|
U.S. Stock, Fixed Income & Cross-Asset Closing Daily
Monday, August 24, 2026 — Full Market Close Report | Data as of: ~4:00 p.m. ET close (Fed-probability cards timestamped 25 Aug 2026 12:35 a.m. EDT)
Prepared for Institutional Investors. Not Personalized Investment Advice; Verify Independently before Acting.
|
| | The tape in one paragraph. The long end finally got what the Treasury has been promising it, and the equity index fell anyway — on a day when three stocks rose for every two that fell. Treasury Secretary Scott Bessent told CNBC the buyback operations "could be even larger" than the new $4bn minimum, and CNBC's Steve Liesman reported that two senior Treasury officials are considering funding them out of the near-$1tn General Account — Bessent's own phrase for it is a "Treasury Twist." The official par curve bull-flattened on that: the 30-year fell 4 bp to 5.23%, the 20-year 4 bp to 5.21%, the 10-year 4 bp to 4.70%, while the 2-year did not move at all and the 1-month bill fell a basis point. Yet equities could not use it. The S&P 500 fell 0.28% to 7,652.86 even though 306 of its members rose against 185 that fell — a 1.65-to-1 advance line under a red index, which is the cap-weighted divergence in its purest form. The Dow rose 140.15 points, or 0.26%, to 53,417.16, its first back-to-back gain since 5 August, while the Nasdaq Composite fell 0.76% to 25,980.19 and the Nasdaq 100 0.97% to 29,023.18. The damage was one industry. SOX fell 2.70% to 11,423.2, a three-week low, and it did so on a report that should have been bullish for half of it: Bloomberg said over the weekend that Nvidia has notified customers of price increases above 15% on Vera Rubin and Grace Blackwell servers because memory costs are soaring. The memory complex sold the news — Seagate −6.51%, SanDisk −6.45%, Micron −5.83% to $910.43, Western Digital −5.24% — and Nvidia fell 2.91% to $208.48 on 135.19m shares, its seventh consecutive decline, the first such run since September 2022, two sessions before it reports. VIX rose 4.76% to 15.85. On macro, no release rated "Very high" landed in the past twelve hours — Monday's calendar carried only the New York Fed SCE Labor Market Survey at 11:00, rated Low — and none is due in the next twenty-four hours; Tuesday's heaviest item is Consumer Confidence at 10:00, rated Medium, and the first Very-high print is the PCE deflator on Wednesday at 8:30 (§7). The second-order tells are unusually clean. The dollar rose on the session the long end rallied — DXY +0.16% to 98.99 — and gold rose with it, +0.78% to $4,697.80, which is not a discount-rate trade. The Fed strip went the other way from the long bond: the September hike moved from 39.0% to 41.4% on Investing.com and 39.9% to 41.9% on CME, so the curve flattened from both ends at once (§8). Utilities rose 0.81% and real estate 0.54%, precisely inverting Friday, when they were the only two groups to fall. And in the plumbing, the Twist showed up before the bills did: SOFR rose 2 bp to 3.65%, exactly level with IORB, with tri-party and broad general collateral both 3 bp higher at 3.63% (§9). Finally, the man who taught Bessent the trade wrote against it. Stanley Druckenmiller, in a Wall Street Journal op-ed, called the buyback plan a mistake: "Governments defending prices against fundamentals always lose." |
| | Index | Close | Chg | %Chg | Note |
|---|
| S&P 500 | 7,652.86 | −21.51 | −0.28% | Range 7,638.17–7,670.30. 306 advancers vs 185 decliners on Investing.com's component board — the index fell on a 1.65-to-1 up market. 1.87% below the 13 August closing record of 7,798.99 | | Nasdaq Composite | 25,980.19 | −200.27 | −0.76% | Range 25,910.82–26,109.96. Still on track for its first positive month in three (CNBC) | | Dow Jones Industrials | 53,417.16 | +140.15 | +0.26% | Range 53,261.95–53,508.18. First back-to-back gain since 5 August; on track for a fifth straight positive month | | Nasdaq 100 | 29,023.18 | −285.68 | −0.97% | Range 28,875.97–29,151.97. Underperformed the S&P by 69 bp | | Russell 2000 | 2,995.08 | −22.79 | −0.76% | Range 2,991.24–3,015.06. Back below 3,000. Investing.com's baseline is 3,017.87 against the 3,016.21 in Friday's edition; on the latter the move is −21.13, −0.70% (Data Notes) | | VIX | 15.85 | +0.72 | +4.76% | Range 15.61–16.06. Gave back 82% of Friday's 5.50% decline | | PHLX Semiconductor (SOX) | 11,423.2 | −317.2 | −2.70% | Opened at the session high of 11,513.9, low 11,263.3. Lowest in three weeks; CNBC's tally for the same index is −2.64% | | UST 2Y (official par) | 4.24% | 0 bp | — | Did not move, on a day the 30-year fell 4 bp | | UST 10Y (official par) | 4.70% | −4 bp | — | Bloomberg's evening board marked 4.71% | | UST 20Y (official par) | 5.21% | −4 bp | — | −9 bp on the week — the best-performing tenor on the curve, and the one Treasury named | | UST 30Y (official par) | 5.23% | −4 bp | — | CNBC quoted 5.234% intraday; −8 bp on the week | | WTI (Oct, NYMEX front) | $85.01 | −$1.63 | −1.88% | Settle from the CNBC @CL.1 board. Fell on the day the Iran sanctions actually landed | | Brent (Oct, ICE front) | $92.17 | −$1.76 | −1.87% | Brent–WTI spread $7.16, in 13 cents | | Gold (Comex Dec) | $4,697.80 | +$36.20 | +0.78% | Highest close since mid-May; CNBC put the intraday high at 4,728.3, the best since 13 May | | Silver (Comex Sep) | $68.594 | −$0.416 | −0.60% | The precious complex split: gold up, silver and platinum down | | Copper (Comex Sep) | $6.6050 | +$0.0250 | +0.38% | A second consecutive gain; Freeport-McMoRan hit an intraday all-time high | | DXY | 98.99 | +0.16 | +0.16% | Investing.com's dollar-index board. Rose on a bull-flattening curve |
| | - The buyback trade worked, and it worked only where Treasury pointed. Bessent told CNBC the operations could exceed the new $4bn minimum, and CNBC reported that the near-$1tn General Account may fund them — the "Treasury Twist." The curve answered precisely: 30-year −4 bp to 5.23%, 20-year −4 bp to 5.21%, 10-year −4 bp to 4.70%, 5-year and 7-year −2 bp, 2-year and 3-year unchanged, 1-month bill −1 bp. That is a clean bull flattener with the pivot inside two years, and on the week it is emphatic: the 20-year is 9 bp richer and the 30-year 8 bp, against a 2-year 5 bp cheaper and a 3-year 6 bp cheaper. 2s30s flattened 13 bp on the week to 99, breaking below 100 for the first time in this reporting window. The announcement effect this report has tracked since Wednesday has now converted from relative value into level. Forward catalyst: the operation schedule and the first execution on 9 September; the funding mechanics, because bills-for-bonds is showing up in repo already (§9).
| - The memory complex sold a memory-cost story, and that is the day's most interesting mispricing. Bloomberg reported over the weekend that Nvidia has told customers servers with Vera Rubin and Grace Blackwell chips will cost more than 15% more on early-2027 shipments, because memory prices have surged. The direct read is that DRAM and NAND suppliers have pricing power. The tape did the opposite: Seagate −6.51% to $794.65, SanDisk −6.45% to $1,493.12, Micron −5.83% to $910.43 on 30.01m shares, Western Digital −5.24% to $435.38. Wedbush's Dan Ives called the increase bullish for the group. Four of the five worst Nasdaq 100 performers disagreed on the same morning. Either the market is pricing demand destruction at the buyer — a 15% server price rise is a tax on the hyperscaler capex line — or a cohort that has run for months took a confirmed cost story as the exit. Forward catalyst: Nvidia, Wednesday 26 August after the close (§5), which is the only event that can arbitrate.
| - Breadth and the index disagreed by more than at any point this month. 306 S&P 500 members rose and 185 fell — and the index closed down 0.28%. Six of eleven S&P sectors were higher, consumer staples leading at +1.4%, with information technology −1.3% (CNBC). The leadership list is a rate-sensitive and defensive roll-call: Church & Dwight +3.78%, Altria +3.60%, American Water Works +3.34%, Edison International +3.32%, BXP +3.43%, PG&E +2.90%, Costco +2.50%, Mastercard +3.31%, Visa +3.06% to $382.41 at a 52-week high. Against them, four semiconductor and storage names each took more index points out than the entire staples complex put in. When the equal-weighted market rises and the cap-weighted market falls on the same session, the correct statement is not "stocks fell" — it is that one industry is now large enough to be the index. Forward catalyst: Nvidia Wednesday; any broadening confirmation from the PCE deflator.
| - The dollar and gold rose together on a bull-flattening curve. DXY closed at 98.99, +0.16%, and gold settled at $4,697.80, +0.78%, its best close since mid-May, having traded as high as 4,728.3 intraday (CNBC). A long-end rally that lowers the discount rate should, textbook, weaken the dollar and help gold; a stronger dollar should hurt it. Getting both at once means neither is being priced off the real rate. The Bloomberg framing on the same evening was "Bitcoin Tops $80,000 on Debasement Trade, Sets Three-Month High," and the coin did exactly that after the U.S. close. What ties the three together is the credibility question the buyback has opened, not the level of yields. Forward catalyst: the PCE deflator; Chair Warsh at Jackson Hole on Friday.
| - Druckenmiller wrote against his own protégé, in public, on the day the trade paid. Stanley Druckenmiller — who employed Bessent at Duquesne — published a Wall Street Journal op-ed calling the buyback programme a mistake, arguing that "Governments defending prices against fundamentals always lose" and that "Every basis point of artificial yield suppression is a subsidy to procrastination." Mohamed El-Erian made the institutional version of the argument on CNBC the same morning, noting that intervention "makes sense when you can identify either a market failure or an institutional trouble" and that neither applies. Two of the most-followed macro voices in the market attacked the mechanism on the session it delivered a 4 bp long-end rally. That is a setup, not a verdict: the trade is now consensus-contested rather than consensus-owned. Forward catalyst: the 9 September operation, where execution replaces rhetoric.
| - The Canada trade breakdown repriced the transport complex and left the metal alone. Trump posted that from 1 January 2027 tariffs on all Canadian cars, trucks, automotive parts and steel go to 50%, after talks collapsed; Prime Minister Carney's retaliation starts 8 September. Truckers took it: J.B. Hunt −5.65% to $259.88, Old Dominion −2.42% to $201.48, Knight-Swift more than 3% lower intraday. BorgWarner fell 5.05% and Ford was among the worst S&P names. But the supposed beneficiaries barely moved by the close — Nucor +0.41% to $244.64 and Steel Dynamics +0.26% to $229.27, after both traded up more than 1.5% early and the materials ETF (XLB) printed its first intraday all-time high since 12 February before turning red. BofA's Carlos Capistran put the macro damage at "modest" with tariffs touching just over 5% of Canadian exports, and named the real channel: "Renewed uncertainty is likely to weigh on investment, and potentially hiring, more than the tariffs themselves." Forward catalyst: 8 September retaliation; USMCA renewal, which UBS now sees slipping into 2027.
| - Oil fell on the day the sanctions actually arrived. Bessent unveiled "Operation Economic Outcast," which he had trailed on X as "an economic D-Day — the single greatest financial offensive ever marshaled against an adversary," with countries retaining ties to Tehran to be "removed from the US dollar system" and no exemption signalled for China. WTI settled at $85.01, −1.88%, and Brent at $92.17, −1.87%. This is the third distinct session in a fortnight where an Iran escalation has been sold, and this one had the most content. The product tape says where the risk actually sits: heating oil −4.37% to $4.2677, more than twice crude's decline, while RBOB fell only 1.54%. On a per-barrel basis the distillate crack collapsed $6.56 to $94.23 while the gasoline crack gave up just 52 cents to $52.36. Distillate is the barrel most exposed to a sanctioned-flow disruption, and it is the one that unwound. Forward catalyst: enforcement actions on third countries, particularly Chinese refiners.
| - The front end will not follow the long end down, and the Fed strip moved the other way. While the 30-year rallied 4 bp, the 2-year and 3-year did not move a basis point and the September hike probability rose — 41.4% from 39.0% on Investing.com and 41.9% from 39.9% on CME. Cumulative hike by October is now 57.6% against 53.1% a session earlier, and by December 74.9% against 71.2%, with the probability of a cut at any 2026 meeting still 0.0%. A curve that flattens because the long end rallies and the policy path is priced higher is not a growth-scare flattener; it is a market accepting an intervention in the term premium while refusing to change its view of the funds rate. Forward catalyst: PCE 8/26 08:30; Warsh at Jackson Hole 8/28.
| - The funding tape cracked, and it cracked in the place the Twist would hit first. SOFR rose 2 bp to 3.65% for the 21 August effective date, exactly level with the 3.65% IORB after two sessions through it, with the 75th percentile at 3.70% from 3.68% and the 99th at 3.73%. Tri-party and broad general collateral each rose 3 bp to 3.63%, having been pinned at 3.60% for the prior three sessions, on roughly $1.2tn a day each. The overnight reverse repo facility took $380m, up from $200m. Selling bills to buy bonds is, mechanically, a drain of collateral-free cash into collateral — and this is what that looks like on day two of the idea, before a single operation has been run. Forward catalyst: month-end on 31 August; the 9 September operation and the bill auction sizes that fund it.
| - America rallied alone, after three sessions of cheapening alone. On Bloomberg's board the U.S. 10-year fell 4 bp on the official par curve while Germany, the United Kingdom, France, Spain, the Netherlands, Portugal and Greece were all unchanged and Italy and Switzerland moved less than a basis point. For three consecutive sessions this report flagged an American move of 3–5 bp against a flat bund as evidence of an American-specific duration problem. Monday is the same fact with the sign reversed, which strengthens rather than weakens the reading: the problem is being priced as domestic in both directions, and European duration is simply not participating in either leg. Forward catalyst: the euro-area flash HICP at the start of next week, and whether a U.S. PCE surprise transmits at all.
| | | Sector | 1-Day | 1-Week | YTD |
|---|
| Consumer Defensive | +1.68% | +1.76% | +9.12% | | Communication Services | +0.92% | +1.03% | −0.90% | | Financial | +0.84% | +0.02% | +8.29% | | Utilities | +0.81% | −2.27% | +0.13% | | Real Estate | +0.54% | +1.08% | +11.62% | | Basic Materials | +0.18% | +6.13% | +23.10% | | Consumer Cyclical | −0.14% | +0.92% | −2.26% | | Healthcare | −0.16% | +4.06% | +12.19% | | Energy | −0.86% | +0.74% | +37.64% | | Industrials | −1.00% | −5.04% | +11.81% | | Technology | −1.71% | −4.86% | +20.23% |
| 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 — CNBC's GICS tally put consumer staples at +1.4% and information technology at −1.3%. Communication services prints +0.86% in the source table and is shown here at that value in the 1-Day column. | | YTD reconciliation. Compounding each group's 21 August YTD by Monday's one-day move reproduces the published YTD to within 0.031 percentage points across all eleven groups. Worked examples: technology 1.2233 × 0.9829 = 1.2024 → +20.24% against a published +20.23%, deviation 0.008 pp; consumer defensive 1.0731 × 1.0168 = 1.0911 → +9.11%, published +9.12%, deviation 0.007 pp; energy reconciles to 0.004 pp. The largest deviation is real estate at 0.031 pp, the same group that drifted twice last week, which suggests a membership effect rather than a data error. (The 1-Day column above rounds communication services to +0.92% in one vendor read and +0.86% in another; the reconciliation uses +0.86%.) | | Read the table as two tables. Nine of eleven groups moved less than 1% and the two that did not are technology and industrials. Technology at −1.71% carries the entire index decline: Micron, Seagate, Western Digital, SanDisk, AMD (−3.49%), Broadcom (−2.63%), Marvell (−3.27%), Oracle (−2.74%) and Nvidia all sit in that Finviz bucket, and Nvidia alone traded 135.19m shares. Industrials at −1.00% is the Canada tariff, concentrated in freight. Everything else went the other way, and it went the way the bond market told it to: utilities +0.81% and real estate +0.54% are the two highest-duration equity sectors and they are the two that inverted Friday's move exactly, when they were −1.97% and −0.05% on a rising index. That symmetry is the cleanest evidence available that the equity market is, at the sector level, trading the long end correctly even when the index does not. | | The weekly column is the sharper document again. Industrials −5.04% and technology −4.86% against basic materials +6.13% and healthcare +4.06% describes a five-session rotation out of the cycle and the AI complex and into hard assets and defensives. Note the composition trap: basic materials is +23.10% year to date and its members hit an all-time high intraday on the Canada steel news before fading, while energy at +37.64% YTD fell 0.86% tracking a crude price down 1.88%. Utilities are back above water for 2026 at +0.13%, having been −0.68% on Friday — a single 0.81% session flipped the year, which is how thin the sector's 2026 has been. | | 4 · Movers & Single-Name Catalysts |
| | Higher | - Expedia (EXPE) +5.44% to $339.13 on 1.56m shares — range $323.55–$341.09, and a fresh 52-week high, its second of August. Evercore ISI's Mark Mahaney raised his target to $430, the Street high, reiterating Outperform. The best S&P 500 performer on the day.
- Dollar Tree (DLTR) +4.01% to $136.75 — range $131.57–$137.07, three sessions before it reports (§5).
- Church & Dwight (CHD) +3.78% to $102.53, Altria (MO) +3.60% to $68.47 on 10.12m shares, Costco (COST) +2.50% to $971.40 — the consumer-staples block that CNBC's GICS tally put at +1.4%, the best sector of the session.
- Mastercard (MA) +3.31% to $599.86 and Visa (V) +3.06% to $382.41 — Visa at a 52-week high and the largest Dow contributor, worth roughly +71 index points on an $11.37 per-share gain at the current divisor.
- American Water Works (AWK) +3.34%, Edison International (EIX) +3.32%, PG&E (PCG) +2.90%, BXP (BXP) +3.43%, Duke Energy (DUK) +1.77%, Southern (SO) +1.30% — the duration-sensitive block, on a 4 bp rally at 20 and 30 years.
- Walmart (WMT) +2.69% to $106.49 on 35.62m shares — range $104.00–$106.60. It finally bounced. Friday's edition led on the fact that Walmart could not rally the day after a 9.15% collapse even with JPMorgan publicly buying it; two sessions later it has recovered 2.7% and JPMorgan's $125 target now sits 17.4% above the price.
- GoDaddy (GDDY) +3.55%, Gartner (IT) +3.51%, VeriSign (VRSN) +3.19%, Arthur J. Gallagher (AJG) +3.10%, Yum! Brands (YUM) +2.85%, eBay (EBAY) +2.83%, FedEx (FDX) +2.76%, Chipotle (CMG) +3.04% on 18.2m shares.
- Ulta Beauty (ULTA) +3.32% to $538.76 — three sessions before it reports; UnitedHealth (UNH) +2.22% to $398.76; JPMorgan (JPM) +1.37% to $356.39.
- Meta (META) +1.66% to $559.02, Amazon (AMZN) +1.33% to $262.07, Alphabet A (GOOGL) +0.94% to $348.06, Microsoft (MSFT) +0.84% to $487.31, Apple (AAPL) +0.32% to $310.34 — all five megacap platforms rose while the semiconductor complex fell, the widest platform-versus-silicon split of the month.
- Strategy (MSTR) +2.83% to $122.63 on 41.3m shares — the best Nasdaq 100 performer, after the company set up a reserve that can be used to buy bitcoin (Bloomberg). Freeport-McMoRan (FCX) +1.49% to $77.80, at an intraday all-time high and pacing a four-day run of more than 16%.
- RUM Group (RUM) +5% intraday — a $13.7bn deal to supply an unnamed cloud customer with AI chips, per an SEC filing. Not an S&P 500 member.
| | Lower | - Seagate (STX) −6.51% to $794.65, SanDisk (SNDK) −6.45% to $1,493.12, Micron (MU) −5.83% to $910.43, Western Digital (WDC) −5.24% to $435.38 — the memory and storage complex, on the Nvidia server price-hike report. CNBC flagged Sandisk as the worst S&P 500 performer at −6.7% on its own tally and Seagate, Micron and Western Digital as the worst three in the Nasdaq 100, all down at least 5%.
- Amentum (AMTM) −6.65% to $20.06 — range $19.99–$21.65, closing at the low; the worst performer on Investing.com's component board.
- J.B. Hunt (JBHT) −5.65% to $259.88 — range $256.75–$275.68, a 7.2% high-to-close fade, on the 50% Canadian auto and steel tariff. Old Dominion (ODFL) −2.42% to $201.48; BorgWarner (BWA) −5.05% to $64.36.
- Super Micro (SMCI) −5.56% to $35.17 on 39.93m shares; Axon (AXON) −4.80% to $597.59; Akamai (AKAM) −4.31%.
- Moderna (MRNA) −4.30% to $138.89 on 31.59m shares — range $130.00–$145.50, an 11.2% intraday span, the fourth consecutive session of double-digit intraday range after +176.97%, −23.55% and +8.86%.
- Robinhood (HOOD) −4.17% to $103.62 and Coinbase (COIN) −3.76% to $179.48 — the crypto equity complex gave back part of last week's move while bitcoin itself held, closing the U.S. day near $78,800 before topping $80,000 overnight. Strategy went the other way.
- Tesla (TSLA) −3.83% to $348.95 on 39.19m shares — range $348.26–$363.24, closing at the low, giving back three-quarters of Friday's 5.14% gain.
- AMD −3.49% to $456.75, Marvell (MRVL) −3.27% to $229.29, Intel (INTC) −3.12% to $87.26 on 96.9m shares, Nvidia (NVDA) −2.91% to $208.48 on 135.19m shares, Broadcom (AVGO) −2.63% to $358.76, Taiwan Semiconductor ADR (TSM) −2.11% to $410.12. Bloomberg reported that Taiwan has indicted an Nvidia manager following a chip-smuggling probe.
- Oracle (ORCL) −2.74% to $142.45 on 14.16m shares — gave back nine-tenths of Friday's 3.10% rebound. Palantir (PLTR) −2.25% to $175.89.
- Caterpillar (CAT) −2.04% to $811.02 — the worst Dow component, and worth roughly −105 index points, which the Dow overcame anyway. Boeing (BA) −1.75% to $210.46; Merck (MRK) −1.24% to $150.66; Chevron (CVX) −1.06% to $203.09.
- HP Inc. (HPQ) −3.80% to $28.58 and Agilent (A) −3.50% to $153.43 — both report Wednesday. Quanta Services (PWR) −3.53%, CDW −3.45%, LyondellBasell (LYB) −3.45%, Enphase (ENPH) −3.70%.
- Applied Optoelectronics (AAOI) −11% intraday on a $600m at-the-market equity programme filed with the SEC through Raymond James and Needham. IREN −4.94% to $39.81; Rocket Lab (RKLB) −5.91%; XPeng (XPEV) −8.53%; NIO −5.83%; Alibaba ADR (BABA) −0.73% to $118.47 after pricing an HK$80bn ($10.2bn) share placement at HK$112.70, a 8.4% discount to Friday's HK$123, with all net proceeds earmarked for AI infrastructure. None of these six is an S&P 500 member.
| | Analyst actions | - Evercore ISI raised Expedia (EXPE) to a Street-high $430 target, Outperform reiterated — 26.8% above Monday's $339.13 close. The stock rose 5.44% to a 52-week high, the cleanest analyst-to-price transmission of the session.
- Wolfe Research upgraded PulteGroup (PHM) to Outperform from Peer Perform, target $158, on built-to-order mix. Trevor Allinson: built-to-order at "45% of 2Q Orders" against a 60% company target implies "a ~50 bp idiosyncratic tailwind" to gross margins over time.
- Baird upgraded Darden Restaurants (DRI) to Outperform, target $250, "implying a 13% upside from Friday's close" on Baird's own framing, calling it "a core casual dining holding" with a decade-long total-shareholder-return algorithm above 15%.
- Wells Fargo cut Canada Goose (GOOS) to Underweight from Overweight on an El Niño-driven warm winter, taking the target to $10 from $16 and modelling "a modest -600bps impact to Q3 comps." A weather downgrade two notches in one step is unusual; it is also the second Canada-linked negative of the session.
- Jersey Mike's (JMKE) drew a wave of initiations after its 30 July debut. Morgan Stanley Overweight, $29 (roughly 22% upside from Friday); Jefferies Buy, $29; Loop Capital $40, implying 68%. Morgan Stanley's Brian Harbour: "a robust franchised business and long-standing share taker."
- Set the analyst tape against the price tape. On Friday BMO initiated Nvidia at Outperform with a $340 target; Nvidia has fallen in every session since and is now 63.1% below that target at $208.48, having declined seven days running. The sell-side and the tape have been pointing in opposite directions on the AI complex for a full week.
| | 5 · S&P 500 Earnings Calendar — Current & Next Week |
| | 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 24–28) — remaining |
| | Tue 8/25. AMC: Intuit (INTU). | | Wed 8/26. BMO: J.M. Smucker (SJM), Williams-Sonoma (WSM). AMC: Nvidia (NVDA), Salesforce (CRM), CrowdStrike (CRWD), Synopsys (SNPS), Agilent Technologies (A), Veeva Systems (VEEV), HP Inc. (HPQ). | | Thu 8/27. BMO: Dollar General (DG), Dollar Tree (DLTR), Best Buy (BBY), Hormel Foods (HRL). AMC: Marvell Technology (MRVL), Autodesk (ADSK), Workday (WDAY), Ulta Beauty (ULTA). | | Fri 8/28. No S&P 500 reporter on either bucket. | | Next week (Aug 31 – Sep 4) |
| | Mon 8/31. No S&P 500 reporter on either bucket; the Nasdaq capture returns three names in total 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 (8/21 report): | - No additions and no removals for 8/25 through 8/28. Every S&P 500 name on the prior roster reappears on this session's independent Nasdaq capture in the same before-open / after-close bucket. Williams-Sonoma before the open on 26 August is now confirmed by a third consecutive capture.
- Next week is new to this report; nothing to diff against.
- Dual listings deduped: Brown-Forman appears as both BF.A and BF.B on the source calendar and is carried once, as the class B share, which is the index line.
- Non-members on the same dates, listed so nobody mistakes their absence for an omission: Bank of Montreal (BMO), Bank of Nova Scotia (BNS), Heico (HEI/HEI.A), Zoom (ZM), Dick's Sporting Goods (DKS) and Semtech (SMTC) on 8/25; Trip.com (TCOM), Li Auto (LI), Okta (OKTA), Nutanix (NTNX), Dycom (DY), Donaldson (DCI) and Everpure (P) on 8/26; Royal Bank of Canada (RY), Toronto-Dominion (TD), CIBC (CM), Affirm (AFRM), Rubrik (RBRK), Burlington (BURL), IREN and Harmony Gold (HMY) on 8/27; Credo (CRDO), MongoDB (MDB), NIO and GitLab (GTLB) on 9/1; Snowflake (SNOW), Five Below (FIVE) and Argan (AGX) on 9/2; Ciena (CIEN), Zscaler (ZS), Samsara (IOT), Guidewire (GWRE), DocuSign (DOCU), UiPath (PATH) and Planet Labs (PL) on 9/3; Frontline (FRO) and KT Corp (KT) on 9/4. Borderline membership cases are listed in Data Notes and conservatively excluded.
- What the forward calendar hands the desk. The concentration has not moved: eighteen S&P 500 names across three days, and nine of them on Wednesday 26 August, seven after the close. What has changed is the price the calendar is being met at. Nvidia arrives having fallen in seven consecutive sessions and 2.91% on Monday alone, on 135.19m shares, with the memory complex that supplies it down 5–6.5% the same day on a report about its own input costs (§2). Agilent and HP Inc. both fell more than 3.5% on Monday, two days ahead of their prints. Then the calendar rolls straight into a second AI week: Broadcom on Wednesday 2 September, which is the other side of the financing question this report has been tracking, and Palo Alto Networks and Dell on Tuesday 1 September. For the reaction function and the dispersion inside it, see §4 and §2.
| | 6 · U.S. Treasury Yields — Official Par Curve |
| | Source: U.S. Department of the Treasury daily par yield curve, 24 August 2026, read from the official Text View. All fourteen tenors were extracted; the ten published below are the report's standing tenor set. Changes are versus the 21 August row (1-day) and the 17 August row (1-week). | | Tenor | 24 Aug | 1-Day | 1-Week |
|---|
| 1 Mo | 3.79% | −1 bp | 0 bp | | 3 Mo | 3.87% | −1 bp | 0 bp | | 1 Yr | 4.04% | +1 bp | +4 bp | | 2 Yr | 4.24% | 0 bp | +5 bp | | 3 Yr | 4.31% | 0 bp | +6 bp | | 5 Yr | 4.41% | −2 bp | +3 bp | | 7 Yr | 4.55% | −2 bp | +1 bp | | 10 Yr | 4.70% | −4 bp | −2 bp | | 20 Yr | 5.21% | −4 bp | −9 bp | | 30 Yr | 5.23% | −4 bp | −8 bp |
| | Spread | 24 Aug | 1-Day | 1-Week |
|---|
| 2s10s | 46 bp | −4 bp | −7 bp | | 3M10Y | 83 bp | −3 bp | −2 bp | | 2s30s | 99 bp | −4 bp | −13 bp | | 20s30s | 2 bp | 0 bp | +1 bp |
| | The shape and its diagnostic: a bull flattener with the pivot inside two years. Everything from five years out richened — 5s and 7s by 2 bp, 10s, 20s and 30s by 4 bp each — while the 2-year and 3-year did not move at all and the 1-year cheapened a basis point. Bills went the other way in miniature, with the 1-month and 3-month each 1 bp richer. A flattener produced by the long end rallying, rather than the front end selling off, is a term-premium event, and it is the exact opposite of Friday's parallel policy repricing. The proof that it is term premium and not growth is in §8: on the same afternoon the market added 2.4 points to the September hike probability. Markets do not simultaneously price more tightening and a weaker economy; they price an intervention in the long end while leaving the funds-rate view alone. | | The weekly table is now unambiguous, and it vindicates the tenor Treasury named. Across five sessions the 20-year has richened 9 bp and the 30-year 8 bp, while the 2-year has cheapened 5 bp, the 3-year 6 bp and the 5-year 3 bp. 2s30s has flattened 13 bp to 99, through the 100 bp level for the first time in this reporting window, and 2s10s 7 bp to 46. One week ago this report noted that the 20-year was the only tenor on the curve that had given nothing back; it is now the tenor that has taken the most back. The buyback programme has bought nothing — the first operation is 9 September — and it has moved 9 basis points at the point of the curve it is aimed at. | | Off-table bills, where the funding story lives. The dropped tenors did move, and they moved in a way that complicates the picture rather than confirming it. The 1.5-month rose 1 bp to 3.78%, the 2-month was unchanged at 3.80%, the 4-month unchanged at 3.90% and the 6-month rose 1 bp to 3.96% — so while the two published bills richened a basis point each, the tenors either side of them cheapened or held. On the week the shape is a clean bear-steepening of the bill curve: 1.5-month −2 bp, 2-month −2 bp, 1-month and 3-month flat, 4-month +1 bp, 6-month +1 bp. That is a market beginning to price more bill supply at three-to-six months, which is precisely what a bills-for-bonds "Treasury Twist" would require, and it belongs with the repo data in §9 rather than with the term-premium argument above. The front-end kink this report flagged for three sessions has also resolved: the 1-month at 3.79% now sits 1 bp above the 1.5-month at 3.78%, against a 3 bp inversion on Friday. | | Vendor gap, explained rather than disputed. Bloomberg's board marked the 10-year at 4.71% and CNBC quoted the 30-year at 5.234% and the 10-year at 4.704% intraday, against official par closes of 4.70% and 5.23%. The morning CNBC read had the 10-year at 4.7120%, −2 bp, and the 30-year at 5.2497%, −2 bp, so the move extended through the afternoon rather than fading. The official par curve is struck near 3:30 p.m. ET on bid-side quotes and carries its own smoothing; real-time vendor marks run into the evening electronic session, where by 1:08 a.m. ET Tuesday WSJ's board already had the 10-year back at 4.716%, +1.5 bp, and the 30-year at 5.240%. Levels agree within a basis point at every tenor; the differences are timing, not disagreement. | | 7 · U.S. Macroeconomic Calendar |
| | Source: Federal Reserve Bank of New York official Economic Indicators Calendar for August and September 2026. Times ET. Sensitivity is this report's rating, and it drives what §1 must name. | | Current week (Aug 24–28) — remaining | | Date | Time (ET) | Release | Sensitivity |
|---|
| Tue 8/25 | 08:30 | Philadelphia Fed Non-Manufacturing Survey | Low | | Tue 8/25 | 10:00 | Consumer Confidence (Conference Board) | Medium | | Tue 8/25 | 10:00 | New Residential Sales | Medium | | Tue 8/25 | 10:00 | Richmond Fed Survey of Manufacturing Activity | Low | | Wed 8/26 | 08:30 | Advance Durable Goods | Medium | | Wed 8/26 | 08:30 | Gross Domestic Product, 2nd release (Q2) | Medium | | Wed 8/26 | 08:30 | Personal Income and the PCE Deflator | Very high | | Wed 8/26 | 10:00 | Corporate Bond Market Distress Index (NY Fed) | Low — but see §9 | | Thu 8/27 | 08:30 | Initial Jobless Claims | High | | Thu 8/27 | 10:00 | Multivariate Core Trend Inflation (NY Fed) | Medium | | Thu 8/27 | 11:30 | Weekly Economic Index | Low | | Thu 8/27 | 14:00 | R-Star (Laubach–Williams estimates) | Low | | Thu 8/27 – Sat 8/29 | — | Jackson Hole Economic Policy Symposium (Kansas City Fed); Chair Kevin Warsh's keynote, Friday morning 28 August | Very high (event) | | Fri 8/28 | 10:00 | Michigan Consumer Survey (final) | Medium | | Fri 8/28 | 12:45 | New York Fed Staff Nowcast | Low | | Fri 8/28 | 14:00 | R-Star (Holston–Laubach–Williams estimates) | Low |
| | Next week (Aug 31 – Sep 4) | | Date | Time (ET) | Release | Sensitivity |
|---|
| Mon 8/31 | 10:30 | Dallas Fed Manufacturing Survey | Low | | Tue 9/1 | 10:00 | ISM Manufacturing | High | | Tue 9/1 | 10:00 | JOLTS | Medium | | Tue 9/1 | 10:00 | Construction Spending | Low | | Tue 9/1 | 10:30 | Dallas Fed Texas Retail Outlook Survey | Low | | Wed 9/2 | 08:15 | ADP National Employment Report | Medium | | Wed 9/2 | 09:00 | Labor Market Tightness Index (NY Fed) | Low | | Wed 9/2 | 10:00 | Manufacturing, Shipments and Orders | Low | | Thu 9/3 | 08:30 | Initial Jobless Claims | High | | Thu 9/3 | 08:30 | Advance International Trade in Goods | Low | | Thu 9/3 | 08:30 | Trade Balance | Low | | Thu 9/3 | 08:30 | Productivity & Costs (revised) | Low | | Thu 9/3 | 10:00 | ISM Non-Manufacturing | High | | Thu 9/3 | 11:30 | Weekly Economic Index | Low | | Fri 9/4 | 08:30 | Employment Situation (August payrolls) | Very high | | Fri 9/4 | 10:00 | Global Supply Chain Pressure Index (NY Fed) | Low | | Fri 9/4 | 12:45 | New York Fed Staff Nowcast | Low |
| | Markets are closed Monday 7 September for Labor Day, so the week after next opens on Tuesday. | | The look-ahead — thirty-six hours of nothing, then eleven days that decide the September meeting. From Monday's close to Wednesday's 8:30 there is exactly one release with the capacity to move a rate: Consumer Confidence on Tuesday at 10:00, rated Medium. Then it does not stop. The PCE deflator lands Wednesday at 8:30 into a strip that now prices 41.4% for a September hike, 57.6% cumulative by October — a majority for a second consecutive session — and 74.9% by December, with zero probability of a cut at any 2026 meeting (§8). The Cleveland Fed nowcast has tracked core PCE near 3.3% year on year, and the specific upside risk this month is the 6.5% surge in portfolio management fees that feeds the services line directly. Twelve hours after the print, Nvidia reports (§5). Thursday brings claims, rated High because the four-week average sits near 204,000 and Thursday's 206,000 removed the labour-market route to easing; and Friday brings Chair Warsh's first Jackson Hole address, into a bond market where the Treasury has just moved the 20-year 9 bp in a week without buying a single security. That address is not a data release and it is rated Very high anyway, because it is the only scheduled event that can re-anchor a terminal rate now drawn at 4.085%. Then next week hands the desk ISM manufacturing on 1 September, ISM services and claims on the 3rd, and August payrolls on Friday 4 September — the last labour print before the 16 September FOMC. The asymmetry has changed shape since Friday. A week ago the dovish surprise had more room to travel because the market had been taking easing out; now the long end has already rallied 8–9 bp on a fiscal-plumbing story rather than a data story, so a soft core PCE gets less help from duration than it would have and would instead show up in the front end, where nothing has moved in two sessions. A hot print, by contrast, hits a 2-year that is pinned at 4.24% and a September contract that is already 41.4% priced — and it arrives with Warsh forty-eight hours behind it. The concrete hooks, in the order they can move the Fed card: Consumer Confidence 8/25 10:00 → PCE deflator 8/26 08:30 → claims 8/27 08:30 → Warsh 8/28 morning → ISM manufacturing 9/1 10:00 → ISM services and claims 9/3 → payrolls 9/4 08:30. |
| | 8 · Fed Funds Futures & Rate Path |
| | CME FedWatch — the four-column headline (September 16 meeting) | | Meeting information. Meeting date 16 September 2026; contract ZQU6; expires 30 September 2026; mid price 96.3175; prior volume 36,441; prior open interest 246,151. Current target range 3.50–3.75%. | | Target rate (bp) | NOW | 1 day (21 Aug 2026) | 1 week (17 Aug 2026) | 1 month (24 Jul 2026) |
|---|
| 350–375 (current) | 58.1% | 60.1% | 63.9% | 18.0% | | 375–400 | 41.9% | 39.9% | 36.1% | 55.3% | | 400–425 | 0.0% | 0.0% | 0.0% | 26.6% | | EASE / NO CHANGE / HIKE | 0.0% / 58.1% / 41.9% | 0.0% / 60.1% / 39.9% | 0.0% / 63.9% / 36.1% | 0.0% / 18.0% / 81.9% |
| | Provenance of every column, stated — including a correction to the prior edition. All four columns were read live and numerically off CME FedWatch's Current view, whose footer stamps the data as of 24 Aug 2026 11:46:18 CT; the adjacent interest-rate panel on the same page carries 24 Aug 2026 11:59:50 PM CT, so this is a late-evening read, indicative rather than a settlement snapshot. No column here is chart-read, and the 1 MONTH column carries a reference date of 24 July 2026, exactly one calendar month back, so it is usable in calculations. Correction: Friday's edition published CME's live NOW column as 59.9% hold / 40.1% hike; CME's settled 1 DAY column for the same date now reads 60.1% / 39.9%, a 0.2 percentage point difference. The settled figures are used in every calculation below, and the standing caveat about post-5:00 p.m. reads is the reason the discrepancy exists. | | The CME-versus-Investing.com gap, reconciled and quantified. CME puts the September hike at 41.9%; Investing.com's Fed Rate Monitor, timestamped 25 Aug 2026 12:35 a.m. EDT, puts it at 41.4% — a gap of 0.5 percentage points, and the smallest in this reporting window. The mechanical cause is again the quoted contract price: CME's mid is 96.3175 and Investing.com's 96.320, a difference of 0.25 basis points, with the sign reversed from Friday. Because the 16 September decision is effective from 17 September, only 14 of September's 30 days carry the post-meeting rate in the ZQ average, so the contract weights the new rate at 14/30 = 46.7%. A 0.25 bp price difference therefore scales to 0.25 × 30/14 = 0.54 bp of implied post-meeting rate, or 0.54 / 25 = 2.1 percentage points of hike probability. The price difference over-explains the observed 0.5 pp gap by a factor of four; the residual is each vendor's assumed effective-rate base, and CME's can be checked. Back-solving from a price of 96.3175 gives an implied September average of 3.6825%, which net of 41.9% × 25 bp × 14/30 = 4.89 bp implies a pre-meeting effective rate of 3.634% — against a New York Fed EFFR of 3.63% for 21 August (§9). The two vendors are pricing the same distribution and rounding it differently. | | Multi-day momentum, and a week that has now gone one way. The hike has been bid for eight sessions without a single reversal: 36.1% → 39.9% → 41.9% on CME across 17, 21 and 24 August, and 32.3% → 35.3% → 39.0% → 41.4% on Investing.com's independent columns. That is +5.8 points in a week and +2.0 points in a session on CME, +2.4 points on Investing.com. What makes Monday's increment notable is the company it kept: it happened on the same afternoon the 30-year rallied 4 bp and the 20-year 4 bp. The one-month column still reverses the sign of the longer arc — on 24 July the market priced 81.9% cumulative hike including 26.6% on a full 50 bp, against 41.9% and 0.0% today — so the correct statement remains a large dovish repricing through late July and early August that has now retraced roughly a third of itself in six sessions. The retracement continues to run against the data: it has now survived a 206,000 claims print and a Philadelphia Fed price series at a six-month low. The market is trading the supply of duration and the identity of the Chair, not the inflation path. | | (a) Current-year meeting distributions — current / [prev-day] / [prev-week] | | Source: Investing.com Fed Rate Monitor, updated 25 Aug 2026 12:35 a.m. EDT. Current target range 3.50–3.75%. | | Meeting | Contract | 3.50–3.75 (hold) | 3.75–4.00 (+25) | 4.00–4.25 (+50) | 4.25–4.50 (+75) | Cumulative hike |
|---|
| 16 Sep 2026 | 96.320 | 58.6% [61.0] [66.0] | 41.4% [39.0] [34.0] | 0.0% | 0.0% | 41.4% [39.0] [34.0] | | 28 Oct 2026 | 96.255 | 42.4% [47.0] [50.8] | 46.2% [44.1] [41.4] | 11.4% [9.0] [7.8] | 0.0% | 57.6% [53.1] [49.2] | | 9 Dec 2026 | 96.115 | 25.1% [28.9] [31.0] | 44.6% [45.2] [45.0] | 25.6% [22.5] [20.9] | 4.7% [3.5] [3.0] | 74.9% [71.2] [68.9] |
| | Cut probability is 0.0% at every 2026 meeting, unchanged for the whole of this reporting window. Two thresholds moved on Monday. September's hold fell below 60% for the first time, to 58.6%. And October's hold fell below the +25 cell — 42.4% against 46.2% — so the single most likely outcome for the October meeting is now a hike, not a hold, having merely been a majority in aggregate on Friday. Read the momentum across the three meetings: September +2.4 points, October +4.5, December +3.7 in a session, and +7.4, +8.4, +6.0 across the week. The whole 2026 strip is moving together with the middle moving most, which is a level shift in the expected path rather than a single-meeting reassessment — and it is now eight sessions old. | | (b) Next-year path — modal range, cumulative and the contracts that draw the terminal | | Meeting | Contract price | Modal range | Modal prob. | Cumulative above 3.50–3.75 | Cumulative below |
|---|
| 27 Jan 2027 | 96.085 | 3.75–4.00 | 40.9% | 79.8% | 0.0% | | 17 Mar 2027 | 96.010 | 3.75–4.00 | 34.9% | 85.5% | 0.0% | | 28 Apr 2027 | 95.970 | 4.00–4.25 | 32.9% | 87.2% | 0.0% | | 9 Jun 2027 | 95.930 | 4.00–4.25 | 32.9% | 88.7% | 0.0% | | 28 Jul 2027 | 95.915 | 4.00–4.25 | 32.8% | 89.0% | 0.0% | | 15 Sep 2027 | 95.915 | 4.00–4.25 | 32.3% | 87.9% | 0.4% | | 27 Oct 2027 | 95.920 | 4.00–4.25 | 32.0% | 87.2% | 0.7% | | 8 Dec 2027 | 95.940 | 4.00–4.25 | 30.9% | 84.8% | 1.7% |
| | The terminal moved another basis point higher and the modal range stepped forward two meetings. The cheapest contracts on the strip are July and September 2027 at 95.915, implying 4.085% — against 95.925 and 4.075% on Friday and 4.020% a session before that. The modal range now flips from 3.75–4.00% to 4.00–4.25% at the April 2027 meeting, two meetings earlier than Friday's June crossover, and it no longer flips back in December — the December 2027 modal cell is 4.00–4.25% at 30.9%, marginally ahead of 3.75–4.00% at 30.2%. Cumulative pricing above the current range peaks at 89.0% in July 2027. The strip prices a cut at some point in 2027 at 1.7% by December, up from 1.3% a day earlier but still down from 3.2% a week earlier. The December-2027 contract richening to 95.940 from the 95.915 trough is the only shape on the strip that resembles an eventual cut, and it is worth 2.5 bp. | | (c) Year-end probability ladders | | Year-end 2026 — the 9 December meeting. Current target range 3.50–3.75%. | | Outcome | Range | Current | [prev-day] | [prev-week] |
|---|
| −25 bp | 3.25–3.50 | 0.0% | 0.0% | 0.0% | | Hold | 3.50–3.75 | 25.1% | [28.9] | [31.0] | | +25 bp | 3.75–4.00 | 44.6% | [45.2] | [45.0] | | +50 bp | 4.00–4.25 | 25.6% | [22.5] | [20.9] | | +75 bp | 4.25–4.50 | 4.7% | [3.5] | [3.0] | | +100 bp | 4.50–4.75 | 0.0% | 0.0% | 0.0% |
| | Year-end 2027 — the 8 December meeting. | | Outcome | Range | Current | [prev-day] | [prev-week] |
|---|
| −75 bp | 2.75–3.00 | 0.0% | — | 0.0% | | −50 bp | 3.00–3.25 | 0.1% | [0.0] | [0.2] | | −25 bp | 3.25–3.50 | 1.6% | [1.3] | [3.0] | | Hold | 3.50–3.75 | 13.6% | [14.0] | [18.6] | | +25 bp | 3.75–4.00 | 30.2% | [30.9] | [33.3] | | +50 bp | 4.00–4.25 | 30.9% | [30.7] | [28.1] | | +75 bp | 4.25–4.50 | 17.1% | [16.7] | [12.9] | | +100 bp | 4.50–4.75 | 5.5% | [5.3] | [3.4] | | +125 bp | 4.75–5.00 | 1.0% | [1.0] | [0.5] | | +150 bp | 5.00–5.25 | 0.1% | [0.1] | 0.0% | | +175 bp and beyond | 5.25–5.50+ | 0.0% | 0.0% | — |
| | Cumulative for year-end 2027: above the current range 84.8% [84.7] [78.2]; below 1.7% [1.3] [3.2]. The modal cell flipped from +25 to +50 basis points on Monday — 30.9% against 30.2% — the first time in this reporting window that the year-end 2027 distribution has centred a full two hikes above the current range. The two cells remain within 0.7 percentage points of each other, so the flip is a tie-break rather than a conviction, but the direction has been one-way for six sessions. | | Rounding, stated transparently. The Investing.com rows sum to 100.0% for September, October and December 2026 and for April, July, September 2027; 100.1% for January 2027, October 2027 and December 2027; and 99.9% for March and June 2027. Cumulative figures in this section are computed from the published cells without renormalising, so a cumulative may inherit up to 0.1 point of that rounding. | | | (a) IG and HY credit spreads | | As-of date, stated rather than implied. FRED's ICE BofA series publish with a one-business-day lag; all three carried 21 August as their last observation when read this session, so Monday's session is not in these numbers and publishes on 25 August. Same-day direction was cross-checked against the cash-ETF tape: LQD closed +0.25% at $106.18 and HYG +0.11% at $79.70 on a day the 10-year richened 4 bp. IG cash outperformed HY cash by 14 bp and both closed higher — a duration outcome, exactly as the curve would predict, with no credit signal in either direction. | | Series | Level | As-of | 1-day | 1-week | YTD (31 Dec 2025) |
|---|
IG — ICE BofA US Corporate OAS (BAMLC0A0CM) | 81 bp | 21 Aug | −1 bp (82) | +1 bp (80 on 14 Aug) | +2 bp (79) | HY — ICE BofA US High Yield OAS (BAMLH0A0HYM2) | 270 bp | 21 Aug | −5 bp (275) | +3 bp (267) | −11 bp (281) | CCC & lower OAS (BAMLH0A3HYC) | 1,037 bp | 21 Aug | +2 bp (1,035) | +25 bp (1,012) | +152 bp (885) | | CCC minus HY differential | 767 bp | 21 Aug | +7 bp (760) | +22 bp (745) | +163 bp (604) | | CDX IG 5y | see retrieval note | — | — | — | — | | CDX HY 5y | see retrieval note | — | — | — | — |
| | Mark the prior edition's written test, because this one resolved cleanly and in a direction worth noting. Friday's report set two conditions: "a 24 August IG print at 84 bp or wider would say Friday's 5 bp parallel coupon shift reached the core and would make the widening a trend rather than a beta; a CCC print inside 1,025 bp would say the tail move was a summer-liquidity artefact and end this thread." The 24 August publication delivered IG at 81 bp — one basis point tighter, not wider, and CCC at 1,037 bp — 2 bp wider, not inside 1,025. Neither condition triggered, and the failure mode is the informative part: on a session when the entire coupon curve cheapened 4–5 bp, the IG index tightened and HY tightened 5 bp, so the core absorbed a duration shock without widening at all. Meanwhile the tail did the opposite. CCC has widened 25 bp in a week against HY's 3 and IG's 1, the differential stands at 767 bp, +163 bp on the year, and it has now widened in five of the last six published sessions — including two in which the index itself tightened. The core-versus-tail divergence this report identified on Friday is not merely intact; it is the only part of the credit market carrying information. | CDX retrieval note — the six-step ladder, worked and named. (1) Bloomberg in Chrome: /markets and /markets/rates-bonds both rendered fully, including the Americas, EMEA and Asia-Pacific 10-year yield boards; a programmatic search of the rendered page text returned no CDX string, and the same search on the 24 August high-grade issuance story returned none either. (2) WSJ Market Data: wsj.com/market-data/bonds rendered, with its U.S. Treasurys table live at 1:08 a.m. ET 8/25, and a full-text search returned no CDX match. (3) Cbonds / ICE / S&P Global: the Cbonds CDX.NA.IG 5Y page rendered again this session, and it is now demonstrably stale as well as masked — the level renders masked and the previous value is dated 20/08/2026, four calendar days back, so even a subscription read would not have carried Monday. CDX.NA.HY 5Y and the iTraxx family are identically masked; ICE's Markit CDX.NA.IG product page and the S&P Dow Jones CDX family page publish methodology and index-news documents, not levels. (4) FT Markets Data and Reuters credit wraps: no instrument page and no wire wrap quoting a level in the reviewed material. (5) TradingView / Barchart / CME CDS index products: documentation only. (6) Cash-market proxy: LQD +0.25% and HYG +0.11%, published as a proxy for the direction of cash credit and not as a CDX level. Quoting conventions restated: CDX IG 5y is quoted in basis points of spread; CDX HY 5y in price points, where a rising price means tightening credit spreads. This is the fourth consecutive edition in which the full ladder has been worked and the field remains a gap. | | (b) Money-market and funding plumbing | | NY Fed reference rates publish at roughly 8:00 a.m. ET for the prior business day; the overnight series below carry a 21 August effective date. The SOFR averages and the index carry 24 August, and the reverse-repo take-up is 24 August. | | Measure | Level | Detail |
|---|
| SOFR | 3.65% | 21 Aug, up 2 bp from 3.63% and level with IORB. Volume $2,952bn; 1st pct 3.59%, 25th 3.63%, 75th 3.70%, 99th 3.73% | | EFFR | 3.63% | 21 Aug, unchanged. Volume $96bn, down from $102bn; 1st pct 3.60%, 25th 3.62%, 75th 3.63%, 99th 3.69% | | OBFR | 3.63% | 21 Aug, unchanged. Volume $224bn; 1st pct 3.53%, 99th 3.69% | | TGCR | 3.63% | 21 Aug, up 3 bp from 3.60%. Volume $1,172bn | | BGCR | 3.63% | 21 Aug, up 3 bp from 3.60%. Volume $1,198bn | | SOFR − IORB | 0 bp | IORB 3.65%. Back level with the administered rate after two sessions through it | | 30-day average SOFR | 3.64485% | 24 Aug effective date; 90-day 3.64256%, 180-day 3.65994%; SOFR index 1.25604095 | | ON RRP take-up | $0.380bn | 24 August, from $0.200bn on 21 August. Still effectively empty against a $2.4tn peak | | Reserve balances (week to 19 Aug) | $2,935.3bn | No new weekly print — the next release is 27 August. −$207.4bn from the 15 July peak of $3,142.7bn | | 6-month bill (par, 24 Aug) | 3.96% | +1 bp on the day and +1 bp on the week, while the 1-month and 3-month each richened; off-table under this report's tenor set (§6) |
| | The funding question reopened, and it reopened on exactly the terms this report wrote down. Friday's edition set the test: "a SOFR print back at 3.65% or above with the 75th percentile through 3.70% would reopen it as a bill-supply problem." The 24 August publication carried the 21 August effective date at 3.65%, up 2 bp and level with IORB, with the 75th percentile at exactly 3.70% — the level condition is met outright and the distribution condition is met at the threshold rather than through it. The 99th percentile went to 3.73% from 3.71% and the 1st-to-99th band widened to 14 bp from 13. But the more important number is the one that is not in the test. Tri-party and broad general collateral both rose 3 bp to 3.63% after three sessions pinned at 3.60%, on roughly $1.2tn a day each. That is the secured core moving, not the tail — and the secured core is the part of the market that a bills-for-bonds swap would hit first, because it exchanges reserves for collateral. Read alongside it: the unsecured tail did nothing — EFFR unchanged at 3.63% on lower volume, OBFR unchanged, both 99th percentiles at 3.69% — so this is a collateral event, not a bank-funding event. The reverse repo facility took $380m, still a rounding error, and reserves sit $207.4bn below their July peak with no fresh print until 27 August. The bill curve is the confirming evidence: on the week the 1.5-month and 2-month richened 2 bp each while the 4-month and 6-month cheapened 1 bp each, which is the term structure of an anticipated bill programme. Bessent has described the plan as a "Treasury Twist"; the repo market appears to have started pricing it before the first operation on 9 September. | | (c) Rates volatility and swap spreads | | The MOVE series updated, and for the first time in three editions the ratio is a matched-date figure. Investing.com's board carries 73.98, +0.79%, dated 24/08, with a session range of 73.40–73.98, implying a 21 August value of 73.40. Against VIX at 15.85, the matched-date MOVE/VIX ratio is 4.67×, up from 4.57× on 20 August, the last previously matched pair. The composition of the move is what matters. VIX rose 4.76% on Monday while MOVE rose 0.79%, so equity volatility is repricing the event calendar roughly six times faster than rates volatility is — into a Wednesday that contains the PCE deflator and Nvidia within twelve hours of each other. Set that against what actually happened to rates: an 8–9 bp weekly move at 20 and 30 years, a 13 bp weekly flattening in 2s30s, and 5.8 points of September hike probability added in a week, all delivered by a rates market whose own volatility index has moved less than 1% in a day and is −2.18% on the week. A market that repriced the long end that far without its volatility index moving is a market where the level is being repriced and the distribution is not — the same configuration this report flagged on Friday, now with a fresh 4 bp of evidence behind it, and it is what makes Jackson Hole gamma cheap (§12). | | No verified 2-year, 10-year or 30-year swap-spread level was obtainable this session. The vendors publishing them sit behind entitlement walls and no wire quoted a level in the reviewed material. That is a gap and is recorded as one for a fourth consecutive edition. | | (d) Issuance, leveraged loans and private credit | | The supply story got a forward number on Monday, and a large one. Kelsey Berro of JPMorgan Asset Management told Bloomberg Television that September IG issuance expectations run to $175bn–$250bn — against an August that has already set a record at $145.2bn by 17 August, itself above August 2020's $136bn full-month mark and the third consecutive monthly record. Berro's argument is that the demand side is being underestimated: retail demand for IG this year already exceeds the full-year total for any year back to 2010, and if September's calendar clears without disruption, sidelined investors "could move quickly." Roughly $1.4tn of U.S. IG notes have been sold in 2026, about 9% above the 2020 pace, and the Street's projection is that gross issuance tops $2tn against $1.7tn in 2025. Set that against the pipeline this report has been tracking: the Broadcom package at more than $60bn of senior debt plus roughly $30bn of junior financing through a special-purpose vehicle, with Blackstone and Apollo in discussions, and BofA's flag of $370bn of AI-related debt across the complex. Monday added a fresh equity-side data point on the same theme — Alibaba raised $10.2bn in Hong Kong at an 8.4% discount, entirely for AI infrastructure, nine days after disclosing a 75% profit decline driven by a 75% jump in capex to RMB67.7bn. SoftBank came to the Japanese retail market for ¥1tn at an indicative 4.30%–4.90% and its shares fell 3.8%. No verified HY primary volume, Morningstar LSTA leveraged loan index level or bank CDS level was published in the reviewed material this session; those three fields are gaps and are recorded as gaps. | | The credit take — the core is pricing nothing, the tail everything. The funding market is now the swing variable. On the last published day IG tightened 1 bp to 81, HY tightened 5 bp to 270 and CCC widened 2 bp to 1,037, on a session when the whole coupon curve cheapened 4–5 bp. An index that tightens into a duration shock is not a market with a credit problem; it is a market with a bid. That bid has a name and a number: $175bn–$250bn of September supply that JPMorgan says the market can absorb, on retail demand already at a sixteen-year record. But look at what the same market will not fund. CCC has widened 25 bp in a week and 152 bp on the year while HY itself is 11 bp tighter than it started 2026, and the differential at 767 bp is +163 bp year to date. The cohort that needs the financing pays more every week while the cohort that does not gets a record calendar cleared at 81 bp. The new variable is the plumbing. SOFR at 3.65%, level with IORB, the 75th percentile at 3.70%, and tri-party and broad GC each 3 bp higher at 3.63% is the first session in which a Treasury policy aimed at the long end has left a mark on overnight collateral — and the programme has not bought a bond yet. What breaks it, in order of likelihood: a 25 August IG print at 84 bp or wider would say Monday's tightening was a one-day artefact and reinstate the widening trend; a CCC print inside 1,020 bp would end the tail thread; a SOFR print at 3.67% or above with tri-party through 3.65% would convert the funding story from a curiosity into the dominant driver of front-end pricing and would arrive first at the 6-month bill, now 3.96%. The NY Fed Corporate Bond Market Distress Index on 26 August at 10:00 is the one scheduled reading that can arbitrate between the first two. |
| | | Quote basis — read this before the table. The TradingEconomics currency board had already rolled its daily boundary when captured at approximately 1:07 a.m. ET Tuesday, so its own %Chg column measures only the few hours of the new Asian session and is not reproduced. Levels below are that board's prints; the 1-Day column is computed by this report as the change from the prior edition's TradingEconomics levels for the same pairs, and therefore spans Friday's close through the capture, not a clean Monday session. Worked examples are in Data Notes. Where a clean Monday close exists it is quoted in the Read column as a cross-check: Investing.com's dollar-index board closed at 98.99, +0.16%, and its Monday closes were USD/JPY 159.11, USD/CAD 1.3845 and USD/KRW 1,382.89. | | Pair | Level | 1-Day (computed) | Weekly | Monthly | YTD | Read |
|---|
| DXY | 99.068 | +0.25% | −0.59% | −2.42% | +0.76% | Investing.com's clean Monday close is 98.99, +0.16%. The dollar rose on a 4 bp long-end rally | | EUR/USD | 1.16554 | −0.20% | +0.69% | +2.53% | −0.73% | Fell on a day the bund did not move at all | | USD/JPY | 159.296 | +0.20% | −0.20% | −2.72% | +1.62% | Investing.com's Monday close 159.11. The yen has now moved less than 0.25% on four consecutive sessions | | GBP/USD | 1.36271 | −0.11% | +0.69% | +2.54% | +1.24% | The smallest G10 move; gilts were unchanged | | USD/CHF | 0.80379 | +0.31% | −1.05% | −1.86% | +1.37% | The franc weakened on a risk-off equity day — the wrong sign for a haven | | USD/CAD | 1.38590 | +0.73% | −0.27% | −1.88% | +1.01% | The largest major move of the session. 50% Canadian auto and steel tariffs from 1 January 2027; CNBC had the loonie at 0.723 per dollar by 3:10 a.m. ET | | USD/CNY | 6.72315 | +0.04% | −0.34% | −0.62% | −3.63% | Managed, and barely moved on a 1.21% CSI 300 decline | | USD/KRW | 1382.79 | −0.23% | −2.09% | −5.57% | −3.97% | The won strengthened on a 3.12% Kospi collapse — see the take | | USD/TWD | 31.8980 | +0.19% | −0.13% | −1.24% | +1.75% | Weakened with SOX down 2.70%, the orthodox response | | AUD/USD | 0.71510 | −0.25% | +0.88% | +2.28% | +7.16% | Gave back a third of Friday's 0.79% gain with copper still up | | NZD/USD | 0.59578 | −0.16% | +1.36% | +3.19% | +3.49% | Half the Australian move; the antipodean pair moved together again | | USD/MXN | 16.9522 | +0.35% | −0.65% | −2.86% | −5.92% | Still the strongest major EM currency of 2026, but it gave back on the tariff tape | | USD/SGD | 1.27102 | +0.17% | −0.56% | −1.56% | −1.18% | The Asian funding cross-check; moved with the dollar, not with the region |
| | The take — the dollar rose on a session designed to weaken it, and that is the section's finding. Monday delivered a 4 bp rally at 10, 20 and 30 years driven by an explicit official intention to suppress long-term yields, funded potentially by drawing down a $1tn cash balance. The textbook response is a weaker dollar: lower real yields, more liquidity, an activist Treasury. DXY rose 0.16% to 98.99 on Investing.com's clean Monday close and is +0.25% on the wider basis above. Now hold that against gold at $4,697.80, +0.78%, its best close since mid-May, and bitcoin through $80,000 after the U.S. close. A dollar that rallies while gold rallies is not a real-rate trade in either leg. The most coherent reading is that both are being bid against everything else — the dollar because the alternatives are worse and the tariff tape is actively damaging one of them, gold because the credibility of the issuer is what is being questioned. Druckenmiller's op-ed and El-Erian's CNBC interview both landed on Monday and both argue the same mechanism (§2). The currency market did not sell the intervention; it sold everything the intervention is aimed at competing with. | | The Korean cross broke its rule again, and this time in the informative direction. Two editions ago this report proposed that Korean equity flows are domestically levered, so a Kospi drawdown forces repatriation and lifts the won. Friday broke it — the Kospi rose 0.88% and the won strengthened — and the revision offered was that the won is simply a high-beta dollar expression, not a Kospi proxy. Monday tests the revision and it fails too: the Kospi fell 3.12%, the dollar rose, and the won strengthened 0.23%. Neither the equity rule nor the dollar-beta rule survives. What does survive is the original mechanism in its narrow form: on the two largest Kospi drawdowns in this window — −5.80% and −3.12% — the won strengthened both times, by 1.82% and 0.23%. On the rallies it has gone both ways. The honest statement is that the repatriation effect is real but asymmetric: it shows up in drawdowns and disappears in rallies, which is what forced deleveraging looks like and is not tradeable symmetrically. Size it as a crash hedge, not as a directional view. | | Read the Canadian dollar as the day's only genuine information trade. USD/CAD rose 0.73% on the computed basis and closed at 1.3845 on Investing.com's clean Monday mark, against roughly 1.3759 on Friday. Trump's post committed to 50% tariffs on all Canadian cars, trucks, parts and steel from 1 January 2027; Carney's retaliation begins 8 September. Note the calibration: BofA put the direct macro hit at "modest" — tariffs touch just over 5% of Canadian exports to the United States — yet the currency moved more than any other major pair. That gap between a modest quantified impact and a large price response is the market pricing the confidence channel BofA itself named, and UBS's read that a USMCA deal now slips into 2027. Against that, USD/JPY moved 0.20% and EUR/USD 0.20% on a day with a 3% Korean drawdown, a 2.7% semiconductor decline and a 4 bp curve move. The two largest pairs in the world are, for the fourth session running, refusing to price anything at all. | | | Commodity | Settle / level | Chg | %Chg | Weekly | Monthly | YTD | Driver |
|---|
| WTI (Oct, NYMEX front) | $85.01 | −$1.63 | −1.88% | +1.19% | +2.97% | +48.14% | Open interest 278,061; expiry 22 September. Fell on the day the Iran sanctions landed | | Brent (Oct, ICE front) | $92.17 | −$1.76 | −1.87% | +1.17% | +4.22% | +51.34% | Open interest 270,303. Brent–WTI $7.16, in 13 cents | | Gold (Comex Dec) | $4,697.80 | +$36.20 | +0.78% | +6.69% | +13.43% | +7.06% | Open interest 330,113. Best close since mid-May; intraday high 4,728.3, the highest since 13 May | | Silver (Comex Sep) | $68.594 | −$0.416 | −0.60% | +7.14% | +16.16% | −4.79% | Broke ranks with gold after four sessions in lockstep | | Platinum (Oct) | $1,889.20 | −$2.20 | −0.12% | +7.46% | +14.15% | −9.99% | Ended a three-session run as the complex's largest gainer | | Copper (Comex Sep) | $6.6050 | +$0.0250 | +0.38% | +1.49% | +3.68% | +15.79% | A second consecutive gain; Freeport-McMoRan hit an intraday all-time high | | Palladium | $1,346.00 | — | — | +4.02% | +3.74% | −18.53% | The weakest precious metal of 2026 | | Natural gas (Henry Hub, Sep) | $2.7820 | +$0.0270 | +0.98% | — | — | — | Expiry 27 August; the only energy contract higher on the day | | Gasoline (RBOB, Sep) | $3.2708 | −$0.0512 | −1.54% | −0.40% | −1.17% | +92.21% | Fell less than crude; the gasoline crack gave up 52 cents to $52.36 a barrel | | Heating oil (Sep) | $4.2677 | −$0.1948 | −4.37% | −4.03% | +3.87% | +101.31% | The largest decline in the complex. The distillate crack collapsed $6.55 to $94.23 a barrel |
| | Basis caveats, stated before the analysis. Settles for WTI, Brent, gold, silver, platinum, copper, natural gas, gasoline and heating oil are the CNBC front-contract boards' prior-close fields, read at approximately 1:00 a.m. ET Tuesday, which is the 24 August settlement for each named contract month. Percentage changes are computed against the prior edition's settles for the same contracts. Weekly, monthly and YTD columns are TradingEconomics, captured at 1:02 a.m. ET Tuesday — after that board rolled its daily boundary — so those three columns include a few hours of the new Asian session and should be read as approximations rather than as clean 24 August figures. Three further disclosures. First, gold's basis: TradingEconomics prints spot gold at $4,626.48 against the Comex December settle of $4,697.80 and Bloomberg's board at $4,682.10. The Comex settle agrees with Bloomberg to within 0.34%, inside tolerance; the TradingEconomics spot level sits 1.52% below the December contract, which is carry rather than disagreement, so the TradingEconomics gold level is not quoted and only its ratio columns are used. Second, crude's basis: TradingEconomics' live crude at $85.040 implies a prior settle of $85.01, matching the CNBC October settle exactly; Bloomberg's evening board showed $85.10, +0.11%, on the same basis. All three agree, so nothing is withheld — and note that the Investing.com continuous WTI board has re-based again, printing $87.06 for 21 August against the $86.64 October settle published Friday, which is why the Investing.com board is not used for energy this session. Third, the heating-oil move is unusually large and is double-sourced: CNBC's September contract prior close of $4.2677 and TradingEconomics' live $4.2722, +0.11% (implying $4.2675) agree to four decimal places, so the −4.37% is a real settlement move and not a contract roll — September heating oil does not expire until 31 August. | | The take — the sanctions arrived and the barrel that unwound was distillate, which is the tell. Bessent unveiled Operation Economic Outcast on Monday afternoon with the most aggressive framing of the campaign to date, and both benchmarks fell about 1.9%. That in itself is now familiar; this is the third Iran escalation in a fortnight to be sold. What is new is where it was sold. Heating oil fell 4.37%, more than twice crude, and the distillate crack collapsed $6.55 a barrel to $94.23, while gasoline fell only 1.54% and its crack gave up 52 cents to $52.36. Distillate is the product most directly exposed to sanctioned Iranian and Russian flows, and it carries the fattest crack in the complex at $94 a barrel against gasoline's $52. If the market believed Operation Economic Outcast would actually interrupt barrels, distillate is the last thing it would sell. Selling it hardest says the market is pricing enforcement risk to buyers, not supply risk to the barrel — Chinese refiners being cut off from dollar clearing is a demand event for Iranian crude, not a supply event for the world. Positioning frames the size of the unwind: crude is +48.14% year to date and Brent +51.34%, so a 1.9% day is noise against the year and a $6.55 crack move is not. | | The precious complex split for the first time in five sessions, and gold won. Through last week gold, silver and platinum moved together — +1.97%, +1.33% and +2.84% on Friday alone. Monday broke it: gold +0.78% to a mid-May high, silver −0.60%, platinum −0.12%, palladium still the year's laggard at −18.53%. When the whole complex moves the read is industrial or reflationary; when only the monetary metal moves, on a day the dollar also rose and bitcoin ran through $80,000 after the close, the read is narrower and more specific. This is the debasement bid concentrating rather than broadening — and it concentrated on the session an ex-Duquesne principal published an op-ed arguing the Treasury is suppressing a price signal. Copper is the honest counterweight: at +0.38% it rose for a second day, with Freeport at an intraday record and a four-day gain above 16%, so the industrial story has not gone away, it is simply no longer the same story. Friday's report set the test — "a reflation copper holds its gain through a hawkish PCE print and a debasement copper does not." That test is still live and it now has a cleaner control, because silver has already stepped out of the trade. | | | Desk-style ideas for institutional investors. Each carries an explicit catalyst and an explicit invalidation. These are not personalized investment advice; verify independently and size to your own mandate before acting. | | 1. The rates trade — the calendar spread has now paid four basis points; take a second quarter and raise the stop again | | Mark on the prior edition's rates idea first. The book is long ZQU6 (September 2026) against short ZQZ6 (December 2026), DV01-matched one-for-one at $41.67 per basis point per contract, entered Thursday at 96.325 / 96.160 for a spread of 16.5 bp, marked Friday at 18.5 bp, and cut to three-quarters at that level. Monday's mark: ZQU6 96.320, ZQZ6 96.115, a spread of 20.5 bp — another +2.0 bp, worth $83.34 per contract pair on the retained size and 4.0 bp, $166.68, from entry. No invalidation triggered: the spread never approached 15.5 bp; the September cumulative hike printed 41.4% on Investing.com and 41.9% on CME, both under the 50% stop; and no 2026 meeting shows a non-zero cut probability. The mechanism keeps working for the same reason: December absorbs more of every hawkish repricing than September does — cumulative hike went 71.2% → 74.9% at December against 39.0% → 41.4% at September — because the back contract has more meetings to price. | | The modal path, the base case and the tails. Modal path: hold on 16 September (CME 58.1%, Investing 58.6%, ease 0.0%, and the hold cell below 60% for the first time); hike by 28 October, where +25 bp at 46.2% is now the single most likely outcome, ahead of hold at 42.4%, with cumulative hike 57.6%; one 25 bp hike delivered by 9 December, modal 3.75–4.00% at 44.6% with cumulative hike 74.9%; the modal range steps to 4.00–4.25% from the April 2027 meeting, two meetings earlier than Friday, and stays there through December 2027; terminal 4.085%, drawn by the July and September 2027 contracts at 95.915; year-end 2027 modal cell flipped to +50 bp at 30.9%, with cumulative easing of just 1.7%. Base case: the front end is now the only part of the curve the Treasury cannot reach. Monday delivered a 4 bp bull flattener from five years out while the 2-year and 3-year did not move a basis point and the September hike probability rose 2.4 points. A buyback programme can compress term premium; it cannot change the funds rate. Every session that separates those two facts widens this spread. Tail one, dovish: a core PCE deflator at or below +0.2% m/m on 26 August compresses the spread from the front, and it would land on a strip that has now taken easing out for eight consecutive sessions. Tail two, hawkish: a Warsh keynote on 28 August that treats 3.50–3.75% as a floor takes September through 50% and the spread through 24 bp. Practical implication: the spread has captured 4.0 of the roughly 7.5 bp that separate the entry from the top of its plausible band, and it has done so entirely before the two events that matter. Take a second quarter off at 20.5 bp, carry a half, and raise the stop from 15.5 bp to 18.0 bp so the trade cannot give back more than a quarter of what it has made into a Friday keynote with no session left to fade it. | | Expression: long ZQU6 / short ZQZ6, DV01-matched one-for-one, half the Thursday size. Catalyst: PCE deflator 8/26 08:30; Jackson Hole 8/27–8/29 with Warsh Friday morning; claims 8/27 08:30; ISM manufacturing 9/1; payrolls 9/4 08:30; the first buyback operation 9 September; the 16 September FOMC. Invalidation: the spread back through 18.0 bp; or the September cumulative hike printing above 50% on either vendor; or any 2026 meeting showing a non-zero cut probability. Sizing: one-for-one DV01 at $41.67 per basis point per pair; risks 2.5 bp to make a further 3.5. | | 2. Long the 20-year against the 30-year — close it. The thesis was right and the expression was wrong | | Expression: long the 20-year Treasury point against short the 30-year, DV01-matched, small. Mark: entered Wednesday with 20s30s at +2 bp; +3 bp Thursday; +2 bp Friday; +2 bp Monday. Flat on the day, +1 bp on the week, which at DV01-matched sizing is roughly nothing. The honest reading, which is the point of this entry: the underlying call was correct and it was correct emphatically. Treasury named the 10-to-20 and 20-to-30-year buckets; the 20-year has richened 9 bp in a week and was one of the two best tenors on the curve on Monday. A simple long at the 20-year point would have made 9 basis points. This trade made one, because the 30-year richened 8 bp alongside it — the buyback narrative moved the whole long end, not the named bucket against its neighbour. Relative value was the wrong container for a level view. Action: close the position and take the lesson rather than the loss. The 20s30s expression can only be re-struck on evidence that operations are actually concentrated — that is, on the 9 September execution detail — not on the announcement. Invalidation as written: 20s30s back through −2 bp; it never came close, which is precisely the complaint. | | 3. New — long equal-weight against short cap-weight U.S. large cap | | Expression: long an equal-weighted S&P 500 vehicle against short a cap-weighted S&P 500 vehicle, dollar-neutral, small. Thesis: Monday produced the cleanest cap-weight divergence of the month. 306 S&P 500 members rose and 185 fell — a 1.65-to-1 advance line — and the index closed down 0.28%. Six of eleven GICS sectors were higher, staples led at +1.4%, and the entire decline was carried by one industry: technology −1.71% on Finviz's grouping, with Seagate −6.51%, SanDisk −6.45%, Micron −5.83%, Western Digital −5.24%, AMD −3.49%, Marvell −3.27%, Nvidia −2.91% and Broadcom −2.63%. Nvidia alone traded 135.19m shares. The setup is that the index's concentration risk is now a two-way risk, and it has an event: Nvidia reports Wednesday after the close having fallen seven sessions running, the first such run since September 2022. If the print disappoints, the cap-weighted index carries the loss and the equal-weighted one does not. If the print is strong, the historical pattern is a narrow rally that this pair loses — which is why it is sized small and why the invalidation is tight. Catalyst: Nvidia 8/26 after the close; Broadcom 9/2 after the close; the PCE deflator 8/26 08:30, where a soft print helps the rate-sensitive breadth that led Monday. Invalidation: the cap-weighted index outperforming the equal-weighted by more than 2.5% over five sessions; or a session in which the S&P rises while decliners outnumber advancers, which would say the divergence has inverted; or SOX recovering 11,740 — Friday's close — which would mean the semiconductor drag is simply over. Sizing: small. This is a concentration hedge dressed as a relative-value trade, and it should be held for the event, not for the quarter. | | 4. Long memory and storage against short megacap platform technology — stopped out. Close it | | Expression, as struck Thursday: long an equal-weighted memory and storage basket against short an equal-weighted megacap platform basket (Apple, Amazon, Alphabet, Microsoft), dollar-neutral, small. Mark: the memory basket averaged roughly −6.0% on Monday — Seagate −6.51%, SanDisk −6.45%, Micron −5.83%, Western Digital −5.24% — against a platform basket at +0.86% (AAPL +0.32%, AMZN +1.33%, GOOGL +0.94%, MSFT +0.84%). That is −6.9 points in a single session and roughly −7.8 points from entry across two sessions. The written invalidation was "the megacap basket outperforming the memory basket by more than 4% over five sessions." It triggered on day two. Action: close the position. The honest reading: the trade was stopped by a headline that confirmed its thesis. Nvidia is raising server prices more than 15% specifically because memory costs have surged — that is pricing power at the supplier — and the supplier equities fell 5–6.5% on it while the platforms rose. Either the market is pricing the demand destruction that a 15% server price rise causes upstream, or a crowded cohort used a confirmed cost story as an exit. Both readings argue for standing aside until Nvidia on Wednesday rather than averaging into a stop. | | 5. Protection on the CCC cohort funded in IG — hold at a half; both legs paid | | Expression: long CCC-exposed credit protection (or short a levered-loan / CCC-heavy vehicle) against long IG cash. Mark: the structure paid on both sides for the first time. The protection leg gained — CCC +2 bp to 1,037 bp on 21 August, another monthly wide — and the funding leg gained too, with IG tightening 1 bp to 81. Over five sessions the ratio is what the trade exists to harvest: CCC +25 bp against IG +1 bp, with the CCC-minus-HY differential at 767 bp, +22 bp on the week and +163 bp on the year. The strategic case strengthened on Monday: JPMorgan's Kelsey Berro put September IG supply at $175bn–$250bn and argued the market absorbs it on record retail demand — which is exactly the configuration in which the index stays tight and the tail does not. Catalyst: the 25 August FRED update, carrying the first post-Monday credit spread; the NY Fed Corporate Bond Market Distress Index on 26 August at 10:00; the September calendar clearing. Invalidation, unchanged: the differential back through 735 bp, or IG widening beyond 85 bp — which would say the funding leg has become the risk. Sizing: a half, unchanged. No add before the CMDI print. | | 6. On-balance-sheet AI funding against off-balance-sheet AI funding — hold the half into Marvell's print, then decide | | Expression: long customer-funded suppliers against short self-funded ones, half size. Mark: Marvell −3.27% against Broadcom −2.63% is another 0.64-point loss, the fifth losing session in six. Invalidation test status: the written stop is "a customer-funded structure trading below its announcement price within ten sessions." Marvell at $229.29 remains roughly 6.4% above its pre-announcement level, down from about 10% on Friday — the stop is intact but the cushion has more than a third gone. Action: hold the half into Marvell's print on Thursday 27 August, which is the only thing that can settle whether a customer-funded structure earns a different multiple, and cut on any print that does not. Catalyst: Nvidia 8/26, Marvell 8/27, Broadcom 9/2, and the pricing terms of the Broadcom package. | | 7. Short the utility complex against the S&P 500 — cut to a half; the mechanism ran in reverse on day one | | Expression: short an equal-weighted large-cap utility basket against long the S&P 500, beta-adjusted, small. Mark: utilities +0.81% against an S&P −0.28% is a 1.09-point loss on the first session. The honest reading: the thesis was that the discount rate applied to a regulated 30-year cash flow goes up in either the hawkish branch or the fiscal branch. Monday produced a third branch the thesis did not contain — an official intervention that pushes the 30-year down 4 bp and the 20-year down 4 bp while the policy path is priced higher. In that branch the utility sector wins and the trade loses, and the sector duly flipped its 2026 return positive at +0.13%. Note also that real estate rose 0.54% as the corroborating long-duration confirmation, exactly mirroring Friday. Action: cut to a half rather than close, because the invalidation as written has not triggered: the 30-year is at 5.23%, well outside the 5.10% stop; the basket has outperformed by 1.09% against a 2% five-session limit; and the September hike probability is 41.4%, far above the 30% floor. Catalyst: PCE 8/26; Warsh 8/28; the 9 September operation, whose execution detail determines whether the long-end rally survives contact with reality. Sizing: a half. | | Prior closes, marked forward. The bank-complex short, closed Friday at a token loss, would have cost another 1.12 points on Monday — financials +0.84% against the S&P's −0.28% — so that close was correct for a second consecutive session. The healthcare-versus-semiconductors pair, closed Thursday at its written invalidation, would have paid again: healthcare −0.16% against SOX −2.70%, +2.54 points, on top of Friday's +1.76. That is now +4.3 points foregone across two sessions on a stop that was correct as written. It is being recorded, not re-entered. | | The vol note. VIX closed at 15.85, +4.76%, giving back 82% of Friday's decline, while MOVE rose 0.79% to 73.98 — the first matched-date pair in three editions, at a ratio of 4.67× against 4.57× on 20 August. Equity volatility is repricing roughly six times faster than rates volatility into a Wednesday that contains the PCE deflator at 8:30 and Nvidia after the close, followed by claims on Thursday and Chair Warsh's first Jackson Hole address on Friday morning — and then, eight days later, August payrolls on 4 September, the last labour print before the September FOMC. The market it is pricing that into has just seen its long end move 8–9 bp in a week on a fiscal-plumbing announcement, its terminal rate reach 4.085%, its September hike probability rise for eight straight sessions, and 306 of 500 index members rise underneath a red tape. A 15.85 handle asks for roughly a 1.0% daily move. The single largest position in the index reports in forty-eight hours having fallen seven days in a row. The argument from last week is unchanged and is now better funded: own gamma dated 26 August to 4 September, and finance it out of Tuesday — the calendar in front is empty and the calendar behind it is not. | | The crowded consensuses to stress-test, with the numbers. - "The buyback is working." On the day's evidence it is: the 20-year is 9 bp richer on the week and the 30-year 8 bp, 2s30s has flattened 13 bp to 99, and Treasury has not bought a single security — the first operation is 9 September. But the two most credentialled voices on the tape said the opposite on the same day. Stanley Druckenmiller, who employed Bessent, published a Wall Street Journal op-ed calling it a mistake — "Governments defending prices against fundamentals always lose" — and Mohamed El-Erian told CNBC that intervention "makes sense when you can identify either a market failure or an institutional trouble," and that neither is present. The stress test is mechanical, not rhetorical: the programme may be funded by drawing down the $1tn General Account and selling bills, and the repo market has already moved — SOFR +2 bp to 3.65%, level with IORB; tri-party and broad GC each +3 bp to 3.63% (§9). A term-premium suppression paid for with front-end supply is a transfer, not a reduction. The 9 September execution is where the claim becomes checkable.
- "The front end is anchored." It genuinely was on Monday — the 2-year and 3-year did not move a basis point while the long end rallied 4. But the strip disagrees with the anchor at every horizon: the September hike went to 41.9% on CME from 39.9%, October's +25 cell overtook hold at 46.2% against 42.4%, December cumulative hike is 74.9%, and the year-end 2027 modal cell flipped to +50 bp. Eight consecutive sessions of hawkish repricing, on a fortnight containing a 206,000 claims print and a Philadelphia Fed price series at a six-month low. Either the coupon curve is right that policy is done being interesting, or the strip is right that it is not. The PCE deflator on Wednesday is the first arbiter and payrolls on 4 September is the last one before the meeting.
- "The index tells you what the market did." On Monday it did not. 306 members rose, 185 fell, and the S&P closed down 0.28%. Six of eleven sectors were higher; staples led at +1.4%; utilities, real estate, financials and communication services all rose. The entire decline was technology at −1.71%, and inside that, four memory and storage names down 5.2% to 6.5%. Concentration risk is usually discussed as an upside phenomenon; Monday is what it looks like as a downside one, and Nvidia reports on Wednesday having declined seven sessions running with 135.19m shares traded on the eighth-worst of them. The number to watch is not the index level; it is whether the advance line holds through the print.
- "Memory has pricing power." Nvidia's own actions say it does — servers with Vera Rubin and Grace Blackwell chips are going up more than 15% on early-2027 shipments because memory costs surged, and Wedbush called it bullish. The tape said otherwise by 5.2% to 6.5% across four names on the same morning, and Korea said it first with a 3.12% Kospi decline. One of those two readings is wrong and the resolution is scheduled. If the market is right, the AI capex line has a cost problem that shows up first at the hyperscalers and second in the $370bn of AI-related debt BofA has flagged. If Ives is right, Monday was the best entry in the complex this month.
- The two-sided geopolitical tape, now with content. Bessent unveiled Operation Economic Outcast — countries retaining ties to Tehran to be "removed from the US dollar system," with no exemption signalled for China — and crude fell 1.9%. The composition is the risk. Heating oil fell 4.37% and the distillate crack collapsed $6.55 to $94.23 a barrel, which is the market pricing enforcement risk to buyers rather than supply risk to the barrel. That is a coherent position and it is a fragile one: an actual interruption of Iranian exports, or a Chinese refiner cut off from clearing, revalues the fattest crack in the complex from the wrong side. Separately, the Canada breakdown — 50% tariffs on all cars, trucks, parts and steel from 1 January 2027, retaliation from 8 September — moved USD/CAD 0.73% and cost J.B. Hunt 5.65% on a macro impact BofA calls "modest." Confidence, not arithmetic, is the transmission, and UBS now sees USMCA slipping into 2027.
- "Credit is fine." The index is: IG tightened 1 bp to 81 and HY 5 bp to 270 on a session the coupon curve cheapened 4–5 bp, and JPMorgan says a $175bn–$250bn September calendar clears on retail demand at a sixteen-year record. The tail is not. CCC widened 25 bp in a week to 1,037 and is +152 bp on the year while HY is 11 bp tighter than it started 2026; the differential at 767 bp is +163 bp year to date and has widened in five of the last six published sessions, twice on days the index tightened. A market that will fund a record calendar at 81 bp and will not fund the CCC cohort at any price is not benign; it is bifurcated. CDX IG 5y and CDX HY 5y remain unobtainable after a fourth full working of the six-step ladder, which is itself a transparency risk worth naming.
- Structural watch items. The AI capital cycle is now visibly raising its own costs — Nvidia's 15% server increase is memory inflation passed forward — while Alibaba raised $10.2bn at an 8.4% discount purely for AI infrastructure and SoftBank came for ¥1tn at up to 4.90%. Applied Optoelectronics fell 11% on a $600m at-the-market equity programme, the small-cap version of the same funding pressure. The bill curve is now pricing the Twist — 1.5-month and 2-month 2 bp richer on the week against a 4-month and 6-month 1 bp cheaper — with reserves $207.4bn below their July peak and the reverse repo facility holding $380m. And America rallied alone: the U.S. 10-year richened 4 bp while Germany, the UK, France, Spain, the Netherlands, Portugal and Greece were all unchanged, the mirror image of three sessions of cheapening alone. The duration problem is priced as domestic in both directions, which means there is no external bid to rely on if the 9 September operation disappoints.
- What VIX is and is not pricing. At 15.85 it is pricing a roughly 1.0% daily move. It is not pricing: the PCE deflator and Nvidia arriving within twelve hours of each other on Wednesday; a new Chair's first Jackson Hole address on Friday; August payrolls on 4 September, the last labour print before the FOMC; a 9 September buyback operation that two of the most-followed macro investors alive have publicly called a mistake; a repo market that moved 3 bp on tri-party before the programme has bought anything; a 50% tariff regime with Canada taking effect in a retaliation sequence that starts on 8 September; or the possibility that a 1.65-to-1 advance line under a red index resolves downward rather than upward. There is one Medium-rated release between now and Wednesday's 8:30. The optionality is cheap because of the shape of the calendar, not because of the level of the market.
|
| Sources used this session: CNBC (live blog, front-contract quote boards, GICS sector tally), Bloomberg (/markets quote board, /markets/rates-bonds global yield boards, Markets Wrap, high-grade issuance coverage), The Wall Street Journal (Market Data bonds page; Druckenmiller op-ed), U.S. Department of the Treasury (official daily par yield curve, Text View), CME Group FedWatch, Investing.com (Fed Rate Monitor, index boards, S&P 500 component board, dollar index, MOVE), Finviz (sector group screener, Performance table view), TradingEconomics (commodities and currencies boards), Federal Reserve Bank of New York (Economic Indicators Calendar; reference-rates API), FRED (ICE BofA OAS series, RRPONTSYD, WRESBAL), Nasdaq earnings calendar API, StockAnalysis.com (ETF closes), Reuters and Yahoo Finance wraps. | | Full Source Links, the Overnight / Asia & Europe read-through and the complete Data Notes & Conflicts appendix are in the companion file US_CrossAsset_Daily_2026-08-24_DataNotes.txt and in the canonical Markdown report US_CrossAsset_Daily_2026-08-24.md. | |
|
U.S. Stock, Fixed Income & Cross-Asset Closing Daily — Monday, August 24, 2026. Prepared for institutional investors. Not personalized investment advice; verify independently before acting. All figures sourced as stated in-text; vendor discrepancies are reconciled in the companion Data Notes file.
|
|