Austrian Federal Economic Chamber/Division Bank and Insurance
We would like to express a strong rejection of Option 1 (i.e. as soon as one of the components of threshold is breached), for non-retail portfolios a support for Option 2 (i.e. recognition of default after both thresholds are breached) and for retail portfolios a support for an absolute threshold only.
The suggested ranges for the absolute and relative thresholds are so low, that if only one threshold is breached it would capture minimal amounts of past due exposures that are no real defaults and typically have nothing to do with the deteriorating credit worthiness of the clients. We do not see it appropriate to treat past due exposures that are solely based on an absolute threshold for the non-retail segment as a default, or that are based solely on the relative threshold for the retail segment – especially not with such a low threshold as 2%.
The option of both thresholds being breached (option 2) provides balanced default recognition. As mentioned the suggested range of the absolute or the relative thresholds are already so low, that further requiring the either-or" combination is only resulting in the identification of more technical defaults. These defaults would be either negligible small amounts on its own (in retail), or negligible when compared to the total outstandings (in non-retail), so we assume that the majority of the identified default cases will be technical defaults - ie defaults that are not related to the credit standing of the customer and would only increase the administrative efforts without improving the early detection of real defaults.
In addition, for typically ‘mass-treated’ portfolios, like retail, where the implementation of the dpd-threshold should be easy to operationalize due to the large number of cases to be checked and due to completely automated processes, we suggest going for a simple definition like absolute threshold only."
We agree on the maximum level of the absolute threshold, but suggest raising the maximum amount of the relative threshold. The maximum 2% threshold is very restrictive and would not identify real default cases. We consider it appropriate that it is raised to 4% of the total amount of all credit obligations of the borrower.
However, we believe that for portfolios characterised by a large number of individual files, such as retail portfolios, the threshold should be easy to implement due to the burden represented by the numerous checks. Therefore, for such cases banks may have the option to fix absolute thresholds, in the spirit of the RTS and of the internal risk management processes of the institution.
We regard as the shortest realistic time horizon to implement such changes in 7 years.
Banks have to base their recalibration of risk parameters upon a five year data period. It would be extremely burdensome or possibly impossible for some of them to search through the mass data of their transactions from 2010 on to identify past due periods according to the new materiality thresholds. All accounts of a costumer must be included and checked against the advised limit at that day.
Furthermore, even if a bank had overcome those difficulties the purely statistical recalculation would not provide a correct picture. Banks have implemented procedures to remind customers being in danger to breach their limit. These procedures cannot be simulated in retrospect and data cannot be adjusted to the changed criteria.
Supervisory authorities have to accept the recalibrated risk parameters before a bank is allowed to use them. Because banks cannot control the necessary time for supervisory review and approval the transition period should refer to the date of application by the bank and a preliminary approval should be stipulated until the definite decision. Otherwise a bank would have to step back to the standardised approach when the authorities don’t manage to work through all concurrent applications in time.