To paraphrase Professor Henkin’s classic aphorism about international law – most parties respect most international arbitration agreements most of the time. And likewise, the international arbitral process works smoothly for most parties most of the time. Still, pathological cases arise in international arbitration, as in other contexts. Unfortunately, those aberrations command disproportionate attention, sometimes obscuring the underlying health and efficacy of the arbitral process.
One recent example of a pathological international arbitration involves the by-now infamous Dallah case. There, some ten years after a distinguished arbitral tribunal seated in Paris applied French principles of international arbitration law to conclude that the Government of Pakistan was bound by a contract, the UK Supreme Court reached the opposite result – applying the same principles of French law to deny enforcement of an arbitral award against Pakistan on jurisdictional grounds. Then, only months later, the Paris Court of Appeal reached the opposite conclusion – confirming the tribunal’s award and rejecting Pakistan’s jurisdictional objection under French law. This series of developments, involving two very eminent and experienced national courts, is very unhappy; those developments, and particularly the UK Supreme Court’s decision, contradict both the New York Convention and the objectives of the international arbitral process.
On its facts, the Dallah case is straightforward. Dallah Real Estate and Tourism Holding Company (“Dallah”), a Saudi Arabian company, entered into a Memorandum of Understanding in 1995 with the Government of Pakistan to provide housing in Saudi Arabia for Pakistani pilgrims to Mecca. Thereafter, various Pakistani Government ministers negotiated the terms of an agreement with Dallah to implement the Memorandum of Understanding. In connection with those negotiations, the President of Pakistan issued an ordinance establishing the Awami Hajj Trust (“Trust”), a separate legal entity with independent legal personality. In September 1996, the Trust entered into an agreement with Dallah (“Agreement”), containing the terms previously negotiated by the Government – including an ICC arbitration clause, but no choice of law clause. Despite its previous involvement in negotiations with Dallah, Pakistan was not a signatory to the Agreement.
The Agreement was ill-fated, lasting only four months or so. During that period, Pakistani Government ministers wrote to Dallah (on Government letterhead), addressing issues that had arisen under the Agreement and directing Dallah on how to perform the Agreement. In December 1996, three months after the Agreement was concluded, the Trust ceased to exist (because the Pakistani Government did not renew its existence). One month later, in January 1997, a Pakistani Government official wrote to Dallah (again on Government letterhead) purporting to terminate the Agreement. The Trust then sued Dallah for breach of the Agreement in Pakistani courts. Those courts eventually dismissed the Trust’s claims (on the basis that the Trust no longer existed) – after which Dallah commenced an ICC arbitration against Pakistan (in May 1998), seeking to recover substantial costs it had incurred in connection with the Agreement. Pakistan resisted, among other things, on grounds of jurisdiction.
In a jurisdictional award (in June 2001), an arbitral tribunal composed of Lord Michael Mustill, Nassim Hasan Shah and Ghaleb Mahmassani declared that Pakistan was bound by the arbitration clause in the Agreement. Sitting in Paris, the tribunal looked to French international arbitration law (which it characterized as incorporating “the transnational general principles and usages reflecting the fundamental requirements of justice in international trade and the concept of good faith in international business”). Applying these principles of French law, the tribunal held that Pakistan was the alter ego of the Trust and thus bound by both the Agreement and its arbitration clause. The tribunal subsequently made a final award (in June 2006) awarding Dallah $20 million plus legal costs.
Dallah sought enforcement of the award in England under the New York Convention and the English Arbitration Act, 1996, and, subsequently, also sought exequatur of the award in France. For its part, Pakistan resisted enforcement of the award in England, arguing under Article V(1)(a) of the Convention that there was no valid arbitration agreement between it and Dallah. In August 2009, exequatur of the award was granted in France, while the English enforcement proceedings lasted until November 2010, when the UK Supreme Court denied enforcement of the tribunal’s award ( UKSC 46). (See also my previous posts of 21 August 2009 and 12 April 2010.)
The UK Supreme Court’s decision denying recognition deserves careful attention. The Court is highly respected in the field of international arbitration, with its members enjoying distinguished reputations in international matters, commercial and otherwise, while its judgment in Dallah was clearly the product of careful thought.
The Supreme Court reached its decision to deny recognition of the arbitral award following what it termed an independent investigation of whether the tribunal had jurisdiction: “[the Court must] revisit the tribunal’s decision on jurisdiction” and is “neither bound nor restricted by” the tribunal’s conclusions. Explaining that it was applying French law reflected in Dalico and related French Cour de cassation decisions, the UK Supreme Court held that there had been no “common intention” for the Government of Pakistan to be a party to the arbitration agreement. Focussing narrowly on the formal signatories and literal terms of the Agreement, the Court found that “there was no material sufficient to justify the tribunal’s conclusion” that the Government was a party to the agreement to arbitrate.
Only months after the UK Supreme Court’s decision (on 17 February 2011), the Paris Court of Appeal rejected Pakistan’s application to annul the awards against it under Article 1502(1) of the French Code of Civil Procedure (Case No. 09/28533, 09/28535 and 09/28541). Like the UK Supreme Court, the Paris Court of Appeal applied French international arbitration law (looking to the parties’ “common intentions” under Dalico and similar decisions). But, unlike the UK Supreme Court, the Paris Court of Appeal had no difficulty in concluding that Pakistan had been intended to be a party to the Agreement. The French court took a broad, pragmatic view of the parties’ conduct, focussing in particular on the Government’s (sole) involvement in negotiating the Agreement, in implementing its terms and in terminating the Agreement: the Government “behaved as if the Contract was its own; … this involvement [of the Government], in the absence of evidence that the Trust took any actions, as well as [the Government’s] behaviour during the pre-contractual negotiations, confirm that the creation of the Trust was purely formal and that [the Government] was in fact the true Pakistani party in the course of the economic transaction.”
It remains to be seen how the French Cour de cassation will ultimately decide any appeal from the Paris Court of Appeal. Considering matters as they currently stand, however, the regrettable course of the Dallah case and conflict between the French and English decisions are pathological: they are contrary to both the purposes and specific terms of the New York Convention and they produce a potentially serious injustice. The most fundamental objectives of the Convention include ensuring uniform treatment of arbitral awards, and facilitating the effective enforcement of such awards, in the Convention’s Contracting States. Those goals are undermined when, a decade after an arbitral tribunal decides that parties concluded a binding agreement, courts in different Contracting States reach conflicting conclusions as to the correctness of the tribunal’s award – with a foreign court disagreeing with the courts of the arbitral seat over the application of its own law. For at least three reasons, outlined below, this is not what the drafters of the New York Convention intended, nor what parties concluding international commercial contracts and arbitration agreements intend.
First, the terms of the Convention provide a mechanism specifically designed to avoid conflicts between annulment and enforcement decisions. In particular, Article VI of the Convention (and section 103(5) of the English Arbitration Act) provides enforcement courts with authority to stay (or adjourn) decisions on enforcement pending the outcome of annulment proceedings in the arbitral seat. Dallah was a text-book example of a case where Article VI should have been applied: the Paris Court of Appeal was about to decide almost precisely the issue before the UK Supreme Court – an issue, moreover, that was governed by French law, that had already been decided by a French-seated arbitral tribunal and that involved an arbitral award that had already been granted exequatur by a first instance French court. In these circumstances, the arguments for staying English enforcement proceedings pending the French court’s decision were overwhelmingly powerful.
Article VI grants authority to recognition courts to stay enforcement actions in appropriate cases (where they “consider [it] proper”). This provides a mechanism, designed to further the Convention’s objectives of uniformity by avoiding conflicting decisions in different Contracting States, which is especially appropriate and useful in cases where the courts of the arbitral seat have particular competence on an issue. In Dallah, the decisive issue was one of French arbitration law (the law of the putative arbitral seat, indisputably applicable under Article V(1)(a)’s choice-of-law rules), which had already been decided by a French-seated tribunal; moreover, French courts had unique expertise on the relevant issues (of their own law), were seised of the issue and about to render a decision and they were undoubtedly neutral and objective. Basic principles of common sense and judicial prudence counselled that the UK Court should have awaited the imminent outcome of proceedings in the arbitral seat.
Despite this, the UK Supreme Court refused to grant a stay of recognition proceedings under Article VI, commenting in passing that “since Dallah has chosen to seek to enforce in England, it does not lie well in its mouth to complain that the Government ought to have taken steps in France.” Although the explanation is not entirely clear, it appears that the UK Supreme Court faulted Dallah for not having itself first sought exequatur in French courts before seeking recognition in England – hence, the UK Court’s refusal to stay English recognition proceedings pending Dallah’s exequatur action. That rationale ignores both the specific language and underlying objectives of Article VI – which aim to avoid exactly the conflicting decisions that Dallah produced. More fundamentally, the UK Supreme Court’s apparent rationale is impossible to reconcile with the New York Convention’s deliberate elimination of any requirement that award creditors obtain double exequatur (previously required under the Geneva Convention). Given that, an award creditor like Dallah is entirely free to seek to enforce its award abroad without first seeking exequatur in the arbitral seat. Contrary to the UK Court’s suggestion, Article VI of the Convention, and the policies of efficiency and uniformity it furthers, remain fully applicable in such circumstances.
Second, the New York Convention also provides that, in an enforcement proceeding, the party resisting enforcement bears the burden of proof, both under Article V(1)(a) and otherwise. This is made express in the introductory provisions of Article V(1) and is a fundamental element of the Convention’s basic purpose – again, specifically altering the position under the Geneva Convention. Although these principles are non-controversial, the UK Supreme Court’s decision in Dallah is very difficult to reconcile with them – with the Court instead imposing on the award creditor (Dallah) the burden of producing “material sufficient to justify the tribunal’s conclusion.” That holding misunderstands the Convention and its burden of proof: critically, it was not properly for Dallah to prove the existence of a valid agreement to arbitrate under Article V, but for Pakistan to disprove the existence of such an agreement.
Despite this, the UK Supreme Court declared that: “[t]he scheme of the New York Convention … may give limited prima facie credit to apparently valid arbitration awards based on apparently valid and applicable arbitration agreements, by throwing on the person resisting enforcement the onus of proving one of the matters set out in Article V(1) …. But that is as far as it goes in law. Dallah starts with advantage of service, it does not also start fifteen or thirty love up.” This reasoning misapprehends the meaning and purpose of Article V of the Convention.
Article V establishes the basic rule that it is for the party resisting recognition of an award to prove the applicability of an exception to the Convention’s general obligation (under Articles III and IV) to recognize foreign awards. Importantly, Article V prescribes a substantive burden of persuasion – not merely a procedural allocation of pleading roles – which can have significant consequences in many categories of cases. It is beyond debate that the Convention’s allocation of the burden of proof applies fully to all of Article V’s exceptions, including cases involving claims that there was no valid arbitration agreement under Article V(1)(a). Again importantly, this reverses the allocation of the burden of proof which exists at the stage of enforcing agreements to arbitrate, where the burden of proving the existence of such an agreement is on the party seeking to require arbitration.
The Dallah Court’s tennis analogy cannot be reconciled with the New York Convention’s allocation of the burden of proof in enforcement proceedings. Under Article V, Dallah does not merely enjoy the “advantage of service,” or even the advantage of a couple of points in one game of tennis. Rather, if the metaphor is to be pursued, Dallah had already won an entire match, which concluded with a presumptively valid arbitral award — and the question was whether the outcome of that match should be ignored in recognition proceedings. Under Article V, only where the award debtor (here, Pakistan) itself affirmatively proves that there is no valid arbitration agreement should this exceptional result be permitted. The UK Court’s conclusion that that Dallah had nothing more than the “advantage of service” and was required to provide material demonstrating the existence of a valid agreement to arbitrate is fundamentally contrary to both the plain language and obvious purpose of Article V.
Third, the Convention also requires Contracting States, like the United Kingdom, to apply the law specified in Article V(1)(a) to the validity of agreements to arbitrate – and not to apply their own local law to this issue. In Dallah, that law was indisputably French law (because, in the absence of a contrary choice by the parties, it was the law of the putative arbitral seat). Critically, however, the UK Supreme Court recited the words of the French international arbitration principles articulated in Dalico and other French decisions, but appears not to have applied the real substance of the French standards when evaluating the parties’ actual conduct and agreements.
That conclusion is confirmed by a comparison of the substantive analyses of the UK Supreme Court and of the French courts. Thus, the UK Supreme Court largely ignored factual elements that were central to the Paris Court of Appeal’s (and arbitral tribunal’s) decisions. In particular, the UK Supreme Court largely discounted the fact that, until the day before the execution of the Agreement, all negotiations and formal correspondence (on Government letterhead) took place exclusively between Dallah and the Pakistani Government. Similarly, the Supreme Court ignored the fact that the Prime Minister of Pakistan presided over meetings of the Trust (despite holding no position in it) and that it was the Pakistani Government which both created and later terminated the Trust – the latter act dissolving Dallah’s nominal contractual counter-party. And the UK Supreme Court devoted only passing attention to the fact that the Pakistani Government was actively involved in directing performance of the Agreement and formally terminating the Agreement (again, in correspondence from Government officials on Government letterhead), while the Trust had not been involved at all in either performance or termination of the Agreement. In contrast, the Paris Court of Appeal placed substantial weight on all these circumstances, holding that the Government’s actions both before and after conclusion of the Agreement could only be explained by its status as a real party to the Agreement.
One might debate the evidentiary weight of these various factors – though they point fairly decisively against the UK Supreme Court’s conclusions. The more fundamental point, however, concerns Article V(1)(a)’s choice of¬ law rule for the law governing the arbitration agreement – often described as one of the Convention’s crowning achievements. That rule requires Contracting States not merely to formally recognize foreign standards for the validity of arbitration agreements, but also to apply the substance of those standards in practice, just as the relevant foreign courts would do so.
In Dallah, it is difficult to avoid the conclusion that the UK Supreme Court ultimately failed to appreciate the substance of French law and – to an extent, understandably – applied what amounted to a classically English approach to contract law. Lord Mance hinted at this, referring with evident discomfort to the French standard: “It is difficult to conceive that any more relaxed test would be consistent with justice and reasonable commercial expectations…”
This discomfort is not surprising. The reluctance of English courts to consider precontractual negotiations – as contrasted to the approach in many other civil and common law jurisdictions – is familiar. Equally familiar is the English courts’ emphasis on express terms of contractual agreements and hesitations to embrace notions of good faith. Those rules are fair enough in English settings – indeed, that is why parties frequently agree to English law, applied by English courts or English-seated arbitral tribunals, to govern their commercial contracts. Critically, however, these are not rules of French law – and, as the competing decisions in Dallah illustrate, the application of these English approaches to contract law can, expressly or otherwise, produce very different results from those which obtain under French law. Under Article V(1)(a), it is essential that courts not merely apply just the words, but also the substance and spirit of the legal rules specified by the Convention’s choice-of-law standards. More fundamentally, the challenges of applying foreign law confirm the wisdom of Article VI’s provisions for stays of enforcement proceedings when annulment proceedings are underway in the courts of the country where the award was made.
In sum, the Convention did not contemplate a process that permits a jurisdictional objection to be relitigated effectively from scratch in a foreign court – ten years after the arbitrators’ jurisdictional decision and fifteen years after the events in question. Instead, like commercial parties, the drafters of the New York Convention intended that international arbitration be speedy, efficient and effectively enforced; the drafters of the UNCITRAL Model Law had similar objectives, including by requiring prompt challenges to jurisdictional awards (in Article 16(3)). Regrettably, Dallah does not achieve any of these objectives: it misapplies the Convention’s provisions on burden of proof, stays of enforcement and choice of law, producing a result that frustrates the most basic objectives of the arbitral process.
By Gary B. Born and Michal Jorek