1 This is a tragic workplace accident case which involved the deaths of two construction workers, Mr Ratan Roy Abinash Roy (“Ratan”) and Mr Rajib Md Abdul Hannan (“Rajib”). The appellant, Mr Nurun Novi Saydur Rahman (“Nurun”), was alleged to have instructed Ratan, Rajib and three other workers to load an air compressor onto an uninstalled loading platform. At the material time, the loading platform was suspended at the edge of the 7th floor of a Tower under construction called “Tower A”. ¹ The manner in which the air compressor was loaded onto the loading platform was in breach of a whole slew of safety regulations. One crucial breach was that the loading platform was suspended by a tower crane via four lifting chain slings, instead of being properly secured to the side of Tower A. When the air compressor was loaded onto the loading platform, the air compressor rolled away from the edge of the building. Unfortunately, Ratan and Rajib were standing on the loading platform, in the path of the air compressor. They could not move away in time and fell out of the loading platform together with the air compressor. Ratan and Rajib landed on the ground level of the construction worksite. They were pronounced dead at the scene by responding paramedics.

2 Nurun was charged under s 15(3A) of the Workplace Safety and Health Act (Cap 354A, 2009 Rev Ed) (“WSHA”), for instructing the team of workers under his charge to load the air compressor onto the loading platform when it was unsafe to do so.

3 Nurun claimed trial to the charge. He was convicted and sentenced to a fine of $15,000. The grounds of decision of the district judge is reported at Public Prosecutor v Nurun Novi Saydur Rahman [2017] SGDC 263 (“the GD”). Nurun now appeals against his conviction and sentence. The Prosecution cross-appeals against the sentence.

4 Having considered all the available material as well as the submissions of the parties and Mr Kevin Tan (“Mr Tan”), the Young Amicus Curiae, I dismiss Nurun’s appeal against conviction and sentence and allow the Prosecution’s cross-appeal against sentence. I give my reasons below starting with a summary of the background facts.

5 Nurun was an employee of GS Engineering & Construction Corp (“GSE”), a Korean construction company. At the material time he was deployed at a construction worksite at Fusionopolis Way, Ayer Rajah Avenue, where GSE had been engaged to construct two towers, Tower A and Tower B. Tower A’s structural works were subcontracted to a company identified as Zhang Hui Construction Pte Ltd (“Zhang Hui Construction”). ²

6 On 22 January 2014, the day of the accident, Zhang Hui Construction sought the assistance of GSE to move an air compressor which was located on Level 7 of Tower A to Level 8 of Tower A via the use of a loading platform. The site supervisor for Tower A was an employee of GSE by the name of Miah Rashed (“Rashed”). On the day of the incident, Rashed instructed Nurun not to install the loading platform at Level 7. ³

7 A loading platform is a drawer-like platform used in multi-story construction sites to move heavy loads between different multi-storey buildings or different floors on the same building. ⁴ The loading platform can be transported to different floors or locations on a construction site through the use of a chain sling connected to a tower crane. Once placed on a particular floor, the proper procedure before loading heavy objects onto the loading platform would be to first install the loading platform. Installation involves resting the loading platform on the floor, fixing certain props to the ceiling of that particular floor, and removal of the chain sling from the loading platform thereafter. ⁵ Only after the loading platform is properly installed should heavy loads be shifted onto the platform. When loading is done while a loading platform is suspended from a chain sling instead of being properly installed, there is a risk of the loading platform swinging or tilting during the loading process. A shift of the object to be loaded may cause the loading platform to tilt and the object to drop from height. ⁶

8 The fatal accident occurred at about 12.30pm on the same day. At the time, a team of six GSE employees, including Nurun and the two deceased persons, were tasked with rolling the air compressor onto a loading platform at the seventh floor of Tower A. The other members of the team were:

(a) Mr Kamrul Hassan Mohammad Ali (“Kamrul”);

(b) Mr Kashem Abdul (“Kashem”); and

(c) Mr Saiful Islam Sadat Ali (“Saiful”).

I shall refer to Ratan, Rajib, Kamrul, Kashem and Saiful collectively as “the five workers”.

9 The loading platform in question was initially located on the 10th floor of Tower B. Nurun, Ratan and Kashem first proceeded to Tower B to rig the loading platform to the tower crane, while the rest of the workers went to Tower A to remove barricades to enable the loading platform to be lifted to the 7th floor of Tower A. After the loading platform was shifted to Tower A, the three employees at Tower B joined the rest at Tower A. ⁷ Unfortunately, at Tower A, the loading platform was not properly installed and was suspended by the tower crane instead. Due to the fact that it was not properly installed, the loading platform was slanted away from the tower, and the portion of the platform outside the tower was tilted lower than the portion within the tower. ⁸ Just prior to the accident, the two deceased were standing on the loading platform, pulling the air compressor onto the loading platform. Meanwhile the four other employees were attempting to move the air compressor onto the loading platform from outside the loading platform. ⁹ There were wheels at the bottom of the air compressor. ¹⁰ When the air compressor was moved onto the platform, it rolled away from Tower A and off the platform. Ratan and Rajib, who were in the path of the air compressor, could not move away in time and fell out of the loading platform to their deaths seven floors below. All six employees of GSE were not wearing safety harnesses and were not anchored safely to prevent falls from height at the relevant time. ¹¹

10 The following charge under s 15(3A) of the WSHA was brought against Nurun: ¹²

11 GSE was also charged and convicted of an offence under s 12(1) read with s 20 of the WSHA in relation to the same accident. GSE was eventually sentenced to a fine of $250,000 (see Public Prosecutor v GS Engineering & Construction Corp [2017] 3 SLR 682 (“PP v GSE”) at [94]).

12 The district judge found that there were four issues to be decided (see GD at [13]):

(a) Whether, at the material time, Nurun was acting as a foreman and in charge of the group of workers tasked with moving the air compressor at the material time;

(b) Whether Nurun was trained in the installation, dismantling and use of a loading platform, and whether he realised that a loading platform could not be used if it was not installed;

(c) Whether the fact that Nurun was following Rashed’s instruction to refrain from installing the loading platform amounted to reasonable cause for Nurun to direct the workers under his charge to move the air compressor without first installing the loading platform; and

(d) Whether there was a conspiracy by GSE to pin the blame on Nurun for the accident.

13 In relation to the first issue, the district judge found that Nurun held the appointment as a GSE foreman at Tower A of the worksite and he was acting in this capacity at the material time (see GD at [39] and [90]). Nurun was the one who had given the specific instructions to the workers to load the air compressor onto the suspended loading platform in an obviously unsafe manner. Kamrul, Kashem, Saiful and Ratan had all expressed concerns with the method of loading the air compressor, but Nurun insisted that the workers continue, and they complied because Nurun was their foreman (see GD at [122]).

14 The district judge found that the team of workers were initially unsuccessful in pushing the air compressor onto the loading platform. At that point, Nurun made a single call on his mobile phone to ask one Mr Latifur Rahman (“Latifur”), for extra manpower. However, Latifur did not provide any additional manpower. Thereafter, Nurun directed the team to continue with their attempts to move the air compressor onto the uninstalled loading platform (see GD at [182]). It was then that Ratan and Rajib went onto the loading platform and pulled the air compressor whilst the other workers pushed. The workers finally managed to get the air compressor onto the loading platform, and that was when the accident occurred (see GD at [122]). He found that Nurun’s evidence, that there were conversations over a walkie talkie during the loading process between Nurun and Rashed or between Kashem and Rashed, was a fabrication to show that either Rashed or Kashem was the directing mind when the team was moving the air compressor (see GD at [183]).

15 On the second issue, the district judge found that Nurun was formally trained on the installation, dismantling and proper use of a loading platform (see GD at [175]). Alternatively, even assuming the absence of formal training, Nurun was sufficiently experienced in the installation and use of a loading platform, such that he was capable of training other workers under his charge on the proper method to install and use a loading platform (see GD at [176]). As such, Nurun was aware of the danger that Ratan and Rajib were subjected to at the relevant time and was negligent in insisting that the team should continue to load the air compressor onto the uninstalled loading platform (see GD at [180]).

16 On the third issue, the district judge found that the fact that Nurun was given instructions by Rashed not to install the loading platform did not amount to reasonable cause as he knew of the inherent danger of such a task. If he feared repercussions for disobeying Rashed, Nurun should have informed the safety team of this unsafe direction from Rashed (see GD at [184]).

17 On the fourth issue, the district judge did not accept Nurun’s allegation that GSE had conspired to prosecute Nurun for the offence, nor did he accept the contention that Kamrul, Kashem and Saiful were all part of a conspiracy to falsely implicate Nurun (see GD at [187]–[188]).

18 In sentencing the accused, the district judge noted that the Prosecution had submitted for a custodial sentence of at least four weeks’ imprisonment (see GD at [197]). He highlighted that the Prosecution relied on the precedent set out in Public Prosecutor v Hue An Li [2014] 4 SLR 616 (“Hue An Li”). The district judge agreed with the Prosecution that the primary sentencing consideration in cases involving s 15(3A) of the WSHA was general deterrence (see GD at [205]). However, he did not think that the precedents for causing death by a negligent act under s 304A(b) of the Penal Code (Cap 224, 2008 Rev Ed), including Hue An Li were applicable (see GD at [208]). Although he assessed that the potential for harm in the present case was high, he found that Nurun’s culpability was lower than GSE, which was found in PP v GSE to be medium to high (see GD at [209]). The district judge agreed with the defence’s submission that no custodial sentence had been imposed in all prior cases under s 15(3A) decided after Hue An Li which involved death. Hence, he concluded that the custodial threshold had not been crossed. Based on the state court precedents cited, he then sentenced Nurun to a fine of $15,000, in default two months’ imprisonment (see GD at [210]–[211]).

19 Section 15(3A) of the WSHA reads:

20  As seen from the above, there are three elements to the offence. First, the accused must do an act which endangers the safety or health of himself or others. Second, the act must be negligent. Third, the act must be done without reasonable cause.

21 Counsel for Nurun, Mr Anil Narain Balchandani (“Mr Balchandani”) has launched a sweeping attack on various aspects of the district judge’s decision which touches on all three elements of the offence. In the interest of clarity, I have distilled the objections into three broad issues:

(a) Whether the district judge erred in finding that the five workers were under Nurun’s charge and Nurun had instructed the workers to load the air compressor onto the loading platform;

(b) Whether the district judge erred in finding that Nurun was negligent;

(c) Whether the district judge erred in finding that there was no reasonable cause.

22 The Court of Appeal in Haliffie bin Mamat v Public Prosecutor and other appeals [2016] 5 SLR 636 (“Haliffie”) at [31] has cautioned that the role of the appellate court “is not to re-assess the evidence in the same way a trial judge would”. As highlighted in Haliffie at [32], the appellate court is restricted to considering:

(a) whether the assessment of witness credibility is “plainly wrong or against the weight of the evidence”;

(b) whether the “verdict is wrong in law and therefore unreasonable”; and

(c) whether the “decision is inconsistent with the material objective evidence on record”, bearing in mind that an appellate court is in as good a position to assess the internal and external consistency of the witnesses’ evidence, and to draw the necessary inferences of fact from the circumstances of the case.

23 It is with these principles in mind that I turn to consider the first issue.

24 Mr Balchandani argues that the district judge was wrong to accept the testimony of the “employee-witnesses of GSE”, ¹³ in making his findings of fact. Therefore, the findings of the district judge, which were primarily based on the testimony of these witnesses, ought to be overturned. The “employee-witnesses” refers to Kamrul, Kashem and Saiful. This argument has two prongs:

(a) First, the district judge did not pay proper attention to the argument that there was a “ploy” by GSE to implicate Nurun; ¹⁴ and

(b) Second, the district judge did not address his mind to the inconsistencies in the testimonies of the witnesses. ¹⁵

25 I will deal with each of these arguments in turn.

26 Mr Balchandani’s argument is that Kamrul, Kashem and Saiful had colluded to give false evidence as part of a conspiracy engineered by GSE to implicated Nurun. The court in XP v Public Prosecutor [2008] 4 SLR (R) 686 at [21] set out the law on collusion in these terms:

27 Hence, the first question in considering an allegation of collusion is whether there is sufficient evidence of a motive to falsely implicate the accused such that a reasonable doubt has been raised. In my view, the district judge did not err in finding that there was no motive on the part of Kamrul, Kashem and Saiful to falsely implicate Nurun and he was correct in disregarding the allegation of collusion.

28 The premise of Mr Balchandani’s argument is that GSE had something to gain by engineering a conspiracy to implicate Nurun. However, by the time of Nurun’s trial, GSE had already pleaded guilty and admitted to a statement of facts in relation to the case against it. In fact, judgment for the appeal in PP v GSE was delivered before the trial had concluded. Findings made on the extent of Nurun’s involvement in the matter had no bearing on GSE’s criminal liability. Hence, it is unclear what GSE stood to gain from convincing the witnesses to give false testimony.

29 Leaving aside GSE’s motive, it is even less clear what motive Kamrul, Kashem or Saiful would have for colluding to falsely implicate Nurun. To the extent that Mr Balchandani was suggesting that GSE was exerting some influence as the employer of the three workers in inducing them to give false testimony, it must be highlighted that Kashem and Saiful were no longer working for GSE at the time they gave evidence. ¹⁶ Moreover, there was no evidence raised that showed that any of the witnesses bore any personal grudge against the accused. In fact, the evidence available appeared to suggest that their relationship with Nurun was respectful or at the very least amicable. Kamrul, Kashem and Saiful all addressed Nurun with the honorific “bhai”, which meant “brother”. ¹⁷

30 Mr Balchandani’s response is to suggest that the ploy to implicate Nurun was hatched and engineered by GSE at the start of investigations, and Kamrul, Kashem and Saiful merely decided to continue on with the “charade” at trial. ¹⁸ I do not think this unsupported allegation is a sufficient basis to ground a motive to falsely implicate Nurun on the part of Kamrul, Kashem and Saiful. Bearing in mind that GSE no longer had anything to gain by the time of trial, it would be unusual that the witnesses, some of whom had already left the employ of GSE at the time of trial, would decide to collude and risk perjury in order to support a cause that had expired. In the absence of any supporting evidence of motive to lend weight to the suggestion that there was a conspiracy, I find that the district judge did not err in rejecting the argument that there was a ploy on the part of GSE.

31 Mr Balchandani points to several inconsistencies in the evidence of Kashem, Kamrul and Saiful, including:

(a) Who they were with prior to engaging in the task of moving the air compressor; ¹⁹

(b) Who had conveyed the initial instruction to gather at Tower A and Tower B and when this instruction was conveyed; ²⁰

32 The difficulty with this argument is that these inconsistencies do not go towards the key aspects of the charge, but rather, events that happened prior to the accident. The inconsistencies do not relate to Nurun’s act of instructing the five workers, on the 7th floor of Tower A, to move the air compressor onto the uninstalled loading platform. These inconsistencies also do not relate to whether the workers were under Nurun’s charge. It is not disputed that the evidence of Kashem, Kamrul and Saiful was consistent on the key aspects: that Nurun was in charge of the workers and was the individual who instructed the workers to move the air compressor onto the uninstalled loading platform. ²¹

33 Therefore, Mr Balchandani’s challenge is how to link these inconsistencies to the charge faced by Nurun. He does so by suggesting that these inconsistencies “speak to a larger plot to absolve Rashed from instructing them and to frame [Nurun] for the Accident”. ²² I reject this argument. In the first place, there is no evidence of a motive for Kashem, Kamrul and Saiful to frame Nurun (see [28]–[30] above). Even more tenuous is the suggestion that the witnesses wanted to absolve Rashed from responsibility. Mr Balchandani did not point to any evidence that suggested Kashem, Kamrul and Saiful had this desire. It is also unclear why absolving Rashed from responsibility was an achievable or desirable outcome. It is stated in the agreed statement of facts tendered at the start of trial that it was Rashed that gave the first instruction to Nurun not to install the loading platform. Rashed’s culpability in this regard was clearly established. Moreover, it appears that Rashed had escaped from Singapore sometime before the trial. Hence, there was nothing to be gained by them from an attempt to absolve Rashed.

34 These problems with establishing the motivation of Kashem, Kamrul and Saiful strike at the heart of Mr Balchandani’s argument. Without evidence of a motive to frame Nurun for the accident, the inconsistencies are equivocal at best in suggesting that the witnesses had colluded to frame Nurun, and could very plausibly be explained by the fact that the witnesses’ memory on what they perceived as unimportant incidents had faded with the passage of time. I would add that the district judge, who had the benefit of hearing the accounts of the witnesses first-hand, assessed Kamrul, Kashem and Saiful to be credible witnesses (see GD at [190]) and his assessment of their credibility does not appear to me to be plainly wrong or against the weight of the evidence.

35 In addition, the district judge noted that Nurun was not a credible witness (see GD at [191]). I agree with the district judge that Nurun’s evidence in respect of his involvement appeared inconsistent. Nurun’s case was that he was not in charge of the five workers at all and he was merely a general worker at the time. However, his testimony vacillated between suggesting that Kashem was the individual in charge of directing the task of loading the air compressor, ²³ and having no one in charge of the task. ²⁴ Aside from his vacillating account, Nurun also maintained that he was in contact with Rashed during the task of loading the air compressor, and that on Rashed’s instructions, he had contacted Latifur to seek additional manpower when the task appeared too difficult. ²⁵ It was also an undisputed fact that Nurun received the instruction from Rashed that he was not to install the loading platform. ²⁶ If it were true that Nurun was merely a general worker at the time, it would be unlikely that he would have been the worker liaising with Latifur and Rashed. Additionally, at the time of the accident, Nurun was wearing a white hat (as opposed to the yellow hats worn by general workers) and received extra pay from GSE, which strongly indicated that he was acting in the role of a foreman rather than the run-of-the-mill general worker. ²⁷

36 At this juncture, I highlight that Mr Balchandani also argues that the Prosecution had breached its disclosure obligation (“Kadar disclosure obligation”) as set out in the cases of Muhammad bin Kadar and another v Public Prosecutor [2011] 3 SLR 1205 (“Kadar I”) and Muhammad bin Kadar and another v Public Prosecutor and another matter [2011] 4 SLR 791 (“Kadar II”).

37 At the trial below, it emerged during the testimony of the investigation officer for the case, Ms Chew Siew Huang (“IO Chew”), that Rashed had given two statements. The first statement by Rashed was a denial that he had instructed Nurun not to install the loading platform while the second statement by Rashed included an admission that he was the one who had instructed Nurun not to install the loading platform. ²⁸ Mr Balchandani hence made an application before the district judge for the Prosecution to disclose Rashed’s second statement pursuant to the Prosecution’s Kadar disclosure obligation. This application was rejected, and Mr Balchandani now argues that the district judge erred in applying Kadar I and Kadar II. ²⁹

38 In Lee Siew Boon Winston v Public Prosecutor [2015] 4 SLR 1184 (“Winston Lee”) at [158], in the context of a similar argument on appeal that the Prosecution had breached its Kadar disclosure obligation, I examined Kadar I and Kadar II and made the following observation:

39 I noted that in light of constitutional status of the office of the Attorney-General, there was a presumption of legality or regularity in the context of prosecutorial decisions, including in relation to the Prosecution’s duty of disclosure under its Kadar disclosure obligation (see Winston Lee at [167]–[169]). However, the presumption of legality or regularity could be displaced in appropriate circumstances. The threshold to displace the presumption is that there must be reasonable grounds to believe that the Prosecution has in possession material which should be disclosed (see Winston Lee at [170]–[175]).

40 In the present case, the presumption of legality or regularity has not been displaced. The only indication that Rashed’s second statement includes “material that tends to undermine the prosecution’s case or strengthen the defence’s case” is that IO Chew had stated that Rashed admitted to instructing Nurun not to install the loading platform. Crucially however, the fact that Rashed was the individual that instructed Nurun not to install the loading platform was not in dispute. In fact, this concession was part of the agreed statement of facts tendered to the court at the beginning of trial. ³⁰ As such, it appears to me that the second statement is at best neutral to Nurun’s defence and there were no reasonable grounds to believe that the second statement was material which ought to have been disclosed.

41  Therefore, I find that the district judge’s findings of fact in relation to the material sequence of events that led to the accident (see [13]–[14] above) were not against the weight of the evidence. He did not err in finding that Nurun was the one in charge of the five workers, and had instructed them to move the air compressor onto the uninstalled loading platform, despite the safety concerns expressed by some of the workers.

42 Mr Balchandani submits that the district judge erred in finding that Nurun was formally trained in the installation of the loading platform. ³¹

43 There is some force in this argument. In coming to his finding that Nurun was formally trained, the district judge placed significant reliance on two training attendance forms marked as P25 ³² and P27 ³³ respectively (see GD at [164]). P25 and P27 were ostensibly copies of the same training attendance form that demonstrated that Nurun had attended training for the installation, dismantling and proper use of a loading platform. The district judge found that IO Chew had obtained P25 from GSE sometime in February 2014 during the course of investigations. P27 was obtained during trial from Mr Daniel Woo Chin Chern (“Daniel”), the corporate safety manager of a company called Sante Machinery Pte Ltd (“Sante”). ³⁴ Sante was the company involved in the training of GSE employees in the installation, dismantling and proper use of a loading platform.

44 Mr Md Rakibul Hasan Late Golam (“Rakibul”) gave evidence in relation to training attendance forms such as P25 and P27. Rakibul was a supervisor and trainer with Sante. Rakibul stated that after training was given in relation to loading platforms, a blank training attendance form would be filled up with the name and details of the attendees. The trainer would then fill in his name and sign the form, keep the original copy of the form, and pass a carbon copy of the form to GSE. ³⁵ I note that in Daniel’s evidence, he stated that Sante would keep two copies and pass the original to GSE. ³⁶ Regardless, the witnesses were consistent in that a completed form would be evidence of a particular employee’s attendance in a training session.

45 While Nurun’s name and work permit number can be found on both P25 and P27, there were several troubling issues with P25 and P27.

46 First, even though there was a signature column next to the column reserved for the attendees’ details, the signature column corresponding to Nurun’s details in both P25 and P27 was left blank. The section of the form that was meant for the trainer’s name and signature was also left blank. Rakibul steadfastly insisted during cross examination ³⁷ and during questioning from the district judge ³⁸ that without a signature by the purported attendee, Rakibul would not regard that individual as having attended the training course. Rakibul also highlighted that a form without the trainer’s name and signature also was “not a complete form” and could not be relied upon. ³⁹ Presumably this was a suggestion that if those portions were not filled in, it would mean that no training had taken place. The district judge was cognisant of these difficulties but accepted the evidence of Daniel in relation to this issue. Daniel speculated that the trainer for the training session in question might have mistakenly failed to sign the form, but the fact that the names of the participants were filled in, GSE stamps were found on the forms and Sante had a copy of the form in its possession, showed that training had been conducted (see GD at [168]–[169]). Thus the assumption was that the trainer, who was never identified, was merely careless in filling up the form, and Nurun was similarly neglectful in failing to insert his signature but nevertheless Nurun’s training had taken place.

47 Second, leaving aside the gaps common in both forms, it was immediately apparent that P25 and P27 were not identical copies of each other. I list some of the differences between the two forms below: ⁴⁰

(a) There was the addition of Rashed’s name and signature in P25. The section where Rashed’s name and signature was added in P25 was empty in P27.

(b) P25 was only single sided, but P27 was double sided. The reverse face of P27 was a brief on the safety procedures that had to be undertaken when using the loading platform. It appears that P25 was merely a colour photocopy of the original form that would have been given to GSE by Sante after the completion of the training (see GD at [165]).

(c) The signature of one GSE representative by the name of “Jalal” was missing from P27 but was found in P25.

48 None of the witnesses could provide an explanation as to why these differences existed. In light of these differences in what were supposed to be identical copies, there is a distinct possibility that at least one of the forms had been subsequently amended sometime after the training had concluded, although I would not speculate as to the possible reasons for doing so.

49 In light of the difficulties highlighted above, I find that P25 and P27 cannot be relied upon to establish that Nurun had undergone formal training for the installation of the loading platform. However, this does not mean that the district judge erred in finding that Nurun was negligent.

50 The test for negligence is objective. The court must “consider whether a reasonable man in the same circumstances would have been aware of the likelihood of damage or injury to others” (see Ng Keng Yong v Public Prosecutor and another appeal [2004] 4 SLR (R) 89 at [88]). In the present case, it is clear that in the circumstances, a reasonable man would have been aware of the likelihood of death or injury to others. I highlight five points in particular:

(a) First, the air compressor was of significant heft. Six construction workers had great difficulty moving the air compressor. The district judge’s finding was that the workers had initially failed at the task of moving the air compressor onto the loading platform.

(b) Second, the loading platform was suspended by a chain sling seven storeys above ground level and was tilting away from Tower A;

(c) Third, wheels were attached to the bottom of the air compressor;

(d) Fourth, none of the workers were wearing safety harnesses to prevent falls from height; and

(e) Fifth, some of the workers had already expressed safety concerns with loading the air compressor onto the uninstalled loading platform.

51 Put together, the points above suggest that it was patently clear that the act of instructing the five workers to load the heavy air compressor onto an uninstalled loading platform was an unsafe act and would risk death or injury to others. There was a myriad of ways in which an accident could have occurred. Quite apart from the risk of death or injury in relation to the five workers, which did eventuate in the actual deaths of Ratan and Rajib, there was also a likelihood of the air compressor (with wheels attached at the bottom) rolling off the tilted loading platform and landing on the ground below, injuring or even killing other construction workers.

52 I would add that Nurun himself admitted at trial that he had loaded objects onto loading platforms on prior occasions. In those instances, the loading platform in question had always been installed. Hence, he knew that a loading platform must be properly installed before loading objects onto the platform. ⁴¹ He also admitted that he knew that it was not safe to use an uninstalled loading platform when it is simply being suspended via chain slings connected to a crane. He even gave details of the process of how an installed loading platform would have to be certified by a professional engineer before the workers would be allowed to use the platform for loading. ⁴² He conceded that it was obvious that loading objects onto an uninstalled loading platform was very dangerous. Hence, leaving aside the issue of his formal training, the evidence is clear that Nurun was sufficiently well informed to understand the grave danger of loading heavy objects onto an uninstalled loading platform.

53 Therefore, the district judge did not err in finding that Nurun was negligent. Nurun ought to have known of the risk of death or injury and this was sufficient to make out negligence on his part. Above and beyond that, Nurun did in fact know of the risk of causing death or injury from his actions.

54 Mr Balchandani’s main objection in relation to this issue is that Nurun was “not acting on his own volition” and “conducted himself because he was told to do so by [Rashed]”. ⁴³ The argument is that Nurun was pressured by Rashed to act as he did and hence this amounted to reasonable cause. The district judge hence erred in disregarding this pressure. In my view, the factual and legal premise with this objection is flawed.

55 The evidence on record clearly contradicts the argument that Nurun was not in a position to make his own decisions and was bound to follow the instructions of Rashed. At trial, Nurun agreed that if a foreman received an unsafe work order from a supervisor, it was the duty of the foreman to stop work and inform the safety team. ⁴⁴ Nurun also stated that the foreman had the responsibility to assess whether a particular work procedure was safe. ⁴⁵ Nurun’s own evidence thus demonstrated that he was aware that a foreman could and should refuse to follow unsafe work orders, even if the orders were given by a supervisor. His only excuse was that he was not acting in the capacity of a foreman at the relevant time. Given the finding that Nurun was the foreman in charge of the workers at the relevant time (see [41] above), there is no basis to suggest that Nurun regarded himself as bound to follow the patently unsafe orders of Rashed.

56 Even taking Mr Balchandani’s case at its highest, and operating on the assumption that Nurun was under the misapprehension that he was bound to follow Rashed’s orders, the suggestion that as a matter of law, pressure from a superior can amount to reasonable cause is also questionable. Mr Balchandani does not point to any statutory provision or portion of the relevant parliamentary debates which suggests that Parliament intended for such effect. Section 15 of the WSHA is silent on what amounts to reasonable cause. However, the other provisions of the WSHA and the relevant parliamentary debates suggest that Parliament did not intend for superior orders to amount to reasonable cause. Section 10 of the WSHA states:

57 In the parliamentary debates on the enactment of WSHA, there were various indications that the intention behind the enactment of the WSHA was to ensure that all stakeholders, including rank-and-file workers, would be held responsible for workplace safety (see Singapore Parliamentary Debates, Official Report (17 January 2006), Vol 80 at cols 2206, 2211 and 2217 (Dr Ng Eng Hen, Minister for Manpower)):

58 From the above, it is clear that Parliament intended to effect a “cultural change” by expanding the liability framework to render even rank-and-file workers responsible for unsafe work practices. This responsibility is not “diminished or affected” by the acts of another employee or superior. In my view, it would be inconsistent with this parliamentary intention to find that an employee would be completely exonerated from liability under s 15(3A) of the WSHA, merely because the employee was following work orders. This is especially the case when the work orders in question were patently unsafe, and the employee in question knew that the work orders were patently unsafe. Hence, the district judge did not err in finding that Nurun had no reasonable cause to act as he did.

59 Therefore, for the reasons given above, I find that there is no basis to overturn the conviction and dismiss the appeal against conviction. I now turn to the issue of sentence.

60 This is the first time an offence under s 15(3A) of the WSHA has been brought before the High Court. As a result, this appeal raises several fresh issues of law, particularly in relation to sentencing. To assist the court in coming to its decision, Mr Tan was appointed as Young Amicus Curiae to answer the following questions:

(a) Whether the principles laid down in PP v Hue An Li [2014] 4 SLR 661 (“Hue An Li”) apply to an offence under s 15(3A) of the WSHA;

(b) When the custodial threshold for s 15(3A) of the WSHA offence is crossed; and

(c) What is the appropriate sentencing framework to assist in calibrating the sentence once the custodial threshold for the offence has been crossed?

61 I thank Mr Tan for his detailed research and helpful submissions which have greatly assisted the court. The parties have also provided their own helpful submissions to address the questions posed, and I thank them for the same.

62 As highlighted, Nurun was sentenced to a fine of $15,000, in default two months’ imprisonment. Mr Balchandani submits that the sentence is manifestly excessive. The main thrust of his argument is premised on the past sentencing practice in relation to s 15(3A) of the WSHA. ⁴⁶ Based on these precedents, Mr Balchandani submits for a sentence in the amount of an $8,000 fine, with a three weeks’ imprisonment term in default.

63 The Prosecution submits that the sentence is manifestly inadequate. It argues that the current State Court precedents need to be reviewed as the precedents do not sufficiently utilise the available sentencing range and do not give effect to the legislative intent behind the enactment of s 15(3A) of the WSHA. It invites the court to provide appropriate sentencing guidelines in light of this. ⁴⁷ For reasons largely similar to the Prosecution, Mr Tan also takes the view that the sentencing practice ought to be relooked. ⁴⁸

64  The Prosecution and Mr Tan have each proposed a separate sentencing framework for s 15(3A) of the WSHA. I will discuss both proposals in greater detail in the appropriate section below. Based on the sentencing framework proposed, the Prosecution submits for an imprisonment term of at least 12 months.

65 In light of these arguments, there are three issues to consider:

(a) Whether the existing sentencing practice ought to be reviewed;

(b) If so, what would be the appropriate sentencing framework; and

(c) The appropriate sentence in the present case.

66 I now turn to address these issues in turn.

67 I am in agreement with the Prosecution and Mr Tan that the existing sentencing practice ought to be reviewed.

68 The statutory maximum sentence stipulated for an offence signals the gravity with which Parliament views any individual offence. Hence, a sentencing judge ought to take the maximum sentence into account when calibrating the appropriate sentence, and apply his mind to determine precisely where the accused person’s conduct falls within the entire range of punishment devised by Parliament (see Poh Boon Kiat v Public Prosecutor [2014] 4 SLR 892 at [60]). The courts have cautioned against a situation where only a segment of the possible sentencing range has been utilised. This would run the risk of inadvertently usurping the legislative policy in setting down a particular range of sentences (see Ong Chee Eng v Public Prosecutor [2012] 3 SLR 776 at [24]).

69 The Prosecution and Mr Tan have helpfully compiled a list of all the sentences that have been passed in the State Courts in relation to offences charged under s 15(3A) of the WSHA that have resulted in death. ⁴⁹ The provision was enacted seven years ago and there have been at least 17 such fatal accident cases since the enactment of the provision. Despite this, there has not been a single case where a custodial sentence has been imposed for a s 15(3A) offence.

70 However, this fact alone does not inexorably lead to the conclusion that the full sentencing range has not been utilised. The possibility exists that the conduct of the accused persons in all these 17 cases might have fallen within the lower range in terms of the appropriate sentence. Upon investigation, I conclude that this is not the case. Both the Prosecution and Mr Tan highlight several cases where the accused persons appeared to have a relatively high degree of culpability, committed breaches which created a high potential for harm, and where the breaches resulted in death, and yet all accused persons received fines in the region of $5,000 to $12,000. Thus I agree that the full sentencing range for this particular offence has not been fully utilised.

71 On a related note, the fact that the sentences for this offence have clustered around the lower end of the sentencing range provided by Parliament also does not give effect to the legislative intent behind the enactment of s 15(3A) of the WSHA. The legislative intent of a statutory provision is an important factor in considering the appropriate sentence (see PP v GSE at [44] to [48]).

72 The legislative intent of the WSHA, expressed when it was originally enacted in 2006, is to deter risk-taking behaviour and protect workers and the members of the public. I highlight the relevant portions of the Minister for Manpower’s speech at the second reading of the Workplace Safety and Health Bill (No 36 of 2005) below (see Singapore Parliamentary Debates, Official Report (17 January 2006), Vol 80 at cols 2206 and 2214–2216 (Dr Ng Eng Hen, Minister for Manpower)):

73 I highlight three points from this speech:

(a) First, the WSHA was enacted to improve safety at the workplace and effect a “cultural change”, in part through ensuring that penalties for non-compliance are sufficiently high to deter risk-taking behaviour.

(b) Second, the penalty regime of the WSHA has a wide ambit. It goes beyond situations where actual harm has been done. This was a lacuna in the previous legislation which Parliament expressly wished to address. Hence, the penalty regime under the WSHA provides for meaningful penalties where there are severe lapses but no accidents have occurred.

(c) Third, it was also expressly recognised in Parliament that the penalty to be applied in any particular case would take into account the culpability of the offender, the potential harm that could have been caused, and the harm actually done although a single maximum penalty is prescribed under the law.

74 Section 15 of the WSHA originally did not contain a provision for negligent acts of endangerment. Section 15(3A) was only introduced in 2011. The Prosecution and Mr Tan both submit that there is nothing in the relevant parliamentary debates that suggests that Parliament intended for a departure in any of the general legislative aims of the WSHA when it enacted the s 15(3A) offence. ⁵⁰ Having reviewed the relevant parliamentary material, I agree with this submission.

75 Therefore, this is an opportune moment to review the sentencing practice under s 15(3A) of the WSHA and provide a sentencing framework for guidance. The current sentencing practice has resulted in a clustering of sentences at the lower end of the sentencing range. The full sentencing range has not been considered. Additionally, this clustering of sentences does not accord with Parliament’s intention to effect a cultural change and deter risk taking behaviour. I now turn to address the question of the appropriate sentencing framework.

76 The Prosecution proposes a sentencing framework largely adapted from the framework set out in PP v GSE in relation to an offence under s 12(1), read with s 20 and punishable under s 50(b) of the WSHA (“s 12(1) offence”). The framework has two stages. The first stage involves determining a starting point for sentencing based on the principal factual elements of an offence under s 15(3A), namely, the potential harm resulting from the offender’s act and the culpability of the offender. The levels of potential for harm and culpability are divided into three categories: low, medium and high. The second stage involves adjusting the starting point based on aggravating and mitigating factors. The Prosecution sets out a table of sentencing ranges for offenders being punished for a s 15(3A) offence in a situation where the offender has claimed trial as follows: ⁵¹

77 The values in the table have been calibrated by transposing the ranges set out in the case of PP v GSE for a s 12(1) offence onto the sentencing range available under s 15 (3A) of the WSHA. Based on this table, potential for harm has been given greater weight in influencing the sentence starting point as opposed to culpability.

78 Mr Tan also proposes a similar two stage approach. However, he provides a slightly different list of factors for potential for harm and culpability. Unlike the Prosecution, he provides for equal weightage to be given to potential for harm and culpability. I set out his table of sentencing ranges below: ⁵²

79 I broadly agree with the two stage approach. In particular, I agree that potential for harm and culpability should be the two principal factual elements in determining a starting sentence before being further calibrated based on the aggravating and mitigating factors. However, I do not agree with the proposed tables of sentencing ranges. Before discussing the appropriate table of sentencing ranges, I will first discuss the relevance of actual harm in the calibration of the appropriate sentence, and set out some non-exhaustive factors to consider in assessing potential for harm and culpability.

80 Although s 15(3A) of the WSHA criminalises the risk of harm, ie, the endangerment of others, instead of the materialisation of the risk itself, Mr Balchandani, the Prosecution and Mr Tan all agree that actual harm should be factored into the calibration of the sentence. ⁵³ This is consistent with the express statements made in Parliament that actual harm ought to be taken into account in sentencing (see [73(c)] above).

81 This is also consistent with the approach taken in other offences which also criminalise the risk of harm as opposed to actual harm. In Stansilas Fabian Kester v Public Prosecutor [2017] 5 SLR 755 in the context of an offence under s 67 of the Road Traffic Act (Cap 276, 2004 Rev Ed) (“RTA”), the court emphasised at [40] that the fact that an offender had caused property damage, injury or even death would be an aggravating factor, and the weight of the factor would depend on the nature and magnitude of the damage or harm caused and all other relevant circumstances. Notably, s 67 of the RTA criminalises the risk posed to other road users by a driver who is incapable of having full control of his vehicle, instead of criminalising actual harm itself.

82 The theoretical underpinning for taking the actual harm caused into account in calibrating a sentence, even where an offender is negligent, is the fact that Singapore courts have endorsed the position that the outcome materiality principle trumps the control principle in the context of criminal negligence. The outcome materiality principle is the “intuitive moral sense that outcomes do matter”, and hence the outcome of any criminal act, ie the actual harm caused, should be taken into account in sentencing (see Hue An Li at [70]). The control principle is the “intuitive moral sense that people should not be morally assessed for what is not their fault” (see Hue An Li at [69]). This tension between the two principles is particularly pronounced in situations where negligence is criminalised, because an offender is generally in less control of the result of his/her negligent act as opposed to situations where criminal acts are committed intentionally. The court in Hue An Li gave three reasons why outcome materiality trumps the control principle in the context of criminal negligence. The first was that the provisions in the Penal Code which criminalise negligent conduct are predicated on outcome materiality (see Hue An Li at [71]). Hue An Li was a case involving causing death through a negligent act under s 304A of the Penal Code (Cap 224, 2008 Rev Ed). The second and third reason was stated as such (see Hue An Li at [73] and [74]):

83 The district judge was of the view that “there was no correlation between the sentencing principles in Hue An Li and the present case” (see GD at [208]). I respectfully disagree. While, as rightly noted by the district judge, the first reason given by the court in Hue An Li does not apply in the present case, the second and third reasons apply with full force.

84 Hence, in the context of s 15(3A) of the WSHA, actual harm is relevant as an aggravating factor. The nature and magnitude of the harm caused is relevant in determining the weight to be ascribed to it as a factor in sentencing. In cases where serious harm or death is caused, the sentence should be correspondingly increased to mark the gravity of the offence and the social and economic costs inflicted on society.

85 At this juncture, I highlight that actual harm is to be factored into the two stage framework at the second stage instead of treating it as one of the factors to be taken into consideration when calibrating the appropriate level of potential for harm in the first stage. This is because the first stage of the analysis is concerned with the principal factual elements of s 15(3A), which means that the elements in the first stage analysis must be present in all s 15(3A) offences under consideration. Section 15(3A) criminalises the potential for harm as opposed to actual harm and it is conceivable that not every s 15(3A) offence would result in actual harm. Therefore, actual harm cannot be described as a principal factual element of the offence and it properly belongs to the second stage of analysis.

86 I set out below a non-exhaustive list of factors to be considered in assessing the level of potential harm:

(a) Seriousness of the harm risked;

(b) Likelihood of that harm arising; and

(c) Number of people likely to be exposed to the risk of that harm.

87 As for culpability, an inquiry into an offender’s culpability will involve a holistic assessment of all the circumstances, having particular regard to the nature and circumstances surrounding the unsafe act. Some relevant factors, again not meant to be exhaustive, are as follows:

(a) Nature of the unsafe act;

(b) The number of unsafe acts committed by the offender;

(c) The level of deviation from established procedure.

88 In relation to the tables of sentencing ranges proposed by the Prosecution and Mr Tan (see above at [76] and [78]), I do not agree with the tables proposed for three reasons.

89 First, in relation to the Prosecution’s proposed table, there is a jump in the proposed starting points when one moves from low to medium and from medium to high potential for harm. Taking the example of two offenders who both have the same low culpability, and considering the position if one offender has low potential for harm, and the other offender has medium potential for harm, it can be seen that the sentence for each offender does not smoothly increase, but suddenly leaps from a maximum starting point of a $6,000 fine to a minimum starting point of an $18,000 fine. I am of the view that such a gap arbitrarily restricts the sentencing court from providing certain sentences as starting point. Additionally, as a matter of conceptual clarity, a slight increase in potential for harm, which increases the level of potential for harm from the upper limits of low potential harm to the lowest point of medium potential for harm, should not result in a sudden jump in the starting sentences.

90 Second, in relation to Mr Tan’s table, I am of the view that the Prosecution is right to give greater weight to potential for harm, as opposed to equal weights to both principal factual elements. This is an acknowledgment to the policy behind the WSHA which seeks to deter risk-taking behaviour and give meaningful penalties where there are severe lapses (see [73(a)]–[73(b)] above). An unsafe act done negligently in a workplace and with a low degree of culpability but yet exposes many persons to the risk of very serious injuries should be regarded as far more serious and therefore calls for much greater deterrence than an unsafe act done similarly negligently but with a high degree of culpability and which exposes very few persons to the risk of only minor injuries.

91 Third, the tables provided do not, in my view, adequately consider the impact of the s 15(3) of the WSHA. Section 15(3) of the WSHA reads:

The punishment provision for s 15(3) can be found in s 50(a) of the WSHA which reads:

As can be seen, s 15(3) also provides for the same maximum imprisonment term of two years, for the exact same act, endangerment of the safety and health of others. However, s 15(3) criminalises the far more culpable mens rea of wilful or reckless behaviour. This raises an issue of how to ensure proportionality in sentencing between offenders who commit the same acts, but with very different mens rea.

92 Therefore, after careful consideration, I have crafted a table of sentencing ranges for s 15(3A) taking into account the three issues highlighted above. In Annex A-1 and A-2, I have also provided a preliminary table for the offences of reckless endangerment and wilful endangerment under s 15 of the WSHA. I will explain the table after I reproduce it below in landscape format for ease of reference:

93  Let me elaborate on how the table may be used in sentencing.

94 As a broad description of the table:

(a) The table is subdivided into a set of three broad bands of “low”, “medium” and “high” to mark out generally the various continuously increasing levels of potential for harm. These three bands are intersected by another set of three broad bands of “low”, “medium” and “high” again to delineate broadly the various continuously increasing levels of culpability. The result is a grid of nine large boxes as shown on the table above.

(b) The factors of potential for harm and culpability are regarded as two continuous independent variables that each separately and independently affect the sentence. Any increase in culpability is represented by a rightward movement on the table (or rightwards along the “x” axis), and any increase in potential for harm is represented by an upward movement on the table (or upwards along the “y” axis). Since the sentence should increase for any given increase in potential for harm and culpability, rightward and upward movements in the table (ie further along the “x” and “y” axis away from the point of zero culpability and zero potential for harm point) also necessarily represent an increase in sentence.

(c) The extreme edges of the table (ie the “y” axis at the extreme left of the table versus the extreme right of the table; the “x” axis at the bottom of the table versus the extreme top end of the table) indicate the polar opposites of no sentence and the maximum sentence respectively as a starting point. The bottom and left edges of the table (representing all the points falling on the “x” axis or “y” axis respectively) represent a situation where the conduct of the accused has no potential for harm or no culpability respectively. Where there is no potential for harm or no culpability, no offence has been committed and accordingly, there would be no sentence (which is represented by the value “0” weeks for all points falling on the “x” axis or “y” axis itself). At the right and top edges of the table, I have set limits within which the vast majority of the cases that are likely to come before the courts would belong (ie most cases will fall within the grid of nine large boxes). However, if the offence committed is extremely egregious in terms of the potential for harm or culpability, the starting point can extend to the maximum sentence provided by law. This is reflected by the shaded portions of the table.

(d) The sentences in brackets represent the sentencing ranges within each of the nine large boxes.

(e) The “X” marks generally the midpoint of each of the nine large boxes within the grid of three broad bands for potential for harm intersected by three broad bands for culpability. The sentences for each midpoint is reflected within the ovals directly above the “X”.

95 The table is only meant to assist in determining the starting point for the sentence based on the appropriate levels of potential for harm (and not actual harm caused) and culpability at the first stage. The final sentence will be calibrated based on adjusting the starting point reached with appropriate weight given to the various aggravating and mitigating factors at the second stage. Hence, in this table, the impact of actual harm, eg, death and serious injuries caused, has not yet been taken into account. Actual harm, if caused, and the seriousness of that harm caused will only be taken into account as an aggravating factor at the second stage. Any early plea of guilt from the offender is to be taken into account as a mitigating factor at the second stage after the starting point for the sentence has been determined. This is the reason why the table is only meant to reflect a situation where the offender has claimed trial.

96 The sentencing ranges in the brackets are listed in terms of weeks. However, the sentences of fines and imprisonment are interchangeable with a notional conversion rate of 1 weeks’ imprisonment being convertible to a fine of $5,000. This notional conversion rate is based on the following:

(a) Section 319(d)(i) of the Criminal Procedure Code (Cap 68, 2012 Rev Ed), which applies to s 15(3A) as well as s 15(3) of the WSHA, provides that the maximum term of imprisonment in default of a fine cannot exceed one half of the maximum term of imprisonment for the offence.

(b) This means that if the maximum fine of $200,000 is imposed for a s 15(3) offence, the maximum term of default imprisonment that can be given is one year.

(c) Taking a year to be approximately 52 weeks and assuming the maximum fine is given, every failure to pay $5,000 in fines can result in a maximum of 1.3 weeks’ imprisonment in default. I have thus chosen a notional conversion rate somewhere in the same ballpark, but slightly lower, to reflect the fact that in default imprisonment terms are distinct from imprisonment terms.

(d) I have decided to use the figure for s 15(3) instead of s 15(3A) to ensure consistency between the two offences which have the same maximum imprisonment term and criminalise the same act, albeit with different mental elements. Additionally, I regard the equivalent figure derived from the maximum fine of s 15(3A) as too low to form a useful notional conversion rate.

I have indicated a notional conversion rate between the sentence of a fine and an imprisonment term to provide some guidance to the courts with the aim of achieving more consistent decision making should the courts exercise greater flexibility in sentencing by giving a combination of a fine and an imprisonment term, which is a punishment that is possible under s 15(3A).

97 The custodial threshold will generally be crossed for offences under s 15(3A) when the appropriate sentence crosses the threshold of a maximum fine of $30,000, which is notionally convertible to a sentence of 6 weeks’ imprisonment. I derive this point for the s 15(3A) offence based on the maximum amount of fine that can be imposed under the provision. The same custodial threshold is applied under s 15(3) as well. This is to ensure consistency between the offences, and to avoid an unjust situation where the same level of potential of harm and culpability with a more culpable mens rea for the same unsafe act would be given a more lenient custodial threshold.

98 For the reasons stated above at [90], greater weight has been given to potential for harm as opposed to culpability. This is reflected in the table by the fact that a similar increase in potential for harm results in a greater increase in the sentence when compared to an increase in culpability.

99 As a brief guide on how the table is to be used:

(a)  First, the sentencing judge will have to determine what level of potential for harm and culpability the offender falls under and hence which of the nine large boxes he generally falls within.

(b) Once this has been determined, the sentencing judge can proceed by starting at the midpoint of the appropriate large box, marked by “X”. The sentencing judge can then direct his mind to where the accused conduct lies within each large box. The more culpable the offender’s conduct is, the more the sentencing judge should move rightwards within the large box. The higher the potential for harm, the more the sentencing judge should move upwards within the large box. The opposite applies for less culpable and less potentially harmful conduct respectively.

(c) Based on the position within the large box, the sentencing judge can derive a figure for the sentence in terms of the number of weeks of imprisonment. If the position is above and to the right of the midpoint, the sentence should be higher than the sentence indicated in the midpoint. Conversely, if the position is below and to the left of the midpoint, the sentence should be lower than the sentence indicated in the midpoint. The sentence stated in each of the small boxes situated at each of the four corners of every large box is the indicative sentence to be applied as the starting point should the court decide that particular corner point of the large box to be the most appropriate point to position the potential for harm and culpability of the offender in a particular case. It can be seen that the range for the sentence in each large box is governed by (a) the minimum sentence as indicated by the sentence stated in the small box at the left bottom corner of the large box; and (b) the maximum sentence as indicated by the sentence stated in the small box at the right top corner of the large box.

(d) After deriving the sentence in terms of weeks of imprisonment, the sentencing judge can then consider whether it is appropriate to impose a fine, an imprisonment term or some combination of the two. As a general guide, the custodial threshold is crossed when more than 6 weeks’ imprisonment is appropriate. The notional conversion rate of a $5,000 fine per week of imprisonment is a useful guide where a fine only or a combination of a fine and an imprisonment term is to be imposed.

100 I highlight that the figures in the table are not completely arbitrary. The term of imprisonment for the upper limit of the most severe bands for both potential for harm and culpability (ie the top large box at the extreme right), has been adjusted based on a ratio I had previously derived for offences with the same criminal acts but with different mens rea elements in the case of Abdul Ghani bin Tahir v Public Prosecutor [2017] 4 SLR 1153(“Abdul Ghani”) at [105]. The ratio derived was 10:5:2 for intentional acts, reckless acts and negligent acts respectively. As such, the notional upper limit for the starting point for wilful acts under s 15(3) of the WSHA is the maximum term of imprisonment of 104 weeks, the notional upper limit for reckless acts under s 15(3) is 52 weeks’ imprisonment and the notional upper limit for negligent acts under s 15(3A) of the WSHA is 20.8 weeks’ imprisonment (see Annex A–1 and Annex A–2).

101 For the rest of the figures within the table, I arrive at these figures by finding a ratio between the notional upper limit derived above, and the rest of the points in the grid as shown in the table. With this ratio, and an actual sentence for the notional upper limit corresponding to the respective offences, I could derive actual sentences for all the points in the grid. I derive this ratio by plotting a graph which I reproduce below:

102 The y-axis of the graph represents the “sentence scale”. It does not reflect actual sentences, but is used to derive relative differences in sentences, as will be explained below (see [104(a)]–[104(b)] below). The x-axis of the graph represents different “units” of culpability. The three vertical lines plotted on the graph represent three values of culpability that correspond to the upper limit of low, medium and high culpability respectively. Six curves have been plotted, representing a constant value of potential for harm in each curve. The constant values assigned for each curve is based on the upper limits of potential for harm that is classified as low, medium and high (the bolded lines), as well as the midpoint of potential for harm that is classified as low, medium and high (the dotted lines). Although only six curves were plotted for the sake of clarity, in theory, there can be an infinite number of curves representing every value of potential for harm possible.

103 As can be seen from the graph, as one moves rightwards on each individual curve, the sentence increases. This represents the fact that a higher level of culpability for the same potential for harm will lead to a higher sentence, and the sentence will smoothly increase upwards as the level of culpability increases. Similarly, for potential for harm, each curve is located above the other, this represents that as potential for harm increases, the sentence will increase. The smoothness of the increase cannot be displayed because not every curve representing every value of potential for harm is drawn. But from the graph it is clear that if every curve were drawn, each curve representing a higher value of potential for harm would be located above the curve representing a lower value of potential for harm, which indicates that there is also a smooth increase in terms of sentence for every incremental increase in potential for harm.

104 Using the graph, I produce a scale by which every figure in the table of starting sentencing ranges can be derived. As an illustration:

(a) The intersection between H1 and the curve labelled H2 Upper would give a “sentence scale” of 29 at that point. The intersection between H1 and M2 Upper would give a “sentence scale” of 20 at that point.

(b) Hence, the ratio between the sentence for the upper limit of a high culpability, high potential harm situation and the upper limit of a high culpability, medium potential harm situation must be 29:20. A ratio for all the values as required to populate the table will be calculated in this manner.

(c) Since a notional upper limit corresponding to an actual sentence in weeks has been derived for each individual offence with a different mens rea (eg 20.8 weeks for negligent acts under s 15(3A) of the WSHA), this notional upper limit can be used to derive actual values for each point on the tables as required for offences involving negligent, reckless or wilful acts using the same graph.

105 By embarking on this exercise, I hope to provide some coherence and consistency in the conceptual design of such sentencing tables. While sentencing is often said to be an art, not a science, it is equally important that sentences are seen to increase in tandem and in a logical and coherent fashion with the severity of the criminal conduct in question. There should be no sudden unexplainable jumps or gaps in either the sentence or the sentence range when the severity of the criminal conduct has only increased very slightly as one moves from one point to the next immediate point on either of the continuous scales of increasing potential for harm or increasing culpability, as two separate independent variables. At the same time, the full sentencing range as provided by the law should be used. Additionally, such consistency may allow for computer algorithms or artificial intelligence based programs to be developed to assist in the sentencing exercise in the future. It goes without the need for further re-emphasis in this judgment that the sentencing judge is never deprived of his full discretion in deciding the appropriate sentence. The table is merely a tool to assist the sentencing judge in reaching what he perceives is the most appropriate sentence having regard to the sentencing judge’s sense or judgment of the extent of the potential for harm and the degree of culpability based on all the relevant facts and circumstances of each case.

106 After a sentencing starting point has been derived based on the analysis in stage one, the sentencing judge can then consider the relevant aggravating and mitigating factors to calibrate the sentence accordingly.

107 The aggravating factors, which are not meant to be exhaustive, includes the following:

(a) The actual harm caused. This includes a consideration of the severity of the harm caused, which could be serious injury or even death as well as the number of people that have been harmed. In my view, where death is caused, an additional 8 to 40 weeks’ imprisonment should be added to the sentence depending on the number of fatalities. Where serious injury is caused, an additional sentence of up to 10 weeks’ imprisonment (or the equivalent proportion in fines where appropriate) may be added depending on the severity of the injuries and the number of persons injured;

(b) Whether the unsafe act was a significant cause of the harm that resulted;

(c) Whether the offender obtained financial gain from the breaches;

(d) The existence of relevant antecedents;

(e) Deliberate concealment of the illegal nature of the activity; and

(f) Obstruction of justice.

108 A non-exhaustive list of mitigating factors will include:

(a) A high level of co-operation with the authorities;

(b) A timely plea of guilt; and

(c) Whether the offender has voluntarily taken steps to remedy the breach or prevent future occurrences of similar breaches.

109 To summarise the approach set out above:

(a) First, the sentencing judge must consider the level of culpability and potential for harm based on the factors set out above at [86]–[87].

(b) Second, the sentencing judge will then use the table provided at [92], based on the guidance in [99] to derive a starting point for the sentence.

(c) Third, the sentencing judge should calibrate the sentence by taking into account all the relevant aggravating factors and mitigating factors in the case.

110 I now apply the sentencing framework set out above to the facts of the present case.

111 There was a high potential for harm in this case. All the workers were not wearing safety harnesses when working from height. The loading platform was not properly installed, and tilting away from Tower A, seven floors above ground level. The air compressor was so heavy that six men had great difficulty moving it. There was a high risk that any of the workers on the uninstalled platform could have fallen off the platform, which would most likely lead to death, given the height involved. In the circumstances, the harm risked was very serious and the likelihood of the harm arising was also high. In addition to the risk to the six workers, there was also a real possibility that the air compressor could have landed on other workers, resulting in further casualties or fatalities. Hence, the number of people exposed to the harm was also significant.

112 Mr Balchandani submits that Nurun “stands to the lower end of culpability”, primarily because he was instructed by Rashed, his supervisor, to work without installing the loading platform. ⁵⁴

113 The Prosecution submits that Nurun’s culpability was medium to high as he told the workers under his charge to continue loading the air compressor onto the suspended loading platform, even after they expressed concerns about the safety of the task. The Prosecution also submits that the fact that Nurun was initially directed by Rashed to work without installing the loading platform has no bearing on his culpability because he had known it was unsafe. ⁵⁵

114 In my judgment, the fact that Rashed instructed Nurun to install the loading platform has a bearing on his culpability. However, I do not think that Nurun’s culpability belongs in the low culpability band, but rather in the medium culpability band.

115 The question of whether the acts of third parties would have any bearing on the sentence of an offender was considered in the case of Guay Seng Tiong Nickson v Public Prosecutor [2016] 3 SLR 1079 (“Nickson Guay”). Nickson Guay was a case involving a road accident that claimed the life of a two-month old infant. The accused failed to keep a proper lookout while driving and encroached into the path of another car which was travelling in the opposite direction and had the right of way. This caused the fatal accident. The infant, who was in the other car, was not properly restrained by way of an approved child seat (see Nickson Guay at [1]–[2]). One issue before the court was whether the failure to properly restrain the child should have been taken into account in reducing the sentence (see Nickson Guay at [25]). The court clarified that the conduct of third parties which materially contributed to the outcome for which the offender is charged will only affect the sentence to be imposed if it has a bearing on his culpability (see Nickson Guay at [70]). The court held that the failure to properly secure the infant in an approved child seat did not have bearing on the negligence of the accused, because it did not impact the assessment of whether the accused was more or less negligent in failing to meet the standard of care which is expected of all drivers (see Nickson Guay at [72]).

116 In contrast, I regard the fact that Rashed had come up with the plan to use an uninstalled loading platform and directed Nurun to work in that manner as relevant to Nurun’s culpability. Unlike the accused in Nickson Guay, the third party conduct in the present case was directly relevant to the negligent acts that Nurun had undertaken. It was under the direction of Rashed that Nurun decided to instruct the workers under his charge to move the air compressor in an unsafe manner. If Nurun had come up with the idea himself in an effort to cut corners to save time and effort, I would have found Nurun’s culpability to be higher. It stands to reason that because Nurun was acting under the directions of his direct superior, he was less culpable than his superior. Hence, I do not think that Nurun’s conduct can be classified as being of high culpability, as that would probably be the level of culpability I would ascribe to his supervisor who had concocted the plan to cut corners and directed Nurun to do so.

117 Nevertheless, Nurun cannot be described as possessing low culpability as argued by Mr Balchandani. He had admitted that he was clearly aware of the danger of using the loading platforms in such a manner, and was experienced in the task of using loading platforms (see [52] above). The direction he had given to the workers was also a blatant deviation from established safety procedures. Additionally, he continued to instruct the workers to move the air compressor, despite the concerns they raised to him as to the safety of the task (see [41] above).

118 Therefore, I find that Nurun possessed medium culpability, tending slightly towards the higher portion of the medium culpability range.

119 Nurun is situated in the high potential for harm and medium culpability bands. This has a midpoint of 11.8 weeks. Given the fact that his culpability tended to the higher portion of the medium culpability range, I find that the appropriate starting sentence for Nurun is an imprisonment term of 13 weeks’ imprisonment.

120 There were no mitigating factors in the present case. As for the aggravating factors, the key aggravating factor is that Nurun’s unsafe acts caused the deaths of two individuals. Hence taking into account this aggravating factor, I am of the view that the sentence should be calibrated upwards by an additional 12 weeks imprisonment. This leads to a total of 25 weeks’ imprisonment.

121 Given the fact that Nurun is being sentenced under a new sentencing framework, the question arises as to whether this is a case suitable for prospective overruling.

122 In Adri Anton Kalangie v Public Prosecutor [2018] 2 SLR 557 at [70], the court clarified the principles relevant in determining the applicability of the doctrine of prospective overruling:

123 Having regard to these principles, I do not think this is an “exceptional case” that warrants the invocation of the doctrine of prospective overruling. I have come to this view mainly for the reason that there is no entrenched precedent in relation to the offence under s 15(3A) of the WSHA. The existing sentencing practice was based on unreported State Court decisions. Hence the lack of authoritative pronouncements in this regard is fatal to any attempt to suggest that applying the sentencing framework laid down in this case will result in serious and demonstrable injustice to Nurun.

124 For the reasons given above, I dismiss Nurun’s appeal on conviction and sentence and allow the Prosecution’s cross-appeal on sentence. Accordingly, the fine of $15,000 has been enhanced to a sentence of 25 weeks’ imprisonment.