High Court Judgment on Criminal Law Appeal

IN THE GENERAL DIVISION OF THE HIGH COURT OF THE REPUBLIC OF SINGAPORE

[2026] SGHC 62

Magistrate’s Appeal No 9196/2024/01 Between

… Appellant And

… Respondent

Magistrate’s Appeal No 9197/2024/01 Between

… Appellant And

… Respondent

JUDGMENT

[Courts and Jurisdiction — Court judgments] [Criminal Law — Appeal] i

Public Prosecutor Pek Lian Guan Public Prosecutor Pay Teow Heng

TABLE OF CONTENTS

INTRODUCTION............................................................................................ 1

BACKGROUND FACTS ................................................................................ 6

THE PARTIES’ ARGUMENTS BELOW .................................................... 8

DECISION BELOW ...................................................................................... 10

THE SCOPE OF THIS JUDGMENT .......................................................... 13

THE PARTIES’ ARGUMENTS .................................................................. 14

THE PROSECUTION’S ARGUMENTS ................................................................ 14

PEK’S ARGUMENTS ....................................................................................... 15

PAY’S ARGUMENTS ....................................................................................... 17

ISSUES TO BE DETERMINED .................................................................. 18

THE APPLICABLE LEGAL FRAMEWORK........................................... 19

THE IMPORTANCE OF PROCEDURAL JUSTICE.................................................. 22

SITUATING A COMPLAINT WITHIN THE NATURAL JUSTICE FRAMEWORK ........ 28

Judicial copying ....................................................................................... 30

Inadequacy of reasons ............................................................................. 39

Excessive judicial interference................................................................. 45

Conclusion ............................................................................................... 52

THE APPLICABLE THRESHOLD FOR ASSESSING AN ALLEGED BREACH OF

......................................................................................... 53 NATURAL JUSTICE

THE APPROPRIATE REMEDIAL CONSEQUENCES FOLLOWING A FINDING

.................................................................. 64 OF BREACH OF NATURAL JUSTICE

The prejudice threshold in the arbitration context .................................. 67

The guiding principle of public confidence in the judicial context .......... 72

THE APPROPRIATE ANALYTICAL FRAMEWORK .............................................. 76

HAS THERE BEEN A BREACH OF NATURAL JUSTICE IN

THIS CASE? .................................................................................................. 77

THE NATURE OF THE PROSECUTION’S COMPLAINT ........................................ 78

WHETHER THE DJ FAILED TO APPLY HIS MIND TO THE MATTERS .................. 83

Overarching issues with the GD .............................................................. 83

Issue 1: Reliability of the respondents’ CPIB statements ........................ 94

Issues 2 and 3: The respondents’ mens rea ........................................... 123

THE REMEDIAL CONSEQUENCES .................................................................. 128

CONCLUSION ............................................................................................ 129

ANNEX 1: NON-EXHAUSTIVE LIST OF SIMILARITIES BETWEEN THE GD AND THE RESPONDENTS’ TRIAL

SUBMISSIONS ............................................................................................ 131

This judgment is subject to final editorial corrections approved by the court and/or redaction pursuant to the publisher’s duty in compliance Reports.

Public Prosecutor v Pek Lian Guan and another appeal

[2026] SGHC 62

General Division of the High Court — Magistrate’s Appeals No 9196 of 2024 and No 9197 of 2024 Sundaresh Menon CJ 22 July 2025 23 March 2026 Judgment reserved.

Sundaresh Menon CJ: Introduction

1 The Judiciary is entrusted by the public to state the law and to resolve disputes on this basis. The Judiciary’s effectiveness in discharging its constitutional function rests on its legitimacy in the eyes of the society it serves. This is so because the willingness of the public to abide by its decisions stands in large part on whether it reposes trust and confidence in the Judiciary.

2 This is what underlies the cardinal principle that justice must not only be done, but must also be seen to be done. It is also why, for example, s 8(1) of the Supreme Court of Judicature Act 1969 (2020 Rev Ed) prescribes the general rule that court proceedings be held in “an open and public court to which the

with the law, for publication in LawNet and/or the Singapore Law

public generally may have access”; justice should, at least in general, be done in the open, rather than hidden from the public eye (see, for example, Chua Yi

Jin Colin v Public Prosecutor [2022] 4 SLR 1133 at [34]). As Lord Atkinson

explained more than a century ago in Scott v Scott [1913] AC 417, while the hearing of a case in public may cause discomfort to both the parties and witnesses, it is “on the whole, the best security for the pure, impartial, and efficient administration of justice, [and] the best means for winning for it public confidence and respect” (at 463).

3 By and large, our judicial officers have consistently upheld the highest standards in their work, an observation that is reflected in the privilege the Singapore Judiciary enjoys of a high level of public confidence. But there is no room for complacency. Each time a matter comes before the court, it is incumbent on each Judge and judicial officer to conduct it in a manner that reaffirms the justification for that confidence. And when, occasionally, this standard is not met, it is incumbent on the appellate courts to call it out.

4 These appeals, unfortunately, present such a case. They are the latest in a small number of cases where the headline submission before the appellate court is not solely that the judge below erred on some point of legal principle or fact, but that the judge abdicated his duty to judiciously consider the evidence and submissions by reproducing a party’s submissions in the judgment and presenting it as the judge’s reasons for his decision. In this Judgment, I use “judicious consideration” as a compendious term to refer to a judge’s duty to arrive at a decision founded on reason following a fair, diligent and careful consideration of the material placed before the court; and “judicial copying” to

refer to the reproduction of a party’s submissions as the reasons for a judge’s decision.

5 As I explain in this Judgment, judicial copying does not invariably result in the judgment in question being set aside. That depends on a range of considerations. But it is almost always decried as a poor practice for judges to engage in and I myself have said as much (see Newton, David Christopher v

Public Prosecutor [2024] 3 SLR 1370 (“Newton”) at [3]). This is so for a

number of reasons, which I set out at [6] below; but an unsatisfactory judicial practice can cross the line and become a failure of the judicial function, which is when the affected judgment becomes liable to be set aside. The question in this case is how and where that line is to be drawn. The Prosecution contends that the district judge who heard the matter below failed to discharge his judicial duty, in that the judgment he issued explaining his decision to acquit the respondents reflects such a degree of copying and adoption of the respondents’ written submissions that it gives rise to a reasonable suspicion that he unthinkingly adopted the respondents’ submissions without giving judicious consideration to the case advanced by the Prosecution and the evidence on record. The judge in question, District Judge Soh Tze Bian (“DJ”) who has since retired, has been the subject of similar complaints on at least two prior recent occasions – first, before me in Newton, and more recently, before a different Judge of the General Division of the High Court in Ler Chun Poh v Public

Prosecutor [2024] 6 SLR 410 (“Ler Chun Poh”).

6 It should be quite obvious why judicial copying is unacceptable from at least three perspectives. First, it gives rise to the risk of the losing party feeling that it has not had a fair hearing when the judge appears essentially to have

unthinkingly adopted the case mounted by the other side. Second, in the eyes of the public, allegations of such copying embarrass the Judiciary and can engender a potential loss of confidence in the integrity of the judicial process. Third, from the perspective of the appellate court, there is a needless waste of the judicial resources that must be marshalled to examine not only the substantive merits of the decision below, but also the propriety of the process by which that decision was reached.

7 The present case illustrates these points, but I emphasise the last point here. This was a case of some factual complexity. The trial before the DJ lasted more than 40 days spread over a period in excess of two years. The matter came before me with a record in excess of 18,000 pages. The general rule in such a case would have been for me to approach the DJ’s decision with the usual measure of deference accorded to a trial court, especially on matters of fact, as it is not the role of an appellate court to retry a matter before it (see Soh Chee

Wen v Public Prosecutor [2025] 2 SLR 176 (“Soh Chee Wen”) at [1]). But that

paradigm only holds where there is no real concern over the procedural fairness of the decision or proceedings below such that the trial court’s decision is one which it is appropriate to accord deference to. Here, the Prosecution has challenged that paradigm. It argues that, due to the nature and extent to which the DJ allegedly copied from the respondents’ submissions, there was a serious failure of the judicial process below such that I should set aside the DJ’s grounds of decision and consider the matter afresh. Because this complaint against the

process by which the DJ arrived at his decision may affect the approach I would

take to the Prosecution’s substantive complaint against that decision, it has become necessary for me to deal with the former as a preliminary issue in this appeal. In short, an already complex appeal has been made more so.

Accordingly, this Judgment does not address the respondents’ guilt or innocence on the charges against them. Instead, it focuses on the anterior issue of the lens through which the appellate court should approach the appeals.

8 However, this case does present the opportunity to dispel any lingering misconception that judicial copying may be a legitimate or acceptable judicial practice. It is not. And that is so even though such copying is not in and of itself a ground on which a decision will or must be impugned. As I explain below,

Newton and other decisions that have touched on the implications of judicial

copying establish that judicial copying is not in itself a basis for displacing a decision. Instead, they emphasise that much will depend on the circumstances, and such copying may be, and often is, a symptom of an underlying problem. Given this, in considering whether to set aside a judgment that is infected with judicial copying (or for that matter any other process failing), it will often be helpful to approach this in terms of whether the copying (or other failing) occasions a breach of natural justice. In my judgment, the framework of natural justice is an appropriate prism through which complaints of process failings should generally be viewed because the principles of natural justice embody the baseline standards of due process that litigants are entitled to expect from the courts. In turn, the question of whether there has been a breach of natural justice in a given case must be answered based on whether a fair-minded and informed observer would reasonably suspect or apprehend that there has been a failure in due process. Where that threshold is crossed, it is taken that public confidence in the Judiciary may be undermined, and the appellate court must set that right by responding with the appropriate consequence(s): this may include remitting the tainted matters back to the same judge, or setting aside the impugned

decision and then either ordering a retrial, or reconsidering the evidence and the submissions afresh.

9 Applying this framework, I have come to the view that the Prosecution’s complaint is well-founded. In my judgment, it would reasonably be apprehended in the circumstances of this case that the DJ failed to judiciously consider the submissions and evidence before him, in particular the case advanced by the Prosecution. The effect of this finding is that the DJ’s decision should be set aside. The parties are agreed that if I came to that view, I should not order a retrial but should consider the matter afresh. This stems from the fact that no complaint has been raised with the way the evidence was taken. In the circumstances, the appropriate course is for me to consider the matter afresh based on the record as it stands, and I will proceed to hear submissions on this basis.

Background facts

10 As this Judgment does not address the merits of the charges brought against the respondents, a brief overview of the background facts will suffice.

11 The respondents are Mr Pek Lian Guan (“Pek”) and Mr Pay Teow Heng (“Pay”). At the material time, Pay was a director of Tiong Seng Contractors (Pte) Ltd (“TSC”), a wholly owned subsidiary of Tiong Seng Holdings Ltd (“TS Holdings”). Pek was the managing director of TSC, as well as the chief executive officer and a director of TS Holdings.

12 Pay and Pek each faced two charges under the Prevention of Corruption Act (Cap 241, 1993 Rev Ed) (“PCA”) for their involvement in extending two

loans to Mr Foo Yung Thye Henry (“Foo”). Foo was an employee of the Land Transport Authority of Singapore (“LTA”) at the material time.

13 The alleged impropriety arose from the LTA’s engagement of TSC as the contractor for the construction of the Great World Station and tunnels (“T220 Project”) on the Thomson-East Coast Line (“TEL”) of the Mass Rapid Transit system. Foo was the employee of the LTA responsible for supervising the execution of the T220 Project in varying capacities. When Foo encountered some personal financial difficulties, he reached out to Pay on two occasions seeking his help to arrange some loans. Pay did not immediately agree, but took Foo’s requests to Pek, who facilitated the approval of two staff loans from TSC to Pay, to enable Pay to use the money to extend two loans to Foo in June 2017 and March 2018 (referred to collectively as the “Loans”).

14 In October 2018, the Corrupt Practices Investigation Bureau (“CPIB”) received an anonymous complaint that Foo had been soliciting loans from contractors involved in LTA projects under his charge. Foo had also requested loans from other contractors of the LTA (see, for another recent case involving Foo, Public Prosecutor v China Railway Tunnel Group Co Ltd (Singapore

Branch) [2025] 4 SLR 69). Following investigations, the Prosecution brought

seven charges against Foo for corruptly obtaining or attempting to obtain gratification under s 6(a) of the PCA. Foo pleaded guilty to these charges and was sentenced on 2 September 2021 to a total of 66 months’ imprisonment.

15 In the proceedings below, the Prosecution brought two charges each against Pay and Pek for their respective roles in arranging and providing the Loans to Foo:

(a) For providing the Loans to Foo, Pay faced two charges under s 6(b) read with s 7 of the PCA, of corruptly giving gratification to Foo to induce Foo to do acts in relation to the LTA’s affairs which advanced TSC’s interests in relation to existing and future contracts with the LTA. (b) For facilitating the making of the Loans to Foo, Pek faced two charges under s 6(b) read with ss 7 and 29(a) of the PCA, of abetting Pay’s corrupt giving of gratification to Foo by causing TSC to extend staff loans to Pay with the knowledge that Pay would use the money to extend loans to Foo.

16 The CPIB recorded two statements from Pay. Pay’s 1st statement was recorded by Investigation Officer Tay Wenxun (“IO Tay”) on 11 September 2019 (“Pay’s 1st Statement”), and his 2nd statement was recorded by Chief Special Investigator Lim Eng Cheung (“IO Lim”) on 12 September 2019 (“Pay’s 2nd Statement”).

17 Two CPIB statements were also recorded from Pek. Pek’s 1st statement was recorded on 12 September 2019 (“Pek’s 1st Statement”) by Chief Special Investigator Tan Kian Tat (“IO Tan”). Following Pek’s indication on 13 November 2019 to Chief Special Investigator Johnston Kan (“IO Kan”) that he wished to clarify his 1st Statement, IO Tan recorded a 2nd statement from Pek on 3 February 2020 (“Pek’s 2nd Statement”).

The parties’ arguments below

18 Without oversimplifying matters, there was no dispute that the staff loans were advanced to Pay, who in turn extended the Loans to Foo. The key

question at trial was the intent with which these had been done. In this regard, s 8 of the PCA provides that where a gratification is given to a person who is an employee of the Government or a public body by a person who has or seeks to have dealings with the Government or any public body, the gratification is presumed to have been given and received corruptly as an inducement or reward. The Prosecution’s case was that this presumption applied, and the Loans were deemed to have been given by Pay as an inducement or reward to advance TSC’s business interests with the LTA. The presumption of a corrupt intent under s 8 of the PCA was also invoked against Pek such that he was deemed to have intentionally aided Pay with the requisite knowledge. The Prosecution 1 contended that the respondents had not been able to rebut the presumption of the Loans having been advanced corruptly, and it therefore maintained that it had proved its case beyond a reasonable doubt. The alleged corrupt purpose was said to be “to prevent Foo from causing trouble in the T220 [Project]”. To 2 establish its case, the Prosecution relied primarily on admissions in the respondents’ CPIB statements and some objective evidence. 3 19 The crux of Pay’s defence was that he had extended the Loans to Foo because he had a genuine friendship with Foo and for no other reason. Pay 4

Prosecution’s Written Submissions at the Close of Trial dated 25 June 2024 (“DC 1 Prosecution”) at para 4 (ROP 8987). DC Prosecution at para 5 (ROP 8987–8988). 2 DC Prosecution at paras 5 and 70 (ROP 8987 and 9025). 3 Pay Teow Heng’s Closing Submissions dated 25 June 2024 (“DC Pay”) at para 53 4 (ROP 16931).

challenged the reliability of his 2nd Statement and submitted that the objective evidence did not prove that he had made the Loans with a corrupt intent. 5 20 Pek made a preliminary argument that the presumption under s 8 of the PCA did not apply to his charges of abetment. His main defence was that he 6 had approved the staff loans to Pay (and thus the Loans indirectly) because he believed that Pay was helping Foo as a friend in need, and that this was for a good cause. Pek disputed the reliability of both of his CPIB statements and 7 8 pointed to other evidence, such as Foo’s evidence that he (Foo) had no corrupt intent when he asked for and received the Loans and that he did not believe that Pay or Pek had any corrupt intent in extending the Loans to him, to establish 9 his innocence.

Decision below

21 At the close of the Prosecution’s case, the respondents made a submission of no case to answer. The first section of the DJ’s grounds of decision (“GD”) dealt with this issue. The DJ accepted that the presumption of corruption under s 8 of the PCA applied to both Pay and Pek (GD at [19] and [22]). The DJ accordingly dismissed the respondents’ submission that there was no case to answer and called on the respondents to present their respective defences. It follows from what has been set out above that there was no dispute

DC Pay at para 55 (ROP 16932–16933). 5 Pek Lian Guan’s Closing Submissions dated 25 June 2024 (“DC Pek”) at para 78 (ROP 6 13106). DC Pek at paras 184–186 (ROP 13161–13162). 7 DC Pek at paras 494–495 (ROP 13446–13447). 8 DC Pek at paras 217 and 235 (ROP 13217 and 13228). 9

that the Loans had been made, and the key issue was the motive with which this had been done, or in other words, whether the presumption had been rebutted, and therefore, whether the Prosecution had made out its case on the totality of the evidence.

22 The DJ ultimately acquitted the respondents of all their charges and set out his reasons for the acquittals in the second section of the GD. The DJ considered that the entire case could effectively be determined based on three key issues: (a) whether the respondents’ incriminating CPIB statements were inaccurate, unreliable and unsafe to rely on (“Issue 1”); (b) whether Foo’s evidence as to his and the respondents’ states of mind was relevant and, if so, whether this was consistent with the respondents’ account that the Loans had been made innocently (“Issue 2”); and (c) whether a series of WhatsApp messages between Pay and Foo showed that Foo’s agency relationship with the LTA had been suborned (“Issue 3”) (GD at [28]).

23 On Issue 1, the DJ found that Pay’s 2nd Statement and Pek’s CPIB statements were unreliable (GD at [30]). Even though Pay had signed his 2nd Statement, the DJ considered that there were numerous deficiencies in the process by which IO Lim had recorded the statement that rendered the statement unreliable. In the DJ’s view, IO Lim had, among other things, predetermined Pay’s guilt, selectively recorded inculpatory evidence, retyped in his own words the amendments which Pay wished to make to his statements and provided false or incorrect explanations of certain words in the statement to Pay who was already exhausted at that time, thereby making it impossible for Pay to verify the accuracy of the amendments (GD at [30(a)(i)] and [30(a)(iv)]–[30(a)(viii)]).

24 Similarly, although Pek signed both his CPIB statements, there were several deficiencies in the process by which IO Tan recorded the statements that rendered them unreliable. According to the DJ, IO Tan had, among other things, admitted to trying to “frame” Pek, used a cut-and-paste method to assemble Pek’s 1st Statement which resulted in the repetition and emphasis of incriminating evidence and took the liberty of drafting Pek’s 1st Statement in his own words (GD at [30(d)(i)]–[30(d)(xiv)]). The DJ considered that Pek may not have been aware of the incriminating content of his 1st Statement because of the false sense of security engendered by the manner in which the 1st Statement was recorded and Pek’s possible inability to comprehend the recorded statement and therefore to correct any inaccuracies (GD at [30(d)(viii)] and [30(d)(x)]). Similar issues were also considered as having compromised the reliability of Pek’s 2nd Statement (GD at [30(d)(xv)]).

25 On Issue 2, the DJ held that Foo’s account that the Loans were given without corrupt intent was consistent with the respondents’ evidence, and the Prosecution was bound to accept the truthfulness of Foo’s evidence because it had not sought to impeach his testimony (GD at [34]–[38]). I note in passing here that even in this context, the DJ did not address the Prosecution’s contentions in respect of matters arising from the cross-examination of the respondents.

26 On Issue 3, the DJ found that the WhatsApp messages between Pay and Foo, which pertained to tenders or proposals for future LTA projects, did not indicate that Foo’s agency relationship with the LTA had been suborned or that Foo was beholden to or felt indebted to Pay (GD at [44]–[47]).

27 The DJ found on this basis that the Prosecution had failed to prove beyond a reasonable doubt its case against Pay and Pek that the Loans were given corruptly (GD at [49]). Given the DJ’s view that these three issues were dispositive of the case, he considered it unnecessary to decide on the other issues which arose from the trial, including the existence of objective evidence which was said to demonstrate the respondents’ corrupt intent (GD at [50]).

The scope of this Judgment

28 In this Judgment, I deal only with the preliminary issue which arose from the Prosecution’s contention that the DJ failed to discharge his duty to consider the matter judiciously. This contention was based, among other things, on the nature, extent and manner in which the DJ had or appeared to have copied from, reproduced, or adopted the submissions of the Defence.

29 To briefly explain, I took this course because the Prosecution submits that, if I were to conclude that the DJ failed to judiciously consider the matter below, I should give no weight to his decision and findings and proceed as if I were a first instance judge without exercising the usual restraint observed in terms of appellate intervention in a trial court’s decision, especially on findings of fact. Given that the resolution of this complaint would affect how I (and the parties) should approach the issue of the respondents’ guilt or innocence on the charges against them, I decided, after discussion with the parties at a Case Management Conference, to deal first with the issues of (a) whether the DJ had failed to consider the matter judiciously; and (b) if so, the consequences that should follow. It is with these two issues that this Judgment is concerned, and which I invited the parties to file supplementary skeletal arguments on.

The parties’ arguments

30 The differences between the parties’ arguments centre around three points: (a) first, the appropriate legal test to be applied to determine the Prosecution’s contention that I should have no regard to the DJ’s findings because of his alleged failure to discharge his judicial duty; (b) second, whether the DJ had failed to discharge his judicial duty; and (c) third, what, if any, consequences should follow upon the conclusion to the second question.

The Prosecution’s arguments

31 The Prosecution submits that the relevant question is whether there is a reasonable apprehension or suspicion, judged from the vantage point of a fair-minded and informed observer, that the DJ failed to discharge his duty to judiciously consider the material before him. The threshold of a “reasonable suspicion” is drawn from the principle articulated in cases involving allegations of bias, and the Prosecution contends it is equally applicable when determining whether a judge had failed to discharge his duty to consider the material. 10 32 Applying this test, the Prosecution submits that there is a reasonable suspicion that the DJ failed to discharge his judicial duty. The Prosecution’s case is that the nature and extent of judicial copying in this case, together with the DJ’s failure to consider certain key issues the Prosecution had raised, form the evidentiary basis upon which a fair-minded observer would reasonably apprehend that there might have been such a failure. Thus, the Prosecution contends that no weight should be accorded to the GD, and I should consider

Prosecution’s Skeletal Submissions dated 21 July 2025 (“Prosecution’s Skeletal 10 Submissions”) at paras 6–7.

afresh the evidence that is on the record and come to a conclusion as to the respondents’ guilt on the charges. 11

Pek’s arguments

33 Pek disagrees and submits that it is insufficient for the Prosecution to establish a reasonable suspicion of such a failure. Pek contends instead that the Prosecution must prove on a balance of probabilities that the DJ failed to discharge his judicial duty. According to Pek, the threshold of “reasonable suspicion”, which stops short of an actual finding of fact, is too low and inconsistent with the standard of proof in criminal proceedings that requires any doubts to be resolved against the Prosecution. However, Pek concedes that the 12 criminal standard of proof beyond reasonable doubt is too high, and submits that the lower standard of the balance of probabilities is an appropriate compromise and balance between the rights of the accused person and the integrity of the judicial system. 13 34 Pek further submits that the “reasonable suspicion” threshold should be confined to cases concerning apparent bias, and to the extent that there are precedents which have apparently applied it to allegations that a judge had failed to discharge his duty to judiciously consider the material, those cases are distinguishable because they involved allegations by the accused person rather than by the Prosecution. According to Pek, there is nothing anomalous in 14

Prosecution’s Skeletal Submissions at paras 20–21. 11 Pek Lian Guan’s Skeletal Submissions in HC/MA 9196/2024/01 dated 21 July 2025 12 (“Pek’s Skeletal Submissions”) at paras 2–5. Pek’s Skeletal Submissions at paras 6–8. 13 Pek’s Skeletal Submissions at paras 12–14. 14

holding the Prosecution to the higher standard of the balance of probabilities because the criminal standard of proof means that the Prosecution is fundamentally required to do more in terms of proof than an accused person. 15 35 Applying the approach he contends for, Pek submits that it has not been shown on the balance of probabilities that the DJ failed to discharge his duty to consider the material advanced by all the parties judiciously. In the main, Pek argues that it is unduly simplistic to focus on the fact that the DJ may have adopted parts of the respondents’ written submissions in the GD, given that there are other indicia which suggest that the DJ did apply an independent and judicious mind to the matter before him. These include: (a) the DJ’s own 16 identification of three issues on which the case turned; (b) the DJ’s 17 rearrangement and reorganisation of the respondents’ submissions in the GD; 18 (c) the supplementation of the respondents’ submissions with the DJ’s own observations and analysis; (d) the DJ’s references to the Prosecution’s 19 arguments in the GD; and (e) the DJ’s engagement with counsel, including the 20 Prosecution, in oral closing submissions. Pek submits that, having regard to 21 all the circumstances, it cannot be said that the DJ had failed to apply his mind judiciously to the issues and evidence, or that he had adopted the respondents’ submissions unthinkingly, and there is therefore no basis for the appeal to

Pek’s Skeletal Submissions at para 15. 15 Pek’s Skeletal Submissions at para 23. 16 Pek’s Skeletal Submissions at paras 24–28. 17 Pek’s Skeletal Submissions at paras 29–35. 18 Pek’s Skeletal Submissions at paras 36–51. 19 Pek’s Skeletal Submissions at paras 52–71. 20 Pek’s Skeletal Submissions at paras 72–74. 21

proceed on the basis of ignoring the DJ’s findings or of not according due deference to them. 22 36 Finally, if I were to conclude that the DJ had failed to apply his mind with due diligence to the matter, Pek advances several points on how I should proceed. First, he contends that it would not necessarily follow that I should consider the material wholly ignoring the DJ’s findings because it remains the case that this court is exercising appellate (as opposed to original) jurisdiction. Second, in line with the first point, Pek argues that any fresh consideration of the issues without regard to the DJ’s findings should be confined to the points that the Prosecution has raised on appeal. In any case, Pek also contends that I may nonetheless consider the GD and the findings set out there, but subject to such circumspection as I consider appropriate. 23

Pay’s arguments

37 Pay’s positions on the issues above are broadly aligned with those of Pek. Pay too disputes the Prosecution’s reliance on the “reasonable suspicion” test and submits that this should not be taken outside the context of cases on apparent bias. Instead, Pay argues that, in a case like the present where apparent bias is not alleged, the relevant inquiry is whether a reasonable person apprised of the circumstances would conclude (and not merely suspect) that the judge below had failed to discharge his judicial duty by failing to apply his mind to the evidence and the issues. Pay’s “reasonable conclusion” test entails proof 24

Pek’s Skeletal Submissions at paras 73–75. 22 Pek’s Skeletal Submissions at paras 78–88. 23 Pay Teow Heng’s Skeletal Submissions in HC/MA 9197/2024/01 dated 21 July 2025 24 (“Pay’s Skeletal Submissions”) at paras 7–35.

on the balance of probabilities, and his position on the law is therefore in substance identical to Pek’s. Indeed, counsel for Pay, Mr Tan Chee Meng SC (“Mr Tan”), confirmed this at the hearing before me.

38 On this basis, Pay argues that it cannot be said that the DJ had failed to discharge his judicial duty. Among other things, Pay disputes the Prosecution’s claim that the DJ had unthinkingly copied Pay’s written submissions and highlights the DJ’s exchanges with counsel in oral closing submissions as evidence of the DJ’s understanding of the case and the parties’ arguments. In 25 any case, even if I were to adopt the “reasonable suspicion” test, Pay relies on the same factors to contend that the DJ did not fail to discharge his judicial duty.

39 Finally, as to what should follow if I were to conclude that the DJ had failed to discharge his judicial duty, while Pay does not go as far as Pek to suggest that I should remain open to placing such weight as I think fit on the DJ’s findings in the GD, he contends that the extent to which I should disregard the DJ’s findings in the GD should be guided by the issues that have been framed and identified by the parties in their submissions. 26

Issues to be determined

40 In this Judgment, I deal with the following issues: (a) first, the applicable legal framework and principles for assessing the Prosecution’s claim that the DJ had failed to discharge his duty to consider the materials advanced by all the parties judiciously;

Pay’s Skeletal Submissions at paras 36–42. 25 Pay’s Skeletal Submissions at paras 45–49. 26

(b) second, based on the relevant principles determined at [40(a)] above, whether the DJ had indeed failed to discharge his duty to consider the matter judiciously; and (c) third, the consequences that should follow as a result of my conclusions on the foregoing issues.

The applicable legal framework

41 This is not the first case involving allegations that a judge copied a party’s submissions unthinkingly or that a first instance judge failed to apply his or her mind judiciously to the matter. An argument of a similar nature was made before me in Newton, as well as before another Judge in Lim Chee Huat v Public

Prosecutor [2019] 5 SLR 433 (“Lim Chee Huat”) and Ler Chun Poh. I have

also previously dealt with a case where the source of the alleged copying by the lower court judge in that case was a decision rendered by that judge in a different case (rather than from a party’s submissions) (see Yap Ah Lai v Public

Prosecutor [2014] 3 SLR 180 (“Yap Ah Lai”)). Most recently, the Singapore

International Commercial Court and Court of Appeal encountered a variation on this theme in the context of arbitration, where an arbitral tribunal was alleged to have incorporated into its award material that had been substantially copied from awards rendered in two different arbitrations which were chaired by the same presiding arbitrator (see DJP v DJO [2025] 1 SLR 576 (“DJP”) and DOI

v DOJ [2025] 4 SLR 657).

42 While the issue is not new, the legal framework by which such matters may be addressed remains a matter of some controversy. This is apparent from the way the present parties have taken fundamentally divergent positions even

on the framing of the relevant inquiry: while the Prosecution submits that the question before me is whether there is a reasonable suspicion that the DJ had failed to discharge his duty, the respondents argue that any such allegation must be proved at least on a balance of probabilities. What underlies this are three points. First, there is a difference of view as to whether the “reasonable suspicion” threshold, which was developed in the context of cases concerning bias, should be extended beyond that context to cases such as the present where there is no allegation of bias against the DJ. Second, while the Prosecution claims that the DJ failed to discharge his judicial duty, it has not always been clear what the specific content of the alleged “duty” is. Indeed, in its initial submissions for the appeal, the Prosecution advanced its argument in two ways: (a) one, that the DJ failed to apply his mind to the matter; and (b) two, that the DJ failed to give proper reasons for his decision. Third, there is a difference 27 of view as to whether the test varies when it is raised in the context of a decision that is in favour of an accused person in criminal proceedings.

43 On the first point, the Prosecution’s approach is perhaps a consequence of the prevailing case law, where the same circumstance of judicial copying has been analysed in different ways. In Newton, for example, the alleged copying of the Prosecution’s written submissions was relied on to mount a complaint of apparent bias. A different approach was taken in Lim Chee Huat, where the appellate court’s focus was on whether the nature and extent of the copying indicated that the district judge below had failed to apply his mind to the matter. Although it might be thought that these are inconsistent approaches, in my judgment, the better view is that expressed by Aidan Xu J in Ler Chun Poh,

Prosecution’s Submissions dated 6 May 2025 (“Prosecution’s Appeal Submissions”) 27 at paras 42 and 44.

which is that an alleged failure by a judge to apply his mind to the matter, on the one hand, and bias, on the other, may be “related, but separate, grounds for setting aside a trial judge’s decision” emphasis added.

44 While these may be separate grounds, there is often a common thread that underlies these cases, and that is the concern that there has been a breach

of natural justice which taints either the proceedings or the decision with the

colour of procedural unfairness. To use judicial copying as an example, the reason that essentially the same circumstance was capable of supporting different analyses in Lim Chee Huat and Newton is that the focal point of the analysis was not the copying itself, but the underlying problem of which the copying was a symptom or manifestation. That problem ultimately was a breach of natural justice. Procedural fairness is most starkly expressed in the rules of natural justice.

45 There are at least three consequences of situating a given complaint within the framework of natural justice. (a) First, depending on which rule of natural justice is alleged to have been breached, the relevant considerations and consequences may vary. As I will discuss below, this explains what might on first impression appear to be inconsistent outcomes in prior cases. Further, because a complaint that a judge manifested an appearance of bias is of a different nature to one that the judge failed to apply his or her mind to the matter, the analysis should be specifically tailored to each complaint. (b) Second, and following from the first point, the remedies or consequences that follow may differ depending on the rule of natural

justice that is implicated, the gravity and implications of the alleged breach, and its effect on the proceeding or decision. In some cases, a minimally invasive approach may be appropriate, and the matter may be remitted to the decision-maker to cure any unfairness; in others, the actual or perceived unfairness may be more pervasive, and the possibility of a retrial may have to be considered. (c) Third, once it is understood that the question is ultimately one of natural justice, it follows that the correct inquiry is whether there is a

reasonable suspicion that a breach of natural justice has occurred,

because the “reasonable suspicion” threshold focuses on the appearance of justice and gives effect to the fundamental principle that justice must not only be done, but it must be seen to be done.

The importance of procedural justice

46 In most cases, an appeal against a first instance decision will take the form of a challenge against its merits. The appellate court’s task in such cases is to assess the correctness of the lower court’s decision in law and/or in fact, with the extent of intervention in relation to matters of fact being circumscribed by the well-established principle that an appellate court will be slow to overturn a lower court’s findings of fact unless these can be shown to be plainly wrong or against the weight of the evidence (see, for example, Soh Chee Wen at [65]).

47 This contrasts with the supervisory role of the courts in the arbitral ecosystem at least in our jurisdiction and in most others that adopt some variant of the UNCITRAL Model Law on International Commercial Arbitration (“Model Law”), where the merits of the tribunal’s decision are beyond the limits

of the court’s jurisdiction. The court’s focus in applications to set aside or resist enforcement of an award, is not whether the tribunal arrived at the correct decision, but whether there is some jurisdictional or procedural defect that taints the tribunal’s decision (see, for example, AKN v ALC [2015] 3 SLR 488 at [37]– [39]; BLC v BLB [2014] 4 SLR 759 at [53]; and Swire Shipping Pte Ltd v Ace

Exim Pte Ltd [2024] 5 SLR 706 (“Swire Shipping”) at [1]–[4]).

48 Because there is no bar against an appellate court intervening on the ground that it considers the decision below to be wrong, it is not surprising that the distinction between substantive and procedural justice has not featured to the same degree in litigation as opposed to arbitration. In litigation, if a party is dissatisfied with a decision, it has the option of simply challenging it on its merits and inviting the appellate court to arrive at a different conclusion upon discharging the appropriate burden. Because of this, it may often be thought unnecessary, whether for tactical reasons or otherwise, to challenge a first instance court judgment on procedural grounds even if this may potentially be available. In arbitration, on the other hand, the typical inability to challenge a tribunal’s decision on the ground of an error of fact or law means that procedural injustice may be the primary or even the only available angle of attack.

49 This does not mean that procedural fairness is less important in court proceedings than in arbitration. The courts are as much concerned with ensuring that they arrive at their decisions in a procedurally just manner as they are with the substantive correctness of their decisions. In United Overseas Bank Ltd v

Ng Huat Foundations Pte Ltd [2005] 2 SLR(R) 425 (“Ng Huat”), Andrew

Phang Boon Leong JC (as he then was) encapsulated the interplay between substantive and procedural justice in the following observations (at [4]–[9]):

Procedural and substantive justice 4 It is axiomatic that every party ought to have its day in court. This is the very embodiment of procedural justice. The appellation “procedural” is important. Procedural justice is just one aspect of the holistic ideal and concept of justice itself. In the final analysis, the achievement of a substantively just result or decision is the desideratum. It is more than that, however. It is not merely an ideal. It must be a practical outcome – at least as far as the court can aid in its attainment. 5 However, the court must be extremely wary of falling into the flawed approach to the effect that “the ends justify the means”. This ought never to be the case. The obsession with achieving a substantively fair and just outcome does not justify the utilisation of any and every means to achieve that objective. There must be fairness in the procedure or manner in which the final outcome is achieved. 6 Indeed, if the procedure is unjust, that will itself taint the outcome. 7 On the other hand, a just and fair procedure does not, in and of itself, ensure a just outcome. In other words, procedural fairness is a necessary but not sufficient condition for a fair and just result. 8 The quest for justice, therefore, entails a continuous need to balance the procedural with the substantive. More than that, it is a continuous attempt to ensure that both are integrated, as far as that is humanly possible. Both interact with each other. One cannot survive without the other. There must, therefore, be – as far as is possible – a fair and just procedure that leads to a fair and just result. This is not merely abstract theorising. It is the very basis of what the courts do – and ought to do. When in doubt, the courts would do well to keep these bedrock principles in mind. This is especially significant because, in many ways, this is how, I believe, laypersons perceive the administration of justice to be. The legitimacy of the law in their eyes must never be compromised. On the contrary, it should, as far as is possible, be enhanced. 9 It is true, however, that in the sphere of practical reality, there is often a tension between the need for procedural justice on the one hand and substantive justice on the other. The task of the court is to attempt, as I have pointed out in the preceding paragraph, to resolve this tension. There is a further task: it is to actually attempt, simultaneously, to integrate these two

conceptions of justice in order that justice in its fullest orb may shine forth. [emphasis in original]

50 While these observations were made in the context of civil proceedings, they are equally applicable to criminal proceedings (see Mohammed Ali bin

Johari v Public Prosecutor [2008] 4 SLR(R) 1058 (“Mohammed Ali bin Johari”) at [119] and Public Prosecutor v Goldring Thomas Nicholas [2014] 1

SLR 586 at [40]). Indeed, given that criminal proceedings implicate not only private interests, but the public interest and the life and liberty of the accused person, it stands to reason that courts will be all the more anxious to ensure the fairness of the process.

51 Although the concept of “procedural justice” might seem amorphous, its core resides in the principles of natural justice. It is well-established that there are two foundational pillars of natural justice, namely, that: (a) the parties should be given a proper opportunity to be heard (audi alteram partes); and (b) the adjudicator should be disinterested and unbiased (nemo judex in causa sua) (see, for example, Re Shankar Alan s/o Anant Kulkarni [2007] 1 SLR(R) 95 (“Shankar Alan”) at [42] and DJP at [36]). I refer to the former rule as the “fair hearing rule” and to the latter as the “rule against bias”. These rules of natural justice and their practical outworking set down the baseline standard of due and fair process.

52 Notwithstanding that the same factual matrix may engage both rules, it is important to bear in mind that they are “separate concepts and are governed by separate considerations” (see Tang Kin Hwa v Traditional Chinese Medicine

Practitioners Board [2005] 4 SLR(R) 604 at [10], citing B Surinder Singh

Kanda v Government of the Federation of Malaya [1962] AC 322 at 337). As I

explain below, the conceptual distinction between the fair hearing rule and the rule against bias is significant because the inquiry and the relevant factors that go into it may differ.

53 It follows that it is open to a litigant in court proceedings to challenge a decision on the basis that there has been a failure in the process that led to it, by showing that there has been a breach of either or both of the rules of natural justice. While this is not often pursued because of the possibility of seeking relief simply on the basis of error, such challenges are nonetheless raised from time to time.

54 An example is the argument – often encountered in civil appeals – that the trial judge’s finding concerned a point that was inadequately pleaded or was an impermissible departure from the parties’ cases. In V Nithia v Buthmanaban

s/o Vaithingam [2015] 5 SLR 1422 (“V Nithia”), the Court of Appeal articulated

the general principle that “the parties are bound by their pleadings and the court is precluded from deciding on a matter that the parties themselves have decided not to put into issue”, subject to the exception that a party may be permitted to depart from its pleadings if doing so would not cause irreparable prejudice to the other party (see V Nithia at [38]–[40] and How Weng Fan v Sengkang Town

Council [2023] 2 SLR 235 at [18]).

55 The common thread which runs through both the rule and the exception is procedural fairness (see V Nithia at [37]). As noted in Sheager s/o TM Veloo

v Belfield International (Hong Kong) Ltd [2014] 3 SLR 524, “pleadings serve

the important function of upholding the rules of natural justice” by requiring a

party to give his or her opponent fair notice of the case he or she has to meet (at [94]). The specific rule of natural justice engaged in this context is the fair hearing rule.

56 There is a parallel to this in arbitration where the claim made is that the tribunal adopted an unforeseeable chain of reasoning. The sting of such a complaint is that there has been a breach of natural justice because (a) the parties did not have reasonable notice of the tribunal’s intended reasoning; or (b) the tribunal’s reasoning did not have a sufficient nexus to the parties’ arguments (BZW v BZV [2022] 1 SLR 1080 (“BZW”) at [60(b)]), as a result of which the parties did not have the opportunity to address the tribunal on a point that proved decisive as far as the tribunal was concerned. In Vietnam Oil and Gas Group v

Joint Stock Company (Power Machines – ZTL, LMZ, Electrosila Energomachexport) [2025] 2 SLR 273 (“Vietnam Oil”), the Court of Appeal

explained that such a complaint involves the fair hearing rule, because “[t]he overriding concern … is fairness in that a tribunal ought to give the parties a reasonable opportunity to present their cases and to respond to what is being raised against them” emphasis added.

57 Where the gravamen of the complaint lies in procedural unfairness, if the failure is of sufficient gravity, it may even displace arguments that the judge’s finding may be correct in law or on the evidence. As Phang JC observed in Ng Huat, “[t]he obsession with achieving a substantively just and fair outcome does not justify the utilisation of any and every means to achieve that objective”, and “[t]here must be fairness in the procedure or manner in which the final outcome is achieved” (at [5]). A corollary to this is that the court may on occasion be forced to reject a claim even if it considered that the claim might

have succeeded if argued differently (Al-Medenni v Mars UK Ltd [2005] EWCA Civ 1041 at [21]; Satyam Enterprises Ltd v Burton [2021] 2 BCLC 724 at [38]).

Situating a complaint within the natural justice framework

58 Hence, the Prosecution’s case in these appeals implicates both the procedural and substantive aspects, though these are related. As a substantive matter, the Prosecution contends that the DJ’s findings were incorrect as a matter of fact and law. But it also complains that this failure stemmed from a serious procedural failure, which in my judgment falls within the framework of the rules of natural justice. The heart of the Prosecution’s complaint on the procedural justice point is that the DJ failed to apply a judicious mind to consider its arguments, and that he adopted the case mounted by the Defence without due regard to the Prosecution’s submissions. This “goes not to whether the decision is correct on the merits having regard to the evidence and the law, but to whether the process by which it was reached is procedurally fair” (see

Cojocaru v British Columbia Women’s Hospital and Health Centre [2013] 2

SCR 357 (“Cojocaru”) at [12]). I digress to make a brief observation on a separate point that is a feature of Canadian law and was dealt with in Cojocaru at [14]–[29]. That is the presumption of judicial integrity and impartiality. I do not think it is either necessary or appropriate to consider its applicability in the present case which, as I explain below, is concerned with whether there is a reasonable basis for apprehending a failure of natural justice, rather than with whether there has in fact been a judicial failure on the part of the DJ. In my view, the said presumption will not be relevant in the former case, and I therefore say no more about it.

59 Returning to the main thread of my analysis, in my view, the duty of a judge to consider a matter judiciously is an aspect of the fair hearing rule. More specifically, it is a correlative duty to the parties’ right to be heard, which consists of: (a) a positive aspect, in the opportunity to present the evidence and advance the propositions of law which support one’s case; and (b) a responsive aspect of having notice of one’s opponent’s case and the opportunity to respond to it (see Swire Shipping at [79], citing JVL Agro Industries Ltd v Agritrade

International Pte Ltd [2016] 4 SLR 768 at [146]–[147]). Both aspects would

only have meaning if the opportunity to present one’s case is met with active, impartial, fair and diligent judicial consideration. If the position were otherwise, and all that the right to be heard entailed was the judge physically reading or listening to the parties’ cases, the fair hearing rule would be devoid of any meaningful content. It follows that a failure of a judge to consider the matter before him or her judiciously may found a complaint of breach of natural justice based on the fair hearing rule.

60 As prefaced at [43]–[45] above, situating a complaint of procedural unfairness within the framework of natural justice in this way will often afford a more principled and doctrinally satisfactory analysis as opposed to an approach that treats specific instantiations of procedural unfairness as independent grounds of challenge that are governed by their own idiosyncratic rules and standards. For example, to ask how much judicial copying is “too much” begs the question of the applicable legal yardstick that determines when a judge would be found to have crossed the line. Similarly, where it is alleged that the judge gave “inadequate reasons”, it is difficult to assess in the abstract what amounts to “adequate” or “inadequate” reasons without knowing the purpose for which the question is being asked. The same could be said in the

context of excessive judicial intervention. In my view, the principles of natural justice can and, in this case, do provide the context in which complaints of procedural unfairness fall to be considered and the legal yardstick against which they should be answered. This, as I have noted above, provides insight into how seemingly irreconcilable cases can stand with one another upon closer analysis. In each case, one must first identify what the gravamen of the complaint really is. I illustrate this below with reference to three types of allegations of poor judicial conduct that have emerged in the case law and which I have referred to above: (a) allegations of judicial copying; (b) allegations that a judge has failed to give adequate reasons for his or her decision; and (c) allegations of excessive judicial interference.

Judicial copying

61 Cases of judicial copying have arisen in two contexts which are differentiated by the source of the copying: (a) first, where the judge is alleged to have copied or adopted a party’s submissions (see, for example, Lim Chee

Huat and Newton); and (b) second, where the judge is alleged to have copied

from another decision, whether of his or her own or of another judge (see, for example, Yap Ah Lai and DJP (albeit the latter was in the context of arbitration)). However, even within each category, there are divergences in the focal point of the courts’ reasoning. This is explicable by the different bases upon which the complaint was framed in each of these cases – whether under the fair hearing rule or the rule against bias.

62 The complaint levelled against the district judge in Lim Chee Huat was focused on the fair hearing rule. There, the district judge’s grounds of decision bore such a degree of similarity to the Prosecution’s submissions that they even

reproduced a typographical error made in those submissions. After a survey of foreign authority, Xu J opined that what was critical was that “the judge must have exercised some judgment or thought” and, therefore, the question was whether, in light of the alleged copying, it appeared that “the trial judge exercised his mind on the facts and circumstances of the case before him” (at [49]). In my view, the question in Lim Chee Huat was whether there was a breach of the fair hearing rule by reason of the district judge not having applied his mind to the matter. The court’s analysis was consistent with this, as it held that the combination of the extent and manner of the district judge’s reproduction of the Prosecution’s submissions – described as an “unprocessed adoption” – and the lack of any indication of an assessment of the submissions of both sides and weighing between them, meant that it could not be said that the district judge had properly applied his mind to the matters before him, particularly the Defence’s arguments (at [51]–[52]). The fair hearing rule was breached because the right of the Defence to be heard had been rendered illusory, and on this basis, the court examined the matter afresh without deference to the district judge’s findings (at [56]–[59]). Implicit in this, is the point I have noted earlier that the right to a fair hearing extends beyond the literal “hearing” all the way through to the consideration, evaluation and determination of the matter.

63 In contrast, the complaint in Newton was advanced on the basis of bias based on prejudgment. The appellant alleged that a “fair-minded and informed observer would reasonably apprehend that the [district judge] was biased against his client” based on a confluence of (a) the district judge having already prepared his grounds of decision before hearing oral submissions from either side; and (b) the reproduction of substantial portions of the Prosecution’s

submissions in the grounds of decision (at [22]). Even if the facts of the case

could conceivably have disclosed grounds for a complaint based on a breach of

the fair hearing rule in similar terms to that mounted in Lim Chee Huat, that was not how the appellant proceeded, and the principal question before me was whether the district judge’s decision ought to be set aside on the ground of

apparent bias (at [29]). After considering the totality of the circumstances, I

answered this in the negative (at [43]). Notwithstanding the district judge reproducing almost the entirety of the Prosecution’s written submissions in his grounds of decision and, at least initially, intending to give his decision without hearing oral submissions, he did accede to the request of the appellant’s counsel to make oral submissions. The exchanges between the district judge and the appellant’s counsel during these submissions indicated that the district judge had digested the parties’ cases and considered their merits (at [45]–[47]).

64 The differences between Lim Chee Huat and Newton go beyond the fact that the complaint was made out in the former but not the latter. The two cases demonstrate the importance of a proper appreciation (a) first, of whether the complaint is grounded in the fair hearing rule or the rule against bias; and (b) second, of how the rule of natural justice that has been invoked is specifically alleged to have been breached. To illustrate, consider, for example, the factor of the judge’s engagement with the parties at the hearing. I accorded this significant weight in Newton, and indeed, it outweighed the undisputed fact that the district judge had reproduced almost the entirety of the Prosecution’s written submissions and analysis (at [44]). But this has to be seen in light of the complaint in Newton being that the district judge appeared to have been biased in having prejudged the matter, and to have been unwilling or seemingly unable to hear the appellant out. In those circumstances, the fact that the district judge

agreed to hear and then engaged with the appellant’s oral submissions undermined the complaint because the district judge had done precisely what the appellant claimed he was either not willing or able to do. It does not follow from Newton, however, that the judge’s engagement with the parties or their counsel would invariably be a decisive (or even necessarily a relevant) factor for other types of complaints grounded in the fair hearing rule. In the final analysis, context is everything and much will depend on the facts of the case and the specific complaint in question.

65 In cases like Lim Chee Huat which involved complaints that were in substance grounded in the fair hearing rule, it may not be the case that a single factor, such as engagement with the parties during the hearing, would necessarily bear considerable weight in resolving the matter. This is reflected in the decision of the Hong Kong Court of Appeal in Lo Kai Shui v HSBC

International Trustee Ltd [2023] 6 HKC 411 (“Lo Kai Shui”), where Kwan VP

(with whom Yuen JA and Lam JA agreed) observed as follows (at [14]):

The good impression created by the display of judicial temperament and active engagement with counsel throughout the hearing cannot effectively displace the impression given by a judgment of extensive copying that the judge has subsequently abdicated his core judicial responsibility to think through for himself the issues he has to decide, and has uncritically adopted the submissions of the winning side, particularly when the judgment was given some considerable time after the hearing.

66 I agree with these observations and reiterate – context is everything. For one thing, the complaint that the judge below had not properly applied his mind to the matter, as was the case in Lim Chee Huat, is wider in scope than the targeted complaint in Newton, that the district judge had prejudged the matter before the hearing by reason of the fact that he had come prepared to deliver

judgment and did subsequently deliver a judgment that reproduced the submissions. As I also observed in Newton, in our system of litigation where in most matters comprehensive written submissions will have been filed in advance of the oral hearing, a judge would be expected to come prepared for the hearing having read those submissions and in so doing, having formed certain provisional views (especially where such matters do not involve unduly complex issues or difficult questions concerning the oral evidence) (at [50(c)]). The judge should remain open to being persuaded otherwise at the hearing, but there is no expectation that the judge should come to the hearing with an empty mind or a blank slate; only an open mind (see Prometheus Marine Pte Ltd v

King, Ann Rita [2018] 1 SLR 1 (“Prometheus Marine”) at [39]). In Newton, the

matter concerned sentencing upon a plea of guilt pursuant to an admitted statement of facts, and written submissions had already been filed and had evidently been digested. This again was quite different from the much more contentious setting in Lim Chee Huat, where the appellant claimed trial and raised disputes of fact that called for a proper appreciation and assessment of evidence. Similarly, Lo Kai Shui involved a hotly contested civil action.

67 Further, the nature of the complaint in Lim Chee Huat called for an examination of the judge’s conduct throughout the entirety of the proceedings, so that the judge’s engagement with the parties and their counsel was unlikely to carry the same weight, because that, at best, could show that the judge had been exercising his mind at the hearing. Where a judge reserves his decision, that would suggest a need to consider and resolve points of some factual or legal difficulty. If the judge then subsequently hands down a decision that appears on its face to be an uncritical adoption of one side’s submissions, the fact that he might have demonstrated an application of his mind at an earlier point in time

provides no assurance that he had continued to do so after the hearing and, most importantly, when he arrived at his decision (see Lo Kai Shui at [65] above).

68 That said, notwithstanding the conceptual differences between the fair hearing rule and the rule against bias, they are not mutually exclusive and there may often be an overlap between them such that the same set of facts could potentially found a complaint on either basis. Depending on which line is taken, it is possible that a complaint put one way may be made out while a different complaint put another way may fail on a similar set of facts. Simply put, since the question is different, so too may the answer be. This is demonstrated by a comparison between Yap Ah Lai and DJP, which both involved allegations that the decision-maker’s decision was copied from a different decision rendered in a different case.

69 In Yap Ah Lai, it was alleged that certain key paragraphs of the district judge’s decision, in which he set out the reasons for the sentences he imposed on the appellant, were identical to three paragraphs of another decision he issued in a different case at about the same time (at [4]). In these circumstances, I considered the question before me to be whether there was an appearance that the district judge had failed to properly apply his mind to the facts and circumstances of the case before him (at [70]). Given the material differences between the facts of the two cases, I was satisfied that there was a reasonable basis for thinking that in adopting the same reasons given in a different case, the district judge had failed to fully appreciate and evaluate the material in the case that was before him (at [73]).

70 The focus in Yap Ah Lai was on the fair hearing rule. The concern was of a similar nature to the complaint in Lim Chee Huat, in that the appellant’s right to a fair hearing (which would, as I have noted at [59] above, include the right to a judicious consideration of the specific facts, evidence, and arguments advanced in his or her case) had been denied given the manner in which the district judge seemed to have approached the matter. In Lo Kai Shui, Kwan VP described this, from the perspective of the court, as the “core judicial responsibility [of a judge] to think through for himself, the issues he has to decide” (at [14]). In my view, that is a fair summary. Even if it might have been possible in Yap Ah Lai for the complaint to be made on the basis of prejudgment or apparent bias, that was neither how it was put nor how I approached the matter. However, I was satisfied that it could reasonably be apprehended that the district judge had failed to consider the matter judiciously given the identical framing of the reasons and evaluation of the facts despite the seemingly different circumstances in the two cases.

71 On the other hand, in DJP, an application was brought to set aside an arbitral award for breach of natural justice, with the headline complaint being that the tribunal had copied substantial parts of the award from awards in parallel arbitrations that involved similar facts and similar underlying contracts (albeit with some material differences). Notably, the same presiding arbitrator had sat in all three arbitrations, although the co-arbitrators in the impugned arbitration were not involved in the parallel arbitrations.

72 There were distinct facets to the complaint in DJP. One key facet was put in terms of bias; it was argued that there was a reasonable suspicion that the tribunal had prejudged the matter by using the outcome in the parallel

arbitrations as the starting point for its decision (DJP at [65(a)]). Specifically, it was contended that the way in which the tribunal developed its award meant that its decision-making process was tainted by both an anchoring bias and a confirmation bias (DJP at [34]). The issue was not so much whether the tribunal had given effect to the applicant’s right to be heard by applying its mind to the matter before it (as was the case in Yap Ah Lai), but whether the tribunal had approached the arbitration in a manner that gave rise to a reasonable apprehension that it had failed to consider the matter with a truly open mind. There is of course some overlap between these two questions given that one element of failing to keep an open mind to the case at hand is a failure to apply a judicious mind to it. Thus, for example, the fact that the tribunal in DJP had gone as far as to reproduce portions of the awards in the parallel arbitrations that were plainly not applicable to the dispute before it could, apart from showing that the tribunal had not approached the specific circumstances of the case before it with an open mind, equally demonstrate a failure on its part to consider the matter judiciously. But the two inquiries are ultimately distinct. While the focus in Yap Ah Lai was on whether the district judge could be said, in a meaningful sense, to have made a decision on the matter before him, the focus in DJP was on a logically anterior matter: the starting point of the tribunal’s decision-making (namely, using the awards in the previous arbitrations as a base for deciding the matter before it) and whether that starting point resulted in the parties being on an unequal footing (at [65] and [67]).

73 The Court of Appeal held that there were clear grounds for reasonably apprehending that the presiding arbitrator may have been materially influenced by the earlier decisions that he had been a party to in the parallel arbitrations. It was not permissible for the presiding arbitrator (or, for that matter, any member

of the tribunal) to draw on these earlier decisions or to use them as a starting point, as it was incumbent on the tribunal to consider the matter before it afresh (DJP at [70]). While the tribunal had made some adjustments to the awards in the parallel arbitrations to account for what were thought to be the specificities of the arbitration and could, at least to some extent, be said to have applied its mind to the matter, this did not address the problem that it appeared predisposed towards the same outcome, findings, and reasoning that had obtained in the parallel arbitrations. This predisposition, as the court explained, manifested in the two forms noted above: (a) first, anchoring bias, referring to an unconscious tendency to rely on a conclusion earlier made without regard to new information and fresh analysis; and (b) second, confirmation bias, referring to the difficulty of persuading a decision-maker who has come to an initial view to then change its view (DJP at [74]).

74 In short, whether or not the tribunal had actually applied its mind to the dispute before it, it appeared to have approached the dispute with a coloured mind, and once this was shown to be the case, it could not have made a difference how diligently the material was analysed. As a result, even if the fair hearing rule might not have been breached in the same manner as in Yap Ah Lai, the tribunal had breached the rule against bias, which did give rise to a serious breach of natural justice.

75 For completeness, it should be noted that the Court of Appeal also found that the tribunal did breach the fair hearing rule, albeit in a different way. The fair hearing rule encompasses a number of aspects, of which the duty to consider matters judiciously is but one. In DJP, the Court of Appeal held, separately, that there had been a breach of the fair hearing rule, in that the tribunal’s knowledge

of and access to material from the parallel arbitrations were extraneous considerations in the tribunal’s decision that the parties were unaware of and so were given no opportunity to address (DJP at [80]–[82]). This underscores the importance of focusing not just on the rule of natural justice invoked, but also on the specific way in which the rule is alleged to have been breached.

Inadequacy of reasons

76 A different strand of complaint is the contention that the judge failed to give adequate reasons for his or her decision. In my view, while the cases have not situated such complaints within the framework of natural justice, there may be instances where it may be possible to do so, and in those instances, doing so may present a more principled means of analysing the issue, and help avoid the dangers of attempting to assess the adequacy or sufficiency of reasons in the abstract.

77 In the judicial context, the leading decision on this issue is Thong Ah Fat

v Public Prosecutor [2012] 1 SLR 676 (“Thong Ah Fat”), where the Court of

Appeal highlighted four reasons that underlay the judicial duty to give reasons: (a) First, the process of engaging with the parties’ submissions and analysing the evidence before the court leads to sounder decisions (at [20]). (b) Second, the provision of reasons to the parties ensures that they know why they have won or lost (at [21]).

(c) Third, the provision of reasons assists the appellate court by giving it sufficient material to understand and do justice to the decision taken at first instance (at [22]). (d) Fourth, the giving of reasoned decisions is a facet of open justice in that it curbs arbitrary exercises of power by allowing decisions to be scrutinised, attacked and defended (at [23]).

78 The failure to give adequate reasons will not typically afford a freestanding basis for setting aside a decision. One instance where it may do so is where it is relied on to establish a sufficiently serious breach of natural justice based on the fair hearing rule.

79 It is important here to emphasise that while judges – and, as will be seen, arbitrators or quasi-judicial decision-makers to a more limited extent – are generally subject to a duty to give reasons for their decisions, the content of the duty is context-sensitive, as are the consequences of failing to adhere to it. This follows from the fact that the four justifications identified by the Court of Appeal in Thong Ah Fat as undergirding the duty to give reasons are not uniformly applicable to different types of decision-makers. Indeed, this point was emphasised by the Court of Appeal in CVV v CVB [2024] 1 SLR 32 (“CVV”), which concerned an application to set aside an arbitral award. Although the court accepted that an arbitral tribunal was under a general duty to give reasons under Art 31(2) of the Model Law, it cautioned that it was wrong to import the principles laid down in Thong Ah Fat into arbitration without appropriate modification because “Thong Ah Fat sets out the standards applicable to judges in court cases, and different considerations are in play in a court case as opposed to arbitration” emphasis in original.

80 Turning to the justifications identified by the Court of Appeal in Thong

Ah Fat for the judicial duty to give reasons, the third reason – the facilitation of

appellate scrutiny – is not generally relevant to arbitration because there is generally no right of appeal against the merits of an arbitral tribunal’s decision. The facilitation of appellate scrutiny over a decision is grounded in the fair hearing rule because it ensures that the right of appeal is meaningful (see R v

Sheppard [2002] 1 SCR 869 at [66]). The point, as one commentator has put it,

is that “[t]he judge whose decision is being revisited should not be able, through the provision of no reasons, or inadequate reasons, to deny a party its positive right to effective judicial review or appeal” (see B V Harris, “The Continuing Struggle with the Nuanced Obligation on Judges to Provide Reasons for Their Decisions” (2016) 132 LQR 216 at p 221).

81 Thus, in Thong Ah Fat, the Court of Appeal set aside the appellant’s conviction on a capital charge of drug trafficking and ordered a retrial because it found itself “unable to affirm or overrule” the trial judge’s decision because it did not understand how he had arrived at certain findings that led him to reject the appellant’s defence (at [58]). These deficiencies included, for example, the lack of clarity as to what the judge had found in respect of the appellant’s mens

rea, in that it was unclear whether the appellant had been convicted based on a

finding of actual knowledge or wilful blindness as to the nature of the drugs in his possession or based on the statutory presumption of knowledge not having been rebutted. The problem with such ambiguity, as the court observed, was that the appellant could not “exercise his right of appeal effectively” without knowing what the judge had found his precise mens rea to be (at [48]).

82 An argument of this kind will generally hold no water in arbitration because, as noted above, the court does not usually sit on appeal from an arbitral tribunal’s decision (see CVV at [34]). And while a breach of the fair hearing rule is a ground for setting aside an arbitral award on the ground of breach of natural justice, the focus is on a fair hearing before the tribunal and not a fair hearing of an appeal before the court since no such right of appeal exists in the first place. Similarly, such an argument would hold less weight in final appellate proceedings because there is strictly no further right of appeal in these cases. Thus, the Court of Appeal in Public Prosecutor v Takaaki Masui [2022] 1 SLR 1033 held at [50] that a High Court hearing a Magistrate’s Appeal may dismiss the appeal without providing independent reasons where the grounds of appeal have been considered in detail and rejected by the trial judge.

83 The fourth reason in Thong Ah Fat – open justice – is another point of divergence between litigation and arbitration. Given arbitration’s nature as a private and consensual mode of dispute resolution, an arbitral tribunal is not accountable to the public in the same way that the courts and judges are. Instead, as party autonomy is a cornerstone of the legal framework on arbitration, “an integral feature and consequence of party autonomy is that parties choose their arbitrators and are bound by the decisions of their chosen arbitrators” (see COT

v COU [2023] SGCA 31 (“COT”) at [2]). The arbitrators are principally

accountable to the parties whose consent is the fount of their authority, and, from the parties’ perspective, this means that their avenues of complaint against what they consider to be inadequate reasoning are far more limited.

84 It is unnecessary and indeed, inappropriate, for me to say much more about the relevance of the duty to give reasons in the arbitral context, given the

ruling in CVV. I do observe that the relevance of a tribunal’s reasoning has featured only in two narrow situations, both giving rise to a potential breach of natural justice: first, as noted at [56] above, where the tribunal adopts a chain of reasoning that was not reasonably foreseeable to the parties in that it does not bear a sufficient nexus to the arguments the parties actually put forward; and second, where the absence of any reasoning on a point, when viewed in the totality of the circumstances, leads to the inescapable inference that the tribunal did not consider that point at all. And in the latter respect, the core question is whether the tribunal wholly failed to consider the point and not whether it accurately understood, analysed and dealt with it (see DKT v DKU [2025] 1 SLR 806 at [8(c)]).

85 Returning to the judicial setting, the reasons provided by a judge or decision-maker will often constitute the primary (if not the only) evidence by which a judge can satisfy the parties and the public that the court has given due consideration to the matter before it. The importance of reasons in this context stems from the fact that a degree of opacity is inherent in the judicial decision- making process. A judge’s consideration of the material takes place typically in private and, as a matter that occurs in the mind of the judge, it will not be visible to anyone. To be sure, the privacy of the decision-making process is underpinned by sound policy reasons, some of which were outlined by the Singapore International Commercial Court in CZT v CZU [2023] 5 SLR 241 (at [44]), in a case concerning an unsuccessful application for orders requiring the members of an arbitral tribunal to produce the records of their deliberations:

(a) It allows a decision-maker to reflect, draw conclusions, and if he or she is so minded, change those conclusions in the course of thinking through the matter. (b) It promotes frank and unreserved discussion when there is more than one decision-maker. (c) It protects the decision-maker from outside influence.

86 This lack of transparency, although justified as a matter of policy, can give rise to the concern of lack of accountability, where it results in a party reasonably apprehending, from either the absence of reasons or from the seemingly uncritical adoption of the other party’s case, that its own case has not been judiciously considered. From the Judiciary’s point of view, aside from the duty to the parties who have come to the court seeking justice, there are two additional dimensions: first, of accountability to the public because the legitimacy of the Judiciary is dependent on public confidence; and second, of the heightened duty of a judge to adequately apply his or her mind, such that it could be said that he or she exercised the discretion and judgment required by the judicial office (see Lim Chee Huat at [48]).

87 Even so, context will invariably be significant. I make two points. First, to the extent the issue is framed in terms of whether the judge has given judicious consideration to the matter, a failure to state reasons will not necessarily be conclusive. An appellate court should consider the totality of the circumstances and other indicia to assess holistically whether grounds exist for reasonably apprehending a failure to judiciously consider the matter. A judge’s reasons may not always be expressed in the form of a written or oral decision

delivered at the end of the proceedings: the judge may, for example, have in the course of the proceedings expressed views on the merits of the parties’ cases. Second, there will be gradations in the standard of reasoning that would be expected of a judge across the wide range of matters that come before the courts. Thus, “the more profound the consequences of a decision are, the greater the necessity for detailed reasoning” (Thong Ah Fat at [33]). In general, a more exacting standard of reasoning would be applied in criminal cases, and a fortiori in capital cases where the Court of Appeal went as far as to say in Thong Ah Fat that “detailed reasoning and grounds justifying the outcome ought to be given … as a matter of course” (at [14] and [47]).

Excessive judicial interference

88 Finally, turning to allegations of excessive judicial interference, there is a body of local authority which may be interpreted as having treated this as an independent ground of challenge against a judicial decision (see, for example,

Shankar Alan at [105]; BOI v BOJ [2018] 2 SLR 1156 (“BOI”) at [111]–[113]; Muhammad Nabill bin Mohd Fuad v Public Prosecutor [2020] 1 SLR 984 at

[164]–[166]). However, in my judgment, such complaints can be properly rationalised as a breach of either the fair hearing rule or the rule against bias.

89 This might, at least at first blush, appear to depart from what I said almost two decades ago in Shankar Alan. There I suggested that excessive judicial interference was a “separate basis” from apparent bias for my decision that the disciplinary tribunal’s decision should be quashed (at [105]). Shankar

Alan has been referred to by subsequent courts, including the Court of Appeal

in BOI, as suggesting that excessive judicial interference stands apart from the rule against bias. But there is no inconsistency.

90 The point I was making in Shankar Alan by distinguishing excessive judicial interference from bias was not that excessive judicial interference was necessarily to be considered as a separate matter from bias (or more generally from the natural justice framework). Rather, it was that on the facts of Shankar

Alan, where I had already held that the applicant had successfully made out his

complaint grounded in apparent bias, there was a separate and further basis for impugning the decision of the disciplinary tribunal on the grounds of excessive judicial interference, which would have amounted to a breach of natural justice based on the fair hearing rule. Thus, I stated that, quite apart from prejudgment – which is a species of bias – a complaint of excessive judicial interference could be premised on “the risk of a fair trial being compromised because of the failure of the tribunal to observe its proper role and its duty not to descend into the arena” (see Shankar Alan at [110]). This derived from two decisions of the Court of Appeal which I referred to as illustrations, namely, Singapore Amateur

Athletics Association v Haron bin Mundir [1993] 3 SLR(R) 407 (“Haron bin Mundir”) and Roseli bin Amat v Public Prosecutor [1989] 1 SLR(R) 346

(“Roseli bin Amat”) (see Shankar Alan at [109]).

91 In Haron bin Mundir, the respondent was a track and field athlete who was subjected to a disciplinary inquiry by the appellant’s management committee after he returned to Singapore prematurely from overseas training. He was found guilty of misconduct by a disciplinary subcommittee which recommended that he be suspended from active competition for two years. Subsequently, the report of the disciplinary subcommittee was distributed to the management committee, which invited the respondent to appear at the premises of the appellant to receive the management committee’s verdict. The respondent did not attend the meeting, and the management committee decided at the end

of the meeting to accept the findings of the disciplinary subcommittee and imposed a suspension of 18 months on the respondent. The respondent brought proceedings to set aside the decision of the management committee alleging, among other things, that the proceedings against him had been conducted in breach of natural justice.

92 The Court of Appeal agreed that there had been a “gross breach of the rules of natural justice” (at [67]). Apart from the respondent not having been given proper notice of the case he had to meet, in that the disciplinary subcommittee was “working to a much broader brief than had been revealed to [the respondent]” (at [67]–[70]), the court also expressed concern with the way in which the disciplinary subcommittee had conducted the hearing. In particular, the court observed that the hearing “consisted entirely of a cross-examination … designed to extract admissions” to “preconceived views”; there was no effort to “understand the matter from [the respondent’s] point of view”; and the respondent was “subjected to a humiliating interrogation and personal attack” during which he was “ridiculed and bullied” (at [71]–[73]). In the premises, the opportunity to be heard given to the respondent was “realised more in form than in substance” (at [74]). It is clear from this that the primary focus of the court’s analysis on excessive interference was the respondent’s right to a fair hearing.

93 In Roseli bin Amat, the appellants appealed against their convictions on rape charges not based on the judge’s reasoning but rather on his conduct during the proceedings, which the appellants claimed (a) interfered with their presentation of their defences; and (b) indicated that he had predetermined their guilt before considering the evidence. The former complaint was put along the lines of the fair hearing rule (which encompasses the right to present one’s case),

whereas the latter complaint invoked the rule against bias – specifically, apparent bias in the form of prejudgment. The Court of Appeal rejected the fair hearing complaint but held that the complaint of bias was well-founded. As far as the former was concerned, L P Thean J (as he then was) considered that, even if the trial judge had asked some questions that he ought to have refrained from asking, the conduct of trial was not unfair because the trial judge’s interruptions had not hampered the appellants’ counsel from examining or cross-examining witnesses or presenting their cases (at [10]–[11]). As for the latter, Thean J agreed that several of the trial judge’s comments gave the impression that “at the early stage of the trial [he] had made certain findings or reached certain conclusions on the facts which had yet to be established” (at [13]). In particular, it was wrong for the trial judge to have appeared to have come to a conclusion on the credibility of the complainant based purely on her own evidence, before he had heard all the other evidence adduced by the Prosecution and the Defence (at [20]). Indeed, the trial judge had essentially admitted that he had not kept an open mind throughout the trial when he stated in his judgment that he had been convinced that the complainant was a truthful witness “[a]t the end of her cross- examination” (at [20]). Given the trial judge’s prejudgment of the appellants’ guilt, the Court of Appeal set aside the appellants’ convictions and acquitted them of the charges (at [23]).

94 Haron bin Mundir and Roseli bin Amat demonstrate that a complaint of excessive judicial interference can be analysed along the lines of a breach of either the fair hearing rule or the rule against bias. Roseli bin Amat is especially striking because it was precisely on the conceptual distinction between the two rules of natural justice that the Court of Appeal undertook a nuanced analysis of each complaint and concluded that while one line of attack (based on the fair

hearing rule) was unmeritorious, the other (based on the rule against bias) was made out. Nothing in Shankar Alan casts doubt on this.

95 Analysing these complaints using the framework of natural justice also affords greater clarity in identifying the relevant considerations and the weight to be placed on them. An example of this is the extent to which the judge’s reasons for his or her decision can be invoked as a counterweight to a complaint that the judge had excessively interfered in the proceedings. In In re G (Child) [2015] EWCA Civ 834 (“In re G”), the respondent argued that the fact that the judge had “evaluated critically the evidence that she had heard and arrived at findings which were not a wholesale acceptance of either party’s case” offset the appellant’s complaint that the judge had repeatedly interrupted her counsel’s cross-examination and directed several intemperate remarks at her counsel. The English Court of Appeal rejected this submission (at [52]). Black LJ (as she then was), with whom Sir Colin Rimer and Sullivan LJ agreed, emphasised that the judge’s intervention had “largely prevented [the appellant’s counsel’s] cross- examination from getting off the ground” and “[taken] up a disproportionate amount of the limited time available to [the appellant’s counsel]” (at [47]). It was no answer to this that the judge had produced a cogent analysis of the evidence because “[a] careful and cogently written judgment cannot redeem a hearing in which the judge had intervened to the extent … of prejudicing the exploration of the evidence” (at [52]). By isolating the particular and specific complaint, namely the interference that prejudiced the adducing of the evidence, the court was able to reject, as irrelevant, the argument that the judge had carefully considered such evidence as had been led.

96 The reasoning in In re G was later affirmed by the UK Supreme Court in Serafin v Malkiewicz [2020] 1 WLR 2455 (“Serafin”). Lord Wilson JSC, delivering the unanimous judgment of the court, held that the judge had not allowed the claimant’s case to be properly presented and could not have fairly appraised it given the “barrage of hostility towards the claimant’s case, and towards the claimant himself acting in person, fired by the judge in immoderate, ill-tempered and at times offensive language at many different points during the long hearing” (at [49]). His Lordship rejected the defendants’ reliance on the “quality of the judge’s reserved judgment” as a factor redeeming the proceedings and decision below, because even though the judgment was “on any view a remarkable document” and “intricately constructed and beautifully written”, the claimant’s complaint “[did] not relate to the judge’s judgment and [was] not affected by its ostensible quality” (at [44]).

97 The complaint in In re G and Serafin was that there had been a breach of the fair hearing rule in that a party’s right to present his or her case – specifically, to lead evidence in support of his or her case – was affected by the extent and manner of the judge’s interventions. In both cases, as the complaint was directed at the inputs that would have gone into the formulation of the judgment – namely, the evidence and the parties’ submissions – and as this was found to be prejudiced, it could not be saved by the output, however “remarkable” or impeccable it appeared to be given that that output was itself built on a defective foundation.

98 By the same token, In re G and Serafin do not stand for the proposition that the quality of the judge’s reasoning is never a relevant factor in cases of alleged excessive judicial interference. That would depend on why and how the

interference is alleged to be excessive. Thus, if a complaint of excessive judicial interference is presented as evidence of bias – such as in Roseli bin Amat – a different view might be taken of the relevance of the cogency of the judge’s reasons. Given that the judge’s reasons would be part of the corpus of material that a fair-minded and informed observer would consider when forming a view on the fairness and impartiality of the proceedings (In re AZ (A Child) [2022] 4 WLR 78 at [95]), any appearance of bias arising from the judge’s conduct would have to be weighed along with the seemingly cogent analysis of the parties’ arguments and the evidence in the judge’s reasons.

99 Finally, despite the conceptual distinction between the fair hearing rule and the rule against bias, there will be factors that are relevant regardless of which rule of natural justice the complaint of excessive judicial interference is based on. For example, when assessing whether the judge’s conduct has crossed the line, it would be necessary to weigh the adverse impact of the judge’s interference on a party’s conduct of its case, or any impression of bias, against the reality that the judge might be assisting the parties through his participation.

100 From a fair hearing perspective, a judge may justifiably think it appropriate to give some indication of his or her provisional thoughts or even views, especially where it departs from the parties’ focus or cases, to avoid the possibility that the eventual decision could take the parties by surprise. In DJP, the Court of Appeal noted that the fair hearing rule requires that “an adjudicator must not decide the case on a basis that was neither submitted nor contemplated by the parties without, at least, giving them the opportunity to consider the point and to make their submissions upon it” (at [39]). To this end, a judge should make appropriate interventions.

101 It has similarly been observed that a judge’s expression of views while proceedings are ongoing can be helpful because “counsel are often assisted by the chance to peek into the judicial mind as they then have the opportunity to persuade the court to come to a different view if they so wish” (BOI at [110];

Shankar Alan at [114]; Prometheus Marine at [39]). A judge’s expression of a

preliminary view – even if strongly held – would not by itself establish a complaint of bias, as long as the judge keeps an open mind (Bubbles & Wine

Ltd v Lusha [2018] EWCA Civ 468 at [34]).

Conclusion

102 I summarise some key points from the foregoing analysis in these terms: (a) A judicial decision may be challenged on the basis of a substantive error and/or on procedural grounds. If the particular ground of challenge being considered is that there has been a failure in the process by which the decision was arrived at, the overarching question will typically be whether there has been a breach of natural justice. (b) Whether there has been a breach of natural justice requires a proper appreciation of whether the complaint is based on the fair hearing rule or the rule against bias, and the particular and specific nature of the complaint should be identified as far as possible. (c) The conceptual distinction between the fair hearing rule and the rule against bias should be noted because it brings into focus the source of the alleged procedural unfairness and helps to identify the relevant considerations as well as the weight to be given to them in the analysis.

(d) At the same time, it is important not to adopt an unduly rigid approach to the rules of natural justice. Notwithstanding the conceptual differences, they are not mutually exclusive, and the same set of facts can give rise to a breach of natural justice based on either or both the fair hearing rule and/or the rule against bias.

103 Applying these general principles to the present case, and given that the Prosecution does not level any allegation of bias against the DJ, the question before me is whether there has been a breach of natural justice based on the fair hearing rule, in that the DJ failed to give judicious consideration to the matter before him and instead relied essentially on the case mounted by the Defence.

The applicable threshold for assessing an alleged breach of natural justice

104 I turn to the appropriate test or threshold to be applied. It will be recalled that while the Prosecution draws on the case law on apparent bias and argues for a test of “reasonable suspicion”, the respondents contend that the “reasonable suspicion” test should not be applied outside of the context of bias and it is incumbent on the Prosecution to prove its allegations on the balance of probabilities (see [31], [33]–[34] and [37] above).

105 In my judgment, the answer lies in the rationale of the importance of procedural justice, which is that justice must not only be done but it must be seen to be done so that public confidence in the administration of justice can be maintained. Put simply, not only must procedural justice be done in fact, a

fair-minded and informed observer should be satisfied that it has been done.

The “reasonable suspicion” test, which enforces compliance with the rules of natural justice by focusing on the appearance of things to the fair-minded and

informed observer, meets this purpose. As I will elaborate, there is also no principled basis for confining the “reasonable suspicion” test to complaints of apparent bias or for limiting its application in criminal proceedings to cases where the party alleging a breach of natural justice is the accused person, as the respondents have suggested.

106 In the context of apparent bias, the test of “reasonable suspicion” has been articulated in terms of whether there are circumstances that would give rise to a reasonable suspicion or apprehension of bias in the fair-minded and informed observer (Shankar Alan at [91]; BOI at [103(a)]). This is an objective test that calls for “a hypothetical inquiry into the perspective of the observer and what the observer would think of a particular set of circumstances” (BOI at [103(b)]). The threshold of “suspicion” is lower than proof on the balance of probabilities and is satisfied if the fair-minded and informed observer would think that bias is possible in the circumstances. The qualifier that the suspicion is “reasonable” means no more than that the reasons for harbouring a belief of the possibility of bias can be articulated by reference to the evidence presented, which is assessed objectively (BOI at [103(c)]; Shankar Alan at [48]–[49]).

107 In my judgment, these principles on the “reasonable suspicion” test are equally apposite to challenges based on the fair hearing rule. In Shankar Alan, I explained that the driver behind the “reasonable suspicion” test was “the strong public interest in ensuring public confidence in the administration of justice” (at [75]). There is no sensible reason for concluding that concerns over protecting public confidence in the administration of justice stop at the rule against bias and do not extend to the fair hearing rule. As I observed in Shankar Alan, apart from the direct interests of the parties to the dispute in securing a process that

in fact accords with their legitimate expectations, there is an overriding public

interest that no issue should be allowed to arise with respect to the way in which justice appears to have been administered (at [59]).

108 Behind this emphasis on the appearance of justice is the heralded principle which I have already referred to, namely, that justice should not only be done but it must be seen to be done. This was articulated in the celebrated case of R v Sussex Justices, ex parte McCarthy [1924] 1 KB 256 (“Sussex”). Mr McCarthy had been charged with dangerous driving before the Sussex Justices. The deputy clerk to the justices on the day of the trial turned out to be a partner of the firm of solicitors who were acting for a victim of Mr McCarthy’s alleged dangerous driving in a civil claim against him in relation to the collision that was the subject matter of the charge. At the end of the trial, the deputy clerk retired with the justices, after which the justices returned to give their verdict convicting Mr McCarthy. After learning of the deputy clerk’s employment with the law firm, Mr McCarthy applied to have his conviction quashed.

109 Although the justices deposed to affidavits stating that they had reached their decision to convict Mr McCarthy without consulting the deputy clerk, the court found this irrelevant to the point and resolved that the conviction should be set aside. Lord Hewart CJ, delivering the lead judgment, explained that even if the deputy clerk was actually not involved in the decision-making process, this was not determinative of the issue because courts had to be concerned with the appearance of justice (Sussex at 259):

… a long line of cases shows that it is not merely of some importance but is of fundamental importance that justice should not only be done, but should manifestly and undoubtedly be seen to be done. The question therefore is not whether in this case the deputy clerk made any observation or

offered any criticism which he might not properly have made or offered; the question is whether he was so related to the case in its civil aspect as to be unfit to act as clerk to the justices in the criminal matter. The answer to that question depends not upon what actually was done but upon what might appear to be done. Nothing is to be done which creates even a suspicion that there has been an improper interference with the course of justice. Speaking for myself, I accept the statements contained in the justices’ affidavit, but they show very clearly that the deputy clerk was connected with the case in a capacity which made it right that he should scrupulously abstain from referring to the matter in any way, although he retired with the justices …

Likewise, Lush J, in his concurring judgment, said that “it is irrelevant to inquire whether the clerk did or did not give advice and influence the justices”, as his presence at the justices’ deliberation was objectionable in and of itself (Sussex at 260).

110 Although Sussex involved the rule against bias, there is nothing in the decision which limits the relevance of the appearance of justice to that context. On the contrary, Lord Hewart CJ stated in general terms that “[n]othing is to be done which creates even a suspicion that there has been an improper interference with the course of justice”. In my view, that formulation is broad enough to include interferences with the course of justice stemming from a breach of the fair hearing rule.

111 Elsewhere, the decision of the High Court of Australia in Cesan v The

Queen (2008) 236 CLR 358 also supports the proposition that compliance with

the fair hearing rule should be tested not only from the perspective of what actually occurred, but what might reasonably appear to have occurred. An accused person appealed against his conviction on the ground that there had

been a miscarriage of justice at trial as the judge had fallen asleep on a number of occasions, some of which caused members of the jury to be distracted.

112 The court unanimously allowed the appeal. For present purposes, the reasons of French CJ are especially pertinent as these highlighted the relationship between procedural justice and public confidence in the Judiciary, and emphasised that the maintenance of public confidence required that the parties before the court be given, and be seen to be given, a fair hearing (at [71]):

There are elements of the judicial process which can be said, at least in a metaphorical way, to play a part in maintaining public confidence in the courts irrespective of their relationship to the actual outcome of the process. The appearance of impartiality is one such. … The somewhat elusive criterion of “public confidence” is in some cases, such as the appearance of bias, subsumed in what a fair and reasonable observer would think. The courts nevertheless depend in a real sense upon public confidence in the judicial system to maintain their authority. The maintenance of that authority depends, inter alia upon that element of the judicial process which requires that parties before the court be given and be seen to be given a fair hearing. It is necessary to a fair hearing that the court be attentive to the evidence presented by the parties and to the submissions which they make. The appearance of unfairness in a trial can constitute a “miscarriage of justice” within the ordinary meaning of that term. [emphasis added]

113 That passage encapsulates a number of important points which I will unpack in sequence: (a) First, a court must look beyond the “actual outcome of the process” and concern itself with “public confidence”. This clearly encapsulates the principle that justice must be seen to be done.

(b) Second, the reason for this is that public confidence is critical to the maintenance of the authority of the courts, a point which I have made at the start of this Judgment (see [1] above). (c) Third, the view of the fair-minded and informed observer is taken to be representative of the state of public confidence in and expectation of the Judiciary and its processes. This is consistent with what I considered to be the rationale underpinning the “reasonable suspicion” test in Shankar Alan (see [107] above). (d) Fourth, apart from the appearance of bias, public confidence in the judicial process is equally affected by the appearance of unfairness in the conduct of the hearing.

114 Something of a different view might appear to have been taken in my decision in Shankar Alan, which the respondents rely on in support of their argument that the “reasonable suspicion” test should be confined to the context of apparent bias. In Shankar Alan, I stated, in the context of an allegation of excessive judicial interference, that (at [117(b)]):

… the resolution of such a complaint depends not on appearances or what impressions a fair-minded observer might be left with, but rather on whether the reviewing court is satisfied that the manner in which the challenged tribunal acted was such as to impair its ability to evaluate and weigh the case presented by each side. [emphasis added]

115 On its face, that might be read, as counsel for the respondents have done, to mean that, when considering allegations of excessive judicial interference, the court’s focus should be on actual unfairness rather than on the appearance of unfairness. However, the statement must be understood in its proper context.

The specific complaint that had been made in that case was not that there was an appearance of an unfair trial, but that actual unfairness had resulted from the Disciplinary Committee’s interference during the examination of witnesses (Shankar Alan at [106]). When speaking of the “resolution of the complaint”, it is important to note that the specific complaint before me then was not that the fair hearing rule had been breached based on the appearance of an unfair hearing, but that it had in fact resulted in an unfair hearing. I did not have to consider the possibility of a decision being challenged on the basis of an appearance. I do not therefore regard Shankar Alan as having decided the issue or foreclosed this avenue.

116 Indeed, decisions after Shankar Alan have applied a test based on the appearance of an unfair hearing. In Mohammed Ali bin Johari, the Court of Appeal had occasion to consider the law on allegations of excessive judicial interference and referred approvingly to the following statement of principle in the Ontario Court of Appeal decision in Regina v Valley (1986) 26 CCC (3d) 207 which counsels that focus be trained on the appearance of unfairness rather than actual unfairness (see Mohammed Ali bin Johari at [138]):

… Interventions by the judge creating the appearance of an unfair trial may be of more than one type and the appearance of a fair trial may be destroyed by a combination of different types of intervention. The ultimate question to be answered is not whether the accused was in fact prejudiced by the interventions but whether he might reasonably consider that he had not had a fair trial or whether a reasonably minded person who had been present throughout the trial would consider that the accused had not had a fair trial … [emphasis added in italics and bold italics]

117 After setting out that test, the court went on to apply it to the facts and concluded that the complaint of excessive judicial interference was not made

out because the judge’s conduct did not “give rise to prejudice (or the

appearance of prejudice)” [emphasis added] (see Mohammed Ali bin Johari at

[182]). Mohammed Ali bin Johari was subsequently cited by the High Court in

Nim Minimaart v Management Corporation Strata Title Plan No 1079 [2010] 2

SLR 1 where Steven Chong JC (as he then was) framed the question in terms of “whether ‘a reasonably minded person’ having read the remarks made by the trial judge would consider that there was at least an appearance that the Consent Order was brought about by judicial pressure” original emphasis in italics; emphasis added in bold italics. I emphasise that, in both of these cases, the issue of excessive judicial interference was not raised in the context of alleging apparent bias against the judge but was concerned with whether the complainant’s right to a fair hearing had been prejudiced.

118 I therefore do not accept the respondents’ contention that the “reasonable suspicion” test is confined to the context of apparent bias. There is no logical justification for this. The rule against bias exists alongside the fair hearing rule as the pillars of procedural justice. For public confidence to be maintained, procedural justice must be done and seen to be done in both aspects. At the hearing, when I put to counsel for the respondents, Mr Cavinder Bull SC (“Mr Bull”) and Mr Tan, that public confidence in the administration of justice would surely be affected if a fair-minded and informed observer reasonably suspected that a judge had not conducted the matter fairly, and specifically that he had not applied his mind judiciously to the matter before him, neither Mr Bull nor Mr Tan was able to offer a satisfactory response as to why this was not the case or why a judicial decision made in such circumstances should stand unimpeached. In my view, the point is intuitive to begin with and, quite frankly, unanswerable, when seen in the context of its rationale, and of what is at stake:

the upholding of public confidence in the administration of justice. The fair- minded observer does not have the opportunity to question the judge and to put his reasonable apprehensions to him for these to be addressed. And we cannot tolerate a situation where the observer fairly sees what unfolds and leaves believing that justice may not have been served.

119 This leads me to a related and further point. One aspect of the fair hearing rule, as I have already noted, is that the judge will judiciously apply his or her mind to the material raised at the trial (see [59] above). A failure to do this is a failure to perform the core judicial duty. It seems implausible to me that a reviewing court can only move on this basis if satisfied of this fact on a balance of probabilities, when it will invariably be the case that the judge in question will not be party to the proceedings or have given evidence in his or her own defence.

120 Pek next contends that the “reasonable suspicion” test is inapposite where the complaint of procedural unfairness is levelled by the Prosecution rather than by the accused person. According to Pek, because the general rule in criminal proceedings is that doubts should be resolved in favour of the accused person, the “reasonable suspicion” test is impermissible because it may involve resolving a doubt – as to whether a judge has discharged his or her duty – against the accused.

121 In my judgment, this submission is built on sand, in that it is flawed on at least two levels. First, by conceiving of the “reasonable suspicion” test as a

standard of proof of guilt such that it appears to stand in conflict with the

criminal standard of proof beyond reasonable doubt, Pek fundamentally

misunderstands its nature and what it entails. Although it is true that “suspicion” entails a lower degree of probability than proof on the balance of probabilities (Shankar Alan at [48]–[51]), the “reasonable suspicion” test is not concerned with proof of a fact and is thus not correctly understood as a lower standard of proof than the balance of probabilities (Shankar Alan at [74]). It is not concerned with what has actually occurred, but what appearance has been conveyed to the fair-minded and informed observer. Thus, for example, where it is alleged that there is a reasonable suspicion that the judge is biased, the inquiry on apparent bias is “not directed to an exploration of the actual state of mind of the particular judge” but an “assessment of the possibilities in the context of the objective facts disclosed by the material in evidence” (see Webb

v R (1994) 122 ALR 41 at 62, cited in Shankar Alan at [72]). A finding of

apparent bias does not, by the same token, entail a finding that it is likely that the judge was actually affected by bias (Shankar Alan at [81(c)]).

122 Second, to the extent that the “reasonable suspicion” test can be seen (wrongly) as a standard of proof, Pek’s invocation of the criminal standard of proof as a reason to supplant it with the standard of balance of probabilities is in any event misconceived. As I pointed out to Mr Bull on a number of occasions during his submissions, the criminal standard of proof is relevant when the court is making findings of fact that go towards establishing the guilt of an accused person on the charge(s) against him or her (Public Prosecutor v Mohammed

Liton Mohammed Syeed Mallik [2008] 1 SLR(R) 601 at [34]–[35]). Those are

the findings that must be proved beyond reasonable doubt. In line with this, the Court of Appeal stated in Public Prosecutor v GCK [2020] 1 SLR 486 that the standard of proof beyond reasonable doubt “reduces the risk of convictions arising from factual error” emphasis added. If, however, an acquittal

at first instance is set aside on the basis of a breach of natural justice in the proceedings below, nothing in that entails a finding as to whether the accused is guilty or innocent of the charge(s) against him. Instead, the reversal of the acquittal is simply a byproduct of the finding that the process by which the outcome was obtained is defective, such that the outcome itself cannot stand. The same result would obtain even if the result below was a conviction. It is therefore only where the court proceeds to consider the issue of substantive guilt or innocence, which is not before me at this stage of the appeals, that the requirement of proof beyond reasonable doubt would be the controlling rule.

123 For these reasons, I hold that, irrespective of whether the crux of the complaint of procedural unfairness stems from the fair hearing rule or the rule against bias, the “reasonable suspicion” threshold is applicable.

124 For completeness, I deal briefly with Pay’s suggestion that there is case law supporting a different test of a “reasonable conclusion” (which Pay too interprets as proof on the balance of probabilities). Pay refers in this regard to

Lim Chee Huat where the court stated that the question was whether “a

reasonable person apprised of the circumstances would conclude that the judge did not put her mind to the evidence and the issues” emphasis added, as well as Yap Ah Lai, where I stated my findings on the facts in terms of the fair-minded and informed observer “com[ing] to the conclusion” that the district judge had approached the two cases before him by extracting what he thought were the similarities between them and deciding them as if the issues raised were one and the same (at [70]). It suffices to say that in neither of these cases was the question of the appropriate threshold put in issue, and it would be wrong to attribute to the courts in question an intention that the use of the words

“conclude” and “conclusion” were to carry legal significance as terms of art when there is no indication that this was in fact what was intended.

The appropriate remedial consequences following a finding of breach of natural justice

125 The next question pertains to the consequences which should follow upon the conclusions on the law that I have set out thus far.

126 The Court of Appeal observed in Management Corporation Strata Title

Plan No 301 v Lee Tat Development Pte Ltd [2011] 1 SLR 998 (“Lee Tat (2010)”) at [56] that “not every breach of natural justice is equally serious” and

that “not every procedural wrong entitles the aggrieved party to be reheard and/or to have a final judgment set aside”. In determining the appropriate remedy (if any), the Court of Appeal contrasted breaches of the rule against bias with breaches of the fair hearing rule and opined that a decision tainted by bias

must be set aside and reheard; whereas a more nuanced analysis is warranted

for decisions affected by the fair hearing rule (at [56]–[60]). These passages raise some interesting points and bear reproducing in full:

56 … Of course, not every breach of natural justice is equally serious. For this reason, we agree with the Judge that there are “different grades of procedural wrongs” …, and that not every procedural wrong entitles the aggrieved party to be reheard and/or to have a final judgment set aside. 57 As a matter of principle, where a decision is made in breach of the bias rule, that decision must be set aside and the matters dealt therein reheard. … because if such a judgment is allowed to stand, that may affect public confidence in the Judiciary. 58 However, the same consideration does not apply to a decision made in breach of the hearing rule. …

59 As a matter of principle, a breach of the hearing rule does not entail that the tainted decision must be set aside. Instead, it merely requires that: (a) the aggrieved party be given, in appropriate cases, the opportunity to be heard on the issues on which he was not heard; and (b) the court thereafter considers whether its previous decision was correct and, if it finds that that decision was not correct, either sets aside or rectifies (depending on what the circumstances of the case require) that decision … However, if a hearing on those matters will not change the ultimate outcome of the case (ie, if the aggrieved party will still lose the case even if he is given a hearing on the matters on which he was not heard), then the hearing would be in vain and an exercise in futility. In such a situation, the court would be entitled to exercise its discretion not to grant the aggrieved party a hearing on the matters on which he was not heard … 60 Assuming that it is appropriate for the court to grant the aggrieved party a hearing on the issues on which he was not heard, the court has to go on to consider, after the hearing, whether to uphold, set aside or vary its earlier decision. In this regard, we find the following comments of this court pertinent even though they were made in the context of a breach of the hearing rule in arbitration proceedings (see Soh Beng Tee & Co Pte Ltd v Fairmount Development Pte Ltd [2007] 3 SLR(R) 86 at [91]): If … the same result could or would ultimately have been attained, or if it can be shown that the complainant could not have presented any ground-breaking evidence and/or submissions regardless, the bare fact that the arbitrator might have inadvertently denied one or both parties some technical aspect of a fair hearing would almost invariably be insufficient to set aside the award. [emphasis in original in italics; emphasis added in bold italics]

127 A number of significant points emerge from this extract. As a preliminary point, I agree that not every procedural wrong would necessitate a decision being set aside for breach of natural justice. But I would hesitate to frame the justification for this in terms that “not every breach of natural justice is equally serious” because the court’s subsequent elucidation of the distinction between the fair hearing rule and the rule against bias may give the impression

that a breach of the rule against bias is intrinsically graver. It seems to me that this may be an oversimplification; the true reason why a breach of the rule against bias would lead to the setting aside of the decision almost as a matter of course while a breach of the fair hearing rule would not is not because the former is graver, but because the vice that the fair hearing rule guards against is more amenable to cure in a manner that safeguards public confidence. As I explain below, the amenability of the breach to cure is one of the key factors that should guide the fashioning of the appropriate remedy to correct the breach of natural justice.

128 Indeed, the Court of Appeal was also expressly concerned about the need to uphold “public confidence in the Judiciary” (Lee Tat (2010) at [57]). This, as I will demonstrate, should be the lodestar in guiding the court’s determination of the consequences to follow from a breach of natural justice, regardless of whether the breach concerns the rule against bias or the fair hearing rule. Indeed, the use of public confidence as the touchstone is logical and sound as that is what underlies the “reasonable suspicion” test which applies to allegations of breach of natural justice in court proceedings for the reasons just examined in the previous section at [104]–[128] above.

129 Further, I accept that there are various factors which feed into the analysis of what the appropriate remedy to a breach of natural justice ought to be. Some of these factors were identified by the Court of Appeal in the extract at [126] above. However, to the extent it might be suggested that the Court of Appeal was introducing a threshold requirement of prejudice that applies to arbitration cases to be equally applicable in the judicial context (see Lee Tat

(2010) at [59]–[60]), I disagree.

130 In this section, I first examine the juridical basis of the prejudice requirement in arbitration, before explaining why it would be wrong to conclude that the Court of Appeal imported this into the judicial context. I will then explain why the guiding principle in determining the appropriate remedy in the judicial context should focus on what consequences should follow in order to restore public confidence in the judiciary. In this regard, there are several remedies within the judicial arsenal, and which of these should be deployed depends on several factors that I examine below.

The prejudice threshold in the arbitration context

131 I begin with the requirement that prejudice be shown when seeking relief in arbitration. In that context, the prejudice requirement is the principal safeguard against unconstrained natural justice complaints. And this requirement is rooted in statute. As a result, it is not enough that a breach of natural justice be established in order to set aside an award. The complainant must further demonstrate a real as opposed to fanciful prospect that the breach of natural justice could reasonably have made a difference to the outcome of the arbitration (see L W Infrastructure Pte Ltd v Lim Chin San Contractors Pte Ltd [2013] 1 SLR 125 at [54]; Soh Beng Tee & Co Pte Ltd v Fairmount Development

Pte Ltd [2007] 3 SLR(R) 86 (“Soh Beng Tee”) at [91]).

132 It is important to appreciate the statutory roots of the prejudice requirement in arbitration because of the distinctive nature of arbitration as a mode of private dispute resolution. This stands in contrast to the open nature of justice administered by the courts. As was observed in DJP at [54]:

[One] difference between arbitration and litigation proceedings is that arbitration is a mode of private dispute resolution: the

details of each arbitration and the reasoning contained in each award remain confidential to those who are privy to the arbitration. On the other hand, the judicial process is underpinned by the principle of open justice; court-issued judgments are generally available for public scrutiny, alongside the facts, submissions and evidence material to each dispute. A judgment functions not only to inform parties of why they have won or lost a case, but also to develop the corpus of the law and to assure the public that justice has been served: [Newton] at [40(d)(i)]; Lim Chee Huat at [22]. [emphasis in original]

133 In arbitration, the rules of natural justice exist primarily to safeguard the

private rights of the parties to the dispute. Three key features of arbitration stand

out in this regard: (a) the private and confidential nature of arbitral proceedings (see AAY v AAZ [2011] 1 SLR 1093 and ss 22–23 of the International Arbitration Act 1994 (2020 Rev Ed)); (b) the liberty of the parties to choose their decision-maker by their agreement (see [83] above); and (c) the manifest emphasis on finality of the decision so made (CVV at [1]; COT at [1]). These three features collectively limit scrutiny of arbitration and displace judicial concerns of public accountability. Breaches of natural justice in arbitration, while regrettable, generally do not have broader implications beyond the immediate parties, and as far as the parties are concerned, their rights to recourse are constrained by the need to show prejudice. By contrast, litigation proceedings are conducted in the public sphere where natural justice concerns are predicated not only on protecting the private rights of the parties but also on maintaining public confidence in the administration of justice. The public nature of judicial proceedings means that breaches of natural justice may undermine confidence in the judicial system even where they do not demonstrably prejudice the outcome for the parties. As mentioned above, this dovetails with the rationale of the “reasonable suspicion” test, which is to ensure that public

confidence in the administration of justice is preserved (see [107]–[113] and [118] above).

134 I therefore do not accept the suggestion that the Court of Appeal in Lee

Tat (2010) intended to lay down a broad threshold of prejudice for any natural

justice challenges in the judicial context that is founded on the fair hearing rule. In this regard, the Court of Appeal’s decision not to grant a declaration that it had the power to set aside an earlier decision of its own for breach of natural justice where it would be “pointless” to do so (at [67]), must be seen in context. To properly understand the significance of this holding, it is necessary to briefly set out the background facts in that case.

135 The case stemmed from a protracted dispute between the management corporation of the Grange Heights condominium (“Grange Heights MC”) and Lee Tat Development Pte Ltd (“Lee Tat”) over the right of way over a narrow strip of land (“Servient Land”) which occupiers of the neighbouring lots used to access Grange Road. Grange Heights sits on Lot 687 (which is an amalgamation of Lot 111-34 and Lot 561) while Lee Tat owns the neighbouring Lots 111-32 and 111-33. By conveyances in 1919, a right of way over the Servient Land was granted in favour of Lots 111-32, 111-33 and 111-34 as dominant tenement (but not in favour of Lot 561). The central issue was whether the Grange Heights residents could use this right of way, given that their condominium sat not only on the original dominant tenement (Lot 111-34) but also on Lot 561 which had not been granted any easement rights. The various competing plots of land are best depicted pictorially, with a useful sketch provided in the Annex to Lee Tat Development Pte Ltd v Management

Corporation Strata Title Plan No 301 [2009] 1 SLR(R) 875 (“Lee Tat (2008)”).

136 In 1989, Punch Coomaraswamy J ruled in favour of the residents’ right of way (see MCST Plan No 301 v Lee Tat Development Pte Ltd [1990] 2 SLR(R) 634 at [8]–[11]), and this was upheld by the Court of Appeal in 1992 (see Lee

Tat Development Pte Ltd v MCST Plan No 301 [1992] 3 SLR(R) 1). At the time,

Lee Tat was not yet the owner of the Servient Land.

137 In 2005, the Grange Heights MC sought a declaration that it was entitled as against Lee Tat to repair and/or maintain the right of way. In response, Lee Tat (which by then had become the owner of the Servient Land) sought to relitigate the question of the Grange Heights residents’ right of way. The Court of Appeal (“2005 CA”) held, by a majority, that Lee Tat was estopped by its earlier 1992 decision from relitigating the issue (Lee Tat Development Pte Ltd

v MCST Plan No 301 [2005] 3 SLR(R) 157 (“Lee Tat (2005)”) at [14]–[26]).

138 However, in an unusual turn of events in 2008, a differently constituted Court of Appeal (“2008 CA”) reversed Lee Tat (2005) and held that the Grange Heights’ residents did not have a right of way over the Servient Land. For present purposes, what is relevant is its finding that the 2005 CA erred in applying issue estoppel against Lee Tat. Contrary to Lee Tat (2005), the 2008 CA held that the main question, which was whether the residents of Grange Heights could use the right of way, had not been definitively resolved as against Lee Tat in its capacity as owner of the Servient Land (Lee Tat (2008) at [43(f)]). In the 2008 CA’s view, this was an egregious error which allowed it to overrule the 2005 CA by applying an exception to res judicata established in the decision of the House of Lords in Arnold v National Westminster Bank Plc [1991] 2 AC 93 (“Arnold exception”).

139 Grange Heights MC then applied in SUM 3446/2009 to set aside Lee

Tat (2008) for having been decided in breach of natural justice – because it had

not had the opportunity to address the 2008 CA at all on the Arnold exception. The Court of Appeal in Lee Tat (2010) (“2010 CA”) held that it would be “pointless” to set aside its 2008 decision and grant the Grange Heights MC a hearing on the applicability of the Arnold exception because the 2008 CA’s decision rested on three separate grounds – the Arnold exception being one of them. On this basis, even if the 2008 CA’s reliance on the Arnold exception was discounted, the other two grounds would nonetheless have bound the Grange Heights MC, such that the substantive outcome would have remained unchanged (at [66]–[67]).

140 It was in this context that the 2010 CA referred to the relevance of prejudice that had been recognised in Soh Beng Tee at [91] for arbitration proceedings (see [126] above; Lee Tat (2010) at [60]). To the extent the 2010 CA was correct that the 2008 CA’s decision would be unaffected even if the decision on the Arnold exception was set aside, there was no need to incorporate a further requirement of prejudice. The reference to Soh Beng Tee was, in my view, nothing more than a useful illustration of the broader point that the court would be extremely slow to make a vain or pointless order. The 2010 CA did not address its mind to the policy considerations underlying natural justice in arbitration, which are not identical to those underlying litigation. As I have already noted, in the arbitration context, natural justice is fundamentally concerned with protecting the private rights of the parties to a fair process of hearing. However, in litigation, natural justice serves not only to protect parties’ rights but also to maintain public confidence in the administration of justice. The public nature of litigation places the spotlight squarely on how the breach

of natural justice may impact public confidence in the judicial machinery – and, as I have noted, this may be damaged even if the breach in question does not demonstrably prejudice the parties’ rights.

141 In any case, the 2010 CA’s discussion on the requirement of prejudice in Lee Tat (2010) was not part of its decision. The court there was not concerned with the substantive natural justice challenge itself. Instead, it was merely considering whether to grant certain declarations to the effect that the Court of Appeal had the jurisdiction and power to reopen and correct its previous decisions. The 2010 CA declined to exercise its discretion to grant these declarations because it was of the view that they would “serve no useful purpose” (Lee Tat (2010) at [67]). The prejudice analysis only featured in the context of the court’s jurisdiction to grant the declarations, and not in relation to the determination of the substantive natural justice challenge which, as a result, was rendered moot.

The guiding principle of public confidence in the judicial context

142 If prejudice is not an essential condition before a remedy may be provided in the judicial context, what then is the operative principle? In my judgment, the answer lies in the need to restore public confidence in the judiciary. To this end, there are at least three broad remedial options available to the court. In ascending order of intrusiveness, they are: (a) to address the issue through the normal appellate process, and in such a case, the court might note the breach, but its primary focus will be on the substantive question of whether the outcome reached in the

court below should be interfered with in the exercise of its appellate jurisdiction; (b) at least in criminal matters, remit the tainted matters back to the court below for it to rectify the breach and reconsider the matter; or (c) set aside the defective decision (in whole or in part), and then do one of three things: (i) engage in a de novo review of the (defective portion(s) of the) decision; (ii) order a retrial; or (iii) acquit the accused in a criminal context.

As I emphasised earlier, the choice between these remedies must ultimately be guided by a principled and holistic assessment of what is necessary to restore public confidence in the administration of justice. Without being unduly prescriptive, I briefly canvass each of these options.

(1) Appellate correction 143 The first option represents the least intrusive remedy. It involves the appellate court simply correcting the defect within the boundaries of its usual appellate powers (see Tat Seng Machine Movers Pte Ltd v Orix Leasing

Singapore Ltd [2009] 4 SLR(R) 1101 at [41]). This will most commonly be the

case where the defect in question involves discrete legal errors, such as a judge’s failure to consider an essential legal argument put forth by a party. In essence, such matters are dealt with in the usual way that appeals are.

(2) Remittal 144 Under s 390(1) of the Criminal Procedure Code 2010 (2020 Rev Ed) (“CPC”), an appellate court dealing with a criminal matter may “remit the matter, with the opinion of the appellate court, to the trial court”. Unlike a retrial which typically involves a complete rehearing of the evidence before a different trial judge, remittal is usually a more targeted remedy where specific issues are sent back to the same trial judge for reconsideration (see AOF v Public

Prosecutor [2012] 3 SLR 34 (“AOF”) at [299]). To be sure, this is a power that

is typically resorted to in various circumstances that do not involve an alleged breach of natural justice. But I see no reason to exclude its application in such a context, and our courts have indeed exercised its power to remit in analogous contexts. For instance, in Goh Chin Soon v Public Prosecutor [2021] 2 SLR 308, the Court of Appeal set aside a conviction and remitted the matter back to the lower court (at [87(c)]), after it found that the trial judge wrongly prevented the accused person from calling additional witnesses to respond to certain amended charges (at [51]–[52]). Although this was not couched in the language of natural justice, the conduct in question, in substance, amounted to a breach of natural justice to the extent that the accused was denied the opportunity to fairly present his case.

145 However, for remittal to be an appropriate remedy to address a breach of natural justice, the breach must logically be one that is capable of being cured by the court to which remittal is contemplated. A key factor in this analysis is whether the breach would undermine confidence in the errant judge’s ability to afford a fair process to the parties when reconsidering the remitted issue(s). The Court of Appeal recently expounded on this principle in Vietnam Oil (at [113]– [124]) in the context of remittal in arbitration under Art 34(4) of the Model Law.

Another relevant factor in determining the appropriateness of remittal might be the degree of pervasiveness of the breach. Remittal is usually ordered on a targeted basis, “in a matter which lends itself to a clear and clean outcome in a discrete area of the case” (Imran bin Mohd Arip v Public Prosecutor [2021] 2 SLR 1198 at [91]). Thus, where the breach is pervasive and infects several aspects of the case, remittal may not be an appropriate remedy (see AOF at [302]). And returning to my observation at [127] above, it would seldom, if ever, be appropriate to remit where the complaint is one of any form of bias. The reason for this is that a complaint of bias is targeted at the decision-maker rather than at the process. Whereas concerns over the latter may in some situations be capable of being addressed and overcome, this will not typically be the case in the former.

(3) Setting aside 146 In cases where the breach is pervasive or for any reason is not amenable to being addressed in either of the foregoing ways, the defective decision may have to be set aside in whole or in part, and the reviewing court will then decide what further orders to make. As I have mentioned earlier, there are generally three available options: (a) order a retrial; (b) consider the matter afresh based on the record; or (c) in criminal cases, enter an acquittal.

147 Again, the choice between these options must ultimately be guided by the overarching principle of doing what is necessary to restore public confidence in the Judiciary, subject to any overriding concern of fairness to the accused person. In making this determination, some of the factors identified in the case law concerning criminal matters include: (a) the seriousness of the offence; (b) the expense and length of time required for a new trial, and whether the

testimony of any vulnerable victims may need to be retaken; (c) the availability of evidence given the lapse of time since the offence; (d) the relative strengths of the parties’ cases; and (e) the public interest in ensuring that due process of law is observed (Public Prosecutor v Chua Siew Wai Katheen [2016] 2 SLR 713 at [58]; AOF at [277(d)]).

148 Ler Chun Poh (at [77]), Lim Chee Huat (at [58]), and Lim Tion Choon

(Lin Changchun) v Public Prosecutor [2024] 6 SLR 480 (“Lim Tion Choon”)

(at [48]) illustrate when the reviewing court might consider it appropriate to take it upon itself to reconsider the matter. In these cases, the gravamen of the complaint did not call into question the propriety of the process by which the evidence was adduced. In Ler Chun Poh and Lim Chee Huat, the complaint was concerned with judicial copying. In Lim Tion Choon, the complaint was concerned with the judge’s failure to give sufficient reasons. In all three cases, the breach took place in the course of the judge’s written decision, and there was nothing to suggest that the breach extended beyond that to infect the evidence on the record. There was therefore no compelling need to go through the expense and trouble of a retrial.

149 By contrast, where the breach of natural justice infects the underlying evidence because of the judge’s conduct during the trial, a de novo review based on the record may be inappropriate (see, for example, the cases discussed at [95]–[98] above, including Roseli bin Amat, In re G, and Serafin).

The appropriate analytical framework

150 Pulling these strands together, I adopt the following analytical framework when assessing breaches of natural justice in the judicial context:

(a) first, what is the specific rule of natural justice that is alleged to be breached; (b) second, what are the specific types of complaints that are raised to make good the alleged breach of natural justice; (c) third, on the totality of the material, whether there is a sufficient basis to find that there is a breach of natural justice, which will typically be assessed as a matter of appearances using the lens of the “reasonable suspicion” or “reasonable apprehension” test; and (d) fourth, what, if any, remedy is appropriate to restore public confidence in the administration of justice.

Has there been a breach of natural justice in this case?

151 To recapitulate, the Prosecution’s case comes down to the argument that a fair-minded and informed observer would reasonably apprehend that the DJ deprived it of a fair hearing below by failing to judiciously consider the material before him given the manner and extent of his reliance on the respondents’ written submissions in the GD. To substantiate its case that there has been a breach of the fair hearing rule, the Prosecution points to: (a) the substantial similarities between the respondents’ submissions at trial and the DJ’s GD; and 28

Prosecution’s Appeal Submissions at para 42. 28

(b) the DJ’s alleged failure to deal with several material issues, weigh the Prosecution’s submissions and address the conflicting underlying evidence. 29 152 It is useful, before turning to examine the DJ’s decision, to unpack the two aspects of the Prosecution’s case and consider how they bear on the evaluation of whether there has been a breach of the fair hearing rule.

The nature of the Prosecution’s complaint

153 As explained above, the judicious consideration and evaluation of the facts, evidence and arguments before a judge is a central component of the right to a fair hearing (see [59] and [62] above). I reiterate that the parties’ right to be heard does not stop at the judge physically hearing the evidence or the submissions; it requires the judge to then apply his or her mind to them, adopting a process of reasoned inquiry and deliberative decision-making to reach an outcome that in his or her judgment accords with the merits and justice of the case (see See Kee Oon, Fact-Finding and Reality: A Judicial

Decision-Making Primer (Academy Publishing, 2022) at pp 42 and 49; Mohammad Faizal bin Sabtu v Public Prosecutor [2012] 4 SLR 947 at [27]).

154 As noted above, the duty to consider these issues judiciously is part of the fair hearing rule (see [59] above). It follows that the inquiry is directed at whether there is evidence that gives rise to a reasonable apprehension of a failure on the part of the judge to discharge that duty. Because there will

Prosecution’s Skeletal Submissions at paras 13 and 15–16. 29

typically be no evidence of what precisely the judge did or did not do, it will almost invariably be a case of drawing inferences from the known facts.

155 Evidence of judicial copying from one party’s submissions can give rise to a reasonable suspicion of the unthinking adoption of that party’s position (see [61]–[62] above; see also DJP at [41]), and that the judge had abdicated his or her duty of making a “considered resolution of the controversy at hand” based on a weighing and evaluation of the parties’ cases and the material before the court (Newton at [40(d)(ii)]). In this regard, much will depend on the extent and

manner of the copying.

156 The extent of copying has both a quantitative and a qualitative dimension. As a matter of common sense, the greater the degree of reliance on the material supplied by only one party, the greater the risk that the judge will in fact have, or appear to have, failed to do justice to the other party’s case (see

IG Markets Ltd v Crinion [2013] EWCA Civ 587 (“Crinion”) at [16]). Apart

from considering how much has been copied, it will also be necessary to consider what has been copied. As the judge’s primary function is to decide areas of controversy between the parties, a judge’s substantial adoption of a party’s summary of undisputed background facts or the reproduction of a party’s submission to provide the context of that party’s case will typically give no cause for concern. Contrariwise, where a judge fails to evaluate the merits of both parties’ submissions on an issue that is heavily disputed and instead adopts one party’s submissions on the point without evident critical evaluation, it may more easily suggest the inference that he or she may have abdicated the responsibility of dealing with the parties’ arguments even-handedly.

157 The manner in which the copying has been done may also be probative. This could require assessing whether the copying extended to reproducing errors from the party’s submissions, which may give rise to the apprehension that the judge was indiscriminate in his or her copying and did not exercise any judicious consideration of the material (see, for example, Lim Chee Huat at [52] and Ler Chun Poh at [30]). Of course, the nature of the reproduced error is also relevant. Where the error is one of substance – for example, an inaccurate statement of fact or of the evidence – it is more likely to suggest that the judge did not apply a judicious mind (Cojocaru at [57]–[58]). Another related consideration is whether the judgment contains internal contradictions and inconsistencies which are not explained. Thus, the Hong Kong Court of Final Appeal in Nina Kung v Wang Din Shin (2005) 8 HKCFAR 387 (“Nina Kung”) noted that the first instance judge had made inconsistent findings and “inexplicably reversed himself”, which gave rise to a legitimate concern that the judge had “reproduced the copied material without giving any genuine thought to the issue at hand”, because if he had given the issue independent consideration, “he would surely not have made findings wholly incompatible with considered positions he had previously taken unless he was able to articulate grounds for changing his stance” (at [453]–[454]).

158 It will also typically be instructive to look beyond the evidence of judicial copying and consider, for example, what the judge has or has not addressed. This corresponds with the second plank of the Prosecution’s complaint, which is that the DJ failed to deal with material issues, significant aspects of its case and conflicting evidence.

159 Where a judge has copied the submissions of one party and ignored or failed to deal with the opposing party’s case – and even more where this is an important aspect of that party’s case – without at least explaining the omission, this would often fortify the apprehension of a breach of the fair hearing rule. I digress to observe that one of the core purposes of the judicial duty to give reasons is that the parties, especially the losing party, should be left in no doubt as to why the judge reached the decision he or she did, regardless of whether that decision is right or wrong on the merits (Flannery v Halifax Estate Agencies

Ltd [2000] 1 WLR 377 at 381). It will usually be more important for a judge to

focus on the reasons why the losing party’s case is being rejected (GLAS SAS

(London Branch) v European Topsoho SARL [2025] 1 WLR 5343 at [21], [29]

and [32(d)]). The reason for this was well explained extrajudicially by Sir Robert Megarry in Robert Megarry, “Temptations of the Bench” (1978) 16

Alberta Law Review 406 at 410:

… Sometimes I ask students to say whom they consider to be the most important person in a court room. … My answer, given unhesitatingly, is that it is the litigant who is going to lose. Naturally he will usually not know this until the case is at an end. But when the end comes, will he go away feeling that he has had a fair run and a full hearing? Some litigants, of course, are so unreasonable that nothing will satisfy them, even if they win. But take the reasonable defeated litigant (you will all have known many of these), and see whether he feels that he has had a fair crack of the whip. One of the important duties of the courts is to send away defeated litigants who feel no justifiable sense of injustice in the judicial process. [emphasis added]

160 I reiterate that the assessment should be made holistically and with common sense, because it is well-established that a judge need not address every point raised by the parties; what is necessary is that decisive or material issues ought to be considered and dealt with (Lim Chee Huat at [21]; Crinion at [37];

Cojocaru at [70]). Natural justice does not require that the parties be given

responses on every submission made (SEF Construction Pte Ltd v Skoy

Connected Pte Ltd [2010] 1 SLR 733 at [60]; Haide Building Materials Co Ltd v Ship Recycling Investments Inc [2024] SGHC 222 (“Haide”) at [36]–[37]). A

judge is not expected to address a point if it is devoid of merit or self-evidently bad and may deal with a point implicitly by making findings that render it moot (Lim Chee Huat at [23]; Haide at [38]–[39]).

161 Nor does it follow from the mere fact that the judge has referred to both parties’ arguments that the parties’ right to be heard has been given effect to

meaningfully. Thus, in BZV v BZW [2022] 3 SLR 447, the High Court placed

no weight on “general and self-serving” statements in an award that the tribunal “ha[d] considered in detail” the material and arguments and “[a]ny omission to specifically mention any of the arguments advanced by the Parties or any submission, document or correspondence d[id] not suggest that these ha[d] not been considered” in its analysis as to whether the tribunal had breached the fair hearing rule, as the question of breach was “a matter of substance, not form” (at [127]–[129]).

162 Finally, it will always be wise for the reviewing court to take a step back to examine the quality of the judge’s decision as a whole. An abjectly poor judgment may buttress the argument that the judge has not applied his or her mind, as would an egregious or wholly inexplicable error. The focus here is not on the substantive correctness of the decision but whether the quality of the decision suggests a reasonable apprehension that the judge failed to consider the matter judiciously to such a degree that a party may be said to have been deprived of a fair hearing (BZW at [56]). The focus is not on the existence of the error but whether the error is of a nature that it points to a more fundamental

defect in the process by which the decision was arrived at (Swire Shipping at

[116]–[120]).

163 These considerations are not meant to be exhaustive. Nor can they be applied in a mechanistic way. Ultimately, the relevant considerations that give rise to a reasonable suspicion that a judge has failed to give the parties a fair hearing by applying a judicious mind to the material before him or her will depend on the specific circumstances of each case, on how the complaint in question is framed, and on the sensible assessment of the reviewing court. On this basis, I turn to the facts before me.

Whether the DJ failed to apply his mind to the matters

164 In my judgment, a fair-minded and informed observer would harbour a reasonable suspicion or apprehension that the DJ failed to judiciously consider the material before him. I broadly agree with the Prosecution that this reasonable suspicion arises in light of (a) the extent and manner of the copying of the respondents’ written submissions; and (b) the DJ’s failure to deal with significant aspects of the Prosecution’s case and the evidence before the court (see [151] above).

Overarching issues with the GD

165 To begin with, it is indisputable that significant sections of the GD were copied from the respondents’ written submissions. The Prosecution has documented examples of this in a non-exhaustive list of 18 paragraphs where the GD reproduces the reasoning, structure, wording and stylistic choices of the

respondents’ submissions. I agree with the Prosecution that the DJ appears to 30 have substantially reproduced the respondents’ submissions as his own analysis in these paragraphs. A close review of the GD discloses even more instances of judicial copying. In this regard, I have set out a non-exhaustive list of the similarities between the DJ’s reasoning and the respondents’ submissions in Annex 1 to this Judgment. Although I will also consider this point in greater detail when I examine the specific sections of the GD below, I provide a few particularly troubling examples here:

GD at [30(a)(v)]: [630]–[632]: [IO Lim]’s approach in how he dealt … As Mr Pay testified, [IO Tay]’s approach to amendments on his 1 raises serious concerns. According to st different from [IO Lim]’s. With [IO Tay] amendments to his 1 CPIB st transparent than the process he went Statement, and [IO Tay] explained through with [IO Lim] for the second that Mr Pay could sign next to these. invited to make amendments directly marked-up his 2 CPIB nd Statement with changes that he intended to make, [IO Lim] would next to these amendments, ensuring another room, re-attributable to Pay … In contrast, [IO statement in his own words, and with Pay’s suggested amendments Lim] return it to him. Smugly, [IO Lim] appears to have done this on the Pay’s testimony, the process he CPIB Statement was entirely pretext of ‘neatness’. … experienced with IO Tay, who When Pay marked up his second recorded his first CPIB statement, , Mr Pay was invited to make That is an explanation as convenient was markedly different and more as it

With IO Tay, Pay was Prosecution’s Appeal Submissions at para 42 and Annex A. 30On the other hand, when Mr Pay on his first CPIB statement, and IO Tay explained that Pay could sign

84 take Mr Pay’s statement to that the changes were clear and type the

handled Pay’s amendments to the second CPIB statement in a much less transparent manner. statement with the changes he is contrived. Looking at the intended to make, [IO Lim] took

retyped it in his own words (which appears similar to PP v Dahalan bin worn-out Mr Pay, as [IO Lim]’s method of amending Mr Pay’s 2 Ladaewa where the recording officer nd had used words that were not uttered CPIB Statement meant had no indication whatsoever how or by the accused), and then returned where [IO Lim] had taken in Mr Pay’s it to Pay. [IO Lim] amendments to his 2 CPIB this process was done for the sake nd of “neatness” … Statement … [IO Lim] had integrated his changes, italics]making it impossible for Pay to verify accurately reflected transparency could prevent Pay from changes he made from those made by [IO Lim] …

italics]

GD at [30(a)(viii)]: [598], [637] and [664]: [IO Lim] had a “blatant disregard for the truth” when recording Pay’s [598]: … The most key ingredient in second CPIB statement. He provided Mr Pay’s 2 CPIB Statement was [IO nd Lim]’s blatant disregard for the truth, and explanations to an exhausted Mr Pay words recorded in his statement. His approach was as to the meaning of the words he Mr Pay’s 2 CPIB Statement. … nd the statement to another room, whole picture, [IO Lim]’s intention [637]: … We submit that [IO Lim]’s was clearly to muddle an already incriminating 2 ndomitting exculpatory evidence. His that Mr Pay selective and false explanations to an exhausted costs”, regardless of the actual truth claimed that Pay regarding the meaning of the his deception in providing false or fairness of the process, by costs, with little regard to the truth. Pay was left mischievous and without any clear indication of how conveniently selective as he was recording and agreeing to in selectively recorded only those parts of Pay’s statements that whether his amendments were 85 could be construed as . This lack of approach to the taking of Mr Pay’s while ignoring or CPIB Statement was identifying or distinguishing the mischievous, conveniently actions appeared to be driven by a intended purely to desire to secure a conviction “at all secure the convictions by extracting an incriminating statement at all engaging in underhanded

been exacerbated by the fact that Mr Pay was exhausted, contending with in a manner that was difficult for Pay to notice or contest, especially Foo’s true nature, and having to deal after [IO Lim] with [IO Lim] explanations about any ambiguities in the statement. … amendments to Mr Pay’s 2 CPIB nd for Mr Pay to track, italics] his bone-weary state. [emphasis italics]

166 In these examples, the DJ was setting out what was presented in the respondents’ submissions as his conclusions. In doing so, he reproduced the submissions even to the extent of adopting the adversarial tone in parts of the respondents’ submissions. For instance, lifting from the respondents’ 31 submissions, the DJ found that the relevant CPIB officers had conducted the statement-recording process in a way that was “mischievous and conveniently selective”, “driven by a desire to secure a conviction ‘at all costs’” and/or that they had a “blatant disregard for the truth” (see GD at [30(a)(viii)]). Such a tone and choice of expression, even if some might view it as excessively subjective, might not be remarkable in a party’s advocacy for its position. But in a judicial decision, where findings are made and expressed, there is an expectation that the court will carefully weigh its choice of words, conscious of their adverse, even dire, effects (see Newton at [40(d)(ii)]). A court is expected also to be mindful of the need to be sure that its conclusions have been put to those

practices, such as making [664]: These would naturally have

“surreptitious or ambiguous

amendments” to Pay’s statement Prosecution’s Skeletal Submissions at para 2. 31 the various revelations about Mr provided misleading ’s underhanded practice of making surreptitious 86

Statement that were impossible especially in

affected by them, that their responses have been considered, and the reasons for rejecting those responses are clearly explained.

167 This has particular relevance in the present case because the substance of some of the allegations levelled by the respondents against the relevant CPIB officers (and echoed by the DJ in the GD) either impute or insinuate bad faith and dishonesty on their part, which the CPIB officers strenuously denied in their evidence. Apart from the findings set out at [166] above, the DJ also found that IO Lim had “deliberate[ly] attempt[ed] to obscure the process and potentially manipulate the content of the statement” by retyping Pay’s 2nd Statement after Pay had made amendments to it (see GD at [30(a)(v)]) and that, motivated by a desire to secure a conviction at all costs, he had “[engaged] in underhanded practices” by providing misleading explanations of the meaning of the words recorded in Pay’s 2nd Statement to Pay (see GD at [30(a)(viii)]). By dint of these acts, the DJ found that there were defects with IO Lim’s statement- recording process that rendered Pay’s 2nd Statement unreliable (see GD at [30(c)]). As for IO Tan, the DJ concluded that his statement-recording process was motivated by the primary aim of creating material that would support his preconceived belief in Pek’s culpability by, among other things, selectively recording details and embellishing and inserting his own words into Pek’s statements (see GD at [30(d)(v)] and [30(d)(xiv)]). IO Tan’s conduct thus undermined the accuracy and reliability of both of Pek’s CPIB statements (see GD at [30(d)(xvi)]–[30(d)(xvii)]).

168 It is trite that allegations of dishonesty are treated with particular caution and require cogent proof given their serious nature (see, for example, Tang Yoke

Kheng v Lek Benedict [2005] 3 SLR(R) 263 (“Tang Yoke Kheng”) at [14],

explaining that in the context of the civil standard of proof, the more serious the allegation, the more the party bearing the burden of proof would have to do to establish his case on a balance of probabilities; Parti Liyani v Public Prosecutor [2021] 5 SLR 860 at [128]–[131], applying Tang Yoke Kheng to allegations of improper conduct against the Prosecution). It is no small thing for a judge to make findings against the CPIB officers in the terms that the DJ did, which could not only have consequences on the officers themselves, but also occasion public disquiet against the CPIB. The DJ was of course entitled to make these findings if he considered them justified on the evidence. But, in the face of the kind of allegations made by the respondents, it was incumbent on the DJ to demonstrably have examined the evidence and dealt with the contrasting accounts and explanations of those involved. As will be seen below, however, this was just not done.

169 The reasonable apprehension that the DJ may not have judiciously considered the material, which begins with the reproduction of the respondents’ written submissions in the GD, is then strengthened by significant lapses in the DJ’s reasoning, in particular, his failure to address material aspects of the Prosecution’s case. To recapitulate, the Prosecution’s primary contention was that the Loans had been given with corrupt intent. The Prosecution sought to demonstrate this with three main points: (a) the presumption under s 8 of the PCA that the Loans had been given corruptly as an inducement or reward; (b) the purportedly incriminating statements given by the respondents to the CPIB that demonstrated their corrupt intent; and (c) the other corroborative objective evidence (see [18] above). However, an analysis of the DJ’s reasoning reveals the lack of any proper consideration of, much less engagement with, substantial parts of these arguments. I mention two points in particular.

170 First, there is an omission of any evident consideration of the presumption of corruption under s 8 of the PCA in the DJ’s analysis on conviction. Section 8 of the PCA presumes, upon proof of the predicate fact 32 that gratification has been given to a person employed by the Government or a public body by a person who has or seeks to have dealings with the Government or any public body, that the gratification had been given as an inducement or reward with the requisite corrupt intent. If the presumption applies, the burden is on the accused person to prove the contrary. The DJ’s failure to squarely address this point is conspicuous because he had, earlier in the proceedings, dismissed the respondents’ submission that there was no case to answer after the close of the Prosecution’s case, having held that the presumption applied to both Pay and Pek. On this basis, the burden was on the respondents to prove on a balance of probabilities that the Loans were not paid or given corruptly as an inducement or reward (see, for example, Tey Tsun Hang v Public Prosecutor [2014] 2 SLR 1189 (“Tey Tsun Hang”) at [29]; Teo Keng Chuan v Public

Prosecutor [2001] SGHC 49 at [12]). Indeed, the DJ expressly recorded in his

reasons for his decision that the Prosecution had established a case the respondents had to answer, that this was because the respondents had to rebut

the presumption under s 8 of the PCA (GD at [19] and [22]):

19 In any case, as noted by the Prosecution, section 8 of the PCA clearly applies to Pay and the presumption is that the gratification in the form of the loans that Pay gave to Foo were corrupt and he has to rebut the presumption and hence, his defence must be called. … 22 … contrary to Pek’s submissions, the presumption in s 8 of the PCA applies to Pek and he must rebut the presumption and his defence must be called. …

Prosecution’s Skeletal Submissions at para 14(a). 32

171 In the circumstances, one would have expected the presumption under s 8 of the PCA to have been the starting point in the DJ’s analysis in the GD. This was a point of importance for two reasons. First, because there was no doubt that the Loans were made, the only remaining issue was whether they were made by the respondents with a corrupt intent. Since s 8 of the PCA operated as a presumption of the requisite mens rea (see Tey Tsun Hang at [27]), it ought to have been a central feature of the DJ’s discussion on conviction. And second, the presumption also affected the standard of proof on this issue because if the presumption were applicable, the Defence had the burden of proving on a balance of probabilities that they had not acted corruptly (see, for instance,

Jumaat bin Mohamed Sayed v Attorney-General [2025] 1 SLR 1287 at [41]–

[51] for a detailed discussion on the nature and operation of statutory presumptions). However, the DJ did not mention or address the presumption. In fact, he did not even state that he had found that the respondents had discharged their burden of proving that the Loans were not made with a corrupt intent.

172 At the hearing, Mr Bull directed me to the DJ’s statement that the consistency between Foo’s evidence and the respondents’ evidence was such “as to prove that there is no corrupt element and no corrupt intention in the [Loans]” emphasis added. Mr Bull suggested that the DJ’s use of the word “prove” meant that he was cognisant that the burden of proof lay on the respondents in the light of s 8 of the PCA. I disagree. In the first place, it is telling that the only reference in the entire GD that Mr Bull could locate to support an inference that the DJ had applied his mind to the presumption under s 8 of the PCA was a single word in a single sentence, which made no mention of the presumption at all. It seems to me that Mr Bull was placing more weight

on the single word than it could reasonably bear. But aside from this, if the DJ was at this point expressing a conclusion in relation to the rebuttal of the presumption, then it defies logic or comprehension that the DJ did not consider the objective evidence that the Prosecution relied on, and which the DJ had waved away as irrelevant (see [172] below).

173 In my view, the DJ’s omission to apply (or even mention) the presumption in his analysis on conviction is inexplicable, and it calls to mind the error of the trial judge in Nina Kung who was described as having “inexplicably reversed himself” (see [157] above). What is material is the existence of a significant internal inconsistency within a judgment, rather than its particular manifestation, which may vary from one case to the next. Where this is the case, especially when taken with other considerations, such as a significant degree of judicial copying, it may weigh in the assessment of whether it might reasonably be apprehended that the DJ did not judiciously apply his mind to the matter because such a fundamental error seems implausible if he had done so (see [162] above).

174 Second, the DJ’s analysis proceeded on the basis that the three issues he had identified were dispositive of the case, and it was thus unnecessary to consider any of the other issues that had arisen from the trial. The DJ prefaced his analysis by asserting as follows (GD at [28]):

Having carefully read and considered the voluminous submissions of the parties at the close of the trial comprising more than 1500 pages, I took the view that there are 3 key issues arising from the trial for my determination before I can decide on whether Pay and Pek should be convicted or acquitted of their respective charges. …

175 However, even a cursory survey of the table of contents of the parties’ written submissions before the DJ would reveal that there were several other areas of dispute which the DJ did not address. The DJ did not explain or elaborate on how he had concluded that these other issues were not relevant and did not need to be addressed, or how exactly he determined that the three issues he had identified were all that mattered. For example, the Prosecution’s case below was that the Loans had been advanced as an inducement for Foo to further the interests of TSC in its dealings with the LTA in the T220 Project. To 33 establish this, apart from the presumption under s 8 of the PCA, the Prosecution relied on the combination of (a) purportedly incriminating portions of the respondents’ CPIB statements; and (b) the objective evidence from which 34 inferences of corrupt intent could be drawn, including the commercial 35 significance of the T220 Project to TSC, TSC’s allegedly unsatisfactory 36 performance on the T220 Project, and the process by which the Loans were 37 procured and given to Foo. In these circumstances, it would have been 38 expected that even if the DJ concluded that the respondents’ CPIB statements were unreliable or did not support the Prosecution’s case, he would nonetheless have addressed – even if only briefly – whether the objective evidence relied on by the Prosecution independently established that the Loans had been extended

Prosecution’s Appeal Submissions at para 33; DC Prosecution at para 11 (ROP 8990). 33 DC Prosecution at paras 14–17 (ROP 8991–8993). 34 DC Prosecution at para 70 (ROP 9025). 35 DC Prosecution at paras 71–74 (ROP 9025–9026). 36 DC Prosecution at paras 75–81 (ROP 9026–9031). 37 DC Prosecution at paras 126–127 (ROP 9057–9058); paras 131–140 (ROP 9059–38

9066)

with a corrupt intent. After all, the Prosecution clarified during oral closing 39 submissions that their case was founded on the totality of the surrounding circumstances, and not merely the CPIB statements: 40

DPP LOH: … The last point I want your Honour to remember is our case is not that the statements are the only pieces of evidence or source of evidence against Mr Pay and Mr Pek. They are the main pieces of evidence, but there are two points. … The first point is that when your Honour looks at the statements and you see that they have admitted to a corrupt purpose and then when your Honour steps back and you look at the surrounding circumstances, there are plenty of facts that corroborate what they said in their statements. So it is not correct to say our case stands and falls only by the statements. We are saying the statements present our main evidence but when you look at all the surrounding circumstances, it corroborates. [emphasis added]

176 However, the DJ did not address the objective evidence and simply dismissed it in a single sentence at the end of the GD after resolving the three issues he had identified in the respondents’ favour (at [50]):

As I took the view that my decisions on the above 3 key issues arising from the trial were sufficient for me to find that the Prosecution has failed to prove its case against Pay and Pek beyond a reasonable doubt, it was unnecessary for me to decide on other issues arising from the trial, such as whether there is objective evidence to corroborate the Prosecution’s case that the reason behind the Loans was corrupt, or show that the Loans to Foo were intended to advance TSC’s interests, as well as various other disputed issues of law and facts as framed by Pek’s counsel. [emphasis added]

177 While a judge is not obliged to address every point put forward by the parties (see [160] above), unless the reason for a significant omission speaks for itself or can reasonably be inferred from the circumstances, a judge should

Prosecution’s Skeletal Submissions at para 14(b). 39 NE (19 Aug 2024) at p 142:3–20 (ROP 5468). 40

generally explain why he or she has chosen to disregard it. In this case, it is not evident why the DJ considered that he did not have to address the objective evidence relied on by the Prosecution, given that it was one of the three main planks of the Prosecution’s case (alongside the presumption and the respondents’ CPIB statements). Nor does the DJ’s assertion that he had “carefully read and considered the voluminous submissions of the parties at the close of the trial comprising more than 1500 pages” dispel the reasonable impression that the DJ had not in fact applied his mind judiciously to the issues in the case (see GD at [28]).

Issue 1: Reliability of the respondents’ CPIB statements

178 Against the backdrop of these overarching observations, I turn to each of the issues. In my judgment, the DJ’s analysis of Issue 1, which pertains to the reliability of the respondents’ CPIB statements, raises the strongest suspicion that the DJ failed to apply his mind judiciously to the matters before him. Significantly, the DJ’s finding on this issue underpinned the rest of his analysis on whether the respondents possessed the requisite mens rea (see [219] below).

(1) Substantial degree of copying 179 To begin, as between the three issues addressed in the GD, the analysis of Issue 1 discloses the greatest degree of similarity to the respondents’ submissions. By my estimate, well over half of the DJ’s analysis for this issue consisted of substantially the wholesale adoption of the respondents’ submissions (see Annex 1 to this Judgment). To illustrate this, I set out a comparison of two portions of the GD and the respondents’ submissions below:

GD at [30(a)(vii)]: [615]: [IO Lim] As previously submitted, [IO Lim] investigation, particularly in how he recording of Mr Pay’s 2 CPIB nd Statement with a certain tunnel-statement. He had a “tunnel-vision” approach, focusing solely on evidence or aspects of the case that suggested Pay’s guilt. aspects of the case that he believed indicated there was “more than meets the eye,” implying that these aspects suggested Pay’s guilt. However, at the aspects of the investigation that [IO Lim] elements of the investigation that could have potentially indicated Pay’s relieving Mr Pay of guilt. Clearly, [IO innocence. [IO Lim] himself Lim] acknowledged there were aspects of objectivity. demonstrated a lack of the investigation that could have been [emphasis objectivity during the significant in potentially relieving Pay appeared to conduct his entire of guilt, yet he chose not to consider italics] recorded Pay’s second CPIB them. In my view, [IO Lim]’s conduct during the investigation, vision, insisting on one hand that characterized by selective focus and there were various aspects of Mr a lack of impartiality, shows a clear Pay’s case where there was “more He emphasized certain lack of objectivity which could have than meets the eye” which would suggest Mr Pay’s guilt, but on the other hand ignoring other crucial into Pay’s case and undermines the integrity of the investigation process. same time, he ignored other critical himself accepted would have been important as potentially italics] was completely lacking in GD at [30(d)(v)]: [588]: By [IO Tan’s] drafted Pek’s first statement with the intention to “frame” Pek, focusing shows that [IO Tan] had approached

affected the fairness and thoroughness of the investigation own admission, he The evidence before the Court also

an undue focus on Pek’s culpability. In particular, [IO Tan]: establishing the facts of the case. [IO (a) drafted Pek’s 1 Statement in stTan] crafted Pek’s first statement impressions about Pek, including about Pek, such as portraying him originator of a corrupt scheme; scheme. He (b) recording matters that supported balanced and factual account. Additionally, [IO Tan’s] approach of the matter; and (c) events memory of events. reliability of Pek’s first statement as italics] said during the interview.

italics]

180 The comparison between the GD and the respondents’ submissions reveals clear parallels in both the structure of the reasoning and the analysis of the evidence. Other parts of the DJ’s reasoning on Issue 1, as set out in Annex 1 to this Judgment, reveal a similar degree and manner of reproducing the substance of the respondents’ submissions. This suggests that the DJ had transplanted the respondents’ submissions and presented them as his own almost exclusively on recording the statement recording process with

information that supported Pek’s analysis and conclusions after (seemingly intentionally) incorporating some culpability rather than objectively

cosmetic and inconsequential editorial changes.

a way which created certain to create specific impressions by trying to portray Pek as the as the originator of a corrupt selectively recorded focused almost exclusively on details that would imply Pek’s guilt, instead of providing a Pek’s culpability rather than objectively establishing the facts appears to have been aimed at 96 influencing Pek’s memory of attempted to influence Pek’s , further undermining the an accurate reflection of what was

(2) The DJ’s failure to address points raised by the Prosecution 181 The DJ also failed to address several aspects of the Prosecution’s submissions on Issue 1, namely the reliability of the respondents’ CPIB statements, which was a heavily contested point concerning a critical piece of evidence which the Prosecution itself referred to as “the cornerstone of the Prosecution’s case”. A court considering that issue would have been expected 41 to undertake careful consideration and evaluation of the parties’ arguments and the evidence before it. Taken together with the very significant degree of copying, the DJ’s failure to analyse the key aspects of the Prosecution’s case strengthens the reasonable apprehension that he failed to judiciously consider the material before him in his determination of Issue 1.

182 The Prosecution’s submissions on the reliability of the respondents’ CPIB statements broadly comprised four main points: (a) the alleged deficiencies in the statement-recording process were contradicted by IO Tan’s and IO Lim’s evidence, which should be preferred because they were credible witnesses who had no reason to lie; (b) the respondents’ own explanations for 42 having provided the incriminating statements, such as their belief that they were “victims” of Foo’s indiscretions, being exhausted, or having a poor 43 44

DC Prosecution at para 14 (ROP 8991). 41 Prosecution’s Reply Submissions at the Close of Trial dated 31 July 2024 (“DC 42 Prosecution’s Reply”) at paras 26–39 (ROP 9443–9456). DC Prosecution at paras 24–35 (ROP 8997–9005); DC Prosecution’s Reply at para 21 43 (ROP 9442). DC Prosecution at paras 37(a) and 40 (ROP 9006–9007). 44

command of the English language, were undermined by the other evidence; 45 46 (c) the consistency between the respondents’ CPIB statements added to their reliability; and (d) there was compliance with the statutorily prescribed 47 procedural safeguards for statement recording. The DJ’s analysis neglected to 48 deal with significant parts of these arguments. I take each point in turn.

183 First, the DJ failed to consider the CPIB officers’ evidence that the statement-recording process had been conducted properly. Where witnesses present conflicting accounts, a judge would typically be required to assess the credibility of each witness based on the internal and external consistency of the witness’s evidence and the witness’s demeanour and come to a reasoned view on which account is more believable. This is all the more so when the conflict relates to serious allegations, such as those made by the respondents against the CPIB officers, as noted at [168] above. The safeguard of fairness that the rule in Browne v Dunn (1893) 6 R 67 upholds, by requiring a party to put it to a witness if the party intends to submit that the witness’s evidence should be disbelieved (see, for example, Hong Leong Singapore Finance Ltd v United

Overseas Bank Ltd [2007] 1 SLR(R) 292 at [42]), is seemingly seriously

compromised if the judge does not appear to have considered the explanation or conflicting account that the witness may have offered before making an adverse finding against that witness. The DJ’s analysis suggests a singular consideration

DC Prosecution at paras 63–69 (ROP 9020–9025). 45 DC Prosecution’s Reply at paras 20–25 (ROP 9442–9443). 46 DC Prosecution’s Reply at paras 41–49 (ROP 9456–9463). 47 DC Prosecution at paras 22–23 (ROP 8996–8997). 48

of the respondents’ version of events, in the face of contrary accounts by the CPIB officers. I address each of the respondents’ cases separately.

184 I turn first to Pay’s case, which rested primarily on his contention that IO Lim had retyped Pay’s amendments to his 2nd Statement in different words than those used by Pay before returning the statement to him, leaving Pay in doubt as to whether and how his amendments had been incorporated. 49 According to Pay, IO Lim’s practice of retyping Pay’s amendments led Pay to be unable to identify whether the changes had been made by himself or by IO Lim, and this enabled IO Lim to make surreptitious amendments to Pay’s 50 2nd Statement that portrayed Pay in an incriminating manner. However, this 51 account was categorically rejected as “ridiculous” by IO Lim, who gave evidence that Pay did not make any amendments to his 2nd Statement and that IO Lim did not retype the statement. Despite the clear conflict in the evidence 52 of Pay and IO Lim on the same issue, the DJ held, lifting extensively from Pay’s submissions, that while “[IO Lim] claimed that this process was done for the 53 sake of ‘neatness’”, it appeared that it was “instead a deliberate attempt to obscure the process and potentially manipulate the content of the statement” (GD at [30(a)(v)]). The DJ did not acknowledge that IO Lim had given a contrasting account of events, much less explain why he preferred Pay’s account over IO Lim’s. Additionally, the concern that the DJ had unthinkingly adopted

NE (7 March 2024) at pp 119:10–120:11 (ROP 3322–3323). 49 DC Pay at paras 632–634 (ROP 17193–17195). 50 DC Pay at para 664 (ROP 17210). 51 NE (14 June 2022) at p 130:1–16 (ROP 780); Prosecution’s Appeal Submissions at 52 para 123. DC Pay at paras 630–632 (ROP 17192–17194). 53

Pay’s submissions without reviewing the underlying evidence is strengthened because the DJ appears not to have realised that the claim that IO Lim had retyped Pay’s 2nd Statement for “neatness” was made by Pay (conveying what IO Lim had supposedly told him), and not by IO Lim, who denied Pay’s claim 54 altogether.

185 Pek challenged the reliability of his 1st Statement on account of IO Tan’s failure to record the statement objectively and his purported intention to emphasise Pek’s culpability. This was said to be demonstrated by, among other things, IO Tan’s admission during cross-examination that he had attempted to “frame” Pek and that he drafted the statement in a particular manner to “paint” Pek as devising the idea to give the staff loans to Pay, who would then give the Loans to Foo. The DJ willingly accepted Pek’s 55 contentions as to the alleged shortcomings in the statement-recording process and substantially reproduced extracts of Pek’s closing submissions in his decision (GD at [30(d)(iv)]–[30(d)(v)], and [30(d)(xiv)]). In so doing, the DJ 56 omitted to consider or address IO Tan’s clarification during re-examination that he had not tried to frame Pek and had only recorded the facts stated by Pek. 57 Nor did the DJ address the Prosecution’s contention that IO Tan’s earlier apparent admission was the result of him not having fully comprehended the

NE (7 March 2024) at p 58:6–17 (ROP 3261); NE (21 March 2024) at pp 41:18–42:8 54 (ROP 3975–3976). DC Pek at paras 588–590 (ROP 13500–13501); DC Pek at para 614 (ROP 13514). 55 DC Pek at para 622 (ROP 13517); DC Pek at para 588 (ROP 13500); DC Pek at para 56 497 (ROP 13448). NE (3 June 2022) at p 61:12–21 (ROP 441); NE (3 June 2022) at p 63:15–22 (ROP 57 443); Prosecution’s Skeletal Submissions at para 18(g).

questions that had been posed to him by Mr Bull. Similarly, the DJ’s reliance 58 on IO Tan’s purported admission that he had recorded parts of Pek’s 1st Statement in a particular manner in order to “paint” Pek as someone who intentionally aided the commission of corruption (GD at [30(d)(iv)]–[30(d)(v)]) appears to be an adoption of Pek’s submissions, whilst failing to mention or 59 engage with IO Tan’s clarification in re-examination that he “must have heard the question wrongly” and “missed out the word ‘paint’”, and the 60 Prosecution’s explanation that IO Tan had not appreciated the negative undertones of the word “paint” that was used by Mr Bull in cross-examination. 61 186 Second, the DJ’s analysis of the respondents’ explanations for their incriminating statements also suggests that he may have been perfunctory in his treatment of the Prosecution’s submissions. I illustrate this by reference to two examples. One basis for the DJ’s finding that Pek’s 1st Statement was unreliable was that the statement was “interwoven with the recording officer’s wording choices, which Pek may not have fully understood or noticed” (GD at [30(d)(xiii)]). This again appears to be a reproduction of Pek’s closing submissions. Moreover, it appears to be premised in part on Pek’s submissions 62 that his poor comprehension of the English language “severely handicapped his ability to review and correct the inaccuracies in the CPIB statements that [IO Tan] had crafted”. This entirely disregarded the Prosecution’s contrary 63

DC Prosecution’s Reply at paras 34–35 (ROP 9447–9453). 58 DC Pek at para 622 (ROP 13517); DC Pek at para 588 (ROP 13500). 59 NE (3 June 2022) at p 60:2–5 (ROP 440). 60 DC Prosecution’s Reply at para 37 (ROP 9454–9455). 61 DC Pek at para 539 (ROP 13472). 62 DC Pek at para 539 (ROP 13472). 63

submission that Pek had a sufficiently good command of the English language to understand the contents of his 1st Statement based on his educational background, employment and leadership appointments as well as his apparent familiarity with the English language during trial. 64 187 Indeed, based on my review of the trial transcripts, it is indisputable that the issue of Pek’s competency in the English language arose repeatedly when he was cross-examined on the contents of his CPIB statements. The Prosecution made it clear beyond peradventure that it was disputing Pek’s claim that his difficulties with the English language led to his inability to comprehend the incriminating aspects of his 1st Statement. For example, when asked by the lead counsel for the Prosecution, Deputy Public Prosecutor Alan Loh (“DPP Loh”), to clarify if he had reviewed, amended and signed his 1st Statement, Pek confirmed that he had done so, but that he had difficulties understanding the statement’s contents as it was not “plain and simple English for [him]”: 65

1. Nevertheless, even though you are now saying all these
   things, do you agree that you did review the statement, did make amendments and sign it?

2. I did review the statement, I did make amendments.
   But, again, I’m not able to –

3. Wait, wait. And signed, you signed off on it?

4. And I signed. But I’m not able to read them thoroughly,
   to totally able to process it, to understand what the interpretation of this statement is how you would interpret it right now, your Honour.

5. Mr Pek, this statement is actually in plain English.
   Unfortunately it is, to me, it is plain English. It is not a very complicated document in English or a literary work,

DC Prosecution at para 63 (ROP 9020). 64 NE (17 April 2024) at p 119:10–120:5 (ROP 4563–4564). 65

it is just you describing a conversation or conversations that you had with [Pay] and then things that happened. That’s all it is.

1. Your Honour, although this is a plain English to you, it may not be a plain and simple English for me.

Other instances where DPP Loh crossed swords with Pek on the latter’s fluency in the English language included: (a) DPP Loh’s suggestion to Pek that it was “incredible” that he could not understand certain aspects of his statement due to his Chinese background, when he graduated from a university in the United Kingdom with First Class Honours; (b) DPP Loh’s reference to text messages 66 between Pek and Pay as evidence that Pek “ha[d] no problems communicating in English”; and (c) a line of questioning that was aimed at casting doubt on 67 Pek’s claim that he had difficulties with the English language based on his credentials and success as a businessman. After what appears to have been an 68 exasperated comment by Pek that DPP Loh “[did not] understand the pain of people … who have language ability issue [sic]”, DPP Loh put the 69 Prosecution’s case to Pek that he had no problems in reviewing his statement as he had reviewed minutes of meetings in the course of his day-to-day work. It 70 is troubling, to say the least, given how deeply the Prosecution pressed Pek on this issue, that the DJ paid no heed to its case that Pek had understood the contents of his 1st Statement.

NE (18 April 2024) at p 87:11–24 (ROP 4659). 66 NE (18 April 2024) at p 88:20–24 (ROP 4660). 67 NE (18 April 2024) at pp 89:19–93:25 (ROP 4661–4665). 68 NE (19 April 2024) at p 7:1–13 (ROP 4721). 69 NE (19 April 2024) at p 12:1–7 (ROP 4726). 70

188 Another example relates to the DJ’s finding that IO Lim “provided false explanations to an exhausted Pay regarding the meaning of the words recorded in his [2nd] statement” (GD at [30(a)(viii)]). This is reproduced almost entirely from Pay’s closing submissions. More importantly, in making this finding, the 71 DJ appears to have accepted Pay’s submission that he was “drained and troubled” during the taking of his 2nd Statement because he had been detained overnight at the CPIB and had only been offered the table, chair or floor of the interview room to rest on. The difficulty with this is that whether this had 72 occurred and whether Pay was “exhausted” during the taking of his 2nd Statement were disputed issues. This is clear from the Prosecution’s contrary submission that Pay was in a good condition when his 2nd Statement was being recorded as he had been given proper meals, rest breaks and sufficient time to rest at night, and the Prosecution’s cross-examination of Pay, in the course of 73 which the Prosecution had repeatedly put their case to Pay that he was feeling well enough when his 2nd Statement was recorded and that he had not indicated otherwise to the CPIB officers: 74

1. Now, I put it to you that although you stayed overnight
   at CPIB you were sufficiently alert when you were making your second statement; do you agree?

2. I disagree. It was more than 30 hours.

3. Now, in fact in this statement there is – at paragraph 30
   it is stated that: “I wish to state that I am feeling well. I was feeling well and I am able to continue to assist in investigations.” Do you see that inside the statement?

DC Pay at para 598 (ROP 17178). 71 DC Pay at paras 584–586 (ROP 17172–17174). 72 DC Prosecution at para 40 (ROP 9006). 73 NE (21 March 2024) at pp 30:19–32:3 (ROP 3964–3966). 74

1. This was written by Chris Lim. I wasn’t sick and he want
   to continue. Yes, I am able to continue, but I was very tired.

2. You didn’t tell the IO that you were unable to continue
   to record the statement; correct?

3. Do I have the right to tell the IO said I want to stop and
   don’t – I really don’t know what is my right. I did not tell him. I thought when he want to continue I have to continue. He is an investigating officer.

4. In fact, you didn’t make any complaint in the course of
   your second statement; correct?

5. What kind of complaints, your Honour?

6. You didn’t say that you were not feeling well; right?

7. Yes, I wasn’t not feeling well so I didn’t say I wasn’t
   feeling well.

8. So you were feeling well enough to record the statement;
   right?

9. Like I said, I was very tired. I couldn’t really focus. I was
   eager to go home and I continued with the --- to assist the investigation, yes. …

10. I put it to you that Johnston, IO Johnston did not tell
    you that you could sleep on the chair, the table or the floor; do you agree or disagree?

11. I totally disagree. He said that and that is very clear
    because I would remember that and he was very sarcastic.

The absence of any explanation for the DJ’s finding that Pay was “exhausted” despite the Prosecution’s contrary position, and of any consideration of whether this was borne out on the evidence, is again troubling in that it fuels the apprehension that the DJ may have adopted Pay’s submissions indiscriminately.

189 Third, the DJ completely failed to mention or address the Prosecution’s submission that the existence of material commonalities between Pay’s and

Pek’s statements spoke to the accuracy and truthfulness of their contents. If 75 such commonalities existed, this might have been a compelling point in favour of the Prosecution’s case that the respondents’ CPIB statements were reliable. The DJ’s silence on this suggests a failure to consider and engage with the Prosecution’s arguments. 76 190 Lastly, the DJ disposed of the Prosecution’s submission that the respondents’ CPIB statements were recorded in compliance with the procedural safeguards by just echoing the respondents’ submissions. There is a marked failure to address the contrary position adopted by the Prosecution. According to the Prosecution, it followed from the fact that the respondents were allowed to read through their statements, make amendments to their statements, and eventually signed off on their statements confirming them to be “true and correct”, that the respondents were satisfied with the contents of their 77 statements. In response, Pay contended that IO Lim’s adherence to the 78 procedural safeguards were “meaningless in the face of [IO Lim’s] trickery” and his willingness to “stretch the truth”, which was said to be demonstrated 79 by IO Lim retyping Pay’s 2nd Statement after Pay had made amendments to it 80 and providing misleading explanations of the meaning of certain words in the

DC Prosecution’s Reply at paras 41–44 (ROP 9456–9461). 75 Prosecution’s Skeletal Submissions at para 15(f). 76 DC Prosecution at para 22 (ROP 8996). 77 DC Prosecution at para 23 (ROP 8997). 78 Pay Teow Heng’s Reply Closing Submissions dated 31 July 2024 (“DC Pay Reply”) 79 at para 62 (ROP 17751). DC Pay at paras 631–634 (ROP 17193–17195). 80

statement to Pay. The DJ completely reproduced Pay’s arguments as his own 81 findings:

GD at [30(a)(x)]: Pay’s Reply Closing Submissions at [62]: As submitted by Pay’s counsel, the According to the Prosecution, all of meaningless in the face of [IO Lim]’s conduct based on his own evidence him to be happy to stretch the truth and the Prosecution has offered no explanations for his conduct. statement with a pen, the Prosecution [emphasis added in italics] believe the evidence of [IO Lim] unfortunately has been demonstrated

inconsistencies in [IO Lim]’s evidence. [emphasis in original omitted; emphasis added in italics]

191 The DJ had made these findings in the face of IO Lim’s categorical denial that Pay had made amendments to his 2nd Statement, that he had retyped Pay’s 2nd Statement (see [184] above), or that he was not interested in the truth during the statement-recording process: 82

1. The fact of the matter, [IO Lim], is that you were not interested in the truth. You were putting general propositions to Mr Pay, you were amending his changes by incorporating it in words that you want, and by the procedural safeguards were

the procedural safeguards in respect

DC Pay Reply at paras 64–68 (ROP 17752–17754); DC Pay at paras 598 (ROP 17178) 81of Mr Pay’s CPIB Statements were

and 702 (ROP 17229–17230). which unfortunately demonstrated adhered to. In alleging that Mr Pay was given the chance to make NE (14 June 2022) at p 148:17–25 (ROP 798). 82 amendments to his 2nd CPIB is asking this Honourable Court to , who 107

to be happy to stretch the truth. In doing so, the Prosecution has offered no explanations for any of the

time he was supposed to sign he was just exhausted. Would you not agree? You were not interested in the truth? …

1. It is not true. 192 The DJ’s failure to address the Prosecution’s arguments is also clear from his analysis of Pek’s case. Again, the DJ substantially adopted Pek’s submissions that he had signed his 1st Statement without a full understanding of its contents due to his trust in IO Tan and that an accused person’s ability to review a statement was an insufficient indicator of the statement’s reliability:

GD at [30(d)(viii)]: Pek’s [675]: Pek’s first interview was conducted in a way that was intended to create a When all of Pek’s 1 Interview was conducted in st a manner designed to induce Pek into reviewing his first statement. This approach affected Pek, leading him to a false sense of security and to avert an earnest review of Pek’s 1 st Statement by Pek. These factors had [IO Tan] review of Pek’s 1 Statement, trusted professionalism st [IO Tan], and took [IO Tan’s] Closing Submissions at their conversation. As a result, Pek professionalism as an article of faith. Pek ended up signing a statement the above facts are false sense of security for Pek, viewed in totality, it is clear that potenti and without realising there would discouraging him from thoroughly repercussions. … personally. … rush through the review of his italics] Pek placed his trust in italics] an effect on Pek, who rushed his , relying on his perceived and assuming that the statement accurately reflected signed the statement without 108 fully understanding its that he evidently did not fully implications or realizing the understand the ramifications of, al personal be repercussions on him

GD at [30(d)(xiii)]: [539]: … In the premises, from [IO Tan’s] embellishments which are significant because they alter the true meaning of what Pek intended to convey.

correct the statement and fix the inaccuraciesdocument accurate or reliable, presently where

wording choices, not have fully understood or noticed. that [IO Tan] had crafted Pek’s statements did not fai what Pek had in fact said to [IO Tan] The statement contained on the subject's review to ensure the at the CPIB interviews. Instead, numerous inaccuracies that stem an inaccurate accuracy of a statement. In my view, statement as drafted up by the own phrasing and into [IO Tan’s] recording officer for an accused reliability based on the fact that embellishments cannot be taken as reliable failed to spot and correct. simply because the accused was given the chance to review the Simply allowing Pek the italics] opportunity to review and sign statement and signed it, if the accused lacked the ability to the statement does not automatically render the the recording officer. … . This is the case particularly when the Pek’s language italics] disability severely handicapped his discrepancies are interwoven ability to review and correct the with the recording officer's inaccuracies in the CPIB statements which Pek may The DJ did not engage with the Prosecution’s counterargument that Pek . As admitted that he was able to go through the draft of his 1st Statement twice, 83explained below, Pek’s CPIB failure to detect and correct these inaccuracies highlights the thfully reflect

problematic nature of relying solely

DC Prosecution at para 23 (ROP 8997); NE (26 March 2021) at p 8:1–8 (ROP 4199). 83

numerous inaccuracies were woven it is not enough to presume phrases and , which Pek plainly an accused person reviewed and 109 signed the statement, especially when the accused may not have been fully aware of the nuanced changes or errors introduced by

and that Pek had in fact made amendments to parts of his 1st Statement and countersigned against them. 84 193 Viewed holistically, the DJ’s analysis of Issue 1 reveals extensive reproduction and adoption of the respondents’ submissions, invariably framed in partisan terms, coupled with a failure to address the material points raised by the Prosecution. There was little, if any, assessment or weighing of the arguments made.

194 Against this, Mr Bull contends that the DJ did consider the Prosecution’s arguments but dismissed them because the Prosecution did not provide specific responses to or otherwise address the concerns that had been raised in relation to IO Lim’s or IO Tan’s conduct of the statement-recording process (see GD at [30(a)] and [30(d)]). In effect, Mr Bull contends that the DJ did not refer to the Prosecution’s submissions in his analysis of Issue 1 because the Prosecution failed to proffer any meaningful arguments on Issue 1.

195 I do not accept this. As I have noted above, the Prosecution advanced several lines of argument in support of its contentions as to the reliability of the statements. As far as the case on what actually transpired during the statement- taking process is concerned, there were contested assertions. It cannot seriously be suggested that each of the points made by the Prosecution which I have outlined above that the DJ did not address were so bereft of merit that the DJ did not need to engage with any of them. On the contrary, it seems to me that a considerable portion of the Prosecution’s arguments were of sufficient relevance and substance that a judge considering the matter would have had to

DC Prosecution at paras 55–57 (ROP 9016). 84

grapple with them. This was especially so since the DJ considered that the Prosecution had established a case for the respondents to answer. Instead, the DJ did not demonstrate any attempt to engage with various aspects of the Prosecution’s case, and simply concluded that Pay’s and Pek’s CPIB statements were unreliable due to an “absence of specific explanations from the Prosecution”, in almost identical terms for both of the respondents (GD at [30(c)] and [30(f)]):

30(c) Hence, I found that [IO Lim]’s aforesaid conduct rendered Pay’s second CPIB statement inaccurate and unreliable as they are clearly incompatible with the “uncompromising need for accuracy and reliability” as mandated by the Court of Appeal in Kadar (at [60]) with many inconsistencies and ambiguities in the aforesaid statement relied upon in the Prosecution’s case, and it would be improper for this Court to resolve those doubts in the Prosecution’s favour, particularly in the absence of specific explanations from the Prosecution on [IO Lim]’s aforesaid conduct in the statement recording process. … 30(f) Hence, I found that [IO Tan]’s aforesaid conduct rendered Pek’s CPIB statements, particularly his first statement, inaccurate and unreliable as they are clearly incompatible with the “uncompromising need for accuracy and reliability” as mandated by the Court of Appeal in Kadar (at [60]) with many inconsistencies and ambiguities in the aforesaid statement relied upon in the Prosecution’s case, and it would be improper for this Court to resolve those doubts in the Prosecution’s favour, particularly in the absence of specific explanations from the Prosecution on [IO Tan]’s aforesaid conduct in the statement recording process.

196 Given that there were substantial aspects of the Prosecution’s case which the DJ did not so much as acknowledge, let alone engage with, it is not clear what the DJ meant in saying that there was an “absence of specific explanations

from the Prosecution”. Even if the DJ’s point was that the Prosecution failed to provide a point-by-point riposte to every aspect of the CPIB officers’ conduct which the respondents took issue with, the fact is that the Prosecution did lead evidence disputing the respondents’ account of the statement-taking process and advanced legal arguments on why the alleged lapses in the CPIB officers’ conduct did not prejudice the reliability of the statements’ contents, and these should have been taken into account and evaluated alongside the respondents’ attack on the integrity of the statement-taking process. The DJ may well have considered, as Mr Bull suggests, that the evidence relied on by the Prosecution was not persuasive and that their legal submissions were misconceived. But if that was his view, he certainly did not say so.

197 In my judgment, having regard to these factors, I am satisfied that the informed and fair-minded observer would reasonably apprehend that the DJ had not discharged his obligation to judiciously consider and assess the evidence and arguments before him.

(3) Other factors do not negate the appearance of a failure to judiciously consider the matter 198 The respondents invite me to consider three other factors which they say evidence the DJ’s application of a judicious mind to the evidence and the parties’ cases: (a) the DJ’s allegedly independent observations in the GD; (b) the DJ’s paraphrasing and reorganisation of the parties’ submissions; and (c) the DJ’s engagement with the Prosecution’s case during oral closing submissions. However, none of these suffice to dispel the reasonable suspicion that the DJ failed to judiciously consider the matters before him. I explain.

(A) PURPORTED INDEPENDENT ANALYSIS BY THE DJ 199 According to the respondents, the DJ included some independent analysis and observations in the GD. Pek suggests that the DJ formulated his own legal test, namely, “whether the incriminating CPIB statements of Pay and Pek are inaccurate, unreliable and unsafe to rely on by the Prosecution to prove its case against Pay and Pek” [emphasis in original omitted; emphasis added]. 85 I cannot see this as persuasive in any way since there is nothing novel or original in examining the accuracy or reliability of witness statements. The courts have consistently applied these metrics in assessing the appropriate weight to be placed on witness statements (see, for example, Muhammad bin Kadar v Public

Prosecutor [2011] 3 SLR 1205 at [55]–[56], [60] and [130]). In any event, it

can hardly be said that the DJ intended or saw himself as crafting a legal test. It is clear from [30] of the GD that in framing the issue this way, the DJ was adopting or responding to the respondents’ submissions which, on examination, did feature the terms “inaccurate”, “unreliable” and “unsafe”. 86 200 There are also other portions of the GD which the respondents say reflect the DJ’s independent analysis. However, a careful examination of the GD 87 shows that the DJ’s analysis in most of these paragraphs was, in fact, taken from the respondents’ submissions. For example, contrary to Pek’s submissions, 88

Pek Lian Guan’s Written Submissions dated 6 May 2025 (“Pek’s Appeal 85 Submissions”) at para 699. See, for example, DC Pek at paras 6–7 (ROP 13068–13069). 86 Pek’s Skeletal Submissions at paras 41–48, 59 and 62; Pay’s Skeletal Submissions at 87 para 41(a); Pay Teow Heng’s Written Submissions dated 6 May 2025 (“Pay’s Appeal Submissions”) at para 722. Pek’s Skeletal Submissions at paras 59 and 62. 88

[30(a)(ix)]–[30(a)(x)], [30(d)(x)] and [30(f)] of the GD were materially copied from the respondents’ closing submissions. I provide a few examples:

GD at [30(a)(ix)]: Pay’s Reply Closing Submissions at [60]: As pointed out by Pay’s counsel, the officers, [IO Kan] and [IO Lim], had no frame either of them”, so [IO Kan] evidence should be given more weight and [IO Lim] than Pay’s is an oversimplification and an attempt to sidestep addressing the over-simplification, and an attempt possibility that the CPIB officers as their by the Prosecution to avoid having witnesses may have been dishonest in to engage with the evidence of their investigation and court testimony. The Court. Most importantly, it Prosecution’s stance overlooks a more proposition need be true – while situation. The truth does not Prosecution’s stance that the CPIB According to the Prosecution, the necessarily require that [IO Kan] CPIB officers had “no reason to and [IO Lim] Pay, the truth can be that [IO reason to frame Pay, and that their intention to frame Pay. Rather, the Kan] and [IO Lim] were biased ’s evidence should be and overzealous investigators, preferred over Mr Pay. This is an hypothesis of ill intent on the basis intending to frame Pay, were biased or, as the Court of Appeal in Kadar of the loans having been given (as was submitted in Mr Pay’s Closing their testimonies or biased in their own witnesses’ dishonesty in Submissions at Section XII). investigation. This overzealousness could explain the officers’ conduct [emphasis ignores the fact that neither nuanced understanding of the intent to frame anyone. [IO Kan] and italics] they may not have had any [IO Lim] could have developed a biased specific intention to ‘frame’ Mr had a specific and formed a hypothesis that the loans indicated ill intent or wrongdoing, … possibility exists that these who had formed their own officers, though not explicitly italics]

put it, “overzealous” in their

without requiring a deliberate 114

perspective during their investigation

GD at [30(d)(x)]: [678]: Two significant weaknesses in the approach adopted by [IO Tan] in recording a statement needed to be enforcement officer, [IO Tan]

The first point of failure is [IO recording to 2 critical points of Tan] failure: (a) [IO Tan] during the interview. … The second Pek’s ability to comprehend the point of failure is Pek’s ability to inaccuracies. In contrast, if [IO Tan] had any intention of taking an inaccuracies. This assumes that Pek is fully aware of what was discussed, understands the statement’s wording, Even as a trained law laptop or writing materials into the interview room to take notes during highlighted despite him being a adopted an approach that comprehension issues or stress during Pek’s 1 Interview itself. st trained law enforcement officer. subjected the statement the interview can impact Pek’s ability [IO Tan] himself conceded that a to effectively person ’s dependence on his own ’s ability to statement. A more effective and reliable memory to recall what Pek said recall what Pek said; and (b) method for ensuring the accuracy and person wished to be accurate. [IO Tan] was unable to provide any statement and correct any would have been for [IO Tan] comprehend the statement as notes directly during the interview. This recorded and to correct any accurate and comprehensive capture of Pek statement, it would have been a reliance on memory. Even [IO Tan] simple matter for him to bring a and is capable of identifying and why this was his practice. … correcting mistakes. However, is a priority. However, [IO Tan]’s [emphasis In fact, simply that it was “not [his] practice.” italics] would usually be review and correct the This is an insufficient explanation, as it expected to take notes if that does not address why this practice is in place, nor does it provide a rationale comprehensiveness of the statement explanation for why this was not to take done. All he could offer was his repeated refrain that it was “not would allow for a more accurate [his] practice” to take notes before ’s words and reduce drafting the statement. No explanation was provided as to admitted that taking notes is 115 generally expected when accuracy justification for not taking notes was

for deviating from what would be

documentation. …

italics]

201 Where the observations were not reproduced from the respondent’s submissions, they often either repeated the points which had earlier been copied from the respondents’ submissions or were generic and impressionistic statements that added nothing to the assessment of the evidence before the DJ. Pay, for instance, relies on certain remarks made in the GD at [30(a)(iv)], that IO Lim’s approach to the recording of Pay’s 2nd Statement raised “serious questions about the fairness and thoroughness of the investigation” due to IO Lim’s “selective attention to facts that support a preconceived narrative, while neglecting those that might be in Pay’s favour”, and thus “IO [Lim] may have been biased toward [sic] confirming Pay’s guilt rather than objectively assessing all the relevant information”, to show that the DJ had independently evaluated the evidence. However, these observations merely paraphrased 89 earlier points that were lifted from the respondents’ submissions, such as IO Lim having a “preconceived notion that Pay had committed some form of offence” (see GD at [30(a)(i)], reproduced at Annex 1 to this judgment).

202 In a similar vein, Pek submits that the last few lines of [30(d)(xiv)] of the GD evidence the DJ’s own analysis because, notwithstanding that Pek did not suggest bias on IO Tan’s part, the DJ “felt more strongly about it” and concluded that IO Tan’s conduct “demonstrates a bias that undermines the

Pay’s Skeletal Submissions at para 41(a); Pay’s Appeal Submissions at para 722. 89

considered a more reliable method of

reliability and credibility of the statements”. It may be that the DJ’s conclusion 90 here went further than what had been submitted by Pek below, but the evidential and factual analysis on which this conclusion was founded is extensively sourced from Pek’s closing submissions:

GD at [30(d)(xiv)]: Pek’s [497]:

undermined by the way in which it CPIB statements is magnified by the was drafted. fact that statement was prosecutorial expediency by for prosecutorial convenience rather whom Pek guilelessly trusted. [IO Tan] himself admitted during cross- what Pek actually said. This issue is examination that he tried to “frame” exacerbated by Pek’s trust in [IO Pek as [IO Tan] drafted Pek’s 1 st Tan] Statement. [IO Tan] also embellished drafting Pek’s 1 Statement. Instead During cross-examination, [IO Tan] st of objectively enquiring into the facts “frame” Pek when drafting Pek’s first of the matter, [IO Tan] Closing Submissions at find ways to emphasise Pek’s statement The probative value of Pek’s first suggests an intention to portray Pek CPIB statement is severely The probative deficiency of Pek’s in a manner that could imply guilt 2 Statement. nd that [IO Tan] did not approach the it was patently drafted for It appears that the construed primarily an IO interview. Furthermore, [IO Tan’s] with factual accuracy as his priority. approach included embellishing and Instead, [IO Tan’s] primary focus was than as an accurate reflection of statements, deviating from the actual belief in Pek’s culpability. … , whom he believed was responses. recording his statements faithfully. and added his own words when uncover the facts objectively, [IO Tan’s] admitted that he attempted to continued to . This choice of words culpability even when drafting Pek’s Pek’s Skeletal Submissions at paras 41–42. 90 The evidence shows rather than providing an impartial and objective account of the recording of Pek’s CPIB statements

inserting his own words into Pek’s on crafting material that could be used to buttress a predetermined language and context of Pek’s Instead of seeking to actions reveal a continued effort to highlight Pek’s culpability,

statement. This consistent focus on

credibility of the statements. …

The paragraph therefore does not displace the reasonable apprehension that the DJ did not judiciously consider and weigh the evidence before him.

203 Finally, the respondents also point to several case references in the GD as evidence of the DJ’s exercise of independent judgment. According to Pay, the DJ cited three cases that were not mentioned in Pay’s closing submissions, indicating the DJ’s exercise of independent judgment. Pek asserts that the DJ’s comparison of the instant case to Public Prosecutor v Dahalan bin Ladaewa [1995] 2 SLR(R) 124 (“Dahalan”) demonstrates that the DJ had applied a judicial mind to the facts and evidence before him. 91 204 I disagree. Of the three cases Pay refers to, two can be found in, and were clearly lifted from, Pek’s reply closing submissions. As for the DJ’s 92 reference to the remaining case of Leck Kim Koon v Public Prosecutor [2022] 2 SLR 595 and his analysis of Dahalan, these do not displace the impression

Pek’s Skeletal Submissions at paras 37–40; Pek’s Appeal Submissions at paras 682–91

684.

For Chong Hoon Cheong v Public Prosecutor [2022] 2 SLR 778, see Pek Lian Guan’s 92 Reply Submissions dated 31 July 2024 (“DC Pek’s Reply”) at 40 against the GD at [29(d)]. For Parti Liyani v Public Prosecutor [2020] SGHC 187, see DC Pek’s Reply at [38]– 39 against the GD at [29(c)].

even when drafting Pek’s second 118

proving Pek’s guilt, rather than accurately recording the facts, demonstrates a bias that undermines the reliability and

that arises from his extensive lifting of the respondents’ submissions and his failure to engage with the Prosecution’s case that he had not judiciously applied his mind to the matter. A judge’s application of his or her mind to the relevant legal principles, whether it be part of the judge’s pre-existing corpus of knowledge or through reading the authorities cited by the parties, is only part of the judge’s duty to apply his or her mind to the matter before him or her. The judge is required, in the exercise of the judge’s duty of adjudication of the

dispute before him or her, to consider the evidence that has been led and then

determine how the existing legal principles apply to the facts of the instant case. Put simply, the fact that a judge has read the relevant cases and apprised himself or herself of the legal principles set out therein may not be sufficient to discharge the judge’s duty of adjudication, because, as I have already noted, everything needs to be considered in the round.

(B) THE DJ’S REORGANISATION OF THE RESPONDENTS’ TRIAL SUBMISSIONS 205 Similarly, Pek suggests that the DJ’s consideration of the evidence and issues before him is evidenced by the fact that the DJ reorganised the respondents’ submissions and adopted “discrete and non-consecutive paragraphs” of Pek’s closing submissions. I first observe that this reflects a 93 mechanistic approach to the analysis, which is something I have eschewed.

206 A judge’s reorganisation of a party’s submissions might indicate that the judge synthesised the parties’ arguments and it might, in some circumstances, allay suspicions that the judge had failed to apply his or her mind. However, the crux of the inquiry is whether the DJ has accorded the parties their right to a fair

Pek’s Skeletal Submissions at paras 29–32. 93

hearing. Where the central issue is whether the judge considered, engaged with and weighed the relative strengths of the parties’ cases and the evidence, the fact that the judge may have reassembled significant portions of one party’s submissions into the judgment may not suffice to displace a reasonable apprehension that the judge failed to examine the evidence and issues judiciously. Much will turn on the nature of the issues before the judge. But where, as is the case here, the issues concern contentious disputes of fact, it is less likely that a judge will be found to have satisfactorily engaged with all the material just by paraphrasing one party’s narrative.

207 In Newton, in considering whether there was a reasonable apprehension of prejudgment amounting to apparent bias, I did place some weight on the fact that the district judge had reorganised the Prosecution’s written submissions as evidence that the district judge had digested the material and considered how these influenced his views of the issues (at [48]). However, this observation cannot be considered in isolation. The crux of the analysis there turned largely on the district judge’s engagement with the submissions by the appellant’s counsel at the hearing (at [47]), instead of the fact that he had reorganised the material, which was a point that was corroborative of a lack of prejudgment.

208 Here, although the DJ drew his analysis from different parts of the respondents’ submissions, the general sequence of the DJ’s analysis mirrors the chain of reasoning in the respondents’ arguments. There is also insufficient evidence of the DJ’s engagement with relevant aspects of the Prosecution’s case and the evidence (see above at [181]–[197]). Since the case as a whole, and Issue 1 specifically, is the subject of significant factual disagreements, one

would fairly expect that a court dealing with a matter of such factual controversy would weigh up the evidence and the parties’ competing accounts.

(C) THE DJ’S ENGAGEMENT WITH THE PROSECUTION’S CASE DURING ORAL

SUBMISSIONS

209 Finally, contrary to Pek’s submission, the DJ’s engagement with the Prosecution’s case during oral closing submissions is insufficient to counter the reasonable apprehension of a breach of natural justice. Although this point cuts 94 across all three issues, I address it here given that Pek’s argument is primarily focused on the exchange between the DJ and the parties on Issue 1. To support his argument, Pek refers to parts of the hearing where the DJ had inquired into the Prosecution’s position on certain disputed issues such as Foo’s credibility as a witness, or called upon counsel on both sides to respond to specific aspects 95 of the opposing party’s arguments. According to Pek, these exchanges “leave 96 no room for doubt that the [DJ] engaged deeply with the issues in dispute”; 97 the DJ had concluded that the CPIB statements were unreliable as a result of the Prosecution’s failure to deal with many aspects of the respondents’ cases, instead of simply adopting the respondents’ submissions. 98 210 As previously explained, since the complaint in the present case is that the fair hearing rule has been breached by the reasonable apprehension that the DJ had failed to consider the material judiciously when he was forming his

Pek’s Appeal Submissions at para 715. 94 Pek’s Appeal Submissions at para 715. 95 Pek’s Appeal Submissions at paras 717 and 719. 96 Pek’s Appeal Submissions at para 715. 97 Pek’s Appeal Submissions at para 722. 98

conclusion, the DJ’s engagement with the parties and their counsel prior to this

at the hearing is unlikely to carry the same weight as it did in Newton where the complaint was that the DJ appeared to have prejudged the matter (see [63]–[67] above). Further, as explained at [195]–[196] above, even assuming that the Prosecution had not engaged with the respondents’ cases head-on, this did not mean the DJ was justified in disregarding the points and the evidence that the Prosecution did advance in his overall decision as to the reliability of the respondents’ CPIB statements.

(D) THE DJ’S FOOTNOTE REFERENCES TO THE EVIDENCE 211 I make a final point, raised by the Prosecution, that the DJ’s references to the evidence in the footnotes of his GD reinforce the suspicion that the DJ had decided the matter without proper consideration of the underlying primary evidence before him. The DJ repeatedly indicated in the footnotes that he had 99 derived his findings from parts of the respondents’ closing submissions “and the various references cited therein to the relevant supporting NEs, emails, WhatsApp messages and/or other documentary evidence”.

212 Indeed, a closer examination of the DJ’s footnote references reveals that the cited sections of the respondents’ closing submissions do not even contain some of the underlying evidence referred to by the DJ. For instance, while the DJ made consistent reference to the “emails” allegedly relied on by the respondents, none of the portions of the respondents’ submissions which the DJ cited actually contained any discussion of any e-mails (see footnotes 65–68, 70– 71, 77–78, 82–83 and 85–88 in the GD). These references to non-existent pieces

Prosecution’s Skeletal Submissions at para 18(a). 99

of evidence add to the appearance that the DJ failed to consider the material before him judiciously.

(4) Conclusion on Issue 1 213 Having regard to all of the foregoing factors, I am satisfied that a fair-minded and informed observer would reasonably apprehend that the DJ did not apply his mind judiciously to the evidence and the submissions of the parties, especially that of the Prosecution, in his determination of Issue 1. Accordingly, I find that there has been a breach of the fair hearing rule on this.

Issues 2 and 3: The respondents’ mens rea

214 Turning to Issues 2 and 3, I propose to take them together, and first explain my doing so with reference to how these issues were framed.

(1) The framing of Issues 2 and 3 is ambiguous 215 As I have noted at [174]–[177] above, the DJ did not explain how he arrived at the three issues that he considered to be dispositive of the case. This lack of visibility is especially noted in relation to Issues 2 and 3. In contrast to Issue 1 (the reliability of the respondents’ CPIB statements), which the parties clearly joined issue on, it is far from clear how Issues 2 and 3 came into being.

216 Issues 2 and 3 were both matters going to the mens rea of the charges against the respondents, namely, whether the Loans had been given with corrupt intent. Indeed, Issue 2 makes express reference to the mens rea as it is put in terms of whether Foo’s evidence as to his state of mind and the respondents’ intentions behind the giving of the Loans was consistent with the respondents’

claim that “the Loans were given innocently and not corruptly so as to prove that there is no corrupt element and no corrupt intention in the [Loans]”. However, Issue 3 is framed in terms that appear detached from the mens rea, because the DJ considered the question to be whether certain WhatsApp messages between Pay and Foo “showed that Foo’s agency relationship with [the] LTA had been suborned and proved Foo was beholden to or felt indebted to Pay for the Loans”. This suggests that it is an element of the charges against the respondents that Foo’s agency relationship with the LTA must have been suborned; it might be that this was the DJ’s view of the law (the correctness of which I express no comment on), but if so, he did not articulate this. More importantly, for present purposes, Issue 3 as framed by the DJ appears to be a distortion of the Prosecution’s case. The Prosecution’s submission was that “Pay’s various requests for information [in his WhatsApp messages to Foo], coupled with [Foo’s] acquiescence, suggests that Pay knew that Foo’s agency relationship with LTA had been suborned” [emphasis added]. The 100 Prosecution’s focus was on Pay’s (corrupt) intention rather than on whether Foo’s agency relationship with the LTA had been suborned. Therefore, the issue that the DJ ought to have examined was whether the WhatsApp messages showed that Pay had provided the Loans with corrupt intent, rather than whether Foo’s relationship with the LTA had been suborned. The DJ however concluded Issue 3 by finding that Foo’s agency relationship with the LTA had not been suborned (GD at [48]):

For all the above reasons, I took the view that the above selected Whatsapp messages between Pay and Foo as relied upon by the Prosecution did not show that Foo’s agency relationship with

DC Prosecution at para 146 (ROP 9073); see also DC Prosecution’s Reply at para 10(f) 100 (ROP 9436).

LTA had been suborned and failed to prove that Foo was beholden to or felt indebted to Pay for the Loans.

217 Leaving aside the fact that Issue 3 was framed in a way that seemed to distort or, at least, misunderstand the Prosecution’s submission, it is not clear why the DJ segmented Issues 2 and 3 as if they were separate points rather than different aspects of the overarching question as to whether the respondents had the requisite mens rea. And if he did see them as part of the question of the respondents’ mens rea, as I elaborate below (at [220]–[221]), he then ought to have considered them in the light of the different strands of the Prosecution’s case, including the presumption of corrupt intent under s 8 of the PCA, with its implications on the burden and standard of proof, and the objective evidence that the Prosecution maintained went towards demonstrating the respondents’ corrupt intent. In any case, I approach Issues 2 and 3 together, as different strands going to the same overarching question of the respondents’ state of mind.

(2) The problems with the DJ’s approach to Issues 2 and 3 218 In my judgment, leaving aside any contention of judicial copying (as to which, see Annex 1 to this Judgment), a fundamental flaw in the DJ’s approach is that he arbitrarily narrowed the field of evidence relevant to the issue of the respondents’ mens rea to the two pieces of evidence that were referred to in the framing of the issues (namely, Foo’s evidence and the WhatsApp messages between Pay and Foo). No other evidence that was potentially relevant to the respondents’ mens rea, that was relied on by the Prosecution, was taken into account. Seen in this light, it amplifies the impression of the DJ’s consistent failure to deal with material aspects of the Prosecution’s case and the evidence. The difficulty can be expressed in two ways.

219 First, at a narrower level, given that the admissions in the respondents’ CPIB statements were a cornerstone of the Prosecution’s case on mens rea, the fact that I have found that the DJ’s decision under Issue 1 to disregard the respondents’ CPIB statements was infected by a breach of natural justice, raises the possibility that a major piece of evidence was wrongly shut out by the DJ in his analysis on the respondents’ mens rea. If that is the case, the DJ’s findings on the respondents’ mens rea cannot stand because the entire calculus would have to be revisited should a different conclusion be reached on the reliability of the respondents’ CPIB statements.

220 Second, more generally, even if it turns out that the DJ’s decision on the unreliability of the respondents’ CPIB statements was substantively correct, Issues 2 and 3 as framed by the DJ would nevertheless be an underinclusive analysis of the respondents’ mens rea because only Foo’s evidence and the WhatsApp messages between Pay and Foo were taken into account. As I have noted at [175] above, the Prosecution had also relied on other objective evidence to support the inference of corrupt intent on the respondents’ part; these included the commercial significance of the T220 Project to TSC, TSC’s allegedly unsatisfactory performance on the T220 Project, and the process by which the Loans were procured and given to Foo. The apparent failure to consider the effect of the presumption under s 8 of the PCA within the analysis further compounds the problem (see [170]–[173] above). It would also have been necessary to evaluate the evidence that the respondents gave. To the extent, therefore, that the DJ intended his decisions on Issue 2 and also Issue 3 to be decisions on the mens rea element of the charges, these were arrived at without consideration of potentially material pieces of evidence and arguments by the

Prosecution, and were therefore decisions arrived at in breach of the fair hearing rule.

221 To briefly elaborate on this, in so far as Issue 2 was framed by the DJ in terms of whether Foo’s evidence corroborated that of the respondents, that did not necessarily mean that there was no need to critically examine and consider the evidence that the respondents had themselves given. For example, while the DJ referred to Foo having “testified that there was no expectation of favours in exchange for the Loans” and that “at no point did Pay request any favours” (GD at [35]), it was emphasised by the Prosecution that Foo could not speak to the respondents’ states of mind and his evidence was “purely speculative and [did] not mean that the Loans were not corrupt from Pay and Pek’s perspective” since “[w]hat is in issue here is Pay and Pek’s mens rea, not Foo’s”. Even if, as the 101 DJ seems to have thought, Foo’s evidence was relevant to the question of the respondents’ mens rea, the Prosecution had expressly stated that this could not be conclusive and that it was necessary to consider the respondents’ own accounts of their states of mind. In the circumstances, it is noteworthy that the DJ nonetheless went on to frame Issue 2 as he did. In any event, to the extent that the DJ focused on using Foo’s evidence to corroborate the respondents’ accounts, this exercise wholly overlooked the fact that the Prosecution had disputed the respondents’ own accounts of their states of mind. To give one example, the Prosecution had put to Pay that he had known, when making certain inquiries of Foo as to a tender that TSC had submitted for a LTA project, that “[Foo] was willing to exercise influence and … to help [TSC]”. It was 102

Prosecutions’ Reply Submissions at para 170(a) (ROP 9518). 101 NE (20 March 2024) at p 94:5–17 (ROP 3895). 102

inherently flawed for the DJ to disregard these challenges to the respondents’ own evidence and to focus on matching Foo’s evidence with theirs.

The remedial consequences

222 Having found that a fair-minded and informed observer would reasonably apprehend that the DJ had failed to apply his mind to the matters before him, I turn to consider the appropriate remedial consequences, having regard to the need to restore public confidence in the administration of justice, subject to any overriding concerns of fairness to the respondents, as to which it was not suggested there were any.

223 In my view, the DJ’s decision in the GD should be set aside in full, with no weight to be accorded to it. For the reasons explained above, no part of the DJ’s analysis on Issues 1, 2 or 3 is untouched by a breach of natural justice. Issue 1 is perhaps the most egregious in so far as it is affected by the combination of copious copying of the respondents’ submissions and a failure to engage with the evidence and the Prosecution’s case. Even if Issues 2 and 3 appear to be affected by a lesser degree of copying, a consequence that may perhaps be attributed to how they were framed quite narrowly, the fact remains that the DJ had not judiciously addressed the Prosecution’s case on the respondents’ mens rea. In the circumstances, the correct consequence is for the entire decision to be set aside.

224 Finally, as for how the appeals should proceed in light of my foregoing conclusion, it is common ground between the parties that if I were to come to the view that the GD is tainted by a breach of natural justice, I should deal with the matter based on the record because no issue is taken with the way the

evidence was adduced. The parties agreed that I should not order a retrial because of the considerable practical difficulties this would give rise to. The trial below involved a substantial investment of both legal and judicial resources – spanning over 40 days and stretching across a duration of more than two years. A retrial of similar scope would subject the parties to a considerable amount of additional expenses and delay through no fault of their own, without any corresponding benefit. Accordingly, the appeals will proceed with a consideration of the merits as a fresh hearing based on the record of the proceedings below.

225 I note that Pay and Pek have submitted that, even if I were to consider the matter on a de novo basis, I should nonetheless give some consideration to the DJ’s findings in the Judgment, and the appeals should be confined to the points raised by the Prosecution, as I am ultimately exercising appellate (as opposed to original) jurisdiction (see [36] and [39] above). I have difficulty understanding this submission, which appears to be something of a halfway house between disregarding the GD and the usual approach to appellate intervention, and I reject it. The short point is that once the appellate court has set aside the decision below, it will disregard the decision below. With respect, it makes little sense to give some consideration, still less any weight, to the GD when I have concluded that it may reasonably be apprehended to be a product of the DJ having abdicated his duty to consider the matter before him with a judicious mind.

Conclusion

226 For the foregoing reasons, I set aside the DJ’s decision in the GD in its entirety, with the appeal to proceed along the parameters set out at [224] above.

I will convene a Case Management Conference in due course, if necessary, to give further directions.

Sundaresh Menon

Alan Loh, Tan Ben Mathias, Kelvin Chong and Andrew Chia (Attorney-General’s Chambers) for the appellant; Cavinder Bull SC, Yap Han Ming Jonathan, Chua Ying Ying Erin and Melvinder Singh s/o Ajaib Singh (Drew & Napier LLC) for the respondent in HC/MA 9196/2024/01; Tan Chee Meng SC, Paul Loy Chi Syann and Samuel Navindran (WongPartnership LLP) for the respondent in HC/MA 9197/2024/01.

Chief Justice

Annex 1: Non-exhaustive list of similarities between the GD and the respondents’ trial submissions S/N Summary of the DJ’s analysis in the GD Evaluation Issue 1: Reliability of Pay’s and Pek’s CPIB statements

Pay’s 2nd Statement 1 [29]: The DJ summarises the key authorities The DJ substantially reproduces the on the legal principles and cases governing summary of these cases from Pek’s unreliable statements which were cited by the closing submissions at [507]–[510] respondents. (

Prosecutor [2011] 3 SLR 1205

(“Kadar”)) and [503]–[505] (Raj

Prosecutor [2022] 2 SLR 676); and

[39] (

Prosecutor [2021] 5 SLR 860) and

[40] (Chong Hoon Cheong v Public

Prosecutor [2022] 2 SLR 778

(“Chong Hoon Cheong”)). 2 [30]: The DJ summarises the parties’ cases. - [30(a)]: The DJ states his conclusion that he 3 accepted Pay’s submissions that Pay’s 2nd the DJ’s conclusion on the reliability of Pay’s 2nd S unsafe to rely on because, in the absence of substantially similar to his summary any specific responses by the Prosecution to of his conclusion on the reliability of Pek’s statements (GD at [30(d)]). sufficient Statement was accurately recorded. 4 [30(a)(i)]: T admitted that he had submissions at [616]– Pay had committed some offence: some minor paraphrasing:

During the trial, [IO Lim] 616. [IO Lim] admitted that he approached the telling the Court that he interview with a preconceived approached his interview notion some form of offence starting point that Mr Pay when he realized that Pay’s first had committed an offence of some sort …

Muhammad bin Kadar v Public he DJ found that IO Lim

approached the 131 interview with a preconceived notion that [618] with This paragraph which summarises Kumar s/o Aiyachami v Public began by Pek’s reply closing submissions at Statement was inaccurate, unreliable and tatement is Parti Liyani v Public with Mr Pay with the that Pay had committed

explain IO Lim’s conduct, there was . However,

doubt as to whether Pay’s 2nd

CPIB statement did not clearly

he attempted to retreat from his earlier position. Despite this, [IO Lim] anything in Mr Pay’s 1 st would clearly support this preconceived notion of Mr Pay, [IO Lim] then Pay in exchange for leniency on

statement was “not good enough” evidence … to implicate P 618. After all, [IO Lim] offence. In my view, this scenario raises significant concerns about Pay’s 1 CPIB Statement st given [IO Lim] exchange for leniency on recorded. italics and bold italics] his 1 st

offence. omitted; emphasis added in italics and bold italics]

[30(a)(ii)]: T The DJ’s analysis reproduces Pay’s 5 credibility statement-taking process, was questionable because he took contradictory positions as to support this preconceived belief, 617. When [IO Lim]

whether he was influenced by a “hypothesis” realised that he would struggle to point to in his statement-recording process. acknowledged that there was nothing in Pay’s first CPIB CPIB Statement that [30(a)(iii)]: T6 statement suggesting that investigation was inconsistent and imprecise gratification had been given by submissions at [622]– some minor paraphrasing. which Pay had given one of the Loans to Foo. TSC’s projects. Furthermore, he attempted to backpedal admitted that Pay’s first CPIB from his earlier

objectivity of the investigation. ay in a corruption himself admitted that [30(a)(iv)]: The DJ found that IO Lim had 7 there was nothing in Mr the reliability of Pay’s second taken a myopic approach to the investigation the DJ’s analysis and Pay’s closing CPIB statement, particularly which suggested that ’s apparent gratification had been predetermination of Pay’s guilt given by Mr Pay in before the statement was [emphasis added in TSC’s projects, and that

132 the things Mr Pay had

said at various points of CPIB Statement

he DJ found that IO Lim’s were indeed “not good enough” to implicate Mr he DJ found that IO Lim’s Pay of a corruption closing submissions at [619] with because he was unclear about the timing , and the objectivity of his [627] with

This raised serious concerns about the reliability of the evidence and the overall submissions at [628]. In particular,

facts. that IO approach” to the investigation. [30(a)(v)]: The DJ found that IO Lim handled 8 Pay’s amendments to his 2nd Statement in a less transparent manner (than how IO Tay submissions at [630]– some paraphrasing Statement), by analysis. For a side-by-side comparison, see above at [ 165]. As explained above (at [184words before returning it to Pay. did not consider IO

9 [30(a)(vi)]: T evidence on whether Pay had independently submissions at [653]– loans” was inconsistent.

10 [30(a)(vii)]: The DJ found that IO Lim had a “tunnel- submissions at [615] minor paraphrasing. For a side-by- side comparison, see above at [ 179]. indicated Pay’s innocence. 11 [30(a)(viii)]: The DJ found that IO Lim had provided false explanations of the meaning of the words recorded in Pay’s 2nd Statement to submissions at [584]–[586], [598], an exhausted Pay. [637] and [644], and Pek’s reply closing submissions at [40] and [38] (for a summary of the authorities of

Chong Hoon Cheong and Parti Liyani v Public Prosecutor [2020]

SGHC 187) observations. 12 [30(a)(ix)]: T reproduces Pay’s reply closing officers had no reason to frame Pay was an submissions at [60] with some minor oversimplification and that the CPIB officers could have been “overzealous” observations. For a side-by-side investigation comparison, see above at [200].

133 [632] with

handled Pay’s amendments to his 1st and additional he DJ found that the taking the statement to Prosecution’s argument that the CPIB another room and retyping it in his own vision” approach towards the he DJ found that IO Lim’s ]), the DJ investigation and only focused on the with some paraphrasing and additional , with some minor Lim’s contrary and did not record potentially exculpatory the DJ reiterates Pay’s submission arrived at the phrase “more than friendly inculpatory aspects of the case, while [658] with account that he did not retype the ignoring other aspects which potentially paraphrasing and additional in their Lim took a “myopic

13 [30(a)(x)]: The DJ found that the procedural safeguards were meaningless in the face of Pay’s reply [62]. For a side-by-side comparison, IO Lim’s conduct. see above at [190]. 14 [30(b)]: T Prosecution had failed to discharge its burden Pay’s 2nd Statement outweighs its prejudicial [30(e)]). effect. 15 [30(c)]: The DJ concludes that IO Lim’s The DJ’s conclusion that Pay’s 2nd S unreliable is substantially similar to reliability” as required in Kadar, rendering Pay’s 2nd S unreliable. [30(f)]). For a side-by-side comparison, see above at [ 195]. Pek’s statements [30(d)]: The DJ states his conclusions that he As indicated at S/N 3, this paragraph 16 statements were analysis of Pay’s 2nd Statement (GD unsafe to rely on because, in the absence of at [30(a)]). any specific responses by the Prosecution to statements were accurately recorded. 17 [30(d)(i)]: T Statement contained self-incriminating remarks that were repeated four times, which was minor paraphrasing: interview but instead IO emphasise Pek’s culpability: contrived the self- incriminating material The first statement given by Pek, which was recorded by [IO Tan], Statement that the Loans contains several self- to [Foo] repeated 4 times in the document. project simpler throughout Pek’s 1 what Pek actually said during st Statement. he DJ found that Pek’s 1st further sign that Pek’s

submissions at [515] with some not a faithful representation of Pek’s Tan’s attempt to [IO Tan] not only

recorded in Pek’s 1st

accepted Pek’s submissions that Pek’s is substantially similar to the DJ’s were to make incriminating remarks that were Pay’s work in the T220 conduct was incompatible with the tatement was inaccurate and inaccurate, unreliable and , but also This paragraph is substantially “uncompromising need for accuracy and he DJ concludes that the Such repetition is not a replicated it 4 times faithful representation of similar to the DJ’s conclusion his conclusion regarding Pek’s

of establishing that the probative value of regarding Pek’s statements (GD at statements for the same (GD at tatement inaccurate and explain IO Tan’s conduct, there was closing submissions at This is a the interview but rather sufficient doubt as to whether Pek’s indicates an attempt to 1st Statement emphasize Pek's culpability.

Tan's] admission under cross- inculpation of Pek over examination that he used a cut- being a faithful record and-paste method to compile the statement, and it did not matter to Tan], and him whether Pek in fact repeated himself in that manner, implying

interview responses. In my view, Interviewreason for Pek to incriminate

Pek’s 1st Statement. As lacks reliability. [IO Tan] italics and bold italics] cross- cut-and-

manner. [ original omitted; bold italics]

18 [30(d)(ii)]: T November 2019 cohered with the evidence that IO Tan had minor paraphrasing. questions into a single narrative, repackaging accurately reflect Pek’s responses. 19 [30(d)(iii)]: The DJ found that Pek’s 1st Statement appeared to be more a product of IO Tan’s authorship than an accurate account of what Pek actually communicated, due to additional observations: IO Tan’s rearranging of Pek’s responses and This is further supported by [IO prioritised the copying Statement reflects [IO Tan’s] incriminating: of what Pek told [IO cannot be regarded as reliable. Interview be more a product of [IO Tan's] There is no rational own admission, he relied that the repetition might not reason why Pek would reflect the true nature of Pek's have repeated himself communicatedin such a fashion [IO Tan] took several liberties in during Pek’s 1st there would be no rational , let alone incriminate himself himself multiple times in the submissions at [687] with some multiple times within same statement, reinforcing the idea that the statement admitted under [emphasis in On the whole, Pek’s 1st 135 and pasting sections of the examination, he statements that appeared to be more pasted portions authorship more of Pek’s 1st Statement than Pek’s responses he DJ found that Pek’s 13 and it did not matter to during Pek’s 1st Letter was reliable and Pek’s first statement appears to him whether Pek in fact . By [IO Tan’s] submissions at [520] with some repeated himself in that authorship than an accurate combined Pek’s answers to different emphasis in on memory to type out account of what Pek actually Pek’s 1st Statement . This is because Pek’s responses in a way that may not emphasis in italics and hours later. On top of this, [IO Tan] also constructing the statement. He

exhaustive):

whenever he saw fit; context that could have provided clarity to Pek's answers recorded in the statement. Furthermore, been discussed with Pek; [IO Tan] compounded these

questions; more incriminating (d) chose to omit much of

have actual requirement for an IO to record a and statement word-for-word as held in Leck Kim Koon v PP [2022] Pek’s words by [IO Tan] appears similar to PP v portions of the statements Ladaewa where the High Court that appeared to be more noted that in “expanding” what incriminating. recording officer had used words accused. [emphasis added] added]

20 [30(d)(iv)]: T confirmation that he understood the meaning submissions at [622]. substituted Pek's words as he (without being attempted to “frame” Pek or that Pek’s 1st deemed appropriate, rearranged the order of Pek's responses, (a) substituted Pek’s representation of Pek’s words. combined and repackaged words frequently, answers to separate questions,

21 [30(d)(v)]: Tand chose to omit significant

admitted that he drafted Pek’s 1st Statement (b) rearranged the

with the intention to “frame” Pek, by sequence of what had

minor paraphrasing. For a side-by- Pek’s guilt. side comparison, see above at [179]. (c) combined and inaccuracies by copying and repackaged Pek’s pasting sections of the [30(d)(vi)]: T22 answers to separate statements that appeared to be intentionally drafted , thereby Statement to portray Pek as the initiator of the with some additional observations. skewing the overall content to reflect a more incriminating the context that would narrative than what Pek might have explained the ly conveyed. In my answers recorded in view, while there is no legal Pek’s CPIB statements;

136 (e) compounded

inaccuracies in Pek’s SGCA 42, the substitution of CPIB statements by copying and pasting Dahalan bin

he DJ found that IO Tan’s he DJ found that IO Tan

was said by the accused, the of the word “paint” indicated that he he DJ found that IO Tan submissions at [588] with some Pek’s closing submissions at [604] portions of Pek’s 1st that were not uttered by the Statement was not a truthful or accurate selectively recording details that implied omitted; emphasis

scheme rather than accurately recording what

interview. 23 [30(d)(vii)]: The DJ found that IO Tan had conceded that he had predetermined that the staff loans were being used to conceal TSC’s minor paraphrasing. drafted Pek’s 1st Statement in a manner to suggest th process. 24 [30(d)(viii)]: The DJ found that Pek’s first interview was conducted in a way that was Pek’s closing submissions at [675] . intended to create a false sense of security for For a side-by- Pek, and Pek therefore signed his 1st above at [192]. S implications. 25 [30(d)(ix)]: T failure to take notes during the pre-statement interview and the lapse of time between the submissions at [676]– said pre- recording of Pek’s 1st Statement suggested that he may have prioritised CPIB’s needs over ensuring the accuracy of the statement during the interview. 26 [30(d)(x)]: The DJ found that there were two weaknesses in IO statement- memory to recall what Pek had said during additional observations. For a side- the interview; and (b) the presumed ability of by- Pek to comprehend the statement as recorded [200]. . and to correct any inaccuracies 27 [30(d)(xi)]: The DJ found that IO Tan took the liberty of drafting Pek’s statement in his own words, and did not consider it important submissions at [680]– to capture the exact words used by Pek in the interview Pek neither understood nor used. IO Tan’s approach to statement-recording was heavily inaccuracies.

he DJ found that IO Tan’s [677], with [681] with Pek’s closing submissions at [678] statement interview and his Tan’s approach to submissions at [605] with some with some minor paraphrasing and , instead introducing phrases that recording: (a) using his own involvement in the Loans to Foo and had side comparison, see side comparison, see above at was communicated by Pek during the as a faithful record of Pek’s actual words reliant on Pek to identify and correct tatement without fully understanding its at Pek had orchestrated this

28 [30(d)(xii)]: The DJ found that information from Pek’s 1st Statement. submissions at [705]. [30(d)(xiii)]: T29 Statement contained numerous inaccuracies that stemmed from IO Tan’s own phrasing and Pek’s reply closing submissions at [41] Although Pek was offered the opportunity to observations. For a side-by-side review and sign his 1st Statement, this did not comparison, see above at [192]. reliable. 30 [30(d)(xiv)]: T prosecutorial convenience rather than as an accurate reflection of what Pek actually said. IO Tan had an intention to create material that observations. For a side-by-side would support a preconceived belief in Pek’s comparison, see above at [202]. culpability by embellishing Pek’s responses. 31 [30(d)(xv)]: T which undermined the integrity of Pek’s 1st Statement were also evident in the recording submissions at [720]. of Pek’s 2nd Statement as IO portions of Pek’s 1st Statement into Pek’s request or authorise this. 32 [30(d)(xvi)]: The DJ found that IO Tan’s conduct against in Kadar about the “real possibility” submissions at [510]. of an “overzealous police officer”. [30(d)(xvii)]: The DJ found that there was no 33 assurance that the CPIB statements faithfully submissions at [511]. and there was similarity between the present case and Court of Appeal in Kadar. 34 [30(e)]: T Prosecution had failed to discharge its burden paragraph is substantially similar to Statement (GD at [30(b)]).

he DJ found that Pek’s 1st Pek’s closing submissions at [539] he DJ found that the issues he DJ found that Pek’s 1st and embellishments which altered the Statement was drafted primarily for with some minor meaning of what Pek intended to convey. paraphrasing and additional submissions at [497] with some reflected what Pek had actually disclosed, Tan copied minor paraphrasing and additional As indicated at S/N 14, this he DJ concludes that the exemplified the conduct warned IO Tan mean that the document was accurate or 2nd Statement, even though Pek did not of establishing that the probative value of the DJ’s analysis of Pay’s 2nd the situation forewarned by the decided to omit significant contextual

prejudicial effect. 35 [30(f)]: T submissions at [769]. need for accuracy and reliability” as required by Kadar. paragraph is substantially similar to Statement (GD at [30(c)]). 36 [31]: T - Statement and both of Pek’s statements are inaccurate, unreliable and unsafe to rely on.

Issue 2: Relevance and corroborative value of Foo’s evidence

37 [32]–[33]: The DJ summarises the key case summary of these cases from Pek’s authorities referred to by the parties. closing submissions at [218]–[221]; citations listed at submissions at [243]. 38 [34]: The DJ concludes that he agreed with Pek’s submissions that having chosen not to impeach Foo, the Prosecution was bound to submissions at [225]. accept the truthfulness of Foo’s evidence. 39 [35]: The DJ found that the respondents were corroborated the evidence of Pay and Pek that submissions at [235]– corruptly unequivocal about the innocence of Pay and Pek. 40 evidence was thoroughly consistent with the respondents’ case and the evidence of Pay and Pek that the Loans were given innocently minor paraphrasing: and not corruptly : [Foo] I agreed with Pek’s submissions that Foo’s aforesaid evidence is and evidence of Pay and Pek (as set out in the rest Defence’s case and the evidence

139 [36]: The DJ agreed with Pek that Foo’s

submissions at [239] with some he DJ concludes that IO Tan’s conduct rendered Pek’s statements The DJ largely reproduces the correct to rely on Foo’s evidence which incompatible with the “uncompromising ’s evidence is [238] with thoroughly consistent As indicated at S/N 15, this the Loans were given innocently and not with the Defence’s case , and that Foo was completely and/or block quotes from the case he DJ concludes that Pay’s 2nd thoroughly consistent with the Pek’s CPIB statements outweighs their the DJ’s analysis of Pay’s 2nd Pek’s closing of these submissions) of Pay and Pek that the Loans

corruptly as highlighted in Pek’s given corruptly. In submissions as follows: particular: (a) (a) were extended as personal loans, intended to help Foo as a friend [Foo] to benefit TSC. Pek and out of goodwill, rather than lend money to [Foo] to to benefit TSC. benefit TSC. Pek was also money to Foo with the intention of benefiting TSC. He was clear that beneficiary of the Loans, and TSC and TSC did not give any did not provide any loans to Foo. loans to [Foo]. (b) (b) offered or promised any benefits showing favour to him or TSC. He [Foo] in return for [Foo] emphasized that he did not ask showing favour to Pay or for any favours from Foo in return TSC. Pay did not ask for for the Loans and did not believe any favours from [Foo] in return for the Loans, and him because of the Loans. Pay did not think that [Foo] also confirmed that Foo did not Pay because of the Loans, were given innocently and not that the Loans were not and [Foo] did not give and favours. Pek testified that he did not perceive Pay as wanting any favours. Pek testified that favour from Foo and was “quite Pay testified that the Pay testified that the Loans sure” that Pay had no intention of Loans were not given to obtaining any benefits from Foo. any favour from [Foo] and … testified that he did not was “quite sure” that Pay Pek similarly testified that he did not lend unequivocal that TSC from [Foo]. was not the ultimate … TSC was not the ultimate beneficiary of the Loans

added] Pay testified that he Pay testified that he never had never offered or

41 [37]: The DJ agreed with Pek to Foo in exchange for Foo promised any benefit to

evidence undermined the Prosecution’s case Pek’s closing submissions at [233]– but the Prosecution failed to impeach Foo or [234] with some minor paraphrasing and additional observations. that the court should ascribe greater weight to

that Foo would feel indebted to

respondents’ case that the Loans were given

would feel indebted to innocently. give or attempt to give him any did not try to give Pay he did not understand that Pay wanted to get 140

that Foo’s had no intention of

obtaining any benefits

challenge his credibility. The DJ accepted Foo’s evidence where it cohered with the

omitted; emphasis

42 [38]: The DJ found that Foo, Pay and Pek were all consistent that there was no quid pro

quo, submissions at [240]–[

associated with the Loans.

43 submissions at [249]. guilty plea was an corrupt intention and a belief that Pay or Pek was corrupt. 44 [40]: The DJ found that Foo’s evidence was - clearly that he had asked Pay for a personal loan as he believed that Pay was well to do and had the capacity to loan him money. 45 - friendship between Pay and Pek to warrant the Loans. 46 - Prosecution’s submissions that the evidence showed that Foo’s agency relationship with the LTA had undeniably been suborned. 47 [43]: The DJ summarises his above findings, - stating that “[f]or all the above reasons, [he] rejected the Prosecution’s submissions that Foo’s evidence on the mens rea of Pay and Pek is purely speculative”, and that “Foo’s evidence as the recipient of his own mens rea and the mens relevant and must be accepted as truthful by the Prosecution which did not impeach him”.

Issue 3: Whether the WhatsApp messages between Pay and Foo showed that Foo’s agency relationship with the LTA had been suborned

48 [44]: The DJ reasoned away Foo’s WhatsApp message to Pay: “No worries. Anything I can Pay’s closing submissions at [360]– help, I will. For you and Tiong Seng”, on the [362].

[39]: The DJ disagreed with the Prosecution’s arguments that Foo’s [42]: The DJ disagreed with the admission of feeling “corrupted” and his [41]: The DJ rejected the Prosecution’s corrupt intent or dishonest advantage 242] with submissions that there was no genuine rea of Pay and Pek … is acceptance of his own basis that this was Foo’s unprompted communication and Foo had provided a

“For you and Tiong Seng”. 49 [45]: The DJ reasoned away Foo’s WhatsApp Pay’s closing submissions at [408]– [415] evaluation on the basis that this message was submissions at [305]–[310] with some additional observations. never cross-examined by the Prosecution on the meaning of his responses and there were reasonable explanations for Pay’s responses. [46]: The DJ dealt with the fact that Foo had 50 the DJ’s analysis and Pay’s closing screen on 14 August 2019 which contained submissions at [421]–[428].

that at the time Foo sent the message, the information was not confidential. Pay testified that the image provided no valuable or new information to him and there was no evidence to suggest that Pay had requested Foo to send the image in a way that bypassed usual protocols. [47(a)]: The DJ dealt with Pay’s WhatsApp There is a very 51 message to Foo, in which Pay asked whether between one sentence of this TSC prequalified for certain projects. The DJ paragraph and Pay’s reply closing submissions at [286]. between the first loan made by Pay and the alleged favour sought from Foo due to the length of time between the loan and the time Pay sought for the purported favour. 52 [47(b)]: The DJ found that Pay asking Foo for There is a very slight similarity the name of the person in charge of the “LTA between one sentence of this paragraph and Pay’s reply closing innocuous and there was nothing improper submissions at [302]. about the message. 53 [47(c)]: T contract was not uncommon or unusual, and believed Pay’s clarification that his questions observations. confidential information.

sent Pay a photograph of his work laptop

an internal document relating to TSC’s slight similarity performance in a tender. The DJ reasoned message which mentioned his inability to influence the outcome of the tender and Pay’s reply closing he DJ found that Pay’s question about TSC’s chances for the T316 LTA Pay’s closing submissions at [381] accepted that there was no connection and [388] with some additional a unilateral action on Foo’s part. Pay was HQ redevelopment, project J120” was reasonable explanation for why he had said were not intended to extract specific or

54 [47(d)]: T There are slight similarities between two sentences of this paragraph and and [340]. practice. Foo’s response did not reveal any confidential information. 55 - Prosecution’s interpretation of Foo’s use of the term “we” in his WhatsApp message to Pay that “[w]e should have arranged for Lotte to meet us earlier” was never put directly to Foo, and Pay had clarified that the term “we” referred to LTA rather than TSC and was not evidence that Foo had been “captured”. 56 [48]: The DJ concludes that the WhatsApp - messages between Pay and Foo did not show that Foo’s agency relationship with the LTA had been suborned.

143 [47(e)]: The DJ reasoned that the

he DJ found that Pay’s question about TSC’s chances in the tender for the J101 project stemmed from personal Pay’s closing submissions at [337] curiosity and that this was a common