Response to consultation on Guidelines PD estimation, LGD estimation and treatment of defaulted assets
Go back
• It is expected that most, if not all, models will require rebuild and approval as a result of the IRB repair programme which will also need to take into account local competent authority changes to methodology for example the UK PRA CP29/16 proposals for residential mortgage risk weights.
• We think that this question should be assessed in the light of the quantitative impact study that the EBA is currently conducting.
• Nevertheless, some indications are that:
o The LGD in-default guidelines could have a material impact;
o ELBE could be potentially material in many banks especially if the indirect approach is used; and
o In terms of PD and LGD, the methodological impact is limited but the operational cost of enhancements to documentation, justifications and changes to processes is significant.
Timeline for implementation
With regard to the proposed deadline for implementation of end-2020, we think that this date should only be fixed after the EBA has:
1. Had opportunity to review the comments sent in response to the CP;
2. Assessed the results of the qualitative survey to understand the impact of the proposed requirements on the rating systems; and
3. Received feedback on the yet to be published draft RTS on the nature, severity and duration of economic downturn to be developed in accordance with Article 181(3)(a) of Regulation (EU) 575/2013. As these RTS will be closely related to the estimation of downturn LGD some additional changes may be introduced in the final guidelines on the basis of the feedback received during these consultations.
We are particularly concerned about the impact of the change on competent authorities’ ability to review and approve revised models to enable a consistent implementation date.
We note that the Basel Committee is soon expected to publish its guidance on the use of IRB models. This may also have an impact upon the priority that institutions assign for the redevelopment of IRB models.
Our view is that the concerns expressed by the EBA with regard to the variability and comparability of the risk weights do not apply uniformly across all portfolios across the EU. So competent authorities should be given discretion to determine the priority for the redevelopment of models.
We suggest that a prudent objective may be to prioritise a target of c. 75% of IRB risk weighted assets being compliant by a certain date instead of a full implementation for all IRB model by end-2020.
Taking into account the scope of the changes, we urge the EBA to take a proportionate approach in the implementation of the changes to ease the burden on institutions and competent authorities.
Operational impacts
• We do not foresee any operational limitations.
• In general, one-year default rates are already being calculated at least on a quarterly basis.
• We agree with the requirement to calculate the default rates at least quarterly basis.
• However, it could be an issue for Low Default Portfolios (LDP).
• As far as Retail exposures are concerned, we agree with the proposed treatment of additional drawings for both CCF and LGD. Our members’ models are aligned with CRR and EBA guidelines.
• The approach concerning fees and interest is very important for an appropriate LGD computation. All the fees are considered in the economic loss as well as all the other direct costs. On the other side the interests can be further divided in two categories:
• Contractual interest: These interests have not been considered in the loss rate computation since their inclusion would result in a double counting with respect to the discounting process (whose section is separately treated in the guidelines). In case of perfect alignment between contractual interest rate and discounting rate this treatment would not determine any distortion in the loss rate computation. Nevertheless banks tend to apply a current rates approach for the discounting process as also suggested by the BCBS Working Paper 14: “their use allows the consideration of all available information and facilitates the comparison between LGD estimates from different portfolios”. This approach can determine negative loss rates for the different consideration of the value of money over time. More specific comments on this topic should be provided in the dedicated section of the guidelines;
• Unpaid late fees: These interests are included in the exposure of the denominator of the loss rate. But the guidelines ask banks to consider that, in case of recovery of late interest that have not been previously capitalised, the moment of recovery should be considered a moment of capitalisation. Does this lead to excluding the receipt of unpaid late fees interests and exceeding the amount included in the EAD for the loss rate computation? If the EBA thinks so then we do not agree with the proposal. Our opinion is that a receipt of unpaid late fees should not distort the economic loss estimation. Our opinion is that all the receipts should be considered without any specific treatment for the case of unpaid late fees.
A question for further discussion:
Article 115 states that additional recovery cash flows should be added to the calculation at the date of the return to non-defaulted status in the amount that was outstanding at the date of the return to non-defaulted status and this additional recovery cash should be discounted.
This approach is different from the approach normally used by the banks to discount these recoveries analogously to the other cash flows. Therefore a clarification on this point is requested.
• We do not support the mandated use of benchmarks for use in a bank’s internal risk management processes.
• As the EBA will be aware, the Basel Committee has mandated a fixed grading structure for the publication of Pillar 3 disclosure. Thus banks are now required to map internal grading systems to a benchmark grading system. We are of the opinion that this does not necessarily improve comparability.
• Our view is that the number of pools and grades should reflect the ability of banks to rank-order risk and should remain so.
• The approach to taking into account economic conditions will vary for each portfolio depending on the availability of historical data and the extent to which the portfolio was affected by the downturn. Banks find it a particular challenge to assess the parameters for low default portfolios.
• The number of grades is determined by each bank to take into account its risk management practices. Some banks have grades and descriptions that mirror external ratings in order to assist with rank-ordering and calibration. At the risk-grade level, banks seek to have a grading system in which the TTC PDs exhibit a high degree of stability over the credit cycle and a smoothness of change over time, disturbed only by estimation errors.
• The definition of the long-term average of default rates includes an observed downturn in the economic cycle or if this has not occurred or because the bank is new and the downturn occurred prior to its data series, then as set out in our response to Q5.3. Banks simulate and adjust data to include a downturn. Most banks that use TTC PDs seek to reflect a firm’s long-term credit risk trend to filter out cyclical effects.
• For the purposes of modelling PD our members consider that a default could occur at any time within a year. Some short-term contracts are rolled over and if this is the case then this is also taken into account.
• Our members - where permitted by the regulation - utilise the maturity adjustment - to reduce the Risk Weighted Asset value.
• However, there is no common way of describing the “PiT-ness” of a rating system
• We note that the Bank of England has recently proposed an approach to measure the PiT-ness in its Consultation Paper CP29/16 Residential mortgage risk weights. The EBA may wish to consider the appropriateness of the Bank’s proposals.
• As regards the model philosophy, we support continued flexibility in modelling Point-in-Time (PiT) and Through-the-Cycle (TTC) practices; in fact, the directive and regulations allows different possibilities.
• Banks that have a structured way of determining the rating philosophy should have the capacity to determine the most appropriate modelling choice. We think that the EBA should strike the right balance between reducing variability and allowing a certain degree of methodological freedom of choice (PiT, TTC) and that it should be the competent authority that should have the ultimate decision as to which approach or degree of PiT-ness is appropriate for the models within its oversight.
• It is important to distinguish between rating philosophy and calibration philosophy. A bank may have a PiT rating and a TTC calibration in place; the type of approach (customer versus product) also deserves consideration.
• The frequency of calibration should be set according to the relevance of the model and the changes made.
We note that in the IRTF Final Report there is a wide range of practises:
• 66.7% for LDPs, 62.5% for other non-retail, and 47.6% for retail reported having PDs that are TTC.
• However, 79.2% for LDPs, 87.5% for other non-retail, and 81% for retail portfolios reported having either a hybrid or a PIT rating.
• We support the establishment of a consistent approach to the categorisation of the MoC, while noting there may be difficulties in separating MoC applied to a model into the various categories.
• We agree with the definitions set out in categories A, B and D, but do not support the inclusion of category C. The two examples given are addressed at the time a model is developed, calibrated, finalised, reviewed and approved for use.
The proposal requires banks to quantify ‘the estimation error that results from the identified deficiency in order to justify the level of MoC at least for every calibration segment for categories A, B and D’. The proposals state that where ‘more than one trigger occurs, a higher aggregate MoC should be applied’. We understand this will require the calculation of RWA and EL before and after the adjustment of each category of MoC, and in aggregate. This would be operationally too onerous for banks to implement and maintain.
• Furthermore, given the human judgment in model development and application of risk parameters it is our opinion that there may already be MoC within the modelled parameters.
• We think that the EBA may not have considered the complexity of quantifying the offsetting changes to the estimations of PD and LGD.
• We think that documentation requirements set out in section 4.4.4 is sufficient without the need for detailed quantification of the impact of each MoC.
The suggestion is that the MoC adjustments will ‘result in a more accurate estimate of the risk parameter, where this adjustment can have both positive and negative effect on the risk parameter’. We believe this is difficult to prove and may result in the introduction of unjustified RWA variance. We think that it is unlikely that the outcome will ever have a positive effect on the risk parameter as suggested in the proposal.
We urge the EBA to bear in mind that a MoC should only be applied until the errors are fixed. The MoC is meant not to be permanent but an interim step. We recommend that the EBA guidelines be clearer on the fundamental reason and purpose of the MoC.
In conclusion we do not support the estimation of the impact on the MoC for each category. We think that an overall estimation is sufficient.
Question 1: Do you agree with the proposed requirement with regard to the application of appropriate adjustments and margin of conservatism? Do you have any operational concern with respect to the proposed categorization?
• The number of models requiring redevelopment and competent authority approval cannot be determined on the basis of this guideline alone.• It is expected that most, if not all, models will require rebuild and approval as a result of the IRB repair programme which will also need to take into account local competent authority changes to methodology for example the UK PRA CP29/16 proposals for residential mortgage risk weights.
• We think that this question should be assessed in the light of the quantitative impact study that the EBA is currently conducting.
• Nevertheless, some indications are that:
o The LGD in-default guidelines could have a material impact;
o ELBE could be potentially material in many banks especially if the indirect approach is used; and
o In terms of PD and LGD, the methodological impact is limited but the operational cost of enhancements to documentation, justifications and changes to processes is significant.
Question 2: Do you see any operational limitations with respect to the monitoring requirement proposed in paragraph 53?
We note that the EBA CP does not include any questions about the impact on the implementation of the proposals. We have included comments in our covering letter that are repeated below:Timeline for implementation
With regard to the proposed deadline for implementation of end-2020, we think that this date should only be fixed after the EBA has:
1. Had opportunity to review the comments sent in response to the CP;
2. Assessed the results of the qualitative survey to understand the impact of the proposed requirements on the rating systems; and
3. Received feedback on the yet to be published draft RTS on the nature, severity and duration of economic downturn to be developed in accordance with Article 181(3)(a) of Regulation (EU) 575/2013. As these RTS will be closely related to the estimation of downturn LGD some additional changes may be introduced in the final guidelines on the basis of the feedback received during these consultations.
We are particularly concerned about the impact of the change on competent authorities’ ability to review and approve revised models to enable a consistent implementation date.
We note that the Basel Committee is soon expected to publish its guidance on the use of IRB models. This may also have an impact upon the priority that institutions assign for the redevelopment of IRB models.
Our view is that the concerns expressed by the EBA with regard to the variability and comparability of the risk weights do not apply uniformly across all portfolios across the EU. So competent authorities should be given discretion to determine the priority for the redevelopment of models.
We suggest that a prudent objective may be to prioritise a target of c. 75% of IRB risk weighted assets being compliant by a certain date instead of a full implementation for all IRB model by end-2020.
Taking into account the scope of the changes, we urge the EBA to take a proportionate approach in the implementation of the changes to ease the burden on institutions and competent authorities.
Operational impacts
• We do not foresee any operational limitations.
• In general, one-year default rates are already being calculated at least on a quarterly basis.
• We agree with the requirement to calculate the default rates at least quarterly basis.
• However, it could be an issue for Low Default Portfolios (LDP).
Question 3: Do you agree with the proposed policy for calculating observed average default rates? How do you treat short term contracts in this regard?
• The key point is the consistency between EaD and LGD.• As far as Retail exposures are concerned, we agree with the proposed treatment of additional drawings for both CCF and LGD. Our members’ models are aligned with CRR and EBA guidelines.
• The approach concerning fees and interest is very important for an appropriate LGD computation. All the fees are considered in the economic loss as well as all the other direct costs. On the other side the interests can be further divided in two categories:
• Contractual interest: These interests have not been considered in the loss rate computation since their inclusion would result in a double counting with respect to the discounting process (whose section is separately treated in the guidelines). In case of perfect alignment between contractual interest rate and discounting rate this treatment would not determine any distortion in the loss rate computation. Nevertheless banks tend to apply a current rates approach for the discounting process as also suggested by the BCBS Working Paper 14: “their use allows the consideration of all available information and facilitates the comparison between LGD estimates from different portfolios”. This approach can determine negative loss rates for the different consideration of the value of money over time. More specific comments on this topic should be provided in the dedicated section of the guidelines;
• Unpaid late fees: These interests are included in the exposure of the denominator of the loss rate. But the guidelines ask banks to consider that, in case of recovery of late interest that have not been previously capitalised, the moment of recovery should be considered a moment of capitalisation. Does this lead to excluding the receipt of unpaid late fees interests and exceeding the amount included in the EAD for the loss rate computation? If the EBA thinks so then we do not agree with the proposal. Our opinion is that a receipt of unpaid late fees should not distort the economic loss estimation. Our opinion is that all the receipts should be considered without any specific treatment for the case of unpaid late fees.
A question for further discussion:
Article 115 states that additional recovery cash flows should be added to the calculation at the date of the return to non-defaulted status in the amount that was outstanding at the date of the return to non-defaulted status and this additional recovery cash should be discounted.
This approach is different from the approach normally used by the banks to discount these recoveries analogously to the other cash flows. Therefore a clarification on this point is requested.
Question 4: Are the requirements on determining the relevant historical observation periods sufficiently clear? Which adjustments (downward or upward), and due to which reasons, are currently applied to the average of observed default rates in order to estimate the long-run average default rate? If possible, please order those adjustments by materiality in terms of RWA.
• No.• We do not support the mandated use of benchmarks for use in a bank’s internal risk management processes.
• As the EBA will be aware, the Basel Committee has mandated a fixed grading structure for the publication of Pillar 3 disclosure. Thus banks are now required to map internal grading systems to a benchmark grading system. We are of the opinion that this does not necessarily improve comparability.
• Our view is that the number of pools and grades should reflect the ability of banks to rank-order risk and should remain so.
Question 5: How do you take economic conditions into account in the design of your rating systems, in particular in terms of: d. definition of risk drivers, e. definition of the number of grades f. definition of the long-run average of default rates?
• Introduction: The IIF RWA Task Force (IRTF) reported that an initial (rating) model may perform as a PiT, TTC or hybrid depending on the factors taken into account or forecasted. It noted that discriminating between systemic and idiosyncratic risk at the obligor level is very difficult.• The approach to taking into account economic conditions will vary for each portfolio depending on the availability of historical data and the extent to which the portfolio was affected by the downturn. Banks find it a particular challenge to assess the parameters for low default portfolios.
• The number of grades is determined by each bank to take into account its risk management practices. Some banks have grades and descriptions that mirror external ratings in order to assist with rank-ordering and calibration. At the risk-grade level, banks seek to have a grading system in which the TTC PDs exhibit a high degree of stability over the credit cycle and a smoothness of change over time, disturbed only by estimation errors.
• The definition of the long-term average of default rates includes an observed downturn in the economic cycle or if this has not occurred or because the bank is new and the downturn occurred prior to its data series, then as set out in our response to Q5.3. Banks simulate and adjust data to include a downturn. Most banks that use TTC PDs seek to reflect a firm’s long-term credit risk trend to filter out cyclical effects.
Question 6: Do you have processes in place to monitor the rating philosophy over time? If yes, please describe them.
• Yes we agree with the two alternate approaches. We recognise that each could be biased.• For the purposes of modelling PD our members consider that a default could occur at any time within a year. Some short-term contracts are rolled over and if this is the case then this is also taken into account.
• Our members - where permitted by the regulation - utilise the maturity adjustment - to reduce the Risk Weighted Asset value.
Question 7: Do you have different rating philosophy approaches to different types of exposures? If yes, please describe them.
• Banks vary in their rating philosophies, which can be measured by analyzing migration matrices. The higher the “average migration drift”, the higher the “PiT-ness” of the rating system• However, there is no common way of describing the “PiT-ness” of a rating system
• We note that the Bank of England has recently proposed an approach to measure the PiT-ness in its Consultation Paper CP29/16 Residential mortgage risk weights. The EBA may wish to consider the appropriateness of the Bank’s proposals.
Question 8: Would you expect that benchmarks for number of pools and grades and maximum PD levels (e.g. for exposures that are not sensitive to the economic cycle) could reduce unjustified variability?
We support the comments made by the EBF in its response to this CP as follows:• As regards the model philosophy, we support continued flexibility in modelling Point-in-Time (PiT) and Through-the-Cycle (TTC) practices; in fact, the directive and regulations allows different possibilities.
• Banks that have a structured way of determining the rating philosophy should have the capacity to determine the most appropriate modelling choice. We think that the EBA should strike the right balance between reducing variability and allowing a certain degree of methodological freedom of choice (PiT, TTC) and that it should be the competent authority that should have the ultimate decision as to which approach or degree of PiT-ness is appropriate for the models within its oversight.
• It is important to distinguish between rating philosophy and calibration philosophy. A bank may have a PiT rating and a TTC calibration in place; the type of approach (customer versus product) also deserves consideration.
• The frequency of calibration should be set according to the relevance of the model and the changes made.
We note that in the IRTF Final Report there is a wide range of practises:
• 66.7% for LDPs, 62.5% for other non-retail, and 47.6% for retail reported having PDs that are TTC.
• However, 79.2% for LDPs, 87.5% for other non-retail, and 81% for retail portfolios reported having either a hybrid or a PIT rating.
Question 9: Do you agree with the proposed principles for the assessment of the representativeness of data?
• We are encouraged to read: “the Guidelines do not prescribe any specific method for the quantification of MoC as the appropriate approach will depend on the character of the deficiency and the available data”. However, we are concerned that competent authorities may interpret this as a requirement for banks to have a MoC for each category in each model, and that Annex IV becomes more of an example. We welcome the EBA’s agreement at the open hearing to consider the need for further clarity.• We support the establishment of a consistent approach to the categorisation of the MoC, while noting there may be difficulties in separating MoC applied to a model into the various categories.
• We agree with the definitions set out in categories A, B and D, but do not support the inclusion of category C. The two examples given are addressed at the time a model is developed, calibrated, finalised, reviewed and approved for use.
The proposal requires banks to quantify ‘the estimation error that results from the identified deficiency in order to justify the level of MoC at least for every calibration segment for categories A, B and D’. The proposals state that where ‘more than one trigger occurs, a higher aggregate MoC should be applied’. We understand this will require the calculation of RWA and EL before and after the adjustment of each category of MoC, and in aggregate. This would be operationally too onerous for banks to implement and maintain.
• Furthermore, given the human judgment in model development and application of risk parameters it is our opinion that there may already be MoC within the modelled parameters.
• We think that the EBA may not have considered the complexity of quantifying the offsetting changes to the estimations of PD and LGD.
• We think that documentation requirements set out in section 4.4.4 is sufficient without the need for detailed quantification of the impact of each MoC.
The suggestion is that the MoC adjustments will ‘result in a more accurate estimate of the risk parameter, where this adjustment can have both positive and negative effect on the risk parameter’. We believe this is difficult to prove and may result in the introduction of unjustified RWA variance. We think that it is unlikely that the outcome will ever have a positive effect on the risk parameter as suggested in the proposal.
We urge the EBA to bear in mind that a MoC should only be applied until the errors are fixed. The MoC is meant not to be permanent but an interim step. We recommend that the EBA guidelines be clearer on the fundamental reason and purpose of the MoC.
In conclusion we do not support the estimation of the impact on the MoC for each category. We think that an overall estimation is sufficient.