The known unknowns of central clearing
Speech by Benoît Cœuré, Member of the Executive Board of the ECB,
at the meeting on global economy and financial system hosted by the University of Chicago Booth School of Business Initiative on Global Markets,
Coral Gables, 29 March 2014
Ladies and gentlemen ,
Today, I want to share some thoughts with you on central clearing.
At the Pittsburgh Summit in 2009, G20 leaders committed to have all standardised OTC derivatives contracts cleared through central counterparties (CCPs). The mandatory use of central clearing as a regulatory tool for mitigating systemic risk is a fairly new feature in financial market regulation. Central clearing definitely has the capacity to promote financial stability, provided it is appropriately used and any potential side effects are well understood and effectively mitigated.
In that context, there are still a number of questions that we need to look at. Let me call them the “known unknowns” of central clearing.
I want to touch upon the following issues:
Systemic effects of mandatory clearing
Recovery versus resolution of central counterparties
Optimal design of loss allocation tools
Evolution of global clearing structures
Systemic effects of mandatory clearing
Let me start with the systemic effects of mandatory central clearing. Clearing houses have done their job very well in the crisis, contrary to other financial institutions, and successfully faced the challenges of a rapidly growing volume of transactions. We want to make sure that the same holds true in the new world of mandatory clearing.
The systemic effects of central clearing are numerous and, although the overall balance is undoubtedly positive, some of them are ambiguous and should be addressed carefully.
On the one hand, mandatory clearing reduces systemic risk very significantly as it
leads to the collateralisation of trades that might have remained uncollateralised otherwise,
enhances risk management by means of sophisticated margining methods,
allows for mutualisation of risk, which promotes the resilience of the financial system, especially in the event of catastrophic scenarios,
offers multilateral netting so that less collateral is needed to manage risks, and
facilitates risk management as it helps to overcome information asymmetries that participants in non-centrally cleared markets face.
On the other hand, some consequences of central clearing may increase systemic risk:
as a result of growing risk concentration in CCPs and in a limited number of large global dealers, which act as general clearing members as a gateway for other financial institutions to CCPs,
as a consequence of mutualisation, risks may spread among participants and markets more widely and the distribution of losses is more difficult to predict,
in order to avoid mandatory clearing, we might see migration of risk outside of the regulated universe through the artificial creation of bespoke (and hence risky) products that are not suitable for central clearing.
Moreover, the systemic effect of clearing depends on whether risk is redistributed to those who are more able to deal with it, which is difficult to determine. Finally, the overall systemic effect depends on the incentives for risk-taking and central clearing that accompany risk redistribution.
To promote central clearing and avoid regulatory arbitrage, a number of reforms have been adopted including: (i) mandatory clearing, (ii) margin requirements for non-centrally cleared trades, and (iii) higher capital requirements for non-centrally cleared trades. This highlights the crucial complementarity between the regulation of banks and the regulation of financial market infrastructures (FMIs), with the overall objective that regulation reduces overall risk, and does not shift it around the financial system. Cooperation between the Committee on Payments and Settlement Systems and the International Organisation of Securities Commissions (CPSS-IOSCO) on the one hand, and the Basel Committee on Banking Supervision (BCBS) on the other hand, is key in this respect.
A macroeconomic impact assessment led by the BIS focused on the effects of these reforms and found net benefits of 0.12% of GDP per year. These are estimates of the long-run consequences of the reforms and are expected to apply once those reforms have been fully implemented and produced their full economic effects.
However, central clearing also involves risk redistribution and potentially the creation of new risks. The overall impact is therefore less certain: there are significant benefits which need to be balanced against potential new risks and uncertain risk redistribution effects.
Ultimately, it seems to me that the success of central clearing as a tool for promoting financial stability critically depends on two factors: (i) the ability of CCPs to handle the risks they are concentrating and thus the design of their corporate governance, risk management and recovery frameworks and (ii) the ability of the official sector to address effectively any unintended side effects and “new” risks that may result from the push towards more central clearing.
Recovery versus resolution of CCPs
A second issue that I should like to highlight today concerns the relationship between recovery and resolution of CCPs. As central clearing involves risk concentration in what we may see as “super systemically relevant” institutions, it is crucial that such institutions can recover from severe shocks or be resolved if necessary.
Now there are several questions that are typically raised in this context:
At what point in time does a CCP need to be resolved because it cannot recover?
Given that CCPs provide critical services and are required to cover any losses or liquidity shortfalls that may arise, is resolution of a CCP a realistic concept at all? Is it ever needed or appropriate?
What makes CCPs different from other financial institutions? Do these differences require different rules for their recovery and resolution?
The CPSS-IOSCO Principles for financial market infrastructures (PFMIs) require CCPs to have a plan for allocating any losses and liquidity shortfalls that the regular “waterfall” – that is, a sequence of prefunded financial resources – does not cover. Alternatively, the respective CCP might have to be resolved. The CPSS-IOSCO and the Financial Stability Board (FSB) will provide guidance on the recovery and resolution of CCPs by mid-2014.  The question is: should resolution only be performed if and when recovery has failed? Or could there be circumstances in which the resolution of a CCP is preferable to its recovery, even if recovery was still feasible?
Let me first explore some features that make financial market infrastructures different from other financial institutions:
Service continuation (“critical services”) is more important in the case of FMIs. They are utilities at the heart of the financial system. Their smooth operation is essential for all other types of financial service.
In most cases, there are only a few, if any, substitutes or alternative providers.
Unlike banks, FMIs have ex ante loss allocation rules that they can draw on for the purpose of recovery.
Not all FMIs are exposed to credit risk (but they are still systemically important).
The size and composition of their balance sheets (e.g. rarely subordinated debt instruments) is different, which makes bail-in as a recovery or resolution tool difficult.
To the extent that the use of a CCP is mandatory, it is absolutely essential that the relevant clearing services are preserved regardless of the circumstances.
In conclusion, the recovery of CCPs in order to maintain service continuation is much more important for CCPs than for other financial institutions. That’s why the PFMIs require CCPs to be able to allocate any losses that are not already covered by regular risk management.
So if the recovery of CCPs is so vital, is there any scope for resolution at all? I think there is, for a number of reasons:
First of all, the attempt to recover may simply fail: the ex-ante recovery plans may prove inadequate ex post. Recovery plans have fortunately not yet been tested, and it remains to be seen how they work in practice.
Most importantly, however, recovery is not an end in itself, but a means to preserve financial stability. Entry into resolution should be possible if the recovery plan has failed to return the FMI to viability or if its implementation would otherwise compromise financial stability, for example if the application of a particular loss allocation tool turns out to overexpose certain (groups of) participants.
Moreover, resolution may also become necessary if participants do not want to use the CCP any longer and prefer to wind it down.
In conclusion, the resolution authorities have the responsibility (and are in a better position than CCPs) to do what is systemically needed in a crisis situation, which may mean in extreme cases that the application of resolution procedures is preferable to the activation of a recovery plan.
Of course, much will depend on authorities doing the right thing at the time of the crisis. This in turn assumes that authorities have enough information at any point in time to "connect the dots" of a complex network of CCPs and global dealers. Hence the critical importance of enforcing the new reporting requirements to trade repositories (TRs), and further to this, of being able to aggregate data collected by TRs across products, institutions, and jurisdictions, and conduct analytical work based on this data. 
Optimal design of loss allocation tools
Let me now turn to the third known unknown, which is the subject of much debate: what is the optimal design and composition of a CCP’s recovery tools?
In the field of recovery, we are entering new territory, which is why the authorities have wisely taken a very cautious approach to giving guidance to FMIs on the design of their recovery plans.
Ideally, recovery tools should be comprehensive, effective and controllable; they should create appropriate incentives for risk management and minimise any negative impacts. However, the challenge is that no individual tool can equally meet all of these criteria because of trade-offs, for example:
Uncapped cash calls are comprehensive, but may create disincentives for central clearing.
Variation margin haircutting can be comprehensive and effective (as it reduces pay-outs rather than requiring additional pay-ins), but the loss allocation and hence the systemic effect is uncertain.
Initial margin haircutting may help achieve comprehensiveness, but increases the potential for contagion risk.
The optimal choice of recovery tools is still evolving and remains a controversial issue. The regulatory approach by CPSS-IOSCO is therefore non-prescriptive and aims to provide guidance that should help FMIs to choose the appropriate (mix of) recovery tools. This approach, however, might need to be adapted over time.
Evolution of global clearing structures
Let me now turn to the fourth known unknown, which is the optimal design of the global clearing structure.
Such a design depends on numerous factors and questions:
How many CCPs do we need at global level?
How many markets should a CCP serve?
What are the implications of competitive clearing where there are several CCPs serving the same trading venue?
Should CCPs be linked and how? Should links be limited to cash-market CCPs or also apply to derivatives and commodities CCPs?
What degree of tiering is optimal and how should participants get access to CCPs when needed? This question has gained in importance with the establishment of the clearing obligation.
What we see today is a globalisation process that has triggered a trend towards global clearing with increasingly large global CCPs. Moreover, it appears that many banks prefer indirect access to clearing services so as to comply with the clearing obligation. Client clearing seems thus to be dominated by a few large global intermediaries. This concentration creates a higher degree of dependency on this small group of firms.
The trend towards the globalisation of clearing has benefits and risks. On the one hand, there is greater scope for netting that allows for collateral savings, the lower cost of direct access and using CCPs, and an increase in transparency for both regulators and CCPs. On the other hand, there is a potential for systemic risks as CCPs may become too big to fail. Also, market power may push up clearing fees and restrict entry. This could create disincentives for central clearing and could undermine the efficiency gains reaped in the first place by eroding the pricing power of global dealers. There could also be cross-border regulatory frictions in the case of multiple (and inconsistent) national regulations, and last but not least the cross-border liquidity provision in several currencies becomes more complex.
This leads me to the last and maybe most challenging “unknown”. Given CCPs' role as financial market utilities that has been reinforced by mandatory clearing requirements, and their increasing systemic relevance, would the world be a safer place, if CCPs had been established as non-for-profit institutions? My tentative answer is no. I have outlined the importance of good risk management in CCPs, and the ability to adapt to a changing market environment in terms of counterparties, products, and technologies. I am convinced that this will be better done in a market environment. But this certainly requires strong supervision, effective cooperative oversight arrangements, and setting up the right incentive structure so that CCPs don’t lose sight of public interest, starting with the interest of their indirect members.
Let me conclude. Central clearing will make the global financial system safer. However, it also has the potential to redistribute risk. Its overall systemic effect will thus depend on the incentives for risk-taking and central clearing that accompany risk redistribution. Hence the need for the official sector to provide the right incentive framework ex-ante, produce data that can help understand the distribution of risk, and tightly supervise CCPs in a spirit of cooperation across jurisdictions.
Thank you for your attention.
I would like to thank Andreas Schönenberger for his contributions. I remain solely responsible for the opinions contained herein.
See Financial Stability Board. “Consultation Document: Application of the Key Attributes of Effective Resolution Regimes to Non-Bank Financial Institutions”, 12 August 2013, and CPSS-IOSCO. “Consultative Report: Recovery of Financial Market Infrastructures”, 12 August 2013. See also P. Tucker. “Central counterparties in evolving capital markets: Safety, recovery and resolution”. Banque de France Financial Stability Review, No. 17, April 2013.
See Financial Stability Board. “Consultation Paper: Feasibility Study on Approaches to Aggregate OTC Derivatives Data”, 4 February 2014.