Final draft RTS on the materiality threshold for credit obligations past due

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE EBA/RTS/2016/06 28/09/2016 Final Report Draft Regulatory Technical Standards on the materiality threshold for credit obligations past due under Article 178 of Regulation (EU) No 575/2013

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE Contents

Executive Summary 3
Background and rationale 5
Draft regulatory technical standards on the materiality threshold for credit obligations past
due under Article 178 of Regulation (EU) No 575/2013 9
Accompanying documents 17
4.1 Impact assessment 17
Introduction 17
Problem definition 17
Objectives 18
Baseline scenario 18
Assessment of the technical options 20
Impacts of the technical options 25
4.2 Feedback on the public consultation 33

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE

Executive Summary
Article 178 of Regulation (EU) No 575/2013 (the Capital Requirements Regulation – CRR) specifies
the definition of default that is used for the purpose of the IRB Approach according to Chapter 3 of Title II in Part Three of the CRR, as well as for the Standardised Approach in line with Article 127 of the CRR. The definition specifies, inter alia, that a 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. The materiality threshold for such obligations past due is set by the competent authority and reflects a level of risk that the competent authority considers to be reasonable. In this regard, Article 178(6) of the CRR mandates the EBA to specify the conditions according to which a competent authority shall set the threshold. These final draft RTS consequently specify the conditions according to which a competent authority shall set the materiality threshold. This will help ensure consistency in the setting of the materiality threshold across the entire EU. Due to the wide range of practices currently used by institutions with regard to the materiality thresholds, in part as a consequence of different requirements in this regard set by national competent authorities, these RTS set conditions in relation to both the structure and the application of the materiality threshold. This harmonisation is necessary in order to ensure a consistent use of the materiality threshold and will help reduce the burden of compliance for cross-border groups. The conditions set out in these RTS in particular require that competent authorities set a materiality threshold that is composed of both an absolute and a relative threshold. The absolute threshold refers to the total amount of the credit obligation past due understood 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. The relative threshold is 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. In the case where both of those limits are breached for 90 consecutive days (or 180 days if the competent authority has decided to replace the 90 days with 180 days in accordance with Article 178(1)(b) of the CRR) a default would be considered to have occurred. Based on the requirements included in these RTS, competent authorities are required to set a threshold for retail and for all other (‘non-retail’) exposures, which will apply to all institutions in a given jurisdiction. The absolute threshold cannot be higher than EUR 100 for retail exposures or EUR 500 for non-retail exposures. It is suggested that the relative threshold should be set at the level of 1% for both retail and non-retail exposures. However, if a competent authority considers that this suggested level of the materiality threshold does not reflect a reasonable level of risk it may set a relative
threshold at a different level, which in any case must be lower than or equal to 2.5%. The 1% unified level of the relative threshold and the upper bounds specified on all

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE thresholds ensure sufficient conservatism and harmonisation with regard to the levels of the thresholds across jurisdictions. It is expected that the implementation of these final draft RTS may have a significant impact on the operations of some institutions. In particular for those institutions that use the IRB Approach and where the threshold will change significantly, the implementation of the necessary adjustments may require some time. Hence it is recognised that institutions will need sufficient implementation periods that will allow the changes to be introduced in an efficient manner. However, those periods should be reasonably limited.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE
2. Background and rationale
The definition of default including the concept of a materiality threshold set by competent authorities was introduced by Directive 2006/48/EC of 14 June 2006 (which formed part of what was known as the Capital Requirements Directive – CRD), later replaced by Regulation (EU) No 575/2013 (the Capital Requirements Regulation – CRR). The materiality threshold is not only used for the purpose of the IRB Approach; it applies also to institutions that use the Standardised Approach. However, in the absence of specific rules on the structure and application of the materiality threshold, various approaches have been adopted across jurisdictions. As a consequence a wide range of practices has been observed. In the majority of jurisdictions specific rules have been adopted concerning the materiality threshold. These rules are usually in the form of hard limits, in relative terms, in absolute terms or in a combination of the two. Relative thresholds do not usually differentiate between types of exposures or obligors. The limits range from 1% to 10% of the exposure. Absolute thresholds differ significantly between countries and range from EUR 0 to as high as EUR 50 000. Furthermore, practices differ substantially with regard to the structure of the threshold. Firstly, some competent authorities set different absolute thresholds for retail and non-retail portfolios. Secondly, in some jurisdictions the materiality threshold is not used, and if any amount is past due more than 90 days the exposure is considered defaulted. Thirdly, some competent authorities have not set explicit limits but allow institutions to define their own limits and make case-by-case evaluations, which has led to significant variation across institutions even within the same jurisdiction. Finally, some jurisdictions have also applied a threshold specifically to remove socalled technical defaults from data series used to estimate IRB models. Hence a common understanding of the concept of a materiality threshold is necessary in order for competent authorities to set the threshold. Apart from different structures and levels of thresholds there are also significant differences with regard to the reference amount that is compared with the threshold and the counting of days past due. In a few jurisdictions the days past due may be counted only from the day that the amount past due becomes material, regardless of contractual obligations. In that case the whole amount past due is taken into account. Following other approaches the reference amount compared with the materiality threshold may be either the amount past due more than 90 days or the whole credit obligation. Each of these approaches may be applied either to the total exposure of an obligor or separately to every facility of the obligor. Given the variety of approaches it was clear that any solution adopted in these RTS would require significant adjustments in many institutions. Therefore, the development process focused mainly on
finding the best possible approach that would serve the materiality threshold’s objective of avoiding treating as real defaults such cases where the past due exposure is not a result of materialisation of credit risk but occurs due to other circumstances. The materiality threshold

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE should prevent the recognition of too many defaults that will be cured in a short timeframe, but at the same time the threshold should not prevent timely identification of real default cases. The structure and level of the materiality threshold may have a significant impact on own funds requirements. In particular, in the case of institutions that use the IRB Approach, the classification of exposures as defaulted impacts not only the calculation of risk weights and expected losses for defaulted exposures, but indirectly also other exposures through its impact on PD and LGD models. In general, a lower materiality threshold results in more defaults being identified and consequently the total of the expected and unexpected loss estimations is higher. Under the Standardised Approach the unsecured parts of defaulted exposures are grouped in a dedicated exposure class and receive a conservative risk weight. Due to the potentially high impact of the structure and level of the materiality threshold as described above, it is important to ensure a level playing field across institutions, and within and across jurisdictions. Therefore, the final draft RTS are based on a common structure for the threshold and require that a single threshold should be applied to all institutions in a certain jurisdiction. It is expected that the harmonisation of practices will reduce the burden for crossborder institutions of complying with different requirements in different Member States. A basic decision that had to be taken in the development of these final draft RTS was the definition of the reference amount that the threshold should be compared with. Unification of approaches in this respect is crucial to ensure a minimum level of comparability across institutions and jurisdictions. The reference amount is also a basis for the meaningful calibration of the level of the threshold. It is proposed that in the assessment of the materiality of credit obligations past due all past due amounts related to the credit obligations of the borrower towards the institution, the parent undertaking or any of its subsidiaries should be taken into account. In order to mitigate the risk of splitting the credit obligations into smaller portions or of selective repayment of the obligations by the obligor in order to avoid the default being triggered, all amounts past due, irrespective of which credit obligation of the obligor they are related to, should be summed and the sum should be assessed against the materiality threshold. This approach also ensures that the application of the materiality threshold will be to a large extent independent of the payment allocation scheme used by an institution (i.e. LIFO, FIFO or any other approach). The threshold should be structured as a combination of an absolute and a relative limit. The absolute component should be used as described in the previous paragraph. The relative
component is the sum of all past due amounts as a percentage of the total on-balance-sheet exposures to the obligor excluding equity exposures. In the case of retail exposures where the definition of default is applied at the level of the individual facility, the sum of the amounts past due related to a single credit obligation (facility) of the obligor should be taken into account. For the purpose of the relative threshold this sum should be considered as a percentage of the value of on-balance-sheet exposures related to this single credit obligation. The use of on-balance-sheet exposures as the denominator of the relative threshold provides a simple and comparable

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE solution. As only the outstanding exposures, unlike unused credit lines, can in fact become past due, it ensures consistency between the numerator and denominator of the ratio. Furthermore, it prevents the impact of the relative threshold being diminished by the inclusion in the denominator of off-balance-sheet exposures that cannot in practice be drawn by an obligor and do not have credit characteristics. It is proposed that the obligor should be considered defaulted whenever both of the components of the threshold, i.e. the absolute and the relative limits, are breached for 90 consecutive days (or 180 days if the competent authority has decided to replace the 90 days with 180 days in accordance with Article 178(1)(b) of the CRR). This approach is balanced and proportionate, as it takes into account the exposure value and materiality is assessed in relation to it. Not only the structure but also the required level of the materiality threshold will impact on own funds requirements and may lead to an uneven playing field where thresholds are established by different competent authorities. While some differences in the levels of the thresholds are justified by the local particularities of each jurisdiction, a minimum level of harmonisation across the Union is desirable to ensure that the rules adopted in each jurisdiction are sufficiently prudent. Therefore, according to the final draft RTS competent authorities have flexibility in setting the level of the threshold up to a maximum level defined in the RTS. However, in order to achieve the maximum possible harmonisation a unified level of the relative threshold is proposed in the RTS. This proposed level of the threshold should be taken as a starting point for the analysis to be performed by the competent authorities for the purpose of setting the threshold. In this analysis they should take into consideration various factors relevant for the institutions in their jurisdictions, including the common characteristics of the obligors and transactions. Where this analysis leads to a conclusion that the proposed level of the threshold does not reflect a reasonable level of risk, competent authorities may set a relative threshold at a different level, but not at a level higher than the specified cap. It might be argued that the reasonable level of the materiality threshold depends on the characteristics of the obligors and their exposures; hence, several thresholds could be introduced for different types of exposures. In order to ensure a reasonable balance between the simplicity of the framework, which needs to be sufficiently operational from an implementation perspective, and risk sensitivity, it is proposed to allow the competent authorities to set different absolute thresholds for retail and all other, i.e. non-retail, exposures. It is between retail and all other exposure classes that the most significant differences in
average income and exposure values are observed, and the classification is readily available for all institutions, irrespective of the approach used for the purpose of the own funds requirements calculation. The proposed structure of the threshold is assessed to take into account the substantial differences between retail and non-retail exposures, while at the same time not adding too much complexity. It is expected that the implementation process might be operationally cumbersome especially for institutions that use the IRB Approach. The change in the materiality threshold will result in a change in the definition of default used for the development of IRB models. Consequently, this will entail an adjustment to the risk parameters, which will have to be recalibrated to reflect the

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE changes. Additionally, any changes in the definition of default are considered to be material changes to the rating systems; therefore, an approval of a competent authority will be required for the changed rating systems. Furthermore, the implementation process might be more burdensome for institutions that currently use a significantly different approach with regard to the materiality of past due exposures. It is therefore expected that such considerations will be taken into account by competent authorities when defining the individual timelines for the implementation of the threshold for particular institutions in their jurisdictions. However, in order to prevent excessive delays in the implementation of the threshold across the EU, such transitional periods for firms should be reasonably limited.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE
3. Draft regulatory technical standards
on the materiality threshold for credit obligations past due under Article 178 of Regulation (EU) No 575/2013

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE COMMISSION DELEGATED REGULATION (EU) No …/.. of XXX […] supplementing Regulation (EU) No 575/2013 of the European Parliament and of the Council with regard to regulatory technical standards for the materiality threshold of credit obligation past due under Article 178 (6) THE EUROPEAN COMMISSION, Having regard to the Treaty on the Functioning of the European Union, Having regard to Regulation (EU) No 575/2013 of the European Parliament and of the Council of 26 June 2013 on prudential requirements for credit institutions and investment firms and amending Regulation (EU) No 648/2012 1 , and in particular the third subparagraph of Article 178(6) thereof, Whereas:
(1) Due to similar market and economic conditions in the same jurisdiction, it would be appropriate for competent authorities to set a single threshold for the assessment of materiality of a credit obligation past due for all institutions incorporated in their respective jurisdictions. Such a threshold, that is expected to remain consistent over time, would bring the added benefit of increased comparability of capital requirements among institutions in the same jurisdiction. (2) On the one hand, given that the materiality threshold is dependent on the level of risk that the obligations past due represent, for individual obligors or exposures, the amount that can be considered material depends on the level of the overall credit obligation. On the other hand, in practice, institutions tend to consider all amounts below a certain level as immaterial, regardless of their relation to the overall credit obligation. As a consequence, the threshold to be set by a competent authority should be based on both of the above considerations and consist of two components: an absolute component (i.e. an absolute amount) and a relative component (i.e. the percentage of the whole credit obligation that the amount past due represents), whereby exceeding the limits of both of those components would result in the obligation past due being considered material. (3) There are significant differences in the average income and average amounts of credit obligations between various types of obligors; as a result the materiality of amounts past due in terms of the level of risk they represent, should also be differentiated accordingly. In order to reflect this need for risk sensitivity and 1 OJ L 176, 27.6.2013, p. 1.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE combine it with simplicity of the framework, it would be appropriate to set the absolute component of the threshold at different levels for retail exposures and for all other exposures, separately, where the determination of retail exposures is made in the manner referred to in Article 147 of Regulation (EU) No 575/2013, for banks applying the IRB approach, and in the manner referred to in Article 123 of that Regulation for institutions that apply the Standardised Approach. (4) While there is a need to ensure that the materiality threshold adapts to the local particularities of each jurisdiction, an adequate level of harmonisation across the Union is also desirable. It has been recognised that while the differences in economic conditions including the levels of prices in various jurisdictions may justify different levels of the absolute component of the materiality threshold, such justification can rarely be used where the threshold is expressed in percentage terms. As a result, rules on the setting by competent authorities of the materiality threshold for credit obligations past due for the purpose of identification of default should specify a unified level of the relative component of the materiality threshold while retaining some flexibility for setting the threshold at the level appropriate for specific conditions in the jurisdiction up to a specified maximum level for the threshold. (5) The materiality threshold may have a significant impact on the calculation of capital requirements and expected losses. Further, the threshold to be set by a competent authority is expected to affect all institutions in that jurisdiction irrespective of the method used for calculating their capital requirements. For these reasons, the level of the materiality threshold should be considered by competent authorities, based on a variety of factors. As a result, rules on the setting by competent authorities of the materiality threshold for credit obligations past due for the purpose of identification of default should also specify that competent authorities should consider separately the relevant risk characteristics of retail and all other exposures in their calibration of the threshold. (6) As the level of the threshold set by a competent authority will be applied among others by institutions that operate on a cross-border basis the levels of the thresholds set by other competent authorities may be an important factor when analysing whether the level of risk reflected by the level of the threshold is reasonable. In order to allow full assessment of the appropriateness of the level of threshold by competent authorities it is necessary that the levels of the thresholds set by other competent authorities are transparent to all market participants, hence it would be appropriate to achieve the desired level of transparency by requiring competent authorities to notify the EBA about the
levels of the thresholds that they set in their respecitive jurisdiction so that these can be made public. (7) Although the desired outcome of this Regulation is to unify to large extent the application of the materiality threshold across the institutions it is also recognised that some differences between the levels of the thresholds applicable in different jurisdictions may remain that will reflect different levels of risk that are perceived as reasonable by relevant competent authorities. In such situations, in the case of institutions that operate on a cross-border basis, it is expected that the application of the appropriate level of the materiality threshold will be discussed by relevant colleges of supervisors. In any case it is considered desirable that the applicable

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE level of the threshold is consistent with the internal risk management practices of the institution. (8) Article 178(2)(d) of Regulation (EU) No 575/2013 requires competent authorities to set the materiality threshold at such level that reflects the level of risk that they consider to be reasonable. As the level of the threshold that reflects the reasonable level of risk depends on the way the threshold is applied in the default identification process it is necessary to also specify conditions related to how the threshold will be used during the default identification process, such as the calculation of the measure that will be assessed against the threshold and the stage of the default identification process at which the threshold will apply. This could serve to avoid unintended consequences where institutions might implement the threshold in a manner different from what was intended by a competent authority, and therefore in a way that does not reflect the reasonable level of risk. It would also ensure the harmonised application of the threshold across the Union. In that context, and with reference to the timing of the application of the materiality threshold, it would be appropriate to apply it in order to identify the material credit obligation past due before the start of counting of the required number of days past due in accordance with Article 178(1)(b) of Regulation (EU) No 575/2013. This approach would have the advantage of being able to identify those obligors that pose significantly higher risk because they are characterised by systematically late payments even where partial or irregular payments are made. (9) Taking into account that the materiality threshold for past due exposures forms a
part of an overall definition of default and that in the case of institutions that use the
IRB approach any change of the definition of default leads to material changes in the rating systems that are used for the purpose of calculation of own funds requirements for credit risk it is desirable that the levels of the thresholds set by competent authorities remain stable over time. Therefore competent authorities should not change the levels of the thresholds unless it is necessary due to significant distortions in the default identification processes at the institutions caused by an inadequate level of the materiality threshold in the changed market or economic conditions. (10) Due to the comprehensive analysis that needs to be carried out by competent authorities in order to define the appropriate level of materiality thresholds, adequate time should be allowed to them for setting the threshold. (11) In order to apply the materiality threshold set by competent authorities, some institutions that use the IRB approach might require changes to their IRB models. Such changes are expected to be material changes to the models, requiring prior approval by the competent authority. For other institutions, the implementation of the materiality threshold set by competent authorities might be burdensome if their previous approach for determining the materiality of exposures past due is significantly different from that threshold. Moreover, in the case institutions that use the IRB approach apply the Standardised Approach to a part of their exposures on the basis of Article 148 or 150 of Regulation (EU) No 575/2013 it would be desirable to align the time of application of the new materiality threshold for all exposures of the institution. As a result, it would be desirable for competent authorities to provide, in their respective jurisdictions, longer periods for the

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE application of the threshold to certain categories of firms, based on the principle of proportionality. On the other hand, and in order to prevent excessive delays in the implementation of the threshold across the Union, such longer periods should be limited. (12) This Regulation is based on the draft regulatory technical standards submitted by the European Banking Authority to the Commission. (13) The European Banking Authority has conducted open public consultations on the draft regulatory technical standards on which this Regulation is based, analysed the potential related costs and benefits and requested the opinion of the Banking Stakeholder Group established in accordance with Article 37 of Regulation (EU) No 1093/20102 , HAS ADOPTED THIS REGULATION:
Article 1
Competent authorities shall set the threshold referred to in paragraph 2(d) of Article 178 of Regulation (EU) No 575/2013 in accordance with the conditions specified in Article 2 for retail exposures and in Article 3, for exposures other than retail.
Article 2
Conditions for setting the threshold for retail exposures

For the purposes of setting the threshold referred to in Article 1 for retail exposures,
all of the following shall apply:
(a) competent authorities shall set a single threshold for all institutions in the respective jurisdiction of the competent authority; (b) competent authorities may also set a separate single threshold for all institutions in the respective jurisdiction of the competent authority that apply the definition of default at the level of individual credit facility in accordance with the second subparagraph of Article 178 (1) of Regulation (EU) No 575/2013.
For the purposes of setting the threshold referred to in paragraph 1(a), competent
authorities shall ensure that the threshold consists of an absolute and a relative component whereby:
(a) the absolute component of the threshold is set as a limit to 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 (‘credit obligation past due’), and that such a limit is lower than or equal to 100 EUR or the equivalent of that in the relevant national currency; 2 Regulation (EU) No 1093/2010 of the European Parliament and of the Council of 24 November 2010 establishing a European Supervisory Authority (European Banking Authority), amending Decision No 716/2009/EC and repealing Commission Decision 2009/78/EC (OJ L 331, 15.12.2010, p. 12).

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE (b) the relative component of the threshold is set as a ratio, expressed as a percentage, of the credit obligation past due as referred to in point (a), versus the total amount of all on-balance sheet exposures to the obligor excluding equity exposures, and that such a percentage is equal to 1%, unless such a level does not reflect a level of risk that the competent authority considers to be reasonable in accordance with Article 4(3); (c) where there’s evidence that the relative component of the threshold at the level of 1% as specified in point (b) does not reflect a level of risk that the competent authority considers to be reasonable in accordance with Article 4(3) competent authorities shall set the relative component of the threshold at the level that is lower than or equal to 2.5%.
3. In the course of defining the limit referred to in paragraph 2, competent authorities
shall take into account that the breach of the limit results in the obligor being defaulted, in accordance with the following:
(a) where both of the limits are breached for 90 consecutive days, except for the cases referred to in point (b); (b) where both of the limits are breached for 180 days, in the case where all of the exposures included in the calculation of the credit obligation past due are exposures secured by residential or small and medium enterprise (‘SME’) commercial real estate and the competent authority has replaced the 90 days with 180 days in accordance with Article 178(1)(b) of Regulation (EU) No 575/2013 for these exposures.
4. For the purposes of setting the threshold referred to in paragraph 1(b), the threshold
shall be set according to the same criteria as those referred to in paragraph 2 and shall function in the same manner as referred to in paragraph 3, with the sole difference that ‘credit obligation past due’ and ‘total amount of all on-balance sheet exposures to the borrower excluding equity exposures’ shall refer to amounts of the credit obligation of the borrower that result from a single credit facility granted by the institution, the parent undertaking or any of its subsidiaries.
Article 3
Conditions for setting the threshold for exposures other than retail

For the purposes of setting the threshold referred to in Article 1 for exposures other
than retail, competent authorities shall set a single threshold for all institutions in the respective jurisdiction of the competent authority.
For the purposes of setting the threshold referred to in paragraph 1, competent
authorities shall set a threshold consisting of an absolute and a relative component whereby:
(a) the absolute component of the threshold is set as a limit to the sum of all past due amounts related to the credit obligation past due, and that such a limit is lower than or equal to 500 EUR or the equivalent of that in the relevant national currency; (b) the relative component of the threshold is set as a ratio, expressed as a percentage, of the credit obligation past due as referred to in point (a), versus

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE the total amount of all on-balance sheet exposures to the obligor excluding equity exposures; and that such a percentage is equal to 1%, unless such a level does not reflect a level of risk that the competent authority considers to be reasonable in accordance with Article 4(3); (c) where there’s evidence that the relative component of the threshold at the level of 1% as specified in point (b) does not reflect a level of risk that the competent authority considers to be reasonable in accordance with Article 4(3), competent authorities shall set the relative component of the threshold at the level that is lower than or equal to 2.5%.
3. In the course of defining the limit to the credit obligation past due as referred to in
paragraph 2, competent authorities shall take into account that the breach of both components of the limit results in the obligor being defaulted, in accordance with the following:
(a) where both of the limits are breached for 90 consecutive days, except for the cases referred to in point (b); (b) where both of the limits are breached for 180 days, in the case where all of the exposures included in the calculation of the credit obligation past due are exposures to public sector entities, and the competent authority has replaced the 90 days with 180 days in accordance with Article 178(1)(b) of Regulation (EU) No 575/2013 for these exposures.
Article 4
Setting the level of the threshold

In the course of setting the threshold referred to in Article 2, competent authorities
shall take into account the risk characteristics of retail exposures.
In the course of setting the threshold referred to in Article 3, competent authorities
shall take into account the risk characteristics of exposures other than retail.
For the purpose of Article 2(2)(b) and Article 3(2)(b) competent authorities shall
consider that the level of risk is reasonable where both of the following conditions are met:
(a) the level of the materiality threshold does not lead to recognition of an excessive number of defaults that result from other circumstances than financial difficulties of an obligor; (b) the level of the materiality threshold does not lead to significant delays in the recognition of defaults that result from financial difficulties of an obligor.
Competent authorities shall notify the EBA of the levels of the thresholds that they
set in their respective jurisdiction. Where competent authorities set the relative component of the materiality threshold at the level different than specified in
Article 2(2)(b) and Article 3(2)(b) they shall provide the justification for this
different level of the threshold to the EBA.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE
Article 5
Updating of the threshold
Where the absolute component of the threshold referred to in Article 2(2)(a) and Article 3(2)(a) is set in the relevant national currency other than Euro and as a result of a volatility of currency exchange rates the equivalent of either or both of those components in Euro is higher than the limit specified respectively in Article 2(2)(a) and in Article 3(2)(a) the previously set threshold should remain unchanged unless there’s evidence that the previously set threshold no longer reflects a level of risk that the competent authority considers to be reasonable.
Article 6
Application of the threshold
In the course of setting the threshold referred to in Article 1, where competent authorities define timelines after which such threshold shall apply to institutions, competent authorities shall ensure that institutions using the Standardised Approach in accordance with Part Three, Title II, Chapter 2 of Regulation (EU) 575/2013 will apply this threshold no later than by 31 December 2020.
Article 7
Entry into force
This Regulation shall enter into force on the twentieth day following that of its publication in the Official Journal of the European Union. It shall apply from [instructions to the OJ: please insert the date that corresponds to 90 days after the date of publication in the OJ]. This Regulation shall be binding in its entirety and directly applicable in all Member States. Done at Brussels, For the Commission The President [For the Commission On behalf of the President [Position]

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE
4. Accompanying documents
4.1 Impact assessment
Introduction
Article 178(6) of Regulation (EU) No 575/2013 (CRR) requires the EBA to develop draft regulatory
technical standards to specify the conditions according to which competent authorities shall set the threshold against which the institutions shall assess the materiality of a credit obligation past due (RTS). This threshold shall reflect a level of risk that the competent authorities consider reasonable.
Article 10(1) of the EBA Regulation (Regulation (EU) No 1093/2010 of the European Parliament
and of the Council) provides that when any regulatory technical standards developed by the EBA are submitted to the Commission for adoption they should be accompanied by an analysis of ‘the potential related costs and benefits’. This analysis should provide an overview of the findings regarding the problem to be dealt with, the options identified to remove the problem and the potential impacts of the proposed solutions. This section presents the Impact Assessment (IA) with a cost-benefit analysis of the provisions included in the final draft RTS. Problem definition The primary problem that the final draft RTS aim to address is the lack of common practice and variations in interpretation across Member States when competent authorities specify and institutions apply the materiality threshold. The provisions under Article 178 of the CRR on the application of the materiality threshold are stated in relatively broad terms, and therefore are open to interpretation. As the materiality threshold is directly related to the definition of default of an obligor its common and consistent application is crucial for the comparability of capital requirements. In other words, the threshold that institutions apply and which defines the materiality of credit obligation has a direct impact on the calculation of own fund requirements of institutions. This is true for all banks using the Internal Ratings Based (IRB) Approach or the Standardised Approach (SA). More precisely, the materiality threshold has an impact on SA institutions’ own fund requirements through the classification of exposures as defaulted or nondefaulted. For IRB banks the materiality threshold impacts institutions’ own fund requirements also through the internal estimates of risk parameters and the IRB shortfall computation. Further

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE sections provide a more detailed discussion on the impact of the materiality threshold on the calculation of own funds requirements of the institutions. Lack of common and consistent application of the materiality threshold may further lead to an uneven playing field across Member States and institutions. For example, two institutions located in different jurisdictions with similar risk profiles may be subject to different regulatory treatment if the materiality threshold for the definition of default is not consistent between jurisdictions. Similarly, different treatment of various entities belonging to the same cross-border groups due to different supervisory practices may lead to a significant operational burden for the group and an uneven playing field in the EU banking sector. Objectives The objective of the RTS is to establish convergence of supervisory practices regarding the application of the materiality threshold for past due credit obligations. Harmonisation of the current practices, which vary across Member States and institutions, is expected to enhance comparability of own funds and own funds requirements and to reduce the burden for crossborder institutions in complying with different regulatory frameworks. The RTS in addition aim to set the conditions for the materiality threshold in such a way as to address situations where past due credit obligations are not the result of the materialisation of credit risk but due to other circumstances. In particular, the materiality threshold should help identify such cases where the fact that the past due credit obligation is small indicates that the delay in payment is due to other than credit risk-related circumstances and eliminate the recognition of a large number of defaults that will return to a non-defaulted status in a short timeframe. At the same time the materiality threshold should allow the timely identification of default cases that are generated by the materialisation of credit risk. In particular, the materiality threshold should not prevent the identification of a default on the basis of the indicator that the payment has been delayed for more than 90 days. In the case of institutions that use the IRB Approach the objective of the materiality threshold is also to contribute to accurate estimates of the risk parameters. If the defaults that occur from circumstances other than the materialisation of credit risk are effectively cut off by the materiality threshold, then the quality of data used for modelling is expected to be higher and the model development process is expected to be more transparent, with no further data adjustments required. On the other hand, all defaults due to the materialisation of credit risk should be effectively identified to ensure a sufficient level of data quantity and the accuracy of estimates. Baseline scenario

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE The EBA conducted a Qualitative and Quantitative Impact Study (QIS) to assess the impact of the regulatory proposals to harmonise the definition of default proposed in the consultation papers on the RTS on the materiality threshold for credit obligations past due under Article 178(6) of Regulation (EU) No 575/20133 and on the Guidelines on the application of the definition of default under Article 178(7) of Regulation (EU) No 575/20134 . A total of 72 institutions participated in the study. Detailed results of the QIS are presented in a Report on the results of the data collection exercise on the proposed regulatory changes for a common EU approach to the definition of default (QIS report) published alongside these RTS on the EBA website. The QIS contains two parts, a qualitative questionnaire to gather information on institutions’ current practices and a quantitative survey aimed at quantifying the impact of the proposed technical options around the definition of default. The baseline information from the qualitative questionnaire is the benchmark for assessing the potential costs and benefits that European institutions will be subject to under the technical options. In other words, if the current practices of the institutions are the same as or similar to the elements that are specified in the RTS, the expected costs and benefits will be smaller than if the current practices are very different from the practices resulting from the policy decisions taken under the RTS. The findings of research5 conducted by the EBA among competent authorities showed that institutions in most of the Member States differentiate the threshold for retail exposures from that for non-retail exposures. For this reason the QIS analysis of practices in the area of the materiality threshold is separated between non-retail and retail exposures. This section aims to describe the current practices across the institutions with regard to the materiality threshold for retail and non-retail exposures. The results of the QIS confirm that substantial variability exists in the approaches taken across institutions in all areas related to the materiality threshold both for retail and for non-retail exposures. Section 2.6 of the QIS report presents an overview of the current practices across institutions in relation to the technical options considered in the RTS, including the structure of the threshold, the reference amount for credit obligations past due, the application of the materiality threshold and the current levels of absolute and relative thresholds as a benchmark for the cap thresholds specified in the RTS. The thresholds are defined either in absolute or in relative terms, or they involve a combination of the two criteria. The most popular structure of the threshold is a simple absolute threshold and the next most popular is the technical option proposed by the RTS, i.e. a combination of absolute
and relative thresholds where the breach of both of them triggers default. 3 https://www.eba.europa.eu/documents/10180/878549/EBA-CP-2014- 32+%28CP+on+RTS+on+Past+Due+Materiality+Threshold%29.pdf 4 https://www.eba.europa.eu/documents/10180/1198203/EBA-CP-2015- 15+%28CP+on+GL+on+the+application+of+the+definition+of+default%29.pdf 5 Surveys carried out among competent authorities in April 2014 and June 2013.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE The results of the QIS show a great range of variability regarding the levels of the thresholds in use across institutions. Almost half of institutions use only an absolute threshold and the levels of absolute thresholds range from EUR 0 to EUR 50 000. Relative thresholds, where specified, range from 1% to 10% for both retail and non-retail exposures. It is therefore necessary to limit the range of variability of the materiality threshold by eliminating outliers. However, the levels of the thresholds used by the majority of institutions are within the ranges specified in the RTS:
 In the case of non-retail exposures more than half of the institutions use a threshold that is below EUR 500. Moreover, around 60% of the institutions use a relative threshold that is below the proposed 1% level and more than 85% use a relative threshold that is below the 2.5% cap.  In the case of retail exposures half of the institutions use an absolute threshold that is below EUR 200, of which around half use an absolute threshold that is below EUR 100. Moreover around 65% of the institutions use a relative threshold that is below the proposed 1% level and more than 90% use a relative threshold that is below the 2.5% cap. Apart from different structures and levels of thresholds there are also significant differences with regard to the reference amount that is compared with the threshold and the counting of days past due. The technical option proposed in the RTS, where the reference amount is the sum of all amounts past due, is the most popular among institutions and is used by 48% and 42% of institutions for non-retail and retail exposures respectively. In some other cases the materiality of exposures past due is defined on the basis of the past due amounts with a variety of options with regard to the counting of days past due and the treatment of individual facilities (26% and 32% of institutions for non-retail and retail respectively). Only in 7% of cases is materiality defined in relation to the total obligations of the obligor for non-retail exposures, and this percentage goes down to 4% for retail exposures. In terms of the application of the threshold, 60% of the institutions assess the reference amount against the materiality threshold in an aggregate manner for all exposures to an obligor for nonretail exposures, as proposed by the RTS. For retail exposures, however, 47% of the institutions apply the materiality threshold at the individual facility level, whereas 37% of the institutions perform the assessment in an aggregate manner for all exposures of an obligor. This is consistent with the possibility, granted by Article 178(1) of the CRR, of applying the definition of default at the facility level for retail exposures. Assessment of the technical options This section presents an assessment of the technical options considered in the RTS. Under each option, the potential advantages and disadvantages of the options together with the potential costs and benefits are discussed.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE a. Definition of credit obligation past due Option Advantages Disadvantages 1 The whole single credit obligation if any part of it is past due more than 90 days (or 180 days if applicable) Very conservative approach towards larger exposures and therefore encourages close monitoring and management of exposures. Simple and easy to implement. Default identification process does not account for small credit obligations. Too restrictive in the case of large exposures. Material default is not identified if a large exposure is divided into many small exposures each falling below the threshold. It is a rare practice among Member States, i.e. only two Member States are currently using the approach. 2 The sum of all amounts past due if any of the amounts is past due more than 90 days (or 180 days if applicable) More prudent approach than option 3 that allows the default to be identified sooner because it takes into account all amounts past due even before they reach 90 days past due. Cliff effect might occur where an obligor has a very small unpaid amount on some account. In such a case a delay in payment of an instalment of a credit obligation even of one day would trigger default. This approach might raise doubts about the calculation of days past due: if the calculation of days past due starts after the materiality threshold is achieved then the cliff effect is avoided but the calculation is not compliant with contractual obligation and other purposes for which the number of days past due is used. 3 The sum of the amounts past due more than 90 days (or 180 days if applicable) The cliff effect is effectively avoided (see explanation of the disadvantages of option 2). Effective in the identification of technical defaults that result from errors in IT systems or misunderstandings with clients rather than realisation of credit risk, and therefore should enhance the quality of internal estimates by avoiding overly high cure rates. Less prudent approach in comparison with options 1 and 2. Problematic where only interest is paid on a monthly basis. It would be possible for an exposure to be past due for many months before the materiality threshold was breached. 4 The sum of all amounts past due, but the calculation of days past due starts when the materiality threshold is breached Same arguments as under option 3. Approach that is commonly used across Member States and supported in the feedback received during consultations. The outcomes are independent of the payment allocation scheme (i.e. LIFO, FIFO or any other approach). Gives similar results to option 3 under the LIFO approach and hence is more conservative. Same arguments as under option 3. The counting of days past due is not compliant with contractual obligation and other purposes for which the number of days past due is used, in particular internal monitoring of the credit portfolio. 5 The amount past due more than 90 days
(or 180 days if applicable) on an Same arguments as under option 3. Same arguments as under option 3. Reflects the risk of a facility rather than an

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE individual credit facility obligor. In the case of a large overall exposure, a material default might not be identified if it is split into several facilities. 6 The total amount past due on an individual credit facility if any part of this amount is past due more than 90 days (or 180 days if applicable) Same arguments as under option 2. Same arguments as under option 2. Reflects the risk of a facility rather than an obligor. In the case of a large overall exposure, a material default might not be identified if it is split into several facilities. Given the advantages and the disadvantages of the options, the preferred option is technical option 4. The preferred option seems to reflect the intention of Article 178 of the CRR and it is the approach that is in practice used by the majority of institutions that participated in the QIS. It is therefore expected that the application of this option would generate on aggregate the lowest costs for institutions. Moreover, as the calculation would include all amounts past due, regardless of the length of the past due period, the outcomes are in most cases independent of the payment allocation scheme (i.e. LIFO, FIFO or any other approach). As a result a large degree of harmonisation is achieved without the need to prescribe a specific payment allocation approach. b. Structure of the threshold Option Advantages Disadvantages 1 Absolute threshold Conservative approach, as all amounts above this threshold are considered material regardless of the size of the obligor or total exposure. An absolute threshold is effective in cutting off technical defaults. Simple approach, easy to implement. Does not respect proportionality: the approach is more conservative for larger exposures, where default would typically be identified immediately after 90 days past due. On the other hand in the case of small exposures default might not be identified even after many days past due. 2 Relative threshold Accounts for basic characteristics of specific obligors. Approach generous for large exposures and for obligors that use numerous facilities. In that case, failure to repay one facility might not cause the default of the obligor. 3 Combination of absolute and relative thresholds Allows for effective cut-off points for technical defaults. More risk sensitive, as it accounts for the characteristics of particular exposures. Relatively complex approach and hence more difficult to implement. 3a Breach of only one component of the threshold triggers default All exposures above each component of the threshold are considered material. More prudent approach. Some technical defaults might not be cut off, for example in the case of very small exposures such as unpaid fees the whole amount is past due and therefore default would be triggered by the relative component of the threshold. 3b Breach of both components of the threshold triggers default All exposures
below each component of the threshold are considered immaterial. Approach generous for large exposures and for obligors that use numerous facilities. In that case, failure to repay one facility might

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE Ensures that very small exposures are not taken into account in the identification of default. In line with feedback received during the consultation. not cause the default of the obligor until the relative component of the threshold is breached. Some material defaults might not be identified in a timely manner and cause underestimation of risk by the institution. The analysis of the advantages and disadvantages of the possible structures of the threshold shows that none of these approaches is free from certain weaknesses. Nevertheless, it is impossible to assess the materiality of each exposure individually and therefore some degree of simplification is inevitable. Taking into account the feedback received in the consultation process technical option 3b is selected as the preferred option; it is in practice used by almost a third of the institutions participating in the QIS. This option seems to be in line with the proposed harmonisation approach that is based mainly on a relative threshold and is the most risk sensitive, as it accounts for characteristics of particular exposures. In any case, it is possible to identify the default earlier than the breach of the materiality threshold on the basis of unlikeness to pay considerations.
c. Differentiation criteria
Option Advantages Disadvantages
1 No differentiation: single threshold criterion for all exposures Reasonable under the assumption that regardless of the type of client or product certain very small amounts are immaterial for the institution. Simple and easy to implement. Lack of risk sensitivity (see rationale for options 2 and 3). 2 Differentiation of the threshold for retail and non-retail exposures As opposed to option 1 it could be argued that from the perspective of the institution immaterial amounts are those for which the average cost of collecting would exceed the amount. However, in the case of retail exposures the amounts are individually smaller but there is a significant number of similar exposures. The collection processes are less individual and therefore the costs of recovery of a single exposure are on average lower. That would justify differentiation between retail and non-retail exposures. The average exposure values of retail and non-retail clients are significantly different. The approach is still simple and easy to implement. Classification into retail and non-retail exposures is readily available for all institutions, regardless of whether they use the The approach does not take into account differences between the product types used by retail clients (in particular mortgage loans and other loans) where the definition of default is applied at the level of the individual facility.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE IRB Approach or the Standardised Approach. 3 Differentiation of the threshold for retail and non-retail exposures with additional differentiation of product types for retail exposures Enhanced risk sensitivity by taking into account significant differences in characteristics between retail product types. In the case of retail exposures the application of the definition of default at the level of the individual facility is a common approach among institutions; therefore, the specificities of product type can be considered to be an important risk factor. The recognition of product types required under the IRB Approach and under the Standardised Approach is significantly different. The class of exposures secured by mortgages on immovable property include only those parts of exposures that are fully and completely secured by such mortgages. This class does not refer to the type of obligor and therefore may contain exposures to both retail and non-retail clients. Under the IRB Approach, exposures secured by immovable property are recognised only within the retail exposure class and they include the entire credit obligation regardless of the part of the obligation that is secured fully and completely. Those exposures are therefore incomparable. Additionally, under the IRB Approach qualifying revolving retail exposures are recognised, whereas under the Standardised Approach this category does not exist. Option 2 offers a reasonable balance between simplicity and risk sensitivity by taking into account the most important factor for differentiation, which is easy to implement and comparable for all institutions regardless of the approach used to calculate own funds requirements. The preferred option is therefore technical option 2. d. Cap threshold Option Advantages Disadvantages 1 No cap threshold No need for calibration of a cap threshold at European level. Full flexibility for competent authorities. Limited harmonisation and less comparability of own funds requirements. Possibility of significant outliers. 2 Cap threshold specified in the RTS as a specific amount (equivalent in euros) or percentage Simple and easy to apply. Enhanced harmonisation and comparability of own funds requirements by avoidance of significant outliers. Difficult to calibrate a cap threshold that would be equally suitable for all jurisdictions. In the case of non-euro area countries high volatility of foreign exchange rates may lead to non-compliance of the threshold set in a local currency. 3 Cap threshold defined as a formula that takes into account the characteristics of a certain jurisdiction

Better adjustment of the cap
threshold to the specificity of a certain jurisdiction.
Difficult to define a formula that would
take into account the most important factors.
Difficult to implement the threshold and
to monitor ex post compliance.
Possible need to change the threshold
once the cap ceases to be met.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE Option 2 is the preferred option because it allows the main goal to be served by the cap threshold to be achieved in a simple fashion and is easy to implement. The option is sensible in granting the competent authorities flexibility in calibrating the levels of the threshold to accommodate the characteristics of the banking sector and at the same time ensuring the implementation of sufficiently prudent rules and avoiding significant outliers. During the public consultation process the industry called for a unified European approach as a
part of the efforts to harmonise the definition of default. This gives a rationale for the proposed
1% unified level of the relative threshold, which, coupled with a low cap for the absolute threshold, should ensure a level playing field for institutions across the EU and will significantly increase the comparability of risk parameters. Impacts of the technical options Under the SA, Article 127 of the CRR groups the unsecured parts of defaulted exposures as a specific asset class. Within this asset class the risk weight is assigned according to the ratio between the unsecured part of the exposure value and specific credit risk adjustments. If the specific credit risk adjustments are equal to or greater than 20% of the unsecured part of the exposure value then the assigned risk weight is 100%. The assigned risk weight is 150% if the specific credit risk adjustments are less than 20% of the unsecured part of the exposure value. In other words, the lower the materiality threshold the higher the level of defaulted exposures in the asset class where higher risk weights are assigned, and higher risk weights affect the calculation of own funds requirements calculation. Under the IRB Approach the materiality threshold has an impact on the classification of exposures as defaulted (with a PD that is equal to 100%), which therefore has an impact on the calculation of expected loss and own funds requirements, which represent unexpected loss. Assets classified as not defaulted exposures are subject to a standard formula with regard to risk weight calculation and, for exposures classified as defaulted, lower risk weights are usually applied. Under the Foundation IRB (FIRB) Approach the risk weight of defaulted exposures is set to zero. However, the calculation of expected loss is based on a PD that is equal to 100%; therefore, it is much higher than if the exposure was not classified as defaulted. The lower the threshold, the higher the expected loss. If the expected loss is not fully covered by the credit risk adjustments then the difference is deducted from own funds. Under the advanced IRB (AIRB) Approach the logic is the same except that the risk weight for defaulted exposures is not zero but calculated on the basis of the LGD parameter. In addition, the definition of default, and in particular the materiality threshold that is defined under it, influences the own funds requirements of IRB institutions through the application of PD and LGD parameters. These parameters influence the risk weights applied to all non-defaulted exposures. Lower thresholds will result in a higher default rate and higher PD estimates. However, not all identified defaults generate losses, so this can decrease the LGD estimates of the AIRB

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE institutions. Therefore, lower thresholds will not always be more conservative, as they may lead to lower LGD estimates that subsequently result in lower risk weights. Furthermore, also for nondefaulted exposures, the amount of expected loss calculated based on PD and LGD parameters is compared with the value of credit risk adjustments for these exposures. If the expected loss is not fully covered by the credit risk adjustments then the difference is deducted from own funds. The quantitative part of the QIS was designed to quantify the impact of the proposed technical policy options taking into account the various dimensions of impact as described above. The analysis was performed on a sample of 64 institutions (of which 22 used the SA, 32 used the IRB Approach and 10 reported on both SA and IRB exposures). These institutions account for 44% of the total of EU institutions’ credit risk-weighted exposures6 . The estimation of the impact was performed by institutions on selected representative samples of specified portfolios of retail and corporate SME exposures. The representativeness of the samples selected for the analysis by the institutions may impact the accuracy of the analysis. It should also be noted that the subjectivity of institutions in interpreting and estimating the impact may affect the results of any analysis undertaken. Moreover, institutions were requested to provide their estimates on a best-effort basis, which could affect the accuracy of their estimates in some cases. The detailed results of the analysis are presented in Section 4 of the QIS report. In theory, the lower the threshold is, the greater the number of defaults identified. A higher level of defaults generates higher expected loss and in the case of the SA the credit risk adjustments are higher. The effect on capital requirements, however, is not straightforward and depends on the method used by the institution to calculate capital requirements:
 In the case of institutions that use the SA, a higher rate of credit risk adjustments results in lower risk weights for defaulted exposures (100% or 150%)7 . As these are the highest levels of risk weights used in most of the other exposure classes it is reasonable to expect that the lower the threshold, the higher the risk-weighted exposure amounts. In other words, the lower the materiality threshold, the higher the level of defaulted exposures in the asset class where higher risk weights are assigned, and higher risk weights are used in the calculation of own funds requirements.  In the case of institutions that use the FIRB Approach the risk weight of defaulted exposures is zero. However, the calculation of expected loss is based on a PD that is equal to 100%; therefore, it is much higher than if the exposure was not classified as defaulted. The lower the threshold, the higher the expected loss. If the expected loss is not fully covered by the credit risk adjustments then the difference is deducted from own funds. Moreover, the materiality threshold also impacts the risk weights of non-defaulted 6 ECB statistics on consolidated banking data. 7 Under the SA, Article 127 of the CRR groups the unsecured parts of defaulted exposures as a specific asset class. Within this asset class the risk weight is assigned according to the ratio between the unsecured part of the exposure value and specific credit risk adjustments. If the specific credit risk adjustments are equal to or greater than 20% of the unsecured part of the exposure value then the assigned risk weight is 100%. The assigned risk weight is 150% if the specific credit risk adjustments are less than 20% of the unsecured part of the exposure value.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE exposures through PD estimates. A lower threshold results in a higher default rate, higher PD estimates and higher risk weights for non-defaulted exposures.  In the case of institutions that use the AIRB Approach the impact on capital requirements is complex. The risk weight for defaulted exposures is not zero but calculated on the basis of the best estimate of expected loss (ELBE) and LGD in-default estimates, and should represent unexpected loss in the recovery process. It is not explicit whether the risk weight calculated in this way is higher or lower than the risk weight for non-defaulted exposures. This depends largely on the methodologies used by particular institutions. The materiality threshold also impacts the risk weights of non-defaulted exposures through PD and LGD estimates. With regard to PD it is clear that the lower the threshold is the higher are PD estimates and risk weights. In the case of LGD, however, the impact would most likely be the reverse, because a lower threshold might result in more defaults that would be cured within a short period of time. This effect would decrease LGD estimates and risk weights for non-defaulted exposures. In extreme circumstances, if there were no materiality thresholds (level of threshold set to zero) and all exposures that were past due more than 90 days were treated as defaulted, the overall impact on expected loss and capital requirements would be conservative. However, if the materiality threshold is too low then the main objective of identifying and cutting off exposures where delays in payment result from non-credit related circumstances would not be met and this would have a negative impact on the quality of internal risk estimates. In that situation, institutions might have an incentive to perform additional adjustments and use cleaning techniques on historical data. The quantitative analysis of the QIS is therefore performed separately for IRB and SA institutions. In the case of the SA the analysis is based on the expected reclassifications of exposures to and from the exposure class ‘exposure in default’. In this respect, the lower the materiality threshold, the higher the level of defaulted exposures in the asset class to which higher risk weights are assigned, and higher risk weights adjust the own funds requirements calculation. For IRB institutions, on the other hand, the analysis of the impact is based on the estimated changes in five simplified risk parameters, namely default rate (DR), cure rate (CR), recovery rate (RR), share of defaulted assets within a given type of exposure and ELBE for defaulted exposures. Based on these parameters the impact on risk-weighted assets, expected loss and own funds requirements is calculated according to some simplifying assumptions. QIS results: calibration of the relative threshold to 1% The scenario specified for the purpose of the QIS corresponds to the
provisions included in the RTS regarding the reference amount and the application of the materiality threshold, where:
 The reference amount for the threshold (credit obligation past due) is defined as the sum of all amounts past due, and the counting of 90 days (or, where relevant, 180 days) begins at the moment when this amount breaches the threshold.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE  For non-retail exposures the threshold is applied at the obligor level; for retail exposures the threshold is applied at the obligor or facility level depending on the level of application of the definition of default. With regard to the structure and the level of the materiality threshold the QIS tests a scenario that differs from the proposals of the RTS, in particular:
 For retail exposures the scenario of a simple absolute threshold level of EUR 200 is tested against the requirements included in the RTS of a combination of an absolute threshold capped at EUR 100 and a unified relative threshold suggested at 1% and capped at 2.5%;  For non-retail exposures the scenario assumes the combination of an absolute threshold level of EUR 1 000 and a relative threshold level of 2.5% against the requirements included in the RTS of a combination of an absolute threshold capped at EUR 500 and a unified relative threshold suggested at 1% and capped at 2.5%. These differences allow us to use the QIS results, splitting them between retail and non-retail exposures, in order to arrive at an approximate estimation of the impact for the two bounds of the threshold. The 0% lower bound of the threshold is in fact the main driver for capital impact on all retail exposure classes, where considering only an absolute threshold is equivalent to applying a 0% relative threshold. The 2.5% upper bound of the threshold is the driver of the impact for the non-retail exposure class. According to the results of the QIS there would be a small impact on banks that use the Standardised Approach, i.e. a slight decrease in the capital adequacy ratio (0.005 p.p.) caused by an increase in risk-weighted exposure amounts (RWA) by 0.031 p.p.. No particular difference is observed between the impact on RWA and that on capital ratios coming from the retail and corporate exposure classes. Thus the differences in impact of the two scenarios tested for retail and non-retail exposures are evaluated mainly on the basis of the QIS data for IRB banks. The impact observed in the QIS results for non-retail exposures (corporate SME) gives, as specified above, the impact of the upper bound of the threshold criteria specified in the RTS. The effect of the 2.5% relative threshold (acting as the cap in the RTS) is evaluated in combination with the breach of an absolute threshold of EUR 1 000. While it is true that the absolute threshold tested (EUR 1 000) is above the maximum level specified in the RTS (EUR 500) this is considered to have a limited impact. A EUR 1 000 absolute threshold combined with the proposed relative threshold of 2.5% would only come into play for credit obligations below EUR 40 000. This is also true for the maximum absolute level of EUR 500, which in combination with the suggested 1% relative threshold will come into play for credit obligations below EUR 50 000. As a consequence the QIS results can be read as indicating that by applying the conditions for the materiality threshold specified in the RTS the decrease in DR and CR and the increase in LGD should not be greater than those observed for corporate SME exposures. The results of the QIS for corporate SME exposures show slight decreases in the default rate (0.06 p.p.) while the LGD increases by 1.36 p.p. (mainly driven by a decrease in the cure rate by

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE
2.21 p.p.), suggesting that most banks currently adopt stricter approaches8
. The combination of the two offsetting effects on PD and LGD produces a slight increase in RWA (0.133 p.p.) and a slight increase in the capital adequacy ratio (by 0.024 p.p. based on own funds and by 0.153 p.p. based on CET1), as shown in Table 1.
Table 1: Overall capital impact of the materiality threshold (IRB)
∆𝐶𝐶𝐶𝐶𝑅𝑅𝑂𝑂𝑂𝑂 ∆𝐶𝐶𝐶𝐶𝑅𝑅𝐶𝐶𝐶𝐶𝐶𝐶1 %∆RWA All exposures -0.062% -0.069% 0.681% Retail -0.019% -0.031% 0.626% Corporate SME 0.024% 0.153% 0.133% ∆𝐶𝐶𝐶𝐶 𝑂𝑂𝑂𝑂– percentage points change in total capital ratio ∆𝐶𝐶𝐶𝐶 𝐶𝐶𝐶𝐶𝐶𝐶1– percentage points change in CET1 capital ratio %∆RWA – relative change in RWA The figures shown are weighted averages, where we used as a weight the sum of the institutions’ total exposures in the exposure classes under consideration in the QIS. Regarding retail exposure classes9 the QIS tested the application of a simple absolute threshold that would in practice mean the application of a 0% materiality threshold. The results of the QIS for retail exposures could therefore be used to provide an overview of the possible impact of the lower bound of the threshold as specified in the RTS. The effect of the ‘0% relative threshold’ is evaluated in combination with the breach of an absolute threshold of EUR 200, which is above the maximum level proposed in the RTS (EUR 100). This level of the absolute threshold in combination with the 1% relative threshold would be binding only for exposures below EUR 10 000. Hence it could be argued that for most exposures the increase in DR and CR and the decrease in LGD should not be greater than those observed for retail exposures in the QIS. The results of the QIS show a slight increase in DR and a slight decrease in LGD for all retail exposure classes except for other retail SME exposures, where an increase in both DR and LGD is observed, and qualifying retail revolving exposures10. The combination of the two offsetting effects on PD and LGD produces an increase in RWA (0.626 p.p.) and a corresponding slight decrease in the capital adequacy ratio (by 0.019 p.p. based on own funds and by 0.031 p.p. based on CET1), as shown in Table 1. The application of a unified relative threshold at 1% is expected to lead to average levels of risk parameters remaining similar to current levels; however, the range of necessary adjustments may differ between individual institutions. The overall impact on capital and RWA will most probably lie somewhere between the impacts for retail and non-retail exposure classes. In general, the overall level of capital requirements should not change significantly; however, in the case of 8 The results of the qualitative survey in fact show that around 40% of institutions adopt only an absolute threshold, and more than half of them have absolute threshold levels that are below the tested EUR 1 000 level. 9 QIS retail exposure classes under consideration are: other retail SME and non-SME, qualifying retail revolving exposures, and secured by immovable properties SME and non-SME. 10 The results for qualifying retail revolving exposures are based on a very small sample of observations and may be biased by data quality issues; the analysis will therefore focus on all other retail exposure classes.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE outliers the individual impact may be more significant and may lead to either an increase or a decrease in RWA in individual institutions. In consequence it is expected that risk parameters and capital requirements should be much more comparable between institutions, while at the same time their average levels should not change significantly. Costs for institutions The baseline scenario shows that substantial variability exists in the approaches taken across institutions. In fact, under each technical option some regulatory changes will have to be introduced by Member States, either with regard to the level of the threshold, its structure, its application, the reference amount or the approach towards the assessment taken by institutions. Currently most institutions already use some materiality threshold. The thresholds are set either on the basis of regulations imposed by a competent authority or on the basis of the institution’s own analysis, subject to assessment by a competent authority on an individual basis. Due to the fact that the range of practices with regard to the application of the threshold is very wide most institutions will have to introduce some changes. Changes to the level and in particular the structure of the threshold may have a significant impact on the operations of institutions. The impact and costs for particular institutions will depend on the currently implemented thresholds as well as on the approach used in the calculation of own funds requirements. The QIS generally shows that the introduction of the harmonised definition of the materiality threshold specified in the RTS will lead to a modest capital increase, corresponding to an overall reduction of the capital adequacy ratio of around 0.005 p.p. based on own funds (0.003 p.p. based on CET1) for the sample of SA institutions, and of around 0.062 p.p. based on own funds (0.069 p.p. based on CET1) for the sample of IRB institutions. However, the QIS could not capture the whole spectrum of implementation costs for institutions. Any changes in the materiality threshold will affect in particular those institutions that use the IRB Approach. The risk parameters are estimated on the basis of historical data collected with the assumption of a certain materiality threshold. The consistency of the historical data with the definition of default is crucial for the correct calibration of risk parameters. All historical data will therefore have to be adjusted to the new threshold and the parameters will have to be recalibrated to reflect the current definition of default. The adjustment of data and recalibration of risk parameters may impose a significant operational burden on banks that use the IRB Approach. In particular for those institutions that use numerous rating systems and where the concept of the threshold will change significantly, the process of implementing the necessary adjustments may be
costly and time consuming. In the case of institutions that use the SA the impact of the change in the materiality threshold will be relatively less significant as there will be no need to adjust historical data unless the institution decides to apply to use the IRB approach in the future. Nevertheless, in the case of those institutions for which the concept of the materiality threshold will change significantly the costs may still be material. The change in the threshold may require changes to current risk

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE management processes and possibly also to IT systems used to collect data and calculate capital requirements. Costs for national supervisory authorities
Article 178(2)(d) of the CRR requires national supervisory authorities to define a threshold that
reflects a level of risk that is considered to be reasonable, and they need to account for various considerations arising from the regulatory framework. The analysis of the various types of exposures, obligors and products offered by institutions will require the collection of relevant data and significant analytical work. In order to grant the supervisory authorities sufficient time to conduct this analysis the date of application of these RTS is planned to be postponed by three months. It is however expected that the analysis will be a one-off assignment. Modifications of the level of the threshold should not be introduced too frequently, to avoid regulatory uncertainty for institutions. An additional impact for national supervisory authorities will result from applications from institutions that have been granted permission to use the IRB Approach for material changes to rating systems related to the change in the materiality threshold for the definition of default. Competent authorities will have to review and agree on individual implementation plans with these institutions and later verify the timely implementation of those plans. There are currently 131 banking groups (counted as per the highest EU level of consolidation) that apply the IRB Approach. If it is assumed that the majority of the institutions belonging to these groups will have to seek the approval of a competent authority for material changes to rating systems that result from the modification of the definition of default, granting such approvals in a timely manner may cause a significant operational burden for national supervisors. Benefits The results of the QIS indicate that there is significant variability of practices in the area of the materiality threshold, in particular with respect to both the structures and the levels of the materiality thresholds in use. By establishing harmonised criteria for setting the materiality threshold for past due exposures, greater comparability of own funds requirements for credit risk will be ensured. The proposed harmonisation approach defined in the RTS is based mainly on the level of the relative threshold that is uniformly applied to retail and non-retail exposures. The proposed level of 1% is considered to rightly balance the benefits of timely recognition of default (maximised with a 0% threshold) and the cost of identifying the default too early in terms of the impact on PD and LGD parameters, and in particular on cure rate (minimised by the upper bound of the relative threshold of 2.5%). The common unified threshold was also advocated for by several industry respondents during the RTS consultation period; they pointed out that the RTS should not propose overly complex and varied rules even if simplification would lead to lesser accuracy. Such varied setting is more

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE operationally costly than a warranted unified level, which would bring more certainty to the operation of institutions. Moreover, as the change in the materiality threshold, which is part of the definition of default, leads to material changes to rating systems and might require adjustments to historical data, the level of the threshold should be relatively stable over time. Most institutions will have to implement some changes that will result from the changes in the concept and structure of the materiality threshold. Therefore, the RTS proposal of aligning and harmonising the threshold level will avoid causing an additional burden on the industry where any future change in the threshold would lead to the associated implementation costs being incurred again. Hence the proposed harmonisation to a 1% level seems to fit the objective of simplicity and harmonisation as well as increasing the comparability of own funds requirements. Having a unified threshold level as well as aligning practices regarding other aspects of the materiality threshold such as its structure, reference amount and application will also reduce the administrative and operational burdens for cross-border institutions in complying with different regulatory frameworks in different Member States. Implementation of the changes Detailed explanations on how to carry out the implementation of the provisions included in these RTS and other regulatory products in the most efficient manner are included in the EBA’s opinion on the implementation of the regulatory review of the IRB Approach, published on 4 February 201611. The final deadline for the implementation of all the changes resulting from the review of the IRB Approach has been specified as end-2020. It is expected that this is a sufficiently long period to allow the most cost-efficient implementation, in particular by avoiding multiple applications for material changes to rating systems. The EBA’s opinion is based on Article 146 of the CRR and is addressed only to institutions that use the IRB Approach. In the case of institutions that use the SA the same implementation deadline has been specified directly in the RTS. As a consequence SA institutions will have sufficient time to adequately plan and perform the implementation of the changes and IRB institutions will be able to align the timelines for implementation of the changes in the definition of default for portfolios covered by the IRB Approach and for those under roll-out plans and permanent partial use of the SA. 11 https://www.eba.europa.eu/documents/10180/1359456/EBA-Op-2016-01+Opinion+on+IRB+implementation.pdf

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE
4.2 Feedback on the public consultation
The EBA publicly consulted on the draft proposal contained in this paper. The consultation period lasted for 3 months and ended on 31 January 2015; 31 responses were received, of which 22 were published on the EBA website. This paper presents a summary of the key points and other comments arising from the consultation, the analysis and discussion triggered by these comments and the actions taken to address them if deemed necessary. In many cases several industry bodies made similar comments or the same body repeated its comments in response to different questions. In such cases, the comments and the EBA’s analysis are included in the section of this paper where the EBA considers them most appropriate. Changes to the draft RTS have been incorporated as a result of the responses received during the public consultation. Summary of key issues and the EBA’s response Responses to the consultation showed a general approval with regard to the need to clarify and harmonise the definition of default in order to decrease RWA variability. The respondents asked for additional clarifications and requested that the rules be unequivocal in order to avoid varying interpretations. The EBA agrees with this principle and provides additional clarifications either in these RTS or in other regulatory products as indicated in the feedback table. The RTS contain detailed rules on the specification and application of the materiality threshold, which should prevent differences in interpretation. In addition, several respondents asked for a single threshold across the entire EU, with no discretion allowed at jurisdiction level. Although full harmonisation was not possible due to the discretion given by the CRR to competent authorities to set the threshold at the level that they consider reasonable, the suggestions of the respondents have been partly included by setting in the RTS a recommended unified level of the relative component of the threshold. The majority of respondents commented on the proposed structure of the threshold. There was a clear preference among the respondents for the absolute and relative components of the threshold to be combined in such a way that only breach of both of those components would trigger default. These suggestions have been taken into account and relevant changes have been introduced to the final draft RTS. Many respondents commented on the counting of days past due for the purpose of the application of the materiality threshold. There was general agreement among those respondents that the counting of days past due should start only after the materiality threshold was breached. This suggestion was also incorporated in the final draft RTS. This decision was additionally

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE motivated by the fact that under this interpretation the application of the materiality threshold and identification of default will be to a large extent independent of the payment allocation scheme used by the institution (i.e. FIFO, LIFO or any other approach). It has been agreed that this solution will be most effective in ensuring comparability. Requirements regarding the allocation of payments are often included in national legislation, including consumer protection regulations, as well as in contracts with clients. Direct specification of requirements on payment allocation would not be possible in these RTS, as this was not included in the mandate for these RTS. The respondents expressed various views regarding the maximum levels of the threshold and their granularity. The final calibration included in the final draft RTS takes into account these opinions but also the modified structure and application of the threshold, as well as the results of the qualitative and quantitative impact study that has been performed in order to inform the final decisions on the definition of default. In addition, many respondents requested greater flexibility in the application of the threshold and asked in particular for the possibility to use lower thresholds than those set by competent authorities. It has been clarified that this is already possible, as a default should be considered to have occurred with regard to a particular obligor when the institution considers that the obligor is unlikely to pay its credit obligations to the institution. Therefore, a days past due criterion should be understood as a backstop; institutions may recognise defaults earlier. In particular, they can use a lower threshold if they can demonstrate that it is a relevant indication of unlikeliness to pay and will not lead to an excessive number of cures or a reduction in capital requirements. Finally, many respondents were concerned about the costs of implementation, which are expected to be significant in particular in the case of institutions that use the IRB Approach and pointed out that a long implementation period will be necessary in order to introduce the required changes to IT systems, internal processes and rating systems. The EBA understands these concerns and envisages a phase-in approach with sufficiently long implementation period. These expectations have been expressed in detail in the EBA’s Opinion on the implementation of the regulatory review of the IRB Approach published on 4 February 2016.

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE Summary of responses to the consultation and the EBA’s analysis General comments Aspects related to the harmonisation of the definition of default Responses to the consultation showed a general approval with regard to the need to clarify and harmonise the definition of default in order to decrease RWA variability. Around a quarter of respondents agreed that there was a need for further harmonisation of the definition of default, and for clarification of concepts such as ‘multiple defaults’, ‘transfer of information between separate legal entities’ and ‘the treatment of guarantees’. The EBA appreciates the industry’s support for the work on the definition of default. These RTS are based on a mandate defined in Article 178(6) of the CRR and can only include conditions for competent authorities to set the materiality threshold referred to in paragraph 2(d) of this article. However, the EBA is also planning to specify broader guidelines on the application of the definition of default as mandated in Article 178(7) of the CRR, where more detailed clarification will be provided on other aspects of the definition of default. In addition, the EBA will clarify aspects related to the estimation of risk parameters, including aspects such as multiple default and treatment of guarantees, in separate dedicated guidelines. N/A Several respondents asked for a single threshold across the entire EU with no discretion allowed at jurisdiction level. One respondent also asked for clarification about the treatment applied to banks under ECB supervision. The RTS specify the conditions for competent authorities in as detailed a manner as possible and recommend a unified level of the relative component of the threshold. Furthermore, the levels of both the relative and absolute thresholds are limited by a specified cap. The objective was to reduce variability in the application of the materiality thresholds across jurisdictions as far as possible. Nevertheless, in accordance with
Article 178(2)(d) of the CRR the threshold shall
reflect the level of risk that the competent authority
Article 2(2)(c)
Article 3(2)(c)

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE considers to be reasonable. Therefore, the responsibility for setting an appropriate level of the threshold is with the competent authorities. Due to the differences across EU countries with regard to economic conditions and purchasing power, it would be difficult to find one level of absolute threshold that would reflect a reasonable level of risk in all jurisdictions. Additionally, different currencies would result in additional operational difficulties. Therefore, at this stage it was decided not to unify the level of the absolute threshold across the EU. With regard to banks under ECB supervision the competent authority to set the threshold is the ECB. Around a third of respondents argued that the final RTS should try to reduce the burden of implementation. One proposal was to allow institutions which apply a more conservative treatment (i.e. a lower threshold) not to change their model or their definition of default. One respondent argued that competent authorities should allow banks to employ the absolute threshold only. The EBA agrees on the need to reduce the burden of implementation. Furthermore, it acknowledges that institutions should be allowed to put in default an obligor before breach of the threshold of their jurisdiction. This is already possible, as under
Article 178(1)(a) of the CRR, a default shall be
considered to have occurred with regard to a particular obligor when the institution considers that the obligor is unlikely to pay its credit obligations to the institution. Therefore, institutions may recognise defaults on the basis of a lower threshold if they can demonstrate that this lower threshold is a relevant indication of unlikeliness to pay and will not lead to an excessive number of cures or a reduction in capital requirements. N/A

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE Several respondents proposed allowing the use of expert judgement to classify an exposure as defaulted and to exclude technical defaults. Some respondents also proposed defining technical defaults in order to exclude them explicitly. One respondent provided an example from the leasing industry of the situation where payment is suspended by the obligor due to the failure of the leasing object rather than financial difficulties. It was argued that in such a case expert judgement would have to be applied. Expert judgement used to exclude technical default is out of the scope of these RTS. However, the EBA will clarify the definition and treatment of technical defaults in the guidelines on the application of the definition of default as mandated in Article 178(7) of the CRR. Further clarification in the EBA guidelines on definition of default Some respondents pointed out the need to be in line with future Basel developments and current work on the accounting framework (IFRS 9) and non-performing exposures (NPE). The EBA is monitoring closely and taking into account developments in the global regulatory framework as well as in the accounting framework. The materiality threshold specified in these RTS will be relevant also for the recognition of NPE. N/A Requests for additional clarification Some respondents requested that the rules be very clear in order to minimise the number of interpretations and suggested providing some examples. In particular, some clarifications were requested concerning the computation of the absolute and relative (numerator and denominator) threshold. One respondent asked for an increase in disclosure in order to provide information on how banks interpret the RTS. The EBA agrees that the rules should be unambiguous and therefore the RTS specify the conditions for competent authorities in as detailed a manner as possible. In particular, the computation of the absolute and relative thresholds has been further clarified. The EBA expects that these RTS will lead to a unified interpretation of the relevant requirements of the CRR.
Article 2 and 3
Two respondents asked for the recitals to be more precise in relation to the definition of the retail category (Article 123 for the Standardised Approach and
Article 147(5) for the IRB Approach).
Recital 3 clarifies the application of the definition of retail exposures. Recital 3

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE Many respondents asked for clarification on the level of application of the materiality threshold with regard to total credit obligations. The level of application of the materiality threshold is specified in accordance with Article 178(1)(b) of the CRR, i.e. it refers to all credit obligations to the institution, the parent undertaking or any of its subsidiaries.
Article 2(2)(a)
One respondent asked if the amount past due were to be computed according to a FIFO or a LIFO approach. The specification of the payment allocation scheme to be applied by the institutions is outside the mandate of these RTS. However, the specification of the calculation of the days past due for the purpose of the materiality threshold has been changed in accordance with the feedback received during the consultation period. It is specified in these RTS that the materiality threshold should apply to the total amount past due, and that the counting of 90 days or, where relevant, 180 days should commence only when the materiality threshold is breached. In comparison with the initial proposal included in the consultation paper, under this interpretation the payment allocation scheme will have much lesser impact on the identification of default and hence the harmonisation of this element is considered not necessary in order to achieve sufficient comparability of the default identification processes.
Article 2(3)
Article 3(3)
One respondent asked for a general clarification on the notion of ‘amounts past due’ for the factoring industry. Further clarifications on how to apply the materiality threshold in the case of factoring will be provided in the guidelines on the application of the definition of default as mandated in Article 178(7) of the CRR. Further clarification in the EBA guidelines on definition of default

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE Application of the threshold Around a quarter of respondents asked for the possibility to compensate for unused credit lines with overdue amounts, as this would be closer to risk management practices and would help prevent regulatory arbitrage. It is expected that when amounts are past due the bank and its client will communicate in order to optimise the repayment of the past due amounts. In particular, available credit lines can be used to repay the past due amounts on other credit facilities of this obligor. If those amounts remain unpaid for 90 days after the amounts past due have become material then a default should be triggered. Therefore, no compensation should be allowed between unused credit lines and past due amounts. No change Some respondents asked for clarification of the link between the denominator of the relative threshold and its accounting treatment. In particular, some clarification was required on the treatment of off-balance-sheet exposures (e.g. unused credit lines). Two respondents argued that the denominator of the relative threshold should include unused amounts of credit facilities (‘total amount of credit obligation’ versus ‘carrying amount’). It has been clarified in the RTS that for the purpose of the relative threshold only on-balance-sheet exposures should be taken into account in the denominator. As explained above, unused credit lines can be used to pay the past due amounts on other credit facilities, but they should not be included in the calculation of the relative threshold. The relative threshold should be calculated in a consistent way. The numerator includes all past due amounts. Therefore, as the unused parts of credit facilities cannot be past due, only on-balance-sheet exposures should be included in the denominator.
Article 2(2)(b)
Article 3(2)(b)
Several respondents disagreed with taking into account only the part of the amount that is more than 90 days past due (and not the overall amount if a part of it is more than 90 days past due). There was general agreement that the counting of days past due (DPD) should start only after the materiality The suggestion received from the respondents has been taken into account and the relevant changes have been introduced to the text of the RTS. The proposed approach has the additional advantage that the results of the application of the materiality threshold will be to a large extent independent of the payment allocation scheme used by an
Article 2(3)
Article 3(3)

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE threshold was breached. institution (i.e. LIFO, FIFO or any other approach). One respondent believed that the counting of DPD should be stopped if one payment was made. Another respondent also suggested not recognising a default of an obligor if a new payment was made. An obligor should be considered defaulted whenever the materiality threshold has been breached for 90 or, where applicable, 180 consecutive days, regardless of whether any partial payments have been made during that time. More detailed criteria on when the obligor can be reclassified to nondefaulted status will be specified in the guidelines on the application of the definition of default as mandated in Article 178(7) of the CRR.
Article 2(3)
Article 3(3)
Procyclicality
One respondent argued that the requirements proposed in the consultation paper were procyclical. It referred in particular to the proposal that default be identified after only one component of the threshold was breached, as this rule would not be flexible enough to react to changes in the economic cycle in terms of credit supply. Taking into account the feedback received during the consultation period it is currently proposed that default should be identified after both the relative and the absolute components of the threshold are breached.
Article 2(3)
Article 3(3)
Responses to questions in Consultation Paper EBA/CP/2014/32 Question 1. Do you agree with the approach proposed in the draft RTS (option 1) that default should be recognised as soon as one of the components of the threshold (absolute or relative limit) is breached? Or would you rather support the alternative option, i.e. recognition of default after both thresholds are breached (option 2)? Preferred option There was general agreement that option 2 is the preferred option. A majority of respondents claimed, often providing illustrative examples, that the relative threshold should be binding for large exposures and the absolute threshold should be binding for small exposures (i.e. the absolute threshold should act as a floor). Some The EBA agreed with many of the arguments provided by the respondents and introduced relevant changes to the draft RTS to reflect the support for option 2. However, it has to be mentioned that, although the proposed rules are not expected to lead to
Article 2(3)
Article 3(3)

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE respondents also pointed out that the absolute thresholds are useless for large exposures due to the criterion of ‘unlikeliness to pay’. An analysis provided by one of the respondents showed that the application of option 1 would result in a greater increase in exposure amounts in default than the increase in the number of defaults. Some respondents argued that option 1 would increase the number of NPE and thus the number of ‘false positives’ (technical defaults). Three respondents believed that the cure rate should be minimised because it was directly linked to technical defaults. It was suggested that the rules should aim to minimise the change in the number of NPE across the EU, which would not be achieved under option 1. Some banks were concerned about the increase in administrative costs related to the treatment of a higher number of defaulted exposures. significant changes in capital requirements at EU level, the harmonisation of the threshold may lead to either an increase or a reduction in capital requirements and NPE in some individual institutions. These adjustments are necessary in order to achieve greater comparability across institutions. Regarding the objective of the threshold, it has to be noted that it cannot be set with the sole objective of minimising the cure rate, as it also has to ensure that real defaults will be detected in a timely manner. Other proposals Five respondents believed that the relative threshold should be deleted at least for retail exposures. One reason is that it increases the complexity of the framework. One respondent argued that competent authorities should also allow banks to employ the absolute threshold only. This proposal has been taken into consideration, but it was decided that the final proposal should be based on option 2, which is considered a good compromise between the wide varieties of European practices. Adoption of a different structure of the threshold for retail and non-retail exposures would lead to cliff effects in the case of SME exposures which would move between the retail and corporate exposure classes.
Article 2(3)
Article 3(3)
Some respondents also pointed out that the relative threshold might not be effective when additional Where the obligor has been granted a credit line it may be used for the purpose of the repayment of No change

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE amounts were drawn by the obligor. In that case the risk of the bank clearly increases but the relative threshold also increases. Therefore, the recognition of default might be avoided not by repayment of the past due amounts but by granting additional credit. the other credit facilities. This is considered a part of the ongoing management of the obligations. However, it is also expected that where the institution sees unlikeliness to pay on the part of the obligor then the credit line will not be increased and may not even be extended to the subsequent period. The increase of the credit line in that situation could only increase the potential losses of the institution. It is expected instead that in a period of financial difficulties on the part of an obligor the institution will be inclined to restructure the obligations. This aspect will be addressed in the guidelines on the application of the definition of default as mandated in Article 178(7) of the CRR. Around a quarter of respondents pointed out that the limits of absolute thresholds cannot be constant over time, as they are disturbed by the fluctuation of exchange rates and are not sufficiently predictive with regard to the default of obligors. Therefore, it was argued that only relative thresholds should be used. As the objective of the materiality threshold is to eliminate from the recognition of default exposures where the small amount of the past due obligation indicates that the delay is not related to financial difficulties on the part of the obligor, it is considered important to maintain the reference also to an absolute threshold. However, the issue of the fluctuation of exchange rates has been addressed in
Article 5 of the RTS.
Article 5
Question 2. Do you agree with the proposed maximum levels of the thresholds? Relative threshold Many respondents argued that the maximum relative threshold proposed in the consultation paper at the level of 2% should be increased; the proposals ranged from 2.5% to 5%. However, at the same time it was argued by The proposals received in the consultation process were carefully considered. The final calibration took into account that the final draft RTS are based on option 2 as described in question 1, so that for larger
Article 2(2)
Article 3(2)

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE many institutions that they should be allowed to use lower thresholds than those set by the competent authorities. Some respondents agreed on the level proposed in the consultation paper. One respondent believed that the thresholds should be decreased given that they apply only after the exposures are already 90 days past due. It has to be noted that the appropriate level of the threshold is highly dependent on the option that has been chosen in question 1. Unfortunately, it was not always clear with regard to which option scenario the comments were made. exposures the relative threshold will have, in practice, a binding effect. Moreover, the results of the QIS were taken into account, including the range of current practices and the estimated impact. The maximum level of the relative threshold has been moderately increased to 2.5%. In addition, a unified recommended level of the threshold of 1% has been proposed in the RTS with the possibility for the competent authorities to deviate from this level up to 2.5%. Absolute threshold Many respondents argued that the maximum absolute thresholds proposed in the consultation paper at the level of EUR 200 for retail exposures and EUR 500 for all other exposures should be increased. Some of those respondents linked the need to increase the threshold to the adoption of option 1 in question 1 in order to address large exposures. They argued that this would not be necessary if option 2 were chosen instead. There were a few proposals to increase the absolute threshold for retail exposures, ranging from EUR 250 to EUR 100. For nonretail exposures the proposals to increase the threshold ranged from EUR 1 000 to as high as EUR 10 000, with EUR 1 000 being the most frequent proposal. Several respondents agreed with the proposed levels of maximum thresholds. The proposals received in the consultation process were carefully considered. The final calibration took into account that the final draft RTS is based on option 2 as described in question 1. As under option 2 the absolute threshold serves only as a floor to the relative threshold, it is considered important to keep it at a low level. In addition, an adequate relationship between the maximum thresholds for retail and all other exposures should be preserved. Moreover, the results of the QIS were also taken into account in the final calibration. The final proposal for the maximum thresholds is EUR 100 for retail exposures and EUR 500 for all other exposures.
Article 2(2)
Article 3(2)

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE Some respondents noted that the proposed levels of maximum thresholds are higher than the current practices and therefore could be decreased. It was also mentioned that the proposed levels of the thresholds should reflect adequately the differences between retail and all other exposures and that the distance between the maximum thresholds proposed in the consultation paper might not be sufficient. Again, it has to be noted that the appropriate level of the threshold is highly dependent on the option that has been chosen in question 1. Unfortunately, it was not always clear with regard to which option scenario the comments were made. Other proposals A few respondents proposed that the threshold should be linked to the instalments or monthly repayment amounts. One respondent argued that the default should only be triggered when three monthly repayments had been missed. The policy decisions included in the RTS are based on the requirements of Article 178 of the CRR and in particular on the provision of Article 178(1)(b) that refers to 90 or, where applicable, 180 days past due. The provisions included in the RTS have to be universal and applicable to all types of credit obligations. As not all credit products are based on monthly payments the proposal to link the threshold to monthly repayments could not be incorporated. In addition, even in the case of products with monthly instalments, the non-payment of three instalments would not be fully consistent with the 90 days requirement of the CRR. No change Granularity Several respondents argued that the granularity of the materiality threshold should be increased, especially to The respondents suggested many different proposals for additional dimensions of granularity No change

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE take into account specificities related to:

central government and central banks, regional
governments or local authorities, and public sector entities;
consumer credit, mortgage loans;
asset finance (leasing, hire-purchase, credit
consumption, commercial loans);
Some respondents pointed out the need for the level of granularity to be in line with the exposure classes as defined in the CRR. Some respondents proposed distinguishing the threshold for large corporates which operate across different jurisdictions. Other respondents proposed distinguishing the thresholds for large and small non-retail exposures. One respondent proposed distinguish the relative threshold for retail and non-retail exposures. Several respondents argued that the level of granularity should be decreased if option 2 is chosen, as two thresholds provide enough sensitivity to the different kinds of exposures. One respondent agreed with the level of granularity proposed in the consultation paper. It has to be noted that the appropriate level of granularity of the thresholds is highly dependent on the option that has been chosen in question 1. Unfortunately, it was not always clear with regard to which option scenario the comments were made. for the materiality threshold but all of them seem problematic in terms of consistent implementation and would significantly increase the complexity of the RTS. With regard to the proposals based on the types of obligors and exposure classes it has to be kept in mind that the RTS is applicable to both the IRB Approach and the Standardised Approach. As the exposure classes for these approaches are defined in a completely different manner a consistent application of the threshold would not be possible. Regarding the proposals based on the types of products the following aspects have to be taken into account:
the definition of default is applied in general at the
obligor level and therefore the materiality threshold should also apply at the obligor level; the only exception has been defined for retail exposures, where institutions have a choice whether to apply the definition of default at the obligor or at the facility level;
there is no uniform taxonomy of the types of credit
products across the EU, the classification may be based on a specific business model and may not be stable over time in a dynamic market circumstances; hence the application of the materiality thresholds based on the type of product would not be comparable across institutions and specification of a detailed taxonomy of the types of products could have the unintended consequence of hindering

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE certain types of business models. Taking into account the above considerations it was decided that the level of granularity should remain unchanged from that proposed in the consultation paper. Question 3. How much time is necessary to implement the threshold set by the competent authority according to this proposed draft RTS? Given current practices, what is the scope of work required to achieve compliance? One third of respondents explicitly stated that two years was not sufficient for the implementation of the changes, without taking into consideration the need for data adjustments. At least half of them thought that three years was enough, taking into account that supervisory approval may take up to one year. The concerns of the respondents have been taken into account. The expectations with regard to the implementation of the changes in the definition of default have been expressed in the EBA Opinion on the implementation of the regulatory review of the IRB Approach published on 4 February 2016. The implementation of the changes should be based on individual timelines agreed between an institution and its competent authority with the final deadline at the end of 2020. This timeframe should include in particular the time necessary for supervisory assessment and approval of the changes. As this EBA opinion refers specifically to the IRB Approach the same deadline for the institutions that use the Standardised Approach has been included in
Article 6 of the RTS.
Article 6
A majority of respondents believed that it will not be possible to adjust all historical data in order to recalibrate risk components. In particular, it will be difficult for conversion factors and LGD estimates. Therefore, those respondents argued that they needed five to seven years It has been specified in the EBA Opinion on the implementation of the regulatory review of the IRB Approach published on 4 February 2016 that it is not necessary to adjust all historical data in order to recalibrate risk parameters. Where appropriate Further clarifications in the EBA Opinion on the implementation of the regulatory

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE more in order to collect new data to recalibrate their models. adjustments will not be possible institutions may apply an adequate margin of conservatism to address this additional uncertainty of the estimates until a sufficient set of data with a homogeneous definition of default is reached. review of the IRB Approach Around a third of respondents argued that there should be a ‘grandfathering’ for banks using internal models. As explained above, in accordance with the EBA Opinion on the implementation of the regulatory review of the IRB Approach published on 4 February 2016 institutions using the IRB Approach will have sufficient time for the implementation of the changes in the definition of default, including recalibration of risk parameters. Further clarifications in the EBA Opinion on the implementation of the regulatory review of the IRB Approach Three respondents argued that the transition period should start when the competent authority sets the threshold and not when the RTS are published. Two other respondents proposed letting each competent authority decide on the length of the transition period for its own jurisdiction. In accordance with Article 7 of the RTS competent authorities will set the threshold within 90 days of the date of publication of the RTS in the Official Journal of the European Union. The preparations for the implementation should start as soon as possible, and in accordance with the EBA Opinion on the implementation of the regulatory review of the IRB Approach published on 4 February 2016 the final deadline for implementation will be agreed individually between the institution and its competent authority. Further clarifications in the EBA Opinion on the implementation of the regulatory review of the IRB Approach Question 4. Do you agree with the assessment of costs and benefits of these proposed draft RTS? A majority of respondents who gave an assessment of the costs agreed that the implementation of the RTS would lead to significant costs related to adjustments to IT The EBA welcomes the opinions presented by the respondents and agrees that the proposal may lead to significant implementation costs for certain N/A

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE systems, data warehouses and internal processes, and recalibration of risk parameters (including a possible period of parallel running in case adjustments to historical data are not possible). Many respondents argue that the costs of implementation will exceed the benefits, all the more because the administrative burden would not be reduced, as different thresholds would still apply in each jurisdiction. Furthermore, two respondents pointed out that the benefits would be close to zero for banks with a local business. institutions, especially where the currently applied materiality threshold is significantly different from the proposal included in the draft RTS. However, these costs will be borne just once and as a result greater comparability will be achieved that will last for all subsequent periods. It is considered crucial from the EBA’s perspective to eliminate these sources of variability of capital requirements that are not based on the differences in risk profiles. Several respondents agreed with the analysis provided in the consultation paper. However, two of them agreed with the analysis of the costs but did not see the benefits. The benefits of the harmonisation are of a more qualitative nature and hence they are not quantifiable. However, it is considered crucial from the EBA’s perspective to eliminate these sources of variability of capital requirements that are not based on the differences in risk profiles, also in the context of restoring trust the of market participants in internal models where the IRB Approach is used for the purpose of capital requirements. N/A Several respondents stressed the need for a quantitative assessment of the costs that will be incurred by the institutions in relation to the changes. It is clear that the costs of implementation will be significant for many institutions, in particular those where the currently applied concept of the materiality threshold is significantly different from the one proposed in the RTS. It was however not possible to perform a quantitative assessment before the final policy decisions were specified. In addition, the costs will be distributed over a long implementation period that was specified as up to end-2020, and they will be related to the N/A

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE implementation of various changes in the definition of default and rating systems that will stem not only from these RTS but also from other changes introduced within the review of the IRB Approach. It was therefore decided in the EBA Opinion on the implementation of the regulatory review of the IRB Approach published on 4 February 2016 that the timeline for implementation should be aligned for all changes resulting from the review of the IRB Approach to allow better cost efficiency and avoid multiple changes in the definition of default and in the models. Several respondents argued that the need to obtain the approval of the supervisory authority will be an important contributor to the implementation costs. It was suggested by some of them that in the case of changes in the materiality threshold such approval should not be required. It has been specified in Regulation (EU) No 529/2014 that all changes in the definition of default should be considered material and hence a change in the materiality threshold is a material change that is subject to approval by a competent authority. However, explanations on how to carry out this process in the most efficient manner have been provided in the EBA Opinion on the implementation of the regulatory review of the IRB Approach published on 4 February 2016. No change Question 5. What is the expected impact of these proposed draft RTS? Several respondents argued that the number of defaults recognised on the basis of past due amounts will significantly increase, and this will have a significant impact on risk parameters. An increased number of defaulted exposures could also lead to increased The EBA believes that the adoption of option 2 as described in question 1 will reduce the expected increase in defaults. It should also be taken into account that customer behaviour will change based on the changes in the policies of the institutions, N/A

FINAL DRAFT RTS ON THE MATERIALITY THRESHOLD FOR CREDIT OBLIGATIONS PAST DUE provisions. Many respondents were also concerned by the quality of data, which might decrease due to the higher number of technical defaults. especially given that they will have a sufficiently long implementation period to educate clients about the consequences of non-payments. However, the EBA recognises that some changes in risk parameters are necessary in order to increase the comparability between models. Some respondents believed that it was not possible to estimate the joint impact on RWA and provisions at this stage. The impact of the changes in the definition of default was assessed also on the basis of a dedicated QIS. It is however clear that the estimation could not be fully accurate before the final policy decisions were taken. N/A Several respondents indicated that the costs and capital requirements would depend on the final policy decision and noted in particular that the impact would be higher if option 1 as described in question 1 was used. As the final draft RTS are based on option 2 as described in question 1, it is expected that the costs and impact on capital requirements will be lower than they would have been if option 1 was chosen.
Article 2(3)
Article 3(3)
Two respondents pointed out that own estimates of LGD for exposures secured by immovable property under the IRB Approach cannot be lower than the limits of 10% and 15% respectively provided for in Article 164(4) and (5) of the CRR. This means that the proposed changes could be very material for credit institutions specialising in business secured by residential property as the potential increase in PD would not be compensated for by the decrease in LGD. The changes introduced in the final draft RTS as compared to the consultation paper should result in a much lower increase of PD. Hence the impact on the portfolio secured by the immovable properties should not be as significant as assumed by the respondents.
Article 2(3)
Article 3(3)

Lineage: No known changes

Similar documents from other regulators