Response to consultation on Draft Guidelines on the methodology to estimate and apply credit conversion factors under the Capital Requirements Regulation
1. Question 1: How material are the cases for your institution where you would have to assign an SA-CCF to exposures arising from undrawn revolving commitments and thus restrict the use of own estimates of LGDs within the scope of application for IRB-CCF in the CRR3? For which cases would you not have enough data to estimate CCFs but have enough data to es-timate own estimates of LGDs?
Paragraph 5 of section 3, “Background and rationale” of the EBA consultation paper is intended to clarify that institutions, which do not meet the requirements for estimating IRB-CCFs, should apply the regulatory proscribed LGDs based on the IBA approach. In effect, this means applying the IRB basic approach to the corresponding exposures.
In our view, this would result in the EBA exceeding not only the requirements of CRR (Article 166(8b)) but also those of the Basel Committee on Banking Supervision (paragraph 105, Basel III: Finalising post-crisis reforms), which do not provide for this legal consequence. This requirement is also contradicted by the fact that there are already cases today where RWA must be calculated using SA-CCF and, at the same time, internal LGDs. This applies in the case of a non-revolving product such as a classic instalment loan (e.g. from non-large entities, for special financing exposures or in retail business). Ultimately, institutions not offering revolving products may not apply the A-IRBA in its entirety. In our opinion, this statement also runs counter to the intention of the legislator for institutions that are able to meet the regulatory requirements for internal parameter estimation to also apply them – in this case for the LGD – in order to achieve risk-appropriate capital adequacy. Particular problems would arise in retail banking, for which no LGDs are specified for supervisory purposes.
The requirement should therefore be removed and it should be clarified that even when applying SA-CCF, including in the case of applying Article 166(8b)(b) CRR, A-IRBA LGDs will continue to apply, as is already the case for instalment loans, for example.
In paragraph 132 of the draft guidelines, the EBA would like to give institutions that do not meet the minimum requirements for estimating IRB-CCF due to a lack of data availability the option of using a sufficiently conservative margin of conservation (MoC) instead. In practice, this would lead to a CCF of at least 100% having to be applied.
In our opinion, this requirement also contradicts the stipulations of Article 166(8b) CRR. In particular, this applies if the institution does not have the option to choose. Furthermore, the requirement also contradicts other requirements for the application of MoCs. As the EBA outlines in paragraph 140, the MoC is, according to Basel Committee guidelines, supposed to be commensurate with the anticipated estimation error (CRE36.90). In addition, according to paragraph 168 of the draft guidelines, institutions should ensure that capital requirements are not distorted by the need for excessive adjustments in connection with the MoC. However, where there are too few data points, this is often not possible and therefore runs counter to the requirement outlined above. The situation is similar with regard to the requirement in the PDLGD Guidelines (on which the provisions of the CCF Guidelines now under consultation are based) to develop a plan to address the sources of uncertainty (paragraph 50). Due to the lack of sufficient data, it is not expected that the corresponding MoC will be reduced within a reasonable time frame. Also, it is our understanding that the required backtesting of such estimates can only aim to determine whether these are sufficiently or excessively conservative due to the limited data basis. However, auditing experience shows that justifying compliance with regulatory requirements solely on the basis of (excessive) conservatism does not meet the expectations of auditors, users or the intention of a risk-sensitive RWA calculation.
As a result, it is to be expected that user acceptance of such an estimated parameter would fall and ultimately the desired control effect would be limited or even disappear entirely. We are therefore in favour of this requirement being removed. In connection with the rule in Article 166(8b)b) CRR, in our opinion, there is no need for this requirement for the CCF.
2. Question 2: Do you have any comments related to guidance on the identification of a relat-ed set of contracts which are connected such that they constitute a facility?
Our understanding is that defining a facility at a set of contracts is a very specific use case. As stated in chapter 3 (Background and rationale) paragraph 7 a prerequisite for determining a set of contracts constituting a single facility is a structural interconnection between the contracts. This implies that such an interconnection is subject matter of the respective contracts. This is true e. g. for the example given in chapter 3 (Background and rationale) paragraph 21 d. However, in general contracts are not interconnected in such way, and if such an interconnection is not given, defining a facility at a set of contracts is not appropriate from our point of view. This results in our understanding that defining a facility at a set of contracts is a very specific or even non-existing case. In particular considering retail exposures.
6. Question 6: Have you identified any unwarranted consequences of including fully drawn revolving commitments in the scope of the IRB-CCF. How material are these cases for your institution?
According to paragraph 22 of the consultation paper, all amount increases occurring between the reference data and the default date (as well as beyond that date, where applicable) are to be taken into account in calculating the realised CCF and therefore also for the CCF estimate. This shall explicitly include increases in credit limits based on new credit decisions.
We consider this requirement to be contradictory to the CCF definition in Article 4(1)(56) CRR, which refers to the amount of an existing commitment from a single facility that has not been drawn down at the reference date. In our opinion, however, limit increases based on a new credit decision do not constitute part of the original commitment within the meaning of Article 5(10) CRR. This interpretation is consistent with the ECB’s interpretation according to paragraph 313 letter a) of the EGIM (“This higher (unadvised) credit limit may be disregarded if its availability is subject to a further credit assessment by the institution...”).
The requirement also appears inappropriate from an economic viewpoint. Limit increases are new credit decisions, comparable with new business, and should not therefore be mapped through the CCF model. The corresponding risk only arises when the actual credit decision and commitment are made, and is assessed when it occurs according to the applicable parameters. Taking it into account again as part of the CCF estimate therefore leads to a systematic overestimation of the CCF.
Against this background, limit increases linked to a new credit decision should not be taken into account in either the ex-post measurement or the CCF estimate.
Furthermore, the EBA expects that possible future limit increases and payments should be included, even for exposures that are fully drawn on the reference date (i.e. with no open lines). In these cases, a ‘limit factor’ is to be applied. This seems to us to be problematic, both in methodological as well as in regulatory terms. According to Article 166 CRR, the exposure value shall be sum of the accounting value and the undrawn amount multiplied by the IRBCCF. The introduction of a limit factor is not provided for in the CRR and would also require extensive technical adjustments without creating any discernible added value for risk assessment.
We are therefore firmly opposed to the introduction of this requirement.
According to paragraph 61(b) of the Guidelines, in such cases, realised CCFs should also be calculated and included in the estimate where the open credit line is equal to zero on the reference date (one year before default). As a result, this would also necessitate CCF estimates for fully drawn facilities.
We see a clear contradiction here with the CRR. According to Article 166(1) CRR, “...the exposure value of on-balance sheet exposures shall be the accounting value...”. An estimate of the exposure at default is only envisaged for off-balance sheet items (Article 166(8) ff. CRR). If no open credit line is available on the reference date, by definition there is no off-balance sheet item and therefore an IRB-CCF cannot be applied.
It is also unclear how the anticipated EaD is to be determined in these cases without deviating from CRR requirements. There is no basis in the CRR for introducing a ‘limit factor’ for fully drawn credit lines, as proposed in the consultation paper.
Against this background, no CCF estimates should be implemented for fully drawn facilities on the reference date. Realised CCFs and CCF estimates should continue to apply only in cases where an off-balance sheet item actually exists in the form of an open credit line.
on question 7:
Using a development sample and a testing sample naturally implies that not all available data is used for parameterization. While analysing stability (which is also done in validation) is comprehensible, it should be kept in mind that using all data relevant for risk quantification also for parameterisation is an evident option.
9. Question 9: Do you have any concerns with the requirements introduced to analyse and mitigate a lack of representativeness for CCF? Do the requirements on the different data samples when observing a lack of representativeness impede your ability to model CCF portfolios?
Referring to question 7 we would like to emphasise that using development and testing sample might lead to the result that the development sample as well as the sample used for risk quantification is representative for the application sample, but the testing sample is not. There are options to mitigate the lack of representativeness of the testing sample, but the overall development becomes more cumbersome.
11. Question 11: Are there any concerns with requiring consistency in the analysis of changes in the product mix with the institution’s definition of facility? Are institutions able to identify and link contracts (partially) replacing other contracts where the closing or repayment of one contract is related to the origination of a new contract? Are institutions able to link new contracts that are originated after the reference date to related contracts existing at refer-ence date? In particular, is it possible in the case contracts that are revolving commitments are replaced by contracts that are non-revolving commitments (e.g. by a term loan)?
It is hardly possible to identify contracts (partially) replacing other contracts. Moreover, we want to highlight that a change in product mix by itself is not an indicator that contracts are (partially) replaced by other contracts. As highlighted in question 2 a structural interconnection between such contracts must exists. In our opinion, identifying a structural interconnection is therefore more appropriate than scrutinising a change in the product mix in general.
12. Question 12: Do institutions consider it proportionate to the risks of underestimation of CCF to perform the identification analysis and allocation procedure? If it is deemed not propor-tional, what would be an alternative approach that is still compliant with Article 182(1b) CRR?
Referring to question 11: The change in product mix is generally considered within risk management and lending standards. Therefore, apart from rare cases where a structural interconnection between contracts exists, the requirements set out in 182 (1b) CRR are already addressed by risk management and lending standards.
15. Question 15: Do you agree with the three principles for the calculation for realised CCF in the context of consumer product mix, and their implications for the cases mentioned as ex-amples? In case of disagreement, what is the materiality of the cases with unwarranted re-sults, in particular in relation with the definition of facility applied in your institution? In case of material unwarranted results, can you describe your alternative practice to this CP?
According to paragraph 63, the EBA expects that, when calculating the realised CCF in the numerator, amounts drawn down from non-revolving commitments after the reference date should be taken into account, in addition to amounts drawn down from revolving commitments. The denominator should include only the undrawn amounts of revolving commitments as at the reference date.
We consider this approach inappropriate. Non-revolving commitments are already assessed according to supervisory requirements using SA-CCF when they are created. In our opinion, the additional consideration of drawn amounts in the numerator, without at the same time including the corresponding undrawn amounts in the denominator, leads – depending on the constellation – to considerable distortions in the calculation of the realised CCF. This would have direct and non-intended effects on the IRBA-CCF estimate, creating false control stimuli and would considerably reduce user acceptance due to this inappropriate combination.
We therefore suggest that the use of and commitments for non-revolving products that are already assessed using SA-CCF should be completely excluded in the IRBA-CCF context.
According to paragraph 79 in section 3, “Background and rationale” of the EBA consultation paper, Example box 3 - consumer product mix in Case I indicates that future credit decisions and/or credit approvals should be taken into account in the CCF calculation. The result would be that even the expectation of future business would have to be backed by own funds.
Systematically taking future business into account is not provided for in the CRR. Although there are off-balance sheet exposures in this constellation, nevertheless, the estimation approach is intended to cover an additional off-balance sheet exposure that did not exist at the time of the estimate. However, Article 166 CRR does not provide for this.
Subsequent increases in loan commitments are generally normal business transactions between the customer and the institution and do not pose any risk concerns (e.g. business expansion and changes in the customer relationship, increase in turnover). Only in exceptional cases, when restructuring occurs due to creditworthiness, would a subsequent increase in credit commitments be a circumstance that could be attributed to the existing credit commitment.
In example 3 (Case I), the Draft Guidelines also stipulate that if the limit is increased after the reference date, the additional drawdown amount is included in the numerator of the realised CCF, while the original, lower limit amount remains in the denominator. This can result in realised CCF values of over 100%. We consider this approach problematic.
Firstly, it leads to an inappropriate determination of RWA for existing customers, as future credit decisions or approvals are included in the CCF estimate even though they did not exist on the reference date. This implies unequal treatment of new customers: While a possible future increase in limits is anticipated in the model for existing customers, this is not the case for new customers – which makes it clear that future business is in fact included in the estimate here.
Secondly, we see no legal basis in the CRR for this approach. According to Article 166 CRR, the exposure value of on-balance sheet exposures is the accounting value, while an estimate of EaD is only required for actual off-balance sheet exposures. The proposed inclusion of limit increases in the Draft Guidelines means, however, that an off-balance sheet exposure, which did not exist at the time the estimate was calculated, should nevertheless be included. In our opinion, this is not compatible with the CRR system.
This approach also appears inappropriate from an economic viewpoint. Limit increases are usually an expression of expanding customer business or a change in the customer relationship and are therefore considered safe from a risk perspective. Only in exceptional cases – for example, if a customer’s creditworthiness deteriorates significantly – can a subsequent increase be closely linked to the original credit line. In our view, the blanket inclusion of all limit increases in the estimate clearly goes beyond what is objectively necessary.
We therefore suggest clarifying the guidelines to state that limit increases based on new credit decisions should not be included in the calculation of realised CCF and the CCF estimate.
16. Question 16: Are there any concerns related to the allocation mechanism described in these GL?
Yes, in particular referring to case IX in Example Box 3. In such a scenario an observation would enter the samples where no revolving facility at default date exists.
17. Question 17: Where credit lines are kept open even if the facility is in default, the alterna-tive option described in this consultation box could lead to high realised CCF values. Is this a relevant element for your institution and if yes, why and how material are these cases with-in the scope of IRB-CCF models?
According to paragraph 86 of the draft guidelines, the long-run average (LRA) of the CCF is to be determined as the arithmetic mean of the realised CCFs over the entire observation period, weighted by the number of facilities. It is expressly stipulated that no aggregations across subsets (in particular, no annual averages or averages above obligor level within a grade) may be used.
This requirement directly contradicts existing ECB supervisory practice. The ECB’s Guide on Internal Models (EGIM, paragraph 322(2)) states that, “When the historical observation period is considered to be representative of the LRA, the average realised CCFs should be computed as the arithmetic average of the yearly averages of realised CCFs in that period.” The EGIM explicitly requires the formation of an LRA as the average of the annual average values.
This means that institutions are faced with conflicting requirements from two relevant supervisory sources, which in practice cannot be met simultaneously. This leads to considerable uncertainty in modelling practice and can also result in different findings in regulatory audits.
We therefore suggest harmonising the guidelines with the existing EGIM requirements or at least establishing clear priorities. In particular, it should be made clear whether the LRA must be calculated as a facility-level calculation (as described in paragraph 86) or as the average of yearly averages (as provided for in the EGIM). Only clear and consistent guidelines can ensure consistent and practical implementation by the institutions.
23. Question 23: Do you think that, for the facilities in the region of instability, and/or for fully drawn revolving commitments, a single approach should be prescribed (e.g. one of the ap-proaches above defined in the Basel III framework), or that more flexibility is necessary for institutions to use different approaches they deem most appropriate for these facilities?
According to paragraph 61(b) of the Draft Guidelines, CCFs should also be estimated if the open credit line is equal to 0 one year prior to default. This clearly contradicts the CRR. According to Article 166(1) CRR, “...the exposure value of on-balance sheet exposures (current account, overdraft, credit cards, in this case) shall be the accounting value...”. There is no way to estimate the EaD of the balance sheet portion of the contract. This is only possible for offbalance sheet exposures (see Article 166(8) ff. CRR). However, when the open credit line is 0 there are no off-balance sheet exposures. As a result, there should be no CCFs to estimate in such cases.
27. Question 27: Do you have any comments on the condition set to use the simple approach to estimate additional drawings after default. Do you consider that the simple approach is also relevant for retail portfolios?
CRR 3 provides the discretion to include drawings after default in the CCF estimation also for retail exposures. Taking this option would result the same treatment of drawings after default for retail and non-retail exposures. Consequently, the options, meaning simple approach or modelling approach, to estimate additional drawings should be equivalent for retail and nonretail exposures
31. Question 31: For CCF estimation, do you use estimation methods that incorporate portfolio-level-calibration of the estimates? What are the main reasons to use a calibration at a level that is higher than the grade-level calibration?
C.f. Q 17: With regard to paragraph 86 of the Draft Guidelines, there is a contradiction between the requirements for determining the long-run average CCF (LRA CCF) in the EBA Draft Guidelines and those in the ECB Guide to Internal Models (paragraph 322(d) credit risk chapter), published as recently as July 2025.
EBA: “… institutions should calculate the long-run average CCF as an arithmetic average of realised CCFs over a historical observation period weighted by the number of facilities. Institutions should not use for that purpose any averages of CCFs calculated on a subset of observations, in particular any yearly average CCFs or averages over facilities of the same obligor within one grade.”
ECB: “When the historical observation period is considered to be representative of the LRA, the average realised CCFs should be computed as the arithmetic average of the yearly averages of realised CCFs in that period.”
This means that, according to the EBA requirement, annual average CCF values should not be used; in contrast, in fact, the ECB requirement stipulates that annual average CCF values should be used. We support the EBA’s approach here. Following publication of the final EBA guidelines, the ECB requirement should be adjusted accordingly.
34. Question 34: Are there examples where the haircut approach should be considered the most appropriate approach for estimating the downturn CCF?
Considering a scenario where a macroeconomic risk-driver is incorporated in the CCF model, it would be natural to apply the haircut approach.
37. Question 37: The possibility to have no downturn effect on CCF estimates is restricted to the case where observations are available during a downturn period. Which alternative meth-odologies could be used to prove the non-existence of a downturn effect on CCF estimates, in the case where no observation is available during a downturn period?
In the case where no observation is available during a downturn period, a downturn effect would rely on a statistical dependency between realized CCFs and economic indicators. As already questioned in Question 36 it is not unlikely that no statistically significant correlation between economic indicators and realized CCFs exists. Meaning there is no statistical indication for a downturn effect on CCF. Consequently, in such a case it would be natural to have a downturn effect (in terms of addon) of 0.