Amundi has an experience of the cost resulting from the new regulation under EMIR without being able to express it in monetary terms. We are aware that more specifically collateral management for non-centrally cleared OTC transactions will bring the following:
• Legal and documentation cost, as agreements must be negotiated or amended to take on board new requirements as far as collateral is concerned; legal teams are overbooked with other new developments mainly linked to the implementation of EMIR and the central clearing process but also to the future MIF requirements in terms of distribution and some asset managers may have to solicit external advice which tend to increase the cost of compliance;
• Operational cost exposed in order to establish and then run processes for orderly collateral management; operational costs could be reduced if the suggestions expressed in the answer to the following question were accepted.
• Opportunity cost : for example, the securities held in the portfolio which are to be used as collateral will no longer be available for securities lending activities that improve the return for the fund and its client holder; the reduced facility to sell assets transferred as collateral makes the process to arbitrage positions more difficult and might prevent the manager from seizing opportunities.
Grand fathering clause: we insist on the necessity to fully establish the fact that the regulation will only apply to transactions conducted after its implementation date; it should not only be mentioned under recital 18 but also included in the text of the regulation itself. Furthermore, we consider that a precision should be added to make sure that “genuine amendments” as referred to in the BCBS/IOSCO final report (§ 8.9 and footnote 20, p. 24) made to existing derivative contracts shall not be in the scope of the new regulation. An explicit mention in article 1 FP § 4 would bring the necessary clarification. As a matter of fact, investment funds often amend existing contracts to reduce their notional amount to adjust their exposure, as a consequence of redemptions. Thus, it is very important for them to keep these existing contracts outside the scope of the regulation.
Reverse the presumption that IM will not be collected : in article 1FP §3, the possibility to exempt transactions between counterparties under the threshold requires a prior formal agreement ; we think that the rule should be that in absence of a specific prior agreement among the parties the rule should be that there is no collateral posted; as mentioned legal services are overworked and this suggestion would simply make things workable.
With respect to concentration rules we have the following 3 comments to share:
1. we strongly suggest that there should be a realistic threshold of 100 million € under which diversification rules should not apply; not only because it is not workable to ask for a split collateral over different issuers on smaller amounts than 100 million, but also because in terms of systemic risk a de minimis principle is totally appropriate;
2. All funds (UCITS and AIFs subject to equivalent rules) that comply with the ESMA guidelines published in December 2012 on ETF and other issues relating to UCITS should be exempted from any other type of collateral management rules and specifically any diversification rule; it would be highly counterproductive to suggest to the UCITS investors that existing regulation is inadequate and has to be modified ; furthermore the balance between increased risk mitigation and investor protection (at best minimal ) and resulting cost (at the least very high) does not justify another layer of regulation for UCITS and other funds complying with ESMA guidelines.
3. ESMA guidelines ask for UCITS to spread received collateral in order not to exceed an exposure higher than 20% of the NAV of the fund on one single issuer. The denominator of the ratio is what is important in the fund, i.e. the net asset value or the net capital available in the fund; the proposed regulation refers to the total amount of the collateral when computing the diversification ratio; it is inconsistent to consider that there is a risk that could be mitigated with a diversification rule on a collateral of small amount; it is not workable to ask counterparties to split small amounts of collateral on several issues for even smaller amounts, uneasy to manage and costly to transfer; thus we recommend to compute the diversification ratio as a percentage not of the collateral itself but as a proportion of the NAV for funds, and/or if it is necessary to maintain one single standard as a proportion of the notional amount of the underlying derivative;
Derivative close out: funds are required in Europe to negotiate a possibility to ask for an exit out of a derivative at any time. It implies the possibility to put to an end the initial operation with the initial counterparty. The point is that if, for the sake of best execution, the asset manager has a better price with another counterparty for an opposite transaction it will have to post collateral for two opposite exposures as long as a netting is not accepted between counterparties (and there is no possibility to force such a netting). This is not an issue if no initial margin is exchanged but may become an obstacle to best execution if there is. Funds are heavily regulated and closely monitored and could benefit from an exemption to exchange initial margin; as a consequence they would not suffer any conflict when closing a derivative position.
Minimum Transfer Amount (MTA): the proposed regulation suggests a MTA of 500 000 €. This amount includes the net variation of IM and VM exchanged between 2 counterparties. Our organization, and we understand that it is shared by other entities, monitors IM and VM separately and we would prefer to have two MTAs. We hence suggest that be added it is possible to follow two MTAs instead of one.
As an asset manager, Amundi is not directly concerned with the hedging of a pool refinanced by covered bonds. However as an investor in covered bonds, we strongly believe that the pool behind the bonds is the guarantee of the investor and should not be encumbered at all. Thus we consider that no collateral should be posted by the pool even in relationship with derivative contracts signed to cover its risks.
The question of internal models should not be addressed only with reference to internal rating but more generally. In both instances, transparency and advance notice of changes in the models are a necessity. Properly calibrated models should be stable and should not frequent modifications. Counterparties should hence discuss the models as well as their expected stability.
Internal models : Amundi concludes derivative contracts on behalf of the funds it manages with large counterparties, usually the most prominent investment banks. We understand that should these counterparties ask for IM they would calculate the appropriate level based on an internal model that is approved by the supervisory authority. The same will happen when calculating daily VM or reasonable haircut levels. Amundi has not so far gone through the process of asking the AMF, its supervisory authority, to approve an internal model. We have however a large experience of discussing their models with investment banks when negotiating derivatives or challenging their prices.
We approve the requirement that banks communicate on the models they use and the data they take when running these models. Our responsibility towards our clients includes the verification of the prices of the financial instruments we use and the level of collateral is within the scope of our controls. We would advocate for a specific requirement to discuss the key structure of the models and the source of data used and suggest they should be as often as possible public data. As we fear that it might be difficult to receive the necessary information, we would consider as a good leverage that the client be recognized the possibility to seize the competent authority if he cannot obtain satisfactory explanations.
Internal rating based approach: when assessing credit quality of collateral, fund managers tend to refer to their large experience as direct investors. Managing money for third parties implies not only to comply with supervisory authorities but also to ensure that customers’ money is managed within a defined risk framework. Therefore one must ensure that both the client needs and the supervisory rules are understood. Sufficient information and appropriate transparency are necessary to allow :
- Due diligence made by the Asset Manager on the IRB used by the counterparty. When managing money for third parties, the validation of the IRB by a supervisory body is an important step that should be supplemented by the Asset Manager’s due diligence before selecting eligible counterparties to trade OTC derivatives.
- Regular audit made in order to check that the counterparty is using its IRB in an appropriate manner. The Asset Manager should be able to demonstrate through an audit trail that he has complied with his duty to assess the quality of collateral.
The collateral is used as a guarantee in case the counterparty fails to meet its commitment. The quality of collateral rated by a “defaulting” counterparty may make its sale very difficult (and thus limit considerably the purpose of posting collateral). Typically, if the IRB is used for non-rated securities it can make the sale of the collateral in a stress period more difficult in the absence of an external rating. We feel IRB approach could lead to wrong way risk and should not be encouraged.
For this reason, combined with the fact that small/medium size asset managers will not be able to have approved IRB models nor to monitor their counterparties IRB models, we feel that the advantage given to IRB models as opposed to external ratings is not fair and that the acceptable CQS should be positioned at the same level, level 3 for usual bonds, in both cases.
Before discussing concentration limits it is of prime importance to open the discussion about eligible collateral.
Amundi shares the view that the list of eligible collateral should be large, diversified and include funds. The larger the list, the lesser the risk for procyclicalilty and collateral squeeze. The trouble with a large list of eligible collateral is that it may include assets of lesser quality. Quality of the collateral is assessed on the basis of its liquidity and credit worthiness.
Liquidity: the purpose of collateral is to be able to convert it into cash when the counterparty defaults in order to cover the difference between the price on which the derivative contract was settled and the last level of the variation margin. Liquidity tends to fade or increase very quickly. However, it is a fact that large cap stocks are among the most liquid assets in any circumstances and that government bonds are usually very actively trades as well. But the recent experience has shown that govies might become totally illiquid overnight. Liquidity should be assessed including the possibility to use securities lending and Repo markets to gain it.
Credit quality: if we agree with the fact that investors should not rely overly on credit rating agencies analysis and should develop internal credit assessment capacities, we do not consider that it is fair to introduce a difference of one notch of CQS when determining the eligible collateral depending on whether it is based on an internal rating or an external one provided by a CRA. This causes a particular concern on our part as we will rely on models developed by banks and as such taking into consideration their internal credit estimates. In order to challenge these models we want to be able to take rating published by CRA as reference, especially because the CRAs produce extensive research on credit performance of rated issues over a long period of time.
In general, we consider that the entire field of the investment grade (as defined by CRAs) issues should be considered as eligible. Since they do not show the same level of safety, it is clear that appropriate haircuts have to be provided. We feel that the proposed schedule can be used as a reference.
Amundi does not share the idea to include shares of components of the main indices in the 40% limit that is suggested for convertible, shares and other instruments. First, this proposal is not consistent with the fact that shares are highly liquid and remain one of the few actively traded instruments in periods of stress, probably because of the diversity of types of shareholders. Secondly, this limitation would profoundly alter the investment strategy of many funds that are required to invest exclusively or to a very large proportion in shares. Two examples will bring an illustration:
• classical equity funds that are benchmarked to equity indices have no choice but giving equities as collateral (knowing that UCITS are forbidden to reuse collateral and thus cannot use collateral transformation services)
• and PEA funds in France have to show a minimum investment of 75% in eligible shares; they might be prevented to use derivative contracts and dramatically reduce the investment universe of their holders. This is particularly true if the 40% ratio were computed on the basis of total collateral and not NAV for funds.
For concentration limits, we already expressed our view in question 2 that they should not apply below a threshold of 100 million € in collateral. We also suggest that the percentage be computed on the basis of the NAV for funds, both UCITS and AIFs. If introducing a specific case for funds seems difficult, however we strongly think that it is justified, we could establish the ratio with reference to the Notional Amount of the derivative contracts and then apply a much lower figure. This reference would be totally consistent with the thresholds for applicability of the IM requirement (from 3000 billion to 8 billion €). Lastly we advocate that funds complying with ESMA December 2012 guidelines should be exempted from further diversification rules. Their compliance to ESMA rules should imply their compliance with EMIR requirements.
Amundi has the view that BCBS/IOSCO principles have taken a prudent and pragmatic approach when strongly limiting the possibility to re-use or rehypothecate received collateral. We think that there is no reason not to follow these principles. OTC derivatives are traded between professionals. If a counterparty agrees that the re-use of collateral will be positive and made for the benefit of its clients, re-use should not be prohibited.
An example of circumstances where re-use will be helpful is to be found in the case of back to back transactions where a counterparty A manages the exposition it took from a derivative contract with a fund through a reverse transaction with counterparty B; the collateral received from the fund by A may be re-used and posted in favour of B that finally takes the market risk. The more so if A is a financial institution that provides the fund with a guarantee.
The required prior explicit approval by the poster of the collateral (transaction by transaction), the fact that it can check that the re-use is made in conjunction with a further transaction aiming at managing the risk initially taken from the fund and not to take further exposure, the possibility to earmark the collateral…all is made to limit the re-use to a very few cases when professionals willfully agree.