1 The plaintiff brings this application under s 10(3) of the International Arbitration Act (Cap 143A, 2002 Rev Ed) (“the Act”). By this application, the plaintiff invites the court to declare ¹ that a three-member tribunal appointed by the Singapore International Arbitration Centre (“SIAC”) to arbitrate a dispute between the defendants and the plaintiff lacks the jurisdiction to do so.

2 The parties’ dispute arises out of a contract which the plaintiff entered into with the first defendant in 2012. That contract is known as the Takeout Agreement. ² In 2013, ³ all three parties agreed to modify the Takeout Agreement by entering into an addendum to it. Under the addendum, the second defendant took over from the first defendant all of the first defendant’s obligations to the plaintiff under the Takeout Agreement. The result of the addendum was that the first defendant immediately ceased to owe any primary performance obligations to the plaintiff under the Takeout Agreement. But the addendum expressly provided that the first defendant was nevertheless to be liable to the plaintiff, jointly and severally with the second defendant, for any failure by the second defendant to perform its newly-acquired obligations under the Takeout Agreement.

3 The critical provision of the Takeout Agreement for the purposes of this application is Article 14. Article 14 serves two purposes. First, it records the parties’ express choice of the law of the People’s Republic of China (“the PRC”) to govern the Takeout Agreement. Second, it contains the parties’ arbitration agreement. Article 14 provides as follows: ⁴

4 The addendum expressly provides that it constitutes an indivisible part of the Takeout Agreement, which is to remain in full effect to the extent that the addendum has not modified it. ⁵ The result is that Article 14 governs a single indivisible agreement comprising both the Takeout Agreement and the addendum. The addendum thus turned the Takeout Agreement in general, and Article 14.2 in particular, from a bipartite agreement between the plaintiff and the first defendant into a tripartite agreement between the plaintiff and both defendants. The addendum also extended the scope of Article 14.2 to encompass disputes not only under the original Takeout Agreement but also under the addendum. I shall henceforth use the term “Takeout Agreement” to refer to the single indivisible agreement comprising both the original Takeout Agreement and the addendum.

5 The parties’ arbitration agreement manifests the parties’ unambiguous and unqualified intention to arbitrate “any and all disputes arising out of or relating to” the Takeout Agreement. The plaintiff’s case on this application, however, is that the tribunal lacks jurisdiction to resolve all such disputes because the arbitration agreement is invalid under its proper law, being PRC law. The question before me is therefore whether the parties nullified their manifest intention to arbitrate disputes under the Takeout Agreement by making PRC law its proper law.

6 The defendants commenced the underlying arbitration in 2016 by lodging a notice of arbitration ⁶ against the plaintiff with the SIAC under the Arbitration Rules of the SIAC (5th Edition, 2013). ⁷ In its response to the notice of arbitration, ie from the very outset of the arbitration, the plaintiff challenged the tribunal’s jurisdiction. ⁸

7 The SIAC appointed a tribunal comprising three arbitrators: Mr Hee Theng Fong as chair with Mr Philip Yang and Ms Teresa Cheng SC on the wings. The tribunal gave directions for the plaintiff’s jurisdictional challenge to be determined. ⁹ The parties exchanged written submissions on the issue of jurisdiction. ¹⁰

8 In due course, the tribunal handed down its decision on jurisdiction. A majority of the tribunal, comprising Mr Hee Theng Fong and Mr Philip Yang, held that the tribunal had jurisdiction in the arbitration because: (i) the arbitration is seated in Singapore; (ii) the arbitration agreement is thereby governed by Singapore law; and (iii) PRC law is therefore irrelevant on the question of jurisdiction. ¹¹ Ms Teresa Cheng SC, dissenting, held that the tribunal lacked jurisdiction because: (i) the proper law of the parties’ arbitration agreement is PRC law; (ii) the parties’ dispute is classified in PRC law as a domestic dispute; and (iii) PRC law prohibits a foreign arbitral institution from administering the arbitration of a domestic dispute. ¹²

9 The plaintiff now applies under s 10(3) of the Act to have it declared, contrary to the majority decision of the tribunal, that the tribunal has no jurisdiction in the arbitration. The plaintiff makes this application in Singapore because the tribunal has decided, albeit by a majority, that the arbitration is seated in Singapore. The parties have submitted to the jurisdiction of the courts of Singapore for the purposes of this application. ¹³

10 It is common ground between the parties that I am to determine the question of jurisdiction afresh, by undertaking a hearing de novo. An application under s 10(3) of the Act is neither an application to set aside the tribunal’s majority decision on jurisdiction nor an appeal against that decision. In order to succeed on this application, therefore, the plaintiff need not satisfy the grounds for setting aside an award specified in the Act or even show that the tribunal fell into error in its majority decision. As Steven Chong J (as he then was) said in BCY v BCZ [2017] 3 SLR 357 (“BCY”) at [36]:

11 The plaintiff’s case on this application is that the tribunal lacks jurisdiction to arbitrate the parties’ dispute because their arbitration agreement is invalid under its proper law. ¹⁴ The plaintiff’s case proceeds as follows: ¹⁵

(a) PRC law is the proper law of the parties’ arbitration agreement.

(b) PRC law classifies the parties’ dispute as a domestic dispute because it does not satisfy PRC law’s “foreign elements” test. PRC law prohibits a foreign arbitral institution such as the SIAC from administering the arbitration of a domestic dispute. The parties’ arbitration agreement is therefore invalid under its proper law.

(c) Alternatively, even if the parties’ dispute does satisfy PRC law’s “foreign elements” test, PRC law prohibits an arbitration which has its seat in the PRC from being administered by a foreign arbitral institution such as the SIAC The parties’ arbitration agreement provides that the seat of any arbitration between the parties is to be Shanghai. Accordingly, on this alternative ground also, the parties’ arbitration agreement is invalid under its proper law.

The plaintiff therefore submits that the only dispute-resolution mechanism open to the defendants is litigation in the courts of the PRC. ¹⁶

12 The defendants’ case on this application is that the tribunal has jurisdiction to arbitrate the parties’ dispute because the arbitration agreement is valid and, insofar as PRC law is relevant, does not contravene PRC law. The defendants’ case proceeds as follows: ¹⁷

(a) The seat of the arbitration is Singapore and not the PRC;

(b) The parties’ implied choice as the proper law of their arbitration agreement is Singapore law and not PRC law; and

(c) It is common ground between the parties that, if the proper law of the parties’ arbitration agreement is Singapore law, the arbitration agreement is valid and the tribunal does have jurisdiction. ¹⁸

13 Four critical questions therefore lie at the heart of this application. First, what is the proper law of the parties’ arbitration agreement? Is it PRC law, given that that is their express choice to be the proper law of the Takeout Agreement? Second, what is the seat of the parties’ arbitration? Is it the PRC, being the jurisdiction in which Shanghai – the only geographical location expressly referred to in the arbitration agreement – is situated? Third, what is the interplay between the proper law of the Takeout Agreement and the arbitral law of the seat chosen by the parties in their arbitration agreement when it comes to ascertaining the proper law of the parties’ arbitration agreement? Finally, if the proper law of the parties’ arbitration agreement is PRC law, is the plaintiff correct that the agreement is invalid under PRC law?

14 In determining the proper law of an arbitration agreement, a distinction is drawn between: (i) a free-standing arbitration agreement, ie an arbitration agreement which is contractually separate from the substantive agreement between the parties under which the dispute arises; and (ii) an integrated arbitration agreement, ie an arbitration agreement which is integrated contractually into that substantive agreement. Article 14.2 of the Takeout Agreement is, of course, an integrated arbitration agreement. For that reason, I shall in this judgment analyse only the principles relevant to ascertaining the proper law of an integrated arbitration agreement. Any reference to an “arbitration agreement” in the analysis which follows should therefore be read as a reference to an integrated arbitration agreement, being the only type of arbitration agreement raised by the application before me.

15 There are, broadly speaking, two approaches to choosing the starting point for ascertaining the proper law of an arbitration agreement. One approach proceeds bottom up and the other proceeds top down. The bottom up approach begins with the arbitration agreement itself, construes it to ascertain the seat which the parties have chosen for their arbitrations and identifies the law of that seat as the starting point for ascertaining the proper law of the arbitration agreement. The top down approach begins with the substantive contract into which the arbitration agreement is integrated, construes that contract to ascertain its proper law and then identifies that law as the starting point for ascertaining the proper law of the arbitration agreement. Each approach acknowledges that the starting point which it yields may be displaced upon further analysis. But, to the extent that each approach does identify a starting point, each approach inclines towards that starting point as the end point of the inquiry.

16 It is common ground before me that Singapore law has adopted the top-down approach. ¹⁹ The parties are therefore agreed as regards the principles I am to apply in order to ascertain the proper law of this arbitration agreement. ²⁰ Both parties cite and rely upon the decision of the English Court of Appeal in Sulamérica Cia Nacional de Seguros SA and others v Enesa Engelharia SA and others [2013] 1 WLR 102 ²¹ (“Sulamérica”), of Steven Chong J (as he then was) in BCY and of Belinda Ang J in BMO v BMP [2017] SGHC 127 (“BMO”).

17 I can summarise the applicable principles as follows:

(a) Determining the proper law of an arbitration agreement is to be approached in the same way as determining the proper law of any other contract. In determining the proper law of a substantive contract between two parties, the court is seeking to give effect to the parties’ express or implied choice of proper law. It is only if the parties have made no express or implied choice of proper law that the proper law will be “the system of law with which the contract has the closest and most real connection”: Sulamérica at [9]. This approach is commonly referred to as the three-stage inquiry.

(b) As applied to an arbitration agreement, the three-stage inquiry asks the following three questions. First, have the parties expressly chosen the proper law of their arbitration agreement? Second, have the parties impliedly chosen the proper law of their arbitration agreement? Finally, with what system of law does their arbitration agreement have its closest and most real connection? Sulamérica at [25]; BCY at [40]; BMO at [35].

(c) The three stages of the inquiry are to be embarked upon separately, and in that order: Sulamérica at [25].

(d) In determining the proper law of an arbitration agreement, the doctrine of separability must always be borne in mind. That doctrine holds “that an arbitration agreement forming part of a substantive contract is separable, in the sense that it has an existence separate from that of the contract in which it is found”: Sulamérica at [9].

(e) The natural consequence of the doctrine of separability is that the parties’ choice as the proper law of an arbitration agreement is not necessarily the same law which is their choice to be the proper law of their substantive contract. Having said that, the doctrine of separability does not insulate an arbitration agreement entirely from the substantive contract in which it resides. If the parties have not expressly chosen a proper law for their arbitration agreement but have expressly chosen a proper law for their substantive contract, in the absence of any indication to the contrary, that is a strong indication at the second stage of the three-stage inquiry that they have impliedly chosen the same law to be the proper law of their arbitration agreement: Sulamérica at [11], [25] and [26]; BCY at [44(b)], [49] and [65].

(f) An indication to the contrary at the second stage can arise from the terms of the arbitration agreement itself, or from the fact that the arbitration agreement would be ineffective under the proper law which the parties have expressly chosen for their substantive contract: Sulamérica at [26]; BCY at [44(b)].

18 The three-stage inquiry is an approach specifically and expressly tailored to construing an arbitration agreement in order to ascertain its proper law. I bear in mind, however, that the three-stage inquiry does not operate in a vacuum. It operates against the backdrop of the general principles of contract law which apply when construing any contract. And in particular, it operates against the body of case law which has applied those general principles of specifically to arbitration agreements.

19 I have avoided until now a threshold issue on the first question before me. That question is what law I should apply in order to ascertain the proper law of the parties’ arbitration agreement. The plaintiff submits that I should ascertain the proper law of the parties’ arbitration agreement by applying Singapore’s conflict of laws rules. ²² I accept that submission, and proceed on that basis.

20 In any event, the plaintiff’s alternative submission ²³ is that PRC law approaches the task of ascertaining the proper law of an arbitration agreement in much the same way as Singapore law, at least on the first and second stages of the three-stage inquiry. In other words, PRC law too attempts first to ascertain the parties’ express or implied agreement as to the proper law of their arbitration agreement. It is only if there is no such express or implied agreement that PRC law diverges from Singapore law. In that situation, PRC will hold that the law of the seat is the proper law of the parties’ arbitration agreement whereas Singapore law undertakes the multi-factorial analysis mandated by the third stage of the three-stage inquiry to ascertain the proper law to be imputed to the parties in the absence of an express or implied agreement.

21 The analysis which follows will show that this divergence between Singapore law and PRC law on the first question before me makes no material difference to the outcome of this application. I shall, therefore, proceed on the basis that the first question is to be decided by applying Singapore law and that there is no material difference between Singapore law and PRC law on this question.

22 The leading case in Singapore on how the principles of contractual construction apply to arbitration agreements is the Court of Appeal’s decision in Insigma Technology Co Ltd v Alstom Technology Ltd [2009] 3 SLR(R) 936 (“Insigma”). The Court of Appeal made three points in Insigma which are relevant to the task before me. Indeed, I consider these three points to be more than mere observations – as the Court of Appeal characterised them in Insigma – and in fact to be three guiding principles of primary importance to any court or arbitral tribunal which has to construe an arbitration agreement for any purpose.

23 The first principle which I draw from Insigma is that the principles for construing an arbitration agreement are assimilated with those applicable for construing any other commercial agreement. The fundamental objective of construing a commercial agreement is to give effect to the parties’ intention as they have manifested it objectively in that agreement. So too, the fundamental purpose of construing an arbitration agreement is to give effect to the parties’ intention as they have manifested it objectively in their arbitration agreement. As the Court of Appeal said at [30]:

24 The second principle which I draw from Insigma is that the court should, as far as possible, construe an arbitration agreement so as to give effect to a clear intention evinced by the parties to settle their disputes by arbitration. As the Court of Appeal said at [31]:

25 This second Insigma principle gives rise to two subsidiary principles. First, the courts should not construe an arbitration agreement restrictively or strictly (Insigma at [32]). And second, the courts should prefer a commercially logical and sensible construction over one which is commercially illogical (Insigma at [33]).

26 The third and final principle which I draw from Insigma is that a defect in an arbitration agreement does not render it void ab initio unless the defect is so fundamental or irretrievable as to negate the parties’ intent or agreement to arbitrate (at [37]–[39]). In support of this observation, the Court of Appeal cited the following passage from Fouchard with approval (at [39]):

27 The second and third Insigma principles, although expressed in specific connection with construing an arbitration agreement, are simply corollaries of the first principle: an arbitration agreement is to be construed in accordance with the same principles as any other commercial contract, with the ultimate objective always being to interpret the words which the parties have chosen in their proper context in order to ascertain and give effect to the parties’ intentions, objectively ascertained, rather than to defeat those intentions.

28 Before I turn to apply the three-stage inquiry to the parties’ arbitration agreement, I deal with four preliminary points. These four points are: (i) the defendants’ reliance on extrinsic evidence of pre-contractual negotiations; (ii) the defendants’ invocation of the effective interpretation principle; (iii) the defendants’ invocation of the validation principle and (iv) the plaintiff’s submissions on the doctrine of separability.

29 The defendants submit that it is legitimate for me, as an aid to construing the parties’ arbitration agreement, to take into consideration evidence extrinsic to the Takeout Agreement. This evidence comprises: (i) previous drafts of the arbitration agreement; (ii) evidence of the pre-contractual negotiations between the parties which led to the arbitration agreement in its final form; and (iii) affidavit evidence filed by a witness for the defendants as to his intention in the course of those pre-contractual negotiations. ²⁴

30 I reject the defendants’ submission. For the following reasons, I disregard the extrinsic evidence which the defendants proffer as an aid to construing the arbitration agreement.

31 Singapore adopts the contextual approach to contractual construction. The goal of the contextual approach is to ascertain “what the parties, from an objective viewpoint, ultimately agreed upon in their contract”: Zurich Insurance (Singapore) Pte Ltd v B-Gold Interior Design & Construction Pte Ltd [2008] 3 SLR(R) 1029 (“Zurich Insurance”) at [127]. This is an important point: the goal of the contextual approach is not to ascertain the parties’ subjective intentions, whether those intentions are held unilaterally or shared bilaterally, but their objective intentions, ie, their intentions as manifested by the words of their written contract as those words would be construed by an objective observer aware of the relevant context. All references to the parties’ intention in the remainder of this judgment are references to the parties’ intention ascertained objectively in this way.

32 Where the parties’ contract is a written contract, the parol evidence rule prohibits the court from considering, even under the contextual approach, evidence which is extrinsic to the contract unless certain conditions are satisfied in order to bring the evidence within an exception to the rule. Evidence of the parties’ pre-contractual negotiations is a subset of evidence extrinsic to a written contract which is ordinarily excluded by the parol evidence rule for the purpose of construing that contract.

33 The parol evidence rule and its exceptions form, simultaneously, a part of Singapore’s substantive law of contract, a part of Singapore’s statutory law of evidence and a part of Singapore’s common law of evidence. To the extent that the parol evidence rule and its exceptions are part of Singapore’s law of evidence, they are found in ss 94 to 99 of the Evidence Act (Cap 97, 1997 Rev Ed) (the “Evidence Act”) in Singapore’s statutory law of evidence and are mirrored by analogous rules in Singapore’s common law of evidence.

34 The defendants’ argument on the first preliminary point relies on the decision of Quentin Loh J in BQP v BQQ [2018] 4 SLR 1364 (“BQP”). ²⁵ In that case, Loh J dismissed a plaintiff’s challenge to an arbitral tribunal’s decision on jurisdiction under the same s 10(3) of the Act (at [1]). The plaintiff then applied to Loh J under s 10(4) of the Act for leave to appeal against his decision to the Court of Appeal (at [3]). The plaintiff argued that leave to appeal should be granted because: (i) its application under s 10(3) of the Act raised a question of law as to whether evidence of pre-contractual negotiations is admissible for the purposes of construing a written contract under Singapore law; and (ii) that question was either a question of principle to be decided for the first time or a question of importance upon which further argument and a decision of a higher tribunal would be to the public advantage.

35 Loh J refused the plaintiff’s application for leave to appeal. He held that there is no rule of evidence applicable in arbitrations in Singapore which excludes evidence of pre-contractual negotiations for the purpose of construing a written contract. Loh J concluded therefore that the question of law identified by the plaintiff did not satisfy the test for granting leave to appeal to the Court of Appeal. He arrived at that conclusion for both a statutory and a contractual reason.

36 Loh J’s statutory reason is that ss 94 to 99 of the Evidence Act simply do not apply to arbitration. Section 2(1) of the Evidence Act expressly disapplies Part I, II and III of that Act to “proceedings before an arbitrator” (BQP at [124]). It is Part II of the Act which contains ss 94 to 99 of the Evidence Act. Singapore law’s statutory parol evidence rule therefore has no application to arbitration, whether domestic or international.

37 Loh J’s contractual reason is that the parties in the case before him had contracted out of Singapore’s domestic law of evidence, including the parol evidence rule. The parties had incorporated by reference the Arbitration Rules of the Singapore International Arbitration Centre (5th Edition, 2013) (“the SIAC Rules”) into their arbitration agreement (at [127]). Rule 16.2 of the SIAC Rules ²⁶ provides that it is entirely within the tribunal’s power to determine the admissibility, materiality and weight of any evidence and may receive evidence which is not admissible in law. ²⁷

38 Loh J pointed out that many international arbitral institutions around the world include a rule analogous to Rule 16.2 of the SIAC Rules (at [128]). They do so in order to uphold rather than defeat the procedural expectations of parties involved in arbitration, and in particular in international arbitration (at [126]). Parties to an international arbitration, particularly those from civil law jurisdictions, expect their disputes to be resolved subject to an evidential principle of free admissibility, shorn of any technicalities of the seat’s domestic law of evidence, with factors which would otherwise go to the admissibility of the evidence in question going instead only to the weight to be attached to that evidence.

39 Loh J also observed that the concern in domestic litigation that receiving evidence of pre-contractual negotiations would expand the scope of discovery and the volume of evidence without any appreciable improvement in the accuracy of the decision was of significantly less force in international arbitration. Arbitral tribunals have tools at their disposal, such as the International Bar Association’s Rules on the Taking of Evidence in International Arbitration (“the IBA Rules”), to keep discovery within reasonable and proportionate bounds and to ensure that the arbitration proceeds expeditiously and cost-effectively (at [129]). In the arbitration underlying the application before Loh J, the parties had expressly adopted the IBA Rules.

40 The defendants argue before me, on the basis of BQP, that evidence of the parties’ pre-contractual negotiations is admissible on the application before me. ²⁸ Just as in BQP, the parties before me have expressly incorporated by reference the SIAC Rules into their arbitration agreement, including Rule 16.2. Just as in BQP, the parties before me have expressly agreed that their arbitration is to be conducted in accordance with the IBA Rules. ²⁹ The defendants submit, further, that evidence of the parties’ pre-contractual negotiations is clearly relevant and material since the evidence shows that the plaintiff and the first defendant agreed that any arbitration was to be seated, not in the PRC, but in Singapore as a neutral and third-country venue. ³⁰

41 I do not accept that I can have recourse to evidence of the parties’ pre-contractual negotiations as an aid on the question of construction which lies at the heart of this application. I arrive at that conclusion because the Takeout Agreement contains an entire agreement clause. Article 16.3 of the Takeout Agreement ³¹ reads as follows:

42 It is well-established that parties to a written contract retain the freedom to exercise control, through the terms of their contract, over the interpretive method which a court or tribunal is to apply to their contract in the event of a dispute: Zurich Insurance at [131]. In other words, the parties to a written contract may agree in their contract the extent to which a court or a tribunal may have regard to extrinsic evidence as an aid to construing the contract by way of exception to the parol evidence rule.

43 That is precisely what the parties to the Takeout Agreement did in Article 16.3. The parties contracted out of the parol evidence rule as a rule of Singapore’s evidence law when they incorporated by reference the SIAC Rules into their arbitration agreement. But they reintroduced a contractual analogue of the parol evidence rule by incorporating Article 16.3 into the Takeout Agreement. The parties have, by contract, precluded each other from adducing evidence of their pre-contractual negotiations, including previous drafts of the Takeout Agreement, as an aid to construing their arbitration agreement. That is the contractual effect of Article 16.3 whether the legitimacy of recourse to this evidence were to arise in arbitration or in litigation.

44 I have thus far considered the availability of the extrinsic evidence proffered by the defendants as an aid to construing the parties’ arbitration agreement under Singapore law. But I accept the plaintiff’s submission that that evidence is excluded even under PRC law. Under PRC law, there is a blanket rule denying an adjudicator any power to receive extrinsic evidence of this type as an aid to construction. Thus, the plaintiff’s expert, Professor Chen, opines in her third Report ³² that:

I accept that this proposition represents PRC law on the first preliminary point and does so in terms which arrive at the same result on this point as Singapore law, albeit for different reasons.

45 The second preliminary point is the defendants’ reliance on the principle of effective interpretation. This is a principle drawn from the civil law and which is now said to form part of international arbitration law. It is set out in bold italics in the extract from Fouchard which the Court of Appeal cited at [31] in Insigma (see [24] above). As formulated by Fouchard, the principle is that “where [an arbitration] clause can be interpreted in two different ways, the interpretation enabling the clause to be effective should be adopted in preference to that which prevents the clause from being effective”.

46 It is significant to me that Court of Appeal in Insigma did not endorse a principle of effective interpretation as part of Singapore law, either in terms of the substance of the principle I have cited above or by that label. The Court of Appeal in Insigma was careful to say only that the second Insigma principle (see [24] above) was “similar to the ‘principle of effective interpretation’ in international arbitration law” (emphasis added, at [31]) rather than being identical to or coterminous with it.

47 The fact that the Court of Appeal expressly took the position that an arbitration agreement is to be interpreted as any other commercial contract would means to me that Insigma is not authority for any special principle of “effective interpretation” which applies only to arbitration agreements. To the extent that the principle of “effective interpretation” in international arbitration law differs from the second Insigma principle, I consider that the Court of Appeal’s decision in Insigma prevents me from accepting it as part of Singapore law. I therefore do not accept the defendants’ submission that there exists a specialised or unique principle of effective interpretation applicable only to arbitration agreements.

48 This is an important point. There are two significant differences between the principle of effective interpretation as formulated in Fouchard and the second Insigma principle. The first difference is that the principle of effective interpretation requires, as a prerequisite for its application, that an arbitration agreement have at least two competing interpretations. The second Insigma principle does not. The second difference is that the explicit objective of the principle of effective interpretation is, as its name suggests, the nakedly instrumental objective of ensuring that the arbitration agreement is effective. The explicit objective of the second Insigma principle is to give effect to the parties’ intention as to how their disputes are to be resolved.

49 It therefore appears to me that the second Insigma principle, quite apart from the high authority from which it emanates, is framed in a manner which is explicitly and directly rooted in the fundamental concept of advancing party autonomy whereas that is not necessarily the case for the principle of effective interpretation principle.

50 The third preliminary point is the defendants’ reliance on what is called “the validation principle”. Like the effective interpretation principle, the validation principle is also a principle drawn from international arbitration law. However, unlike the principle of effective interpretation – which operates generally when construing an arbitration agreement – the validation principle operates specifically when ascertaining the proper law of an arbitration agreement.

51 The defendants rely on Professor Gary Born’s formulation of the validation principle in the following terms: “Where the parties have subjected their underlying contract to a law that would, if applied to their arbitration agreement, invalidate that agreement, the separability presumption provides sound analytical reason not to apply that law to the parties’ arbitration agreement.” ³³ (Gary Born, International Commercial Arbitration (Kluwer Law International, 2nd Ed, 2014) at pp 24-27) That formulation is in negative terms: it tells us what the proper law of an arbitration agreement is not, but does not tell us what it is. Professor Born has also put the validation principle in positive terms in his article “The Law Governing International Arbitration Agreements: An International Perspective” (2014) 26 SAcLJ 814 at [51]: The “validation principle provides that, if an international arbitration agreement is substantively valid under any of the laws that may potentially be applicable to it, then its validity will be upheld, even if it is not valid under any of the other potentially applicable choices of law”.

52 The validation principle may well form a part of international arbitration law. It may even form part of the domestic arbitration law of other jurisdictions. But I do not accept that the validation principle forms a part of Singapore’s arbitration law. I come to that conclusion for four reasons.

53 First, like the principle of effective interpretation, the objective of the validation principle is nakedly instrumental. Its explicit purpose is to achieve a prescribed outcome, ie the validation of an arbitration agreement. That, to my mind, fundamentally misstates the objective of the exercise which a court undertakes when construing an arbitration agreement in order to ascertain its proper law. That objective, insofar as the parties have made it possible by the words they have chosen, is to ascertain and give effect to the parties’ intention. The purpose is not, and should not be, to divert the parties to arbitration come what may, without addressing directly the intentions of the parties.

54 The parties’ intention is fundamental in the law of contract but doubly so for an arbitration agreement. Upholding the parties’ intention is the only legitimate basis on which to deprive a respondent of its right of access to the public courts to resolve contractual disputes and to bind that party instead to resolve those disputes through arbitration, privately and without a right of appeal.

55 Second, the validation principle is inconsistent with authority. Both Sulamérica and BCY make clear that what lies at the root of the three-stage inquiry is not a prescribed outcome in favour of arbitration but a desire to give effect, as far as the language chosen by the parties has made it possible, to the parties’ intention. In other words, an arbitration agreement is to be construed like any other contract, without fear and without favour.

56 That is why it is only at the third stage of the three-stage inquiry that the court is justified, in effect, in taking the extraordinary step of imputing a choice of a proper law to the parties which they did not intend to choose for themselves. That judicial imputation is justified because, and only because, by the third stage of the three-stage inquiry, the court has explicitly considered and found that the parties have entirely failed to select a proper law for themselves, whether expressly or impliedly. And even then, the third stage proceeds on the fundamental premise that the parties would, if they had addressed their mind to choosing a proper law for their arbitration agreement, selected the law which has the closest and most real connection to their arbitration agreement. In other words, the third stage simultaneously permits and obliges the court to impute to the parties only that law which can plausibly be imputed to the parties. The third stage does not direct the court to achieve the instrumental objective of ensuring validity.

57 Of course, even at the third stage, whether the parties’ arbitration agreement would be valid under a hypothetical proper law is a weighty factor to be considered. But it is a weighty factor not because the court is giving effect to some policy factor to divert the parties to arbitration. It is a weighty factor because, as a matter of general contract law and as a matter of common sense, parties are presumed to have intended their contracts to be binding.

58 It is true that the three-stage inquiry contains in its third stage a clear departure from party autonomy and the consensual basis of arbitration. But that departure is justifiable because the three-stage inquiry is precisely the same test which is applied to ascertain the proper law of any other contract. In other words, by applying the three-stage inquiry, Singapore arbitration law gives parties to an arbitration agreement exactly the same measure of party autonomy in choosing the proper law of their arbitration agreement as Singapore contract law gives to parties to a substantive contract in choosing the proper law of their substantive contract.

59 In both cases, the parties’ consent to the proper law chosen by the court is real, even at the third stage. The Singapore court, to whose jurisdiction the parties have submitted, finds the proper law objectively, on well-established contractual principles which at every stage has regard to the parties’ intentions, and which bind both parties in equal measure. There is no special “validation principle” in Singapore’s arbitration law which operates only in the domain of arbitration agreements and only to validate an arbitration agreement without a foundation in our general law of contract and without addressing the parties’ intentions.

60 The defendants argue that BCY is authority for the “validation principle”. ³⁴ I do not agree. The validation principle in the terms formulated by Professor Born is incompatible with the three-stage inquiry for ascertaining the proper law of an arbitration agreement which Chong J was at pains to set out in BCY. The validation principle would operate to exclude a candidate for the proper law of the arbitration agreement at any stage of the three-stage inquiry as soon as it was revealed that that candidate invalidated the parties’ arbitration agreement, regardless of the parties’ intent. Only the three-stage inquiry allows the court to ascertain the proper law of an arbitration agreement in a manner which places the parties’ intentions front and centre in the inquiry.

61 I therefore do not consider that BCY is any authority for a validity principle in the terms formulated by Professor Born. Of course, as I have made clear above, both Sulamérica and BCY do acknowledge that the court must, at every stage, have regard to the consequence of invalidity. But that is not as an end in itself but only as a means to the end of giving effect to the parties’ intention objectively ascertained. The difference may be subtle, but the distinction is important.

62 My third reason for rejecting the validation principle as part of Singapore law is that it is unnecessary. There is already in Singapore law a general principle of contractual construction which requires every contract to be construed fairly and broadly, in order to preserve the subject-matter of the contract rather than to destroy it. As Belinda Ang J said in Wartsila Singapore Pte Ltd v Lau Yew Choong and another suit [2017] 5 SLR 268 (“Wartsila”) at [165]:

63 The contracts in question in Wartsila, and in the two cases which Belinda Ang J cited in the passage set out above, were not arbitration agreements. Nevertheless, the “ut res magis” principle (as I shall call it) applies to all contracts, and therefore applies to arbitration agreements under the first Insigma principle.

64 Indeed, the ut res magis principle is especially suited to arbitration. It is simply an aspect of the policy manifest in the second and third Insigma principles to uphold the reasonable commercial expectations of counterparties to an arbitration agreement wherever possible and as far as possible, rather than to defeat them. Further, the ut res magis principle is party-oriented rather than outcome-oriented. It places the emphasis correctly on ascertaining and giving effect to the parties’ intention rather than on achieving a prescribed outcome without regard to their intention. The ut res magis principle is therefore aligned with the primacy of party autonomy in arbitration.

65 Fourth, applying a nakedly instrumental “validation principle” simply stores up problems for the future, at the enforcement stage. Article V(1)(a) of the New York Convention provides that “Recognition and enforcement of [an] award may be refused … if … the [arbitration] agreement is not valid under the law to which the parties have subjected it…”. Article V(1)(a) operates at the enforcement stage to focus attention on the law to which the parties subjected their arbitration agreement. The validation principle operates at the jurisdictional stage to advance an instrumental desire to find jurisdiction by validating an invalid arbitration agreement, without any necessary regard to the parties’ choice of the law to which it is to be subject. Applying the ut res magis principle when construing an arbitration agreement to determine its proper law has the benefit at least that it takes the same party-oriented approach at the jurisdiction stage as the New York Convention mandates at the enforcement stage.

66 The plaintiff rejects the validation principle on the grounds that no arbitration agreement could ever be invalid if the principle were to be accepted whereas our arbitration law framework expressly envisages that some arbitration agreements will be invalid. I do not reject the validation principle for that reason. The validation principle excludes a particular law as the proper law of an arbitration agreement if the result of adopting it would be to invalidate the arbitration agreement. If, in a particular case, the validation principle operates to exclude every proper law that may be potentially applicable, the arbitration agreement would then undoubtedly be invalid.

67 The fourth preliminary point is the plaintiff’s submission ³⁵ that the defendants cannot rely on the doctrine of separability to protect the parties’ arbitration agreement from invalidity. The plaintiff’s argument proceeds as follows. Separability operates to protect an arbitration agreement which would otherwise be valid from being rendered invalid only because the substantive contract into which it is integrated is itself invalid. The parties’ substantive contract being invalid is therefore a condition precedent to the doctrine of separability operating. Separability cannot operate to protect an arbitration agreement from invalidity where the parties’ substantive contract is valid, but subjects the arbitration agreement to a proper law which renders the arbitration agreement invalid.

68 The plaintiff submits that there is no authority in Singapore for the doctrine of separability being used to protect an arbitration agreement from its own inherent invalidity. Indeed, the plaintiff cites the following passage from Sulamérica, at [9] and [26] as authority directly against any such view of separability:

69 I make three points about these two passages from Sulamérica.

70 First, Moore-Bick LJ’s reference to s 7 of the English Arbitration Act 1996 is significant. Section 7 provides as follows:

This section is a statutory statement of the doctrine of separability in English arbitration law. It expressly makes the invalidity or ineffectiveness of the substantive contract a condition precedent to s 7 applying. To the extent that Moore-Bick LJ stated the doctrine of separability narrowly, he was constrained by a controlling statute to do so. We have no equivalent statutory provision in Singapore. There is therefore in Singapore law no equivalent statutory constraint on the scope of the doctrine of separability or on its development.

71 Second, the context in which Moore-Bick LJ made these remarks was to reject the bottom-up approach to determining the proper law of an arbitration agreement in favour of the top-down approach. In that context, it is clear why he made the point that the doctrine of separability, expressed as it is in English law in s 7 of the English Arbitration Act, does not insulate the parties’ arbitration agreement from their substantive contract for all purposes. He did so to explain why the governing law of the parties’ substantive contract is able to reach through the doctrine of separability to form the starting point for ascertaining the proper law of the parties’ arbitration agreement under the three-stage inquiry. That is a proposition which I entirely accept. But that does not mean that the doctrine of separability cannot operate to prevent an express choice of proper law in a substantive contract from invalidating the parties’ arbitration agreement.

72 Finally, the passages from Sulamérica make clear that, even in English law, the doctrine of separability is not an end in itself but simply a means to the ultimate end of giving effect to the parties’ manifest intention to arbitrate their disputes. The objective of the doctrine of separability, as framed in s 7 of the English Arbitration Act 1996, is to ensure that an arbitration agreement is not tainted by the fact alone that the substantive contract into which it is integrated is invalid. The doctrine of separability achieves its objective through the fiction of treating an integrated arbitration agreement as though it were a free-standing arbitration agreement. That fiction is possible only because contractual performance is separable from contractual dispute resolution. The doctrine of separability is not, in that sense, a novel principle created for arbitration agreements and which stands outside the law of contract. It is simply the consequence, in the special context of agreements about dispute resolution, of the ut res magis principle.

73 As Moore-Bick LJ said at [26] of Sulamérica, cited at [68] above, the ultimate source of the concept of separability is the desire to give effect to a presumed intention of the parties that their arbitration agreement should remain effective even if their substantive contract is ineffective. But I do not accept that that is the limit of the doctrine of separability. It is equally legitimate to presume that the parties intend their arbitration agreement to remain effective if a provision of the substantive contract into which it is integrated could, in certain circumstances of fact or law, operate to render their arbitration agreement invalid.

74 To my mind, the only limit on the doctrine of separability is that it should go no further than is reasonable to give effect to the parties’ intention to arbitrate their disputes. Thus, its scope would not go so far as to supply a manifest intent to arbitrate where the parties have failed themselves to make that intent manifest in the words they have chosen to express their arbitration agreement. The core of the doctrine’s scope is no doubt insulating an integrated arbitration agreement which makes manifest the parties’ intention to arbitrate their disputes from invalidity arising only from the invalidity of the substantive contract into which it is integrated. But the scope must also include insulating any such arbitration agreement from invalidity arising from the manner in which a provision in the substantive contract into which it is integrated operates on the arbitration agreement.

75 The plaintiff also cites BCY at [60]–[61] as authority for its submission on the narrow scope of the doctrine of separability:

76 I do not read these passages as an attempt by Chong J to define the limits of the doctrine of separability. I read them instead as describing the situation in which the doctrine is most commonly invoked. It is simply the case that the doctrine of separability is most commonly invoked where the parties’ substantive contract is invalid, in order to avoid that invalidity nullifying the arbitration agreement. But there is no reason in principle why the doctrine of separability cannot have a broader scope, consistent with the ut res magis principle, operating to give effect to the parties’ manifest intention to arbitrate their disputes when a provision of the parties’ substantive contract might operate to defeat that intention. Indeed, Chong J in BCY accepts that the doctrine of separability can operate in tandem with the three-stage inquiry to allow the court to select a proper law for the parties’ arbitration agreement which is not the proper law of the parties’ governing contract precisely because the proper law of the substantive contract would operate to render the arbitration agreement invalid (at [74]).

77 Although I have, for the reasons already given, rejected the validation principle in the broad terms formulated by Professor Born cited at [50] above, I do accept his conceptual point that the doctrine of separability is broad enough in itself to be a sufficient analytical basis for a principle – whatever its content and whatever its label – which operates to uphold an arbitration agreement even when the substantive agreement into which it is integrated is valid but an operation of the substantive agreement could operate to nullify the parties’ manifest intention to arbitrate their disputes.

78 Having dealt with those four preliminary points, I now turn to apply the three-stage inquiry to the Takeout Agreement and to the parties’ arbitration agreement integrated into it.

79 It is beyond dispute that the parties expressly chose PRC law to apply to their substantive contract. Thus, Article 14.1 of the Takeout Agreement provides expressly that: “This Agreement shall be governed by the laws of the People’s Republic of China.”

80 But the parties’ arbitration agreement is not set out in Article 14.1. It is set out in Article 14.2. It is apparent on its face that Article 14.2 does not contain an express choice of law. I do not consider that the parties’ express choice of PRC law to govern their substantive contract in Article 14.1 amounts to an express choice of PRC law to govern the parties’ arbitration agreement in Article 14.2.

81 The plaintiff submits that it does. ³⁶ The plaintiff points out that Article 14.2 is as much a part of the Takeout Agreement as any of the substantive clauses which form part of the Takeout Agreement and which precede Article 14. And it is true that there is nothing in Article 14 or anywhere else in the Agreement to suggest that PRC law should not be the proper law of the parties’ arbitration agreement.

82 I do not accept the plaintiff’s submissions. The mere fact that the parties expressly chose PRC law to govern their substantive contract in Article 14.1 of the Takeout Agreement is not enough, in itself, to constitute that choice an express choice of law for the arbitration agreement integrated into their contract. The first stage would be satisfied only if the parties had, in Article 14.2, expressly provided that the proper law of their arbitration agreement was to be PRC law.

83 The substantive contracts in both the leading cases on the three-stage inquiry had express choice of law clauses. Yet both cases were decided on the second stage, not the first stage.

84 Thus, in BCY, the parties’ substantive contract was expressly governed by New York law. Chong J held that there was a rebuttable presumption that New York law governed the parties’ arbitration agreement (at [72]). The language of presumption means that Chong J was analysing the second stage of the three stage inquiry, having held at the first stage that the parties had made no express contractual choice as to the proper law of their arbitration agreement. If he had found that they had, he would have been obliged to give effect to the contractual choice because they would have been bound contractually by that choice. There would have been no basis for holding that the choice of New York law as the proper law of the arbitration agreement was only a rebuttable presumption. That holding in itself indicates that Chong J was past the first stage of the three stage inquiry and considering the second stage.

85 So too, in Sulamérica, the parties’ substantive contract provided expressly that its proper law was to be Brazilian law. The parties’ arbitration agreement immediately followed the choice of law clause and made no express reference to any law as the proper law of the arbitration agreement (at [5]). The parties accepted that their substantive contract was governed by Brazilian law. But neither party argued that that that choice of law amounted to an express choice of law for their arbitration agreement. As a result, the arguments before the English Court of Appeal in Sulamérica were only on the second and third stages of the three-stage inquiry (at [27]).

86 I therefore accept the defendants’ submission ³⁷ that the parties’ express choice of law in Article 14.1 does not amount to an express choice of law for the parties’ arbitration agreement so as to satisfy the first stage of the three-stage inquiry.

87 It is therefore necessary to go on to consider the second and third stages.

88 At the second stage of the three-stage inquiry, the proper law of the parties’ substantive contract is the starting point as to the parties’ implied choice as the proper law of their arbitration agreement (see BCY at [49]–[50] and [59] –[65]). ³⁸ Indeed, Sulamérica goes so far as to say that if the parties have not expressly chosen a proper law for their arbitration agreement but have expressly chosen a proper law for their substantive contract, in the absence of any indication to the contrary, that is a strong indication at the second stage of the three-stage inquiry that they have impliedly chosen the same law to be the proper law of their arbitration agreement: Sulamérica at [11], [25] and [26]; BCY at [44(b)], [49] and [65].

89 The purpose of the three-stage inquiry is to set out a consistent and principled approach to construing an arbitration agreement in order to ascertain its proper law. It is in that context that this latter observation from Sulamérica ought to be read. The three-stage inquiry adopts the proper law of the parties’ substantive agreement as the starting point on the second stage in order to give effect to what will ordinarily be the intention of the parties in the bulk of cases.

90 Parties do not ordinarily make express provision for the proper law of their arbitration agreement. That is because parties do not ordinarily draw a distinction between the proper law of their substantive contract and the proper law of their arbitration agreement. Therefore, parties who have made express provision in broad and general terms as to the proper law of their substantive contract without making express provision as to the proper law of their arbitration agreement can legitimately be expected, as a starting point, to intend by implication that that the proper law of their substantive agreement is to be the proper law of their arbitration agreement (BCY at [59]).

91 I accept that the starting point on the second stage of the three-stage inquiry is, therefore, that PRC law is the parties’ implied choice as the proper law of their arbitration agreement. The plaintiff acknowledges that this is only a starting point and can be displaced by indications to the contrary. But the plaintiff submits that there are no indications in this case.

92 The plaintiff makes this submission for two reasons. The first reason is that the seat selected by the parties in their arbitration agreement is the PRC. So, the plaintiff argues, PRC law is both the governing law of the parties’ substantive contract and the law of the seat chosen by the parties in their arbitration agreement. There is therefore no alternative to PRC law as the proper law of the parties’ arbitration agreement. The plaintiff’s second submission ³⁹ is that, even if the parties’ arbitration agreement is construed as agreeing that the seat of any arbitration is Singapore, the parties’ choice of a seat in a jurisdiction different from that of the proper law of the substantive contract is not in itself an indication to the contrary and cannot, without more, displace that starting point (BCY at [65]). ⁴⁰

93 I take these two submissions in turn.

95 The plaintiff construes these words ⁴¹ as providing that the seat of any arbitration is the PRC. The plaintiff’s argument proceeds as follows. The arbitration agreement makes no reference whatsoever to Singapore law generally, to Singapore’s arbitration laws specifically or to Singapore even as a geographic location, let alone as a seat. ⁴² The only express reference in the arbitration agreement to any geographic location is the reference to “arbitration in Shanghai”. That express reference suffices in itself to constitute the PRC as the seat. ⁴³ The parties’ express reference to the SIAC and the SIAC Rules in the arbitration agreement indicates merely an intention that the SIAC should administer future arbitrations seated in the PRC and provide institutional support for those arbitration in accordance with the SIAC rules.

96 The defendants, on the other hand, rely ⁴⁴ on Insigma ⁴⁵ (at [30]) to argue that the fundamental objective of construing an arbitration agreement is to give effect to the intention of the parties as expressed in the words of their agreement. The defendants accept that a reference to a specific geographical location has been interpreted in many decided cases as signifying an intention to select that location as the arbitral seat. But, the defendants submit, ascertaining the seat of a particular arbitration agreement depends ultimately on construing the arbitration agreement in question. ⁴⁶ And as a matter of construction of this arbitration agreement, the defendants submit that the parties have agreed on Singapore as the seat of their arbitration and have agreed on Shanghai merely as the venue.

97 It is well-established – and indeed common ground ⁴⁷ between the parties – that the venue of an arbitration is a distinct concept from its seat. The venue of an arbitration is merely the geographic location in which the arbitral hearings and other proceedings take place. ⁴⁸ The seat of an arbitration – or the “place” of the arbitration as it is called in Article 20 of the Model Law – is the jurisdiction whose law governs the arbitral process. As the Court of Appeal held in PT Garuda Indonesia v Birgen Air [2002] 1 SLR(R) 401 (“PT Garuda”) (at [23]–[24]): ⁴⁹

Thus, the defendants submit, ⁵⁰ the selection of the venue or location of hearings does not, in itself, constitute a choice of the arbitral seat.

98 The lynchpin of this aspect of the plaintiff’s case is its submission ⁵¹ that, “[u]nder Singapore law, when the place of the arbitration is expressly stated in an arbitration agreement, that place is considered to be the seat of the arbitration”. The plaintiff submits further that: “The geographical reference in an arbitration agreement…refers to the seat of the arbitration and not to the “venue of hearing’.” ⁵² Thus, the plaintiff submits, the reference to Shanghai in the parties’ arbitration agreement is enough in itself to constitute the PRC as the seat of the parties’ arbitration. I do not accept either of these propositions as being correct. To my mind, these propositions improperly convert what should be a question of construction into a question of precedent.

105 The second case which the plaintiff cites is Hilton International Manage (Maldives) Pvt Ltd v Sun Travels & Tours Pvt Ltd [2018] SGHC 56 (“Hilton”). ⁵³ The plaintiff cites Hilton for the proposition that “the mere reference to the ‘Singapore International Arbitration Centre’ – without stating a geographical place – does not automatically make Singapore the seat of the arbitration”. ⁵⁴ In Hilton, the plaintiff applied in Singapore for a permanent anti-suit injunction to restrain the defendant from commencing litigation in another jurisdiction in breach of an arbitration clause. The question before Belinda Ang J was whether the Singapore court had in personam jurisdiction over the defendant in order to issue the anti-suit injunction. Belinda Ang J held (at [28]) that that turned on whether Singapore was the seat of the arbitration. The difficulty was that the arbitration agreement in Hilton did not make express provision for the seat of any future arbitration. Indeed, it did not make any express reference to any geographical location at all.

107 Hilton too does not stand for the proposition for which the plaintiff cites it. That case stands for the proposition that, where parties have entered into an arbitration agreement which makes no provision as to the seat of any future arbitration but does stipulate that the arbitration will be conducted in accordance with a set of arbitration rules, the parties will be taken to have agreed to the seat which is fixed in accordance with those rules and to have submitted to the jurisdiction of the courts of that seat in any litigation connected with that arbitration.

108 That proposition has no application to the arbitration agreement in the present case for three reasons. First, the arbitration agreement in Hilton (at [8]) made no reference to any geographical location. For the reasons I have given (see [104] above), the parties’ arbitration agreement in the present case makes reference to two geographical locations: Singapore and Shanghai. Second, the arbitration agreement in Hilton made no express provision as to the seat, and expressly referred to the SIAC as merely the “venue of the arbitration”. The SIAC Rules incorporated by reference into the parties’ arbitration agreement in the present case stipulates that Singapore is to be the seat in the absence the parties’ contrary agreement. Finally, although the arbitration agreement in Hilton referred expressly to the SIAC, albeit merely as the venue, it did not make the parties’ arbitration subject to the SIAC Rules. There was therefore no scope for Rule 18.1 of the SIAC Rules to operate in Hilton. Rule 18.1 of the SIAC Rules does operate in the present case to constitute Singapore as the seat, subject to the parties’ contrary agreement.

109 The cases cited by the plaintiff being of no assistance on the fundamental question of construction before me, I return to the words of the parties’ arbitration agreement. Those words expressly incorporate by reference a set of arbitration rules which expressly constitutes Singapore the seat of any future arbitration, albeit subject to the parties’ contrary agreement. The arbitration agreement also refers to Shanghai, but does so without expressly stating whether Shanghai is to be a seat or merely a venue. To my mind, the express reference to the SIAC Rules, a set of rules which the parties expressly agreed was to govern their arbitration and which makes express provision as to the seat, is the clearest possible manifestation of their intention to have all future arbitrations under their arbitration agreement seated in Singapore. Because an arbitration cannot have two seats simultaneously, and because there is nothing in the words chosen by the parties to refer to Shanghai which compels the construction that the PRC is to be the seat, I have no hesitation in holding that the parties’ arbitration agreement, properly construed, constitutes Singapore the seat and constitutes Shanghai merely the venue for the hearings in the arbitration.

110 It is also significant to me 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. Where 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. It is not to the point that PRC law may look at a reference to a geographical reference in an arbitration agreement as the seat, as the plaintiff submits. ⁵⁵ That submission does not take into consideration the express reference in Rule 18.1 of the SIAC Rules to Singapore as the seat of the arbitration in the absence of agreement.

111 On the second stage of the three-stage inquiry, therefore, I have found that the starting point is that the proper law of the arbitration agreement is PRC law, being the proper law of the parties’ substantive contract, but that the parties have also chosen Singapore to be the seat of the arbitration. The question which arises is whether the strong indication in favour of PRC law as the proper law of the parties’ arbitration agreement which arises from the first stage of the three-stage inquiry is to be displaced at the second stage in favour of Singapore law. I hold that it is.

112 In so holding, I reject the defendants’ submission that it is appropriate for me to receive and consider evidence extrinsic to the Takeout Agreement, ie, evidence of the parties’ pre-contractual negotiations and of the defendants’ subjective intention in entering into the arbitration agreement. I do so because of the entire agreement clause found in Article 16.3 of the Takeout Agreement, and for the reasons already given.

113 The plaintiff submits that the law of the seat will be the parties’ implied choice as the proper law of their arbitration agreement only if the parties have failed to make express provision as to the proper law of their substantive contract. ⁵⁶ In support of this proposition, the plaintiff cites dicta from [55] and [64] of Chong J’s judgment in BCY. At [55] of BCY, Chong J refers to two English cases and says:

At [64] of BCY, Chong J says:

114 I can deal with these two dicta quickly. The dictum of Chong J at [55] of BCY expressly incorporates within it the qualifier “on its own”. I therefore do not accept the plaintiff’s submission, without any qualifier, that the law of the seat can be found to be the parties’ implied choice as to the proper law of their arbitration agreement only if there is no express choice of law clause in the parties’ substantive contract. It can be so found if there are indications from which a contrary intent can be gathered. As for the dictum of Chong J at [64] of BCY, he was not there referring to the application of the three-stage inquiry at the jurisdictional stage. He was instead talking about the test to be applied at the setting-aside stage under Article 34(2)(a)(i) of the Model Law and at the enforcement stage under Article 36(1)(a)(i) of the Model Law.

115 The plaintiff has stated the tests to be applied at the second stage of the three-stage inquiry too dogmatically. The ultimate objective of the three-stage inquiry, and of each of its stages, is to give effect to the parties’ intention. Both BCY and Sulamérica make clear that it is a sufficient indication to the contrary on the second stage that the parties’ arbitration agreement will be invalid under the proper law of the substantive contract.

116 I accept the plaintiff’s submission that it is likely that the parties’ arbitration agreement is invalid if PRC law is its proper law. That is either because the parties’ dispute does not satisfy the foreign-elements test in PRC arbitration law ⁵⁷ or because PRC arbitration law does not allow a foreign arbitral institution to administer an arbitration in the PRC. ⁵⁸ Both parties have adduced extensive expert evidence on these two points. It is clear from the expert evidence that PRC law on these two points is uncertain, fraught with difficulty and rapidly evolving. I am deeply conscious of the difficulties which an outsider to a system of law faces in attempting to reach a conclusion on such fraught and difficult questions under that system of law. Those difficulties are compounded when the system of law is fundamentally different from the outsider’s own system of law, and have to be understood only through translation. Having peered cautiously through the window on PRC law which has been offered to me by the expert evidence adduced by both parties, it appears to me that the wiser course is to adopt the more conservative view of PRC law and to concede in the plaintiff’s favour that its position on these two points of PRC law is likely to be the correct position in PRC law.

117 That finding suffices in itself to displace PRC law as the proper law of the parties’ arbitration agreement and to constitute Singapore law as the parties’ implied choice at the second stage of the three-stage inquiry. The contextual approach to construing contracts and the ut res magis principle both permit me legitimately to adopt the law of the seat as the proper law of an arbitration agreement if hewing to the starting point of PRC law would defeat the parties’ manifest intention to resolve their disputes through arbitration.

118 Having concluded on the second stage of the three-stage inquiry that Singapore law is the proper law of the parties’ arbitration agreement, it is not necessary for me to go on to the third stage and determine the law with which the parties’ arbitration agreement has its closest and most real connection.

119 In any event, if it were necessary for me to analyse the third stage, I would hold that Singapore law is the proper law of the parties’ arbitration agreement at the third stage. At the third stage, the parties’ arbitration agreement has its closest and most real connection with Singapore, that being the seat of the arbitration chosen by the parties. That also suffices to fix Singapore law as the proper law of the parties’ arbitration agreement even under PRC law.

120 For all the foregoing reasons I have dismissed the plaintiff’s application with costs.

121 I have also given the plaintiff leave to appeal to the Court of Appeal against my decision. As I have held, Singapore has adopted the three-stage inquiry and rejected instrumental approaches to ascertaining the proper law of an arbitration agreement. The objective of the three-stage inquiry is to give effect to the parties’ commercial intention, whether express, implied or imputed. However, the facts of this case illustrates two ways in which the three-stage inquiry could be said to operate arbitrarily.

122 First, if the SIAC Rules did not include Rule 18.1, it is likely that the outcome of the inquiry as to the seat of the arbitration would have pointed to the PRC. I would have been obliged to invalidate the parties’ arbitration agreement, despite their manifest intent to arbitrate their disputes. It may be said that that outcome is not arbitrary but serves simply to show that there is only so much which the law can do to save an inapt and inept arbitration agreement. But it does seem somehow arbitrary that the mere choice of the arbitral rules could be decisive on this issue when it is almost certainly an unintended effect.

123 Second, the three-stage inquiry can be said to be targeted at giving effect to the parties’ intention – whether express, implied or imputed – in the same way as other contractual rules only if it focuses exclusively on the parties’ intention at the time they entered into the arbitration agreement. Assume, however, that PRC law had changed between the time the parties entered into this arbitration agreement and the time the parties’ arbitration commenced, with the effect that the parties’ arbitration agreement would no longer be invalid under PRC law if it were to be seated in Shanghai but administered by the SIAC. In that scenario, it is virtually certain that there would have been no jurisdictional challenge, and the arbitration would have proceeded on the basis that the seat of the arbitration is the PRC and the proper law of the parties’ arbitration agreement is PRC law. To that extent, it appears to me that the three-stage inquiry permits the court to give effect to the parties’ express, implied or imputed intention when the arbitration is commenced, not their intention when they entered into the arbitration agreement. It might then be said that the three-stage inquiry is simply Professor Born’s wider validation principle in disguise, with the latter at least having the merit of being honest about its objective and transparent in its operation.