IFRS 9 significantly changes the methodology required for provisioning against accounts receivable and other debt investments. Overall, IFRS 9 is designed to replace the “incurred loss” model required by IAS 39 with an “Expected Loss” model. One might think of IAS 39 as backwards looking such that under IAS 39 one would ask “Has a loss occurred?” whereas IFRS 9 is forward looking “Has a loss occurred or might it occur in the future?” IFRS 9 was born out of 2008’s financial crisis where it was viewed that banks were reporting losses too late and the intent was to recognize such losses sooner as conditions are forecast to deteriorate. However, the IASB is not in the habit of writing industry standards and therefore all non-financial institution companies were painted with the same brush in moving to this model, albeit with some simplifications applicable to such non-financial institutions. 

In this publication we discuss the top nine issues a non-financial institution might face in its transition to the new impairment approach under IFRS 9 which is required for years beginning on or after January 1, 2018. It is important to note this publication does not cover all aspects of the impairment model under IFRS 9, but focuses on what in our current experience are the most relevant issues to non-financial institutions. 

Generally, trade receivables will be carried at amortized cost for those entities that are not in the habit of selling such receivables in a way that qualifies for de-recognition. However, the IFRS 9 impairment model also applies to those receivables that are carried at fair value through other comprehensive income (FVOCI) which might be applicable where entities have a factoring program qualifying some receivables for de-recognition. Entities should determine the appropriate classification for such receivables before applying the impairment model. 

In consideration of the complexity of implementing an expected loss model, the standard mandates a simplified approach for trade receivables without a significant financing component. Generally, normal trade receivables with terms of one year or less would not have a significant financing component. The simplification requires that an entity use lifetime expected credit losses to measure its provision against such receivables rather than the more complex staged approach applied to longer term assets (see discussion of “Investments in Debt or Long-Term Receivables”).

IFRS 9.B.5.35 and illustrative example 12 discuss the use of a “provision matrix” as one possible way to implement this simplified approach. A provision matrix essentially applies an expected loss rate to every ageing category of receivables including the current category. The example provided in the standard is:

The example illustrates the major difference between IAS 39’s incurred loss model and IFRS 9’s expected loss model. Under IAS 39 companies generally would wait to establish a provision until the receivable became significantly past due. For example, some companies might not have a provision until the “61-90” day category. Under IFRS 9, essentially all receivables require a provision even where they are not past due based on the expectation that certain receivables will be uncollectible. The IASB notes in IFRS 9’s basis of conclusions that “to recognize a loss allowance on a more forward looking basis before trade receivables become past due would improve financial reporting”. Whether one agrees with this view or not, it is apparent that allowances against all receivables is the intended impact of IFRS 9. 

The gross carrying amount of the receivables would be “written off” when the entity has no reasonable expectations of recovering anything for the financial asset. Presumably, this would be where the expected loss for an individual receivable is close to 100%. Therefore, an entity should consider whether any receivables in the more than 90 day category should be written off to reduce the gross amount of the receivable and provision. 

Note, the above model also applies to “contract assets” within the scope of IFRS 15 such as accruals of revenue or certain variable consideration. Thus, this model applies even before a customer is invoiced and such accruals need to be considered even if they are outside the traditional receivables ageing reports. Generally, such accruals will attract approximately the same provision as the billed receivables that are not yet due. However, ageing of unbilled receivables may be important in the context of “IFRS 9 versus IFRS 15 Considerations” as discussed below. 

The first step is likely to gather the information on past history of uncollectible accounts and how much has ultimately been lost on those accounts migrating through the ageing categories. 

Companies would be expected to stratify their receivables into different groupings before applying the provision matrix. For example, a company might sell to both government and non-governmental organizations. The government accounts may have a significantly different loss experience from accounts with non-governmental organizations. Examples of factors that might be considered in such stratification would include geographical regions, product type, customer ratings, collateral, and the nature of the customer (e.g. wholesale vs. retail). 

In considering stratification, it is important to first understand the drivers of credit risk for the underlying receivables. In some cases, the driver of credit risk may be geographic, in other cases it may be based on the nature of the customer or the nature of the goods/services. In a complex business there will be multiple matrices. For example, product A in region B being sold to customer type C. The level of segmentation required is often a matter of significant judgment and in developing segments an entity should consider where further segmentation would be expected to lead to immaterial changes. 

Stratification may go down to the individual customer level in some cases. For example, where a particular customer is known to be in financial difficulty, it may require an increased provision compared to historical averages over all ageing categories. It is important to consider and avoid any double counting of losses in these situations. 

For most companies it is likely accounts where losses are experienced would eventually make their way to the “more than 90 days past due” category. Assuming this assumption is valid one might ask “historically what percentage of sales makes its way to that category and what percentage is lost from such receivables?” 

For example, an entity might conclude that 1% of its sales will migrate to the over 90 day category and for those accounts that are in default that historically 30% of the face value of such receivables is never collected. This would imply that the starting point for the first ageing category is 30% * 1% = 0.3%... That is 0.3% of the dollar amount in current receivables historically have not been collectible. Similarly, one might ask historically what percentage of the “1-30 day past due category” has migrated to the 90 day category?” Put another way, what proportion of receivables that are 1-30 days past due are expected to be paid in full and what proportion migrates to the 90 day category. Likely, there is a higher percentage of such accounts that would migrate since they are already past due… For example, an entity might conclude that 5.33% of such accounts migrate to the 90 day category and the company has already established that of those accounts migrating the historical loss ratio is 30%... Therefore, based on historical data the loss ratio for the 1-30 day past due category is approximately 30% * 5.33% = 1.6%. Of course, this is only one possible method of determining the historical loss percentages; there may be other methods that also align with IFRS 9’s principles. 

Even if an entity has never experienced historical defaults, a provision based on expected losses will still be required. These entities may look to derive probabilities of default and loss rates from credit ratings, industry or other data. 

Income taxes may be impacted by the change in methodology for impairment and in particular this change could create new or additional temporary differences to account for. 

Establishing the historical loss ratio is only the starting point and entities need to consider:

Systems may not have been established to gather data in the above manner and therefore some entities will require system changes or require the generation of new analytical reports to establish the matrix. Furthermore, ongoing controls over such estimates would need to be established as discussed in “Ongoing controls and procedures”. 

When looking to historical data, it is important to segregate losses between those relating to customer disputes or other discounts provided versus credit losses. Only those losses that are due to credit risk would be within the scope of the IFRS 9 provision. Other losses due to disputes / discounts / inefficiencies are subject to guidance in IFRS 15 which should be applied prior to IFRS 9. For example, accrued receivables at a company billing by hour might be subject to a provision (due to variable consideration constraints or historical discount patterns) prior to the IFRS 9 model being applied only to the net amounts recorded under IFRS 15. 

In terms of income statement presentation, the adjustments required under IFRS 15 will directly impact the “revenue” line and therefore will have a corresponding impact on gross margins where such figures are reported. The IFRS 9 provision will generally impact an expense line such as “provision for bad debts”, but will not impact revenue. The provision matrix will only work well if the data being used is of high quality and accurately reflects credit loss experience. Therefore, where entities have historically not tracked credit losses separately from losses due to other factors historical data may need to be re-examined in order to parse the data between different types of losses. Furthermore, companies should consider different tracking going forward in order to simplify the provision matrix process. 

As noted above, the simplified model applies to trade receivables and contract assets without a significant financing component. 

Some entities may have long-term financing arrangements with customers for goods or services sold. For long-term trade receivables and IFRS 15 contract assets containing a significant financing component the entity has a policy choice whether to use the simplified approach or the full IFRS 9 model. 

Furthermore, entities may have invested funds not immediately required for the business in longer-term debt investments or made loans to strategic investees (or in standalone unconsolidated accounts to other companies within the group). The simplified model would generally not apply to these types of instruments. 

Where an entity cannot apply or has chosen not to apply the simplified model, the full IFRS 9 impairment model would apply where such instruments are measured at amortized cost or at fair value through other comprehensive income. Generally, a corporate entity would not originate or purchase credit impaired assets, so discussion of such matters is beyond the scope of this paper. 

The full IFRS 9 impairment model has three categories or “stages”. Generally speaking, a purchased or originated receivable would fall into the first stage:

Assets would be re-considered each reporting period to determine if they remain in the “performing stage” or had moved to one of the other stages. These stages are relative and not absolute. That is, moving between categories is considered based on whether there has been an increase in credit risk relative to when it was originated or purchased. There is no bright line in the standard for considering this, but rebuttable presumptions are that an instrument 30 days or more past due falls in Stage 2 and an instrument overdue by more than 90 days falls into Stage 3. 

A simplification may be applied to low credit risk assets. These are generally equivalent to “investment grade” assets such that they can be presumed to be in Stage 1 while they remain investment grade. This simplification does not mean that no provision is required against such assets. The simplification merely avoids the need to consider migration of such assets to Stage 2 or Stage 3 while they remain investment grade.

Notice that the provision for Stage 2 and Stage 3 are both lifetime expected credit losses; the difference between the two categories relates to how interest is recognized. In Stage 2 interest is based on the gross amount of the asset before deducting the provision whereas in Stage 3 interest is computed based on the net amount of the asset after deducting the provision.

The full IFRS 9 model is complex to implement because it requires an entity to consider:

Income taxes may be impacted by the change in methodology for impairment and in particular this change could create new or additional temporary differences to account for. 

Lessors are not immune to the IFRS 9 impairment requirements despite the fact that leases are recognized under IAS 17 or IFRS 16. 

However, IFRS 9 provides a policy choice for lessors to either apply the simplified lifetime expected model or the full model. This policy choice can be made separately for operating vs. finance lease receivables. 

Often companies may provide financial guarantees to unrelated parties such as customers, strategic investees, or even related entities within the same group in standalone accounts. IFRS 9 requires that unless the guarantee is measured at fair value through profit and loss, a provision be established using the expected credit loss model.

IFRS 9 amends IFRS 7 disclosures concerning credit risk. More extensive disclosure is required about an entity’s credit risk management practices and how they relate to the recognition and measurement of expected credit losses. 

For instruments measured using the “full IFRS 9” model there are detailed quantitative and qualitative disclosures required about the categorization of the instruments and how the provisions have been calculated and how they have changed over the period.

IFRS 7.35N allows for entities to disclose some of the information based on the provision matrix where that simplified model has been used. 

Implementing such disclosures will likely require significant efforts and entities should ensure that the implementation project is designed to establish systems and processes are capable of generating the required quantitative information for such disclosures.

Since the IFRS 9 impairment model is substantially different than IAS 39 and requires different estimates and judgments to be applied, new and different internal control procedures may be required. 

Entities should consider the need for new internal controls over the impairment model and where these controls will be performed within the group. Some of the controls may need to be implemented outside of the financial reporting group (e.g. in the collections or credit management group of different business units) and other controls may be implemented centrally. 

Such controls would likely also involve back-testing of experienced losses to the established provisions and determining whether changes to the provisioning methodology, additional correlations to macro-economic indicators or enhanced segmentation of customers is required. The aim of back-testing is a continuous improvement in the robustness of the model. However, back-testing does have inherent limitations due to the continuous changes in the macroeconomic environment.

IFRS 9 allows for two transition approaches: “full retrospective” and “modified retrospective”. For an entity with a calendar year-end and quarterly reporting the difference between the approaches is as follows:

The full retrospective transition method is only available where an entity can restate prior periods without the use of hindsight. Ideally, this would mean that an entity wishing to use the full retrospective approach would have its impairment provision calculated for January 1, 2017 before the end of Q1 2017. However, as long as an entity can show that it has not relied on hindsight (i.e. has used only the data available at January 1, 2017) in establishing its provision it may not be necessary to have actually calculated the provision during the 2017 year.

There are certain simplifications applicable where the full impairment model is applied. These simplifications are centered around the fact that it may be difficult to evaluate whether there has been a significant increase in credit risk to establish whether an instrument has migrated between categories. For non-investment grade assets, of undue cost or effort is required to establish whether such a re-categorization has occurred, the entity may recognize lifetime expected credit losses on such instruments. 

Certain impairment disclosures are required upon transition to IFRS 9 by IFRS 7. An entity is required to disclose information that would permit the reconciliation of the ending impairment allowances recognized in accordance with IAS 39 to the opening loss allowances recognized under IFRS 9. 

As discussed, we intended this publication to cover issues of most relevance to non-financial institutions, but there are a number of other issues that may apply to such entities beyond the scope of this paper which should be considered where applicable. Some of these other issues include:

For further information please contact: