Response to joint Consultation on draft RTS on risk-mitigation techniques for OTC-derivative contracts not cleared by a CCP
Go back
Documentation
This letter has already suggested that NFC- equivalent non-EU counterparties ought not to be subject to VM requirements. However, if this requirement is retained in the RTS, HSBC and its counterparties would be required to renegotiate thousands of collateral agreements in advance of trading after the start of December 2015. Further, as many such clients would not have the liquidity to meet future margin calls, they may require special financing arrangements in order to fund the calls. They will also need to establish processes to receive the funds and place them back as margin.
In respect of financial counterparties (“FC”), which are covered by the VM requirements in most instances, the revised documentation process could be slightly less onerous if covered by the protocol approach suggested by ISDA. However, in HSBC’s experience with many counterparties in relation to past protocols, it can take a number of months for FCs to perform relevant legal diligence and to gain management approval to adhere to such protocols. In the meantime, trading would have to cease. HSBC also notes that, protocols are not always applicable to certain master agreements (such as the German Ramenvertrag) or bespoke collateral arrangements and, in those situations, separate negotiation is likely to be required with those counterparties.
Even where protocols are capable of being applied, if the counterparty does not adhere, there would be a need to negotiate a new agreement, which would be a manually intensive and time consuming process involving positive action and cooperation from the counterparty. The counterparty would also have to establish internal processes both for giving or pledging assets, and recognising assets pledged to it, and releasing assets. These are made more onerous when assets are given as collateral in systems in which the recipient has no pre-existing accounts.
Threshold Application
HSBC welcomes the concept of a Threshold Amount for IM to reduce systematic liquidity strain. We believe the Threshold Amount determination should apply at the level of the legal agreement which would usually be at the contracting party level only. Though the application of Threshold for IM is optional, it will have an effect on pricing and on the operational processes that are likely to be necessary in the early stages of application.
As the option to apply a Threshold Amount has financial value (or expense) if used by different subsidiaries, it would actually be required to be applied for pricing equivalence, but it is not obvious how the ESAs would like this to work in situations where a fiduciary duty is owed by the holding company to a number of different entities within a group. Each could reasonably demand to apply the threshold for transactions with their entity, for example, this would apply to publically listed companies within a financial group.
HSBC does not believe that it is reasonable to limit the threshold to the consolidated group level. Such an approach penalises particularly groups which have subsidiary banks which are separately capitalised for risk mitigation purposes and function on an arms-length basis from other entities in the group, over those which have a more operational branch structure. As both counterparties would have to agree the Thresholds, where this spanned different entities in the different groups this would be another area requiring negotiation.
Legal Opinions
The ISDA response highlights some of the practical constraints on the number, nature and timing of legal opinions being required. HSBC shares these concerns.
Notional Limit Determination
Without public attestation or separate and continuing bilateral representation, firms cannot have certainty as to the total notional outstanding in derivatives from any other company or group of companies (where relevant). The ESAs might consider, as an alternative means of achieving the policy objective, holding a list for all derivatives counterparties which exceed each of the successive thresholds or allow a longer lead time from the announcement of IM eligible counterparty groups and the application of the regulations.
It would be helpful if the final RTS confirmed the comments made by ESAs at the public hearing, to the effect that, where a counterparty exceeds a threshold, only new transactions from that date need necessarily to be subject to the IM process. As this is not likely to be practicable immediately after triggering a threshold, perhaps a reasonable time for compliance with this requirement (for example, 4 months), ought to be permitted to ensure that the new counterparty has appropriate documentation, is fully integrated in the operational process and would have had sufficient time to undertake preparatory testing.
Process for IM collection and release
Under some EU country laws (Belgian law, for example) total title transfer is required to perfect a pledge. As a result, though the economic interest is retained by the pledgor, the pledgee can use the pledged assets as they please (unless controlled and agreed under ancillary arrangements). Indeed this is the normal practice under so-called tri-party collateral agreements within Euroclear. A legal question then arises as to whether the imposition of ancillary restrictions on the use of pledged assets prevent the construction that transfer requirements for perfection have been met. The processes for pledging under the proposed IM arrangements are therefore to some extent breaking new ground. The legal position has also not been tried on what might be described as an industrial scale before. This interplay between national laws, new simultaneous bi-party IM pledging documentation and possible operational options HSBC believes is likely to make it difficult to achieve legal certainty over the arrangements in a timely way. It may be prudent for the ESAs to introduce milestones post-RTS publication so the ESAs can assess readiness. The ESAs should ensure that they are apprised of the new operational risks being introduced under these regulations.
The timeframes for agreeing trade populations, derivative valuations, and future exposures is extremely short for the process to be anything other than fully automated. HSBC is concerned that the timeframe for creation and delivery of such systems after the RTS is published and other international regulations are drafted and published is likely to be too short for a vendor solution to be created, tested and implemented. Standalone automated processes within each firm would present more connectivity issues including time to ensure connections exist between counterparties and the risks of breakages in links between counterparties. ESAs can reduce this impact by publishing RTS as soon as possible, perhaps indicating a preference for common market systems in the text. A simple process should be made available whereby on proof that material operational risk would result from compliance, the relevant NCA could agree that capital requirements would be deemed sufficient risk mitigation for a limited period.
It is likely that counterparties will need to pledge or give lien over assets held in a range of depositaries. The processes for interoperation between the depositaries are not well established. HSBC is therefore concerned that the processes for granting IM and releasing IM will not be sufficiently well established to ensure that the operation poses no additional credit, legal or other risk on failure of the counterparty. HSBC understands the ESA approach is to require firms’ procedures to be sufficiently comprehensive to ensure compliance, but this is not likely to be possible where the underlying external operational structures are not established.
Model approval
The use of initial margin models is economically important. Companies which are only able to offer standardised approach computation are unlikely to find trading counterparties at the same market price as those with models. The capacity and competence of the NCAs to respond to, or not reject notifications, of initial margin models will be critical for firms. However, the draft RTS is unclear as to what the competent authority is approving, if anything. If NCA approval or acquiescence is required this may lead to an uneven playing field within the EU. HSBC recommends that ESAs clarify that the notification of models to the National Competent Authority is for information only.
Financial institutions in the EU are already covered by capital requirements for any shortfalls of margin under CRR. With such cover, the need for precision of estimation and requirement for margin itself is relatively slight. The final RTS in isolation should not dictate the total risk or capital or funding, but is just one component in these calculations. The overall requirements are computed using existing more precisely computed, regulatory approved and often modelled capital figures. In HSBC’s view, this should mean that the RTS does not need to be prescriptive about the degree of testing and validation applied to the models.
Model details
It is important that there should be some convergence of model outcomes and therefore presumably of models themselves in order that reasonable levels of agreement on valuation and future risk are reached in a timely way. It is also important for the industry to understand how the ESAs intend choices to be made when the results of models used by a firm and their counterparty disagrees. The ESAs could specify, for example that model approach could be agreed between the parties, that the receiving party model always prevails, or the posting party, or that there could be an agreed tolerance between these and the higher or lower should be pledged. It is not sufficient to determine that if there is failure to agree, the standardised computation is to be applied as this will change the price of the business reinforcing market fragmentation.
The model requirement to use at least 25% of stressed data is confusingly worded. The period for the financial stress is unclear, and bank and counterparty modelled IM requirements are unlikely to agree when using different periods. HSBC believes that setting fixed stressed periods or ESA support for industry wide harmonised period would help to reduce the mismatches.
Historical data is required to be at least three years, but this phrasing can lead to issues, if one party chooses to use 20 years and another 4 years, dramatic differences would be observed. Two counterparties may use the same model but have different results when using different data periods.
A separate risk factor for each equity or commodity that is significant may have unforeseen issues, because, for example, an equity deemed insignificant for one netting set may be significant for another.
Allocation to the appropriate asset class based upon primary risk factor has an element of subjectivity. The classification of hybrids is unclear and may be treated differently by the counterparties.
Although this is aligned to Basel, the initial margin model is based upon siloed risk, and no offset is allowed between asset classes. This differs from the real risk. Particularly on close out, offsetting will exist at least to some extent, so IM will be far higher than necessary. HSBC suggests that ESAs should consider allowing offsetting across risk factors.
For the annual validation of the model by “suitably qualified and independent parties”, if this is intended to refers to third parties, in HSBC’s view this is unnecessarily onerous. ESAs should clarify that this refers to internal processes and not third parties.
Collateral
Any restrictions on permitted collateral should be strictly limited to permit substitution which is in turn a key element of reducing potential adverse market liquidity impact.
The standardised add-on factors are aligned with Basel proposals but in HSBC’s view the factors are too high; and appear to be inconsistent and high compared with the standardised approach for measuring counterparty credit risk (“SA-CCR”) when scaled for appropriate margin period of risk. HSBC suggests that the scaling proposed in SA-CCR (paragraph 164) should be applied with no maturity dependency.
EU Competent Authorities are presently taking quite differing views on what constitutes sufficiency of modelling data. With the Prudential Regulation Authority, for example, suggesting that for sovereigns, public sector entities and financial sector entities IRBA (Internal Ratings Based Advanced) models are likely to be rejected. This would be likely to lead to quite different haircuts between counterparties.
The collateral has to be credit quality assessed each day, but it is unclear which internal model would be used in this assessment. The ESAs could clarify the precedence for such validation.
Should a margin giver’s model determine that it is over-collateralising, this could result in an additional capital requirement inferring that systemic risk is actually being created. HSBC suggests that the ESAs should introduce a prudent systematic requirement that capital regulation haircut collateral should not be required to be posted in excess of the Exposure at Default (“EaD”). The risk of the collateral posted changing in value is already capitalised on the posting firms’ balance sheet. If applied symmetrically, the system collateralisation would remain sufficient. For collateral received, the collateral receipt would be haircut in any case in the computation of collateral offset against EaD, so the system remains protected and the issue of excess posting or asymmetry need not arise.
HSBC believes that the approach to collateral haircutting is over-engineered and, depending on the meaning of settlement currency, would create risk (as determined by regulatory measures) in the system. Less prescription from the ESAs would be useful and would not materially impact stability.
HSBC is concerned at the potential for market fragmentation which could arise where counterparties reasonably choose each other based on the relative cost efficiency, and their costs are lowest where IM rehypothecation is permitted, so the market could become fragmented on regional lines. This is problematic, not so much because of the direct end user transaction, but because of the cost of the various market transactions used to hedge the portfolio in which the end user risk is managed.
HSBC believes that the ESAs should discuss their proposed approach with the wider Basel committee to better understand the cross border impact of failing to allow for rehypothecation.
Question 2. Are there particular aspects, for instance of an operational nature, that are not addressed in an appropriate manner? If yes, please provide the rationale for the concerns and potential solutions.
Yes, there are such operational concerns in addition to concerns around implementation and timing. It is not only the appropriateness of approach though - entire areas of operation and operational interaction appear not to have been addressed.Documentation
This letter has already suggested that NFC- equivalent non-EU counterparties ought not to be subject to VM requirements. However, if this requirement is retained in the RTS, HSBC and its counterparties would be required to renegotiate thousands of collateral agreements in advance of trading after the start of December 2015. Further, as many such clients would not have the liquidity to meet future margin calls, they may require special financing arrangements in order to fund the calls. They will also need to establish processes to receive the funds and place them back as margin.
In respect of financial counterparties (“FC”), which are covered by the VM requirements in most instances, the revised documentation process could be slightly less onerous if covered by the protocol approach suggested by ISDA. However, in HSBC’s experience with many counterparties in relation to past protocols, it can take a number of months for FCs to perform relevant legal diligence and to gain management approval to adhere to such protocols. In the meantime, trading would have to cease. HSBC also notes that, protocols are not always applicable to certain master agreements (such as the German Ramenvertrag) or bespoke collateral arrangements and, in those situations, separate negotiation is likely to be required with those counterparties.
Even where protocols are capable of being applied, if the counterparty does not adhere, there would be a need to negotiate a new agreement, which would be a manually intensive and time consuming process involving positive action and cooperation from the counterparty. The counterparty would also have to establish internal processes both for giving or pledging assets, and recognising assets pledged to it, and releasing assets. These are made more onerous when assets are given as collateral in systems in which the recipient has no pre-existing accounts.
Threshold Application
HSBC welcomes the concept of a Threshold Amount for IM to reduce systematic liquidity strain. We believe the Threshold Amount determination should apply at the level of the legal agreement which would usually be at the contracting party level only. Though the application of Threshold for IM is optional, it will have an effect on pricing and on the operational processes that are likely to be necessary in the early stages of application.
As the option to apply a Threshold Amount has financial value (or expense) if used by different subsidiaries, it would actually be required to be applied for pricing equivalence, but it is not obvious how the ESAs would like this to work in situations where a fiduciary duty is owed by the holding company to a number of different entities within a group. Each could reasonably demand to apply the threshold for transactions with their entity, for example, this would apply to publically listed companies within a financial group.
HSBC does not believe that it is reasonable to limit the threshold to the consolidated group level. Such an approach penalises particularly groups which have subsidiary banks which are separately capitalised for risk mitigation purposes and function on an arms-length basis from other entities in the group, over those which have a more operational branch structure. As both counterparties would have to agree the Thresholds, where this spanned different entities in the different groups this would be another area requiring negotiation.
Legal Opinions
The ISDA response highlights some of the practical constraints on the number, nature and timing of legal opinions being required. HSBC shares these concerns.
Notional Limit Determination
Without public attestation or separate and continuing bilateral representation, firms cannot have certainty as to the total notional outstanding in derivatives from any other company or group of companies (where relevant). The ESAs might consider, as an alternative means of achieving the policy objective, holding a list for all derivatives counterparties which exceed each of the successive thresholds or allow a longer lead time from the announcement of IM eligible counterparty groups and the application of the regulations.
It would be helpful if the final RTS confirmed the comments made by ESAs at the public hearing, to the effect that, where a counterparty exceeds a threshold, only new transactions from that date need necessarily to be subject to the IM process. As this is not likely to be practicable immediately after triggering a threshold, perhaps a reasonable time for compliance with this requirement (for example, 4 months), ought to be permitted to ensure that the new counterparty has appropriate documentation, is fully integrated in the operational process and would have had sufficient time to undertake preparatory testing.
Process for IM collection and release
Under some EU country laws (Belgian law, for example) total title transfer is required to perfect a pledge. As a result, though the economic interest is retained by the pledgor, the pledgee can use the pledged assets as they please (unless controlled and agreed under ancillary arrangements). Indeed this is the normal practice under so-called tri-party collateral agreements within Euroclear. A legal question then arises as to whether the imposition of ancillary restrictions on the use of pledged assets prevent the construction that transfer requirements for perfection have been met. The processes for pledging under the proposed IM arrangements are therefore to some extent breaking new ground. The legal position has also not been tried on what might be described as an industrial scale before. This interplay between national laws, new simultaneous bi-party IM pledging documentation and possible operational options HSBC believes is likely to make it difficult to achieve legal certainty over the arrangements in a timely way. It may be prudent for the ESAs to introduce milestones post-RTS publication so the ESAs can assess readiness. The ESAs should ensure that they are apprised of the new operational risks being introduced under these regulations.
The timeframes for agreeing trade populations, derivative valuations, and future exposures is extremely short for the process to be anything other than fully automated. HSBC is concerned that the timeframe for creation and delivery of such systems after the RTS is published and other international regulations are drafted and published is likely to be too short for a vendor solution to be created, tested and implemented. Standalone automated processes within each firm would present more connectivity issues including time to ensure connections exist between counterparties and the risks of breakages in links between counterparties. ESAs can reduce this impact by publishing RTS as soon as possible, perhaps indicating a preference for common market systems in the text. A simple process should be made available whereby on proof that material operational risk would result from compliance, the relevant NCA could agree that capital requirements would be deemed sufficient risk mitigation for a limited period.
It is likely that counterparties will need to pledge or give lien over assets held in a range of depositaries. The processes for interoperation between the depositaries are not well established. HSBC is therefore concerned that the processes for granting IM and releasing IM will not be sufficiently well established to ensure that the operation poses no additional credit, legal or other risk on failure of the counterparty. HSBC understands the ESA approach is to require firms’ procedures to be sufficiently comprehensive to ensure compliance, but this is not likely to be possible where the underlying external operational structures are not established.
Model approval
The use of initial margin models is economically important. Companies which are only able to offer standardised approach computation are unlikely to find trading counterparties at the same market price as those with models. The capacity and competence of the NCAs to respond to, or not reject notifications, of initial margin models will be critical for firms. However, the draft RTS is unclear as to what the competent authority is approving, if anything. If NCA approval or acquiescence is required this may lead to an uneven playing field within the EU. HSBC recommends that ESAs clarify that the notification of models to the National Competent Authority is for information only.
Financial institutions in the EU are already covered by capital requirements for any shortfalls of margin under CRR. With such cover, the need for precision of estimation and requirement for margin itself is relatively slight. The final RTS in isolation should not dictate the total risk or capital or funding, but is just one component in these calculations. The overall requirements are computed using existing more precisely computed, regulatory approved and often modelled capital figures. In HSBC’s view, this should mean that the RTS does not need to be prescriptive about the degree of testing and validation applied to the models.
Model details
It is important that there should be some convergence of model outcomes and therefore presumably of models themselves in order that reasonable levels of agreement on valuation and future risk are reached in a timely way. It is also important for the industry to understand how the ESAs intend choices to be made when the results of models used by a firm and their counterparty disagrees. The ESAs could specify, for example that model approach could be agreed between the parties, that the receiving party model always prevails, or the posting party, or that there could be an agreed tolerance between these and the higher or lower should be pledged. It is not sufficient to determine that if there is failure to agree, the standardised computation is to be applied as this will change the price of the business reinforcing market fragmentation.
The model requirement to use at least 25% of stressed data is confusingly worded. The period for the financial stress is unclear, and bank and counterparty modelled IM requirements are unlikely to agree when using different periods. HSBC believes that setting fixed stressed periods or ESA support for industry wide harmonised period would help to reduce the mismatches.
Historical data is required to be at least three years, but this phrasing can lead to issues, if one party chooses to use 20 years and another 4 years, dramatic differences would be observed. Two counterparties may use the same model but have different results when using different data periods.
A separate risk factor for each equity or commodity that is significant may have unforeseen issues, because, for example, an equity deemed insignificant for one netting set may be significant for another.
Allocation to the appropriate asset class based upon primary risk factor has an element of subjectivity. The classification of hybrids is unclear and may be treated differently by the counterparties.
Although this is aligned to Basel, the initial margin model is based upon siloed risk, and no offset is allowed between asset classes. This differs from the real risk. Particularly on close out, offsetting will exist at least to some extent, so IM will be far higher than necessary. HSBC suggests that ESAs should consider allowing offsetting across risk factors.
For the annual validation of the model by “suitably qualified and independent parties”, if this is intended to refers to third parties, in HSBC’s view this is unnecessarily onerous. ESAs should clarify that this refers to internal processes and not third parties.
Collateral
Any restrictions on permitted collateral should be strictly limited to permit substitution which is in turn a key element of reducing potential adverse market liquidity impact.
The standardised add-on factors are aligned with Basel proposals but in HSBC’s view the factors are too high; and appear to be inconsistent and high compared with the standardised approach for measuring counterparty credit risk (“SA-CCR”) when scaled for appropriate margin period of risk. HSBC suggests that the scaling proposed in SA-CCR (paragraph 164) should be applied with no maturity dependency.
Question 4. In respect of the use of a counterparty IRB model, are the counterparties confident that they will be able to access sufficient information to ensure appropriate transparency and to allow them to demonstrate an adequate understanding to their supervisory authority?
HSBC supports the detailed responses from the industry and notes that:EU Competent Authorities are presently taking quite differing views on what constitutes sufficiency of modelling data. With the Prudential Regulation Authority, for example, suggesting that for sovereigns, public sector entities and financial sector entities IRBA (Internal Ratings Based Advanced) models are likely to be rejected. This would be likely to lead to quite different haircuts between counterparties.
The collateral has to be credit quality assessed each day, but it is unclear which internal model would be used in this assessment. The ESAs could clarify the precedence for such validation.
Should a margin giver’s model determine that it is over-collateralising, this could result in an additional capital requirement inferring that systemic risk is actually being created. HSBC suggests that the ESAs should introduce a prudent systematic requirement that capital regulation haircut collateral should not be required to be posted in excess of the Exposure at Default (“EaD”). The risk of the collateral posted changing in value is already capitalised on the posting firms’ balance sheet. If applied symmetrically, the system collateralisation would remain sufficient. For collateral received, the collateral receipt would be haircut in any case in the computation of collateral offset against EaD, so the system remains protected and the issue of excess posting or asymmetry need not arise.
HSBC believes that the approach to collateral haircutting is over-engineered and, depending on the meaning of settlement currency, would create risk (as determined by regulatory measures) in the system. Less prescription from the ESAs would be useful and would not materially impact stability.
Question 6. How will market participants be able to ensure the fulfilment of all the conditions for the reuse of initial margins as required in the BCBS-IOSCO framework? Can the respondents identify which companies in the EU would require reuse or re-hypothecation of collateral as an essential component of their business models?
HSBC notes that the BCBS-IOSCO paper permits single rehypothecation which appears to be aimed at financial intermediaries with limited inventory and also intended to reduce somewhat the liquidity impact on the markets of such margin. Rehypothecation may also somewhat defray the potential expense of IM, but this expense should not be a determining consideration for the ESAs over the reduction of operational risk.HSBC is concerned at the potential for market fragmentation which could arise where counterparties reasonably choose each other based on the relative cost efficiency, and their costs are lowest where IM rehypothecation is permitted, so the market could become fragmented on regional lines. This is problematic, not so much because of the direct end user transaction, but because of the cost of the various market transactions used to hedge the portfolio in which the end user risk is managed.
HSBC believes that the ESAs should discuss their proposed approach with the wider Basel committee to better understand the cross border impact of failing to allow for rehypothecation.
Upload files
HSBC Response - MRUD Final.pdf
(286.34 KB)