Dynamic risk management

Date recorded:

Cover note (Agenda paper 4)

AP 4A contains the project background and a summary of the Board’s discussions to date. See the June IAS Plus meeting summary for more details.

AP 4B analyses the measurement of imperfect alignment and how an entity should communicate the effects of imperfect alignment in financial reporting.

AP 4C analysis how an entity, which is applying the dynamic risk management (DRM) accounting model, should treat a change in risk management strategy and how such a change would affect profit or loss and other comprehensive income (OCI). The paper focuses on changes in risk management strategy that would require a change in the entity’s target profile.

Summary of discussions to date (Agenda Paper 4A)

Background

The aim of the proposed DRM model is for financial statements to represent faithfully the effect of dynamic risk management activities undertaken by an entity. It aims to help users assess management’s performance by focusing on how well management was able to align the asset profile with the target profile using derivatives.

In November 2017 the Board tentatively decided that the DRM accounting model should be developed based on the cash flow hedge mechanics.

In December 2017 the Board tentatively decided the staff should develop the accounting model for DRM in two phases. In the first phase the IASB would focus on ‘core areas’ such as: (i) asset profile; (ii) target profile; (iii) derivative instruments used for DRM purposes; and (iv) performance assessment and recycling. In the second phase the focus would be on extensions of the concepts.

The first two of the ‘core areas’ were discussed during the February 2018, March 2018 and April 2018 Board meetings. The Board identified the qualifying criteria for financial assets in the asset profile and financial liabilities in the target profile.  The Board also tentatively decided on the allowable designations, reasons for de-designation and the documentation that should be completed before an entity starts applying the DRM model. The third and fourth ‘core areas’ were discussed during the June 2018 Board meeting. The Board tentatively decided on the types of derivatives that can be designated within the DRM model, their designation and de-designation and which derivatives will be addressed within the first phase of the DRM project. It also tentatively decided what information should be provided in the statement of profit or loss regarding DRM activities. The results should reflect the entity’s target profile in the case of perfect alignment and the reclassification should occur over the time horizon of the target profile (even in the case of voluntary discontinuation of the DRM accounting model). The Board tentatively decided that in order to apply the DRM accounting model, an entity must demonstrate that there is a continuing economic relationship.

Staff recommendation

This paper was only educational and did not contain a staff recommendation.

Imperfect Alignment (Agenda Paper 4B)

Measurement of imperfect alignment

The staff highlight the difference between the measurement and assessment of imperfect alignment. Measurement is the quantification of the actual difference, if any, between the benchmark and designated derivatives. It is focused on what has occurred rather than expected behaviour i.e., it is retrospective in nature.

The staff view is that requiring entities to measure imperfect alignment will quantify the strength of the link between the items designated within the DRM accounting model. It will provide users with information, in the context of DRM, about how efficient and effectively the reporting entity’s management has discharged its responsibilities to protect the entity’s economic resources from unfavourable events and how the management will use the entity’s economic resources in future periods.

In this paper, the staff presented three scenarios when the imperfect alignment can arise. When there is:

  • excess notional when compared with the benchmark derivatives
  • excess term when compared with the benchmark derivatives
  • insufficient notional when compared with the benchmark derivative

The staff think these scenarios are sufficient to develop a principle regarding the information content of imperfect alignment. Furthermore, these scenarios highlight that when cash flows designated within the DRM accounting model are in excess to those required to achieve alignment, the effects of imperfect alignment are contractual and therefore the cash flows in question will flow to or from the entity. However, when insufficient cash flows are designated (i.e. when caused by cash flows that will not occur), measuring imperfect alignment provides information about the effect on the entity’s future economic resources had the benchmark derivatives been executed. The latter is similar to the concept of opportunity cost.

Because the asset and target profiles are subject to change over time (e.g. origination of new loans or issuance of new financial liabilities), and if an entity does not update the portfolio of designated derivatives in response to those changes, this can result in imperfect alignment. Therefore, the staff think that the entity should measure alignment immediately prior to updating the asset profile, target profile, or designated derivatives (i.e. prior to each change in inputs), or at a minimum, at each reporting date.

For the purpose of the DRM model, changes in inputs are different from changes in assumptions, such as changes in prepayment assumptions. When a change in prepayment assumption occurs, it indicates that management’s estimation about when a loan (or portfolio of loans) will mature was inaccurate. Therefore, some degree of imperfection should be captured by measurement when changes in prepayment assumptions occur because the assumption can have an impact on the degree on alignment to date, not just on a forward looking basis. I.e. the entity should measure imperfect alignment based on the asset profile, target profile and designated derivatives immediately after the change in prepayment assumption and considering the new benchmark derivative. The issue of change in prepayment assumptions is illustrated in the paper via an example.

Similarly to change in prepayment assumptions, an entity should measure imperfect alignment immediately after the breach of qualifying criteria.

Communication of imperfect alignment

The staff used the same scenarios to discuss and illustrate where the imperfect alignment should be presented, i.e. whether it should be presented in the statement of profit or loss or in other comprehensive income.

The staff indicate that, in the case of over-hedging, presenting the difference between the change in clean fair value of the designated derivative and the change in clean fair value of the benchmark derivative as imperfect alignment in the statement of profit or loss:

  • provides relevant information to users of financial statements as it has predictive valu
  • does not provide entities with the opportunity to claim derivatives are executed for risk management purposes and account for them as such, even when the true intent could be trading in nature
  • will not be a deviation from the current key principle of hedge accounting requirements within IFRS 9, i.e. application of ‘lower of’ test and recognition of ineffectiveness in profit or loss

In the case of under-hedging, similarly to IFRS 9 and after applying the ‘lower of’ test, the staff are of the preliminary view that the ineffectiveness should not be recognised in financial statements. Recognising gains or losses within the statement of profit or loss for an asset or liability that does not exist (i.e. the benchmark derivative) is inconsistent with the Conceptual Framework. In this particular case, disclosures should be required to inform users about the impact of imperfect alignment.

Target profile as a range

The staff consider the advantages and disadvantages of defining the target profile as a range, for example to limit the variation in net interest income from a change in market interest rates to less than +/- 15% of interest income reported in the previous period. They highlight that defining the target profile as the range could complicate the mechanics for deferral and more specifically reclassification and could cause potential abuse. The staff conclude that the target profile within the DRM accounting should be defined as a single possible outcome.

Assessment of imperfect alignment

Assessing imperfect alignment aims to ascertain whether financial assets, financial liabilities and derivatives designated within the DRM model can be expected to meet the risk management strategy for which they have been designated. The objective of the assessment is to set a minimum level of alignment to apply the DRM accounting model.

The staff considered two alternatives for defining a minimum performance threshold:

  • Quantitative threshold
  • Combination of a quantitative assessment with qualitative thresholds

The staff view is that a prospective assessment would be more appropriate, as this would provide users with an indication of whether the risk management strategy will be achieved in the future. A quantitative threshold would:

  • increase the comparability of information reported in the financial statements (as it sets a minimum requirement for preparers to be eligible for the DRM model and all preparers would follow the same threshold
  • strength discipline around the application of the DRM model
  • avoid designation of relationships below the threshold from qualifying for the DRM model

The staff view is that there should not be a ‘bright line test’ within the DRM accounting model. That would be inconsistent with the objectives of the DRM accounting model to improve information provided regarding risk management and how risk management activities affect an entity’s current and future economic resources and it could be viewed as governing rather than reflecting risk management.

An entity could use quantitative assessment such as a sensitivity analysis with multiple scenarios of interest rates to demonstrate the existence of an economic relationship. This sensitivity analysis must consider only financial assets, financial liabilities and derivatives designated within the DRM model, and the quantity of designated derivatives (defined in terms of time and notional) should be the quantity that the entity actually uses for asset transformation. However, the entities should be not allowed, similarly to IFRS 9, to designate items within the DRM model reflecting an imbalance between the weightings of the asset profile and the designated derivatives that would create imperfect alignment. This could result in an accounting outcome inconsistent with the purpose of the DRM accounting model.

The staff view is that an entity should assess imperfect alignment, at a minimum, at each reporting date or upon changes in inputs, changes in assumptions or breach in qualifying criteria.

If an entity fails the assessment, the DRM model should be discontinued from the last date on which the requirement was met and the amount previously recognised in other comprehensive income should be reclassified to profit or loss over the life of target profile.

Staff recommendation

The staff recommend that:

  • entities should be required to measure imperfect alignment on an on-going basis
  • measuring imperfect alignment provides information about the extent to which an entity has not achieved its risk management strategy and, therefore, quantifies the potential impact on the entity’s future economic resources
  • in the case of over-hedging, the difference between changes in fair value of the designated and benchmark derivatives should be presented in the statement of profit or loss as imperfect alignment
  • the ‘lower of’ test should be retained within the DRM accounting model. As a result, to fully communicate the impact of imperfect alignment, disclosures will be required in the case of under-hedging
  • the target profile within the DRM accounting should be defined as a single outcome
  • the DRM model should require a minimum performance threshold in the form of qualitative thresholds supported by quantitative analysis

Discussion

One of the Board members indicated that, in accordance with the paper, measuring imperfection should be done when the inputs or assumptions change,  However, there may be an in-between situation—e.g., a prepayment occurred during the period that  was not noticed. Management will not have taken a conscious action to deal with the prepayment, so neither the assumptions or inputs will have changed. Yet, in substance, something changed because the asset was derecognised. How would such an element of imperfection be captured? The staff responded that this difference will be captured via a change in inputs because the asset was derecognised.  

The staff confirmed that the timing of the testing is not optional and it should be done either prior to the change in inputs/assumptions or at the reporting date, whichever happens first. In respect of the question about the materiality of the changes the staff indicated that because the term ‘significant’ was not defined, the paper suggests the measurement at each change. However, the staff expect that if the change is insignificant then entities, for practical reasons, need not re-measure. The staff said that possible simplifications and how to bring dynamic changes to the model will be addressed in the future. One Board members said that the paper should have consistent principles for the terms ‘on-going basis’ and ‘change in inputs/assumptions’.    

The question was raised as to whether it will be visible what is happening in the model during the year or whether one net number of imperfection will be visible at the end. The staff responded that the disclosures will be addressed in the future.

One of the Board members ask how the benchmark derivative will be monitored to ensure that it is a ‘perfect’ derivative and how frequently this will be assessed. The staff responded that the benchmark derivative is a derivative that accomplishes the risk management strategy and hence it is a matter of fact rather than a matter of opinion and it can be proved mathematically (i.e. the benchmark derivative is the difference between the asset profile and target profile). It was challenged that the aim of the project is to see whether the risk management is sound enough and that investors want to know the quality of risk management, which depends on how the asset profile and the target profile were defined. The staff responded that the aim of the project is not to provide guidance on what the appropriate risk management strategy is because this is a specific decision of each entity and cannot be dictated by accounting. The aim is to develop a model that will capture the entity’s strategy so the economics are reflected.  Transparency of the risk management strategy is the task of disclosures and those will be developed later.

The staff confirmed that the model will only capture imperfection in respect of designated items and not all items that may be managed economically. For example, risk is managed on a daily basis with a continuous change of derivatives but those derivatives are only designated quarterly. In this case, derivatives that are outside the model are not taken into account when measuring imperfection. The question was raised about the possible inconsistency of the information provided to investors and the different level of imperfection captured depending on how often the items are designated (i.e. economic hedging—daily versus accounting hedging— quarterly). The staff explained that based on the mechanics of the hedge accounting, if the items are brought into the DRM model later, that is done at cost because the derivatives are ‘off-market’ and that should rectify the issue. Again, the possibility of disclosing items that are not within the model is a different matter and it may be discussed at later stage.

One of the Board members suggested including a caveat that measuring imperfect alignment provides information regarding the extent of achieving (or not achieving) the risk management strategy by an entity only in respect of the designated derivatives. Furthermore, it was suggested to state clearly that the measurement of imperfection should be done by comparing the changes in fair value of the benchmark and actual derivative, i.e. indicating that this is ‘the’ way to be done and not ‘a’ way. The staff confirmed that the delta fair value gives the measure of ineffectiveness similarly to other hedge accounting models. However, there may be other methods than comparing the benchmark and the actual derivative and those methods have not been explored yet.  The staff agreed to clarify the measurement, as the delta fair value, in the paper.

In respect of presentation of imperfection, specifically over-hedging, the Board members agreed that it should be presented in profit or loss but the question was raised as to what ends up in P&L and in which line. The staff responded that this will be developed in a later stage.

In respect of paragraphs 63 to 65 of the paper, one Board member indicated that in the case of change in assumptions this should not call into question the estimates from the previous period and definitely should not indicate that those previous period assumptions were inaccurate.

The staff indicated when an entity breaches qualifying criteria because the risk management strategy was changed (and for example assets are no longer measured at amortised cost and only at fair value through profit or loss (FVTPL)), that should be treated as if there was a breach of qualifying criteria and not as a change in the risk management strategy.

The Board agreed with the ‘lower of’ test, however, it was suggested to include a caveat that if the entity is setting up the hedge for 1,000 and using only 750 it does not indicate that this is good enough to get into the hedge accounting model. Furthermore, the disclosure requirements should provide information about the measurement of imperfection in the case of under-hedging. The staff should consider how to communicate under-hedging because the information should be provided equally for both sides, i.e. for over- and under-hedging.

The concern was raised that the banks usually set up the target profile as the range and if they need to define a figure, there will be a tendency to set this at the upper limit on the basis of treatment of under-hedging.  There is no recommendation or guidance in respect of the mid-point and there is no expectation that this will be developed in a later stage. The staff expect that entities will usually have a number as the target profile and a range around that number. Nevertheless the target profile, for DRM model purposes, should be defined/quantified as one number (a point estimate). Furthermore, it was indicated that disclosures may be useful for explaining the risk management strategy when there is deviation from the point estimate, but the measured imperfection is within the acceptable range set up by an entity.

The Board members indicated that the specifics of the qualitative assessment and minimum thresholds (i.e. what tests should be done on an on-going basis) should be developed in the concept of the DRM model and not based on the IFRS 9 hedge accounting model. The staff agreed to come back on with the proposal of exact tests and the thresholds in the future Board meeting.

Change in Risk Management (Agenda Paper 4C)

While an entity can change its risk management strategy for various valid reasons, the staff expect such changes to occur infrequently. For the purpose of the DRM accounting model, a change in risk management strategy is when management makes a decision that requires a change in the entity’s target profile. The following are not a change in risk management strategy:

  • A change in inputs
  • Alteration of the scope of the DRM accounting model (e.g. an entity that decides to de-designate items from the asset profile)
  • A change in assumption e.g. a change in prepayment speeds

In the paper, the staff consider how the change in risk management strategy should be reflected in financial reporting. More specifically, whether a change in risk management strategy should be treated as a termination event or a continuation of the DRM accounting model and how the accumulated amount in other comprehensive income should be reclassified to the statement of profit or loss after the change in risk management strategy. The staff considered three possible approaches to address this matter:

  • Alternative 1: Reclassify accumulated other comprehensive income immediately
  • Alternative 2: Amend the reclassification pattern to align with the revised target profile
  • Alternative 3: Maintain the previous reclassification pattern

The staff think that Alternative 1 could lead to abuse in order to achieve specific accounting outcome that would be inconsistent with the purpose of DRM accounting model and such treatment would be inconsistent with cash flow hedge accounting under IFRS 9.

Similarly, in the case of Alternative 2, the staff indicate that entities may be changing their risk management strategy to achieve a specific accounting outcome that is inconsistent with the purpose of the DRM accounting model. Furthermore, reclassifying over the life of the amended target profile could result in other comprehensive income being deferred beyond the period over which risk was managed in the first place.

For the reasons indicated above, the staff do not support either of these approaches.

Alternative 3 indicates that a change in risk management strategy would effectively be treated as a termination event because the previous reclassification pattern for the amounts recorded in other comprehensive income before a change in risk management strategy takes place. Furthermore, that would prevent any abuse of the DRM accounting model by entities. Because of that the staff support Alternative 3.

Staff recommendation

The staff recommend that when a change in risk management strategy requires a change in the entity’s target profile, the accumulated amount in other comprehensive income should be reclassified to profit or loss over the life of the target profile as defined prior to the change in risk management strategy.

Discussion

A question was raised regarding paragraph 13 of the paper as to whether the offsetting swap will be measured at FVTPL and be outside the DRM model. The staff clarified that an entity may either de-designate the original swap and then both swaps will be outside the model at FVTPL or the offsetting swap may be designated into the model and then the cancellation will happen within the DRM model. Both swaps will be still included in measuring imperfection and if they do not cancel each other entirely then that will be caught by the imperfection.   

The Board discussed the timing of the reclassification from OCI in the case of de-designation of old derivatives. The staff provided further explanation regarding the mechanics and the rules for keeping in or reclassifying from OCI and re-iterated that the deferral in OCI is justified if the economics of target profile and asset profile are still there and that is also the reason for the timing of the reclassification from OCI. In substance, this is similar to the discontinuation rules in IFRS 9.

The request was raised to provide a clear definition of a change in risk management strategy (e.g. when an entity changes the type of derivatives they are using, whether that is a change in risk management strategy or not), with examples.

Decision

The Board unanimously approved the staff recommendation.

Correction list for hyphenation

These words serve as exceptions. Once entered, they are only hyphenated at the specified hyphenation points. Each word should be on a separate line.