Response to cP on Guidelines on Credit Risk Mitigation for institutions applying the IRB approach with own estimates of LGDs
Go back
• §29.a) (i) “Modelling approach” and (ii) “Substitution approach” (for all the UFCP techniques and typically for guarantees issued by “Mutual cooperatives guarantees fund for SME loans”, named Confidi);
• §29.b) Substitute SA in defining of risk weight (typically) to recognize the effect of public funds counter - guaranteed by State (Fondo Centrale di garanzia per le PMI o Fondo di garanzia per la prima casa).
We would like to draw attention on the requirement to recognize the UFCP in the Substitution approach in accordance to Chapter 4 of CRR (see § 36 b). These requirements – the same for Foundation IRB Approach – look like too strict and limit the range of possible UFCP which could be recognized, with the consequence to create a strong penalization for the AIRB institutions.
We propose to consider only the requirements in accordance to Chapter 3 of CRR.
Another point we would like to draw attention is when it comes to unfunded credit protection and rating transfer, to the link existing between provisions of Section 6.2 of this EBA CP and section “5.2.3 Treatment of ratings of third parties”, par. 62 of “Guidelines on PD estimation, LGD estimation and the treatment of defaulted exposures” (EBA/GL/2017/16).
On the one hand, based on the section 6.2 “The effects of unfunded credit protection”, par. 29 of this EBA CP, institutions may recognize the credit risk mitigation effects of unfunded credit protections using the adjustment of PD or LGD estimates in accordance with Article 160(5), 161(3) and 164(2) of CRR, on the basis of the criteria specified by institutions in accordance with Article 183(2) and (3) of CRR, as further specified in paragraphs 35 and 36.
In particular, one of the approaches currently used by IRB banks is the “substitution approach”.
The necessary condition for the “substitution approached” application is the eligibility requirement for unfunded credit protection. According to the section 5.2, the unfunded credit protection (e.g., guarantees) is eligible if it is legally effective and enforceable in all relevant jurisdictions.
On the other hand, based on the section “5.2.3 Treatment of ratings of third parties”, par. 62 of “Guidelines on PD estimation, LGD estimation and the treatment of defaulted exposures” it is stated that:
“Institutions should have clear policies specifying the conditions under which the rating of a third party who has a contractual or organizational relation with an obligor of the institution may be taken into account in the assessment of risk of the considered obligor”.
Afterwards, the EBA lists the possible manners in which the rating of such a third party may be taken into account in the assessment of risk of the considered obligor. In particular, under par.62(a) the EBA requires that:
“…the rating of such third party being transferred to a relevant obligor (‘rating transfer’), where there is no difference in risk between the obligor and the related party because of the existence of an appropriate guarantee and the rating of a third party is assigned internally in accordance with the rating system for which the institution has received permission in accordance with Article 143(2) of Regulation (EU) No 575/2013”.
According to the above, a prerequisite for “rating transfer’” is “the existence of an appropriate guarantee”.
What par. 62 implies is that while it is on the intermediary to specify the conditions of a clear policy on the use of the rating of a third party in the assessment of risk of the considered obligor, the guarantee required for risk transfer must be “appropriate”.
Therefore, it does not need to be “eligible” under Article 183 of Regulation (EU) CRR and, therefore, the definition of the “clear policy” set by the institution should also include and support a definition of “appropriate” guarantee.
It is therefore not clear if the PD “substitution approach” in EBA CP is considered to be equivalent to the rating transfer approach (EBA/GL/2017/16) and, if so, if the eligibility requirements for the CRM purposes (EBA CP) are considered to be equivalent to the appropriateness requirements for the rating transfer purposes (EBA/GL/2017/16).
The process of rating transfer should prevent any fully consolidated and controlled subsidiary from being granted a lower riskiness than its consolidating parent company, as there is no difference in risk between such counterparties. Therefore, when the consolidating parent company has a low credit standing with a high PD, its subsidiary should be given the same creditworthiness level through the “rating transfer” approach. On the contrary, the application of the credit risk mitigation techniques through the PD and/or LGD substitution between Guarantor and Guaranteed would result only in a reduction of the risk profile of the Guaranteed counterparty (otherwise, the PD and/or LGD substitution is not applied, keep its stand-alone rating grade).
We would like in addition to make sure that the qualification of “organizational relation” (e.g., the strong connection between a parent company with its fully consolidated and controlled subsidiaries) in par. 62 of EBA/GL/2017/16 would be qualified as an “appropriate” guarantee, thereby supporting the rating transfer. Should not this be the case, par 62 of EBA /GL/2017/16 would prove of a very limited use in actual practice.
Also, the recognition of the “conditional guarantees” is important if the empirical evidence/data collected by lenders may support this possibility in term of reduction of risks.
However, as already said before, CP introduces strong limitations in the collateral effects under the AIRB approach, foreseeing the tighter FIRB requirements in case of not eligibility of the collateral. Indeed, the only eligible collaterals are those for which the effects on the LGD are possible to model. This requirement will probably lead to the application of the FIRB approach to most of the secured exposures, that will also be used when considering the effect on LGD floors, as the haircuts to apply are considered according to the comprehensive approach (under the foundation approach).
Actually, we would support an approach as much as possible aligned to the GL on PD and LGD and on Downturn component, especially considering the Basel 3 framework on Input floor and haircut on AIRB models: the downturn component potentially double counts the effect derived by the implementation of not negligible floors to the LGD (both secured and unsecured post collateral haircut application are already very punitive).
Regarding the LGD floor imposed to the secured part of the exposure, aimed at embedding in the model the risk of reduced realizable value of the collateral, we deem that it duplicates the effect of both the regulatory and internally estimated haircuts, which already cover this event. In addition, as far as the unsecured portion is concerned, the LGD estimate already considers the downturn effect. Consequently, this would lead to higher haircuts and floors double-counts stressed scenario variables which have been already considered in the LGD estimates.
In addition to the ones on LGD, the requirements used in FIRB and SA would create a strong disincentive for banks to model internal haircuts given that the collaterals subject to the largest devaluation would be penalized while the ones with better quality would not benefit from it (given that the input floors are applied at single transaction level). This basically means that banks are encouraged to apply the standardized and FIRB parameter for the collateral evaluation even under the AIRB approach.
In the end, we do not agree with the content of § 30. Institutions should be free to apply the “substitution effect” without limitations, as if institutions were exposed directly to guarantor (and so considering the guarantor’s portfolio).
As a matter of fact, some institutions have invested and adopted the Advanced IRB Approaches in order to assess credit risk more deeply and carefully, basing on the effective experience (cash flows during the foreclosure procedures) in a medium/long term. So, they should take into account all the possible “empirical” CRM effects, regardless if the CRM techniques fulfil “theorical” requirements.
We propose that the EBA Guidelines allow institutions to perform appropriate adjustment of PD and/or LGD even relying on ineligible collateral o guarantees, provided that the internal model is overall consistent.
Furthermore, it is not clear whether the substitution approach can be applied only if the execution costs are expected to be negligible, or it is necessary to shift to the modelling approach. We, therefore, suggest adding some clarification in the final guidelines on the following sentence: “the institution may reasonably expect that the direct costs of exercising the unfunded credit protection are negligible with respect to the amount covered by the unfunded credit protection.”.
In particular, we wonder whether it should be considered the effect of the other existing unfunded credit protections on the same part of the exposure (in accordance with paragraphs 40) or not. Furthermore, we deem not clear the reported example when although more collaterals (funded and unfunded) are able to cover the whole exposure, an unsecured exposure portion still remains (see Figure 4). We would therefore recommend adding more detail in the final guideline in order to better explain how to consider the effect of credit risk mitigation.
In any case, the requirements/methodologies to assess the LGD in case of multiple collaterals/guarantees could not take into account the internal assessment of the institution. They could be too strict and penalising, since they create a limitation in the recognition of CRM techniques even if institutions have empirical evidence of the effect of them on LGD.
We propose that the value of LGD should be based on the internal track record of cash flows recorded by institutions.
a) in case the guarantee is provided to a single loan, the guarantee
isk mitigation is no longer active, and it terminates its coverage effect
b) whereas, when the guarantee is provided to a portfolio of loans, the occurred loss of a single loan does not imply the termination of the risk mitigation effects because the contractual guarantee continues to be operative and effective until its legal maturity (that shall not be shorter than of the covered portfolio Weighted Average Life). This is the specific feature of all securitization transactions:
As a matter of fact, within art 4(61) the CRR states that a securitization is “a transaction or scheme, whereby the credit risk associated with an exposure or pool of exposures is tranched, having both of the following characteristics: (a) payments in the transaction or scheme are dependent upon the performance of the exposure or pool of exposures; (b) the subordination of tranches determines the distribution of losses during the ongoing life of the transaction or scheme”.
It is possible to consider the portfolio guarantee as eligible Unfunded Credit Protection, based on rules defined under Chapter 4, for own synthetic transactions portfolio. In particular, for Synthetic securitizations it is possible to divide the risk in one or more tranches, transferring it to a third party and reflecting in the credit risk parameter the CRM effects of the UFCP.
As a general comment with respect to this issue, we would like to point out that the UFCP schemes described in the draft guidelines, shall operate and be used both on STA and A-IRB portfolio. The effect of the UFCP should be, indeed, independent of the eligibility rules applicable on the guarantor, and we recommend allowing the adjustment of the relevant parameters in case of both STA and IRB underlying portfolios.
Question 1: Do you agree with the proposed clarifications on eligibility requirements in accordance with Article 181(1)(f) of the CRR?
NoQuestion 2: Do you agree with the proposed clarifications on the assessment of legal certainty of movable physical collateral? How do you currently perform the assessment of legal effectiveness and enforceability for movable physical collateral?
NoQuestion 2: Do you agree with the proposed clarifications on the assessment of legal certainty of movable physical collateral? How do you currently perform the assessment of legal effectiveness and enforceability for movable physical collateral?
NoQuestion 4: Do you have specific concerns related to the recognition of collateral in the modelling of LGD? How do you currently recognise collateral in your LGD estimates?
NoQuestion 5: What approaches for the recognition of the unfunded credit protection do you currently use? What challenges would there be in applying approaches listed above for the recognition of unfunded credit protection?
In general, the Italian banking industry uses especially the approaches defined in:• §29.a) (i) “Modelling approach” and (ii) “Substitution approach” (for all the UFCP techniques and typically for guarantees issued by “Mutual cooperatives guarantees fund for SME loans”, named Confidi);
• §29.b) Substitute SA in defining of risk weight (typically) to recognize the effect of public funds counter - guaranteed by State (Fondo Centrale di garanzia per le PMI o Fondo di garanzia per la prima casa).
We would like to draw attention on the requirement to recognize the UFCP in the Substitution approach in accordance to Chapter 4 of CRR (see § 36 b). These requirements – the same for Foundation IRB Approach – look like too strict and limit the range of possible UFCP which could be recognized, with the consequence to create a strong penalization for the AIRB institutions.
We propose to consider only the requirements in accordance to Chapter 3 of CRR.
Another point we would like to draw attention is when it comes to unfunded credit protection and rating transfer, to the link existing between provisions of Section 6.2 of this EBA CP and section “5.2.3 Treatment of ratings of third parties”, par. 62 of “Guidelines on PD estimation, LGD estimation and the treatment of defaulted exposures” (EBA/GL/2017/16).
On the one hand, based on the section 6.2 “The effects of unfunded credit protection”, par. 29 of this EBA CP, institutions may recognize the credit risk mitigation effects of unfunded credit protections using the adjustment of PD or LGD estimates in accordance with Article 160(5), 161(3) and 164(2) of CRR, on the basis of the criteria specified by institutions in accordance with Article 183(2) and (3) of CRR, as further specified in paragraphs 35 and 36.
In particular, one of the approaches currently used by IRB banks is the “substitution approach”.
The necessary condition for the “substitution approached” application is the eligibility requirement for unfunded credit protection. According to the section 5.2, the unfunded credit protection (e.g., guarantees) is eligible if it is legally effective and enforceable in all relevant jurisdictions.
On the other hand, based on the section “5.2.3 Treatment of ratings of third parties”, par. 62 of “Guidelines on PD estimation, LGD estimation and the treatment of defaulted exposures” it is stated that:
“Institutions should have clear policies specifying the conditions under which the rating of a third party who has a contractual or organizational relation with an obligor of the institution may be taken into account in the assessment of risk of the considered obligor”.
Afterwards, the EBA lists the possible manners in which the rating of such a third party may be taken into account in the assessment of risk of the considered obligor. In particular, under par.62(a) the EBA requires that:
“…the rating of such third party being transferred to a relevant obligor (‘rating transfer’), where there is no difference in risk between the obligor and the related party because of the existence of an appropriate guarantee and the rating of a third party is assigned internally in accordance with the rating system for which the institution has received permission in accordance with Article 143(2) of Regulation (EU) No 575/2013”.
According to the above, a prerequisite for “rating transfer’” is “the existence of an appropriate guarantee”.
What par. 62 implies is that while it is on the intermediary to specify the conditions of a clear policy on the use of the rating of a third party in the assessment of risk of the considered obligor, the guarantee required for risk transfer must be “appropriate”.
Therefore, it does not need to be “eligible” under Article 183 of Regulation (EU) CRR and, therefore, the definition of the “clear policy” set by the institution should also include and support a definition of “appropriate” guarantee.
It is therefore not clear if the PD “substitution approach” in EBA CP is considered to be equivalent to the rating transfer approach (EBA/GL/2017/16) and, if so, if the eligibility requirements for the CRM purposes (EBA CP) are considered to be equivalent to the appropriateness requirements for the rating transfer purposes (EBA/GL/2017/16).
The process of rating transfer should prevent any fully consolidated and controlled subsidiary from being granted a lower riskiness than its consolidating parent company, as there is no difference in risk between such counterparties. Therefore, when the consolidating parent company has a low credit standing with a high PD, its subsidiary should be given the same creditworthiness level through the “rating transfer” approach. On the contrary, the application of the credit risk mitigation techniques through the PD and/or LGD substitution between Guarantor and Guaranteed would result only in a reduction of the risk profile of the Guaranteed counterparty (otherwise, the PD and/or LGD substitution is not applied, keep its stand-alone rating grade).
We would like in addition to make sure that the qualification of “organizational relation” (e.g., the strong connection between a parent company with its fully consolidated and controlled subsidiaries) in par. 62 of EBA/GL/2017/16 would be qualified as an “appropriate” guarantee, thereby supporting the rating transfer. Should not this be the case, par 62 of EBA /GL/2017/16 would prove of a very limited use in actual practice.
Question 6: Do you have any specific concerns related to the issues excluded from the scope of the Guidelines?
Considering the importance of the SA guarantor for the Italian banks’ portfolios, we support the possibility to use the SA RW substitution in all the 29.a) approaches, giving to banks the flexibility for its adoption (without any mandatory principle), according to internal empirical evidence/best practices.Also, the recognition of the “conditional guarantees” is important if the empirical evidence/data collected by lenders may support this possibility in term of reduction of risks.
However, as already said before, CP introduces strong limitations in the collateral effects under the AIRB approach, foreseeing the tighter FIRB requirements in case of not eligibility of the collateral. Indeed, the only eligible collaterals are those for which the effects on the LGD are possible to model. This requirement will probably lead to the application of the FIRB approach to most of the secured exposures, that will also be used when considering the effect on LGD floors, as the haircuts to apply are considered according to the comprehensive approach (under the foundation approach).
Actually, we would support an approach as much as possible aligned to the GL on PD and LGD and on Downturn component, especially considering the Basel 3 framework on Input floor and haircut on AIRB models: the downturn component potentially double counts the effect derived by the implementation of not negligible floors to the LGD (both secured and unsecured post collateral haircut application are already very punitive).
Regarding the LGD floor imposed to the secured part of the exposure, aimed at embedding in the model the risk of reduced realizable value of the collateral, we deem that it duplicates the effect of both the regulatory and internally estimated haircuts, which already cover this event. In addition, as far as the unsecured portion is concerned, the LGD estimate already considers the downturn effect. Consequently, this would lead to higher haircuts and floors double-counts stressed scenario variables which have been already considered in the LGD estimates.
In addition to the ones on LGD, the requirements used in FIRB and SA would create a strong disincentive for banks to model internal haircuts given that the collaterals subject to the largest devaluation would be penalized while the ones with better quality would not benefit from it (given that the input floors are applied at single transaction level). This basically means that banks are encouraged to apply the standardized and FIRB parameter for the collateral evaluation even under the AIRB approach.
In the end, we do not agree with the content of § 30. Institutions should be free to apply the “substitution effect” without limitations, as if institutions were exposed directly to guarantor (and so considering the guarantor’s portfolio).
Question 7: Do you agree with the proposed clarification regarding the parallel treatment of ineligible UFCP and ineligible FCP? How do you currently monitor the cash flows related to ineligible unfunded credit protection and how do you treat such cash flows with regard to the PD and LGD estimates?
We do not agree with the limitations/requirements included in § 31 which look very burdensome and penalising. Institutions should be free to use PD and LGD which are internally assessed relying on the regular and empirical track record of cash flows (even recorded during the foreclosure procedures), taking into account every UFCP o FCP, without any limitations (especially if limitations are based on stricter requirements of Chapter 4).As a matter of fact, some institutions have invested and adopted the Advanced IRB Approaches in order to assess credit risk more deeply and carefully, basing on the effective experience (cash flows during the foreclosure procedures) in a medium/long term. So, they should take into account all the possible “empirical” CRM effects, regardless if the CRM techniques fulfil “theorical” requirements.
We propose that the EBA Guidelines allow institutions to perform appropriate adjustment of PD and/or LGD even relying on ineligible collateral o guarantees, provided that the internal model is overall consistent.
Question 8: Do you agree with the proposed rules for the application of the substitution approach? Do you see any operational limitations in excluding the guaranteed part of exposure to which substitution approach is applied from the scope of application of the LGD model for unguaranteed exposures?
We understand the rationale of the proposed rules for the application of the substitution approach.Furthermore, it is not clear whether the substitution approach can be applied only if the execution costs are expected to be negligible, or it is necessary to shift to the modelling approach. We, therefore, suggest adding some clarification in the final guidelines on the following sentence: “the institution may reasonably expect that the direct costs of exercising the unfunded credit protection are negligible with respect to the amount covered by the unfunded credit protection.”.
Question 9: Do you agree with the proposed rules for the application of the modelling approach?
We agree with the proposed rules for the application of the modelling approach, where the unfunded credit protection is considered as a risk driver in the model development.Question 10: What challenges would you envisage for back-testing the substitution approach? Do you agree that the back-testing should be performed rather at Expected loss level? Do you have any approach currently in place for the back-testing of substitution approach?
A backtesting might be hypothetically done by treating the Expected Loss of the Guarantor as a sort of “LGD Secured” for the tranche of exposure covered by the UCP contributing to the overall estimated LGD (combining Secured and Unsecured) in order to compare this with the Realized LGD observed on the facility supported by personal guarantees. This kind of back-testing leads to consider the substitution approach similar to the modelling one (since it would be as consider the mitigation effect in the LGD modelling).Question 11: Do you agree with the proposed guidance for the estimation of the LGD of comparable direct exposure towards the guarantor? What concerns would you have about the calculation of the risk weight floor?
We would have some concerns on the feasibility to satisfy the requirement foreseen by article 37.C related to the calculation of risk weight floor where the original exposure benefits from multiple unfunded credit protection and where two or more are providing protection to the same part of the original exposure. Specifically, in case of multiple unfunded credit protection on the same part of the exposure it is not clear the criteria to use for the calculation of each risk weight direct exposures to the guarantor.In particular, we wonder whether it should be considered the effect of the other existing unfunded credit protections on the same part of the exposure (in accordance with paragraphs 40) or not. Furthermore, we deem not clear the reported example when although more collaterals (funded and unfunded) are able to cover the whole exposure, an unsecured exposure portion still remains (see Figure 4). We would therefore recommend adding more detail in the final guideline in order to better explain how to consider the effect of credit risk mitigation.
In any case, the requirements/methodologies to assess the LGD in case of multiple collaterals/guarantees could not take into account the internal assessment of the institution. They could be too strict and penalising, since they create a limitation in the recognition of CRM techniques even if institutions have empirical evidence of the effect of them on LGD.
We propose that the value of LGD should be based on the internal track record of cash flows recorded by institutions.
Question 12: Do you consider portfolio guarantees as a form of eligible UFCP? Do they include cases where the guarantee contract sets a materiality threshold on portfolio losses below or above which no payment shall be made by the guarantor? Do they include cases where two or more thresholds (caps) either expressed in percentages or in currency units are set to limit the maximum obligation under the guarantee? How do you recognise the portfolio guarantees’ credit risk mitigation effects in adjusting risk parameters?
Unfunded Credit Protection (UFCP) provided to a portfolio of loans rather than to individual exposure is subject to a different mechanism, even in case the credit worthiness of the guarantor remains unchanged. In particular, once the loss is realized and the guarantee activated, there are two possible occurrences:a) in case the guarantee is provided to a single loan, the guarantee
isk mitigation is no longer active, and it terminates its coverage effect
b) whereas, when the guarantee is provided to a portfolio of loans, the occurred loss of a single loan does not imply the termination of the risk mitigation effects because the contractual guarantee continues to be operative and effective until its legal maturity (that shall not be shorter than of the covered portfolio Weighted Average Life). This is the specific feature of all securitization transactions:
As a matter of fact, within art 4(61) the CRR states that a securitization is “a transaction or scheme, whereby the credit risk associated with an exposure or pool of exposures is tranched, having both of the following characteristics: (a) payments in the transaction or scheme are dependent upon the performance of the exposure or pool of exposures; (b) the subordination of tranches determines the distribution of losses during the ongoing life of the transaction or scheme”.
It is possible to consider the portfolio guarantee as eligible Unfunded Credit Protection, based on rules defined under Chapter 4, for own synthetic transactions portfolio. In particular, for Synthetic securitizations it is possible to divide the risk in one or more tranches, transferring it to a third party and reflecting in the credit risk parameter the CRM effects of the UFCP.
As a general comment with respect to this issue, we would like to point out that the UFCP schemes described in the draft guidelines, shall operate and be used both on STA and A-IRB portfolio. The effect of the UFCP should be, indeed, independent of the eligibility rules applicable on the guarantor, and we recommend allowing the adjustment of the relevant parameters in case of both STA and IRB underlying portfolios.