Response to consultation on draft Guidelines on IRRBB and CSRBB
Go back
Internal models are based on historical analysis and a deep understanding of the existing products, which reveal that NMD profiles and client’s behavioural differ from one country to another, doing that the average repricing maturity could differ significantly, and that one size does not fit all. Banks offer different types of deposit whose specific characteristics have to be considered in the estimation of parameters to properly measure the risk and manage them.
However, we understand that this requirement may represent a homogeneous measure for the benchmarking of regulatory metrics, especially in economic areas like the Eurozone, facing low interest rates and higher elasticities to the interest rates than other markets. Nevertheless, this restriction may lead to very different results when these banks are compared with banks that belong to economies with high interest rates and very low elasticity to the movements of the interest rates, such as some Latin-American countries. In these markets, historical analysis shows growing and very stable non-maturity deposits volumes. Therefore, compliance with this restriction would lead these banks to force their internal models and metrics and not reflect a realistic risk profile.
Moreover, setting a CAP to weighted average repricing of NMDs is not necessarily a prudent approach, and it may be equally imprudent to assign a too low as a too high repricing period for NMDs. Please note that a shorter duration on NMDs does not necessarily imply being less risky than a longer duration NMDs. On the contrary, this could widen the existing mismatch between the asset and liability sides, therefore impacting internal action management, forcing to shorten duration in the asset side or entering hedging transactions. This could even affect the prices of products.
As previously raised, we are of the view there should not be limitation to the development of behavioural models and this restriction should not affect internal metrics. Having said that, whether a restriction was to be set up for regulatory benchmarking only, then:
• A 5-year CAP is a commonly extended maturity reference for NMDs.
• We welcome the proposal to apply the CAP to the total amount and not to the core deposits, as it makes it more compatible with different internal models. We consider it is not convenient to limit individually specific parameters of the models.
• It should be applied to the duration instead of the average repricing maturity to better fit to different internal model approaches and ease the comparison of results among Europe (economies with low interest rates) and another jurisdiction as LATAM.
• The segmentation of NMD and the application of the caps should be focused on an adequate internal segmentation of the interest rate sensitivity and not refer to liquidity or funding aspects, like “transactional” and “non-transactional”.
In Art.98 (5a) of the CRDV the EBA is mandated to develop regulatory technical standards, specifying in letters b) and c) that these standards are subordinated to specific elements with common modelling and parametric assumptions, but excluding behavioural assumptions. Based on this, we understand that the cap of 5 years on NMD has been eliminated directly for the calculation of SOTs in the new RTS on SOTs.
But we would appreciate a clarification about why this restriction not applying in SOTs is now requested in paragraph 111 of the new guides for internal metrics. If any, setting a CAP would make sense for benchmarking purposes only and never for internal management purposes.
In many cases, implementing the standardised methodology may not lead to a more accurate risk assessment. Corrective measures to solve the identified deficiencies, or a partial application of selected items of the standardized method should be taken instead. Thus, in the case of minor vulnerabilities (not material) in the IMS, that do not invalid the risk measurement, we consider that it could be more suitable to establish temporary add-ons to internal metrics until deficiencies are corrected, instead of directly imposing the standard methodology.
In addition, it should be clearly stated that the standard methodology is not considered as a minimum requirement to evaluate an internal model as “satisfactory”.
The use of a run-off assumption for EVE should not be understood as a limitation, but a complementary approach to NII measures. The guidelines would benefit from additional clarifications regarding the assessment of some risks such as optionality. This risk has a significant sensitivity to valuation models and parameterizations that can lead to different results difficult to compare between banks and to assess the level of risk.
CSRBB scope of products: Considering the definition given in page 14 of the CP, it is understood that CSRBB targets on market perceptions and relates to market tradable products in the banking book. However, paragraph 124 refers to all banking book positions, referring also to not tradable products in markets. We would like to encourage the use of the definition given by Basel and adopted in EBA GL/2018, where this risk is only applicable for products accounted at fair value with active markets in place.
CSRBB Definition: In order to avoid double-counting with the credit risk, we support excluding from CSRBB the migration and default risk by eliminating the idiosyncratic risk. However, the extraction of the market credit risk factor may become problematic for those products without observable market prices.
CSRBB Methodology: According to the CSRBB definition, the market credit spread, and liquidity spread should be calibrated separately for each credit rating level. This implies that the resulting credit spread, and its volatility, rises as credit quality worsens. In consequence, different shocks should be applied to each credit rating category.
[Please find the graph regarding this question in the uploaded PDF document]
This approach, as it is defined, will be hardly feasible to implement for other instruments out of the market traded assets, as it requires to classify all the banking book transactions according to their level of creditworthiness in order to apply the corresponding market credit spread and volatility to each credit quality level. Such calculation would imply feeding ALM tools with detailed credit risk information at transaction level, regarding the customer credit rating (internal/external), collaterals, guarantees, etc.
Therefore, we strongly recommend limiting the scope of CSRBB to fair value assets with observable market prices. Whether the evaluation of CSRBB is extended to other items out of the fair value assets (e.g. to assess the impacts on NII), we suggest applying a simplified approach and setting a single (common) market credit shock by currency for all the balance sheet instruments (assets, liabilities, and off-balance).
Furthermore, we consider that the measurement of CSRBB impacts on NII overlaps with the measurement of Business Risk, which considers, among other factors, the margin compression. Therefore, it should be clearly stated that under no circumstance should it be allocated any capital charge derived from CSRBB from a NII perspective. Otherwise, it would produce a double counting.
We would recommend the scope to be limited to fair valued (FV) assets, and credit derivatives, tradable in-depth markets and under a business model of HTC&S. This approach would be consistent with Basel standards on IRRBB 2016.
We consider that the scope of instruments included in the CSRBB perimeter should be limited, by default, to those assets, and credit derivatives, accounted at fair value with potential impact on Equity, through P&L or OCI.
As explained in the Guidelines accompanying documents, institutions are expected to include in the perimeter “instruments sensitive, or expected to be sensitive, to volatility in credit spreads that may potentially impact the institutions income and / or capital”. Once migration and default risk is excluded from CSRBB, the effect of market credit spread variations for other banking book instruments, accounted at amortised cost and intended to be held to maturity, will have no impact on Banks’ Equity.
Consider the intention is to identify how risk is affected by changes in the price of tradable market products. In that case, the positions under an HTC model are out of scope since the risk of losses might materialize only in the case of sales, and it is a common understanding that products under this business model are not envisaging sales.
Therefore, we recommend that the potential extension of the perimeter to other items out of the fair value assets should be left to the criteria of the bank, according to the identification/existence of CSRBB in other products (e.g. specific loans portfolios intended to be traded). Thus, we suggest modifying the wording of the Guidelines accordingly, and limiting by default the scope of CSRBB to fair value assets, with the potential inclusion of other relevant products only in the case of existence of sensitivity to credit spread risk. Otherwise, banks will have to justify, one by one, the exclusion of a great number of balance sheet items.
From an NII perspective, in contrast to the EVE approach, the scope of instruments may be interpreted to apply to other Banking book items subject to CSRBB, due to the fact that a change in the margins of the new volumes will have an effect in NII projections (Income). However, as explained in the previous question, the intention of the NII impact assessment should be clarified, as it overlaps with business risk. Besides, we strongly recommend applying a simplified approach by setting a single market credit spread shock by currency to all the balance sheet items exposed to CSRBB.
We suggest monitoring and controlling the effect in fair value instruments in a separate (secondary) metric and not replacing the standard NII definition.
The inclusion of market value changes of fair value instruments in the banking book would change the commonly accepted NII perspective, making the risk exposure very dependent on the projection period and the accountancy criteria, which may lead to a misinterpretation of the results. Furthermore, it introduces a much higher complexity in the calculation, and overlaps with the EVE approach.
Positions at fair value are generally sovereign bonds or derivatives used in the banking book as stable hedging instruments that do not affect NII following the accounting rules. Additionally, these portfolios are already incorporated in EVE metrics, and including them in NII would lead to a double accounting. Besides, It could lead to undesirable incentives for maintaining positions based on accounting criteria, such as HTC instead of economic motivations.
Lastly, operationally, the calculation of the broader NII metric is highly complex. There is an intersection between the NII and the FV changes during the projection period (e.g., 12M). The calculation of FV changes requires excluding the instruments maturing during the first year and estimating the future MtM impact at the end of the projection period on the fair value instruments with longer maturities (>1Y) which will be computationally very demanding. Likewise, the inclusion of fees and commissions would be very burdensome.
Inflation treatment: We agree on the general proposal to use prudent assumptions concerning inflation behaviour. However, we consider that restricting the inflation modeling by setting a scenario-independent premise would be a fatal flaw. As a component of nominal interest rates, the evolution of inflation is highly correlated to the interest rates levels, especially in those jurisdictions with high market interest rates. Considering constant inflation in combination with high-interest rate shocks is unrealistic and not necessarily prudent. By introducing this assumption, nominal rates volatility will be fully applied (passed-through) to real rates, which are generally less volatile. We strongly recommend eliminating from the text the reference to scenario-independent assumptions, and do not prescribe the inflation treatment which should be done according to the banks’ internal models.
Additionally in case CSRBB refers to all products, we would appreciate whether the EBA could clarify if other restriction such as the cap on the average repricing maturity of NMD could apply also to CSRBB or only to IRRBB or how this restriction could impact in the reinvestment assumptions for CSRBB.
[Please find some additional comments regarding this question in the uploaded PDF document]
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.
We do not support setting limitations to behavioural modelling, furthermore when they are not supported by empirical observations, and even these restrictions could disagree with local regulations in jurisdictions out of Europe. We are more in favour of establishing guidance about how NMDs must be modelled, according to customers’ data.Internal models are based on historical analysis and a deep understanding of the existing products, which reveal that NMD profiles and client’s behavioural differ from one country to another, doing that the average repricing maturity could differ significantly, and that one size does not fit all. Banks offer different types of deposit whose specific characteristics have to be considered in the estimation of parameters to properly measure the risk and manage them.
However, we understand that this requirement may represent a homogeneous measure for the benchmarking of regulatory metrics, especially in economic areas like the Eurozone, facing low interest rates and higher elasticities to the interest rates than other markets. Nevertheless, this restriction may lead to very different results when these banks are compared with banks that belong to economies with high interest rates and very low elasticity to the movements of the interest rates, such as some Latin-American countries. In these markets, historical analysis shows growing and very stable non-maturity deposits volumes. Therefore, compliance with this restriction would lead these banks to force their internal models and metrics and not reflect a realistic risk profile.
Moreover, setting a CAP to weighted average repricing of NMDs is not necessarily a prudent approach, and it may be equally imprudent to assign a too low as a too high repricing period for NMDs. Please note that a shorter duration on NMDs does not necessarily imply being less risky than a longer duration NMDs. On the contrary, this could widen the existing mismatch between the asset and liability sides, therefore impacting internal action management, forcing to shorten duration in the asset side or entering hedging transactions. This could even affect the prices of products.
As previously raised, we are of the view there should not be limitation to the development of behavioural models and this restriction should not affect internal metrics. Having said that, whether a restriction was to be set up for regulatory benchmarking only, then:
• A 5-year CAP is a commonly extended maturity reference for NMDs.
• We welcome the proposal to apply the CAP to the total amount and not to the core deposits, as it makes it more compatible with different internal models. We consider it is not convenient to limit individually specific parameters of the models.
• It should be applied to the duration instead of the average repricing maturity to better fit to different internal model approaches and ease the comparison of results among Europe (economies with low interest rates) and another jurisdiction as LATAM.
• The segmentation of NMD and the application of the caps should be focused on an adequate internal segmentation of the interest rate sensitivity and not refer to liquidity or funding aspects, like “transactional” and “non-transactional”.
In Art.98 (5a) of the CRDV the EBA is mandated to develop regulatory technical standards, specifying in letters b) and c) that these standards are subordinated to specific elements with common modelling and parametric assumptions, but excluding behavioural assumptions. Based on this, we understand that the cap of 5 years on NMD has been eliminated directly for the calculation of SOTs in the new RTS on SOTs.
But we would appreciate a clarification about why this restriction not applying in SOTs is now requested in paragraph 111 of the new guides for internal metrics. If any, setting a CAP would make sense for benchmarking purposes only and never for internal management purposes.
Question 2: Do respondents find that the criteria to identify non-satisfactory IRRBB internal models provide the minimum elements for supervisors’ assessment?
The high-level principle of being compliant with the Guidelines is simply too broad and may be disproportionate if applied literally, without considering the materiality of the identified gaps. It should be clearly stated that the materiality/relevance of the potential weaknesses of the IMS will be considered in order to require the application of the standardised methodology.In many cases, implementing the standardised methodology may not lead to a more accurate risk assessment. Corrective measures to solve the identified deficiencies, or a partial application of selected items of the standardized method should be taken instead. Thus, in the case of minor vulnerabilities (not material) in the IMS, that do not invalid the risk measurement, we consider that it could be more suitable to establish temporary add-ons to internal metrics until deficiencies are corrected, instead of directly imposing the standard methodology.
In addition, it should be clearly stated that the standard methodology is not considered as a minimum requirement to evaluate an internal model as “satisfactory”.
The use of a run-off assumption for EVE should not be understood as a limitation, but a complementary approach to NII measures. The guidelines would benefit from additional clarifications regarding the assessment of some risks such as optionality. This risk has a significant sensitivity to valuation models and parameterizations that can lead to different results difficult to compare between banks and to assess the level of risk.
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?
Yes, we would welcome more clarity about some key issues: CSRBB scope of products: Considering the definition given in page 14 of the CP, it is understood that CSRBB targets on market perceptions and relates to market tradable products in the banking book. However, paragraph 124 refers to all banking book positions, referring also to not tradable products in markets. We would like to encourage the use of the definition given by Basel and adopted in EBA GL/2018, where this risk is only applicable for products accounted at fair value with active markets in place.
CSRBB Definition: In order to avoid double-counting with the credit risk, we support excluding from CSRBB the migration and default risk by eliminating the idiosyncratic risk. However, the extraction of the market credit risk factor may become problematic for those products without observable market prices.
CSRBB Methodology: According to the CSRBB definition, the market credit spread, and liquidity spread should be calibrated separately for each credit rating level. This implies that the resulting credit spread, and its volatility, rises as credit quality worsens. In consequence, different shocks should be applied to each credit rating category.
[Please find the graph regarding this question in the uploaded PDF document]
This approach, as it is defined, will be hardly feasible to implement for other instruments out of the market traded assets, as it requires to classify all the banking book transactions according to their level of creditworthiness in order to apply the corresponding market credit spread and volatility to each credit quality level. Such calculation would imply feeding ALM tools with detailed credit risk information at transaction level, regarding the customer credit rating (internal/external), collaterals, guarantees, etc.
Therefore, we strongly recommend limiting the scope of CSRBB to fair value assets with observable market prices. Whether the evaluation of CSRBB is extended to other items out of the fair value assets (e.g. to assess the impacts on NII), we suggest applying a simplified approach and setting a single (common) market credit shock by currency for all the balance sheet instruments (assets, liabilities, and off-balance).
Furthermore, we consider that the measurement of CSRBB impacts on NII overlaps with the measurement of Business Risk, which considers, among other factors, the margin compression. Therefore, it should be clearly stated that under no circumstance should it be allocated any capital charge derived from CSRBB from a NII perspective. Otherwise, it would produce a double counting.
Question 4: As to the suggested perimeter of items exposed to CSRBB, would you consider any specific conceptual or operational challenge to implement it?
Regarding scope, we believe it should be conditional to the CSRBB definition and only positions with tradable markets should be considered.We would recommend the scope to be limited to fair valued (FV) assets, and credit derivatives, tradable in-depth markets and under a business model of HTC&S. This approach would be consistent with Basel standards on IRRBB 2016.
We consider that the scope of instruments included in the CSRBB perimeter should be limited, by default, to those assets, and credit derivatives, accounted at fair value with potential impact on Equity, through P&L or OCI.
As explained in the Guidelines accompanying documents, institutions are expected to include in the perimeter “instruments sensitive, or expected to be sensitive, to volatility in credit spreads that may potentially impact the institutions income and / or capital”. Once migration and default risk is excluded from CSRBB, the effect of market credit spread variations for other banking book instruments, accounted at amortised cost and intended to be held to maturity, will have no impact on Banks’ Equity.
Consider the intention is to identify how risk is affected by changes in the price of tradable market products. In that case, the positions under an HTC model are out of scope since the risk of losses might materialize only in the case of sales, and it is a common understanding that products under this business model are not envisaging sales.
Therefore, we recommend that the potential extension of the perimeter to other items out of the fair value assets should be left to the criteria of the bank, according to the identification/existence of CSRBB in other products (e.g. specific loans portfolios intended to be traded). Thus, we suggest modifying the wording of the Guidelines accordingly, and limiting by default the scope of CSRBB to fair value assets, with the potential inclusion of other relevant products only in the case of existence of sensitivity to credit spread risk. Otherwise, banks will have to justify, one by one, the exclusion of a great number of balance sheet items.
From an NII perspective, in contrast to the EVE approach, the scope of instruments may be interpreted to apply to other Banking book items subject to CSRBB, due to the fact that a change in the margins of the new volumes will have an effect in NII projections (Income). However, as explained in the previous question, the intention of the NII impact assessment should be clarified, as it overlaps with business risk. Besides, we strongly recommend applying a simplified approach by setting a single market credit spread shock by currency to all the balance sheet items exposed to CSRBB.
IRRBB only
NII Definition: We recommend maintaining the NII definition aligned to the common (narrow) definition of net interest income and not replacing it by an earnings metric including market value changes of fair value instruments. The narrow perspective of NII is more aligned with the internal management and control of IRRBB.We suggest monitoring and controlling the effect in fair value instruments in a separate (secondary) metric and not replacing the standard NII definition.
The inclusion of market value changes of fair value instruments in the banking book would change the commonly accepted NII perspective, making the risk exposure very dependent on the projection period and the accountancy criteria, which may lead to a misinterpretation of the results. Furthermore, it introduces a much higher complexity in the calculation, and overlaps with the EVE approach.
Positions at fair value are generally sovereign bonds or derivatives used in the banking book as stable hedging instruments that do not affect NII following the accounting rules. Additionally, these portfolios are already incorporated in EVE metrics, and including them in NII would lead to a double accounting. Besides, It could lead to undesirable incentives for maintaining positions based on accounting criteria, such as HTC instead of economic motivations.
Lastly, operationally, the calculation of the broader NII metric is highly complex. There is an intersection between the NII and the FV changes during the projection period (e.g., 12M). The calculation of FV changes requires excluding the instruments maturing during the first year and estimating the future MtM impact at the end of the projection period on the fair value instruments with longer maturities (>1Y) which will be computationally very demanding. Likewise, the inclusion of fees and commissions would be very burdensome.
Inflation treatment: We agree on the general proposal to use prudent assumptions concerning inflation behaviour. However, we consider that restricting the inflation modeling by setting a scenario-independent premise would be a fatal flaw. As a component of nominal interest rates, the evolution of inflation is highly correlated to the interest rates levels, especially in those jurisdictions with high market interest rates. Considering constant inflation in combination with high-interest rate shocks is unrealistic and not necessarily prudent. By introducing this assumption, nominal rates volatility will be fully applied (passed-through) to real rates, which are generally less volatile. We strongly recommend eliminating from the text the reference to scenario-independent assumptions, and do not prescribe the inflation treatment which should be done according to the banks’ internal models.
CSRBB only
Considering the CSRBB definition and scope is subject to credit risk in markets and marketable positions, the separation of IRRBB and CSRRBB is clear. However, if it refers to all products in the banking book, whether the products are traded or not in financial markets, we believe it is not so clear enough given that doble accounting may exist for products already under credit internal models.Additionally in case CSRBB refers to all products, we would appreciate whether the EBA could clarify if other restriction such as the cap on the average repricing maturity of NMD could apply also to CSRBB or only to IRRBB or how this restriction could impact in the reinvestment assumptions for CSRBB.
Both IRRBB and CSRBB
Entry into force: EBA proposal represents a significant change in how risk must be assessed and monitored, especially regarding the wider NII and CSRBB definitions. We consider that entities should be given enough time, at least from 6 months to 1 year, to undertake preparatory work to implement the new approaches and definitions that require the adaptation of information systems, models, and controls.[Please find some additional comments regarding this question in the uploaded PDF document]