This was the reason for the introduction of an exemption for pension provision and retirement products in the framework Regulation.

Transcription

1 ABI response to the joint Discussion Paper on Draft Technical Standards on risk mitigation techniques for OTC derivatives not cleared by a CCP under the Regulation on OTC Derivatives, CCPs and Trade Repositories The ABI The Association of British Insurers (ABI) is the voice of insurance, representing the UK s insurance, investment and long-term savings industry, the largest in Europe and the third largest in the world. It was formed in 1985 to represent the whole of the industry and today has over 300 members, accounting for some 90% of premiums in the UK domestic market and who, as institutional investors, have some 1.7 trillion of funds under management Summary The ABI welcomes the opportunity to comment on the joint discussion paper on draft technical standards on risk mitigation techniques for OTC derivatives not cleared by a CCP under the Regulation on OTC Derivatives, CCPs and Trade Repositories. It is imperative that full and careful consideration is given by the ESAs to the draft technical standards as there is potential for costly unintended consequences for the providers of long-term savings products and their customers if this is not worked out correctly. As representatives of both pension funds and life insurance products (that both provide income for retirement), the ABI closely followed EMIR during the Level 1 negotiations. The ABI welcomes the objectives of EMIR as a means of supporting more robust and safer financial markets. As investors and product providers in these financial markets ABI members want this outcome too. However, it was widely acknowledged that long-term savings institutions (pension funds and pension-related insurance funds) would be disproportionately hit by the costs by the mechanism established under EMIR. These costs are of such a size they cannot be absorbed by the companies and will have to be passed on to pensioners and savers. As a consequence an exemption for certain pension schemes from clearing was agreed, as long as they were subject to proper risk mitigation techniques. If this exemption is to have any worth to these pension arrangements, it is important that the risk mitigation requirements are not over loaded with excessive restrictions nor try to import elements from the rest of the EMIR Regulation, in particular on margin and collateral which are the main drivers of the disproportionate costs for long terms investors (and the need for the exemption).

2 ANNEX Q1. What effect would the proposals outlined in this discussion paper have on the risk management of insurers and institutions for occupational retirement provision (IORPs)? The effect could be severe. IORPs and other providers of retirement provision, including insurers, will be severely disadvantaged by having to post Initial Margin for non-centrally cleared trades, as the exemption in EMIR does not extend beyond the requirement to centrally-cleared trades, especially if the type of collateral required is upgraded - for example, by requiring cash for collateral requirements for initial and variation margin. This would mean disinvesting out of bonds and other long-term investments, and/or potentially being charged a fee imposed by the banks for converting bonds into cash on a client s behalf. Both are costly. As long term investors, insurance companies and pension funds invest considerably in long-term assets, such as government and corporate bonds. The only way to minimise the costs on long-term investors would be to ensure corporate and government bonds are accepted as collateral for Initial and Variation Margin. In addition, these assets would be held separately as margin rather than being fully invested for the ultimate benefit of pensioners so as to provide a return. This was the reason for the introduction of an exemption for pension provision and retirement products in the framework Regulation. Q2. What are your views regarding option 1 (general Initial Margin requirement)? We do not believe any of the three options are appropriate. All three options have been drafted with banks in mind rather than the clients, such as insurers and pension funds. Currently, certain funds which are identified as low risk and have a minimal chance of default including pension and insurance funds do not pay any Initial Margin. Insurance companies are prudentially regulated and are well funded and stable institutions. Furthermore, insurance companies are only allowed (under Solvency I and Solvency II) to use derivatives for the purposes of hedging or for efficient portfolio management. Insurers will primarily use derivatives to hedge against interest rate movement and inflation risks. Insurance companies are not allowed, under EU legislation, to use derivatives for speculation. In terms of being a counterparty to a derivatives trade, insurance companies are exceptionally low risk. It is distinctly disadvantageous (because insurers invest in the long term and will therefore be disproportionally hit by margin costs) and unnecessary (due to their low counterparty risk) for these particular funds to have these margin requirements imposed on them. This is in addition to the fact that client funds are then not being best utilised to the ultimate pensioners benefit. We believe there should be an exemption for those funds which are already regarded as low risk. Such an exemption should be formulated under the agreed definition of pension schemes arrangements in EMIR Article 2(7a)(a)-(d). The burdens that the phase-in of central clearing for pension schemes in EMIR would also apply if Initial Margin becomes mandatory for such funds, particularly if the list of eligible collateral is too limited. Q3. Could PRFCs adequately protect against default without collecting Initial Margins? 2

3 Yes, we believe they are in the business of extending credit lines to others whilst being responsible for their own risk mitigation. This particularly applies to very stable counterparties like pension and insurance funds, who are extremely unlikely to default. Q4. What are the cost implications of a requirement for PRFC, NPRFC and NFCs+ to post and collect appropriate Initial Margin? If possible, please provide estimates of opportunity costs of collateral and other incremental compliance cost that may arise from the requirement. Please see our response to Q1. Pension and insurance funds will be severely disadvantaged by having to put up Initial Margin which are fund holder assets that can therefore not be fully utilised to obtain the best return for pensioners. Due to the short time frame of the consultation we are unable to provide cost estimates. Q5. What are your views regarding option 2? We find Option 2 extremely disproportionate. Whilst we recognise the underlying desire to ensure stability within the financial system, we would seriously question using highly rated assets to support those of lesser creditworthiness. This would be to the serious detriment of the end client, including pension funds and insurers and their customers. Pension and insurance funds will be severely disadvantaged by having to put up Initial Margin which are fund holder assets that can therefore not be fully utilised for the best return for pensioners. The criteria for institutions to post IM should not only be limited to their systemic nature but also the likelihood of default. Q6. How in your opinion - would the proposal of limiting the requirement to post Initial Margin to NPRFCs and NFCs+, impact the market / competition? This proposal creates incentives for NPRFCs and NFCs+ to deal with each other more frequently and hence by-pass the PRFCs. We believe this poses a whole set of other risks to the financial system. Q7. What is the current practice in this respect, e.g. - If a threshold is currently in place, for which contracts and counterparties, is it used? - Which criteria are currently the bases for the calculation of the threshold? See answer to questions 1 & 2. Currently, certain funds which are regarded as low risk and unlikely to default including pension and insurance funds do not pay any Initial Margin. It is therefore distinctly disadvantageous to these funds to have these margin requirements imposed on them, as the client funds are then not being best utilised to the clients benefit. We believe there should be an exemption for those funds which are already regarded as low risk. Q8. For which types of counterparties should a threshold be applicable? If the option to apply a threshold is to be applied, it should not only take into account the exposure of the bank to the counterparty, but also the likelihood of default of that counterparty, particularly 3

4 extremely low-risk clients using derivatives for hedging purposes like pension funds and life insurers. The ESAs have correctly identified the importance of credit quality in Paragraph 34, including that the threshold should not only be calculated using external credit ratings, but further due diligence and understanding of the business model is necessary. Q9. How should the threshold be calculated? Should it be capped at a fixed amount and/ or should it be linked to certain criteria the counterparty should meet? We would prefer any threshold to be more closely linked to the counterparty credit quality and underlying risk profile, than the bank s exposure to the counterparty. Q10. How in your opinion - would a threshold change transactions and business models? An appropriately calibrated threshold should allow pension and insurance funds to continue their directional use of derivatives for hedging purposes without too great market dislocation. On the other hand, an inappropriately designed Initial Margin obligation would raise costs for pension funds and insurance funds to the point where they would have to consider whether to take a hedge at all, with consequences for the investors in the funds. Q11. Are there any further options that the ESAs should consider? As we said in our answer to Q1, the ESAs should strongly consider the possibility of permanently exempting pension schemes from the mandatory posting of Initial Margin. This would be a simpler solution than applying complex threshold figures based on the credit quality and risk profile of the counterparty. Q12. Are there any particular areas where regulatory arbitrage is of concern? No areas of concern regarding regulatory arbitrage come to mind. We do, however, have concerns about the compatibility of global regulation, most notably between EMIR and Dodd-Frank (US). We are concerned that EMIR and Dodd-Frank and other international regulation will not be as sufficiently interoperable as required in order to avoid a disruption of market flows and liquidity. We would request the ESAs consider the global implications of incompatibility both in its own right but also as we believe this is where regulatory arbitrage is most likely to occur. Q13. What impacts on markets, transactions and business models do you expect from the proposals? Please see our answers to Q1, Q4 and Q10. If Initial Margin is mandated for pension and insurance funds we would expect to see fewer transactions which will impact on overall market liquidity, and lower returns from such investment funds which may have knock-on effects to certain parts of investment businesses, and a subsequent overall rising of the levels of risk being taken. 4

5 Q14. As the valuation of the outstanding contracts is required on a daily basis, should there also be the requirement of a daily exchange of collateral? If not, in which situations should a daily exchange of collateral not be required? Yes, we agree that collateral should be exchanged daily. However, the ESAs should also consider the value of Minimum Threshold Agreements (MTAs) as they currently exist in ISDA CSAs. They are currently in use by low-risk clients like insurers and pension funds to set a certain threshold amount below which the client does not have to post Variation Margin. Q15. What would be the cost implications of a daily exchange of collateral? As this has been market practice in the UK for some time, we do not foresee any particular additional cost implications, except perhaps for some smaller non-financial participants. In terms of daily exchange of collateral, the only significant cost increase would occur if insurance companies and pension funds were unable to post government and corporate bonds as collateral, or if existing MTAs were removed. Q16. Do you think that the Mark-to-market method and/or the Standardised Method as set out in the CRR are reasonable standardised approaches for the calculation of Initial Margin requirements? We believe that some may find the CRR mark-to-market requirements onerous, particularly those entities which are not currently subject to CRR. We believe those that are in a position to create more risk-sensitive modelling should be allowed the ability to use internal models, subject to approval by the National Competent Authority. We believe they are unlikely to use potentially lower margins as a competitive advantage as any such advantage would need to be weighed against the increased risks. It is difficult to envisage how an insurance company, as the client counterparty, regulated under Solvency II would be able to adapt its practice to accommodate valuation methods built for banks. Care needs to be taken in this area. The more eligibility is limited centrally, rather than allowing participants to bilaterally decide and monitor in the context of market conditions, the more there is a risk that in times of market distress entities which could collateralise may be prevented from doing so because of the nature of the assets they have available (even if they had appropriate haircuts applied). For example, if corporate bonds were not eligible and there was a market event, is it preferable that an entity failed its collateral call because of overly restrictive rules than that it collateralised with corporate bonds with appropriate haircuts? We understand the need to set absolutely minimum standards, or suggest appropriateness but care must be taken to avoid potential unintended and counterproductive consequences of trying to prescriptively set the bar too high. Q17. Are there in your view additional alternatives to specify the manner in which an OTC derivatives counterparty may calculate Initial Margin requirements? Current market practice is weighted towards counterparty risk. We believe that counterparties should be allowed to determine the specific Initial Margin requirement within certain parameters. 5

6 We recognise the underlying desire to be more prescriptive, but would suggest a set of criteria that create parameters within which counterparties can reach consensus as to the appropriate value of Initial Margin requirements. We would propose that one of those boundaries should be zero, for those with near-zero default risk. Q18. What are the current practices with respect to the periodic or event-triggered recalculation of the Initial Margin? Approaches to calculating Initial Margin requirements are periodically reviewed. Q19. Should the scope of entities that may be allowed to use an internal model be limited to PRFCs? No. We believe there are NPFRCs which are fully capable of using internal models. As with PRFCs, we would expect these to have been approved by the national competent authority. Q20. Do you think that the Internal Model Method as set out in the CRR is a reasonable internal approach for the calculation of Initial Margin requirements? No. As insurers we will be using the Solvency II definitions of Internal Model Method. The definitions are different. To have to run two different model approaches would be excessively burdensome, time consuming and costly. However, we recognise that the methodology proposed in the CRR is the most sensible one for our usual counterparties the banks to use. For non-pension-related funds, we would anticipate discrepancies in the calculation for Initial Margin. Both the CRR and Solvency II models have been tailored for industry-specific issues. We propose that insurers use the Solvency II model for the calculation, and that banks our usual counterparties - should use the CRR model. We believe that each side should post symmetrical Initial Margin of the lower of the two calculations. The counterparty that calculated the higher margin should be required to hold the additional amounts (ie. the difference between the lowest and highest margin calculations) in a separate account. Q21. Do you think that internal models as foreseen under Solvency II could be applied, after adequate adjustment to be defined to the internal model framework, to calculate Initial Margin? What are the practical difficulties? What are the adjustments of the Solvency II internal models that you see as necessary? Yes, we do believe that internal models as determined by Solvency II could be applied. As insurers, we would expect to use the Solvency II model as these will be being used internally. However, we would not necessarily expect our usual counterparties the banks to want to use a Solvency II model. We propose that banks use the CRR model, that insurers use the Solvency II model. We believe that each side should post symmetrical Initial Margin of the lower of the two calculations. The counterparty with the extra regulatory burden should be required to hold this in a separate account. Q22. What are the incremental compliance costs (one-off/on-going) of setting up appropriate internal models? 6

7 For one-off costs, the November 2010 PwC survey indicates that the industry-wide Solvency II implementation cost is on course to exceed the European Commission s estimate of EUR3 billion. Lloyds of London is expecting to spend GBP250 million in total on implementation, with annual ongoing Solvency II related expenses of about GBP60 million to GBP70 million. Multinational insurers in the UK have set aside roughly GBP100 million for Solvency II implementation. The expenses incurred by regulatory bodies in enforcing Solvency II will, naturally, be passed on to the industry. In the UK, for example, the Financial Services Authority (FSA) will levy over GBP34 million in fees against insurers in 2011 and 2012 to cover its costs of implementing the Solvency II regime. The fees aim to cover the expense of approving insurers' internal models as well as other related costs. The total FSA cost of implementing Solvency II over the lifetime of the program remains in the anticipated range of GBP100 million to GBP150 million. The FSA s proposed annual funding requirement for regulated firms in 2011 and 2012 is GBP500.5 million. Q23. To what extent would the mark-to-market method or the standardised method change market practices? The market will not be harmonised, due to differing methodologies being employed. Those sophisticated enough will be able to use a more risk-adjusted method, to the potential detriment of those using the mark-to-market method. Consequently there will likely be disagreement on which Initial Margin amount to use. Q24. Do you see practical problems if there are discrepancies in the calculation of the IM amounts? If so, please explain. Please see our responses to Q20 and Q21. We would anticipate discrepancies as both the CRR and Solvency II models have been tailored for industry-specific issues. Given our proposal that banks use the CRR model and that insurers use the Solvency II model, and that each side should post symmetrical Initial Margin of the lower of the two calculations. the counterparty with the extra regulatory burden should be required to hold this extra amount in a separate account. We propose this would provide a practical, measurable and auditable solution. We believe this would provide a practical, measurable and auditable solution. Q25. Would it be a feasible option allowing the party authorised to use an internal model to calculate the IM for both counterparties? Possibly, though we do see the possibility of the non-authorised party being forced to put up more Initial Margin that may really be required, depending on the variable parameters of the internal model being used. Should this option be pursued we would want to see sufficient safeguards to ensure that any disagreement could be addressed using the dispute resolution regime. Q26. Do you see other options for treating such differences? We have no comment. 7

8 Q27. What kinds of segregation (e.g., in a segregated account, at an independent third party custodian, etc.) should be possible? What are, in your perspective, the advantages and disadvantages of such segregation? In the event Initial Margin is pursued, and if it cannot be re-hypothecated, for it to be meaningful it would have to be held in segregated accounts by third party custodians in order to ring-fence the assets and give the client as much protection as possible. However, we recognise that this will result in additional costs which will end up being paid by the end customer. We would therefore request that the ESAs ensure that these charges are not so excessive that the cost of this protection becomes prohibitive. As outlined in the preliminary Level 1 agreement (Article 37.5) clients must be given the choice of full segregation of their assets in segregated accounts with ring-fencing of the assets, and omnibus accounts. The former will be more costly, however some clients may decide this increased level of protection is the most appropriate for them. Most importantly clients must be offered the choice between the two, and that it is clear what levels of protection/segregation are offered in each case. Q28. If segregation was required what could, in your view, be a possible/adequate treatment of cash collateral? We cannot see cash in a triparty arrangement being plausible. In this kind of arrangement, the insurer would be at risk in some way and presumably the triparty agent would receive some sort of fee without having the concern for the risk taken. This risk would presumably fall within some predetermined limit of where the tri-party agent invests in the cash. Q29. What are the practical problems with Tri-Party transactions? See Question 28; we cannot see it working for cash. Q30. What are current practices regarding the re-use of received collateral? We believe that asset owners should not have any restrictions over how they choose to manage those assets. Current market practice is to re-use received collateral in order to maximise returns to investors. Q31. What will be the impact if re-use of collateral was no longer possible? We are concerned that if considerable amounts of collateral is posted and not reusable, that there may be cascading effect in the liquidity of available lack of collateral which could leave the market effectively grid-locked. Banning re-use would have the additional consequence of lower returns and increased costs to our clients. Q32. What are, in your view, the advantages and disadvantages of the two options? We strongly prefer option 2 which gives all parties more scope and flexibility to cover differing preferences and circumstances. We note, however, that the consultation does not seem to differentiate sufficiently between Initial Margin and Variation Margin, which might require different type of collateral. Do the ESAs intend to specify different collateral eligibility for IM and VM? The disadvantages of Option 1 include a potential lack of available collateral as well as the inequality of banks and CCPs as to types of collateral accepted. The reason a temporary 8

9 exemption for pension scheme was built into EMIR was to recognise the burden of cash or nearcash Variation Margin had on such schemes. Q33. Should there be a broader range of eligible collateral, including also other assets (including non-financial assets)? If so which kind of assets should be included? Should a broader range of collateral be restricted to certain types of counterparties? Yes, we believe there should be a broader range of eligible collateral, which for insurers are currently set out in ISDA CSAs. In line with current practice, we believe that government and corporate bonds and asset-backed securities should certainly be included, and that other assets should be acceptable subject to agreement on both sides, with appropriate haircuts and or being priced in. More broadly, however, if the ESAs wish to make the financial system more robust then they must ask whether the counterparties are sufficiently protected by the collateral that the banks post to them, and not merely what collateral the counterparties are required to post with the banks. Q34. What consequences would changing the range of eligible collateral have for market practices? We believe that we would see more transactions particularly from corporates, which could help to reduce economic risk. Additionally, we would avoid the potential gridlock arising from a lack of near-cash collateral. By allowing the use of high quality corporate bonds and ABS, it will not just incentivise but it will allow those, such as long term investors, to clear transactions via CCPs who would otherwise be priced out of clearing as an option. We would not expect increased risk due to increased competition as we would expect counterparties to exercise good judgement as to what assets were deemed acceptable collateral for specific transactions. Q35. What other criteria and factors could be used to determine eligible collateral? Any collateral that is directly correlated with the underlying for the transaction should not be deemed eligible. Q36. What is the current practice regarding the frequency of collateral valuation? Our Members undertake collateral valuations on a daily basis. Q37. For which types of transactions / counterparties should a daily collateral valuation not be mandatory? We believe there may be some corporates with a very illiquid and/or highly bespoke hedging contracts which should not be mandated to undertake daily collateral valuation. However, we do agree with the proposal that all financial institutions and particularly PRFCs should have to undertake valuations daily. Q38. What are the cost implications of a more frequent valuation of collateral? 9

10 We agree with the ESAs that valuation should be undertaken daily. We are not convinced of any merits of valuing collateral more frequently than daily. This would result in increased costs for personnel as well as systems. Q39. Do you think that counterparties should be allowed to use own estimates of haircuts, subject to the fulfilment of certain minimum requirements? We prefer the option offered whereby counterparties that meet certain minimum requirements will be allowed to use their own estimation of appropriate haircuts. They are usually closest to the details and information and are best able to determine the appropriate levels to apply. Q40. Do you support the use of own estimates of haircuts to be limited to PRFCs? No, we believe there are certain NPRFCs (eg. UCITS funds) who are capable of determining appropriate haircuts. Q41. In your view, what criteria and factors should be met to ensure counterparties have a robust operational process for the exchange of collateral? We support the proposal for draft technical standards specifying the risk management procedures for the exchange of collateral and, in particular the operational process relating to the exchange of collateral, and we agree that such processes should be robust. Q42. What incremental costs do you expect from setting up and maintaining robust operational processes? As our members already carry out daily collateral valuation, we would expect minimum if any extra costs given their currently robust systems. However, there may be some details created by the technical standards that may require some parts of the system to be amended slightly which would incur some costs. Extra costs are undoubtedly going to be faced by those who do not already carry out daily valuations. Q43. What are your views regarding setting a cap for the minimum threshold amount? How should such cap be set? We support the proposals for a cap for a minimum threshold amount. We believe this threshold should be zero, given the potential for nugatory collateral exchange, and the cap should be at a low, but appropriate, value. We also urge consideration to be given to what kind of collateral is eligible within this cap. Q44. How would setting a cap impact markets, transactions and business models? We believe a zero threshold and low value cap would give all market participants the opportunities to better mitigate risk by hedging as required, without there being the additional burden of collateral requirements. Q45. In your views, what should be considered as a practical or legal impediment to the prompt transfer of own funds or repayment of liabilities between the counterparties? 10

11 We think practical or legal impediments should be restricted to where there are specific legal restrictions on transferring own funds (e.g. exchange controls); where there are regulatory restrictions to transferring own funds (ideally this should not include where regulatory permission is required to transfer, only where the regulator has stated that own funds cannot be transferred, if for example transferring the own funds would reduce local solvency below a certain level); and practical impediments e.g. on insolvency of one counterparty. Q46. What is the current practice regarding the collateralisation of intragroup derivative transactions? It is not current market practice to provide collateral for intragroup transactions. Intragroup transactions affect only the parties involved without any market impact, are measures of internal risk management and the overall risk profile of the group does not change. The credit risk is managed centrally and there is therefore no requirement for collateral exchange. If collateral requirements were introduced for intragroup transactions, these would create additional administrative burdens and cause substantial operational costs for cash management. Q47. What is the impact of the presented options on the capital and collateral requirements of the counterparties affected by the relevant provisions and the span of time necessary to comply with the Regulation? We do not have any comment at the moment, but would hope to be able to get further information to the ESAs in due course. ABI, 2 April