Dynamic risk management

Date recorded:

Potential refinements to the Dynamic Risk Management (DRM) model—Risk Limits (Agenda Paper 4A)

During the April 2021 meeting the Board discussed challenges identified during the outreach that are key to the viability and operability of the DRM model. This paper sets out the staff’s preliminary views on potential refinements to the DRM model in order to address one of these challenges and to align closer the DRM model to entities’ risk management practices by incorporating the concept of risk limits into the target profile.

Issues being addressed by the potential refinements to the DRM model

  • Hedge objective—risk management practices look at the liabilities together with the assets the entity originates from the funds received and hence the assets and liabilities must be considered in a combined/net view instead of using gross designation
  • Dynamic nature—portfolios of assets and liabilities are not static and often require changes to both hedged items and hedging instruments. Frequent de-designation and re-designation, and tracking associated amortisations under IAS 39 is complex, costly and susceptible for operational mistakes
  • Optionality in the underlying portfolios—the interest rate risk is managed on the portfolio level and not on individual loan level. Each loan that contains optionality creates uncertainty when cash flows in the underlying portfolio arise. Current activities to mitigate this risk as proxy hedging, layering techniques or frequent de- and re-designation are based on the aggregation of individual underlying loans rather than on portfolio as the whole; hence there is a difference between the level at which the risk is mitigated and hedged
  • The meaning of effectiveness in portfolio hedging—management of portfolios is on the basis of expected behaviour, hence the acceptance of the uncertainty of cash flows (due to optionality) is a part of the risk management strategy. The changes in cash flows would typically cause adjustments through use of new derivatives. However, the unhedged position is also accepted if it is within risk limits set up by the management. Therefore, staying within risk limits should not result in misalignment that affects profit or loss. Splitting portfolios into tranches or layers is not necessarily consistent with the actual risk management strategy
  • A risk mitigation approach to macro hedge accounting—there are several approaches to implement an accounting model for DRM. The difference between them is in relation to the treatment of the unhedged portion of the underlying portfolios but all of them present issues and challenges. Based on the feedback received, preparers favour a hedge accounting model that is based on risk mitigation, i.e. the extent of risk exposure that is actually hedged and not including the portion that is left unhedged

Potential refinements to the DRM model

One of the key elements of the DRM model is the target profile. Based on the outreach feedback received, the target profile should be a net open risk position, should be dynamic and should constitute a range of acceptable outcomes. Therefore, the refinements to the target profile and benchmark derivative elements of the DRM model will focus on the following three elements:

  • Definition and objective of the target profile—the risk management strategy and risk objective elements of the target profile should be separated. The staff is of the opinion that the target profile should only represent the risk management strategy and the new element could be included in the DRM model, which would represent the risk management objective element. Then, the target profile could be defined as the “acceptable open risk position” given an entity’s risk management strategy. This position represents the acceptable range (risk limits). The refinement to the definition and role of target profile is more intuitive, closer aligned with risk management strategies and it is also consistent with an example in IFRS 9:B6.5.24(a)
  • Inclusion of a risk mitigation intention—this will fulfil a similar role in the DRM model as the risk management objective in the general hedge accounting model in IFRS 9 and it will be subject to certain boundaries, namely it cannot create new risks and it should transform the current net open risk position to a residual risk position that is within the target profile. The risk mitigation intention is a single-outcome element, i.e. a fixed amount of risk to be mitigated through derivatives within a set period of time that would depend on the frequency of changes to the underlying portfolio. The risk mitigation intention can be expressed as the portion of the current net risk exposure that the entity intends to mitigate and be determined as PV01 or in nominal terms. It is directly linked to an entity’s target profile that would mandate how much risk the entity is willing to accept or leave open. The changes to the risk mitigation intention will not trigger discontinuation of DRM model, however they will trigger retrospective assessment of performance against the risk mitigation intention on a single outcome basis and that will impact the measurement of misalignment that is reflected in the profit or loss
  • Construction of the benchmark derivative—after the proposed revision to the target profile, the construction of the benchmark derivative should be based on the risk mitigation intention in order to be used as a theoretical single outcome derivative and be used for measurement purposes. The benchmark derivative cannot impute the terms of the designated derivatives if they are not reflecting the risk mitigation intention. There may be several reasons why the risk mitigation intention might be different from designated derivatives e.g. different benchmark rate, difference in tenor basis, or maturity or volume mismatch

Retrospective performance assessment

The potential refinements to the DRM model would require adding two retrospective performance assessments for the purpose of determining the impact of changes to the current net open risk position:

  • Assessment whether an entity mitigated interest rate risk—i.e. assessment whether the effect of unexpected changes to the current net open risk positions during the period resulted in the entity creating risk, namely being in an over-hedged position. The ‘lower of’ test would be applied in order to define whether the misalignment will be reported in the profit or loss. The test will compare the risk mitigation intention (represented by the benchmark derivative) with the current net open risk position at the end of the period under assessment
  • Assessment whether the target profile has been achieved—i.e. assessment whether the risk mitigation intention has transformed the current net open risk position to a residual risk position that falls within the target profile. To the extent that the residual risk position falls within the target profile, no misalignment will be reported in the profit or loss

Operability of the DRM model—required inputs

Following inputs for the DRM model will be required in order to make the potential refinements operable:

  • Conversion of various risk metrics into a target profile by maturity buckets—in practice, many different risk metrics are used (e.g. changes in economic value of equity, changes in net interest income or interest rate gaps). These metrics would need to be converted into a set of uniform risk limits to be used as the basis for the performance assessment. However, if an entity only sets one overall risk limit, instead of identifying risk limits for each maturity bucket, the risk mitigation intention may be not specific enough and potentially misused. The staff plan to do further research and analysis in this matter
  • Capability to distinguish existing positions from new business—entities should be able to track new business so it could be excluded from retrospective assessment in the DRM model whether the risk mitigation intention has been achieved
  • Capability of determining the current net open risk position—entities should be able to separate the effect from derivatives as it is in the current net open risk position from assets and liabilities that drives the assessment for risk mitigation purpose.

Question for the Board

The staff was not looking for any decisions based on this paper. Instead they asked only for questions or comments on the potential refinements presented. Such comments will be considered by staff and the proposed refinements will be presented for the Boards tentative decision at the future meeting.   

Illustration of potential refinements to the DRM model—Risk Limits (Agenda Paper 4B)

In this paper, the staff illustrated how the potential refinements within the DRM model would work. The presentation includes:

  • A diagram summarising the interaction of the key elements of the DRM model before and after refinements
  • The DRM cycle including prospective and retrospective actions required and DRM boundaries
  • Application and mechanics of the DRM model with a numerical example
  • DRM model performance when the risk managers decide to trade positions outside of the DRM boundaries (i.e. example of over-hedging)

This agenda paper should be read in correspondence with Agenda Paper 4A.

Board discussion (Agenda Papers 4A and 4B)

Overall, the refinements to the DRM model were well received by the Board, especially the facts that despite of introducing the range and risk mitigation intention, the model still meets the overall objectives discussed in the past and that entities will be still able to measure the performance (i.e. there is only one reference point).

The following comments and challenges were raised by the Board:

  • What mechanism can be included in the model to provide protection against changing the range as part of managing the misalignment or to achieve an accounting outcome? Staff responded that the range is not the main contributor to the ineffectiveness. The ineffectiveness is the difference between changes in fair value of the actual derivatives and benchmark derivatives hence the misalignment will be captured. Furthermore, the model is reflecting the risk management strategy of the entity. The purpose of the limits is to manage the business of the entity and the risks that the entity is facing. Hence, changing the limits is not that easy as they will affect the business or capital charges. Changing the limits is not an accounting decision but a business decision. The staff considered including potential disclosures when the entity is moving outside the range or if the window range is really wide so it would be visible for the users of the financial statements (or the regulators) to what extent the entity is managing its risk within the limits.
  • Board members challenged that there will be a potentially different outcome for two identical banks doing exactly the same but having a different range or different risk mitigation intention. The staff responded that this is the case with any hedge accounting model as it depends how much risk an entity wants to hedge. It was reiterated that the risk limits are aligned to risk management strategy and not with an accounting outcome. Entities may wish to hedge a portion of the risk and the refined DRM model is designed to capture that. One Board member suggested that the other information should be provided (disclosed) to assist with the comparability and understanding the differences between banks’ risk limits and the choices on their risk mitigation intention.
  • It was raised that because there is no single outcome, only a range, that the documentation of such a range is key, especially if it relates to management intention. The range should be documented at the start of the hedging. One Board member asked whether the result of the hedge would be within reserves (akin to a cash flow hedge) or within P&L (akin to a fair value hedge). The staff responded that this will still be looked at and it will be brought to the Board for discussion in a future meeting. A further question was raised in relation to ineffectiveness: whether an entity has to measure and recalibrate the hedge prospectively and whether the ineffectiveness is retrospective which will affect P&L. The staff responded that ineffectiveness is the measure to see whether the risk was mitigated and hence it is a retrospective test. However, it also has an impact on what should be done prospectively to mitigate the risk.
  • Questions were raised regarding the interaction between the risk management strategy and the profile specifically, (i) how much entities can change it, (ii) how it will be reflected and what will be disclosed around it, (iii) how staff envisage entities applying this, i.e. to different portfolios, (iv) how it will be broken down, i.e. is it a risk management strategy for the company or for different portfolios and finally (v) how all the changes will tie together. One Board member indicated that the level of disaggregation will affect the level of misalignment and how much will be recognised in P&L. The staff responded that if an entity changes its risk limits, this would be a discontinuation of the hedge and the entity would have to start a new one. In the case of one model overall versus one model per portfolio, the staff believes that from the risk management perspective, it is assessed at the company level. However, there is a problem because currently the different currencies or assets have different characteristics and cannot be mixed in the model. The staff admitted that this would need to be looked at further in order to align it with the entities’ risk management.
  • It was noted by one Board member that aligning with the risk management strategy and introducing the range created a double level of ineffectiveness, i.e. the range (as being outside the range) and a single point represented by “risk mitigation intention”. However, as the Board member stated, the single point in the measurement of ineffectiveness is a non-negotiable part of package. Furthermore, the question was raised whether the risk management intention must be defined before start of the hedging, i.e. before the derivatives are put in place. The staff responded that it should be clear from the start, but given this is a dynamic model, the intention may change quite frequently. The staff reiterated that risk mitigation intention has to be within risk limits but also risk mitigation intention cannot exceed the current net open position, i.e. an entity cannot hedge more than it has. The current net open position can be only moving towards zero and this forms the discipline, boundary, making sure that the entity is not introducing risk. On the question of whether an entity will assess during the period or at the period end when performing the range and risk mitigation intention assessment, the staff responded that yes, it is the case, but the period may be as short as a day because it depends on how often an entity is updating its risk management intention.
  • Board members asked to make it clear that introducing the range does not mean that there will be no ineffectiveness. In their view, the range should be reasonable and not driven by accounting incentives. Furthermore, at this point, disclosures on what the risk limits are should not be discarded on grounds of this may be sensitive and commercial data, because the Board needs to think about the discipline when applying the ranges.
  • Board members also challenged how the preparers benefit from this model. The staff responded that they would have the ability to change the risk mitigation intention without discounting the hedge. The staff added that they received feedback from one of the banks that this model better reflects their risk management strategy, but it is limiting their accounting options. One Board member added that if the banks will be able use their risk management strategy as the basis for accounting designation, this would provide better alignment and that would then help facilitations with the investors.
  • There was a challenge as to whether the risk management should be optional or should become mandatory. Hedge accounting is an option. It was questioned whether the risk management is faithfully portraying what the entity is doing, i.e. does an entity really have an option not to portray that. One Board member indicated that the answer for this question has a knock-on effect on what was done in Phase 1 (classification and measurement) of IFRS 9, specifically whether an entity has something that is measured at fair value or amortised cost? It was also questioned why there was so much emphasis on Day 1 accounting, if the Day 1 assessment is immediately overridden by Day 2 accounting i.e. by dynamic risk management. Furthermore, the Board member reminded the staff that whatever will be decided, including risk management disclosures, this should not contradict what was said and what is disclosed by an entity in respect of classification and measurement.
  • Board members brought to the staff’s attention that the paper mentions ‘over-hedging’ but it is silent about under-hedging. If the risk limits are missed out, that can be in both directions and it causes ineffectiveness in both cases. It was reiterated that over-hedging means creating a risk position and not mitigating risks. It was suggested to change the current terminology of ‘over-hedging’ or ‘under-hedging’ by saying that an entity is “outside the risk limits”.
  • There were also comments regarding the following wording: (i) “acceptable open risk position” – Board members suggested changing this into something more neutral like “agreed or target open risk positions”, and (ii) the term “intention” is a subjective term in the Board’s view and hence it was proposed to change it to “risk mitigation objective” to indicate that this is a single outcome measure. Furthermore, there were some concerns about a new terminology introduced as “risk mitigation intention” that is nowhere used in IFRS Standards.

The staff said that they will take into account the Board’s comments. The Board will decide on the refinements to the DRM Model during the next meeting. 

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.