Austrian Federal Economic Chamber, Division Bank and Insurance
Yes, we agree with the proposed definition of technical default.
Yes, we belive other situations should be included in this definition. However, when an IRB institution needs to make a change to the definition of default it should include a margin of conservatism due to inconsistency in historic data. If this inconsistency was due to technical defaults not being defined so far and only being defined now, we would recommend this to be explicitly excluded from the reasons for application of the margin of conservatism.
We have some concerns that not all non-financial circumstances such as commercial disputes are resolved within the first 90 days. Ideally, some expert judgment should be introduced in determining the technical defaults or for exceptional circumstances. In particular public sector entities often delay payments, which is not associated with default. Considering technical defaults only due to personal or system errors could be realistic.
Moreover, the definition of technical default is in a way too restrictive: all non-payments not due to credit reasons are considered defaults with the exception of the cases under para. 20 (a), (b) of the draft GL.
This would imply in many cases major changes from the current practice of non-retail activities where no definition is given and where expert judgment is used to determine whether a past due status relates to a technical default or not.Thus, non-consideration of technical defaults should be left as an option available to financial institutions.
If the Past Due Criterion is not caused by a lack of creditworthiness of the customer but rather caused by a mistake it should not be considered as default in accordance with Art 178 CRR.
The definition is sufficiently clear.
However, the differentiation made on page 7 of the Consultation Paper (Point 3.2.3 Factoring, para 1, 2nd sentence) is according to Austrian law not valid. We therefore suggest that this differentiation is optional but not mandatory.
Concerning Point 3.2.3 Factoring, para 2, we suggest that the timeframe shall be prolonged as usual for working capital loans, i.e. limit breach + 90 days. Here we also suggest that the receivable due date should be prolonged to + 90 days.
General comment on factoring: According to Austrian local GAP all factoring exposures (recourse and non recourse) are booked on the debtor risk and should therefore be treated as purchased receivables and the counting of days past due should commences when the payment for a single receivable becomes due.
In general we agree with the harmonisation of the definition of default between IFRS 9 and the regulatory definition.
But it is of utmost importance that the new default definition is not applied before mandatory date of IFRS 9. This would avoid for banks to adapt the new definition of default to current accounting practices and shortly after to the revised IFRS 9 principle.
Generally, Stage 3 of IFRS 9 includes exposures that are credit-impaired (e.g. in case of significant financial difficulties of the debtor, breach of contract, etc.). However, the condition that all exposures classified as ‘Stage 3’ under IFRS 9 are treated as defaulted should be applied automatically. Otherwise it may prevent from reclassifying in performing status some asset typologies, which according to the accounting principles remain in Stage 3 also after having being restructured. Hence, we would suggest that the EBA’s guidelines embed a Stage 3 definition avoid an automatism between Stage 3 and default.
More detailed descriptions of ‘Sale of credit obligation’ in the context of Austrian banking practice would be helpful.
We are not convinced of the indicator. The sales of a credit obligation at a loss could happen for other reasons than credit risk deterioration, such as liquidity management or regulatory capital.
Also the threshold seems to be too restrictive, as a simple reduction of 5% of the nominal value would lead to a default of the debtor. This approach could have contagion effects linked to its other exposures within a bank. Moreover, recent markets’ history has showed that when financial markets are highly volatile, some bonds could be under 95% of their value only because markets expectations. For instance, even the case of a rating downgrade does not automatically mean the exposure is in default: there is still a large room to manoeuvre within the investment grade (or non-investment grade) until it reaches default.
A sale below 95% of value can easily occur on non-defaulting exposures, e.g. for long-term, low-interest assets, if the counterparty default risks cannot be estimated precisely by the contractual partner.
Basically, the effective interest rate at the moment before signing seems more appropriate – depending, however, on the composition of the interest elements that may be considered and interest elements that must not be considered.
Cash flows should be discounted with the customer’s original effective interest rate. In any case, financial institutions which do not apply an effective interest rate under the accounting standards relevant to them should be given an opportunity to use other discounting rates in line with the institutions' economic management instead of the effective interest rate.
If the contractual cash flows on a financial asset have been renegotiated or modified according to IFRS, an entity shall assess whether there has been a significant increase in the credit risk of the financial instrument. This assessment does not automatically lead the asset to be in Stage 3 or to be considered as defaulted. We believe that the default criteria should be aligned with IFRS 9 and not by setting artificial thresholds to evidence default.
Ad para 38: Under conservative interpretation the 1st forbearance measure triggers the examination obligation. Exceeding the formula threshold leads to a default with a minimum default period of 1 year. This policy differs from the current obligation to trigger a default in case that after the 1st forbearance measure 'under probation' a default only occurs after the 30-days-past-due criterion or after the implementation of the 2nd forbearance-measure. This would lead to an increase in NPL-ratios and to an inappropriate extension of default periods. We have serious concerns about the above mentioned approach and therefore believe that the current methodology should be maintained.
EGB intends to follow the modification / de-recognition standards of the IFRS 9. However, we consider the proposed limit at 1% very conservative. In addition, the rational of applying materiality thresholds for different default drivers is not clear.
As paragraph 33 prescribed a 5% threshold and paragraph 40 a 1% threshold we propose, for simplicity, to align the sale and forbearance default trigger-related materiality thresholds at 5% for both in retail exposures, and we also suggest including an absolute threshold (EUR 250 or in line with the new proposal of the EBA) and 5% for both for non-retail exposures.
In relation to the calculation of the materiality for forbearance related default definition, paragraph 40 proposes to use customers’ original effective interest rate (vs. the interest rate used in forbearance). This is technically not feasible especially in case of long term credit obligations where several interest rate changes may occur and not all of these are kept in the system. We propose to use the interest rate applicable at the time the forbearance is decided (instead of the new (forborn) interest rate).
The purchase or origination at a discount price of an asset shall not be a criterion for unlikeliness to pay. The material discount could for example be the result of negotiations or changes to market conditions, so this proposal might trigger a default for reasons that are independent from the credit risk of the counterparty.
However it is not clear how paragraph 47 (g) differs from paragraph 30 ff ‘Sale of Credit obligation’. Please provide us with further explanations and please indicate what you consider as ‘material discount’.
In general we can accept the 12 months probation period before the return from default to non-default for distressed restructuring and 3 months probation period for all other indication of default.
But should there be no probation period for a return from default to non-defaulted status instead a symmetrical treatment should be provided for exposures that are moving from non-defaulted category into default category and in the reversed scenario. F.i. in IFRS 9 an asset can be transferred form stage 3 to stage 2 based on significant decrease in credit risk without probation period. By applying a probation period, financial instruments would move into default quicker than back to non-defaulted status. This can result in exposure being classified as defaulted but not credit impaired under IFRS 9 (bucket 2 exposures) or the exposure would be classified as defaulted and credit impaired (in bucket 3) but with no loan loss allowance.
For purposes of para 59 the application of criteria in para 58 seems more appropriate than a minimum 1-year-default period: If an obligor is defaulted and should be reclassified as non-defaulted it would be better to apply the assessment provisions in para 59(a) to (c) – thus enabling a restructuring before 1 year and still meeting the CRR’s aims.
The postponement of a due instalment and/or due interest and/or due fee towards the end of the credit period is not considered as an extension of the 'grace period' towards the end of the credit period. Should such postponement be considered as 'grace period' this would indicate that the 1-year-minimum period starts at the end of the credit period.
We disagree; the credit risk may be significantly different in different consumer products. For example a credit card receivable may be past due but the loan of the same customer can be fully performing. In our opinion a pulling effect cannot be automatically assumed if 20% threshold is exceeded but definition of default at obligor level requires expert judgement.
Rather, institutions could be asked to either demonstrate that a default on a single facility does not have an impact on the probability of default for other facilities to the same counterparty.
We do not support the proposed definition of application of materiality threshold to joint credit obligations.
Credit institutions should have the opportunity to deviate in justified situations. An automatic process should therefore envisage judgments and expert opinions.
If a joint obligation towards an institution defaults, the individuals taking part in the joint obligations (and their individual obligations, respectively) should not be automatically considered as defaulted. This mechanism is even more problematic when applied to joint obligations consisting of a large number of individuals in which case considering all the individuals involved in the joint obligation automatically as defaulted may not be economically justified at all.
As the change of default definition is a material change we do not see any issues for involvement of the management body for approval of implementation of default definition and as well with the regular internal audit review. We, however, do see possible obstacles with use test.