Response to consultation on Guidelines on proportionate retail diversification methods
Q1. What is the percentage of exposures within your retail portfolio that are part of a group of connected clients?
ABI does not answer this question, as the abovementioned information can only be provided by banks. Anyway, we take the occasion of this question, regarding a basic element of the diversification test, to highlight the need to clarify other basic elements, on how the proposed diversification test should be performed and on the scope of the assessment.
A key aspect on which clarification is needed is the definition of “exposure value” that shall be used for the test. Indeed, the consultation paper clarifies certain aspects, such as the fact that the exposure value shall be considered at the level of the group of connected clients. Anyway, other elements are not mentioned, e.g. whether the exposure value shall be considered gross or net of provisions and of the effects of credit risk mitigation. Also the use, in the Explanatory box (page 7), of “EAD” as symbol for the exposure value appears ambiguous, as it might be read as meaning that the exposure value shall be determined as EAD.
In this regard, ABI would recall that, for the same purpose of the classification in the retail portfolio, banks already determine an exposure value that shall be compared with the absolute threshold of 1 million euro pursuant to Article 123 (1)(b).
ABI would argue that using the same definition of exposure value used for the latter, also for the purpose of the diversification test, would represent the best suited solution in terms of consistency of the framework but also with a view to proportionality (cost efficiency and reduction of complexity).
As regards the scope of the retail exposures to be assessed, ABI would seek confirmation that the diversification test shall consider all the exposures potentially classified as regulatory retail, irrespective of the applicable risk weight, i.e. also exposures qualifying as “transactor exposures”, and loans granted to pensioners or employees as per Article 123(5), shall be included in the “original portfolio” (and shall be assessed against the 0.2% threshold).
Q2. Do you identify any implementation issue in implementing the diversification method?
As regards the proposed diversification method, ABI would observe that it appears complex and might be operationally burdensome, depending on the number of exposures and composition of the original portfolio. In this regard, the alternative approach appears preferable, as it only requires the calculation to be performed once.
Anyway, such approach appears restrictive, and ABI would suggest carefully assessing the impact of its application (once defined the exposure value that shall be used), and possibly reconsidering the level of threshold - from the currently envisaged 5% - in order to ensure that it does not result in a material impact that would alter the overall impact of the new regulations decided by the legislators. In this regard, it is worth noting that the stylised example in the Consultation Paper shows that, for certain portfolio structures, the 5% threshold would determine a very huge impact (in the example, around 35% of the overall amount would be disqualified from the retail portfolio). Consideration of the impact is all the more important given that, in case of a significant increase in the cost of regulatory capital, the affected clients would be represented by households and SMEs.
As to implementation challenges, it has also to be noted that both the proposed approaches would result in uncertainty of exposures classification and risk weight, given that:
- the classification of an exposure would not always be predictable ex ante (at the moment of the loan origination), as the actual possibility to assign it to the retail portfolio will depend on the composition of the portfolio at the moment of the next assessment
- the classification of an exposure could change over time, because the draft Guidelines require the test to be performed at each reference date (e.g. a rise in the size of the retail portfolio before exclusions would increase the 0.2 threshold accordingly, and exposures might become eligible that were not eligible at the previous assessment).
Among other things, clarification would be needed about the interplay of such uncertainty with the provisions of the EBA Guidelines on loan origination and monitoring (EBA/GL/2020/06, LOM), that require the portfolio classification and the cost of the regulatory capital to be taken into account in the pricing of loans. More precisely, ABI would suggest clarifying that it is sufficient that the pricing considers the classification expected at the moment of the loan origination, and such pricing remains LOM compliant even in case of changes in consequence of the application of the retail diversification method.
Last but not least, ABI would highlight that the request to perform the assessment at each level for which the own funds requirements are calculated (i.e. both at individual and consolidated level) makes complexity and operational burden considerably higher. The double calculation would imply that the same exposure might be eligible for retail classification or not depending on the level of the assessment. In ABI’s view, banks should be required to perform the calculation only once, at consolidated level (at legal entity level for institutions that are not part of banking groups). This solution appears very much preferable both in terms of cost efficiency (which is needed to ensure that the proposed approach remains proportionate, as required by the CRR) and correct representation of risk (provided that adequate diversification is achieved at consolidated level, an exposure does not become more or less risky depending on the size of the portfolio of the specific subsidiary where the loan is booked).
Q3. Which methods do you currently use to assess retail diversification? Please elaborate.
No comment.
Q4. Under the proposed approach, in the first step of the calculation before any exclusion, what is the share in terms of exposure value of the large eligible retail exposures as defined under the proposed approach compared to all the eligible retail exposures?
No comment.