153bn | | BGCR | 3.63% | 0 bp | 3.56% | 3.68% | $1,173bn | | SOFR - IORB | 0 bp | -1 bp | — | — | IORB 3.65% |
| | The turn is now fully cleared, and the answer to the question this report has carried for three sessions is yes. SOFR printed 3.65%, exactly at the 3.65% IORB, ending a two-day run through the administered rate and completing the normalisation that Monday's 3 bp and Tuesday's 1 bp spread began. The distribution tightened with the mean: the 99th percentile fell to 3.73% from 3.74% and volume to $2,882bn from $2,912bn, both the lowest of the post-turn window. Tri-party and broad general collateral held at 3.63% with volumes at $1,153bn and $1,173bn, and the unsecured market has not moved for four effective dates: EFFR and OBFR unchanged at 3.63%, EFFR volume steady at $114bn, OBFR volume rising to $216bn from $203bn. The one thing moving against the calm is the facility: reverse repo take-up rose to $702m on 3 September from $525m on 2 September, the first increase since the 31 August turn print of $6.726bn — a small absolute number, but the direction reversed. Reserve balances still read $2.9249tn for the week ended 26 August; the H.4.1 covering the turn was not yet reflected in the release read this session, and it remains the number to watch. | | The bill strip is where the funding story sits, and it has moved from four months to two. Section 6 records a 9 bp inversion inside a fortnight of bill maturity — the 1.5-month at 3.82%, down 5 bp, against the 2-month at 3.91%, up 2 bp, the only tenor on the whole fourteen-point curve to rise on a day the market repriced a Fed meeting by thirteen points. Both are off-table under the trimmed tenor set. No policy expectation generates a 9 bp step between two bills maturing two weeks apart; that is settlement, auction supply and dealer balance-sheet capacity. The 4-month at 3.99% now sits 4 bp above the 6-month at 3.95%, wider than Wednesday's 2 bp, so the anomaly this report flagged twenty-four hours ago has extended rather than resolved. Read together with a reverse repo facility that took in $702m and reserve balances $68bn below their 5 August peak, the picture is one of collateral rather than credit — worth watching into the 9 September buyback operation, and not yet worth trading. | | (c) Rates volatility and swap spreads | | Metric | Level | Vintage | Read | | ICE BofA MOVE | 74.68 | Delayed vendor series, 3 September stamp | -5.03 points, -6.31%, on an internally consistent day range of 74.68-79.71 with an open of 79.71 | | VIX | 14.32 | 3 September close | -5.79%, low 14.23, the lowest close of the reporting window | | MOVE / VIX | 5.22 | Both 3 September | Against roughly 5.1 on mixed vintage Wednesday | | 10y Treasury-swap spread | ≈38 bp | 25 August (Bloomberg) | No update published in the reviewed sources since | | 30y Treasury-swap spread | Narrowest since February | 25 August (Bloomberg) | Level not published by the source; the ranking is |
| | The divergence closed, and it closed from the rates side. For a fortnight this section has flagged a MOVE index rising while VIX fell — the configuration Bloomberg's Cameron Crise associated with equity downdrafts. On Thursday both fell together: MOVE -6.31% to 74.68 and VIX -5.79% to 14.32, the first session in the window in which rate volatility and equity volatility declined in step, and the MOVE level is now 5.03 points below the 79.71 that this series' own open implies for 2 September. The caveat is unchanged and is stated rather than smoothed: the series prints a "previous close" of 95.74, roughly twenty-one points outside its own day range, for a seventh consecutive session. That field is disqualifying and the change published above is computed against the series' own open of 79.71, which is internally consistent with the day range and reconciles with the 77.88 this report published for 1 September plus one unpublishable session. The swap-spread basis is unchanged in substance and in vintage: Treasuries have outperformed equivalent-maturity swaps since the 19 August buyback announcement. | | (d) Issuance, leveraged loans and private credit | | • | The September calendar is the live variable and Thursday improved its terms materially. Bloomberg puts expected U.S. corporate issuance at about $215bn this month against dealer ranges of $175-250bn, following August's record $145.2bn IG month and roughly $1.4tn of 2026 IG notes sold to date, about 9% above the 2020 pace. That supply now meets a 4.77% ten-year rather than a 4.79% one, and — more usefully for an issuer — a 4.34% two-year and a 4.52% five-year, each 5 and 2 bp cheaper to swap into. Bank of America's Meghan Swiber and Eleanor Xiao, carried forward and still the cleanest statement of the underlying problem: "Despite Treasury buybacks and other recent policy actions, investors remain reluctant to add duration. A shrinking official-sector bid leaves the market increasingly dependent on price-sensitive private demand to clear ongoing Treasury supply." |
| | • | The demand side is still being tested in concession rather than in spread. Issuers have been paying roughly 5 bp in new-issue concessions on books covered about 2x, with order-book attrition near 40% — and the IG index has now stopped widening, printing 81 bp unchanged after two consecutive widening updates. |
| | • | Private credit and the AI financing chain. Carried forward and still live: Blue Owl leading a $2.4bn debt financing for IREN to buy Nvidia chips secured against accelerators; Brookfield's $600m niche-credit payout and its agreement to subordinate CDK debt to extend maturity; and Broadcom's disclosure that it may provide residual value guarantees that are contingent liabilities to two AI laboratories, which its finance chief Amie Thuener described as "empowering two of our most strategic customers, the leading AI labs, to bridge the gap between their current cash flow and the significant upfront investments required for their businesses." New this session and structurally relevant: Nvidia's agreed $13bn acquisition of Hugging Face is being funded on balance sheet rather than through the vendor-financing channel, which is the first large AI transaction in weeks that does not add a contingent liability to the chain. |
| | • | Structural and regulatory: the SEC is preparing a plan to widen investor access to private markets (Bloomberg), and S&P Global is weighing a spin-out of Capital IQ Pro — the parent rose 4.37% on Thursday. Leveraged loans and bank CDS remain uncovered: no dated Morningstar LSTA index level or bank-CDS print was obtainable this session. Named watch items stand: Brightline's $350m Assured-backed loan arranged in case of bankruptcy; Guggenheim Investments' disclosure that affiliates may buy its hard-hit loan; and PG&E's strategic and financial review with its $2bn capital-spending deferral, undertaken explicitly to curb "the need for higher-cost borrowing" — the stock rose 4.73% on Thursday on a Truist downgrade. |
| | The credit take. The configuration is now sharper than at any point this reporting window, and it is no longer explicable as a lag. CCC at 1,053 bp is 27 bp wider across three updates and 165 bp wider on the year, while IG at 81 bp is unchanged and 2 bp wide of the 2026 open and HY at 266 bp is 17 bp through it — a 787 bp CCC-minus-HY differential, a fresh extreme. What makes it uncomfortable is the corroboration failure repeating: HYG +0.13% and LQD +0.14% on Thursday after +0.01% and +0.12% on Wednesday, so two sessions of tail widening have drawn two sessions of nothing from the tradeable market. Add the equity side: VIX at 14.32 is the lowest close of the window and MOVE fell 6.31% with it, so both volatility markets are now pricing calm alongside a widening credit tail — the precise configuration this section exists to flag. The funding leg still supports the benign reading — SOFR back level with IORB at 3.65%, general collateral steady at 3.63%, a $702m reverse repo print — so there is no plumbing catalyst near term, though the 9 bp bill inversion inside a fortnight of maturity is a small unexplained residual. What breaks the calm: Friday's payroll hot enough to put the September hike back above 65%, which widens HY faster than it tightens IG; CPI on 11 September, now the deciding print for the meeting and one that will carry a barrel up 9.9% in a week; or the $215bn September calendar clearing into a market whose concessions are already 5 bp. What confirms it: CCC back inside 1,000 bp, HYG holding through the calendar, SOFR staying at or below IORB, and reserve balances stabilising above $2.90tn when the H.4.1 covering the turn publishes. |
| | | Levels from the TradingEconomics currency board read after the U.S. close. The vendor's daily boundary had already rolled at capture — DXY, EUR, GBP, JPY, CHF, CAD, KRW, CNY, SGD, HKD, MXN and INR carry a live 10:41 clock and AUD, NZD and TWD a Sep/04 date stamp — so the vendor's own %Chg column measures only the new session and is not reproduced. Every change below is computed as the 24-hour move against the prior edition's levels for the same vendor, with worked examples in Data Notes. Quote basis: USD per unit for EUR, GBP, AUD and NZD; units per USD for JPY, CHF, CAD, KRW, TWD, CNY, MXN, SGD and INR. | | Pair | Level | 24h Chg | Context | | DXY | 98.989 | -0.59% | The largest single-day dollar decline of the reporting window. Bloomberg's Dollar Spot gauge at its lowest since May. +0.68% year to date | | USD/JPY | 155.985 | -1.73% | The yen's best session of the window, roughly double Wednesday's 0.91%. Intraday extreme 155.30; 2.74 yen below Wednesday's close | | NZD/USD | 0.58964 | +0.82% | The best major, and a full reversal of Wednesday, when it was the worst at -0.74% | | USD/CHF | 0.80772 | -0.64% | The franc's first gain in four sessions, ending a run this report flagged twice | | AUD/USD | 0.72080 | +0.53% | A second consecutive gain; +8.02% year to date, the strongest major on the board | | USD/CAD | 1.37890 | -0.41% | A second consecutive loonie gain on a crude settle that barely moved | | GBP/USD | 1.35318 | +0.37% | Ends a three-session decline, on the day 10-year gilts richened 10 bp | | EUR/USD | 1.16284 | +0.36% | Reverses two sessions of losses; -0.95% year to date | | USD/KRW | 1355.47 | -0.18% | The won barely participated, on a day the Kospi rose and the dollar fell 0.59% | | USD/TWD | 31.7110 | -0.14% | Also muted, and moving with the won rather than against it for the first time in three sessions | | USD/CNY | 6.71611 | -0.02% | Effectively unchanged, so the yuan took none of the dollar move. -3.73% year to date | | USD/MXN | 16.9018 | New row this session | -6.20% year to date, the strongest emerging-market currency on the board | | USD/SGD | 1.26668 | New row this session | -1.52% year to date | | USD/INR | 94.5380 | New row this session | +5.19% year to date, the weakest currency on the board against the dollar |
| | The take: the yen did the job twice as hard as it did on Wednesday, and this time the mechanism is two-sided. USD/JPY fell 1.73% to 155.985, roughly double Wednesday's 0.91% and a cumulative 2.74 yen in a session, with an intraday extreme of 155.30 — a level that puts the pair a long way from the 160 handle intervention watchers were treating as live three days ago. Half the move is American: the 1-year fell 5 bp, the 2-year 5 bp, and the September meeting became a coin flip. The other half is now Japanese, and it is new. Bloomberg reports traders lifting bets on Bank of Japan rate increases with Nomura saying three consecutive hikes are possible in an extreme case, and 10-year JGBs richened 10 bp to 2.89% in the following session after holding above 3% earlier in the week. A carry pair that had been moving on one leg is now moving on both, which is why a 0.59% dollar decline produced a 1.73% yen gain — a 2.9-to-1 amplification. Bloomberg's own framing of the follow-through was a "Carry Trade Exodus." | | Sterling is the cross that finally behaved, and it took a 10 bp gilt rally to do it. GBP/USD rose 0.37% to 1.35318, ending a three-session slide, on the day 10-year gilts fell 10 bp to 5.13% — twice the move in any other market on Bloomberg's board and a full reversal of the three-day cheapening to 5.23%. For three sessions this report argued sterling was trading a fiscal risk premium rather than rates, because it fell both when gilts sold off and when they stabilised. Thursday is the first observation that cuts the other way: gilts rallied hard and the currency rallied with them. One session does not settle it — the currency's gain is 0.37% against a 10 bp bond move, which is a low beta for a rate-driven currency — but it removes the cleanest evidence for the fiscal-premium reading and puts the November budget back to being a scheduled risk rather than a daily one. | | The Asian block is the second-order read, and it is the same anomaly for a fourth session. On a day the dollar posted its worst session of the window, USD/KRW fell only 0.18%, USD/TWD 0.14% and USD/CNY 0.02% — so three Asian currencies captured between a thirtieth and a third of the move that the yen captured triple. Bloomberg reports an Asian currency index reaching levels last seen in October 2024 in the following session, which is the point: the region's response came after the New York close, not during it. For the won this is now the fourth consecutive session of behaviour that portfolio flow cannot explain — it weakened on record chip exports, strengthened on a 4% equity collapse, and now declines to strengthen on a broad dollar break. The corporate-repatriation explanation this report has carried remains the only one that fits all four observations. The yuan's near-total non-response, at -0.02% on a 0.59% dollar move, is a managed-fix statement rather than a market one. | | | Settlement basis, stated, and reconciled to the prior edition. All rows are quoted from the Investing.com per-contract settle series — Thursday's settle is the contract quote's "Prev. Close" field read after the electronic session reopened, and Wednesday's is the same contract's historical row. WTI, RBOB, heating oil and natural gas are on the October contract; gold, silver and copper on December; Brent on the November contract, front since the 23 August rollover. Three of Wednesday's published settles are corrected here rather than carried forward — WTI to $91.01 from $90.63, gold to $4,414.60 from $4,433.85, silver to $65.463 from $65.950, and natural gas to $2.956 from $3.003 — each confirmed independently by Investrade's session recap; the reconciliation is in Data Notes. Weekly and year-to-date columns are TradingEconomics spot returns on the vendor's Sep/04 date stamp. | | Contract | Settle | Chg | %Chg | Week | YTD | Driver | | Silver (Comex Dec) | $67.704 | +$2.241 | +3.42% | +0.97% | -6.27%* | The best move on the board, and the first close above $67 in the window | | Gold (Comex Dec) | $4,539.90 | +$125.30 | +2.84% | +0.53% | +3.63%* | A record dollar move for the window, on a 0.59% dollar decline | | Copper (Comex Dec) | $6.6710 | +$0.0780 | +1.18% | +0.51% | +16.08%* | Joined the metals bid after refusing it on Wednesday | | RBOB gasoline (Oct) | $3.1349 | +$0.0311 | +1.00% | +3.04% | — | Reversed Wednesday's 1.12% decline | | WTI (Oct, NYMEX) | $91.30 | +$0.29 | +0.32% | +9.91% | +59.64%* | A weaker dollar against Saudi price restraint | | Brent (Nov, ICE) | $95.52 | -$0.11 | -0.12% | +8.72% | +57.40%* | The only lower energy contract; see Data Notes on the contract basis | | Natural gas (Oct, NYMEX) | $2.928 | -$0.028 | -0.95% | +1.07% | -20.81%* | Gave back Wednesday's first $3 handle in a session | | Heating oil (Oct) | $4.5987 | -$0.0835 | -1.78% | +7.83% | — | The worst move on the board, on a day crude rose |
| | *\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 took the whole move, and the leverage to the dollar is the number that matters. Gold settled $4,539.90, up $125.30 or 2.84%, and silver $67.704, up 3.42%, against a DXY decline of 0.59% — so the precious complex moved 4.8 and 5.8 times the currency. A gold move five times the dollar's is not a currency trade; it is a real-rate trade, and Thursday supplied the mechanism directly: the 2-year fell 5 bp and the September hike became a coin flip on a governor's remarks, which lowers the nominal side of the real rate without touching the inflation side — Waller explicitly said he does "not see elevated energy prices and tariffs now as a significant source of ongoing inflation pressure." The gold-silver ratio fell to 67.05 from 67.44, so silver outperformed for a second consecutive session. And copper finally joined at +1.18%, after refusing Wednesday's materials rally at +0.08% — though the equity did not follow, with Freeport-McMoRan falling 1.85% on the day its metal rose. That divergence is the cleanest reminder available that the metals bid is monetary rather than industrial. | | The energy complex broke internally again, and this time it broke the other way. WTI settled $91.30, up 29 cents, while Brent fell 11 cents to $95.52 — the only lower energy contract and the first Brent-under-WTI relative session of the window, narrowing the Brent-WTI differential to $4.22 from $4.62. Underneath it, the products inverted Wednesday exactly: heating oil fell 1.78% and RBOB rose 1.00%, against Wednesday's -0.31% and -1.12%. Two named catalysts sit behind the crude tape and they point opposite ways: Bloomberg reports Iran claiming fresh strikes on U.S. bases, and against it Saudi Arabia raising its flagship crude prices by less than expected, which is a producer telling the market it does not believe in the premium. A weaker dollar did the rest. The distillate collapse is the tradeable part and it is covered below. | | The crack spreads, on a consistent October basis against October WTI: | | • | Distillate crack: $4.5987 × 42 - $91.30 = $101.85, down $3.79 from $105.64 on the restated 2 September settle basis. |
| | • | Gasoline crack: $3.1349 × 42 - $91.30 = $40.37, up $1.02 from $39.35. |
| | • | The differential narrowed $4.81 to $61.48 from $66.29 — the largest single-session narrowing since the position was opened, and more than five times Wednesday's $0.86 widening in the other direction. |
| | Read the two sessions together and the distillate thesis has failed its own test. The trade was a closed-strait premium in middle distillate. It has now lost money outright on two consecutive sessions in which crude rose — 0.45% then 0.32% — which is precisely the invalidation condition this report wrote down on Wednesday. What has changed underneath is not sentiment but supply signalling: a post-war record of 17m barrels through Hormuz on Monday, the President's "short-lived" characterisation, and now Saudi Arabia declining to raise its official selling prices as much as the market expected. Three separate confirmations that the physical barrel is moving. On the TradingEconomics spot basis heating oil is still +115.97% year to date against gasoline's +83.70%, so the structural argument survives; the tactical one does not, and Section 12 closes the position rather than defending it. Natural gas gave back its $3 handle at $2.928, down 0.95%, one day after the first close above three dollars in the window — a round trip nobody was watching either way. | | | 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 ZQU6/ZQZ6 spread was stopped out; the replacement is a 2026-versus-2027 spread | | Mark first, and it is a stop. Long ZQU6 (September 2026) against short ZQZ6 (December 2026), DV01-matched one-for-one at $41.67 per basis point per contract, entered twelve sessions ago at 96.325 / 96.160 for a spread of 16.5 bp, trimmed to a quarter, stop raised to 24.5 bp on Tuesday. Thursday's mark: ZQU6 96.313, ZQZ6 96.085 — a spread of 22.8 bp. That is -3.2 bp on the day and through the 24.5 bp stop, so the position is closed at the stop for +8.0 bp from entry, worth +$333.33 per contract pair before costs, on a quarter. | | Why it stopped, and the mechanism was written down in advance. The prior edition named the loss scenario precisely: "this spread can move 3 bp in an hour — against the position, because December would rally more than September." It moved 3.2 bp, and December did rally more — ZQU6 richened 1.3 bp against ZQZ6's 4.5 bp. The forecast was right about the mechanism and wrong about the trigger: the tail was bought back by Waller rather than by a payroll, a day early. December's +75 bucket fell to 7.2% from 10.4% and its cumulative-above to 82.7% from 88.2%, while September's contract was pinned because a coin-flip meeting has nowhere left to go. Recorded as a loss of 1.5 bp against Wednesday's mark and a completed trade at +8.0 bp from entry. | | The replacement — long ZQZ6 against short ZQZ7, DV01-matched one-for-one, quarter size. Entry: 96.085 / 95.820, a spread of 26.5 bp. Thesis: the 2027 strip moved a near-uniform 4.0 to 6.0 bp at every one of eight meetings while the modal range stayed 4.00-4.25% from March onward — the market moved the hikes later, it did not remove them, so the terminal rate is anchored and the 2026 path is not. December 2026 still carries 82.7% cumulative-above and a 32.8% two-hike bucket after a session that flipped its mode to one hike; if the hold Waller described happens, that is where the remaining probability has to leave from. The modal path, base case and tails. Modal path: a 25 bp hike on 16 September is now a coin flip at 50.4% on Investing.com and 50.2% on CME, with ease at 0.0%; one hike is modal at October (50.2%) and at December (42.7%); the 2027 strip modes at 4.00-4.25% at seven of eight meetings. Base case: one hike into year-end 2026 with the timing pushed from September to October or December, and a terminal rate that has not moved. Tail one, and it is Friday: payrolls at +55,000 consensus against a July print that fell 23,000 — a hot number puts September back above 65% and the spread narrows against the position. Tail two, and it is the bigger one: CPI on 11 September, which Waller has now made the deciding print; a hot August inflation number reverses the whole of Thursday and then some. Practical implication: the leverage is unchanged at roughly ten percentage points of probability per basis point of ZQ price, so this spread is a low-delta way to own the 2026 timing question without owning the terminal rate. Catalyst: payrolls 9/4 08:30; the buyback operation 9 September; PPI 9/10; CPI 9/11; the 15-16 September FOMC. Invalidation: the spread through 22.0 bp; or the September cumulative hike back above 65% on either vendor; or the December 2026 mode reverting to two hikes. Sizing: a quarter, at $41.67 per basis point per pair. | | 2. Long the October distillate crack against short the October gasoline crack — invalidated on its own criterion; CLOSED | | Mark, on the restated settle basis. The two legs stand at $101.85 and $40.37 for a differential of $61.48, against $105.64 and $39.35 for $66.29 on Wednesday's corrected settles — a session loss of $4.81 on the relative trade and $3.79 on the distillate leg outright. The honest reading, and there is nothing to defend. Wednesday's invalidation was written as "a second consecutive session in which the distillate leg loses money outright while crude rises." That is exactly what happened: crude settled +0.32% and heating oil -1.78%. Three separate physical signals now argue against the thesis — a post-war record 17m barrels through Hormuz, the administration's "short-lived" framing, and Saudi Arabia raising official selling prices by less than expected. Action: close the position at $61.48, banking +$5.48 against the $56 entry level rather than waiting for the $60 stop to be touched. The structural argument — heating oil +115.97% year to date against gasoline's +83.70% — is intact and can be re-expressed later on a fresh catalyst; the tactical one is finished. Recorded as closed on thesis invalidation, not on a stop. | | 3. Long the 20-year against the 30-year, on the November refunding — hold the half | | Expression: long the 20-year bond against short the 30-year, DV01-neutral, half size. Mark: 20s30s at 0 bp with both tenors at 5.25% — unchanged on the day for a seventh consecutive session, and 1 bp flatter on the week. What changed: both tenors richened 2 bp together on a day the front end richened 5, so the long end participated least in a rally it had also led least on the way down. That is the buyback floor holding through a directional move in the other direction, which is a stronger test than Wednesday's flat tape. Against it sits a $215bn September corporate calendar and an operation on 9 September whose size Treasury has described only as "at least double." The dislocation this trade waits for still needs the calendar to arrive. 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. | | 4. Protection on the CCC cohort funded in IG — a third consecutive gain; hold the half | | Expression: long CCC-exposed credit protection (or short a levered-loan / CCC-heavy vehicle) against long IG cash. Mark: CCC 1,053 bp, +4 bp; HY 266 bp, +1 bp; IG 81 bp, unchanged on the 2 September FRED update, taking the CCC-minus-HY differential to 787 bp from 784 — a 3 bp gain, on top of 5 and 13 in the two prior updates for a cumulative 21 bp. The honest reading: three consecutive paying updates is a good run and the reason not to press it is unchanged and now doubled. HYG closed +0.13% and LQD +0.14%, a near-exact repeat of Wednesday's +0.01% and +0.12%, so the same-day cash market has now declined to confirm a widening tail twice in a row. A position paid only by a one-day-lagged series is a position to hold. Action: hold the half; do not add. Catalyst: payrolls 9/4; the September IG calendar clearing into a $215bn month; CPI 9/11. Invalidation: the differential back through 750 bp, or IG widening beyond 90 bp, which would mean the aggregate has joined the tail and the pair no longer isolates anything. Sizing: a half, unchanged. | | 5. Short the debasement complex against long the dollar — both invalidation triggers fired; CLOSED | | Expression: short an equal-weighted basket of Comex gold, Comex silver and a bitcoin proxy against long the dollar index, dollar-notional matched. Mark: the short basket returned gold +2.84%, silver +3.42%, bitcoin +3.70% for an average of +3.32%, against DXY -0.59% — a 3.91-point loss on the day on a token position, the second consecutive session in which both legs lost. The honest reading: the invalidation was written in two parts and both fired on the same session. Gold was to be closed on "reclaiming the 31 August settle of $4,481.50" — it settled $4,539.90, $58 through it. The dollar leg was to be closed on "DXY back below 99.11" — it closed 98.989. There is no discretion left. Action: close the position. Cumulative outcome across its life is recorded rather than quietly dropped: the trade paid while the front end was cheapening, and it lost every session after the front end began richening on 2 September. The lesson to carry: a short-debasement trade is a short-real-rate-decline trade wearing a currency costume, and the currency leg is the smaller half — the metals moved 4.8 and 5.8 times the dollar on Thursday. | | 6. Long volatility into the consumer prints — the Lululemon leg paid in full; take it, keep the semiconductor leg | | Mark. The straddle leg was bought on Lululemon into Thursday's close at $121.77, up 1.42% on the session. The stock then fell as much as 20% after hours on a second consecutive full-year guidance cut — revenue to $10.35-10.5bn, a 5-7% decline, and adjusted earnings to $9.48-9.73, on second-quarter revenue down 4% with comparable sales down 9%. The invalidation was "Lululemon beating and rallying more than 5%"; the opposite happened by a factor of four. Action: take the Lululemon leg off in the pre-market rather than holding the gamma through Friday's payroll, where the remaining premium is event risk you are no longer paid for. Keep the October volatility on the semiconductor proxy. Thursday strengthened that leg rather than weakening it: SOX rose 0.11% against a 1.16% Nasdaq 100, and Broadcom fell 2.74% on the day Macquarie upgraded it to outperform at $490, about 37% above the close. A complex that will not rally on a 13-point collapse in hike odds and will not rally on a 37%-upside upgrade is a complex with a dispersion problem, and dispersion is what long vol owns. Catalyst: Oracle 8 September (Section 5), the next AI capital-expenditure read; payrolls 9/4; CPI 9/11. Invalidation: SOX outperforming the Nasdaq 100 on two consecutive up sessions, which would say the non-participation was positioning rather than doubt. Sizing: a quarter, expressed in premium rather than in delta. | | 7. New: short the packaged-food cohort against long consumer defensive — a dividend cut is a re-rating trigger, not a single-name event | | Thesis. Campbell's cut its dividend 36%, announced $500m of savings and reported a 37% earnings decline with a sales miss — and the read-across was immediate and indiscriminate: Tyson -7.26%, Campbell's -6.98%, Lamb Weston -3.76%, Conagra -3.46%, General Mills -3.25%, Kraft Heinz -3.20%, McCormick -2.27%, JM Smucker -2.23%, Bunge -2.23%. Nine names down an average of 3.6% on a session the S&P rose 1.06% is a 4.6-point relative move confined to one sub-industry, on a day nothing macro touched food. Meanwhile consumer defensive as a whole closed +0.17%, because Walmart rose 2.20% — so the group index conceals the break entirely. Expression: short an equal-weighted basket of the packaged-food names above against long the consumer-defensive sector, dollar-neutral, half size. The pair isolates the sub-industry from the staples beta and from the rate move. Why now rather than after the next print: a dividend cut at a name with Campbell's payout history changes the discount-rate assumption for every levered food balance sheet on the same day, and the market has just demonstrated it will mark the whole cohort within one session. Catalyst: Kroger on 11 September (Section 5) as the channel read; CPI on 11 September for food-at-home; any second payout cut in the cohort. Invalidation: the basket recovering more than half of Thursday's underperformance within three sessions, which would mark it a one-day sympathy move; or a takeover approach for any basket constituent, which is the obvious asymmetric risk in a cohort trading at these multiples. Sizing: a half, dollar-neutral. | | Prior closes, marked forward. The AI-halo basket against long Nvidia, closed on 1 September, would have lost ground on Thursday for the first time since it was closed: Palo Alto +1.05%, ServiceNow +6.49%, Fortinet +1.18%, Adobe +2.13% and Synopsys unquoted, for a four-name basket average of +2.71%, against Nvidia +1.80% — a 0.91-point loss on the pair, against Wednesday's 7.22-point gain. Cumulatively the closed pair has still cost more than it saved and the record stands. The gasoline-versus-distillate crack pair, closed nine sessions ago, would have gained $4.81 on Thursday, its largest favourable session yet and the mirror image of the position Section 12 item 2 has just closed — a reminder that this book has now been on both sides of the same spread and made money on neither. The long-silver-against-short-gold pair, closed on thesis invalidation, would have gained 0.58 points: silver +3.42% against gold +2.84%, its best session since being closed. | | The vol note. VIX closed 14.32, down 5.79%, with a low of 14.23, on a session the index gained 1.06% — a 5.5-to-1 ratio of volatility decline to index gain, far tamer than Wednesday's 15-to-1 and the sign of a market that bought the rally rather than merely stopped hedging. Three sessions have taken the index -0.71%, +0.46%, +1.06% and VIX 16.34 → 15.20 → 14.32, so the surface has now fully unwound a spike and then some, closing at the lowest level of the reporting window. A 14.32 handle asks for roughly a 0.90% daily move against a market that has just delivered 1.06%, into a payroll and a three-day weekend — so the near-dated surface has moved from cheap on Wednesday to cheaper, and the reason to own it has strengthened rather than weakened. Rates volatility confirms rather than contradicts for the first time in a fortnight: MOVE fell 6.31% to 74.68 on an internally consistent day range, so the divergence this book was hedging has closed from the rates side. That changes the instruction: the hedge is no longer a cross-market divergence trade, it is a straightforward event-gamma trade. Keep the index gamma dated across 4 September, keep the 10-11 September inflation block — which is now the deciding print rather than a secondary one, following Waller — and keep financing both in the 18-19 September meeting expiry, where a coin-flip meeting is still the richest premium on the curve. | | | Crowded consensuses to stress-test, with the numbers. | | 1. | "The Fed hikes in September." The strip says 50.2% on CME and 50.4% on Investing.com, down from a settled 63.2% and 60.1%. Stress test: the meeting lost 13.0 points in one afternoon on remarks from a single governor who is not the chair, with Kevin Warsh's hawkish stance the reason the market was positioned the other way in the first place. The 3 August column reads 67.2%, so the market is 17.0 points below where it sat a month ago. At roughly ten percentage points of probability per basis point of ZQ price, a 1.3 bp move in the September contract did all of this — and can undo it. The remaining risk is now CPI on 11 September, not Friday's payroll, because Waller said his decision would be "heavily influenced by what we learn about August inflation." |
| | 2. | "The terminal rate is falling." It is not, and the strip is unusually clear about it. Every 2027 contract richened 4.0 to 6.0 bp with almost no curvature, and the modal range stayed 4.00-4.25% at seven of eight meetings — only January flipped down a bucket, to 3.75-4.00% at 36.6%. Stress test: a parallel richening with an unchanged mode is the market moving hikes later, not removing them; cumulative-above at December 2027 is still 91.2% and the cut probability is 0.9%. Anyone reading Thursday as a dovish repricing of the destination is reading a repricing of the schedule. |
| | 3. | "The long end is the problem." Still not, and the evidence keeps accumulating. 20s30s printed exactly 0 bp for a seventh consecutive session with both tenors at 5.25%, and the long end richened only 2 bp on a day the 2-year richened 5 — so it participated least in the rally exactly as it had cheapened least in the selloff. Stress test: that symmetry is what a buyback floor looks like, and it is about to be tested by a $215bn September corporate calendar and an operation on 9 September of unspecified size. Ben Emons of FedWatch Advisors remains the counterweight: the market prices roughly 60 bp of hikes through year-end while the combined move in forwards and long-term spot yields "implies a rate-hike path of roughly 120 basis points." |
| | 4. | "Credit is fine because credit is tight." IG at 81 bp is unchanged and 2 bp wide of the 2026 open and HY at 266 bp is 17 bp through January — but CCC has widened 27 bp across three updates to 1,053, taking the CCC-minus-HY differential to 787 bp, a window extreme. Stress test: the same-day cash market has now registered nothing twice in a row — HYG +0.13% and LQD +0.14% after +0.01% and +0.12% — so the tail is widening in a lagged index the tradeable market will not confirm, on two consecutive observations rather than one. And the newest AI credit exposure remains off the indices entirely: Broadcom's contingent residual-value guarantees to two AI laboratories are vendor financing that no spread series captures. |
| | 5. | "Volatility is right to be calm." VIX at 14.32 is the lowest close of the reporting window, asking for a 0.90% daily move into a payroll, a three-day weekend, PPI and the CPI print that now decides the meeting. Stress test: MOVE fell 6.31% to 74.68 alongside it, so for the first time in a fortnight both volatility markets agree — which removes the divergence signal and replaces it with a simpler problem, that the surface is cheapest at precisely the moment the calendar is heaviest. The MOVE series has now published a "previous close" outside its own day range for seven consecutive sessions, so the cross-market comparison rests on the series' open rather than on its own close field. |
| | The two-sided geopolitical tape. Escalation returned to the headlines even as the barrel ignored it: Bloomberg reports Iran claiming fresh strikes on U.S. bases, and the crude complex answered with a 32-cent WTI gain and an 11-cent Brent decline, because Saudi Arabia raised its official selling prices by less than expected — a producer signalling it does not believe in the risk premium. Against that, the de-escalation evidence from Wednesday stands: 17m barrels through Hormuz on Monday, a post-war record, with regional exports above pre-war levels once the Saudi and Emirati bypass pipelines are counted, and the President's "short-lived" characterisation. Canada's retaliation on $20bn of U.S. goods lands 8 September, four days from now and the day Oracle reports. Domestically the balance-sheet risk continues to move faster than the political one: PG&E rose 4.73% on the day Truist cut it to hold, one session after a strategic and financial review and a $2bn capital-spending deferral. | | Structural watch items. SOFR returned to exactly the 3.65% IORB, ending the through-IORB run, while reverse repo take-up rose to $702m from $525m — a small number moving the wrong way; reserve balances at $2.9249tn with the H.4.1 covering the turn still to appear in the release read this session; a 9 bp bill inversion inside a fortnight of maturity between the 1.5-month at 3.82% and the 2-month at 3.91%, with the 4-month now 4 bp above the 6-month against 2 bp on Wednesday; $215bn of expected September corporate supply into a 4.77% ten-year; Japan's 10-year 10 bp lower at 2.89% with the yen 1.73% stronger at 155.985 and Nomura telling Bloomberg three consecutive Bank of Japan hikes are possible in an extreme case, with the decision on 18 September; 10-year gilts 10 bp richer at 5.13%, the largest move on Bloomberg's global board; OpenAI's GPT-6 release and Nvidia's agreed $13bn purchase of Hugging Face; and Adobe's new chief executive Anil Chakravarthy taking over five days before the company reports. | | What VIX is and is not pricing. At 14.32 — the lowest close of the reporting window — VIX is pricing roughly a 0.90% daily move into August payrolls Friday at 08:30, then Labor Day, then PPI on the 10th and CPI on 11 September, the print a Fed governor has just made the deciding input for a meeting priced at a coin flip. It is not pricing the fact that the decision function changed hands on Thursday: two data beats — ISM services 55.4 against 54.2 and claims 206,000 against 205,000 — were overwritten inside an afternoon by remarks from one governor, which means the reaction function is now personnel rather than data, and personnel is not something an index option knows how to price. It is not pricing a semiconductor complex that refused a 1.16% Nasdaq 100 rally at +0.11% and sold Broadcom 2.74% through a 37%-upside upgrade. It is not pricing a credit tail 27 bp wider across three updates against a cash proxy that has moved a combined 0.14% in two sessions. It is not pricing nine packaged-food names down an average of 3.6% on a +1.06% index day. And it is not pricing the calendar shape, which is the same trap as Wednesday and one day closer: a payroll on Friday morning followed by three days in which nobody can trade the answer, into an inflation print that now carries the meeting. The expression follows and it has simplified: own the 4 September gamma, own the 10-11 September inflation block outright rather than as a hedge, and fund both in the 18-19 September meeting expiry, where a 50/50 meeting keeps the premium rich. |
| | Sources used this session: Investing.com (major-indices board, Nasdaq 100, PHLX Semiconductor, S&P 500 component board, Fed Rate Monitor, per-contract commodity quote and historical boards, economic calendar, bitcoin), Finviz group screener in Performance table view, U.S. Department of the Treasury Daily Treasury Par Yield Curve Text View, CME FedWatch, Bloomberg.com (Markets, Rates & Bonds, Markets Wrap), WSJ Market Data (Bonds & Rates, U.S. economic calendar), TradingEconomics (commodities and currency boards), FRED (ICE BofA OAS series), the New York Fed reference-rates and reverse-repo APIs, the Federal Reserve H.4.1, the Nasdaq earnings calendar API, CNBC, Reuters, Yahoo Finance, Investrade, stockanalysis.com and The Motley Fool. | | Full Data Notes & Conflicts, the Overnight / Asia & Europe read-through and the complete categorised Source Links are in the companion files US_CrossAsset_Daily_2026-09-03_DataNotes.txt and US_CrossAsset_Daily_2026-09-03.md. |
|