1 The desire to do justice is at the heart of the legal process. In the context of criminal justice, that objective is commonly understood to mean the proper adjudication of guilt and the determination of truth. Yet, as we observed in Kho Jabing v Public Prosecutor [2016] 3 SLR 135 (“Kho Jabing”) at [47], we must never lose sight of another equally important function of justice, which is the attainment of finality. After all, to permit “an endless repetition of inquiry into facts and law in a vain search for ultimate certitude implies a lack of confidence about the possibilities of justice that cannot but war with the effectiveness of the underlying substantive commands” (see Paul M Bator, “Finality in Criminal Law and Federal Habeas Corpus for State Prisoners” (1963) 76(3) Harv L Rev 441 (“Bator”) at 452). There must therefore come a point where a concluded court decision may legitimately be left in a state of repose, unencumbered by the prospect of further judicial review (see Mackey v United States 401 US 667 (1971) (“Mackey”) at 683).

2 The balance between the search for truth and the attainment of finality is one which every legal system will inevitably have to negotiate. A system that leaves no room for corrigibility risks the ultimate injustice of condemning the innocent and allowing the guilty to go unpunished. On the other hand, a system that caters to a perpetual and unreasoned anxiety of error would do violence to the search for closure that the pursuit of every legal contest awaits (see Kho Jabing at [50]). In Kho Jabing, we examined the balance between truth and finality in the context of legal arguments that were belatedly raised after the appellant had been convicted and his appeal, dismissed. The applications before us raised the related but distinct issue of when a court may reopen an earlier decision on the basis of subsequent changes in the law. We therefore considered that these applications presented an opportunity to clarify the conditions under which such a change can form the basis for reopening a previous court decision that was properly made in accordance with the law at the material time, a subject which has hitherto not been considered in our jurisprudence. In this judgment, when we refer to a subsequent change in the law, we mean one that is effected by a judicial decision, unless otherwise indicated.

3 We begin by setting out the relevant background facts. The backdrop to these applications is our decision in Saravanan Chandaram v Public Prosecutor and another matter [2020] 2 SLR 95 (“Saravanan”), in which we considered the permissibility of the Prosecution’s “dual charging practice”. Under the “dual charging practice”, where a single compressed block of cannabis-related plant material was certified by the Health Sciences Authority as containing (a) cannabis and (b) fragmented vegetable matter containing cannabinol (“CBN”) and tetrahydrocannabinol (“THC”), the Prosecution would consider preferring both (a) a charge in respect of the portion certified to consist purely of cannabis, and (b) a charge in respect of the portion consisting of fragmented vegetable matter that had been found to contain CBN and THC. We held in Saravanan (at [197] and [198(c)]) that the Prosecution’s “dual charging practice” was impermissible. More recently, we reaffirmed the impermissibility of the Prosecution’s “dual charging practice” in Abdul Karim bin Mohamed Kuppai Khan v Public Prosecutor [2021] 1 SLR 1390 (“Abdul Karim”) at [36].

4 The respondents in these applications had been charged with and convicted pursuant to the Prosecution’s “dual charging practice”. Following our decision in Saravanan, the Prosecution filed these applications inviting us to review our previous decisions in relation to the respondents, set aside the cannabis mixture charges that had been preferred against them in accordance with the “dual charging practice” and consequently reconsider the sentences imposed on them. However, by way of a letter dated 15 February 2021 (“the PP’s Letter”), the Prosecution sought leave to withdraw all four applications. The Prosecution took the stance that Saravanan did not automatically apply to the respondents as their cases pre-dated Saravanan and had been decided in accordance with the prevailing law then.

5 After hearing the parties, we granted the Prosecution leave to withdraw the applications. We saw no juridical basis for refusing to allow the Prosecution to do so and the respondents did not contend otherwise. The Public Prosecutor has carriage of prosecutions and is constitutionally vested with prosecutorial power, exercisable at his discretion, in the conduct of any criminal proceedings (see Art 35(8) of the Constitution of the Republic of Singapore (1985 Rev Ed, 1999 Reprint) (“the Constitution”)). That discretion necessarily extends to the continuance or withdrawal of the applications before us. There was also no suggestion that the withdrawal of the applications would be tantamount to an abuse of either the judicial process or prosecutorial power.

6 The Prosecution’s withdrawal of the applications nevertheless leaves open the possibility of the respondents filing their own applications to seek to reopen their convictions and/or sentences. We therefore indicated to the parties that we would issue these written grounds to elaborate on the applicable threshold for the court to revisit its prior disposal of a case where there has been a change in the law. We also highlighted that, following our review of the authorities, the threshold was that of substantial injustice and would not easily be met. It is to this discussion that we now turn.

7 The starting position of our analysis is that every judgment of the court is final. This proposition applies with stronger force to decisions in concluded appeals. As we explained in Kho Jabing at [49], “[a]s we venture further along the criminal process, we must give greater presumptive weight to the veracity of the findings already made and accord greater prominence to the principle of finality”. In the same vein, we recently observed in Iskandar bin Rahmat v Public Prosecutor [2021] SGCA 89 at [2] that “the issuance of a final judgment by this [c]ourt brings an end to the legal process available to parties in relation to a criminal conviction or sentence”.

8 Given that concluded criminal matters, and particularly concluded criminal appeals, are final and cannot be reopened on their merits, it follows that a matter ordinarily cannot be reopened just because there has been a subsequent change in the law. Any reconsideration of a concluded matter based on a subsequent change in the law is necessarily concerned with the merits of that matter, which the court typically has neither the jurisdiction nor power to re-assess (see Koh Zhan Quan Tony v Public Prosecutor and another motion [2006] 2 SLR(R) 830 (“Koh Tony”) at [29]). The judicial reluctance to undo decisions that were properly made in accordance with the law as it was then understood is grounded principally in respect for the finality of judgments, which is the principle that litigation must at some definite point be brought to an end (see FTC v Minneapolis-Honeywell Regulatory Co 344 US 206 (1952) at 213). Once the trial and appellate processes have run their course, a presumption of finality and legality attaches to the conviction and sentence (see Barefoot v Estelle 463 US 880 (1983) at 887).

9 The principle of finality has been described, perhaps unfairly, in some quarters as the notion that “in most matters, it is more important that the applicable rule of law be settled than that it be settled right” (see Burnet v Coronado Oil & Gas Co 285 US 393 (1932) at 406). To those who subscribe to such a view, the principle of finality may appear to be at odds with the interests of justice; specifically, the correction of error and the search for truth. In our view, however, there are three main reasons why the principle of finality is not only compatible with, but also integral to, justice.

10 First, respect for finality maximises scarce judicial resources by channelling them towards more productive ends. Applications “litigating the validity under present law of criminal convictions that were perfectly free from error when made final” take up valuable resources which should instead go towards the disposal of cases being heard for the first time (see Mackey at 691). In contrast, the subjects of concluded criminal appeals have had two opportunities, before the trial and appellate courts, to defend their positions. Attempts to reopen concluded criminal appeals are wasteful because they essentially afford litigants a third bite of the cherry (see Ryan W Scott, “In Defense of the Finality of Criminal Sentences on Collateral Review” (2014) 4(1) Wake Forest Journal of Law and Policy 179 (“Scott”) at 185–186). This is not merely a logistical concern but one with profound implications for access to justice by the large number of other litigants.

11 Second, the finality of court judgments is crucial to the effectiveness of the deterrent and rehabilitative functions of the criminal justice system. It is essential to the deterrent function of the criminal law that “we be able to say that one violating [the] law will swiftly and certainly become subject to … just punishment” (see Bator at 452; see also Teague v Lane 489 US 288 (1989) at 309). The endless reopening of concluded criminal appeals is also injurious to the rehabilitation of offenders. After all, rehabilitation begins with an offender’s acceptance that he has been justly sanctioned and that he stands in need of re-education. The rehabilitative process cannot possibly begin if “the cardinal moral predicate is missing, if society itself continuously tells the [offender] that he may not be justly subject to [re-education]” in the first place (see Bator at 452).

12 Third, and most importantly, the finality of court decisions is fundamental to the very integrity of the judicial process. As we emphasised in Kho Jabing at [47], “[n]othing can be as corrosive of general confidence in the criminal process as an entrenched culture of self-doubt engendered by abusive and repetitive attempts to re-litigate matters which have already been decided”. An unbounded willingness to reopen concluded criminal appeals denies cases the closure that they deserve, thereby undermining the legitimacy of the judicial process itself.

13 For these reasons, final judgments, particularly those issued by an appellate court, will not be readily unsettled. After finality attaches, a judgment stands even if the law subsequently changes. However, there are two ways in which the court’s revisionary powers may be invoked to permit a departure from the default position of finality. These are: (a) the Court of Appeal’s inherent power to reopen a concluded criminal appeal; and (b) an appellate court’s statutory power to review its earlier decision under s 394I of the Criminal Procedure Code (Cap 68, 2012 Rev Ed) (“the CPC”). We now proceed to examine each of these in turn.

14 We first consider the Court of Appeal’s inherent power to reopen concluded criminal appeals. That such an inherent power exists has not always been settled law. Prior to 2010, the Court of Appeal held in several decisions that once it had delivered its judgment in a criminal appeal, its appellate jurisdiction ceased and it was functus officio. It was reasoned on this basis that upon the conclusion of a criminal appeal, the Court of Appeal did not have any statutorily-conferred jurisdiction to reopen a matter and reconsider its substantive merits: see, for example, Abdullah bin A Rahman v Public Prosecutor [1994] 2 SLR(R) 1017 at [10] and [13]; Lim Choon Chye v Public Prosecutor [1994] 2 SLR(R) 1024 at [8] and [12]; Jabar bin Kadermastan v Public Prosecutor [1995] 1 SLR(R) 326 at [58]; and Vignes s/o Mourthi v Public Prosecutor [2003] 4 SLR(R) 518 at [4]–[8]. In these cases, the Court of Appeal held that the conclusion of the appellate process meant that it was jurisdictionally foreclosed from hearing applications to adduce fresh evidence, staying a sentence of death or hearing a second appeal. Those decisions are but the logical extension of the default position set out at [7]–[8] above.

15 An exception to the default position was first identified in Koh Tony. The applicants there had been charged with murder but were convicted by the High Court of the lesser offence of robbery with hurt. Upon the Prosecution’s appeal in CA/CCA 2/2005 (“CCA 2/2005”), the Court of Appeal substituted the convictions for robbery with hurt with convictions for murder. The applicants then filed criminal motions to the Court of Appeal, contending that it did not have the jurisdiction to hear the substantive merits of CCA 2/2005 in the first place (“the Jurisdictional Issue”). According to the applicants, the Supreme Court of Judicature Act (Cap 322, 1999 Rev Ed) (“the 1999 SCJA”) only permitted the Prosecution to appeal against acquittals, whereas they had been convicted by the High Court, albeit of a lesser offence. The Prosecution argued that the Court of Appeal had no jurisdiction to consider the criminal motions as it had already disposed of the substantive merits of CCA 2/2005 and was thus functus officio.

16 The Court of Appeal held that it had both the jurisdiction and the power to hear the criminal motions. The Jurisdictional Issue ought to have constituted an integral part of CCA 2/2005 and the applicants were not estopped from raising that issue in the criminal motions (see Koh Tony at [23]). The Court of Appeal therefore concluded (at [23]) that it remained “properly seised of the present case in so far as the question of jurisdiction [was] concerned” [emphasis in original]. There was also no question that the Court of Appeal had the power to decide the Jurisdictional Issue by virtue of s 29A(4) of the 1999 SCJA (see Koh Tony at [23]–[24]). In summary, the Court of Appeal held that if an application to reopen a concluded criminal appeal was premised on a challenge to the Court of Appeal’s jurisdiction to hear the appeal in the first place, rather than a challenge to the substantive merits of the appeal, the court would remain seised of jurisdiction to hear the matter even if it had already ruled on the merits of the appeal (see Kho Jabing at [14]).

17 Koh Tony, however, did not go so far as to authorise the Court of Appeal to review the merits of an already concluded criminal appeal. The Court of Appeal only held that it had the jurisdiction and power to determine the Jurisdictional Issue; indeed, it took great pains to clarify that it remained functus officio in so far as the substantive merits of CCA 2/2005 were concerned (see Koh Tony at [22] and [29]). As the Court of Appeal emphasised (at [29]), its decision was “not a carte blanche for this court to review its previous decisions when it is truly functus officio. In particular, this court has neither the jurisdiction nor power to review the substantive merits of the case”.

18 After Koh Tony came our decision in Yong Vui Kong v Public Prosecutor [2010] 2 SLR 192 (“Yong Vui Kong”). The applicant in Yong Vui Kong had been convicted on a capital drug trafficking charge. He filed a notice of appeal but subsequently withdrew his appeal. He then filed a criminal motion petitioning the Court of Appeal to treat the previous withdrawal of his appeal as a nullity and to restore his appeal for hearing. The Prosecution argued that the Court of Appeal had no jurisdiction to hear the criminal motion because the applicant’s appeal, once withdrawn, was deemed to have been dismissed on its merits, with the result that the court was functus officio and could not hear any further appeal against his conviction.

19 The Court of Appeal held (at [26]–[28]) that the applicant’s withdrawal of his appeal was a nullity as he had laboured under a fundamental mistake at the material time. The applicant was thus allowed to proceed with his appeal. More importantly, the Court of Appeal observed (at [15]–[16]), albeit in obiter, that even if the substantive merits of the appeal had already been heard and decided, it might nonetheless have the jurisdiction to hear further arguments on those substantive merits:

As was noted in Kho Jabing (at [18]), the operative question in Yong Vui Kong was not whether the court could reconsider its previous decision in a concluded criminal appeal (in the sense of whether it had the jurisdiction to do so), but whether it ought to do so (in the sense of whether it ought to exercise its power to that end).

20 A gradual shift in the Court of Appeal’s approach can thus be discerned. Although it was held in Koh Tony that any review of an appellate decision was confined to the specific question of whether the court had the jurisdiction to hear the appeal, the court in Yong Vui Kong intimated its willingness to review the merits of its earlier decision if the interests of justice so required. This change in judicial attitude culminated in Kho Jabing, in which the Court of Appeal held (at [77(a)]) that, as the final appellate court in Singapore, it had the inherent power to reopen a concluded criminal appeal in order to prevent a miscarriage of justice. In line with the point raised by Chan Sek Keong CJ in Yong Vui Kong (at [16]), the Court of Appeal held that its inherent power to reopen a concluded criminal appeal was a facet of the judicial power vested in it by virtue of Art 93 of the Constitution. When the court exercised this “power of review”, it was acting within the scope of its statutorily-conferred appellate jurisdiction, which was not completely exhausted upon rendering a decision on the merits of the appeal (see Kho Jabing at [77(a)]).

21 When, then, might the court exercise its inherent power to reopen a concluded criminal appeal? An applicant for such relief had to satisfy the court that there was sufficient material on which it could conclude that there had been a miscarriage of justice (see Kho Jabing at [77(b)]). Material tendered in support of a review application would only be “sufficient” if it was “new” and “compelling” (see Kho Jabing at [52]). The Court of Appeal accepted (at [53]) that new legal arguments could form the basis of a review application if such arguments were “new”, in the sense that they: (a) had not been considered at an earlier stage of the proceedings; and (b) were material which could not, even with reasonable diligence, have been presented to the court before the filing of the review application. The Court of Appeal further observed (at [58]) that new legal arguments would, by and large, only be able to constitute the basis for a review application if such arguments were made following a change in the law.

22 In order for the material in support of a review application to be “compelling”, the material had to be “reliable, substantial and powerfully probative” such that it could show “almost conclusively” that there had been a “miscarriage of justice” (see Kho Jabing at [59]–[61]). The Court of Appeal indicated (at [60]) that a new line of legal authorities would be objective and, hence, potentially reliable material. Material would be considered “substantial” and “powerfully probative” if it was logically relevant to the precise issues which were in dispute (see Kho Jabing at [61]).

23 As mentioned at [21] above, the court would only exercise its inherent power of review if there was a “miscarriage of justice”. In general, the court would only find that there was a “miscarriage of justice” in the following situations (see Kho Jabing at [65], [69], [70] and [77(e)]):

(a) First, the court might find that a decision on conviction or sentence had been shown to be “demonstrably wrong”.

(i) In relation to a decision on conviction, the applicant had to show that it was apparent, based on the evidence tendered in support of the application alone, that there was a powerful probability that the decision was wrong.

(ii) In relation to a decision on sentence, the applicant had to show that the decision was based on some fundamental misapprehension of the law and was thus blatantly wrong on the face of the record.

(b) Second, the court might find that the impugned decision was tainted by fraud or a breach of natural justice, such that the integrity of the judicial process had been compromised.

24 The Court of Appeal ultimately declined to exercise its inherent power of review in Kho Jabing. This was because the legal arguments raised by the applicant at the eleventh hour were neither new nor compelling and thus failed to satisfy the prerequisites for the court’s exercise of its inherent power of review.

25 That is an overview of the Court of Appeal’s inherent power of review, as it has developed over time. We now turn to an appellate court’s statutory power of review under s 394I of the CPC.

26 Section 394I of the CPC sets out the procedure by which an appellate court may review its earlier decision. While only the Court of Appeal, as the final appellate court in Singapore, possesses the inherent power of review (see Kho Jabing at [77(a)]), the statutory power of review may be exercised by both the General Division of the High Court and the Court of Appeal, so long as the decision sought to be reviewed is a decision of an appellate court on the merits of an appeal (see s 394G(1)(a) of the CPC).

27 Section 394J of the CPC codifies the requirements for an appellate court’s exercise of its power of review as set out in Kho Jabing and provides as follows:

28 The speech of the then Senior Minister of State for Law, Ms Indranee Rajah (“the Minister”), at the second reading of the Criminal Justice Reform Bill is useful in shedding light on the genesis of the court’s statutory power of review. The Minister expressed that “[t]he re-opening procedure, including the relevant threshold tests, is largely a codification of a number of considered decisions made by the Court of Appeal in balancing the interests of finality against the need to prevent a miscarriage of justice” (see Singapore Parliamentary Debates, Official Report (19 March 2018) vol 94 (Indranee Rajah, Senior Minister of State for Law)). The decisions that the Minister referred to would have included our decisions in Kho Jabing and, perhaps to a lesser extent, Yong Vui Kong.

29 We make four observations on the court’s statutory power of review. First, s 394J(1)(b) of the CPC makes it clear that an appellate court’s statutory power of review does not affect its inherent power to review an earlier decision on its own motion. The same point was underscored by the Minister during the relevant Parliamentary debates.

30 Second, while an applicant may challenge the court’s decision in a concluded criminal appeal by invoking either the court’s statutory power of review under s 394I of the CPC or its inherent power (per Kho Jabing), the substance of any such application is typically unaffected by the choice of remedial avenue. This follows from the fact that s 394J of the CPC codifies the requirements set out in Kho Jabing, a fact which the Minister noted in Parliament. It is unsurprising that the requirements under s 394J of the CPC mirror those laid down in Kho Jabing; indeed, it might be somewhat arbitrary if the success of a review application were contingent on one’s choice of remedial avenue.

31 Third, and following from the point we have just made, even though the substance of a review application remains the same regardless of which remedial avenue is utilised, the two avenues are not duplicative. An applicant may make only one review application under s 394I of the CPC in respect of any decision of an appellate court (see s 394K of the CPC). This means that where “sufficient material” on which an appellate court may conclude that there has been a miscarriage of justice only emerges after a prior review application brought under s 394I of the CPC has been heard and dismissed, an applicant may have further recourse to the court’s inherent power of review but not to its statutory power.

32 Our fourth and final point is of particular importance for present purposes: Parliament had clearly contemplated that a subsequent change in the law could constitute the basis for a review application under s 394I of the CPC. This is made explicit in s 394J(4), which provides that any material consisting of legal arguments will only be “sufficient” within the meaning of s 394J(2) if, in addition to satisfying all the requirements set out in s 394J(3), it is based on “a change in the law that arose from any decision made by a court after the conclusion of all proceedings relating to the criminal matter in respect of which the earlier decision was made” [emphasis added].

33 To consolidate the foregoing analysis, judgments of the court are final and judgments of an appellate court cannot ordinarily be reopened on their merits. A subsequent change in the law without more would not constitute sufficient grounds on which the court may exercise either its inherent or statutory power of review (see [7]–[8] above).

34 We have consistently held that a mere change in the law would not justify the exercise of our revisionary powers: see Ang Poh Chuan v Public Prosecutor [1995] 3 SLR(R) 929 (“Ang Poh Chuan”) at [24] and Ng Kim Han and others v Public Prosecutor [2001] 1 SLR(R) 397 (“Ng Kim Han”) at [22]. The question that then arises is: under what conditions may an appellate court reopen the merits of an earlier decision following a change in the law? It is to this question that we now turn.

35 There is a dearth of local authorities to guide us. This is perhaps unsurprising: the Court of Appeal’s inherent power to reopen concluded criminal appeals was only expressly recognised in Kho Jabing in 2016, and the court’s statutory power of review came into effect later still on 31 October 2018. From our review of the case law, it appeared that since 2016 and until the present criminal motions were filed, only one review application had been made on the basis of a subsequent change in the law. That application culminated in our decision in Gobi a/l Avedian v Public Prosecutor [2021] 1 SLR 180 (“Gobi”), a case which we return to at [97]–[100] below. Given the paucity of local authorities on the issue at hand, some guidance may be usefully drawn from the English authorities concerning the analogous issue of granting leave to appeal out of time following a change in the law. It should be noted that the objection to granting such an extension of time is surely less weighty than that to reopening a concluded matter after the appellate process has been exhausted. Nonetheless, the principles in the English authorities are instructive.

36 As it turns out, one of the relevant English authorities concerns the proper interpretation of the term “cannabis”. This is very similar to the issue that we were confronted with in Saravanan, the decision upon which these criminal motions were founded. In R v Mitchell [1977] 2 All Er 168 (“Mitchell”), the appellant had been convicted on two counts of possession of cannabis with intent to supply, among other charges. The cannabis in question comprised the leaf, stalk and clean seeds of the cannabis plant. The appellant was granted leave to appeal against his sentence. Following the appellant’s conviction but before his appeal against sentence was heard, the English Court of Appeal held in a separate matter that the leaf and the stalk of the cannabis plant did not fall within the mischief targeted by the Misuse of Drugs Act 1971 (“the 1971 Act”). The appellant then sought an extension of time to appeal against his conviction. The central question in Mitchell was whether the clean seeds of the cannabis plant were “cannabis” within the meaning of s 37(1) of the 1971 Act. The English Court of Appeal (Criminal Division) held (at 170J–171B) that only those parts of the flowering or fruiting tops of the cannabis plant that contained cannabis resin were caught by the 1971 Act. Although the clean seeds were part of the fruiting tops of the cannabis plant, they did not contain cannabis resin and were hence not “cannabis” for the purposes of s 37(1) of the 1971 Act.

37 But that was not the end of the matter. The court went on to consider (at 171C) whether to grant the appellant an extension of time to file an appeal against his conviction, given that “[t]he appellant ha[d] been sentenced to a term of imprisonment of three years for an offence which on the facts was not a crime”. As the court explained, the starting position was that a mere change in the law would not justify granting an extension of time to file an appeal:

38 On the facts, the court allowed the appellant’s application for an extension of time and granted him leave to appeal against his conviction for two reasons. First, the court noted (at 171G) that the appellant had already been granted leave to appeal against his sentence. If it refused to grant the appellant an extension of time to appeal against his conviction, it would be faced with “the totally unreal task of endeavouring to determine what the correct sentence was for an offence which had not been committed”. The second factor that made the case “something of an extraordinary case” was the fact that the appellant was in prison. Although the appellant was serving a concurrent sentence for other offences, that concurrent sentence ought to have been over or nearly over as at the date on which his application was heard, assuming he had earned his full remission. Should the court refuse to grant him an extension of time to appeal against his conviction, it would effectively be keeping him in prison for an offence he had not committed (see Mitchell at 171H–171J).

39 The court was thus satisfied that the matter was a “very rare case” that warranted the exercise of its discretion to grant the appellant an extension of time and leave to appeal against his conviction (see Mitchell at 171J). Nonetheless, the court took care to clarify (at 172A) that its decision ought not to be construed as an invitation to every offender who had been convicted of similar offences to seek leave to appeal out of time, and emphasised that such applications “[would] not be greeted with very much enthusiasm”.

40 In arriving at its decision, the court also referred to the unreported decision of R v Ramsden [1972] Crim LR 547 (“Ramsden”), in which it was stated that “[w]here a subsequent decision of a superior court has produced an apparent change in the law, that coupled with other circumstances may be a factor which will induce the court to grant leave to appeal out of time. Nevertheless … this must in every case be a matter of discretion.” [emphasis added in italics and bold italics] (see Mitchell at 171D–171E).

41 The defendant in Ramsden had been convicted of dangerous driving before it was subsequently held that fault was a necessary element of that offence. Refusing to grant the defendant leave to appeal out of time, the court observed that alarming consequences would flow if it were to permit the general reopening of old cases on the ground that a subsequent decision had dispelled a widely held misconception about the law underpinning the defendant’s conviction. Had the court granted the extension of time sought, “[n]o doubt … everyone convicted of dangerous driving over a period of several years could have advanced the same application” (see also R v Jawad (Mohid) [2013] EWCA Crim 644 (“Jawad”) at [29]). In other words, the court in Ramsden had regard to the potential impact on other decided cases in deciding not to grant the extension of time sought.

42 The case of R v Hawkins [1997] 1 Cr App Rep 234 (“Hawkins”) is also instructive. The applicant there had pleaded guilty to and was convicted of five counts of obtaining property by deception under s 15(1) of the Theft Act 1968 (“Theft Act”), among other charges. Thereafter, it was separately held that a chose in action created by an electronic bank transfer was not “property belonging to another” that could be “obtained” by deception under s 15(1) of the Theft Act. The applicant cited this change in the law and sought leave to appeal against his conviction.

43 The court expressed that it had to be persuaded that there was a “good reason” for granting leave to appeal out of time, particularly where a defendant had pleaded guilty (as the applicant had) or had made a conscious decision not to appeal. Citing authorities such as Mitchell and Ramsden, the court noted that a change in the law after the date of conviction or plea of guilty would not usually be regarded as a good reason for granting leave to appeal out of time. The court further stated as follows:

Nonetheless, the court noted that despite the general rule against reopening a conviction where a defendant had pleaded guilty, it would “eschew undue technicality and ask whether any substantial injustice has been done” [emphasis added]. Having concluded that the applicant had not suffered any substantial injury, the court refused the extension of time sought.

44 What is of note is that in arriving at its conclusion, the court found it “arguable” that, based on the admitted facts, the jury could have convicted the applicant for other property offences. In our judgment, this is a significant factor that differentiates Hawkins from Mitchell. In Mitchell, the applicant had only been in possession of parts of the cannabis plant that did not contain cannabis resin and were hence not “cannabis” within the meaning of s 37(1) of the 1971 Act. It follows that he could not have been convicted on any alternative charges and that substantial injustice had arisen from his wrongful incarceration. On the other hand, the applicant in Hawkins had admitted to dishonest or fraudulent conduct which could have supported convictions for other offences. Any injustice that he had suffered was therefore less readily apparent and, in any event, did not appear to be “substantial”.

45 In the same vein, it was held in R v Ramzan and others [2006] EWCA Crim 1974 (“Ramzan”) that whether a defendant could be said to be guilty of other offences was a relevant factor when considering whether substantial injustice had been established. Ramzan concerned several appeals and applications by defendants who had been convicted of conspiracy to commit money-laundering offences. Following their convictions, there was a change in the law regarding the mens rea for conspiracy to launder money. The defendants appealed against or sought leave to appeal against their convictions.

46 The court stressed (at [39] and [54]) that even though the convictions for conspiracy could no longer stand, on the defendants’ own cases taken together with the juries’ verdicts, all the defendants must have committed one or more substantive offences of money-laundering for which the mens rea remained unchanged. Hence, even if the subsequent change in the law had been known at their trials, they would nonetheless have been convicted of “substantive offences of great gravity”. Accordingly, subject to any individual grounds for finding substantial injustice, “these [were] strong cases for refusal of leave”. In considering the individual cases, the court repeatedly stated that those defendants who must nonetheless have committed substantive money-laundering offences would not have suffered substantial injustice by virtue of their convictions for conspiracy. Leave to appeal was accordingly refused in those cases (see Ramzan at [70], [71], [73], [75] and [78]).

47 We next consider the decision of the House of Lords in R v Benjafield and others [2003] 1 AC 1099 (“Benjafield”). Benjafield was a set of consolidated appeals and applications by defendants who had been convicted of drug and property offences and against whom confiscation orders had been made. The Human Rights Act 1998 (“the 1998 Act”) came into force after those confiscation orders had been made. One of the defendants (“Mr Rezvi”) applied for leave to appeal against his sentence five months out of time. All the defendants contended that the statutory bases for the confiscation orders were incompatible with the right to a fair hearing and the presumption of innocence enshrined in Arts 6(1) and 6(2) of the Convention for the Protection of Human Rights and Fundamental Freedoms, which was scheduled to the 1998 Act.

48 The House of Lords granted Mr Rezvi’s application for leave to appeal out of time because it considered that his appeal raised issues of considerable general interest (see Benjafield at [46]). Nonetheless, the House of Lords expressly confirmed (at [46]) that “[i]t is not usual to grant leave to appeal out of time where the grounds of appeal are based on post-trial changes in the law”.

49 R v Ballinger [2005] EWCA Crim 1060 (“Ballinger”) is another case that illustrates the strictness of the principle that leave to appeal out of time will not ordinarily be granted where there has been nothing more than a change in the law post-conviction. The applicant in Ballinger had been convicted of indecent assault at a naval court martial presided over by a judge advocate. It was held in two subsequent decisions (one of which was R v Dundon [2004] EWCA Crim 621 (“Dundon”)) that courts martial were incompatible with the right to a fair hearing, a right protected under Art 6(1) of the European Convention on Human Rights (“ECHR”) and given effect to by the 1998 Act. The applicant sought leave to appeal against his conviction out of time. The Prosecution conceded that, if leave were granted, it would have no grounds for resisting the appeal.

50 Counsel for the applicant submitted that the applicant had suffered a substantial injustice, namely, a wrongful conviction by virtue of the Art 6(1) breach. The court rejected that argument and found that the applicant had failed to establish substantial injustice for three broad reasons. First, the court seemingly took the view that the applicant had, for the most part, not perceived himself to have suffered any substantial injustice. At no stage did he criticise the conduct of the judge advocate, and although he did not plead guilty, he did not seek to appeal against his conviction or challenge its safety until the decision in Dundon was issued (see Ballinger at [21]). Second, the court observed that the longer the delay in filing an appeal, the more compelling the reason had to be before it would grant leave to appeal out of time. The delay in this case was substantial, and even after the decision in Dundon had been rendered, there was a further delay of more than three months (see Ballinger at [22]–[23]). The extended delay reflected a calculated risk on the applicant’s part and also rendered a re-trial less feasible (see Ballinger at [23]). Third, the court stated that if the applicant were granted an extension of time to appeal, there would be no grounds for refusing to grant a similar extension of time to every other defendant who had been convicted before a court martial.

51 The court concluded that an applicant seeking leave to appeal out of time had to show “more than that there has been a breach of Article 6 and that in consequence his conviction is unsafe; he must also show that he has suffered a substantial injury or injustice”. The applicant had failed to do so and was refused leave to appeal out of time.

52 The implications of changes in law on previous convictions were squarely dealt with in R v Cottrell; R v Fletcher [2007] 1 WLR 3262 (“Cottrell”). The court started from the premise (at [42]) that “[i]n reality, society can only operate on the basis that the courts … apply the law as it is. The law as it may later be declared or perceived to be is irrelevant” [emphasis added]. The court then cited Mitchell and Benjafield and made the following observations (at [46]):

53 Subsequent to Cottrell, the United Kingdom Supreme Court handed down its seminal decision in R v Jogee; R v Ruddock [2016] 2 All Er 1 (“Jogee”). Jogee ushered a sea change in the established legal position on the mens rea for parasitic accessorial liability. Under the prevailing law prior to Jogee, if two people (“D1” and “D2” respectively) set out to commit an offence, and in the course of that joint enterprise D1 committed another offence (“crime B”), D2 would be guilty as an accessory to crime B if he foresaw the possibility that D1 might act as he did. It was held in Jogee (at [9], [10] and [83]) that the requisite mens rea for parasitic accessorial liability was not foresight but an intention to assist or encourage the commission of crime B; foresight was evidence of but not synonymous with intent.

54 Mindful that Jogee would likely spur a torrent of applications by offenders who had been found to be accessories to crimes on the basis of foresight rather than intent, the Supreme Court cautioned as follows (at [100]):

55 As anticipated, Jogee precipitated a flurry of applications for leave to appeal out of time based on the change in the law on parasitic accessorial liability. A series of these applications was addressed by the court in R v Johnson and other appeals [2017] 4 All Er 769 (“Johnson”). The court reiterated (at [18]) the principle that an applicant who had been “properly convicted on the law as it then stood” had to establish “substantial injustice” before leave to appeal would be granted. It also observed (at [18]) that the requirement of substantial injustice “takes into account the requirement … for a court to be able to alter or correct the law upon which a large number of cases have been determined without the consequence that each of those cases can be re-opened”.

56 Importantly, Johnson suggests that whether the change in the law would have made a difference to the outcome of the concluded case in question is strongly probative of whether substantial injustice has arisen. The court stated (at [21]) that “[i]n determining whether that high threshold [of substantial injustice] has been met, the court will primarily and ordinarily have regard to the strength of the case advanced that the change in the law would, in fact, have made a difference” [emphasis added].

57 In refusing to grant the applicants leave to appeal out of time, the court found that “there was no injustice, let alone substantial injustice” because the jury’s verdicts “would have been no different post Jogee” (see Johnson at [57], [83], [96], [123], [160] and [191]). In respect of two of the applicants, the court observed (at [84]) that it would ordinarily have granted them leave to challenge their convictions because their trial lawyers had reserved their position on the law on parasitic accessorial liability pending the determination of Jogee. However, the court ultimately refused to grant leave because a post-Jogee direction would not have made a difference to the jury’s verdicts (see Johnson at [83]). Put simply, the safety of those convictions was not in question. The overriding concern of the court was thus whether the change in the law would have had a bearing on the outcome of the concluded case in question.

58 R v Crilly (John Anthony) [2018] EWCA Crim 168 (“Crilly”) was the first instance of a court quashing a murder conviction on account of the change in the law brought about by Jogee. In brief, the applicant and two others (including one “Flynn”) had intended to commit a burglary in what they had expected to be an empty house. When the house was found to be occupied, the applicant unsuccessfully tried to persuade Flynn to leave. The applicant then searched the property for money. During the robbery, Flynn attacked and killed the occupant. The applicant was convicted of murder and robbery even though he did not inflict any violence upon the occupant.

59 Following the decision in Jogee, the applicant sought an extension of time of 11 years and three months to appeal against his murder conviction. He contended that the trial judge had misdirected the jury and that he had been wrongly convicted of murder on the basis of foresight, not intent. The court granted exceptional leave, quashed the applicant’s murder conviction and ordered a re-trial on the murder charge (see Crilly at [42], [43] and [50]). Leave was granted primarily because it would have made a difference to the applicant’s case if the jury had been properly directed in accordance with the law as set out in Jogee. The case against the applicant at trial was that he had foreseen that Flynn might cause grievous bodily harm, not that he had intended that Flynn cause such harm. Furthermore, the evidence was insufficient for the court to conclude that the jury would have found the requisite intent on the applicant’s part, had it been properly directed by the trial judge (see Crilly at [42]). In the circumstances, the court held that the issuance of Jogee-compliant directions to the jury would have made a difference, that the applicant’s conviction was unsafe, and that a refusal of exceptional leave to appeal out of time would result in substantial injustice. The decision in Crilly is thus in keeping with the court’s emphasis in Johnson (at [21]) that “[i]n determining whether that high threshold [of substantial injustice] has been met, the court will primarily and ordinarily have regard to the strength of the case advanced that the change in the law would, in fact, have made a difference” [emphasis added].

60 While “substantial injustice” is an undoubtedly high threshold, it is sufficiently expansive to permit consideration of injustice arising not only from the impugned conviction or sentence but also from the attendant consequences thereof. R v McGuffog (Correy Rowan) [2015] EWCA Crim 1116 (“McGuffog”) illustrates this point well. The applicant in that case had intended to make a right turn around a mini roundabout to enter a certain lane. At the material time, one Mr Duffill had been cycling along a separate lane towards the same roundabout. Mr Duffill was visible to the applicant, who had the right of way. Assuming that Mr Duffill would stop to afford him that right of way, the applicant very slowly commenced his right turn and, in so doing, likely cut the mini roundabout slightly. Unfortunately, Mr Duffill did not stop and collided with the applicant. Mr Duffill fell off his bicycle and suffered a fatal head injury.

61 It was undisputed that the applicant did not possess a valid driving licence at the material time. The applicant pleaded guilty to an offence of causing death by driving a motor vehicle on a road while unlicensed under s 3ZB of the Road Traffic Act 1988. Under the then prevailing law in R v Williams (Jason John) [2011] 1 WLR 588 (“Williams”), an offence under s 3ZB could be made out even if the defendant’s manner of driving was faultless. The applicant received a suspended sentence and was disqualified from driving for a year. He served the suspended sentence and the disqualification period. Two years later, the United Kingdom Supreme Court overruled Williams and held in R v Hughes [2013] 1 WLR 2461 (“Hughes”) that in order to establish a s 3ZB offence, there had to be something open to proper criticism in the defendant’s driving and which had contributed in some more than minimal way to the death. The applicant then sought leave to appeal against his conviction retroactively.

62 The applicant highlighted that, as a result of his conviction, it was impossible for him to obtain motor insurance for under £2,700. Furthermore, his wife was entitled to a car under a certain scheme on account of their daughter’s disability. However, he was prohibited from driving that car due to his conviction. If his conviction were quashed, the driving ban would be lifted, he would be able to obtain a full driving licence, and the motor insurance costs would be met by the scheme.

63 The court noted (at [14]) that even though the applicant had already served his sentence, he “continues to suffer the remoter consequences of his conviction”. It went on to explain (at [15]) that “it is not merely the change in the law that is creating an injustice to this applicant but also the remoter consequences of it with which he continues to live”. Accordingly, the court granted the extension of time sought and quashed the applicant’s conviction.

64 Crucially, there was no doubt that following the decision in Hughes, the applicant could not be said to have committed either the s 3ZB offence of which he had been convicted or any alternative offence. The Prosecution conceded that the applicant’s conviction was unsustainable in the light of the decision in Hughes. Furthermore, the court explicitly observed (at [13]) that it would likely not have granted an extension of time if the applicant “could as easily have been charged with another offence”.

65 Drawing the above threads together, one may distil the following general principles:

(a) The mere fact that the law has changed without more will not justify the court granting leave to appeal out of time. This is so even if the change in the law corrected misconceptions about the meaning of a statute or declared the previous law to be erroneous (see Mitchell at 171C and Jogee at [100]).

(b) An applicant who relies on a change in the law in seeking leave to appeal out of time must show that he has suffered substantial injustice (see Hawkins; Ramzan at [70], [73], [76] and [78]; Ballinger at [18] and [21]; Cottrell at [46]; Jogee at [100]; and Johnson at [18] and [21]). If the application is based on nothing other than a change in the law but the applicant’s conviction was properly made at the time, the court is unlikely to find that substantial injustice has arisen (see Cottrell at [46] and McGuffog at [8]).

(c) A key consideration when evaluating if an applicant has suffered substantial injustice is whether the change in the law would have made a difference to his case (see Jogee at [100]; Johnson at [21]; and Crilly at [42]).

(d) Refusal of leave is not limited to cases where the applicant could have been charged with a different offence had the true position in law been appreciated earlier (see Jogee at [100] and Jawad at [29]). Nonetheless, the fact that an applicant could have been charged with or convicted of a different offence is a factor that strongly militates against a finding of substantial injustice (see Ramzan at [39] and [54] and McGuffog at [13]; see also Hawkins).

(e) Applicants will be expected to act promptly and without undue delay in seeking leave to appeal out of time following a change in the law (see Ballinger at [23]; see also Ang Poh Chuan at [44] and [46]). Additionally, the court may be more circumspect about making a finding of substantial injustice if the applicant pleaded guilty or made a conscious decision not to appeal (see Hawkins; see also McGuffog at [8]).

(f) The court will also consider the extent to which the law was changed. If the change in the law was far-reaching such that granting an applicant leave to appeal out of time would open the floodgates to similar applications by many other similarly situated defendants, the court may be less inclined to grant an extension of time (see Jogee at [100] and Johnson at [18]; see also Ballinger and Ramsden).

66 Although most of the relevant English authorities concern the review of convictions, the same standard of “substantial injustice” should and does apply to the review of sentences. This is readily apparent from R v Howe (Paul Alfred) [2009] EWCA Crim 2707 (“Howe”) in which two appellants sought leave to appeal against their sentences of imprisonment for public protection (“IPP”). IPP sentences (which were abolished in 2012) were designed to ensure that dangerous offenders remained in custody for as long as they presented a risk to society. An offender sentenced to IPP had to first serve his or her minimum term of imprisonment; thereafter, the offender would only be released upon satisfying the Parole Board that he or she no longer posed a risk to society. The minimum term would typically be fixed at half the length of the determinate sentence that would have been imposed for the offence in question. Both appellants had been sentenced to IPP for downloading, possessing and/or accessing obscene photographs of children. The first appellant was sentenced to a minimum specified term of two and a half years, based on a determinate term of five years. The second appellant was sentenced to a minimum term of six months, ostensibly based on a determinate sentence of a year.

67 IPP sentences would only be imposed if it was shown that the offender’s re-offending would cause serious harm. After the appellants were sentenced, it was held in a separate decision that the link between the downloading of obscene images of children and the harm that children might suffer was too remote to satisfy that requirement. The appellants subsequently sought to set aside their sentences. The court applied the test of “substantial injustice” and found (at [4], [5] and [11]) that the test was satisfied. The first appellant remained in custody and, although the second appellant had been released, he remained on a long licence liable to recall. The court therefore quashed the IPP sentences and left the original determinate sentences undisturbed.

68 The courts have also consistently applied the yardstick of “substantial injustice” to cases involving the review of confiscation orders, even though confiscation orders are not sentences in the strict sense: see, for example, Jawad at [29] and [30]; R v Meghrabi (Diana) [2014] EWCA Crim 197 at [9], [11], [12] and [15]; and R v Bestel and others [2013] EWCA Crim 1305 (“Bestel”) at [11]–[15], [23], [24], [31], [73] and [79]. There is thus no reason why reviews of convictions and sentences should not be governed by the same standard and considerations. Even in an application for review of a sentence, the starting point remains that the matter was determined on its merits and decided on the basis of the law as it then stood. Accordingly, whether a review application concerns a conviction or a sentence, something more than a mere change in the law or a mere error is required to displace the earlier decision.

69 Aside from the foregoing point of principle, there are also practical considerations that call for the high threshold of “substantial injustice” to be applied to reviews of sentences. In common with many jurisdictions, most criminal matters in Singapore are disposed of by way of guilty pleas rather than trials. Consequently, “[n]ew rules of sentencing procedure … threaten to disrupt far more cases than new rules of trial procedure” (see Scott at 181). The court’s reluctance to upset final judgments therefore cannot be any less pronounced where reviews of sentences are concerned.

70 Having surveyed the English authorities, we consider that an appellate court may only reopen an earlier decision, whether pursuant to its inherent or statutory power of review, when it has been presented with new material that gives rise to a powerful probability that substantial injustice has arisen in the criminal matter in respect of which the earlier decision was made. This proposed threshold borrows the requirement of “substantial injustice” found in English authorities such as Jogee and Johnson, and is aligned with the well-established threshold of “serious injustice” that must be met whenever our courts’ revisionary powers are invoked (see Public Prosecutor v Yang Yin [2015] 2 SLR 78 at [25]). We add that the same test of substantial injustice should be adopted when deciding whether to grant leave to appeal out of time. As we are presently concerned with “change in the law” cases, the “new material” in question will likely take the form of new legal arguments based on changes in the law brought about by a judicial decision issued after the appellate court had rendered the decision that is sought to be reviewed.

71 The test of substantial injustice comprises two essential components. Not only must an injustice have arisen, but the injustice must also be substantial to warrant the court’s exercise of its power of review.

(a) An injustice may be said to have arisen in one of two ways:

(i) Where an applicant seeks to set aside his conviction, an injustice will only have arisen if the new material strikes at the soundness of the conviction in a fundamental way.

(ii) Where an applicant seeks to challenge his sentence, an injustice will only have arisen if the new material shows that the earlier decision was based on a fundamental misapprehension of the law.

(b) An injustice may be said to be substantial under the following circumstances:

(i) An applicant may establish substantial injustice if the new material points to a powerful probability that his conviction is unsound and if the facts do not disclose any other offence of comparable gravity.

(ii) Where the new material shows that the earlier decision was based on a fundamental misapprehension of the law, any resulting injustice will only be considered substantial if the said misapprehension had a significant bearing on the sentence imposed.

72 Our approach is informed by the considerations set out in the English authorities (see [65] above) and comports with both s 394J of the CPC and our decision in Kho Jabing. The language of “powerful probability” and “a fundamental misapprehension of the law” can be found in ss 394J(6)(b) and 394J(7) respectively while the requirement of “substantial injustice” is mirrored in the substantive requirement of a “miscarriage of justice” in ss 394J(2) and 394J(5). Furthermore, since s 394J codifies the requirements set out in Kho Jabing, it follows that our approach is also consistent with that decision.

73 The alignment of s 394J of the CPC, our decision in Kho Jabing and the approach that we have set out at [71] above reinforces our point that the substance of any review application is typically unaffected by an applicant’s choice of remedial avenue (see [30] above). Hence, a mere change in the law without more will not constitute sufficient material justifying either the grant of leave (under s 394H of the CPC) to make a review application or the exercise of the court’s inherent power of review. In this connection, we echo the well-established position in English law that the court will not reopen a concluded decision for no other reason than that the law has changed, even if the change in the law declared the previous legal position to be wrong or rectified misconceptions about the meaning of a statute (see [65(a)] above). Instead, it is of paramount importance that an applicant demonstrates that he has suffered substantial injustice (see [65(b)] above).

74 We begin by elaborating on the requirement that an injustice has arisen. Where an applicant seeks to set aside his conviction, an injustice will only have arisen if the new material strikes at the soundness of the conviction in a fundamental way. If, in the light of a subsequent change in the law, it appears that the applicant has, quite simply, not committed any offence, then there would be an air of unreality in insisting that he nonetheless be held guilty. Where the new material strikes at the soundness of the conviction in a fundamental way, the change in the law would likely have made a difference to the outcome of the applicant’s case, which is the court’s overarching consideration when applying the “substantial injustice” test (see [65(c)] above).

75 Our approach mirrors that adopted in Ng Kim Han, in which several petitioners sought to have their convictions set aside by way of a criminal revision. In allowing the petition, Yong Pung How CJ observed (at [10]) that while there was no clear test of what amounted to “serious injustice”, it could not be disputed that such injustice would exist when a person had been convicted despite the obvious absence of an essential component of the offence in question. It is for the same reason that the court granted the appellant in Mitchell exceptional leave to appeal against his conviction out of time (see [36]–[39] above). The court candidly acknowledged (at 171G) that, were it to decide otherwise, it would have the “totally unreal task” of determining the correct sentence for “an offence which had not been committed” [emphasis added].

76 New material that strikes at the soundness of a conviction in a fundamental way must, in essence, give rise to a powerful probability that no offence has in fact been committed. Such new material will typically arise in the light of subsequent changes in the law as to the constituent elements of an offence and/or the proper interpretation of statutory provisions that generate criminal liability. The requirement that the new material must relate to the fundamental propriety of a conviction serves to sift out unmeritorious applications that seek to re-litigate matters based on subsequent changes in the law that are only tangentially relevant to the conviction.

77 The need for the new material to strike at the soundness of a conviction in a fundamental way also explains the decision in Ballinger. It will be recalled that the application in Ballinger arose following decisions such as Dundon, which held that naval courts martial were incompatible with the right to a fair trial protected under Art 6(1) of the ECHR. In Dundon, the court held (at [89]) that the judge advocate’s status as a naval officer meant that he was not institutionally independent of the Prosecution and accordingly quashed the defendant’s conviction. However, the applicant in Ballinger unsuccessfully relied on the decision in Dundon to challenge his conviction. In our view, what distinguishes Ballinger from cases such as Mitchell and Crilly is the fact that the applicant in Ballinger had unquestionably committed an offence – he challenged his conviction only on grounds of procedural legality. Although the conviction in Dundon was set aside on this basis, it bears reiterating that the proceedings in Ballinger, unlike those in Dundon, had already been concluded. This meant that the overriding interest in Ballinger was finality; something more than a procedural error was needed to displace that interest.

78 Matters of procedural fairness (or criminal procedure) therefore do not typically go towards the soundness of a conviction in the sense of establishing that an offence has not in fact been committed. In any event, we take the view that they are more properly classified as potential breaches of natural justice (see s 394J(5)(b) of the CPC and Kho Jabing at [69]). We also note that an applicant will likely be unable to demonstrate that such a change in the law would have yielded a different outcome, which is a critical element of the “substantial injustice” test (see [65(c)] above).

79 We next consider instances where an applicant seeks to challenge his sentence. In such cases, an injustice may only be said to have arisen if the new material shows that the earlier decision was based on a fundamental misapprehension of the law. The reason for this is relatively straightforward: the courts commonly lay down new sentencing guidelines and sentencing frameworks, and it would be alarming if every new sentencing framework could unravel all previous sentencing decisions for the offence in question (see [65(f)] above). As an example, the sentencing framework for rape offences laid down in Ng Kean Meng Terence v Public Prosecutor [2017] 2 SLR 449 (“Terence Ng”) would not suffice as grounds on which an applicant may seek to have his sentence reviewed. Individuals who were convicted for rape offences under the sentencing framework set out in Public Prosecutor v NF [2006] 4 SLR(R) 849 (“NF”), which was the applicable sentencing framework for rape offences before Terence Ng was decided, cannot be said to have been convicted based on a fundamental misapprehension of the law. The impetus for the Terence Ng framework was principally the desire to devise a comprehensive and conceptually coherent sentencing framework for rape offences (see Terence Ng at [12(a)] and [12(b)]). Even though the NF framework was eventually replaced, it cannot be said to have been founded on a misapprehension of the law.

80 An example of a fundamental misapprehension of the law might be the converse of what transpired in Abdul Nasir bin Amer Hamsah v Public Prosecutor [1997] 2 SLR(R) 842 (“Abdul Nasir”). Until that decision was rendered, life imprisonment was understood to mean 20 years’ imprisonment. The basis for this understanding was s 57 of the Penal Code (Cap 224, 1985 Rev Ed) (“the 1985 PC”), which provided that “[in] calculating fractions of terms of punishment, imprisonment for life shall be reckoned as equivalent to imprisonment for 20 years”. In Abdul Nasir, however, we highlighted (at [30]) that on a proper construction of s 57, life imprisonment ought to be equated with 20 years’ imprisonment only when calculating fractions of terms of punishment. Having considered the legislative intent underlying offences that were punishable with life imprisonment, the plain meaning of the term “life” as well as s 45 of the 1985 PC, we concluded (at [32]) that life imprisonment ought to be understood as a sentence of imprisonment for the whole of the remaining period of the convicted offender’s natural life. Of course, the change in the law occasioned by Abdul Nasir did not give rise to any substantial injustice because the previous erroneous understanding had worked to the advantage of offenders. But had an error of such a nature had the opposite effect such that offenders had previously been sentenced to more than what they should have been, this could arguably be said to be a fundamental misapprehension of the law.

81 It is, however, insufficient that an injustice has arisen in the criminal matter in respect of which the earlier decision was made. That injustice must be substantial. Such a rigorous standard accords due recognition to the fact that the impugned decision has already undergone at least two rounds of independent scrutiny – once by the court exercising original criminal jurisdiction and another by the appellate court. The requirement that the injustice be “substantial” also better distinguishes the function of an appeal, which is the correction of error, from that of a review, which is the correction of miscarriages of justice (see Kho Jabing at [49] and [65]).

82 Where an applicant successfully establishes a powerful probability that his conviction is unsound and no comparable offence has been committed, he will have established substantial injustice. After all, “[t]here is no doubt that the conviction of an innocent person is an injustice – perhaps the ultimate substantive injustice, proof of which would justify reopening a concluded criminal appeal” [emphasis in original] (see Kho Jabing at [64]). Conversely, this means that even where an applicant demonstrates that his conviction is unsustainable, he will nevertheless have failed to establish substantial injustice if he could have been charged with or convicted of other offences of comparable gravity despite the change in the law (see [65(d)] above and Ramzan at [39] and [54]).

83 Where the impugned decision is a decision on sentence, any injustice (arising from a fundamental misapprehension of the law) will only be considered substantial if the said misapprehension had a significant bearing on the sentence imposed (see, for example, Howe at [66]–[67] above). Sentencing involves an exercise of discretion and potential adjustments to a sentence that are of marginal significance cannot justify interference with the sentencing court’s exercise of discretion (see Public Prosecutor v Siew Boon Loong [2005] 1 SLR(R) 611 at [22]). Moreover, given that an appellate court will only entertain an appeal against sentence where (among other things) the sentence is manifestly excessive or manifestly inadequate (see Public Prosecutor v Mohammed Liton Mohammed Syeed Mallik [2008] 1 SLR(R) 601 at [82]), the threshold for intervention by an appellate court in a review application cannot be any less stringent.

84 We conclude this section by making three observations. First, there can be no question that even in cases where it is subsequently found that substantial injustice has arisen following a change in the law, those cases were correctly decided in accordance with the prevailing law as it was understood at the time. Under the approach that we have set out above, however, the legal correctness of a decision at the time that it was made is tempered by the overall justice of the case. In other words, while a matter might have been properly decided in line with the then prevailing jurisprudence, the “substantial injustice” test effectively engages considerations of prejudice to the applicant or the unconscionability of tying him to the old legal position.

85 In this vein, we stress that applicants should act with all due urgency should they be of the view that a change in the law affords them strong grounds for seeking a review of their convictions and/or sentences (see [65(e)] above). The longer the delay between the change in the law and the filing of the review application, the more it can presumptively be said that the applicant has deliberately chosen not to challenge the earlier decision against him, and the more difficult it will likely be for him to prove that he has suffered substantial injustice (see Ballinger at [22]–[23]; see also [50] above). Hence, where an applicant has, by his wilful and persistent inaction, assumed the risk of any prejudice that he may suffer, finality interests should assume priority. Of course, a court confronted with this issue will have to consider the length of time that has passed from when the applicant ought reasonably to have known of the opportunity to seek a review and also the reasons advanced to explain any delay.

86 Second, our approach is consonant with the cases in which an appellate court revisited an earlier decision following a change in the law. In Gobi, for example, we exercised our statutory power of review under s 394I for the first time as there was a powerful probability that the applicant had been wrongfully convicted on a capital charge when the necessary elements of the offence had not been made out. In Public Prosecutor v Ong Say Kiat [2017] 5 SLR 946 (“Ong Say Kiat”), the court granted the respondent leave to appeal out of time on account of the substantial injustice that he had suffered by being imprisoned for a significantly longer duration than he otherwise would have been. The court intimated that the disproportionately heavy sentence had been imposed on the respondent pursuant to a fundamental misapprehension of the law. We consider both of these cases in greater detail at [92]–[95] and [97]–[100] below.

87 Third, and most importantly, the criteria that must be fulfilled before an appellate court may review an earlier decision following a change in the law are exacting and are intended to forestall endless re-litigation (see [65(f)] above). The stringent requirements that we have set out above cumulatively serve as a filter to ensure that only meritorious applications in which there is a powerful probability that substantial injustice has arisen will be entertained. Where such applications are concerned, “[t]he floodgates argument should not be allowed to wash away both the guilty and the innocent” (see Yong Vui Kong at [15]). The possibility of error does not however mean that we should disregard the interests in certainty and repose. Indeed, those interests animate the strict requirements that must be met in order to establish substantial injustice.

88 In this connection, we note that it will be exceedingly rare for applicants to successfully establish substantial injustice based on the change in the law occasioned by Saravanan. This is for several reasons. First, in capital cases or cases in which the Prosecution has reduced the weight of the drugs stipulated in the charge to an amount falling just short of the capital threshold, the Prosecution’s “dual charging practice” will make little to no difference to the eventual sentence imposed. We reiterate that in so far as reviews of sentences are concerned, substantial injustice will only be established if it is shown that, among other things, the change in the law would have had a significant bearing on the sentence imposed (see [83] above).

89 Second, in cases where accused persons elect to plead guilty, the Prosecution will usually proceed on the cannabis charge and apply for the cannabis mixture charge (in respect of the same block(s) of cannabis-related plant material) to be taken into consideration (“TIC”) for the purpose of sentencing. Where the proceeded charges and the TIC charges pertain to the same block(s) of cannabis-related plant material, the courts tend not to accord the TIC charges significant aggravating weight. Abdul Karim is a case in point. The appellant’s appeal against sentence was dismissed because the High Court judge had not regarded the TIC charge (which concerned cannabis mixture) as aggravating, among other reasons.

90 Third, any application for us to reopen a concluded decision based on the change in the law brought about by Saravanan essentially invites us to retrospectively alter one part of the factual matrix – in other words, to assume that the position in Saravanan was already the applicable law when the case at hand was decided. The difficulty, however, is that any retrospective view of events must also take into account the full range of factors, including how the Prosecution might have acted had it appreciated the legal position in Saravanan at the material time. It seems to us that in the vast majority of cases pre-dating Saravanan where offenders had been charged pursuant to the Prosecution’s “dual charging practice” and so convicted, the Prosecution could have easily proceeded on charges other than the impugned cannabis mixture charges, such that there would have been no appreciable difference in the aggregate sentence imposed. This is a point of considerable importance, which we elaborate on at [116]–[119] and [133] below.

91 Having addressed when a court may revisit an earlier decision following a change in the law, the question that follows is how a court may do so. In our view, there are three mechanisms at the court’s disposal. First, where a decision has not been appealed against previously, an appellate court may grant an offender leave to appeal out of time. Second, where an appeal against a lower court’s decision has already been disposed of, the appellate court may either exercise its inherent power to review the concluded appeal or grant the offender leave (under s 394H of the CPC) to make a review application. Third, the Court of Appeal may exercise its power to review any matter on its own motion if so warranted. We proceed to discuss each of these avenues in turn.

92 Where a decision has not previously been the subject of an appeal, an appellate court may grant an offender leave to appeal out of time. This course of action was adopted in Ong Say Kiat. The respondent in that case (“Ong”) had been sentenced in 2014 to five years’ corrective training (“CT”) for an offence of theft in dwelling with common intention. The decision in Sim Yeow Kee v Public Prosecutor and another appeal [2016] 5 SLR 936 (“Sim Yeow Kee”) was handed down some 21 months later. The Prosecution then filed a criminal revision asking that the sentence of five years’ CT imposed on Ong be set aside. The Prosecution contended that in the light of Sim Yeow Kee, the sentence of five years’ CT appeared to be unduly disproportionate compared to the likely term of regular imprisonment that would otherwise have been imposed, even though the imposition of CT was justified given the prevailing jurisprudence at the time of Ong’s sentencing.

93 The court held (at [32] and [37]) that the case properly called for the exercise of its appellate rather than revisionary jurisdiction. It further held (at [34]) that the factual and legal substratum that underlay the Sim Yeow Kee sentencing approach to CT had already existed at the time of Ong’s sentencing. The sentencing approach in Sim Yeow Kee was devised because of several changes in the operating conditions which affected the sentence of CT and which had already taken effect at the time of Ong’s sentencing. Accordingly, the court held (at [34]) that the sentence of five years’ CT imposed on Ong was “wrong in law” and manifestly excessive, notwithstanding the prevailing jurisprudence pre-Sim Yeow Kee.

94 What is of note is the court’s finding (at [33]) that the law laid down in Sim Yeow Kee ought to have applied at the time of Ong’s sentencing because “the relevant factual and legal matrix” [emphasis added] was the same in both cases. Reading between the lines, it appears that the court was intimating that Ong’s sentence was based on a fundamental misapprehension of the law (see [71(a)(ii)] and [79] above). Moreover, the injustice that Ong had suffered was doubtlessly substantial. By the time the criminal revision was heard, and without taking into account his time spent in remand, Ong had spent over two and a half years in prison (see Ong Say Kiat at [38]). Even after considering his antecedents, the likely imprisonment term for his offence would have been only nine months at the very most (see Ong Say Kiat at [13]). Given the substantial injustice that had arisen, the court granted Ong leave to appeal out of time, treated the appeal as having been heard, allowed the appeal, and sentenced him to a term of imprisonment of time already served (see Ong Say Kiat at [44] and [50]).

95 Notwithstanding the fact that approximately 21 months had elapsed between Ong’s sentencing and the decision in Sim Yeow Kee, the court’s decision to grant Ong leave to appeal out of time is unimpeachable. As the English Court of Appeal expressed in Johnson (at [21]), “[i]t is not … material to consider the length of time that has elapsed. If there was a substantial injustice, it is irrelevant whether that injustice occurred a short time or a long time ago. It is and remains an injustice.” [emphasis added].

96 It goes without saying that when a case has already been heard and disposed of by an appellate court, the dissatisfied litigant is precluded from seeking a second appeal. Instead, such a litigant may have recourse to either the appellate court’s statutory power of review under s 394I of the CPC or this court’s inherent power of review.

97 Gobi was the first (and thus far, only) instance of a review application that was successfully brought under s 394I of the CPC. The applicant in Gobi claimed trial to a capital charge of importing not less than 40.22g of diamorphine. The trial judge held that the applicant had rebutted the presumption of knowledge under s 18(2) of the MDA and acquitted him of the capital charge. The trial judge instead convicted the applicant of a non-capital charge of attempting to import a Class C controlled drug and sentenced him to 15 years’ imprisonment and ten strokes of the cane. The Prosecution’s appeal against the trial judge’s decision was allowed in CA/CCA 20/2017 (“CCA 20/2017”). We held that the trial judge erred in finding that the s 18(2) presumption had been rebutted and convicted the applicant of the capital charge.

98 The applicant sought a review of our decision in CCA 20/2017, pursuant to s 394I of the CPC. The application was brought on the basis of the following new legal arguments (“the Legal Arguments”): (a) applying our holdings in Adili Chibuike Ejike v Public Prosecutor [2019] 2 SLR 254 (a decision issued after CCA 20/2017 had been decided) to the presumption of knowledge under s 18(2) of the MDA, the s 18(2) presumption should not encompass wilful blindness; and (b) the Prosecution could not have invoked the s 18(2) presumption because its case at the trial was not one of actual knowledge but, at most, one of wilful blindness.

99 We allowed the application and exercised our statutory power of review as the Legal Arguments provided sufficient material on which it might be concluded that there had been a miscarriage of justice in CCA 20/2017 (see Gobi at [44], [49] and [50]). We held (at [56]) that the knowledge presumed under s 18(2) of the MDA was confined to actual knowledge of the nature of the drugs in an accused person’s possession and did not encompass knowledge of matters to which an accused person had been wilfully blind. As the Prosecution’s case at the trial was that the applicant had been wilfully blind to the nature of the drugs in question, it was not entitled to rely on the s 18(2) presumption (see Gobi at [105], [109] and [121]). Furthermore, the applicant had not been wilfully blind to the nature of the drugs in the first place (see Gobi at [123]–[125]). Accordingly, we set aside his conviction on the capital charge and reinstated the conviction on the reduced charge (see Gobi at [133]).

100 It should be noted that we exercised our statutory power of review in Gobi even though the s 18(2) presumption did encompass the doctrine of wilful blindness under the law pre-Gobi (see, for example, Masoud Rahimi bin Mehrzad v Public Prosecutor [2017] 1 SLR 257 and Tan Kiam Peng v Public Prosecutor [2008] 1 SLR(R) 1). In other words, the applicant had been properly convicted under the law as it then stood. However, Gobi is readily reconcilable with the approach that we have laid out at [70]–[71] above. In Gobi, we held (at [56]) that only actual knowledge of the nature of the drugs in an accused person’s possession could be presumed under s 18(2) of the MDA. However, the Prosecution’s case at the trial was one of wilful blindness even though it had invoked the s 18(2) presumption. On appeal, the Prosecution instead ran its case on the basis of actual knowledge, and the Defence did not pick up on this change of position. The court’s decision on conviction in CCA 20/2017 was premised on the holding that the applicant had failed to rebut the presumption of knowledge (see Gobi at [121]). But, by reason of the Prosecution’s change of position on appeal, the fact and significance of which were not appreciated at the time, the Prosecution had in fact failed to establish the necessary mens rea for the offence of drug importation at the trial. In the circumstances, there was a powerful probability that an injustice (or, to use the language of ss 394J(2) and 394J(5) of the CPC, a miscarriage of justice) had arisen in the sense that the soundness of the applicant’s conviction on the capital charge was doubtful. This injustice was unquestionably “substantial” since it was a matter of life and death.

101 Aside from s 394I of the CPC, an applicant may also seek to have an earlier decision of this court reopened pursuant to our inherent power of review. Although this power was expressly set out in Kho Jabing, we ultimately declined to exercise our inherent power of review in that case. As explained at [24] above, the arguments raised by the applicant in Kho Jabing were neither new nor compelling and therefore did not warrant the exercise of our inherent power of review.

102 Thus far, we have only exercised our inherent power of review to reopen and reverse a concluded decision in the case of Ilechukwu Uchechukwu Chukwudi v Public Prosecutor [2021] 1 SLR 67 (“Ilechukwu”). However, as we exercised that power on the basis of new psychiatric evidence rather than new legal arguments following a change in the law, Ilechukwu is irrelevant for present purposes.

103 Finally, the Court of Appeal has the power to review any matter on its own motion if so warranted. This power is a facet of the judicial power vested in the Court of Appeal by virtue of Art 93 of the Constitution. Although s 60D of the Supreme Court of Judicature Act (Cap 322, 2007 Rev Ed) (“the SCJA”) sets out the matters that the criminal jurisdiction of the Court of Appeal consists of, the SCJA does not expressly state when the jurisdiction of the Court of Appeal in a criminal appeal ends. To adopt the reasoning of Chan CJ in Yong Vui Kong at [16] (with which we agreed in Kho Jabing at [77(a)]), in the absence of any statutory provision providing for the termination of its jurisdiction in a criminal appeal, there is no reason why the Court of Appeal should circumscribe its own judicial power to preclude itself from reviewing any matter on its own motion.

104 That this power should not be lightly exercised is perhaps most evident from the fact that it has yet to be exercised by this court. The role of the court is to adjudicate disputes that are before it; a court that is too eager to review matters on its own motion not only risks encroaching upon the territory of the Public Prosecutor, who has sole responsibility for the conduct of prosecutions, but also risks jeopardising its perceived legitimacy. We therefore leave it to a suitable occasion in the future to discuss the precise contours of the Court of Appeal’s power to review any matter on its own motion.

105 Having considered the circumstances in which an applicant may seek to reopen a concluded decision, as well as the avenues by which he may do so, we now deal squarely with the applications before us. We begin by providing an overview of the parties’ respective positions.

106 As mentioned, the Prosecution filed the present applications, all of which pertain to the soundness of the respondents’ convictions and/or sentences in relation to their respective cannabis mixture charges. The table below sets out the order(s) sought in each application:

107 In the PP’s Letter, the Prosecution sought leave to withdraw the four applications for the following reasons:

(a) Saravanan does not automatically apply to decisions concluded as of 29 April 2020, when that case was decided. Instead, Saravanan presumptively applies only to cases that have come before the courts on or after 29 April 2020.

(b) The respondents’ cases were properly decided in accordance with the prevailing law then. As held in Johnson, where a change in the law has occurred, there is no automatic reopening of concluded cases that were properly decided under the law as it then stood.

(c) The Prosecution and law enforcement authorities had relied on the established legal position prior to Saravanan in engaging in the “dual charging practice”. Offenders who were convicted pursuant to the “dual charging practice” pre-Saravanan had made a conscious decision to deal in compressed blocks of cannabis-related plant material, despite being aware of the permissibility of the “dual charging practice” pre-Saravanan. The interests of justice, particularly the need to uphold legal certainty and to protect the finality of decisions under established law, militate against the reopening of the concluded cases in question.

108 The respondent in CM 11, Pang Chie Wei (“Pang”), unreservedly consented to the Prosecution’s withdrawal of the application. Prior to the hearing, the respondent in CM 12, Shanmugam a/l Applanaidu (“Shanmugam”), objected to the Prosecution’s withdrawal of CM 12. His counsel subsequently informed us at the hearing that he had since discussed the matter with Shanmugam and that Shanmugam, too, gave his unreserved consent to the withdrawal of CM 12.

109 Shalni, the respondent in CM 14, initially consented to the withdrawal of the application. However, her counsel, Mr Amarick Gill (“Mr Gill”), indicated at the hearing that she had changed her position. Mr Gill submitted that if Shalni’s cannabis mixture charge were set aside, that could potentially reduce her aggregate sentence by five years (see [132] below). On that basis, he urged us to deny the Prosecution’s withdrawal of CM 14. As we stated at the hearing, however, there was simply no juridical basis for us to stop the Prosecution from withdrawing the application (see [5] above). Mr Gill thereafter indicated that he reserved the right to file an application on Shalni’s behalf to seek substantially the same relief as that originally sought by the Prosecution in CM 14. We observed that it was open to him to do so but that the merits and likely success of any such application fell to be considered against the high threshold of substantial injustice which we outlined then and which we have explored in detail in these grounds of decision.

110 That left us with Suventher, the respondent in CM 13, who objected to the withdrawal of the application. Although he accepted that Saravanan did not “automatically” apply with retroactive effect, he contended that there had been a “miscarriage of justice”. He argued that the Prosecution had erred in preferring separate charges involving cannabis and cannabis mixture against him, that the High Court judge had erred in treating the TIC charge (which pertained to cannabis mixture) as an aggravating factor in sentencing, and that we had erred in affirming the sentence imposed by the High Court judge.

111 Having set out the parties’ respective positions, we now examine the implications of Saravanan on each of the four cases out of which the present applications arise. Before doing so, we note that Pang and Shanmugam unreservedly consented to the withdrawal of CM 11 and CM 12 respectively. We are nonetheless cognisant that this is the first time that we have expressly considered the threshold for invoking the court’s revisionary powers based on a subsequent change in the law. Hence, and as we informed the respondents at the hearing, if it appears to any of them, after having perused these grounds of decision, that there is a basis for invoking the court’s powers of review or revision, it remains open to them to do so. That said, we reiterate once again that the applicable threshold of substantial injustice is a very high bar that will not commonly be crossed (see [6] above).

112 Pang, the respondent in CM 11, pleaded guilty to three charges under the MDA. Three other MDA charges were taken into consideration for the purpose of sentencing.

113 All six charges related to the same act of importation in furtherance of the common intention of Pang and his co-accused. The High Court judge found Pang guilty of the three proceeded charges and convicted him accordingly. The table below summarises the drugs that formed the subject matter of each charge as well as the sentence imposed for each proceeded charge:

114 The sentences for the first and second charges were ordered to run consecutively while the sentence for the third charge was ordered to run concurrently. The aggregate imprisonment term imposed was thus 25 years. The aggregate sentence of caning was limited to 24 strokes, pursuant to ss 328(1) and 328(6) of the CPC.

115 Before subsequently seeking leave to withdraw CM 11, the Prosecution sought an extension of time to appeal against Pang’s sentence on account of his conviction on the third charge. The first and third charges (which are highlighted in the table at [113] above) related to the same four blocks of cannabis-related plant material. As mentioned at [108] above, Pang did not object to the withdrawal of CM 11.

116 In our judgment, the facts did not show that there was a powerful probability that substantial injustice would arise if the Prosecution was granted leave to withdraw CM 11. It was true that Pang could not be said to have committed some other offence in respect of the cannabis mixture charge (see Hawkins; Ramzan at [39] and [54]; and McGuffog at [13]). However, one could not blindly ignore the fact that if the cannabis mixture charge was disregarded, the Prosecution could have proceeded with a different combination of three charges – or indeed, on all five remaining charges – such as would yield at least the same aggregate sentence of 25 years’ imprisonment and 24 strokes of the cane. In our view, this was a strong factor that militated against a finding of substantial injustice.

117 We highlight that the courts have considered how the Prosecution might have otherwise proceeded when assessing whether substantial injustice has been established in “change in the law” cases. In Cottrell, for example, the first defendant was convicted of two counts of indecent assault for having had sexual intercourse with a 15-year-old girl on two occasions. Subsequently, the House of Lords held that where the Prosecution only relied upon evidence of sexual intercourse with a girl under the age of 16, it could not charge an accused person with indecent assault if 12 months or longer had elapsed since the alleged sexual intercourse. The first defendant then sought leave to appeal against his conviction out of time on the basis that he had been charged with indecent assault more than 12 months after the acts of sexual intercourse had occurred.

118 The English Court of Appeal refused to grant leave. The court emphatically stated (at [59]) that “[i]t would be a manifest injustice to the complainant if [the first defendant] were able to take advantage of that part of the change of law which suited him, without having to accept the inevitable consequences of the process which would have applied to this case if the erroneous practice had been recognised earlier, and the necessary adaptations to it adopted” [emphasis added]. In addition to the acts of sexual intercourse, the complainant had also alleged that the first defendant had touched her breasts. Although the first defendant was convicted on the two counts of indecent assault that were based on the acts of sexual intercourse, the jury could not agree on the separate count of indecent assault arising from the allegation that he had touched the complainant’s breasts. In the light of the first defendant’s convictions, the Prosecution did not pursue the count of indecent assault on which the jury was unable to agree (see Cottrell at [12]). Clearly, if the Prosecution had appreciated the subsequent change in the law earlier, it could and would have pursued the remaining count of indecent assault. As put by the court in Bestel (at [24]), “[i]n Cottrell the [first defendant] would have been convicted of sexual assault founded purely upon acts committed before the act of unlawful sexual intercourse on which the charge had been laid”.

119 It was hence an eminently relevant consideration in our context that the Prosecution could and likely would have proceeded on at least three out of the five remaining charges against Pang, such as would yield either the same or a similar aggregate sentence. The fourth and fifth charges carried a mandatory minimum sentence of five years’ imprisonment while the sixth charge attracted a mandatory minimum sentence of three years’ imprisonment. In addition, the fourth, fifth and sixth charges all carried a mandatory minimum of five strokes of the cane. This meant that any combination of three proceeded charges would have resulted in an aggregate imprisonment term of 23 or 25 years as well as 24 strokes of the cane. The Prosecution could also have opted to proceed on more than three proceeded charges, in which case Pang might have received an even harsher aggregate sentence. Another possibility was that the Prosecution might have proceeded on a capital charge against Pang. In this regard, we noted that the first charge was initially a capital charge of importing not less than 915.6g of cannabis. Much like the first defendant in Cottrell, we saw no reason why Pang ought to be allowed to take advantage of the decision in Saravanan without having to take into account how the Prosecution might have acted under a different set of circumstances.

120 For these reasons, we concluded that no substantial injustice had arisen in Pang’s matter.

121 The respondent in CM 12, Shanmugam, claimed trial to two charges of drug importation. The subject matter of both charges was as follows:

(a) first charge: six blocks containing not less than 1,969.3g of cannabis; and

(b) second charge: six blocks containing not less than 3,584.2g of cannabis mixture.

122 Both charges related to the same six blocks of cannabis-related plant material. The High Court judge found Shanmugam guilty of both charges and convicted him accordingly. As the judge found that Shanmugam was a courier and as the Prosecution extended a Certificate of Substantive Assistance, the judge exercised his discretion under s 33B(1)(a) of the MDA and imposed a sentence of life imprisonment and the mandatory 15 strokes of the cane per charge. The aggregate sentence imposed was thus life imprisonment and 24 strokes of the cane. Shanmugam appealed against his sentence but subsequently withdrew his appeal.

123 By way of CM 12, the Prosecution originally sought an extension of time to appeal against the sentence imposed on Shanmugam. Before us, Shanmugam consented to the withdrawal of CM 12 (see [108] above). It was clear beyond peradventure to us that any appeal against Shanmugam’s sentence would be hopeless. As Shanmugam was convicted of importing a capital amount of not less than 1,969.3g of cannabis, life imprisonment was the lowest possible sentence that could have been imposed under s 33B(1)(a) of the MDA. Disregarding the cannabis mixture charge would have made no difference to the sentence of life imprisonment that was imposed in relation to the cannabis charge. The 15 strokes of caning imposed on Shanmugam would also have been administered by now, given that he withdrew his appeal against sentence in November 2017 (see ss 327(1)(a) and 327(2) of the CPC). Accordingly, setting aside the cannabis mixture charge would not have made any difference at all in Shanmugam’s case, and it followed that no substantial injustice had arisen.

124 Suventher, the respondent in CM 13, pleaded guilty to a single charge of importing two blocks containing not less than 499.9g of cannabis. Another charge was taken into consideration for the purpose of sentencing. The TIC charge concerned the same two blocks of cannabis-related plant material, save that they were alleged to contain not less than 999.9g of cannabis mixture.

125 The High Court judge found Suventher guilty of the proceeded charge and convicted him accordingly. Although the judge acknowledged that the proceeded charge and the TIC charge pertained to the same blocks of cannabis-related plant material, he nonetheless found that the substantial weight of the cannabis mixture in the TIC charge was an aggravating factor that warranted an enhancement in sentence (see Public Prosecutor v Suventher Shanmugam [2016] SGHC 178 at [25]–[26]). The judge ultimately sentenced Suventher to 23 years’ imprisonment and the mandatory 15 strokes of the cane.

126 In Suventher Shanmugam v Public Prosecutor [2017] 2 SLR 115 (“Suventher (CA)”), we dismissed Suventher’s appeal against sentence. We also laid down sentencing guidelines for the unauthorised import or trafficking of cannabis and held (at [21] and [29]) that the sentence imposed in such cases ought to be proportional to the quantity of drugs in the offender’s possession.

127 In CM 13, the Prosecution initially sought to have Suventher’s TIC charge (which pertained to cannabis mixture) set aside and urged us to exercise our inherent power to review our decision in Suventher (CA). At the hearing, Suventher initially protested against the Prosecution’s withdrawal of CM 13. He contended that the TIC charge ought to be set aside and argued that this would in turn reduce his sentence for the proceeded charge.

128 We found Suventher’s argument to be wholly devoid of merit. It will be recalled that he pleaded guilty to a charge of importing not less than 499.9g of cannabis. Having regard to the sentencing guidelines in Suventher (CA), the appropriate indicative starting sentence would have been 29 years’ imprisonment, given that the quantity of cannabis involved was at the very highest end of the weight range. There was nothing remarkable about his culpability to warrant an adjustment to the indicative starting sentence; nor were there any other aggravating factors or any valid mitigating factors. In other words, and as we made clear to Suventher at the hearing of these applications, he could have been sentenced to 29 years’ imprisonment and 15 strokes of the cane based on the sheer weight of the cannabis stated in the proceeded charge alone, even if the TIC charge had not been accorded any aggravating weight. Indeed, we expressly noted in Suventher (CA) at [41] that “having regard to the guidelines … as the charge was for a quantity of drugs that was at the very top of the weight range, the sentence could in fact have been much more severe”. After we explained this to Suventher at the hearing, he no longer maintained his objections to the Prosecution’s withdrawal of CM 13.

129 In summary, given that the sentence imposed on Suventher was very much on the lenient side, it was patently untenable for him to argue that he had suffered any injustice, let alone substantial injustice, by virtue of the TIC charge.

130 Shalni, the respondent in CM 14, pleaded guilty to three charges of drug importation relating to the same transaction. The High Court judge found Shalni guilty of all charges and convicted her accordingly. The table below summarises the drugs that formed the subject matter of each charge as well as the sentence imposed per charge:

131 The sentences for the first and third charges were ordered to run consecutively while the sentence for the second charge (which was the cannabis mixture charge) was ordered to run concurrently. The aggregate sentence imposed was therefore 30 years’ imprisonment. Shalni’s appeal against sentence was dismissed.

132 The second and third charges (which are highlighted in the table at [130] above) related to the same block of cannabis-related plant material. CM 14 concerned the integrity of Shalni’s conviction on the second charge. The Prosecution originally applied to set aside Shalni’s conviction and sentence in respect of the second charge and invited us to exercise our inherent power to review our earlier decision to dismiss Shalni’s appeal against sentence. As earlier mentioned at [109] above, Mr Gill submitted that we ought to reconsider our dismissal of Shalni’s appeal against sentence. According to him, if the Prosecution had erred in preferring the cannabis mixture charge, then there would have been only two proceeded charges. Pursuant to s 306(2) of the CPC, the sentences for those two charges (namely, the first and third charges) might but need not have run consecutively. Mr Gill argued that had those two sentences been ordered to run concurrently, Shalni would have been sentenced to an aggregate imprisonment term of 25 rather than 30 years.

133 As we made clear to Mr Gill, the difficulty in trying to undo Shalni’s conviction on the cannabis mixture charge, which was properly rendered at the time, was that we would inevitably have had to consider how the Prosecution might have proceeded had it appreciated the legal position in Saravanan then. Were we to consider only the first part of that equation and ignore the realities of how the Prosecution might have otherwise proceeded, we would in effect be selectively altering only one part of the factual matrix on hindsight. The first charge against Shalni was originally framed as a capital charge involving more than 15g of diamorphine. Had the Prosecution been earlier apprised of the fact that it could not have proceeded with the cannabis mixture charge, it might well have exercised its prosecutorial discretion differently in deciding whether to reduce the capital charge. With respect, it seemed to us that the perspective that Mr Gill put forward on his client’s behalf did not take into account the full range of factors relevant to the reconstruction of past events. Therefore, for substantially the same reasons canvassed at [90] and [116]–[119] above, we were of the view that there was no powerful probability that substantial injustice had arisen in Shalni’s matter.

134 Finally, as the Prosecution quite fairly pointed out in the PP’s Letter, they had engaged in the “dual charging practice” pursuant to our decision in Public Prosecutor v Manogaran s/o R Ramu [1996] 3 SLR(R) 390 (“Manogaran”), in which we held (at [43]) that the Prosecution could prefer dual charges of cannabis and cannabis mixture arising from the same compressed block of cannabis-related plant material. Manogaran was good law for nearly 24 years, until we departed from that established legal position in Saravanan. If the principle of finality is to have any real significance, then the Prosecution, law enforcement authorities and convicted offenders alike must be entitled to the belief that cases that were concluded pre-Saravanan have been set at rest for good.

135 For these reasons, we were satisfied that Shalni had not suffered substantial injustice on account of her conviction and sentence for the cannabis mixture charge.

136 The finality embodied in every judgment of the court is what instils in us the confidence that justice has been done and may be treated as having been done. Accordingly, the court’s revisionary powers should not be exercised in a manner that “paves the way for a flood of re-litigation every time the criminal law gets changed” (see Ng Kim Han at [23]). Changes in the law after the conclusion of a criminal matter, and especially after the conclusion of a criminal appeal, do not warrant overriding the finality of adjudication.

137 Nevertheless, we recognise that there will be exceptional circumstances where a subsequent change in the law extends to an offender a colourable claim of innocence or calls for us to revisit the soundness of an offender’s sentence. We consider that the threshold of substantial injustice, while high, nonetheless allows for the vindication of the rights of the minority of litigants who have been grievously wronged, and that this threshold is where the delicate balance between truth and finality properly lies.