Response to consultation on draft ITS amending Commission Implementing Regulation (EU) 2024/3117 regarding Credit Risk and IFRS 9 Benchmarking reporting
General 1: Are the instructions and templates clear to the respondents? If not, what should be clarified?
Overarching comments:
Timing: For the December 2026 benchmarking exercise, among the available alternatives, we propose that the December 2026 benchmark data — to be delivered in April 2027 — should continue to follow the current reporting format. Given the implementation timeline, institutions will not have sufficient time to adapt their systems to accommodate the extensive changes proposed, including the additional cost allocation requirements.
In the case of IFRS 9, the perimeter expansion represents a substantial increase in scope, which will also require significant time to implement and adjust.
Our rationale is that, if FINREP/COREP changes will not become mandatory until September 2027, it is difficult to justify why such a substantial and structural change for benchmarking is being requested already for the year-end closing. This would also imply a retroactive application to December 2025.
Considering the changes in the EBA Stress Test, it would be necessary to clarify the treatment of the obligor and guarantor perspectives: under IFRS 9 is required on a pre-mitigation basis, and under credit risk on a post-mitigation basis.
General 1: Are the instructions and templates clear to the respondents? If not, what should be clarified?
Credit Risk:
We welcome the move toward a simplification in the granularity of the portfolios and the use of the COREP asset class definitions, as this reduces the risk of misalignment.
The main difficulty we foresee concerns the calculation of the five year default and loss rates (items 0214 and 0301 in template C102.02) at the required level of detail, particularly the new asset classes introduced by CRR III.
We propose using the existing two year historical series under CRR III as a baseline and adding one additional year each subsequent year, so that a full five year history will be available after three years.
For column 0020, Exposure Class, of template 101.01, the following is stated: “Each counterparty shall be assigned to one of the following Exposure Classes.” However, when the different categories that the variable may take are detailed, in the section on the advanced method, under retail, it appears that a counterparty may be assigned to more than one category. For example, a mortgage loan to finance the acquisition of a first home by an individual should be assigned to both value A40 and value A44.
Clarification is needed on whether each counterparty is required to be assigned to one, and only one, category, or whether assignment to multiple categories is permitted.
Additional clarification is needed on whether column 0051 of template 101.01 corresponds to the same rating scale used in column 0005 of COREP template C08.02With regard to template C 105.00, it would be appreciated further clarification on whether the requested metrics are expected to be reported on a pre-credit risk mitigation basis or on a post-credit risk mitigation basis. Such clarification would help ensure a consistent interpretation and application of the reporting requirements across institutions..
Clarification would be appreciated regarding whether the PDs to be reported in the template 102.1 should be considered on a pre-credit risk mitigation basis or on a post-credit risk mitigation basis, in order to ensure a consistent interpretation and application across institutions.
In the case of the template 104, clarification would be needed regarding two aspects:
- The treatment of changes in the definition of default over the observation window. In particular, where institutions have transitioned from a previous default definition to the current CRR/EBA definition of default, it is unclear whether historical observations should be reported according to the definition applicable at each point in time, or whether institutions are expected to reconstruct historical observations in line with the assumptions and methodology underlying the current approved model. This clarification would be essential to ensure consistency, comparability and operational feasibility across institutions.
- Whether the LGD required should be in accordance with the EBA Guidelines on PD and LGD estimation or following an approach similar to that required for columns 300 and 301 of template C 102.01. This last approach could imply significant implementation efforts, as such methodology differs from the approaches typically used to calculate realised LGD within institutions’ current LGD modelling, monitoring and validation frameworks.
IFRS9:
An additional clarification is needed regarding whether the exposure amounts reported for IFRS 9 in C 101.02 and C 102.02 should be determined before or after the application of credit risk mitigation. The ambiguity arises because the Instructions, Part I, I.2 IFRS 9 Information, paragraph 14 refer to actual exposure existing at the reference date in the form of either Original Exposure or Exposure after credit risk mitigation, whereas the specific IFRS 9 exposure columns define the reported amount as the exposure considered for the application of IFRS 9 impairment requirements: C 101.02, column 0200 “Exposure value - IFRS 9”, and C 102.02, columns 0400, 0401 and 0407 for Stage 1, Stage 2 and Stage 3 exposure values. In addition, paragraph 21 states that, for guaranteed exposures, the PD should remain that of the original obligor and the effect of the guarantee should be reflected in LGD and ECL. We therefore ask the EBA to clarify explicitly whether IFRS 9 exposure values should be reported on an accounting IFRS 9 impairment basis, including off-balance sheet exposures after accounting conversion factors, rather than on a prudential after-CRM substitution basis.
The instructions should also include a concise mapping from the current IFRS 9 templates C.111-C.118 to C 101.02/C 102.02, identifying retained, deleted and transformed data points.
The EBA should provide a dedicated IFRS 9 portfolio matrix for C 102.02, identifying the applicable exposure classes, active dimensions and column blocks for each reporting population. This should clearly state which dimensions apply to IFRS 9 reporting, such as geography, collateral type, sector and total performing portfolios, and which dimensions from C 102.01 should not be extrapolated to IFRS 9, such as IRB model split or full rating segmentation.
General 2: Do the respondents identify any discrepancies between these templates and instructions and the calculation of the requirements set out in the underlying regulation?
No, we do not.
General 3: Do the respondents agree that the amended ITS on Supervisory Reporting allows to fulfil the requirements and purpose of the Article 78 of the CRD?
Yes, we agree.
General 4: what is the information (please refer to single columns) that you see as particularly costly to be provided? Please provide detailed information about the implementation and running costs of the proposed changes, also compared to the current total costs.
Credit Risk:
The highest implementation and operational costs are expected to be associated with columns 0203 and 0204 of template C 102.01.
In order to report these columns, institutions would generally need to adopt one of the following approaches:
- Populate the relevant databases with risk parameters excluding MoCs, supervisory measures, add-ons, etc., and subsequently enhance the calculation engines to generate multiple RWA outputs reflecting the different adjustments (e.g. RWAs excluding MoCs; excluding both MoCs and supervisory measures; excluding MoCs, supervisory measures and add-ons, etc.).
- Alternatively, institutions could rely on the databases currently used for simulations and stress-testing purposes, incorporating there the parameters excluding MoCs and supervisory measures, and performing dedicated ad-hoc engine runs on an annual basis specifically for the benchmarking exercise.
Under the first approach, the main costs would stem from the technological developments required to store, manage, and process the additional parameter information within production systems and reporting infrastructures. This would likely involve significant implementation efforts across data architecture, calculation engines, and reporting processes.
Under the second approach, the costs would be more operational in nature, as institutions would need to perform additional annual calculations and ad-hoc executions that are not currently part of the regular reporting framework. This would increase operational complexity and require additional resources for execution, validation, reconciliation, and governance activities.
Compared with the current reporting framework, both alternatives imply a non-negligible increase in costs and operational burden, particularly given that the requested information is not typically produced through existing standard reporting processes.
In addition, regarding the template 104, institutions may be required to reconstruct the new CRR3 exposure classes retrospectively over a 15-year historical horizon. This may require complex historical mappings, reclassification exercises and reconciliation processes, or the use of assumptions to reconstruct the new exposure classes for historical periods.
Furthermore, institutions may be expected to apply the current approved LGD model framework to defaults observed over the last 15 years. This could imply significant implementation efforts, particularly considering the potential changes in portfolio segmentation over time.
Finally, regarding template C 102.01 and, in particular, metrics 300 and 301, the current requirement to calculate loss rates based on credit risk adjustments and write-offs — including adjustments recognised prior to the default event — differs from the methodologies typically used to calculate realised LGD within institutions’ current LGD modelling, monitoring and validation frameworks. Adapting existing calculation frameworks to comply with the requested methodology may therefore require the development of additional calculation, aggregation and data reconstruction processes.
On this regard, institutions would welcome the possibility of allowing these metrics to be calculated following a realised LGD approach for the exposures defaulting during the relevant observation period. With regard to the one-year loss rate metric, it should be noted that, as the metric is intended to capture exposures defaulting during the last year, recovery processes for a significant proportion of those exposures may still be ongoing at the reporting reference date. In this context, institutions would welcome the consideration of alternative approaches for the calculation of the one-year loss rate (which could also be extended to the five-year metric), such as:
- calculating the metric based on recovery processes closed during the relevant observation period; or
- assigning to newly defaulted exposures the LGD estimates derived from the current approved model framework.
Similar approaches could also be considered for the calculation of the five-year metric where the proportion of recovery processes still open remains significant.
General 5: what is the reporting level for banking groups that you deem more appropriate?
Credit Risk:
Given that the aim of the benchmarking exercise is to compare risk parameters and RWAs, as well as to assess their variability across institutions, we consider consolidated level reporting to be the most appropriate approach, since it captures the parameters arising from the different models and geographies within the banking group.
Furthermore, the proposed templates already introduce more granular breakdowns, including obligor geography and model information in Template C 102.01, already provides a holistic view of the group’s risk profile. Therefore, we do not consider it necessary to require reporting at the level of individual entities, regardless of their size, when they belong to a group whose ultimate parent entity is located in a Member State.
The migration to the new benchmarking framework should not automatically replicate the current perimeter of reporting entities. Subsidiaries already covered by the consolidated submission should not be required to submit stand-alone benchmarking reports unless a specific supervisory need justifies otherwise. Separate reporting should only be required where it provides clear incremental supervisory value beyond the consolidated submission; otherwise, it would duplicate data collection, controls, DPM/XBRL implementation, and would undermine the EBA’s simplification and burden-reduction objectives.
IFRS9:
Sub-consolidated or individual reporting should be exceptional and limited to cases where it is necessary for supervisory comparability, for example where a material subsidiary has distinct IFRS 9 models, materially different staging or overlay practices, or a significant SA portfolio not sufficiently visible at group level. A broad extension to subsidiary-level reporting would materially increase implementation and reconciliation costs and would weaken the simplification objective.
CR 1: Do you foresee any issues with supplementing the sample of counterparties listed in C 100.00 by the Large Exposures counterparties reported in the Template C 27.00 at the same reference date in order to expand the sample of low default counterparties for the reporting for template C 101.01/.02?
We believe that supplementing the sample of counterparties in template C 100.00 with Large Exposures counterparties reported in template C 27.00 could undermine the original objective of the benchmarking exercise, namely ensuring an appropriate level of comparability across counterparties.
In particular, counterparties derived from template C 27.00 may be concentrated in a limited number of institutions, whereas the likelihood of such concentration is lower for the names included in the EBA reference list. As a result, the additional counterparties may not be sufficiently representative across the participating banks.
Therefore, rather than enhancing the benchmarking exercise, this additional data source could introduce noise into the analysis, as those counterparties may exhibit higher variability in the reported parameters due to being associated with exposures held by only a small subset of institutions.
In addition, this approach would appear to go against the simplification objective of the benchmarking exercise. If information from template C 27.00 were to be used, institutions would still need to maintain the EBA reference list C100.00, which is updated annually, while also developing a consistent process within the EBA benchmarking calculation engine to combine both sources of information. This would increase implementation and operational complexity, rather than simplifying the reporting process.
If the use of C 27.00 information is nevertheless retained, we believe that it should be subject to minimum safeguards. These should include a minimum number of reporting institutions per counterparty, clear materiality criteria, appropriate outlier treatment, consistent identification and mapping rules, and the exclusion of counterparties reported by only one or two institutions.
CR 2: Do you identify any errors in the list of counterparties in C 100.00? If so, please report it to us.
We would like to raise a potential classification inconsistency under Article 107(3) of the CRR. Certain third-country credit institutions or financial institutions located in third-country jurisdictions that do not have an equivalent prudential and supervisory framework are classified as Large Corporate rather than Institution. In this respect, the EBA list appears to classify these counterparties only as Institutions, without taking into account the CRR equivalence criterion. This could create inconsistencies between the regulatory classification applied by institutions and the counterparty classification used in the low default portfolio benchmarking exercise.
CR 3: Do you foresee any issues with the transfer of the portfolio/model split from the former template C 105.02 to the template C 101.01 and the new template C 102.01? Do you agree with the EBA’s assessment that the portfolio/model split in the new template C 102.01 is the preferred option compared to designing a separate template increasing the reporting details of the former templates C 105.01 and C 105.02?
We agree with the integration of the information currently reported in template C 105.02 into the new template resulting from the merger of the former templates C 102.00 and C 103.00.
As acknowledged in the cost-benefit assessment, this approach effectively increases the number of data points in the new template, since the model dimension becomes part of the new primary key. However, from a longer-term perspective, we see clear benefits in this integration and in the removal of template C 105.02.
In particular, template C 105.02 previously provided a breakdown of EAD and RWAs at a level of granularity that will now consistently apply to the full set of information contained in the new template. In our view, this results in a more coherent and streamlined reporting framework, while also avoiding the maintenance of a separate template dedicated exclusively to the portfolio/model split.
CR 4: Do you think it is possible to provide the details of the impacts of the RWEA add-ons and MoCs (columns 0203, 0204, 0205 of Template C 102.01) at the required segmentation level?
Although it would be technically possible to incorporate such information, this would require significant IT developments for institutions. These developments would not be limited to a simple data extraction exercise, but would involve changes to the data model, parameter storage, calculation engines, reporting processes, controls and reconciliation framework.
Regarding add-ons applied directly to risk weights rather than to the underlying parameters, the relevant information could be already available at granular level for institutions. Therefore, it could be possible to calculate their impact at any aggregation level required.
CR 5: Do you agree with the proposed simplification in the reporting for column 0202 (RWEA Standardised) of the template C 102.01? Do you deem the segmentation level for which the information is requested sufficient to highlight the key differences between the IRB approach and the Standardised Approach?
The cost of reporting a given column does not materially decrease simply because the number of rows for which that column has to be completed is reduced. Since institutions generally maintain the information at granular level, once a column is required, the additional effort needed to populate it for all rows rather than for only a subset of rows is not substantially higher.
Therefore, when considering simplifications, the regulator should place greater emphasis on the selection of the columns to be included, rather than on limiting the reporting population through a reduced subset of rows.
That being said, instead of — or in addition to — introducing a breakdown by geography for all exposures, it could be more meaningful to introduce rating as an additional segmentation axis. Rating is likely to be a key driver of the differences observed between capital requirements under internal models and under the Standardised Approach.
CR 6: The split by collateral type (column 0061 of new template C 102.01) differs from the split used in the current Benchmarking portfolios. Do you foresee any issues or disadvantages with the new definition and instructions as compared to the ones currently in use?
In principle, we consider the proposed approach to be appropriate. However, we would appreciate clarification as to whether the references made to the columns for Credit Derivatives are correct. The instructions refer to columns 0040 and 0150 of COREP template C 08.02, whereas the columns containing information on credit derivatives are actually columns 0050 and 0160.
In addition, we believe that further guidance would be beneficial regarding the type of exposures that should be reported under category 60 “Exposures with partial collateralisation”, in order to ensure a consistent interpretation and application across institutions.
CR 7: Do you see material obstacles in providing the information required by the column 0302 of the template C 102.01 (Long-run average realized loss rate of assigned facility grades)?
We do not foresee material conceptual difficulties in providing the information required under column 302, as the long-run average LGD is already calculated within the current LGD modelling framework.
Nevertheless, some implementation challenges may arise from a data architecture perspective.
CR 8: Do you see material obstacles in providing the information required by the new loss rate template (C 104.00)? Would alternative measures of loss rates be more appropriate?
The implementation of the C 104.00 – Loss rates template may entail a significant operational and implementation burden for institutions. In particular, the requirement to report realized LGD figures on an annual basis for defaults originated over the last 15 years implies the need to perform yearly estimations of realised LGD.
While the underlying historical information is generally expected to be available within institutions’ systems, the main challenge lies in extracting, reconstructing and aggregating the data in the specific format and level of granularity required by the template. This requirement could therefore force institutions to redesign data aggregation and LGD estimation processes, as multiple realised LGD estimations may need to be performed for different reference dates and recovery process outcomes.
In this respect, it is assumed that the LGD required is the one calculated in accordance with the EBA Guidelines on PD and LGD estimation. However, if the metric is expected to be calculated following an approach similar to that required for columns 300 and 301 of template C 102.01, this could imply significant implementation efforts, as such methodology differs from the approaches typically used to calculate realised LGD within institutions’ current LGD modelling, monitoring and validation frameworks. Further clarification on this point would therefore be appreciated.
In addition, the implementation of this template raises a number of important methodological and operational challenges. First, institutions may be required to reconstruct the new CRR3 exposure classes retrospectively over a 15-year historical horizon, despite the fact that historical data repositories and legacy reporting frameworks were not originally structured according to the new CRR3 portfolio segmentation. This may require complex historical mappings, reclassification exercises and reconciliation processes.
Second, the template appears to implicitly assume that the current portfolio segmentation and approved IRB model framework can be consistently replicated over the full 15-year observation period. However, institutions have typically undergone multiple changes in portfolio segmentation, model architecture, calibration approaches and supervisory-driven redevelopments over time. In practice, this may require institutions to retrospectively reconstruct historical portfolios according to the current segmentation criteria and to assign the corresponding LGD estimates under the current approved model framework to historical default observations. Performing this exercise consistently over a long historical horizon may involve significant implementation complexity.
Finally, clarification would be needed regarding the treatment of changes in the definition of default over the observation window. In particular, where institutions have transitioned from a previous default definition to the current CRR/EBA definition of default, it is unclear whether historical observations should be reported according to the definition applicable at each point in time, or whether institutions are expected to reconstruct historical observations in line with the assumptions and methodology underlying the current approved model. This clarification would be essential to ensure consistency, comparability and operational feasibility across institutions.
CR 9.1: The split by facility type (column 0062 of template C 102.01) is designed to collect information on the functioning of the CCF models. The classification into buckets (facility types) of off-balance-sheet items is based on Annex I of Regulation (EU) No 575/2013. Do you think there exist other classifications (for example product types) of off-balance-sheet items that would be more useful for representing the CCF models?
We believe that the current axis based on “type of facility” functions relatively well and provides a greater level of granularity than the new proposal.
In addition, for IRB exposures, the classification based on Annex I buckets appears to work adequately for FIRB exposures, where the applicable CCFs are aligned with those under the Standardised Approach. However, for exposures for which institutions use internally estimated CCFs, the current classification may in some cases be more suitable and better aligned with the actual functioning and differentiation of internal models than the proposed approach.
Therefore, while the proposed classification provides a reasonable and harmonised framework, maintaining sufficient granularity and flexibility for internally modelled exposures would be important to ensure the reporting remains representative of institutions’ risk management practices.
CR 9.2: Template C 101.01 and C 102.01 propose two alternatives to collect the information on the CCF, either by one column (column 0110 in C 101.01 and column 0120 in C 102.01) or by two separate columns (columns 0111/0112 in C 101.01 and columns 0121/0122 in C 102.01). Which of these reporting alternatives do you prefer?
We do not identify a clearly superior alternative between the two proposed approaches. From a simplification and operational burden reduction perspective, the use of a single column appears more appropriate, as it would likely reduce the overall number of reported data points.
On the other hand, from a reporting granularity and information richness perspective, maintaining two separate columns could be beneficial, particularly as this would be consistent with the approach currently followed in COREP after the introduction of differentiated columns for Standardised Approach (STD) CCFs.
Therefore, we believe that both alternatives have merits, and the preferred option ultimately depends on the balance sought between reporting simplification and the level of analytical detail required.
CR 10: Do you support the use and the proposed granularity of the common master scale to limit the number of reported rating grades? What alternative would you propose.
We do not identify a clearly superior alternative between the two proposed approaches. From a simplification and operational burden reduction perspective, the use of a single column appears more appropriate, as it would likely reduce the overall number of reported data points.
On the other hand, from a reporting granularity and information richness perspective, maintaining two separate columns could be beneficial, particularly as this would be consistent with the approach currently followed in COREP after the introduction of differentiated columns for Standardised Approach (STD) CCFs.
Therefore, we believe that both alternatives have merits, and the preferred option ultimately depends on the balance sought between reporting simplification and the level of analytical detail required.
IFRS9 1: Do you agree with the proposed definition of SA institutions as relevant for the benchmarking reporting?
Yes, considering a proportionate materiality criterion for all institutions, SA and IRB institutions. We agree with the objective of including relevant SA institutions in IFRS 9 benchmarking where they are subject to IFRS9 and classified as Large Institutions. This avoids limiting the exercise to IRB institutions only, while also preventing an excessive extension of the reporting requirement to all SA institutions applying IFRS 9. However, for banking groups that already report benchmarking information because they include IRB institutions, the inclusion of SA portfolios should be subject to a materiality assessment.
This is particularly relevant because adding SA portfolios or SA entities with limited contribution to the group’s IFRS9 exposure or ECL may increase the reporting population, portfolio mapping effort and reconciliation burden without materially improving the supervisory benchmarking outcome.
We therefore recommend that the final ITS clarify the application of the SA perimeter within banking groups that already report benchmarking information as IRB groups. The materiality threshold could be based on IFRS9 exposure, ECL amount, or contribution to the group’s total IFRS9 impairment perimeter.
The reporting framework should focus on information that is material enough to improve comparability and supervisory analysis, rather than requiring immaterial SA data whose cost of production may outweigh its analytical benefit.
IFRS9 2: Do you agree with the proposal to refer to SA exposure classes in defining the portfolios for SA institutions?
Yes, provided that the level of SA exposure class granularity is clarified. The final ITS should specify the exact level at which SA asset classes must be reported, including whether the requirement refers only to broad SA exposure classes or to more granular subcategories.
This clarification is particularly important for real estate exposures. For example, it is not sufficiently clear whether mortgage exposures should be reported only at the level of residential versus commercial immovable property. This level of detail materially affects portfolio generation, mapping from accounting IFRS 9 data, reconciliation with prudential reporting and implementation cost.
We also understand that this question is not relevant only for standalone SA institutions. The instructions indicate that IFRS 9 information may also be required from IRB institutions for exposures under the SA approach. Therefore, the final ITS should clarify the SA exposure class structure both for relevant SA institutions and for SA portfolios within IRB banking groups. This would ensure consistent reporting and avoid different interpretations of the SA IFRS 9 perimeter across institutions.