The October consumer price report has a hole where its most quoted number should be. In the table that normally opens with the year's inflation rate, the 12-month column for all items, for core prices, for food, for energy and for shelter carries a dash.
U.S. consumer prices rose 0.4 per cent in October, seasonally adjusted, and core prices rose 0.3 per cent, the Bureau of Labor Statistics said on 10 November 2026. It published no annual rate because it never collected October 2025 prices. Using the bureau's approximation method, we estimate annual inflation at about 4.0 per cent and core at 2.8 per cent.
Why is there no annual inflation rate for October?
Because there is nothing to compare October 2026 with. "BLS did not collect survey data for October 2025 because of a lapse in appropriations," the release says, and so it "cannot publish 12-month percent changes for October 2026 for indexes where October 2025 data are unavailable." The October 2025 cell in the bureau's all-items series is simply empty, sitting between September's 324.800 and November's 324.122.
Only three of the headline series escaped. Gasoline, new vehicles and used cars and trucks have October 2025 values, so the release prints their 12-month changes: gasoline up 37.4 per cent, new vehicles up 0.7 per cent, used cars down 3.3 per cent.
The monthly figures were unaffected. Energy rose 0.7 per cent in October after 5.2 per cent in September, food rose 0.5 per cent after no change, and shelter rose 0.3 per cent. Used cars fell 1.2 per cent. Electricity rose 0.9 per cent and motor vehicle insurance 0.7 per cent.
How we rebuilt the missing month
The bureau does not leave readers empty-handed. Its fact sheet on missing data says a missing index can be approximated by the geometric mean of the months on either side: multiply September by November and take the square root. It is blunt about the status of the result. Approximated values "cannot be interpreted as official CPI series," and percent changes built on them "should also be considered as unofficial estimates."
We applied that method to the unadjusted CPI-U indexes, the series the bureau uses for 12-month changes. For all items, the square root of 324.800 times 324.122 is 324.461. October 2026's index was 337.511. The ratio gives a 12-month rise of 4.02 per cent. The same steps for every major group produce the table below.
| Index | Sep 2025 | Nov 2025 | Approx. Oct 2025 | Oct 2026 | Approx. 12-month change | Published Sep 2026 change |
|---|---|---|---|---|---|---|
| All items | 324.800 | 324.122 | 324.461 | 337.511 | 4.0% | 3.7% |
| Core (less food and energy) | 330.804 | 330.425 | 330.614 | 339.921 | 2.8% | 2.6% |
| Energy | 283.589 | 279.588 | 281.581 | 340.061 | 20.8% | 20.4% |
| Food | 342.405 | 341.721 | 342.063 | 352.359 | 3.0% | 2.5% |
| Food at home | 316.042 | 313.954 | 314.996 | 322.876 | 2.5% | 1.8% |
| Shelter | 418.598 | 418.593 | 418.595 | 432.745 | 3.4% | 3.2% |
| Core goods | 168.083 | 166.693 | 167.387 | 169.011 | 1.0% | 0.7% |
| Core services | 434.852 | 435.344 | 435.098 | 449.901 | 3.4% | 3.1% |
| Electricity | 296.887 | 294.939 | 295.911 | 307.673 | 4.0% | 4.4% |
On these estimates, annual inflation went up in October, not down. About 4.0 per cent would be the highest reading since May, when the published rate was 4.2 per cent, and core at 2.8 per cent would also be the highest since May. Food is the largest move in the table, with grocery prices up about 2.5 per cent against 1.8 per cent a month earlier.
Can the estimate be trusted?
Partly, and we can measure how far. The three series that kept their October 2025 data let us run the method and check it against the real answer. The fact sheet's own worked example uses a Washington-area index from 2022, not these.
| Index | Actual Oct 2025 | Approximated Oct 2025 | Published 12-month change | Approximated 12-month change | Miss |
|---|---|---|---|---|---|
| New vehicles | 178.723 | 178.672 | 0.7% | 0.8% | 0.0 points |
| Used cars and trucks | 186.962 | 186.744 | -3.3% | -3.2% | 0.1 points |
| Gasoline (all types) | 277.021 | 280.834 | 37.4% | 35.5% | 1.9 points |
For slow-moving prices the method is close to exact. For gasoline it fails in a predictable direction. Pump prices fell hard from September to October 2025 and only a little from October to November, so the midpoint sits too high and the estimated 12-month rise comes out too low. Gasoline is part of the all-items index, so our 4.0 per cent is, if anything, more likely to be a little low than high. That is our inference from the test, not a figure the bureau has published.
There is a second caution in the data itself. The shelter index barely moved between September and November 2025, from 418.598 to 418.593. The approximation takes that flat stretch at face value, and shelter is the largest single component of core prices.
Two years instead of one
There is a way to avoid the estimate altogether: skip October 2025 and measure over two years. October 2024 was collected normally. From 315.664 then to 337.511 now, all-items prices rose 6.9 per cent, an average of 3.4 per cent a year. Core rose 5.6 per cent over the same two years, or 2.8 per cent a year.
That comparison says something the one-year figure cannot. Core inflation has run at about the same annual pace over two years as our estimate for the latest year, so the underlying rate has not obviously sped up. The all-items rate has, because the second year carries most of the energy shock: our energy estimate is a 20.8 per cent rise in a year, against an average of under 12 per cent a year over two.
The 2026 path, from the bureau's published 12-month rates, shows how much of this year's inflation has followed oil. The all-items rate was 2.4 per cent in January and February, rose to 4.2 per cent in May, eased to 3.4 per cent in July and August, and reached 3.7 per cent in September. We covered that September report when gasoline alone was up 34.2 per cent.
What happens next?
The Federal Reserve meets on 8 and 9 December, according to its calendar, with new economic projections. At its October meeting it held rates at 3.75 to 4 per cent, and two officials voted for an increase. The November price report is not due until 10 December, the day after the decision, so October's figures, with their blank annual column, are the last consumer price index officials will see before they vote.
November's report will not have the same gap. November 2025 was collected, so the bureau can publish a normal 12-month change again. The question for that release is the one this estimate cannot settle: whether the oil price, which the Energy Information Administration expects to hover near $100 a barrel for the rest of the year, keeps the annual rate near 4 per cent. Households have already priced some of it in, with expected inflation of 3.9 per cent in the New York Fed's latest survey.
The October monthly changes, the statement that October 2025 data were not collected, the three published 12-month changes and the date of the next release are from the Bureau of Labor Statistics' Consumer Price Index news release of 10 November 2026. The approximation method and its caveats are from the bureau's fact sheet on approximating missing data. All index levels are from the bureau's published CPI-U series (U.S. city average, not seasonally adjusted, and the seasonally adjusted series for monthly changes), retrieved on 11 November 2026. The Federal Reserve's meeting dates are from its 2026 calendar, and the oil-price outlook from the Energy Information Administration's Short-Term Energy Outlook of 10 November 2026. Every approximated October 2025 index, every approximated 12-month change, the two-year annualised rates and the test of the method are this publication's arithmetic and are unofficial estimates, as the bureau says any such figure must be. The analysis is our own.
Topics moneyinflationconsumer price indexBureau of Labor StatisticsCPIgasolineFederal Reserveeconomic data
