Skip to content
Grounds publishedDecided 23 Mar 2023

General Division of the High Court

COT v COU and others and other matters

Feedback & support

[2023] SGHC 69Case no. Originating Summons No 482 of 2021

What happened, in plain English

There is no plain-English summary of this judgment yet.

Ask about this judgment

Laws this judgment relies on

Grounds of decision

Official text from eLitigation

  1. Introduction

  2. [1] Before me are three applications to set aside an arbitration award. The award is in favour of the sole claimant in the arbitration and against all three respondents in the arbitration. Because of a divergence of interests between the respondents (see [14] below), each respondent was separately represented in the arbitration and has brought its own application to set aside the award.

  3. [2] To avoid confusion, and also to avoid revealing the parties’ identities, I shall refer to all of the parties by their position in the arbitration and not by their position in the three applications before me or by their anonymised names in the title to these applications.

  4. [3] I have dismissed all three applications. Each respondent has appealed against my decision. I therefore now set out the grounds for my decision.

  5. Anonymisation

  6. Background to the parties’ dispute

  7. [5] I begin by introducing the parties and setting out the background to their dispute.

  8. The parties in the arbitration

  9. [6] The claimant is a company incorporated in Arnor. It produces and supplies Modules worldwide. The Modules are high technology and high value components essential for a certain class of large and big budget public infrastructure projects (“Infrastructure Projects”).

  10. [7] At all material times until October 2016, all three respondents were members of the same multinational group of companies. I shall refer to this group as “the Rohan Group”. Another member of the Rohan Group is a company which, for reasons I will explain (see [12] below), I shall refer to as “the Procurement Company”. The Procurement Company was not a party to the arbitration and is therefore not a party to these applications. But the subject matter of the arbitration from which these applications arise is the respondents’ alleged liability to the claimant for the debt which the Procurement Company owes to the claimant for Modules sold and delivered.

  11. [8] The Rohan Group comprises hundreds of companies around the world. The business of the Rohan Group is to develop, finance, install, own and operate Infrastructure Projects worldwide.

  12. [9] The first respondent is a company incorporated in Singapore. It is the Rohan Group’s holding company for its business interests in two regions. Gondor is in one of those two regions.

  13. [10] The second respondent is a company incorporated in Gondor. It is an engineering, procurement and construction (“EPC”) contractor. Its business is constructing and commissioning Infrastructure Projects for the Rohan Group in Gondor. The arbitration arose out of one such project (“the Project”).

  14. [11] The third respondent is a company incorporated in Gondor. It is a special purpose vehicle. The first respondent and the Procurement Company incorporated the third respondent for the sole purpose of owning and operating the Project.

  15. [12] The Procurement Company is a company incorporated in Singapore. As its anonymised name suggests, the Procurement Company is the Rohan Group’s centralised procurement arm. Its role is to procure goods for the Rohan Group from vendors around the world and to supply those goods to members of the group after applying an intragroup markup.

  16. [13] Until October 2016, the first respondent held 99.99% of the shares in both the second and third respondents. Another member of the Rohan Group held the remaining 0.01% of the shares in the second respondent. The Procurement Company held the remaining 0.01% of the shares in the third respondent. In October 2016, the Rohan Group sold the third respondent to an unrelated group (“the Sauron Group”). In March 2017, the Rohan Group sold the second respondent to another unrelated group of companies.

  17. [14] From March 2017, therefore, each respondent has been under separate ownership: (a) the Rohan Group continues to own the first respondent; (b) another unrelated group has owned the second respondent since March 2017; and (c) the Sauron Group has owned the third respondent since October 2016. This divergence of ownership, and therefore of interests, is why each respondent was separately represented in the arbitration and has brought a separate challenge to the award.

  18. The contractual chain for supplying Modules to the Project

  19. [15] The Modules which the second respondent needed to complete the Project for the third respondent were supplied under a chain of contracts entered into in 2015 and 2016. Under this chain: (a) the claimant sold Modules to the Procurement Company; (b) the Procurement Company sold Modules to the second respondent; and (c) the second respondent sold Modules to the third respondent.

  20. [16] The contract between the claimant and the Procurement Company is called the “Module Supply Agreement” (“the MSA”). They entered into the MSA in August 2015. Under the MSA, the claimant agreed to sell Modules to the Procurement Company for use in the Rohan Group’s Infrastructure Projects worldwide. This included, but was not limited to, the Project.

  21. [17] It appears that there was no formal contract between the Procurement Company and the second respondent. But the Procurement Company invoiced the second respondent for the Modules which it supplied for all of the projects for which the second respondent was the EPC. The second respondent accepts that it was contractually bound to pay these invoices to the Procurement Company.

  22. [18] The contract between the second respondent and the third respondent is called the Equipment and Material Supply Contract (“the EMS Contract”). They entered into the EMS Contract in March 2016. Under the EMS Contract, the second respondent was obliged to procure Modules for the Project and to supply them to the third respondent.

  23. [19] The invoicing chain for the Modules followed the contractual chain. Where the Modules were intended for the Project, the claimant delivered the Modules directly to the second respondent in Gondor. The second respondent then used the Modules to perform its obligations to the third respondent under the EMS Contract.

  24. The Procurement Company fails to pay the claimant

  25. [20] From August 2015 to March 2016, the claimant issued 13 invoices to the Procurement Company for Modules which the claimant had delivered or was due to deliver to the second respondent for the Project. The Procurement Company was obliged to pay the claimant under these invoices the total sum of ₴29.40m.

  26. [21] In March 2016, ₴16.72m out of this ₴29.40m was already due for payment. A substantial part of this ₴16.72m was not just due but overdue for payment. The claimant therefore suspended delivery of Modules for the Project until the Procurement Company paid for all Modules already delivered for the Project, even if payment was not yet due.

  27. [22] In an effort to persuade the claimant to resume delivering Modules for the Project, senior executives of the Rohan Group entities doing business in Gondor engaged in written and oral negotiations with the senior management of the claimant in March 2016. I describe these negotiations in more detail at [67]–[107] below.

  28. [23] On 18 March 2016, the claimant resumed delivering Modules for the Project.

  29. [24] In April 2016, the Rohan Group’s ultimate holding company sought protection from creditors in rehabilitation proceedings commenced under Chapter 11 of the United States Bankruptcy Code (“Chapter 11”). At the same time, the Procurement Company as well as certain other Rohan Group companies also sought protection under Chapter 11.

  30. [25] As a result of certain transactions in April and May 2016 (see [106]–[107] below), the sum which the Procurement Company owed to the claimant was reduced to ₴7.35m.

  31. [26] In June 2016, it became publicly known that the Sauron Group was to acquire a portfolio of the Rohan Group’s assets in Gondor. This portfolio included all the shares in the third respondent.

  32. [27] Towards the end of October 2016, representatives of the Sauron Group entered into negotiations with the claimant in an effort to resolve the claimant’s claim for ₴7.35m. These negotiations failed to reach an agreement. The result was a three-way standoff. The Sauron Group’s position was that: (a) the second respondent’s remaining liability for Modules was no more than ₴3.84m as set out in the second respondent’s books; and (b) in any event, Gondor’s foreign exchange laws obliged the second respondent to pay that sum to the Procurement Company (as the second respondent’s contractual counterparty) and prohibited the second respondent from paying that sum directly to the claimant (with whom the second respondent had no contract of any kind). The Rohan Group’s position was that no money whatsoever should be paid to the Procurement Company, as it was by then subject to the Chapter 11 proceedings. The claimant’s position was that it was a company under public ownership and therefore could not accept a mere part payment in the sum of ₴3.84m to discharge an indisputable claim for ₴7.35m for Modules sold and delivered.

  33. [28] At the end of October 2016, the Rohan Group transferred its shares in the second respondent and the third respondent to the Sauron Group.

  34. [29] In November 2016, representatives of the Rohan Group again attempted to persuade the claimant to accept a part-payment of ₴3.84m from the Sauron Group in exchange for issuing the manufacturers’ warranties for the Modules to the third respondent, on the basis that the claimant could reserve its right to recover the balance of the ₴7.35m from the Procurement Company. It was again the claimant’s position that, as a company under public ownership, it could not accept part-payment for an undisputable claim.

  35. [30] The following facts are therefore not in dispute. The claimant sold and delivered Modules worth ₴7.35m to the Procurement Company who sold them on to the second respondent who in turn sold them on to the third respondent. The second respondent used the Modules to complete the Project for the third respondent. The Project is operational, and the third respondent is earning substantial revenue from it. The Procurement Company still owes the claimant ₴7.35m for the Modules. According to the second respondent, it owes the Procurement Company only ₴3.84m for the Modules. According to the third respondent, it has paid all sums due to the second respondent for the Modules.

  36. The arbitration and the award

  37. [31] The claimant commenced the arbitration in April 2017. The arbitration was administered by the Singapore International Arbitration Centre (“SIAC”) but took place under the Rules of Conciliation and Arbitration of the International Chamber of Commerce (“the ICC Rules”). The tribunal was constituted in February 2019. The evidential hearings in the arbitration took place over two days in August 2020. The tribunal issued its final award in February 2021.

  38. [32] In its final award, the tribunal held that the respondents were liable jointly and severally to pay the claimant the principal sum of ₴7.35m as damages for breach of contract. In arriving at that decision, the tribunal made three sets of findings that are material for present purposes:

  39. (a) First, the tribunal found that the claimant resumed delivering Modules to the second respondent on 18 March 2016 pursuant to a contract with the respondents. The tribunal referred to this contract as “the…Modules Delivery Agreement”. For brevity, I shall refer to the contract found by the tribunal as “the MDA”.

  40. (b) Second, the tribunal found that it had jurisdiction to hear the claimant’s claim against the respondents arising under the MDA.

  41. (c) Third, the tribunal found that the respondents were in breach of the MDA.

  42. [33] I now set out the details of each of these three sets of findings.

  43. Findings about the MDA

  44. [34] The tribunal made the following findings on the MDA.

  45. [35] The claimant and all three respondents were parties to the MDA. They concluded the MDA on 18 March 2016 (the precise date is significant). Some of the terms of the MDA are oral; others are in writing. The oral terms are found in discussions between the claimant and agents for each of the three respondents between 15 March 2016 and 18 March 2016. The written terms are set out in a non-disposal undertaking dated 17 March 2016. I shall refer to this non-disposal undertaking as “NDU-3”.

  46. [36] Under the MDA, the respondents jointly and severally agreed to pay the claimant the sum of ₴7.35m. The respondents also agreed not to dispose of 24% of the equity in the third respondent until that sum of ₴7.35m had been paid to the claimant.

  47. [37] The consideration for the MDA was the claimant’s resumption of delivery of Modules to the second respondent. The respondents entered into the MDA because the second respondent needed the Modules urgently to complete the Project for the third respondent and to avoid the substantial loss that the Rohan Group would suffer if there was any delay.

  48. Findings on jurisdiction

  49. [38] The tribunal made the following findings on jurisdiction.

  50. [39] The tribunal had jurisdiction to determine the claimant’s claims for breach of the MDA as against all three respondents. All three respondents were parties to the MDA. All three respondents were also therefore parties to an arbitration agreement contained in NDU-3. The respondents were thereby bound to submit disputes arising under the MDA to arbitration.

  51. Findings as to breach of the MDA

  52. [40] The tribunal made the following findings on breach.

  53. [41] All three respondents were in breach of the MDA. That breach had caused the claimant to suffer loss and damage in the sum of ₴7.35m. The respondents were therefore jointly and severally liable to pay ₴7.35m to the claimant as damages for breach of contract.

  54. The grounds for setting aside the award

  55. [42] The respondents have applied to have the award set aside on three broad grounds:

  56. (a) First, the award should be set aside under Art 34(2)(a)(i) of the UNCITRAL Model Law on International Commercial Arbitration (“the Model Law”) as it has been enacted in Singapore by s 3(1) of the Act. None of the respondents concluded any contract whatsoever with the claimant at any time, whether by way of the MDA (as the tribunal found) or otherwise. There was therefore no valid arbitration agreement between the parties within the meaning of the second limb of Art 34(2)(a)(i).

  57. (b) Second, the award should be set aside under Art 34(2)(a)(iii) of the Model Law. The tribunal acted either ultra petita, alternatively infra petita:

  58. (i) The dispute which the parties submitted to arbitration was whether they had concluded a contract in writing on 17 March 2016 in NDU-3 (“the NDU-3 Agreement”), not whether they had concluded the MDA consisting of oral and written terms on 18 March 2016. By finding that the parties had concluded the partly-oral, partly-written MDA on 18 March 2016, the tribunal had therefore acted ultra petita: it had either dealt with a dispute not falling within the terms of the submission to arbitration or had decided a matter beyond the scope of the submission to arbitration within the meaning of Art 34(2)(a)(iii).

  59. (ii) Alternatively, the dispute which the parties submitted to arbitration comprised a number of important matters raised by the respondents which went directly to the fundamental issue of whether the claimant had concluded any contract whatsoever with the respondents at any time. The tribunal acted infra petita in relation to these matters: it either failed to decide these matters or failed to exercise the authority that the parties had granted to the tribunal by not deciding these matters. This failure caused substantial prejudice to the respondents. The award is therefore within the scope of Art 34(2)(a)(iii) of the Model Law when that provision is read with CRW Joint Operation v PT Perusahaan Gas Negara (Persero) TBK [2011] 4 SLR 305 at [34].

  60. (c) Third, the award should be set aside under s 24(b) of the Act or Art 34(2)(a)(ii) of the Model Law. The claimant fundamentally changed its case in the arbitration by advancing the argument that the contract which the parties had concluded was the MDA rather than the NDU-3 Agreement. Further, the claimant made this fundamental change at a very late stage. The tribunal accepted the claimant’s changed case in its award without allowing the respondents an opportunity to adduce further evidence or to make further submissions to address the change. The tribunal thereby either breached the rules of natural justice within the meaning of s 24(b) of the Act or rendered the respondents unable to present their case on the MDA within the meaning of Art 34(2)(a)(ii) of the Model Law.

  61. Issues

  62. [43] The four questions I must answer to determine the respondents’ application to set aside the award track the grounds on which the respondents rely to challenge the award. Those four questions are:

  63. (a) Does an arbitration agreement exist between the claimant on the one hand and any one or more of the three respondents on the other?

  64. (b) Did the tribunal act ultra petita, ie, did it deal with a dispute not falling within the terms of the submission to arbitration when it held that the claimant and the respondents concluded the MDA?

  65. (c) Did the tribunal act infra petita, ie, did it either: (a) fail to decide a number of important matters raised by the respondents which went directly to the fundamental issue of whether the claimant had concluded any contract whatsoever with the respondents at any time; or (b) fail to exercise the authority that the parties had granted to the tribunal by not deciding those matters?

  66. (d) Did the tribunal breach the rules of natural justice when it found that the contract between the claimant and the respondents was the MDA concluded on 18 March 2016 rather than the NDU-3 Agreement concluded on 17 March 2016?

  67. [44] I can deal with the infra petita ground briefly. This ground was, in my judgment, nothing more than a disguised attempt to challenge the merits of the tribunal’s decision. I therefore need not consider it further.

  68. [45] I take the remaining three questions in turn.

  69. The arbitration agreement

  70. The law

  71. [46] The respondents’ first challenge to the award is under the second limb of Art 34(2)(a)(i) of the Model Law. That limb provides that “[a]n arbitral award may be set aside … if the party making the application furnishes proof that … the arbitration agreement … is not valid under the law to which the parties have subjected it …”.

  72. [47] On its face, the second limb of Art 34(2)(a)(i) appears to be premised on an arbitration agreement that has been formed between the parties (as indicated by the use of the definite article in the phrase “the arbitration agreement” in Art 34(2)(a)(i)) but which is subject to some internal or external factor which vitiates it or renders it unenforceable (as indicated by the words “is not valid” in Art 34(2)(a)(i)). The second limb does not, on its face, appear apt to cover an argument that no arbitration agreement has ever existed between the parties.

  73. [48] This appears especially to be the case when the language of the second limb of Art 34(2)(a)(i) is contrasted with the language of Art 16(1) of the Model Law. Article 16(1) empowers the tribunal to decide whether an arbitration agreement exists or is valid. The second limb of Art 34(2)(a)(i), on the other hand, allows the court to set aside an award only if the arbitration agreement “is not valid”. The Model Law thereby appears to recognise a distinction between existence and validity as concepts and appears to have taken a conscious decision to rest the second limb of Art 34(2)(a)(i) only on the concept of validity.

  74. [49] It is nevertheless well established that the second limb of Art 34(2)(a)(i) extends to a challenge to an award on the grounds that no arbitration agreement between has ever existed between the parties (PT First Media TBK (formerly known as PT Broadband Multimedia TBK) v Astro Nusantara International BV and others and another appeal [2014] 1 SLR 372 (“PT First Media TBK”) at [156]; Jiangsu Overseas Group Co Ltd v Concord Energy Pte Ltd [2016] 4 SLR 1336 (“Jiangsu Overseas”) at [46]).

  75. [51] A court hearing a challenge under the second limb of Art 34(2)(a)(i) must therefore determine for itself de novo whether an arbitration agreement exists between the parties and is valid (PT First Media TBK at [163]). Both the parties and the court are therefore free of any and all constraints arising from the manner in which the parties and the tribunal dealt with the same issue in the arbitration, whether any such constraint is said to arise from the arguments presented or not presented to the tribunal, from the evidence that the parties placed or did not place before the tribunal, or from the findings of fact and holdings of law that the tribunal made or did not make in arriving at its conclusions on this issue and on the ultimate issue of its own jurisdiction (Jiangsu Overseas at [48]).

  76. [52] The court approaches a challenge under the second limb of Art 34(2)(a)(i) de novo because it is properly the court’s role to determine whether an arbitration agreement exists and is valid. It is true that the Act and the Model Law give the tribunal jurisdiction to decide this issue (s 10(2) of the Act and Art 16(1) of the Model Law). And it is also true that a court will allow the tribunal to have the first opportunity to decide this issue even if only a prima facie case is made out on this issue (Tomolugen Holdings Ltd and another v Silica Investors Ltd and another appeal [2016] 1 SLR 373 at [67]). That is the doctrine of kompetenz-kompetenz as enacted under s 10(2) of the Act and as applied in Singapore.

  77. [53] But the doctrine of kompetenz-kompetenz rests on a legal fiction: that a tribunal whose jurisdiction rests on consent can itself determine whether that consent exists. The unalterable logical reality is that a tribunal cannot pull itself up by its own bootstraps. A determination whether that consent exists can only be made by the national courts of a state. Only those courts exercise a mandatory jurisdiction over the parties, ie, a jurisdiction that does not rest on consent but instead on the coercive powers of the state.

  78. [56] When a party fails before a tribunal on this issue and places it before a court via a permissible avenue (s 10(3) of the Act and Art 34(2)(a)(i) of the Model Law), the court therefore determines that issue in the exercise of its original civil jurisdiction. That is why a court does not approach an application under either of these avenues as a setting aside application or even as an appeal from the tribunal’s decision. In so far as the heading to s 10 of the Act suggests that an application under s 10(3) is an appeal, in my view it goes too far.

  79. [57] What I have described thus far is the court’s conceptual approach to determining whether an arbitration exists and is valid. This must be kept separate from the court’s procedural approach to any such determination. The court determining a jurisdictional challenge de novo remains free to exercise its procedural discretion in the usual way. That discretion extends to what evidence it will receive, whether it is content to rely on the evidence presented to the tribunal or wishes to receive evidence anew and whether it receives the evidence on affidavit alone or viva voce, with or without cross-examination. Thus, the fact that the court’s conceptual approach to this issue is to treat the application as a hearing de novo does not give a party a right to insist on the court undertaking a full rehearing of the evidence led before the tribunal (AQZ v ARA [2015] 2 SLR 972 at [57]).

  80. The law on contract formation

  81. [59] The claimant’s submission on jurisdiction is that a contract arose out of negotiations in March 2016 that incorporated a term obliging the claimant and the respondents to resolve disputes arising from that contract by arbitration. The principles of law which I must apply to analyse that submission are well-established and are not in dispute. I therefore begin by setting out those principles:

  82. (a) A contract is formed when there is an identifiable agreement that is complete and certain, supported by consideration, and made between parties who intend to create legal relations (Gay Choon Ing v Loh Sze Ti Terence Peter and another appeal [2009] 2 SLR(R) 332 (“Gay Choon Ing”) at [46]).

  83. (b) The inquiry into whether a contract has been formed is an objective one (R1 International Pte Ltd v Lonstroff AG [2015] 1 SLR 521 (“R1 International”) at [51]). The court ascertains the parties’ objective intention as disclosed by their correspondence and interactions seen in the light of the relevant background. This background includes the industry the parties are in, the character of the documents alleged to constitute the contract as well as the course of dealings between the parties.

  84. (c) In conducting this objective exercise, the court will consider the whole course of the parties’ negotiations to ascertain whether an agreement was reached at any given point in time (Gay Choon Ing at [53]; China Coal Solution (Singapore) Pte Ltd v Avra Commodities Pte Ltd [2020] 2 SLR 984 at [26]).

  85. (d) In accordance with the objective approach, once both parties have, to all outward appearances, manifested their agreement on the same terms on the same subject-matter, one party generally cannot rely on some unexpressed qualification or reservation which it continued to hold to show that he did not in fact agree to the terms to which he appeared to agree. The subjective reservations of one party cannot prevent the formation of a contract (Lim Koon Park and another v Yap Jin Meng Bryan and another [2013] 4 SLR 150 at [66], citing Aircharter World Pte Ltd v Kontena Nasional Bhd [1999] 2 SLR(R) 440 at [30]).

  86. (e) There is an identifiable agreement when the parties have reached a consensus ad idem. This consensus is usually identified by reference to an offer made by the offeror which is then accepted by the offeree.

  87. (i) An offer must consist of a definite promise to be bound, provided that certain specified terms are accepted (Gay Choon Ing at [47], citing M P Furmston, Cheshire, Fifoot and Furmston’s Law of Contract (Oxford University Press, 15th Ed, 2007) at p 40).

  88. (ii) An acceptance is a final and unqualified expression of assent to the terms of an offer (Gay Choon Ing at [47], citing Edwin Peel, Treitel: The Law of Contract (Sweet & Maxwell, 12th Ed, 2007) at para 2-015). Even if an offeree does not communicate express acceptance of the offer, the offeree’s positive, negative or even neutral conduct is capable of evincing an intention to accept the offer (CIMB Bank Bhd v Italmatic Tyre & Retreading Equipment (Asia) Pte Ltd [2021] 4 SLR 883 at [20], citing Midlink Development Pte Ltd v The Stansfield Group Pte Ltd [2004] 4 SLR(R) 258 at [50] and [52]).

  89. (f) The parties’ subsequent conduct can be considered in determining whether a contract has been formed (Simpson Marine (SEA) Pte Ltd v Jiacipto Jiaravanon [2019] 1 SLR 696 at [78]).

  90. (g) It is possible for parties to form a contract on a set of essential terms at a given point in time, even if negotiations continue past that point on other terms of the contract which are non-essential. The fact that the non-essential terms are not agreed “does not prevent the contract based on the essential core terms from coming into existence” at the earlier point in time (R1 International at [52]). The crucial question is whether the parties, by their words and conduct, objectively ascertained, demonstrated that they intended to be bound by the essential terms from that point in time even though their negotiations on the non-essential terms continued (Rudhra Minerals Pte Ltd v MRI Trading Pte Ltd [2013] 4 SLR 1023 at [27]). If so, the contract is formed at that time and binds the parties, unless the continued negotiations show that the parties agreed to revoke the agreement (Gay Choon Ing at [53]).

  91. (h) Finally, and as an overarching point, a court does not analyse contract formation in a rigidly mechanical or dogmatic manner. Although the correct approach remains to examine the whole course of the parties’ negotiations to isolate and identify offer and acceptance, this must be done while having regard to the entire background and context (Gay Choon Ing at [63]).

  92. [60] I now turn to examine the negotiations in March 2016 to determine whether the claimant can establish its submission that a contract arose out of negotiations in March 2016 that incorporated a term obliging the claimant and the respondents to resolve disputes arising from that contract by arbitration (see [59] above).

  93. The negotiations in March 2016

  94. [61] The negotiations in March 2016 involved a number of executives from the claimant and from the respondents. It is useful at this stage to identify these executives by their anonymised names.

  95. The executives who participated in the negotiations

  96. [62] Two executives from the claimant participated in the March 2016 negotiations:

  97. (a) Legolas was the claimant’s Chief Executive Officer and General Manager.

  98. (b) Gimli was the claimant’s sales operation manager.

  99. [63] Five senior executives of the Rohan Group entities that were responsible for its business in Gondor participated in the March 2016 negotiations:

  100. (a) Gandalf was the President, Chief Executive Officer and a director of the Rohan Group’s ultimate holding company.

  101. (b) Aragon was a director of the first respondent from June 2011 to April 2017.

  102. (c) Boromir was a director of both the first respondent and the third respondent from March 2016 to October 2016. He was also the General Counsel for the Rohan Group’s business in Gondor from June 2010 to October 2016.

  103. (d) Frodo was a director of the second respondent from February 2015 to April 2017. He was also an employee of the second respondent.

  104. (e) Samwise was an employee of the Procurement Company and a senior director at a wholly owned subsidiary of the Procurement Company in Arnor.

  105. [64] Of these seven executives on both sides, the tribunal heard evidence in the arbitration only from Legolas, Gimli and Samwise. Legolas and Gimli gave evidence, quite obviously, for the claimant. Unusually, Samwise also gave evidence for the claimant, even though he had been involved in the March 2016 negotiations as an employee of the Rohan Group.

  106. [65] Each respondent called only one witness in the arbitration. None of the respondents called any of the witnesses I have listed at [63] above. As a result, all three of the respondents’ witnesses in the arbitration had no involvement in these negotiations and therefore had no personal knowledge of the March 2016 negotiations.

  107. [66] I now describe those negotiations.

  108. Aragon drafts NDU-1

  109. [67] The Procurement Company’s failure in March 2016 to pay the claimant’s invoices as they fell due was the result of serious cash flow difficulties affecting the Rohan Group as a whole at that time. The Rohan Group feared that these difficulties would cause suppliers to stop supplying the goods which the group needed to complete its Infrastructure Projects around the world.

  110. [69] As a first step in this process, Boromir – together with members of his in-house legal team – drafted a template non-disposal undertaking (“NDU-1”). The intent was to offer an agreement in the terms set out in NDU-1 to each essential supplier as a reassurance to that supplier that it would be paid if it continued to supply goods to the Rohan Group entities.

  111. [70] NDU-1 was drafted as an undertaking given by the first respondent in favour of a “Contractor” not to dispose of the first respondent’s shares in the third respondent. The term “Contractor” is left blank, to be filled in as necessary once negotiations with a particular supplier concluded in an agreement.

  112. [71] NDU-1 defines a number of material terms. The plural term “Parties” is defined as both the first respondent and the Contractor collectively; with the singular term “Party” defined as either one of them individually. The term “Company” is defined as the third respondent. The term “Payment Obligations” is defined as the second respondent’s payment obligations to the Contractor pursuant to unpaid invoices to be listed in an annex to NDU-1. The term “Invoices” is defined to mean the unpaid invoices listed in the annex. The term “Security Interest” is defined as a security interest of any kind over the shares in the third respondent.

  113. [72] The recitals to NDU-1 record the following as the factual background leading up to it:

  114. (a) the third respondent is in the process of constructing the Project in Gondor.

  115. (b) the first respondent holds 99.99% of the shares in the third respondent.

  116. (c) the third respondent engaged the second respondent as its contractor for the Project; and the second respondent had “sub-contracted a part of the supply” for the Project to the Contractor.

  117. (d) The Invoices were pending payment from the second respondent.

  118. Notably, the recitals do not even mention the Rohan Group’s desire to maintain continuity of supply as part of the factual background to NDU-1, let alone impose it as an obligation on the Contractor.

  119. [73] The contractual heart of NDU-1 is found in cll 2.1 to 2.3. These three clauses set out the first respondent’s undertaking to the Contractor to retain 24% of the shares in the third respondent until either the first respondent or the second respondent pays the sums due to the Contractor on the Invoices, subject only to the first respondent’s power to create a future Security Interest in favour of lenders to raise financing for the Project. Clause 2.1 provides as follows:

  120. The reason NDU-1 covered only 24% of the shares in the third respondent was that, at the time Boromir drafted it, the first respondent had already pledged 76% of the shares in the third respondent to a consortium of lenders to secure project finance.

  121. [74] Clause 2.2 of NDU-1 provides as follows:

  122. [75] Clause 2.3 provides as follows:

  123. [76] In the course of the negotiations, the respondents referred to NDU-1 as creating a “pledge”. That is obviously not correct. NDU-1 did not create a pledge of the first respondent’s shares in the claimant in favour of the Contractor in any sense of the word. It did not even claim to do so. NDU-1 was merely the first respondent’s undertaking not to dispose of 24% of the shares in the third respondent until the second respondent paid the Contractor in full all sums due on the Invoices.

  124. [77] Clause 8 of NDU-1 provides that it is to be governed by and construed in accordance with the laws of Gondor. Clause 9 provides that disputes under NDU-1 which cannot be resolved amicably are to be submitted to arbitration in Gondor.

  125. [78] The signature block of NDU-1 records that Boromir is to execute it as a deed for and on behalf of the first respondent. Notably, NDU-1 makes no provision whatsoever for the Contractor to execute it. That is no doubt because the contractual content of NDU-1 is entirely unilateral. NDU-1 does not even attempt to impose any obligation of any kind on the Contractor.

  126. 13 March 2016 – the claimant suspends delivery of Modules

  127. [79] On or around 13 March 2016, Legolas asked Samwise to convey to Aragon that the claimant would suspend all further deliveries of Modules for the Project until the claimant received payment for Modules already delivered.

  128. [80] This created a significant risk that the Project could not achieve commercial operation by 30 March 2016. That would expose the third respondent – and therefore both of its shareholders, the first respondent and the second respondent and the Rohan Group as a whole – to a significant risk of substantial financial loss. That risk led to the March 2016 negotiations.

  129. 15 March 2016 – phone call

  130. [81] The negotiations began on 15 March 2016. On that day, Aragon spoke to Legolas by telephone to try and persuade him to resume deliveries of Modules to the second respondent. On the call with Aragon and Legolas were Frodo and Samwise.

  131. [82] Aragon’s overarching point to Legolas was that the Rohan Group would not “run away” and fail to pay the claimant. Aragon explained that the Modules were urgently required to complete the Project. He assured Legolas that the Rohan Group was committed to paying the sums due to the claimant for Modules delivered for the Project and that the Project was on track to achieve commercial operation if the remaining Modules were delivered without further delay. Achieving commercial operation would release the bank financing which the Rohan Group had already secured. Out of that financing, the claimant would be paid in full for all Modules sold and delivered.

  132. [83] Legolas would not relent. Aragon then offered to “pledge” part of the first respondent’s shares in the third respondent as security for the debt due to the claimant. Legolas responded that he would consider resuming delivery of Modules for the Project if, in addition to this “pledge”, the claimant received at least a part payment of the sum outstanding. Legolas also asked Aragon to send him details of his proposal.

  133. 15 March 2016 – NDU-1

  134. [84] After the call, but also on 15 March 2016, Aragon sent an email to Legolas, copied to Frodo and Samwise. In an effort to persuade Legolas that the claimant would get paid upon completion of the Project, Aragon attached four documents to his email:

  135. (a) Evidence of the terms on which the third respondent had contracted to sell the output of the Project to a publicly owned entity in Gondor upon completion.

  136. (b) Evidence that the price at which the third respondent could sell the output of the Project to the publicly owned entity would fall drastically if the Project had not achieved commercial operation by 30 March 2016.

  137. (c) Evidence that the third respondent had secured a term loan that it could draw down upon once the Project achieved commercial operation.

  138. (d) A copy of NDU-1.

  139. [85] The copy of NDU-1 attached to this email had not been customised to apply to the claimant. Thus, the annex to NDU-1 was left blank with no Invoices listed. More importantly, the language of NDU-1 had not been modified to cater for the actual contractual chain for the supply of Modules for the Project. NDU-1 as drafted proceeded on the premise that the second respondent owed Payment Obligations to the Contractor under a contract between them.

  140. [86] The actual position, however, was that the Rohan Group had assigned the task of procuring Modules for the Project to the Procurement Company under the MSA. As a result, the invoices which remained unpaid – and which had led the claimant to suspend delivery of Modules to the second respondent – were invoices which the claimant had issued to the Procurement Company, not to the second respondent. When applied to the claimant, therefore, the definition of “Payment Obligations” in NDU-1 and the third and fourth recitals in NDU-1 were factually wrong.

  141. 16 March 2016 – Legolas replies

  142. [87] Legolas replied to Aragon’s email on 16 March 2016, making three points. First, he told Aragon that the claimant would review the documents attached to Aragon’s email. Second, he said that the claimant would require NDU-1 to be executed before it agreed to release a first shipment of Modules. Third, he said that the claimant would require the Modules already delivered to be paid for before it agreed to release a second shipment.

  143. 17 March 2016 – NDU-2

  144. [88] On 17 March 2016, Legolas followed up on his 16 March 2016 email. He insisted that 57% of the shares in the third respondent should be subject to the non-disposal undertaking, because the claimant had delivered 57% of the Modules for the Project but had not yet received payment for any of them.

  145. [89] Attached to Legolas’s email was a copy of NDU-1 marked up with the amendments which the claimant required. I shall refer to this markup as “NDU-2”. The claimant required six material amendments to be made:

  146. (a) The third recital was to be amended to make express reference to the Procurement Company and the purchase orders that it had placed with the claimant under the MSA. This was no doubt intended to address the issue I have identified at [85]–[86] above.

  147. (b) Clause 2.1 was to be amended by increasing the percentage of the first respondent’s shares in the third respondent which were to be subject to the non-disposal undertaking from 24% to 57%.

  148. (c) Clause 2.2 to be amended to give the claimant priority of payment out of any funds which the first respondent raised from project finance lenders by granting a Security Interest to the lenders over the first respondent’s shares in the third respondent.

  149. (d) Clause 2.3 was to be amended to provide that NDU-2 would terminate once the claimant had received complete payment of the sums arising from the purchase orders that the Procurement Company had placed with the claimant under the MSA. This too was no doubt intended to address the issue I have identified at [85]–[86] above.

  150. (e) The governing law of NDU-2 was to be Singapore law rather than Gondorian law.

  151. (f) Disputes were to be resolved by arbitration at the SIAC in Singapore rather than by arbitration in Gondor.

  152. [90] Frodo replied to Legolas within seven minutes, informing him that the Rohan Group’s Gondorian legal team would review NDU-2. He also offered Legolas an opportunity to resolve any open items on a conference call. Legolas responded just over a half-hour later telling Frodo to go through the changes so that the parties could move forward.

  153. [91] Later that day, Samwise called Gimli to explain that the first respondent could not subject more than 24% of the shares in the third respondent to the non-disposal undertaking because all of the remaining shares were already subject to a Security Interest. Samwise confirmed this in a follow-up email to Gimli sent in the afternoon of 17 March 2016.

  154. [92] Gimli replied to Samwise’s email agreeing that NDU-2 could subject only 24% of the shares in the third respondent to the non-disposal undertaking, but insisting that all of the other amendments be accepted:

  155. 17 March 2016 – NDU-3

  156. [93] At 10.05 pm on 17 March 2016, Frodo sent the following email to Gimli:

  157. [94] Attached to this email was the third version of the non-disposal undertaking. I shall refer to this version as “NDU-3”. NDU-3 differed from NDU-2 in the following respects:

  158. (a) It had the claimant’s name inserted in place of a blank both on the cover page and in the definition of “Contractor” in the body of the agreement.

  159. (b) It was signed by Boromir.

  160. (c) As agreed by Gimli, the claimant’s amendment to cl 2.1 in NDU-2 (see [89(b)] above) was deleted and 24% was reinstated as the proportion of the first respondent’s shares in the third respondent which were subject to the non-disposal undertaking.

  161. (d) The claimant’s amendment to cl 2.2 in NDU-2 (see [89(c)] above) was deleted and the original cl 2.2 from NDU-1 (see [74] above) was reinstated. The effect was that the claimant would not get any priority of payment out of any funds raised from project finance lenders against a Security Interest over the first respondent’s shares in the third respondent.

  162. (e) The claimant’s amendment to cl 2.3 in NDU-2 (see [89(d)] above) was deleted and the original cl 2.3 from NDU-1 (see [75] above) was reinstated.

  163. (f) Four invoices totalling ₴9.43m issued by the claimant to the Procurement Company for Modules delivered to the second respondent were listed in the annex.

  164. [95] NDU-3 accepted the following amendments which the claimant had made in NDU-2:

  165. (a) The reference to the Procurement Company and the MSA in the third recital (see [89(a)] above).

  166. (b) The governing law being Singapore law (see [89(e)] above).

  167. (c) Arbitration being at the SIAC in Singapore (see [89(f)] above).

  168. 18 March 2016 – NDU-4

  169. [96] On 18 March 2016, Gimli sent an email to Samwise objecting to the changes to NDU-2 which were introduced in NDU-3:

  170. [97] Gimli went on to ask Samwise to make three amendments be made to NDU-3 “so as to reach a fair agreement”:

  171. (a) Clause 2.2 should reinstate the language from NDU-2 (see [89(c)] above) giving the claimant priority of payment out of any funds raised from project finance lenders against a Security Interest to over the first respondent’s shares in the third respondent.

  172. (b) Clause 2.3 should at least provide that the non-disposal undertaking terminates when the claimant receives full payment of the Invoices listed in the annex to NDU-3. This was no doubt intended to go further to address the issue I have identified at [85]–[86] above than the amendments to the third recital (see [89(a)] above) which had been accepted (see [95(a)] above).

  173. (c) Three further invoices totalling ₴7.29m should be listed in the annex. The claimant had issued these invoices to the Procurement Company only in March 2016. The time for the Procurement Company to pay these invoices had therefore not yet expired. These further invoices were due but not yet overdue.

  174. [98] Gimli followed up with a second email to Samwise attaching a fourth version of the non-disposal undertaking document incorporating the claimant’s amendments to NDU-3. I shall call this version “NDU-4”.

  175. [99] NDU-4 amended cl 2.3 of NDU-3 as Gimli had required (see [97(b)] above) and included the three March 2016 invoices in the annex (see [97(c)] above). Notably, NDU-4 did not include the amendment to cl 2.2 which gave the claimant priority of payment out of any funds raised from project finance lenders against a Security Interest over the first respondent’s shares in the third respondent (see [89(c)]) as the claimant had required (see [97(a)] above).

  176. [100] Gimli’s email to which NDU-4 was attached reads as follows:

  177. Legolas releases the Modules

  178. [101] It was at this point that Gandalf intervened personally in the negotiations.

  179. [102] On 18 March 2016, Gandalf asked Samwise to convey to Legolas Gandalf’s personal promise to Legolas that all of the claimant’s invoices would be settled. Samwise duly conveyed the promise to Legolas.

  180. [103] Samwise also told Legolas that the changes that Gimli had asked for in NDU-4 were unnecessary because NDU-3’s language made sufficiently clear that the claimant’s invoices to the Procurement Company would be paid in full, and that the claimant would be entitled to the benefit of the non-disposal undertaking over 24% of the shares in the third respondent until the claimant received full payment.

  181. [104] Legolas agreed to leave NDU-3 unamended.

  182. [105] On the same day, Legolas authorised the release of the undelivered Modules.

  183. [106] On 22 March 2016, the second respondent paid ₴5.06m to the Procurement Company under the EMS. On 25 March 2016, the Procurement Company paid ₴5.06m to the claimant under the MSA. This reduced the sum due from the Procurement Company to the claimant by ₴5.06m, ie, from ₴16.72m (see [19] above) to ₴11.66m.

  184. [107] In May 2016, the claimant agreed that the Procurement Company could return Modules worth ₴4.31m to the claimant. This reduced the sum due from the Procurement Company to the claimant by ₴4.31m, ie, from ₴11.66m (see [22] above) to ₴7.35m. This sum of ₴7.35m remains due from the Procurement Company to the claimant for Modules which the claimant had sold and delivered for the Project.

  185. What the claimant must establish

  186. [108] NDU-3 contains the arbitration clause that the claimant relied on to commence the arbitration. The burden lies on the claimant to establish that this arbitration clause binds each respondent. All of the respondents accept that, if I find that the March 2016 negotiations gave rise to a contract, the arbitration clause in NDU-3 is a term of that contract.

  187. [109] To succeed on the first issue, therefore, the claimant must establish two points. The first point is that the March 2016 negotiations gave rise to a contract. The second point is that each respondent is a party to the contract. To establish the second point, the claimant must show that a servant or agent of each respondent bound it to the contract said to have arisen from the March 2016.

  188. [110] It is more convenient to deal with contract formation before authority. In dealing with contract formation, I will assume that each respondent is a party to the contract that I have found the March 2016 negotiations gave rise to, before explaining why I have found that each respondent is indeed a party to that contract.

  189. A contract was formed on basic or essential terms on 17 March 2016

  190. [111] The claimant’s submissions on contract formation are as follows. The claimant formed a contract with the respondents by words and conduct on “basic or essential terms” on 17 March 2016 and then on full terms on 18 March 2016. 17 March 2016 was when Frodo sent Gimli a copy of NDU-3 signed by Boromir (see [93] above). 18 March 2016 was when Gimli released the Modules after Gandalf had spoken to Legolas (see [96] to [104] above). Legolas agreed to resume delivery of the Modules because he accepted Gandalf’s personal request not to require any of the changes to NDU-3 that Gimli had proposed in NDU-4 as a condition for resuming delivery. These communications on 18 March 2016 gave rise to a full terms contract.

  191. [112] For the reasons which follow, I accept that a holistic assessment of the March 2016 negotiations shows that the parties did indeed conclude a contract on “basic or essential terms” on 17 March 2016 and a contract on full terms on 18 March 2016.

  192. Two observations

  193. [113] In arriving at this conclusion, I bear in mind the following two observations about contract formation.

  194. [114] First, whether a contract has been formed is a question of law. And this question of law is to be determined on the usual objective approach applicable in the law of contract. A party’s subjective belief about contract formation is of no legal effect if that belief is not made manifest at the time. So too, a party’s subjective belief about whether a particular communication is an offer, a counteroffer, an acceptance of an offer or a rejection of an offer is of no legal effect if it is not made manifest at the time. Thus, it is perfectly possible for parties to have satisfied all the requirements to conclude a contract as a matter of law without realising it when it happens. They are nevertheless bound by that contract. That is so even if both parties have a subjective and privately held belief that they have not concluded a contract and therefore conduct themselves after that point in time as though they are continuing to negotiate a contract yet to be formed. The parties’ subjective and privately held beliefs cannot override the legal effect of their conduct in satisfying the requirements for contract formation at the earlier point in time.

  195. [115] There is therefore a need for extreme caution when receiving evidence as to the negotiating parties’ subjective and privately held beliefs during their negotiations. That is so when the evidence is contemporaneous, eg communications while the negotiations are taking place, but which are internal to one party. That is especially so when the evidence is given after a dispute has arisen. Evidence of the former type is immaterial. Evidence of the latter type will be tainted by the inevitable conscious and subconscious incentives to give evidence of past beliefs self-servingly, with the benefit of hindsight and to fit neatly within the legal requirements of each party’s case.

  196. [116] Second, although it is true that contract formation is an event rather than a process, it is not always easy to say exactly when that event occurs in a fast-paced commercial negotiation involving multiple participants (Tribune Investment Trust Inc v Soosan Trading Co Ltd [2000] 2 SLR(R) 407 at [40]). It is often the case that commercial entities continue their negotiations even after agreeing what they consider to be the essential terms of the contract.

  197. Analysis of the parties’ negotiations

  198. [117] I start my analysis of the March 2016 negotiations by concentrating on the negotiations from 15 March 2016 to 17 March 2016 (see [81]–[95] above). I will then consider the negotiations on 18 March 2016.

  199. [118] In my view, Aragon’s phone call (see [81] above) and follow up email (see [84] above) on 15 March 2015 amounted to an offer by the respondents to the claimant. The terms of this offer were as follows: in exchange for the claimant resuming delivery of the Modules, the respondents would give the claimant an undertaking in the form of NDU-1 and assure the claimant that full payment for all Modules would be made upon the Project achieving commercial operation.

  200. [119] Legolas did not accept this offer. But he did not reject it either. He clearly saw Aragon’s offer as a starting point for further negotiation. Legolas therefore said, as an initial indication, that he would need “the executed undertaking document for releasing the first shipment of modules … [and] the full payment for the current two shipments [of the Modules]”. He then followed up on 17 March 2016 by sending NDU-2 to Frodo (see [89] above).

  201. [120] NDU-2 was the claimant’s counteroffer. The respondents accepted the claimant’s counteroffer save only for one term. That term was Legolas’s request to increase the shares in the third respondent which were to be subject to the first respondent’s non-disposal undertaking from 24% to 57%. But Gimli withdrew this request after Samwise explained to him why it could not be accepted (see [91] above). The claimant’s position was that all other terms of the counteroffer should be accepted.

  202. [121] I find that, at this point and on the objective approach, there was an agreement on the basic or essential terms of the contract.

  203. [122] The respondents submit that the parties reached no agreement whatsoever, not even on basic or essential terms, on 17 March 2016. The respondents’ submission proceed as follows. The respondent’s conduct in sending NDU-3 to the claimant on 17 March 2016 (see [94] above) and the claimant’s conduct in sending NDU-4 to the respondents on 18 March 2016 (see [98] above) shows that no agreement had been reached even on basic or essential terms on 18 March 2016, let alone on 17 March 2016. Further, even on 18 March 2016, the claimant’s email sending NDU-4 to the respondents amounts to a rejection of the respondents’ offer by way of NDU-3 on 17 March 2016. It destroyed the respondent’s offer and rendered it incapable of acceptance.

  204. [123] I do not accept the respondents’ submissions for the following reasons.

  205. [124] First, as I have said (see [59(b)] above), whether a contract has been formed is determined on the objective approach. The respondent’s considered response to the claimant’s counteroffer of 17 March 2016 was to reject only the increase in the shares to be subject to the non-disposal undertaking from 24% to 57%. Implicit in this is the respondents’ acceptance of all of the other terms of the claimant’s counteroffer. The claimant agreed to withdraw that single term from its counteroffer on 17 March 2016. In doing so, the claimant insisted that the respondents must accept all the other terms of its counteroffer. In response, the respondents did not qualify in any way their earlier acceptance of those other terms of the claimant’s counteroffer.

  206. [125] At the moment of Gimli’s email to Samwise on 17 March 2016, therefore, the parties had, on the objective approach, reached consensus ad idem on the basic or essential terms of their contract (Independent State of Papua New Guinea v PNG Sustainable Development Program Ltd [2019] SGHC 68 at [103]). Even if the respondents at that time had already formed an intention to resile from the acceptance or to alter other terms of the claimant’s counteroffer – as they arguably did the next day in NDU-3 – that does not detract from my finding. When the claimant sent its email on 17 March 2016, any such intention on the respondents’ part would at best relate only to a subjective and privately held qualification or reservation. That cannot affect my finding, on the objective approach, that a contract was formed.

  207. [126] Second, I do not accept the respondents’ submission that they resiled from their acceptance of the claimant’s counteroffer by sending NDU-3 to the claimant. On my findings, the respondents could not resile from their acceptance because they were already bound by a contract on basic or essential terms. That is so even if they subjectively believed at that time that no contract had been formed and that they were continuing to negotiate a contract yet to be formed.

  208. [127] Further, NDU-3 did not in fact qualify the basic or essential terms agreed. The law recognises (see [59(g)] above) that parties may agree a set of basic or essential terms, thereby forming a contract at the moment of that agreement. The fact that the parties go on to agree other terms after that moment does not dissolve the contract which was formed when the basic or essential terms were agreed. If they wish to dissolve that contract, they must do so in accordance with the law of contract. It is not suggested that the emails which the parties exchanged on 18 March 2016 amount to any such dissolution of the contract on basic or essential terms already formed.

  209. [128] Third, I find that none of the terms from NDU-2 which the respondents failed to incorporate into NDU-3 relate to the basic or essential terms agreed on 17 March 2016. I examine first Gimli’s email of 18 March 2016. It is true that Gimli protests that NDU-3 “is NOT what we both agreed on the phone”. But the email is not framed as a dissolution of a contract already formed. It is not even framed as a non-contractual disavowal of everything that has gone before it. Instead, Gimli makes clear that his complaint is confined to two amendments in NDU-2 which the respondents failed to carry through into NDU-3: the amendments to cll 2.2 and 2.3.

  210. [129] Neither cl 2.2 nor cl 2.3, as amended by the claimant, is a basic or essential term.

  211. [130] The amended cl 2.2 is not a basic or essential term. The effect of the claimant’s amendment to cl 2.2 was to give the claimant priority of payment out of any funds which the first respondent may raise from project finance lenders. It is true that this amendment changed the substantive content of an existing obligation in NDU-2. It is also true that Gimli initially insisted on the amended cl 2.2 being incorporated into the parties’ agreement. But the claimant’s own conduct thereafter demonstrates that this amendment was not a basic or essential term. When the claimant sent NDU-4 to the respondents later that day, the claimant itself abandoned the amendment and reinstated cl 2.2 to the version found in NDU-2. I therefore accept the claimant’s submission that Gimli’s decision not to insist on the amended cl 2.2 in NDU-4 establishes that this term was not a basic or essential term to either party.

  212. [131] The amended cl 2.3 is also not a basic or essential term. Clause 2.3 governs when the obligations under the non-disposal undertaking come to an end. I accept the claimant’s submission that its proposed amendment to clause 2.3 goes only to the form in which this contractual obligation is recorded and not to the substance of this obligation. In both NDU-2 and NDU-3, the parties’ obligations under the non-disposal undertaking terminate when the claimant receives full payment for its unpaid invoices from the first respondent or the second respondent. The substance of this obligation is therefore unchanged from NDU-2 to NDU-3.

  213. [132] The claimant’s amendment in NDU-4 to include three invoices additional to those listed in the annex to NDU-3 is also not a basic or essential term. As with the amended cl 2.3, the inclusion of these invoices is a matter of form rather than substance, purely for the avoidance of doubt. The inclusion merely clarified that the scope of the assurance of payment to the claimant covered these invoices. Even without this amendment, all of the invoices issued by the claimant to the Procurement Company came within the substantive obligations in NDU-3 as they fell due for payment. That is so whether or not they were expressly listed in the annex to NDU-3. Including these three invoices does not in any way affect the substance of the basic or essential terms agreed.

  214. [133] Finally, I accept that the claimant’s conduct in releasing the Modules for the Project on 18 March 2016 was causally connected to the contract which the parties had reached on basic or essential terms. Indeed, I go further and find that the release amount to both performance of the contract and the consideration for the contract.

  215. [134] I do not accept the respondents’ submission that the release of the Modules was unconnected to the March 2016 negotiations and was merely the result of a commercial decision by the claimant to deepen its relationship with the Rohan Group in anticipation of future business. This submission is not supported by any contemporaneous evidence. This submission also runs counter to the inherent probabilities and the commercial expectations. If the claimant decided on 13 March 2016 to withhold delivery of Modules for the Project to reduce its exposure to the risk of the Procurement Company defaulting on its payment obligations under the MSA, it makes no commercial sense for the claimant to do an about-face just five days later and increase its exposure to the Rohan Group by delivering more Modules without having obtained any payment for invoices already due and overdue and without any assurance of payment in the future. It makes even less sense for the claimant to do this unilaterally, without asking for or securing any benefit in return.

  216. [135] I therefore accept that the claimant released the Modules on 18 March 2016 in performance of its obligations under and as consideration for the contract concluded on basic or essential terms on 17 March 2016.

  217. [136] For the foregoing reasons, I accept the claimant’s submission that the parties concluded a basic and essential terms contract on 17 March 2016 and then on full terms on 18 March 2016. This conclusion means that there is a valid arbitration agreement between the parties. The contract I have found incorporates NDU-3. Clause 9 of NDU-3 contains an agreement to arbitrate disputes.

  218. [137] Although this was not put in issue by the respondents, I also accept that this arbitration agreement fulfils the formalities set out in s 2A of the Act:

  219. (a) The arbitration agreement is in writing. It appears in cl 9 of NDU-3, the written component of the contract (see ss 2A(3) and 2A(4) of the Act). The fact that the MDA is partly in words and partly by conduct is therefore not to the point.

  220. (b) Further, since the MDA incorporates NDU-3, cl 9 is part of the MDA (see s 2A(7) of the Act).

  221. [138] The claimant’s alternative submission is that, even if no contract was formed on 17 March 2016, NDU-3 was the respondents’ offer to the claimant in the terms set out in NDU-3 and Legolas accepted this offer on 18 March 2016. Given my finding in favour of the claimant on its primary submission, I need not consider its alternative submission.

  222. The parties to the contract

  223. [140] The claimant does not have to establish this second point as against the first respondent. That is because it is common ground that Boromir had the first respondent’s authority to participate in the March 2016 negotiations and to bind the first respondent to any contract that may arise as a result. Boromir also signed NDU-3 expressly for and on behalf of the first respondent. The first respondent therefore accepts that, if I find that the March 2016 negotiations did give rise to a contract, as I have, the first respondent is a party to that contract.

  224. [141] The second and third respondents, on the other hand, each submit that no person who participated in the March 2016 negotiations had any authority whatsoever to negotiate with the claimant on each of their behalf or to bind either of them to any contract that may have arisen as a result.

  225. [142] It is clear on the face of NDU-3, of course, that neither the second respondent nor the third respondent are signatories to NDU-3. But I consider that feature to be a point of form rather than a point of substance. It is true that a person who signs a contract containing an arbitration agreement is, prima facie at least, a party to the arbitration agreement (Gary B Born, International Commercial Arbitration (Wolters Kluwer 2nd Ed, 2014) at pp 1521–1522). That is merely the application of a general point of contract law to the specific case of an arbitration agreement. But it does not follow that a party who has not signed a contract containing an arbitration agreement is not a party to an arbitration agreement incorporated into that contract: “[t]he sterile formality of a signature is not always necessary in law to breathe life into contractual undertakings” (Midlink Development Pte Ltd v The Stansfield Group Pte Ltd [2004] 4 SLR(R) 258 at [55]). More importantly, given my findings on the parties’ contract, only part of the contract is found in NDU-3. Accordingly, the fact that the second and third respondents are not signatories to NDU-3 is not conclusive as to whether they are parties to the contract of which NDU-3 forms a part.

The judgment continues past this page. Read the rest at eLitigation.