The Singapore High Court applied this three-stage analysis in its recent decision in BNA v BNB and another  SGHC 142 (“BNA”). The Court in BNA also dealt with other intriguing legal issues. In particular, it affirmed the view that in the absence of an express choice of the proper law of the arbitration agreement, the implied choice of law should presumptively be the proper law of the underlying contract (and not the law of the seat). It also rejected the “Validation Principle” thesis advanced by Professor Gary Born that the court, in determining the law governing an arbitration agreement, should always apply the law that would validate the arbitration agreement instead of other potentially applicable choices of law that would invalidate the arbitration agreement. In this post, however, I will critically examine only the Court’s application of the three-stage analysis.
The decision in BNA
The dispute in BNA arose out of a contract that I will refer to as the “Agreement”. Article 14.1 identifies the law of the People’s Republic of China (the “PRC”) as the governing law of the Agreement. Article 14.2 provides that “[…] any and all disputes arising out of or relating to this Agreement […] shall be finally submitted to the Singapore International Arbitration Centre (SIAC) for arbitration in Shanghai, which will be conducted in accordance with its Arbitration Rules […]”.
The defendants commenced arbitration against the plaintiff pursuant to Article 14.2 and in accordance with the SIAC Rules 2013. A three-member tribunal was constituted. The plaintiff challenged the tribunal’s jurisdiction on the ground that the arbitration agreement is invalid under its proper law, being PRC law, because: (a) the dispute in question is considered a “domestic dispute” under PRC law, and PRC law prohibits foreign arbitral institutions from administering arbitrations of “domestic disputes”; and (b) PRC law also prohibits foreign arbitral institutions from administering PRC-seated arbitrations. The tribunal heard the jurisdictional challenge as a preliminary question, and determined (Ms Teresa Cheng SC dissenting) that it had jurisdiction because the seat of the arbitration is Singapore, the proper law of the arbitration agreement is Singapore law, and hence the arbitration agreement is valid. Thereafter, the plaintiff brought an application under s 10(3) of the Singapore International Arbitration Act (Cap 143A, 2002 Rev Ed) to seek the High Court’s de novo determination of the tribunal’s jurisdiction.
The Court held that the tribunal had jurisdiction over the dispute. In so finding, the Court determined, among other things, that:
- the three-stage analysis applied to determine the proper law of the arbitration agreement (at  and ); and
- in construing an arbitration agreement, the principles set out by the Court of Appeal in Insigma Technology Co Ltd v Alstom Technology Ltd  3 SLR(R) 936 (“Insigma”) should apply, and these include the principles that: (i) an arbitration agreement should be construed like any other form of commercial agreement; (ii) an arbitration agreement should be construed so as to give effect to a clear intention evinced by the parties to settle their disputes by arbitration; and (iii) a defect in an arbitration agreement should not render it void ab initio unless the defect is so fundamental as to negate the parties’ intention to arbitrate (at –);
- there was no express choice of the proper law of the arbitration agreement (at –);
- in the absence of an express choice of the proper law, the implied choice of the proper law should presumptively be the proper law of the underlying contract (at ,  and –);
- Singapore is the seat because: (i) the parties’ reference to the SIAC Rules 2013 in Article 14.2 meant that they have expressly agreed for the arbitration to be seated in Singapore, given that Rule 18.1 of the SIAC Rules 2013 states that in the absence of a contrary agreement by the parties or a contrary determination by the tribunal, the seat of any arbitration under the SIAC Rules 2013 shall be Singapore; and (ii) the parties’ reference to “arbitration in Shanghai” in Article 14.2 merely constitutes Shanghai as the venue for the hearings in the arbitration because Shanghai is a city (and not a law district) (at –);
- the law of the seat, being Singapore law, should displace the proper law of the Agreement, being PRC law, in order to constitute Singapore law as the parties’ implied choice of the proper law of the arbitration agreement, given that the parties’ arbitration agreement will be invalid under the proper law of the Agreement (at –); and
- Singapore law should in any event be considered the law that has the closest and most real connection with the arbitration agreement because Singapore is the seat (at  and ).
I respectfully suggest that the Court’s application of the three-stage analysis to the facts leaves much to be desired. In my view, the Court’s key error lay in constituting Singapore as the seat of the arbitration. Article 14.2 of the Agreement provides for all disputes to be submitted to arbitration in Shanghai in accordance with the SIAC Rules 2013. Rule 18.1 of the SIAC Rules 2013 states: “The parties may agree on the seat of arbitration. Failing such an agreement, the seat of arbitration shall be Singapore, unless the Tribunal determines, having regard to all the circumstances of the case, that another seat is more appropriate”.
The Court considered Article 14.2 to be a reference to two geographical locations (i.e., Singapore and Shanghai), as it found that “[t]he effect of the parties’ choice of the SIAC Rules  (and in particular Rule 18.1) […] is that they have expressly agreed both that Singapore should be the seat for their future arbitrations subject to contrary agreement and also that there should be ‘arbitration in Shanghai’” [emphasis in original] (at ). This construction of Article 14.2 is untenable for two reasons.
First, the reference to “arbitration in Shanghai” should have been considered a contrary agreement as to the seat as contemplated by Rule 18.1. The Court observed that “the parties’ arbitration agreement does not refer to the PRC, which is a law district, but to Shanghai, which is a city but not a law district”, and reasoned that “[w]here an arbitration agreement constitutes a law district such as Singapore expressly as the seat of any future arbitration, a reference in the same arbitration agreement to a geographical location which is not a law district is much more naturally construed as a reference to a venue rather than as a reference to a seat” [emphasis added] (at ).
It appears to me to be logically incongruent for the Court to assume that Article 14.2 expressly constitutes Singapore as the seat of the arbitration and rely on this assumption to bolster its construction of the reference to Shanghai as a venue (and not the seat) of the arbitration. By doing so, the Court has succumbed to tautology: the Court’s assumption that Article 14.2 constitutes Singapore as the default seat depended first on a finding that there is no contrary agreement as to the seat, but the Court inexplicably relied on that assumption to arrive at its finding of a lack of contrary agreement.
A plain reading of Article 14.2 would show that the reference to Shanghai in Article 14.2 is the only geographic location specified in the arbitration agreement. The Court in BNA itself noted that “if an arbitration agreement provides for any future arbitration to take place in a single geographic location, that location will be the seat of the arbitration unless the parties otherwise agree” (at ). That being the case, Shanghai should have been considered the seat.
Secondly, even if the reference to “arbitration in Shanghai” is not considered a contrary agreement on the seat, this should not necessarily lead to the conclusion that the default seat of Singapore will apply in the absence of party agreement on the seat. First, Rule 18.1 of the SIAC Rules 2013 states that the tribunal may still select a seat besides Singapore if it considers it appropriate, having regard to all the circumstances of the case. In other words, Rule 18.1 gives the tribunal a veto over the default choice of seat, such that the reference to the SIAC Rules 2013 without an express choice of seat cannot be taken to be an express selection of the default seat of Singapore. Further, the SIAC Model Clause (revised as of 1 September 2015) recommends that parties specify the seat of arbitration of their choice on top of stating that all disputes are to be referred to arbitration administered in accordance with the SIAC Rules. This shows that even the SIAC did not consider a reference to the SIAC Rules 2013 in an arbitration agreement to be sufficient in itself to constitute an express selection of Singapore as the seat in the absence of party agreement on the seat.
Accordingly, the Court should have held that the arbitration agreement, properly construed, constitutes Shanghai, the PRC, as the seat. That having been said, this would not have been the end of the inquiry. Finding that the arbitration agreement would be invalid under the proper law of the underlying contract is sufficient to rebut the presumption that the proper law of the underlying contract is the implied choice of the proper law of the arbitration agreement (BNA at ). Here, given that the proper law of the Agreement is PRC law, and the arbitration agreement would be ineffective under PRC law, the presumption that the proper law of the Agreement applies to govern the arbitration agreement is rebutted. But the law of the seat also may not apply as it too is PRC law. Hence, it simply cannot be inferred that the parties impliedly chose PRC law as the proper law of the arbitration agreement. The Court should have proceeded to the third stage of the three-stage analysis and considered with which system of law the arbitration agreement has the closest and most real connection. At this stage, even though the SIAC Rules 2013 are selected, the fact that the proper law of the Agreement is PRC law and the seat of the arbitration is Shanghai, the PRC, should lead the Court to conclude that PRC law is the system of law with which the arbitration agreement has the closest and most real connection.
The proper application of the three-stage analysis would have led to the unhappy result that the arbitration agreement is invalid. That should not however dissuade the Court from arriving at this conclusion, for it is simply the logical outcome of an inquiry that does not seek to divert parties to arbitration come what may, but instead has, as its sole objective, the ascertainment of the parties’ intentions in entering into the arbitration agreement. Ultimately, as the Court in BNA sensibly observed in its closing remarks, “there is only so much which the law can do to save an inapt and inept arbitration agreement” (at ). Courts should not be in the business of rewriting poorly drafted arbitration agreements in the name of giving effect to an intention to arbitrate where doing so would “result in an arbitration that is not within the contemplation of either party” (Insigma at ).
The author would like to express his gratitude to Foo Yuet Min for her helpful comments. All errors are the author’s alone.