Impairment of Assets Under IAS 36: Indicators, Cash-Generating Units, Value in Use and Goodwill

"The objective of this Standard is to prescribe the procedures that an entity applies to ensure that its assets are carried at no more than their recoverable amount. An asset is carried at more than its recoverable amount if its carrying amount exceeds the amount to be recovered through use or sale of the asset. If this is the case, the asset is described as impaired and the Standard requires the entity to recognise an impairment loss."

Two words in that paragraph do the work. "Use or sale" is the reason recoverable amount has two limbs rather than one: the entity is not required to write down an asset it can still use profitably merely because nobody would pay much for it, nor to write one down because it is unprofitable in use if it could be sold for more. And "carried at more than" fixes the direction. IAS 36 is a ceiling test only. It never writes an asset up above cost, and outside the reversal rules in IAS 36.114 to 36.123 it never writes one up at all.

"If, and only if, the recoverable amount of an asset is less than its carrying amount, the carrying amount of the asset shall be reduced to its recoverable amount. That reduction is an impairment loss." The loss is recognised immediately in profit or loss, unless the asset is carried at revalued amount under another standard, in which case IAS 36.60 treats it as a revaluation decrease.

The "if, and only if" is not decorative. It removes any scope for a general or precautionary write-down, and it removes any scope for smoothing a loss across periods. An impairment is recognised in full in the period in which recoverable amount falls below carrying amount, and it is measured by that shortfall, not by management's assessment of how much of the shortfall is permanent.

"After the recognition of an impairment loss, the depreciation (amortisation) charge for the asset shall be adjusted in future periods to allocate the asset's revised carrying amount, less its residual value (if any), on a systematic basis over its remaining useful life."

This is the step that gets skipped. The impairment journal is posted to the ledger, the fixed asset register keeps the original depreciation profile, and the following year's charge is wrong on every impaired asset. It also has a presentation consequence. Where the impairment charge is presented as an adjusting or non-underlying item, the depreciation saving it creates in later years flows through underlying profit unless it is tracked, which flatters every subsequent period by exactly the amount the write-down was supposed to cost.

"This Standard shall be applied in accounting for the impairment of all assets, other than: (a) inventories (see IAS 2); (b) contract assets and assets arising from costs to obtain or fulfil a contract that are recognised in accordance with IFRS 15; (c) deferred tax assets (see IAS 12); (d) assets arising from employee benefits (see IAS 19); (e) financial assets that are within the scope of IFRS 9; (f) investment property that is measured at fair value (see IAS 40); (g) biological assets related to agricultural activity within the scope of IAS 41 that are measured at fair value less costs to sell; (h) contracts within the scope of IFRS 17 that are assets and any assets for insurance acquisition cash flows as defined in IFRS 17; and (i) non-current assets (or disposal groups) classified as held for sale in accordance with IFRS 5."

IAS 36.3 explains why the list exists: each of those items sits in a standard that already contains its own recognition and measurement requirements. The practical value of the list is therefore negative. It tells you what not to test, and it is the answer to a large share of the questions people ask about impairment, because most of them are about receivables, loans or inventory and none of those is an IAS 36 question.

IAS 36.4 states that the Standard applies to financial assets classified as subsidiaries under IFRS 10, associates under IAS 28 and joint ventures under IFRS 11, adding that "For impairment of other financial assets, refer to IFRS 9."

That is the inclusion people miss. An investment in a subsidiary held in the parent's separate financial statements looks like a financial asset and is not tested like one. It is tested as an individual asset under IAS 36 against its recoverable amount, and IAS 36.12(h) gives it an indicator that exists for nothing else in the standard. Where a parent has been taking dividends out of a subsidiary faster than the subsidiary earns them, that alone is an indicator.

IAS 36.5 brings revalued assets in as well, noting that the only difference between an asset's fair value and its fair value less costs of disposal is the direct incremental cost of disposal. If disposal costs are negligible, a revalued asset's recoverable amount is necessarily close to or above its revalued amount, so impairment is unlikely. Where disposal costs are significant, a revalued asset can still be impaired.

"An entity shall assess at the end of each reporting period whether there is any indication that an asset may be impaired. If any such indication exists, the entity shall estimate the recoverable amount of the asset." (IAS 36.9)

IAS 36.10 then adds the annual tests, and specifies that the annual test of an indefinite life intangible or an intangible not yet available for use "may be performed at any time during an annual period, provided it is performed at the same time every year", with different intangibles tested at different times. An intangible first recognised in the current annual period must be tested before the end of that period.

IAS 36.11 explains the logic for the third category: an intangible not yet available for use carries more uncertainty about generating enough benefit to recover its carrying amount than one already in use, which is why it earns a mandatory annual test even though it has no goodwill attached.

IAS 36.15 applies the ordinary materiality filter: an indicator that would not lead to a material adjustment does not force a full calculation of recoverable amount. This is a real relief and it is regularly over-claimed. The materiality judgement is against the effect on the financial statements, not against the cost or inconvenience of running the model.

IAS 36.17 carries a requirement that is not about impairment at all and is missed more often than any other sentence in the standard. Even where no impairment loss is recognised, the presence of an indicator signals that the remaining useful life, the depreciation or amortisation method, and the residual value may need to be reviewed and adjusted under the standard applicable to the asset. An indicator that produces no write-down should still produce a review.

The external indicators are: "(a) there are observable indications that the asset's value has declined during the period significantly more than would be expected as a result of the passage of time or normal use. (b) significant changes with an adverse effect on the entity have taken place during the period, or will take place in the near future, in the technological, market, economic or legal environment in which the entity operates or in the market to which an asset is dedicated. (c) market interest rates or other market rates of return on investments have increased during the period, and those increases are likely to affect the discount rate used in calculating an asset's value in use and decrease the asset's recoverable amount materially. (d) the carrying amount of the net assets of the entity is more than its market capitalisation."

Indicator (c) is the one that catches companies out in a rising rate environment, because nothing has happened to the business at all. The forecast is unchanged, the assets are unchanged, and the test still fails because the denominator moved. Indicator (d) is the one that is hardest to act on, because it is an observation about the whole entity while the test is performed on units.

The internal indicators are evidence of obsolescence or physical damage, significant adverse changes in the extent or manner in which an asset is or is expected to be used, which IAS 36.12(f) expands to include "the asset becoming idle, plans to discontinue or restructure the operation to which an asset belongs, plans to dispose of an asset before the previously expected date, and reassessing the useful life of an asset as finite rather than indefinite", and internal reporting indicating that economic performance is or will be worse than expected.

IAS 36.14 gives the internal reporting evidence content: cash flows for acquiring, operating or maintaining the asset significantly higher than budgeted; actual net cash flows or operating results significantly worse than budgeted; a significant decline in budgeted net cash flows; or operating losses or net cash outflows when current period amounts are aggregated with budgeted future amounts.

The last of those deserves attention. It is not "the unit lost money this year". It is the aggregation of the current year with the budget, which means a unit forecasting a recovery is still caught if the recovery does not cover the losses already incurred.

A separate indicator applies to an investment in a subsidiary, joint venture or associate: the investor recognises a dividend from the investment and evidence is available that either "the carrying amount of the investment in the separate financial statements exceeds the carrying amounts in the consolidated financial statements of the investee's net assets, including associated goodwill", or "the dividend exceeds the total comprehensive income of the subsidiary, joint venture or associate in the period the dividend is declared".

In plain terms, taking cash out of a business faster than the business earns it is a statement about the value of the investment. This indicator sits in the parent's own accounts, where nobody usually looks, and it triggers on a transaction the group treats as routine.

"This Standard defines recoverable amount as the higher of an asset's or cash-generating unit's fair value less costs of disposal and its value in use." IAS 36.19 then adds: "It is not always necessary to determine both an asset's fair value less costs of disposal and its value in use. If either of these amounts exceeds the asset's carrying amount, the asset is not impaired and it is not necessary to estimate the other amount."

IAS 36.19 saves more audit and preparer time than any other sentence in the standard, and it is the reason a well-run impairment process starts by asking which of the two measures is easier to support rather than by building a discounted cash flow model by reflex. For a property-heavy unit with a recent valuation, fair value less costs of disposal answers the question in an afternoon.

IAS 36.20 permits value in use alone to serve as recoverable amount where fair value less costs of disposal cannot be measured "because there is no basis for making a reliable estimate of the price at which an orderly transaction to sell the asset would take place between market participants at the measurement date under current market conditions".

IAS 36.21 permits the opposite where there is no reason to believe value in use materially exceeds fair value less costs of disposal, and notes this "will often be the case for an asset that is held for disposal", because the value in use of such an asset consists mainly of the net disposal proceeds anyway.

Note what IAS 36.20 does not say. It does not permit fair value less costs of disposal to be ignored because measuring it is difficult, or because no valuation has been commissioned. The condition is that there is no basis for a reliable estimate, which is a statement about the market rather than about the entity's budget for valuers.

"Recoverable amount is determined for an individual asset, unless the asset does not generate cash inflows that are largely independent of those from other assets or groups of assets. If this is the case, recoverable amount is determined for the cash-generating unit to which the asset belongs." IAS 36.22 then gives two exits: the individual test can still be used where the asset's fair value less costs of disposal is higher than its carrying amount, or where value in use can be estimated to be close to fair value less costs of disposal and the latter can be measured.

IAS 36.67 explains when the individual test fails: where value in use cannot be estimated to be close to fair value less costs of disposal, for example because the asset's continuing use generates cash inflows that are not largely independent, or where the asset does not generate largely independent inflows at all.

"As defined in paragraph 6, an asset's cash-generating unit is the smallest group of assets that includes the asset and generates cash inflows that are largely independent of the cash inflows from other assets or groups of assets. Identification of an asset's cash-generating unit involves judgement."

The standard's own example, sitting immediately after IAS 36.68, is the cleanest illustration in the document. A bus company operates five routes under a contract with a municipality requiring a minimum service on each. Assets and cash flows for each route are separately identifiable, and one route runs at a significant loss. Because the entity has no option to curtail any single route, the lowest level of largely independent inflows is the five routes together, and the cash-generating unit for each route is the bus company as a whole.

Read that carefully, because it is easy to invert. Separate revenue data does not create a separate unit. What creates a separate unit is the ability to generate, or to stop generating, cash independently. IAS 36.69 adds that inflows means inflows from parties external to the entity, and directs attention to how management monitors operations, by product line, business, individual location, district or region, and how it makes decisions about continuing or disposing of assets and operations.

IAS 36.70 deals with internal markets: where an active market exists for the output produced by a group of assets, that group is a cash-generating unit even if some or all of the output is used internally, and the cash flows are built on management's best estimate of arm's length prices.

IAS 36.72 requires units to be "identified consistently from period to period for the same asset or types of assets, unless a change is justified". Where the aggregation changes and a loss is recognised or reversed, IAS 36.130(d)(iii) requires disclosure of the current and former way of aggregating assets and the reasons for the change.

Those two paragraphs together are the control on the single most effective way of avoiding an impairment, which is redrawing the unit. It is rarely deliberate, because the wider unit usually corresponds to a genuine management reporting line that has itself changed. The disclosure requirement makes it visible anyway.

"The carrying amount of a cash-generating unit shall be determined on a basis consistent with the way the recoverable amount of the cash-generating unit is determined." IAS 36.76 then specifies that the carrying amount "includes the carrying amount of only those assets that can be attributed directly, or allocated on a reasonable and consistent basis, to the cash-generating unit and will generate the future cash inflows used in determining the cash-generating unit's value in use", and "does not include the carrying amount of any recognised liability, unless the recoverable amount of the cash-generating unit cannot be determined without consideration of this liability".

This is the symmetry rule, and it is the single sentence in IAS 36 that is broken most often. Every argument about what belongs in the carrying amount is in substance an argument about what is in the cash flows. Get the two out of step and the comparison is meaningless, in either direction: an asset in the carrying amount whose cash flows are excluded produces a false impairment, and a liability deducted from carrying amount whose payments stay in the cash flows produces a false pass.

IAS 36.78 sets out the exception. Where the disposal of a unit would require the buyer to assume a liability, and only one measure of recoverable amount is available for the unit and the liability together, the liability is deducted from both sides. The standard's own illustration is a mine with a legal obligation to restore the site: a buyer would not take the mine without the obligation, so the fair value less costs of disposal of the unit is measured net of it, and the restoration provision is deducted from the carrying amount too.

IAS 36.79 adds a practical concession for the same problem in reverse: for practical reasons, recoverable amount is sometimes determined after considering assets that are not part of the unit, such as receivables, or liabilities that have been recognised, such as payables or pensions, and in those cases the carrying amount is increased or decreased by the same amounts.

Both paragraphs say the same thing in different directions. Adjust both sides or neither.

IAS 36.30 requires five elements to be reflected: "(a) an estimate of the future cash flows the entity expects to derive from the asset; (b) expectations about possible variations in the amount or timing of those future cash flows; (c) the time value of money, represented by the current market risk-free rate of interest; (d) the price for bearing the uncertainty inherent in the asset; and (e) other factors, such as illiquidity, that market participants would reflect in pricing the future cash flows."

IAS 36.32 then makes the practical point that elements (b), (d) and (e) can be reflected either as adjustments to the cash flows or as adjustments to the discount rate. That is the origin of the double-counting problem in unit 9. Risk goes in one place or the other. Putting a conservative haircut on the cash flows and a risk premium in the rate charges for the same risk twice, and the result is an impairment the standard does not require.

"In measuring value in use an entity shall: (a) base cash flow projections on reasonable and supportable assumptions that represent management's best estimate of the range of economic conditions that will exist over the remaining useful life of the asset. Greater weight shall be given to external evidence. (b) base cash flow projections on the most recent financial budgets/forecasts approved by management, but shall exclude any estimated future cash inflows or outflows expected to arise from future restructurings or from improving or enhancing the asset's performance. Projections based on these budgets/forecasts shall cover a maximum period of five years, unless a longer period can be justified. (c) estimate cash flow projections beyond the period covered by the most recent budgets/forecasts by extrapolating the projections ... using a steady or declining growth rate for subsequent years, unless an increasing rate can be justified. This growth rate shall not exceed the long-term average growth rate for the products, industries, or country or countries in which the entity operates, or for the market in which the asset is used, unless a higher rate can be justified."

Three separate constraints sit in that one paragraph. The starting point must be the approved budget, not a model built for the impairment test. The explicit period is capped at five years. The terminal growth rate is capped at long-run average growth. All three are breachable with justification, and very few justifications survive contact with an audit file. A terminal growth rate above long-run growth for the relevant market is a claim that the unit outgrows its economy in perpetuity, and that claim needs evidence rather than a footnote.

IAS 36.34 adds the discipline nobody applies: management should assess the reasonableness of its current assumptions "by examining the causes of differences between past cash flow projections and actual cash flows".

"Future cash flows shall be estimated for the asset in its current condition. Estimates of future cash flows shall not include estimated future cash inflows or outflows that are expected to arise from: (a) a future restructuring to which an entity is not yet committed; or (b) improving or enhancing the asset's performance."

IAS 36.45 makes both limbs symmetrical: value in use does not reflect the cost savings or benefits expected from an uncommitted restructuring, and it does not reflect the outflows that would improve performance or the inflows expected to arise from them.

The enhancement rule is the one the FRC named in its 2024/25 review. You are valuing the asset you own today, not the asset the investment plan will eventually create. Maintenance capital expenditure stays in, because it sustains the current standard of performance. Growth capital expenditure comes out, and the growth comes out with it.

Preparers find this hard for a structural reason. IAS 36.33(b) requires the approved budget as the starting point, and the approved budget is almost always a growth plan. The adjustment from board budget to IAS 36 cash flows is therefore a real reconciliation, and it should exist as a document rather than as an assertion that the budget was adjusted.

IAS 36.39 defines what is in: inflows from continuing use, outflows necessarily incurred to generate those inflows including outflows to prepare the asset for use, and the net cash flows on disposal at the end of the useful life. IAS 36.50 defines what is out: "(a) cash inflows or outflows from financing activities; or (b) income tax receipts or payments."

IAS 36.51 explains why: "Estimated future cash flows reflect assumptions that are consistent with the way the discount rate is determined. Otherwise, the effect of some assumptions will be counted twice or ignored. Because the time value of money is considered by discounting the estimated future cash flows, these cash flows exclude cash inflows or outflows from financing activities. Similarly, because the discount rate is determined on a pre-tax basis, future cash flows are also estimated on a pre-tax basis."

IAS 36.40 completes the consistency set for inflation. A nominal discount rate requires nominal cash flows. A real rate requires real cash flows, which still carry specific price increases and decreases even though general inflation has been stripped out.

"The discount rate (rates) shall be a pre-tax rate (rates) that reflect(s) current market assessments of: (a) the time value of money; and (b) the risks specific to the asset for which the future cash flow estimates have not been adjusted."

IAS 36.56 defines that rate as "the return that investors would require if they were to choose an investment that would generate cash flows of amounts, timing and risk profile equivalent to those that the entity expects to derive from the asset", estimated from the rate implicit in current market transactions for similar assets, or from the weighted average cost of capital of a listed entity holding a similar single asset or portfolio. It closes with the warning: the rate "shall not reflect risks for which the future cash flow estimates have been adjusted. Otherwise, the effect of some assumptions will be double-counted."

Appendix A gives the practical route. IAS 36.A17 permits the entity's weighted average cost of capital determined using techniques such as the capital asset pricing model, its incremental borrowing rate, or other market borrowing rates as a starting point. IAS 36.A18 requires those rates to be adjusted to reflect how the market would assess the specific risks of the asset's cash flows and to exclude risks not relevant to them or already reflected in the cash flows, naming country risk, currency risk and price risk.

IAS 36.A19 states that the rate "is independent of the entity's capital structure and the way the entity financed the purchase of the asset". A company with cheap debt does not get an easier impairment test.

IAS 36.A20 is the sentence that causes all the trouble: "Paragraph 55 requires the discount rate used to be a pre-tax rate. Therefore, when the basis used to estimate the discount rate is post-tax, that basis is adjusted to reflect a pre-tax rate." It requires an adjustment. It does not specify the method.

IAS 36.A21 permits a single rate normally, and separate rates for different future periods where value in use is sensitive to a difference in risk across periods or to the term structure of interest rates.

IAS 36.28 defines costs of disposal as the direct incremental costs attributable to the disposal of the asset, excluding finance costs and income tax expense, and excludes costs already recognised as liabilities so that they are not deducted twice. Fair value itself is measured under IFRS 13, which means market participant assumptions rather than the entity's own intentions.

That distinction is the point of having two measures. Value in use is entity-specific and captures synergies and plans that only this entity has. Fair value less costs of disposal is what a buyer would pay, and it is not constrained by IAS 36.44, so a market participant's expectations about restructuring or expansion can be reflected where a buyer would price them in. Those are genuinely different questions with genuinely different answers, and the standard takes the higher.

Where recoverable amount is fair value less costs of disposal, IAS 36.130(f) requires the level of the IFRS 13 fair value hierarchy in which the measurement falls, the valuation technique used for Level 2 and Level 3 measurements, and any change in technique with the reasons. IAS 36.134(e) adds, for units carrying significant goodwill or indefinite life intangibles, each key assumption, how management determined it, and whether it reflects past experience or external sources.

Those requirements are what stop the fair value route being a way of disclosing less. A Level 3 fair value less costs of disposal model built on management's own budget with a market participant label attached is the version that does not survive scrutiny, and the disclosure requirements are drafted precisely to expose it.

"An impairment loss shall be recognised for a cash-generating unit (the smallest group of cash-generating units to which goodwill or a corporate asset has been allocated) if, and only if, the recoverable amount of the unit (group of units) is less than the carrying amount of the unit (group of units). The impairment loss shall be allocated to reduce the carrying amount of the assets of the unit (group of units) in the following order: (a) first, to reduce the carrying amount of any goodwill allocated to the cash-generating unit (group of units); and (b) then, to the other assets of the unit (group of units) pro rata on the basis of the carrying amount of each asset in the unit (group of units)."

The order is not a presentational convention. Because IAS 36.124 prohibits reversal of a goodwill impairment, the sequence determines how much of the loss is permanently locked in. A unit that takes its loss entirely against goodwill can never recover any of it, whatever happens next. The same unit taking the same loss against plant can recover up to the IAS 36.117 ceiling.

"In allocating an impairment loss in accordance with paragraph 104, an entity shall not reduce the carrying amount of an asset below the highest of: (a) its fair value less costs of disposal (if measurable); (b) its value in use (if determinable); and (c) zero. The amount of the impairment loss that would otherwise have been allocated to the asset shall be allocated pro rata to the other assets of the unit (group of units)."

IAS 36.106 explains why the pro rata mechanism exists at all: where it is not practicable to estimate the recoverable amount of each individual asset, the standard accepts an arbitrary allocation between the assets of the unit other than goodwill, because all assets of a unit work together. The allocation is therefore not a valuation of each asset. It is a convention, with a protective floor bolted on.

The floor is the step most often skipped, and skipping it produces a file where the total loss is right and every individual line is wrong. That survives review because the control total ties, and it corrupts the depreciation charge on every affected asset for the rest of its life.

"For the purpose of impairment testing, goodwill acquired in a business combination shall, from the acquisition date, be allocated to each of the acquirer's cash-generating units, or groups of cash-generating units, that is expected to benefit from the synergies of the combination, irrespective of whether other assets or liabilities of the acquiree are assigned to those units or groups of units. Each unit or group of units to which the goodwill is so allocated shall: (a) represent the lowest level within the entity at which the goodwill is monitored for internal management purposes; and (b) not be larger than an operating segment as defined by paragraph 5 of IFRS 8 Operating Segments before aggregation."

The two conditions pull in opposite directions and that tension is the design. The internal monitoring test pushes the level down to wherever management actually watches the acquisition. The operating segment ceiling stops it being pushed up into a group so large that no acquisition could ever fail. IAS 36.81 explains why groups of units are permitted at all: goodwill often contributes to the cash flows of several units and sometimes cannot be allocated to individual units on a non-arbitrary basis.

IAS 36.82 adds a point that is useful in a conversation with management: applying paragraph 80 results in goodwill being tested at the level that reflects the way the entity manages its operations, so the development of additional reporting systems is typically not necessary. If the answer requires a new reporting system, the level is probably wrong.

Note also the phrase "irrespective of whether other assets or liabilities of the acquiree are assigned". Goodwill from an acquisition can and often should sit in a legacy unit that benefits from the synergies, even though none of the acquired assets went there.

IAS 36.84 requires the initial allocation to be completed before the end of the first annual period beginning after the acquisition date where it cannot be completed in the year of acquisition itself. IAS 36.86 deals with disposals: where an operation within a unit is disposed of, the goodwill associated with that operation is included in the carrying amount of the operation when determining the gain or loss, measured on relative values unless the entity can demonstrate a better method. IAS 36.87 requires goodwill to be reallocated on a relative value basis when an entity reorganises its reporting structure in a way that changes the composition of units.

Those two paragraphs are the maintenance requirements, and they are the ones that decay. Goodwill from an acquisition eight years ago, in a group reorganised twice since, with no reallocation on either occasion, sits in a unit that is now large enough that it can never fail. Nothing in the accounts flags that, because each individual year looks unremarkable.

IAS 36.90 sets the test: "A cash-generating unit to which goodwill has been allocated shall be tested for impairment annually, and whenever there is an indication that the unit may be impaired, by comparing the carrying amount of the unit, including the goodwill, with the recoverable amount of the unit."

IAS 36.88 covers unallocated goodwill: where goodwill relates to a unit but has not been allocated to it, that unit is tested first without the goodwill whenever an indicator exists, and the larger group of units including the goodwill is then tested as well.

IAS 36.98 fixes the order of testing within a unit. Where there is an indication that an individual asset inside the unit is impaired, that asset is tested and written down first, before the unit containing the goodwill is tested. The same bottom-up sequence applies between a unit and a group of units. Getting the order wrong shelters an impaired asset behind the unit's headroom.

IAS 36.96 permits the annual test at any time in the year provided it is at the same time each year, with different units tested at different times. IAS 36.99 allows a prior period calculation to be reused only where the assets and liabilities of the unit have not changed significantly, the last calculation gave a substantial margin, and the likelihood that a current calculation would fall below carrying amount is remote.

"Corporate assets include group or divisional assets such as the building of a headquarters or a division of the entity, EDP equipment or a research centre. The structure of an entity determines whether an asset meets this Standard's definition of corporate assets for a particular cash-generating unit. The distinctive characteristics of corporate assets are that they do not generate cash inflows independently of other assets or groups of assets and their carrying amount cannot be fully attributed to the cash-generating unit under review."

IAS 36.102 sets the two-branch test. Where a portion of the carrying amount can be allocated to the unit on a reasonable and consistent basis, allocate it and compare the enlarged carrying amount with recoverable amount. Where it cannot, compare the unit's carrying amount with its recoverable amount without the corporate asset, identify the smallest group of units that includes the unit and to which the corporate asset can be allocated on a reasonable and consistent basis, and then compare the carrying amount of that group, including the allocated portion, with its recoverable amount.

The structure mirrors the unallocated goodwill test in IAS 36.88, for the same reason. An asset serving several units has to be tested against all of them, and the second comparison can produce a loss the first did not.

"An entity shall assess at the end of each reporting period whether there is any indication that an impairment loss recognised in prior periods for an asset other than goodwill may no longer exist or may have decreased. If any such indication exists, the entity shall estimate the recoverable amount of that asset."

IAS 36.111 mirrors the impairment indicators in reverse: observable indications that value has increased significantly, favourable changes in the technological, market, economic or legal environment, decreases in market interest rates likely to increase recoverable amount materially, favourable changes in the extent or manner in which the asset is used, and internal evidence that economic performance is or will be better than expected.

The obligation is symmetrical and it is regularly ignored. Companies that impaired quickly during a downturn are often slow to run the reversal assessment when conditions recover, and IAS 36.110 gives them no discretion about whether to look.

"An impairment loss recognised in prior periods for an asset other than goodwill shall be reversed if, and only if, there has been a change in the estimates used to determine the asset's recoverable amount since the last impairment loss was recognised."

IAS 36.115 gives examples of what counts: a change in the basis for recoverable amount, a change in the amount or timing of estimated cash flows or in the discount rate where value in use was used, or a change in a component of fair value less costs of disposal.

What does not count is the passage of time. Unwinding of the discount increases present value mechanically as the cash flows come closer, and IAS 36.116 confirms that this alone does not support a reversal even though recoverable amount has technically risen. A reversal booked because the model produced a higher number a year later, with no change in estimate, has no basis.

IAS 36.117 sets the ceiling: "The increased carrying amount of an asset other than goodwill attributable to a reversal of an impairment loss shall not exceed the carrying amount that would have been determined (net of amortisation or depreciation) had no impairment loss been recognised for the asset in prior years." IAS 36.118 explains what happens above the ceiling: that would be a revaluation, accounted for under the standard applicable to the asset.

IAS 36.119 recognises the reversal immediately in profit or loss, unless the asset is carried at a revalued amount, in which case it is treated as a revaluation increase. IAS 36.121 requires the depreciation charge to be adjusted in future periods over the remaining useful life, mirroring IAS 36.63.

For a unit, IAS 36.122 allocates the reversal to the assets of the unit, except goodwill, pro rata with their carrying amounts, and IAS 36.123 caps each asset at the lower of its own recoverable amount if determinable and its own never-impaired carrying amount, reallocating any excess to the other assets except goodwill.

"An impairment loss recognised for goodwill shall not be reversed in a subsequent period." IAS 36.125 gives the reason: IAS 38 prohibits recognising internally generated goodwill, and any increase in the recoverable amount of goodwill after an impairment is likely to be internally generated goodwill rather than a recovery of the acquired goodwill.

IFRIC 10.8 closes the remaining gap: "An entity shall not reverse an impairment loss recognised in a previous interim period in respect of goodwill." A goodwill impairment booked at 30 June cannot be reversed at 31 December of the same year, even where the conditions that caused it have gone completely. That is the deliberate outcome, and it is the reason the sequencing in IAS 36.104 matters so much.

IAS 36.126 requires, for each class of assets, the amount of impairment losses recognised in profit or loss during the period and the line items in which they are included, the amount of reversals recognised in profit or loss and their line items, and the amounts of losses and reversals recognised in other comprehensive income for revalued assets. IAS 36.129 requires the same information by reportable segment where the entity reports segment information under IFRS 8.

This layer is almost always delivered, because it is a schedule rather than a narrative.

For each material impairment loss or reversal on an individual asset, including goodwill, or on a cash-generating unit: the events and circumstances that led to it; the amount; for an individual asset, the nature of the asset and the reportable segment it belongs to; for a unit, a description of the unit such as whether it is a product line, a plant, a business operation, a geographical area or a reportable segment, the amount by class of asset and by reportable segment, and where the aggregation of assets has changed since the previous estimate, a description of the current and former basis and the reasons for the change; the recoverable amount and whether it is fair value less costs of disposal or value in use; and where it is fair value less costs of disposal, the fair value hierarchy level and the valuation technique with any change in technique.

The clause on changed aggregation is the one companies overlook, and it is the disclosure that makes a redrawn cash-generating unit visible to a reader.

IAS 36.134 applies to each unit or group of units for which the carrying amount of goodwill or of indefinite life intangibles allocated to it is significant in comparison with the entity's total. For each such unit it requires the carrying amount of goodwill allocated, the carrying amount of indefinite life intangibles allocated, and the basis on which recoverable amount has been determined. Where value in use is used, it requires each key assumption on which management has based its cash flow projections, a description of management's approach to determining the value assigned to each key assumption and whether that value reflects past experience or external sources and if not how and why it differs, the period over which cash flows have been projected with an explanation where longer than five years, the growth rate used to extrapolate beyond that period with a justification where it exceeds long-run average growth, and the discount rate applied.

IAS 36.134(f) then requires, "if a reasonably possible change in a key assumption on which management has based its determination of the unit's (group of units') recoverable amount would cause the unit's (group of units') carrying amount to exceed its recoverable amount": the amount by which recoverable amount exceeds carrying amount, the value assigned to the key assumption, and the amount by which that value must change, after incorporating any consequential effects on the other variables, for recoverable amount to equal carrying amount.

Read the last limb again. It asks for a number, not a comfort statement. In worked example 1 that number is 0.4 percentage points on the discount rate and 0.6 percentage points on the terminal growth rate, alongside headroom of 1,174.

IAS 36.135 covers the fallback, where goodwill or indefinite life intangibles are spread across multiple units and no individual allocation is significant: disclose that fact together with the aggregate carrying amount. It exists for genuinely diffuse balances, not as a route around unit-level disclosure.

US GAAP splits the problem across three models rather than running one. Goodwill sits in ASC 350-20 and is tested at reporting unit level, being an operating segment or one level below, in a single quantitative step comparing the reporting unit's fair value with its carrying amount, preceded by an optional qualitative screen. Indefinite-lived intangibles sit in ASC 350-30 and are tested individually. Long-lived assets and asset groups sit in ASC 360-10 and are tested in two steps: a recoverability test comparing undiscounted expected future cash flows with carrying amount, and only where that fails, a measurement step using fair value.

The sequence between the three models also matters. Indefinite-lived intangibles are tested before the long-lived asset group, which is tested before goodwill, so a charge taken lower down changes the carrying amount that the goodwill test then works with.

The undiscounted first step has no equivalent anywhere in IAS 36 and it is the single largest source of difference. An asset group whose undiscounted cash flows cover carrying amount is not impaired under ASC 360-10 at all, regardless of what a discounted calculation would produce.