2026-10-12 → 2026-10-16

Core CPI, which excludes food and energy, is expected at 0.2% m/m against 0.3% in August. This is the number that decides the week.

0.2% monthly compounds to 2.43% annualized, which is essentially at target. RBC expects exactly that - a more modest +0.2% m/m - and reads the annual rate's nudge higher as arithmetic rather than acceleration. The consensus, in other words, is that an energy shock large enough to put the headline at 3.6% has not yet reached anything that is not energy.

The hawkish scenario is 0.3% or 0.4%. That is the print that converts an energy shock into an inflation problem, because it means pass-through has reached airfares, freight, insurance and the service categories where crude arrives on a lag of months rather than days. It is also the only plausible print this week that makes 27-28 October live again - MUFG is explicit that core must surprise to revive an October hike - and it would hand the four participants who wanted two more hikes their evidence eleven days before the meeting.

The dovish scenario is 0.1%. Combined with a 0.6% headline it would describe the cleanest possible version of a supply shock: large, concentrated and contained. It would also strengthen the case the minutes show Vice Chair Jefferson and President Williams making, that waiting for more data costs little. December's 74.5% is where that would register.

It is worth stating plainly what the consensus implies, because it is a strange pair of numbers. The street wants a 0.6% headline and a 0.2% core in the same report - a 7.4% annualized top line and a 2.4% annualized underlying rate, in the same economy, in the same month. Both have been roughly true for three months. The genuine risk this week is not that one of them is wrong, but that the gap between them finally begins to close from the wrong end.

Core CPI at 2.5% y/y from 2.4% is the quietest line on this week's calendar and the one that would most change the Fed's problem if it came in wrong.

Estimates cluster tightly, 2.4% to 2.5%, and the increase consensus expects is mechanical: a 0.2% monthly print replacing a weaker month a year ago lifts the annual rate without any change in the current run rate. RBC makes exactly this point - the annual pace nudges higher on base effects, not momentum.

The level matters more than the change. At 2.5%, core CPI is 0.5pp above target while the policy rate sits at 3.75-4.00%, which is to say the Committee is already restrictive against this number and has been for some time. Several participants in the September minutes judged current policy only mildly restrictive, or not restrictive at all; a 2.5% core is the strongest available argument against them, and it is the argument Jefferson and Williams are reported to be making when they say they may wait for more data.

The tail is 2.7% or above. That would break the 2.4-2.5% range core CPI has held through the entire energy move, and it would mean the thing the Fed hiked in September to prevent has started happening regardless - second-round effects. It would reprice October, December and the 2027 path together, and it is the one print this week with the power to move the curve by double digits in either direction. Nobody expects it. That is usually when a number is worth naming.

September headline CPI is expected at 0.6% m/m against 0.4% in August, and it is the largest consensus monthly gain anywhere on this week's calendar. Compounded, a 0.6% monthly rate is 7.4% annualized. That is not a number any committee can look through indefinitely, and it is why this release rather than Thursday's producer prices is the week's event.

Almost all of it is energy, and the mechanism is documented rather than inferred. Brent averaged $114 a barrel in September on the EIA's Short-Term Energy Outlook and is still above $100 - $102.91 on Monday, after briefly passing $105 last week on Houthi strikes against Saudi infrastructure. One forecaster models gasoline at +10% m/m seasonally adjusted against +7.4% unadjusted, which means a part of the expected strength is seasonal arithmetic layered on top of a genuine price shock rather than the shock alone.

The hawkish scenario is 0.7% or above. That hands the four participants who wanted two more hikes this year rather than one a concrete number, and it drags the 16.5% currently priced for 27-28 October higher. The front end would cheapen fastest, the dollar index would extend from 102.21 and an 18-month high, and gold at $4,196 would meet again the rate-support problem that has dogged it for five weeks.

The dovish scenario is 0.4% or below, and its effect is asymmetric in a way worth being explicit about. Because October is already only 16.5% priced, a soft headline has almost nothing left to remove from this month; it would instead work on December's 74.5%, which is the live number. The week's risk sits in December, not October, in both directions.

In isolation, though, the headline is close to noise. A central bank cannot tighten away a supply shock it did not cause, and this Committee has said as much; what it can respond to is that shock leaking into everything else. The headline tells you how large the shock is. Core, published in the same release, tells you whether it has leaked. Only the second one changes policy.

The annual headline rate is expected at 3.6% from 3.4% in August. The Cleveland Fed's nowcast puts it at 3.60%, street estimates run as high as 3.7%, and MUFG's framing of it is the useful one: at 3.6% the headline tells the Committee nothing it does not already know, and only a core surprise revives an October hike.

At 3.6% the annual rate would sit 1.6pp above the 2% target and 1.1pp above the 2.5% expected for core, against a 1.0pp gap in August. A headline-core spread that is still widening is evidence that the energy shock is still arriving rather than spreading. That is mechanically reassuring and politically useless.

Politically useless is the operative phrase. The y/y headline is the number that reaches voters, newspapers and the Committee's own credibility constraint, and it is the number Chairman Warsh invoked when he committed the Fed to a 'timelier return' to 2% at the September press conference. A 3.6% print, with one hike delivered and 16 of 19 participants projecting at least one more this year, is awkward in a way a 2.5% core is not. That asymmetry is why the hawkish tail here is fatter than the arithmetic alone suggests.

The scenarios compress accordingly. 3.5% or below and the story becomes that the shock is peaking, which argues for December's 74.5% coming down and takes pressure off the long end that has just started to rally. 3.7% or above - the top of the street range - and the question stops being whether energy passes through to the consumer and becomes whether a central bank that has publicly tied itself to a timelier return can hold at 3.75-4.00% through two more meetings while its headline measure accelerates. The 10-year's five-week refusal to rally was already an answer of sorts to that question. A 3.7% print would make it louder.

Australian employment is expected to rise 20,000 in September after a 39,500 gain in August, and the composition of that August number is why a soft headline here would not automatically be read as softness.

August added 39,500 jobs against a 20,000 consensus and the unemployment rate rose anyway, to 4.6%, because the gain was part-time: full-time employment fell 6,300 while the labour force grew faster than employment did. The ABS has separately flagged a methodology change that reduces the reliability of August's seasonally adjusted figures, a caveat worth carrying into any September comparison.

The policy context is the most hawkish in this brief. The Reserve Bank raised its cash rate 25bp to 4.60% in September, unanimously - its fourth increase of 2026 and a 15-year high - and the next decision is 3 November. Minutes of that September meeting are published on Tuesday 13 October, outside the red-folder calendar but inside this week, and they will be read for whatever the statement did not say.

A strong print - 40,000 or more with full-time work leading - puts a fifth hike into November pricing and lifts AUD/USD against a dollar index that has spent the quarter making highs at 102.21. A weak print does the opposite, but less cleanly: the Bank's own August Statement on Monetary Policy already expects the unemployment rate to drift up to 4.8% by end-2028, so gradual loosening is in the forecast and does not by itself stop a bank that has hiked four times this year.

The release lands at 02:30 CEST, which means European desks meet it as a settled fact at the open and the price action happens in the Asian session. For a brief written in Amsterdam that makes it context rather than an event.

The unemployment rate is expected to hold at 4.6%, and a flat forecast on a rate that has been rising is itself a statement.

4.6% arrived in August in the same report that added 39,500 jobs - the signature of a labour force expanding faster than employment, and on secondary accounts the highest seasonally adjusted rate since late 2021. The RBA's August Statement on Monetary Policy described the jobs market as strong overall while acknowledging the rate has risen a little, and projects a gradual climb to 4.8% by end-2028. September at 4.6% would therefore sit exactly on the path the Bank has drawn, which is the least interesting outcome and the most likely one.

The hawkish surprise is 4.5% or lower. With the cash rate at 4.60% after four hikes and underlying inflation the Bank's stated concern, a labour market that tightens again removes the last argument for patience before 3 November.

The dovish surprise is 4.8%, which would arrive two years early against the Bank's own projection and would be the first genuine evidence that 4.60% is biting. That is the print that stops the hiking cycle. It is also why this line rather than the employment change is the one to watch: Australian employment is volatile month to month and gets revised, while the unemployment rate is the series the Reserve Bank has committed to a published forecast path for, and deviations from a published path are what central banks actually respond to.

UK monthly GDP for August is expected at -0.1% after +0.4% in July. A contraction would set up the most uncomfortable central-bank meeting anywhere in this brief.

The arithmetic of the Bank of England's position, plainly: Bank Rate has stood at 3.75% for six consecutive meetings; the September vote was 6-3, with Pill, Mann and Greene wanting 4.00%; August CPI was 3.1%; and the Bank's own projection has inflation rising to roughly 3.75% in the fourth quarter and slightly above 4% in early 2027 on mid-September energy prices. Markets price close to 80% odds of a 25bp increase on 5 November, a meeting that also carries a new Monetary Policy Report.

Into that, a shrinking economy. -0.1% after +0.4% is not a recession signal on its own - monthly UK GDP is noisy, and July's 0.4% was strong - but the direction removes the one thing that would make a November hike comfortable. The three dissenters argued for a pre-emptive rise as a risk-management strategy against second-round effects; risk management is a harder case to make against a contracting print, even though the logic of it does not actually depend on growth.

The market consequence runs through gilts rather than sterling. At 5.385% the 10-year gilt is the most expensive 10-year money in this brief - 14.5bp above the Treasury and 188.5bp above the Bund on the latest published fixes - and the 30-year has been reported above 6%. A weak GDP print that softens November pricing would normally rally the front of the gilt curve and steepen it. The complication is that the UK long end has been trading on fiscal supply and inflation expectations rather than on the policy path, which is how a country ends up with a 6% thirty-year yield under a 3.75% policy rate, and it is why a dovish repricing would not necessarily help the part of the curve that matters.

Two caveats on the forecast itself. The -0.1% consensus is not universal - Trading Economics' own model has August nearer +0.1% - and the release lands at 08:00 CEST on the same morning as the American PPI and retail sales cluster, which gives a surprise in either direction about six and a half hours of clean air before Washington takes the session.

Core PPI is expected at 0.3% m/m against 0.2%, which would make it the faster of the two core measures published this week - above Wednesday's expected 0.2% core CPI.

That ordering is the point of the release. Core PPI strips food and energy from the producer basket, so an acceleration here cannot be attributed to gasoline. It has to come either from genuine upstream cost pressure in goods and services, or from the second-round effects the Fed hiked in September to prevent. 0.3% monthly compounds to 3.66% annualized, which is nowhere near target.

The scenario that matters is the combination rather than the level. 0.3% core PPI alongside a 0.2% core CPI says producers are absorbing the shock while consumer inflation stays contained - the benign reading, and the one the market will take. 0.3% core PPI alongside a 0.3% core CPI says pass-through is already under way, and that pairing, not either number alone, is what revives 27-28 October.

It is also the better leading indicator of the two, and it receives a fraction of the attention. Producer core prices move before consumer core prices, so a reading of 0.4% or above here would be the earliest hard evidence of leakage available this quarter - several weeks ahead of the October CPI that would confirm or deny it. Against a front end that has already marked October down to roughly 16.5%, that is a number with room to hurt.

Core retail sales - the headline measure excluding autos - are expected at 0.5% m/m after 1.4%, which means the consensus has vehicles subtracting from a total forecast at 0.3%.

A 0.5% ex-autos forecast sitting above a 0.3% total is unusual enough to unpack. It implies vehicle sales fell hard enough in September to take roughly 0.2pp off the top line, which after August's 1.4% ex-autos gain reads as payback in a category that had been running hot. Autos are also the most interest-rate-sensitive line in the entire report, and the policy rate has just risen for the first time since 2023.

The caveat from the headline applies here too, and with more force: gasoline stations remain inside the ex-autos measure. So this 0.5% forecast absorbs the same roughly 10% seasonally adjusted pump-price increase that is expected to drive Wednesday's CPI. Take autos out and leave gasoline in, and for this month the series behaves closer to a price index than a volume one.

What would genuinely inform is a miss below 0.2%. That would mean the gasoline arithmetic failed to carry even the ex-autos measure, which requires real contraction somewhere in discretionary spending - and it would arrive in the same 14:30 CEST slot as a PPI print expected at 0.5%. Prices up and volumes down, released simultaneously, is the cleanest stagflation signal this calendar can produce. It is a live possibility this week rather than a tail.

Producer prices are expected to rise 0.5% m/m in September after 0.4% in August, with the annual rate forecast at 5.9% against 5.4%.

A 5.9% producer rate against an expected 3.6% consumer rate is the most arresting pair of numbers on the week's calendar. Producer prices sit upstream of consumer prices and they are running 2.3pp faster. Either that gap closes through pass-through to the consumer, or it closes through compression of corporate margins. PPI is the series that eventually tells you which, and the gap has now been open long enough that the answer is overdue.

Energy drives the monthly figure again, and the same Brent path that lifts Wednesday's CPI headline - a $114 September average on the EIA's accounting, above $105 intraweek last week, $102.91 on Monday - works through producer prices with less dampening. The PPI basket carries a heavier goods weight and has no owners' equivalent rent to dilute it, so the shock arrives more or less undiluted.

What makes this second-order rather than first-order is sequencing. It lands a day after CPI, so its information content is conditional. A hot PPI following a hot core CPI is confirmation and moves the front end. A hot PPI following a soft core CPI is a margin story and moves equities instead. On its own, 0.5% against 0.4% would be a non-event, and treating it as one is the correct default.

Retail sales are expected at 0.3% m/m in September after 1.2% in August: a sharp deceleration that would still leave the American consumer growing.

The series is nominal, and this month that is the whole difficulty with it. Retail sales are reported in dollars rather than units, and gasoline stations sit inside the basket. With pump prices modelled up roughly 10% on a seasonally adjusted basis, a 0.3% headline gain is consistent with flat or falling real volumes - the same money buying less. Reading this print as a demand signal without that adjustment is the most common mistake available this week.

The context against which to read it has deteriorated. September payrolls added 29,000 against an 84,000 consensus with 60,000 of downward revisions attached, the unemployment rate rose to 4.2%, and Canada - the nearest comparable economy, and the first to absorb this year's tariff regime - shed 68,000 jobs in September against a consensus of a 10,000 gain, with its unemployment rate at 6.5%. A consumer still spending through all of that is the strongest argument the hawks have. A consumer that stops is the recession the long end has refused to price.

The scenarios split along the Fed's two mandates rather than along one. 0.6% or better and the growth-and-inflation case for December's 74.5% hardens, the 2-year cheapens, and the dollar extends from 102.21. 0.0% or negative, following a hot CPI the day before, and the week delivers the stagflationary pair - prices rising, demand stalling - which is the configuration with no good monetary answer, and the one that cheapens the long end without doing anything for the front.

One note on the forecast distribution: Trading Economics' in-house model sits at 0.1% rather than the 0.3% consensus, so the downside tail here is better populated than the headline forecast suggests.

Governor Bailey speaks at 04:00 CEST on Friday. The hour tells you this is a scheduled appearance at an overseas event rather than a policy set piece, and that any text will be read in London several hours after it is delivered.

He speaks into the tightest corner of any governor in this brief. Bank Rate has been held at 3.75% for six meetings, the last vote was 6-3 with three members wanting 4.00%, August CPI was 3.1%, and the Bank's own forecast has inflation reaching roughly 3.75% by the fourth quarter and above 4% in early 2027. Markets price close to 80% odds of a hike on 5 November. He is, in other words, two weeks from being outvoted or from validating a hike he has not yet endorsed.

And two days before he speaks, the August GDP print is expected to show the economy contracting 0.1%. Bailey has to either defend November pricing against a shrinking economy or walk it back and risk the long end, where the 10-year gilt sits at 5.385% and the 30-year has been above 6%. There is no comfortable third option, which is precisely why an off-hours speech is worth reading: the useful signals in this kind of situation tend to arrive in the venues nobody staged.

Three things to listen for. Any reference to second-round effects - the phrase the three dissenters used to justify a pre-emptive rise - is hawkish and would be a tell about which camp he joins on 5 November. Any characterisation of the energy shock as temporary or as something to look through would be the clearest signal yet that he intends to sit through a 4% headline. And any mention of gilt market functioning or term premium would be the most consequential of the three, because it would mean the Governor has started treating the long end as a constraint on policy rather than a consequence of it.

Chairman Warsh speaks at 05:30 CEST on Friday, 90 minutes after Bailey, eleven days before the FOMC convenes on 27-28 October - and, more to the point, on the last day before the Committee's communications blackout opens. The Federal Reserve's own blackout calendar runs 17-29 October. Whatever Warsh says on Friday is the final word from the Chairman before the decision.

He will have seen everything this week produced. He committed the Fed to a 'timelier return' to 2% at the September press conference, a phrase CNBC reported had Wall Street recalculating how far the Committee would go, and he presided over a unanimous 12-0 hike with 16 of 19 participants projecting at least one more this year. The minutes released last Wednesday showed many participants calling a higher path prudent on risk-management grounds, and several judging current policy only mildly restrictive or not restrictive at all.

Against that, the pricing. October sits at roughly 16.5% and December at roughly 74.5% on Polymarket. The market has accepted the Chairman's guidance and rejected his urgency: it believes the hike is coming and does not believe it is coming this month. Warsh is the one person who can correct that in either direction, and the last scheduled appearance before a blackout is exactly where it would be done.

The two tails are not symmetric. If he signals October is live, roughly 58 points of probability have to migrate from December back to October, and the front end cheapens into it - the 2-year richened 4bp across last week and gave 5bp of that back on Friday alone, which is a fair indication of how thin the conviction is. If he endorses waiting for more data, he confirms what is already priced and the move is small. A chairman who used Jackson Hole in August to nearly double the odds of a September hike is not a speaker whose dovish scenario should be assumed.

The hour is the practical problem. 05:30 CEST is 23:30 Thursday on the US East Coast, which is an unusual slot for a Fed chairman and means the European open on Friday will be pricing a text most of the market has not read in full. Thin liquidity and a hawkish paragraph is how gaps happen.