China's producer prices rose at their fastest pace in more than four years in September, and the stated driver was raw materials — higher oil and commodity costs pushing up factory-gate prices. Consumer inflation, meanwhile, barely moved. The pressure in the Chinese industrial chain is entering through the input door, and that is a very different animal for the commodity complex than broad demand-pull inflation.
Here is why the gap between those two numbers matters more than either number alone. When producer prices run ahead of consumer prices, the manufacturer eats the difference. Margins compress at the smelter, the refinery, the converter — and compressed processing margins are the single most reliable leading indicator of rationed throughput I know. A Chinese refiner facing dearer crude and a soft domestic sales book does not place the same incremental cargo order next month. The input cost shows up as strength in the commodity print today and as a subtraction from physical demand a quarter out. That is the cycle, and it has run this way through every raw-material squeeze I have watched from the physical side.
Which brings us to the metal that is supposed to be the instrument for exactly this moment. XAU/USD closed at $4,147.04, down $17.40 or 0.42%. It opened at $4,164.48 — essentially flat against the prior close of $4,164.44 — ran as high as $4,169.97, and then spent the balance of the session leaking lower to a $4,143.76 low. The full range was $26.21. That is a market with its hands in its pockets, and I am not going to pretend otherwise.
The detail that nags at me is the dollar. Yesterday's Americas session had the dollar retreating ahead of the U.S. CPI release and crude firming — the textbook configuration for a bullion bid. Gold declined anyway, its fourth consecutive red close, extending a 1.1% slide over five sessions and 3.3% over thirty. The close landed in the lower third of the day's range. One session's close location is texture, not evidence; four red closes in a row with a softer dollar underneath them is at least a pattern worth naming.
Gold sits 25.9% below its 52-week high of $5,597.23 — a long way down from the peak, and still a long way above anything that looks like distress.
The counterargument, and where I have been wrong
The honest counter is that none of this is a signal yet. A $26 range the day before a U.S. CPI print is positioning, not conviction. The question posed in the market wrap was whether inflation "remain[s] concentrated in energy, or will the data show price pressures spreading more broadly" — and until that is answered, nobody with size is expressing a view in bullion. The Chinese PPI beat is one month of data from one economy; it supports an observation about input costs, nothing larger.
| Symbol | Close | Change | Day range | 52-week range |
|---|---|---|---|---|
| XAU/USD | $4,147.04 | -17.40 (-0.42%) | $4,143.76–$4,169.97 | $3,900.00–$5,597.23 |
I also owe a correction. I called for gold above $4,350 within seven sessions and it never came close; I called for a break below $4,112.09 and that expired too. Both were wrong in opposite directions, which tells you I was reading momentum into a market that has mostly been drifting. What changed since those calls is that the drift acquired a direction — the metal is now 11.0% below its 90-day high and only 3.0% above the 90-day low of $4,027.92, and the balance of that range has shifted. My downside level at $4,120 from Monday remains live and the market has moved toward it, not away.
So the levels. A daily close below $4,120 within five sessions would confirm the slide is more than pre-data positioning and put the 90-day low back in play. A close above the session high of $4,169.97 within three sessions ends the four-day streak and tells me the selling was calendar-driven rather than structural. That second level is where the argument gets settled fastest, because it is the cheapest thing for the market to reject. What would change my mind on the downside case: a CPI print that shows price pressure broadening beyond energy, with the dollar continuing to give ground. That combination has historically been the one configuration bullion cannot ignore, and it would make the last four sessions look like noise.
The thing I keep circling back to is the mismatch in the Chinese numbers. Commodity costs are climbing hard enough to lift factory-gate prices at a four-year clip, and the metal whose entire purpose is to price that risk has gone down four days running. Either the physical cost inflation is real and gold is mispriced, or the inflation is confined to a narrow energy channel that will not broaden and gold is correctly bored. Tomorrow's data arbitrates one version of that question for the United States. But does a U.S. consumer price print actually tell us anything about whether a Chinese converter keeps buying tonnage at these input costs?