RDP 2013-12: Central Counterparty Links and Clearing System Exposures 2. CCP Links and Exposure

Clearing is the process of transmitting, reconciling and confirming payment obligations after a trade has been negotiated. Traders in financial markets face the risk that their counterparty may default after a trade has been negotiated but before it has settled, leaving them exposed to a loss in replacing the trade if market values change adversely. To reduce exposure to this ‘replacement cost’ risk, trading counterparties in many securities and derivative markets clear their trades through a CCP, which intermediates between all participants in a given market through novation.[3] Novation replaces the contract between a buyer and a seller with two contracts: one between the buyer and the CCP and one between the seller and the CCP. Novation protects traders from the risk of their counterparty defaulting, as the CCP becomes responsible for meeting any defaulting trader's obligations.

To novate trades to a CCP, traders must both be participants of that CCP, subject to the CCP's rules and requirements. These typically involve the CCP monitoring participants' financial health and trading activities, and participants posting collateral (initial margin) to cover the CCP's potential losses. By centralising all transactions through a single entity, CCPs concentrate a complex network of bilateral exposures into a single set of exposures, subject to strict risk management standards. This multilateral netting of bilateral exposures aims to reduce overall exposure in the system and economises on collateral, since less collateral is required to cover potential losses.

CCP links facilitate novated trades between participants of different CCPs. Here, novation replaces the contract between a buyer and a seller with three contracts: one between the seller and its CCP, one between the buyer and its CCP, and one between the two CCPs. Interoperability allows participants of one CCP to trade with participants of other CCPs without the costs of maintaining multiple CCP memberships. Interoperability can broaden market access, allowing trade across a wider range of products and venues, while promoting competition by lowering barriers to entry for new CCPs and allowing participants to choose their preferred CCP. By enabling participants to hold all their positions in a single CCP, interoperability avoids the loss of netting that occurs when positions are fragmented across multiple clearing venues.

Figure 1: Unlinked and Linked CCPs
Figure 1: Unlinked and Linked CCPs

However, CCP links also generate exposures that could pose financial stability risks. In particular, links introduce credit exposure between linked CCPs. This exposure arises because all linked positions in a defaulting CCP are taken on by linked CCPs, potentially leading to very large losses. Although the probability of a CCP default is small, such an event could threaten the solvency of any surviving linked CCPs if the number of trades cleared across the link and the adverse movement of prices were significant enough. CCP links could then become the cause of significant disruption to the financial system.

It is therefore important that any inter-CCP exposures are well managed, and that linked CCPs provide sufficient collateral to each other to ensure that a default by one would not threaten the solvency of the others. As the range of financial products cleared by CCPs becomes broader and proposals for interoperability increasingly common, the effect of CCP links on exposures has become a topic of significant policy interest. In Section 3, we estimate the effect of interoperability on exposures using a stylised model in which market participants conduct centrally cleared trades through linked and unlinked CCPs.

2.1 Literature Review

There is a small but informative literature on the impact of different clearing market structures on the financial system. Both qualitative and quantitative approaches have been used to examine the effects of recent developments in market structure on collateralisation and systemic risk.

CPSS (2010) provides a broad overview of clearing market structure developments over the past decade. It identifies competition among CCPs as a key trend, noting that this could lead to the fragmentation of clearing, increasing the number of outstanding contracts and decreasing netting efficiency. While interoperable links may reduce these inefficiencies, there is a contagion risk that arises from inter-CCP exposures, particularly if linked CCPs are unable to monitor other CCPs' risk management practices effectively.

There has been some academic interest in modelling different types of market structures and examining their impact on market functioning, exposures and risk. Jackson and Manning (2007), for example, investigate the implications for risk of three alternative clearing structures: bilateral clearing, ‘ring’ clearing, and CCP clearing. They analyse both the magnitude and distribution of replacement cost losses across trading counterparties. By simulating trading positions and asset price changes using defined distributions, their model demonstrates that multilateral netting reduces replacement cost losses, and that CCP clearing distributes those losses more evenly across participants than the alternative arrangements considered.

Duffie and Zhu (2011) use a similar approach to Jackson and Manning (2007) to consider the implications of central clearing for netting efficiency. They find that central clearing could actually increase counterparty risk in OTC derivatives markets if netting sets of multiple products become fragmented across different CCPs. Importantly, they show that using multiple CCPs to clear different products always increases exposure relative to clearing all products through a single CCP.

An implication of higher exposures in any market is that more collateral is required to cover these exposures. This is noted by Duffie and Zhu (2011) and assessed quantitatively by Heller and Vause (2012). They estimate that a prudent CCP with a monopoly on OTC derivative clearing would need US$43 billion of initial margin for clearing interest rate swaps (IRS), and up to US$107 billion for clearing credit default swaps (CDS). The model sheds light on how these values could be affected by market structure: if, for example, three regionally focused CCPs merge, large dealers would be required to post 25 per cent less initial margin. Similar results are obtained if clearing is consolidated across products: if a single IRS CCP merged with a single CDS CCP, initial margin requirements would fall further, by a similar percentage.

While consolidation of clearing appears to be beneficial from a risk management perspective, it is unlikely, at least in the short-to-medium term, that consolidation will take place on the scale modelled in Heller and Vause (2012). Legal and regulatory frameworks differ between jurisdictions, potentially inhibiting large-scale cross-border clearing, while commercial incentives may well support a multi-CCP environment for many products in large jurisdictions.

Interoperability has been proposed as a way of avoiding fragmentation in a multi-CCP environment. Singh (2013) argues that interoperability is a desirable way of proxying a ‘first-best’ solution in which all trades are cleared by a single CCP, because trades in different CCPs can be netted against each other. However, Singh also suggests that inter-CCP exposures would need to be covered through larger default fund contributions; whether interoperability is desirable depends on the relative size of these two effects.

Collateralisation under interoperability is modelled by Mägerle and Nellen (2011). Using a model of margin requirements in a multi-CCP environment, they show that interoperable CCPs would collect the same initial margins as a single, consolidated CCP, but that these margins may not collateralise total system exposure sufficiently if they fail to cover inter-CCP exposure appropriately. This is especially likely if the linked CCPs have different coverage levels for initial margin, or if more than two CCPs are linked.

Anderson, Dion and Pérez Saiz (2013) extend the Duffie and Zhu (2011) model to examine the exposure impact on small jurisdictions of introducing a global CCP. They demonstrate that having a single global CCP clearing all trades is the most efficient from an exposure perspective, and examine an alternative scenario in which there is a domestic and a foreign CCP. In this case they assess the impact on exposure of creating a link between them. Where the number of participants in the domestic CCP is small relative to that of the foreign CCP, their model finds that total system exposure increases. This is because the inter-CCP obligations created by the link outweigh the netting benefits to domestic participants.

The underlying model used in our paper is similar to that of Anderson et al (2013), in that it is a straightforward application of the Duffie and Zhu (2011) model of CCP links. However, we make different assumptions about the nature of participation and exposure that allow us to generalise the results to CCPs of any size. We also consider a number of useful extensions to the model – such as more general distributional assumptions, a different model of participation, and allowing more than two CCPs to be linked – to demonstrate its applicability to a wide variety of scenarios. Similarly to Anderson et al (2013), our focus is on system-wide levels of exposure, and whether linking CCPs in an otherwise fragmented clearing market increases or reduces that exposure.

Footnote

Three CCPs currently operate in Australia: two CCPs in the ASX Group, ASX Clear and ASX Clear (Futures), and the London-based CCP, LCH.Clearnet Ltd. [3]