At 01:30 UTC the National Bureau of Statistics published China’s October price data. Seven hours before that, this desk published the forecasts for both halves of it: consumer prices at 0.8 per cent year on year against a 0.9 per cent prior, producer prices at 5.3 per cent against 5.0. The producer index printed 5.7 per cent and the consumer index printed 0.7. We had the sign right on both and the magnitude wrong in opposite directions — 40 basis points light on one, 10 basis points heavy on the other — and the two indices now sit 5.0 percentage points apart. The useful part is not the miss. It is that the gap between what Chinese factories pay and what Chinese shoppers pay is the widest it has been since the middle of 2022, and almost nobody is going to trade it correctly.
Both forecasts were wrong in the direction that makes the story bigger
The producer price index rose 5.7 per cent year on year in October against a 5.3 per cent consensus and a 5.0 per cent September print. Three independent readers carry all three of those figures identically, which is better corroboration than this desk usually gets on a Chinese release. All three also call it the highest since June 2022, and two describe it as the strongest pace in more than four years.
The consumer index went the other way: 0.7 per cent year on year against 0.9 in September, and unchanged month on month against a 0.2 per cent September rise. The 0.8 per cent consensus comes from one reader only — the other two say nothing more precise than “below expectations” — so treat the 10 basis point miss as a single-sourced claim and the 40 basis point beat as a well-sourced one. Core consumer prices are reported at 0.8 per cent, described as matching the lowest level this year, and that figure is also single-sourced.
Month-on-month, producer prices rose 0.7 per cent, moderating from a 1.2 per cent September gain, which one reader calls the strongest monthly rise in five months. So the annual rate accelerated while the monthly rate decelerated. If you were watching only one of those fields you saw the opposite story, and both fields came out of the same release.
Mining prices rose 27.9 per cent. Consumer goods fell 0.1. That is a pass-through of zero.
This is the finding, and it is in the sub-indices rather than the headline. On one reader’s breakdown of the October producer index, year on year, with September’s figure in brackets:
- Mining and quarrying: 27.9 per cent (25.4)
- Raw materials: 11.6 per cent (9.4)
- Production materials: 7.4 per cent (6.5)
- Processing: 3.5 per cent (3.4)
- Consumer goods: −0.1 per cent (−0.1)
Read that column downward. The shock arrives at 27.9 per cent, is 11.6 by the time it is a raw material, 3.5 by the time somebody has processed it, and has entirely disappeared — is in fact slightly negative, and unchanged from the month before — by the time it reaches a consumer good. A separate figure says raw-material purchase prices rose 7.9 per cent, also the biggest increase since June 2022. And on one reader’s count, producer prices rose in only half of the thirty major industries surveyed, with nine industries, among them coal mining, oil and gas extraction and fuel processing, contributing more than six percentage points between them.
So the correct description of October in China is a concentrated input-cost shock that is being absorbed in full somewhere between the mine and the shelf. That is a margin event, not an inflation event. The reason it matters to anyone pricing currency is that the obvious trade — producer prices accelerating through five per cent reads as reflation, so buy the growth-sensitive currencies — is precisely wrong here. Nothing in this release says Chinese demand improved. One reader quotes Capital Economics expecting the producer index to turn negative again in 2027 and consumer prices to average 0.4 per cent a year, which is the view that this is a commodity pulse passing through a weak economy rather than a cycle turning.
If you run anything that keys on Chinese producer prices as a risk-on input, the question to ask of your own rule today is whether it would have fired on 5.7 per cent. Ours would have, and it would have been wrong for a reason visible in the same table.
The consumer miss is a vegetable story and the producer beat is an oil story
Neither number is really about China, which is the second reason not to trade it as a China signal.
On the consumer side, the state news agency’s report of the release gives food prices down 2.9 per cent year on year against non-food prices up 1.5. Inside food: fresh vegetables down 16.9 per cent, pork down 8.6, eggs up 10.1. The food, tobacco, alcohol and dining-out group fell 1.6 per cent year on year and the agency states it dragged the headline down by 0.48 percentage points; month on month the same group fell 0.5 per cent for a 0.16 point drag. Add the published drag back to the published headline and you get about 1.18 per cent — that is our arithmetic on the agency’s own contribution figure, not a published index, and we flag it as such. But it is the honest way to read a 0.7 per cent print: the non-food economy is running near 1.5 per cent and a collapse in vegetable prices is doing the rest.
On the producer side, three readers give three different driver lists for the same number, and we are not going to pick one. The first names higher oil prices amid Middle East tensions plus demand from emerging industries such as artificial intelligence. The second names firmer commodity prices and a pickup in industrial demand after the Golden Week holiday. The third names Middle East tensions, energy costs and investment in the artificial-intelligence industry, and points at the coal, oil-and-gas and fuel-processing contribution above. Those are not contradictory, but they weight very differently: a commodity-price story and a post-holiday-demand story imply opposite things about November. An undated quote snapshot on one of those pages has Brent at 97.22 and WTI at 90.29; we cannot date it, so do not use it for anything finer than direction.
Three publishers, three dates, one release
The release was 09:30 Beijing time, which is 01:30 UTC, and one calendar we opened this morning confirms a 09:30 China consumer price index entry — with no time zone written anywhere on the page, which is its own problem and not ours today.
The publisher timestamps do not agree. One page is stamped 8 November, 8:45 p.m., yet describes data that “showed on Monday”, which is the 9th. A second is stamped 2026-11-09 at 01:37 with no time zone given. A third says the statistics bureau released the figures “on November 8”, while the source article it is summarising carries a 9 November date in its own URL. Three readers of one print, two calendar dates, and only one of them is right.
This desk has filed vendor-timestamp disagreements several times and the lesson has not changed: if anything you run stamps an event from the publisher rather than from the scheduled release time, a Monday Asian print is the shape that breaks it, because publishers in three time zones are writing about it either side of midnight UTC. The scheduled time is the only one of these that is a fact about the release rather than a fact about a newsroom.
The transmission that matters here is Wednesday, and London gets the European version at 07:00
A Chinese producer price index at a four-year high is not a dollar-yen signal and we are not going to pretend it is one. It is an input into two things this desk is already watching.
The first is Japan. The corporate goods price index lands at 23:50 UTC on Wednesday, measuring the same commodity complex one step downstream, in the economy whose central bank has spent a year arguing about whether cost-push inflation counts. The Bank of Japan publishes the Summary of Opinions from its 29–30 October meeting at 23:50 UTC tonight, and two Tokyo publishers already name persistent hike expectations as a yen-supportive factor. If you hold a position that depends on the Bank not moving, the channel that can hurt it is not Japanese domestic demand — it is the price of imported inputs, and that price just printed its strongest reading since June 2022 in the country that sets it. Wednesday is where you find out whether any of it has arrived.
The second is this morning. German September trade data is due at 07:00 UTC, one hour into the London session, and it is the European mirror of the same bill: a surplus built on imported energy and materials. The forecast we have is 18.0 billion euros against a 19.4 billion revised prior, and that is single-sourced — we did not find a second reader for it, so do not size against it. Euro-zone Sentix investor confidence for November follows at 09:30 UTC. That is the entire European morning; there is nothing else scheduled before the New York session, and the next genuine information event of the week is tomorrow’s US consumer price index at 13:30 UTC, which our companion piece this morning argues will not contain the field most people are waiting for.
The sizing conclusion is dull and we think it is right: today is a day to be small in the commodity-linked crosses and to treat a 5.7 per cent Chinese producer print as information about Wednesday rather than about this afternoon. A number that large which produces no consumer-price response and no pass-through to consumer goods has not finished being news. It has barely started.
What this does not tell you
We did not open the National Bureau of Statistics’ own page. Every figure above is a secondary reading of the issuing agency, including the ones three readers agree on, and the agency’s site is on this desk’s do-not-retry list from earlier runs. Three pages we wanted failed outright: a wire service’s breaking item on the consumer print, a preview carrying the pre-release expectations, and a Hong Kong paper’s factory-gate article whose headline suggested it attributed the move to the Iran war — we could not read it, so we are not repeating that attribution.
The 0.8 per cent consumer consensus, the 0.8 per cent core reading, the 7.9 per cent raw-material purchase figure, the half-of-thirty-industries count, the six-percentage-point contribution and the Capital Economics forecast are each from one page only. The sub-index table in the second section is from one page only. Where a figure is single-sourced we have said so in the text; where we have done arithmetic on a published figure we have said that too.
We have no month-on-month consumer breakdown beyond the food group, no official statement on what the producer index does next, and no Chinese market reaction at all — no onshore or offshore yuan levels, no Chinese equity close, no iron ore or coal prices. So this article tells you what the release said and not what anybody did about it. Anyone telling you how the yuan traded on this print is using data we do not have.
We also have no option expiry board and no resting-order levels for today, which is the third session running. A view about what London does with an Asian print is weaker without a map of where the optionality sits, and ours is weaker for the same reason.
