Full explanation
Overview
Every asset on a balance sheet carries a promise. IAS 36 is the standard that checks whether the promise is still true.
Your balance sheet says a machine is worth eight hundred thousand euros. The market says six hundred and twenty thousand. Which number is a lie? IAS 36 answers that — and its answer is neither of them.
Is the asset still worth what the books say?
The whole standard rests on one comparison. On one side, the carrying amount — what the books say the asset is worth today. On the other, the recoverable amount — what you could actually get out of it. If the carrying amount is higher, the asset is overstated, and the difference is an impairment loss. That's it. Everything else in IAS 36 is detail about how you measure those two numbers.
Where IAS 36 applies — and where it steps back
First, where it applies. IAS 36 is the safety net under most non-financial assets — property, plant and equipment, intangibles, goodwill, right-of-use assets you lease in under IFRS 16, and investments in subsidiaries, associates and joint ventures. But it deliberately steps back where another standard already handles falling values: inventories under IAS 2, contract assets under IFRS 15, deferred tax under IAS 12, employee benefits under IAS 19, financial assets under IFRS 9, investment property carried at fair value under IAS 40, and anything classified as held for sale under IFRS 5. If a standard already remeasures the asset downward, IAS 36 doesn't need to.
So how do you work out what an asset is really worth? There are two answers, and you take the better one.
Take the higher of the two
Recoverable amount is the HIGHER of two figures. One: fair value less costs of disposal — what a buyer would pay, minus what it costs you to sell it. IFRS 13 tells you how to measure that fair value; IAS 36 only tells you which amount to use. Two: value in use — the present value of the cash the asset will generate if you keep using it. And the logic behind taking the higher is simple. A rational company will either sell the asset or keep it, whichever leaves it better off. So the asset is worth the better of those two outcomes. You only need both numbers when the first one you calculate comes out below the carrying amount.
Indicators — and the three you test anyway
You don't test everything every year. At each reporting date you look for indicators. External ones: the market value has dropped sharply, interest rates have risen, technology or the market has moved against you, or your net assets now exceed your own market capitalisation. Internal ones: physical damage, obsolescence, the asset is idle, plans to dispose of it early, or internal reporting shows it performing worse than expected. If any indicator exists, you test. But three things are tested every single year regardless of indicators: goodwill, intangibles with an indefinite useful life, and intangibles not yet available for use.
One asset, three numbers
Meridian Limited's machine sits at eight hundred thousand euros. A competitor launches better technology — that's an indicator, so Meridian tests. A buyer offers six hundred and fifty thousand, but disposal costs are thirty thousand, so fair value less costs of disposal is six hundred and twenty thousand. Discounting the cash flows from continuing to run the machine gives a value in use of six hundred and eighty thousand. Recoverable amount is the higher of the two: six hundred and eighty thousand. Carrying amount eight hundred thousand, less recoverable six hundred and eighty thousand, is an impairment loss of one hundred and twenty thousand euros, straight to profit or loss.
Same loss. Different destination.
One variation worth knowing. If the asset had been carried under the revaluation model, the impairment is treated as a revaluation DECREASE. It goes to other comprehensive income first, absorbing any revaluation surplus already sitting in equity for that asset, and only the excess beyond that surplus hits profit or loss. Same loss, different destination — and if you've watched the IAS 16 episode, that's exactly the mechanism you already know.
Value in use sounds straightforward. It isn't — and the constraints are where most people get it wrong.
You can't rescue an asset by assuming you'll fix it
Value in use has rules that stop it becoming wishful thinking. The discount rate must be PRE-tax, and it must reflect the risks specific to that asset. Cash flows are based on the asset in its CURRENT condition — so you exclude the benefit of any future restructuring you haven't committed to, and you exclude improvements that would enhance the asset's performance beyond its present standard. Projections normally run no more than five years on detailed budgets, and beyond that you extrapolate with a growth rate that shouldn't exceed the long-term average for the market. The point of every one of those constraints is the same: you cannot rescue a failing asset by assuming you'll fix it later.
The cash-generating unit
Now a problem. Most assets don't generate cash on their own. A single oven in a bakery earns nothing by itself — the shop earns. So when you can't measure recoverable amount for an individual asset, you move up to its cash-generating unit: the SMALLEST identifiable group of assets that generates cash inflows largely independent of other assets. Smallest matters. The bigger the unit you choose, the easier it is for a strong asset to hide a weak one — which is precisely why the standard forces you down to the smallest level that works.
Goodwill absorbs it first
Goodwill never generates cash on its own, so it's allocated to the cash-generating units expected to benefit from the acquisition — tested at the level management monitors it, and never at a level larger than an operating segment. Then, when a unit is impaired, the loss is allocated in a strict order. Meridian's northern division: goodwill two hundred thousand, a building nine hundred thousand, equipment three hundred thousand — one million four hundred thousand carrying. Recoverable amount is one million fifty thousand, so the loss is three hundred and fifty thousand. Goodwill absorbs it FIRST — all two hundred thousand, wiped to nil. The remaining one hundred and fifty thousand goes pro rata across the other assets: one hundred and twelve thousand five hundred to the building, thirty-seven thousand five hundred to the equipment. That leaves the building at seven hundred and eighty-seven thousand five hundred, with the equipment at two hundred and sixty-two thousand five hundred. And there's a floor — you never write an individual asset below the highest of its own fair value less costs of disposal, its own value in use, and zero.
Now the part that trips people up in exams and in practice. What happens when the asset recovers?
You may undo it. You may not profit from it.
Impairments can reverse — with one absolute exception. If the indicators that caused the loss have gone, you reverse it, but only up to a ceiling: the carrying amount the asset WOULD have had, net of normal depreciation, if it had never been impaired at all. Meridian's machine was impaired to six hundred and eighty thousand with four years left, so depreciation is one hundred and seventy thousand a year, leaving five hundred and ten thousand after one year. Had it never been impaired, eight hundred thousand over four years is two hundred thousand a year, so it would have stood at six hundred thousand. Recoverable amount has recovered to seven hundred thousand — but the ceiling is six hundred thousand. So the reversal is ninety thousand, not one hundred and ninety thousand. You may undo an impairment; you may never use one to create a gain.
Goodwill impairment is never reversed
And the exception: goodwill impairment is NEVER reversed. Ever. The reasoning is that any apparent recovery is internally generated goodwill — and IAS 38 forbids recognising that. There's a sharp corollary in IFRIC 10: if you recognise a goodwill impairment in a half-year interim report, you cannot reverse it at the year end, even if on full-year figures you would never have booked it. The interim number stands. It's a rule that catches out companies with volatile first halves.
The board looked again — and kept the model
This is also the most actively debated standard the board maintains. Critics argue goodwill impairments arrive too late to be useful — that the test is complex, expensive, and that management optimism delays the write-down. In March two thousand and twenty-four the board published an exposure draft on business combinations, goodwill and impairment. Its headline decision: the impairment-only model STAYS. Goodwill amortisation is not coming back. Instead it proposes targeted relief — removing the bar on cash flows from future restructurings and enhancements, and permitting post-tax inputs — plus new disclosures on whether an acquisition actually delivered what management promised. One thing that stays exactly as it is: goodwill still can't be tested above the level of an operating segment. It's still in redeliberation, so nothing has changed yet.
Which brings us to what actually reaches the person reading the accounts.
The assumptions matter more than the number
The disclosures are where impairment becomes readable. For each material loss you show the amount, which line of the income statement it sits in, and the events that caused it. For cash-generating units carrying goodwill or indefinite-life intangibles you disclose the key assumptions, the discount rate, and the growth rate — plus sensitivity, if a reasonably possible change would trigger a further impairment. A 2013 amendment tightened this: after IFRS 13 was issued, entities were briefly required to disclose recoverable amount for every unit carrying goodwill. That was an unintended consequence, and it was narrowed back to units actually impaired, with fair value hierarchy disclosures where recoverable amount is fair value based.
IFRS 18 — impairment lands in operating
From the first of January twenty twenty-seven, IFRS 18 changes where the loss appears. Income and expenses split into operating, investing and financing categories — and an impairment of an asset used in the business sits in OPERATING. So it lands inside operating profit, the very subtotal IFRS 18 finally defines. It also means an impairment can no longer be quietly presented as something below the operating line. And because IFRS 18 applies retrospectively, the 2026 comparatives are already being restated.
What a write-down is really telling you
So what should an investor do with all this? Three things. A large goodwill impairment is an admission that an acquisition didn't deliver what was promised — read it as a verdict on management's capital allocation, not as a one-off accounting entry. Watch the assumptions, not just the number: a discount rate that drifts down or a growth rate that drifts up, year after year, is a company working hard to avoid a write-down. And remember it's asymmetric — goodwill impairments can only ever go one way, so an unimpaired balance is not evidence that the acquisition worked, only that it hasn't yet been admitted otherwise.
Every euro traceable to a rule
On Meridian's accounts: the machine written from eight hundred thousand down to six hundred and eighty thousand, a one hundred and twenty thousand euro charge in profit or loss. The northern division's goodwill eliminated entirely — two hundred thousand gone — with a further one hundred and fifty thousand spread across the building and equipment, leaving them at seven hundred and eighty-seven thousand five hundred and two hundred and sixty-two thousand five hundred. And a year later, ninety thousand of the machine's impairment reversed — capped, deliberately, at the value it would have held had the loss never happened.