Super central bank week; hike pricing at 87–90% with the dot plot as the swing factor

0. Weekly Arc

The past week completed the hawkish consolidation that began with Chair Warsh’s Aug 28 Jackson Hole keynote. The Sept 4 payrolls beat re-leaned September pricing toward ~60%; the Sept 10 PPI and Sept 11 core CPI broke the market through, lifting hike odds from ~70% to an 85–90% band and pulling the 10-year toward 5% and the 30-year to ~5.35%. The ECB hiked on Sept 10 and signalled more. The live question has migrated from whether the Fed hikes to what the dot plot and Warsh’s press conference imply about a second move.

1. Policy Narrative & Expectations

The past ~24h brought no fresh repricing impulse but a decisive shift in what the market is trading. September hike pricing has consolidated rather than extended: 87.3% on CME FedWatch ^[1][2][3], 88% per Soochow’s read ^[4], 90% per BOC International Securities ^[5] and a >90% reading via CME after the CPI print ^[6], with Refinitiv-linked money-market pricing at 82.5% ^[7], OIS swaps at 90% ^[8], and both El-Erian and Guosen Securities citing a >80% band ^[9][10]. Because the first hike in three years is now treated as “near-certain” ^[11], the marginal uncertainty has moved to three places: whether the simultaneously released dot plot and economic projections hint at more than one hike this year; the wording of Warsh’s press conference; and whether the move is framed as preventive calibration or as the opening of a consecutive-hike cycle ^[1][5][4]. That is a materially different question set from yesterday’s “how hot was the CPI,” and it is the one that will drive the term premium.

1.1 FOMC Officials’ Remarks

No sitting FOMC official made a fresh monetary-policy remark in the past 24h; the batch consists of restatements of the existing anchor.

• [ONGOING] Hawkish: Chair Kevin Warsh — the Jackson Hole frame still governs: “if we cannot become confident fast enough that underlying inflation is moving toward target,” the Fed “will have work to do” ^[11]; the 2% PCE target remains the anchor and policy adjustment is contingent on underlying inflation returning to 2% “clearly and at a sufficient speed” ^{[1][8][3][12]}, with the Fed prepared to act through hikes if necessary ^[7]. SWS Research adds that Warsh has not yet given clear guidance on monetary policy in the AI era and that the framework is currently turning conservative ^[13].

1.2 Policy Signals & Institutional Communication

• [NEW] The calendar itself is now the event: the FOMC decision lands Sept 16 local time (early hours of Sept 17 Beijing time) with Chair Warsh’s press conference at 02:30 Beijing on Sept 17, the BOE decides at 19:00 Beijing on Sept 17, and the BOJ concludes its September meeting on Sept 18 with Governor Ueda’s press conference at 14:30 ^[14][1][7]. Gelonghui frames the US, Japan and UK meeting in succession as a “stress test” for global assets ^[14] and notes G7 major central banks rarely stood at the threshold of the same tightening round ^[1].
• [NEW] Bessent testimony teed up: Treasury Secretary Bessent is expected to appear before the House Financial Services Committee on Sept 15 ^[14].
• [ESCALATED] The “one-off hike” question is now the market’s central uncertainty: WSJ’s Nick Timiraos wrote that investors have basically concluded the Fed will hike next week — the first hike in three years — and that the harder question is what happens afterward, noting that since the 1990s the Fed has delivered only one “one-off” hike ^[15]. Gelonghui similarly frames the biggest remaining uncertainty as “how hawkish” the post-decision messaging is ^[1].
• [NEW] Credibility trap flagged explicitly: Bloomberg economists Anna Wong, Andrew Sacher and Eliza Winger wrote that market signals are unambiguous — investors want and expect the FOMC to hike — and that a failure to hike would cost Warsh credibility in the eyes of market participants ^[11]. The July FOMC already saw three officials dissent in favour of a hike ^[11].
• [NEW] The two-path framework: BOC International Securities argues Fed hikes carry strong policy inertia — the 2022–24 cycle delivered 11 hikes in total and kept going even with the policy rate at 4% — and lays out two September paths: a lone 25bp preventive move with a data-dependent statement and no sharp dot-plot upgrade, versus a hike paired with an upward-shifted dot-plot median and hawkish tone that the market would read as the start of a new consecutive cycle with total hikes potentially reaching 3 ^[5].
• [ONGOING] Dot-plot expectations: Soochow’s base case is a unanimous 25bp September hike with the median dot guiding one more 25bp within the year ^[4].
• [NEW] A 1988–89 analogue is being introduced: Citi Research’s quant macro strategy report argues the current macro environment increasingly resembles the 1988–89 tightening cycle — rising hike fears, resilient growth and re-accelerating inflation — though its KNN model remains in the “Normal” regime rather than the “financial conditions tightening” regime ^[15]; the same note relays Tianfeng Securities’ reconstruction of that cycle, in which the FOMC raised the funds rate 25bp to 6.75% on 03/30/1988 and hiked a total of 16 times to 9.8125%, a cumulative 331.25bp ^[15].
• [ONGOING] Structural rotation: the next FOMC is Oct 28, and hawkish voting members rotate out later, which combined with the hike being priced should ease next year’s hike worries ^[2].
• [ONGOING] Treasury buyback arithmetic: the Treasury raised the size of its long-bond buybacks from $2bn per operation to at least $6bn ^[7], but the Bessent-led intervention came in under $6bn, below market expectations and providing only limited support to long bonds ^[8]; SWS Research argues a Bessent-style twist operation is constrained by fiscal discipline, making a below-expectation buyback entirely natural ^[13].

2. Key Data & Market Read

• [ONGOING] August CPI (released 9/11): headline matched expectations at +0.4% m/m and +3.4% y/y, while core rose 0.3% m/m above the 0.2% consensus — the largest monthly gain since April ^{[2][4][3][12]}.
• [NEW] Composition is where today’s batch adds value: core CPI m/m was +0.29% versus +0.2% expected, the first core beat since January 2025 once the April 2026 statistical distortion is excluded ^[4]; super core (including hotel accommodation within housing services) rebounded to +0.51% m/m from +0.19% ^[4]; ex-housing services accelerated to +0.54% m/m with the y/y rate rising to 3%, driven mainly by airfares and wireless communication services ^[3]. The headline rebound was energy: energy goods CPI rose 4.13% m/m ^[4], gasoline rebounded 3.9% m/m after a near-3% July decline ^[12], and fuel prices rose 4.3% m/m and 28% y/y, lifting the energy subindex to +2.1% m/m ^[3]. Offsetting: shelter cooled, with OER measures falling below 0.2% m/m ^[4] and housing CPI at +0.3% m/m with the y/y rate easing to 3% ^[3].
• [NEW] Stripping out the noise changes the read: Soochow calculates that excluding high-volatility and technical distortions, core CPI m/m would fall back to +0.09%, with the shelter cooldown the trend contribution ^[4]; the same report shows inflation breadth continuing to narrow, with the share of items rising faster than their past-5-year average falling to a record-low 17.6% and the share rising more than 4% the lowest since May 2020 ^[4]. Chen Li makes the same point from the contrarian side, arguing the wireless-services spike was an unexpected disturbance and underlying core inflation was roughly 0.2% ex-it ^[16].
• [ESCALATED] The PCE conversion has turned more hawkish: a Reuters survey of economists now expects August core PCE m/m to accelerate by about 0.1pp from July to 0.3%, implying the underlying disinflation trend has stalled ^[3]. Working the other way, BEA’s Sept 30 annual PCE recalibration is widely expected to revise core PCE y/y down by 0.1–0.3pp ^[4].
• [ONGOING] August PPI (released 9/10): +5.4% y/y, above the 5.3% expected and accelerating from 4.7%, reinforcing the case for a hike next week ^[2].
• [ONGOING] August nonfarm payrolls (released 9/4): +162k against roughly 53k expected, with June and July revised up a combined 55k and the unemployment rate steady at 4.1% ^[16].
• [NEW] Next week’s US data is Fed-light and consumer-heavy: August retail sales on Sept 16 (20:30 Beijing) with attention on consumer resilience, initial jobless claims for the week of Sept 12 on Sept 17, plus housing starts and industrial production ^[14][11][7]; Japan’s August national CPI lands Sept 18 with y/y expected at 2%, and UK August headline inflation is expected at 3.1%, a five-month high ^[11].
• [NEW] Narrative impact: the data does not decisively confirm either path. The hawkish reading — core above expectations, a stalled PCE conversion, an oil impulse still feeding through — justifies the Fed delivering ^[4][3][12]. The dovish reading — a single, record-low breadth print with core ex-noise near 0.09% and a mechanical PCE downgrade two weeks later — means the Fed could hike while privately treating the move as a credibility repair rather than the start of a cycle ^[4]. That ambiguity is precisely why the dot plot carries more weight than the decision this time.

3. Financial-Conditions Signals

• [ONGOING] Rates: the 10-year ended flat at 4.97% and the 2-year rose 4bp to 4.62% after the CPI print ^[4]; as of Sept 11 the 2/5/10/30-year curve sat near 4.63%/4.78%/4.96%/5.35% ^[16]; WSJ reports that “U.S. Treasury yields ended the week higher, with the 10-year yield almost reaching 5% before next week’s Federal Reserve meeting” ^[6].
• [NEW] The long-end driver decomposition has shifted to fiscal/supply/policy-uncertainty: Soochow attributes the rapid rise in US long-bond yields not to monetary policy alone but to three pressures jointly lifting the term premium — fiscal concerns, supply-demand mismatch and Warsh policy uncertainty ^[8]; SWS Research adds that US fiscal discipline is weak and that the mid-cycle problem behind high Treasury yields currently has no path to resolution ^[13]. The counterweight is that high yields are starting to attract allocation money and resistance to further long-end increases is building ^[16].
• [ONGOING] Financial conditions: Citi Research’s gauge has tightened slightly but remains about 0.55 standard deviations below its long-term average ^[15].
• [ONGOING] Auctions and buybacks: this week’s poor auction results were part of what pushed Treasury yields higher, with next week’s calendar carrying $13bn of 20-year notes and $19bn of 10-year TIPS ^[7].
• [NEW] Dollar and positioning: the dollar index oscillated at high levels on the week ^[7], and Citi’s model now names the dollar as its top currency pick, with the expected Sharpe ratios of the pound, yen and euro against the dollar all negative and market enthusiasm for the yen visibly faded ^[15].
• [ESCALATED] Oil moved the whole week: WTI front-month rose 9.37% on the week to $100.05/bbl and Brent 8.65% to $104.61 ^[7]; Brent reached a year-high $109 and at one point approached $110 ^[2][8]; Bloomberg’s framing is that energy markets are signalling a winter crisis and rising interest rates ^[17]. Offsetting news: a temporary navigation arrangement in the Strait of Hormuz pushed oil notably lower and helped restore risk appetite ^[3], and the IEA cut its 2026 global oil demand forecast, now expecting demand to fall 2.5mn bpd over the year versus 1.6mn bpd last month ^[7].
• [NEW] Gold: COMEX September gold fell 1.44% on the week to $4,366.20/oz and silver 1.26% to $64.55/oz ^[7]; real rates remain the traditional headwind, but long-end Treasury supply and the large US fiscal deficit continue to provide bottom support ^[7]; Industrial Securities reads gold’s post-CPI rise as more a correction of the prior day’s PPI-driven overshoot than a new defensive bid ^[3].
• [NEW] Equities and flows: US indices ended the week lower — Dow -1.57%, Nasdaq -0.66%, S&P 500 -0.80% ^[7] — yet Friday’s session gapped up and traded strong, with the Nasdaq +0.96% and the Philadelphia Semiconductor Index +1.81% ^[2]; Asian markets fell collectively on Friday after the PPI release, with the Nikkei -1.93% and KOSPI -1.76% ^[2].
• [NEW] Reserve diversification and allocation competition: global central banks are diversifying reserve assets into gold and US equities, and AI-chain debt financing is diverting global allocation funds ^[13]; Japan’s economic fragility means policy intervention would generate US Treasury selling pressure ^[13].
• [NEW] Copper whipsaw: expectations of Trump tariffs on refined copper pushed copper to a record high, and rumours that the White House tariff plan had stalled triggered a plunge ^[2].

4. Global Central-Bank Linkages

• [ESCALATED] ECB: the Sept 10 meeting raised the deposit facility, main refinancing and marginal lending rates to 2.50%, 2.65% and 2.90% respectively, in line with expectations, with Lagarde saying the decision was unanimous ^[18]. Lagarde added that the neutral-rate range is hard to define accurately amid external disturbances and is not a strong constraint on the hiking endpoint, leaving further hikes open ^[18]; the ECB’s own projections put 2026–28 headline inflation at 3.0%/2.5%/2.1%, core at 2.5%/2.6%/2.3% and real GDP growth at 0.9%/1.4%/1.5% ^[18]. Market-implied odds of an October hike sit at 61%, with roughly 49% attached to two more hikes within the year ^[18]; Refinitiv showed money markets pricing a 65% probability of an October hike with a hike fully priced by year-end ^[7], and BNP Paribas now expects one more hike in December with rates held at a mildly restrictive level until 2027 ^[7].
• [ESCALATED] BOJ: the Bank is expected to hike 25bp on Sept 18, taking the policy rate from 1.0% to 1.25% — the highest since 1995 and, at three months after June, the shortest interval since the hiking cycle began in March 2024 ^[1][11]. A Reuters survey shows nearly all surveyed economists expect the hike and over 90% expect another before January ^[1]. Three pressures underpin it: potential inflation indicators above 2%, corporate goods prices up over 7% for three consecutive months, and a weak yen plus higher oil pushing up import costs ^[1]. The stated core question is not excessive inflation but whether inflation and wage growth form a virtuous cycle, and whether normalization again triggers an unwind of global carry trades ^[1].
• [ONGOING] BOE: the Bank is expected to hold at 3.75% on Sept 17 with attention on the vote split; UBS expects the 6:3 split to persist with three members voting to hike, and markets have almost fully priced a November hike with about three hikes expected by mid-2027 ^[1][7]. Bailey has said current pricing reflects a “risk premium” on energy prices rather than a certain signal that a hike is imminent ^[1]; Refinitiv data show a 32% probability of a hike at this meeting ^[7], and ING expects a hold given the policy rate is already restrictive, employment growth weak and wage growth moderate ^[7].
• [NEW] BoC and Brazil: the Bank of Canada held rates earlier this month while stressing inflation risks, with Wednesday’s minutes expected to reveal its policy lean ^[11]; Brazil’s central bank is expected to deliver a fifth consecutive 25bp cut, taking the Selic rate to 13.75% ^[11].
• [NEW] PBoC: Guosheng’s monetary-direction factor for the PBOC is above zero and issuing a bullish signal, indicating an easing policy stance ^[19].

5. Asset Implications

Quadrant	Current probability tilt	Key asset implication	Anchoring narrative
Growth↑ + Inflation↑	Rising	Energy/commodities and TIPS remain the cleanest expressions — WTI above $100 and Brent above $104 on the week [7] — while nominal long bonds stay blocked by a term premium being lifted by fiscal concerns, supply-demand mismatch and Warsh policy uncertainty [8][13]	§2 / §3
Growth↑ + Inflation↓	Falling	The “noise” path: core CPI ex-distortions near 0.09% and record-low inflation breadth [4], plus an expected 0.1–0.3pp downward PCE recalibration on Sept 30 [4], would leave the Fed with cover for a one-off move and favour front-end/belly carry over long-end duration	§1.2 / §2
Growth↓ + Inflation↑	Rising (tail)	The energy pass-through leg is live and policy-insensitive — Haitong International explicitly flags the time lag from producer to downstream prices as a brake on core CPI and PCE cooling [12] — and gold is the hedge, but real rates keep suppressing it [7] and the IEA now sees demand falling 2.5mn bpd in 2026 [7]	§2 / §3 / §6
Growth↓ + Inflation↓	Falling	The “credit repair” path: a delivered hike read as the Fed withstanding political pressure and upholding policy independence would be supportive of long-end Treasuries [12], with a rate-hike landing as a small-scale risk clearing and a possible rebound window after the September meeting [13]	§1.2 / §2

Stock-bond correlation call: the regime remains the inflation/policy-driven positive-correlation configuration, but Friday’s tape carried an important nuance. Equities, long-end Treasuries and gold all fell quickly after the CPI release and then recovered together, with the major US indices closing up nearly 1% ^[3] and the 10-year ending flat ^[4] — that is the market trading convergence of uncertainty rather than a joint repricing on a shared inflation shock. That negative-correlation print is interpretive and fragile. The structural backdrop is still positive-correlation: the 30-year sits near 5.35% ^[16], the term premium is being driven by fiscal and supply factors that monetary policy cannot offset ^[8][13], and the week’s cross-asset pattern — US equities down, oil up almost 9%, gold down on the week — is the signature of an inflation-and-policy-driven regime rather than a growth-driven one ^[7]. The escape hatch is narrow and event-specific: if the Fed hikes but the dot plot points to only one move this year and Warsh frames it as calibration, the policy-uncertainty component of the term premium can compress, and long duration can partially recover its hedge value ^[4][12]. If the dot plot shifts the median up and the tone is hawkish, both legs of the portfolio are repriced by the same variable again.

Risk-budget implication: underwrite the FOMC as a two-sided event and do not take the 87–90% pricing as a directional signal. Keep commodities and energy at or above benchmark: energy is the only sleeve whose driver is genuinely orthogonal to the Fed’s reaction function, and it is the cheapest hedge against the stagflation quadrant. In rates, stay underweight long-end nominal duration into the decision but pre-commit to rebuilding it on a calibration read, and prefer curve and relative-value expressions to outright shorts given the two-sided term-premium risk. In credit, keep duration short: the AI-chain financing channel is competing for global allocation money ^[13] and record-tight spreads leave little protection. In equities, retain moderate beta but express it through quality and AI-hardware durability — Friday’s Philadelphia Semiconductor Index +1.81% versus a negative week for the broader market is the clearest evidence that the rate-sensitive and rate-insensitive parts of the index have decoupled ^[2][7]. Hold gold as the structural fiscal/dollar-credit hedge, sized for continued real-rate whipsaw rather than as equity insurance.

6. Contrarian & Tail Risks

• Consensus fragility: the 87–90% pricing is a market construct whose meaning is unpriced. The falsifiable pillars are narrow. First, the entire core exceedance rests on a handful of noisy categories — hotels, airfares and a telecom price hike amplified by BLS methodology — and stripping them out leaves core CPI m/m near 0.09%, with inflation breadth at a record low; a Fed targeting PCE could plausibly treat this as noise. Second, the dot plot is the swing variable: one camp expects the median to guide one more hike this year, another expects a preventive one-and-done with a data-dependent statement, and one house explicitly expects only a lone preventive move — one of these is wrong. Third, the one clearly contrarian house call in this batch is that the Fed will not hike next week, which the author himself acknowledges is the low-probability scenario in market pricing. Fourth, the pre-meeting divergence was the largest in nearly a decade, with the average hike probability across Aug 24–Sept 11 at only 56%, an unusually balanced expectation set for a meeting now priced near-certain. Fifth, the credibility argument cuts both ways — if the Fed cannot skip a hike the market expects, a delivered hike that is then not followed through would damage credibility more than a hold would have.
• Second-order transmission: three chains are not in market pricing. The energy pass-through chain is only starting — the July gasoline decline has reversed, fuel is up 28% y/y, and the producer-to-downstream lag will support downstream goods and services prices and brake core CPI and PCE cooling into year-end; if shipping chokepoints are further blocked, oil can surge again and simultaneously lift inflation expectations, short-rate expectations and long-end term premium, re-forming the oil-inflation-Treasuries-risk-asset negative feedback loop. The carry-trade chain is the cross-asset tail: another BOJ normalization step raises the risk of a fresh yen carry unwind hitting global equities, credit and emerging-market currencies, and Japan’s own fragility means policy intervention would in turn create Treasury selling pressure. The allocation-competition chain is the equity transmission vector: AI-chain debt financing is diverting global allocation funds and central banks are diversifying reserves into gold and US equities, which changes who the marginal buyer of duration is. Layered on top, the political channel is active — a Wednesday hike would run directly counter to the President’s wishes, and the Fed is operating under strong political interference, which itself argues for acting when the data supports it, to reduce doubts about independence.
• Source quality control: hike odds are a band across instruments, vendors and snapshot times — 82.5% (Refinitiv-linked money markets), 87.3% (CME FedWatch, cited by three separate houses), 88% (Soochow’s read), 90% (BOC International Securities), 90% (OIS swaps after the CPI print), and “>80%” (El-Erian, Guosen) — sequencing and instrument matter more than any single reading. Direct internal conflicts: the central CPI fact is genuinely disputed across this batch — some sources describe the core print as above expectations, others as fully inside the median forecast range, and one explicitly says the core m/m reading it uses was in line — so treat the “hot core” claim as contested rather than settled. The oil week is also reported inconsistently (a year-high $109 print versus a $104.61 weekly close). Single-source and unverified items include El-Erian’s “lose-lose dilemma” framing and his >80% pricing figure (social post), Chen Li’s no-hike call (authored view, not a survey), the Iran attack claim whose results require distinguishing Iranian statements from US confirmation and independent verification, and any model-derived positioning output, which its own publisher warns can fail if the market environment changes materially. One stale item in the batch is dated July 3 and is not usable for this week’s Fed judgment.

---

Appendix: Additional Sources

• ^[20] CITIC Securities — range-bound market call; a September hike as a signal the July correction is ending
• ^[21] Wallstreetcn — super-week preview: hike versus hold “evenly matched”; BOJ hike and BOE hold expected
• ^[10] Guosen Securities — preventive-versus-anti-inflation hike taxonomy; price the hike probability above 80%
• ^[8] Soochow Securities — three pressures lifting the term premium; buyback intervention below expectations
• ^[19] Guosheng Securities Financial Engineering Team — A-share timing radar; PBOC monetary-direction factor bullish
• ^[13] SWS Research — fiscal discipline, reserve diversification and AI-chain debt as structural long-end factors
• ^[18] Industrial Securities — ECB meeting details, projections and October pricing
• ^[12] Haitong International — energy pass-through lag; a delivered hike as supportive of long-end Treasuries
• ^[22] SWS Quantitative (dated 2026-07-03) — stale, excluded from this week’s assessment

---

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