Response to consultation on Guidelines for the estimation of LGD appropriate for an economic downturn
Go back
In this case, a simplified procedure can make sense. An add-on could be derived from the calibration of the model for performing exposures, which is transferred to the default portfolio. Validation would ensure that the add-on used is appropriate. Such a procedure should be expressly authorised in the Guidelines.
It should be borne in mind that this huge effort serves only to determine a relevant downturn phase. The actual methodological modelling work, i.e. the modelling/derivation of an LGD downturn model which explains the analysed effects, as well as the validation of the results and the comparison of the results with the reference value, involve an additional large amount of work for the institutions.
The lion’s share of the work carried out has no relevance at all for the modelling results, for example the analytical effort for downturn periods which do not lead to the maximum downturn LGD. Limiting the analyses to relevant downturn periods should be expressly authorised in the Guidelines.
By far the most important indicator for a relevant downturn period is a peak in the observed loss rates (in so far as the higher values are not attributable to a so-called “sonar effect” from the realisation processes). If no peak can be identified in the observed loss rates, it can be assumed that no relevant downturn phase exists. A criterion for the need to consider a downturn period could be expressed as follows:
Firstly, the institutions examine the time in default to determine a “maximum realisation period” (for example by means of a definition based on a specific percentile of the distribution of the realisation period). In order now to explain an observed peak in the observed loss rates, the institutions examine from the time of this peak in the loss rate those downturn periods not exceeding the “maximum recovery period” before the occurrence of this peak. Earlier downturn periods, as well as periods after the occurrence of the peak, cannot be responsible for the observed peak.
Downturn periods, for which the influence on loss rates can be examined only using methods in accordance with Section 7, are not relevant either, in so far as there are further downturn periods for which the influence can be investigated using methods in accordance with Section 5 or 6. For the final downturn LGD, the results of methods in accordance with Section 7 should not be used according to the specifications, and the margin of conservatism (MoC) can also be determined otherwise. An exemption of these downturn periods from the analysis requirement should therefore be expressly authorised in the Guidelines.
A distributional approach is therefore preferred. The objection concerning the variance of the time series, which possibly does not yet contain an economic downturn phase, can be taken into consideration for example by choosing a correspondingly high percentile.
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?
A large number of calibration segments may come into being especially for models in which the portfolio to determine the LGD in-default is divided into different pools derived from the internal realisation processes (e.g. according to time in default, step in the realisation process, type of product). The proposed approach would in these cases call for considerable effort – irrespective of whether or not the segments can be expected to react differently to macroeconomic shocks.In this case, a simplified procedure can make sense. An add-on could be derived from the calibration of the model for performing exposures, which is transferred to the default portfolio. Validation would ensure that the add-on used is appropriate. Such a procedure should be expressly authorised in the Guidelines.
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?
The requirement to examine each individual downturn period for each macroeconomic factor analysed, for each individual calibration segment or even for granular sub-segments, gives rise to an immense effort for the institutions. The analyses would have to be carried out at four different levels (observed LGDs, annual realisation cash flows, period in default and probability of derecognition) and would have to include the various time lag effects over a horizon of the average realisation period.It should be borne in mind that this huge effort serves only to determine a relevant downturn phase. The actual methodological modelling work, i.e. the modelling/derivation of an LGD downturn model which explains the analysed effects, as well as the validation of the results and the comparison of the results with the reference value, involve an additional large amount of work for the institutions.
The lion’s share of the work carried out has no relevance at all for the modelling results, for example the analytical effort for downturn periods which do not lead to the maximum downturn LGD. Limiting the analyses to relevant downturn periods should be expressly authorised in the Guidelines.
By far the most important indicator for a relevant downturn period is a peak in the observed loss rates (in so far as the higher values are not attributable to a so-called “sonar effect” from the realisation processes). If no peak can be identified in the observed loss rates, it can be assumed that no relevant downturn phase exists. A criterion for the need to consider a downturn period could be expressed as follows:
Firstly, the institutions examine the time in default to determine a “maximum realisation period” (for example by means of a definition based on a specific percentile of the distribution of the realisation period). In order now to explain an observed peak in the observed loss rates, the institutions examine from the time of this peak in the loss rate those downturn periods not exceeding the “maximum recovery period” before the occurrence of this peak. Earlier downturn periods, as well as periods after the occurrence of the peak, cannot be responsible for the observed peak.
Downturn periods, for which the influence on loss rates can be examined only using methods in accordance with Section 7, are not relevant either, in so far as there are further downturn periods for which the influence can be investigated using methods in accordance with Section 5 or 6. For the final downturn LGD, the results of methods in accordance with Section 7 should not be used according to the specifications, and the margin of conservatism (MoC) can also be determined otherwise. An exemption of these downturn periods from the analysis requirement should therefore be expressly authorised in the Guidelines.
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?
See reply to question 1.Question 4: Do you consider the description of the approaches to be sufficiently clear?
No. It is not clear whether it is possible to use external time series from observed values (for example, from observed realisation proceeds) within the approaches described in Section 5 or 6.Question 6: Do you expect that the total exposure amount or share which is treated with the policy proposed in Section 7 is material?
From the present perspective, for Bausparkassen portfolios, an approach at least in accordance with the extrapolation approach seems possible.Question 7: Do you have specific examples of types of exposures which will fall under the policy proposed in Section 7?
No.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?
The proposed MoC is in no way risk-sensitive and affects all portfolios equally, irrespective of the type of product, the collateral or other decisive characteristics of the portfolios. A minimum add-on of 20% is disproportionate for a retail portfolio consisting of mortgage financing secured by real estate and is unjustifiable.A distributional approach is therefore preferred. The objection concerning the variance of the time series, which possibly does not yet contain an economic downturn phase, can be taken into consideration for example by choosing a correspondingly high percentile.