SENSE AND SEPARABILITY

Abstract

This article explores the doctrine of separability, as understood in particular in the English legal tradition. It does so by reference to the decisions in Sulamérica Cia Nacional de Seguros SA and others v Enesa Engelharia SA and others and ENKA İnşaat ve Sanayi A.Ş. v OOO ‘Insurance Company Chubb’ & Ors that explore the relevance of the concept when determining the law applicable to the arbitration agreement. These decisions largely treat the doctrine as irrelevant to the determination of the law governing the arbitration agreement. They do so because of the way in which English law views separability as tied inimically to the concept of enforcement of the arbitration agreement. This is unsurprising given the content of section 7 of the Arbitration Act 1996 and the position of the doctrine of separability as a legal fiction that must be restricted to its defined purpose. Viewed against the potential reform of the Arbitration Act 1996, the author asks whether a broader view of separability can be adopted. The author's view is that there are cogent and compelling reasons for adopting a broader view, that would promote certainty and consistency in a way that is not best served by the current approach.

I. INTRODUCTION

The doctrine of separability is a key tenet of arbitral law and practice. In the English system, it was not always so, but its statutory footing in section 7 of the Arbitration Act 1996 (1996 Act)¹ has left little doubt as to its importance for some time now. That provision provides as follows:

Separability of arbitration agreement.

Unless otherwise agreed by the parties, an arbitration agreement which forms or was intended to form part of another agreement (whether or not in writing) shall not be regarded as invalid, non-existent or ineffective because that other agreement is invalid, or did not come into existence or has become ineffective, and it shall for that purpose be treated as a distinct agreement.

As it presently stands, section 7 of the 1996 Act treats the doctrine of separability as a device to insulate the arbitration agreement from matters that would potentially invalidate the contract of which it forms a part. Viewed in this way, separability is designed to protect and promote the enforcement of the arbitration clause. The judgments of the majority and the minority in the Supreme Court in ENKA İnşaat ve Sanayi A.Ş. v OOO ‘Insurance Company Chubb’ & Ors ² provided English practitioners with a reminder of the limits of the doctrine, closely linking it to section 7. In doing so, the Supreme Court expressly distanced English law from the more expansive approach to separability provided by the Court of Appeal below.³

Given the clear wording of section 7 of the 1996 Act, in particular the use of the words ‘for that purpose’, the Supreme Court's approach and more narrow definition of separability are undoubtedly correct. Nevertheless, viewed against the backdrop of a potential reform of the 1996 Act, this does raise the following question: statute aside, is there a principled basis for arguing for a broader view of separability, that takes the doctrine beyond the concept of enforcement of the arbitration agreement and views the agreement as separable for all, or most, purposes? In this arena, the author's view is that the Court of Appeal's approach does carry the day. The judgment of Popplewell LJ provided a convincing and well-reasoned argument as to why the doctrine of separability should be viewed more broadly. Building on these reasons, the author adopts the same approach. Should the 1996 Act (and with it, section 7) be reformed, it is argued that now is the time to recognise, as the Departmental Advisory Committee on Arbitration (DAC) originally did, that the buck does not—and should not—stop with enforcement where separability is concerned.

II. SULAMéRICA

To set the decision of the Supreme Court (and, indeed, of the courts below it) in context, it is necessary first to consider the Court of Appeal's earlier decision in Sulamérica Cia Nacional de Seguros SA and others v Enesa Engelharia SA and others ⁴ and, in particular, the judgment of Moore-Bick LJ.

The case concerned a dispute between certain insured interests, Enesa and others, and their insurers, Sulamérica and others, arising out of certain policies of insurance connected with the construction of a hydroelectric generating plant in Brazil. The matter before the Court of Appeal concerned an appeal against an order of the Commercial Court granting the insurers’ application for an anti-suit injunction, to restrain the insured interests from pursuing proceedings before the courts in Brazil.

The policies of insurance were governed by the law of Brazil. As to dispute resolution, they contained both an exclusive jurisdiction clause in favour of the courts of Brazil and a London arbitration clause. The insurers relied on the existence of the arbitration clause to support their application for an anti-suit injunction. The insured interests’ position, however, was that, as a matter of the law of Brazil, the arbitration clause could only be exercised with their consent (which they had not provided). As a matter of English law, there was no such requirement.

This gave rise to an important question; namely: did the law of Brazil, or English law, govern the arbitration clause (as distinct from the underlying contract)? In circumstances where the policies did not contain an express choice of law clause specifically relating to the arbitration agreement, it was for the Court of Appeal to consider the appropriate test to determine that question.

In paragraph 25 of his judgment, Moore-Bick LJ set out the now well-known ‘Sulamérica test’, holding that the answer to the question of what the proper law of an arbitration agreement is involves two key propositions.⁵ The first is that the proper law of an arbitration agreement may not be the same as that of the substantive contract. This is so even if the agreement forms part of the substantive contract. The second proposition is the three-stage inquiry which constitutes the Sulamérica test itself. This inquiry involves looking, first, for an express choice, secondly for an implied choice and, failing that, looking for the law with which the arbitration agreement has the closest and most real connection.

As to the manner in which the inquiry under the Sulamérica test was to be undertaken (and of particular relevance to this article), Moore-Bick LJ held that in the absence of any indication to the contrary, there would be an assumption that the parties intended the whole of their relationship to be governed by one system of law.⁶ Thus, even though an express choice of law clause relating to the substantive contract might not be considered suitably worded enough to cover the arbitration agreement, it may well result in an implied choice of the same law.

In Sulamérica itself, the Court of Appeal acknowledged the strength of the argument in favour of Brazil serving as an implied choice of the proper law of the arbitration agreement,⁷ but was persuaded by two factors in particular to find that the proper law of the arbitration agreement was English law (being the law of the seat of the arbitration agreement). One was that the requirement, as a matter of the law of Brazil, for the insured interests’ consent before being able to exercise the arbitration clause undercut and was contrary to the broad wording of the arbitration clause itself.⁸

The other, and somewhat more curious, factor was the choice of a different procedural law (ie English law) to govern the arbitration.⁹ This, it was said, indicated that the parties intended that all matters pertaining to the arbitration agreement would be governed by the law so chosen. The reason why the author describes this aspect of the decision as curious is because, were one to adopt a broad view of separability, stemming beyond enforcement of the arbitration clause, this would make sense: if the arbitration clause is separable for all (or most) purposes, then a choice of law governing the substantive contract is arguably not as relevant as a choice to adopt a particular procedural law for the arbitration agreement itself. One relates to the arbitration agreement, whereas the other simply does not. Moore-Bick LJ, however, had expressly eschewed any suggestion that a broad view of separability was relevant or helpful to the discussion earlier in his judgment, stating as follows:

The concept of separability itself, however, simply reflects the parties' presumed intention that their agreed procedure for resolving disputes should remain effective in circumstances that would render the substantive contract ineffective. Its purpose is to give legal effect to that intention, not to insulate the arbitration agreement from the substantive contract for all purposes.¹⁰

Thus, separability was viewed by Moore-Bick LJ as a vehicle through which the parties’ presumed intention was to be given effect. Put another way, one assumes that where parties enter into a contract, they intend that even if their contract fails for some reason, the dispute resolution mechanism to which they agreed will continue to govern any disputes arising from the failure of the contract or the relationship underpinning it. That, according to Moore-Bick LJ, was the purpose and the limit of the separability doctrine. A different approach was adopted in ENKA İnşaat ve Sanayi A.Ş. v OOO ‘Insurance Company Chubb’ & Ors in the Court of Appeal, but not in the Supreme Court.

III. ENKA

The ENKA case concerned a fire that broke out at a Russian powerplant. This fire caused a great degree of damage to the powerplant. The damage to the powerplant was covered by a policy of insurance. After paying out to the insured, the insurer, a Russian entity called OOO ‘Insurance Company Chubb’, then sought to recover its losses from various subcontractors, including the Turkish company ENKA İnşaat ve Sanayi A.Ş. (which had been involved in the construction of the plant). Chubb brought these proceedings before its home courts in Russia. As with many of these sorts of cases, ENKA expressed its dissatisfaction with the forum chosen by Chubb by bringing proceedings in another forum; specifically, the Commercial Court in England. ENKA alleged that the dispute was governed by an arbitration clause providing for International Chamber of Commerce arbitration proceedings in London. To enforce this obligation, ENKA also sought an anti-suit injunction, the effect of which would have been to put a stop to the proceedings in Russia.¹¹

The main contract did not contain a governing law clause (although it was common ground that the law governing that contract was Russian law), and the arbitration clause did not contain an express choice of law either. Two important issues arose for consideration. The first was whether or not forum conveniens considerations are relevant to the English Court's decision to exercise its jurisdiction to grant an anti-suit injunction to put a stop to foreign court proceedings allegedly brought in breach of an arbitration agreement. The second question concerned the principles to be applied when determining the law governing an arbitration agreement as a matter of English law.

A. The Commercial Court

In the Commercial Court, Andrew Baker J dismissed ENKA's claim. He did so principally on the basis that the appropriate forum for determining the validity and scope of the arbitration agreement was Russia and not England.¹² As to the issue of the law governing the arbitration agreement, where separability is concerned Andrew Baker J noted that ‘the separability of the agreement to arbitrate is or may be an important concept (enshrined in this jurisdiction by s.7 of the Arbitration Act 1996), but it remains the case that there is a single contract, one term of which is that agreement with that particular characteristic’ and, further, that ‘[t]he separability of an arbitration agreement makes it a natural candidate for at least the possibility that it might be governed by a system of law different to that which governs the contract generally’.¹³ Equally, Andrew Baker J did not consider an express choice of seat to be relevant, describing it as not of any moment for the purposes of determining the proper law of the arbitration agreement.¹⁴

B. The Court of Appeal

The Court of Appeal overturned Andrew Baker J's decision. In doing so, it took a rather different approach to the question of the proper law of the arbitration agreement. This departure from Andrew Baker J's approach was justified in part by Popplewell LJ, with whom Males and Flaux LJJ agreed, by a need to re-examine the authorities in this area, which his Lordship described as doing ‘no credit to English commercial law’ and the requirement ‘to impose some order and clarity on this area of law’.¹⁵

The Court of Appeal achieved this clarity by dispensing with its previous guidance in Sulamérica,¹⁶ instead directing that the Court must first look at whether there is an express choice of law in the main contract and, as a matter of construction, determine whether that choice of law is sufficiently broad to cover the arbitration agreement. In the Court of Appeal's view, it would only be in a minority of cases that one could conclude that the express choice of law in the matrix contract was intended to cover the arbitration agreement as well.¹⁷ Indeed, in the majority of cases, there was a strong presumption that the law of the seat acts as an implied choice of law for the arbitration agreement.¹⁸

Separability was considered to play a very important role in determining whether such an express choice of law governing the arbitration agreement existed. In the Court of Appeal's judgment, there was no principled basis for treating the law governing the main contract as a significant source of guidance for determining the law governing the arbitration agreement in circumstances where the arbitration agreement was subject to a different curial law (ie the law of the seat).¹⁹ The stated reason for this was because of the existence of the doctrine of separability, and the fact that it sought to treat the arbitration clause as an agreement distinct from the main contract.²⁰ Thus, if the arbitration agreement were a separate agreement, there would be no reason to look to the governing law in the matrix contract to inform the governing law of that separate agreement.

The Court of Appeal was not willing to go so far as to say that the doctrine of separability would treat the arbitration agreement as separate for all purposes, but said that there were good reasons for doing so where the choice of law of the arbitration agreement was concerned.²¹ Specifically, where the parties have expressly chosen a different curial law, they have already gone some way to distancing the arbitration agreement from the matrix contract precisely because the express choice of law governing the contract would then have nothing to say about the curial law.²² In such circumstances, it was not a far leap to suggest that the closely related aspect of the choice of law governing the arbitration agreement itself (which would be used to determine matters such as the validity, existence and effectiveness of the arbitration agreement) would equally be separate from the main contract (and any choice of law under it).²³

The Court of Appeal also felt compelled to provide three further arguments in support of its analysis. The first was to address the argument that rational businessmen would not choose one law to govern the matrix contract, with another law governing the arbitration agreement. The Court of Appeal observed that, in this respect, this argument simply had nothing to say where such rational businessmen had already chosen more than one law to govern their relationship: the curial law and the law governing the underlying agreement.²⁴

Secondly, the Court of Appeal noted that the degree of overlap between the curial law and the law governing the arbitration agreement suggested that they should be the same. In particular, there were various provisions of the 1996 Act which took the curial law beyond simply governing the procedural rights of the parties, and moved it into affecting their substantive rights. By way of example, section 6 of the 1996 Act defines what, under the 1996 Act, constitutes an arbitration agreement. Thus, the curial law would determine this issue—and not the law governing the arbitration agreement itself.²⁵

Thirdly, it was held that where the issue was one of implied choice—as opposed to the closest and most real connection—if the curial law could determine the law governing the matrix contract, it would be anomalous if it could not also determine the law governing the separate arbitration agreement.²⁶

In every instance, the Court of Appeal's decision is strongly influenced by an understanding of separability that extends beyond the concept of enforcement. Separability is seen here as affecting—or influencing—the essence of what an arbitration agreement is, as opposed to a procedural device that allows the clause to continue to govern the resolution of disputes when the validity or existence of the contract in which it sits is called into question.

C. The Supreme Court

The Supreme Court, by a majority, reached the same conclusion as the Court of Appeal. It did so, however, by a different route and by rejecting the reasoning applied by the Court of Appeal in coming to its decision. In the majority decision, Lords Hamblen and Leggatt (with whom Lord Kerr agreed) held that, generally, where the parties have chosen the law governing the main contract, but have not chosen (expressly or impliedly) the law governing the arbitration agreement, the law governing the matrix contract would also apply to the arbitration agreement.²⁷ Lords Hamblen and Leggatt explained this conclusion on the basis that it would provide a degree of certainty, achieve consistency, avoid complexities and uncertainties, avoid artificiality and ensure coherence.²⁸

In coming to this conclusion, Lords Hamblen and Leggatt notably felt compelled to reject the Court of Appeal's interpretation of the doctrine of separability. This is unsurprising given, as stated above, that it formed a key—if not the central—part of the analysis which led the Court of Appeal to conclude that a choice of law in the matrix contract was largely irrelevant to determining the choice of law applicable to the arbitration agreement.

Although the majority was willing to accept that the doctrine of separability (i) certainly made the arbitration agreement more amenable to the application of a different governing law than other aspects of the matrix contract; and (ii) may well be relevant where applying the governing law of the main contract would render the arbitration clause ineffective, the majority thought that the Court of Appeal's view that the doctrine meant that a choice of law governing the matrix contract was largely irrelevant when viewing the law governing the arbitration agreement was to put the principle ‘too high’.²⁹ In particular, the majority did not agree that the separability principle generally meant that the arbitration agreement was to be viewed as a separate agreement when it came to applying the choice of law governing the matrix contract.³⁰ Thus, separability could be counted in the balance—but it did not serve to dictate, once and for all, the issue of the appropriate governing law. In so holding, the majority cited, with approval, Moore-Bick LJ's description of separability in Sulamérica,³¹ closely linking the doctrine to enforcement and the presumed intention that a dispute resolution mechanism would remain effective notwithstanding the invalidity of the matrix contract.³²

Having said all of this, the majority came to the same view as the Court of Appeal, ie that the law governing the arbitration agreement in this case was English law. The majority was able to do so because it held that although the law governing the main contract was Russian law, this was not an express or implied choice.³³ Thus, when looking at the arbitration agreement, one was required to ask with which law did the arbitration agreement have its closest and most real connection? Here, the express choice of seat as London was determinative: generally speaking (in the absence of choice), the law with which the arbitration agreement would be most closely connected was the curial law (here, English law).

Turning briefly to the minority, Lords Burrows and Sales agreed with the majority that an express or implied choice of law governing the main contract would generally apply to the arbitration agreement as well. They disagreed on the position where there was no express or implied choice of law, being willing to hold that, in those circumstances, the law governing the arbitration agreement would still be that of the law governing the underlying contract, as that was the law with which the arbitration agreement had its closest and most real connection.³⁴ In this respect, it is worth noting that Lords Burrows and Sales were more emphatic than the majority in their rejection of a wide-ranging or expansive view of separability. They noted that the ‘rationality’ of the doctrine was ‘to ensure that the arbitration agreement is effective despite the non-existence, invalidity, termination or rescission of the main contract’, that as was clear from section 7 of the 1996 Act the doctrine was devised for a specific (and limited) purpose, and that such purpose did not extend to working out the conflict of laws rules applicable to an arbitration agreement, with the result that, for that latter purpose, the arbitration agreement was to be regarded as a part of the main contract.³⁵ It is also notable that the minority expressly referenced the status of the doctrine as a legal fiction—an important matter to which this article must now turn.³⁶

IV. SEPARABILITY AS A LEGAL FICTION

The Swedish international lawyer and academic, Dr Gillis Wetter, once observed that the concept of separability ‘contains a large element of legal fiction’, being driven to a great extent by developed arbitral jurisdictions’ recognition that, without the doctrine, ‘the arbitral process would be ineffective’.³⁷ This observation is reflected in the minority judgment of the Supreme Court referred to above. The concept of a ‘legal fiction’ is important to the analysis of the doctrine of separability and, indeed, to its limitations.

What do we mean, then, when we describe a concept such as separability as a ‘legal fiction’? In his text, Legal Fictions, Lon Fuller began by describing a fiction as an ‘expedient, but consciously false, assumption’,³⁸ before going on to describe a legal fiction as either (i) ‘a statement propounded with complete or partial consciousness of its falsity’; or (ii) ‘a false statement recognised as having utility’.³⁹ A statement must be false before it can be a fiction, and consciousness of the falsehood is necessary to its utility (and renders it safe). As Fuller went on to say, ‘it is precisely those false statements that are realized as being false that have utility’.⁴⁰

The doctrine of separability falls into Fuller's second category; namely, a false statement that has some utility. Separability is a ‘legal fiction’ because, in the majority of cases, an arbitration clause is not a separate agreement. Rather, it is simply one clause in a much larger contract that was drafted by the same people, at the same time, with other clauses falling before and after it. However, without separability, and as Wetter noted,⁴¹ the process of arbitration would become ‘ineffective’. This is the utility aspect of the fiction: the arbitration agreement is treated as a separate agreement because to do otherwise would risk rendering the arbitration clause ineffective as soon as the validity of the main contract was called into question. In circumstances where the main contract is only likely to be called into question where there is a dispute, not treating the arbitration clause as ‘separable’ would have serious ramifications for dispute resolution mechanisms—and dispute resolution more generally.

V. THE LIMITS OF SEPARABILITY

Taking the decisions analysed above, with the exception of the Court of Appeal in ENKA, these authorities clearly limit separability's purpose to ensuring that the arbitration agreement remains enforceable, even if the contract of which the arbitration clause forms a part is invalid or unenforceable.

It is arguable that English law's focus on separability as a mechanism to ensure enforcement of the arbitration agreement to some extent stems from its initial resistance to the doctrine, as characterised by the decision of the House of Lords in Heyman v Darwins Ltd.⁴² In this case, Lord Wright famously declared that an arbitration agreement was ‘merely procedural and ancillary’,⁴³ with the result that its fate was determined with that of the main contract of which it formed a part. Some 50 years later, Steyn J (as he then was) went on to record that English law had moved to ‘a more advanced state’, recognising that an arbitral clause conferred on a tribunal the power to determine matters such as rescission, frustration and whether a contract was invalid.⁴⁴ Steyn J recognised, however, that there was further work to be done as ‘no English Court has yet been asked to take the final step of ruling that an arbitration clause, which forms part of a written contract, may be wide enough to cover a dispute as to whether the contract was valid ab initio’.⁴⁵ Two years later, the Court of Appeal completed this evolution of the English common law towards a complete doctrine of (narrow) separability in Harbour Assurance Co. (U.K.) Ltd. v Kansa General Insurance Co. Ltd.,⁴⁶ holding that a dispute as to whether a contract was void ab initio was capable of falling within an arbitration clause forming a part of the contract so impinged. The focus of these decisions, however, was very much on the procedural nature of the arbitration agreement, and the practical utility of ensuring that the arbitration agreement could be enforced.

Shortly after Harbour, English law gave separability a statutory footing, codifying it in section 7 of the 1996 Act. As was observed by Lords Burrows and Sales in ENKA, the 1996 Act closely ties separability to the concept of enforcement.⁴⁷ The 1996 Act expressly provides that the arbitration agreement shall be treated as a ‘distinct agreement’ but only ‘for that purpose’. The ‘purpose’ to which section 7 refers is found in the preceding part of section 7, which states that ‘an arbitration agreement which forms or was intended to form part of another agreement (whether or not in writing) shall not be regarded as invalid, non-existent or ineffective because that other agreement is invalid, or did not come into existence or has become ineffective’.⁴⁸

The use of the wording ‘for that purpose’ makes clear that the scope and limit of separability under the 1996 Act is to ensure that the underlying contract's invalidity does not affect the arbitration agreement's validity. Indeed, as noted in the DAC Report on the Arbitration Bill, section 7 was drafted ‘to make clear that the doctrine of separability is confined to the effect of invalidity etc of the main contract on the arbitration agreement, rather than being … a free-standing principle’.⁴⁹

When viewing separability as a matter of English law as a legal fiction, this qualification is important. As noted above, the essence of a legal fiction is to make a statement that we know to be false because it is useful in the particular context to do so. It follows from this that it is not appropriate or proper to extend beyond the particular use. In English law, the use or purpose of the doctrine of separability is defined in statute.⁵⁰ Viewed in this way, it is quite clear why the Court of Appeal's understanding of the doctrine in ENKA was unsustainable, and equally clear why the Supreme Court's approach (and, indeed, that of Moore-Bick LJ in Sulamérica) was not simply consistent with orthodoxy, but correct as a matter of English law. This is not to say, however, that the doctrine should be so limited as a matter of principle. This author's view is that, were English arbitration law to be reformed,⁵¹ a broader view of separability is merited. Arguments of principle are addressed in the section that follows.

VI. A BROADER VIEW OF SEPARABILITY

As can be seen from the judgment of Moore-Bick LJ in Sulamérica (and, indeed, in section 7 of the 1996 Act), the justification for the legal fiction of separability lies in the fact that an arbitration clause would not be of much use if it could be abandoned at the drop of a hat, as soon as one party calls into question the validity of the contract of which the arbitration clause forms a part. At that stage, it becomes convenient to treat the arbitration clause as a separate contract.

Having said all this though, does this mean that separability should only come into play when the enforcement of the arbitration clause is in jeopardy? Arguably not. There are several reasons why both consistency and legal certainty dictate that separability should extend beyond enforcement, and should have something to say about the arbitration clause more generally.

A. Consistency

The issue of consistency builds upon the point made by the Court of Appeal in ENKA; namely, that, as a matter of English law, the curial law does not simply make an impact upon issues of procedure in the arbitration. To the contrary, it can also affect issues of substance—such as providing the answer to the question: what is an arbitration agreement? Indeed, in the same way, whether an arbitration agreement is enforceable at all is not the sole preserve of the law governing the arbitration agreement. Certain provisions of the 1996 Act, designed to protect consumers, extend the curial law to the question of whether the arbitration agreement is enforceable at all. Thus, the combined effect of sections 89 to 91 of the 1996 Act⁵² and the Unfair Arbitration Agreements (Specified Amount) Order 1999⁵³ is that an arbitration agreement in the case of a domestic arbitration is unenforceable against a consumer where the amount in dispute does not exceed £5,000.

Where the curial law creeps into matters of substance, the argument that the curial law has a closer connection to the arbitration agreement than the law governing the main contract becomes quite compelling. Indeed, this is where consistency becomes important: instead of carving up different issues and attempting to determine whether the matter is governed by one law (the curial law) or another (the law governing the arbitration agreement, if different), it would make sense simply to say that the effect of separability is that the arbitration agreement is to be treated as a distinct agreement, with its own distinct governing law (and that the law with which the arbitration agreement has its closest connection is the law of the seat).

B. Legal Certainty

A separate—and distinct—issue is that of legal certainty. In ENKA, Popplewell LJ bemoaned the state of English law on the question of which law governs the arbitration agreement in the absence of an express choice. The decision of the Supreme Court, in effect, takes the state of English law back to that with which Popplewell LJ took issue; namely, the test in Sulamérica. However, it is easy to see why Popplewell LJ was critical of the state of the law, stemming from Sulamérica. In that case, whereas Moore-Bick LJ noted the importance of the doctrine of separability and the somewhat unique nature of the arbitration agreement, the approach taken was to promote a test that favoured the governing law of the main contract being the law with which the arbitration agreement was most closely connected. Nevertheless, and in the result, the arbitration agreement was held to be governed by English law, as the law of the seat. Similarly, in ENKA, the Supreme Court held that the law with which the particular arbitration agreement was most closely connected was the law of the seat—here on the different basis that the law governing the main contract, Russian law, was neither an express nor implied choice.

Were one simply to say that, in the absence of an express choice or clear and compelling reasons, the law governing the seat would also govern the arbitration agreement, legal niceties would be avoided and certainty would be promoted. If the only reason for saying that the law governing the main contract will be determinative is because separability has its limits, this is no reason at all. Separability itself is a fiction, and a respectable purpose of that fiction may well be to achieve legal certainty.

As to this point, it is worth noting that such an approach would not be particularly novel. It has already been adopted in the rules of one of the major arbitration institutions, the London Court of International Arbitration (LCIA) and just next door to England, in Scotland.

Taking Scotland first, the Arbitration (Scotland) Act 2010⁵⁴ provides under section 6 as follows:

Law governing arbitration agreement

Where—

1. (a) the parties to an arbitration agreement agree that an arbitration under that agreement is to be seated in Scotland, but

2. (b) the arbitration agreement does not specify the law which is to govern it, then, unless the parties otherwise agree, the arbitration agreement is to be governed by Scots law.

This, of course, means that the position in Scotland is very different to the position in England. Moreover, Scotland's equivalent of section 7 of the 1996 Act does not appear to limit separability to issues of enforcement, in that it simply provides that ‘[a]n arbitration agreement which forms (or was intended to form) part only of an arbitration agreement is to be treated as a distinct agreement’.⁵⁵

As to the LCIA, its 2020 Rules⁵⁶ make specific provision for the governing law of the arbitration agreement. Article 16.4 provides as follows:

Subject to Article 16.5 below, the applicable law to the Arbitration Agreement and the arbitration shall be the law applicable at the seat of the arbitration, unless and to the extent that the parties have agreed in writing on the application of other laws or rules of law and such agreement is not prohibited by the law applicable at the arbitral seat.

Article 16.5 goes on to provide that ‘[n]otwithstanding Article 16.4, the LCIA Rules shall be interpreted in accordance with the laws of England’. It is not entirely clear what this provision is intended to mean or signal,⁵⁷ but Article 16.4 at least is clear that unless the parties have agreed to apply a specific law to the arbitration agreement, it will be governed by the law applicable at the seat of the arbitration. This is also quite different from the ENKA position, and places at the forefront of the conversation the law at the seat, as opposed to the law governing the underlying contract. The Rules could only do so if they were to be based upon a view of the arbitration agreement as a separate—and separable—agreement.

As things currently stand, the Law Commission's consultation, proposing certain reforms to the 1996 Act, does not have as its focus the concept of separability. However, if there is a convincing argument (and this author believes that there is) that extending separability to cover the arbitration agreement more generally will serve to benefit legal certainty, the provisions referred to above demonstrate how straightforward, relatively speaking, any statutory amendment would be.

C. The Practical Benefits

In his recent Gaillard lecture,⁵⁸ Lord Hoffmann highlighted a particularly important benefit of linking the law governing the arbitration agreement to the seat of the arbitration (as opposed to the underlying contract): that of avoiding satellite disputes in arbitration. The difficulty with the approach in ENKA is that it requires the parties to look to the law governing the main contract, before they can determine the law governing the arbitration agreement. As any student of private international law will know, determining the law governing the underlying contract is not always straightforward (if, of course, the law is not set out in the matrix contract itself). The law of the seat is perhaps more straightforward, particularly in the context of international arbitrations. Several of the major institutional rules either provide for a default seat of the arbitration,⁵⁹ or make specific reference in their model clauses to the seat.⁶⁰ Separability promotes this linkage to the seat—because a broad view of separability does not view the choice of law of the underlying agreement (a different contract) as particularly or necessarily relevant to the law underlying the arbitration agreement. Separability, in this sense, promotes the practical benefit of minimising procedural disputes whereas, arguably, a narrow view of separability encourages such disputes.

D. Principle

A final reason for considering a more expansive view of separability is one of principle. Specifically, the risk of a narrow approach to separability is that it can—and does—lead to results that seem contrary to principle and difficult to defend. An illustration of this comes in the form of the Court of Appeal's recent judgment in DHL Project & Chartering Limited & Anor v Gemini Ocean Shipping Co Limited (Newcastle Express).⁶¹ The principal issue in dispute in Newcastle Express was a challenge to an arbitrator's jurisdiction under section 67 of the 1996 Act. The English Court had been asked to determine whether a charterparty that was expressly stated to be ‘subject shipper/receivers approval’ contained a binding arbitration agreement that conferred jurisdiction on the arbitrator to determine whether the charterparty had ever been concluded. The arbitrator determined that he had jurisdiction, whereas the High Court and the Court of Appeal determined that he did not.

The broader issue of principle in the case, addressed in the judgment of Males LJ in the Court of Appeal, was whether an arbitration agreement may be considered separable where questions of contract formation, as opposed to contract validity, are concerned. The Court of Appeal determined that the arbitration could not be considered separable where formation (as opposed to validity) of the matrix contract is in issue—in those circumstances, the arbitration agreement would stand or fall with the matrix contract.⁶²

The Court of Appeal's conclusion, and its reasoning, is problematic. A central aspect of the Court's reasoning was that the argument that ‘I never agreed to that’ must necessarily apply as much to the matrix contract as it does to the arbitration agreement.⁶³ To support this argument, the Court of Appeal placed much reliance on Steyn J's judgment in Harbour Assurance Co. (U.K.) Ltd. v Kansa General Insurance Co. Ltd. (in circumstances where the reasoning of Steyn J was approved by the Court of Appeal in that case).⁶⁴ It is questionable, however, whether that early judgment on the principle of separability can truly support an argument that issues of contract formation are necessarily outwith the scope of an arbitrator's jurisdiction.

To illustrate why, the author refers to the decision of the House of Lords in Fiona Trust and Holding Corporation & Ors v Yuri Privalov & Ors ⁶⁵ (which is also referenced in the Court of Appeal's judgment). The key passage from that case for present purposes (which is also quoted in Males LJ's leading judgment in Newcastle Express) is paragraph 17 of Lord Hoffmann's judgment. It is worth setting out that paragraph in full:

The principle of separability enacted in section 7 means that the invalidity or rescission of the main contract does not necessarily entail the invalidity or rescission of the arbitration agreement. The arbitration agreement must be treated as a ‘distinct agreement’ and can be void or voidable only on grounds which relate directly to the arbitration agreement. Of course there may be cases in which the ground upon which the main agreement is invalid is identical with the ground upon which the arbitration agreement is invalid. For example, if the main agreement and the arbitration agreement are contained in the same document and one of the parties claims that he never agreed to anything in the document and that his signature was forged, that will be an attack on the validity of the arbitration agreement. But the ground of attack is not that the main agreement was invalid. It is that the signature to the arbitration agreement, as a ‘distinct agreement’, was forged. Similarly, if a party alleges that someone who purported to sign as agent on his behalf had no authority whatever to conclude any agreement on his behalf, that is an attack on both the main agreement and the arbitration agreement.⁶⁶

This passage does not suggest that separability is irrelevant as soon as questions of contract formation are in issue. Quite the opposite. Where a person claims that their signature was forged, he is not saying that the contract is invalid: he is saying ‘I never agreed to that’. In the specific case of forgery, the claim that a party ‘never agreed to that’ equally affects the arbitration agreement, because a signature cannot be forged for one purpose and genuine for another. The reason why the matrix contract and the arbitration agreement never came into existence is the same. But, importantly, the agreements remain to be viewed as separate agreements. Separability, as it were, is the reason why the question (did a contract come into existence?) is asked twice. The reasoning in the Court of Appeal's judgment is problematic because, although it acknowledges the importance of asking the question twice,⁶⁷ it ignores it in the result. If the question of whether a contract came into existence can be asked twice, there seems to be no reason why the arbitrator is incapable of determining whether the matrix contract ever came into existence. It is also questionable whether section 7 of the 1996 Act necessitates the conclusion drawn in Newcastle Express, given that section 7 extends the principle of separability to situations where the matrix contract is ‘non-existent’. Nevertheless, section 7 undoubtedly adds to the confusion that results in decisions such as Newcastle Express, because it provides that separability applies for some purposes and not others. There would arguably be less confusion if a broader approach to separability were to be adopted.

VII. CONCLUDING REMARKS

Separability serves as a bedrock of modern arbitral law and practice. In its traditional form (at least in England), its role is limited to ensuring that the arbitration agreement remains enforceable in the face of an argument that the main contract of which it inevitably forms a part is invalid or otherwise unenforceable. To this extent, separability is a legal fiction. As a legal fiction, it must remain within the limits of its purpose, as defined. In England, that purpose is set out in the 1996 Act and, again, ties the doctrine to the concept of enforcement of the arbitration agreement itself.

Viewed against this backdrop, the decision of the Supreme Court in ENKA ⁶⁸ is unsurprising. Indeed, both the majority and the minority sought to reject a broader approach to separability, as had been advocated for by Popplewell LJ in the Court of Appeal below. But does this mean that separability must be so limited? Or put another way, were the 1996 Act to be reformed, must section 7 continue to limit separability to concepts of enforcement? The DAC originally thought not—and perhaps for reasons of legal certainty and consistency and, indeed, as a matter of principle, it may be appropriate to revisit that issue again.