Quantitative standards for measuring the capital requirement
Interest rate risk in the banking book measurement system track record
An ADI’s IRRBB measurement system must have a reasonable track record in measuring this risk. Accordingly, before the ADI receives an IRRBB approval, the ADI must demonstrate to APRA that the system has undergone a sufficiently long period of initial internal monitoring prior to its use for the calculation of the capital requirement. The length of this monitoring period will depend upon the performance of the ADI’s IRRBB management framework and its track record in managing and measuring IRRBB.
Data
An ADI must have transparent and verifiable processes for collecting relevant data inputs (e.g. exposure information, details about customer behaviour and product and wholesale rate histories) on an ongoing basis. These processes must be consistent, timely and comprehensive across the ADI.
An ADI's IRRBB exposure data must be comprehensive in capturing all material exposures from all appropriate business activities, banking book items and geographic locations. The ADI must be able to justify that any excluded exposures, both individually and in aggregate, would not have a material impact on the overall estimate of its IRRBB capital requirement.
Interest rate data
For the purpose of determining the IRRBB capital requirement, the interest rates used and modelled by an ADI must include at least one yield curve in each material currency. Immaterial currencies may be combined into one or more groups and modelled using a single yield curve for each group based on a currency, or composite of currencies, broadly reflective of the interest rate characteristics of the group.
An ADI’s IRRBB measurement approach must model each yield curve using a model that appropriately captures variation in the volatility of interest rates along the curve.
Regardless of the type of model used, interest rate assumptions will depend to a significant degree on the ADI’s observation history of interest rates. The observation period must be at least six years, except where the ADI provides evidence that the use of a shorter observation period would provide a more appropriate distribution of future interest rates. Interest rate assumptions are not required to directly or exactly reflect the characteristics of the observation period but, where they do not, the basis for the difference must be explained and justified.
An ADI must update its interest rate data sets regularly and reassess its data sets whenever the interest rate environment is subject to major change.
Correlations between interest rates on the same and different yield curves must be recognised and included in an ADI’s internal model. Correlation assumptions must be sound, justifiable and implemented with integrity.
Part Attachment B - Banking (prudential standard) determination No. 4 of 2024 - Attachment B | Zoe
An ADI must document the source of the interest rates used. If a derived or modified yield curve is used, the basis of derivation or modification must be documented.
Data policies
An ADI must document its IRRBB data management policies and procedures. These policies and procedures must cover:
the collection of data;
processes for ensuring integrity, completeness, consistency and accuracy;
data storage;
application purposes; and
an outline of all data flows between systems, including whether any manual processes are involved in such flows.
Modelling repricing and yield curve risks
For the purpose of determining the IRRBB capital requirement for repricing and yield curve risks, an ADI’s internal model must measure the maximum potential change in the economic value of the banking book (EVBB), as a consequence of changes in interest rates, for a 99 per cent confidence level and over a one-year holding period (consistent with the soundness standard detailed in paragraph 30 of this Prudential Standard).
The economic value of an individual banking book item is calculated as the net present value as at the beginning of the holding period, of expected future notional principal and interest cash flows. Notional cashflows should be discounted using the actual and simulated wholesale market interest rates as appropriate. EVBB is the sum of the economic value of all banking book items.
An ADI’s internal model must estimate the 99th percentile of the distribution of EV0 minus EV1 where:
EV0 is EVBB minus an earnings offset (refer to paragraph 17 of this Attachment). In this case, EV0 and EVBB are both calculated using the discount rates prevailing at the beginning of the holding period; and
EV1 is EVBB minus an earnings offset (refer to paragraph 17 of this Attachment). In this case, EV1 and EVBB are both calculated using the discount rates simulated as prevailing at the end of the holding period.
EV0 and EV1 are calculated on the same items using the same outstanding balances and repricing dates.[7]
The principal component of the notional cash flows must be derived using an ADI’s repricing assumptions. The interest components relating to a principal payment must be calculated using historical wholesale rates (refer to paragraphs 15 and 16 of this Attachment).[8]
For banking book items with defined inception and principal payment dates, the historical wholesale rate applicable to a principal payment is the wholesale market rate applicable to a payment on the repricing date, taken from the yield curve applicable on the inception date of the instrument incorporating the principal payment. APRA will review the appropriateness of an ADI’s choice of wholesale rates for the purposes of this paragraph and paragraph 12 of this Attachment and may, in writing, determine that the ADI must use different rates for the purpose of determining its IRRBB capital requirement.
For banking book items that do not have defined inception or principal payment dates, assumptions must also be made about inception dates. In such cases, the historical wholesale rate applicable to a principal payment will be determined as described in paragraph 15 of this Attachment but using the inception and principal payment dates specified in the ADI’s repricing assumptions.
For the purposes of paragraph 13 of this Attachment, the earnings offset must be calculated as the economic value, as at the beginning of the holding period, of a notional twelve-month, equally weighted, monthly moving average portfolio of fixed for floating interest rate swaps. For this purpose, the ADI receives fixed and makes floating rate payments and the total principal amount covered by the swaps is equal to the sum of the book value of all banking book items.
The above approach to the measurement of repricing and yield curve risks is based on the use of value-at-risk techniques on a static balance sheet. An ADI that develops new methodologies, such as the use of dynamic simulation, is encouraged to approach APRA to discuss the use of such methodologies within the capital adequacy framework. If APRA has already approved the ADI’s internal model, it may be necessary for APRA to vary its approval to permit the use of a new methodology by the ADI.
Modelling basis and optionality risks
An ADI that does not have APRA’s written approval to exclude basis and/or optionality risks from its IRRBB capital requirement and does not have a risk measurement model for these risks, must use a method approved by APRA, in writing, for determining an appropriate amount of regulatory capital for those risks.
To the extent possible, the data sets and assumptions used for modelling repricing and yield curve risks and optionality and basis risks should be consistent.
An ADI’s internal model for basis risk must explain the historical variation of margins between product interest rates and the implied cost of funds. In addition, the requirements regarding correlation assumptions detailed in paragraph 8 of this Attachment must be met.
An ADI’s internal model for optionality risk must accurately capture the unique risks associated with optionality, including the non-linear price characteristics of implicit option positions.
An ADI’s internal model may require assumptions about the behaviour of its customers when measuring the capital requirement for basis and optionality risks. Some of these assumptions, including those related to prepayment rates on fixed-rate loans, will be embodied within the ADI’s repricing assumptions. Additional behavioural assumptions may be required for the purpose of measuring optionality risk, such as the volatility of prepayment rates or the relationship between those rates and interest rates. The measurement of basis risk will also require assumptions to be made about the variation of the ADI’s customer product rates around the implied cost of funds. These types of assumptions must be clearly documented by the ADI, including reference to the data on which the assumptions rely.
Interest rate risk in the banking book capital requirement
An ADI’s IRRBB capital requirement, as determined by its internal model, must be calculated as:
the amount estimated for repricing and yield curve risks which is the estimated 99th percentile of EV0 minus EV1 (refer to paragraph 13 of this Attachment); plus
the amounts estimated for basis and optionality risks as detailed in paragraphs 19 to 23 of this Attachment, except where the ADI has written approval from APRA to exclude basis and/or optionality risks; less
any amount resulting from assumed diversification benefits between repricing and yield curve risks, basis risk and optionality risk; plus
the embedded loss (or embedded gain if this amount is negative) which is defined as the sum of the book value of all banking book items minus EV0
subject to the capital requirement not being less than zero.
An ADI may only incorporate diversification benefit estimates if the ADI can demonstrate to APRA, using robust quantitative and qualitative analysis, that its correlation estimates are appropriate and take into account the uncertainty surrounding any such estimates (particularly in periods of stress).
Stress testing
An ADI must have in place a comprehensive and rigorous program of stress testing its internal model. Stress testing must include:
consideration of a breakdown in the ADI’s key modelling assumptions, such as its repricing assumptions; and
scenarios based on sudden changes in the level of interest rates and changes in the slope and shape of the yield curve.
An ADI’s policies and limits must reflect the results of stress-testing exercises and these results must be communicated to relevant senior management and the ADI’s Board, or Board committee, on a regular basis.
Validation
An ADI must have a robust and documented system in place to validate the accuracy and consistency of its internal model. The ADI must demonstrate to APRA that its internal validation process enables it to assess the performance of its internal model in a meaningful and consistent manner. As part of the validation of its internal model, the ADI must regularly compare actual loss experience against its estimates for those losses to ensure their reasonableness.
An ADI must have in place a robust process for validating changes to its internal model (including information that flows into that model). This would include a systematic process for reviewing the appropriateness of modelling assumptions and for making changes to those assumptions.