The preference of the majority of our members is that default should be recognised once both components of the threshold have been breached. We recommend that the absolute limit is also balanced by a relative one. This will minimise the chance of false positive technical breaches as our expectation is that an absolute limit breach will be the triggered more frequently than the relative one.
However the technical standard should enable individual competent authorities to permit banks to continue to employ the absolute threshold only, or where this is aligned with their current business practice, lower thresholds in order to prevent unnecessary model changes which would require supervisory approval.
We therefore recommend altering the text of Article 2.2 (c) as follows:
(c) where either or both of the limits referred to in points (a) and (b) is breached, the credit obligation past due shall be considered material, and the obligor shall be considered defaulted. Competent authorities may permit institutions to employ either of the limits, or establish them for their own purposes at a lower level, where doing so would not result in fewer past due obligations being classified as material.
This is recommended on the basis that both the absolute and relative thresholds set by the competent authorities are maximum ones – see our response to question 2.
We support the €200 for retail and €500 for corporate absolute thresholds but wonder however whether it would be helpful if there were different percentage relative thresholds for corporate and retail exposures to reflect different market specificities. For instance retail exposures, and SME exposures managed as retail on a portfolio basis, could be subject to a nationally set relative threshold whereas larger corporates would be subject to a higher percentage limit which would be set at the same level across all EU jurisdictions.
Where banks are already using more conservative absolute and relative thresholds they should continue to be allowed to do so. This approach recognises that portfolios and the ways of managing portfolios are different and banks should therefore have the flexibility to use different thresholds. For example, the same threshold is not considered appropriate across such different retail product types as high value mortgages and low value overdrafts. Banks should be able to adopt a more conservative approach where they consider this appropriate.
As we have noted above these proposals represent a material change to a key input into model design. It will require significant work by our member banks to recalibrate rating models and amend the IT systems and governance processes that support them. But the supervisory community which will likely have a significant number of model changes to approve will also have significant demands placed on it.
We believe that a transitional period of less than two years will prove impossible to implement and that consideration should be given to allowing competent authorities to adopt a phased approach which focuses on approving any necessary changes to more material models first, with others following to a longer time frame of up four years. This problem will be ameliorated if banks that are currently using materiality thresholds that are more prudent than those proposed should be allowed to continue doing so, without being required to migrate to the proposed thresholds.
Furthermore we understand that the EBA will soon be making proposals in relation to the Definition of Default. This is also an aspect of the Basel Committee’s consultation on a revised standardised approach to credit risk. Our strong recommendation is that this suite of changes should be introduced in a harmonised and coherent manner, not in a piecemeal fashion spread over a period of time. This would put pressure on both our members and supervisors resources as we envisage multiple applications for model change approvals will be required, which would better be addressed all at once.
The impact assessment is a commendably thorough assessment of the current landscape in Europe in relation to materiality thresholds and of the technical options that have been proposed. We note however that there is no hard assessment neither of the costs of implementing the changes proposed in the draft RTS for banks nor for the costs to supervisors of doing so, which will ultimately also be borne by banks.
Our broad-brush assessment of the costs and benefits is that there will be a material cost for both banks and supervisors, a probable increase in capital and a somewhat more harmonised approach to the establishment of materiality of default thresholds, recognising that a degree of competent authority discretion will remain.