China’s September PPI rises 5% while CPI gains 0.9%: tracing upstream price pressure

China’s September price data reveal a sharp contrast across industries. Figures published on October 14 by the National Bureau of Statistics website (stats.gov.cn) show factory-gate producer prices, measured by the PPI, rising 5.0% from a year earlier, while consumer prices, measured by the CPI, increased 0.9%. Energy, metals and chemical raw materials recorded substantial increases, but some industrial products continued to become cheaper and the overall household consumption basket rose much less. ¹²

Compared with August, the PPI increased 1.2% and the CPI 0.2%. Industrial purchasing prices rose 6.9% year on year and 1.2% month on month; core CPI, which excludes food and energy, increased 1.0% annually and was unchanged monthly. The annual comparison is with September 2025, while the monthly comparison is with August 2026. They answer different questions: a large annual increase is not the additional cost incurred during September alone. ¹²

Historical photograph: Shanghai, January 8, 2014. Photographer: Livelikerw; work: Nong Gong Shang Supermarket in Huangpu Shanghai, Wikimedia Commons; CC BY-SA 3.0.

Strong upstream increases do not mean every industrial price is rising

Within the PPI, prices of production materials increased 6.5% year on year, contributing about 5.07 percentage points to the headline change. Consumer-goods prices at the factory gate fell 0.1%, subtracting about 0.03 percentage points. Within production materials, mining prices rose 25.4%, raw-material prices 9.4% and processing prices 3.4%. This breakdown shows that upstream products and production inputs drove the headline 5.0% increase; it does not mean every factory raised its prices by the same amount. ¹

Purchased inputs also differed substantially. Prices of nonferrous metal materials and wires rose 20.0% annually, fuel and power 13.6%, and chemical raw materials 12.6%. Their monthly increases were 1.4%, 3.7% and 2.4%, respectively. The category with the largest annual increase was therefore not the one rising fastest during the month. For near-term replenishment, the monthly movement in fuel and power and the annual movement in metals each matter, but require separate interpretation. ¹

Household averages conceal opposing changes in spending categories

Food prices in the CPI fell 1.7% from a year earlier and service prices rose 0.8%. Against August, however, food rose 0.4% and services fell 0.2%. Food becoming cheaper annually but more expensive monthly is not a contradiction: the reference periods differ. A price below its level a year ago does not necessarily make the same purchase cheaper this month than last month. ²

Transport and communications spending also diverged. Energy used by transport vehicles rose 14.9% year on year and 4.5% month on month; passenger-car prices fell 1.7% annually and 0.3% monthly; communications-device prices rose 14.3% annually and 3.6% monthly. Transport energy is a statistical category, so its change must not be relabelled as a gasoline-price change. Nor does the movement in communications devices describe all electronic products. ²

A household that frequently commutes by car, one buying a car and one spending a large share of its budget on food may consequently experience different price movements. The difference depends on what, how much and when they buy. It does not mean every household necessarily experiences a 0.9% increase in spending. The CPI measures prices; changes in the quantity a family buys are not themselves counted as price changes.

Factory-gate, purchasing and consumer indexes use different baskets

The NBS methodology defines the PPI as the movement in industrial products’ factory-gate prices at their first sale. The purchasing-price index measures the prices industrial companies pay for intermediate inputs. The CPI covers consumer goods and services bought by households. Transport, wholesale and retail stages may follow the factory-gate transaction, while many household services have no one-to-one equivalent in industrial output. The PPI–CPI difference is therefore not a ready-made measure of cost pass-through. ¹²

The surveys’ coverage reinforces that distinction. The CPI spans eight major categories and 268 basic classifications, drawing prices from nearly 120,000 collection points in about 500 cities and counties. The factory-gate survey covers 41 industrial divisions and more than 1,300 basic product classifications; the purchasing survey covers nine major categories and more than 800 classifications. Industrial price collection involves more than 40,000 enterprises. These are broad, weighted statistical sets, not the purchase and sales invoices of the same business. ¹²

Both indexes moved to a 2025 base from January 2026, with changes to classifications, samples and weights. The NBS estimates that rebasing affects monthly year-on-year indexes by an average of about 0.06 percentage points for CPI and 0.08 percentage points for PPI. These are average rebasing effects, not adjustments that can mechanically be subtracted from September’s rates to produce a supposedly “true” inflation figure. ¹²

The effect on profit depends on the individual business

Industry factory-gate prices provide more specific clues than the aggregate. In September, automotive manufacturing prices fell 2.0% year on year, while computer, communications and other electronic-equipment manufacturing prices rose 5.8%. Electricity and heat production and supply prices fell 3.9%, and nonmetallic mineral-product prices fell 2.5%. Upstream increases were clearly not transmitted evenly to all downstream industries. Even an industry average cannot be treated as every constituent company’s actual transaction price. ¹

Purchasing-price inflation of 6.9% and factory-gate inflation of 5.0% do not imply a fall in profit margins equal to the difference between them. Their product mixes and weights differ, and they do not match a particular company’s inputs to its outputs transaction by transaction. A manufacturer using large amounts of metals and energy, with limited ability to adjust selling prices promptly, could face pressure if input increases persist. Price-linked contracts, inventory bought earlier or productivity improvements could alter that outcome.

Equally, higher selling prices do not automatically increase an upstream supplier’s profit. Input costs, output volumes, wages, financing expenses and product mix all affect results. Testing whether costs are squeezing profit requires company-level selling prices and costs per unit, gross margins, inventory turnover and cash flow, alongside sales volumes. Price indexes alone cannot establish whether revenue growth comes from charging more or shipping more.

Purchasing decisions depend on exposure and pass-through conditions

For procurement teams, the report is most useful as a guide to costs worth checking: which changes match their own input mix, how often contracts reset and whether suppliers’ quotations have already changed. Buying large inventories simply because the aggregate index is rising could create working-capital pressure and losses if prices subsequently fall. Bringing purchases forward still depends on orders, delivery schedules and financial capacity.

The next useful evidence is whether industry prices keep increasing, whether companies can raise prices without a material loss of sales volume, and whether reported costs, profit and cash flow support the same picture. September’s data establish a divergence between upstream and household prices and suggest that some manufacturers need to watch cost pass-through. They do not yet establish a broad recovery in manufacturing profits, a general strengthening of household demand or an inevitable direction for future policy.

Source notes

1. October 14, 2026, 09:30 Beijing time · Industrial producer prices in September 2026: up 5.0% year on year and 1.2% month on month (original title: 2026年9月份工业生产者出厂价格同比上涨5.0% 环比上涨1.2%) · National Bureau of Statistics website (stats.gov.cn), including industry tables and methodology. Data and definitions are paraphrased and translated; the analysis is the publisher’s. Reuse for free public news is under the NBS terms of service. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

2. October 14, 2026, 09:30 Beijing time · Consumer prices in September 2026 rose 0.9% year on year (original title: 2026年9月份居民消费价格同比上涨0.9%) · National Bureau of Statistics website (stats.gov.cn), including category tables and methodology; data and definitions paraphrased and translated under the same terms. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷