Hot CPI, falling long yields, and a Chair who won't guide: the curve is now a positioning trade

0. Weekly Arc

The week inverted the sequence’s usual logic. Through Oct 12–14 the tape was owned by term premium and supply; September CPI (Oct 14) and a four-and-a-half-year-high PPI (Oct 15) then landed hot, yet the 10-year still finished the week near 5.098% in a bull flattening. Officials never validated the move — Warsh declined to pre-judge at the IMF (Oct 16), Williams saw no need to act consecutively — and October hike odds slid from a 31.6% intraweek peak to roughly 20%. By Oct 17–18 the debate had rotated to positioning, the Nov 3 midterms and the fiscal slope beyond them.

1. Policy Narrative & Expectations

The net change over the past ~24h is that the market’s story about the long end has decoupled from its story about inflation. Money markets now put the probability of a 25bp hike on Oct 28 at only 15%, with December fully priced and three 25bp hikes in total expected over the next 12 months ^[1], while a parallel read has October odds briefly touching 31.6% before settling near 20% ^[2] — a distribution that has barely moved despite an above-expectation CPI and a PPI print described as the largest in four and a half years ^[1]. The reconciliation is not dovish data; it is that policy communication pinned the short end while technicals and positioning did the rest: Guosen Securities attributes the week’s ~15bp decline in the 10-year to 5.098% to four drivers — safe-haven demand, communication that suppresses the short end, another Treasury buyback that lightened the private sector’s duration burden, and underappreciated positioning and convexity effects ^[2]. On timing, the published base case is unchanged at no move in October and one hike in December, with the Oct 28 statement and dot plot and the Nov 3 midterms named as the two observation points ^[2]. The marginal new input is political rather than monetary: GF Securities frames the midterms as a repricing of the next two years’ policy slope, with the base case (Democrats take the House, ~61.8%) cooling fiscal-expansion expectations and pushing long-end term premium and real rates lower, and an ~8.8% Republican sweep as the reverse tail that lifts growth, reflation and long-end yields and raises the odds of future Fed hikes ^[3].

1.1 FOMC Officials’ Remarks

• [ESCALATED] Chair Kevin Warsh: he expressed optimism about US economic strength this week but did not disclose whether he thinks further hikes are needed or when they might come ^[1], and made no forward judgment at the IMF meeting ^[2]. The escalation is structural rather than tonal: Guojin Securities reads the Chair as weakening all three forms of forward guidance at once — no action timing, no data thresholds, and no disclosure of the specific data the Fed will rely on ^[4] — while SWHY Strategy describes the framework as turning conservative with no clear guidance yet for the AI era ^[5]. Marginal difference vs prior briefings: the refusal to guide is now being described as an operating framework rather than a single press-conference choice.
• [ONGOING] Neutral/swing — New York Fed President John Williams: he is in no hurry to act consecutively ^[2].

1.2 Policy Signals & Institutional Communication

• [NEW] GF Securities prices the midterms as a rates trade: in the base case of a Democratic House (~61.8%), cooling fiscal-expansion expectations lower the risk of further deficit and Treasury-supply increases, leaving long-end term premium and real rates biased lower and favouring Treasuries and gold at the expense of the dollar and copper; in the ~8.8% tail where Republicans hold both chambers, the direct result is long-end Treasury pressure and a stronger dollar, with consumer, financial, energy and AI-infrastructure chains relatively favoured, and gold still has upside because the yield rise would come from deficit and term premium rather than real growth ^[3] (projections).
• [NEW] The fiscal impulse behind that tail is quantified and unlegislated: Trump has proposed paying about $5,000 to roughly 240 million adult citizens if Republicans win the midterms — roughly $1.2 trillion, close to 4% of GDP — with no congressional approval and no specified funding source; GF Securities argues this, combined with tighter immigration lifting wage and services inflation from the supply side, raises the probability of future Fed hikes and of real rates staying elevated ^[3].
• [NEW] The anti-guidance doctrine is being institutionalised: the Chair’s working group has recruited former Bank of England Governor Mervyn King to lead improvements in communication style, and the Fed may adopt the BOE’s deliberately lower-information template, whose inflation fan chart widens so fast that the far end spans deflation to marked high inflation ^[4].
• [NEW] Guosen Securities’ map of the next two tests: the October probability holding 20% is the key variable — a rebound to 30–40% would force this week’s bull flattening to be re-read as a liquidity trade — with the Oct 28 statement and dot plot first and the Nov 3 midterms second; its base case remains no move in October and one hike in December ^[2] (projections).
• [NEW] Buybacks are pace management, not a fiscal pivot: the Treasury bought back bonds again, easing the private sector’s duration-absorption burden, but Guosen explicitly warns this should not be misread as a fiscal turn ^[2].
• [ONGOING] December is fully priced with three 25bp hikes expected over the next 12 months ^[1].

2. Key Data & Market Read

• [ONGOING] September CPI: above expectations and higher than the prior reading on a year-on-year basis, with the economy still described as strong ^[3][1].
• [ONGOING] September PPI: the largest increase in four and a half years, read as the latest evidence that inflation has broadened beyond energy products, with core PPI also above its prior reading ^[3][1].
• [ONGOING] Initial jobless claims: essentially unchanged from the prior week at a low level ^[3].
• [NEW] Narrative impact — hot data failed to lift yields: Guosen Securities states plainly that this week’s inflation data should have pushed rates higher and did not, with the 10-year falling about 15bp from above 5.2% to 5.098% as 30-year and 10-year yields declined together in a bull flattening, and October hike odds peaking at 31.6% before returning to roughly 20% ^[2]. El-Erian makes the same observation from the tape — bond yields fell despite hotter inflation, equities showed sharp volatility and dispersion with a tech selloff and accelerating rotation, and oil and geopolitical pressures stayed in play ^[6] (single source / unverified). The market read is that this was a growth-and-haven trade, not a data-driven repricing.
• [NEW] Cross-asset week: US equities fell across the board — the Dow down 0.67%, the Nasdaq down 4.48% and the S&P 500 down 1.95% to 7,659.44, with the Nikkei 225 down 5.37% to 65,324.29 — while Europe’s three main indices rebounded, led by the FTSE 100 up 3.13%, with the DAX up 2.13% and the CAC 40 up 1.97% ^[3][1]. In China, the CSI 300 fell 3.4% and the Hang Seng 0.11% over the same window ^[2].
• [NEW] ING expects the yield decline to be rebuilt against: rates strategist Patrick Garvey said participants will note another good claims reading, strong retail sales and uncomfortably high PPI inflation and may use that as a platform to rebuild higher yields in coming weeks ^[1] (projection).
• [NEW] Next week’s calendar: October PMIs for the US and Europe on Oct 23, plus weekly jobless claims and the final October University of Michigan sentiment index, with the PMIs framed as the week’s focal point for FX and bonds and as a guide to the timing and pace of further hikes ^[1][7].

3. Financial-Conditions Signals

• [NEW] The four-driver decomposition of a rally against hot data: Guosen Securities attributes the move to safe-haven demand, policy communication that pinned the short end, the Treasury buyback, and positioning/convexity effects it considers the most underestimated of the four ^[2].
• [NEW] Positioning is the fragile leg: bond-fund duration is already significantly underweight by the report’s estimate, implying yields biased lower over the next three to six months, while falling mortgage prepayment risk and the cheapest-to-deliver switch force hedgers to sell duration into rallies — roughly $60bn of 10-year equivalent from that channel alone plus another $10–15bn from asset managers; trend-following funds’ duration shorts are likewise crowded and would only be covered with the 10-year at 4.85%–4.95% ^[2].
• [NEW] Dollar: the dollar index fell 0.44% on the week while offshore CNY appreciated 0.11% ^[2].
• [NEW] The fiscal layer under the term premium: US debt has risen from 108.8% to 125.8% of GDP since 2019 while net interest has climbed from 7.6% to 12.6% of fiscal revenue, with the primary deficit still near 3.5% of GDP — a $40 trillion federal debt stock and roughly $1 trillion of annual net interest ^[8]. Wallstreetcn’s Song Xuetao argues the market treats 10-year yields above 5% and 30-year yields at 5.5% as an alarm precisely because developed-economy rates transmit downward while emerging-market rates do not, and because the financing gap and duration absorption now sit in the core economies rather than the periphery ^[8].
• [NEW] Developed-market balance-sheet risk has repeatedly been the stress origin: the 2022 UK pension crisis, the US regional bank problem and France’s fiscal squeeze are cited as a sequence in which high-grade sovereign debt, not weak credit, started the liquidity shock ^[8].
• [NEW] The AI-financing channel is competing for the same capital: concerns are growing about the sustainability of debt-financed AI investment just as long-end yields retreat from multi-year highs ^[1], AI-chain bond issuance is diverting global allocation flows ^[5], and the pricing anchor for tech equities has shifted to earnings and financing conditions ^[2].
• [NEW] Energy is the live inflation channel: WTI’s front month rose 10.41% to $101.41 and Brent 4.91% to $109.86 on the week, with Hormuz daily flows falling from a 17.15 million bpd baseline to 11.62 million bpd in the week to Oct 13 and US crude inventories down 8.56 million barrels in the week to Oct 9, raising freight costs and risk premia for shippers ^[1].
• [NEW] Precious metals split: COMEX October gold fell 0.75% on the week to $4,159.50/oz and COMEX silver fell 6.06% to $56.99/oz ^[1], while spot gold was up only 0.07% and spot silver down 4.15% on Guosen’s measure ^[2] — with ANZ arguing that a pullback in yields and the dollar paves the way for gold above $5,000/oz in 2027 ^[1] (projection).

4. Global Central-Bank Linkages

• [NEW] ECB: Christine Lagarde said at the IMF/World Bank meetings that the high euro-area inflation caused by elevated energy prices has not yet affected the overall economy, with officials still working out how far indirect effects will transmit ^[1], and separately that the bond selloff will slow inflation ^[9] (single source / unverified).
• [ESCALATED] BOE: UK inflation is expected to rise to a nine-month high, with economists looking for 3.4% annual price growth in September, setting the stage for the budget and a likely rate increase ^[10]; money markets put a 93% probability on a 25bp hike at the Nov 5 meeting and fully price at least four 25bp hikes over the coming year ^[1]. Investec’s Sandra Horsfield said MPC members appear to be converging on some tightening but less than expected, and that CPI and wage forecasts close to target could strengthen their conviction at the next meeting ^[1].
• [NEW] Others: the central banks of South Korea, Indonesia and the Philippines meet next week with currency moves and higher energy prices complicating their decisions ^[1]; SWHY Strategy flags Japanese economic fragility and warns that policy intervention would generate Treasury selling pressure ^[5], and notes that the US and Japan conducted their first joint FX intervention in 15 years even as long-end Treasuries were themselves under pressure ^[4].

5. Asset Implications

Quadrant	Current probability tilt	Key asset implication	Anchoring narrative
Growth↑ + Inflation↑	Rising	Commodities↑ TIPS↑ nominal long bonds↓ — CPI and PPI both landed hot with PPI the largest in four and a half years, WTI gained over 10% on Hormuz disruption, and the fiscal tail (a ~$1.2trn cash handout plus tighter immigration) is explicitly read as raising the odds of further hikes and higher real rates	§1.2 / §2
Growth↑ + Inflation↓	Falling	Stocks↑ long bonds↑ gold↓ — the long-end rally was not disinflation: the four named drivers are haven demand, communication, buybacks and positioning, and ING expects yields rebuilt higher on strong retail sales and high PPI; equities fell rather than rose alongside bonds	§2 / §3
Growth↓ + Inflation↑	Rising	Commodities↑ gold↑ stocks↓ — energy supply loss, a $40trn debt stock with net interest at 12.6% of revenue, developed-market sovereign debt as the recurring origin of liquidity shocks, and an equity tape where the Nasdaq fell 4.48% and the Nikkei 5.37% while oil rose	§3 / §6
Growth↓ + Inflation↓	Unchanged at a low probability	Long bonds↑↑ stocks↓ credit spreads↑ — this quadrant owns this week’s price action but not its mechanism: the 10-year fell ~15bp on haven flows and crowded-short covering rather than on cooling activity, and the covering level of 4.85%–4.95% caps the runway	§3 / §1.2

Stock-bond correlation call: this week printed the negative, hedge-paying correlation — hot inflation, a ~15bp decline in the 10-year, and an S&P 500 down nearly 2% with the Nasdaq down 4.5%. But the mechanism is the tell, and the mechanism was flow, not regime. All four identified drivers — safe-haven demand, short-end pinning through communication, Treasury buybacks and convexity/positioning — are technical; none is disinflation. The structural regime remains inflation-, supply- and term-premium-driven positive: the fiscal arithmetic is deteriorating on every line, the energy shock is live, and the Chair has deliberately removed the guidance that would let the market price the path. Directional call: positive correlation is still the base case, and the negative episode lasts only as long as crowded duration shorts need covering and no supply or policy shock arrives. Two verification dates now exist: the Oct 28 statement and dot plot, and the Nov 3 midterms — the first tests whether the Fed re-anchors the short end, the second whether the fiscal tail is priced out.

Risk-budget implication: the bond sleeve’s gain this week was a positioning dividend, not a validation of duration as a hedge, so it should not be capitalised into a larger long-end risk budget. Keep duration expressed in the belly and the curve rather than as a long-end level bet, because the rally’s own fuel — short covering toward 4.85%–4.95% plus hedging flows that sell into strength — is self-extinguishing, and because the buyback that helped is explicitly described as pace management. Treat gold as the fiscal-credibility and reserve-diversification sleeve rather than a rates hedge: official-sector diversification is named as a structural flow, ANZ sees a path above $5,000, and the metal’s weekly softness is a real-rate artefact rather than a thesis break. Keep energy at or above benchmark but book it as a directional inflation position: Hormuz flow loss and inventory draws are simultaneously the CPI story, the term-premium story and the reason December stays priced. In equities, treat index beta as an AI-earnings-plus-financing-conditions bet rather than a broad cyclical bet — the pricing anchor has moved to earnings and financing, and the rotation into financials and industrials belongs only to the ~8.8% fiscal tail. In credit, stay underweight the AI-adjacent issuance that is diverting global allocation and whose debt-financed sustainability is now a named concern.

6. Contrarian & Tail Risks

• Consensus fragility: the baseline is “October is off at 15–20%, December is fully priced, and hot inflation no longer lifts the long end.” Five assumptions are falsifiable. First, the same week carries two different October probabilities — 15% and roughly 20% — and an intraweek peak of 31.6%, three statistics from different vintages that must not be traded as one. Second, the negative stock-bond print rests on a positioning argument the same research shop caps: trend-follower shorts are only covered at a 10-year of 4.85%–4.95%, so the trade is a range, not a regime. Third, the opposite read is published in the same batch — ING expects good claims, strong retail sales and uncomfortably high PPI to become a platform for rebuilding higher yields within weeks. Fourth, the midterm base case is a probability (~61.8%), not an outcome, and its mirror image (~8.8%) is the scenario that lifts real rates, the dollar and hike odds. Fifth, the fiscal impulse underlying that tail is neither legislated nor funded — roughly $1.2trn, close to 4% of GDP — so its inflation content is a scenario weight, not a forecast.
• Second-order transmission: five chains are not in the price. The energy-to-breakeven chain is the most immediate: Hormuz flows fell by roughly a third from baseline and crude inventories drew heavily while the 10-year fell, an inconsistency that resolves either through higher inflation compensation or through demand destruction. The fiscal-dominance chain is the most structural: debt at 125.8% of GDP, net interest at 12.6% of revenue and a primary deficit near 3.5% mean the term premium has a floor no policy rate sets. The developed-market-sovereign chain is the most under-priced: the UK pension crisis, US regional banks and French fiscal strain form a sequence in which the highest-quality collateral, not the weakest credit, starts the liquidity shock. The AI-financing chain is the most reflexive: debt-funded AI capex sustainability is now a named concern, AI-chain issuance is diverting global allocation, and the same complex anchors index earnings. The Japan-and-geopolitics chain is the most dated: Japanese fragility plus intervention implies Treasury selling pressure, a joint US-Japan FX intervention already occurred with the long end under strain, and the US President has said the US may strike Iran’s nuclear facilities again, including the Fordow underground site — an oil-supply tail with no calendar.
• Source quality control: volume is not confirmation. Lagarde’s “bond selloff will slow inflation” line and El-Erian’s weekly framing both arrive as single-source social posts and are carried here as colour, not as reporting; the Bloomberg charting item is a headline and a link with no extractable facts, and the Reuters and WSJ week-ahead items are headline relays naming PMIs as the focus. Two large blocks are dated framework pieces rather than today’s news — Guojin’s analysis of the Chair’s communication doctrine is from August and SWHY Strategy’s mid-cycle framing from September — and are used for structure, not as fresh signals. The two Guosheng Securities notes are A-share timing-radar models with no direct Fed content and an explicit model-failure disclosure; they are excluded from the Fed narrative. Direct conflicts read side by side rather than averaged: October odds of 15% against roughly 20%; gold down 0.75% on COMEX October futures at $4,159.50 against spot gold up 0.07%, which are a contract and a venue difference; and the 10-year at 5.098% alongside a 30-year “5.5%” cited as a stress threshold rather than a closing level. Finally, GF Securities’ midterm probabilities and the handout arithmetic, ANZ’s gold target, and Guosen’s positioning estimates are model- or scenario-derived and frame a distribution rather than a price.

---

Appendix: Additional Sources

• ^[11] Christophe Barraud — Bloomberg link: “Charting the Global Economy: Inflationary Pressures Mount in US” (headline only, no substance)
• ^[12] Christophe Barraud — Reuters week-ahead relay: AI rally wobbles, market spotlight turns to earnings
• ^[13] Guosheng Securities Financial Engineering — A-share six-dimension timing radar; no direct Fed or US macro content
• ^[14] Guosheng Securities Financial Engineering — dated 09/12 A-share radar, model-based; excluded from the Fed narrative
• ^[5] SWHY Strategy — dated 09/2026 mid-cycle framework: weak US fiscal discipline, reserve diversification, AI-chain debt financing

---

This report is a macro-mechanism analysis, not investment advice.