Every register remediation eventually reaches the same fork. You have found that the recorded useful lives, or the assets themselves, do not match reality. The accounting question is not whether to fix it. It is whether fixing it goes forward only or reaches back into periods you already closed and filed.
The answer is not discretionary, and it is not decided by the size of the number. It is decided by a single question: was the information available at the time, and was it used?
The two treatments
A change in accounting estimate is recognised prospectively. ASC 250-10-45-17 expressly prohibits restating or retrospectively adjusting prior-period amounts, or presenting pro forma prior periods. IAS 8.36 requires the same: the effect goes into profit or loss for the period of change, and future periods if they are affected.
A correction of an error is retrospective. Material prior-period errors are corrected by restating the comparative amounts for the periods affected, or by restating opening balances where the error predates the earliest period presented. That is a restatement in the full sense — comparatives move, and the change is disclosed as a correction rather than a refinement.
The gap between those two outcomes is why the classification matters more than almost any other judgement in a register remediation. Our guide to re-assessing useful life covers the prospective case in detail.
How each framework defines an error
ASC 250-10-20 defines an error in previously issued financial statements as an error in recognition, measurement, presentation, or disclosure arising from:
- mathematical mistakes;
- mistakes in the application of generally accepted accounting principles; or
- oversight or misuse of facts that existed at the time the financial statements were prepared.
IFRS frames it through the same lens. IAS 8 defines a prior-period error as an omission or misstatement resulting from a failure to use, or misuse of, reliable information that was available when the financial statements were authorised for issue and that could reasonably be expected to have been obtained and taken into account.
Both definitions point at the same moment in time — the information set that existed when the earlier statements were prepared. Neither asks whether the old number turned out to be accurate. They ask whether it was supportable then.
Worked contrast
Two companies revise a class of production equipment from a ten-year life to a five-year life. Identical adjustment, opposite accounting.
Company A — change in estimate. The ten-year life was set from manufacturer guidance and comparable industry experience. Three years in, a process change put the equipment on continuous duty rather than single-shift, and maintenance records now show accelerated wear. The shorter life reflects developments that did not exist at inception. Prospective. Nothing restates.
Company B — correction of an error. Manufacturer documentation in the original capital file already stated a five-year service life, and the ten-year figure was applied anyway because it matched an existing register convention. Nobody revisited it. The fact existed and was overlooked. That is ASC 250-10-20 language almost word for word. Retrospective restatement, subject to practicability.
Same asset class, same new life, same journal size — and one company files a restatement while the other files a disclosure note.
Why documentation decides the outcome
Notice what separated those two cases: not analysis, but evidence about what was known. That has a practical consequence most teams underestimate.
If your register carries lives that nobody can source — no capital file rationale, no engineering input, no documented convention — you cannot easily demonstrate that the original estimate was reasonable on the information then available. You are not automatically in error territory, but you have lost the ability to prove you are not.
What a defensible reassessment file contains:
- Dated physical verification. Confirmation that the asset exists, where it is, and what condition it is in. You cannot credibly re-life an asset you have not confirmed.
- Nameplate and serial capture. Photographic evidence tying the physical unit to the register record.
- The basis for the new life. Engineering assessment, maintenance history, duty cycle, or manufacturer data — and the date it was obtained.
- What changed since inception. The explicit bridge from the original estimate to the new one. This is the sentence that makes it a change in estimate.
That file is the deliverable, not a by-product of one. It is also what your auditor will ask for first.
A note on immateriality
Not every error triggers a restatement. Both frameworks apply materiality, and immaterial errors are generally corrected in the current period without restating comparatives. But materiality is an assessment to be made and documented, not an assumption to lean on — and it is assessed against the financial statements as a whole, not against the register in isolation.
The bottom line
Most useful-life revisions are legitimate changes in estimate, recognised prospectively, with no comparative touched. That is the normal case and it is well supported by ASC 250-10-45-17 and IAS 8.36.
The exception is narrow but real, and it is defined by what was knowable at the time. The way you stay clearly on the right side of it is to build the evidence file as you go — which is precisely what a documented physical verification produces.
Frequently asked questions
What is the difference between a change in accounting estimate and the correction of an error?+
A change in estimate arises from new information or new developments after a reasonable original judgement, and is recognised prospectively. An error is a misstatement that existed at the time — a mathematical mistake, a misapplication of the framework, or the oversight or misuse of facts already available — and is corrected retrospectively by restating prior periods.
Who decides whether a revision is an estimate change or an error?+
Management makes the initial determination and the auditor evaluates it. The test is objective rather than a matter of preference: it turns on whether reliable information was available at the time the earlier financial statements were authorised, and whether it was used. Contemporaneous documentation is what makes the determination defensible.
Does a large adjustment automatically mean it was an error?+
No. Magnitude is not the test. A very large revision can be a legitimate change in estimate if it genuinely follows new information, and a small revision can be an error correction if the original figure ignored evidence that was already in hand.
What happens if it is impracticable to determine the effect of an error on prior periods?+
IAS 8 permits correction from the earliest period for which retrospective restatement is practicable, with the balance corrected prospectively, and requires disclosure of the circumstances. US GAAP applies similar practicability considerations within its restatement guidance.
How do we stay on the change-in-estimate side of the line?+
Document what you knew and when. A revision supported by dated physical verification, nameplate and serial capture, condition assessment, and engineering or maintenance input reads as a new estimate on new evidence. A revision with no file behind it invites the question of whether the original figure was ever supported.
