Hebei's October PMI at 43.7%: demand collapsed, costs collapsed faster — and that is the story

Hebei's steel sector PMI fell to 43.7% in October from 48.0% in September, according to the Hebei Province Metallurgical Industry Association (HBMIA) — a second consecutive month below the 50 threshold. The headline is easy to read as another data point in a long soft patch. The sub-indices tell a more specific and more interesting story: the collapse is concentrated in prices, not in output, and that distinction matters more for margins than for volumes.

The composition of the fall

The composite dropped 4.3 points, but the damage is unevenly distributed. New orders fell 10.9 points to 34.2% — a level that in diffusion-index terms implies a broad, not marginal, contraction in domestic bookings. Production, by contrast, eased only 2.5 points to 46.3%.

Source: HBMIA monthly releases, as reported by SteelOrbis and Mysteel. September figures from the HBMIA 1 October release; October from the 1 November release.

Three observations follow from the table.

First, the order-to-output gap. New orders at 34.2% against production at 46.3% means mills are producing into a demand hole. The HBMIA attributes the output decline to loss-making mills trimming volumes and stepping up maintenance, but the 2.5-point fall is modest relative to a 10.9-point demand drop. In practice, Hebei producers responded to deteriorating orders with cost-cutting rather than with a decisive supply response — a pattern consistent with the province's structure, where a large share of capacity carries fixed-cost and employment obligations that discourage deep cuts.

Second, inventories are not clearing. The finished steel inventory index held exactly at 50.0%, unchanged for a second month. That is not a neutral reading — it is the diffusion index saying stocks are neither rising nor falling, which, with output above the order index, implies the destocking that falling production should have produced is being offset by weak offtake. The association attributes the flat reading to a small build in construction steel mill stocks offset by faster destocking in flats and plate. The product mix is doing the work here: rebar is accumulating while flat products drain, which is the signature of a property sector that remains the weak link while manufacturing-linked demand holds up better.

Third, and most importantly, the price indices. The ex-works finished steel price index fell 20.7 points to 25.6%, and the raw material purchase price index fell 25.6 points to 35.4%. Both are far below 50 — meaning the overwhelming majority of surveyed mills reported lower selling prices and lower input costs in October.

Why the cost collapse matters more than the demand collapse

The raw-material purchase price index falling 25.6 points in a single month is a violent move. The HBMIA attributes it to four concurrent developments: softer imported iron ore, two rounds of coke price cuts landing during the month, scrap prices falling throughout October, and a weak billet market.

On the published numbers, the cost index (35.4%) sits roughly 10 points above the ex-works price index (25.6%). In diffusion-index logic, that gap is the margin signal — and a 10-point gap in favour of costs means the rate of price decline for steel is steeper than for inputs. Falling input costs are not rescuing margins; they are chasing selling prices down.

The mechanism is straightforward and worth stating plainly, because it explains why the November outlook is for further softening rather than a trough. When mills buy iron ore and coke on the spot market and sell steel into a weak domestic market, a fall in input prices is passed through to product prices within weeks. The association's own framing — falling coking coal and iron ore "pulling down steelmakers' production costs and weakening the support for steel prices" — describes exactly this pass-through. Cost deflation in this configuration is not a cushion; it is a mechanism that removes the floor under product prices.

The one apparent anomaly is the raw material inventory index rising 3.7 points to 43.9%. Read naively, that looks like restocking ahead of demand. The association's explanation is more prosaic: mills bought on dips as raw material prices weakened (a practice Chinese traders call fengdi buku, buying the dip to restock). That is a passive, cost-driven restock, not a demand signal. Note also that 43.9% remains below 50, so raw material inventories are still being drawn down across the sample — just less aggressively than in September.

Exports: the only sub-index above 50, and why it will not save the month

New export orders held at 54.8%, unchanged from September and the only sub-index in expansion. But the composition behind that stability has shifted. The HBMIA notes that as the domestic-international price spread narrowed and export margins thinned, exporters moved "from scrambling for orders to taking orders based on demand." In other words, the index is holding up not because export demand is strengthening but because the marginal exporter has become more selective — quoting on enquiries rather than chasing volume.

That has a direct implication for the November export picture. A narrowing price spread between Chinese and international markets erodes the arbitrage that has driven Chinese steel exports through 2026. China exported 85.64 Mt of steel in January–September 2026, down 2.6% year on year, with September at 10.49 Mt — up 3.3% m/m but only 0.3% y/y. The export engine is decelerating, and a stable new export order index at 54.8% against a collapsing domestic new order index at 34.2% cannot compensate for the domestic shortfall in volume terms. Exports are a partial offset, not a solution.

Reading the composite correctly

The HBMIA's methodology is worth understanding before drawing conclusions from the headline. Under its group standard T/HBMIA 6—2019, the composite weights new orders at 0.35, production at 0.25, employment at 0.15, supplier delivery time at 0.15 (as an inverse indicator) and raw material inventory at 0.10 — a diffusion-index approach with seasonal adjustment, aligned with the national NBS manufacturing PMI and the CFLP steel PMI but weighted for Hebei's industrial structure.

Using the three published weights that are disclosed (new orders 0.35, production 0.25, raw material inventory 0.10), the published sub-indices contribute 0.35×34.2 + 0.25×46.3 + 0.10×43.9 = 27.9 points to the composite. To reach 43.7, the remaining 30% of the weight — employment and supplier delivery time — must contribute about 15.8 points, implying an average of roughly 52.5 on those two components. (This is my own arithmetic, not a published figure; employment and delivery sub-indices are not disclosed in the open summary.) A reading near 52.5 on those two is plausible — employment close to neutral and delivery times lengthening, which registers as expansion on an inverse indicator when demand is weak — but it means roughly a third of the composite is holding the headline above where the demand and price indices alone would put it. The underlying demand picture is weaker than 43.7% suggests.

What to watch in November

The association expects steel prices to soften further in November as demand slackens in the traditional off-season. On the October data, that call is well supported. The sequence to watch is whether the cost index keeps falling faster than the price index — if it does, margin compression accelerates into year-end even as absolute prices decline; if the gap narrows, mills are at least holding pricing power relative to inputs.

The more consequential question is whether Hebei's mills respond to a 34.2% new order index with genuine supply discipline. October's production index at 46.3% suggests they have not yet done so. Steel markets have a habit of finding their floor only after output has been cut hard enough to clear the inventory that the 50.0% reading says is not clearing. Until that happens, the "weak supply-demand equilibrium" the HBMIA describes is likely to resolve downward.

For context on scale: Hebei produced 152.32 Mt of crude steel in January–September 2026, still first among Chinese provinces. The province's steel value-added accounts for roughly a quarter to nearly 30% of provincial above-scale industrial value-added — which is why a Hebei PMI reading of 43.7% is not a regional curiosity but a leading indicator for Chinese steel balances overall.