Same average, different exposure. An Evidence Press briefing.

When governments publish an average debt maturity, what does that tell us about the cost of higher interest rates next year? Less than it might seem.

Imagine one hundred billion pounds of fixed-rate debt. In one portfolio, all of it matures in four years. In another, half matures now and half in eight years. The total and average maturity are identical. But if refinancing rates rise by one percentage point, the first portfolio has no extra interest cost in year one. The second has half a billion pounds, under the stated refinancing model.

This paper computes the lowest and highest costs compatible with a published debt summary. It uses established mathematical methods and exact arithmetic certificates. The package contains four thousand five hundred and seventy-six certified endpoints, not that many new theorems.

The analysis also asks which additional information removes the uncertainty. Annual repayment amounts and their average dates can determine particular annual costs. Other reporting packages contain several numbers, and their collection costs have not been compared.

An important accounting correction limits the interpretation. The historical purchase cost of bonds held by the Bank of England's Asset Purchase Facility is not its current loan balance. Net financing also depends on cash and how the accounts evolve. A dated observation from the seventh of October is usable now; the problem is missing matched information, not an unfinished month.

The revised paper verifies the relevant fields of one hundred and five gilts against an official export and reconciles a separate February accounting example. It does not supply a complete UK fiscal forecast. Its earlier funding scenarios remain explicitly hypothetical.

The paper, code and exact model checks are linked on Evidence Press. This is an unrefereed candidate dated the thirteenth of October, twenty twenty-six. The British narration uses a synthetic voice and is not additional scientific evidence.