Response to consultation on draft Guidelines on IRRBB and CSRBB
Go back
With respect to the IRRBB measurement assumptions, ABI would recommend clarifying that paragraph 105 should not be intended as requiring inflation to be kept constant across the scenarios.
The proposed approach (Chapter 4.4. of the draft GL) seems far too broad, as in practice all instances where a bank deviates from any of the requirements of the Guidelines can fall in the scope of “non-satisfactory” internal models, to be disqualified in favor of standardised methodologies.
As a premise, it should be highlighted that the standardised methodologies represent a useful tool to ensure the proportionality of the framework and might be suitable to provide an acceptable estimate of risk, especially in case of banks of lower size and/or complexity.
However, they necessarily rely on simplistic assumptions which make them by definition less accurate than internal models (but not more conservative, for the reasons said in the response to Question 1). The more complex the bank, the more sophisticated models are needed to manage and measure interest rate risk, thus implying that, where internal models are available, they should normally be considered preferable. Mandatory switching to standardised methodologies should therefore be limited to instances where severe deficiencies are observed in internal models (and are not occasional but lasting over time).
In ABI’s view, the GL should be amended accordingly. The GL should specify that the internal models shall be deemed as “not satisfactory” within the meaning of Article 84(3) CRD only in case the observed deficiencies of the models materially affect the reliability of the results, thus it can be reasonably assumed that the standardised methodologies are actually better suited to measure the IRRBB of the bank than its internal models.
This should in principle exclude from qualification of the internal models as “not satisfactory”, those instances where a bank happens not to be fully compliant with the GL in regard of provisions pertaining to formal requirements and processes. In these cases, repair actions should be required, rather than switching to standardised methodologies.
With respect to Chapter 4.4 of the GL, ABI would also recommend better clarifying how the proportionality principle shall be applied. While such principle is mentioned in paragraph 118, it is not clear how it is expected to be applied in practice. Providing guidance in this respect would be very beneficial.
In addition, the GL should clarify that, in case a bank is required by its supervisor to switch to the standardised methodology, an appropriate timing shall be granted to implement the switch (at least one year). Also, the process for subsequent return to the internal models should be clarified.
The key issue with regard to the proposed definition of CSRBB is that the proposed perimeter of application is not consistent with such definition (see the response to Question 4 below).
Moreover, certain aspects could be clarified, such as consideration of the changes in credit spread affecting bank’s own liabilities, the treatment of credit derivatives hedging credit risk on a single exposure or a portfolio of exposures, as well as interpretation of changes in the NII perspective (provided that increased market credit/liquidity spreads – hence market perception of increased risk - would more often result in higher NII).
More generally speaking, clarification on the intended use of the results of the CSRBB assessment would be very helpful for banks to identify the methodological approaches and assumptions which best fit for purpose.
The CSRBB is defined as risk driven by changes of the market price for risks other than IRRBB and idiosyncratic credit risk. This implies that CSRBB can be consistently assessed only with respect to those items that are exposed to such movements in the market prices, i.e. are actively traded on financial markets.
It is inherently nonsensical – before being impossible in practice - to determine market spreads for instruments for which no active market exists. In other words, banks cannot see how the distinction between the credit spread related to idiosyncratic risk and the market spreads can be operationalized in the absence of a market where the latter can be retrieved.
In any case, it should be borne in mind that such changes would not affect a bank’s equity, unless the item is accounted for at fair value. This is also consistent with the business model under which an item is held by the bank: under the IFRS (accounting standards applied by the vast majority of EU banks) items accounted for at amortising cost are supposed not to be sold before maturity and therefore by definition not exposed to market risk – so they should consistently be excluded from the scope of the CSRBB framework.
Indeed, the limitation of the scope of the CSRBB to items accounted for at fair value is crystal clear in the BCBS standards. The definition of CSRBB illustrated in the Annex 1 of the BCBS standards “Interest rate risk in the Banking book” (Figure 1, reported below) clearly shows that that the Basel Committee intends the CSRBB as referred only to items accounted for at their fair value (mark-to market).
Any enlarged scope would ostensibly goldplate the BCBS standards and raise level playing field issues.
On the other hand, including items, which are not exposed to market spread movements, in the scope of the CSRBB assessment would determine overlap with the credit risk assessment and the credit risk Pillar 1 requirements.
It might be worth reminding that the Directive does not provide guidance on this topic, therefore leaving the EBA to define the scope of CSRBB sensitive instruments consistently with the definition of CSRBB itself. Nor any constraints to possible exclusion of items are provided in the Level 1 text, neither a mandate for a deviation from BCBS standards can be considered rooted in the Directive.
Notably, the draft GL formally allow banks to exclude items from the assessment, upon demonstration that they are not sensitive to CSRBB. Far from being a “pragmatic” approach (as claimed in the GL), this is in practice unfeasible, given that in a bank’s balance sheet there are thousands of products with different features, and for each one the draft GL would require a dedicated assessment as well as appropriate documentation.
Basic application of the principle of proportionality, as well as level playing field consideration, would imply that the EBA identifies classes of products excluded from the CSRBB assessment, rather that requiring thousands of EU banks to assess and document that the same product categories – such as retail mortgages – are not sensitive to CSRBB.
POSITION PAPER 2022
7/7
All in all, ABI recommends defining the perimeter of the CSRBB assessment scope as limited to fair valued and actively traded securities (in line with the Basel definition).
A materiality criterion should in any case be applied and exclusion of non-material portfolios should be allowed.
To this end, more clarity on the expected use of the outcome of the CSRBB assessment would be useful to understand whether the various provisions are really needed or not.
Still on proportionality, it has to be noted that, although the principle is mentioned, no indication can be found as regards its concrete application in the CRSBB framework, e.g. no graduation of the obligations, based on banks’ size and/or complexity, is provided.
To achieve actual proportionality, clear indication of supervisory expectations for banks of different size and complexity is essential, as well as certainty about the provisions that are not binding (in the different cases).
As regards IRRBB, instead, ABI would argue that, to keep consistency with banks’ management practices, the NII perspective should address NII as commonly intended, i.e. including interest income and expenses only. As highlighted in ABI response to the consultation on the RTS on SOTs, the inclusion of changes in fair value – sought in a more general “earnings” perspective - would, among other things, determine an overlap between the NII and the EVE metrics. Finally, it has to be noted that, depending on how the items are classified for accounting purposes, the fair value changes might affect either the earnings or the equity, which raises the question of whether the latter case can actually be consistently included in an “earnings” perspective.
Question 1: In the context of the measurement of the impact of IRRBB under internal systems, paragraph 111 envisages a five year cap repricing maturity for retail and non-financial wholesale deposits without a specified maturity. Would you foresee any unintended consequence or undesirable effect from this behavioural assumption in particular on certain business models or specific activities? If this is the case, please kindly provide concrete examples of it.
The introduction of a five-year cap to repricing maturity, for retail and non-financial wholesale deposits without a specified maturity, is not expected to have major impacts on Italian banks. Anyway, it has to be noted that such limit seems not justified under a prudential perspective, since setting a lower maturity is not in principle a conservative assumption. Indeed, the interest rate risk is symmetric, so that measuring it assuming a lower maturity of certain items, than the actual maturity (estimated with reliable internal models), is exactly as wrong as using a higher maturity, and equally leads to misrepresentation of risk (lower or higher, depending on the maturity of the remaining assets and liabilities). It is also highly questionable whether including a behavioural assumption in the Guidelines (GL) is compliant with the Level 1 regulation, given that such assumption would finally be applied also for the purpose of the SOTs, and the Directive explicitly excludes behavioural assumptions from the common modelling and parametric assumptions for the SOTs that the EBA is mandated to specify.With respect to the IRRBB measurement assumptions, ABI would recommend clarifying that paragraph 105 should not be intended as requiring inflation to be kept constant across the scenarios.
Question 2: Do respondents find that the criteria to identify non-satisfactory IRRBB internal models provide the minimum elements for supervisors’ assessment?
In ABI’s view, the criteria to identify non-satisfactory IRRBB internal models should be specified in more detail and reduced in scope.The proposed approach (Chapter 4.4. of the draft GL) seems far too broad, as in practice all instances where a bank deviates from any of the requirements of the Guidelines can fall in the scope of “non-satisfactory” internal models, to be disqualified in favor of standardised methodologies.
As a premise, it should be highlighted that the standardised methodologies represent a useful tool to ensure the proportionality of the framework and might be suitable to provide an acceptable estimate of risk, especially in case of banks of lower size and/or complexity.
However, they necessarily rely on simplistic assumptions which make them by definition less accurate than internal models (but not more conservative, for the reasons said in the response to Question 1). The more complex the bank, the more sophisticated models are needed to manage and measure interest rate risk, thus implying that, where internal models are available, they should normally be considered preferable. Mandatory switching to standardised methodologies should therefore be limited to instances where severe deficiencies are observed in internal models (and are not occasional but lasting over time).
In ABI’s view, the GL should be amended accordingly. The GL should specify that the internal models shall be deemed as “not satisfactory” within the meaning of Article 84(3) CRD only in case the observed deficiencies of the models materially affect the reliability of the results, thus it can be reasonably assumed that the standardised methodologies are actually better suited to measure the IRRBB of the bank than its internal models.
This should in principle exclude from qualification of the internal models as “not satisfactory”, those instances where a bank happens not to be fully compliant with the GL in regard of provisions pertaining to formal requirements and processes. In these cases, repair actions should be required, rather than switching to standardised methodologies.
With respect to Chapter 4.4 of the GL, ABI would also recommend better clarifying how the proportionality principle shall be applied. While such principle is mentioned in paragraph 118, it is not clear how it is expected to be applied in practice. Providing guidance in this respect would be very beneficial.
In addition, the GL should clarify that, in case a bank is required by its supervisor to switch to the standardised methodology, an appropriate timing shall be granted to implement the switch (at least one year). Also, the process for subsequent return to the internal models should be clarified.
Question 3: Is there any specific element in the definition of CSRBB that is not clear enough for the required assessment and monitoring of CSRBB by institutions?
The definition of CSRBB is in principle rather clear, although it might be difficult to operationalize.The key issue with regard to the proposed definition of CSRBB is that the proposed perimeter of application is not consistent with such definition (see the response to Question 4 below).
Moreover, certain aspects could be clarified, such as consideration of the changes in credit spread affecting bank’s own liabilities, the treatment of credit derivatives hedging credit risk on a single exposure or a portfolio of exposures, as well as interpretation of changes in the NII perspective (provided that increased market credit/liquidity spreads – hence market perception of increased risk - would more often result in higher NII).
More generally speaking, clarification on the intended use of the results of the CSRBB assessment would be very helpful for banks to identify the methodological approaches and assumptions which best fit for purpose.
Question 4: As to the suggested perimeter of items exposed to CSRBB, would you consider any specific conceptual or operational challenge to implement it?
The perimeter identified in the GL, i.e. in principle all banking book items, poses heavy concerns in terms of both conceptual soundness and operational feasibility.The CSRBB is defined as risk driven by changes of the market price for risks other than IRRBB and idiosyncratic credit risk. This implies that CSRBB can be consistently assessed only with respect to those items that are exposed to such movements in the market prices, i.e. are actively traded on financial markets.
It is inherently nonsensical – before being impossible in practice - to determine market spreads for instruments for which no active market exists. In other words, banks cannot see how the distinction between the credit spread related to idiosyncratic risk and the market spreads can be operationalized in the absence of a market where the latter can be retrieved.
In any case, it should be borne in mind that such changes would not affect a bank’s equity, unless the item is accounted for at fair value. This is also consistent with the business model under which an item is held by the bank: under the IFRS (accounting standards applied by the vast majority of EU banks) items accounted for at amortising cost are supposed not to be sold before maturity and therefore by definition not exposed to market risk – so they should consistently be excluded from the scope of the CSRBB framework.
Indeed, the limitation of the scope of the CSRBB to items accounted for at fair value is crystal clear in the BCBS standards. The definition of CSRBB illustrated in the Annex 1 of the BCBS standards “Interest rate risk in the Banking book” (Figure 1, reported below) clearly shows that that the Basel Committee intends the CSRBB as referred only to items accounted for at their fair value (mark-to market).
Any enlarged scope would ostensibly goldplate the BCBS standards and raise level playing field issues.
On the other hand, including items, which are not exposed to market spread movements, in the scope of the CSRBB assessment would determine overlap with the credit risk assessment and the credit risk Pillar 1 requirements.
It might be worth reminding that the Directive does not provide guidance on this topic, therefore leaving the EBA to define the scope of CSRBB sensitive instruments consistently with the definition of CSRBB itself. Nor any constraints to possible exclusion of items are provided in the Level 1 text, neither a mandate for a deviation from BCBS standards can be considered rooted in the Directive.
Notably, the draft GL formally allow banks to exclude items from the assessment, upon demonstration that they are not sensitive to CSRBB. Far from being a “pragmatic” approach (as claimed in the GL), this is in practice unfeasible, given that in a bank’s balance sheet there are thousands of products with different features, and for each one the draft GL would require a dedicated assessment as well as appropriate documentation.
Basic application of the principle of proportionality, as well as level playing field consideration, would imply that the EBA identifies classes of products excluded from the CSRBB assessment, rather that requiring thousands of EU banks to assess and document that the same product categories – such as retail mortgages – are not sensitive to CSRBB.
POSITION PAPER 2022
7/7
All in all, ABI recommends defining the perimeter of the CSRBB assessment scope as limited to fair valued and actively traded securities (in line with the Basel definition).
A materiality criterion should in any case be applied and exclusion of non-material portfolios should be allowed.
Both IRRBB and CSRBB
The distinction between the provisions applicable to the IRRBB and those referred to CSRBB is clear in the draft GL. Anyway, it has to be noted that the CSRBB framework outlined in the draft GL appears in many respects very similar to the IRRBB framework as to governance, processes and controls required. Given that the CRD (Article 84) requires banks a different level of engagement for CSRBB (“assess and monitor”) than for IRRBB (“identify, evaluate, manage and mitigate”), in ABI’s view the principle of proportionality would require that related obligations should be graduated accordingly.To this end, more clarity on the expected use of the outcome of the CSRBB assessment would be useful to understand whether the various provisions are really needed or not.
Still on proportionality, it has to be noted that, although the principle is mentioned, no indication can be found as regards its concrete application in the CRSBB framework, e.g. no graduation of the obligations, based on banks’ size and/or complexity, is provided.
To achieve actual proportionality, clear indication of supervisory expectations for banks of different size and complexity is essential, as well as certainty about the provisions that are not binding (in the different cases).
As regards IRRBB, instead, ABI would argue that, to keep consistency with banks’ management practices, the NII perspective should address NII as commonly intended, i.e. including interest income and expenses only. As highlighted in ABI response to the consultation on the RTS on SOTs, the inclusion of changes in fair value – sought in a more general “earnings” perspective - would, among other things, determine an overlap between the NII and the EVE metrics. Finally, it has to be noted that, depending on how the items are classified for accounting purposes, the fair value changes might affect either the earnings or the equity, which raises the question of whether the latter case can actually be consistently included in an “earnings” perspective.
Upload files
PP_ABI_IRRBB_def GL.pdf
(239.61 KB)