Before turning to the answer to the 8 questions of the consultation, Amundi would like to express the following general comments:
- For the sake of readability, it should be considered to keep incrementing the number of the articles that follow each other instead of letting art 1 CCP in between art 4 GEN and art 5 GEN and having a great confusion with numbers not consistent with sections and subsections. Furthermore, references should be double checked as some appear to be erroneous (see for example in art FP (5) a reference to art 5 GEN (3) which should point to art 6 GEN and in (6) a reference to article 3(a) when it should be to paragraph 3(a)).
- Asset managers are directly impacted by the proposed RTS, even if a vast majority of individual funds will be below the threshold of 8 billion and exempted from the exchange of Initial Margin; Variation Margin will apply to all of them in March 2017 and collateral requirements as expressed in the draft RTS seems unfairly calibrated for them and contradicts ESMA’s guidelines for UCITS.
- Practical questions are not addressed in the consultation and further clarification would be helpful; first example: the daily consequences of the split of MTA between IM and VM knowing that the time to collect or post is different for both; second example : coordination in the application timeline with the expected dates of implementation of compulsory clearing through CCPs (as it would be weird to impose collateral on a contract that will be subject to mandatory central clearing just a couple of months after).
- The use of a model developed by a counterparty and validated by authorities or third parties should be more clearly authorized without requiring from the other counterparty a replication of the calculations but a limited review of the methodology and the principles of the model; we consider that the usage of a common model by both sides is important to ensure consistency and ease reconciliations; we strongly support the initiative of ISDA to develop a model that could be shared by all actors on the market ; we consider that it would largely facilitate all the operational process and enhance the efficiency and safety of transactions and we encourage authorities to favour its emergence.
- Exemption for buyer of pure cash option from both IM and VM is clearly stated in rationale 6 but could be more explicit in the text of the RTS (section 2).
- Eligibility of UCITS as collateral is examined under article 5 LEC and it is of real significance for asset managers to develop clear operational procedures to facilitate the usage of UCITS as collateral.
Amundi is very sensitive to the level playing field that should be maintained internationally on OTC derivatives. As an asset manager, Amundi knows that there is a worldwide competition and that distortion may occur from different regulatory regimes in different countries. However we believe that harmonized and consistent regulation should proceed from effectively enforceable requirements and cannot rely on suggestions to extend unilaterally domestic provisions to third country entities. We doubt that article 3 GEN will be applicable and fear it will lead to more complexity and uncertainty as we consider that a EU firm will not be in a position to impose its requirement to exchange collateral to an entity which shall not be subject to a similar requirement in its own state. If the result is to exclude this entity from the list of counterparties it might create a great disadvantage for the EU firm and, in the case of funds, for the end client-investor.
The draft differentiates between VM and IM. For VM, article 1VM provides for a delay of 3 days to collect margins when the delay is reduced to 1 day for IM and for VM in case of absence of IM. The time schedule is not practical: we reckon that it is possible to calculate on D+1 (i.e. one day) the valuation of the positions and the amount of margin to be collected or posted; that being done, it is important to reconcile this amount with the counterparty and to accept it; afterwards both firms have to agree on eligibility, haircuts and diversification in the case of securities collateral before giving instructions to post or collect collateral; all of that can lead to an instruction at the end of the day and, allowing for time zone differences, to a settlement in the following couple of days. D+1 is too short in practice, except if it is considered as the delay to calculate, reconcile and instruct, knowing that a technical delay will apply to effectively deliver or transfer cash or securities. In such a case, it will nevertheless be very tight and demanding when divergences appear and models and data have to be discussed between experts.
Amundi is highly concerned with the implementation of IM exchanges in relationship with the model approach. We consider that it is important in terms of risk that both counterparties, except for the cases of regulatory exemptions (see recital 6 which is not explicitly expressed in the RTS), exchange amounts based on the same model. The reasons for asymmetry must be very limited and specifically monitored. Amundi does not intend to challenge and replicate all models of its numerous counterparties, but is keen to maintain its present approach with VM, i.e. to discuss and analyse at the start of a transaction the valuation model and the IM calculation. This implies a fair knowledge of the calculation process and of the source of data used, though without demanding access to all proprietary elements of the model. We would, hence, recommend the usage of a common model accepted by most counterparties. In that respect we expect that ISDA work on SIMM will be a success and simplify collateralization in the future. We would like to be established that a model developed and monitored under the oversight of a professional body such as ISDA is a valid model as should also be models agreed upon by NCAs in charge of banks supervision. We expect to be authorized to rely on the internal models of the counterparties knowing that results are monitored and challenged when they diverge from our expectations.
We consider that the common practice on the market is to assess risk according to sensitivities to the applicable risk factors. We do not see any advantage at creating a specific approach for OTC derivatives that would rely on asset classes. We further agree that revisions and modifications of their internal models should be announced and explained by counterparties ahead of time with a possibility to discuss them and get a fair account of the rationale and consequences of these changes in methodology or principles.
Funds are independent entities except for those which are guaranteed by another fund or the fund manager itself. The asset manager has a global view of the portfolio of the fund and is paid to take risks on behalf of the clients investors according to the investment strategy described in the prospectus. When the manager concludes an OTC derivative, he or she knows that there are 4 levels of risk :
- The risk of the underlying and its volatility which should eventually make the profit or loss of the position
- The counterparty (CP) risk that may jeopardize the expected profit if the counterparty cannot pay
- This CP risk is mitigated through collateral, be it variation margin calls or IM; the third level of risk relates to the accessibility to the collateral and its value and easiness to sell on the market (high quality and liquidity tests)
- This risk can be mitigated through appropriate haircuts and diversification.
The impact of the diversification is very remote from the initial risk and it should be given its appropriate level of significance, which is not high. Amundi totally agrees with ESAs when they suggest to care for diversification of collateral for systemically important institutions that will collect in excess of 1 billion € of collateral. Amundi strongly recommends that a filter (or materiality threshold) apply not only for the diversification of govies with a maximum of 50% per country but also for the other two diversification rules mentioned in art 7LEC 1 (a) and (b) limiting to 10% of the collateral a list of securities that are issued by the same group and to 40% of the collateral the total of equities and convertibles. We consider that the level of 1 billion € of collateral is a reasonable one for this filter (even if a figure in hundreds of millions would be acceptable). In other words we suggest to place (2) in art 7 LEC at the beginning of the article in position (1) and transfer (1) in the second place as (2).
With regard to funds, it is not appropriate to refer any limit to a percentage of the collateral. The relevant amount is the asset under management which is the basis for risk taking and is far less volatile than the collateral. From a practical point of view it is a nightmare to require eleven different issuers for a collateral that will amount to a couple of millions or even less. This is about the size of collateral in the vast majority of funds. ESMA published in December 2012 guidelines on ETF and other UCITS issues where it asked for diversification of collateral and introduced the figure of 20% per issuer with reference to the total asset of the UICTS. The lack of threshold made this regulation burdensome and disputable in terms of financial stability but nevertheless it proved workable thanks to the reference to AUM to calculate the percentage.
Equity funds have only equities in their portfolios. The suggestion to limit the collateral that can be posted or collected to 40% in equities will make it very difficult for them to sign OTC derivative contracts. It is an unexpected and hopefully unintended consequence of the proposed RTS and we strongly oppose this view.
Amundi insists on the necessity:
1. to apply a materiality threshold before applying diversification rules on the collateral and the figure of 1 billion € seems properly calibrated in that respect;
2. to calculate the limit not as a percentage of the collateral but of the assets under management in the case of a fund or a mandate;
3. to suppress the limitation of 40% on the total of equities as it is a limitation of eligibility which is contrary to level one text more than a measure of diversification.
As a principle documentation must be completed before signing any transaction at Amundi. However, we are concerned by the terms of art 2OPD (2) that requires an independent annual legal review of the enforceability of bilateral agreements. On one hand we want to clarify that the legal department of Amundi includes in-house lawyers that are independent from all investment activities : they should be considered as able to provide an independent review. On the other hand, we think that professional associations should be encouraged to help their members in that process and make country per country analyses of the adequacy of different agreements. More specifically and as an illustration, opinions made available by ISDA or FBF in relationship with their standard terms and contracts give, in our view, sufficient comfort to consider the legal review as accomplished.
The segregation of IM is of paramount importance in Amundi’s view in order to best ensure investors protection. However the segregation of cash which is fungible by nature is a tricky concept. If we see the possibility to establish individual accounts for different purposes for a same fund, we do not believe that this money will be protected in case of insolvency of the depository. We need the RTS to be workable and to clarify that individual segregation for cash does not mean bankruptcy remote and that is a further motive to impose strict limitations on the entities that can act as depository for funds.
The same comments on independent legal review apply as in question 5 above. We expect ISDA will provide relevant legal opinions on most jurisdictions and up-date them. However we fear that the globalization of the market create new conflicts of law when territoriality is difficult to establish: head office of CP 1, of CP 2, or of its management company in case of a fund, currency of the deal, place of trade, place of settlement, law of the documentation…legal counsel will certainly have to sort it out.
We agree that when cash is posted as initial margin, the collector should maintain it with a view to protect the collateral poster to whom it should be returned at the end of the transaction. We consider that the collected cash should be re-invested in low risk and liquid instruments. We think that the segregation out of its prop trade is essential, that the list of eligible instruments could be limited when compared to art LEC1 and that diversification rules should apply above a materiality threshold. This can easily be done on a contractual basis as the reinvestment of cash is subject to an agreement between the counterparties.
For the re-use of securities collected as collateral, the recent discussion about the SFT regulation has shown that it is difficult not to presume the explicit agreement of the counterparty to re-use if a transfer of collateral was executed through a total transfer of property. Thus, we accept the protective ban of any re-use as a good initiative but we are worried that it could create a problem of international competition and even playing field if this restriction were specific to the EU.
The two concepts of termination currency (for the final settlement) and transfer currency (for regular payments) are not totally clear, their definition being left to the counterparties to decide. It makes it difficult to assess the consequences of the proposed RTS in that subject. Anyway, the 8% additional haircut is in itself very penalizing. The more so because it simply applies to securities and not cash. We expect internal models will show a much lower impact for currency mismatch. As a matter of fact, currency movements are very sharp when specific events occur and in that case 8% might just not be sufficient to offer protection and they are very slow most of the time on a day to day basis. Calibration at 8% is far too high in our view, by 4 to 8 times. Furthermore, such a high percentage would incentivize counterparties to post securities denominated in the basis currency and might reduce their ability to provide highly liquid securities of very high quality.