European Banking Federation

Please see also the general comments of the attached EBF position paper.

Wording of Section 3.2.2 (first bullet point on page 7) and section 5.1 D c (pages 50-52) seem to suggest that only data or system errors caused by the institution are covered in the definition of technical default as opposed to including data or system errors caused by counterparties.

If defaults that are a result of data or system errors of the counterpart are not regarded as technical defaults, modelling of PDs for large corporates will turn into a matter of modelling probability of errors in customer’s data and payment systems. With the strict interpretation of a technical default suggested by EBA, the most relevant risk driver in Large Corporate portfolios would be the size of the company. We believe that payment delays not related to deterioration of the credit quality of the counterparty should not lead to default.

We do not share the view in paragraph 20 that the classification of the obligor to a defaulted status should not be subject to additional expert judgement. Expert judgement has an important place within credit risk management and should continue to be used. Situations may occur that will result in past due exposure of more than 90 days, however not due to a credit deterioration of the counterparty.

For example:

• In leasing business lines, a client could suspend payments not only for a difficulty to reimburse the bank but also for a management decision if a dispute occurs on the leased good (e.g. regarding the quality of goods).
• The same might occur in factoring business. In case of a controversy on a supply, litigation or discussion, the debtor can decide not to pay the invoices, although his creditworthiness is unchanged. This furthermore will often be unknown to the factor and could lead also to damaging contagion effects (please see also our response to question 2 below).
• Commercial dispute on a SBLC (Standby Letter of Credit) 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 the participating EU banks would place the corporate in default, resulting in a possible limitation of the customer’s access to the credit).
• Call of suretyship where the suretyship contest the legitimacy of the call which entail a past due situation (case of unfair/abusive claim).
• Logistic process issues for Energy & Commodities financing or generally for Trade Finance leading to delays in the delivery. For example, merchandise blocked at the customs, prohibition on entering or leaving ports, strike etc.
• Disputes regarding the amount or the nature of collaterals 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. The expert assessment is essential.
• Specific cases of sovereign counterparts, for which default may be assessed at political level.
• Cases of force majeure (environmental disasters, legally imposed measures, riots, strikes, wars…).
• Payments made by debtors to a factor for certain ceded invoices and not yet registered on the right account due to difficulties in the payment reconciliation process.
• Invoices due but not correctly and promptly dispatched to the debtor by the seller.

Some commercial disputes can last for several months or even years meaning that borrower would stay in default during the whole period. We believe it is important to avoid capturing defaults related to exposures that are disputed or waived or that are not related to the decrease in the quality of the credit risk.

Finally, we assume that the technical defaults that do not lead to actual default do not need to be registered in a central risk database and registration in the local data base for audit trails would suffice. The audit trails of technical defaults that did not lead to actual defaults should only be required for new cases.

As a final consideration we would like to stress that should the EBA definition of default (defined among others by way of new and stricter rules on technical default) remain unchanged, floors on the minimum level of LGD have to be reduced or completely removed. In addition, some adjustment will be needed to the general calibration of standardised approaches.
Please see also the general comments of the attached EBF position paper.

We suggest do add the following sentence after §22: When the factor and the client agree a due date for the credit granted to the client, the counting of days past due shall commence from such date.”

In the factoring arrangements where the ceded receivables are not fully transferred to the factor, the timing for reimbursement/regularization of advances is contractually specified, therefore, the counting of days past due should refer to that date. It does not necessarily require a breach in the percentage agreed between the factor and the client.

The treatment of exposures to debtors stemming from IAS/IFRS compliant purchased trade debts within a factoring agreement with a client (i.e. where the risks and benefits related with the assigned receivables are fully transferred to the factor and the factor has exposures to the debtors of the client) should take into account that the reliability of the due date of the invoices may be affected by numerous events related to the trade relationship between the buyer and its supplier.

In such cases, a significant delay of the payment may occur without any sign of deterioration of the situation of the debtor. Such situations may originate from contractual provisions or also from informal communication and exchange between the buyer and the supplier.

In such situations, we believe that a relief should be introduced by way of a rebuttable presumption on the automatic classification as past due of trade debtors or of a suspension of days past due counting when the factor is aware of these events, regardless the degree of formality. These occurrences shall however trigger an analysis of the debtor's situation in order to assess possible indications of unlikeliness to pay.

In particular, when the buyer disputes a receivable (e.g. receivables not existing at all or just partially existing, commercial supply not regular or different to the agreements, etc.), the amount or even the very existence of the invoice may be challenged. It is very uncommon that disputes are brought to a court. While disputing parties are usually try to settle the dispute outside the court, the process can nevertheless be time-consuming and exceed the 90 days. The opportunistic use of disputes in order to hide financial difficulties could easily be detected through the analysis of the debtor’s situation triggered by the occurrence of the dispute."
Please see also the general comments of the attached EBF position paper.

The stage 3 of IFRS 9 includes exposures where the credit risk of a financial asset increases to the point that is considered credit impaired. While most of the defaulted exposures would be classified as ‘Stage 3’ under IFRS 9, we believe it should not be assumed automatically. Also, given that banks must calculate PD by defining what is a default first in order to calculate impairment and not in the reverse order, impairment figures could not be used as an input for default triggers.

While we welcome the clarification made by EBA that the “incurred but not reported losses” (IBNR), which is a current notion of IAS 39 should not be considered as an indication of unlikeliness to pay (§ 26 of the CP), we consider that such a clarification should encompass the stage 2 of the forthcoming IFRS 9. Therefore, it should be stated clearly that the IFRS 9 stage 2 should not be considered as an indication of default. This is all the more important given that § 178.3 b) of the CRR mentions “a significant perceived decline in credit quality” as an indication of unlikeliness to pay that might be misleading and wrongly analogised with the stage 2 of IFRS 9 (“significant increase in credit risk”).
Please see also the general comments of the attached EBF position paper.

In the Consultation Paper (par. 25) it is mentioned that in general, Specific Credit Risk Adjustments (SCRA) should be treated as an indication of unlikeliness to pay if they concern “losses as a result of current or past events affecting a significant individual exposure or exposures that are not individually significant which are individually or collectively assessed”.

This point seems to be in contradiction to recital 12 of Commission Delegated Regulation 183/2014, which clearly states that “For the purpose of the determination of default under point (b) of Article 178(3) of Regulation (EU) No 575/2013, it is necessary to include only Specific Credit Risk Adjustments which are made individually for a single exposure or a single obligor, and not to include Specific Credit Risk Adjustments made for whole groups of exposures. Specific Credit Risk Adjustments made for whole groups of exposures do not identify obligors of exposures belonging to such groups for which a default event is considered to have occurred. In particular, the existence of Specific Credit Risk Adjustments for a group of exposures is not sufficient reason to conclude that default events have occurred for each of the obligor or exposures belonging to this group”.

While we agree that selling of a credit obligation resulting in a loss due to fall in credit quality could be an indication of default, we consider setting a threshold inappropriate. Even a small deterioration in credit rating or changes in interest rate might lead to impact on the market value of an asset greater than the proposed 5%. If an asset would deteriorate from AA to BB, the asset is still quite remote from default, but could be sold at a discount larger than 5% due to credit deterioration. This should however not trigger a default for IRB modelling. The proposed threshold would restrict the selling activity and lead to less liquid financial markets. We would recommend removing a minimum threshold and leave more room to expert judgement.

Moreover, this is not consistent with the incentive to use Basel parameters to assess the credit risks, nor with the notion of “use test”.

For instance, the following revolving credit facility has been recently proposed to financial institutions:

Borrower: xx (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%
(source: xxx)

With a discount up to 16%, the counterparty would have been declared in default accordingly with the proposed definition. Nonetheless, the borrowing company is still rated Baa3/BBB+ by external credit agencies.

The recent history 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. In consequence, the bank may cease granting facilities to the obligor and this could trigger an actual payment default.

Credit obligations could be sold for another reason than the anticipation of a decrease in credit quality of the issuer. A decision to sell participations in loans on performing clients and with a significant loss may be dictated by:

- Regulatory capital savings or employment
- Liquidity management
- Balance sheet management
- Country envelope consumption
- Counterparty exposure management
- Single limit concentration management

A sale price of an asset, which is the fair value, will include other elements besides the credit quality such as liquidity premium; general changes to market conditions, etc and it may not always be straightforward to distinguish which part of the economic loss is related to the deterioration in credit quality. We would therefore suggest to set up objective criteria to identify sales of credit obligations not related to credit risk.

It also has to be taken into account that in case of sales of credit obligations ‘en bloc’, a discount is usually applied compared to one to one evaluation (in order to conclude the deal earlier) with no link to the real risk of the block.

We would propose that the sale of credit obligations is considered as an indication for unlikeliness to pay but should be associated with other indicators and not as a stand-alone criterion.

Finally, we would appreciate a clarification by the EBA on whether securitized credits have to be considered within the “sale of credit obligation” category.
Please see also the general comments of the attached EBF position paper.

We would advocate the discounting with interest rate that is used as an approximation of the original effective interest rate on a best effort basis.

Considering the 1% threshold for the diminished financial obligation proposed in the QIS, we believe the threshold should be removed. In our view, the relevant measure for recognition of default should be set at a level, when the new cash flow (NPV) would no longer be adequate to cover the value at origination of the obligation, regardless of the decline in NPV.

In addition art 178 (3.d) of CRR considers “material forgiveness,…, of principal, interest or, where relevant fees”. The proposed threshold seems not to be consistent with the above mentioned materiality criterion and has therefore to be set at a significantly higher level.

We would also see no need for specifying additional indicators to be considered for identification of default if the net present value of expected cash flows on the distressed restructuring arrangement is higher that the net present value of expected cash flows modifications.

Concerning the formula for calculation of the diminished financial obligation (DO), it is not clear whether the cash flows include the expectation of recovery. If they do not take into account recovery expectation, the threshold would not make much sense as the new restructured loan could include a reinforcement of the collateral value which might mitigate (partially or totally) the diminished financial obligation measured with the proposed formula.

Also the formula provides possibility to hide a distressed situation by sufficiently extending maturity and maintaining an equivalent NPV of cash flows. It is also not clear if the two NPV parameters only include the future contractual cash flows or also PD and LGD associated to those cash flows in each moment. In the case of latter, the approach would suffer from a circularity problem. Some restructuring would not be considered as defaulted under the proposed formula for instance when the credit obligation is turned into a PIK loan (payment in kind) with capitalized interest during the period. The proposed formula gives an economic loss of 0.

Concerning paragraph 43, we see it contradictory to the Chapter 5 part D. Paragraph 43 states that “All exposures classified as forborne non-performing …. should be classified as default and subject to distressed restructuring”.

Chapter 5 Accompanying Documents part D however states that the preferred option of EBA is a “non-obligatory alignment of the definition of default with the non-performing exposures’ given the unintended consequences and high default rate should definition of default be aligned with the non-performing exposures. This would be consistent with EBA answer on 2/10/2015 to a question for “exit criteria NPE” where it is mentioned that “the category of non-performing exposures can be broader that the category of defaulted or impaired exposures. Defaulted or impaired exposures are mandatorily considered as non-performing but non-performing exposures need not be impaired or defaulted.“

Finally, we would like to underline that the concept of distressed restructuring does not apply in case a revision of the conditions is allowed by the contract (e.g. embedded clauses) or by specific laws (e.g. moratoria issued by banking association/government) or to commercial renegotiations (e.g. change of interest rate for commercial purposes or alignment with current market practices).
Please see also the general comments of the attached EBF position paper.

We disagree. Basing the definition of default on the price of purchased asset is in our view not appropriate. The price of an asset ought to reflect its value at that point in time. A material discount can be the result of other than financial distress such as general changes to market conditions and a result of negotiations e.g. settling other transactions.

Therefore it seems unreasonable that the discount as such should be an indicator of unlikeliness to pay. An asset purchased at a material discount should be checked for any other indicators of unlikeliness to pay. Such assessment is normally a part of the due diligence of the asset to be bought, to be able to establish a relevant value/price of the asset. For every asset or portfolio of assets that an institution purchases (with or without material discount) a proper due diligence is made, including testing the 90 days and the unlikeliness to pay criteria. The assessment could trigger a default.

Introduction of rules linking default to the price of purchased (or sold) assets could in effect lead to disincentives for banks to purchase/sell assets at discount to avoid putting its client in default (if a bank already has an exposure to the issuer). This would have negative effect on the role of the banks as intermediaries on the financial markets. Purchased receivables management is an integral part of the banking sector’s activities.

In our view, default should be triggered only for reasons that are directly linked to the credit risk of the counterparty.
Please see also the general comments of the attached EBF position paper.

We believe the institutions are best placed to recognize when a customer is no longer in default and we consider the set probation periods inappropriate. The institutions should determine the probation period in a way that that minimizes the ‘re-defaults” of the clients.

Any probation period from default to non-default status is inconsistent with what is set out in Article 178(5) where it is stated that: “If the institution considers that a previously defaulted exposure is such that no trigger of default continues to apply, the institution shall rate the obligor or facility as they would for a non-defaulted exposure”.

Also, under IFRS 9 favourable changes in credit risk should be recognised symmetrically with unfavourable changes in credit risk (IFRS 9 BC 5.210). 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 which is contra intuitive.

The suggested 3 months’ probation period is considered too long in relation to both large corporate exposures and retail consumers in particular when applied together with the strict definition of the technical default as proposed in the CP. In case of retail customers, payments are not always fully automatized by systematic debit of the customer’s account. Also, in some countries, loans with an undefined maturity are common for SMEs portfolios. Delays in payments of large corporates may be caused by systems or data errors. Such defaults would not necessarily mean a deterioration in the credit quality of the borrower especially if the cure period is short (less than 30 days).

We would suggest that in such cases, customers being classified as defaulted could return to non-default status as soon as the obligation is paid in. The three months period should therefore be eliminated as mandatory provision at least for default trigged by past due.

In the final draft of the Guidelines, having regard to Article 59 of the Consultation Paper, it is opportune to specify which repayment suspensions shall be considered as a “grace period”. In particular, if a restructuring arrangement provides a temporary suspension of the sole interest share of the loan, it is not clear if that suspension shall be treated as a grace period, considering that the principal share of the loan will not be suspended.

Similarly, should a defaulted client of a bank be bought by another client of the bank (client B) that is not in default, the exposures of the client B should not be considered defaulted if there is no decrease in the credit quality of client B (due to the acquisition). The remaining unlikeliness to pay should be the decisive criteria. Depending on the portfolio specific characteristics, there might be different or no probation periods.

Finally, we believe the following should be added at the end of paragraph 60a):
…”or the debtor has otherwise demonstrated its ability to comply with the post forbearance conditions”.

The current EBA Definition of Default (effective from September 2014), Paragraph 157 (c) states that:

“Concerns may be considered as no longer existing when the debtor has paid, via its regular payments in accordance with the post-forbearance conditions, a total equal to the amount that was previously past-due (if there were past-due amounts) OR that has been written-off (if there were no past-due amounts) under the forbearance measures OR the debtor has otherwise demonstrated its ability to comply with the post-forbearance conditions.”

The requirement that the “debtor has otherwise demonstrated its ability to comply with the post-forbearance conditions” has been removed from the EBA Consultation Paper, however we would consider this an important and more reliable indicator that the customer’s situation has fully resolved itself. Therefore we would have a strong preference that the criterion for curing to non-defaulted status read as follows “a material payment has been made (a total equal to the amount that was past due or has been written off) OR the debtor has otherwise demonstrated its ability to comply with the post-forbearance conditions”.
Please see also the general comments of the attached EBF position paper.

To summarize our understanding of the proposed guidance:

The guideline confirms that for retail exposure, the financial institution in accordance with second sub-paragraph of Article 178 may apply the definition of default at the individual credit facility level rather than at the obligor level. Furthermore the choice should reflect the financial institution’s internal risk management practice. This may imply that a financial institution in general apply the definition of default at obligor level, but for some specific types of exposure apply it at facility level.

Under IRB the financial institution is required to ensure that the risk estimates correctly reflect the definition of default applied to each type of exposures.

We supports that the credit institution may decide when to apply the default definition at obligor level and/or facility level.
Please see also the general comments of the attached EBF position paper.

We agree that when an obligor defaults on a significant part of their exposures, the institution will consider this as additional indication of the unlikeliness to pay but it should not automatically indicate the unlikeliness to pay of the remaining credit obligations of this obligor. For example, a mortgage default might result in a higher Probability of Default on other credit obligations but not necessarily the default of them.
Please see also the general comments of the attached EBF position paper.

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.

Moreover, the identification of joint fully liability of retail obligors (i.e. married couple) would expose institutions to unbearable burdens especially when this implies ongoing updates of dynamic information (the marital status) difficult to obtain. As a result, we propose to drop article 85.
Please see also the general comments of the attached EBF position paper.

We agree. The requirements seems to be in line with CRD IV requirements. It should be ensured however that there is also an alignment with the final Basel Committee Guidelines on credit risk management processed to be applied in accounting for expected credit losses.
Denisa Mularova