BC 166 Additionally, respondents, field visit participants and North American round-table participants raised the following concerns about the practicability and costs of applying the proposed two-step approach:
(a) many companies would be required regularly to perform the second step of the impairment test, and therefore would need to determine the fair values of each identifiable asset, liability and contingent liability within the impaired unit(s) that the entity would recognise if it acquired the unit(s) in a business combination on the date of the impairment test. Although determining these fair values would not, for some companies, pose significant practical challenges (because, for example, fair value information for their significant assets is readily available), most would need to engage, on a fairly wide scale and at significant cost, independent valuers for some or all of the unit's assets. This is particularly the case for identifying and measuring the fair values of unrecognised internally generated intangible assets.
(b) determining the fair values of each identifiable asset, liability and contingent liability within an impaired unit is likely to be impracticable for multi-segmented manufacturers that operate multi-product facilities servicing more than one cash-generating unit. For example, assume an entity's primary basis of segmentation is geographical (eg Europe, North America, South America, Asia, Oceania and Africa) and that its secondary basis of segmentation is based on product groups (vaccinations, over-the-counter medicines, prescription medicines and vitamins/dietary supplements). Assume also that:
(i) the lowest level within the entity at which the goodwill is monitored for internal management purposes is one level below primary segment (eg the vitamins business in North America), and that goodwill is therefore tested for impairment at this level;
(ii) the plants and distribution facilities in each geographical region manufacture and distribute for all product groups; and
(iii) to determine the carrying amount of each cashgenerating unit containing goodwill, the carrying amount of each plant and distribution facility has been allocated between each product group it services.
If, for example, the recoverable amount of the North American vitamins unit were less than its carrying amount, measuring the implied value of goodwill in that unit would require a valuation exercise to be undertaken for all North American assets so that a portion of each asset's fair value can then be allocated to the North American vitamins unit. These valuations are likely to be extremely costly and virtually impossible to complete within a reasonable time period (field visit participants' estimates ranged from six to twelve months). The degree of impracticability will be even greater for those entities that monitor, and therefore test, goodwill at the segment level.

The Board's redeliberations
BC 165 Many respondents disagreed with the proposal to adopt a two-step approach to impairment testing goodwill. In particular, the second step of the proposed impairment test and the method for measuring any impairment loss for the goodwill caused considerable concern. Respondents provided the following conceptual arguments against the proposed approach:
(a) by drawing on only some aspects of the SFAS 142 two-step approach, the result is a hybrid between fair values and value in use. More particularly, not measuring goodwill's implied value as the difference between the unit's fair value and the net fair value of the identifiable net assets in the unit, but instead measuring it as the difference between the unit's recoverable amount (ie higher of value in use and fair value less costs to sell) and the net fair value of the identifiable net assets in the unit, results in a measure of goodwill that conceptually is neither fair value nor recoverable amount. This raises questions about the conceptual validity of measuring goodwill impairment losses as the difference between goodwill's implied value and carrying amount.
(b) it seems inconsistent to consider goodwill separately for impairment testing when other assets within a unit are not considered separately but are instead considered as part of the unit as a whole, particularly given that goodwill, unlike many other assets, cannot generate cash inflows independently of other assets. The previous version of IAS 36 is premised on the notion that if a series of independent cash flows can be generated only by a group of assets operating together, impairment losses should be considered only for that group of assets as a whole-individual assets within the group should not be considered separately.
(c) concluding that the recoverable amount of goodwill-which cannot generate cash inflows independently of other assets- should be measured separately for measuring impairment losses makes it difficult to understand how the Board could in the future reasonably conclude that such an approach to measuring impairment losses is also not appropriate for other assets. In other words, if it adopts the proposed two-step approach for goodwill, the Board could in effect be committing itself to an 'individual asset/fair value' approach for measuring impairments of all other assets. A decision on this issue should be made only as part of a broad reconsideration of the appropriate measurement objective for impairment testing generally.
(d) if goodwill is considered separately for impairment testing using an implied value calculation when other assets within a unit are considered only as part of the unit as a whole, there will be asymmetry: unrecognised goodwill will shield the carrying value of other assets from impairment, but the unrecognised value of other assets will not shield the carrying amount of goodwill from impairment. This seems unreasonable given that the unrecognised value of those other assets cannot then be recognised. Additionally, the carrying amount of a unit will be less than its recoverable amount whenever an impairment loss for goodwill exceeds the unrecognised value of the other assets in the unit.

BC 163 After the Board decided on the frequency of impairment testing, it expressed some concern that the proposed test would not be cost-effective. This concern related primarily to the requirement to determine the fair value of each identifiable asset, liability and contingent liability within a cash-generating unit that would be recognised by the entity if it had acquired the cash-generating unit in a business combination on the date of the impairment test (to estimate the implied value of goodwill).
BC 164 Therefore, the Board decided to propose as a first step in the impairment test for goodwill a screening mechanism similar to that in SFAS 142. Under SFAS 142, goodwill is tested for impairment by first comparing the fair value of the reporting unit to which the goodwill has been allocated for impairment testing purposes with the carrying amount of that unit. If the fair value of the unit exceeds its carrying amount, the goodwill is regarded as not impaired. An entity need estimate the implied fair value of goodwill (using an approach consistent with that described in paragraph BC160) only if the fair value of the unit is less than its carrying amount.

BC 161 In developing the Exposure Draft, the Board's discussion focused first on how the recoverable amount of goodwill allocated to a cashgenerating unit could be separated from the recoverable amount of the unit as a whole, given that goodwill generated internally after a business combination could not be measured separately. The Board concluded that a method similar to the method an acquirer uses to allocate the cost of a business combination to the net assets acquired could be used to measure the recoverable amount of goodwill after its initial recognition. Thus, the Board decided that some measure of the net assets of a cash-generating unit to which goodwill has been allocated should be subtracted from the recoverable amount of that unit to determine a current implied value for the goodwill. The Board concluded that the measure of the net assets of a cash-generating unit described in paragraph BC160(b) would result in the best estimate of the current implied value of the goodwill, given that goodwill generated internally after a business combination could not be measured separately.
BC 162 Having decided on the most appropriate measure of the recoverable amount of goodwill, the Board then considered how often an entity should be required to test goodwill for impairment. Consistently with its conclusions about indefinite-lived intangibles, the Board concluded that non-amortisation of goodwill increases the reliance that must be placed on impairment tests to ensure that the carrying amount of goodwill does not exceed its recoverable amount. Accordingly, the Board decided that goodwill should be tested for impairment annually. However, the Board also concluded that the annual test is not a substitute for management being aware of events occurring or circumstances changing between annual tests indicating a possible impairment of goodwill. Therefore, the Board decided that an entity should also be required to test goodwill for impairment whenever there is an indication of possible impairment.

BC 159 As a result, and consistently with the Board's decision to modify its proposal on allocating goodwill when an entity disposes of an operation, the revised Standard requires an entity that reorganises its reporting structure in a way that changes the composition of one or more cash-generating units to which goodwill has been allocated:
(a) to reallocate the goodwill to the units affected; and
(b) to perform this reallocation using a relative value approach similar to that used when an entity disposes of an operation within a cash-generating unit (group of cash-generating units), unless the entity can demonstrate that some other method better reflects the goodwill associated with the reorganised units (groups of units).
Recognition and measurement of impairment losses
(paragraphs 88-99 and 104)
Background to the proposals in the Exposure Draft
BC 160 The Exposure Draft proposed a two-step approach for impairment testing goodwill. The first step involved using a screening mechanism for identifying potential goodwill impairments, whereby goodwill allocated to a cash-generating unit would be identified as potentially impaired only when the carrying amount of the unit exceeded its recoverable amount. If an entity identified the goodwill allocated to a cash-generating unit as potentially impaired, an entity would then determine whether the goodwill allocated to the unit was impaired by comparing its recoverable amount, measured as the 'implied value' of the goodwill, with its carrying amount. The implied value of goodwill would be measured as a residual, being the excess of:
(a) the recoverable amount of the cash-generating unit to which the goodwill has been allocated, over
(b) the net fair value of the identifiable assets, liabilities and contingent liabilities the entity would recognise if it acquired the cash-generating unit in a business combination on the date of the impairment test (excluding any identifiable asset that was acquired in a business combination but not recognised separately from goodwill at the acquisition date).

BC 156 Some respondents to the Exposure Draft suggested that although in most circumstances goodwill could not be identified or associated with an asset group at a level lower than the cash-generating unit or group of cash-generating units to which it is allocated for impairment testing, there may be some instances when this is not so. For example, assume an acquiree is integrated with one of the acquirer's pre-existing cash-generating units that did not include any goodwill in its carrying amount. Assume also that almost immediately after the business combination the acquirer disposes of a loss-making operation within the cash-generating unit. The Board agreed with respondents that in such circumstances, it might reasonably be concluded that no part of the carrying amount of goodwill has been disposed of, and therefore no part of its carrying amount should be derecognised by being included in the determination of the gain or loss on disposal.
Reorganisation of reporting structure
(paragraph 87)
BC 157 The Exposure Draft proposed that when an entity reorganises its reporting structure in a way that changes the composition of cash-generating units to which goodwill has been allocated, the goodwill should be reallocated to the units affected using a relative value approach similar to that used when an entity disposes of an operation within a cash-generating unit.
BC 158 In developing the Exposure Draft, the Board concluded that a reorganisation that changes the composition of a cash-generating unit to which goodwill has been allocated gives rise to the same allocation problem as disposing of an operation within that unit. Therefore, the same allocation methodology should be used in both cases.

BC 153 The Exposure Draft proposed that when an entity disposes of an operation within a cash-generating unit to which goodwill has been allocated, the goodwill associated with that operation should be:
(a) included in the carrying amount of the operation when determining the gain or loss on disposal; and
(b) measured on the basis of the relative values of the operation disposed of and the portion of the cash-generating unit retained.
BC 154 This proposal has been carried forward in the Standard with one modification. The Standard requires the goodwill associated with the operation disposed of to be measured on the basis of the relative values of the operation disposed of and the portion of the cashgenerating unit retained, unless the entity can demonstrate that some other method better reflects the goodwill associated with the operation disposed of.
BC 155 In developing the Exposure Draft, the Board concluded that the proposed level of the impairment test would mean that goodwill could not be identified or associated with an asset group at a level lower than the cash-generating unit to which the goodwill is allocated, except arbitrarily. However, the Board also concluded that when an operation within that cash-generating unit is being disposed of, it is appropriate to presume that some amount of goodwill is associated with that operation. Thus, an allocation of the goodwill should be required when the part of the cash-generating unit being disposed of constitutes an operation.

BC 151 If the initial allocation of goodwill acquired in a business combination cannot be completed before the end of the annual period in which the business combination is effected, the Exposure Draft proposed, and the revised Standard requires, that the initial allocation should be completed before the end of the first annual period beginning after the acquisition date. In contrast, ED 3 proposed, and IFRS 3 requires, that if the initial accounting for a business combination can be determined only provisionally by the end of the period in which the combination is effected, the acquirer should:
(a) account for the combination using those provisional values; and
(b) recognise any adjustments to those provisional values as a result of completing the initial accounting within twelve months of the acquisition date.
BC 152 Some respondents to the Exposure Draft questioned why the period to complete the initial allocation of goodwill should differ from the period to complete the initial accounting for a business combination. The Board's view is that acquirers should be allowed a longer period to complete the goodwill allocation, because that allocation often might not be able to be performed until after the initial accounting for the combination is complete. This is because the cost of the combination or the fair values at the acquisition date of the acquiree's identifiable assets, liabilities or contingent liabilities, and therefore the amount of goodwill acquired in the combination, would not be finalised until the initial accounting for the combination in accordance with IFRS 3 is complete.
Disposal of a portion of a cash-generating unit containing goodwill
(paragraph 86)

BC 149 In deciding not to converge with SFAS 142 on the level of the goodwill impairment test, the Board noted the following findings from the field visits and North American round-table discussions:
(a) most of the US registrant field visit participants stated that the Board's proposals on the level of the goodwill impairment test would result, in practice, in goodwill being tested for impairment at the same level at which it is tested in accordance with SFAS 142. However, several stated that under the Board's proposals, goodwill would be tested for impairment at a lower level than under SFAS 142. Nevertheless, they believe that the Board's approach provides users and management with more useful information.
(b) several round-table participants stated that they (or, in the case of audit firm participants, their clients) manage and have available information about their investments in goodwill at a lower level than the level of the SFAS 142 impairment test. They expressed a high level of dissatisfaction at being prevented by SFAS 142 from recognising goodwill impairments that they knew existed at these lower levels, but which 'disappeared' once the lower level units were aggregated with other units containing sufficient 'cushions' to offset the impairment loss.
BC 150 In considering suggestion (c) in paragraph BC145, the Board observed that aggregating units that constitute businesses with similar characteristics could result in the disappearance of an impairment loss that management knows exists in a cash-generating unit because the units with which it is aggregated contain sufficient cushions to offset the impairment loss. In the Board's view, if, because of the way an entity is managed, information about goodwill impairment losses is available to management at a particular level, that information should also be available to the users of the entity's financial statements.
Completing the initial allocation of goodwill
(paragraphs 84 and 85)

BC 147 In relation to (b), the Board noted that SFAS 142 requires goodwill to be tested for impairment at a level of reporting referred to as a 'reporting unit'. A reporting unit is an operating segment (as defined in SFAS 131 Disclosures about Segments of an Enterprise and Related Information*) or one level below an operating segment (referred to as a component). A component of an operating segment is a reporting unit if the component constitutes a business for which discrete financial information is available and segment management regularly reviews the operating results of that component. However, two or more components of an operating segment must be aggregated and deemed a single reporting unit if the components have similar economic characteristics. An operating segment is deemed to be a reporting unit if all of its components are similar, if none of its components is a reporting unit, or if it comprises only a single component.
* The basis for identifying 'operating segments' under SFAS 131 differs from the basis for identifying segments based on the entity's primary reporting format under IAS 14. SFAS 131 defines an operating segment as a component of an enterprise (a) that engages in business activities from which it may earn revenues and incur expenses, including revenues and expenses relating to transactions with other components of the enterprise; (b) whose operating results are regularly reviewed by the enterprise's chief operating decision maker to make decisions about resources to be allocated to the segment and assess its performance; and (c) for which discrete financial information is available.
BC 148 Therefore, unlike IAS 36, SFAS 142 places a limit on how far goodwill can be 'pushed down' for impairment testing (ie one level below an operating segment).