The month that never was

There is a hole in the Consumer Price Index. Not a delayed release, not a revision — a month for which no index will ever exist, because the prices that would have made it were never observed.

Months in the series

630from January 1974

Permanent gaps

1October 2025

Across the gap

-0.23%two months of movement, one reading

Effect on the COLA

noneoutside the third quarter

Why it cannot be recovered

A price index is not a model output that can be recomputed. It is an aggregation of prices that field staff observed on particular days in particular outlets. If nobody visited the outlet, there is no observation, and there is nothing to reconstruct from. Statistical imputation can fill a missing quote inside a month where most quotes exist; it cannot invent a month where none do.

The affected reference period fell entirely inside a lapse in appropriations. Collection stopped, and by the time it resumed the reference window had closed.

What the gap actually does

The move from September 2025 to November 2025 is -0.23% — but that is two months of price movement compressed into one observable step, not a monthly change. Any chart that plots the series without acknowledging the gap draws a line that means something different from every other segment on it.

The three things it breaks silently

1. Twelve-month comparisons. Every twelve-month rate anchored on the missing month is undefined, and code that indexes by array position rather than by date will happily return the wrong one.
2. Month-on-month changes. The step across the gap spans two months. Treated as one, it overstates the monthly rate substantially.
3. Any quarterly average containing it. The fourth quarter of that year cannot be averaged at all. It is not that the answer is uncertain — there is no answer.

How this site models it

The missing month is stored as an explicit null — not a zero, not an absent key. The function that averages a quarter throws an error if any member is null rather than skipping it, filling it or defaulting it to zero. There is no code path on this site that can quietly average a quarter containing a month that does not exist.

The twelve-month change columns elsewhere on this site render a dash rather than a number wherever the comparison would cross the gap. That is the honest output, and a dash that looks like an omission is better than a number that looks like a fact.

Why the adjustment survived

The cost-of-living adjustment is decided by July, August and September. The missing month sits outside that window entirely, so the quarter that produced the 2.8% adjustment for 2026 was complete. All six index values behind it are on its page.

The structural reason is the same one that explains the gap: September data is collected before the fiscal year turns, so an appropriations lapse beginning on 1 October can delay the September release but cannot prevent it. The same boundary that saved September is exactly what destroyed October. The nine-day delay, and why it was only a delay.

Common questions

Related

• The CPI-W seriesin full
• Data qualitywhat else changed
• Announcement timingthe 2025 delay
• The 2026 COLAunaffected
• Methodologyhow nulls are handled
• The projectionand the gap