Dynamic risk management

Date recorded:

Cover note (Agenda Paper 4)

The cover note provided an overview of the outreach, i.e. who were the participants in the outreach and what was the focus of the main questions asked during outreach.

Appendix A provides a statistical analysis of participants by geographic region.

This paper was not discussed.

Executive summary of feedback from outreach (Agenda Paper 4A)

The paper was structured as follows:

  • Background: the outreach focused exclusively on banks and provided the views and concerns from a banking preparer perspective only and a reminder of the objective of the dynamic risk management (DRM) model
  • Feedback received: analysed separately for each of the topics: objective of the DRM model, key advantages of the DRM model and key areas for improvement in the DRM model

The paper summarises the following key messages received from the participants in the outreach:

  • Almost all participants supported the objective of the DRM model. However, as risk management strategy commonly defines the target profile on a risk limit basis, this should be incorporated to the DRM model to better reflect the risk management view.
  • Most participants indicated that the issue of accounting mismatch is still not fully resolved in the DRM model or is inconsistent with accounting practices. All participants were concerned about the potential impact of recognising changes in fair value of derivatives in OCI on the regulatory capital and volatility of the capital.
  • Many participants also commented on implementation costs and change of current practices, users’ need for information about DRM and disclosure of sensitive information.
  • Almost all participants responded positively to the ability to designate a net open risk position or core demand deposits. However, in respect of prepayable assets, most participants suggested that the designation of the layer of nominal amounts instead of a percentage (proportion) of portfolio should be allowed.

Question for the Board

The staff asked the Board whether the Board have any views, comments or questions on the feedback that would help staff to formulate the plan for the next steps in the project.

Board discussion

Board members discussed the feedback received with regard to the objective of the DRM model. To one Board member it seemed like the participants in the outreach would like the model to demonstrate whether the hedging strategy of the entity is effective. However, the objective should remain devoted to having the most faithful representation of the economic performance of the entity. Therefore, avoiding accounting mismatches should not be the main focus of the model. The Vice-Chair echoed this by saying that no volatility in the income statement would not be achievable. The question was only how much volatility constituents would be willing to accept. One Board member said that the Board would soon have to decide whether it would use the target profile in the model.

A Board member noted that two thirds of the participants were from Europe. The Vice-Chair said that the focus on Europe in the outreach was due to many EU banks having a particular interest in the outcome of the project. This is because of the structural features in the EU, for example a high percentage of fixed-rate assets and the way in which EU banks are funded.

Interest rate risk management strategies—Information from outreach (Agenda Paper 4B)

This paper summarised detailed information about participants’ interest rate risk management strategies and activities that was obtained during the outreach.

The key points are:

  • The participants apply diverse interest rate risk management strategies, processes and techniques.
  • The participants’ risk management strategy is not entirely focusing on stabilising net interest income (NII) but uses different measures like present value of as single basis point change in interest rates based on gap analysis (PV01), economic value-at-risk (EVaR) or economic value of equity based on cash flows (EVE). Almost all the participants said that their interest rate risk management strategies are expressed in terms of the risk to both EVE and change in NII, consistent with the Interest Rate Management in Banking Book framework (IRRBB) as defined by the Bank of international Settlement (BIS).
  • The participants’ interest risk management strategy is expressed by risk limits and not as a single outcome. The risk limits are determined in accordance with banks’ risk appetite and the only external restriction relates to the regulatory requirements, e.g. the European Banking Authority has a guideline indicating banks need to report immediately if the economic value change in the banking book is more than 20% of the Tier 1 capital for a 200bps parallel shift of interest rate.

The participants responded positively to the DRM model being a principle-based model and not prescribing eligible risk strategies, risk aggregation approaches or behaviourilisation methods. They described elements that are typical in their risk management strategies like approvals and changes, risk management levels, risks metrics, range of acceptable risks limits, risk aggregation and time horizon, scope of assets and liabilities (being cconsistent with the principles set out in the IRRBB framework) and behaviouralisation methodologies. Each of these items was explained by participants in greater detail in the paper, including the illustration of the sequence of risk management activities (Appendix A to the paper) and illustration of setting acceptable risk limits (Appendix B to the paper).

A few participants asked the Board to clarify how the DRM model should be applied when the risk management strategy is not explicitly set by the senior management, and instead only set individually for each subsidiary within the group, i.e. a decentralised approach to interest risk management. Specifically, they asked whether hedge accounting designations based on the DRM models implemented at each subsidiary level can be considered as valid designations for the DRM model at a group consolidated level.

This paper was discussed together with Agenda Paper 4C (see below).

Feedback summary: Target Profile (Agenda Paper 4C)

This paper provides a summary of feedback received from the outreach on the target profile element of the DRM model which, based on the Board’s tentative decision, should be defined as a single outcome.

All participants said that their interest rate risk management strategies define risk limits, which is expressed as a range of possible outcomes and not as a single outcome. Furthermore, the participants said that if a net open risk position remains within a bank’s risk limit, then the bank considers its dynamic interest rate risk management activities as being ‘successful’. Anytime that the open risk position is close to the risk limits, the bank may carry hedging activities. As such defining the single outcome will not represent their risk management strategy and activities and it would not be consistent with the regulatory framework. Therefore, almost all participants said that the proposed single outcome will be a huge operational challenge in the DRM model and that the Board should reconsider it. The staff note that an entity usually develops a limited range around a pre-defined desired net open risk position and therefore should be able to articulate that as a single outcome target. Furthermore, incorporation of risk limits would represent significant conceptual challenges (e.g. lower recognition of imperfect alignment in profit or loss).

Many participants asked the Board to clarify the meaning of the target profile and to provide additional application guidance. The participants also asked to provide sufficient guidance that would help them to determine what changes in entity’s risk management strategy would lead to changes in target profile and what accounting impact the frequent changes to the single outcome would have.

Participants also asked the Board to illustrate the application of the DRM model through examples based on different risk management strategies than stabilising net interest income. This is because banks are using many different measures like economic value of equity or interest gap sensitivities.

Appendix A to this paper included a chart that illustrates a potential alternative approach to describe the target profile (using the same components of the DRM model as per the Board’s tentative decisions), contemplated during outreach with participants.

Board discussion

The Board discussed Agenda Papers 4B and 4C together.

Board members discussed risk limits and how it would be difficult to define them. The wider risk limits are set by entities, the less information will be available about the quality of performance as broad risk limits will always look better than narrow limits. One Board member expressed concerns that using risk limits would also mean that derivatives can be included that increase interest rate risk. In his view, this is contrary to the objective the Board is trying to achieve with the DRM model.

A Board member said that a departure from the target profile would likely require more disclosure to explain the performance of the entity. The target profile is used by less and less banks.

The Board would also have to examine how the shifting focus to economic value of equity influences the DRM model, which is based on the stabilisation of net interest income. Board members also said that incorporating the decentralised approach of hedging (which is prevalent in most banks) would be challenging.

Feedback summary: Designation of expected cash flows and impact on imperfect alignment (Agenda Paper 4D)

This paper provided a summary of feedback from outreach on the designation of financial assets and financial liabilities based on expected cash flows in the DRM model and the impact of changes in prepayment assumptions on the recognition and measurement of imperfect alignment.

Participants supported the Board’s tentative decision for the use of expected (i.e. behaviouralised), instead of contractual, cash flows for prepayable financial assets, demand deposits, and future transactions as in their view this allows a better alignment of the entity’s interest risk management and activities.

Many participants indicated that a sizeable portion of their dynamically managed interest exposure represents the equity model book and hence they were supportive of including it as part of the DRM model. The participants also supported the Board’s principles-based approach with regards to the requirements for defining and designating portfolios within the DRM model but they indicated that more application guidance would be required to provide better clarity on key matters.

Almost all participants said that they manage the repricing gap based on expected prepayment amounts and timing and that options not used mainly due to cost reasons. Prepayment risk is generally not the issue as the participants are either compensated for loss of interest income or the loans are variable rate loans. However, the opportunity to incorporate the use of derivatives with optionality in the DRM model was welcomed by participants as this could help managing the net interest income. 

Most participants were of the view that a change in prepayment assumptions should not necessarily lead to the immediate recognition of a gain or loss (i.e. imperfect alignment) because the effect of changes in assumptions generally relate to unhedged positions as only assets considered as stable are hedged. Furthermore, the participants said that they are currently applying either ‘bottom layer’ approach or ‘first payments made/received’ approach which does not reflect the effect of all changes in prepayment assumptions so they requested the Board to consider to what extent changes in prepayment assumptions should be recognised as imperfect alignment within the DRM model.

The participants indicated that the effects of changes in assumptions should be reflected in net interest income (NII) prospectively over time because the effect of these changes ultimately result in change to future NII.

The DRM model allows for designation of a percentage (portion) of a portfolio. However, most participants said that the designation of a layer of nominal amounts of the qualifying assets should be allowed as this would mitigate the risk of unnecessary profit or loss volatility from undesignated risk positions (i.e. an entity would recognise any gain or loss in profit or loss only when the prepayments are high enough to breach the ceiling of the bottom layer).  Furthermore, in their view the ‘bottom layer’ approach is operationally simple because it does not require tracking and amortisation. Some participants argued that this approach is representative of their risk management view. The paper contains two examples illustrating how the ‘bottom layer’ approach works if used either for risk management purposes or if only used for accounting purposes.

The staff indicate that even applying the ‘bottom layer’ approach, the tracking would not be avoided entirely because individual portfolios are being separated onto bottom and upper layers. Hence, some tracking would still be required. The main aim of the approach is to avoid profit or loss volatility. The ‘bottom layer’ approach is also not reflecting the objective of the DRM model.

This paper was not discussed and will be brought back to a later meeting.

Feedback summary: Recognition of fair value changes in OCI (Agenda Paper 4E)

This paper provided summary of feedback received regarding the Board’s tentative decision to recognise the changes in the fair value of derivatives used to align asset profile with the target profile in other comprehensive income (OCI).

Almost all participants expressed concerns about the potential implications of this mechanism on regulatory capital and all participants agreed that recognition of changes in the fair value of derivatives in OCI is worth exploring only if the prudential regulators provided a filter allowing to eliminate any volatility in regulatory capital that might arise from the application of the DRM model (i.e. similarly to the prudential filter currently in place for the cash flow hedge reserve).

Some of the participants were also concerned about the volatility of IFRS equity as being misleading for the users of financial statements because of the potential impact on the key performance indicators such as return on equity or leading, in extreme circumstances, to a negative equity balance. These participants suggested that the Board should consider changing the DRM model to use the fair value hedge mechanism.

The staff noted that most respondents of the 2014 Discussion Paper rejected the fair value approach model and that it was then suggested that a cash flow hedge accounting model reflects DRM activities better when interest rate risk is managed in terms of cash flow variability and it does not introduce significant operational complexity due to onerous tracking and amortisation requirements.

The staff asked Board members whether they have any questions, comments or whether there is any feedback that requires further research.

Board discussion

Board members discussed the concerns raised by constituents about volatility in IFRS equity, caused by OCI fluctuations. The Vice Chair acknowledged these concerns and said that if there is a risk that the equity volatility could lead to a run on the bank, this needs to be carefully considered when developing the model. One Board member suggested to reach out to regulators to understand whether this could happen.

Another Board member said that he could not fully understand the desire for a fair value hedging model as that had been proposed with the portfolio valuation approach and received negative feedback from banks. That is why the Board was now focusing on cash flow hedging, however, if that is also undesirable, it would be difficult to come up with an alternative.

The staff concluded the session by saying they will bring back a project plan at a future 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.