EBF response to the EBA consultation on IRB requirements

13 March 2015

This RTS when finalised, will be a critical document setting out the direction and role of rating systems within banks’ risk management and capital management frameworks. Its importance and impact should not be underestimated. For that purpose, the EBF provides extensive comments in this response.

EBF stated these general remarks:

• This consultation paper attempts to create a uniform interpretation of supervisory assessment methodology and to ensure consistency in internal model outputs and comparability of the risk-weighted exposure amounts across EU member states; however, the consultation paper needs further clarification and detail on a number of points to achieve this.


• A Uniform and consistent interpretation of the IRB approach within the single rulebook is especially important for cross border institutions. Operationalising different rules in each country is difficult, costly and increases regulatory complexity (e.g. different supervisory practices for parameter estimation; reporting requirements; supervisory reference models to the parameters in IRB, among others). Therefore it is important that this consultation paper does not give mandate to further increased general supervisory judgements, such as add-ons or the use requirement, which prevent further harmonisation.


• The proposed text fails to make distinction between the roles of the different lines of defence, including the supervisory regime. It would appear that each line of defence is expected to replicate the role of the one beneath it. In particular the supervisory review ails to leverage the considerable investment by institutions in model validation and suggests an intensive supervisory review of all initial model proposals and subsequent changes. This will require a considerable investment in resources by the Competent Authorities with little supervisory benefit.


• The cost of compliance implied by the draft RTS is considerable. The proposed supervisory review fails to acknowledge that institutions are naturally incentivised to ensure that rating systems discriminate risk and therefore the focus of the supervisory review should not be to ensure model optimality but to ensure that capital distortions are kept to a minimum. Institutions, already incentivised to discriminate risk, are best placed to evaluate the cost and benefits of including or excluding risk factors. However EBF would support an exception where a systemic risk factor is identified for a group of institutions by a National Competent Authority.


• The framework is ill-suited to an environment where risk management of an institution is constantly developing and improving with consequent changes to an institution’s rating systems. The requirement for near-forensic examination by supervisors of model changes is a disincentive to investment in risk management and places an improper responsibility upon supervisors to examine all aspects of the proposed model or change.


• The objectives of the supervisory review are often left unstated and without clear priorities. A clearly-stated objective should be that capital requirements should not be distorted, as distinct from ensuring the optimality of rating systems.


• The framework repeatedly confuses economic cycle with long run average and these in turn with an observed observation period. Careful attention to these interrelated but distinct concepts would assist in ensuring that the framework is clear.


• A very significant concern is that the framework fails to acknowledge the recent economic experience of many Member States which have experienced significant market volatility. In particular, the proposal to exclude estimated data from estimates of the long run average default rate (Article 49(3)) may have significant macroeconomic consequences such as impacting lending capacity within many Member States and present a significant barrier to entry within the Single Market.


• EBF considers that the requirement established in article 49(3)(b) in order to determine the PD when the observed period does not cover a full cycle is against the Through-the-Cycle (TTC) philosophy of the Basel agreement.


• The proposed RTS fails to clarify the desired regulatory treatment for defaulted assets, a noted source of divergence and discretion between jurisdictions and institutions.


• No time limits for supervisory reviews are specified. Untimely supervisory decisions are a disincentive to improving rating systems with implications for the risk management of institutions.


• There is also a separate significant and demanding requirement for supervisory benchmarking (CRR Article 78). No reference to this as a supervisory tool is made within this proposed RTS.

It should be noted that banks have made considerable investments in the control frameworks surrounding the use of internal models within their independent model validation and separate risk-based audit activities. Banks are also naturally incentivised to discriminate between obligors on the basis of credit quality. The key supervisory concern therefore should be to ensure that the overall assessment and measurement of risk by banks is appropriate and consistent. The distinction, therefore, between model discrimination and model calibration should be made with supervisors concentrating efforts on the latter. This should additionally facilitate a more rapid supervisory decision process with real impact in ensuring that banks are correctly and appropriately assessing risks with a strong forward bias. In this respect, EBF would call for:

• A clearly stated strategy and focus for the supervisory review. This would delegate criteria of lesser importance fully to the control frameworks of banks. These latter would remain subject to supervisory review but on a risk-based assessment basis. The more important consideration of the risk assessment should form the core of the supervisory review.


• A clear exposition of the requirements for Long-Run Average PDs and the close but distinct relationship with the Economic Cycle. This must incorporate the legitimate use of estimated data within an appropriate framework so that banks can reflect new lending strategies within their risk assessments.


• A clear exposition of the requirements for downturn and long run estimates for LGD and CCF measures, to ensure consistency and again to ensure that the risk assessment reflects the lending strategies of banks which may be new.


• Clearly stated time limits for supervisory review.


• Elaboration of the principle of proportionality to provide an effective tool for the classification of banks and the consequent determination of supervisory expectations to ensure consistency across jurisdictions.


• Elaboration of clear expectations for independent model validation and for banks’ internal audit functions commensurate with roles appropriate to these functions.


• The opportunity to clarify the treatment of defaulted assets should be taken.

Full response

© EBF