Response to consultation on Guidelines for the estimation of LGD appropriate for an economic downturn
Go back
Our proposal is to replace this 20% add-on with the reference value approach which can instead be disregarded as a benchmark option in case of downturn adjustment computed coherently with prescriptions of Section 5 and 6.
− The average LGD from the two worst years by facility grade / pool can be subject to the low number of facilities. This issue is particularly critical in particular for some segments and the result in that case can be more determined by the low number of facilities than by effective peaks of loss rates. The layer of computation of the reference values should be reconsidered to adequately consider this aspect;
− It is not clear how to deal with correlation structure among the intermediate parameters;
− In case of model component (typically cure rate and LGD for “cured” and “not cured”) each component should be analyzed with the perspective of the two worst years over 20 (10) years. First question: should the two worst years be computed according to each axe of analysis (“cured” losses, “not cured” losses, probability to be cured) or should it be one for all determined by the analysis of the “not cured” losses? In this case how the different components should be considered altogether also taking into account the correlation structure above introduced?
− How to deal with some extraordinary events (such as massive credit disposal) which can strongly influence the reference values but are not result of a downturn situation and are also potentially mitigated within the LGD calibration (see EBA GL on PD estimation, LGD estimation and the treatment of defaulted exposures).
EBA clearly indicated that reference values should be considered as a “soft floor”; nonetheless, it is evident that institutions have the burden of proof to demonstrate potential misalignments with internally obtained results. This consequence appears not appropriate, in particular by considering the new framework proposed by the EBA with the “impact assessment”. The combination of the two analyses seems too burdensome and useless, in particular in case of downturn estimation based on observed losses (Section 5 of the GL). The impact assessment has a clear framework and is strictly related to the definition of the economic factors. The same does not hold for the reference values which are referred to the peaks of the losses and represent a floor difficult to compare with the model results since the peak of the losses cannot be linked to a macroeconomic distress (downturn); for example, the peak of the losses can be determined by an extraordinary event such as a massive credit sale which is the result of managerial policy to reduce NPL likely not performed in a downturn situation but more result of previous downturn situations or regulatory constraints (reduction of NPL ratio).
Reference values approach could be disregarded with the new framework based on the impact assessment since it not necessary to have two backtestings of the internal downturn model. As introduced in answers to Questions 6 and 8, reference value approach can instead substitute the approach proposed in Section 7 of the fixed add-on for the cases with no internal data for downturn adjustment estimation.
Question 1: Do you think that additional guidance around the estimation of LGD in-default, which reflect downturn conditions, is needed? If yes, could you provide examples of sound methodologies for transposing downturn LGD estimates from performing to non-performing exposures?
The proposal to adopt the performing downturn LGD estimation also for the non-performing exposures is shareable. The only exception can be represented by different samples due to peculiar exclusion for defaulted asset model: in this case the analysis should be performed again but still neglecting the reference dates (“vintage” component) which could sensibly increase the complexity of the downturn quantification.Question 2: Do you share the concern that the proposed policy in paragraph 15 could create an undue burden if applied to every downturn period identified? If yes, in order to better balance the accuracy of the estimations and its operational complexity what evidence should be provided by institutions in order to justify the exemption of identified downturn periods from the proposed policy in paragraph 15?
Yes, the analysis should be performed coherently with the available historical series for the LGD estimation and taking into account potential structural breaks (such as for example the introduction of the Euro within the Eurozone). The rule of the twenty years could be adapted according to these intuitive matters in order to avoid an undue burden on institutions. Finally, we think that, if different downturn periods result from the economic factors analysis, the rule should not be “the worst of the worst” but an average of the different downturn periods.Question 3: Do you agree with the proposed level of downturn LGD estimation set out in paragraph 14? In particular, do you support the concept that the downturn LGD estimates of different calibration segments could be based on different downturn periods? Is the policy on the level of downturn LGD estimation as well as the relation between the level of downturn LGD estimation and the relevant downturn periods sufficiently clear?
We think that the proposed level of the downturn LGD estimation and the relationship with the downturn periods (i.e. the economic factors) should be better explained. A critical aspect is the relationship between “type of exposure” introduced in the RTS, the “class of exposures” defined in the CRR IV and “calibration segment” introduced by the GL for downturn LGD estimation: the definitions should be better defined and harmonized and the link has to be clearly indicated. In case of economic factors related to real estate exposures vs. unsecured exposures within the same regulatory segment (i.e. Corporate class of exposure) an example could better clarify: a) how to deal with peaks and troughs deriving from different downturn periods (in the same class of exposure); b) how to compute the downturn impact; c) how to apply the downturn impact to the long run average LGDs estimated by grades / pools where the latter are even more granular than the simple secured vs. unsecured differentiation (due to the other risk drivers). Granularity: the proposed policy in the GL allows to quantify downturn LGD estimates at a more granular level than long-run average LGD estimates where this provides more appropriate final downturn LGD estimates. How to combine this aspect with the above issue on type of exposure vs. calibration segment?Question 4: Do you consider the description of the approaches to be sufficiently clear?
The description of both the “haircut” and the “extrapolation” approaches should be integrated with some further practical examples for a clear comprehension.Question 5: Do you agree to the limitation of approaches for quantification of downturn LGD estimates? If not, which other approaches should be considered? Would you prefer the alternative policy considered – if yes how should a minimum MoC be established in this case?
Yes, we agree but we ask for more details on the alternative approaches provided in Section 6. We do not have specific opinions on further alternative policies already disregarded by the EBA.Question 6: Do you expect that the total exposure amount or share which is treated with the policy proposed in Section 7 is material?
We think that this option can have a direct impact on LDP as long as the IRB approaches are permitted. In general, our opinion is that the 20% add-on is too conservative and not justified. Moreover, it is not clear how it has been estimated since a subjective measure is clearly dangerous on such impacting topics. We understand that the idea is to provide a strong incentive to the institutions for an internal estimation of the downturn adjustment but we think that, in some cases such as LDP / low data portfolios, the unavailability of data is a crucial aspect. Finally, we deem that the 105% cap is not reasonable since there is not a clear rationale for which the Downturn loss should overcome 100%.Our proposal is to replace this 20% add-on with the reference value approach which can instead be disregarded as a benchmark option in case of downturn adjustment computed coherently with prescriptions of Section 5 and 6.
Question 7: Do you have specific examples of types of exposures which will fall under the policy proposed in Section 7?
LDP portfolios can be a clear example.Question 8: Do you agree to require a minimum MoC quantified via a fixed add-on to the long-run average LGD? If not, which of the alternatives should be considered? Do you see reasons for differentiating the fixed add-on according to exposure classes?
As introduced in Question 6, we think that the fixed 20% add-on is too penalizing and not justified. We understand that the idea is to provide a strong incentive to the institutions for an internal estimation of the downturn adjustment but we think that, in some cases such as LDP, the unavailability of data is a crucial aspect. Our proposal is to replace this 20% add-on with the reference value approach which can instead be disregarded as a benchmark option in case of downturn adjustment computed coherently with prescriptions of Section 5 and 6. The adoption of the reference value approach can also allow differentiating the floor among exposure classes.Question 9: Do you agree to the minimum MoC as the max(0,min(20%, 105% - LRAVLGD)?
No, as explained in Question 6 and 8.Question 10: Is the policy regarding the reference value sufficiently clear? Alongside with the potentially limited applicability of the reference value to the downturn LGD estimation according to paragraphs 18-19, for what reasons could the reference value feasibly be omitted? Do you agree to the proposed clarification of the role of the reference value?
The policy on the reference values has some issues to be clarified:− The average LGD from the two worst years by facility grade / pool can be subject to the low number of facilities. This issue is particularly critical in particular for some segments and the result in that case can be more determined by the low number of facilities than by effective peaks of loss rates. The layer of computation of the reference values should be reconsidered to adequately consider this aspect;
− It is not clear how to deal with correlation structure among the intermediate parameters;
− In case of model component (typically cure rate and LGD for “cured” and “not cured”) each component should be analyzed with the perspective of the two worst years over 20 (10) years. First question: should the two worst years be computed according to each axe of analysis (“cured” losses, “not cured” losses, probability to be cured) or should it be one for all determined by the analysis of the “not cured” losses? In this case how the different components should be considered altogether also taking into account the correlation structure above introduced?
− How to deal with some extraordinary events (such as massive credit disposal) which can strongly influence the reference values but are not result of a downturn situation and are also potentially mitigated within the LGD calibration (see EBA GL on PD estimation, LGD estimation and the treatment of defaulted exposures).
EBA clearly indicated that reference values should be considered as a “soft floor”; nonetheless, it is evident that institutions have the burden of proof to demonstrate potential misalignments with internally obtained results. This consequence appears not appropriate, in particular by considering the new framework proposed by the EBA with the “impact assessment”. The combination of the two analyses seems too burdensome and useless, in particular in case of downturn estimation based on observed losses (Section 5 of the GL). The impact assessment has a clear framework and is strictly related to the definition of the economic factors. The same does not hold for the reference values which are referred to the peaks of the losses and represent a floor difficult to compare with the model results since the peak of the losses cannot be linked to a macroeconomic distress (downturn); for example, the peak of the losses can be determined by an extraordinary event such as a massive credit sale which is the result of managerial policy to reduce NPL likely not performed in a downturn situation but more result of previous downturn situations or regulatory constraints (reduction of NPL ratio).
Reference values approach could be disregarded with the new framework based on the impact assessment since it not necessary to have two backtestings of the internal downturn model. As introduced in answers to Questions 6 and 8, reference value approach can instead substitute the approach proposed in Section 7 of the fixed add-on for the cases with no internal data for downturn adjustment estimation.