Response to consultation on draft revised Guidelines on recovery plans indicators

Go back

Question 1: Do you have any comments on the general requirements that should drive the calibration of recovery indicators as proposed in paragraph 27 of these guidelines?

The additional guidance on the calibration of recovery plan indicators’ thresholds focuses a lot on the conservatism of the calibration of regulatory capital and liquidity indicators, suggesting that recovery thresholds are usually set too close to regulatory requirements, hence reducing the alert function of the indicators.
In our opinion, the assessment of the degree of conservatism of a recovery dashboard (i.e. its ability to timely alert the management of the bank and allow recovery options to be implemented) should not be based on the calibration of the maximum alert thresholds only, but more generally consider the structure of the dashboard.
Indeed, a traffic light approach in which each indicator is assigned different levels of alert, with multiple levels ranked by increasing order of severity, hence by decreasing order of conservatism, would prompt a bank to consider its situation and whether it is appropriate to take any actions, because of:
- either a conjunction of widespread moderate stress affecting a larger number of indicators (meaning that several thresholds set well above regulatory requirements have been crossed),
- or a severe stress in one particular area affecting a smaller number of indicators (crossing a threshold equal to or even set lower than the corresponding regulatory requirement).
With this approach, the first alert levels are by definition highly conservative and their progressive deterioration comes to the attention of senior management very early. On the contrary, there is no reason to be too conservative in the calibration of the maximum alert level (i.e. setting it over the regulatory requirement). The structure of a traffic light dashboard guarantees that the management of a bank will have sufficient time to act effectively in a crisis even if the maximum alert thresholds are set at a low level.
We fully agree with the statement (paragraph 27(e)) that an institution should ensure that the calibration of its recovery indicators is consistent with its risk management and Risk Appetite Framework (RAF) (e.g. early warning and limit framework). The Risk Appetite Statement (RAS) escalation process must be consistent with the recovery dashboard escalation process at all times. Therefore, at no time should the maximum alert level of the recovery dashboard for any indicator be above, in a better or even equal situation, than the corresponding risk appetite limit. The European Central Bank (ECB) does not ask RAS limits to be set above supervisory requirements. Hence, banks should be authorized to set maximum alert levels of the recovery dashboard under minimum supervisory requirements level when needed. The Liquidity Coverage Ratio (LCR) or the Net Stable Funding Ratio (NSFR) for instance, and accordingly some of the thresholds of a given indicator, including of course the threshold triggering the start of an escalation process, could be below minimum requirements if relevant.
However, it must be taken into account that the metrics included in the Risk Appetite Framework are very limited and selected by the Board of each entity, so they may not always coincide with the list of metrics included in the EBA minimum list. Therefore, it may be the case that metrics classified as recovery indicators - because they are included in the minimum EBA list - are not part of the entity's RAF, but they will always be risk-managed metrics.
In addition, we signal that the overall recovery capacity (ORC) has not been properly defined by a regulatory framework yet: it was generally introduced by the Commission Delegated Regulation (EU)2016/1075 and is currently used in the field by the ECB. ORC is also related only to the capital and liquidity ratios and not to asset quality or profitability indicators.
The EBA, as a regulator, should not support any concept that does not have a clear regulatory status such as the ORC, which poses arm’s length major issues. Indeed, the computation of the ORC should be clarified. For example, its dependence on the scenario and the change of scenarios year-on-year, together with uncertainty on the term “overall”, make it complicated for banks to use ORC for the calibration of the recovery indicator framework, as required by the Guidelines under par. 27(a).

Question 2: Do you have any comments on the requirement that there should be no automatic recalibration of recovery indicators upon the application of temporary supervisory relief measures, however it could be allowed by competent authorities in those cases specified in paragraph 31 of these guidelines?

The European Banking Industry disagrees with this approach. The calibration of risk appetite metrics and recovery dashboard thresholds should take into account regulatory requirements from time to time.
If we acknowledge the ECB relief measures in reaction to the Covid-19 crisis in 2020 were indeed labelled “temporary”, we suggest a more nuanced view in this context:
- The relaxation of counter-cyclical buffers, taken by supervisors, should be reflected in recovery dashboard thresholds. Counter-cyclical buffers are dynamic by nature and the currently suggested EBA approach defeats the purpose of this buffer, which is to adjust the regulatory requirement according to positions in the cycle, whatever the drivers or circumstances are;
- The relaxation of the P2R components that do not qualify as CET1 should also be reflected in the recovery dashboard thresholds to the extent that this relief measure anticipated a permanent measure initially scheduled to come into effect in January 2021, as part of the latest CRD review;
- Therefore, the currently suggested EBA-approach should only be considered for extraordinary and temporary measures not covered by the examples given above.
More generally, we would welcome additional clarifications on the EBA guidelines with regard to the following:
- At paragraph 43, the Guidelines make an explicit reference to “capital requirements applicable to the institution”, acknowledging that capital requirements should be taken into account for calibration. This should, in our opinion, apply to changes in capital requirements introduced in crisis situations as well, unless specified otherwise. This is all the more true as supervisors will expect, conversely, increases in regulatory requirements to be reflected in the calibration of recovery indicators;
- When referring to the recalibration of the recovery indicators, EBA mentions a supervisory consent that, in our understanding, applies only in case of supervisory relief measures (par. 30) and the situation when institutions want to recalibrate their thresholds (par. 29). In cases of recalibration of the thresholds, the request should meet all the requirements as defined in paragraph 27, without requiring any prior approval from supervisory side, as a notification to the competent authorities should be sufficient. In our view, further clarifications on the description for the process of “requirements for the calibration of the indicators” and separate process “requirements for recalibration in case of supervisory relief measures (when supervisory approval is needed)”, should be provided.

Question 3: Do you have any comments on guidance introduced in relation to actions and notifications upon breaching recovery indicators, including the proposed timelines for internal escalation and notification to the competent authorities?

We believe more flexibility should be provided by EBA with reference to the notification process in case of the recovery indicators breaches, in order to effectively focus on real potential issues and avoid “over notification”. In particular, EBA should consider the following:
- Not all recovery indicators have the same influence on the financial position of the bank (i.e. liquidity and capital indicators are primary, and should be considered as more relevant if compared to secondary indicators as asset quality or profitability indicators, or even more to some market-based indicators or macroeconomic indicators that banks can use as early warning signals);
- Therefore, not all types of recovery indicators breaches should necessarily be notified to the supervisors or, at least, not with the same timeline. In this respect, while a specific timeline for the notification of breaches is welcome, as it removes uncertainty, a longer notification process in case of asset quality, profitability, market-based and macroeconomic indicators should be envisaged;
- For the same reason, alerting the management body of the institution in 24 hours seems too burdensome and should not be always required (i.e. referring to CEO and/or CFO should be sufficient, especially in case of asset quality and profitability, or even more for market-based and macroeconomic indicators).
Nevertheless, it should be highlighted that, if it is decided to activate the recovery plan, this first communication may not include a full detailed action plan or a deep analysis of the solutions to restore the indicators levels. A more comprehensive analysis with the impacts of the selected options to be implemented may be communicated to competent authorities, if needed, at a later stage and in due time.

Question 4: Do you have any comments on introducing a possibility for competent authorities to request institutions to provide a full set of recovery indicators (breached or not)?

Question 4 relates to paragraph 38, about the adequate frequency of production of indicators. Paragraph 38 suggests a monthly frequency at least, even if it admits that values of the indicators may not change. This is somehow contradictory and leaves room for interpretation. In order to ensure proportionality and priorities when imposing the monthly production of indicators, we suggest a differentiated view to the extent that not all indicators should be produced with the same frequency. For example, it makes sense and therefore should be expected that the frequency for liquidity indicators is higher than for asset quality indicators. For some indicators, proxy values for intermediary months should be accepted.
In any case, it should be recalled that the request by the competent authority to provide a full set of recovery indicators should be advanced only in case of activation of a state of crisis, as stated at paragraph 18 of chapter 3.3 “Additional guidance on breaching and monitoring of recovery indicators” (see page 9 of the CP).

Question 5: Do you have any comments on the proposed threshold calibration of regulatory capital indicators at levels above those requiring supervisory intervention and therefore to be generally calibrated above the combined capital buffer requirement while still allowing calibration within buffers only under certain conditions?

Capital indicators should be calibrated in a way where indeed capital buffers can be partially or fully used before a capital maximum alert level is breached. In this regard, we signal that capital conservation buffers, including P2G, are allowed to be used in a stress. This message was communicated by ECB in the 12th March 2020 press release on the usability of capital buffers in the Covid-19 crisis: “The ECB will allow banks to operate temporarily below the level of capital defined by the Pillar 2 Guidance (P2G), the capital conservation buffer (CCB) and the liquidity coverage ratio (LCR).”
It would not be consistent for the ECB to state that banks can use conservation buffers in crisis and at the same time the EBA guidelines to suggest that in the same situation institutions could enter into recovery. Therefore, there should be no condition to the possibility to calibrate capital maximum alert levels within buffers.
In our view, and if we acknowledge banks must be prudent on setting up an alert system which timely takes into consideration a breach of the combined capital buffer requirement, with swift escalation and notification process, and potential activation of the recovery plan, they should maintain the discretion, under specific circumstances, to use capital conservation buffers, or P2G, in case of stress, without triggering the execution of recovery options.
The comment applies also to calibration of MREL and TLAC recovery indicator, if deemed relevant (see Question 9).

Question 6: Do you have any comments on the proposed calibration of the recovery threshold for MREL?

See Questions 5 and 9.

Question 7: Do you have any comments on the proposed threshold calibration of regulatory liquidity indicators (LCR and NSFR) above their minimum regulatory requirement i.e. 100%?

The recovery plan should be and is articulated with the other prudential requirements such as contingency plan, resolution plan, together with their accompanying stress tests and their mitigation actions. (PLEASE SEE THE GRAPH AT PAGE 6 OF THE DOCUMENT ATTACHED)
The articulation is important so that each process is not confused with the other ones.
Hence, in an emerging / actual liquidity crisis, the liquidity contingency plan (LCP) would be first considered and potentially activated with its LCP mitigation actions. The LCP has its own monitoring metrics with accompanying threshold and governance to activate the LCP. The LCP mitigation actions would typically affect only marginally the business model / business franchise of the institution.
To the extent that the LCP mitigation actions would be deemed insufficient to satisfactorily restore the liquidity position of the bank, more impactful actions would be needed and the recovery plan would typically be considered and activated. This would enable to activate the recovery plan mitigation actions, which would cover a larger scope of actions with potential modifications of the business franchise / business model of the institution.
It should be noted that LCP mitigation actions would still be fully or partially available when triggering the recovery plan. Hence, and to the extent that the LCP mitigation actions would be deemed sufficient to restore the liquidity position, there would be no need to activate the recovery plan. The two processes should not be confused.
As a consequence, the recovery plan thresholds should be articulated with the same role of the liquidity contingency plan ones, which is to restore the liquidity position of the bank.
Usually, recovery plan thresholds, as well as liquidity contingency plan thresholds, result from traffic lights approach providing progressivity in the alert mechanism. Within this traffic light approach, it makes sense to have thresholds that are set above regulatory requirements for relevant metrics (e.g. Liquidity Coverage Ratio or Net Stable Funding Ratio).
However, the traffic light thresholds should be consistent with the considered metrics.
For instance, the role of the LCR is to maintain a liquidity buffer to be used in times of stress to offset net outflows. CRR is clear that this buffer should actually be used, if needed, even if it leads to no longer satisfying the 100% regulatory threshold. This has also been reminded by the European Central Bank during the peak of the sanitary crisis in March 2020. The use of the buffer is typically an LCP mitigation action. Other LCP mitigation actions can be implemented to restore the 100% LCR regulatory threshold (for instance, assets beyond the LCR liquidity buffer could be monetized). Hence, when calibrating the recovery plan threshold for the LCR, an institution should be able to consider the LCP mitigating actions. In that perspective, it makes sense to have both above and below 100% recovery plan thresholds for LCR, representing the increasing level of severity resulting from the traffic light approach.
The role of the NSFR is to ensure a sound structural liquidity position as measured at a one-year horizon. The 100% regulatory threshold applied to the NSFR means that there should be more available stable funding than required stable funding both measured at one-year horizon. Considering the structural nature of this metric and its horizon, a range of mitigating actions, usually listed in the LCP, that can be activated and deliver their impacts over time, could be implemented to restore the 100% threshold. To the extent that those mitigating actions would be deemed insufficient, the recovery plan would need to be activated. Hence, it also makes sense to have both above and below 100% recovery plan thresholds for the NSFR.
On a day-to-day basis, we agree that institution must prudently manage their regulatory indicators above 100%. However, we consider that an institution should keep the discretion to judge if, in a specific crisis situation, being below the usual minimum regulatory requirement should or not trigger the implementation of the recovery plan (e.g. a bank running temporarily below the 100% threshold, but able to restore its position via the implementation of LCP mitigation actions). As already underlined under Question 5 when referring to capital indicators and under Question 1 when discussing the benefits brought by a traffic light approach, a prudent set-up in terms of alert system will allow the bank to activate the recovery plan also before breaching the most severe indicators, if necessary and according to the specific circumstances.
Besides, regarding liquidity requirements, they are actually already representative of a stress situation. Therefore, they already integrate the deterioration of the environment before the actual liquidity issue arises. As these liquidity requirements already integrate a risk margin over the actual economic situation of the institution, it would not make sense to impose an additional margin in the thresholds.
As a consequence, the sentence “The thresholds for indicators based on regulatory liquidity requirements (LCR and NSFR indicators) should be therefore calibrated above the minimum requirements of 100%.”, as in paragraph 54 of the Guidelines, should be softened, allowing banks to set indicators based on regulatory liquidity requirements (LCR and NSFR indicators) above or below regulatory minimums, in due considerations of LCP mitigation actions and the flexibility envisaged by CRR.

Question 8: Do you have any comments on the proposed threshold calibration for the indicator of liquidity position?

First of all, EBA should clarify the definition of the “liquidity position” indicator.
Then, the recovery plan indicators should relate to risks being considered, i.e. indicate changes in the risk drivers or in risk measurement.
The risk drivers are identified in the risk identification process of the institution. It would make sense to leverage upon their identification to select the indicators, but changes in risk drivers do not mechanically translate into a higher risk position as it does not take into account the available mitigants. Conversely, risk mitigants may be valuable to monitor, although they do not necessarily convey the actual risk of the institution.
For example, the counterbalancing capacity (CBC), which is a risk mitigant, in isolation conveys only limited information on the actual risk of the institutions, which result from the comparison between the CBC and the potential net cash outflows (NCO’s) that it would help mitigate. Hence, having an increase in CBC while NCO’s would be increasing even more does not provide relevant information, and could even provide a wrong signal as it could be perceived as a less risky position.
That is the reason why we believe that the CBC and the liquidity position, which seems to relate to an extended CBC with central bank eligible assets (that are most probably already covered in the CBC), are not relevant recovery plan indicators.
In our opinion, recovery plan indicators should relate to:
• Risk metrics to measure the actual risk of the institution whereby risks are confronted with risk mitigants. Those metrics could trigger recovery plan. The institution should elect the most relevant of them in their minimum list of recovery plan indicators.
LCR and NSFR are two valid potential risk metrics that may or may not be substituted or complemented with other internal based risk metrics (e.g. Internal Liquidity Stress Test, Economic Liquidity Gap).
The available unencumbered assets central bank’s eligible, the so-called liquidity position and the combination of the two, i.e. the Counterbalancing Capacity, are not relevant risk indicators. Also, the suggested definition of those components is not clear in the paper.
• Risk drivers to identify potential trends in their variations; in isolation, those indicators should not trigger recovery plan as they do not take into account available mitigants. Institutions should elect their most relevant risk drivers indicators as additional recovery plan indicators.
The CBC and so-called liquidity position may or may not be relevant.
The Concentration of liquidity and funding sources, Cost of total funding (retail and wholesale funding), Average tenure of wholesale funding and Cost of wholesale funding may or may not be relevant as measuring risk drivers, (i.e. eligible to additional recovery plan indicators, and are not risk metrics and are not eligible for the minimum list of recovery plan indicators).
The Contractual maturity mismatch is usually neither a risk driver nor a risk metric and would usually not be relevant as a recovery plan indicator.

Question 9: Do you have any comments on the proposed changes to the minimum list of recovery plan indicators?

First, we approve that the proposed evolution of minimum list of recovery plan indicators be supplemented by the possibility left to institutions to substitute an indicator in a category when considered more relevant.
We believe that even if MREL/TLAC is a natural RAS indicator (it is a supervisory requirement the institution must fulfil), it is not an appropriate indicator for a recovery dashboard for at least two reasons, detailed below.
1/ We wish to point out that the purpose of the recovery dashboard is to identify potential recovery situations. MREL/TLAC is a regulatory requirement that is designed to ensure that banks have sufficient loss absorbing liabilities to be able to face up to a resolution situation without need for public solvency support. However, the fact that a bank may at some point fall slightly short of an MREL target is no way indicative of the existence, or potential existence, of a recovery situation. This indicator can be meaningful when approaching resolution only.
Unlike CET1, MREL/TLAC eligible debt must by definition be redeemed at its redemption date, and be re-issued at or around the date of redemption of the instrument reaching its end date. Circumstances completely independent from the health of an institution can lead to difficulties in renewing MREL that is reaching its redemption date, or the date at which it no longer qualifies as MREL. This has been clearly recognized by the legislator in BRRD2, via the introduction of mechanisms through which any application of the MREL Maximum Distributable Amount (MDA) mechanism may be suspended in circumstances such as market disruption.
In other words, MREL/TLAC is not an appropriate indicator for a recovery dashboard, because failing to meet MREL/TLAC requirements is not necessarily indicative of a recovery situation.
Should EBA finally decide to keep TLAC / MREL in its list of indicators, it should be specified which MREL will be considered: total MREL or subordinated MREL.
2/ The only risk specific to TLAC or MREL is the bank losing its access to the market.
In the list of the minimum list of recovery plan indicators in annex II, it is mentioned that “each indicator is subject to the possibility for an institution to justify that it is not relevant for it, however in such a case it should be substituted with another indicator which is more relevant for this institution”. We infer from this paragraph that institutions have some discretion not to provide the MREL and TLAC (where relevant) indicators if deemed not relevant.
Indeed, we signal that there are other indicators in the EBA minimum list or additional list that reflect the market confidence in an institution already. From this perspective, adding an MREL indicator could have the same effect as adding other liquidity indicators: it could unbalance the score of the recovery dashboard, generating the risk of an early activation of the dashboard.
See above (Question 8) the reason why we believe that the Available unencumbered assets central bank’s eligible and the Liquidity position are not relevant recovery plan indicators.
Regarding asset quality indicators, our Members signal that most of the stakeholders that monitor the banking activity usually focus on Non-Performing Exposure (as defined by EBA) indicators instead of Non-Performing Loans indicators. This could be considered in the list of indicators.
Referring to market-based and macroeconomic indicators, we would like to signal that:
1/ As stated in paragraph 62, “Macroeconomic indicators aim to capture signals of deterioration in the economic conditions where the institution operates, or of concentrations of exposures or funding”. Therefore, they should only be considered as an early warning indicator, not triggering the same process as for a liquidity indicator for example.
2/ In terms of notification (see also Question 3), we do not think that the breach of a sole of the above indicators should automatically trigger communication requirements to the management of the bank and/or competent authorities, since:
- Information on these indicators is public;
- For competent authorities, the result is that all banks will be affected similarly;
- Internally, this leads to communications to the top management, and implies at times complex internal process of notification, with limited added value.
In case these indicators are to be captured within recovery indicators, governance should be modified: at least, the sole breach of an indicator should have a different process of notification/escalation, taking also into account that some of the macroeconomic indicators are published by the statistical authorities with a significant time lag (e.g. GDP and Unemployment).

Question 10: Do you have any comments on the impact assessment?

Recalibration of indicators due to supervisory relief measures: the calibration of recovery plan indicators should be adjusted due to supervisory relief measures, as these are typically systemic in nature.
Specific timeline for the notification of the breach to the supervisor: this option grants one business day for the escalation process and one additional business day for the notification to the competent authority, which seems very tight.

Upload files

Name of the organization

European Banking Federation