The new definition of default – How is it going to affect you?

By Hubert Fonteijn, Senior Consultant

and Kristian Lajkep, Regulatory Compliance Officer

The 1 January 2021 will see the European Banking Authority’s (EBA’s) guidelines on the Definition of default come into action. This new definition aims to eradicate different approaches towards the default definition used by different institutions within the EU. It is estimated that the new definition will make a substantial impact on models, capital adequacy ratios and accounting characteristics of Banking institutions. More profoundly, the new definition of default is going to have a considerable effect on the own fund requirements. Depending on the disparity between their currently implemented definition and the new one, the effect on the institutions may be significant. Their nature will also differ depending on the type of approach that a given institution uses to calculate the capital requirements.

To find out more about the implemenation challenges, please consult our recent expert input  "New definition of default: Insight on challenges to implementation."

For the institutions that use the Internal Ratings based (IRB) approach, the new definition may have a considerable impact on the parameters of credit risk and is consequently going to influence the expected loss calculations and, by extension, risk weights. The institutions using the IRB approach are likely to be relatively more affected by the new definition of default. They will need to adjust the historical data fed in their IRB model to comply with the new definition of default. They will also need to include a new margin of conservatism to their calculations, that would offset the possibility of the past data used to calculate the requirements not using the new definition of default. Any changes to the internal models need to be approved by national competent authorities, which may add to the time costliness of the process.

The institutions that use the standardised approach are expected to be affected to a smaller degree as they are not relying on historical data. However, under CRR, the definition of default is a base on which the different exposures are assigned in default classes, so the impact of the new guidelines is still not going to be insignificant.

If the EBA’s impact assessments are to be believed, then the capital requirement is not likely to substantially increase. However, as this is ultimately meant to be a harmonization of frameworks, whilst most of the individual institutions will find their current capital reserves adequate or near-adequate, some institutions will indeed need to increase their available capital substantially. Whilst the decrease of CET1 ratio is most likely going to be negligible for institutions that make use of the standardized approach, the CET 1 ratio is expected to drop by approximately 20bp’s for the institutions that use the IRB approach. It is however notable that re-developing the internal models is going to be a far greater cost for most of the banking institutions than a potential increase in capital requirements.

Ultimately, the burden upon any institution will largely depend on the definition that a given institution has used to date, and how much it differs from the current definition. The EBA is aware that the time, effort and costs of implementation of the new definition of default can be substantial, particularly for the institutions that are using the IRB approach and those whose definition of default was very different from the newly established one. It claims that it is entirely possible that in addition to having to replace their default identification and IT processes, the rating systems will also likely need to be recalibrated. This aligns with Finalyse's experience: although a definition of default is not complex from a conceptual point of view, it is often surprisingly complicated to embed this definition into the IT systems. Moreover, the introduction of a change in the default definition will add a considerable workload on the reporting teams, since there could potentially be a large difference in the number of defaults before and after the introduction of the new definition. This difference will then have to be analysed and explained when preparing regulatory and annual reports.

As is customary with most of the regulatory pieces implemented as part of the single European framework, the EBA claims that the harmonisation of practices will not only lead to an increased comparability of risk parameters and own funds requirements (which is its primary objective), but will also reduce the costs of the institutions that operate in more than one European country and thus currently need to comply with the definitions in use in the different member states.

As a result, firms that have cross-border operations will benefit from the unified framework. It is likely that these institutions are using the IRB approach and, consequently, even if the transformation is likely to be most costly for these, they are also most likely to benefit from it.

Upon drafting the guidelines, the EBA has identified several inconsistencies under the previous regime, throughout the union and among distinct firms, which they seek to harmonise:

• Past due criterion in the identification of default
• Indications of unlikeliness to pay
• Conditions for a return to non-defaulted status and treatment
• The definition of default in external data

This paper goes through each one of these and identifies the major aspects and intuitions presented in the new guidelines.

Past due criterion in the identification of default consists of a number of different elements. One of the most important issues is the harmonisation of the counting of the days past due. The CRR states that:

“default shall be considered to have occurred when the obligor is past due more than 90 days on any material credit obligation to the institution, the parent undertaking or any of its subsidiaries.”

What constitutes the “material” breach is explained in a special RTS (on the materiality threshold for credit obligations past due). The RTS demands the competent authorities to use a twofold materiality threshold:

1. The absolute threshold defined as the sum of all past due amounts related to the credit obligations of the borrower towards the institution, the parent undertaking or any of its subsidiaries, and
2. The relative threshold defined as a percentage of a credit obligation past due in relation to the total on-balance-sheet exposures to the obligor, excluding equity exposures.

If both of these thresholds are breached for more than 90 (or in special cases 180) days, the default is said to have occurred. As for the specifications of the thresholds, both are set by the competent authorities, but the absolute threshold should not surpass EUR 100 for retail exposures or EUR 500 for non-retail exposures. The relative threshold is again set by competent authorities, but must be lower than or equal to 2.5% for retail and non-retail exposures. The EBA suggests a relative threshold of 1% to unify the framework. Competent authorities can use lower thresholds than those specified in the RTS and the institutions are encouraged to use still lower ones for their internal use.

But how exactly does one calculate these 90 days? If the credit arrangements allow the debtor to postpone or suspend the payment in certain circumstances, then those days ought not to be counted. Similarly, by law, some obligations may be postponed or suspended independently of the will of contractors. In this case too, the days are not counted. Nevertheless, this way of stalling may reflect some financial difficulties, so in these circumstances, the institutions are encouraged to assess the debtor’s ability to pay, even if by themselves they do not constitute a default. The countdown is also suspended for obligations undergoing court proceedings.

It may also be noted that a breach of smaller amounts for technical rather than financial reasons (failure of payment system), a so-called “technical default”, should not be treated as a default. Similarly, it may be possible that payments for the exchange of goods and services, from governmental or government-related institutions (This special treatment applies in particular to factoring exposures or similar types of arrangements but does not apply to instruments such as bonds.), will stall if the payments are conditional on a number of administrative procedures enshrined in law that may take longer than expected. Although the EBA notes that the governmental institutions should make every effort to repay their dues, this type of exposure (when there is no other indication of difficulties and obligation is not past due more than 180 days) gets a special treatment.

The factor contracts, by nature, do not lend themselves very easily to the past due criterion. As such, they are divided into two groups depending on whether or not the underlying receivables are recognised on the factor institution’s balance sheet. Since the own funds calculation is based on the accounting figures, the records of individual receivables may lead to somewhat more complete figures and may further simplify the calculation.

This January, IFRS 9 has become applicable for banking institutions, but they will only apply to insurers as of 1/1/2021. Within this regulatory framework, the definition of default follows this rule:

Unless there are other indications of unlikeliness to pay, the exposures classified by IFRS 9 as stage 2 are not considered defaulted, as are the stage 3 exposures, i.e. exposures treated as credit-impaired. The exceptions to this rule are the following:

• Exposures that use 180 days instead of 90 days as past due criterion
• Whenever the materiality threshold above says that an exposure does not belong to default
• Technical past due and governmental exposures described above

With regards to the other indications of unlikeliness to pay, the CRR views material credit-related economic loss from sale of debt as indicating a chance of default. The EBA’s material threshold of this loss is set to 5%. This again is a cap and the use of a smaller threshold by institutions is allowed and encouraged. The loss is defined as the difference between the price for which the obligation was sold and the amount outstanding. It should be noted that if the sale of the debt instruments is demonstratively motivated by other reasons than its declining quality, then it needs not to be counted as constituting towards a default.

The ITS on forbearance and non-performing exposures identify certain kinds of forbearance towards the debtor as constituting a distressed restructuring. A form of financial restructuring that would lead to reduced payment obligations resulting from material forgiveness, or postponement, of principal or interest, indicates an unlikeliness to pay. The exposures on which such forbearance is applied will be considered defaulted if the present value of restructured debt cash flows (using the original effective interest rate is materially lower than the present value of previous cash flows. The materiality threshold on these is very strictly capped at 1%, effectively allowing no more than a relatively slight rounding down. Even if the
difference is smaller than this materiality threshold, the institutions are encouraged to search for other indications of unlikeliness to pay, as any forrbearance is viewed as very telling.

The EBA notes that this list of indicators of unlikeliness to pay is far from being exhaustive. Institutions are encouraged to look on their own for less obvious traces of inability to meet the obligation. In particular, the connection of obligors to their partners is something that may raise a red flag as well as the purchase of an asset at a material discount rate. Last but certainly not least, there is a clear indication of unlikeliness to pay – a bankruptcy. Bankruptcy as a state is well defined in all national laws, but the EBA guidelines provide some unification of this definition.

According to CRR, the obligor is considered defaulted as long as it is not absolutely clear that it is no longer seen as unlikely to repay its obligations in full (without having to surrender the collateral). In order to avoid multiple defaults, the non-defaulted status is only retrieved after a probation period that should take at the very least 3 months (1 year for distressed restructuring) of honouring all obligations and not showing any signs of unlikeliness to pay. Institutions should implement an assessment process and gather enough data as to be able to justify the trust. Institutions are also encouraged to look for the instances of multiple defaults (a default soon after regaining the non-default status) and take them into consideration when judging the length of the probation period as well as possible indicators of inability to pay.

The requirements with regards to external data apply to the institutions that are using IRB only and which use the external data for the estimation of risk parameters. Not only should the external data be reflecting the portfolio that the institution holds, but its definition of default (among other things) should also be consistent with the definition of default used by the given institution. As seen above, there are many material thresholds that leave a lot of discretion to the individual institutions. Institutions need to make sure that the thresholds and policies reflected in the data are consistently reflecting their own approach towards defaults. This may not always be possible and certainly won’t be possible for historical data if an institution changes its definition of default as a result of the very EBA guidelines that this article is about. As a result, the institutions are required to use a larger margin of conservatism in their estimation of risk parameters.