12:21 PM Iris Ng (LL.B., Singapore Management University)

    A closer look at proposed amendments to the International Arbitration Act: How far should parties be allowed to contract out of the grounds for annulling an award?


    Under Proposal 5 of the Ministry of Law’s recent public consultation on Singapore’s International Arbitration Act (“IAA”), parties will be given the option to limit or waive by agreement (collectively, “contract out of”) the annulment grounds in s 24(b) of the IAA and Article 34(2)(a) of the UNCITRAL Model Law on International Commercial Arbitration (“ML”), but not those in s 24(a) of the IAA and Article 34(2)(b) of the ML. Such an agreement can only be made after the award has been rendered. This post examines the rationale for permitting parties to contract out and the approach in various jurisdictions, to argue that Proposal 5 is sound save for the restriction that contracting out is only permitted post-award.

    Rationale for allowing contracting out – party autonomy versus judicial oversight

    The consultation paper identifies the rationales for Proposal 5 as party autonomy, and the avoidance of situations where the seat and enforcement courts reach conflicting decisions on the same grounds. Although the party autonomy rationale is a sound one – party autonomy is a cornerstone of arbitration and a key reason why arbitration is so attractive – the second rationale warrants further thought.

    Unlike the first rationale, the second rationale is not of general application. The issue of conflicting decisions arises in two narrow situations: (a) where enforcement abroad precedes annulment proceedings in Singapore; and (b) where the award has been annulled in Singapore but is nonetheless regarded as enforceable in the enforcement jurisdiction.

    In case (a), the onus should arguably be on the award debtor to seek a stay of enforcement. The counterargument is that the legal device of a stay may not be available in that jurisdiction. However, that situation is unlikely in practice. Article VI of the New York Convention provides that pending an application for setting aside or suspension of the award, the enforcing court “may, if it considers proper, adjourn the decision on the enforcement of the award”. This power is mirrored in Article 36(2) of the ML, which has been reproduced in the national legislation of many of the numerous jurisdictions that have adopted the ML. But a further problem might yet arise from the discretionary nature of this remedy. Summarising the factors relevant to the exercise of discretion under Article 36(2), Born writes that these include the likelihood that the foreign court will annul the award, the amount of time required before annulment proceedings are concluded, and the balance of hardship suffered by each of the parties (G Born, International Commercial Arbitration (Kluwer Law International, 2nd ed, 2014) at p 3722). Significantly, courts do not always have regard to the strength of the case for annulment in deciding whether to grant a stay. In these circumstances, Proposal 5 has residual use in permitting parties the option of agreeing to waive setting aside proceedings to avoid the risk of conflicting decisions. 

    Case (b) arises only where the enforcement jurisdiction subscribes to delocalisation theory, which holds that an award is detached from its seat and can survive annulment. But given that Singapore subscribes to territorialism (see Prometheus Marine Pte Ltd v King, Ann Rita [2018] 1 SLR 1 at [46]), the locus of control should be rooted in the seat, whose findings should take precedence. It should not be for the seat to strive to ameliorate the problems that flow from delocalisation, which is in any case not the predominant view (as opposed to the strict territorialism). Therefore, the party autonomy rationale remains the stronger and more widely applicable justification for allowing contracting out.

    Returning to party autonomy, this, however, has its limits. The countervailing consideration is the need for judicial oversight, given that a legal system retains an interest in preventing its legitimacy from being subverted because it is used as an instrument of injustice. The question is therefore how the balance should be struck. This engages the following concerns: (1) What grounds should parties be allowed to contract out of? (2) Which parties should be allowed to contract out? (3) At what point should parties be entitled to contract out?

    Why legislative reform is needed

    Before addressing the questions identified above, I digress to address why legislative reform is preferable to relying on the courts to develop the law: The issues identified require holistic analysis of all annulment grounds. The common law, being incremental and piecemeal in development, is not well-equipped (or, in some instances, even empowered) to make these changes in a timely fashion. Indeed, common law jurisdictions such as Canada and New Zealand have grappled with the issue of “contracting out” by way of case law, but the results are unsatisfactory as the following cases demonstrate.

    In Noble China Inc v Lei Kat Cheong 1998 CarswellOnt 4386, the Ontario Court of Justice held that Article 34 of the ML was not mandatory because it was not listed as such in the Analytical Commentary and contained the permissive language of “may”. It also rejected the submission that since the list of grounds for setting aside in Article 34 refers to mandatory provisions, Article 34, being the only way of enforcing those principles, must itself be mandatory. Instead, agreements that contravened mandatory provisions (such as Article 18 on equal treatment of parties) would themselves be “ineffective” as waivers, and therefore not preclude the invocation of Article 34, which is itself not mandatory (at [68], [71]–[72]). Kroll has criticised this reasoning because it does not cater to the possibility of the tribunal deviating from mandatory provisions of the lex arbitri without party agreement to that effect: S Kroll, “Setting aside proceedings in Model Law jurisdictions – selected procedural and substantive questions from the case law” (2005) 8 Int ALR 170 at 175.

    Almost two decades later, the Ontario Superior Court of Justice in Popack v Lipszyc 2015 ONSC 3460 adopted a more nuanced interpretation of Article 34: Parties cannot contract out of Article 34 for all purposes; the proper approach is to “consider to what extent the [a]rbitration [a]greement seeks to exclude Article 34, and if it does, to what extent it is effective in doing so given the specific matters at issue”. In each case, the court must consider whether the provision sought to be excluded relates to either a mandatory provision of the relevant arbitration statute or conflicts with public policy (at [47] and [52]). Unfortunately, this approach is vague and leads to uncertainty.

    Methanex Motonui Ltd v Joseph Spellman & Ors [2004] 1 NZLR 95 (“Methanex”) illustrates the piecemeal approach that courts must necessarily take, because the scope of analysis in each case depends on the parties’ arguments. Dealing only with the right to review for breach of natural justice, the court held that contracting out of this was impermissible (at [127]–[132]).

    As the above demonstrates, it is unsatisfactory to rely on the courts to puzzle out the circumstances when parties can contract out of annulment grounds.

    Questions for consideration

    Turning proper to the three questions posed above, different jurisdictions have answered these in various ways. A summary table comparing the French, Belgian, Swiss, Swedish, Colombian and Russian statutes is provided in the Annex. These will be referred to below where relevant.

    What grounds should parties be allowed to contract out of?

    Proposal 5’s approach of explicitly specifying the grounds that parties can or cannot contract out of is a useful one. Although some of the jurisdictions surveyed (France, Belgium, and Russia) do not indicate the grounds that may be excluded, this leads to uncertainty. The courts may well interpret the statute as being subject to the limit of public policy notwithstanding the lack of express restrictions on contracting out, as the Belgian courts have done (see Marc Dal, “National Report for Belgium (2019)” in Lise Bosman (ed), ICCA International Handbook on Commercial Arbitration (Kluwer Law International, 2019) (“ICCA Handbook”) Supplement No 104 at p 52). It is clearer to specify which grounds parties can contract out of.

    There is also much to commend about the designation of the grounds in s 24(b) of the IAA and Article 34(2)(a) of the ML as excludable, and those in s 24(a) of the IAA and Article 34(2)(b) of the ML as non-excludable.

    First, the need for judicial oversight counterbalances the party autonomy rationale and militates against permitting parties to contract out of all annulment grounds (which is the position in Switzerland, Sweden and Colombia). Prima facie, it could be argued that even if no restrictions are placed on contracting out of annulment grounds, an agreement to that effect will remain amenable to judicial oversight and control. Such an agreement, being in the first place a contract, will be subject to the usual contractual doctrines (in particular, vitiating factors like fraud, duress and mistake; or public policy). However, contractual regulation would depend on the outcome of applying the substantive law of the contract, which is not necessarily Singapore law (ie, the issue would be whether the contract is valid by its proper law rather than whether contracting out of certain annulment grounds is impermissible under the lex arbitri). Therefore, to strike the appropriate balance between party autonomy and the need for judicial oversight, any limits should still be imposed as part of the lex arbitri.

    Secondly, it would be consistent with existing judicial authority to carve out fraud/corruption, non-arbitrability and conflict with public policy from the realm of optional grounds. In BAZ v BBA and others and other matters [2018] SGHC 275, the court held that the doctrine of waiver should not generally apply to public policy and non-arbitrability objections. The court referred to the Analytical Commentary of the Model Law, and concluded at [67]–[68] that:

    … The importance of ensuring that an award does not offend the most basic notion of morality and justice outweighs the principle of finality in arbitration that the doctrine of waiver seeks to achieve. Thus, a genuine claim on the ground that an award would offend the public policy of the state cannot be easily waived.

    … Support for this position can be found in the preparatory materials to the Model Law. The Analytical Commentary notes that certain defects such as violation of public policy and non-arbitrability “cannot be cured” by submission to the arbitral proceedings and a failure to raise objections during the proceedings … The rationale for the position is that public policy and arbitrability are fundamental to the forum, so an award that falls foul of the public policy or the arbitrability criteria of the state should not be enforced even if parties waived their objection. …

    [emphasis omitted]

    Thirdly, pursuant to party autonomy, the contracting out of all other grounds should be permitted. This is so even for the natural justice ground despite the controversy surrounding this ground. On one hand, Methanex held that attempts to contractually exclude the right to review for breach of natural justice are always ineffective because it would be a “contradiction in terms to adopt a process which is described as an arbitration but which does not attract a duty to observe the requirements of natural justice” (at [127]–[132]). A similar notion is espoused in the IBA Non-Waivable Red List. On the other, the Singapore case of Kok Seng Chong v Bukit Turf Club and another [1992] 3 SLR(R) 772 takes the contrary view in one brief line (at [101]), though not in the arbitration context. The English and Canadian authorities also suggest that breaches of natural justice can be waived (see eg, Farrelly (M&E) Buildings Services Ltd v Byrne Brothers (Formwork) Ltd [2013] EWHC 1186 at [27]–[29]; Canada v Taylor [1990] 3 SCR 892 at [177]). The difficulty with natural justice has been explained by one commentator as arising from the fact that breaches of natural justice or non-compliance with arbitral procedure are a matter of degree (F Bachand & F Gélinas, The UNCITRAL Model Law After Twenty-Five Years: Global Perspectives on International Commercial Arbitration (Juris Publishing, 2013) at pp 130 and 132). In this connection, it is also relevant to consider that regardless of whether parties may contract out of natural justice requirements per se, it is uncontroversial that “the explicit parameters of … arbitration agreements may implicitly impact the threshold for a breach of natural justice”, as one commentator put it. An illustration of this can be found in China Machine New Energy Corp v Jaguar Energy Guatemala [2018] SGHC 101, where the Singapore court observed that because the tribunal was obliged under the contract to conduct an expedited arbitration notwithstanding the dispute’s scale and complexity, due process had to be followed within the strictures of the arbitration agreement, principally time constraints. Considering the above, Proposal 5’s approach of permitting parties to contract out of the natural justice ground generally is a satisfactory one. The court would retain the power to annul the award where the breach of natural justice is so egregious that it simultaneously engages the public policy ground, which cannot be waived.

    Accordingly, Proposal 5’s approach regarding the grounds that parties may contract out of is a sound one.

    Which parties should be allowed to contract out of annulment grounds?

    The jurisdictions surveyed adopt one of three main approaches: (a) permitting contracting out only in international arbitrations (France); (b) permitting contracting out only where none of the parties are connected to that jurisdiction (Belgium, Switzerland, Sweden, and Colombia); or (c) permitting contacting out for only institutional and not ad hoc arbitration (Russia).

    Proposal 5’s approach of allowing parties to contract out of annulment grounds in respect of Singapore awards to which the IAA applies is a sound one. Approach (b) is too restrictive and is inconsistent with the tenor of s 5 of the IAA, which does not focus on the parties’ nationality or domicile. Further, approach (c) should be rejected as there is no good reason a priori for excluding ad hoc arbitration, which can equally involve sophisticated parties capable of looking after their own interests.

    At what point should parties be entitled to contract out of annulment grounds?

    Contrary to Proposal 5, agreements to contract out of annulment grounds should be permitted regardless of whether they are entered into before or after the award has been rendered.

    Proposal 5 is out of line with the approach of numerous jurisdictions, including France, Switzerland, Belgium, Sweden, and Colombia. Significantly, while French law used to permit waiver only where the agreement was entered into after the award was made, that provision no longer applies: Yves Derains & Laurence Kiffer, “National Report for France (2013 through 2018)” in ICCA Handbook Supplement No 99 at p 82.

    Moreover, setting the cut-off time at after an award is rendered serves no useful purpose. It could be argued that only in a post-award situation can parties be fully apprised of what they are giving up in terms of opportunities to challenge the award. But equally, this promotes opportunistic behaviour, given that parties would favour such agreements where the award is in their favour but afflicted by one or more problems that might otherwise have warranted annulment. Conversely, parties would be more likely to reach agreement on the grounds that they would like to waive if the agreement is entered into prophylactically, before the relationship sours and the parties go for arbitration. It would also be more convenient for parties to be able to agree beforehand, in one shot when entering into the arbitration agreement, on the grounds for which they would like to waive recourse. Thus, the timing when contracting out is permissible should not be restricted.


    This article has argued that Proposal 5 is generally a step in the right direction. Together with two of the other proposed amendments to the IAA – allowing parties to request the tribunal to decide on jurisdiction at a preliminary stage, and to opt into the mechanism of appeals on questions of law arising out of an award – these affirm a robust approach in favour of party autonomy that will, if adopted, increase Singapore’s competitiveness as an arbitral seat.

    * The article is written in the author’s personal capacity, and the opinions expressed in the article are entirely the author’s own views.

    ** This blog entry may be cited as Iris Ng, “A closer look at proposed amendments to the International Arbitration Act: How far should parties be allowed to contract out of the grounds for annulling an award?” (20 August 2019) (

    *** A PDF version of this entry may be downloaded here

    **** The Annex may be seen here

Comment Section