Response to consultation on supervisory reporting changes related to CRR2 and Backstop Regulation (Framework 3.0)

Question 5: The template F39 requests information on the stock of NPEs and related loss allowances/provisions broken-down by the same time buckets as introduced in Article 47c of the CRR and used in the new NPE LC templates of COREP as well. These data allow supervisors to monitor institutions’ NPE coverage strategies more effectively and capture their risk profiles more accurately. They complement, from an accounting perspective, the information provided in COREP on prudential backstop calculation. Which benefit and challenges with regard to the compilation and reporting of this information do you envisage? Template F 39 is very complex, e.g. the granular “of-which”-reporting (especially line 120 “Debt instruments granted after 26 April 2019”) requires an excessive data collection. The whole template F39 will not add any value, but rather confuse readers, as the template contains FINREP as well as COREP information. It is furthermore unclear if capped or uncapped values need to be entered (as in the COREP templates). Additionally, it is not clear to us which risk provisioning types need to be entered (and if those are supposed to comply with the risk provisioning types under the CRR that are used for the calculation of the CRR backstop). Any shortfall seems not to be included, which adds a further gap in comparability of the COREP and FINREP templates. We therefore advocate eliminating this template.

The following additional questions arise for us:
The EBA states the following in chapter 3.5.2, paragraph 34 of the Consultation Paper: “[...] The NPEs have been also broken-down by instrument and some additional details (e.g. the amount of exposures affected by the Backstop Regulation) are provided in separate rows. This amendment facilitates banks to determine the appropriate amounts of specific credit risk adjustments to be included in the backstop calculation. Furthermore, it enhances supervisors’ ability to monitor the accuracy of a bank’s specific credit risk adjustments calculation.” Further, paragraph 36 states that: “[...] In particular, the new Finrep template allows supervisors to conduct reviews, as part of their SREP process, on the accounting impairment coverage levels that are the basis for determining the specific credit risk adjustments included in the backstop calculation.” How do these statements fit into the requirement that FINREP includes intraperiod valuation adjustments, but only audited amounts may basically be included in COREP as specific credit risk adjustments?

CP, page 43 (chapter 5.1 Additional clarifying examples): Why are the time intervals ‘>7 years <= 8 years’, ‘>8 years <= 9 years’ and ‘> 9 years’ missing for V in the clarifying example in chapter 5.1 of the Consultation Paper? In cases where an exposure is secured by immovable property in accordance with Title II of Part 3, different factors apply to these time intervals in accordance with Article 47c(2) of the CRR.

Question 6: Are the instructions and templates C35.01 to C35.03 clear to the respondents? Calculation of the non-performing period is not conclusively addressed. Under Article 47c (2) and (3) of the CRR, this should be based on the period since classification as non-performing. What is not clear is whether this should be based on the date when the debtor or the individual exposure becomes non-performing. In contrast to a debtor, an exposure can only become non-performing after it has been created. This question arises because calculation of the minimum coverage requirements must be based on individual transactions. In this respect, please provide additional instructions in point 8.1.of Annex II.

In which row should accumulated negative changes in fair value for NPE measured at fair value in accordance with the applicable accounting framework be reported?

Question 7: Do the respondents identify any discrepancies between these templates and instructions and the calculation of the requirements set out in the underlying regulation? Additional guidance is given on page 43 of the CP for the calculation of the minimum coverage requirement. Footnote 16 states “In case a deduction is not calculated at exposure but at portfolio level (i.e. IRB shortfall), the total calculated deduction should be allocated to each exposure weighted by the exposure value.” In our opinion, this contradicts the provisions of Article 47c (1) (b) (iv) of the CRR concerning the IRB shortfall, which requires an EL-based redistribution. We ask for clarification that for an EAD based distribution is possible all other items and for inclusion of the clarification in the ITS.

See comments and recommendations under point 8.

Question 8: Do the respondents agree that the amended ITS fits the purpose of the underlying regulation? The suggested separate reporting for non-performing and forborne exposures (templates C 35.02 and C 35.03) seems disproportionally complex especially in view of the secondary practical relevance of this topic. According to Article 47c of the CRR, forborne exposures receive relief compared with non-performing exposures in the form of lower percentages for the calculation of the prudential backstop (and therefore an extended ”phase-in” of any potential capital deduction). The reporting, which only incorporates a shift in the percentages, seems too complex for this matter. We therefore suggest inserting two new lines in template C 35.01: one for the percentages and one for the absolute values for forborne exposures. Templates C 35.2 and C 35.3 can be deleted.

Question 9: Do respondents consider that the new proposed supervisory reporting templates reflect correctly the disclosure requirements, in particular new templates which introduced considerable change? Given that the integration aims at improving consistency, including a standardisation in formats and definitions, do respondents agree that this objective is achieved? Generally, yes.

This applies also for template C 08.05: As a minimum, there is no obvious difference.

The objective should be achieved, as the templates and instructions directly refer to the regulatory basis. However:
a) The starting time of tables 8.5 and 8.5b should take into account both the postponement of EBA\GL\2017\16 and the required one-year observation period. As a result, the planned first reference date for the application of the technical standards (30 June 2021) cannot be met.

b) Because of the high effort with regard to collecting the data and its low readability we propose to:
1. only show aggregated figures for higher exposure class levels such as corporates, retail, etc.
2. only show aggregated figures for the margin of conservatism (only one column, not three)
3. delete the following paragraph on page 113 in Article 3.3.8.1: “In addition to template C 08.05, institutions shall report information included in template C 08.05b in case that they apply Article 180(1)(f) of CRR…”

If it is implied that institutions have to disclose both templates for PD clusters and for their own applied Masterscale, around 430 rows will additionally have to be shown.

Templates C 08.01 und C 08.06 (specialised lending exposures):

Paragraph 48 of the “EBA Report on the credit risk mitigation (CRM) framework” of 19 March 2018 contains the following section on specialised lending exposures:
“…any guarantees that are part of the security package contribute to the factors that enable the slotting. Any additional guarantees, such as those against default by a sovereign bank or another bank, should be treated under the CRM framework, provided they meet the eligibility criteria.”

Neither template C 08.01 nor template C 08.06 provides items for reporting additional guarantees. In our view, the “greyed-out” cells in row r080 and columns c040 to c080 in template C 08.01 should be made available for capturing this data.

Question 10: Are the instructions and templates clear to the respondents? Generally, yes.

However, we are seeking clarification regarding template C 07.00, row 330. According to the relevant instructions exposures – for which a deduction from own funds in accordance with Part Two of the CRR is required – must be reported here. Could you please list in detail which deductions need to be considered. We assume that deductions which arise out of the application of IRB Approaches or of stricter supervisory valuation provisions (such as supervisory shortfall, NPE backstop or PruVal) need not be reported here.

The application of the look-through approach or the mandate-based approach for CIUs (according to Article 132a of CRR 2) may lead to underlying assets that are assigned both to the Standardised Approach and the IRB Approach. Regarding the breakdown of CIU exposures by approaches in template C 07.00, rows 281 to 283, we assume that these rows contain the underlying exposures assigned to Standardised Approaches only. Underlying exposures assigned to IRB Approaches are not reported in this template.

Furthermore, the new column added “Additional value adjustments and other own funds reductions” in C 09.01 should be accompanied by examples.

Question 11: Do the respondents identify any discrepancies between these templates and instructions and the calculation of the requirements set out in the underlying regulation? We are seeking clarification regarding template C 08.07 “Scope of IRB and SA Approaches”. Institutions are required to use the leverage exposure according to Article 429(4) of the CRR as the relevant total exposure value in this context. We would have expected the exposure value according to Article 111 or 166 of the CRR as the basis, naturally before deduction of credit risk adjustments, additional value adjustments, etc. From our perspective, the leverage exposure is not the correct basis for the statement of the scope of IRB and SA Approaches.

Question 12: Do the respondents agree that the amended ITS fits the purpose of the underlying regulation? We have objections to the introduction of column 061 ”AVA / PruVal” in template C 09.01 “Geographical breakdown”. PruVal reporting was only implemented in the last amendment of the ITS on supervisory, which was very burdensome. We also expect the implementation of a geographical breakdown of AVA to be very burdensome. AVA is not calculated at the level of single contracts, whereas reporting a geographical breakdown requires the artificial allocation to single contracts. A clear redistribution to single contracts is not possible, especially for banks using an advanced approach to calculate AVA. Thus the results of the geographical breakdown would not lead to any significant gain in knowledge. On the other hand, implementation would be extremely complex and burdensome.

In any event, we would expect this new requirement to be limited to AVAs that qualify and are used to reduce the EAD according to Article 111 of CRR and in line with reporting in template C 07.00 column 030.

We have objections to template C 08.07 “Scope of IRB and SA Approaches”. We cannot identify any binding CRR requirement that could form the basis for this reporting. Beyond that, we would like to raise the question of whether the template is still necessary for Pillar 3 disclosures and also for reporting. On the one hand, exit thresholds will become less important in the future. On the other hand, both exit thresholds and entry thresholds are not regulated in a uniform way across the EU. In conclusion, the information provided by this template is thus limited and there is no consistent legal basis for reporting it.

Question 14.1: If further proportionality would introduced for this templates, would a threshold be an appropriate way? If yes, which thresholds would respondents recommend to distinguish between institutions that should report on quarterly basis and those that should report with lower frequency? Should it be based on the size of the reporting institution, the size of the derivative business, the total amount of CCR exposure or something else? In general, we consider it sufficient to provide this template semi-annually instead of quarterly. Moreover as for other reports, a threshold for small and non-complex institutions would be appreciated and would reflect the proportionality approach. The amount of reported data should reflect the size of derivative business and the total amount of CCR. Hence, institutions which meet the conditions under Article 4 (1) point 145e of the CRR should report annually.
Additionally, derivatives in the banking book and/or transactions with counterparties within an IPS structure should not be included in template C 34.06.

Question 14.2: Would a semi-annual frequency for small and non-complex institutions be adequate to capture the volatility of these exposures? Reflecting the response above, an annual frequency would – if thresholds were not exceeded – suffice.

Question 16: Are the instructions and templates clear to the respondents? The following matters require clarification in our view:

{C 34.2; r0030; c0030} Notional amounts (for SA-CCR)
It is not clear whether the current notional amount or the notional amount after the adjustment in accordance with Article 279b(2) of the CRR should be disclosed for derivatives with no fixed notional amounts.

{C 34.2; r0030; c0050 and c0060} Variation margin and NICA for SA-CCR
Please confirm that the calculation of these exposures is as follows: Aggregate amount of collateral given less aggregate amount of collateral received, in each case applying applicable haircuts.

{C 34.2; r0030; c0120} Exposure value pre-CRM for SA-CCR and
{C 34.2; r0030; c0130} Exposure value post-CRM for SA-CCR
If the EBA does not make any adjustments to items {C 34.2; r0030; c0120} in line with our suggestion for Question 17, the following issues would then still be open:

Please confirm that the exposure value should be calculated without any variation margin both paid and received and ICA (collateral given is not normally counted towards CRM).
Please confirm that in the case of margined transactions, the EAD should be calculated using the rules for unmargined transactions, meaning that replacement costs are calculated without the minimum transfer amount and the threshold calculation of the PFE uses the maturity factor for unmargined transactions.

C 34.3; c0050 Current market value (CMV)
The CMV is calculated at the netting set level. However, template C 34.3 stipulates that the derivatives must be classified by the criteria of risk category and hedging sets. This means that in the standard case, the CMV of a single netting set must be distributed across several rows in this template. Please explain how this distribution should be performed. One possible solution would be to total up all of the market values of the single derivatives across a row that are allocated to the row in question when netting positive and negative market values (across all netting sets).

C 34.9; r0020, r0050 Index CDS, Other credit derivatives
Should “basket CDSs” (CDSs whose underlying is based on a single basket of reference names, not on a listed index) be reported under Index CDSs {r0020} or under Other credit derivatives {r0050}?

Question 17: Do the respondents identify any discrepancies between these templates and instructions and the calculation of the requirements set out in the underlying regulation? In column c0120, the EAD must be reported uncollateralised, but also without margining according to the instructions. We presume that the EAD must be calculated without variation margins and NICA in the context of the SA-CCR. This would basically mean that the replacement costs for this exposure would be calculated solely from the current market value (CMV). No multiplier would be used in the potential future exposure and the maturity factor may not be calculated in accordance with Article 279c(1)(b) of the CRR. To populate this reporting field, this means that an additional EAD calculation would have to be performed for all netting sets, which would considerably increase the technical effort for automated report generation at this point.

Further, we are not clear about the purpose of this item to be reported in the context of the SA-CCR, because there is no requirement to disclose the EAD in accordance with Article 274(2) of the CRR (as the actual final outcome of the SA-CCR) in connection with item {C 34.2; r0030; c0130} anywhere in the template. We therefore suggest disclosing the EAD under Article 274(2) of the CRR 2 in cell {r0030; c0120} before the application of “other CRM”, i.e. CRM to be taken into account outside the calculation of the EAD (credit protection outside Article 197 of the CRR, unfunded credit protection).

Question 18: Do the respondents agree that the amended ITS fits the purpose of the underlying regulation? Template C 34.01 is just required to prove compliance with the thresholds for use of the Simplified Approach. In order to reduce the regulatory burden, we propose eliminating this template for institutions that do not intend to make use of the Simplified Approach.

Template C 34.01 requires data to be reported as at the last day of month-1, month 2 and month-3 of the quarter. In order to reduce the burden and give institutions enough time for process implementation with regard to the initial reporting, the data reported in the template should be limited to one month (month 3) of the quarterly reporting as at 30 June 2021.

Although we are aware that the EBA is applying Article 273a(3) in the literal sense in template C 34.01, the criteria of short or long position in a derivative cannot always be reasonably applied in practice. Non-trading book institutions using the OEM (and that have to provide evidence of approval to use it in this template) do not split individual transactions into short and long positions when they calculate EAD. These institutions would have to make a corresponding split just to provide this evidence, which would lead to a disproportionately high reporting effort and, in our opinion, contradicts the principle of proportionality.

Question 19.1: Are the structures presented in Section 5.1.2 complete? If not, could respondents provide detailed information on other structures in which a credit institution may have exposures exempted in accordance with Article 429a(1)(d) or (e) of the CRR? Promotional loans to private persons and non-public sector entities are not considered to be beneficiaries eligible for Article 429a(1)(d) exemption. Hence, private persons and non-public sector entities should be added to the graphics as beneficiaries. This exemption requires amendments to the corresponding COREP templates C 40.00 and C 47.00.

IPS structures in which a central institution passes through promotional loans to its group’s credit institutions have not been taken into account. Hence, some leeway to unburden small and non-complex institutions and to increase proportionality remains unused.

Question 19. 2: Do the proposed amendments provide for an adequate reporting on exposures of credit institutions that are involved in these structures? As these additional reporting requirements are only relevant for a small percentage of institutions with a certain business model, we would prefer to exclude those requirements from the overall reporting templates and to introduce a separate template that is only relevant for those specific institutions. Additionally, data collections within QIS could provide information whether these specific requirements have a substantial impact on LRE calculation.

The same applies to cash pooling and settlement/trade date accounting-specific reporting requirements. In some cases, those two topics have an only minor impact on the calculation but would lead to disproportionate operational complexity and costs. At the very least we would appreciate a situation where the reporting requirements for cash pooling arrangements that can be netted prudentially were to be removed from the templates, as they have no effect on the actual calculation of the leverage ratio exposure.

Because CPAs will only have to be disclosed separately in the context of the leverage ratio, meeting this extremely granular reporting requirement could lead to a disproportionately high investment cost and/or ongoing reporting effort. Reasons:

A netting process will have to be implemented in the reporting system in order to generate such granular outcome data. For a client, this netting process needs datasets on all asset and liability accounts with the following minimum content: account balances before netting, CPA contract number, attribute “Eligibility for accounting netting yes/no”, attribute “No eligibility for supervisory netting/Eligibility for supervisory netting under Article 429g(2) of the CRR/Eligibility for supervisory netting under Article 429g(3)”. The individual attributes would have to be generated in the course of extensive upstream processing; in addition, CPA contract datasets would probably have to be generated along the lines of derivatives and SFT netting.

We wish to point out that in the case of CPAs with daily settlement, that settlement is not normally performed at the end of the business day, but continuously over the course of the day. Simulated gross netting would have to be implemented in the account management system in which the CPAs are normally recorded so that the reporting system is fed with the corresponding outcome data.

We would also like to point out that the granularity in r193-198 breaches the systematic approach in template C 47. The model for all other items in C 47 is “1) Report entire exposure before deductions, 2) Report entire deductions“ (i.e. in line with the proposal shown above).

Alternative proposal: Institutions whose CPA business is not material (e.g. exposure value before netting is less than 5% of total assets) do not have to enter data in the cells in C 40.

C 47.00 Leverage Ratio Calculation (LRCalc)
Please delete the sub-items/secondary items 065/081/091/092/093 and 220. The related primary items 061 (Derivatives: replacement cost) and 091 (Derivatives: potential future exposure) must then be reported in accordance with Article 429c of CRR 2.

Compared with the previous calculation for derivatives, the calculation in the SA-CCR is more extensive and more complex. Disclosing parameters when calculating the leverage exposure for derivatives is only possible – if at all – with considerable technical effort. We are unable to discern any supervisory benefit from the separate disclosure of individual subtotals. For this reason, the level of detail in template C 47.00 should be limited to what is strictly necessary.

C 48.02 – Leverage ratio volatility: daily values for the reporting period (LR6.2)
The requirement to provide daily values in this context requires a new form of data storage and processing (temporary storage of daily values to determine mean quarter-end values). This will increase complexity. We therefore oppose the calculation of daily values.

Question 20: Regarding the proposals to include averaging for some components of the leverage ratio in accordance with Article 430(2) and (7) of the CRR, to develop the standards the EBA shall take into account the how susceptible a component is to significant temporary reductions in transaction volumes that could result in an underrepresentation of the risk of excessive leverage at the reporting reference date. What leverage ratio components do respondent consider most and least susceptible to temporary reductions in transaction volumes? Most: Securities finance transactions; however, reporting daily exposures is excessive and would lead to disproportionate operational complexity and costs. Our understanding is that a calculation based on weighted month-end average exposure values would provide the same information as an operationally excessive reporting of daily actuals. In any case, LR6.2 (C 48.01) should be deleted in order to reduce excessive burdens, as LR6.1 (C 48.02) already shows the average result of LR6.2.

Least: Other on-balance-sheet items.
If it remains mandatory to calculate LRE based on daily averages, we would appreciate a change in C 48.02 from reporting 60 daily values to reporting the maximum exposure amount within the specified quarter reporting time frame.

Question 21.1: Would respondents agree to align the information reported by requiring the RWEA in this template without taking into account potential substitution effects due to credit risk mitigation? We would generally agree to aligning RWA and LRE figures and would prefer to report both figures after substitution.

However, we do not see any basis for retaining template C 43.00 and therefore advocate deleting it (see also Q24).

Question 21.2: Would respondents strong reasons based on costs to prefer instead the reporting of both values, the RWA as well as the leverage ratio exposure, after substitution effects? What would be the reasons? The main reasons are operationally based. To report RWA figures before substitution would require additional background processes, leading to increased complexity and costs.

Question 22: Are the instructions and templates clear to the respondents? Generally, yes.

Question 23: Do the respondents identify any discrepancies between these templates and instructions and the calculation of the requirements set out in the underlying regulation? Generally, no. Nevertheless we would ask you to consider the following two issues:

a. Template C 48.01 requires the mean the daily values of the reporting quarter to be reported – in addition, all business days within the reported period must be reported in template C 48.02.
This would lead to the fact that, for the first reporting date as at 30 June 2021, the daily process must be in place by 1 April 2021.
In order to reduce the burden and to give the institutions enough time for process implementation, the daily reporting should be shortened for the first reporting of daily values to one month (June 2021).

b. According to EBA Q&A 2015_1856, the “Other Assets” item should be stated gross. Any offsetting related to tax assets or liabilities should be reversed. We would suggest inserting a separate row in Template C 47.00 “LRCalc” to increase transparency about the amounts reversed.

c. The titles of rows 185 to 189 are very confusing. In order to provide better guidance for reporting institutions, we suggest the following adjustments:
i. row 185 “Regular-way purchases or sales awaiting settlement: Accounting value under trade date accounting”: Change to “Trade date accounting: Regular-way sales awaiting settlement”
ii. row 186 “Regular-way purchases or sales awaiting settlement: Reverse out of accounting offsetting under trade date accounting”: Change to: “Trade date accounting: Reverse out of accounting offsetting regarding regular-way sales awaiting settlement”
iii. row 187 “(-) Regular-way purchases or sales awaiting settlement: offset in accordance with 429(g)(2) of the CRR”: Change to: “Trade date accounting: Regular-way purchases awaiting settlement offset in accordance with 429(g)(2) of the CRR”
iv. row 188 “Regular-way purchases or sales awaiting settlement: Full recognition of assets under settlement date accounting”: Change to “Settlement date accounting: Regular-way purchases awaiting settlement”
v. row 189 “(-) Regular-way purchases or sales awaiting settlement: offset for assets under settlement date accounting in accordance with 429(g)(3) of the CRR”: Change to “(-) Settlement date accounting: Regular-way sales awaiting settlement offset in accordance with 429(g)(3) of the CRR”

Question 24: Do the respondents agree that the amended ITS fits the purpose of the underlying regulation? Especially with regard to reporting requirements in LRCalc (C 47.00): the extensive data collection on regular way purchases or sales and cash pooling leads to a complex and confusing reporting structure with only minor additional insights. This also contradicts the initial idea of an easily comprehensive and comparable ratio/reporting. The relevant reporting requirements should be deleted. At the very least, we would appreciate it if the reporting requirements for cash pooling arrangement that can be netted prudentially were to be removed from the templates as they have no effect on the actual calculation of the leverage ratio exposure.

Further, the additional reporting requirements on public development credit institutions are only relevant for a small percentage of institutions with a certain business model, so we would prefer excluding those requirements from the overall reporting templates and introducing a separate template that is only relevant for those specific institutions. Additionally, data collections within QIS could provide information whether these specific requirements have a substantial impact on LRE calculation (see question 19.2).

There have been no changes in the derivative limits above which additional reporting items must be recorded, although the new definition of small and non-complex institutions has now been introduced. In this case, the derivative limits should be aligned with the same limit used for the application of the OEM (Article 273a(2) of CRR 2) in order to ensure more consistent treatment.

Additionally, we do not see any legal basis for retaining templates C 40.00 and C 43.00. In the same way as C 41.00 and C 42.00, which will be deleted under the present Draft, these templates were introduced for reporting data necessary to produce the report in accordance with Article 511 of the CRR. The EBA already submitted this report in 2016. Retaining data reports that are no longer required for supervisory purposes is excessively time-consuming for the institutions, so we are advocating the deletion of both templates.

Question 25: Are the instructions and templates clear to the respondents? The description for columns 120 – 170 (template C 28.00) “Indirect exposures” relates in each case to Article 403 (substitution approach). To clarify a standardised European approach, a note should be added in line with the underlying reporting methodology illustrated in EBA Q&A 2014_1349 and the extended scope defined by the Federal Financial Supervisory Authority on the basis of this understanding for those German institutions that it supervises that a substitution effect should not only be reported when a recognised credit risk mitigation technique is used for large exposures, but also when exemptions from the large exposures limits are applied, which are based on guarantees.
Furthermore, the reference in columns 120-170 to Article 403 (3) CRR 2 might be misinterpreted without further clarification. We understand it as a reference to specialized triparty business, in which case the limit according to Article 403 (3) letter b CRR 2 could be reported instead of the amount according to Article 403 (3) letter a CRR 2.

The description of columns 240-290 (template C 28.00) in comparison to column 300 is not clear, as the substitution of financial collateral with the issuer is generally mandatory under CRR 2. The description for columns 240 – 310 (template C 28.00) “(-) Eligible credit risk mitigation (CRM) techniques” states that “CRM techniques may have three different effects in the LE regime: substitution effect; funded credit protection other than substitution effect; and real estate treatment.” Our understanding is that under CRR 2, any funded credit protection, i.e. including where the Financial Collateral Comprehensive Method (FCCM) is used, leads to substitution. We are therefore seeking a review of the description, which we believe should be “two” (different effects). As a result, we also believe that the label for column
300 “(-) Funded credit protection other than substitution effect” is no longer correct.
Maybe columns 240-290 only refer to full substitution under the Financial Collateral Simple Method (FCSM), while column 300 is for the Financial Collateral Comprehensive Method (FCCM), including haircuts on the collateral?
As these are the exposure-reducing columns, the haircuts decrease the market value of the collateral, while this haircut-reduced part as a risk mitigation effect should be the indirect exposure to the collateral issuer. The word “increased” should therefore be replaced by “decreased” (instruction for row 300).
A haircut within the FCCM decreases the value of the collateral. An increased value (market value + haircut) for the issuer is not mentioned in Article 401(1) and makes no economic sense, as the potential loss in the collateral cannot be greater than the mitigation effect on the original exposure.
This also does not to fit with the description of columns 120-170, which states that the amount of reducing the direct exposure must be the amount of the exposure of the collateral issuer.

For clarification, the description for column 320 (template C 28.00) “(-) Amounts exempted” should also contain a reference to Article 493(3) of the CRR.

Question 27: Do the respondents agree that the amended ITS fits the purpose of the underlying regulation? The reporting framework 3.0 only contains changes relating to large exposures that can be derived directly from CRR 2. The elimination of reporting the maturity bands for the ten largest loans to an institution and the ten largest loans to unregulated financial sector entities relieve the reporting burden, albeit only for those institutions that are required to submit a group report.

The elimination of the reporting on the “type of connection” in the context of the “group of connected clients” (GCC) in template LE 3 only achieves selective – albeit welcome – relief. No further content-related relief is proposed. Additionally, the reports should still be submitted quarterly. Despite the supervisory commitment to proportionality and the acknowledged need to reduce burdens in the reporting regime, the potential relief offered by a half-yearly report granted by the lawmakers, especially for small and non-complex institutions, is therefore not used.

Additionally, further changes to the reporting requirements will result from the EBA mandates granted by CRR 2. This applies in particular to the mandates under Article 4(4) of CRR 2 (GCCs), Article 390(9) of CRR 2 (determination of the exposure value of the underlying in the case of derivatives) and Article 394(4) of CRR 2 (definition of shadow banks for reporting purposes). In the meantime, it is evident from the EBA roadmap on large exposures published on 21 November 2019 that the EBA will not be able to meet the deadlines stipulated by CRR 2 for the above mandates. For example, the draft RTS to determine the underlying risk for derivatives in accordance with Article 390(9) of CRR 2 will not be submitted by March 2020, but only by December 2020. However, the new large exposure rules already have to be applied as of 28 June 2021, so it is simply not going to be possible to apply them fully as at the first CRR 2 basis reporting reference date of 30 June 2021.
Reporting the ten largest shadow banks will already be mandatory as at 30 June 2021. According to the EBA, the mandate under Article 394(4) of CRR 2 will only be complied with as at December 2021. Although there are guidelines for limiting exposures to shadow banks, it cannot be ruled out that the RTS under Article 390(4) of CRR 2 will not contain a different definition of shadow banks, or that the definition underlying the guidelines might also be revised to reflect international developments and new insights. Multiple adaptations of IT processes must be avoided. It should be clarified in this respect that, until the RTS under Article 394(4) of CRR 2 has been finalised and after the end of an appropriate IT implementation period, the report can be implemented as it is today, i.e. based on the concept of an “unregulated financial entity” anchored in the CRR. The delay’s in the EBA’s work may not be at the institutions’ expense.

Question 28: Paragraph 4 of Article 428d in the CRR2 states: “all derivative contracts referred to in points (a) to (e) of paragraph 2 of Annex II that involve a full exchange of principal amounts on the same date shall be calculated on a net basis across currencies, including for the purpose of reporting in a currency that is subject to a separate reporting in accordance with Article 415(2), even where those transactions are not included in the same netting set that fulfils the requirements set out in Article 429c(1).” Reporting by currency subject to separate reporting is required to be made on a net basis across different netting sets. This might envisage a situation of derivatives across various counterparties with different settlement currencies. There is a need to provide further instructions on which specific currency subject to separate reporting report should capture the net value in these cases. The implication is that the CRR2 requires consistency between ASF and RSF by currency subject to separate reporting on which specific requirements can be set by CAs. It is proposed to look at each netting set and calculate the fair value for each of them in its settlement currency. For all netting sets with matching settlement currencies a net amount shall be calculated in accordance with Article 428k(3) and 428ag(3), and reported in the relevant currency subject to separate reporting. Do respondents agree with this proposal? Would respondents consider it more adequate to look at all payables and receivables related to derivatives and calculate a net amount? Our understanding of the proposal (and the example 1 on page 49) is that, for the NSFR in significant currencies (including ”EUR-only NSFR”), only FX derivatives with a settlement currency equal to the relevant significant currency should be included, and all FX derivatives with different settlement currencies should be exempted in their entirety from separate reporting in significant currencies.

If this understanding is correct, we agree with the proposal and do not need further instructions. Otherwise we would like to see some additional and more detailed examples for reporting of FX derivatives by currency, because it is not then entirely clear the daily exchange of the principal amounts should be treated under Article 428ah(2).

Question 29: Do respondents consider that the “NSFR calculation tool” appropriately translates the use of the different templates for informative purposes? Overall, yes.

Question 30: Are the instructions and templates clear to the respondents? Generally, yes. But we would propose including a field with the legal reference (like the former NSFR ITS) in every row of the template.

Question 31: Do the respondents identify any discrepancies between these templates and instructions and the calculation of the requirements set out in the underlying regulation? General comments
The item ”Coins and banknotes” is missing (compared with the current template). We assume that these assets should be included in c80.00, r0040, c0010 (or c82.00, r0030, c0010). This should be mentioned in the relevant instructions.

There should be a field with the legal reference (like the former NSFR ITS) in every row of the template.

Given that the “accounting value” is requested for the NSFR, further instructions and definitions are needed (keywords, i.e. accrued interests general credit risk adjustments, unrestricted contingency reserves under section 340f of the HGB), especially with regard to any implicit comparability with FINREP (see 5.4.2). Furthermore, a statement that the accounting value must be used throughout the ITS would enhance data quality, comparability and consistency across the respective reports.

We would welcome statements about determining the encumbrance period for assets in the cover pool. There is also a lack of clarity about how the assets in the cover pool are allocated to the non-mandatory overcollateralisation.

Because of the link to the accounting and uncertainties from the Basel III monitoring, clarification that all derivatives must be included in the NSFR (regardless of whether or not they are recognised in the accounting) would be useful.

In the area of derivatives, the BCBS distinguishes between own-name initial margins and initial margins made on behalf of clients. There is no requirement for stable funding for the latter. This distinction is not made in CRR 2 and it is therefore not clear whether they may also be disregarded.

Comments on template C 80.00:
Under CRR 2, non-HQLA level 2B equities are subject to an 85% factor if they are exchange-traded. This factor is used in reporting row 0580 solely for the ≥ 1 year maturity band. If appropriate, it should be clarified that equities must be allocated to the longest maturity band. To our knowledge, there is not a general rule in CRR 2 that assets without a stated maturity must be allocated to the most conservative maturity band (i.e. the longest for the RSF factors). By contrast, a corresponding requirement for liabilities/equity is contained in Article 428k(1) of CRR 2. Alternatively, a separate reporting row could be created for equities.

Whereas CRR 2 does not contain any reference to maturities for initial margins paid for derivatives or central counterparties, the template distinguishes between the three maturities and HQLA in reporting rows 0950 and 0960 (all cells with the same RSF factor). What is the basis for calculating the maturities (maturity of the collateral (if present) or term of the derivative contract)?

Reference is made to the LCR to capture the undrawn portion of committed credit and liquidity facilities. The LCR distinguishes between irrevocable and revocable facilities. CRR 2 does not make an explicitly detailed distinction for the NSFR, but merely talks of “committed facilities within the meaning of the LCR”. Is our assumption correct that only irrevocable (and under certain conditions possibly also revocable) facilities are to be reported in row 1060 with a factor of 5%, whereas revocable facilities are not to be reported, at least in the first instance as long as country-specific factors are not implemented? That would also be consistent with the BCBS rules or the Basel III monitoring instructions.

We request clarification of where the deductions from regulatory capital should be reported.

Comments on template C 81.00
In accordance with Article 428l(a) of CRR 2, operational deposits of central banks are subject to a 50% ASF, and other liabilities to central banks are subject to a 0% factor. However, there is no separate row for reporting operational deposits of central banks. Are central banks deliberately excluded for operational deposits? There is a corresponding row (row 21) in the BCBS template.

Under Article 428k(3)(d) of CRR 2 and because it is not addressed anywhere else, secured liabilities to retail customers and small businesses (unless they can be treated as retail deposits under Article 428i) are subject to a 0% ASF factor. No row is provided for this in the templates proposed by the EBA. Row 43 is provided for this purpose in the BCBS template. Was the separate disclosure of secured liabilities to these counterparties omitted deliberately? To our knowledge, the institutions that participated in Basel III monitoring report some exposures in this row. In order to apply the correct factors, rows 310 “ASF from liabilities provided where the counterparty cannot be determined” and 430 “Other liabilities” would generally be suitable, although the latter is a better fit in terms of substance.

C 81.00; row 0040 to 0060; column 0050:
We see a discrepancy regarding the ASF factor for capital instruments (ex-CET 1) maturing between 6 months and 1 year. Article 428l(d) of CRR 2 stipulates an ASF factor of 50% for ”any other liabilities with a residual maturity of a minimum of six months but less than one year not referred to in Articles 428m, 428n and 428o”. In our opinion, this includes liabilities from maturing capital instruments (6 months to 1 year) because they are not referred to in the named articles. As a departure from the CRR, C 81.00, rows 0040 to 0060, column 0050 stipulates an ASF factor of 0%.

The expectation of an 50% ASF factor would be in line with NSFR QIS (B3M) reporting practice:

NSFR QIS (B3M) is used to report capital instruments with maturities of less than 1 year in the liabilities positions for the relevant investors (generally financial undertakings/institutions, i.e. row 34 of NSFR ASF QIS), because row 8 of NSFR ASF QIS is only viable for maturities greater than 1 year.

The section ”ASF from capital items and instruments” of CRR 2 NSFR (C 81.00) now has additional cells for maturities of less than 1 year (i.e. rows 0040 to 0060, columns 0010 to 0020). However, these additional cells stipulate a lower ASF factor of 0% in the 6-12 months bucket (column 0050) compared with an ASF factor of 50% for unsecured liabilities from financial clients (rows 290 and 300, column 50), and also compared with the reporting in ASF NSFR QIS already referred to.

We see no justification for this discrimination of liabilities from maturing capital instruments against other short-term unsecured liabilities and therefore would expect an ASF factor of 50% for column 0050 of rows 0040 to 0060, which would be in line with our understanding of Article 428l(d) of CRR 2.

Comments on the Instructions (Annex XXVII)
We presume that in No. 4 of Part I (General Instructions), Article 428ah(2) of CRR 2 is meant, and not Article 428h(2) of CRR 2.

Question 32: Do the respondents agree that the amended ITS fits the purpose of the underlying regulation? See Question 31.

Question 33.1: Do respondents agree with the separate presentation of POCIs outside the IFRS 9 Impairment stages? We very much welcome the EBA’s decision to separate POCIs from the IFRS 9 impairment stages. However, this will result in multiple initialisation scenarios. At the moment we are in the implementation process for the DPM 2.9 reporting requirements. The technical switch between different DPMs results in additional effort, since we have to guarantee no data clashes will arise. DPM 3.0 implementation will result in overlapping with DPM 2.9 in terms of project activities. Therefore, integration into DPM 2.9 would have been less costly and time-consuming.

Due to the harmonisation of the default definition, non-performing exposures in stage 3 will be identical to defaulted exposures. Since POCIs are part of stage 3 exposures under DPM 2.9, we expect the sum of defaulted exposures under DPM 3.0 POCIs and stage 3 exposures. We request clarification of this point.

Question 33.2: Are the criteria to distinguish between “non-performing” and “performing” POCIs clear? Which challenges with regard to the practical application of these criteria do you envisage? Are the criteria for distinguishing between “non-performing” and “performing” POCIs identical in comparison to the existent criteria for identifying performing/non-performing other exposures? We request clarification of this point.

Question 34.1: Which challenges with regard to reporting of this information do respondents envisage? This will result in multiple initialisation scenarios. At the moment we are in the implementation process for the DPM 2.9 reporting requirements. The technical switch between different DPMs results in additional effort, since we have to guarantee no data clashes. DPM 3.0 implementation will result in overlapping with DPM 2.9 in terms of project activities. Therefore, integration into DPM 2.9 would have been less costly and time-consuming.

Question 34.2: Do you see any inconsistencies between this data and the data collected in other FINREP templates? If we understand the requirement correctly, the sum of “cash balances at central banks and other demand deposits” should equal the sum of F01.00 row 30 + F01.00 row 40. If so, we do not see any inconsistencies.

Question 35: In template F12.02, additional columns have been added to report the direct transfers between Stage 1 and Stage 3, without considering any intermediate passage through Stage 2. This information is useful in the context of monitoring IFRS 9 post-implementation initiatives and supervisory activities. Which challenges with regard to reporting of this information do respondents envisage? In general, the information on direct stage 1-3 transfers and vice versa exists in the data warehouse. However, it can be expected that exposures are not material, which results in manual, infrequent data collection for internal data analysis only during the reporting year. Hence, the extensive adjustment of technical systems and the establishment of new reporting processes are necessary to report the information. Considering the very low materiality of such cases, the implementation burden outweighs the information benefit.

Question 36: In template F18.00, the information on loss allowances for more than 30 days-pastdue exposures has been added. This information is already reported in template F23.04 by institutions which fulfil both of the conditions referred to in points (i) and (ii) of Article 9(2)(h) of the current ITS on reporting. Since this information is relevant for monitoring IFRS 9 post implementation initiatives and supervisory activities, it has been included in template F18.00 for all institutions, although it may create some overlaps with F23.04. Which challenges with regard to reporting of this information do respondents envisage? How are nGAAP users affected? An exemption to report this information for nGAAP users should be included.

Question 37: Are the instructions and templates clear to the respondents? Since a separate chapter for comments on Asset-Encumbrance templates is not provided, we have decided to aggregate our comments with regard to the new AE templates here:

a) Missing “of which” row for EHQLA in Advanced Templates F 36.01/F36.02:
In templates 36.01 and 36.02, rows 200/230 report the total of (un)encumbered central bank eligible assets. In terms of consistency, we would have expected a corresponding row for (un)encumbered EHQLA. Could you please describe the reason for leaving out the total EHQLA rows in templates F 36.01 and 36.02.

b) Interpretation of EHQLA as asset quality criterion (F 32.01, F 32.02, F 32.03):
We would like to point out in this context that the use of HQLA as an asset quality criterion does not necessarily outweigh the validity of central bank eligibility as a measure for asset quality. In order to properly assess the asset encumbrance disclosure, information on encumberable assets within non encumbered assets would be more useful for external recipients. Encumberable unencumbered assets imply further funding potential and give also insights into the risk profile inherent in encumbered assets. Hence, the higher the volume of encumberable unencumbered assets is, the fewer the risks that might occur in stress situations. Nevertheless, the HQLA and CBE criteria are good indicators for estimating asset quality in the case of a shortfall or financial crisis. However, when it comes to a breakdown of lending business between institutions, central bank funding is the only option.

c) Label amendments Asset-backed securities to Securitisations (F 32.01, F 32.02, F 32.04, F 34.00, F 36.01, F 36.02):
Is there any difference in the definition of asset-backed securities and securitisations, or is it simply an alignment of ITS and labels within the templates. We request clarification of this point.

In line with the principle of proportionality and taking into account the fact that one of the core elements of CRR 2 is to make it easier to lend to SMEs (see 3.5.3 No. 39), we would welcome the exclusion of public promotional loans by development banks, transmitted loans and publicly guaranteed export credits from the calculation of asset encumbrance.

Question 38: Do respondents agree with the proposal to harmonise templates and instructions with regard to the reporting of the information of LEI codes? Comparing the FINREP and COREP templates, we see little evidence of harmonisation. Although it appears that the COREP and FINREP templates have been partially aligned, there are still inconsistencies. COREP requires the reporting of ”LEI Code” and ”National Code”. However, FINREP reporting requires ”LEI code”, ”National Code” and ”Entity Code”. The ITS does not provide a proper definition for ”National Code”. From our point of view, it seems as if reporting is mandatory when there is no LEI Code and voluntarily in the opposite case. In terms of implementation effort, the burden might be low if there is a clear definition and accessibility for all group entities (via the public register) for the ”National Code” reporting. Otherwise, the implementation burden would be multiplied (e.g. LEI code admission is costly and not entirely implemented within group entities on top of additional National Code selection). Therefore, we expect that reporting by LEI Code and Entity Code, consistently aligned with COREP and FINREP, would be more appropriate. Nevertheless, we would very much welcome a clarification of the EBA’s rationale for the ”National Code” as a third code format within FINREP, and of the definition and reporting method (mandatory/voluntarily).

Question 39: The integration between disclosure and reporting aims at improving consistency, including a standardisation in formats and definitions. Do respondents agree that this objective is achieved? F 32.01, F 32.02, F 32.03: Although EHQLA was already implemented in the annual disclosure, data delivery processing has to be additionally implemented for quarterly reporting. Therefore, reporting of both central bank eligibility (CBE) and EHQLA imposes unnecessarily reporting burdens for institutions. Furthermore, we would not have expected double reporting of CBE and EHQLA in terms of alignment. Hence, eliminating the reporting of CBE would be more consistent with asset encumbrance disclosures and would ease the reporting burden.

There are still some inconsistencies between the reporting templates and existing disclosure guidelines. The harmonisation of reporting and disclosures should ensure that only information from the reporting templates is used in banks’ disclosures.