In the exposure draft ”Risk Mitigation Accounting” published on December 3, 2025, the IASB proposes a new model to account for interest rate risk management at portfolio level, particularly for banks. The objectives of the risk mitigation accounting model (RMA model) are to appropriately reflect the economic effects of asset liability management (ALM) and to provide meaningful information on the designated derivatives used for this purpose in IFRS financial statements. These objectives are to be achieved through the greatest possible use of ALM metrics, methods and processes for IFRS purposes, a new mechanism for deriving the accounting to be recognised, a new accounting logic and meaningful disclosures. In addition to an overview of the RMA model, this article discusses the extent to which the RMA model meets the requirements for proper accounting from an interest rate risk management perspective.
Points of criticism of the previous IFRS regulations from the perspective of banks’ interest rate risk management
Due to their business model, banks are exposed to interest rate risk, which they regularly measure and manage at portfolio level as part of their ALM. The previous IFRS regulation for the presentation of these activities – the portfolio fair value hedge for interest rate risks in IAS 39 (IAS 39 PFVH) – has been criticized for years due to a lack of transparency, limited eligibility of underlying and hedging transactions, insufficient consideration of the dynamics of the managed portfolios, and the absence of performance measurement of interest rate management. While interest rate risk management is forward-looking, IFRS accounting naturally focuses more on a retrospective view. Under IAS 39 PFVH, the dynamics of risk management are addressed through regular de- and redesignation, which in turn result in a correspondingly large number of amortization effects that are difficult to explain and forecast or manage.
Brief overview of the RMA model
To overcome these points of criticism, the RMA model takes into account the existing interest rate risk strategy and uses the risk metrics and methods applied by the credit institutions as far as possible. It is designed as an ongoing process consisting of six steps. The cycle of the respective steps is to be aligned with the actual time sequence of the ALM activities and the reporting cycle of the accounting entity, so that the RMA process will not be purely sequential, as illustrated in the following flow chart:
The ALM portfolio forms the basis for deriving the RMA portfolio in the first step, subject to certain eligibility criteria. For example, non-derivative financial instruments that are recognized at fair value through profit or loss are excluded. Internal transactions are also not permitted, even if they play a central role in a bank’s ALM. From the eligible nonderivative transactions included in the underlying portfolios, the current organic net repricing risk exposure (NRRE) is determined using the metric applied for risk management purposes. Other key parameters relevant for risk measurement, such as cash flow modelling and the structure of repricing time bands, are likewise derived from ALM.
In the second step, the risk mitigation objective (RMO) is determined as the portion of the NRRE that is to be mitigated by the ALM activities and offset in the accounts as part of the RMA model. The RMO is specified for each repricing time band by comparing the NRRE with the net mitigation effect of the derivative financial instruments included in the model (designated derivatives), in particular the so-called risk mitigation test.
In the third step, benchmark derivatives are constructed in order to make the RMO determined in the respective risk metric measurable for accounting purposes. Their hypothetical mitigation effect must correspond exactly to the RMO in each repricing time band. The construction methodology depends heavily on the risk metric applied. The benchmark derivatives are not recognised separately; rather, they are hypothetically continued in a manner analogous to the designated derivatives and considered in subsequent process steps and RMA cycles at their updated values.
In the fourth step, the NRRE derived in the first step, usually for several points in time, is recalculated retrospectively for each repricing time band, taking into account updated information regarding previously made assumptions- for example on the expected amount of early repayments. The RMO determined in the second step for the respective point in time is then compared with the corresponding, if necessary adjusted, NRRE and reduced accordingly if it exceeds the adjusted NRRE in terms of amount.
If the RMO has been adjusted in the fourth step, the previously constructed benchmark derivatives must also be adjusted in a fifth step so that the aggregate hypothetical management effects from the benchmark derivatives constructed in steps 2 and 5 corresponds to the adjusted RMO for each repricing time band at each point in time.
While steps 1 to 5 are determined by the respective risk metric, the final sixth step derives the necessary accounting adjustment entries in currency units across all repricing time bands. These entries are derived from the cumulative earnings components of the designated derivatives and the constructed benchmark derivatives since the start of an RMA model as part of two ”lower of” tests.
To ensure that the future compensation effect on net interest income embodied in the recognised risk mitigation adjustment (RMA adjustment) does not exceed the maximum compensation requirement associated with the original net repricing risk exposure, it must be assessed at each reporting date whether indicators exist, such as the premature sale of a significant volume of fixed-interest assets, that suggest the existence of a RMA adjustment excess. If such indicators are present, the present value of the NRRE must be calculated and compared with the recognised RMA adjustment. If the recognised RMA adjustment exceeds the present value of the NRRE, it must be reduced immediately by the excess amount and recognised in profit or loss.
Requirements for the RMA model and their fulfillment from a risk management perspective
The specific requirements for the RMA model can be derived in particular from the points of criticism of the previous IFRS regulations listed above. In our opinion, the key requirements from a risk management perspective can be summarized as follows:
- The RMA model should lead to economically meaningful IFRS accounting
- The RMA model should enable a transparent and comprehensible accounting process
- The results of the RMA model should be easy to predict
Ad 1: Economically meaningful IFRS accounting
The net interest income resulting from the application of the RMA model generally corresponds to the net interest income from the hedged net interest position. Only earnings components from the designated derivatives remain in net trading income, unless they have a risk-mitigating effect overhedging). In contrast, net interest positions not hedged by designated derivatives (underhedging) only lead to a fluctuation in net interest income if such a fluctuation results from the respective subsequent measurement method, i.e. only from variable-interest transactions in the event of a change in the reference interest rate and no recognition of interest-induced fair value changes in the income statement. However, interest rate-induced fair value changes in financial instruments measured at fair value through other comprehensive income are still recognized in other comprehensive income and reported in the balance sheet. The application of the RMA model does not result in a separate, purely accounting-driven result from hedging relationships. Amortisation effects known from IAS 39 PFVH do not exist in the RMA model either, as the realization of the RMA adjustment in subsequent periods follows the realization profile of the benchmark derivatives (ED IFRS 9 B7.4.17). In particular, the income statement resulting from the application of the RMA model is therefore highly informative from an economic perspective.
It should be noted that where financial instruments measured at fair value through other comprehensive income are included, volatility in equity arising from the fluctuation in the revaluation reserve is not offset by the accounting mechanics of the RMA model, even in the case of a full hedge. Furthermore, the necessary adjustments to the realization of the RMA adjustment in subsequent periods after an RMA
excess has been recognized in profit or loss are generally to be derived on a lump-sum basis, if necessary on a straight-line basis (ED IFRS 9 B7.4.23). Due to the eligibility criteria, the basic population of transactions from the ALM portfolio and IFRS accounting are not aligned. This can be due to the inclusion of primary, interest-bearing financial instruments measured at fair value through profit or loss or AT1 bonds classified as equity under IFRS. This means that discrepancies remain between risk management and the presentation in the balance sheet and income statement in accordance with IFRS. This can no longer be resolved through targeted designations or proxy hedges.
Ad 2: Transparent and comprehensible accounting process
In steps 1 to 5, the RMA model is based on the metrics, methodologies and assumptions used in ALM and thus enables a high degree of synchronization between the ALM and RMA processes. The derivation of adjustment postings in step 6 is closely and comprehensibly linked to the results from the previous steps, so that the RMA process is fundamentally transparent and comprehensible from a risk management perspective. The cyclicality of the steps in the RMA process, which must be aligned with the ALM activities, is also a considerable advantage in terms of transparency and traceability. The division of the net risk position into an organic net risk position (NRRE) and a derivative as the starting point for deriving the RMO in step 2 is also easy to understand and is also unavoidable for accounting purposes due to the mixed model in accordance with IFRS.
A retrospective adjustment of values in the past, as provided for by the RMA model in steps 4 and 5, is unusual in risk management. Furthermore, there are no comparable ALM activities to check for the existence of an RMA excess and, if necessary, its immediate recognition in profit or loss.
Ad 3: Good predictability of the results
Similarly, the extensive use of metrics, methodologies and assumptions used in ALM generally lead to good predictability in the context of scenarios. As in the ALM, this is naturally limited by the risk of inaccurate assumptions made, although there are also clear, predictable interdependencies for these factors in the RMA model.
However, the predictability is limited with regard to the future effects of the review for the existence of an RMA excess and, if applicable, its immediate recognition in profit or loss. On the one hand, this is due to the inherently uncertain occurrence of unexpected changes causing an RMA excess, and on the other hand, specifically for the RMA model, to the complex interaction between the retrospective consideration of unexpected changes in steps 4 and 5 and the occurrence or amount of an RMA excess.
Conclusion
Overall, it can be concluded that the RMA model offers significant potential for improving economic relevance, as well as transparency, traceability and predictability compared to the IAS 39 PFVH. In particular through the extensive use of ALM metrics, methods and processes and a comprehensive accounting framework.
Nevertheless, certain conceptual aspects in the RMA model, such as the check for an RMA excess and possibly its effect,may reduce these advantages. Initial trial calculations reveal further conceptual issues, the convincing solution of which depends on their practical operability. It therefore remains uncertain whether the RMA model is advantageous from an overall cost-benefit perspective. It should be the subject to intensive investigations during the consultation phase of the draft standard and associated trial calculations.