Contracts as regulation: the ISDA Master Agreement

This article proposes a rule of contractual interpretation for regulatory contracts defined as contracts (i) used by a large number of market participants, (ii) subject to limitations on deviation and (iii) designed with market problems (such as negative externalities) in addition to transactional problems (such as transaction costs) in mind. The rule states that the outcome of the interpretation of regulatory contracts must not be inconsistent with the objectives of the regulatory framework applicable to the market in which the contract is used. The article suggests that the reliance on that rule is normatively justified in cases where the regulatory objectives are clearly defined and particularly when the alternative outcome is to void the contract or its provision. By adopting the regulatory rule for contract interpretation, courts can preserve the autonomy of private regulatory regimes created through contracts without sacrificing public policy objectives. Several examples of

There are good legal and economic reasons for courts to enforce regulatory contracts per the plain meaning of their terms as long as that meaning is not inconsistent with the objectives of the applicable regulatory framework. However, when the plain meaning of a regulatory contract is inconsistent with those objectives, courts should instead seek to align the contract's meaning with those objectives through regulatory interpretation. Regulatory interpretation of contracts must be based on clearly defined regulatory objectives and a clear understanding of how a more conventional interpretation would undermine the achievement of those objectives. While the regulatory interpretation of a contract reduces the flexibility afforded by the plain meaning interpretation, it can ultimately increase the efficiency and safety of markets and benefit market participants. It can also promote legal and market certainty in particular when the alternative is the invalidation of a contract or its provision-an outcome that should be avoided given the legal and market uncertainty it would create.

Introduction
The master agreement (MA) developed by the International Swaps and Derivatives Association (ISDA) is a standardized or boilerplate contract commonly used by participants in the $544 trillion 1 over the counter (OTC) derivative securities market. The MA's influential role in this sizeable market helps explain why the MA has, over the years, attracted a considerable amount of scholarly scrutiny in the literature on both contracts and regulation. A cursory review of the literature on the MA reveals that the normative evaluation of the economic effects of the MA by scholars of contracts and regulation was different. Contract scholars identified various economic benefits of the MA, such as a reduction of transaction costs and various positive externalities. Positive externalities come in two main forms: learning and network externalities. 2 Learning externalities arise from a historically established use of contracts and contribute to drafting efficiencies, reduced uncertainty over the (judicial) meaning of contract terms, and familiarity with terms among users. Network externalities arise from the widespread use of a contact form and tend to boost some of the learning externalities identified above.
In contrast to contract scholars who zeroed in on the economic benefits, regulatory scholars identified various problems with the MA and its influential role in the markets, such as negative externalities. Negative externalities can be understood as costs suffered by third parties as a result of transactions. Transactions governed by the MA could create costs for third parties, such as increased systemic risk 3 or created transactional precedents that could affect market integrity. 4 Such third-party effects would create apparent issues of regulatory legitimacy of the MA. 5 Regulatory scholars, like regulators, tend to look for risks in markets and contracts. The professional focus of regulatory scholars can account for the difference in their normative evaluation of the MA's economic effects compared to contract scholars. But the difference can also be accounted for at a more analytical level-contract scholars and scholars of regulation focus on different units of analysis when examining the MA's impact. Contract scholars focus primarily on the economic impact of the MA on transactions, whereas regulatory scholars focus on the economic impact of the MA on markets. While markets can be understood as aggregations of transactions, market problems, such as systemic risk or market integrity, are different from transactional problems, such as transaction costs. The MA addresses both types of problems and, in that respect, is a unique type of contract. This article seeks to capture that uniqueness by identifying the MA's regulatory properties and examining their implications for adjudication of contracts with such properties.
The three elements of contracts identified in this article that are jointly constitutive of their regulatory properties are as follows: 1. a large number of market participants uses the contract; 2. the users of the contract are constrained in their ability to deviate from it; and 3. the contract addresses market problems (such as negative externalities) besides transactional problems (such as transaction costs). Of course, it is unusual for a contract to address market problems by design because contracts are designed primarily by lawyers. Lawyers are paid to optimize transactions, not markets. There is no reason to doubt that the lawyers who conceived of the MA in the 1980s viewed it as an instrument addressing transactional problems, such as inconsistencies across the different contract forms used by swaps dealers. However, that has changed over time as MA grew to prominence in the market. Its design increasingly incorporated regulatory elements, such as the various information obligations under the DoddFrank Act. 6 As Robert Pickel, the former CEO of ISDA noted in a series of articles published in the Capital Markets Law Journal, the Global Financial Crisis (GFC) cemented the shift from the transactional to a regulatory paradigm. 7 Nevertheless, the normative implications of that shift remain unclear mainly because of a lack of clarity over how courts (and other bodies) should interpret regulatory contracts such as ISDA MA. There are good legal and economic reasons for courts to enforce regulatory contracts following the plain meaning of their terms as long as that meaning is not inconsistent with the objectives of the applicable regulatory framework. However, I submit that when the plain meaning of a regulatory contract is inconsistent with those objectives, courts should instead seek to align the meaning of the contract with those objectives through regulatory interpretation. Regulatory interpretation of contracts must be based on clearly defined regulatory objectives and a clear understanding of how a more conventional, plain meaning interpretation would undermine the achievement of those objectives.
The recent saga of manufactured defaults is an excellent example of a situation in which a lack of clarity over how courts (and other) bodies should interpret regulatory contracts led to the creation of a regulatory gap and undermined the achievement of the objectives of the regulatory framework applicable to OTC derivatives markets. Until recently, manufactured defaults were (and to some extent still are) technically allowed under ISDA's documentation. While ISDA has recently taken steps to address the problem, those steps may prove insufficient because of a growing concern among regulators about the impact of the practice of manufactured defaults on market integrity. As a result, courts may, in the future, be faced with the question of whether the enforcement of terms undermining OTC derivatives markets' integrity is a desirable outcome from the point of view of public policy.
To single-out the manufactured defaults sage is not to say that ISDA and the regulators worked at cross-purposes on this or other matters. Over the years, essential complementarities have emerged between the objectives pursued by ISDA and those of public regulators. The MA is published by ISDA and can be analysed in the context of ISDA's objectives, but those objectives often overlap with the objectives of public regulators. This article's claim is simply that the interpretation of regulatory contracts must occur against the backdrop of the synergies or complementarities that can be achieved between private and public regulation.
In this article, I examine the emergence of such complementarities between the ISDA MA and the public regulation of OTC derivatives markets in the area of systemic risk and market integrity. The examination suggests that while regulators are best positioned to facilitate the emergence of complementarities between private regulation through contract and public regulation, courts can sometimes play a role in that respect. 8 Courts can play that role by ensuring that the plain meaning of a regulatory contract is not inconsistent with the objective of the applicable regulatory framework. When the plain meaning of the regulatory contract is inconsistent with those objectives, courts should seek to align the contract's meaning with those objectives through regulatory interpretation. I call this the regulatory rule of contract interpretation.
The regulatory rule can be seen as a special case of the plain meaning or textual rule of contractual interpretation identified by contract scholars as the most appropriate for financial boilerplate. 9 Contract scholars argue that the plain meaning rule is the most appropriate for financial boilerplate because it helps courts to strike the right balance between the competing normative considerations of party autonomy and legal certainty. 10 8 For a similar argument in the context of contracts used in multilateral consumer markets, see RE Scott, 'The Paradox of Contracting in Markets' (2020) 100 L & Contemp Problems. Scott argues that the use of contracts in multilateral markets calls for the abandonment of the bilateral contract paradigm. He suggests contracts scholars should instead focus on ways that the state can facilitate the formation of a regulatory network that improves the efficiency of standardized contract terms in multilateral markets. 9 Throughout this article I will use the terms 'plain meaning interpretation' and 'textual interpretation' interchangeably. 10 The literature referred to in (n 2) articulates that view.
In this article, I argue that when a contract performs a regulatory function, the competing normative considerations are different-they are the autonomy of the regulatory regime established through the contract, on the one hand, and public policy, on the other. By balancing these considerations, the courts address an important issue of political economy, namely who should regulate what and how.
In limited circumstances, courts may be well-positioned to address this issue and seek to resolve it by adopting the regulator rule of contract interpretation. As suggested earlier, the reliance on that rule is normatively justified in cases where the regulatory objectives are clearly defined, and courts also have a clear understanding of how a more conventional, plain meaning interpretation would undermine the achievement of those objectives. While the adoption of the regulatory rule of contract interpretation will reduce the flexibility afforded by a plain meaning interpretation, it can ultimately increase the efficiency and safety of markets and benefit market participants. It can also promote legal certainty particularly when the alternative outcome is to void the contract or its provision.
In the remainder of this article: Section 2 develops a conceptual account of the ISDA MA as a regulatory contract by linking a description of the contractual architecture of OTC derivatives markets with a discussion of the structural and functional features commonly associated with regulation; Section 3 discusses the evolution of the structural and functional features of the ISDA MA; and Section 4 discusses examples of cases involving the interpretation of the provision of the ISDA MA and the ISDA Credit Definitions against the backdrop of the applicable regulatory framework.

Regulation through contract in OTC derivatives markets
The relational and regulatory dimension of financial boilerplate Market participants who seek to enter into an OTC derivative trade for the first time commonly 'execute an ISDA'. The 'ISDA' they have in mind is, of course, the ISDA MA. 11 However, while the MA is the central element of the contractual architecture of OTC derivatives markets designed by ISDA (the organization), it is only one of many elements of that architecture. The other elements include (but are not limited to): 14 The perception of increased counterparty risk prompted market participants to start collateralizing their transactions. To facilitate the process, ISDA developed credit support documents under both New York and English law, each reflecting the specific features related to provision of security in the particular jurisdiction.

amendment protocols.
Legal scholars sometimes use the term 'modular' to refer to a contractual architecture comprising a multiplicity of elements. 15 In a modular contractual architecture, each element of the architecture can be designed to perform a different function. For example, one part may be designed to address legal risk and another part economic risks. The allocation of different functions to different elements of the contractual architecture, allows the parties to manage transactional complexity efficiently.
Consider two parties who want to enter into a credit default swap (CDS). The transaction will be governed by the 2014 CDs, which, in turn, will be governed by the MA. The lawyers for the parties can negotiate in the Schedule any desirable departures from the MA's pre-printed form. But they do not have to negotiate the MA from scratch. The same MA can then be used by the parties to govern future CDS entered into between the parties. Traders will only have to execute a confirmation incorporating the CDS, and there is no risk that they will exceed their mandate by changing the legal terms of the deal.
We can immediately see that in ISDA's case, modularity is a cost-effective response to complexity. 16 Market participants frequently enter into many highly customized OTC derivatives trades, which raises the issue of identifying cost-effective ways of dealing with the vast web of complex transactions. The cost-effectiveness of the contractual architecture of OTC derivatives markets revolving around ISDA's documentation helps explain why it is so commonly relied on by market participants.

The structure of a regulatory contract
The crucial structural feature of the modularity of ISDA's contracts is that certain elements of a modular contract are more difficult or costly to adapt than other elements. At a basic level, we can distinguish between elements that can be adapted easily and elements that cannot, at least not without the adaptation being considered 'off-market' and, therefore, costly to make. We can refer to the former as relational and the latter as regulatory elements.
Confirmations, credit support documentation, and provisions of the MA that can be modified through the Schedule are relational in the sense that they can be easily adapted. In contrast, the provisions of the MA that cannot be easily modified through the Schedule, CDs and the amendment protocols are regulatory in the sense that they cannot be easily adapted. Table 1 summarizes the elements of the contractual architecture of OTC derivatives trades with relational (left column) and regulatory features (right column).
The distinction between the relational and regulatory provisions is sometimes lost when legal scholars refer to the MA using the generic term of 'boilerplate'. Jeffrey Golden, one of the authors of the early version of the MA, recently made that point in an article published in the Capital Markets Law Journal and encouraged the use of the term 'relational boilerplate'. 17 The use of the term is meant to suggest that the parties can easily adapt the MA within specific parameters, but not as easily beyond them.
However, while the term 'relational boilerplate' may be well suited to describe terms that can be easily modified, it may be inaccurate to describe the terms that cannot be easily modified. I submit that the term 'regulatory' is better suited for that purpose because it emphasizes (i) that there is a limit to deviation from those terms and (ii) that limit is set not by the party with greater bargaining power, but rather by the market, with the help of ISDA as the preeminent industry association in OTC derivatives markets.
There are several ways in which ISDA can help impose such limits. First, ISDA uses its persuasive authority to encourage the users to adhere to the rules and specific meaning of those rules. As the preeminent global industry association for OTC derivatives markets, ISDA enjoys a considerable level of authority, and there is some evidence that market participants tend to follow ISDA's recommendations. The widespread adoption of ISDA's protocols is an excellent example of market participants' willingness to follow ISDA's recommendation concerning certain amendments of ISDA's documentation.
The legal mechanics of protocol implementation is a second and an even more striking example of how ISDA can help the market impose limitations on the ability of the users of its documentation to deviate from the MA. While market participants are free to decide not to adhere to the amendment protocols, their failure to do that could limit the scope of counterparties willing to trade with them on alternative terms or, at the very least, increase the cost of such trades. Hence, the use of protocols further strengthens the case for the conceptualization of the contractual architecture developed by ISDA as a de facto regulatory structure.
The third element constitutive of ISDA's contractual architecture's regulatory structure is the role of courts in the architecture of enforcement of the MA. Indeed, one of the principal objections that could be raised against a conception of regulatory contracts is that CDs Amendment protocols Credit support documentation they do not have an inherent enforcement mechanism. Neither the distinction between regulatory and relational terms nor the amendment protocol mechanism gives ISDA a role in enforcement per se even though ISDA members have the real capability to determine the meaning of certain of its terms through the Credit Determinations Committees (CDCs). 18 ISDA has, in recent years, created an arbitration mechanism as well.
That view, however, disregards the role of courts. Courts are the critical element of enforcement of any regulation, whether public or private. 19 Contract scholars recognize that implicitly when discussing the MA. They argue that courts should follow the plain meaning of the contract's language and only give the intention of the parties a secondary consideration. 20 That is because the language of the contract is believed to capture the expectations of the market accurately. If these expectations are disappointed, the argument goes, market participants will lose the incentives to rely on boilerplate. As a result, the beneficial economic properties of boilerplate, such as transaction cost reductions and various positive externalities, could be lost.
If courts were to follow the plain meaning rule of contract interpretation, the users' ability to deviate from the contract would be limited. There is evidence that courts do apply that rule, thereby enforcing such limitations. 21 Given these limitations, financial boilerplate does operate like a statute (using the terminology derived from the influential article Contracts as Statue by Stephen Choi and Mitu Gulati), or, consistently, with the terminology used in this article, regulation. In other words, the case for the plain meaning or textual interpretation of financial boilerplate amounts to an explicit recognition that contracts can have the structure of regulation.

The function of a regulatory contract
Contracts can also assume the function of regulation whenever they seek to address a problem that is not merely a transactional problem (such as transaction costs), but also a market problem (such as negative externalities). Contracts are frequently said to govern transactions whenever they seek to optimize transactions, but they can also be said to regulate markets whenever they seek to optimize markets. Contracts are more likely to optimize markets when industry associations design them because of the potentially broader functional outlook of industry associations as compared with the outlook of, say, law firms. The design of the MA incorporates several provisions, which determine not only the economic properties of transactions but also the efficiency and safety of the OTC derivatives markets. The first that comes to mind is, of course, the close-out netting provision. 23 The problem the provision was designed to address was systemic risk. Systemic risk is commonly defined as the risk of adverse consequences of an institutional default for the financial system as a whole. 24 Since OTC derivatives markets are dominated by marketmaking dealers, the risk that such an institution's failure will have systemic consequences was deemed high.
The close-out netting provision sought to address this problem by allowing the nondefaulting counterparties to terminate their contracts with a defaulting institution. The enforceability of the provision relied on exemptions from bankruptcy law for OTC derivatives market participants. Legislators worldwide have granted such exemptions to OTC derivatives market participants based on the systemic risk argument. 25 The second example of a provision with implications for optimization of the OTC derivatives markets that I will discuss in this article is the self-interest provision in CDS. 26 The provision is a less obvious choice for such discussion than close-out netting because it was not designed with market problems in mind. The provision explicitly allows the parties to engage in transactions involving the underlying referenced asset, including 'any action which might constitute or give rise to a Credit Event'. Nevertheless, the self-interest provision's regulatory function became apparent in recent years as the opportunistic practice of manufactured defaults attracted regulatory scrutiny. At that point, market participants and regulators alike called for a change to the design of the self-interest provision to preserve the integrity of the markets. 27 A cursory overview of the two regulatory provisions of ISDA's documentation identified above shows that contracts' regulatory function is effectively enabled by private lawbankruptcy and contract law-which render ISDA's documentation enforceable. In this way, regulation through contract can exist even in the absence of public regulation or in cases where the scope of applicability of public regulation is not straightforward for legal reasons. This is why the interpretation of regulatory contracts, such as the MA, can play an important role in facilitating the achievement of the objectives of the regulatory framework applicable to the markets in which the MA is used as a regulatory instrument. The key issue, of course, is to understand what those objectives are. ISDA may have its objectives, and public regulators may have their own. Furthermore, there could be various groups of public regulators who may have an interest in the regulatory outcome. While the working assumption should be that those objectives are not inconsistent, 28 inconsistencies could arise, as they have in the context of manufactured defaults. In the process of contractual interpretation, courts may sometimes be called to resolve those inconsistencies and they will exercise a considerable amount of discretion in doing so.
In this context, contractual interpretation may entail choices that are informed not only by purely economic considerations but also by considerations of the political economy of regulation-who should regulate what and how? Suppose ISDA, through its documentation, seeks to improve the efficiency of transactions and steer the users of the contract towards specific behaviour. In that case, this can have distributive implications, for example, in terms of the allocation of risks. 29 Suppose further the allocation of those risks seems, in aggregate, problematic from a public policy standpoint. In that case, it is only appropriate for a court to point that out and even solve that problem to the extent legally possible.

Regulatory contracts: a conceptual framework
To summarize, a contract should be viewed as regulatory when (i) a large number of market participants uses that contract to document their transactions, (ii) their ability to deviate from the contract is limited and (iii) the contract itself (or parts thereof) is designed with market problems (such as negative externalities) rather than transactional problems (such as transaction costs) in mind. While the first element is a condition necessary for a contract to regulate a market, as opposed to transactions, the second and third conditions are generally consistent with the economic understanding of the regulatory, respectively, structure and function of regulation. In the next section, I discuss the exact parameters of the MA as a regulatory contract.

ISDA MA as a regulatory contract
Regulatory contracts: structural features The principal structural feature of a regulatory contract is that the contract users are constrained or limited in their ability to deviate from it. The previous section identified three principal institutional sources of such limitations: 1. the regulatory and relational features of the contractual architecture, 28 ISDA's marketing slogan is 'safe, efficient markets', which would be aligned with the objectives of OTC derivatives market regulators. ISDA's stated objective 'to promote the development and maintenance of sound risk management' also points to an overlap with the objectives of public regulation. ISDA, Bylaws, art II, clause (f). 29 Anna Gelpern was to my knowledge the first to make that point. As she noted, 'contract scholars often ask what trade groups do for contracts-how they improve drafting, respond to legal shocks, and resolve interpretation disputes. Reflecting back on the role that contracts play in constituting and motivating industry groups is different. This line of inquiry goes to group incentives, decision-making, legitimacy, and authority, and in turn helps contextualize the form and content of the contract.' Gelpern (n 16).
2. the multilateral amendment mechanism of the contract, and 3. the design of the architecture of enforcement of the contract.
Market participants could, in principle, seek to overcome those limitations by not using the MA. But while this is technically possible, reliance on the MA is often a de facto precondition of market access, particularly in the context of access to OTC derivatives markets in the USA and the UK. The OTC derivatives markets in the USA and the UK are made by broker-dealers who are nowadays unlikely to enter into a transaction without the MA being executed. 30 If an OTC derivatives trade is entered into in connection with a capital markets financing, the documentation governing the capital markets portion of the deal also typically stipulates that the OTC derivatives trade must be on the terms of the MA. Thus, in practice, market participants have little choice other than to go with the MA. What are the consequences of that choice for the definition of their legal rights?

Contract design
As noted above, some aspects of the contractual architecture developed by ISDA are regulatory in the sense that they are of a 'one-size-fits-all' type and cannot be easily adapted or customized by individual users. Other elements are relational in the sense that they can. The emergence of the regulatory and relational features of the contractual architecture can be traced back to the 1987 Interest Rate and Currency Exchange Agreement, the predecessor of the MA, which introduced a structural separation of the regulatory provisions and the relational provisions by publishing a pre-printed form of the MA and, separately, defining the scope of provisions of the MA that could be changed in the Schedule. 31 By doing that, ISDA effectively designated a set of mandatory provisions that cannot be easily deviated from pertaining primarily to counterparty credit risk.
In contrast, the provisions that can be easily deviated from operate as defaults-they can be deviated from but only if the parties have made their alternative choices explicit; otherwise, they would apply without change. This is reflected, for example, in the language of a form of Confirmation describing individual transactions, which states that '[it] evidences a complete and binding agreement' and, further, that until the MA is executed the Confirmation 'shall supplement, form part of, and be subject to, an agreement in the form of the 1992 MA'. Thus, even if the MA had not been negotiated at the time that parties have entered into the transaction, the pre-printed form of the MA would govern any and all transactions between the parties. 32 30 While this would happen on occasion prior to the GFC when many transactions were concluded without the MA in reliance on the so-called long-form Confirmations, nowadays, as Paul Harding notes 'many banks will not trade with a counterparty, particularly an unrated one, unless they have signed the MA first. This, of course, reduced ISDA backlogs where a deal has already been done but tends to infuriate traders who cannot understand why it should take so long to negotiate a MA' Harding (n 12). 31 As Harding notes, many banks compose a general Schedule and base variants for specific entities (eg corporates, building societies, pension funds trustees and hedge funds) upon it. ibid at 421. The changes can be substantial and many banks have their own extensive schedules for different types of products or clients. Finally, the MA's contractual architecture also includes some optional provisions, which will apply only if the parties specifically include them in the MA. Optional provisions are also relational and can be included in the MA as appropriate. The 'Additional Representations' incorporated into the Schedule are an example of an optional provision.

Amendment protocols
The MA's multilateral amendment mechanism is the second institutional source of limitation on the users' ability to deviate from the MA. While historically, ISDA introduced significant amendments through a new version of the MA, in the last two decades, ISDA developed a multilateral mechanism to amend the documentation-the protocol mechanism. 33 As ISDA notes, '[t]he benefit to an adhering party to a protocol is that it eliminates the necessity for costly and time-consuming bilateral negotiations'. 34 Instead, market participants are asked to sign on to a multilateral contract under which all of their past and future agreements with counterparties who have also adhered to the relevant protocol will be affected. 'Rather than bilaterally agreeing to a set of amendments (the combination of which will be specific to the client), clients will adhere to an ISDA protocol, agreeing to contractual amendments published by ISDA and elected on the system.' 35 As such, the protocols establish a contractual link between virtually all market participants. But again, that is only true of the market participants who have adhered to the relevant protocols. [m]arket participants who adhered to an ISDA protocol in recent years are familiar with a process that involved submitting signed and conformed copies of an adherence letter to a designated email address. A new process was established in August 2012, when ISDA, in an effort to provide a more streamlined and efficient method, developed a technical solution to further automate adherence. The adhering party will still need to provide a letter signed by an authorized signatory in order to validate the adherence, but the new process will make this easier and will allow that adhering party to monitor the status of that adherence from the submission stage to the approval stage. 'Unlike with previous ISDA protocols where amendments were effected solely with delivery of an adherence letter by each party to the underlying document to be amended (i.e., a master agreement), the DF protocol included additional bilateral delivery requirements in order to effectuate the amendments . . Each party that submits an Adherence Letter must also deliver a completed Protocol Questionnaire to each relevant counterparty for the amendments to be effective. As a result of these additional bilateral delivery requirements, ISDA together with Markit have developed a technology-based solution to automate the information-gathering process and provide sharing of submitted data and documents to permissioned counterparties.' 36 Anna Gelpern and Mitu Gulat discuss one of the more interesting cases, in which ISDA has made such a determination-the case of the Greek sovereign debt restructuring in the early 2010s. A Gelpern & M Gulati, 'CDS Zombies' (2012) 13 Eur Busi Org LR 347. from the ISDA regime. As Peter Werner, a senior legal counsel at ISDA noted, arguably, this is because ISDA emerged from the sell-side, and there was no need for it. As it started to encompass a broader membership base issues started to arise. 37 In 2013, ISDA published model arbitration clauses for use with the MA. 38 Neither the CDCs nor arbitration are per se enforcement mechanisms, but the MA can be enforced through courts. ISDA seeks to shape the enforcement outcomes either by lobbying for changes in the law that would ensure those outcomes or intervening as amicus curiae in significant litigation. 39 There is evidence that courts attach considerable weight to the arguments made through ISDA's interventions. 40 These non-judicial and judicial mechanisms of enforcement constitute a third source of institutional limitations on the ability of the users of the MA to deviate. Together with the design of regulatory and relational terms and the multilateral amendment mechanism, these three sources of institutional limitations on the ability of the users of the MA to deviate are jointly constitutive of the regulatory structure of the MA. In other words, they speak to the institutional limits of market participants to deviate from the rules set out by ISDA.

Regulatory contracts: functional features
We can now turn to the principal functional feature of regulatory contracts. The conception of regulatory contracts advanced in this article envisages that regulatory contracts will be designed with market problems (such as negative externalities) rather than transactional problems (such as transaction costs) in mind. While transaction cost reduction is an important design principle that appears to have informed ISDA's efforts over the years, it is apparent that in some cases, the design of the MA creates solutions to both transactional problems and market problems. For example, the close-out netting provisions of the MA provide a solution to the transactional problem of counterparty risk, but also the market problem of systemic risk. Furthermore, the self-interest provision of the CDs is a solution to the transactional problem of conflicts of interest, but also to the market problem of market integrity.

Systemic risk
I have discussed the substance of the close-out netting above. I have also noted that exemptions from automatic stays in bankruptcy law were required for the close-out netting provisions to work. As discussed above, those exemptions were granted based on the recognition that the failure of a large derivatives market participant could have systemic effects. The combination of close-out netting and the exemptions from automatic stay created the regulatory function of the MA-systemic risk regulation.
Over the years, several prominent bankruptcy scholars questioned the validity of the systemic risk argument in favour of the exemptions and (indirectly) the MA's role in systemic risk regulation. 41 First, they argued that close-out netting benefits OTC derivatives users at the expense of other creditors even without any systemic risk threats. Secondly, the benefits of close-out netting may attract market participants to OTC derivatives trading, leading to a concentration of financial institutions on that can kind of activity and thereby increase systemic risk. Thirdly, close-out netting can encourage runs on financial institutions that find themselves in financial difficulties.
The GFC validated these counterarguments, at least to some extent. 42 For example, in the years leading up to the crisis, the American Insurance Group (AIG) became a significant player in the market for CDS and faced a run by its most significant CDS counterparty, Goldman Sachs. The run started in early 2007 when Goldman issued a collateral call to AIG under the CDS as a result of the decrease in the value of the collateral provided by AIG. AIG faced difficulties in meeting its counterparty's demand. The transaction was eventually settled between the parties. However, due to the collateral calls, AIG's liquidity position became so dire at a certain point that the US Department of Treasury decided to intervene by effectively 'bailing-out' AIG.
Regulators in the USA took note and, in the wake of the financial crisis, developed a special resolution regime that, among other things, temporarily stayed the exercise of the close-out netting rights under the MA to give resolution authorities time to take actions to stabilize a failing financial institution. 43 While such stays were already part of the banking regulation, 44 they did not apply to other types of financial institutions, such as insurance companies. In the USA, Title II of the Dodd-Frank Act (DF Act) provides for the mechanics of curtailment of those rights for a broad spectrum of systematically important financial institutions. 45 While prima facie, the mechanics seemed to resolve the problem, there remained questions about the effectiveness of the DF Act mechanics, mainly when dealing with crossborder transactions. The mechanics mandated by the DF Act would apply to a US entity such as AIG. However, questions remained about the foreign counterparties of US entities. To the extent that they have not enacted such protections, the efforts of US authorities could be impeded. As Scott O'Malia, ISDA's CEO put it, [i]f a US financial group enters resolution, then Dodd-Frank would apply and a stay would be imposed on terminations by its derivatives counterparties at least, those subject to US law. If that US company has traded with a UK counterparty under English law, however, then there is some doubt as to whether the stay would apply, potentially impeding the efforts of the US resolution authority to deal with the situation. 46 One possible way to address the problem was to have similar stays imposed on derivatives in those jurisdictions. However, the concern was that this could prove to be impracticable as changes in the law of that kind can take a long time.
Another possibility was to have financial institutions agree between themselves that they will stay the exercise of their close-out netting rights under the MA in cases where the defaulting counterparty is a financial institution subject to resolution by the competent authorities. Accordingly, ISDA working with the 18 largest banks and the Financial Stability Board developed the Resolution Stay Protocol that will impose a stay on cross-default and early termination rights within standard ISDA derivatives contracts between G-18 firms in the event one of them is subject to resolution action in its jurisdiction. The stay is intended to give regulators time to facilitate an orderly resolution of a troubled bank. 47 The Resolution Stay Protocol transformed the close-out mechanism provided for under the MA through coordinated efforts of ISDA and public regulators into an explicit regulatory mechanism embedded in a contract. The problem was solved through a contract rather than public regulation because it was easier to implement the regulatory solution through a contract on a transnational basis. The argument could be made that a solution of convenience for regulators should not be regarded as a template for the regulatory process. The solution amounted to forcing parties to agree to something contractually that regulators felt would take too long or be too difficult to achieve through legislation around the world. The effectiveness of the solution also remains to be seen. Nevertheless, the Resolution Stay episode tells us something important about the role the MA now occupies in the broader regulatory architecture of OTC derivatives markets. That role is par excellence, a regulatory one.
Market integrity I now turn to the self-interest provision and its implications for the integrity of the OTC derivatives markets. Gina-Gail Fletcher recently offered a helpful conceptualization of market integrity as encompassing the concepts of market fairness, the absence of market abuse and the protection of investors. 48 As she notes, 'it is an essential characteristic of a well-functioning market, as investors are only likely to participate to the extent they believe the markets are fair and not stacked against them. If, on the other hand, investors believe the markets to be rigged they are unlikely to enter the markets or, should they enter, they will discount the value of all transactions.' 49 Manufactured defaults are a good example of a market practice undermining market integrity. In recent years a series of stories has been published in the financial press discussing the practice of CDS holders prompting the reference entities to trigger bond defaults in situations in which the handsome pay-out under the CDS would allow the CDS holder to compensate the reference entity for its trouble. Perhaps the most notable case involved Hovnanian, a US-based house builder, and a Blackstone owned hedge fund GSO. As the Financial Times reported, in 2018 Hovnanian agreed to default on some of its debt in exchange for favourable financing from GSO. 50 The deal stirred up market controversy. Solus Alternative Asset Management LP (Solus), one of the main protection sellers on Hovnanian's CDSs, unsuccessfully lobbied ISDA to issue a 'clarification' of its rules that would prevent GSO's trade from succeeding. 51 Solus was successful, however, in prompting the Board of ISDA to issue a statement. 52 In the statement, dated 11 April 2018, ISDA's Board made a note of the regulatory context of the DF Act and relevant anti-manipulation and anti-fraud laws and stated that 'narrowly tailored defaults, those that are designed to result in CDS payments that do not reflect the creditworthiness of the underlying corporate borrower (the reference entity in the CDS), could negatively impact the efficiency, reliability and fairness of the overall CDS market.' The Board instructed ISDA staff to consult with market participants and advise the Board whether amendments to the CDs should be considered. Shortly thereafter, on 24 April 2018, the US Commodity Futures Trading Commission (CFTC) also issued a statement, in which it suggested that '[m]anufactured credit events may constitute market manipulation and may severely damage the integrity of the CDS market'. 53 On 6 March 2019, ISDA circulated the first proposed changes to the standard CDS contract. 54 Later, in June 2019, the US Securities and Exchange Commission (SEC), the CFTC and the UK Financial Conduct Authority (FCA) released a joint statement on opportunistic strategies in the credit derivatives markets. 55 The statement noted that the pursuit of manufactured defaults 'adversely affect[s] the integrity, confidence and reputation of the credit derivatives markets, as well as markets more generally. These opportunistic strategies raise various issues under securities, derivatives, conduct and antifraud laws, as well as public policy concern.' However, the statement stopped short of announcing any direct actions by the agencies. Instead, the agencies appear to have implicitly delegated the solution of the problem to ISDA.
Shortly after the publication of the joint statement by the agencies, in August 2019, ISDA published the final version of the Narrowly Tailored Credit Event Supplement to the 2014 ISDA CD (the NTCE Supplement). 56 In the NTCE Supplement, ISDA expanded the definition of 'Failure to Pay' under the MA by excluding from the scope of the definition failures that do not 'directly or indirectly either result from or result in a deterioration in the creditworthiness or financial condition of the Reference Entity'. 57 Pursuant to the NTCE Supplement, the CDCs are responsible for making the determination of whether or not an event constitutes a 'Failure to Pay'. 58 Crucially, in making its determination, the CDC is allowed to rely on various external factors a non-exhaustive list of which is included in the NTCE supplement. 59 Through the NTCE Supplement, the regulatory function of ISDA's documentation has been expanded to incorporate market integrityrelated considerations.

The regulatory context in textual interpretation
What are the implications of the recognition of the regulatory function of contracts for their interpretation? I have suggested earlier that deference to the plain meaning of a regulatory contract may be warranted as long as that meaning is not inconsistent with the objectives of the regulatory framework applicable to the market in which the contract is used as a regulatory instrument. Needless to say, when the plain meaning of the regulatory contract is blatantly inconsistent with that regulatory framework, the regulatory framework will prevail over the plain meaning of the contract. However, in many cases, courts exercise a certain amount of discretion in determining whether the regulatory framework is applicable at all. When they determine that it is not, the regulatory function of contracts is enabled by the legal frameworks of private law. This section reviews a sample of case law to illustrate how courts exercise that discretion and whether their exercise of that discretion is broadly consistent with the rule of contractual interpretation for regulatory contracts proposed in this article.

Systemic risk considerations
Consider first a much-discussed series of cases in which New York and English courts sought to reconcile the meaning of the close-out netting provision with its purported aspiration to regulate systemic risk with the doctrinal requirements of bankruptcy law. The cases arose from lawsuits initiated by the estate of the failed investment bank Lehman Brothers against certain of its OTC derivatives counterparties who, upon Lehman's default, decided to withhold the payments owed to Lehman consistently with a plain meaning of section 2(a)(iii) of the MA. Recall that prior to the GFC the section provided that Each obligation of each party under Section 2(a)(i) is subject to (1) the condition precedent that no Event of Default or Potential Event of Default with respect to the other party has occurred and is continuing.
In the first case before the US Bankruptcy Court for the Southern District of New York (SDNY), Judge James Peck of the SDNY ruled in favour of Lehman finding that through its inaction, Metavante has implicitly waived its rights to terminate the contract. 60 As Ellenberg et al. note in their review of the case: The court noted that while the Bankruptcy Code does not specify that non-defaulting counterparties must act promptly after a filing in order to rely on the protection afforded by its safe harbour provisions, the legislative history of the Bankruptcy Code establishes that Congress intended only to shield parties to financial contracts from the systemic risk that would result from cascading losses due to a counterparty's bankruptcy filing. Because the degree of systemic risk that could result from a single filing diminishes over time, both this decision and existing precedent held that the safe harbour only protects actions that are taken reasonably promptly after the filing date. 61 Interestingly, in the English case Lomas v JFB Firth Rixson, based on a similar set of facts and involving Lehman's European estate, the English Court of Appeals (ECA) rejected Lehman's argument. 62 The ECA observed that it was only the performance of the obligation that was being suspended, and not the obligation itself. This interpretation, the court noted, is in accordance with the express language of section 2(a)(iii), which stated that the condition precedent to payment is to subsist for so long as the event of default or potential event of default 'has occurred and is continuing' As Moller et al note, 'the suggestion that the suspension lasts only for a reasonable period was, according to the judge, contrary to that express provision as to the duration of the payment suspension'. 63 As they further noted, before dealing with each of the various alternative interpretations of s.2(a)(iii) advanced by the Joint Administrators and by the respondents, the judge referred to two general considerations in interpreting s.2(a)(iii). The first was the need, given the widespread use of the ISDA master agreement, for 'clarity, certainty and predictability in its interpretation.' The second concerned the limited circumstances in which an English court will find that a term is implied into a contract . . . There is no scope for the court to find that a term is implied simply because it makes commercial sense or even because reasonable parties to the contract would have adopted the term had it been suggested to them. (ibid) The judgment of the ECA stood in sharp contrast with the earlier decision of the US bankruptcy court for the SDNY. In the English case, the ECA considered the close-out netting provisions to be a purely commercial one and followed the provision's plain meaning. In the USA, the SDNY interpreted the provision in the light of the regulatory framework applicable to the OTC derivatives market. 64 ISDA was initially critical of the outcome in the case from the SDNY. However, subsequently, the ECA outcome attracted a considerable amount of criticism from some market participants and various governmental and regulatory bodies who expressed concerns over the uncertainty caused to the defaulting parties when section 2(a)(iii) is interpreted in the way the ECA interpreted it. Indeed, as early as December 2009, the UK Treasury called upon ISDA to find a solution to limit the operation of section 2(a)(iii) to a 'reasonable period'. ISDA proposed its first amendment in 2011, but it was only the subsequent version published in 2014 that was adopted. 65 The history of the amendment of the MA for use in relation to section 2(a)(iii) suggests that possible inconsistencies between the plain meaning of a regulatory contract and the legal context should ideally be resolved prospectively through coordination between private and public regulators. Nevertheless, in some cases, courts may play a role in facilitating such coordination. That role consists of ensuring that the outcome of the interpretation of a regulatory contract is not inconsistent with the objectives of the regulatory framework applicable to the market in which the contract is used as a regulatory instrument.
The application of the regulatory rule of contractual interpretation may be challenging in the absence of any prior indication from regulators as to what those objectives are. However, if those objectives are clear, as they were in the case of the close-out netting provision, there are few reasons why courts should not interpret the meaning of the regulatory contract in the light of those objectives. This is what Judge Peck did in Metavante.
The argument could be made that the practical significance of the Metavante decision was limited. Judge Peck held that Metavante had waived its right to terminate the contract, but he did not mandate Metavante to make back payments-the section 2(a)(iii) condition still applied. Further, in the case, Metavante was out of the money (it owed Lehman). Its actions may have been different if it was in the money (owing from Lehman) or even out of the money in a two-way trade (because of fluctuations in the value of the underlying asset or index).
But while the practical significance of the Metavante may not have been great, its theoretical significance was. The theoretical significance of the decision consisted of demonstrating that it is possible and desirable to ensure that the plain meaning of a regulatory contract is not inconsistent with the objectives of the regulatory framework applicable to the market in which the contract is used as a regulatory instrument. Otherwise, the use of that contract by market participants could undermine the achievement of those objectives, which would be problematic from a public policy standpoint.

Market integrity considerations
There is no doubt that a regulatory interpretation of a contract takes away from market participants some of the flexibility ensured in a strict textual interpretation. There is a difference between being able to suspend payments for a reasonable period (under the regulatory interpretation of section 2(a)(iii)) and indefinitely (under a strict textual interpretation of section 2(a)(iii)). ISDA represents market participants who generally want flexibility and this, perhaps, explains why ISDA has historically endorsed strict textual interpretations of its documents. The NTCE Supplement was something of an anomaly in that respect. It permitted the CDCs to consider the purpose of a transaction in determining whether a 'Failure to Pay' and therefore a credit event has occurred. 66 While ISDA has not abandoned the textualist approach, it has for the second time, explicitly accounted for the regulatory context in which its contracts operate.
That shift of perspective can be easily explained and justified in line with the reasoning presented in this article. ISDA represents market participants who generally want to be able to use ISDA's documents flexibly. But ISDA also represents market participants who want efficiency and safety. The regulators made it clear that the opportunistic use of the self-interest provision undermines the achievement of those objectives even though, admittedly, they stopped short of taking any actions and let ISDA devise a regulatory solution which will inevitably affect contractual interpretation. That choice prompts two questions. First, is the objective of the regulatory framework applicable to the OTC derivatives markets clear enough to warrant a regulatory interpretation of the self-interest provision? Second, are the CDCs the most appropriate venues to make such determinations?
The answer to the first question is rather straightforward, considering the regulatory developments discussed above. Yes, the regulatory framework applicable to the OTC derivatives markets is clear enough to warrant a regulatory interpretation of the selfinterest provision. Regulators have indicated that the pursuit of manufactured defaults is detrimental to the integrity, confidence and reputation of the credit derivatives markets.
As for the second question, several scholars of regulation, including myself, have in the past made the point about the conflicted nature of CDCs. 67 ISDA has taken important steps to resolve conflicts of interest that could arise from the CDCs' composition in the latest version of the CDC rules. 68 Nevertheless, the delegation of the task of the policing of market integrity to the CDCs raises a more general issue about the CDCs' proper role. The task is regulatory in nature and there could be doubts as to whether, as currently set up, the CDC can perform that role. Fletcher suggests that ISDA could set up a separate adjudicatory panel and entrust it with the task. 69 Conceivably, the Panel of Recognised International Market Experts in Finance could play that role as well. While technically equipped for the task, each of those bodies would inevitably face the political economy challenge that is the running theme of this article-who should regulate what and how?
Consider a recent case also discussed by Fletcher in her article, Good Hill Master Fund LP v Deutsche Bank AG, 70 in which the New York state appellate court addressed the implications of the implied covenant of good faith on a CDS counterparty's self-dealing conduct under the MA given the self-interest provision. In the case, Good Hill was the protection seller and the reference obligation was a series of residential mortgage-backed securities (RMBS) also held by Good Hill, which it subsequently sold to Bank of America triggering an event of default under the CDS. Deutsche refused to return the collateral provided by Good Hill in connection with the CDS. Good Hill sued for breach of contract and return of the collateral. Deutsche Bank defended its actions, alleging that Good Hill breached the implied covenant of good faith and fair dealing under the CDS contract by engineer[ing] [a] commercially unreasonable and untenable . . . purchase price' for the RMBS. The appellate court, however, disagreed with Deutsche Bank, affirming the decision of the trial court.
Fletcher summarizes the courts' findings as follows: the implied covenant of good faith and fair dealing does not impose any obligations beyond the terms of the contract and cannot impose duties that would contradict other terms of the contract. In light of the self-interest provision, Good Hill was permitted to both (i) enter into transactions involving the RMBS underlying the CDS, and (ii) pursue its own interests even if its actions may be detrimental to Deutsche Bank's interests. In concluding, the court stated that Deutsche Bank failed to meet its burden of proving Good Hill acted in bad faith in breach of the CDS contract.
While there is a clear commercial logic to the judgment, in my view, the case illustrates the limitations of a strict textual interpretation in the context of a provision, which has regulatory implications. There are good legal and economic reasons for courts to enforce regulatory contracts in accordance with the plain meaning of their terms as long as that meaning is not inconsistent with the objectives of the applicable regulatory framework. However, when the plain meaning of a regulatory contract is inconsistent with those objectives, courts should instead seek to align the meaning of the contract with those objectives through regulatory interpretation. Otherwise, as already noted, the use of that contract by market participants could undermine the achievement of those objectives, which would be problematic from a public policy standpoint.
The necessary condition that has to be met for a regulatory interpretation of contracts to be normatively justified is the identification of a clear regulatory objective and a clear understanding of how a more conventional, plain meaning interpretation would undermine the achievement of those objectives. Whereas that condition may not have been met in 2017, at the time Good Hill was decided, it was met soon thereafter, in particular following the issuance by the CFTC and, later, by the CFTC together with the SEC and FCA of statements on opportunistic strategies in the credit derivatives markets. These statements should be regarded as the legal basis of future regulatory interpretations of the self-interest provisions of the CDS. While the regulatory interpretation of the self-interest provision in line with that statement would reduce the flexibility afforded by the plain meaning interpretation, it would ultimately increase the integrity of markets and benefit market participants in the long run, consistently with the regulatory rationale put forward by the agencies.

The judicial cost and error objections
The rule of regulatory contractual interpretation proposed here is best understood as an extension of the plain meaning or textual rule. The regulatory rule is not a contextual rule, even though it considers the regulatory context of the contract. 71 But it is subject to some of the same objections typically raised with respect to contextual interpretation, namely that (i) it is costly and impractical to implement 72 and (ii) likely to result in errors. 73 While the cost argument is persuasive in the context of a very large number of commercial contracts litigated in courts, the universe or regulatory contracts litigated in courts is very 71 The argument could be made that the text/context distinction may not be as analytically useful as contract scholars believe it to be. small making the cost argument less persuasive. Similarly, the impracticality objection seems weak in cases where the regulatory objectives that inform the interpretation of a regulatory contract are clear. A high level of clarity concerning those objectives is necessary for a successful application of the regulatory rules of contract interpretation.

Conclusions
Boilerplate contracts can display relational, but also regulatory features. The conceptual difference between the two categories of boilerplate contracts is not only in terms of the degree of modification that is possible but also in terms of their respective functions. While relational contracts allow the parties to optimize their transactions, regulatory contracts are designed to optimize markets. Markets can be understood as aggregations of transactions, but market problems, such as negative externalities, are different from transactional problems, such as transaction costs. The ISDA MA addresses both types of problems and, in that respect, is a unique type of contract.
This article sought to capture that uniqueness by identifying the MA's regulatory properties and proposing a rule for the interpretation of contracts with such properties. The rule states that the outcome of the interpretation of regulatory contracts must not be inconsistent with the objectives of the regulatory framework applicable to the market in which the contract is used as a regulatory instrument. The reliance on that rule is normatively justified in cases where the regulatory objectives are clearly defined and particularly when the alternative outcome is to void the contract or its provision. While the regulatory interpretation of a contract reduces the flexibility afforded by the plain meaning interpretation, it can ultimately increase the efficiency and safety of markets and benefit market participants.

Funding
This article is part of a project that has received funding from the European Research Council (ERC) under the European Union s Horizon 2020 research and innovation programme (Grant Agreement No ERC-2016-CoG 725798 -REVEAL).