CPI Lands Wednesday Night: Headline Set to Jump to 3.7% as Core Cools — The Last Piece Before the October FOMC, and a Forward-Pricing Playbook for MPF Bond Exposure

The US Bureau of Labor Statistics releases September CPI at 8:30pm Hong Kong time on Wednesday. It is the last major inflation print before the October 27–28 FOMC meeting — and forecasts are split in a way rarely seen: headline inflation is expected to jump to 3.6–3.7% year-on-year (+0.6% month-on-month) on surging energy prices, while core inflation is expected to cool to just +0.2% month-on-month. For MPF members, the real question is not “what will CPI print” but “which number will the Fed believe” — and under T+1/T+2 forward pricing, you cannot react after the release anyway.

The macro matrix: risk-off for duration

The data first (week ending October 10):

• Yields: 10-year around 5.24–5.28%, 30-year around 5.65% (a 24-year high), 2-year around 4.83%. The long end kept climbing despite a soft September jobs report — markets are demanding compensation for holding long-duration debt through an energy-driven inflation episode, on top of heavy Treasury supply.
• Policy: Fed funds at 3.75–4.00% (after the September 16 hike, the first since 2023). CME FedWatch prices an 82–83% chance of no change on October 27–28 (only a 17–18% chance of a hike), but a roughly 69% chance of a cumulative 25bp hike by December. In the dot plot, 16 of 18 officials expect at least one more hike this year; the end-2026 median sits at 4.1%.
• Official voices: Williams (September 29) said there was “no need for urgency” and that one more hike “may be appropriate” late this year; Jefferson (October 1) said forming a judgment “may take more time”; the September meeting minutes (released October 7) showed most officials judged at least one more hike appropriate before year-end.
• The real economy: September payrolls rose just 29,000 (consensus: 84,000); unemployment ticked up to 4.2% (mostly as participation rose to 61.8%); hourly earnings grew only 0.1% on the month and 3.0% year-on-year, the slowest since May 2021. Consumer confidence sits at 81.9, the lowest since April 2014 — yet final Q2 GDP printed 2.2%, far above expectations. Soft and hard data keep disagreeing.
• Energy: Brent around US$104.70, WTI around US$91.90 — the culprit behind the expected +0.6% monthly headline jump.
• Technicals: the S&P 500 keeps making record highs; the equal-weight index snapped a seven-week losing streak with a 1.6% bounce; the dollar index stands at 101.9, a six-month high.

Matrix verdict: neutral on equities (AI capex is holding the market up),明确 risk-off on duration. With the 10-year near 5.3%, any upside surprise in core inflation hits long-bond prices directly — and MPF bond fund NAVs reflect it on T+1.

The split forecast: 3.7% versus 0.2% — which one does the Fed believe?

Wednesday’s script has two layers:

Layer one (headline): the Reuters poll expects headline CPI at 3.6% year-on-year and +0.6% on the month (August: 3.4% / +0.4%); some consensus tallies see 3.7%. The driver is almost entirely energy — petrol prices tracking the oil spike. Imported, one-off in nature.

Layer two (core): stripping out food and energy, the call is just +0.2% on the month (August: +0.3%) and 2.5% year-on-year (August: 2.4%). Natixis is more precise — core around +0.19% on the month — and argues August’s overshoot was driven by one-off component spikes that should not repeat in September. If that holds, September’s hike may prove to be the “one-and-done” of this tightening cycle.

Markets do not buy “one-and-done”: December hike odds still sit near 70%. The essence of the disagreement: will the Fed look through the energy spike (trust core), or get boxed in by the headline (trust the whole number)? The Fed claims to watch core, but the lesson of 2022 is that when headline inflation stays elevated long enough, expectations pressure forces action anyway. Warsh has set his tone as “inflation is too high, and has been for too long” — not dovish.

The transmission chain into MPF is short: a hot core print (0.3% or above) → October hike bets revive → long-end yields push higher → bond fund NAVs marked down on T+1. An in-line or cool core → the December script holds, and markets keep “hiking slowly”.

Event study: what bond funds actually did after the last hike

Rather than guessing the number, study the mechanism. Two verified events:

First, the September 16 hike did not clear the air. Manulife’s Hong Kong Bond Fund fell another 0.11% in a single day on September 22; its international bond fund fell 0.07%. A delivered hike is not a bottom for bond funds — because long-end yields kept rising after the hike (the 10-year went from around 5.1% in mid-September to around 5.28% last week), and duration risk is priced daily.

Second, the 2022 precedent. Manulife Hong Kong Bond Fund -9.46%, MassMutual Global Bond Fund -13.57%, HSBC Age 65 Plus Fund -13.21%. When stocks and bonds fell together, the bond “safety leg” was the most painful one.

The duration maths (illustrative): Fidelity’s Hong Kong Bond Fund runs about 3.8 years of duration — every 100bp rise in yields costs roughly 3.8% of NAV. Against the average MPF balance of HK$338,950 with a 30% bond allocation (about HK$101,685 in bonds): a 100bp move erases roughly HK$3,900; a 2022-style -13.57% erases roughly HK$13,800 — more than two months of contributions at HK$5,000 a month.

The BEA lesson: when the Fed hiked on September 16, it was holding inflation numbers that were revised down five weeks later (July core PCE cut from 3.3% to 3.0%). “Data dependence” is itself noisy — the expected value of betting on any single data point is negative over time. That is where discipline starts: you are not trading CPI; you are managing execution risk under forward pricing.

The switching playbook: three scenarios, three member types

The framework below is illustrative, not personal advice. All switches settle at T+1/T+2 unknown prices; nothing here is day-trading guidance.

Scenario A — core at or below 0.2% (in line): the December script holds; the long end stays elevated and choppy. Do nothing. Especially DIS members over 50: the statutory derisking glidepath moves about 6.7 percentage points a year automatically and does not change on one data point.

Scenario B — core at or above 0.3% (hot surprise): October hike bets revive; long-end selloff risk rises. Any already-planned moves (ECA transfers, TVC transfers, fund switches) should wait until after the October 27–28 FOMC for reassessment. Self-directed members with more than half their portfolio in bonds (heavy in Age 65 Plus-type or global bond funds) could consider parking 20–30% of the bond sleeve (about 10–15 percentage points of the total portfolio) tactically in the MPF Conservative Fund, then decide on redeployment after the meeting — at the cost of near-zero return while parked (the prescribed savings rate is just 0.001% a month) plus two rounds of forward-pricing blind windows.

Scenario C — core below 0.2% (cool surprise): the “one-and-done” narrative gains traction; duration could enjoy a relief rally. But do not chase: you cannot buy the rally day’s NAV (T+1), so chasing means paying for noise. Hold the allocation and let time work.

Young accumulators (ages 30–40, 100% equities or the DIS Core Accumulation Fund): do nothing in all three scenarios. CPI is two weeks of noise; your compounding is measured in decades — the illustrative sandbox of HK$5,000 a month at a 7% gross return reaching about HK$6.1 million over 30 years will not be rewritten on a Wednesday night.

Forward-pricing defence: the calendar and the no-go zones

Pin this calendar to the wall (Hong Kong time):

• Monday, October 12: US equities open, bond market closed (Columbus Day) — thin Treasury liquidity can magnify price moves; not an execution day.
• Wednesday, October 14, 8:30pm: September CPI; overnight: the Fed’s Beige Book.
• Thursday, October 15, 8:30pm: September PPI and retail sales.
• October 27–28: FOMC meeting; December 8–9: the year’s final meeting.

No-go rules:
1. Switch instructions submitted on CPI day (Wednesday) or the day before buy the post-release unknown price — your order is already in transit when the number drops. Wednesday is not an instruction day.
2. Split large switches into tranches, no more than half the planned amount per instruction, and keep clear of FOMC week.
3. Before any switch, ask: is this allocation discipline, or an emotional reaction to a number? If the latter, the BEA revision lesson stands — the number you are betting on may be rewritten five weeks later.

The close: at 8:30pm on Wednesday the number will be revealed. But the rules of the MPF game never change: you always buy at an unknown price. The real alpha is not in calling CPI — it is in letting execution discipline keep you clear of the noise during the two weeks when everyone else is staring at the number, and keeping every switch on a day when you can see its cost.

Execution constraints: MPF uses forward pricing — switch instructions settle at T+1/T+2 net asset values, and day trading is not available. The scenarios above are an illustrative framework; tactics must map to real MPF fund classes (North American equity funds, Hong Kong/global bond funds, Age 65 Plus funds, the Conservative Fund, DIS constituent funds) and do not constitute personal investment advice.