EBA’s Latest Amendment to the Definition of Default: Key Changes Explained

The EBA asked for a distinction between performing (PE) and non-performing (NPE) forborne exposures first in its 2013 ITS on reporting. No clear guidance on how to separate between performing and non-performing forbearance was given, resulting in wide differences in practice among banks, as highlighted by the Quantitative Impact Study of 2016. To remediate this variability, the NDoD introduced the NPV test for forbearance measures:

$$DO = \begin{array}{c} NPV_0 - NPV_1 \\ \hline NPV_0 \end{array}$$

Where \(NPV_1\) is the net present value of the contractual cash flows after the concession and \(NPV_0\) is the net present value of the contractual cash flows before the concession. To be close to the IFRS9 accounting norms, the discounting of the cash flows is done using the original contract’s effective interest rate. 

A contract is considered as NPE (and the obligor/facility defaulted) if the NPV test is greater than 1%. In its CP, and now in the amendment of the guidelines, the EBA precises that the goal of the test is to distinguish between the concessions that will result in no diminished financial obligations for the obligor, and those that will mechanically introduce a loss for the bank. The 1% threshold is mainly chosen as a way to avoid any “technical” default due to minor discounting or rounding effects that would result by the adoption of a 0% threshold, such as:

• Tiny NPV loss due to the rounding of the monthly principal and interest payment

• Day count conventions (30/360, Actual/365…)

• “de minimis” or nuisance amount adjustment (rounding of the payments and final payment adjusting the balance)

• Changes in compounding frequency with rounding of the new contractual interest rate

It is interesting to note that there are currently different supervisory practices with respect to the calculation of this NPV threshold. For example, the Danish Supervisor allows for a delay in the NPV calculation if the institutions automatically classify the forborne exposures as defaulted.

In the CP, a 5% threshold has been considered to replace the 1% threshold currently in place. The idea behind such an increase is that the action of putting a facility or obligor into default has stronger effects than solely the recording of the loss, and hence a low NPV test threshold has a disproportionate effect on the banks’ provisioning. Indeed, defaulting a forborne exposure will result in:

• Stage 3 provisioning 

• Contagion to other exposures in case of a default definition at obligor level or if the pulling effect is triggered

• Increase of the recorded number of defaults, which will affect the backtesting and future calibration of IRB PD models

• Increase of RWA (due to the fact to have defaulted exposures on the balance sheet) both standardized and IRB approaches

• Operational and reputational costs of putting a client into default (collateral revaluation, margin call, change in commercial conditions…)   

However, the EBA opposed this change of threshold. First and foremost, it considered it would defeat the purpose of the forbearance framework (Introduced by the EBA/ITS/2013/03 and extended by the EBA/GL/2018/06) which prevents the use of multiple restructurings to financially distressed obligors in order to avoid recording defaulted exposures (a practice also known as “zombie lending”). Indeed, a bank could grant a concession leading to a loss of 4.5%, and then another one, without recording a default, while getting a loss of more than 8.7% on the exposure.

A second argument used by the EBA is that a 5% threshold would create inconsistencies with the relative materiality threshold of 1% for the counting of the days past due (DPD). The idea here is that setting a forbearance materiality threshold higher than 1% creates an opportunity for 'regulatory arbitrage.' It would allow lenders to grant concessions to borrowers who have already breached the 1% DPD materiality threshold and are approaching 90 days past due without classifying them as NPEs, provided the NPV loss remains below the higher forbearance materiality threshold. Such an argument is however weak as it mixes up a threshold calculated at the facility level (the NPV loss test) and a threshold calculated at the obligor level (the DPD materiality threshold). Furthermore, it compares a loss (the write-off in NPV due to the concession) with a delay in repayment, which will potentially be repaid in the future.

A third argument put forward by the EBA is that banks are highly levered with regulatory leverage ratios around 5%, so a loss between 1% and 5% on a material part of their portfolio could almost wipe out their Tier 1 capital. Such an argument is, however, far-fetched in our opinion as it extrapolates the threshold calculated on a single facility with the bank-level capital buffer.

In the CP, the EBA proposed two changes to the forbearance NPE probation period (PP) exit criteria. First the EBA proposed to shorten the probation period of non-performing forborne exposures to 3 months, from one year previously, under the condition that the forbearance was classified as NPE due to a NPV loss above 1% but below 5%. 

In the current forbearance framework, there are three reasons (see figure 1 below) for which a forborne exposure can be (re)classified as NPE:

• An NPV test indicating a loss larger than 1%

• A significant amount of arrears (30DPD) during any probation period

• A new forbearance has been granted on a forborne contract that has been previously classified as NPE

A second proposition is to condition the exit from forborne NPE status to a material payment equal to the amount that has been written-off by the concession granted, capped at 20% of the principal outstanding prior to the concession.

We are of the opinion that aligning the NPE forborne PP to the other default PP would have made granting forbearance measures more attractive. First, the duration of the default period, due to the effect discounting of cash flows (EBA/GL/2017/16 paragraph 143), has a strong influence on the RWA impact of defaults. Second, reducing the length of the NPE period would hasten the exit of the NPE status and hence reduce the effect of NPE prudential backstops on CET1 capital. However, the proposed changes to the framework, conditioning the length and exit of the PP to additional calculations would have complexified unduly the already burdensome forbearance framework. Furthermore, the one-year NPE PP aligns with the length of the prudential backstop pause triggered by the granting of a forbearance to an already NPE exposure.

None of these changes have been retained and hence the forbearance framework remains unchanged: the NPV loss threshold remains at 1% and the NPE probation period stays at 12 months, without any further exit conditions.

 In response to the increasing likelihood of recurring climate-related crises, the EBA considered whether to permanently embed rules for "general payment moratoria" into its Guidelines on the DoD. Such rules would clarify the process guiding exceptional moratoria, such as the one issued during the COVID-19 pandemic. The central regulatory concern regarding such moratoria is that, by allowing automatic concessions to borrowers, they would trigger a mass classification of loans classification as forborne (potentially NPE) under the current DoD.

The EBA rejected the idea of introducing a dedicated exemption related to moratoria as it would enable a Member State to issue a legislative moratorium whenever a specific crisis is affecting its territory. Such possibility would go against the objective of harmonisation of the DoD and would introduce undue RWA variability. 

Some respondents to the CP proposed extending this special treatment of the DPD count to operational leasing, i.e. a commercial rental agreement where a business (the lessee) rents an asset from a leasing company (the lessor) for a specific period, without taking ownership of the asset at the end of the contract. They draw a parallel between non-recourse factoring and operational leasing, as both are invoice-based and have overlapping instalments (new financial commitments before the previous expired). 

The EBA ruled against this extension of the technical DPD to operational leasing as it argues that, whereas non-recourse factoring is a tri-partite arrangement, there is a direct contractual relationship in the case of leasing and hence the credit institution has full control of the payment collection process. 

Are you looking to simplify your IRB model landscape? To put a stop to the sprawling repair programs and countless findings? Drawing on extensive experience with supervisory engagements, Finalyse can support banks in their IRB and standardized approach journey. 

Finalyse has successfully helped clients:

• Ensure compliance with regulatory expectations (CRR and EBA)

• Redevelop their IRB models

• Implement their standardized capital calculation

• Validate the IRB rating chain (models, data and governance)

• Optimize their RWA calculation and remove unnecessary capital burden

Our consultants bring hands-on experience across regulatory, risk, and operational dimensions of credit risk, enabling institutions to move from concept to execution with confidence.