The definition of technical default expressed in the paper is too restrictive: actually, all non-payments not due to credit reasons would be considered as defaults except for the two cases quoted in the Consultative Paper. This would represent a sizeable change compared to the current practice on non-retail activities where expert judgment is largely used to determine whether a past due situation is pertaining to technical default or not.
The approach proposed by the EBA requires a much more extensive definition of a technical default situations in non-retail activities, for example resulting in past due exposure of more than 90 days, but not due to a credit deterioration of the counterparty. Please find below a non-exhaustive list of such situations to be added in our sense to the definition of technical default:
• Commercial disputes with a customer in the case of leasing activity when the quality of the delivered product / service is contested.
• Commercial disputes in the case of factoring activity. For example, when there is a contestation, litigation or discussion on the invoice between the client and the debtor. This will often be unknown to the factor and could lead also to damaging contagion effects.
• Commercial dispute on a “Stand by letter of credit” (SBLC) would put in default a bank or a large corporate with a potential contagion effect in the case of a syndication of the SBLC (i.e.: all participating EU banks would place the corporate in default, resulting in a possible limitation of the customer’s access to credit).
• Call of suretyship where the suretyship contest the legitimacy of the call which entails a past due situation (case of unfair/abusive claim).
• Syndication arrangements: the agent takes routinely more than 90 days to pay lenders; this practice is without any link to the borrower’s current situation.
• Long administrative processes to give payment authorizations. For example, transfers of loans between Local authorities require a period of time way beyond the 90 days period. The current French “territorial reform”, following the simplification objective favoured by European authorities, will lead to numerous transfers and delays in the coming years, independent from the financial situation of the counterparties involved.
• Logistic process issues for Energy & Commodities financing or generally for Trade Finance: all reasons which lead to delivery delays. For example, merchandise blocked at the customs, prohibition on entering or leaving ports, strike actions…
• Disputes regarding the amount or the nature of collateral in case of margin calls.
• As for asset financing long term loans, amendments/waivers or consents are possible due to – for example – a lack of customer responsiveness, maintenance check of products, reality check of the financing according to new market conditions. Expert assessment is essential.
• Specific cases of sovereign counterparts, for which default may be assessed at political level.
• In general, cases of force majeure (environmental disasters, legally imposed measures, riots, strikes, wars…).
We would like to take this opportunity to underline that a too restrictive definition of technical default will only lead to a higher number of multiple defaults and will certainly create some “noise” in the default information due to additional “non-credit” related defaults. Moreover, commercial disputes are solved over long periods (often many years) which would mean that a borrower would stay in default during this period. However, we believe the proposed approach might be acceptable for an application to retail counterparties where a more mechanical stance seems adequate.
A major risk of this too restrictive definition is to decrease the pertinence of models and to reduce the incentive of the “use tests” on non-retail counterparties. This is likely to strongly increase the default rate and consequently, the allowance of credit loss (cost of risk) of the bank at constant perimeter in terms of activity and risk profile.
As already recommended by IIF RWA Task Force, it would be more risk-sensitive to keep flexibility and expert judgement for specific cases.
We appreciate that the EBA gives credence to the specificity of the factoring activity.
We welcome the EBA proposal to anticipate the implementation of IFRS 9 and to leverage as far as practicable on the accounting framework. It is indeed important to avoid inconsistencies with the accounting treatment as much as possible.
Stage 3 of IFRS 9 includes exposures that are credit-impaired (e.g.: significant financial difficulty of the obligor, breach of contract, concession granted due to financial difficulty, probable bankruptcy of the borrower etc.…). Accordingly, we agree that stage 3 exposures will be generally considered as defaulted and stage 2 as non defaulted, though we realise that there may be some exceptions.
However, while defaulted exposures would end up in stage 3 under IFRS 9, the other way around might not be true. There might be situations where some stage 3 exposures would not be defaulted, for instance where national options exist, as rightly pointed out in the consultation paper, but also due to technical defaults (please see answer to question 1 above).
Besides, while we welcome the clarification made by the EBA that the “incurred but not reported losses” (IBNR) should not be considered as an indication of unlikeliness to pay (paragraph 26 of the CP), we consider that such a clarification should encompass stage 2 of IFRS 9.
Indeed, beyond the IBNR which is a current notion of IAS 39, it should be clearly stated that the IFRS 9 stage 2 should generally not be considered as an indication of default. This is all the more important as article 178.3 b) of the CRR mentions “a significant perceived decline in credit quality” as an indication of unlikeliness to pay; this might be misleading and wrongly liken to the stage 2 of IFRS 9 (“significant increase in credit risk”).
The rule related to the sale of credit obligations seems very constraining as a mere [5%] decrease in the nominal value would induce a default of the obligor and related contagion effects on its other exposures within a banking group. This is over-restrictive, does not reflect the reality and would have the following unintended effects:
• Threshold/cliff effect: recent history (2008) has proved that when financial markets are highly volatile, some bonds could be under 95% of their par value because the markets anticipate a future decrease of the credit market without the issuer being itself in default.
• Perverse effect: the bank may cease granting facilities to the obligor who may incur an actual payment default.
• In a deleveraging period or for portfolio management purposes, this may incite the sale of good quality assets only, in order to avoid a defaulted categorization of the other assets issued by the counterparty.
• A bank may sell assets for other reason than the anticipation of a decrease in credit quality of the issuer. A decision to sell participations in loans on performing clients at a significant loss may be dictated by:
o Regulatory capital savings or employment
o Liquidity management
o Balance sheet management
o Country envelope consumption
o Counterparty exposure management
o Single limit concentration management
• A fixed threshold does not take into account the maturity of the credit obligation –the pricing impact of a credit deterioration will not be the same on an obligation if its residual maturity is 1 year or 10 years. Furthermore, the credit obligations market may not always be liquid, hence the decision of a bank to sell a significant exposure on a client may cause a pricing shift not necessarily correlated to that client's creditworthiness.
As a consequence, we consider this approach inappropriate in its spirit as it does not heed a true assessment of the credit risk of the counterparty and might induce too many perverse effects.
We deem that expert judgment should be included in this appreciation. We think that the proposed treatment could be instead considered as another criterion of unlikeliness to pay in association with other indicators.
Finally, the proposed treatment could also be incompatible with the approach proposed in the distressed restructuring chapter: a default could be triggered by a discounted sale of an obligation, while a distressed restructuring respecting the 1% rule would not trigger the default of the same counterparty.
We understand that the EBA aims here at preventing credit institutions from selling assets before they become in default in order to avoid an impact on the internal PD. We would like to propose an alternative approach to this situation in order to avoid defaulting all counterparties for which a discount is higher than 5%. We therefore suggest the EBA to replace paragraph 32 with the following:
Where however the loss on the sale of credit obligations is related to the credit quality of these obligations, in particular where the institution sells the credit obligations due to an anticipation of a certain future default, the institution should consider that the borrower is in default.
Generally speaking, we understand from paragraphs on distressed restructuring that the scope of exposures on which the diminished financial obligation has to be calculated is restricted to forborne exposures, meaning to cases in which concessions have been extended towards a debtor facing or about to face difficulties in meeting its financial commitments. More particularly, we would like the final guidelines to state explicitly that renegotiations in which concessions can be granted due to commercial reasons (as opposed to due to financial difficulties of the client) are not in the scope of these paragraphs. Those renegotiations could lead to diminished financial obligations because of market conditions and should not trigger default for the obligors.
In addition to the example in Q6, please find in the annex an example of investment grade loans quotes illustrating cases where the loan is negotiated below 95% of par without being in or close to default.
On the basis of paragraphs 38 to 41 of the proposed guidelines, we understand that:
• The notion of “forborne exposure” is equivalent to the notion of “distressed restructuring”. By definition, distressed restructuring should be considered to have occurred when forbearance measures have been extended towards a debtor as specified in the ITS on forbearance and non-performing exposures.
• However, a forborne exposure/distressed restructuring is not necessarily non-performing or defaulted; a borrower may repay in full according to the post renegotiation schedule.
• A forborne exposure/distressed restructuring should be considered as defaulted when a material diminished financial obligation is observed. In other words, it should be classified as defaulted if the diminished financial obligation exceeds a certain threshold. This threshold is set by the institution. However, the threshold should not be higher than 1%.
We believe that the cap of 1% is too low. A distressed restructuring could be classified as defaulted with an immaterial diminished financial obligation. We would consider a cap of 5% to be more appropriate.
Inconsistencies with accounting treatment should be avoided as much as possible. Accordingly, the interest rate used should be aligned with the one applied for accounting treatment, i.e.:
• the original interest rate or an approximation thereof for fixed-interest rate loans;
• for variable interest rate loans, the current effective interest rate (consistently with the rate used to project cash flows).
As stated in question 4, basing the definition of default on the price of an asset is not appropriate.
This would indeed have a pernicious effect as banks will have an incentive not to purchase discounted assets in order to avoid to put the borrower in default on their data systems, this would be an issue especially if banks are already exposed to the issuer. From an economic standpoint, we cannot ignore the fact that banks act as intermediaries on markets and that purchased receivables management is an integral part of the banking sector’s activity.
In paragraph 47, G (other indications of unlikeliness to pay), “assets purchased or originated at a material discount” should NOT be considered as a possible indication of unlikeliness to pay.
For instance, the following revolving credit facility has been recently proposed to banks:
• Metals and mining company
• Margin: LIBOR +85
• Maturity: May 15, 2020
• Facility Size: $3.0 billion Senior Revolving Credit Facility
• Offering Amount: $25mm
• Price/Par value: 84%
With a discount up to 16%, the counterparty would have been declared in default according to the proposed definition. Nonetheless, the borrowing company is still rated Baa3/BBB+ by external credit agencies, while recently downgraded further to a technical event on a dam it is operating.
We therefore disagree with this proposal: it would trigger default for reasons which are independent from the credit risk of the counterparty. Moreover, this is not consistent with the incentive to use Basel parameters to assess credit risks nor with the notion of “use test”.
We understand the will to hinder multiple default cases by imposing a probation period. However, the proposed 3 months period might be too long in certain cases and would require important adaptations of the current rules regarding management of defaulted clients.
The rule regarding the return to non-defaulted status will induce a lot of issues, namely regarding modelling and operational monitoring.
We believe that the rule should be refutable in order to allow taking into account certain situations such as the repurchasing of a defaulted client (for example a bank) by a sound counterparty.
The rules regarding the counting of days should be considered independently from the rules regarding distressed restructuring. Thus an obligor could be in default because of a distressed restructuring, while not being past due thanks to the restructuration.
We agree with the proposed approach for retail exposures. Our retail activities strike not only transactional relationships with our clients but also fiduciary and middle/long-term relationships with them.
Nonetheless, we note that the contagion from the professional to the personal accounts (or vice versa) of a given counterpart should not be applied systematically but should be decided on an expert basis.
In some cases, isolating credit facilities and identifying the defaulted exposures amid others is not clear cut, e.g. when a mortgage payment is automatically assigned to another credit facility.
Some banking arrangements stipulate that some overdrafts can be indeed automatically compensated by cash from another account of a given client. Sometimes, these accounts are managed by the same banking group but by different legal entities. We think that legal restrictions linked to confidentiality could prevent the possibility to consolidate the defaults between different legal entities of our institution. The application of this pulling effect approach would be operationally very expensive.
The rules for the calculation of the days past due and the materiality threshold at the consolidated level would be complex to define and implement. In a word, the added value of this approach would be low.
As already stated in the introduction, the leverage ratio seems to be another indication of unlikeliness to pay quite constraining and not economically justifiable if taken in isolation: (c) the borrower’s overall leverage level has significantly increased or there are justified expectations of such changes to leverage". While the borrower’s overall leverage level is a useful risk parameter, it should not be considered in isolation; it would be indeed counter-intuitive to consider that a significant increase in leverage should be automatically considered as an event of default."
We understand that:
• In the event a joint credit obligation is identified as defaulted, all individual obligors linked to this joint obligation are also considered as defaulted.
• Other joint obligations between one of these defaulted obligors and other non-defaulted obligors are not systematically defaulted.
• When one but not all individual obligor(s) is/are defaulted, there is no systematic contagion towards joint credit obligations, although this may be extended on an expert basis.
• When all individual obligors are defaulted, the joint credit obligations is also considered defaulted.
A global reflection about materiality thresholds shall be considered as far as it may raise competitive issues between Member States (e.g. in France, the past due is presumed significant as from the first 1€ in default). In some countries a 200 euros threshold already represents a significant amount. We think that a maximum" threshold would be more efficient than a strictly imposed limit, with the possibility to reduce the threshold based on the loan / context reality."
We agree with these requirements which banks by and large already meet. This seems to be already in line with CRD 4 requirements.
We would like to take this opportunity to draw the EBA’s attention on the following points:
Coordination with the Basel Committee Guidelines:
As a point of attention, we underline that the Basel Committee already launched guidelines on credit risk management processes to be applied further to the application of IFRS 9. It is crucial to ensure that there will be no contradiction between both sets of requirements.
We understand EBA requires banks to incorporate the new default definition into their internal procedures/IT systems and adjust accordingly their rating systems by the end of 2019 (including competent authority model approval). In order to meet this deadline, this leaves us four years to do the below mentioned steps in the following order:
1. Launch the necessary developments in order to incorporate the new definition in the IT system
2. Observe new defaults. A minimum observation period of 12 months is required
3. Assess the impact of this definition on rating models, in order to determine if models should be recalibrated or rebuilt
4. Recalibrate or rebuild the models
5. Obtain approval by the bank’s internal model validation team
Please note that the implementation process is burdensome and may imply material changes from an IT perspective.
LIFO / FIFO approach:
• The daily basis counting begins when the amount past due breaches the threshold.
• An exposure will be considered as defaulted when the materiality threshold will be breached over 90 consecutive days.