1 This is an appeal by Mr Kalaichelven Genesan (the “appellant”) against his sentence. He had initially appealed against his conviction as well, but he subsequently elected to proceed only with the appeal against sentence.

2 The appellant was convicted at trial in a District Court of one amalgamated charge of criminal breach of trust (“CBT”) as an employee under s 408 of the Penal Code (Cap 224, 2008 Rev Ed) (the “Penal Code”), read with ss 124(2) and 124(8)(a)(i) of the Criminal Procedure Code (Cap 68, 2012 Rev Ed). The various acts of CBT spanned 16 November 2020 to 18 September 2021 and involved goods worth $108,339.63 from Lim Siang Huat Pte Ltd (the “employer”), for whom the appellant had worked as a storekeeper or warehouse assistant.

3 The District Judge (the “DJ”) sentenced the appellant to 28 months’ imprisonment. On appeal, among other arguments, the appellant relied on three new sentencing precedents which were not referred to below. He cited Chong Kum Heng v Public Prosecutor [2020] 4 SLR 1056 (“Chong Kum Heng”) and Chua Ya Zi Sandy v Public Prosecutor [2021] SGHC 204 (“Sandy Chua”) in support of his submission for a reduction in sentence, as well as the unreported case of Public Prosecutor v Wan Kam Lan DAC 920285/2018 (11 July 2018) (“Wan Kam Lan”) referred to in Chong Kum Heng at [55] and [62]–[64]. I found these cases to be distinguishable and thus dismissed the appeal and affirmed the sentence below. I now set out my reasons for doing so.

4 The appellant was convicted on 22 July 2024 following eight days of trial. Before the trial commenced, one Muhammad Khairudin Bin Kamis (the “accomplice”) had pleaded guilty on 4 September 2023 to one charge under s 413 of the Penal Code and was sentenced to 30 months’ imprisonment. This was a charge for habitually dealing in stolen property worth $281,725.74 which he had received from the appellant and two other employees. That s 413 charge carried a maximum term of 20 years’ imprisonment. This may be contrasted with the appellant’s amalgamated s 408 charge, which carried a maximum punishment of 30 years’ imprisonment (ie, double the maximum punishment of 15 years for one unamalgamated charge under s 408 of the Penal Code).

5 The employer was an importer and distributor of dry food products to restaurants and other food establishments. The appellant was entrusted with the employer’s goods as a “picker”. He was not authorised to remove or sell the goods, but merely to pick out the goods ordered by customers from the employer’s warehouse and arrange for these goods to be collected from the warehouse staging areas by drivers for delivery to the customers.

6 The prosecution’s case below was that the appellant had committed the offence with the accomplice who was employed as a driver contracted to the employer. The manner of offending involved the appellant selling the employer’s goods to the accomplice when he was not authorised to do so. Within the offending period, he would take small amounts of items from the employer’s warehouse, which did not form part of the daily orders, to avoid detection. Since the employer replenished its inventory on a daily basis, this approach rendered the detection of the unauthorised sales more difficult, due to the comparatively larger quantity of items sold to customers every day and the timeous replenishment of any missing stock within the warehouse.

7 The accomplice testified as a prosecution witness in respect of the methodology of offending between him and the appellant (ie, placing orders with the appellant, communicating via text message on WhatsApp, voice notes, phone calls, etc). The employer’s general manager gave evidence on how she had been tipped off that the employer’s goods were being sold at discounted rates and noticed excess goods not corresponding to any customer orders in the staging areas of the warehouse for pick-up by drivers. Her enquiries eventually led to the appellant’s arrest. Two other employees (“PW6” and “PW7”) testified that they had been contacted by the appellant, who sought to persuade them to give evidence placing the blame for the offence entirely on the accomplice to the exclusion of the appellant.

8 The appellant was unrepresented at the trial until he was found guilty and convicted. Thereafter, he was represented by Mr Ashvin Hariharan, who continued to represent him on appeal. The appellant’s defence, in brief, was that he had never placed extra items in the staging area for pick-up by the accomplice, and that other persons who had borrowed his phone could have exchanged the various incriminating messages via WhatsApp with the accomplice. He also denied in cross-examination that he had reached out to PW6 and PW7 to ask them to change their evidence in his favour, but conceded during his own cross-examination that he never challenged that part of their evidence when he cross-examined them.

9 The DJ rejected the appellant’s evidence and preferred the prosecution witnesses’ evidence. In particular, the DJ accepted the evidence of the accomplice, whom he found to be a credible witness with no apparent motive to lie, and whose evidence was corroborated by the chat log of their WhatsApp messages. In contrast, he found that the appellant’s evidence was difficult to square with several of the WhatsApp messages, and it featured convenient vacillation in various parts.

10 The DJ calibrated the appropriate sentence based on two first instance precedents relied upon by the respondent, viz, Public Prosecutor v Choo Hiang Mui [2021] SGDC 213 (“Choo Hiang Mui”) and Public Prosecutor v Sukumar s/o Munisamy [2014] SGDC 265 (“Sukumar”). The amounts at issue were roughly similar between them: $108,339.63 in the appellant’s case; $126,304.32 in Choo Hiang Mui; and $110,474.10 in Sukumar. All these cases involved first-time offenders convicted at trial who made no restitution.

11 The sentence in Choo Hiang Mui was 30 months’ imprisonment in the first instance, but this was reduced to 26 months on appeal. The sentence in Sukumar was 30 months’ imprisonment. The DJ held that the gravity of the appellant’s offending fell in between Choo Hiang Mui’s and Sukumar’s. It was less serious than that in Sukumar, which had involved a slightly larger sum ($110,474.10 versus $108,339.63) and a much longer period of offending (about four years versus ten months). It was more serious than that in Choo Hiang Mui, which involved a larger sum ($126,304.32 versus $108,339.63) but a shorter period of offending (six months versus ten months). The sentence of 28 months was a middle ground between the 26 months in Choo Hiang Mui and the 30 months in Sukumar.

12 The appellant provided three broad reasons in his submissions on appeal for a lower sentence. First, he sought to rely on two new reported sentencing precedents on appeal, not referred to below, to argue for a reduction in sentence. Citing my decision in Chong Kum Heng (at [62]), he suggested that a 14-month imprisonment term should be considered the starting point for a s 408 CBT offence which involved a sum of around $111,000. In that case, the sentence for the said s 408 charge was reduced on appeal from 21 months to 18 months’ imprisonment. It was contended that the disparity between 18 months in Chong Kum Heng and the 28 months in the present case meant that the sentence was manifestly excessive. The appellant further cited the case of Sandy Chua, which was said (at [9]) to have approved the sentencing approach in Chong Kum Heng.

13 The appellant also relied on Wan Kam Lan, in which the accused had been sentenced to 14 months’ imprisonment for a CBT offence involving $125,796.32. Given the similarity in the quantum, the appellant suggested that his sentence should not differ substantially from both the 14-month term in Wan Kam Lan and the 18-month term in Chong Kum Heng. Using these sentences, he arrived at the range of 18 to 21 months’ imprisonment as the appropriate penalty for his offending.

14 Second, the appellant argued that the DJ erred in fact in finding that PW6 and PW7 had been encouraged by the appellant to tailor their evidence in his favour and to pin all the blame on the accomplice. The appellant pointed to equivocal aspects of their evidence to argue that it could not be confidently determined, beyond a reasonable doubt, that he had in fact attempted to pervert the course of justice in this way.

15 Third, the appellant argued that the DJ was wrong to consider the precedent in Choo Hiang Mui, since the reduced sentence on appeal of 26 months was unreasoned, thereby rendering it an unreported precedent. As for Sukumar, the appellant argued that a greater reduction from the sentence of 30 months in Sukumar was needed, since the offending period of four years was significantly longer than the appellant’s ten months, and the accused there was more senior in the company than the appellant was here.

16 In his written submissions, the appellant further pointed out that another employee – namely, one Muhammad Adham Syahin bin Noor Mohamed (“Adham”), who likewise sold stolen goods worth approximately $163,000 to the accomplice – had been sentenced to 26 months’ imprisonment, which was less than the 28 months meted out to the appellant. This appeared to allude to the parity principle though it was not articulated as such. For the reasons which follow below (at [42]–[43]), I was of the view that Adham’s sentence worked against the appellant – in fact, granting the appeal would have offended the principle of parity, having regard to Adham’s 26-month imprisonment term and the relevant differences between their respective cases.

17 The respondent submitted as follows. First, the threshold for this court to intervene in the DJ’s factual findings on PW6 and PW7 was not met, and the DJ was entitled to accept their evidence as credible. The fact that the appellant had attempted to induce them to commit perjury in his own trial was clearly a weighty offender-specific aggravating factor in the circumstances.

18 Second, the respondent argued the new precedents of Chong Kum Heng and Sandy Chua did not demand a different sentence than that meted out below. Chong Kum Heng had to be confined to its unique facts, viz, where the accused was permitted to sell the company’s scrap materials and obtain the sale proceeds to be used for the good of the company, for which the quantum converted to his own use was not a direct proxy for the degree of harm caused to the company. Moreover, Sandy Chua involved an accused who had pleaded guilty to a s 408 CBT charge in the amount of $41,319.90 and had received ten months’ imprisonment. A sentence significantly higher than the ten months there was warranted for the appellant, since: (a) the sum misappropriated here of over $108,000 was much higher; (b) the offending period of ten months was longer than the five-and-a-half weeks in Sandy Chua; (c) the scheme here was much more sophisticated than that in Sandy Chua; and (d) the accused in Sandy Chua pleaded guilty while the appellant claimed trial and was convicted. At the oral hearing, the respondent highlighted that Wan Kam Lan was an unreported case, and hence, lesser weight should be accorded to the 14-month term there due to the absence of written reasons for that sentence.

19 The respondent adopted the DJ’s approach to Choo Hiang Mui and Sukumar below (ie, calibrating a custodial term which was higher than Choo Hiang Mui’s 26 months and lower than Sukumar’s 30 months), and argued that it should not be interfered with on appeal.

20 Finally, at the oral hearing, the respondent submitted that the appellant’s sentence was, in fact, in line with Adham’s sentence, given that Adham had pleaded guilty and there were no aggravating factors which worked against him, unlike in the appellant’s case where he was convicted at trial and found to have attempted to instigate PW6 and PW7 to lie in court under affirmation.

21 In coming to my decision to dismiss the appeal, I considered the following four points:

22 The appellant failed to satisfy the appellate intervention threshold of showing that the DJ’s decision to accept PW6 and PW7’s evidence as to their having been contacted by the appellant to alter their testimony at trial was plainly wrong or clearly against the weight of the evidence below.

23 First, it should be noted, for completeness, that the appellant accepted in cross-examination that he never challenged this aspect of PW6 and PW7’s evidence during his cross-examination of them (see at [8] above). While the confrontation rule in Browne v Dunn (1893) 6 R 67 may be relaxed for an accused who is self-represented, as a matter of procedural fairness, if one had been falsely accused of contacting a witness to instigate tailored testimony, and no such communication was ever made, one would naturally expect some questions from even a layperson to the effect that PW6 and PW7 had made up their entire account. This is certainly a factor which the DJ was entitled to consider in drawing forensic inferences as a matter of plain common sense.

24 Even putting that aside, the appellant’s main argument rested upon supposed ambiguities in the oral evidence of PW6 and PW7. However, the oral evidence was nowhere near as equivocal as the appellant suggested. PW6 categorically confirmed that the appellant (ie, “Kalai”) had contacted him to ask him to “blame everything on [the accomplice] Muhammad so that he himself will be free” – the appellant makes much of the use of the words “I think” in front of that sentence, but in context, he made far too much of an ordinary filler phrase; the texture and context of the whole of PW6’s evidence, especially the clear terms he used to describe the appellant’s instructions to him and his conversations with him (eg, “I replied him angrily like a high voice but through text”, “if there was an investigation with the police and he has a case, he is not supposed to contact me”, etc) made it crystal-clear that PW6 was not guessing or speculating; rather, he was recounting what happened in his own words, and therefore, the phrase “I think” could not be construed in the overly literal fashion suggested by the appellant.

25 PW7’s evidence was similarly clear – he gave a detailed, vivid account about what the appellant told him at the time. He said that “[i]f Muhammad supported him in taking the blame for the case, then he will support Muhammad’s family monthly financially. He asked if I can help to state that he was not involved”, and there was no question here as to whether PW7 was muddled or confused. PW7 responded in the affirmative that “Kalai tried to contact [him] recently, contrary to the appellant’s submission that there was no proof that it was he who contacted PW7 to make this request.

26 In relation to both PW6 and PW7, the appellant argued that there were no copies of these messages over Facebook Messenger adduced as evidence in the trial. That point was a neutral one at best – the absence of evidence is not evidence of absence. Even if a witness’ credibility could have been stronger, that is not a good reason to reduce the weight one attaches to their evidence. It simply means that one could have accorded greater weight to their evidence if that additional corroboration had been produced. It does not mean that the oral evidence of PW6 and PW7 was incredible or not to be believed.

27 The overarching issue was whether the DJ was plainly wrong in accepting PW6 and PW7’s evidence. That was clearly not the case. The DJ was presented with the evidence of two independent witnesses, both of whom gave clear and categorical evidence to the effect that they had been contacted by the appellant in relation to their evidence at trial. They stated unequivocally that the appellant had attempted to persuade them to tailor the contents of their evidence. Having regard to the typical forensic indicia of “[c]ontent credibility, evidence of partiality, coherence and a need to analyse the evidence in the context of established facts” (see Sakthivel Punithavathi v Public Prosecutor [2007] 2 SLR(R) 983 at [76]), these were two witnesses without any self-interest in the outcome of the trial, no apparent motive to lie or frame the appellant, giving evidence that was clear, coherent, cogent, inherently believable (given the appellant’s evident self-interest in doing what they said he did), and internally consistent. Moreover, the fact that both of them claimed to have been contacted by the appellant, who instructed them to do something similar (ie, tailor their trial testimony in his favour) was an external consistency which lent further credibility to their respective accounts (see Jagatheesan s/o Krishnasamy v Public Prosecutor [2006] 4 SLR(R) 45 at [40] and [79]).

28 All in all, in the absence of anything within the record below to suggest that PW6 and PW7 had some reason to collude or fabricate such a substantially similar story to falsely incriminate the appellant for procuring perjury, I was not persuaded that the DJ could be faulted for accepting their evidence.

29  I considered that none of the new sentencing precedents relied upon by the appellant as enumerated at [3] above had any relevant application to his offending, for the reasons that follow.

30 In Chong Kum Heng, the remark which I had made (at [62]) that the 14 months’ sentence for misappropriating $125,796.32 in the unreported case of Wan Kam Lan was a “starting point” should not be taken out of context or in isolation to suggest a sentencing benchmark of general application. As I had noted at [63], Wan Kam Lan concerned a case featuring an early plea of guilt, evidence of remorse and some restitution. These elements were absent in the present case. The fact that Wan Kam Lan could serve as an indicative starting point in Chong Kum Heng (at [62]) did not mean that it would also serve as a starting point in all cases of an offence under s 408 of the Penal Code. In any event, Wan Kam Lan was an unreported case, and it is settled law that limited precedential value should be accorded to unreported decisions, which may not set out full details concerning the facts and circumstances of the case. Ironically, this was a point which the appellant himself made in submissions in relation to the DJ’s reliance on Choo Hiang Mui below. It applies mutatis mutandis to his reliance on Wan Kam Lan.

31 As for the sentence of 18 months’ imprisonment in Chong Kum Heng for misappropriating $111,000 (at [65]), I agreed with the respondent that there were a number of features in that case that were patently different from the present case. The most prominent among these was the nature of the fraud – indeed, in Chong Kum Heng, the accused was directed to dispose of the wastage (essentially scrap material). In other words, the material was to be disposed of in any event. The breach of trust was therefore confined to misapplying the proceeds of the disposal for his own use and not for the benefit of the company. While I did emphasise (at [60]) that this had “no bearing on the fundamental point that the [accused] was expected to adhere to [the victim’s] policy on the use of the sale proceeds in the first place”, and there was therefore still a breach of trust, that did not mean that this factor became totally irrelevant to sentencing. It would help in appraising the extent and gravity of the breach of trust and the relative sophistication of the fraud.

32 In addition, one of the bases that I had relied on (at [62]) for distinguishing Chong Kum Heng from Wan Kam Lan was the fact that the fraud in the latter case was “more egregious” since it involved “falsification of the company’s records and the issuance of company cheques”. In Chong Kum Heng, the wastage was intended to be disposed of, and the accused was entrusted to do so “without having to properly account for their use” (at [57]). Consequently, the nature and extent of the breach of trust remained relevant. In the present case, it was clear that the nature of the breach was more egregious than in Chong Kum Heng, as the appellant was not even permitted to sell or dispose of the entrusted goods. His responsibility was strictly limited to placing items in staging areas to correspond with customers’ orders and to ready them for pick-up and delivery (see at [5] above). The betrayal of trust was therefore graver than that in Chong Kum Heng, seeing as the employer never consented to or authorised the disposal of the misappropriated goods in the first place. It was also relevant that the appellant took steps to minimise the detection of his fraud, viz, by taking smaller quantities of items from the warehouse each day to allow the missing stock to be replenished via regular supplies (see at [6] above), all in coordination with the accomplice. That too involved a level of planning and sophistication greater than was displayed by the accused in Chong Kum Heng, whose fraud took the simpler form of the retention of a portion of the sale proceeds of the scrap material (at [38]) instead of applying them for the specified purpose.

33 Moreover, in property offences such as CBT, the sum misappropriated is a key indicator of the harm perpetrated by the accused, and consequently, a weighty factor to consider in sentencing (see Wong Kai Chuen Philip v Public Prosecutor [1990] 2 SLR(R) 361 at [18]). As I observed in Chong Kum Heng at [60], the sums in the charges there were not “a direct and accurate proxy for the degree of harm caused to [the victim]” on account of the victim’s policy allowing for the waste to be disposed of by its subcontractors without the sale proceeds being used for the good of the company in that scenario (at [59]). That further reduced the direct comparability of the quantum in Chong Kum Heng with the value misappropriated here.

34 Likewise, it was noted in Chong Kum Heng (at [57]) that the victim had “not perceived itself to have suffered tangible loss or damage”, and while it was emphasised that the victim’s “more forgiving eye” (at [61]) did not absolve the accused of liability or “diminish his culpability”, it changed the texture with which the offending matrix in Chong Kum Heng was viewed when used as a comparator. In the present case, the employer had suffered a very tangible and quantifiable loss indeed, viz, the disposal of goods it never wanted to be disposed of to the accomplice, and worse, for these goods to then be sold at competitive discounted rates undercutting the employer’s rates. When these factual differences were considered, together with the appellant’s attempts to pervert the course of justice at trial, and the fact that the maximum penalty was twice that in Chong Kum Heng on account of the amalgamated charge, the custodial term in Chong Kum Heng had limited applicability in determining the appropriate sentence for the present offending.

35 As for Sandy Chua, the respondent was correct that the 10 months’ imprisonment term there could not shed much light on the appropriate sentence here. The amount of $41,319.90 misappropriated over five-and-a-half weeks (at [1] and [3(b)]) in that case was significantly far apart from the $108,339.63 misappropriated over the course of ten months at issue here. That case also involved an accused who had pleaded guilty at an early stage (at [5(a)]) and, again, lacked the weighty aggravating factor of an attempt to pervert justice at her trial by making active efforts to manipulate witnesses’ testimonies. Given these distinguishing factors, the custodial terms in Chong Kum Heng and Sandy Chua could not offer much guidance as to whether the sentence meted out here was manifestly excessive in the circumstances.

36 The appellant levelled two main criticisms at the DJ’s reliance on Choo Hiang Mui and Sukumar. The first was that the sentence in Choo Hiang Mui was reduced on appeal from 30 months to 26 months’ imprisonment, but no written grounds were furnished by the appellate court for its decision. I agreed, in principle, that less reliance should have been placed on the custodial term in Choo Hiang Mui for the same reasons I gave at [30] above. Unreported precedents have traditionally been accorded little weight because “they are unreasoned” and/or are “without grounds or explanations” to shed light on “what had weighed on the mind of the sentencing judge or why the sentencing judge had approached the matter in a particular way” (see Janardana Jayasankarr v Public Prosecutor [2016] 4 SLR 1288 at [13(b)]). We have the benefit of reasons for the first-instance sentence, but since that was not the sentence being considered as a precedent, the reasons for the appellate court’s decision to calibrate the sentence downwards to 26 months were unknown. In the circumstances, the traditional rationale for according less weight to unreported precedents applied.

37 That being said, that did not get the appellant very far. As I had indicated above, the sentences in Chong Kum Heng, Sandy Chua, and Wan Kam Lan were not useful comparators to determine the appropriate sentence in the present case. Even if one were to disregard the 26-month term in Choo Hiang Mui, the appellant would still have to show that the DJ erred in his approach to the 30-month term in Sukumar to satisfy the appellate intervention threshold.

38 The appellant suggested that a greater downward calibration was warranted because the period of offending was much longer, viz, about four years from 2007 to 2011 compared with ten months here (see Sukumar at [2]). However, that was only one difference. The amounts were about the same ($110,474.10 there, $108,339.63 here), and no attempts were made in Sukumar to pervert the course of justice or influence witness testimony at trial. The appellant further claimed that the accused in Sukumar was much more senior in the company compared to the appellant. Even assuming that was correct, from the broad description of the accused’s role as an “operations manager” who was in charge of “overall daily operations” alone in Sukumar (at [9]–[10]), that did not by itself warrant the inference that there was substantially greater reposal of trust and confidence vis-à-vis the use and application of the property at issue, or that there was graver betrayal of trust and confidence by the accused in Sukumar. This was not a sentencing factor which could be accorded much weight, in my view, in comparing the facts of Sukumar with the present facts.

39 I also saw no relevance in the appellant’s point that the sum misappropriated in Sukumar’s case was the product of five charges whereas the sum here was the product of one amalgamated charge. In determining the global sentence in Sukumar of 30 months (at [85]–[86]), the court, applying the totality principle, would have to compare the overall gravity of the accused’s offending in proportion with the holistic effect of the global imprisonment term meted out there (see Jeffery bin Abdullah v Public Prosecutor [2009] 3 SLR(R) 414 at [16] and Public Prosecutor v Raveen Balakrishnan [2018] 5 SLR 799 at [47(c)], [58], and [71]–[73]). There could thus be no objection in principle to comparing the global 30-month imprisonment term in Sukumar with the appropriate term for the amalgamated CBT charge here.

40 In light of these facts, I was unable to agree that a reduction of two months from 30 months’ imprisonment in Sukumar to the 28 months’ term here was necessarily inadequate. There was no basis shown to intervene in the sentence imposed below, which was well within the ambit of the DJ’s sentencing discretion.

41 For completeness, I turn to Adham’s sentence of 26 months’ imprisonment, and how the principle of parity was engaged.

42 In the proceedings below, the respondent referred to Adham’s sentence in mounting an argument on the parity principle. For context, three employees – Adham and the appellant being among them – sold goods misappropriated from the employer to the accomplice. Adham, for his part, misappropriated goods worth $162,960.81, considerably higher than the $108,339.63 in the appellant’s case and over the same time frame of about 10 months. However, crucially, Adham pleaded guilty and did not have the weighty aggravating factor of trying to instigate perjured testimony at trial. The exact sentencing discount given to Adham for his plea of guilt was not known in the absence of written grounds. He received 26 months’ imprisonment for that charge.

43 In my view, Adham’s sentence was not inconsistent with the sentence received by the appellant. While Adham did misappropriate substantially more than the appellant, the fact that Adham pleaded guilty and did not try and pervert justice by suborning perjury meant that a modest increase from the 26 months in Adham’s case to the 28 months’ term for the appellant was not inconsistent with the principle of equal justice between co-offenders. The fact that the appellant had tried to suborn witnesses was a weighty aggravating factor which was absent in Adham’s case. These differences justified a higher custodial term for the appellant. As I recently reaffirmed in Public Prosecutor v Ng Whye Quan [2025] SGHC 200, the parity principle did not require that all co-offenders be treated alike, but that co-offenders in like situations be treated alike (at [32]), consistent with the principle of equal justice and the maintenance of public confidence in the justice system’s even-handed application of the law. Here, Adham and the appellant were clearly not in like situations. Accordingly, there was no violation of the principle of parity by the DJ below.

44 For the foregoing reasons, I determined that the sentence of 28 months’ imprisonment was not manifestly excessive and did not warrant any appellate intervention. I therefore dismissed the appeal and affirmed the DJ’s sentence.