This article was first published in the March 2011 edition of Accounting and Business magazine.

Studying this technical article and answering the related questions can count towards your verifiable CPD if you are following the unit route to CPD and the content is relevant to your learning and development needs. One hour of learning equates to one unit of CPD. We'd suggest that you use this as a guide when allocating yourself CPD units.

The global financial crisis highlighted the need to review the impairment-accounting framework for financial assets. IAS 39, Financial Instruments: Recognition and Measurement has come under fire for only allowing impairments against incurred losses and not those that are expected.

The general purpose of an impairment test is to ensure that an asset is not carried for financial reporting purposes at an amount that exceeds its recoverable amount; to do so would overstate a reporting entity's financial position and performance.

Currently, International Financial Reporting Standards (IFRS) require that all financial assets, except those that are measured at fair value through profit or loss, are assessed for impairment. As reporting entities adopt the International Accounting Standard Board's recent standard (IFRS 9) on financial instruments, only financial assets measured at amortised cost will be subject to a single model of impairment for financial reporting purposes.

Currently, the IASB recognises the impairment of financial assets using the incurred-loss model in IAS 39. The incurred-loss model requires recognition of an impairment loss when there is objective evidence that an impairment exists for a financial asset or group of financial assets. Under the incurred-loss model, losses, including credit losses, are not recognised where they are expected as a result of future events.

The incurred-loss model has attracted criticism because it does not permit credit losses to be recognised until a trigger event has occurred. To address this, the IASB proposed moving to an expected-loss model in an exposure draft published in November 2009. This model would require entities to determine and account for the expected credit losses on a financial asset when originated or acquired.

Recognition

Impairment will be recognised under the exposure draft based on loss expectations, which will allow entities to start providing for impairment earlier than under the incurred-loss model. The model also recognises that many entities are compensated for credit losses - for example, via credit spread in a loan - and, therefore, an impairment loss should not be recorded upon the initial recognition of a financial asset.

Under the exposure draft, an allowance for the initial expected losses is effectively provided over the life of the asset through a reduction of interest revenue. Therefore the model also addresses the criticism of the incurred-loss approach that too much interest income is recognised in the periods before a loss event occurs.

Under the expected-loss model, the amortised cost of a loan or other financial asset is calculated using the effective interest rate method (EIR), as being the present value of the expected cashflows over the life of the loan, discounted at the EIR. Unlike the incurred-loss model, it does not wait for a loss event before recognising that a certain level of impairment will reduce the recoverable amount of the loans.

In certain circumstances, particularly where there is deterioration in underlying economic conditions, the basis for estimating the expected cashflows may be changed during the life of the loan and the consequent increase in loss provision would be recognised immediately.

The exposure draft proposed that entities should apply a probability-weighted approach to the determination of expected cashflows. This approach works well for large portfolios, but it would be difficult to identify the individual loans which would default. A probability-weighted approach allows a reasonable estimate to be made of the impact of expected impairment on the portfolio. However, it is difficult to see how this would work when considering impairment at an individual asset level.

The expected cashflows collected from these assets would not be the same as their expected cashflows based on a probability-weighted outcome.

Expected-loss example

A portfolio of loans of $70m is initially recognised on 1 January 2010. Each loan carries an interest rate of 6%. Management feels that loans will default at an annual rate. If defaults occur as expected, the rate of return from the portfolio will be approximately 3.5%. In this simple scenario, if the entity used the incurred-loss model, the interest income would be $4.2m and the loan would be stated at $70m at 31 December 2010.

If the expected-loss model were used, the interest income would be $2.45m ($70mx3.5%), the expected-loss adjustment would be $1.75m ($4.2m-$2.45m) and the loans would be stated at $68.25m ($70m-$1.75m).

Operational issues

The proposed model has some operational issues, which have been discussed at the Expert Advisory Panel (EAP). The concerns raised by the EAP included:

  • Most financial institutions manage interest income and credit risk separately. Interest is determined from the loan-accounting systems, and credit losses are determined via the credit systems. These systems are generally not integrated and would find it difficult to provide an EIR net of credit losses.
  • Expected cashflows are generally not stored in any systems. Most entities would calculate an expected-loss rate for a 12-month period, but would not record expected cashflows. It would be challenging to develop credit expectations over the full life of the asset.
  •  Most portfolios are open rather than closed, with new loans constantly being added to the portfolio. The expected loss rates for these portfolios are point-in-time estimates, which change as the portfolios change. Current systems do not track whether the change in the expected-loss rate arises from a new loan or an old one. This makes it operationally challenging to correctly identify and account for changes in original estimates.
  • When a credit loss occurs early in the life of a portfolio, this can lead to an inadequate allowance for the loans remaining in the portfolio.

The responses to the exposure draft indicated that it was operationally difficult to implement. On 31 January 2011, the IASB and FASB published, for public comment, joint proposals on the impairment of financial assets, a supplementary document to the exposure draft. Many respondents to the IASB's original exposure draft agreed with the proposed impairment approach but said, operationally, it was too difficult to apply, especially in the context of open portfolios.

The proposals in the supplementary document are based on suggestions from the EAP and retain many of the principles from the original exposure draft. The supplementary document addresses some of the major operational difficulties identified, particularly for open portfolios. The scope of the supplementary document is narrow because it applies only to financial assets measured at amortised cost and managed in an open portfolio, excluding short-term trade receivables.

The supplementary document sets out the following points for comment. Many financial institutions have two broad groups of financial assets that are monitored differently. They are loans that are considered problematic ('bad book') and those that are not ('good book'). Financial assets in the 'good book' are generally monitored on a portfolio basis, while those in the 'bad book' are managed more closely, often on an individual basis.

The document proposes to replace the incurred-loss-impairment models in IAS 39 and US GAAP with an expected loss-impairment model, including separate approaches to recognising expected losses for performing assets in a 'good book' and for troubled assets in a 'bad book'. Expected credit losses in the 'good book' would be recognised under a time-proportional approach based on the weighted average age and expected life of the assets in the portfolio, but subject to a minimum allowance of at least those credit losses expected to occur in the foreseeable future (not less than 12 months from the reporting date). When assets are transferred from the 'good book' to the 'bad book' the proposals would require the expected credit loss to be immediately recognised.

In some cases, recognising a time-proportional amount for the 'good book' may result in actual losses occurring that exceed the allowance balance at the time of the loss. This might occur if a portfolio has a number of loans that are expected to default early in their life. To address this issue, the board set the minimum allowance balance, mentioned above.

The foreseeable future is the period for which an entity can develop specific projections of events and conditions to estimate expected losses for the portfolio, and is required to be no less than 12 months after the entities' reporting date. The supplement has a 60-day comment period, with comments due on 1 April 2011.

Getting ready

The new expected-loss methodology will mean that entities will need to update their data collection and assumption settings soon. They should start identifying available and reliable data sources that could be used to forecast expected losses. The new approach will require changes to systems throughout the organisation and training will be needed. People will have to be taught to have sufficient insight into the new requirements.

The incurred-loss model is outdated and the expected-loss-impairment methodology will change the impairment landscape. Given that the expected-loss-impairment approach is imminent, the sooner systems, data, resources and other requirements are considered, the more prepared people will be to consider the challenges the new impairment standard will offer.

Graham Holt is an examiner for ACCA and executive head of the accounting and finance division at Manchester Metropolitan University Business School