The Supreme Court recently gave an important judgment (Reliance Industries Ltd v Union of India) on the implied exclusion of Part I of the Arbitration and Conciliation Act, 1996, for arbitration agreements that are governed by Bhatia International (ie, all agreements concluded before 6 September 2012). RIL v UoI is likely to be welcomed by the arbitration community as further evidence of the hands-off approach that the Supreme Court has adopted in recent times. In one sense, this is true: the view the Court takes in RIL, without doubt, mitigates some of the damage that Bhatia International did for agreements that are still subject to the pre-BALCO regime. But the Court, with respect, has made two mistakes in its analysis which are likely to prove troublesome in the future, generating yet more litigation about when Part I is impliedly excluded.
It may be helpful to begin by briefly recapitulating what the courts have said about this and why the problem remains despite BALCO. As readers of this Blog will know, Bhatia International decided that Part I of the 1996 Act applies even if the seat of arbitration is outside India, essentially on the ground that section 2(2) is not exhaustive. But in paragraph 32 of its judgment, the Court held that the parties may, by agreement (express or implied), ‘exclude’ all or any of the provisions of Part I. From the many post-Bhatia cases dealing with implied exclusion, only two propositions became relatively well-established: (1) the choice of a foreign proper law, in the absence of a foreign seat of arbitration, does not impliedly exclude Part I (Citation Infowares, Indtel Technical Services; and (2) the choice of a foreign seat of arbitration combined with the choice of a foreign lex arbitri or (perhaps) foreign proper law does impliedly exclude Part I (eg Dozco v Doosan).
In RIL v Union, the question was whether the Delhi High Court had jurisdiction to entertain a petition under section 34 challenging an award made by a London-seated Tribunal. The basis of the challenge was public policy and that certain disputes under a Production Sharing Contract (‘PSC’) were not arbitrable. Questions of arbitrability are normally resolved by applying the law governing the arbitration agreement (not lex arbitri). But it is important to notice that the question before the Supreme Court was jurisdictional, that is, ‘does section 34 apply’ and not ‘are these disputes arbitrable?’ The answer to the section 34 question depended on whether the PSC impliedly excluded Part I. In this context, three provisions in the PSC are of importance: (1) Article 32 provided that the PSC would be governed by Indian law and that nothing in the contract would entitle the Contractor ‘to exercise the rights, privileges and powers conferred upon it in a manner which will contravene the laws of India’; (2) Article 33.12 provided that the arbitration agreement (note: not the arbitration) would be governed by English law and that the ‘venue’ of the arbitration would be London; and (3) Article 33.6 provided that if the two arbitrators chosen by the parties could not agree on the appointment of the presiding arbitrator, the appointment would be made by the Secretary General of the Permanent Court of Arbitration at the Hague.
Nijjar J observes, at  and , that whether Part I is excluded in these circumstances depends on the intention of the parties, and that the choice of a foreign seat and the PCA appointment mechanism point towards implied exclusion. This is clearly correct. What is, with respect, wrong is the Court’s analysis of the law governing the arbitration agreement, both as a matter of principle and authority. Taking authority first, Nijjar J holds, as did Colman J in A v B, that the choice of the seat of arbitration is akin to an exclusive jurisdiction clause. There is much to be said for this view as a matter of principle (see below) but it is inconsistent with Supreme Court precedent. None of the cases dealing with implied exclusion hold that the mere designation of a seat impliedly excludes Part I: in fact, one could argue that the cases implicitly reject that proposition by relying on some additional factor (other than the foreign seat) to find that Part I does not apply. In other words, in all these cases involving foreign-seated arbitration clauses, the Supreme Court has said that Part I was excluded because there was also a foreign law governing the arbitration agreement (Videocon), the main contract (Dozco) or the conduct of the arbitration (Yograj). It is therefore doubtful whether the observation that a seat is akin to an EJC is correct as a matter of precedent, though it should of course be adopted (see below) should a larger Bench revisit this issue altogether. The second difficulty—as a matter of precedent—with RIL is its reading of Videocon, which it regards as indistinguishable. In Videocon, as we have noted, the Court thought that the parties had chosen English law as the lex arbitri when in fact they had chosen it as the law governing the arbitration agreement. Although  of that judgment cuts both ways, the better view is that Videocon decided (as a matter of law, though the particular facts were not properly analysed) that the choice of a foreign seat plus a foreign lex arbitri excludes Part I. Indeed, Videocon cited with approval a Gujarat HC decision which had taken exactly that view—Hardy Oil). In RIL, however, the Court takes Videocon to have decided that the choice of a foreign law governing the arbitration agreement excludes Part I, without considering why on that hypothesis it was necessary in Videocon to analyse Hardy Oil. For similar reasons, RIL’s analysis of what Yograj decided is also open to question: the effect of the then-current version of the SIAC Rules was to make the Singapore International Arbitration Act the lex arbitri, and it says little, if anything, about the law governing the arbitration agreement.
So much for authority. As a matter of principle, it is respectfully submitted that there are three errors in the case law culminating in RIL. First, the choice of a foreign seat is itself powerful evidence—as Nijjar J correctly observes—that the parties intend their ‘dispute resolution relationship’ to be governed by the law of that country. One could object that the mere choice of a foreign seat cannot be enough to attract the Para 32 exception in Bhatia because that would eviscerate the main rule. However, it is often overlooked that the contract in Bhatia had chosen the ICC Rules, which provided in article 23(2) that the parties could approach ‘any competent judicial authority’ for interim measures: so the choice of a Paris seat in that case was not determinative since the parties had also agreed that they could approach a court outside France for interim measures. Secondly, the Supreme Court appears to accept in Yograj that the choice of a foreign lex arbitri in addition to a foreign seat excludes Part I. This means that Part I is excluded if the parties explicitly choose ‘London seat, English Arbitration Act 1996 to govern the arbitration’. But this is an empty formality, because there is a strong (and rarely rebutted) presumption that the lex arbitri (not the law governing the arbitration agreement) is the law of the country of the seat. So if the Supreme Court is of the view that the choice of a foreign lex arbitri excludes Part I, it is hard to see why the choice of a foreign seat is not enough: surely nothing can turn on whether the foreign lex arbitri is itself chosen expressly or impliedly.
Thirdly, and most importantly, the erroneous analysis in this line of cases has curiously made the law governing the arbitration agreement decisive: this is what RIL decides and this is what RIL takes Videocon to have decided, whatever the facts of that case. This means, it appears, that Part I is not excluded if there is a London seat, but Indian law governing the main contract and the arbitration agreement (potentially even if there is an express choice of English lex arbitri). Yet that cannot be correct: there is no reason why parties should not arbitrate in one seat but under the substantive law of a foreign country, whether that substantive law governs the main obligations or the agreement to arbitrate. For jurisdiction, the crucial link is between the seat and the lex arbitri, not the seat and the law governing the arbitration agreement. Indeed, it is a matter of great controversy whether, in the absence of an express choice of law to govern the arbitration agreement, it follows the law of the seat or the law governing the main contract, and the most recent decision in England takes the view that it is the latter (Arsanovia v Cruz City).
In sum, it is respectfully submitted that the reason this area of law is in disarray is that the courts have not properly analysed the difference between the law governing the arbitration agreement and the lex arbitri, or what it means—as a matter of contract law—to say that the parties ‘impliedly’ intend something. Implication is indistinguishable from construction, as Lord Hoffmann has explained in a number of illuminating speeches, most notably in Mannai Investment and Belize Telecom. It is to be hoped that a sufficiently large constitution of the Supreme Court will revisit this area from first principle when the opportunity next arises.