OTC derivatives clearing: how the agency and principal models compare

In 2009 in what was then thought to be the wake of the financial crisis the derivatives industry came under close scrutiny from regulators. A strong message was conveyed to the industry on various political and regulatory fronts: change your ways or they will be changed for you.

Even before the financial crisis the Federal Reserve Bank of New York had expressed concerns about credit default swaps (CDS) and in particular about the then sizeable confirmation backlog and later the amount of outstanding uncompressed notional. Ironically, the industry had begun to address many of those concerns well before the onset of the financial crisis.

Indeed, prior to the enactment of the Dodd-Frank Act, in response to a meeting in April 2009 between major market participants and an assortment of regulators, in June 2009 a group of the largest dealers and buyside participants committed to achieve buyside access to central clearing for certain standardised OTC derivatives by December 15 2009.

Interdealer clearing of interest rate swaps (IRS) had been available through LCH SwapClear as far back as September 1999, and interdealer clearing of CDS index products had commenced on ICE Trust (ICE) in March 2009. Although industry efforts to provide buyside access to central clearing began several months prior to June 2009, it was really after June and throughout the balance of 2009 and early 2010 that significant efforts to execute got underway.

The thinking on the Street at the time was that if the market did not step up to the plate and voluntarily police itself, the regulators would certainly do so. Notwithstanding the industry's significant efforts to do just that, the Dodd-Frank Act was enacted with its panoply of burdensome requirements, including, among others, mandatory execution and clearing for standardised derivatives.

Three major central counterparties (CCPs) sought to offer two divergent models for buyside clearing, the Chicago Mercantile Exchange (CME) and ICE for CDS and LCH.Clearnet (LCH) for IRS. Buyside clearing has existed on CME since 2010 and on ICE and LCH since December 2009. LCH and ICE put forth back-to-back principal models that mimicked many of the features of the US futures agency model, but relied on different facets of the US insolvency regime than did the US futures model.

This was due, at that time, to problems with the relevant regulatory regime governing US futures (including the definition of commodity contract under the US bankruptcy code and the US Commodity Futures Trading Commission's (CFTC) Part 190 rules) as well as certain shortcomings of the US futures model (like loss mutualisation due to fellow customer risk, which we discuss below, as well as operational and investment risk) that made expanding the then existing futures regime to cleared OTC swaps difficult or undesirable.

The CFTC would later address some of the problems with relying on the futures regime for cleared OTC swaps in its May 2010 amendments to its Part 190 rules and which would ultimately be dealt with more reliably and statutorily under the Dodd-Frank Act and related rules.

CME on the other hand sought to duplicate the US futures style agency model limiting its members to futures commission merchants (FCMs), notwithstanding the then uncertain state of the regulatory landscape with respect to cleared OTC derivatives. CME's original model contemplated cleared OTC swaps being credited to what is known as a 30.7 account or a secured account which is supposed to be used for transactions and associated collateral relating to foreign futures and options.

Other than being a somewhat more intelligent place than the 4d account, also known as the segregated account for US futures transactions and associated collateral, there was no particularly good legal basis for this choice. Indeed, it was an inherent weakness in CME's original cleared OTC swaps model not present in the original ICE and LCH models.

Later, CME switched over from this 30.7 account approach to create, under its rules, a special sequestered account class for cleared OTC swaps that largely mirrored the 4d segregated account class applicable to futures (and which pursuant to the amended 190 rules would have been respected in an FCM insolvency).

Post-Dodd-Frank, the only non-retail client clearing model effectively permitted in the US for OTC swaps is the futures-style agency model through an FCM with a derivatives clearing organisation (DCO). Security-based swaps will be subject to a different US clearing regime, the rules of which have yet to be promulgated by the SEC, but pursuant to the Dodd-Frank Act will require standardised security-based swaps to be cleared.

Elsewhere around the world neither model is mandatory, and those looking to establish new CCPs are faced with the choice of which model to adopt or, in some cases, whether to create platforms permitting both.

In the UK, LCH continues to make the principal model available for client clearing (although per Dodd-Frank it is not able to be used for client clearing for US customers), but it has also launched, on a side-by-side basis, an FCM agency model to service US customers as a registered DCO. It remains unclear whether other CCPs using or that might use the principal model outside of the US will also seek to layer on a parallel or side-by-side FCM platform to service US customers.

A variety of factors may ultimately come into play in making such a choice, and in some cases foreign exchanges and associated clearing organisations may seek to provide access to US persons pursuant to an exemption under the CFTC's foreign board of trade rules, but even in that case the foreign board of trade and its related clearing association, if not a DCO, must essentially demonstrate (the final rules set forth an extensive array of criteria) that they are subject to regulation/relevant international standards in their home jurisdictions comparable to what DCMs and DCOs are subject to in the US.

Principal vs agency? Although largely moot in the US, the question remains relevant given the G20 nations' commitment to centralised clearing made at the 2009 G20 Summit in Pittsburgh and reaffirmed at the 2010 G20 Summit in Toronto, which is supposed to be in place by the end of 2012, as well as the nascent state of CCPs in many of the G20 nations. The G20 nations are in various states of development in creating CCPS, with the US, UK, Australia and Japan leading the way.

Many of the Asian nations other than Japan are facing delays in implementing centralised clearing, and the European Markets Infrastructure Regulation also has many political hurdles to overcome. Accordingly, whether the CCPs in these countries will choose the principal model, the agency model or both remains to be seen.

The principal and agency models differ primarily based on the legal capacity in which the clearing member acts in relation to clearing and the nature of the legal and economic relationships between the CCP, the clearing member and the customer. However, regardless of whether a CCP adopts a principal or agency model, in developing a platform for clearing OTC swaps, there are several factors that must be taken into consideration.

Determining the types of entities that can participate as clearing members, financial resource requirements and default processes are all crucial considerations for a successful OTC swaps clearing platform.

One of the most critical areas of concern is ensuring that there are adequate insolvency protections for end-users upon a clearing member's default, including providing adequate legal protection of customer margin and associated positions cleared through the CCP and the ability to move such positions and related margin to another clearing member if the customer's original clearing member defaults—this is often referred to as portability.

Setting parameters around documentation, margining and the holding and use of collateral in moving from a bilateral to a centralised clearing system must also be addressed. A robust documentation system clearly expressing each party's rights and obligations under the transactions is also vital to a successfully operated clearing system. Each of the principal and agency models have addressed these and a variety of other matters in different, often divergent, ways, as discussed below in more detail.

Key differences between models

It should be apparent that there are some significant differences between the principal and agency models notwithstanding a great number of similarities as well. Although many subtleties and nuances might be identified, there are a number of fundamental points of divergence worth focusing on: the legal capacity in which clearing members act and the basis upon which the customer's position can be said to be cleared. the principal model is based largely on contractual relationships among the parties that determine how clearing is effected, margin is protected and portability is achieved, while the agency model is driven more by statutory and regulatory legal requirements governing the clearing process, which are then complemented or implemented by the rules of the CCP and to a far lesser extent the documentation governing the relationship between the FCM and the customer. as a function of the preceding point, the documentation is very different. The principal model uses Isda style documentation to dictate the relationship between the customer and the clearing member, while the agency model relies on the Customer Agreement and addendum concept. under the agency model, collateral arrangements, including what qualifies as eligible collateral and how such collateral is held and segregated, is largely determined by law and regulation and to a lesser extent CCP rules. Conversely, the principal model provides far greater flexibility for bespoke collateral arrangements, including significantly more flexibility on rehypothecation and the types of collateral the customer can provide to the clearing member as well as in respect of margin financing. post-default netting and settlement as well as portability are achieved differently in the agency and principal models. The principal model relies on a series of contractual arrangements and security interests as well as the relevant CCP rules. The agency model, at least in the US, relies on a statutory and regulatory regime that is only anecdotally interested in the contractual relationships between the parties and is very much designed to facilitate portability.

Under the agency model, the clearing member acts as agent and guarantor of the customer such that, at least in theory, the customer faces the CCP through its agent, the FCM, which then guarantees the customer's performance to the CCP. In many jurisdictions this idea is met with skepticism because market participants find it hard to see how it is an agency relationship when the FCM is guaranteeing the customer's performance to the CCP and when the customer has no real direct contractual remedies against the CCP.

It is, however, typically characterised as such in the US, not only in the vernacular of market participants but also omnipresent throughout the CFTC's regulations and the clearinghouses' rules, which collectively form the basis for the agency model.

For example, Rule 39.12(b)(6) promulgated by the CFTC in its final rulemaking codifying DCO core principles provides "a derivatives clearing organisation that clears swaps shall have rules providing that, upon acceptance of a swap by the derivatives clearing organisation for clearing … the original swap is replaced by an equal and opposite swap between the derivatives clearing organisation and each clearing member acting as principal for a house trade or acting as agent for a customer trade." This agency characterisation is important, because it has significant regulatory capital benefits for FCMs.

The agency model used in the US is referred to as the FCM model, because it assumes that the clearing members will be FCMs registered with the CFTC. Those clearing futures or OTC swaps for customers in the US must be FCMs. The Dodd-Frank Act includes many amendments and reforms to the CEA, including the addition of a requirement that any clearing organisation, wherever located, which clears swaps for participants located in the U.S. must be registered with the CFTC as a DCO. In addition, the CEA as modified by the Dodd-Frank Act effectively now requires any clearing member clearing swaps on behalf of US participants to be registered as an FCM.

Under the US futures model customer property must be segregated at the FCM and DCO levels from the FCM's and DCO's assets and must be treated as customer property. Operational co-mingling of such segregated amounts by account class in omnibus accounts at the FCM and DCO levels was permitted for administrative convenience.

Historically, the CFTC's rules allowed reinvestment of such segregated funds in a list of permitted investments pursuant to Rule 1.25, which included, among other things, interaffiliate repos. The CFTC recently adopted a final rule amending Rule 1.25 and narrowing significantly what types of investments are permissible for customer funds.

In an FCM insolvency resulting in a deficiency in such customer account class, the loss or deficiency is allocated pro rata across all customers in that account class pursuant to the CFTC's Part 190 rules. Thus, for example, in the case of MF Global, if it turns out ultimately that there was a deficiency in segregated funds relating to US futures, all of the customers in the 4d account class will share in such shortfall ratably.

The CFTC recently adopted final rules for cleared OTC swaps implementing the legal segregation with operational comingling (LSOC) model, also referred to as the full legal segregation model, pursuant to Section 4d(f) of the CEA, as amended by the Dodd-Frank Act. Segregation at both the FCM and DCO levels is required just like the futures model. FCMs are also required to prevent DCOs from using, and DCOs are prohibited from using, a swap customer's property as collateral for another swap customer's swap contracts.

Under LSOC, in a clearing member default caused by a default of a customer of that clearing member, a DCO cannot apply the property of non-defaulting swap customers of the defaulting FCM to satisfy such a deficiency, but rather must look only to the property of the defaulting customer and other available financial resources (for example, assets of the defaulting FCM, its own equity, the guaranty fund or unfunded assessments).

This is a major change from the futures segregation model which doesn't distinguish between defaulting and non-defaulting customers in an FCM default scenario. LSOC imposes additional information reporting requirements on FCMs to ensure DCOs have the information they need for proper allocation. LSOC increases the clearinghouse's risk, because it reduces the available amount of financial resources, and as a result, LSOC may increase swap clearing costs due to higher margin, guaranty fund and assessment requirements imposed by DCOs stripped of access to non-defaulting swap customers' margin than is the case under the current US futures agency model.

LSOC reduces but does not completely eliminate the fellow customer risk that exists in the US futures agency model. LSOC also does not address investment risk or operational risk and thus does not solve the MF Global problem.

The relationship between a customer and its FCM along with the creation of a customer account at the FCM level is typically documented with some form of futures account agreement (customer agreement) between the customer and the FCM, which is bounded both by applicable law and the applicable CCP rules (such customer agreements not being CCP specific).

In its original model, CME released documentation to expand the customer agreement to cleared OTC CDS in the form of a CME-specific addendum (the CME addendum) that was intended to be able to be appended to any customer agreement used by an FCM (there is no standardised form of customer agreement, but they all tend to look quite similar). The concept of the CME addendum made a lot of sense, but unfortunately the document itself was a bit difficult to work with.

Nevertheless, with the advent of Dodd-Frank and the mandatory move to the agency model across the board in the US, the Futures Industry Association (FIA) began work on a platform and product agnostic addendum (the FIA addendum) that like the CME addendum was intended to expand an existing customer agreement for use with cleared OTC swaps.

Unlike the CME addendum, however, the FIA addendum was intended to be a simple and easy to understand document that would work across all products and CCPs using the US agency model. Unfortunately, the FIA addendum has yet to be officially published, and there are a number of different versions floating around the marketplace that some FCMs have signed-up with some clients, including bespoke changes.

Meanwhile, Isda and the FIA did publish an execution agreement (the FIA/Isda execution agreement), which is basically a fancy form of give-up agreement for use in the cleared OTC derivatives space, which was the product of significant input from a wide variety of market participants. However, the CFTC responded very negatively to the FIA/Isda execution agreement at the behest of some vocal buyside participants who argued that the voluntary annexes which contemplated making FCMs party to the agreement would impede open access and even though characterised as voluntary would be forced upon the market, effectively becoming mandatory.

The relationship between the CCP and its clearing members as well as their customers (such as it is) is established pursuant to CCP rules and some form of documentation entered into between the CCP and its clearing members. CCP rules tend to be vague and difficult to understand, which leads many market participants to struggle with their interpretation.

This is partly a function of history and partly a function of the fact that such rules have always been drafted to provide significant discretion to the CCP. With the advent of a revamped set of DCO core principles and related rules in the US, we expect to see at least some clarifications being made as rules are revised to comport with the new core principles and related rules.

The principal model is much more like the traditional OTC bilateral swaps market than the agency model is in that the customer faces the clearing member as principal and the clearing member in turn faces the CCP as principal. This in effect creates identical back-to-back trades, but, for the most part, without ever establishing any principal or other relationship between the clearinghouse and the customer.

Rather than being rules and regulations-based, the principal model is largely contractual in nature with the customer to clearing member contract having various links to the happenings under the clearing member to CCP contract (the latter usually being created and existing under the CCP's rules).

The portability and margin protection features that typify the agency model are achieved in the principal model through the use of a security package that, in addition to the typical posting of initial and variation margin, includes the grant of a security interest by the clearing member to each of its customers in its right to the return of collateral, or the corresponding receivable, from the CCP (this can be phrased in different ways with slight or significant variations, but both the LCH model and the original ICE model make use of such a mechanism).

Documentation for the principal model can be implemented in a number of different ways. The original ICE offering and the LCH model both effectively operate to duplicate the salient parts of an existing Isda master agreement (non-clearing master) between the parties that is used for uncleared transactions with a variety of amendments and supplements that will govern the cleared transactions between them (clearing master).

Both the non-clearing master and the clearing master are then effectively subject to umbrella terms that govern the relationship between the clearing member and the customer regarding both agreements. The cleared trades and the uncleared trades are completely separate, although there will be a mechanism to allow for the separate termination of the uncleared trades and the cleared trades on a clearing member default which is necessary to allow close-out netting and ultimately to achieve portability.

The relationship between the CCP and the clearing member will be pursuant to one or more agreements, but for the most part will be dictated by the parameters of the relevant CCP's rules, including product-specific provisions set forth therein. Without diving too deeply into the weeds, the net effect is that the customer gets the benefit of clearing through the CCP without actually facing the CCP. Following a clearing member default there may be certain limited interaction between the customer and the CCP.

As noted above, the question of principal versus agency in the US is largely settled, but in the rest of the world it remains a very live issue. The hope of global harmonisation seems more like a utopian fantasy than anything capable of actualisation at this juncture.

Nevertheless, we remain optimistic that in a few years time only a very few aspects of centralised clearing, if any, will be considered novel and that while distinctions may exist at least some harmonisation will be achieved.

By partner Stan Renas, senior associate Noah Melnick and associate Chris Davis of Linklaters in New York. The views are of the authors only and do not represent the views of Linklaters or of any other party