Publication date: 9 February 2017

The EBF, on behalf of its member banks, welcomes the opportunity to comment on the EBA’s consultation on Guidelines on PD estimation, LGD estimation and the treatment of defaulted exposures (EBA/CP/2016/21). Additionally, we thank the EBA for its interim engagement with the Industry through the Public Hearing held on the 19th of January of 2017. Our member banks found that these discussions were helpful in clarifying aspects of the intentions of the Guidelines. Notwithstanding, this meeting represented only initial feedback and this letter provides the considered response of the EBF and its members. This feedback will both reiterate and build upon the discussions at the Public Hearing. We believe that this feedback is itself more considered in consequence.

The EBF supports the objective of the draft Guidelines, to reduce unjustified variability in the IRB Approach. We understand that the EBA considers these clarifications and harmonisation necessary in order to achieve comparability of risk parameters estimated on the basis of internal models and to restore trust in these models by market participants while at the same time preserving risk sensitivity of capital requirements.

One important observation from the Public Hearing was the depth of analysis and consideration given to the draft Guidelines prior to their publication. It is clear that alternatives have been considered. It is also clear that choices have been made to develop and progress the necessary harmonisation in a considered manner. We ask, however, that consideration is provided to describing more clearly the choices afforded to banks in modelling and the role of the supervisor in evaluating or providing constraints on these choices. Already, individual supervisors are imposing the draft Guidelines as prerequisites to model and material change proposals often with strongly prescriptive interpretations that our member banks believe are contrary to those intended. Without greater detail on the context of the Guidelines, there is a significant risk that they might inadvertently promote greater divergence in modelling outcomes for similar risks. Much of our feedback, seeks to draw out this clarity within the finalised Guidelines.

We understand that the harmonisation of internal modelling practices requires a thorough analysis of current practices. Nevertheless, we would also like to draw the attention of the EBA to the aspect of cost involved. The many triggers introduced for re-development, reestimation and re-calibration of internal models indirectly raises questions about how often this should be made and the burden to constantly recalibrate. The aspect of cost should be taken into account and eventually consider annual reviews instead of constant recalibrations. The results of the impact assessment currently underway should be carefully examined in order to factor in the cost of changes as a relevant variable. In allcases, the objective of MoC, of re-calibrations and of re-developments should be to ensure the adequacy of regulatory capital and not be used to require model optimisation by a supervisor. The optimality of a model should be a choice for an individual bank and here cost versus benefits will be considered.

One possibility to limit the impact on cost would be to apply the Guidelines only to future models. If this is not the case, then we would ask for adequate phase-in periods in order to avoid redeveloping a huge number of models. Some sense of proportionality should be introduced in the final version of the Guidelines or we could end up in a situation where virtually all models will need to be reworked.

As regards the timeline for implementation we would like to draw to EBA’s attention the fact that many concurrent changes are being published about internal models (e.g. “RTS on IRB assessment methodology”, “RTS on materiality threshold for past due exposures”, “GL on the application of the definition of default”, “RTS on the nature, severity and duration of economic downturn”). Somehow these Guidelines wrap up many of those changes thus there should be a comprehensive understanding of them all before implementation. The date of application from 1 January 2021 seems to be challenging.

In order to achieve reduction in the unjustified variability it is important that the guidelines are adhered to by competent authorities without imposing stricter requirements. This is, obviously, an implementation issue, but we want to highlight it for the sake of harmonisation.

In summary, the EBF supports the technically sound and well-informed approach that the EBA is following. The use of models is central to the risk management practices and it deserves an in-depth review like the one proposed in these series of EBA guidelines.

This EBF response paper addresses the specific questions raised at the end of the EBA consultative paper after some broader initial questions to consider.