Neither relational nor discrete: the ISDA Master Agreement as a bimodal contract

1. Introduction

While recent English court decisions have seemed to abjure the concept of relational contract theory,¹ the decisions in Yam Seng² and subsequently in Bates v Post Office³ reified the pioneering work carried out by each of Stewart Macaulay and Ian Macneil in the 1960s and 1970s, respectively.⁴ Coupled with a general judicial movement towards contextualism,⁵ any notions that a contract should be assessed only within its four corners have essentially dissipated.⁶ However, opinions—both academic and judicial—continue to oscillate on both contractual construction and recognition of a class of contracts referred to as ‘relational’.

The theory posits, inter alia, that contracts can be observed on a spectrum, at one end are those referred to as discrete contracts which describe one-off, simple, exchanged-based transactions, and at the other end are those referred to as relational contracts which can be thought of as being dependent on not just the contractual terms but on the wider relationship between the parties. However, as Macneil observes, either end of the spectrum, ‘like the ends of rainbows … are mythical’,⁷ meaning that most contracts sit somewhere between either extremity. Yet, it is proposed by those supportive of the theory that nearly all contracts are, to some extent, at the relational end of the spectrum.⁸

The predominant approach so far has been to designate contracts as either discrete or relational. However, academics, such as Melvin Eisenberg,⁹ Hugh Collins,¹⁰ Elizabeth Mertz¹¹ and even Macneil¹² himself, have recognized that many contracts are, in fact, intrinsically dichotomous, containing both relational and discrete terms. This article seeks to investigate this phenomenon by using the International Swaps and Derivatives Association (ISDA) Master Agreement (MA) as a case study of what is referred to here as an example of a bimodal contract, a contract that is neither entirely relational nor entirely discrete but a combination of the two. In particular, this article evaluates the proposal by ISDA to introduce a smart contract version of its MA which has highlighted that there are some terms in the MA, referred to by ISDA as ‘operational clauses’,¹³ which are highly discrete, acontextual functions which can be automated, while recognizing that there are also ‘non-operational clauses’¹⁴ which may expressly refer to behavioural standards such as ‘good faith’ and ‘commercial reasonableness’.

2. Relational contract theory

In a lot of ways, relational contract theory mirrors developments in behavioural economics, with both concepts criticizing the more classical, formalist approach—whether it be ‘classical contract law or rational choice theory’—and instead focusing on the mercurial and social nature of humans when interacting with each other. There are also connections with the concept of ‘incomplete contracts’ in which it is deduced that contracts cannot cover every eventuality because the parties cannot prognosticate on all possible future events and because the costs of endeavouring to do so would be prohibitive or, to put it in the language of economists, the transaction costs would far exceed the potential utility.¹⁵ In essence, relational contract theory, much like these developments in economics, recognizes that contracts are limited in their capacity to capture the whole nature of the relationship between the parties.

David Frydlinger, Oliver Hart and Kate Vitasek have advocated that, in recognition of the aforementioned concepts, what emerges between the parties is a move away from trying to catalogue each potential scenario, repudiating the self-serving, individualistic, corporate protectionism which creates a tension at the planning stage of every potential business relationship or creating what Oliver Hart and John Moore have described as ‘shading’—the idea that if a party feels aggrieved it will adjust its performance, perhaps only subtly, to the detriment of the relationship.¹⁶ Frydlinger, Hart and Vitasek put forward the concept of ‘formal relational contracts’ which, far too simply put, are written contracts with a more flexible framework based on pre-defined principles and objectives which overlay the more specific operational terms.¹⁷

In terms of English law, these new concepts form part of a continuum of the ideas promulgated by the English judiciary—in a judicial and extrajudicial capacity—that, eg and in the words of Lord Steyn, a ‘thread runs through our contract law that effect must be given to the reasonable expectations of honest men’.¹⁸ Relational contract theory can therefore be seen as an extension—at least judicially given the chronological judicial recognition of the two—of the contextual approach, which was advocated by proponents such as, amongst others, Lord Wilberforce¹⁹ and Lord Hoffmann,²⁰ respectively.

The idea that contracts must be viewed within the ‘matrix of facts’²¹ in which they are situated is developed further by relational contract theory in which there are certain expectations incumbent upon the parties which may not be expressly set out in the legal agreement between them.

Ian Macneil set out ‘four core propositions’ of relational contract theory: (i) ‘every transaction is embedded in complex relations’, (ii) understanding the transaction requires understanding the ‘essential elements of its enveloping relations’, (iii) ‘effective analysis of any transaction requires recognition and consideration of all essential elements of its enveloping relations that might affect the transaction significantly’ and (iv) combining contextual analysis of relations and transactions is more efficient and complete than non-contextual analysis.²²

However, relational contract theory has been predominantly associated with the idea that contracts fit on a spectrum. Macneil originally placed ‘transactional or discrete contracts’ at one end and ‘relational contracts’ at the other. The two are defined as follows: This spectrum or scale, applied to contracts in toto, has been argued to slant heavily towards the relational end, with even ostensibly discrete transactions, such as the one-off purchase of fuel, being described as relational. The argument in this particular example is that the motorist understands that she must pay for the fuel, the attendant accepts the money offered as an acceptable form of payment and must provide change if applicable, and both parties understand the surrounding circumstances around their transaction which allows it to occur without criminal or civil sanctions.²⁵ On this basis, Macneil labelled such transactions as ‘as-if discrete’, since all discrete transactions, he postulated, ‘are embedded in relations’.²⁶ While this sociolegal observation is perfectly rational, this article does not apply such a broad definition of the term relational contracts because it is the author’s opinion that such broad strokes, when aiming for practical application rather than theoretical validity, essentially devalue the vital role that the relational contract theory spectrum can provide in the categorization of contracts, especially in the face of some criticism that discerning between two ends of such a scale to begin with is ‘too simplistic’²⁷ an exercise.

• ‘Discrete transactions are characterized by careful measurement of what is exchanged, specification of time and manner of performance, comprehensive planning regarding the allocation of burdens and benefits, and little expectation of cooperation outside the scope of the exchange.’²³

• Relational contracts ‘involve continuing exchange and interaction between the parties. The open-ended nature of contractual relations prevents detailed planning and necessitates cooperation and compromise regarding the incidence of contractual burdens and benefits.’²⁴

While the distinction between discrete and relational contracts initially gathered pace under English law, particularly with its judicial application in two particular court cases,²⁸ this article investigates the dichotomy ‘within’ particular contracts as well as ‘between’ them. The English courts have so far investigated a ‘class of contract’ referred to as relational.²⁹ This article suggests that such investigations should discern between discrete and relational classes when analysing terms rather than just contracts more holistically. This concept questions the infallibility of proclaiming one contract to be discrete and another to be relational or, as MacNeil initially argued, that all contracts are relational.³⁰ While this article does not seek to undermine the original theory on relational contracts, it does seek to reframe and build upon it. Instead of considering a contract to be categorized as discrete or relational in the macro-sense, the theory can be applied on a micro-level to the very terms of one particular contract which creates an interstice ‘between the terms’ of a contract rather than ‘between contracts’ more widely.

This article uses the ISDA MA as a case study given its ubiquity³¹ in the over-the-counter (OTC) derivatives market and the development by ISDA of a smart contract version of its MA which raises important questions over the automation of some of its terms but not others which, it is argued here, indicates a division within the MA between discrete and relational terms.

3. ISDA

The ISDA MA has been described as ‘probably the most remarkable standard form ever devised’.³² This contract forms an important legal basis between parties dealing in the OTC derivatives market even as regulators have tried to reassert their authority after decades of deregulation.

The MA was first pitched as a compromise to protracted legal wrangling and associated battles of the forms which had arisen between a group of banks dealing in OTC derivatives. In pursuing the standardization of legal terms, they formed what they referred to as a Documentary Committee in May 1984, which, a year later in 1985, evolved into ISDA and in that same year, as described by Simon Firth, the association had published a ‘Code of Standard Wording, Assumptions and Provisions for Swaps, comprising a set of standard definitions and terms that could be incorporated in whole or in part into individually negotiated agreements,’³³ the precursor to the ISDA MA. The MA is in its third iteration, the initial version was published in 1987, followed by a second in 1992, and the latest version was published in 2002. It is accompanied by supporting documents, such as a Schedule, a Credit Support Annex, Definitions and Confirmations—these, along with the MA, form a single agreement. English law and New York law are the two main governing laws of the MA but this has since expanded. For the purposes of this article, references to the MA are to the 2002 English law version.

OTC derivatives are private, bilateral arrangements between counterparties trading financial instruments such as forwards, options, swaps or combinations of the aforementioned. In essence, derivatives are financial contracts deriving their value from some underlying commodity, asset, rate, index or the occurrence or magnitude of an event.³⁴ OTC transactions do not involve an exchange sitting between the parties guaranteeing performance of the contract so they pose additional systemic risk to the market. As John Biggins and Colin Scott describe, ‘OTC derivatives trading is … considered capable of offering highly innovative and socially useful risk management and investment strategies … [however, it is] comparatively more risky than exchange trading.’³⁵

Since the global financial crisis of 2008, there has been a concerted effort by regulators to rail back on a lot of the deregulation which had been so zealously pursued under the neoliberal belief that markets could adequately self-regulate using OTC derivatives to simply and effectively reduce risk. However, in spite of these attempts and in the face of criticism of regulatory efforts to control an already complex market,³⁶ the terms of the ISDA MA still play a major role in framing the legal obligations and responsibilities of parties within this very global market, creating what Jeffrey Golden has described ‘as a kind of global law by contract’.³⁷

The wider implications of ISDA’s role aside, the success of the MA has meant that ISDA has been able to take a leading role in pushing forward a number of initiatives in order to improve standardization and operational risk management across the market. One of those initiatives is the proposed introduction of its MA in smart contract form, a concept which, as TJ de Graaf explains, was originally, ‘at a time when the blockchain did not yet exist … a computerized transaction protocol that execute[d] the terms of a contract’.³⁸

4. Smart contracts

A ‘smart contract’, as defined by Christopher Clack, Vikram Bakshi and Lee Braine, is ‘an automatable and enforceable agreement. Automatable by computer, although some parts may require human input and control. Enforceable either by legal enforcement of rights and obligations or via tamper-proof execution of computer code.’³⁹ ‘Smart contract’ is perhaps a misnomer in the nomenclature, as ISDA admits when it refers to its proposition as a ‘smart derivative contract’,⁴⁰ the distinction being that a smart contract may not be a contract at all, because the idea was originally conceived by programmer Nick Szabo to describe a ‘set of promises, specified in digital form, including protocols within which the parties perform on these promises’.⁴¹

Kristian Lauslahti, Juri Mattila and Timo Seppälä define smart contracts from a legal perspective as ‘digital programs, based on the blockchain consensus architecture, which will self-execute when the terms of the agreement are met, and due to their decentralized structure are also self-enforcing and tamper-proof.’⁴²

Smart contracts predominantly operate on blockchains: digital, peer-to-peer, decentralized platforms which derive their name from the blocks of data which act as a snapshot of transactions and which are linked together through each node (each computer system in the network) to create data chains, providing an unalterable public record.⁴³ These decentralized systems and records give rise to the term ‘Decentralized Ledger Technology’ (DLT).⁴⁴ DLT refers to the decentralization of any database that is required in a network so that the participants, as part of a blockchain, are able to verify and monitor activity. This is done in a way that anonymizes a lot of the data, replacing such information with cryptographic keys. This same technology is used to underpin cryptocurrencies such as Bitcoin. On such networks, smart contracts can form the legal basis of arrangements between parties.

However, while ISDA has published a paper on potential DLT solutions,⁴⁵ the MA as a smart contract would, as originally proposed by ISDA, operate slightly differently. ISDA Assistant General Counsel Ciaran McGonagle, along with academic Christopher D Clack, emphasize that participants in the financial services sector will not necessarily adopt a peer-to-peer solution simultaneously so a centralized technology platform may need to be conceived instead.⁴⁶ ISDA would implement this centralized platform by way of the ISDA Common Domain Model⁴⁷ which allows parties to record trade details, provide settlement information, exchange confirmations, manage collateral and credit risk, and standardize reporting information. It is also governed by an executive committee of ISDA members.⁴⁸ In this sense, the MA as a smart contract would resemble the original conception of the idea, being a computerized transaction which executes certain terms without the decentralized network.

For the purposes of this article, the important point raised by the proposed introduction of a smart contract version of the MA is the description by ISDA of certain clauses within the agreement. In order to establish what will be automated, ISDA has categorized two clauses in its MA, those referred to as ‘operational’ and those which are ‘non-operational’.⁴⁹ This categorization derives from the initial purpose of automation as envisaged by ISDA which is to streamline the process of payment obligations or collateral transfers⁵⁰ under the MA.⁵¹ Transfers of collateral or margin—an amount required by one or either party to mitigate the risk of movements in the price of the underlying instrument—can still be found to be processed between parties via an e-mail exchange which is followed by a bank transfer. Automation, for this purpose, is designed to alleviate the administrative burden and mitigate the risk of certain manual processes. It is indicative of the mechanical nature of certain operational clauses in the MA, referred to here as discrete clauses, which obviate human intervention and the associated susceptibility to error. That does not mean that such clauses are unfettered from judicial intervention, there may still be problems which arise with these clauses as well as the coding which underpins them. As the UK Jurisdiction Taskforce has observed, smart contracts may reduce legal intervention but there are still risks associated with, eg, the code itself or system failures which will still require adjudication.⁵² So the code will be, prima facie, clear and unambiguous and execute as expected but there will be times when a ‘judge will need to look beyond the four corners of the code’.⁵³

This code, when run, will only be able to compute a certain amount of context and far less than what Hugh Collins and David Campbell have referred to as the ‘implicit dimensions of contracts’, such as ‘fair dealing, good faith and co-operation’.⁵⁴ Ostensibly, those operational clauses which have been coded to run as a smart contract will operate in a limited contextual capacity and execute in a discrete manner, for example, when X event occurs and parameter Y is met, then execute Z. The MA as a whole will comprise those discrete clauses which are coded and, contiguously, those relational clauses which may form part of the static code but will not, in the words of ISDA, ‘embed any conditional logic’.⁵⁵

5. Classifications

As alluded to above, relational contract theory has become an extension, or is at least a close relative, of contextualism, the idea most notably propagated in English law by Lord Wilberforce, Lord Steyn and Lord Hoffmann, respectively. Lord Steyn encapsulated contextualism in the case of Total Gas Marketing Ltd v Arco British Ltd⁵⁶ in which he stated that questions of interpretation have to be considered ‘in the light of the contractual language, the contractual scheme, the commercial context, and the reasonable expectations of the parties’.⁵⁷

However, claims that a contract should be considered ‘relational’ are often accompanied by two concepts which trouble English law: implied terms and good faith. More troublingly for English law, the case of Yam Seng raised the prospect of implying terms of good faith, doubling the ante.

Implied terms

There was a brief period in which implied terms were considered more broadly as simply part of the process of construction. This was signalled by Lord Hoffmann’s statement in ‘Attorney General of Belize v Belize Telecom Ltd⁵⁸ in which he stated that implication of a proposed term should be considered if it ‘spell[s] out in express words what the instrument, read against the relevant background, would reasonably be understood to mean.’⁵⁹

However, in the case of Marks and Spencer plc v BNP Paribas Securities Services Trust Company (Jersey) Limited,⁶⁰ the judgment of Sir Thomas Bingham in the case of Philips Electronique Grand Public SA v British Sky Broadcasting Ltd was emphasized:

The courts’ usual role in contractual interpretation was, by resolving ambiguities and reconciling apparent inconsistencies, to attribute the true meaning to the language in which the parties themselves had expressed their contract. The implication of contract terms involved a different and altogether more ambitious undertaking: the interpolation of terms to deal with matters for which, ex hypothesi, the parties had made no such provision. It was because the implication of terms was potentially so intrusive that the law imposed strict constraints on the exercise of that extraordinary power.⁶¹

As David Campbell and Hugh Collins have observed, the courts are ‘reluctant to rewrite contracts in ways that they might regard as more reasonable, or to satisfy what may be regarded as a reasonable expectation which has not been protected by an express contractual agreement.’⁶² However, they put forward an argument that, in certain circumstances, the only way to distinguish between, what Stewart Macaulay refers to as, the ‘real deal’ and the ‘paper deal’⁶³—what the parties have agreed as opposed to what their contract may stipulate—is by incorporating the implicit dimensions of the contract into the analysis.⁶⁴

By classifying the MA as ‘bimodal’, the intention is not to suggest that implied terms could not form a part of contractual construction even though doing so in the MA, given its global use and importance in such a large market, would appear unlikely. However, nor is there a subscription to the idea that describing contracts or contractual terms as relational will reflexively mean that an implied term of good faith will need to apply.⁶⁵

The designation of terms as either discrete or relational could help to demarcate between those clauses (or even within clauses themselves) which, in the case of a dispute, could determine their treatment by the courts. For example, ISDA uses its payment clause—section 2(a)(i) of the MA which will operate alongside specific parameter-setting clauses within the ISDA Schedule and Credit Support Annex⁶⁶—as an example of a clause which is open to automation.⁶⁷ Indeed, payment obligations, such as netting or the exchange of margin, are often cited as responsive to automation because the inputs can be rendered with a high degree of certainty.⁶⁸ Compare this payment obligation as applied to those amounts arising in the periodic course of business under the MA to the more specific close-out amount calculation. The close-out amount is payable on early termination⁶⁹ and forms part of the early termination amount payable under section 6 of the MA. Historically, previous versions of the MA set out various methodologies for the calculation of amounts payable on early termination,⁷⁰ which included the non-defaulting party determining its loss in good faith⁷¹ or obtaining third party market quotations which had to produce a commercially reasonable result.⁷² This caused a number of disputes in both the UK⁷³ and the USA⁷⁴ and, in part, led to the introduction of the 2002 MA which now utilizes a close-out amount methodology under which a non-defaulting party must act in good faith in order to determine any gains or losses using commercially reasonable procedures.⁷⁵

These examples show that, even between amounts payable, there are those terms which will be discrete, routine and uniform in their calculation, whereas there are others which will be far more relational, irregular and ad hoc in their methodology.

Automation can capture a lot of those discrete terms and will, over time, subsume more even perhaps those which we would today consider relational.⁷⁶ However, in the event of a formal dispute, it is proposed here that the court’s role is to limit its interference with such discrete terms and to concentrate on, or fill gaps around, those which, eg, require the parties to act in ways which are not yet, and which may never be, automatable.⁷⁷

Robert Scott has observed that while it can be argued that all contracts are relational, contract law is formal, simple and classical.⁷⁸ In contrast, Ian Macneil suggests that contract law ‘should generally track the relational behaviour and norms found in the relations to which it applies’.⁷⁹ Distinguishing terms within contracts as either discrete or relational, could be used to reconcile, as far as possible, such opinions. This bifurcation of terms could be used in contract law to signpost where judicial intervention may be required and where it is not. However, such distinction, it is conceded, is much easier to make with the kind of smart contract discussed here, with its associated operational and non-operational clauses.

Good faith

Good faith in the performance of contracts has historically been met with ‘hostility’⁸⁰ under English law. In fact, it has been met with similar opprobrium by the UK government which disseminated a paper in which it raised concerns centred around the unpredictability of the term when used in legal proceedings.⁸¹ The concept has been described by Michael Furmston and John Carter as being associated with certain standards such as honesty, reasonableness and fair dealing, with the latter traditionally being unrecognized in contract law.⁸²

Express duties of good faith have not seemed to pose insuperable problems of interpretation. There have even been attempts to standardize the concept when expressed by the parties⁸³ but, more broadly, its inclusion in a contract will convince the courts to look at ‘the context in which the good faith obligation was entered into’.⁸⁴

Implied terms, as a default, are not to be interpolated unless, as was affirmed in Marks and Spencer plc v BNP Paribas Securities Trust Company (Jersey) Ltd,⁸⁵ ‘at the time the contract was made, a reasonable reader of it would consider the term to be so obvious as to go without saying or the term is necessary for business efficacy.’⁸⁶ Moreover, this test, coupled with an apprehension, both in the legal and economic literature, about the imprecision of the concept of good faith leads to a more general argument against, or at least an exhortation for proceeding extremely cautiously with, implying terms of good faith into contracts.⁸⁷

The recognition of a class of contracts referred to as ‘relational’, precipitated by the case of Yam Seng,⁸⁸ has, in Lord Justice Leggatt’s words, ‘provoked divergent reactions’ but ‘there appears to be growing recognition that such a duty [of good faith] may readily be implied in a relational contract’.⁸⁹

Judicial reaction to relational contracts following the decision of Yam Seng has been mixed⁹⁰ but there appears to be a particular impediment to certain classes of contract, such as the MA, being classified as relational and that relates to their standard form as well as their professional drafting. The treatment of standard-form contracts and the need for consistency in their interpretation was addressed by Lord Diplock in the case of The Nema:⁹¹

… it is in the interests alike of justice and of the conduct of commercial transactions that those standard terms should be construed … as giving rise to similar legal rights and obligations in all [cases] in which the events [that] have given rise to the dispute do not differ from one another in some relevant respect. It is only if parties to commercial contracts can rely upon a uniform commercial construction being given to standard terms that they can prudently incorporate them in their contracts without the need for detailed negotiation or discussion.⁹²

Professionally drafted agreements

The ‘strict constraints’ on implying terms and the continued unease under English law to derogate too far from the principle of freedom of contract⁹³ must be assessed in conjunction with the statement by Lord Hodge in Wood v Capita Insurance Services Ltd⁹⁴ that the English courts would, prima facie, adopt a more formalist or textualist approach to the interpretation of sophisticated agreements in comparison to the perhaps more contextual approach they may take to the interpretation of less sophisticated agreements. Even though he qualified that assertion somewhat by saying that there may be circumstances where the division between sophisticated and unsophisticated will be inconsequential.⁹⁵

Furthermore, in two recent English High Court decisions, there has been a judicial reluctance to imply terms into professionally drafted agreements. In UTB LLC v Sheffield United Ltd,⁹⁶ Fancourt J stated that where ‘detailed, professionally-drawn contracts exist, it is more difficult to imply terms because there is a strong inference that the parties have given careful consideration to all the terms by which they agree to be bound’⁹⁷ and in Taqa Bratani Ltd v Rockrose UKCS8 LLC,⁹⁸ HH Judge Pelling QC stated that implying terms must be carried out with particular care in professionally drafted and negotiated agreements between well-resourced parties because ‘where an issue has been left unresolved, it is much more likely to be the result of choice rather than error’.⁹⁹

In terms of judicial treatment of the ISDA MA in particular, it has been described by the English courts as ‘probably the most important standard market agreement used in the financial world’¹⁰⁰ and ‘plays an important role in the efficient functioning of the international financial markets and their financial stability’.¹⁰¹ In addition, it has been stated that it is ‘axiomatic’ that the MA is ‘interpreted in a way that achieves the objectives of clarity, certainty and predictability, so that the very large number of parties using it know where they stand’¹⁰² and that ‘consistency, predictability and certainty are essential and so they are much less susceptible to interpretation by reference to background circumstances or matrix’.¹⁰³ Such descriptions would appear to support the view more generally proposed by Catherine Mitchell that English law is ‘undergoing a formalist—or at least anti-contextualist—shift’.¹⁰⁴

Classification as relational

Despite these signals that the MA should be interpreted narrowly and viewed conceptually as at the discrete end of the spectrum, there are some indications that the MA is not so easily categorized.

First, the agreement expressly refers to certain standards of behaviour expected from parties using the MA. The English law version of the 2002 MA expressly references ‘good faith’ 12 times and ‘commercially reasonable’ 11 times, respectively.¹⁰⁵ In the case of Mid Essex Hospital Services NHS Trust v Compass Group UK and Ireland Ltd,¹⁰⁶ it was held that where specific references are made to concepts such as good faith, ‘care must be taken not to construe a general and potentially open-ended obligation such as an obligation to “co-operate” or “to act in good faith” as covering the same ground as other, more specific, provisions, lest it cut across those more specific provisions and any limitations in them.’¹⁰⁷ While this may ostensibly limit recognition of relational concepts in the MA to those express references in isolation, the inclusion of such relational expectations itself is redolent of something more than a simple exchange.

Secondly, as a theoretical construct, the MA would, in many cases fulfil the requirements of what would be defined as a relational contract. Using Macneil’s 12 axes determining extreme discrete or relational traits in contractual relations or Mr Justice Fraser’s characteristics of relational contracts enumerated in Bates v Post Office,¹⁰⁸ we can see that the MA could fit into some of the relational examples provided: the cumulative transactions underpinned by the MA will often require and expect future cooperation, there may be personal relationships involved albeit under the auspices of large corporations, there is a certain amount of trust and confidence reposed in the other party, there is fidelity involved in the transactions, transferring rights under the MA will often be refused if there is a material impact on credit ratings, and its users, as evinced by some of the express terms in the MA, expect certain behavioural standards from their counterparties.¹⁰⁹

Finally, economist Oliver Williamson has suggested that discrete contracts operate well for high certainty, low complexity, low dependency situations,¹¹⁰ which does not align with the market volatility, high complexity and systemic risk associated with OTC derivatives markets. Moreover, scholars such as Jeffrey Golden see the MA as a relational contract because, inter alia, it is necessarily ‘incomplete’ on the basis that it is designed to ‘establish a framework for a course of dealing over a (usually extended) period of time’ and so ‘the need to fill in meaning may be more to the fore’.¹¹¹ Given that all contracts can be seen as incomplete,¹¹² the question then becomes how to reconcile this assertion that contracts cannot ever fully reflect the relationship between the parties with the reluctance of the courts, for the reasons enumerated above, to interfere too much with the terms of the MA.

6. Conclusion

The proposed introduction of a smart contract version of the MA raises certain questions about how the contract is classified and interpreted. As Melvin Eisenberg has observed, relational contract theory provides a method by which the construction of contracts by the courts can be individualized, open and inductive which stands in opposition to classical contract law which is standardized, axiomatic and deductive.¹¹³ This also feeds into the ideas promulgated by psychologists and economists which have refuted ideas of rational choice theory and formalism in a world of behavioural biases and incomplete contracts.

However, there are convincing arguments, both legal and economic, for why treating the MA as a relational contract, in a practical sense, gives rise to uncertainty and higher back-end transaction costs. In relation to the former, there are legitimate concerns that ‘individualizing’ judicial responses create incertitude in a large and global market which relies on the standard-form MA. In addition, Eric Posner and, more specifically on the complexity of adjudicating financial disputes, Jeffrey Golden, have raised concerns that while judges who understand the complex derivatives market can play an important role in interpreting these contracts in a way that mitigates systemic risk, the responsibility is often placed upon judges without such expertise to interpret these terms which itself presents that very risk.¹¹⁴ In relation to the latter, higher ‘back-end’ transaction costs—a concept discussed by Robert Scott and George Triantis and which relates to the costs of enforcement of the contract as opposed to ‘front-end’ transaction costs which relate to the costs of, for example, negotiating the terms of the contract that will govern the relationship¹¹⁵—arguments abound of how formalism in contractual interpretation reduces costs for both parties in disputes brought before the courts.¹¹⁶

If we maintain a demarcation between contracts, reconciling these methods of interpretation and classification are, prima facie, insoluble because there is too much logic on either side to dismiss the other. On formalism and contextualism alone, the argument for the former is that courts should not interfere with the written terms agreed between the parties while the argument for the latter is that contracts provide an incomplete picture of what has been agreed. The label discrete/relational has the same logical polarity when used reasonably.

The answer proposed in this article, which may already subconsciously form part of judgments on contracts, is to focus on how we view the classification of discrete/relational not by the contract as a whole but by the terms contained therein. I refer to such contracts as ‘bimodal contracts’, containing both discrete and relational terms. While this label maintains a distinction at the contract level, its aim is to focus on how such contracts may vacillate between either end of the discrete/relational spectrum based on the operational or non-operational nature of their terms. Ostensibly, this designation could create more complexity to a theory that has already been met with judicial and academic demurral. However, if the aim of the theory is to more accurately reflect business and societal attitudes and norms, circumstances may dictate an even more nuanced approach to that originally devised by Ian Macneil and Stewart Macaulay.

The reluctance we have seen from the English courts so far has, inter alia, hinged upon designating certain contracts as relational. In the case of Morley v The Royal Bank of Scotland PLC,¹¹⁷ Mr Justice Kerr stated that he rejected the ‘argument that the loan agreement, as extended several times, was a “relational” contract of any kind. It was an ordinary loan facility agreement.’¹¹⁸ It is, prima facie, the classification of the whole contract as relational which appears so unpalatable and, given the exigency of certainty in globally important standard market agreements¹¹⁹ and the purported sophistication of the parties involved in financial transactions¹²⁰ documented by the MA, it stands to reason that courts may be hesitant to class such contracts as relational with all the associated connotations. The solution, in the case of the MA, may be to either omit such classifications altogether or, as is proposed in this article, to designate such contracts as ‘bimodal’ with the aim of interpretation being to distinguish those clauses which are discrete with those that are relational, a task that becomes less abstruse with the advent of a smart contract version of the MA, with its operational and non-operational clauses.

Robert Scott, in an article ominously entitled ‘The Death of Contract Law’, observes that ‘contracting parties may simply prefer to behave under two sets of rules: an explicit (and rigid) set of rules for those parts of their relationship that require legal enforcement and an implicit (and flexible) set of rules for those aspects that respond best to self-enforcement.’¹²¹ When discussing a smart contract version of the MA, we maintain the distinction between rigidity and flexibility described by Scott but the mechanisms for enforcement are upended. Within the MA, there are clauses which fit the discrete, transactional, rule-based logic of automation while there are also those clauses which are relational and necessarily flexible. However, the enforcement, in the first instance, of discrete terms under the smart contract will be self-enforcing, or at least automatic, while the enforcement of relational terms will require more traditional legal enforcement.

The development of smart contracts for wider use in otherwise traditional legal arrangements is still in the nascent stages but the attempt here is to suggest that simple bifurcation of contracts as either discrete or relational may impose too heavy a burden on contracts which are often more nuanced and difficult to categorize. While there are archetypal examples of discrete and relational contracts, perhaps consideration of ‘bimodal contracts’, which contain characteristics of both, presents a more palatable solution to interpreting and categorizing certain contracts which are themselves being designed and executed in different and more complex ways.

Footnotes

Author notes