Discretionary freedom in impairment testing

The degree of discretionary freedom in impairment losses is only interesting if the reporting flexibility is relatively higher or lower than under the previous amortisation method. If the discretionary freedom is supposed to be equal, it is reasonable to expect that both methods will be subject to the same intensity of opportunistic earnings management. However, there are reasons to believe that the impairment-only method offers more discretionary freedom than the amortisation method (Watts 2003, Ramanna 2008, Ramanna and Watts 2009, Kothari et al.

2010). Goodwill-amortisation charges are indeed discretionary in nature, but at a discount relative to goodwill-impairment losses. As stated earlier, it is not possible to observe the consumption of goodwill. Still, goodwill-amortisation charges are believed to be more verifiable and thereby easier to audit. The amortisation plan is generally linear and most accounting regimes require a maximum amortisation

period of 20 or 40 years.⁵ Besides, every change in the amortisation plan and the effects of such changes on accounting numbers must be revealed in additional disclosures. This makes it difficult to employ changes in goodwill amortisation as a reporting strategy. Reported impairment losses, however, are easy to manipulate and very difficult to audit. As managers generally have superior information about the firm’s future prospects, it is difficult for auditors to question estimates and assumptions made by the managers regarding impairment testing, if they are not clearly unreasonable (Benston et al. 2007).

The flexibility of the impairment regulation is easy to demonstrate. Flexibility is given when it comes to the allocation of goodwill to cash-generating units, the frequency of impairment testing and the measurement of impairment losses. First, impairment test of goodwill can be performed at any date during a year, but has to be executed at the same date each year for the same cash-generating unit (IASB 2008a: IAS 36.90). The choice of test-dates can, therefore, be made according to the managers’ reporting strategy. If the managers want to shift earnings from future periods to the present by avoiding goodwill-impairment losses, cash-generating units that operate in seasonal industries should be tested during periods of the year where the cash-generating units’ recoverable amounts are at the highest.⁶ In contrast, if managers intend to shift earnings from the present into the future by overstating goodwill-impairment losses, impairment-test dates should be chosen to minimise the recoverable amounts of the cash-generating units.

5 For instance, US-GAAP required that goodwill should be amortised over a period not to exceed 40 years (APB 1970b), and UK-GAAP has a presumption that goodwill shall not be amortised over more than 20 years (ASB 1997).

6 It could be argued that the testing dates make no difference. However, in practice forecasting periods are short and terminal values are not necessarily defined as perpetuity. Also, when recoverable amounts are measured using other estimates than present values, the testing dates are likely to be relevant.

However, the requirement to test each cash-generating unit at the same date each year, limits managers’ discretionary freedom.

The regulation offers an exception from estimating the recoverable amounts each year. For the exception to take effect, three cumulative requirements must be met (IASB 2008a: IAS 36.99). First, assets and liabilities allocated to the units have not changed significantly since last time the recoverable amounts were estimated.

Second, when the recoverable amounts of the cash-generating units were estimated the last time, they exceeded the cash-generating units’ carrying amounts by substantial margins. Third, an analysis of events and changes in circumstances suggests that the probability that the recoverable amounts have fallen below the carrying amounts of the units is remote. The managers are, therefore, left with discretion to sidestep impairment tests of goodwill. The list of indicators suggesting that assets are impaired also provides some discretionary freedom.

Since a fixed test date for each cash-generating unit might preclude timely recognition of impairment losses, IAS 36 provides a non-exhaustive list of impairment indicators (IASB 2008a: IAS 36.12). An unscheduled impairment test is required when one or more of these indicators suggest that the asset has impaired. Since the list is non-exhaustive, managers are free to find additional impairment indicators. To the extent that overstated impairment losses are consistent with managers’ reporting strategies, the managers have incentives to employ events and circumstances other than those listed as impairment indicators.

The allocation of goodwill to cash-generating units may influence the likelihood of reporting impairment losses in the future. According to IAS 36, goodwill shall be tested for impairment losses at a level of reporting referred to as cash-generating unit or groups of cash-cash-generating units (IASB 2008a: IAS 36.80). The

higher the level of aggregation at which cash-generating units are formed, the larger is the probability that a decrease in goodwill will be compensated by internally-generated goodwill in another cash-generating unit. Therefore, the level at which an entity defines its allocation units for goodwill determines to a large extent the likelihood of reporting goodwill-impairment losses in subsequent periods (Henning, Shaw and Stock 2004, Zang 2008, Ramanna 2008, Ramanna and Watts 2009).

The most significant discretionary element in the testing procedure relates to the estimation of recoverable amounts. The recoverable amount is the higher of the value-in-use and the fair value (IASB 2008a: IAS 36.18). The fair value will be an observed market value of the cash-generating unit or a market value of a similar cash-generating unit. If market values are unavailable, calculating the present value of future net cash flows is the best available method to get an estimate of the recoverable amount. The present-value technique requires estimates of future cash flows, or in more complex cases, expectations about possible variations in the amount and timing of the cash flows. In order to achieve more reliability, external information should be given more weight than internal information (IASB 2008a:

IAS 36.33). The present-value technique provides plenty of room for discretionary freedom. Even when managers try to estimate unbiased recoverable amounts, the problems associated with uncertain future cash flows and risk-adjusted rates are serious. This makes it reasonable to question the relevance and reliability of the recoverable amounts and by that, the impairment losses calculated upon them (Watts 2003).

This section has demonstrated that in most of its facets goodwill-impairment testing is a highly discretionary procedure that allows managers to coordinate

impairment accounting to their reporting strategy. The discretionary freedom can be exploited to understate impairment losses and overstate current earnings and net assets or to overstate impairment losses, understate current earnings and net assets. Assumptions and estimates and other subjective elements are required at all stages (Ramanna 2008, Zang 2008, Ramanna and Watts 2009). This suggests that the impairment-only approach provides managers with opportunities to engage in earnings management, which may impair the decision usefulness of goodwill numbers.

3. Value relevance – some fundamentals and