Deutsche Boerse Group / Clearstream / LuxCSD

The proposed approach refers to a capital charge for operational risk. The method proposed would require splitting the total business into pre-defined business lines and allocating all income to any one of those. The business lines are fully pre-described in Article 317 (4) CRR which does not include a business line “third party custody” but rather refers to “Agency services”.
As such, all custody service income would have to be allocated to that business line. This, however, would put the income from custody for clients a second time under capital charge which cannot be intended. Moreover, third party custody is not creating income but rather cost. All in all, the proposed method does not seem to fit at all. It further needs to be noted, that the BCBS has already proposed to withdraw the current standard method as is has not demonstrated to correctly reflect the various business lines. We therefore reject the proposal.

More generally, we reject to add a dedicated charge for custody risk in whatever form on top of its coverage within the general framework of the operational risk capital charge. To our understanding derived from the consultation paper itself, even EBA in general agrees to this approach for client custody. Consequently, a separate charge for any (sub-)custody with third parties does not make sense:
• Any income form own securities is included in the relevant indicator for the Basic Indicator and Standardised Approach and the risks related to the investments are to be captures appropriately under the AMA.
• Any income from client custody business is to be captured NET for the relevant indicator under the Basis Indicator or Standardised Approach. Moreover, under certain conditions as specified in Article 316 (1) CRR even some outsourcing expenses can be used to reduce the relevant indicator in addition.
• Any related risk for (sub-)custody of assets is to be captures under the AMA.

All in all, this demonstrates that an additional charge is not justified. The level 1 requirements to include (not just) operational risk but also legal and custody risk in our mind is immanently embedded in the operational risk charge of the banking framework and as such does not require any add on for CSDs.

In addition to the general comments made above and the answers to the specific question raised, we do want to comment on dedicated topics regarding Articles 1 to 5 the EBA draft RTS proposal as follows:

Capital definition (Recital 4 / Article 2)
The proposed regulation refers – like the delegated regulation (EU) No. 152/2013 – with regards to the capital requirements for “investment risk” to the capital rules for credit (and market) risk as defined in the regulation (EU) No. 575/2013 (CRR) and also defines operational risk based on CRR rules. However, CRR defines capital and in addition positions to be deducted from capital. Consequently, either the same capital definitions should be used and therefore the treatment of the required capital deductions would be the same or a dedicated treatment of such items under the “investment risk” should be given. Furthermore, Article 46 (4) of CSD-R foresees an element of deduction from capital on its own which needs to be incorporated in the framework. For the majority of CSDs capital may only consist of paid-in capital, retained earnings and reserves and neither additional capital components nor (complex) deductibles are to be recognised. As such we clearly favour an approach to use the capital definitions from CRR only modified by a rule to reflect Article 46 (4) CSD-R for all CSDs.

We therefore propose to reword Recital 4 / Article 2 as from the attached document.

Capital requirements for operational, legal and custody risk
(Recitals 9 and 10, Articles 3- 5)
We clearly agree to the usage of the operational risk capital charge method as derived from the banking framework. CSD-R does not give a definition of “operational risk” but adds the need to cover also legal and custody risk to the requirements. Delegated Regulation (EU) No. 152/2013 includes under similar considerations the “legal” risk as part of the operational risk capital charge for CCPs. We therefore also clearly agree to the same for legal risk in the draft RTS for CSDs.
Related to the custody risk and potentially sub-categories of custody risk, the draft RTS however is not following a consistent approach and we clearly disagree to the approach proposed. The management of any (potentially) additional risk of CSD links is part of the respective rules on CSD-links (Articles 50 to 52 CSD-R and related provisions in the ESMA RTS). However, we do not regard risks from CSD-links to be different from those of general sub-custody. Any “normal” credit institution or investment firm subject to CRR / Basel III offering custody services is using sub-custody and finally holds the client assets in custody with the CSD at the upper end of the custody chain. As there is no dedicated capital charge for this sub-custody, we cannot see a different treatment for the same activity being performed by a CSD. The same argument is true also for the custody of proprietary assets of any credit institution or investment firm. Any such risk is covered in the Basic Indicator Approach (BIA) or with dedicated risks if deemed necessary in the Advanced Measurement Approach (AMA).
In the BIA as well as in the Standardised Approach (TSA) revenues are used as relevant indicator and even some external costs (e.g. commission expenses like the ones for sub-custody and custody for own securities) are deducted from the relevant indicator. It is therefore the custody margin which is included in the relevant indicator.
In summary, already the banking framework is covering any risk of third party custody and this should not be added once more. Custody risk is – like legal risk – part of the operational risk capital charge under CRR, consequently, we disagree to adding additional elements on top of the banking framework for that purpose. However, although the reference in Article 3 (1) (a) draft RTS and the header of Article 4 draft RTS are referring to “custody risk”, the text of Article 4 (1) draft RTS does not include custody risk. We think that this has been done by mistake and kindly ask to include “custody risk” in the wording of Article 4 (1) in the final version of the RTS. Moreover, we clearly disagree to set up a dedicated capital requirement for the risk of sub-custody and custody for own securities (see above) as proposed in Article 5 draft RTS.

The Basel Committee (BCBS) is currently discussing the introduction of floor rules for the advanced / internal approaches based on the relevant standardised methods. BCBS d306 is proposing a general floor for model based approaches based on the standardised approaches of the same risk category. With regards to operational risk that would be the Basic Indicator Approach (BIA). The Basel Committee has not proposed yet any level of the floor and also has not come up with a final framework as the BCBS d306 consultation only closed on 27 March 2015 and has received various comments raising strong concerns. While we in general oppose these floor rules and rather urge regulators to calibrate the internal models , we clearly oppose an introduction of a floor by the draft RTS being different from the BCBS ones in case introduced. Any floor finally set by the BCBS is supposed to be taken over by the EU banking framework (CRD / CRR), which is currently proposed to be used for CSDs as well. As such, in case any floor is implemented, it will be applicable for CSDs as well. Furthermore, the current Basel I floor which is – correctly – not proposed to be put in place for CSDs had been designed to take care that the “new” Basel II rules for differentiated credit risk and newly introduced operational risk (at the time in 2004) do not fall short compared to former rules (Basel I) for credit risk only. Therefore, it does not make sense to introduce a floor on operational risk in order to comply with non-existing past rules for credit risk for CSDs. As such, for the reasons explained above we clearly favour to remove Article 4 (5) from the proposed RTS.

In this context, it is to be noted that due to a market wide increase of legal and compliance matters, we observe a sharp increase of capital charges under the AMA for these types of risk. Contrary, due to the low levels of interest, net interest income is become less a driver for the relevant indicator under both the BIA and the TSA. These two movements clearly contribute to disproportionate levels of capital requirements for CSDs under the AMA compared to the BIA. However, even without these current drivers, the AMA for CSDs is leading to far higher requirements than the BIA would. In the case of Clearstream’s the two CSDs and the Financial Holding Group falling currently in scope of CRR, a rough calculation of BIA as comparison to the current AMA shows – based on different allocation consequences between the two methods – levels of 30 to 45 % under BIA compared to the current AMA capital charge. As such, the introduction of the mandatory operational requirement to calculate the BIA capital charge while the AMA will most likely generate much higher capital charges seems not to be justified.

Finally, we do want to point out that currently both the BIA and the TSA are under review by the BCBS and a proposal to merge the two towards a modified BIA only has been issued in 2014. We also expect a review of the AMA in the near future. As such, any modification / addition to the banking framework may have to be revised once the new rules are implemented in the EU which could be avoided by refereeing to the complete banking framework only.

As such, we propose the following amendments to the Recitals and the draft Articles, as from the attached document.
We agree in general to the proposed elements of the draft RTS with the following exceptions:
(1) The introduction of a capital charge for custody risk limited to items given to third parties in (sub-) custody does not make sense. The general CRR approaches towards operational risk cover in our view custody risk in full. (see above)
(2) We see the need to cover business risk not only with capital but also with (expected) income. At least with regards to the estimation of business risk impact, the related (expected) income needs to be deducted from the gross estimated risk costs.
(3) We disagree to the proposed floor to capture the capital requirements for winding down or restructuring. Here once more the alignment with the EMIR rules for CCPs does not seem to be appropriate. Especially for smaller CSDs the floor of 6 months operation cost for winding down or restructuring may be too high and EBA should consider lowering this. Especially in markets where more than one CSD is available, the period of 6 months seems to be too pessimistic. We however acknowledge the limitation of the level 1 text in this regards. In addition, we doubt that continuing for the full period of restructuring or winding with the average cost of the past is also too pessimistic as the changed business is supposed to be as a general rule of thumb less costly (strategic costs and IT development is supposed to be reduced and general operations in a supposed declining business is also supposed to be lower). As such, the cost base as of the 4th month (in case winding down / restructuring is supposed to take that long) should be assumed to be lower than 100 % of the past (e.g. 80 %).
(4) The proposed floor on operational risk charges when using the AMA is rejected as it does not make sense to introduce it at all and especially not for CSDs only in advance of any (possible different) floor for banks which is currently in discussion.

All in all, it needs to be noted that currently CSDs not operating under a banking license in general have limited rules for capital requirements. Often only a minimum amount is fixed (sometimes as a portion of annual costs) without setting any risk related requirement. As such, EBA needs to consider a phase-in approach for such CSDs not offering banking type of ancillary services currently to implement the necessary tools for the calculation of the requirements and its fulfilments. We propose to have a phase in period of 30 months at least. During that period a simple approach to calculate the capital requirement based on total cost or the volume of assets under custody is proposed. This simple ratio should be fulfilled at least with 60 % at the point in time the application is filled, with at least 80 % six month after application is granted and in full one year after application is granted. Such an approach would allow smaller CSDs to collect any needed capital with sufficient lead time. Of course the application should clearly indicate how the CSD is (a) going to increase the level of capital based on the interim rule in case it falls short at the time of application and (b) how and when it is going to implement the necessary calculation tools. The tools have to be available for calculation purposes at least 18 months after application is granted in order to give additional 12 months for any needed additional capital measure on top of the interim rules.

In addition to the general comments made above and the answers to the specific question raised, we do want to comment on quite a few items in more detail and raise some questions on dedicated topics regarding Articles 6 to 8 of the EBA draft RTS proposal as follows:

Capital requirements for investment risk (Recital 11, Article 6):
In general a CSD does not take assets of its clients onto its balance sheet on a trustee basis. Securities taken as collateral or into custody are not shown in the balance sheet of a CSD and remain in the ownership of the clients. However, CSDs offering banking services are taking cash deposits in order to perform settlement and custody activities and all CSDs (may) take to a small extent also cash as collateral for various purposes (though securities collateral is the majority of any collateral taken). Any such client cash taken as deposit or collateral is placed by the CSDs in their own name and never “on behalf” of the clients as cash becomes part of the CSDs assets and is not segregated as a client position. In case a CSD would act as a trustee for its clients, it would not bear the underlying investment risk. As such, the wording of Recital 11 needs to be adjusted accordingly. In addition, also loans to clients/Participants and other assets should be subject to the “investment risk”.
Regardless of the authorisation to offer banking type of ancillary services, a CSD should be allowed to use all available methods for Credit Risk Mitigation as offered by CRR.

As such, Recital 11 and Article 6 (2) of the draft RTS should be amended as from the attached document.

Capital requirements for business risk (Recital 12, Articles 7, Annex II):
As stated in Recital 12 draft RTS, business risk can lead to reduced revenues or increased cost and subsequently in a reduced net income which may lead even to a loss. However, only in case the business risk is severe or net income is very low, a loss to be borne by equity may be the consequence. As such, a large portion or even all of the business risk is covered by current income and any floor to be covered by equity seems to be more than artificial (in this regards our criticism is already valid for the respective rules on CCPs in Delegated Regulation (EU) No. 152/2013). As such, this clearly needs to be stated in Recital 12 draft RTS and no floor should be introduced in Article 7 draft RTS.
Our argumentation is further substantiated also in the scenarios proposed in Annex II which refer to “a loss of 30 % income” (note, it is not clear what is meant by “income” and this should be further specified. We believe here gross revenues are in scope), “unexpected increase of funding cost of 10 %” etc. which all may or may not be covered in full or parts out of current / forecasted income.

Beside (i) Delegated Regulation (EU) No. 2015/488 as regards to own funds requirements for investment firms based on fixed overheads for the purpose of Article 97 (4) CRR and (ii) Article 4 (2) of Delegated Regulation (EU) No. 152/2013 for CCPs, Article 7 draft RTS is introducing yet another definition of an “appropriate cost base” for more or less the same purpose. We clearly urge the EBA and the EU Commission to come up with a consistent and coherent approach across all areas of capital requirements (i.e. at least CRR, CSD-R and EMIR). As we disagree to have a capital floor for business risk, also the definition of “annual gross expenses” needs to be removed from Article 7 draft RTS and needs to be included in Article 8 draft RTS instead. We therefore propose a unique approach to be implemented consistent for all financial market regulations as an adjustment to Article 8 draft RTS as specified below (in the proposed rewording of Article 8). As such, we clearly favour a definition which is not referencing to IFRS as IFRS does not specify certain categories of expenditure but to refer to directive 86/635/EC which does define categories of expenditure – to our understanding – to a sufficient degree. Such an approach would be in line with the general approach of Article 316 CRR for the relevant indicator for operational risk.

As such, we propose the following changes to the draft RTS as from the attached document.

As regards Annex II (p.68) on business risk scenarios, whereas we agree with the requirement for regular stress tests of business risks, we are of the opinion that the stress scenarios outlined in Annex II are too prescriptive and do not address the main business risk drivers of (I)CSDs. Consequently, this requirement will create additional workload for the CSDs having to perform such tests without tangible benefit in terms of risk management.
Moreover, we are not convinced that predefined business risk scenarios are the most appropriate means to calculate capital requirements for business risk, and we believe that a more flexible approach, similar to that in EMIR, which prefers the adoption of reasonably foreseeable adverse scenarios relevant to the CSD’s business model, as approved by the competent authority, could be more efficient and proportionate.

Capital requirements for winding-down or restructuring (Recitals 6 – 8, Article 8):
As during a restructuring or wind-down of a CSD quite some expenses (IT-development, strategic initiatives, declining activities) may well lead to (substantial) decreased cost, the wording of Recital 6 is in our mind wrongly only indicating in an upwards direction but not taking (the more likely) downwards development of costs into account. The wording therefore should be adjusted.

Furthermore, as CSDs are in general not parties of any given settlement transaction, there is no need to close out pending transactions / open (derivative) positions for CSD’s participants like in the case of a CCP. As such, the winding down / restructuring process for CSDs is totally different from those of any CCP. Consequently also the minimum winding down period should in principle be lower compared to CCPs. Especially within the T2S environment, participants should be in a position – also taking their mandatory business continuity and restructuring plans into account – to move their assets quickly via free of payment transactions to another CSD in case of a winding down event. As it is the duty of the CSD to develop a proper restructuring / winding down plan and it also needs to determine in this plan an appropriate schedule for its implementation in case needed, we do not see the need to define any minimum timeframe for the capital charge. The plans as such are in addition subject to review and agreement of the relevant authority.

However, a timespan of at least 6 months is referred to in CSD-R itself. In order to reflect our arguments above, EBA therefore should at least consider reducing the cost basis for any month beyond a three month period to 80 % of the general cost basis and take into account even lower percentages for longer winding down periods – if applicable - as also income will flow during that period. Furthermore, the timespan for winding-down or restructuring is not strictly depending on the services provided but mainly on the volumes of assets under custody, number of participants and securities administered at the upper tier of the custody chain. Finally, it is also dependent on the possibilities of alternative CSDs being able to offer same kind of core services. Any additional service which does not belong to the core CSD activities may also be stopped without the mandatory need of (immediate) replacement by another services provider.

While we agree on some common guidelines for the estimation of the winding-down or restructuring period and the underlying scenarios, we disagree to set too detailed requirements.
We recommend sorting the Recitals in the sequence of the topics raised in the legal text itself. Therefore we propose to shift Recitals 6 to 8 after Recital 12.

Subsequently, we propose a re-wording as from the attached document.
As explained in section C of our comprehensive response to the consulation, there are several open questions which are in our mind not properly specified in the proposed approach. Furthermore, we see a lot of weaknesses in the approach and overall we clearly see a tendency to massively overcharge capital needs for intraday credits.
As such, we do not agree to the proposed approach.
In case the approach is implemented nevertheless, CSDs already operating today need to get an appropriate phase in time to be compliant with the massive operational requirements and the additional capital needed. In this case, we propose a similar phase in approach as proposed in question 2 with regards to the capital charges for smaller CSDs as follows:

• Grant a general phase in time of 30 months after granting the authorisation for full compliance with the rules.
• Put in place an interim add on of 50 % of the capital charge for overnight credit risk to be reached latest by 6 months after granting the authorisation and 100 % of the capital charge for overnight credit risk to be reached latest by 12 months after granting the authorisation.
• Put an obligation to start calculating the capital requirements not later than 18 months after granting the authorisation.
• Request a clear plan with the application on how the requirements will be implemented and how capital shortages – if any –will be filled.

Moreover, we would prefer a totally different approach to calculate credit risks from intraday credits as explained further along this document.

We believe that such a model is reflecting the intraday credit risk, is generating a reasonable level of capital charge and it is also reflecting the collateralisation appropriately. Furthermore it has the benefit of being easier to implement than the complex proposal of EBA. However, also this method will need some time for implementation.

We therefore propose also in this case a phase in approach:
• Implementation and fulfilment latest 12 months after granting the authorisation.
• Additional capital surcharge of 50 % on the capital requirements for (overnight) credit risk to be reached latest 6 months after granting the authorisation.

In addition to the general comments made above and the answers to the specific question raised, we do want to comment on dedicated topics regarding Articles 9 to 13 of the EBA draft RTS proposal as follows:

Capital surcharge for intraday credits (Recital 13 -15, Article 9)
Article 54 (3) (d) and Article 54 (4) (e) CSD-R respectively ask for an capital surcharge to cover the risk including credit and liquidity risk for the provision of intra-day credits to the participants in its securities settlement systems or other users of CSD services. It is important to note, that only risks coming for (a) the provision of intra-day credits and (b) coming from the active offering of a CSD in our reading of the text are covered by this rule and any other exposure which is not subject to these two conditions is not in scope. In practise this is focussing on if not limited to cash credits towards participants including advances for custody payments.

The view expressed above is reflected in Recital 13 draft RTS but not taken over in full in Article 9 draft RTS.

Furthermore, CSD-R does not define on what basis the capital surcharge for the intraday credit risk is to be calculated. Therefore, EBA has some freedom in the way the rules are defined.

Currently, intraday credits provided to participants are managed by CSD using collateral pools per client (or even per dedicated accounts of a client, i.e. there may be several loans and collateral pools per client).

Credits are granted giving maximum external communicated amounts and (higher) internally approved lines. Their usage is however limited to the collateral value of eligible collateral (market value less haircuts defined on internal rules).

The respective exposure of a given client is monitored real time against the limits and the collateral value which needs any time at least to be equal to but in general exceeds the exposure value substantially. In order to assess this, only the total collateral value of the dedicated collateral pool is assessed and no dedicated allocation of specific collateral occurs and also no structural requirements on the content of the pool (of in principle high liquid assets) exist. As such, the mandatory collateral allocation of Article 18 (1) (a) draft RTS (“maintain collateral segregated from the other securities of the borrowing participant”) in combination with Article 18 (1) (b), 19 and 20 draft RTS (clear sequence of collateral used and limitation for other collaterals only in cases where the highest quality is not available any longer) of the draft RTS would not allow using the total collateral pool any longer but require dedicated collateral (with only the value necessary to cover the current exposure and therefore losing the collateral value of available collateral exceeding the allocated portion). We reject such an approach (see our comments to Articles 18 and subsequent draft RTS) as impact on the calculation of the intraday credit risk capital surcharge is severe. This is true for technical considerations but also with regards to the result of the calculation as such.

In addition, the calculation of the capital charge for credit risk based on CRR already includes risk weights and specifically haircuts. The haircuts foreseen in CRR take into account quality of the collateral, remaining maturity of the collateral and any potential currency mismatch. Moreover, the haircuts are using liquidation periods of up to 20 days and are in general substantially lower than the proposed 10 percent haircut for very short liquidation periods. Especially for the short term duration (intraday only) the standard liquidation period of 5 days would be most appropriate and would be by far lower. As such, we disagree to such a flat haircut which is not taking care of the underlying market risks.

In addition to the flat and very high proposed haircut, the proposal foresees to use the complete CRR model, i.e. to have the CRR haircuts on top of the proposed haircut of 10 percent, which represent a double haircut we clearly reject. In our opinion, this 10% add-on is excessive since: haircuts already incorporate credit, market and liquidity risk factors – among others – and on average, far exceed the proposed stress factor. Therefore, we are of the opinion that the stress factor should be eliminated. Should the collateral stress factor be applied as proposed, it may force CSDs to further increase the haircuts applied to collateral securities, which will have a negative impact on participants – either by higher collateral costs or by reduced settlement efficiency.

While we see benefits for the general approach to calculate the capital charge for intraday exposures based on a “peak” situation and using the general CRR rules for overnight exposures for that purpose, we disagree to a number of elements of the concrete proposal. On top of that, the definition of “peak exposure” lacks clarity in at least the following elements:
1. Should exposures or risk weighted exposures be used?
2. Should exposures / risk weighted exposures be used prior or after taking credit risk mitigation into account?
3. How are “peak exposures” taking the difficulties of clearly timing in- and outflows into account (we refer in this regards to our comments on the BCBS consultation on “Monitoring indicators for intraday liquidity management - BCBS 225” )? Based on the ability of cash correspondents to deliver an exact time stamp of any payment (i.e. an income payment to the account of the CSD’s participant), the exact timing of settlements and the related cash flows (the technical settlement and the legal settlement may occur at different times with up to several hours difference), the different technical cycles to process cash movements (real time, near time, rapid batch, few daily batches only) of CSDs counterparties, the real cash position of a client may not be derived properly. While for credit monitoring and measurement the current book position is usually taken into account, for a capital charge in principle only the real legal risk should be considered. This is difficult – if at all – to be determined at any given time during the day (however, can be easily derived as per end of any given day). In this context we also want to mention back valuations and corrections as well as cancelations of payments.
4. Is this the value of the peak credits used by all clients (total of all clients) at any given point during the day or the total of all peaks per clients (total peaks) of that day?

We see an alternative measure to the proposed treatment as being superior although simpler. This is based on granted / approved (intraday) lines and general collateral criteria by using nevertheless to a large degree the general approach for credit risk under CRR with some slight modifications. This approach is similar to the one used for exposures to investment funds (Article 132 (5) CRR) when looking through based on the mandate as well as the conversion of off balance sheet items base on Credit Conversion Factors (CFFs) as defined in Article 111 (1) in conjunction with Annex I and Article 166 (8) of CRR respectively:

Total amount of (approved) [cash] credit lines being available for intraday usage are converted into exposures using the credit conversion factor for committed but unconditionally cancellable credit lines (currently 0 % as per Article 111 (1) in combination with Annex I and Article 166 (8) (a) CRR respectively; however, we propose to use 5, maximum 10 % for the time being as a floor).

Apply haircuts according to CRR Article 224 on collateral expected (market value to be assumed in the amount of the line granted, no access market value corresponding to haircuts taken to be assumed) according to the internal rules for collateral eligibility whereby the lowest quality and longest maturity is to be assumed and a 5 day liquidation period is to be assumed (note: The EBA proposal does not take undrawn lines into account. We regard this as being appropriate as lines which are not properly collateralised will not allow to be drawn and lines which are collateralised should receive a net exposure value close to zero or even zero).
As collateral taken will have internally set haircuts, the real market values will exceed the credit lines given. As such, we propose not to apply any cross currency haircut.
It is to be noted, that overnight drawn lines will receive in addition the standard capital charge for credit risk. (double counting accepted in this approach but also included in the EBA approach).

We believe that such a model is reflecting the intraday credit risk consistent with the generic requirement as set in the level 1 text, is generating a reasonable level of capital charge and it is also reflecting the collateralisation appropriately. Furthermore it has the benefit of being easier to implement than the complex proposal of EBA. However, also this method will need some time for implementation.

It is to be noted that counterparty risk weights are given for any kind of counterparty in the CRR. As such, the dedicated mentioning of certain counterparties in Recital 15 in our understanding does not make sense. A general reference to the CRR should be sufficient for that purpose.

While the EBA proposal is referring to the most recent calendar year , an approach based on authorised credit lines could capture most recent data and would be far closer to current situations also with regards to collateral arrangements.

Following our proposal, the relevant Recitals and Articles should be modified, based on two possible options as from the attached document.

Finally, EBA should also duly consider a proper time for implementations of these proposal in this RTS by the inclusion of dedicated and reasonably adapted “transitional Articles”.
The financial industry is currently preparing solutions to comply with the requirements on intraday liquidity monitoring set out in BCBS 248. Once these are in place, we expect to be able to have a near time view on our positions with our key cash correspondents. An important feature of BCBS 248 is a materiality threshold, which we think is critically important. While we expect our key cash correspondents to be able to provide real time transaction data, providers in less developed markets (without corresponding regulatory requirements) may not be able or willing to provide the required data.

In addition to the general comments made above and the answers to the specific question raised, we do want to comment on quite a few items in more detail and raise some questions on dedicated topics regarding Articles 10 to 17 of the EBA draft RTS proposal as follows:

• There is to our knowledge no requirement in the CSD-R or the (draft) ESMA technical standard requiring monthly meetings of the risk committees. As such, we disagree to the requirement of sending monthly reports to the risk committees for review as requested in Article 11 (4) as well as Article 33 (2) of the draft RTS. We consider in addition, that the frequency should be proportionate to the business of the CSD and / or the credit institution nominated in accordance with point (b) of Article 54 (2) CSD-R. Moreover, Article 11 (1) – (3) draft RTS do not specify any frequency of that report. However, based on the wording in Article 11 (2) draft RTS the frequency of at least annually may be assumed. We therefore propose to include an additional sentence to add that the policy should also define the frequency of the report to be prepared which should be subject to the risk profile of the CSD and should be “at least quarterly”. Subsequently, the reference to the timing (“monthly”) can be completely taken out form Article 11 (4) of the draft RTS.
• With regards to the measurement on monitoring of overnight credit risk, the question arises if the data to be used and stored is to be value date corrected or not. We propose to leave this to the discretion of the CSD.
• In certain jurisdictions, collateral can only be taken based on written agreements. As any written agreement contains a certain level of commitment, in this circumstance we cannot agree to the term “granting only uncommitted credit lines” as a granting of a line in such circumstances would always contain a commitment.

As such, we kindly ask to rephrase Recital 21 as follows:

Recital 21
As a result, a CSD-banking service provider should be permitted to grant committed lines only on the basis of unconditional revocability at any time to borrowing participants in the course of …


In this context, we strongly oppose to the limitation of granting only uncommitted credit lines as requested by Article 16 (c) draft RTS. We however propose to refer to the need to grant only credit lines which can be cancelled unconditionally at any time.
• Credit limits for secured credits – which is the standard towards participants as also required by Article 59 (3) (c) CSD-R – are only valid up until the amount of available collateral value (i.e. market value less haircut). As such, Article 17 (e) (ii) in this context is not meaningful as the credit limit does not need to be reduced if it is “dynamically” capped on the collateral value of collateral available.
The participant decides on the securities being made available for collateralisation purposes. In general only those securities can be used by the CSD based on its eligibility criteria for collateral purposes. Usually the use of collateral is depending on a written agreement. As such, this should exclude participants’ client assets (unless the use is for the purpose of the participants’ clients) and may be limited only to a part of the whole portfolio of participants’ assets. Therefore it needs to be clarified how the status of “all securities in the account of the borrowing participant” is to be determined. We ask to change this to “all securities of the borrowing participant being available for collateral purposes …”. Furthermore, the sequence of Article 19 (1), 19 (2), 20 (1) and 20 (2) in its current form is not clear.

Article 18
Clearstream is of the opinion that its collateral management policies, procedures and systems already ensure compliance with Article 59 (3) (a) through (e) of Regulation (EU) No. 909/2014 (“CSD-R”) and are generally in line with the provisions of Article 18 (1) 1 of the draft RTS.
However, Articles 18 to 23 of the draft RTS introduce certain monitoring and reporting requirements that are incompatible with the current collateral management concept used by the (I)CSDs i.e. the pooling of collateral in the participants’ accounts. In addition, Article 18 (1) (b) draft RTS introduces a hierarchy of collateral which will require the development of a sequencing algorithm capable of allocating collateral according to pre-determined quality criteria. Such mechanism cannot be supported by the existing “collateral pool” concept and will require significant IT development.

According to Clearstream’s current collateral management system, when Clearstream extends a cash loan to a participant, the amount required to fully secure the loan is reduced from the participant’s collateral pool (after a haircut has been applied to the market value of the securities comprising the pool). Conversely, when a cash loan is repaid, the equivalent amount is made available as collateral to the participant. If the participant has not enough collateral to secure the cash loan, the transaction fails. This flexible methodology – which was designed taking into account the role of Clearstream as a securities settlement system – ensures that no particular security is segregated or blocked from the collateral pool of the participant. As a result, as long as the participant has a sufficient collateral value in the account(s) pledged to Clearstream, it is free to deliver any security from the pool.

As an example, assume that a participant has a collateral pool valued at 1,000€ (after haircuts) and that the pool is comprised of 500€ of German government bonds, 250€ (equivalent) of UK Gilts and €250 (equivalent) of US Treasury bills. Should the participant require a loan of, say, 700€, Clearstream’s collateral management system would reduce the value of the participant’s collateral pool to 300€. However, no specific security would be blocked from the participant’s account. This means the participant would be able to sell / deliver up to 300€ of a combination of the securities – at the participant’s choice – as long as the amount sold / delivered is equal to or lower than 300€. The operational flexibility provided by the concept of “collateral pool” is highly valued by participants and ensures very high levels of settlement efficiency.

We are deeply concerned that the new requirements will force the (I)CSDs to identify and segregate (possibly from the participant’s pledged account to another account) individual securities for the amount of the cash loan. This will require significant IT development to re-design the (I)CSDs collateral system. In addition, we at Clearstream are concerned that, in order to maintain equivalent levels of settlement efficiency as those achieved today by the (I)CSDs, participants might have to immobilize at the (I)CSDs a significant amount of securities of the highest quality. This could worsen the already reduced supply of high-quality securities available for collateral purposes .

As the proposed way is deviating from current processes, technical implementation is necessary which takes approx. 24 month time to be implemented. As such, the draft RTS needs to foresee for existing CSDs a phase-in solution to be compliant with the requirements. We would strongly suggest EBA to consider adding an Article with transitional provisions that would consider the implementation times for the different provisions.

Firstly the requirement under Article 18 (1) (c) (ii) draft RTS in our understanding does not make sense. In case all collateral as referred to in Article 19 and 20 draft RTS are used (requirement of Article 18(1) (b)(ii) draft RTS) this includes collateral as defined in Article 20(2) draft RTS and as such, they are not available in addition as they are already used. We kindly ask the EBA to look into that text and clarify the intention. Possibly, Article 18 (1) (b) (ii) should only refer to Article 20(1) instead of Article 20.
In particular, Article 18(1)(d) of the RTS requires CSDs to monitor on near real-time basis the credit quality, market liquidity and price volatility of each security....". Considering the very strict eligibility requirements imposed by Articles 19 and 20 draft RTS, the need for "real-time" monitoring appears disproportionate. In addition, the RTS would need to provide further clarification on the concept of “market liquidity” and on which metrics will be required to be monitored. Alternatively, and considering the difficulty in establishing one or more metrics for the monitoring of market liquidity, the EBA could consider deleting the requirement for the monitoring of “market liquidity”, as we propose in our review of Article 18 hereunder.

Article 19
As mentioned above, Clearstream does not agree with the quality-driven hierarchy of collateral introduced by Article 18 draft RTS and further detailed by Articles 19 and 20 draft RTS. In Clearstream’s opinion, there should be one eligibility criteria which must be sufficiently flexible to accommodate the complexity of the (I)CSDs businesses, which include the safe-custody and settlement of a wide range of debt and equity instruments, issued by private and public-sector institutions, in a variety of jurisdictions and across many currencies. The segregation of collateral in different “quality buckets” will require a substantial investment to develop new collateral management and reporting systems. Furthermore, from a market liquidity perspective, this fragmentation of collateral will contribute to further increase the attractiveness of (and, consequently, the demand for) high-quality (particularly top-rated government) securities at the expense of other securities.

In Clearstream’s opinion, Article 19 draft RTS is confusing: it is not clear why the text separates between “financial instruments considered as highly liquid collateral bearing minimum credit and market risk” (as per paragraph 1 of Article 19 draft RTS) and “transferable securities and money market instruments [shall be] considered as highly liquid collateral bearing minimum credit and market risk” (as per paragraph 2 of Article 19 draft RTS). Once again, this artificial segregation of securities will create enormous monitoring and reporting challenges for unclear reasons and benefits.

Specifically, Article 19 (1) of the draft RTS introduces an excessively high number of conditions (a total of eight) for a security to be considered as “Highly liquid collateral with minimum credit and market risk”. We oppose, in particular, to the following conditions:
• On 19 (1) condition (a): it is not justified to exclude securities issued by highly-rated private institutions from the first-tier of collateral.
• On 19 (1) condition (b): currently, Clearstream relies on external assessments of the credit quality of issuers of collateral. Securities issued by institutions which do not have an external credit rating are not eligible as collateral. The requirement for CSDs to prepare their own assessment of credit and market risk of securities accepted as collateral will require a significant expansion of the credit and risk departments of those institutions. This requirement appears disproportionate, particularly given that only governments and central banks are eligible issuers, as per condition (a).
• On 19 (1) condition (c): we consider the provisions therein relating to the average time-to-maturity of highly liquid collateral, as not being appropriate and on top of that being unclear content wise.
Firstly it is unclear to whom this requirement is being addressed to, as the reference to “the banking services provider’s portfolio” in our mind does not make sense. It is our understanding that the provisions of this Article 19 (1) (c) relate to the financial instruments which are being provided as collateral by the participant. Hence the reference should be – if at all – to the total amount of collateral delivered by any given counterparty.
Nevertheless, the text is confusing and appears hardly practical to be implemented – as it would entail the requirement for the CSDs to dynamically calculate the average time-to-maturity of the portfolio of securities pledged by participants to ensure it is two years or less. Moreover, the requirement raises the question whether the requirement would apply to portfolios pledged by individual participants or the total collateral portfolio of all participants. Or whether the requirement encompass “securities settled in the trade” i.e. securities which have been purchased but not yet paid for by the participants In addition, condition (c) assumes that medium- and long-term securities are not liquid, which is not accurate.
Furthermore, (negative) impacts of longer maturities due to higher risks of price fluctuations can be sorted out / managed by adequate haircuts. However, at the time of taking the collateral the structure of the total collateral pool may not be known and it may be subject for changes which could subsequently make collateral less liquid just due to the fact, that short term maturity collateral has been withdrawn. It is obvious, that liquidity of any given security cannot rely on the composition of a collateral pool as it is the market and not the collateral pool that defines the liquidity of any given security.

Having said the above, should the Article 19 (1) (c) draft RTS, actually propose restricting the investment portfolio of the CSD to an average time-to-maturity of not more than two years. Then EBA should consider, similarly to the comments we have provided to ESMA, that given the extremely high quality requirements and the experienced scarcity of available securities, such a requirement would make it virtually impossible for CSDs to maintain an appropriate investment portfolio. The investment portfolio is used to build the necessary (and much required) liquidity buffer of a CSD. We understand that the main consideration for this restriction was based on liquidity considerations. We do not see any evidence in the market that securities with a time-to-maturity of more than two years are less liquid. If the restriction cannot be removed completely, we suggest extending the average time-to-maturity restriction to at least five years. Also here, the time to maturity of the investments of CSDs cannot be compared with the EMIR-requirements of a CCP which have a completely different liquidity requirement. As such, it seems to be unreasonable to limit the investment possibilities with a focus on liquidity and not coming from a market risk view (which we tend to refuse). Furthermore, in combination with the discussed prohibition (via the ESMA RTS on CSD-R) for hedging the portfolio, this is reducing the investment options substantially especially in times like the current once where government issue rather long term debt in order to capture the favourable interest levels.
In the light of the above, we are proposing the deletion of such requirement.

Furthermore, still on the same Article 19 (1) draft RTS we have the following remarks:
• On 19 (1) condition (g) requires price data on these securities to be “publicly available on close to a real time basis”. The latter part of the requirement i.e. “close to a real time basis” is problematic for debt instruments.
• On 19 (1) condition (h) requiring on the same day basis liquidation of collateral appears to be conflicting with the T+2 settlement rules as implemented with CSD-R for on-exchange trades. It therefore needs to be clarified that the same-day basis refers to the liquidation trade but not for the settlement which can be done within the timeframe for settlement of on-exchange transactions as defined by the CSD-R. Moreover the requirement is redundant in view of condition (f) of this same Article and, in particular, on the face of the liquidity requirements introduced by the RTS.

With regards to Article 19 (2) draft RTS, the conditions for transferable securities and money market instruments to be considered as “highly liquid collateral with minimum credit and market risk” are very much the same as those outlined above and, hence, our comments also apply to those conditions. However, we have the following remarks on condition (h):
• Condition (h) (ii): it is unclear the reason why a CSD could not accept as collateral securities issued by an institution that is a CSD-banking service provider or an entity that is part of the same group as a CSD-banking service provider. Would this condition mean that instruments issued by Deutsche Börse could not be accepted as collateral by any CSD in Europe because Deutsche Börse owns a CSD / (I)CSD? Please clarify.
• Condition (h) (iii): Clearstream operates a network of agent banks (securities depositories and cash correspondent banks) which includes some of the largest global banks. Does this condition mean that Clearstream would not be allowed to accept as collateral securities issued by those banks even if there is no correlation between the participant pledging the collateral and the agent banks? This appears an excessive and disproportionate measure on top of already very conservative and restrictive measures.
• From the wording of Article 19 (1) and 19 (2) draft RTS it is unclear how the two rules are placed in relation to each other.
Article 20 draft RTS creates another two categories of collateral: the so-called “highly liquid assets – HLA” (paragraph 1) and “securities settled in the trade” (paragraph 2). As mentioned above, Clearstream opposes the segmentation of collateral as proposed in the RTS due to increased complexity in monitoring and reporting – which shall attract not only additional costs but also new, unnecessary operational risks – as well as the likelihood that this segmentation will have a significant adverse impact in settlement efficiency.

Furthermore, Article 20 (1) draft RTS makes reference to Article 18 (1) (b) (i) draft RTS – which, in turn, makes reference to the requirements of Article 19 draft RTS. In our opinion, the text is extremely confusing. Additionally, the Article establishes that CSDs can accept as collateral “securities settled in the trade” provided they have both of the following:
(i) a prearranged funding arrangement as referred to in point (e) of Article 59 (4) of Regulation (EU) No 909/2014 and Article 36, that provides for the liquidation of these instruments within 5 days;
(ii) sufficient qualifying liquid resources as referred to in Article 34 that allow covering the time gap for liquidating such collateral in case of default of the participant.
We consider this requirement disproportionate and strongly oppose to its maintenance.

To summarise:
• Article 19 (1) draft RTS sets out eight conditions for a security to be considered as “Highly liquid collateral with minimum credit and market risk”.
• Article 19 (2) draft RTS, on the other hand, outlines another eight conditions for transferable securities and money market instruments to be considered as “highly liquid collateral with minimum credit and market risk”.
• Article 20 (1) draft RTS establishes another four conditions (which come on top of the conditions already set in Article 19 draft RTS) for financial instruments to be considered “highly liquid assets”.
• Article 20 (2) draft RTS sets out three conditions (which also come on top of the conditions already set in Article 19 draft RTS) for financial instruments to be considered “securities settled in the trade”.

Clearstream does not see any benefit on the proposed segmentation of collateral and proposes the EBA to strongly re-consider the collateral eligibility requirements along simpler lines.

Accordingly, Clearstream proposes to amend Article 18 to avoid the requirement for different hierarchies of collateral and to merge paragraphs 1 and 2 of Article 19, as from the attached document.
In addition to the general comments made above and the answers to the specific question raised, we do want to comment on quite a few items in more detail and raise some questions on dedicated topics regarding sub-section 4 on Articles 20 to 23 of the EBA draft RTS proposal as follows:

Article 20
On Article 20 draft RTS, the reference made by Article 20 (2) to point (ii) of Article 18 (d) is unfortunately wrong as such clause does not exist. We it was EBA intention to perhaps refer to Article 18 (1) (c) (ii) . We also do not understand the phrase “securities settled in the trade” in the introduction sentence of Article 20(2) draft RTS as this is not referred to in the conditions of that provision, for which further clarification should be provided.

Article 21
On the topic of collateral valuation, Article 21 (2) draft RTS requires eligible collateral to be valued “on a real-time basis or on a near to real-time basis”. Clearstream is of the opinion that, taking into consideration the extremely conservative requirements as outlined in Articles 18, 19 and 20 (for eligibility), Article 22 (haircuts) and Article 23 (concentration limits), the requirement for valuation “on a real-time basis or on a near to real-time basis” is unnecessary. In addition, and assuming that “real-time” market prices are available, it is somewhat inconsistent that the RTS requires “real-time” mark-to-market for the best quality securities and, on the other hand, accepts that less quality securities are marked to model. At Clearstream, securities are revalued three times a day through a process which does not interfere with the (I)CSD’s core settlement system. The requirement for “real-time or on near to real-time” valuation will require significant IT developments and, in our opinion, will bring limited practical benefit to the safety of our collateral management practices.

Article 22
With regards to the requirements of Article 22 (4) of the draft RTS we disagree to the requirement to take the country of issuance of the assets into account as this may or may not have an impact and as such in general is not per se a differentiating factor. For example issuance in country A may be combined with trading on an exchange in country B etc.
We therefore urge the EBA to take out this as a mandatory criterion.
Article 22 (5) makes reference to Article 22 (3) (b). We believe this is a mistake and the reference should have been possibly intended to paragraph 4. Furthermore, paragraph 9 requires daily review of haircuts – this appears inconsistent with the need to avoid pro-cyclicality under paragraph 7 and particularly unnecessary, given the strict requirements already imposed by paragraph 4.

Article 23
Also for Article 23 (3) of the draft RTS we disagree with levels proposed for the concentration limits. Due to the limitations proposed to classify for high quality liquid assets, some concentration will be the consequence anyway which cannot be penalised as this would be contradictory. As such, type of issuer and type of asset do not make sense as there will be a high concentration on government debt instruments in any case. In addition, the country of issuer for the same reason as already stated above seems not to be an eligible criterion. Furthermore, the settlement currency is totally inadequate as the draft RTS at various places requires that the currencies must match. With e.g. a high degree on EUR settlement and loans, the concentration in EUR currency for the collateral is a logic consequence. We therefore strongly demand to reduce the number of categories. Furthermore, we disagree to the conditions (h) and (i) as this does not make sense in the light of the arguments raised above.
Article 23 (5) draft RTS creates a requirement which is not compatible with the current concept of “collateral pool” as mentioned earlier in the context of this response."
According to Article 59 (3) (c) of Regulation (EU) No. 909/2014 (“CSD-R”), a CSD authorised to provide banking type ancillary services “… shall fully cover corresponding credit exposures to individual borrowing participants using collateral and other equivalent financial resources”. Article 24 of the EBA/CP/2015/02 (the “CP”) establishes the conditions by which commercial bank guarantees provided by a financial institution referred to in Article 39 (2) of the CP may be considered as other equivalent financial resources".

The following aims at providing a more detailed analysis of the practical impediments for the implementation of the nine conditions as proposed by Article 24 (2):
• Condition (a) establishes a dynamic cap on bank guarantees (the amount of the guarantee cannot exceed 1% of the settlement values of the securities settlement system over one year period). In Clearstream’s view, the maximum amount of a bank guarantee accepted as collateral should not be subject to specific regulation but rather left to the discretion of the management of the CSD offering banking services. However, should the EBA believe it is required to establish a cap on the amount of bank guarantees, such cap should be established as a proportion of the shareholders’ equity of the CSD (e.g. 25% of own funds). In addition, from a practical point of view, it is not clear whether the “one year period” is counted from a calendar year perspective or on a rolling basis. The latter will dramatically increase the complexity for the monitoring of the condition.
• Conditions (b), (c), (d) and (g) appear acceptable .
• Condition (e) requires the bank guarantee to be “honoured, on demand, within the period of liquidation of the portfolio of the defaulting borrowing participant providing it free of any regulatory, legal or operational constraint”. Reference made to “the period of liquidation of the portfolio of the defaulting borrowing participant providing it” appears out of context. Which portfolio of the “defaulting borrowing participant” is to be liquidated? If a participant provides a bank guarantee to secure its obligations, it is assumed it has no securities pledged to the CSD. Additionally, the requirement that the bank guarantee is provided “free of any regulatory, legal or operational constraint” is already implicitly included under condition (d). In Clearstream’s opinion, the text of condition (e) is unclear. We suggest it is shortened to read “it can be honoured on demand”.
• Condition (f) (i) requires that the bank guarantee is not issued by “an entity that is part of the same group as the borrowing participant covered by the guarantee”. In Clearstream’s view, this restriction is unnecessary as long as the CSD considers the credit exposure on the issuer of the guarantee in the aggregation of the credit exposure it has on the consolidated group. As a matter of fact, most bank guarantees accepted by Clearstream are issued by sound banks on behalf of a bank subsidiary which has not sufficient proprietary assets deposited at Clearstream to secure its credit limits. Considering the very small financial impact such guarantees pose to the solvency and liquidity positions of the CSDs, we believe this restriction should be lifted.
• Condition (h) requires the bank guarantee to be backed by collateral that meets three additional conditions. At Clearstream, we believe condition (h) is virtually impossible to be satisfied. The requirement of securing the issuer’s credit risk with additional collateral imposes a third line of defence vis-à-vis the risk of the borrowing participant: the bank guarantee covers the risk of the borrowing participant's default and the collateral covers the credit risk of the issuer. As a result of this requirement, the CSD would benefit from two credit enhancements to cover the same risk (a cash loan to the borrowing participant). The requirement will make the issuance of a bank guarantee extremely expensive as the issuer will have to price in not only the applicant’s risk but also the cost of posting additional collateral to the beneficiary.
• Condition (i) is redundant given the requirement (as outlined in condition (e)) that the bank guarantee is issued on demand.

Based on the above considerations, Clearstream proposes that:
(1) The text of Article 24 draft RTS explicitly exempts from its scope stand-by letters of credit issued by a syndicate of banks on behalf of an operator of a securities settlement system in favour of another operator of a securities settlement system.
(2) The EBA aligns Article 24 (2) with Article 24 (3) draft RTS, which is simpler to understand and, from an operational point of view, to implement.

Therefore, we propose Article 24 (2) should read as follows in its amended form as from the attached document.

In addition to the general comments made above and the answers to the specific question raised, we do want to comment on quite a few items in more detail and raise some questions on dedicated topics regarding Articles 24 to 29 of the EBA draft RTS proposal as follows:
• It does not make sense to add overnight credit exposures of the previous day to intraday exposures of the subsequent day as requested by Article 26 (4) (b) of the draft RTS. Any overnight exposure will form part of the intraday exposure the next day but should not be added as this would include that exposure twice. Moreover, any incoming fund is offsetting any cumulated open exposure and in principle, the incoming funds cannot be allocated to the any dedicated open exposure but to only to the total net open exposure. As such, we kindly ask to change the wording to “that the amount of overnight credit not yet reimbursed is included in the intraday exposures of the next day”.
• We disagree to the requirement in Article 27 (c) draft RTS to have on a mandatory basis a guarantee from the paying agent of issuer agent in case custody payments are advanced. The obligation for the issuer (via its agents) to pay custody proceeds to the holders is established by the Terms and Conditions of the issue. There is no other contractual mechanism such as a guarantee required for intermediaries in the custody chain to make decision to advance custody payments or not. When it comes to the CSD layer, It is to the discretion of the CSD and in the interest of the receiving participant(s) and of the overall market settlement efficiency that such payments may be advanced subject to proper rules, including but not limited to sufficient counterparty credit worthiness and collateralisation. As such, it is the CSD deciding to advance and this decision cannot be made dependent on the intermediary (cash provider) to deliver a guarantee. In general any advance for custody payments should be linked to the legal enforceable right to cancel the payment and to recall the funds credited instead.
• We disagree to reporting frequencies of more frequent than daily as this is only giving a snap shot which is outdated the moment it is sent. As such it is not adding value at all. Therefore we consider the proposal in Article 28 (4) draft RTS as being over demanding and ask to change from “at least daily” to “at least weekly, but not more frequent than daily”. The same is true related to Article 40 (4) of the draft RTS.
• It needs to be clarified, why Article 29 of the draft RTS is necessary on top of part 8 of CRR and how the two are linked. We rather recommend to ask for disclosure within the CRR disclosure report and to add the additional elements which are not already part of the CRR requirements. The same aspect is to be clarified with regards to Article 41 of the draft RTS."
Besides the liquidity risk concerns resulting from the above question 7, our response to this question also includes other concerns regarding Articles 30 to 42 of the draft RTS.

The Liquidity Coverage Ratio (LCR) under CRR has been designed to ensure that financial institutions have the necessary assets on hand to withstand short-term liquidity disruptions. Effectively, financial institutions are required to hold an amount of highly-liquid assets equal to or greater than its net cash flows over a 30 day period (having at least 100% coverage).
In addition to the LCR regulatory requirements, one of the core reasons why financial institutions would build a liquidity buffer is to be able to withstand intraday liquidity disruptions, which may only last a few minutes or at least in most of the cases not extend beyond intraday.
In case of intraday liquidity disruptions, a financial institution would naturally utilize the entire liquidity buffer at its disposal to fix its intraday liquidity issue.
The artificial distinction of liquidity pools for LCR (30-day horizon) and for the RTS (intraday/overnight) is neither useful from a pure liquidity risk management view nor required under CRR or CSD-R.
Liquidity risk of a CSD is only derived from its banking activities. As such, two different and cumulative liquidity pools for the same liquidity risk do not make any sense at all.
Therefore only an integrated liquidity management to fulfil all internal and regulatory requirements with the same pool of liquid assets is in our view an adequate approach.

Article 31
Further on the Section 1 on Measurement of intraday liquidity risks and more specifically on the Article 31 draft RTS on the Measurement of intraday liquidity risks , we would like to raise the following concerns:

An important feature of BCBS 248 is a materiality threshold, which we think is critically important. BCBS 248 suggests considering a currency as “significant”, if the aggregate liabilities in this currency amount to 5% or more of the bank’s total liabilities. While we expect our key cash correspondents in the major economies to be able to provide real time transaction data, providers in less developed markets (without corresponding regulatory requirements) may not be able or willing to provide the required data, although only counterparties of a high quality (credit and operational) are selected. Forcing real-time data delivery on (cash) transactions in less developed currencies may cause service providers to no longer offer services in that currency. Consequently, certain markets may no longer be accessible. Overall, this would possibly put European CSDs in a situation where certain markets could no longer be served although they are overall not material, function well and have not added to any significant risk in the past. This may nevertheless push major clients out to operate with other counterparties which can offer such markets. A consideration of all (currently at Clearstream around 50) currencies in the “real-time” monitoring may place a significant burden on a CSDs IT capacity. As a (I)CSD usually covers currencies from across the world, the time zone difference will make it extremely difficult to ensure a “real-time” monitoring even for remote “exotic” currencies.
Article 31 (1) (a): Liquidity flows are mainly determined by customers’ settlement activity and cash management, which cannot be predicted prior to the customer cash deadline. We therefore propose to measure daily gross liquidity flows rather than expected flows.
Article 31 (1) (c): We do not understand what the “range of potential net funding shortfalls” could be. As all payments – with the exception of time-critical payments – are due at the end of the day, there can only be a potential shortfall at the end of the day.
Article 31 (2):The required metrics can only be calculated ex post. It should be clarified, what “on an ongoing basis” means.
Based on the arguments raised above we propose the following amendments to the EBA draft RTS as from the attached document.


Further on the Section 2 on the Monitoring Liquidity Risks (page 53) , and more specifically on the Article 33 on the Monitoring intraday liquidity risks, we would like to raise the following concerns.

Article 33
Article 33 (2): The data requested under Article 33 (2) (b) – (d) go beyond the scope of BCBS 248. Whereas the concepts of “objectives” and “risk appetite” may make sense in connection with credit risk, the sole objective for liquidity risk management is to fulfil all payment obligations.
Article 33 (3): The requirement goes beyond the scope of BCBS 248. This comparison can only be delivered ex post, rather than “on a near to real-time basis”. In addition, there should be a materiality threshold, as we expect that correspondent banks in less developed currencies will not provide near to real-time information.

Based on the arguments raised above we propose the following amendments to the EBA draft RTS as from the attached document.

Further on the Section 3 on the Managing Liquidity Risks (page 54 onwards), we would like to raise the following concerns on the following specific articles:
Article 35
The requirements of Article 35 draft RTS go beyond the scope of BCBS 248. The scope of the draft requirements overall seems excessive, given that the focus is on intraday liquidity risk.
Article 35 (2) (b) The haircuts referenced in Article 22 do not properly reflect the low risk inherent in the (highest quality) collateral a CSD has at hand. Concentration limits do not make sense for liquidity risk management purposes as long as assets are central bank eligible and there is no cap on usage of the central bank credit facilities.
Article 35 (2) (c) The requirement goes beyond the scope of BCBS 248.
Article 35 (4) The requirement to hold assets covering the default of the two largest participants (“Cover 2”) seems to be adopted from EMIR. While this makes sense for a CCP, who guarantees the fulfilment of all obligations of the defaulting participant, there is no such obligation for a CSD. In case of a customer default, the CSD would simply not execute any further settlement transactions for the defaulted customer. We therefore do not think that a “Cover 2” requirement should be applicable for a CSD.
Art. 35 (5) and (10). As we understand that Article 35 is supposed to govern liquidity risk, we fail to see the connection to financial risk. Due to the very conservative investment policies applied by CSDs, and given the very short placement tenors, a move in market prices, market volatility and price correlation are not expected to pose a material risk on the CSDs liquidity position. The only events that we expect to put stress on the CSDs liquidity position are a deterioration of the CSDs credit standing (“idiosyncratic” event) or an unavailability of money market credit lines (“market disruption” event). A stress test comparing the peak liquidity need with the available liquidity (stressed for a partial unavailability of liquidity providers) could add value.
Art. 35.6. A reasonable test could be a comparison at the end of the day of the peak liquidity need with the available liquidity sources, which could be stressed for a decline in collateral value or unavailability of liquidity providers.
Art. 35 (13) (a) Liquid assets are used in the regular course of business of the CSD. They should not be considered a “contingency” funding source.
Art. 13 (b) We consider the outlined possible conflict in Article 35 (13) (b) draft RTS to remove a hedge as a clear indication that EBA considers hedging interest rate risk on investments as being a legitimate approach of investing excess liquidity. As such, we strongly urge the EBA to align with ESMA to allow for such strategies subject to fulfilment of all liquidity needs under the currently discussed draft RTS. Our current understanding of the ESMA proposal would be that such hedges would not be allowed for CSDs.
Based on the arguments raised above we propose the following amendments to the EBA draft RTS as from the attached document.

Article 36
(Uncommitted) credit lines granted by cash correspondent banks and depositories are a key liquidity source for CSDs. In contrast to lines granted by money market counterparties, these lines have proved highly reliable, even in stressed conditions. Not taking those into account distorts the liquidity position of the CSD.

Based on the arguments raised above we propose the following amendments to the EBA draft RTS as from the attached document.

Article 37
In general, the scope and consequences of the proposed stress testing seems excessive and likely to place an unmanageable burden on CSDs.
Article 37(2) All liquidity providers are banks whose core business is extending credit and who our subject to prudential supervision. A requirement to obtain confidence that the bank is able to manage its liquidity risk should rather be in scope of the prudential supervision than allocated to individual borrowers. We rather propose to get confidence in the availability of the granted credit lines by establishing a regular testing schedule of these lines.

Article 37(3) The likelihood of a breach in stress tests depends on the calibration of the stress tests. Countermeasures should rather be addressing the insufficiency of actual liquidity needs (or early warning indicators) rather than stress test results.

Article 37(4) There should be a distinction between a liquidity shortfall against actual liquidity needs and a shortfall against stress scenarios. Depending on the severity and applied confidence level of stress tests, the likelihood and practical implications of breaches may vary significantly.
Article 37(6) Several of the proposed stress tests seem either unmanageable or not relevant, e.g. price movements are not expected to have a material impact on the liquidity position. In order to have a manageable framework of stress tests, they should focus on the most significant risk factors, which are in our opinion a deterioration of the CSDs credit standing (“idiosyncratic” event) or an unavailability of money market credit lines (“market disruption” event).
Based on the arguments raised above we propose the following amendments to the EBA draft RTS as from the attached document.

Article 38
Article 38(1) Unlike for CCPs, there is no obligation for a CSD to support settlement of a defaulted customer.
Article 38(3) and (6) It is not possible to anticipate customers’ settlement. “Potentially uncovered liquidity shortfalls” are only expected to arise very short term intraday. It seems impractical to set up an ad hoc reporting on these frictional issues to the risk committee and the competent authority.
Based on the arguments raised above we propose the following amendments to the EBA draft RTS as from the attached document.

Article 39
Article 39(6) The key sources of CSD´s liquidity are customer cash and (uncommitted) money market credit lines. For both, no “arrangements” exist. Committed credit lines, for which there would be an “arrangement” are a supplementary source of liquidity, which is not required to manage liquidity in every market the CSD operates in. It will be difficult to define concentration limits on customer cash.
Based on the arguments raised above we propose the following amendments to the EBA draft RTS as from the attached document.

In addition to the general comments made above and the answers to the specific question raised, we do want to comment in more detail and raise some questions on dedicated topics regarding Articles 40 to 41 and the annexes of the draft RTS proposal as follows:

• We disagree to the proposed business risk scenarios s proposed in Annex II of the draft RTS to a substantial degree:
a. CSDs are virtually completely funded by equity. Funding cost for equity cannot really change in the P&L but only in budget assuming a certain dividend expectation.
b. Reduction of income only matters in case current (expected) or past year income is taken into account to cover business risk.
c. Same argument is true related to higher contribution for pension plans.
d. Long client balances can only reflect to cash deposits. These can only occur, in case the CSD offers banking type of services. Even here, the impact beyond a reduction of current (net) interest income is hardly to measure and to be covered by current income.
e. Same is true for short client balances.
• It needs to be clarified, why Article 41 and 29 of the draft RTS is necessary on top of part 8 of CRR and how the two are linked. We rather recommend to ask for disclosure within the CRR disclosure report and to add the additional elements which are not already part of the CRR requirements.
Article 42
While Article 42 draft RTS is placed as section 6 of Title III, Chapter II, we believe such a rule would have its more appropriate place just after Titles I to III, and are therefore suggesting it should be placed as Title IV.
Moreover, Article 42 draft RTS does not contain any rule for its phasing in, and looking into the massive intervention into IT systems the current proposal (but also an adjusted approach) will bring to the CSDs. We strongly believe there is room for an inclusion of transitional provisions and phasing-in arrangements.

This consists of the following elements:
1. Phasing in rules for CSDs that currently do not fall under the capital rules of CRR or national rules adopting the CRR rules to them with regards to the capital requirements.
2. Phasing in rules for CSDs offering banking type of ancillary services with regards to the capital surcharge for the provision of intraday credits to their participants
3. Phasing in rules for CSDs offering banking type of ancillary services with regards to dedicate rule for credit risk.
4. Phasing in rules for CSDs offering banking type of ancillary services with regards to dedicated rules for liquidity risk.

As such, we propose to amend Article 42 along these lines and make the following proposal as a starting point as from the attached document.
Mathias Papenfuss
D