Governing law of an arbitration agreement, and the awkward relationship between English and international law

Abstract

The Arbitration Bill, currently before Parliament,¹ provides an interesting case study in two respects. First, it provides an example, in the context of commercial arbitration, of how conflict of laws rules can be complex in theory and thereby defeat party expectations in practice. Clause 1 of the Arbitration Bill introduces a new default rule, identifying the governing law of an arbitration agreement, which seeks to be clearer and more predictable than the current approach of the common law. And yet clause 1 disapplies its new default rule to the sub-set of investor-state arbitrations. This is the second aspect of the case study. The need to exclude investor-state arbitration from clause 1 reveals the continuing awkward relationship between English law and international law.

The international context

Arbitration is a popular choice for resolving disputes arising out of international commercial contracts. For example, an Indian company and a Chinese company might agree to build a factory in India. If there is a dispute, litigation before the courts in India or China might not appeal to one of the companies, fearing that the other might have a home advantage. Arbitration provides the companies with the opportunity to arbitrate at a neutral seat.²

London is the world’s most popular choice of seat for international commercial arbitration.³ Indeed, international arbitration was a feature of London business as early as the fourteenth century.⁴ What makes international arbitration especially appealing is that, by virtue of the New York Convention,⁵ an arbitral award can be enforced like a court judgment in any state signatory to the Convention. The New York Convention has been ratified by 172 states worldwide,⁶ and one study suggests that arbitral awards are successfully enforced around 73 per cent of the time.⁷ This regime tends to be simpler, more reliable, and more universal than trying to enforce court judgments in foreign countries.

Governing law of an arbitration agreement

An arbitration agreement is usually a clause in a main contract. For example, in a commercial contract to build a factory, one of the clauses might provide that any dispute arising out of the contract will be resolved by arbitration. Despite being a clause in the main contract, an arbitration agreement is deemed to be separable,⁸ or perhaps even separate,⁹ from the main contract.

It is usual for an international commercial contract to contain a governing law clause. It is usual for an arbitration agreement to designate a seat. It is not usual for arbitration agreements to have a governing law clause.¹⁰ The explanation often given is that commercial parties, when negotiating a contract, are not always aware of the technical rule that an arbitration clause is separable and so needs separate provision in addition to the choice of a seat.

It matters which law might govern an arbitration agreement. The governing law will determine who is party to the agreement. It will address questions like arbitrability, that is, what matters can be arbitrated, and scope, that is, whether this dispute falls within the arbitration clause.¹¹

So which law governs the arbitration agreement? One study suggests that, across 80 jurisdictions, 51 per cent have a rule that applies the law of the seat, 34 per cent apply the law of the main contract, 9 per cent adopt a validation approach (like Switzerland),¹² and 6 per cent adopt an approach not aligned with national law (like France).¹³ In England and Wales, governing law currently falls to be determined by the common law. The leading case is the Supreme Court decision in Enka v Chubb,¹⁴ which says as follows.

The arbitration agreement will be governed by the law chosen by the parties,¹⁵ unless that choice is contrary to public policy.¹⁶ If there is no express choice of law for the arbitration agreement, then there is an ‘inference’ or rebuttable ‘presumption’ that the law of the main contract will ‘generally’ apply to govern the arbitration agreement as well.¹⁷ This may be displaced in three situations. First, where the law of the seat provides that the arbitration agreement is governed by the law of the seat.¹⁸ Second, where there is a risk that the identified law might render the arbitration agreement invalid,¹⁹ or not binding on one party,²⁰ or of reduced scope²¹ (this is known as the validation principle). Third, where the choice of England as a seat, in combination with a reference to a local association or practice, implicitly indicates the choice of English law as the governing law.²² Otherwise, the arbitration agreement will be governed by the law with which it has the closest and most real connection.²³ The closest connection, said the majority in the Supreme Court, was the law of the seat,²⁴ but perhaps still subject to the validation principle.²⁵

Some authors endorsed the decision in Enka v Chubb,²⁶ while others criticized it for lacking clarity or certainty.²⁷ The complexity of the principle leads to unpredictability in its application to the facts: it is notable, for example, that the Supreme Court’s decision was split,²⁸ and differed from the decision of the Court of Appeal,²⁹ which in turn differed from the decision of the judge at first instance.³⁰

Clause 1 of the Arbitration Bill provides a new rule: unless the parties expressly agree otherwise, the law of the seat will govern the arbitration agreement. Departing from Enka v Chubb, clause 1 also says that a choice of law to govern the main contract does not count as an express choice of law to govern the arbitration agreement.

This new approach was the recommendation of the Law Commission,³¹ partly because it was clearer, but mainly for the following reason. If London is chosen as the seat of arbitration, then the expectation is that arbitration will happen in London. If the arbitration agreement is governed by English law, that will make it likely that the arbitration will take place in London, because English law is supportive of arbitration. If instead a foreign law governs the main contract, as is often the case, and so also is deemed to govern the arbitration agreement, that foreign law might be less supportive of arbitration, and in some cases will preclude it entirely (as acknowledged by the validation principle).³² In those cases, it would be unsatisfactory if an express choice and expectation of London as a seat of arbitration was nullified by an implied or presumed or inferred choice of foreign governing law. Instead, the new default rule in the Arbitration Bill will likely ensure that, where London is the chosen seat of arbitration, that express choice will be supported by having English law govern the arbitration agreement.

There are two further points to make. First, the new default rule need not be seen as undermining party autonomy.³³ It allows the parties an express choice of law to govern the arbitration agreement. True, it does not allow an implied choice of governing law, but that is to ensure effect is given to the express choice of arbitration in London. It should not be the case that an implied choice (of governing law) ousts an express choice (of arbitration in London). The new default rule also provides greater clarity and predictability, which are other values additional to autonomy which are appreciated by commercial parties.

Second, what happens if the parties have not agreed on a seat for their arbitration? Clause 1 of the Arbitration Bill does not address this situation, and so it seems that the common law and Enka v Chubb would continue to apply. However, in practice, this situation is likely to be rare. Even if the parties have not chosen a seat, some arbitral rules specify a seat,³⁴ or provide for the seat to be determined by the tribunal,³⁵ or by the arbitral institution.³⁶ Under the Arbitration Act 1996,³⁷ a seat can be designated by the parties, or the tribunal, or an arbitral institution, or ‘otherwise determined’, presumably by the court.³⁸ In other words, by the time any absence of seat becomes contentious, there is likely to be an authority which can impose a choice. At that point, clause 1 of the Arbitration Bill would apply, determining the governing law, even retrospectively.³⁹

The question then arises as to why investor-state arbitration should be treated differently.

About investor-state arbitration

If state X wishes to encourage foreign investment from the nationals of state Y, then X and Y might enter into a bilateral investment treaty (BIT).⁴⁰ Each state makes promises about how it will treat investors who are nationals of the other state. What if a state breaks its promises? One state might sue the other state for breach of the BIT. Also, the BIT often includes a standing offer of arbitration to investors, so that if an investor has a complaint, they can sue the host state (the state where the investment was made). The standing offer of arbitration in the BIT is made by the host state, and it is accepted by the investor by their act of bringing an arbitration claim. Together this offer and acceptance counts as an arbitration agreement in writing.

The International Centre for Settlement of Investment Disputes (ICSID) Convention⁴¹ created the International Centre for the Settlement of Investment Disputes. Its purpose was to provide a neutral (‘de-localized’) venue and procedure to settle investor-state disputes by arbitration in Washington. It has been ratified by 158 states.⁴² For those states which have not ratified the ICSID Convention, there is also the ICSID Additional Facility, which allows non-signatories to use ICSID on an ad hoc basis.

Not all investor-state arbitrations are ICSID arbitrations. Rather, it depends on the terms in the BIT. BITs can offer investors a choice of forum which often includes:⁴³ sue in the courts of the host state; or bring arbitration proceedings under ICSID; or bring institutional arbitration proceedings under the (Swedish) Stockholm Chamber of Commerce (SCC), or the International Chamber of Commerce (ICC) or other set of institutional rules; or bring ad hoc arbitration proceedings under UNCITRAL Arbitration Rules.⁴⁴

In English law, the ICSID Convention is given effect by the Arbitration (International Investment Disputes) Act 1966. Section 3 expressly disapplies the provisions of the Arbitration Act 1996,⁴⁵ so that these two regimes are separate.⁴⁶ The Arbitration Bill only amends the 1996 Act. However, a non-ICSID investor-state arbitration is treated like any other commercial arbitration and is governed by the Arbitration Act 1996.⁴⁷ This means that clause 1 of the Arbitration Bill, and its default rule on governing law, would apply to non-ICSID investor-state arbitrations, unless there was an exception.

Governing law of investor-state arbitration agreements

There is an important distinction, as ever, between which law governs the merits of the arbitration (the substantive dispute), and which law governs the arbitration agreement itself.

The law which governs the merits often includes the provisions of the BIT itself, the domestic law of the host state, and general principles of international law.⁴⁸ It might also be necessary to consider domestic law as the background context for interpreting what the BIT is intended to address.⁴⁹ This is a complex picture, and necessarily so. For example, an investor-state dispute might involve the following issues.⁵⁰ Is the claimant a national of the other state party to the BIT? This will involve the domestic law of the other state. Has the investor made an investment in the host state? This might involve the domestic law of the host state (for example, questions of property law).

Furthermore, institutional arbitral rules often provide a wide discretion to a tribunal to decide, in default of agreement by the parties, which law governs the merits. For example, ICSID provides that the tribunal shall decide a dispute in accordance with such rules of law as may be agreed by the parties; and in the absence of such agreement, the tribunal shall apply the law of the host state and such rules of international law as may be applicable.⁵¹ UNCITRAL Arbitration Rules provide that in default of agreement by the parties, the tribunal will apply whatever law the tribunal considers appropriate.⁵² The same approach is taken by the ICC Arbitration Rules,⁵³ and the SCC Arbitration Rules.⁵⁴ What is more, international law might have little to say on some topics, like contract rights. A tribunal needing to decide such issues might need to look beyond international law for guidance—although it might find such guidance from other international sources like UNIDROIT.⁵⁵

However, when it comes to the law governing the arbitration agreement, this will always be international law, at least in the first instance. The arbitration agreement comes into existence when the standing offer to arbitrate is accepted by commencing arbitration according to its terms. The standing offer is contained in the treaty. Treaties are governed by international law. In other words, interpretation of the standing offer of arbitration means adopting the approach set out in the Vienna Convention.⁵⁶ The approach of the Vienna Convention is not the same, for example, as the approach an English court might adopt when interpreting a UK statute or a commercial contract.

The treaty might make reference to national law, saying that national law is applied in some circumstances. When one law re-directs to another law, that is called ‘renvoi’. And so the discussion in Dicey, Morris & Collins says that BIT arbitration agreements are governed by international law,⁵⁷ but can include renvoi to national law.⁵⁸ However, whether there is renvoi to national law depends upon the interpretation of the words of the treaty, and that interpretation is governed by international law. This is why international law always applies at least in the first instance.

The need for an exception

Clause 1 of the Arbitration Bill provides a default rule that the law of the seat will govern the arbitration agreement. If an investor-state arbitration is seated in London, then English law would govern the arbitration agreement. And yet, as we have seen, international law should govern that arbitration agreement, at least in the first instance. That would seem to reveal the need for an exception in clause 1 for non-ICSID investor-state arbitrations. After all, ICSID investor-state arbitrations are already governed by a separate regime. But wait—do the English courts already have a practice of applying international law to the interpretation of treaties? If so, then perhaps no exception is needed. Similarly, legislation in Scotland,⁵⁹ and in Sweden,⁶⁰ applies a default rule that an arbitration agreement will be governed by the law of the seat, without any exception for investor-state arbitration agreements.

Two arguments have been advanced against any need to except investor-state arbitration agreements from clause 1 of the Arbitration Bill. An analysis of those arguments shows how there continues to be an awkward relationship between English law and international law.

The first argument against the need for an exception says that English law already has a conflict of laws rule such that agreements derived from a treaty are governed by international law. This was the approach favoured by the Court of Appeal in Ecuador v Occidental Exploration and Production Co,⁶¹ where the court suggested that treaty-derived investor-state arbitration agreements are governed by international law. That approach in Ecuador v Occidental has been followed in subsequent cases.⁶²

Applying that approach to clause 1 of the Arbitration Bill, the reasoning would look like this. Clause 1 provides that the law of the seat governs an arbitration agreement. Where an arbitration is seated in London, English law will govern the arbitration agreement. English law has a conflict of laws rule which provides that treaty-derived investor-state arbitration agreements are interpreted according to international law. Thus, there is no need to exclude investor-state arbitration from clause 1.

However, this approach does not enjoy universal support as a matter of principle. For example, it has been said that treaties are governed by international law as the only regime ever applicable, and not because a domestic regime of conflict of laws allows it.⁶³

Furthermore, the reasoning above runs into practical difficulties. Clause 1 of the Bill applies a default rule such that English law governs the arbitration agreement. This must mean English law rather than English conflict of laws rules, otherwise it leads to circularity, as follows. An arbitration agreement is interpreted according to its governing law. Governing law is determined by conflict of laws rules. Those rules, currently including Enka v Chubb, ask whether the governing law is designated by a statute. The Bill would designate English law as the governing law. If that means English conflict of laws rules, those rules ask whether the governing law is designated by a statute… To break that loop, the Bill must mean English law other than English conflict of laws rules. A similar rule against circularity can be found in Dicey, Morris & Collins.⁶⁴

The second argument against the need for an exception says that English law has a rule of interpretation (rather than a conflict of laws rule) such that treaties are interpreted according to the principles of international law found in the Vienna Convention.⁶⁵ Alternatively, the Vienna Convention has been said to codify customary international law,⁶⁶ and customary international law can be adopted into English common law. In other words, if English law were to govern an investor-state arbitration agreement, English law would interpret that agreement by applying principles of international law.

Now, when English courts are faced with a justiciable issue which requires the interpretation of a treaty or convention, the courts do indeed tend to apply the Vienna Convention.⁶⁷ This is an approach repeatedly evidenced or acknowledged in the Supreme Court.⁶⁸ But here is the problem: constitutionally, international law is not a directly applicable part of English law. Treaty international law must be incorporated by statute,⁶⁹ and customary international law must be adopted by the common law. It was once the view that customary international law was part of English law, but the modern view is that customary international law is not part of English law, although it can guide the development of common law.⁷⁰ In other words, customary international law becomes part of English law only when a court decision adopts it into the common law.

The Vienna Convention is not itself directly a part of English law: the treaty has not been incorporated into English law by statute. And to the extent that the Vienna Convention does codify customary international law, that customary international law is not directly applicable either, but requires the common law to develop its own matching rule. A matched rule still suggests a parochialism—a need to apply domestic law—when the courts want to apply international law. Furthermore, how precisely a domestic law matches international law is unpredictable and may change over time, and anyway, the courts cannot develop an English rule to match international law if that would be inconsistent with English legislation.⁷¹ In that regard, clause 1 of the Arbitration Bill, which references domestic law (the law of the seat), might be read as inconsistent with applying international law after all.

It may be that we are moving in a direction whereby the English courts will be able to apply international law directly in appropriate cases,⁷² but we are not there yet. Instead, when the English courts apply the Vienna Convention, they tend to leave unaddressed the precise mechanism they are using to achieve that outcome.⁷³ Arguably, this produces the right result, but it is a fudge. The exception to clause 1 of the Arbitration Bill allows that fudge to continue, by excluding investor-state arbitration agreements from the new regime. This is necessary because a stark legislative prescription to apply English law cannot yet provide a constitutional or logical guarantee that international law questions will be properly handled.

Conclusion

Parties choose international arbitration because it provides a neutral venue for dispute resolution and an award which is readily enforceable worldwide. The law which governs the arbitration agreement will determine important questions like who is party to the arbitration agreement, and whether the dispute is arbitrable. The common law approach to identifying the governing law, set out in Enka v Chubb, is complex and unpredictable, as shown in part by the differing judicial approaches in Enka v Chubb itself. It also risks the practical consequence of defeating party expectations by allowing an implied choice of foreign governing law with restrictive attitudes to arbitrability to defeat an express choice to arbitrate.

Clause 1 of the Arbitration Bill contains a new default rule such that the arbitration agreement will be governed by the law of the seat of the arbitration. London is the world’s most popular choice of seat for international arbitration, so where the parties choose to arbitrate in London, the default rule means that the arbitration agreement will be governed by English law. The default rule has the virtue of clarity, and it will increase the likelihood that the parties’ express choice to arbitrate in London will be honoured because English law is supportive of arbitration.

Clause 1 contains an exception for arbitration agreements, such as investor-state arbitration agreements, which derive from standing offers to arbitrate contained in treaties. The terms of those arbitration agreements are set out in the treaty, and so should be governed by international law. If clause 1 were to apply to those arbitrations too, then if an investor-state arbitration were seated in London, the arbitration agreement would be governed by English law. Two arguments, with case law to evidence them, have been advanced as to why that should not be a problem, but an analysis of both arguments reveals the awkward relationship between English law and international law. Specifically, the argument that English law has a conflict of laws rule that applies international law to treaties fails logically because it is circular. The argument that English law has a rule of interpretation that applies international law to treaties reaches the right result but in a way which is constitutionally unsound. What the case law really evidences is a fudged solution to a persistent problem. The exception in clause 1 of the Arbitration Bill allows that fudge to continue.

Part of the virtue of clause 1 is the clarity it brings to the question of governing law. But where international law is at issue, clarity, and the strict consequences it demands, risks upsetting a useful compromise of principle—hence the exception in clause 1 for treaty-derived arbitrations.

Footnotes