01:08 PM Leong Hoi Seng Victor (Justices’ Law Clerk, Supreme Court of Singapore)

    Examining the “special circumstances” test in BLY v BLZ [2017] SGHC 59: when should jurisdictional challenges result in stays of the arbitral proceedings?



    In BLY v BLZ and another [2017] SGHC 59 (“BLY”), the High Court considered a little-considered question arising under section 10(9) of the International Arbitration Act (Cap 143, 2002 Rev Ed) (“IAA”). Two parties had gone to arbitration. The tribunal rendered its finding on jurisdiction, which one party then challenged in the High Court. The challenging party invited the Court to exercise its discretion to stay the arbitration until the jurisdictional challenge was determined. The Court declined to do so.

    In coming to its decision, the Court watered the “jurisprudential seeds” sowed in AYY v AYZ and another [2015] SGHCR 22 (“AYY”), the only previous decision on section 10(9). But the court in BLY explicitly preferred the “special circumstances” test over the “irreparable prejudice” test set out in AYY, and also rejected the “balance of convenience” test argued for by the challenging party in BLY.

    In my view, the “balance of convenience” test should have been adopted in substance instead, although the terminology of “balance of convenience” is arguably unhelpful. On this test, the outcome would also have been the same – the stay should not be granted in this case.


    The approach in BLY v BLZ

    Section 10(3) of the IAA allows parties to apply to the High Court to challenge a tribunal’s jurisdictional ruling. Section 10(9)(a) provides that the application “shall not operate as a stay of the arbitral proceedings…unless the High Court orders otherwise”. Section 10(9)(b) buttresses this by providing that “no intermediate act or proceeding shall be invalidated except so far as the High Court may direct”. BLY concerned only the issue in section 10(9)(a).

    Section 10(9)(a) is sui generis. In the earlier case of AYY, the Assistant Registrar (“AR”) held that a stay of the arbitration would be ordered if an applicant could demonstrate that refusing to stay the arbitration would result in prejudice that could not be compensated by costs, similar to the position in litigation (AYY at [7]).

    Neither party in BLY argued for the application of this “irreparable prejudice” test; nor was it adopted by the Court. The plaintiff (who applied for the stay) argued that the court should apply a “balance of convenience” test instead. This test would balance the prejudice to the applicant if no stay were ordered against the prejudice to the respondent in delaying the arbitration if a stay was ordered (BLY at [6]). The defendants disagreed. They argued that “special circumstances” would be needed to displace the prima facie position that the arbitration would continue, in line with the policy of minimal curial intervention (BLY at [7]).

    The court preferred the “special circumstances” test for two reasons. First, the court noted that the default position in section 10(9) was that the arbitration was to continue. The test to be adopted could not render the default position “meaningless” (BLY at [8], [14]). Second, the “special circumstances” test accurately captured the balance between the court’s need to “have control over a tribunal’s decision on jurisdiction” and the “need to ward against the abuse of such recourse as a dilatory tactic to hold up the arbitration” (BLY at [11]).

    As to what was meant by “special circumstances”, the court held that these were circumstances which were not a “usual and attendant by-product or consequence of a tribunal’s decision to continue” (BLY at [15]). Such circumstances were distinguished from irreparable prejudice in principle. Prejudice refers to a certain outcome. While prejudice can arise from special circumstances, not all special circumstances lead to irreparable prejudice, nor does irreparable prejudice always arise from special circumstances. Accordingly, the court did not grant the stay application since the arbitral tribunal ordered disclosure of information that was not unduly sensitive (BLY at [23]) and any prejudice could be controlled by the parties’ confidentiality clause (BLY at [25]). These were ordinary circumstances which would have arisen out of the arbitration anyway.


    The differences between the three tests

    In order to determine which test best addresses the court’s concerns, we must first understand the differences between them. There are two differences: whether the court considers the outcome or the process, and from whose perspective the court considers the outcome or the process.

    The “irreparable prejudice” looks at the outcome of granting a stay from an applicant’s point of view. In contrast, the balance of convenience test as described by counsel in BLY is to “balance the prejudice that may result from the carrying out of an unnecessary arbitration and the prejudice that would result from a delay in the arbitration” (BLY at [6]). The prejudice that results from an unnecessary arbitration is prejudice to an applicant. The prejudice from a delayed arbitration is prejudice to the respondent. So it similarly looks at the outcome but considers the outcome from both the applicant’s and respondent’s points of view.

    While these two tests consider the outcome, the “special circumstances” test considers the process. “Special circumstances” are usefully contrasted against “irreparable prejudice” in that the latter “is the consequence or outcome that arises given the existence of some special circumstances in the case” (BLY at [18]). Hence, the special circumstances test does not look at the outcome, but only whether the consequence “is to be expected in the usual course of arbitral proceedings” (BLY at [19]). If the consequence flows naturally from arbitration, then it cannot be a special circumstance.


    Revisiting the Court’s reasoning

    We now turn to revisit the two reasons for the Court’s preference of the “special circumstances” test. The first reason given was that the default position of allowing the arbitration to continue should not be rendered meaningless. Presumably, the court is referring to the fact that under either of the alternative tests, stays would likely be “routinely granted” (BLY at [23]). The concern here is that “routinely” granting stays would dilute, and indeed subvert, the Singapore courts’ pro-arbitration stance (see, eg, AKN v ALC [2015] 3 SLR 488). But there are three reasons why this concern may not allow the court to distinguish between the tests. First, it is entirely possible that the “special circumstances” test could equally result in stays being routinely granted. The court defined special circumstances to mean circumstances which do not usually occur during the ordinary course of arbitration. The commonplace situation where the tribunal acts in a procedurally deficient manner but does not prejudice either party could still qualify as a special circumstance, because parties in an arbitration expect that the tribunal will ordinarily comply with all procedural rules.

    Second, the suggestion that the other tests would result in too many stays being granted may not necessarily hold true either. Since 2012, when the IAA was amended to create section 10(9), there have only been two contested cases – AYY and BLY. And this remains the case even after AYY adopted the “irreparable prejudice” test in December 2015. The Court’s concern in this regard appears overstated. And third, even if many stays were granted, this may not mean that the default position is meaningless. What a default position does is not to regulate how many applications should be made or granted. Rather, it regulates which party has the burden of showing that the standard is met. And in this case, the burden of proof is on the party applying for the stay and the standard to be met is on a balance of probabilities. In contrast, where a stay application under section 6 of the IAA is made, all the applicant must show is a prima facie arbitral agreement before the burden shifts for the party resisting the stay to show that the arbitral agreement is null and void, inoperative, or incapable of being performed (see, eg, Tomolugen Holdings Ltd and another v Silica Investors Ltd and other appeals [2016] 1 SLR 373 at [65]). The distinct threshold in section 10(9) already gives meaning to the default position.

    The second reason the court gave to prefer the “special circumstances” test was the need for a balance between judicial control of the arbitration and not allowing stay applications to be used as a delay tactic. The Court said that this balance was already struck by the Model Law drafters in coming to this default position. Hence exercising the discretion, which removes a “fundamental feature” (BLY at [10]), must be compensated for by an equally compelling reason. I make two observations here. First, it is unclear whether the precise balance struck by the Model Law drafters as described by the Court applies with equal force in the modern version of the IAA which came into force in June 2012. The three “fundamental features” that the Court noted were: a short time period for recourse to the court; exclusion of appeals from the decision; and arbitrations continuing while the court hears the challenge (BLY at [11]). But as the Court also noted, the second feature has been removed by section 10(4) of the IAA, which explicitly allows appeals, with the leave of the High Court, to the Court of Appeal. There were also additional features inserted in the 2012 amendments: (1) a tribunal’s ruling that it has no jurisdiction can now be challenged in the High Court; and (2) the court can make an adverse costs order against any party if the jurisdictional challenge fails. These aforementioned amendments suggest that the balance in the IAA has been recalibrated, and is thus different from the previous position under Art 16(3) of Model Law.

    Even if the balance has remained unchanged, the features that have been added to the regime would suggest that section 10(9) does not necessarily point to a preference for the “special circumstances” test. In AYY, the amendment allowing the court to make adverse costs orders was cited by the AR in preferring the prejudice test. The AR noted that section 10(7) allows the court to make a costs order (AYY at [7]). Presumably, the AR’s thinking must have been that since section 10(7) already covers prejudice compensable by costs, the discretion in section 10(9) covers prejudice not compensable by costs. If so, the same reasoning could apply to the “balance of convenience” test, since it also applies the “prejudice” test, but to both the applicant and the respondent.

    Hence, the reference to the Model Law to justify the preference of the “special circumstances” test may be challenged. The reasons given by the Court thus do not convincingly support the adoption of the said test. To the contrary, the Court raised concerns which may arguably point towards an outcome-based test instead. To do so, I consider two examples the court gave of what would constitute “special circumstances”: (a) where the tribunal has acted in a manner that was manifestly and egregiously improper (BLY at [20]); and (b) where sensitive documents have been disclosed, such as information about government programs (BLY at [19], [24]).

    We begin with the second example. In BLY, the plaintiff sought a stay of the arbitration based on certain disclosure orders which the plaintiff said would prejudice its position in the future. The Court disagreed with the plaintiff. But the Court suggested that extra-sensitive information may qualify as a special circumstance. This is puzzling. If we recall, the “special circumstances” test looks at the situations of the parties, and not the outcome reached. If the circumstances ordinarily arise from arbitration, then whether there is prejudice is immaterial – the court should not intervene. A strict application of this test here would give the same conclusion: if the government were party to an arbitration, and the next step was for the tribunal to order disclosure as a natural course of the arbitration, then surely the court should not intervene either. But the Court in BLY expressed the view that this may constitute a special circumstance. This indicates that what the Court was focusing on was actually the outcome, since the only difference between normal disclosure and disclosure of government documents is that the latter outcome is far more prejudicial.

    We can see the Court’s interest in the outcome in the first example as well: where the tribunal acts improperly. Undoubtedly, the tribunal acting improperly would fall afoul of the “special circumstances” test since tribunals are never meant to act improperly in the ordinary course of arbitration. But let us stretch this example a little. Imagine that the breach was slight. Perhaps the tribunal constituted itself in 31 days instead of 30 days. Ordinarily, the parties would not have expected the tribunal to do so because it is not par for the course in arbitration. But this results in little prejudice, if any, to the party seeking a stay. What would the court say? It is likely that the court would refuse to grant the stay. The Court in BLY hinted at this when it tagged on the qualifier “manifestly and egregiously” to the supposed impropriety (BLY at [20]). This is another hint that what matters to the court is the extent of prejudice.

    I recognize that the Court expressly rejected the outcome-based prejudice test for being both over- and under-inclusive. It is over-inclusive because orders given during arbitration such as disclosure orders are always ‘prejudicial’ to one party. It is under-inclusive because it would not capture a situation of a mutually-agreed stay (BLY at [20]). And it is true that the prejudice test is over-inclusive in the sense that most orders give an advantage to one party over another. But it does not follow that simply because the prejudice test is over-inclusive, we should stop looking at the outcome altogether and focus on the process instead. It is equally possible to raise the threshold of what is an unacceptable outcome such that a stay should be granted.

    As for the issue of under-inclusiveness, it is true that the “irreparable prejudice” test does not accommodate the situation of a mutual stay. But arguably neither does the “special circumstances” test: it may be eminently within the ordinary course of arbitration that both parties may wish to hit a pause button where there are parallel proceedings on the same issue. Indeed, a mutually-agreed stay may not be the best example to use, since in such situations there is no need to invoke the court’s discretion under section 10(9). In that sense, a mutual agreement was never meant to be caught by the section.

    Accordingly, the Court’s concerns appear to be directed at the outcome, specifically, that the prejudice test would result in too many stays.


    Advocating a balancing test

    As discussed in the preceding section, the Court’s concerns may be more adequately addressed by an outcome-based test which adopts a higher threshold than the prejudice test. This sounds like the “balance of convenience” test, which balances the respective prejudice to the applicant and the respondent. Presumably, a stay will only be granted where the former outweighs the latter.

    The usefulness of this test is seen when we apply it to disclosure orders. The court would first acknowledge that disclosure orders prima facie prejudice the applicant as it reveals otherwise confidential information. But this potential prejudice can be minimized as the applicant can still persuade the tribunal not to grant the order, and even if ordered, can still be controlled using a confidentiality clause like the one in BLY. This minimized prejudice is balanced against the prejudice to the respondent: (1) the respondent’s expectations of an award are delayed, and (2) even if the court agrees with the respondent that the tribunal has jurisdiction, costs orders are not available because section 10(7) limits costs orders to the court’s finding that there is no jurisdiction. On balance, the respondent would likely suffer more prejudice and no stay would be granted.

    Indeed, these were precisely the same factors considered by the Court in applying the “special circumstances” test. The Court first noted that the plaintiff did not show how the information was “so sensitive and confidential” that it warranted protection by a stay (BLY at [23]). This implicitly acknowledges that there is some (but insufficient) degree of confidentiality. Next, the Court noted the possibility of protecting this confidentiality interest by a confidentiality clause (BLY at [25]). The Court’s analysis ended there because the “special circumstances” test only considers the plaintiff’s position. But if the Court had gone on to look at the defendants’ position, the analysis suggested above may have been adopted.

    Finally, the Court’s reticence to adopt the “balance of convenience” test is understandable. The phrase harks back to the doctrine of forum non conveniens, which has different policy concerns from the policy of minimal curial intervention in international arbitration. But nevertheless, if we consider the Court’s substantive concerns, I suggest that a test that balances the outcomes on both sides is preferable, and would indeed still accommodate any policy concerns that the court may have; although it should be framed in a manner that more accurately captures the Singapore courts’ pro-arbitration stance. A closer parallel, if the court were inclined to draw one, would perhaps be the “strong cause” test that is required to depart from exclusive jurisdiction clauses. This would be more suitable in highlighting the courts’ policy concerns while still allowing the court to consider the prejudice to both parties.

    * The opinions contained in this commentary reflect the authors’ own views and are not to be understood as reflecting the views of the authors’ employers.

    ** This blog entry may be cited as Leong Hoi Seng Victor, “Examining the “special circumstances” test in BLY v BLZ [2017] SGHC 59: when should jurisdictional challenges result in stays of the arbitral proceedings?” (21 April 2017) (

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

Comment Section