1 This is Mr Foo Kok Boon’s application in HC/OA 607/2023 (“OA 607”) to set aside the consent interlocutory judgment entered against him on 2 May 2019 (“the CIJ”). He is the third defendant in HC/S 834/2021 (“the Suit”). For ease of explication, I will refer to him as the “applicant” in this judgment.

2  The present application has arisen because of my decision in Salmizan bin Abdullah v Crapper, Ian Anthony [2023] SGHC 75 (“Salmizan”). In Salmizan, I decided, among others, that a defendant, who has entered into a interlocutory judgment in respect of liability, whether by consent or not, cannot challenge causation at the assessment of damages stage (“AD stage”) in a personal injury arising out of motor vehicle accidents (“PIMA”) claim (see Salmizan at [146(d)]). It follows that for a plaintiff to make out his cause of action in respect of every head of damage he is claiming for, he needs to establish causation for every head of damage before an interlocutory judgment on liability can be entered in respect of his entire claim (see Salmizan at [119] and [146(c)]). However, before Salmizan, an interlocutory judgment on liability may be granted in respect of the plaintiff’s entire claim even if the plaintiff was not able to establish causation for every head of damage at that stage of the proceedings. This was certainly the case at the time that the applicant entered the CIJ.

3 Relying on the principles in Salmizan, the applicant ostensibly applies to set aside the CIJ. The application is made on the ground that Mr Ngow Kheong Shen, who is the plaintiff in the Suit and the first respondent in the present application, did not establish causation for every head of damage when the CIJ was entered, and therefore had no basis to enter into the CIJ. ¹ However, at the hearing before me, Mr Anthony Wee (“Mr Wee”), who appeared for the applicant, explained that the real purpose of the present application is to seek a ruling that Salmizan should only apply prospectively. Mr Wee explained that he had to do this by applying for the CIJ to be set aside because he did not think the applicant has standing to challenge the effect of Salmizan directly. By asking the court to rule that Salmizan should only apply prospectively, the applicant is in effect taking the position that the CIJ should not be set aside. Therefore, while there is a seeming inconsistency between the applicant arguing for the CIJ to be set aside and for Salmizan to apply only prospectively, Mr Wee’s explanation resolves this issue.

4 Having considered the parties’ submissions, I dismiss the applicant’s application to set aside the CIJ. I also hold unequivocally that the doctrine of prospective overruling should apply in relation to Salmizan. The precise effect of this decision would be that a defendant who entered into an interlocutory judgment, whether by consent or not, prior to the date of the decision in Salmizan (ie, 30 March 2023) is entitled to raise issues of causation at the AD stage, even in respect of all the damage that the claimant claims to have suffered.

5 I now set out the reasons for my decision.

6 Before I turn to the applicant’s application properly, I first deal with the correct procedure to set aside the CIJ. Prior to making the present application in OA 607, the applicant had proceeded by way of HC/SUM 1288/2023 (“SUM 1288”), as a summons within the Suit. However, the applicant decided to take out the present originating application instead after the decision of the learned Deputy Registrar Samuel Wee (“DR Wee”) in Muhammad Tirmidzi Bin Misnawi v Agnes Chai Yui Yun [2023] SGDC 100 (“Muhammad Tirmidzi”). ²

7 In my view, the applicant is correct in taking out a fresh application to set aside the CIJ. I gratefully adopt DR Wee’s clear and careful analysis in Muhammad Tirmidzi, which led him to conclude that a party seeking to set aside a consent interlocutory judgment ought to begin fresh proceedings to do so instead of trying to do that by way of a summons in the prior case (see Muhammad Tirmidzi at [40]). It is not necessary to repeat DR Wee’s analysis, save to say that: (a) the legal authorities, such as the Court of Appeal decision of Siva Kumar s/o Avadiar v Quek Leng Chuang and others [2021] 1 SLR 451 at [34], are clear that an application to set aside a consent judgment or order must be commenced as fresh proceedings; (b) the Rules of Court (2014 Rev Ed) (“ROC 2014”) do not provide the court with a residual discretion to set aside contractual consent orders; and (c) parties cannot by consent confer on the court a jurisdiction which the court otherwise lacks (see Muhammad Tirmidzi at [17]).

8 With this preliminary point in mind, I turn to examine the applicant’s position in the present application to set aside the CIJ.

9 By way of background, the Suit arose out of a chain collision on 6 July 2015, which involved seven cars near the Moulmein Exit of the Central Expressway (“the Accident”). ³ At the time the CIJ was entered on 2 May 2019, the plaintiff’s pleaded claim was for, among others, general damages for the following physical injuries: (a) whiplash; (b) shoulder injuries; (c) increased risk of developing osteoarthritis; (d) persistent depressive disorder (dysthymia); and (e) heart attack. ⁴

10 Subsequent to the CIJ being entered, the applicant’s insurers appointed three doctors to examine the plaintiff. ⁵ In their respective reports, the doctors concluded, among others, that: (a) the Accident did not cause the plaintiff’s heart attack; (b) the plaintiff’s alleged shoulder injuries were chronic, and it was difficult to ascertain if the tears were due to degeneration or trauma; and (c) although the plaintiff suffered from dysthymia after the Accident, the symptoms of such were perpetuated by a back-to-back bereavement as well as work stress. ⁶

11 As such, the applicant broadly contends that he is not liable to compensate the plaintiff for all pre-existing injuries or any other alleged damage which was not caused by the applicant’s breach of duty. ⁷ Further, the applicant argues that the existing trial dates in October 2023 can remain as the issues in dispute remain the same. ⁸ At the hearing before me, Mr Wee argued that Salmizan should only apply prospectively as the decision has created practical difficulties in cases where the interlocutory judgment was entered into before Salmizan was decided.

12 As against all these points raised by the applicant, Mr John Vincent (“Mr Vincent”), who appeared for the plaintiff, objected to the CIJ being set aside on the basis that the law was different at the time the CIJ was entered into. ⁹ However, Mr Vincent also aligned himself with Mr Wee’s argument that Salmizan should only apply prospectively.

13 As I have already alluded to, the main issue raised in the present application is whether Salmizan should apply only prospectively, that is, from the date of its decision, and not retroactively, that is, before the date of the decision as well.

14 The applicable law in this regard has been settled by the Court of Appeal decision of Adri Anton Kalangie v Public Prosecutor [2018] 2 SLR 557 (“Adri”), which I am bound by (see also the insightful article by Gary K Y Chan, “Prospective Overruling in Singapore: A Judicial Framework for the Future?” in Comparing the Prospective Effect of Judicial Rulings Across the Jurisdictions (Eva Steiner, ed) (Springer Verlag, 2015) ch 16). As Sundaresh Menon CJ held for a unanimous court in Adri at [27], a judgment pronouncing on a legal issue is traditionally taken to be unbounded by time. Such a judgment would therefore have both retroactive and prospective effect. The practical effect of the retroactive effect of judgments is that a judgment would attach new consequences to events that occurred prior to the pronouncement of the new law. Applied in the present case, if Salmizan is to have retroactive effect, then the CIJ ought to be set aside even if the law was different at the time it was entered into. However, it is possible for a court to restrict the retroactive application of its judgment. This is known as the doctrine of prospective overruling, which was the subject of extensive analysis in Adri.

15 It is not necessary for me to revisit the applicable principles in relation to the doctrine of prospective overruling, as I respectfully agree with (and, in any event, am bound by) the analysis of the doctrine in Adri. It suffices for me to summarise the applicable law along the four principles laid down by Menon CJ in Adri at [39]–[44]:

(a) First, it is only in an exceptional case that a court may exercise its discretion to invoke the doctrine of prospective overruling. Indeed, such exceptionality is likely to be even more prominent in the context of civil cases, for which the Court of Appeal has observed that “in contrast to criminal cases, civil cases presenting exceptional circumstances that justify invoking the doctrine of prospective overruling are likely to be few and far between” (see L Capital Jones Ltd and another v Maniach Pte Ltd [2017] 1 SLR 312 (“L Capital”) at [71]).

(b) Second, it follows from the exceptionality of the doctrine of prospective overruling that it should only be invoked in circumstances where a departure from the ordinary retroactivity of a judgment is necessary to avoid serious and demonstrable injustice to the parties or to the administration of justice. Indeed, Menon CJ observed in Adri (at [40]) that such a high threshold of “necessity”, as well as the need for serious and demonstrable injustice, are consistent requirements across much of the case law in the Commonwealth. In this regard, the four factors identified in the High Court decision of Public Prosecutor v Hue An Li [2014] 4 SLR 661 (“Hue An Li”) at [124] are relevant, viz, (i) the extent to which the pre-existing legal principle or position was entrenched; (ii) the extent of the change to the legal principle or position; (iii) the extent to which the change in the legal principle or position was foreseeable; and (iv) the extent of reliance on the legal principle or position.

(c) Third, as the Court of Appeal alluded to in Hue An Li at [124], judicial pronouncements are by default retroactive. Therefore, until and unless the appropriate appellate court explicitly states that a judicial pronouncement is to take effect only prospectively, that pronouncement should presumptively be taken to be unbounded by time. This third principle raises a separate point that only the “appropriate appellate court” can pronounce on the prospective effect of its judgment. I will return to this point later (see below at [18]–[25]).

(d) Fourth, the onus of establishing that there are grounds to limit the retroactive effect of a decision should ordinarily be on whoever seeks to do so (see the New Zealand Supreme Court decision of Lai v Chamberlains [2007] 2 NZLR 7 at [154]).

16 With the applicable law in mind, I turn now to the present application.

17 Applying the principles laid down in Adri, I conclude that this is an exceptional case for the doctrine of prospective overruling to apply to a civil case.

18 As a preliminary point, I deal with the requirement Menon CJ laid down in Adri (at [43] and more explicitly at [70(a)]) that only the “appellate courts”, which he defined as “the Court of Appeal and the High Court sitting in an appellate capacity”, have the discretion to invoke the doctrine of prospective overruling in exceptional cases.

19 The problem in the present application is that I am no longer sitting in any judicial capacity in relation to Salmizan. Moreover, even within the context of my deciding of Salmizan, I was not sitting in an appellate capacity. This is because Salmizan was an uncontested application pursuant to O 33 r 2 of the ROC 2014 for the General Division of the High Court (“the High Court”) to determine three preliminary questions of law. As such, the issue of prospective overruling was never argued. It was in that context that I expressly reserved the question of whether Salmizan should only apply prospectively in the following manner (see Salmizan at [143]):

20 In my view, notwithstanding that this court is not presently sitting as an appellate court in relation to Salmizan or the underlying case therein, this court possesses the jurisdiction to pronounce on the prospective application of Salmizan. I say this for the following reasons.

21 First, in Salmizan itself, I had expressly reserved the question of whether the conclusions reached in Salmizan should apply prospectively. In so far as it is necessary to decide this point, I hold that the effect of my reservation in Salmizan is to preserve the court’s jurisdiction to decide the point at a more appropriate time, that is, when full arguments are made on it.

22 Second, while I was not formally sitting in an appellate capacity in Salmizan, the nature of the application made this substantively the case. This is because Salmizan (by which I refer to the specific application under O 33 r 2 of the ROC 2014) had originated from MC/MC 8815/2020 (“MC 8815”), which remained live when Salmizan was decided. Indeed, MC 8815 was transferred from the State Courts to the High Court for the determination of the preliminary questions of law before (so the defendant in MC 8815 represented) MC 8815 would be transferred back to the State Courts for the assessment of damages to continue. As such, the legal principles articulated in Salmizan would be binding on the proceedings of MC 8815 below. To that extent, and for the purposes of exercising the jurisdiction to pronounce on the prospective effect of Salmizan, I hold that I was sitting in a capacity akin to an appellate capacity in Salmizan.

23 Third, and in any event, the High Court in Hue An Li did say that the doctrine of prospective overruling can also apply to a “first-time judicial pronouncement” (at [125]). This would include situations where the court making the pronouncement is not sitting in an appellate capacity. As such, the requirement that only an appellate court can invoke the doctrine of prospective overruling is not cast in stone. It can be departed from in appropriate cases, and I find that this is an appropriate case to do so.

24 Finally, leaving aside the legal analysis (which remain undoubtedly of first importance), this appears to be the most practical outcome. If I were to hold that I have no jurisdiction to decide on the prospective effect of Salmizan, then, given the manner in which Salmizan had come to the High Court in the first place, there would be no other court to pronounce on this issue. In this regard, I understand that the defendant in Salmizan has applied for permission to appeal the matter to the Appellate Division of the High Court (“Appellate Division”). ¹⁰ It may well be that the Appellate Division will sit formally as an appellate court if it gives permission for the appeal to proceed. However, as of now, it is not known if the Appellate Division will do so. And even if it does so, it may be some time before it renders a substantive decision. It would not be practical to hold parties, such as those in the present case, in abeyance in the meantime.

25 For all these reasons, I decide that this court possesses the jurisdiction to pronounce on the prospective application of Salmizan.

26 Moving on, I decide that the doctrine of prospective overruling should apply in relation to Salmizan. In coming to this decision, I am conscious that the Court of Appeal had stated in L Capital at [71] that the civil cases which would attract the application of prospective overruling are “likely to be few and far between”. However, “few and far between” does not mean never, and in an exceptionally appropriate case, a court should call out those “few and far between” cases. This is one such case for the following reasons.

27 First, it is clear that the previous approach, where parties could reserve all issues of causation to the AD stage, was well-entrenched. Indeed, that this was the case can be seen from Form 9I of the State Courts Practice Directions 2014, which had, prior to 1 March 2021, permitted parties to leave “the issues of damages and causation to be assessed and costs reserved to the Registrar assessing the damages” [emphasis added]. This was why the learned Deputy Registrar Vince Gui (“DR Gui”) explained in the Magistrate’s Court decision of Kek Lai Quan (Guo Laiquan) v Lim Junyou [2022] SGMC 7 at [8] that, prior to Tan Woo Thian, “it was common practice for parties to motor accident claims to enter interlocutory judgment by consent on the understanding that the defendant would be allowed to challenge the causation of injuries in toto at the assessment of damages”. In practice, DR Gui said that this arrangement was often reflected by an express reservation in the terms of the interlocutory judgment.

28 To similar effect are the statements of the learned Deputy Registrar Kevin Kwek in the Magistrate’s Court decision of Eliora Yow (an infant suing by her father and litigation representative, Yow Tuck Meng Jerry) v Kwa Kian Peng [2020] SGMC 44 at [10], and those of the learned Deputy Registrar Lewis Tan in the District Court decision of Gannison s/o Varimuthu v Choa Beng Teck [2023] SGDC 92 (“Gannison”) at [15]. Moreover, as the applicant submitted in his application, the Ministry of Law’s website had published the “Legal Aid Bureau’s Practitioner’s Guide to Accident & Personal Injury Claims” in October 2021, in which it was stated, in the section on “determining quantum” [emphasis added], that the injuries claimed for needs to be “clearly caused by the accident”. This suggests that under the previous approach, issues of causation were regarded as going towards quantum, and not liability, and could therefore be reserved to the AD stage. Having regard to all these statements, it is safe to conclude that the previous approach was well entrenched.

29 Second, it is clear that the departure from the previous approach has been extensive. This shift began with the Court of Appeal decision in Tan Woo Thian v PricewaterhouseCoopers Advisory Services Pte Ltd [2021] 1 SLR 1166 (“Tan Woo Thian”). In that case, Menon CJ held ex tempore at [8] that in a bifurcated trial, the plaintiff at the liability stage would need “to show that he did, in fact, suffer one or more types of loss that was causally connected to the alleged breach” [emphasis in original]. This “made it impossible to challenge causation in toto at the AD Stage” (see Salmizan at [70]). Thereafter, several State Courts decisions that were decided after Tan Woo Thian (but before Salmizan) interpreted Tan Woo Thian to mean that “causation can reasonably be disputed at the assessment of damages phase as long as at least some damage or loss was caused by the defendant’s negligence” [emphasis in original] (see the Magistrate’s Court decision of Lim Mei Choo (Lin Meizhu) v Muhammad Azham bin Razak (Direct Asia Insurance (Singapore) Pte Ltd, intervener) [2021] SGMC 74 at [21]). When the issue arose before me in Salmizan, I clarified that it was not possible to dispute causation to any extent at all at the AD stage (see Salmizan at [82] and [116]). This short summary shows that the departure from the previous approach, where parties could reserve all issues of causation to the AD stage, has been extensive.

30 Third, it is clear that the departure from the previous approach was not foreseeable. This is principally because even the relevant court form prior to 1 March 2021 provided that parties could reserve all issues of causation to the AD stage. Also, secondarily, before Tan Woo Thian was decided, there was simply no reason for anyone to doubt the incorrectness of the previous approach.

31 Fourth, there would have been much reliance on the previous approach. This is because, in the context of PIMA cases, claimants would have to comply with the “Pre-Action Protocol for Personal Injury Claims and Non-Injury Motor Accident Claims” (“the Protocol”) before commencing court proceedings, which very purpose is to cater for the early settlement of such cases (see the State Courts Practice Directions 2014 at para 38(1)(b) and the State Courts Practice Directions 2021 at para 39(2)). As one possible outcome under the Protocol, there would be many cases in which the parties had agreed on liability and entered into a consent interlocutory judgment on such agreement (see Salmizan at [3]–[6]). In some of those cases, the parties may have, in reliance on the previous approach and the relevant court form, consented for an interlocutory judgment to be entered into but reserved issues of causation to be decided entirely at the AD stage. Separately, even in the situation where the parties did not agree on liability, there would be defendants in PIMA cases who might have decided not to dispute issues of causation at the trial on liability in reliance on their belief that issues of causation could be canvassed at the AD stage.

32 Taking all these factors into account, I am satisfied that a departure from the ordinary retroactive effect of Salmizan is necessary to avoid serious and demonstrable injustice to the parties at hand or to the administration of justice in general (see Adri at [40]). In particular, allowing Salmizan to apply retroactively would mean the unravelling of many interlocutory judgments that were entered into on the basis of the previous approach. This would lead to “gravely unfair and disruptive consequences for past transactions or happenings”, which is what the doctrine of prospective overruling seeks to avoid (see the House of Lords decision of In re Spectrum Plus Ltd (in liquidation) [2005] 2 AC 680 at [40], cited with approval in Adri at [40]). This is because underlying the doctrine of prospective overruling is the broader consideration of giving effect to the rule of law, which requires, among others, that “the law should be such that people will be able to be guided by it” (see Joseph Raz, The authority of law: Essays on law and morality (Oxford University Press, 1979) at p 213). The Court of Appeal elaborated on this in Adri at [30] when it stated that prospectivity in judicial pronouncements can advance the rule of law in so far as “it enables one to arrange his affairs on the basis of the law then prevailing without fear that his rights or obligations would be affected by subsequent judicial pronouncements to the contrary”.

33 Indeed, this consideration was present in Public Prosecutor v Dato’ Yap Peng [1987] 2 MLJ 311, where the Supreme Court of Malaysia (presently known as the Federal Court) applied the doctrine of prospective overruling because the previous legal position in respect of the constitutionality of a criminal provision had been relied on for some 11 years, with convictions and acquittals secured under that provision as a result (at 320–321). Likewise, in Re Manitoba Language Rights [1985] 1 SCR 721, the Canadian Supreme Court granted a declaration of temporary validity in respect of its decision that all existing provincial legislation was unconstitutional for failing to be enacted and published in both official languages, French and English. In order to preserve the rule of law, the Canadian Supreme Court regarded it necessary to temporarily deem valid and effective the Acts of the Manitoba Legislature, which would have been valid if not for this constitutional defect. The court held that this period would run from the date of the court’s judgment to the expiry of the minimum period necessary for translation, re-enactment, printing, and publishing (at [107]–[109]).

34 Just like the above-mentioned cases, I am satisfied that allowing the doctrine of prospective overruling to operate in relation to Salmizan would further the interest of allowing parties to be guided by the law and arrange their affairs on the basis of the law then prevailing. To otherwise allow the principles in Salmizan to take effect retroactively would, as I alluded to above, frustrate the expectations of the defendants in numerous PIMA cases who might have entered into a consent interlocutory judgment or conducted the proceedings in a certain way on their belief that they would be able to challenge causation entirely at the AD stage. Indeed, if the interlocutory judgments were set aside, it would greatly increase the time and complexity of the proceedings. In this regard, the observations of DR Tan in Gannison at [21] illustrate this problem:

35 Further, while not directly relevant to the principles in Adri, it is of practical relevance that the applicant in Salmizan has sought permission to appeal to the Appellant Division. In this regard, the problem of increased time and complexity might be exacerbated if Salmizan operates retroactively, and the Appellate Division eventually decides to overrule Salmizan. This is because if Salmizan operates retroactively, parties who had entered into a consent interlocutory judgment before Salmizan might have reacted to the new approach in Salmizan but subsequently find themselves back to square one if Salmizan is overruled. As such, considering the potentially disruptive effects that the retroactive application of Salmizan might have in view of the potential appeal, this is yet another reason for the doctrine of prospective overruling to apply to that decision.

36 For all the reasons above, I dismiss the applicant’s application in OA 607. The CIJ is to stand, and all timelines are to remain in that case.

37 More broadly, as is required by Adri at [70(d)], I state unequivocally that the doctrine of prospective overruling should apply in relation to Salmizan. The precise effect of this holding is that a defendant who entered into an interlocutory judgment (whether by consent or not) prior to the date of the decision in Salmizan (ie, 30 March 2023) is entitled to raise issues of causation at the AD stage, even in respect of all the damage that the claimant claims to have suffered. In other words, Salmizan would not apply to affected cases where an interlocutory judgment was entered prior to 30 March 2023, but only to cases where an interlocutory judgment was entered into on or after 30 March 2023. For avoidance of doubt, this also means that Salmizan would apply to affect cases that were commenced before 30 March 2023 but in which interlocutory judgment was not entered on or after 30 March 2023.

38 For completeness, I make no order as to SUM 1288.

39 Unless they are able to agree, the parties are to tender written submissions on the appropriate costs order for both OA 607 and SUM 1288 within 14 days of this decision, limited to seven pages each.