The October inflation report came out on Tuesday without the figure most people look for: the inflation rate over the past year. The Bureau of Labor Statistics couldn’t calculate it. Our estimate is that prices rose about 3.8% to 3.9% in the year to October, a little faster than September’s 3.7% and short of the 4.0% that a simple fix gives.

Why October has no 12-month number

A 12-month rate divides this October’s price index by last October’s. Last October doesn’t exist. The government shutdown in the fall of 2025 stopped the price survey that month, so the Bureau says it “cannot publish 12-month percent changes for October 2026.”

The monthly figures are complete. Prices rose 0.4% in October after seasonal adjustment, and core prices, which leave out food and energy, rose 0.3%. What’s missing is the yardstick most people, and many contracts, use.

Two stand-ins, two answers

Anyone who needs the annual number has to invent last October. A Bureau fact sheet describes one way: take the months on either side, September and November 2025, and use the point midway between them. Measured from that stand-in, prices rose 4.02% in a year.

Treasury had to invent the same month a year ago, because inflation-protected Treasury bonds (TIPS) adjust their value with each month’s price index. Its rule took September 2025 and added one month’s share of the past year’s inflation. That index is final for TIPS and won’t be revised. Measured from it, prices rose 3.66%.

The two answers sit 0.37 point apart because of one detail. Before seasonal adjustment, prices slipped 0.21% between September and November 2025, while Treasury’s rule assumed a rise of about 0.25% in October. Set last October too high and this year’s rise looks smaller.

The 12-month rate depends on which stand-in you use.

Stand-in for October 2025	Index	Change to October 2026
Midpoint of Sept. and Nov. 2025 (BLS fact sheet)	324.46	4.02%
Treasury’s TIPS index	325.60	3.66%
October 2026, published	337.51	end point

Source: BLS; Treasury via Tipswatch; MKViz calculation

What four decades of Octobers say

Each fix has a habit, and the past shows it. For every October from 1985 to 2023, we hid the real number, filled it each way, and compared the result with what was actually published.

The midpoint ran high in 31 of 39 years. Before seasonal adjustment, prices tend to peak in October and level off in November, so October usually sits above the middle of its neighbors. This year the midpoint has a second reason to run high: with no rent survey in October 2025, the Bureau held rents flat, which left the November index a little low.

Treasury’s rule ran low on average, and it swung much wider. In 2008, when gas prices collapsed that fall, it was off by 1.4 points.

Take out each method’s usual October error, and the two answers move toward each other.

Corrected for past October misses, both land just above September’s rate.

Source: BLS; MKViz calculation

I’d put October at about 3.8% to 3.9%, with the higher end a bit less likely because of the rent hold. That’s at most a small step up from September, not a jump to 4%.

Core inflation comes out at about 2.4% by Treasury’s rule and 2.7% by the midpoint, around September’s 2.6%.

When the number is written into a contract

For my household budget, the difference between these estimates is noise. Social Security doesn’t need it either: its 3.6% raise for 2027 was set in October from July-to-September prices. The same goes for most pay raises, which no formula binds.

A contract is different. Many union contracts and pension plans, and some commercial leases, raise payments by the change in the unadjusted index over a set year. If yours runs October to October, the clause has nothing official to point to this year.

The Bureau allows its midpoint fill only “if all parties agree.” Before the bill is due, check whether your clause names a fallback, such as another month or a carry-forward rule. If it doesn’t, agree on a stand-in and put it in writing. On a $2,000 monthly rent, the gap between the two is about $7.31 a month.

What the Fed reads instead

The Federal Reserve, which targets a different price index, doesn’t need this annual number to decide. Core prices rose 0.3% in October, and the Fed’s Christopher Waller had said that a reading around that size could put a half-point increase on the table. After the report, traders treated at least a quarter-point rate hike at the Dec. 8-9 meeting as a sure thing.

My mortgage rate is fixed, so a December hike won’t touch it. What did move was the variable rate on my credit line: by September it had climbed more than the central bank’s own increases. That month I paid it off with cash that was earning less than half as much. If you carry a balance at a variable rate, its rate matters more for you before Dec. 9 than any decimal on October’s inflation.

Sources

Method: We filled the missing October 2025 CPI-U (not seasonally adjusted) two ways, with the geometric mean of September and November 2025 as a BLS fact sheet describes, and with Treasury's rule for inflation-protected bonds (September 2025 raised by one-twelfth of the prior 12-month change), then measured the change to October 2026. To test each, we treated every October from 1985 to 2023 as missing, filled it the same way, and compared the 12-month rate with the real one; we subtracted each method's average October error. Data: BLS via our own download. Code and data

4 more sources