The Rube Goldberg Approach to Integrating CCPs: A Recipe for Disaster
As noted in earlier posts (and by others commenting on the proposed Eurex-LSE merger) the main potential benefit to exchange customers* is the capital and margin savings from netting efficiencies between Eurex futures and LCH swaps. However, regulators and others have expressed concerns that the downside is the creation of an bigger too big to fail clearing entity. A couple of weeks back Silla Brush and John Detrixhe reported that the merger partners are trying to square that circle by cross-margining, but not merging the CCPs:
LCH.Clearnet and Eurex held 150 billion euros ($169.5 billion) of collateral on behalf of their members as of Dec. 31, according to the merger statement. The London-based clearer is developing a system that allows traders to offset their swap positions at LCH.Clearnet with their futures holdings at Eurex. The project, which works even though the two clearinghouses are separate, should enable customers to reduce the total amount of collateral they must set aside.
“We can cross margin our over-the-counter clearing with their listed derivatives without merging the clearinghouses, and without comingling the risk-management framework,” LSE Group CEO Xavier Rolet said in a Bloomberg Television interview on Wednesday. Rolet will step aside if the companies complete their merger.
The devil will clearly be in the details, and I am skeptical, not to say suspicious. In order for the separate but comingled system to work, Eurex’s CCP must have a claim on collateral held by LCH (and vice versa) so that deficiencies in a defaulter’s margin account on Eurex can be covered by excess at LCH (and vice versa). (As an illustration of the basic concept, Lehman had five different collateral pools at CME Clearing–interest rate, equity, FX, commodities, energy. There were deficiencies in two of these, but CME used collateral from the other three to cover them. As a result there was no hit to the default fund.)
How this will work legally is by no means evident, especially inasmuch as this will be a deal across jurisdictions (which could become even more fraught if Brexit occurs). Further, what happens in the event that one of the separate CCPs itself becomes insolvent? I can imagine a situation (unlikely, but possible) in which CCP A is insolvent due to multiple defaults, but the margin account at A for one of the defaulters has excess funds while its margin account at CCP is deficient. Would it really be possible for B to access the defaulter’s collateral at bankrupt CCP A? Maybe, but I am certain that this question would be answered only after a nasty, and likely protracted, legal battle.
The fact that the CCPs are going to be legally separate entities suggests their default funds will be as well, and that they will be separately capitalized, meaning that the equity of one CCP will not be part of the default waterfall of the other. This increases the odds that one of the CCPs will exhaust its resources and become insolvent. That is, the probability that one of the separate CCPs will become insolvent exceeds the probability that a truly merged one would become so. Since even the separate CCPs would be huge and systemically important, it is not obvious that this is a superior outcome.
I am also mystified by what Rolet meant by “without comingling the risk management framework.” “Risk management framework” involves several pieces. One is the evaluation of market and credit risk, and the determination of the margin on the portfolio. Does Rolet mean that each CCP will make an independent determination of the margin it will charge for the positions held on it, but do so in a way that takes into account the offsetting risks at the position held at the other CCP? Wouldn’t that at least require sharing position information across CCPs? And couldn’t it result in arbitrary and perhaps incoherent determinations of margins if the CCPs use different models? (As a simple example, will the CCPs use different correlation assumptions?) Wouldn’t this have an effect on where firms place their trades? Couldn’t that lead to a perverse competition between the two CCPs?+ It seems much more sensible to have a unified risk model across the CCPs since they are assigning a single margin to a portfolio that includes positions on both CCPs.
Another part of the “risk management framework” is the management of defaulted positions. Separate management of the risk of components of a defaulted portfolio is highly inefficient. Indeed, part of the justification of portfolio margining is that the combined position is less risky, and that some components effectively hedge other components. Managing the risks of the components separately in the event of a default sacrifices these self-hedging features, and increases the amount of trading necessary to manage the risk of the defaulted position. Since this trading may be necessary during periods of low liquidity, economizing on the amount of trading is very beneficial.
In other words, co-mingling risk management is a very good idea if you are going to cross margin.
It seems that Eurex and LSE are attempting to come up with a clever way to work around regulators’ TBTF neuroses. But it is not clear how this workaround will perform in practice. Moreover, it seems to sacrifice many of the benefits of a merged CCP, while creating ambiguities and legal risks. It also will inevitably be more complex than simply merging the two CCPs. Such complexity creates systemic risks.
One way to put this is that if the two CCPs are legally separate entities, under separate managements, relations between them (including the arrangements necessary for cross margining and default management) will be governed by contract. Contracts are inevitably incomplete. There will be unanticipated contingencies, and/or contingencies that are anticipated but not addressed in the contract. When these contingencies occur in practice, there is a potential for conflict, disagreement, and rent seeking.
In the case of CCPs, the relevant contingencies not specified in the contract will most likely occur during a default, and likely during stressed market conditions. This is exactly the wrong time to have a dispute, and failure to come to a speedy resolution of how to deal with the contingency could be systemically catastrophic.
One advantage of ownership/integration is that it mitigates contractual incompleteness problems. Managers/owners have the authority to respond unilaterally to contingencies. As Williamson pointed out long ago, efficient “selective intervention” is problematic, but in the CCP context, the benefits of managerial fiat and selective intervention seem to far outweigh the costs.
I have argued that the need to coordinate during crises was one justification for the integration of trade execution and clearing. The argument applies with even greater force for the integration of CCPs that cooperate in some ways (e.g., through portfolio margining).
In sum, coordination of LCH and Eurex clearing through contract, rather than through merger into a single entity is a highly dubious way of addressing regulators’ concerns about CCPs being TBTF. The separate entities are already TBTF. The probability that one defaults if they are separate is bigger than the probability that the merged entity defaults, and the chaos conditional on default, or the measures necessary to prevent default, probably wouldn’t be that much greater for the merged entity: this means that reducing the probability of default is desirable, rather than reducing the size of the entity conditional on default. Furthermore, the contract between the two entities will inevitably be incomplete, and the gaps will be discovered, and extremely difficult to fill in-, during a crisis. This is exactly when a coordination failure would be most damaging, and when it would be most likely to occur.
Thus, in my view full integration dominates some Rube Goldberg-esque attempt to bolt LCH and Eurex clearing together by contract. The TBTF bridge was crossed long ago, for both CCPs. The complexity and potential for coordination failure between separate but not really organizations joined by contract would create more systemic risks than increasing size would. A coordination failure between two TBTF entities is not a happy thought.
Therefore, if regulators believe that the incremental systemic risk resulting from a full merger of LCH and Eurex clearing outweighs the benefits of the combination, they should torpedo the merger rather than allowing LSE and Eurex to construct some baroque contractual workaround.
*I say customers specifically, because it is not clear that the total benefits (including all affect parties) from cross margining, netting, etc., are positive. This is due to the distributive effects of these measures. They tend to ensure that derivatives counterparties get paid a higher fraction of their claims in the event of a default, but this is because they shift some of the losses to others with claims on the defaulter.