
Core inflation rose for a second month running — the energy shock has reached everything else.

Written by
Oliver Summers
September US CPI · Data released 14 October 2026. The market reaction and pricing below refer to that release, not a live update.
The call
The numbers. Core inflation rose 0.3% in September, against 0.2% expected — the same pace as August. The annual core rate climbed to 2.6%, from 2.4%. Headline inflation hit 3.7%.
What that does. Core strips out food and energy. It rising twice in a row means price pressure is no longer confined to the petrol pump. That is the thing a central bank cannot wait out.
The rate. 3.75–4.00%. Odds of a rise at the 27–28 October meeting roughly doubled on the release, from 22% to 43%.
The currency. Higher rates make a currency worth more to hold, and this print put a meeting back in play that the market had written off. Dollar stronger.
The print
Core CPI +0.3% m/m, against +0.2% expected and +0.3% previously.
Core CPI +2.6% y/y, against +2.5% expected and +2.4% previously.
Headline +0.6% m/m, in line. Headline +3.7% y/y, against +3.6% expected and +3.4% previously.
The headline was expected. Core was not.
Why it matters
On the surface this is a petrol story. Gasoline rose 9.1% in the month and accounted for more than half the entire increase. Energy is up 20.4% over the year, gasoline up 34.2%. A central bank can look through that, and has said so.
It cannot look through what is underneath it.
Core goods rose 0.3% in the month — against just 0.7% across the whole year. Goods prices have been flat for months, and then moved in a single month by nearly half a year’s worth. Services, excluding energy services, rose 0.4%. Shelter rose 0.4%. Food was flat, so none of this is groceries.
That is the pattern of costs being passed on. Energy and tariffs reach manufacturers first, manufacturers reach shelves second, and this is the month it showed up on the shelves. The argument that American inflation is an oil spike the Fed can sit out did not survive this print.
The case
The market repriced immediately and in the right place. Two-year yields, the maturity that tracks policy expectations most closely, jumped 13 basis points to 4.93% — a larger move than the 10-year’s 9 basis points. The dollar index rose 0.5% to 102.38. Gold fell 1.2%. Stocks fell, with the Russell 2000 down 1.4%.
Odds on an October rise went from 22% to 43%. Note what that is and is not. It is not a market that now expects a rise on 28 October — it still leans against one. It is a market that had almost entirely dismissed the meeting and has been forced to take it seriously again.
That gap is where the dollar’s support comes from. The Fed’s own September projection pointed to one more rise before the end of the year, and this print makes that harder to avoid.
Ideas
Short GBP/USD or short AUD/USD — the clean expressions. The Fed decides on 27–28 October. The Bank of England does not meet until 5 November and the Reserve Bank of Australia not until 3 November. Both give a live dollar event with nothing scheduled on the other side.
USD/JPY, USD/CAD and EUR/USD — liquid, but the dates collide. The Bank of Japan and the Bank of Canada both decide on 28 October, the European Central Bank on the 29th. Any of these can be traded on this print, but the position has to be closed before that week or it becomes a bet on two central banks at once.
Sources
BLS — Consumer Price Index, September 2026 · Benzinga — market reaction, Fed odds, yields, dollar index · Central bank meeting calendar — October and November decision dates
What would prove this wrong
The next check after the CPI release was producer prices, scheduled for 15 October with a 0.5% monthly rise expected. Producer prices are the cost pressure before it reaches the shops — a reading well below that would weaken the pass-through argument immediately. This report does not incorporate that subsequent release. Beyond that, core falling back to 0.1–0.2% next month would break the two-month run and weaken the case for another rise.
