1 Stalking is the label given to a series of acts repeated with unreasonable frequency, with the effect that the recipient of such unsolicited attention feels that his privacy is violated, or, worse still, feels that his personal safety is threatened.

2 To curb such antisocial conduct, the Protection from Harassment Act (Cap 256A, 2015 Rev Ed) (“POHA”) was enacted in 2014, with s 7 of POHA specifically targeted at acts of unlawful stalking.

3 Since then, there have been multiple cases of stalking prosecuted under s 7 of POHA. In each case, the severity of the stalking, as well as the harm caused to the victim(s), were of varying degrees. Having amassed a sufficient number of cases from which a rough sentencing trend may be derived, it is now apposite to provide a sentencing framework so as to ensure greater clarity, coherence and consistency in sentencing offenders convicted for such offences (see Ng Kean Meng Terence v Public Prosecutor [2017] 2 SLR 449 (“Terence Ng”) at [37]).

4 Lim Teck Kim, the appellant in this case, was the jilted ex-boyfriend of the victim. In his fruitless attempt to rekindle his relationship with her, he committed two offences, namely criminal trespass and unlawful stalking. He pleaded guilty to both offences and he was sentenced to a $500 fine and three months’ imprisonment respectively by the District Judge. ¹

5 The appellant appealed against the three months’ imprisonment term which he received for the unlawful stalking charge (under s 7 of POHA) on the basis that it was manifestly excessive.

6 Having considered the cases and developed a sentencing framework, I allow his appeal, and reduce his sentence for the unlawful stalking charge to a $5,000 fine (in default two weeks’ imprisonment).

7 Sometime in December 2016, the victim broke up with the appellant. Later, in mid-2017, the victim informed her condominium security officers not to let the appellant enter the condominium. ²

8 On 12 March 2018, the appellant rode his e-scooter through the ‘In’ driveway reserved for residents of the victim’s condominium. He then waited for the victim in the basement carpark. A security officer went to the basement carpark to confront the appellant. When the appellant saw the security officer, he fled the premises on his e-scooter. ³

9 This formed the substance of his criminal trespass charge under s 447 of the Penal Code (Cap 224, 2008 Rev Ed) (“Penal Code”), for which he was fined $500 by the District Judge. ⁴ He did not appeal against the fine.

10 The unlawful stalking charge under s 7 of POHA related to the following acts committed by the appellant with intent to cause harassment and alarm to the victim during the period of 23 February 2018 to 13 March 2018 (about three weeks).

11 On 23 February 2018, at about 1.30am, at the victim’s residence, the appellant was upset after quarrelling with the victim. He therefore self-inflicted bruises on his face and showed them to the victim, blaming her for making him hit himself. She was appalled and tried to pacify him. She then let him stay over because she was afraid of further backlash if she tried to make him leave. ⁵

12  Later the same day, the victim returned home from work early at the appellant’s insistence. At about 6.30pm, when the victim said that the appellant’s bruises looked better, the appellant threatened to hurt himself if she made him leave the house or if she left him. He then began hitting himself on the face and throwing around items in her house. He also spat the blood in his mouth onto her side table. She felt afraid and told her domestic helper to ask security to call the police. ⁶

13 On the same day, at about 7.45pm, at the victim’s residence, before the police arrived, the appellant shouted at the victim that he would rather kill himself than end their relationship. ⁷

14 Upon the arrival of the police, the appellant was arrested for an offence of attempted suicide under s 309 of the Penal Code; ⁸ such a charge does not however appear to have been preferred against him in this case.

15 After being released on bail, between 23 February 2018 and 6 March 2018, the appellant made and attempted to make communication through mobile messaging applications (iMessage and WhatsApp) to the victim, pleading with the victim to see him in person. ⁹

16 Between 4 March 2018 to 13 March 2018, the appellant also loitered around the victim’s condominium and the vicinity, pleading for her to rekindle her relationship with him. ¹⁰ For example, on 6 March 2018, the appellant was aware that the victim had a chiropractor’s appointment at Bukit Timah Shopping Centre (“the shopping centre”). Armed with such knowledge, he approached her at the carpark of the shopping centre, and attempted to initiate a conversation with her, despite her informing him that she wanted nothing to do with him. ¹¹

17 On 15 March 2018, the victim obtained an Expedited Protection Order pursuant to s 13(1) of POHA against the appellant. ¹² Thereafter, the statement of facts do not disclose any further acts of harassment by the appellant.

18 This appeal relates to the three months’ imprisonment which he received for his unlawful stalking offence that flowed from the above facts.

19 As alluded to above, I have developed a sentencing framework for cases of unlawful stalking, which is an offence punishable with a fine not exceeding $5,000 or with imprisonment for a term not exceeding 12 months or to both (s 7(6) POHA).

20 This framework draws from the two-step sentencing framework developed by the Court of Appeal in Terence Ng. The court first identifies the offence-specific factors to determine the appropriate sentencing band for the particular offence. Thereafter, the court considers the aggravating and mitigating factors specific to the offender to calibrate a sentence that is appropriate for that offender (Terence Ng at [39]). This two-step sentencing framework has since been followed in other criminal cases (see, eg, Kunasekaran s/o Kalimuthu Somasundara v Public Prosecutor [2018] 4 SLR 580 (“Kunasekaran”) and GBR v Public Prosecutor and another appeal [2018] 3 SLR 1048).

21 However, the framework proposed herein seeks to refine the methodology which was developed in Terence Ng.

22 In Terence Ng, the Court of Appeal considered several offence-specific factors which go towards the aggravation of the offence. In the context of rape, which was the offence before the court in Terence Ng, such offence-specific factors included the abuse of position and breach of trust, premeditation, and the use of violence (Terence Ng at [44]).

23 After considering the aggravating offence-specific factors in the case, the sentencing judge would be able to determine the appropriate band which the accused’s offence fell within.

24 Cases which presented no offence-specific factors or where the factors were only present to a very limited extent were classed as Band 1 cases, which were punishable with ten to 13 years’ imprisonment, with six strokes of the cane. Cases with two or more aggravating factors fell under Band 2, and were punishable with 13 to 17 years’ imprisonment, with 12 strokes of the cane. Finally, extremely serious cases of rape, which, “by reason of the number and intensity of the aggravating factors”, were classed as Band 3 cases, for which the appropriate punishment was 17 to 20 years’ imprisonment, with 18 strokes of the cane (Terence Ng at [47]).

25 In Kunasekaran at [47], I observed that the “two-step sentencing bands” approach has much to commend it, and that it is a reliable methodology which improves clarity, transparency, coherence and consistency in sentencing.

26 Nonetheless, a problem which arises from the two-step sentencing band approach is that there is an overt focus on the number of aggravating factors. Hence, in Terence Ng, once there were two or more aggravating factors, the case inevitably tipped out of Band 1 into Band 2. I understand that it is possible to adjust beyond the sentencing range prescribed for the band when considering the appropriate sentence at the second stage of the framework (ie, offender-specific factors stage) (see Terence Ng at [62]). However, this does not curb the problem identified, which relates to the excessive or inadequate weight which may be given to offence-specific factors (the first stage). Given the inevitable focus on the sheer number of offence-specific factors present, the current two-step sentencing band approach implicitly assumes that each offence-specific factor carries the same weightage. This disregards the possibility that each of these offence-specific factors may carry different aggravating weights in influencing the sentence.

27 To more accurately evaluate the appropriate amount of weight to be ascribed to each offence-specific factor, I therefore propose a points system, whereby each offence-specific factor is given a range of points, allowing the sentencing judge to determine the appropriate number of points to be ascribed based on the intensity of the aggravating factor. Once the total number of points from the offence-specific factors are tallied, the indicative starting sentence, based on the sentencing band, can be determined. The sentencing band which the accused’s offence falls under will therefore be determined not primarily by the number of offence-specific factors present, but by the combined intensity of all the offence-specific factors present, as reflected by the aggregated number of points. This takes into account not only the number of offence-specific factors present but also the different weightage that each of these offence-specific factors may have.

28 I now present the sentencing framework for the offence of unlawful stalking under s 7 of POHA to better demonstrate my point.

29 In determining the appropriate band of the offence, the court should first have regard to the offence-specific factors that relate to both the degree of harm caused by the offender and the degree of his culpability.

30 A review of cases in relation to unlawful stalking offences (s 7(1) POHA) discloses the following, non-exhaustive, list of offence-specific factors which independently reflect the increased culpability of the accused or the heightened level of harm suffered by the victim, thereby aggravating the offence:

(a) Duration and frequency of stalking: Cases of unlawful stalking are necessarily disruptive and harmful to the lives of the victims, who live subject to the offender’s watchful eye. Where the stalking continues for a longer period of time and at a high frequency, the harm and disruption caused to the victim will inevitably be of a higher degree: see Public Prosecutor v Tan Khoon Aik Nelson (SC-913209-2016, Magistrate’s Arrest Case No 903858 of 2017 and others) (“PP v Nelson Tan”), Tan Yao Min v Public Prosecutor [2018] 3 SLR 1134 (“Tan Yao Min v PP”) at [91] and s 7(5)(b) POHA.

(b) Degree of intrusion into the victim’s life: There are varying degrees of unlawful stalking. In certain cases, the intrusion may amount to merely loitering around the victim’s premises. Here, lesser adjustment is required by the victim to avoid the accused person. In certain cases, the offender may follow the victim to places he/she frequented, call or message him/her daily, confront the victim inside or outside his/her home, etc: see Public Prosecutor v Tan Boon Wah (SC-905671-2016, Magistrate’s Arrest Case No 904633-2016 and others) ¹³ (“PP v Tan Boon Wah”), discussed in Tan Yao Min v PP at [87]. The latter class of cases would amount to a high degree of intrusion into the victim’s life (see ss 7(5)(b)–(e) POHA). This discloses additional harm suffered by the victim, and is independent from the duration and frequency of stalking. Hence, additional points may be attributed.

(c) Vulnerable victim: Where the targeted victim is a minor, the offender’s unlawful stalking would be considered more egregious, particularly when the offender is significantly older than the victim: see Tan Yao Min v PP at [90] and Public Prosecutor v Ng Han Wei (SC-912985-2016, Magistrate’s Arrest Case No 901757 of 2016) (“PP v Ng Han Wei”). Preying on victims ailing under physical or mental conditions would also be an aggravating factor: see s 7(5)(g) POHA.

(d) Public dissemination of sensitive information or images: In certain cases, the offender may publicly disseminate sensitive information or images of the victim, such as nude photographs which are not otherwise available in the public domain: see Public Prosecutor v Adrian Goh Guan Kiong (SC-902574-2016, Magistrate’s Arrest Case No 902040 of 2016) (“PP v Adrian Goh”), discussed in Tan Yao Min v PP at [83]. The public nature of such acts would exacerbate the degree of harm and embarrassment suffered by the victim, and would clearly increase the offender’s culpability. Such cases are particularly egregious, especially if the harassment crosses into the online domain, whereby the leak of such sensitive information or images could have a “viral and permanent” impact (Singapore Parliamentary Debates, Official Report (13 March 2014), vol 91 (Mr K Shanmugam, Minister for Law).

In fact, the offender’s culpability is also heightened as such cases often involve an abuse of a position of trust, as accused persons with access to such highly sensitive information or images of the victim would often have shared a relationship with the victim: see PP v Adrian Goh, PP v Tan Boon Wah and Public Prosecutor v Lai Zhi Heng (SC-912644-2015, Magistrate’s Arrest Case No 909122 of 2015 and others) (“PP v Lai Zhi Heng”) (discussed in Tan Yao Min v PP at [19] and [82]).

(e) Use of threats against the victim: When the offender uses threats against the victim, the degree of harassment, alarm, fear or distress caused to the victim would be heightened, thereby increasing the harm caused to the victim: see PP v Nelson Tan and s 7(5)(g) POHA.

(f) Harm to victim beyond those ordinarily caused in (a), (b), (d) and (e) (“extraordinary harm”): The harm suffered by the victim of unlawful stalking varies from victim to victim. In most cases, the harm suffered by the victim will be factored into the sentence by factors (a), (b), (d) and (e), which incorporate the harm ordinarily suffered by the victim. For example, a victim who is frequently stalked over a long duration is likely to suffer loss of his “freedom to do any act which he is legally entitled to do” (s 7(5)(f) POHA). Such ordinary harm is already accounted for in factor (a) above, which ascribes points based on the duration and frequency of stalking. To prevent double counting against the accused, no additional points ought therefore to be attributed for such ordinary harm suffered.

In certain cases, however, the victim may suffer harm beyond what is normally caused by factors (a), (b), (d) and (e). For example, while a victim who suffers the public dissemination of his or her sensitive information or images (see factor (d) above) may ordinarily be expected to suffer emotional trauma and damage to his or her reputation, where the victim suffers from permanent emotional or psychological scars or loses his job as a result of the accused’s acts, such would amount to extraordinary harm. If found to be present by the objective circumstances or an unrefuted victim impact statement, such extraordinary harm would be an aggravating factor independent from the other factors considered above, for which points may be attributed without double counting against the accused.

(g) Involving third parties: When the offender also involves persons other than the victim, such as the victim’s next-of-kin or the victim’s partner (the “third parties”), the inconvenience and distress would extend beyond the victim himself, as the third parties affected may also have to employ self-help measures to avoid the offender: see Tan Yao Min v PP at [91]. Where no separate charges are preferred for such inconvenience caused to the third parties, this should be considered as an aggravating factor. If separate charges are preferred, this factor should not be considered aggravating to prevent double counting against the accused: see PP v Nelson Tan.

31 Points are to be ascribed to each of the offence-specific factors based on the degree of aggravation of the offence occasioned by the presence of that factor. Where the offence in relation to the particular factor is more serious, more points are to be given to that factor. No points are awarded if the factor is absent.

32 Furthermore, I propose ascribing a maximum of between three to five points to each of the offence-specific factors identified based on their relative weightages. Accordingly, the maximum number of points allocated to each factor may differ. For example, the factor of public dissemination of sensitive information or images may result in irreversible harm to the victim’s reputation. To reflect its potential greater weightage as a factor (ie, when compared to other factors such as the use of threats against the victim, which are often temporal in nature), the maximum number of points allocated to the factor of public dissemination of sensitive information would be higher so as to allow more points to be eventually ascribed to it during the points assessment process by the judge.

33 At the outset, it is emphasised that the framework as proposed applies to cases where the accused has claimed trial, rather than pleaded guilty to the offence. As discussed at [45(a)] below, a plea of guilt may entitle an accused to a discount in sentence.

34 The points to be ascribed to each offence-specific factor is as follows:

35 Once the severity of the offence, as gleaned from the number of points, has been ascertained, the offence may be placed in the appropriate sentencing band.

36 Cases with one to five points would be of the lowest severity, for which the custodial threshold would not be crossed. Such cases fall under Band 1, for which the indicative sentence would be a fine of up to the maximum of $5,000.

37 Once the case exceeds five points, the custodial threshold would generally be crossed, subject to any offender-specific mitigating factors which may justify a high fine in lieu of a custodial sentence. Once the custodial threshold is crossed, the indicative starting sentence would be 0.8 months’ imprisonment (six points), with a 0.8 months’ increment for each additional point.

38 The sentencing bands are thus as follows:

(a) Band 1 (one to five points): up to $5,000 fine;

(b) Band 2 (above five points to ten points): $5,000 fine, or up to four months’ imprisonment;

(c) Band 3 (above ten points to 15 points): four to eight months’ imprisonment;

(d) Band 4 (above 15 points to 20 points): eight to 12 months’ imprisonment.

(e) Band 5 (above 20 points): 12 months’ imprisonment and above.

39 In detail, the indicative starting sentence for each point is as follows:

40 It can be seen that after 20 points, the indicative starting sentence exceeds the statutory maximum of 12 months’ imprisonment for unlawful stalking offences under s 7 POHA. This is not problematic, as the indicative starting sentence is but a reflection of the appropriate sentence after the consideration of the offence-specific factors only (ie, the first stage). At the next stage, the indicative starting sentence may be calibrated downwards to account for offender-specific mitigating factors, such as a timely plea of guilt. After accounting for both the offence-specific and offender-specific factors, a recommended final sentence would be obtained. This recommended final sentence may fall below 12 months’ imprisonment. If the recommended final sentence continues to exceed 12 months’ imprisonment after considering the offender-specific factors (ie, the second stage), the actual sentence to be imposed must in any case not exceed the maximum imprisonment term of 12 months’ imprisonment, although the court may couple that maximum imprisonment term with a fine not exceeding $5,000 (see s 7(6) POHA). This means that the mandatory statutory cap on the actual sentence must be applied even though the culpability and recommended final sentence assessed under the sentencing framework may be higher than the maximum prescribed by law.

41 The points-based framework serves simply to guide judges in exercising their discretion in arriving at the appropriate indicative starting sentence, having regard to the offence-specific factors of each case. It does not take away the discretion vested in sentencing judges, who must ultimately conduct an “exercise in evaluative ethical judgment” (Terence Ng at [48]) in determining the appropriate number of points (if any) to attribute to each of the offence-specific factors.

42 Such a methodology, which does not depart in substance and in principle from the two-step sentencing band approach developed in Terence Ng, has at least two advantages over the present number-of-aggravating-factors focused approach:

(a) First, it allows for calibration between the offence-specific factors. Certain factors cause far more harm, or reflect the heightened culpability, of the accused. For example, in the context of unlawful stalking, the public dissemination of sensitive information or images may cause irreversible harm to the reputation of the victim, and therefore requires more points to be ascribed to it within the given range to better reflect the greater intensity of the offender’s wrongdoing. Hence, the maximum number of points that can be given for this factor is higher when compared to the maximum number of points that can be given for a factor assessed to be of lower weightage.

(b) Secondly, it allows for calibration within each of the offence-specific factors. As is patent from the table at [34] above, the severity within each offence-specific factor may vary from case-to-case. For example, where the stalking occurs over a longer period and with greater frequency, more points within the given range ought to be allocated to this offence-specific factor. The natural corollary is that an accused with a greater number of aggravating factors (each of a low weightage) may not necessarily fall on a higher band than an accused with a lesser number of aggravating factors (but each of which may be of a high weightage); the appropriate sentencing band depends cumulatively therefore on the degree of severity of each of the aggravating factors present having regard also to their individual relative weightages, as reflected in the aggregate number of points scored in each particular case. In other words, the total number of points scored reflects the cumulative effect of all the aggravating factors present, including their relative degree of importance or significance in terms of both the harm caused and the culpability of the offender.

43 At the second step, the court should have regard to the offender-specific factors (whether mitigating or aggravating). The court has the discretion as to the appropriate amount of weight to place on such factors, and an adjustment beyond the sentencing range prescribed in each sentencing band may be called for in the appropriate case. For example, in cases on the borderline of each sentencing band, the mitigating or aggravating factors may tip the case to the lower or higher sentencing band.

44 Some offender-specific aggravating factors include:

(a) Offences taken into consideration for the purposes of sentencing, especially when the offences taken into consideration are of a similar nature: see PP v UI [2008] 4 SLR(R) 500 at [38].

(b) The presence of relevant antecedents, especially if the offences were committed against the same victim, such that the harm suffered by the victim is compounded.

(i) Enhanced penalty provision: relevance of subsequent similar offences: It is noted that a person who has been convicted under s 7 POHA is liable for enhanced penalty for any subsequent offence. If the subsequent offence is prosecuted under the enhanced penalty provision in s 8 POHA, the accused shall, on conviction, be liable to a fine not exceeding $10,000 or to imprisonment for a term not exceeding two years or to both (s 8(e) POHA). The fine and imprisonment terms for the enhanced offence is therefore twice that of the initial offence, which carries a maximum fine of $5,000 or imprisonment not exceeding one year or both (s 7(6) POHA). In such cases, the accused would already be punished by way of the enhanced penalty charge under s 8 POHA, and no additional aggravating weight should therefore be given to the prior antecedents of the accused in coming to an appropriate sentence for the s 8 POHA charge.

If, however, the Prosecution exercises its discretion and elects to prosecute a subsequent offender with a prior conviction under s 7 POHA (even though s 8 POHA could be used), the relevant antecedents could then be considered as an aggravating factor.

(c) The offender committed the present set of offence(s) while on court bail for other offence(s).

45 Some offender-specific mitigating factors include:

(a) Pleas of guilt: In unexceptional plea of guilt cases, a timely plea would be “indicative of genuine remorse” (Public Prosecutor v NF [2006] 4 SLR(R) 849 at [57]). It also has mitigating value as it “spares the victim the ordeal of having to testify, thereby saving the victim the horror of having to re-live the incident” and “saves the resources of the State which would otherwise have been expended if there were a trial” (Terence Ng at [66]). In such unexceptional cases, considering that the maximum imprisonment term for unlawful stalking offences under s 7 POHA is a relatively short period of 12 months’ imprisonment, an approximate 25% discount from the indicative starting sentence as determined after the first-stage (offence-specific factors) would be appropriate.

The amount of discount for a plea of guilt may be scaled upwards or downwards if other factors are present. For example, if the accused had earlier elected to claim trial and only pleaded guilty on the first day of the trial, the discount ought to be far lesser, as resources would already have been expended in preparing for a trial. This accused is also likely to be less remorseful than one who pleaded guilty at the first available opportunity.

(b) The youth or advanced age of the offender: Terence Ng at [65].

(i) For youths, rehabilitation is a key sentencing consideration, for which a lower or alternate rehabilitative sentence like probation or reformative training may be considered. However, in egregious cases of unlawful stalking (eg, Band 4 and 5 cases), “a condign punishment” would be required “in order to satisfy the needs for deterrence and retribution” (see Terence Ng at [65(b)], citing Public Prosecutor v Mohammad Al-Ansari bin Basri [2008] 1 SLR(R) 449).

(ii) For older offenders, “the imposition of substantial custodial terms deprives the elderly of a larger fraction of their expectation of life. This – a concern for the overall proportionality of punishment … is the real reason for affording leniency on account of advanced age” (Terence Ng at [65(c)]). Nonetheless, given that the maximum imprisonment term for the offence of unlawful stalking is only 12 months’ imprisonment, the custodial term is unlikely to be so substantial such that this factor features weightily.

46 The offender-specific factors stated above are non-exhaustive. Ultimately, each sentence must be calibrated to best fit the accused’s circumstances. In this regard, the classic principles of sentencing, namely retribution, deterrence, prevention and rehabilitation are helpful in determining the appropriate amount that an accused’s sentence ought to be enhanced or reduced at the second stage of the framework.

47 Applying the framework to past cases of unlawful stalking will provide a better illustration of the usefulness of the refined two-stage sentencing framework.

48 As per the framework, the offence-specific factors present in each case will first be analysed, with the appropriate number of points attributed depending on the facts disclosed. An indicative starting sentence (falling within the appropriate sentencing band) would then be disclosed from the aggregate number of points after this first stage.

49 Moving to the second stage, appropriate additions or subtractions to the indicative starting sentence will be made depending on the offender-specific factors present in each case. A final sentence derived by applying the framework would then be disclosed (the recommended final sentence).

50 Later in this judgment, the recommended final sentences disclosed from each case will be back-tested and compared against the actual sentences imposed in the individual cases. This back-testing will demonstrate the efficacy and reliability of the framework in ensuring clarity and consistency in sentencing which is commensurate with the culpability and the harm in each case.

51 

52 As seen from the cases analysed, this points-based sentencing framework allows for a deeper consideration of each offence-specific factor, without giving an undue focus on the sheer number of such factors present. For the avoidance of doubt, the two-step sentencing band approach remains, albeit with refinements as to how the appropriate sentencing band is to be derived.

53 To evaluate the efficacy of the proposed sentencing framework, it has been back-tested against the cases analysed above.

54 In the analysis of the cases above, points were attributed to the offence-specific factors in each case, thereby shedding light on the indicative starting sentence. Thereafter, the offender-specific factors were taken into account, to arrive at the recommended final sentence under the sentencing framework herein.

55 These recommended final sentences were then compared against the actual sentences meted out in the cases, revealing the following:

56 Reviewing the graph above, it will be seen that for cases on the lower end (less points), the recommended final sentence is lower than the actual sentence. As for cases on the higher end (more points), the recommended final sentence is higher than the actual sentence:

57 On its face, the sentencing framework therefore appears to be more lenient to accused persons with lesser points (less serious offending), and stricter to accused persons with higher points (more serious offending).

58 However, the reason for the significant differences between the recommended final sentence and the actual sentence at the lower end and higher end of the framework is because the sentence for cases of unlawful stalking appear to have congregated around three to six months’ imprisonment (the “sentencing trend”) (Tan Yao Min v PP at [81]).

59 Hence, in determining the appropriate sentence for each case, judges refer to the sentencing trend of three to six months’ imprisonment as the starting point. Thereafter, appropriate adjustments upwards or downwards are made if the case at hand is distinguishable from the cases which make up the sentencing trend (see Public Prosecutor v Lim Seow Kiat [2018] SGMC 44 (“PP v Lim Seow Kiat”) at [27]–[29]).

60 For example, in Tan Yao Min v PP, the judge considered seven decisions where the offenders were sentenced to three to six months’ imprisonment for unlawful stalking. After considering the cases, the judge noted the following aggravating factors in the case before him: (a) the victim was a minor and was thus a vulnerable victim, (b) the offender had committed similar offences against the same victim in 2010 and 2015, and (c) the offender’s stalking affected third parties, as the offender had added the victim’s friend on Facebook, and spoke to the victim’s father’s colleague at his workplace (Tan Yao Min v PP at [90]). While the judge noted that, unlike the offender in PP v Lai Zhi Heng, the offender did not disseminate the victim’s nude photographs, he found that in totality, eight months’ imprisonment was not manifestly excessive (Tan Yao Min v PP at [91]–[92]).

61 This approach was also adopted by the District Judge. First, he noted the sentencing trend of three to six months’ imprisonment (Public Prosecutor v Lim Teck Kim [2018] SGMC 77 (“GD”) at [10]). Thereafter, he reviewed the seven cases considered in Tan Yao Min v PP, and found the facts of the present case “similar to the case” of PP v Nelson Tan (one of the seven cases considered in Tan Yao Min v PP), where the offender was sentenced to three months’ imprisonment (GD at [18]). In the circumstances, he also sentenced the appellant to three months’ imprisonment (GD at [20]).

62 As a result of the reliance on the sentencing trend as the starting point in determining the appropriate sentence, cases are often straightjacketed into a narrow sentencing width. This congregation of sentences can be seen when the actual imprisonment terms in the cases analysed were given notional discounts or uplifts depending on the offender-specific factors present. This was so as to arrive at a notional first-stage sentence, which reflects the combined severity of the offence-specific factors only, without any regard being given to the factors in favour or against the particular offender (ie, offender-specific factors).

63 As the severity of the offence should increase with every increase in the number of points, a positive (upward) gradient would be expected for the notional first-stage sentences of the cases, whereby the offender-specific factors are disregarded.

64 The notional sentences for each case analysed were derived as such:

65 Plotted on a graph, the notional first-stage sentences for the cases produced a best-fit line with a very mildly positive gradient:

66 The very mildly positive gradient (showing a total increment in the imprisonment term of only about one month from five points to 25 points) suggests that the sentences in the cases analysed are being decided primarily by reference to the present sentencing trend of three to six months’ imprisonment, with insufficient regard to offence-specific factors and their respective intensity. As a result, the nominal first-stage sentences are bunched together at around seven months’ imprisonment (as can be seen on the best-fit line), even though the points, which reflect the intensity of the offence-specific factors and therefore serve as the proxy for the severity of each individual offence, do vary widely from seven to 22 points.

67 This congregation of the sentences for the analysed cases between three to six months’ imprisonment is plainly unsatisfactory for three reasons.

68 First, it is likely that the full sentencing range, which permits up to 12 months’ imprisonment, will not be properly utilised (see Tay Wee Kiat and another v Public Prosecutor and another appeal [2018] 4 SLR 1315 at [76] and Soh Qiu Xia Katty v Public Prosecutor [2019] 3 SLR 568 at [28(b)]).

69 Secondly, the dispersion of the nominal first-stage sentences of the cases (ie, based purely on the offence-specific factors and with the effects of all offender-specific factors removed) as plotted on the graph (at [65] above) also demonstrates a lack of logical consistency in sentencing. Some cases scoring lower points have higher nominal first-stage sentences than some of those cases scoring higher points, when the opposite should logically have been the case.

70 Thirdly and most importantly, it can lead to dissimilar cases being treated alike.

71 For example, while both PP v Adrian Goh and PP v Lai Zhi Heng involved the public dissemination of the victims’ nude photographs, the duration and frequency of harassment as well as the degree of intrusion in the two cases were diametrically opposite. The offender in PP v Adrian Goh harassed the victim for a short four-day period, and his harassment was limited to electronic and non-physical means. In contrast, the offender in PP v Lai Zhi Heng harassed the victim incessantly for one and a half years, harassing her around her school and house. He also issued threats to the victim, causing her to suffer such extraordinary harm that she had to switch schools to avoid the offender.

72 Considering the intensity of each offence-specific factor, it would be clear that the offence in PP v Lai Zhi Heng (scoring 22 points for the offence-specific factors) was of a far graver nature than that in PP v Adrian Goh (scoring only 12 points for the offence-specific factors). Yet, without a framework to guide the sentencing judge, such superficially similar cases may result in an inevitable congregation of sentences; indeed, both offenders in PP v Adrian Goh and PP v Lai Zhi Heng received the same sentence of six months’ imprisonment. It must be noted that the incongruity cannot be explained by any possible differences between the two cases in terms of the offender-specific factors because there was nothing much in fact to differentiate between them: both offenders had no relevant antecedents and both had pleaded guilty.

73 The two-step sentencing framework therefore seeks to ensure that appropriate weight will be given to the offence-specific factors in each case. While two cases may appear at first blush to be largely similar, a closer analysis may reveal the differing seriousness of the offences in each case.

74 Applying the sentencing framework, “outliers” like PP v Lai Zhi Heng will therefore be minimised, and the sentencing judge would be assisted in arriving at a sentence with more clarity and coherence. Insofar as the offence-specific factors were derived from the analysis of the cases, the sentencing framework does not depart from the precedents, but rather seeks to refine them and detail the sentencing considerations that each court should take into account for cases of unlawful stalking.

75 Applying the framework to the facts of the present case, the offence-specific factors disclose an aggregate of six points:

(a) Short duration, moderate frequency of stalking (two points): The appellant stalked the victim for a short duration of about three weeks (from 23 February 2018 to 13 March 2018). The frequency of stalking was moderate, as the appellant’s stalking did not amount to a daily intrusion into the victim’s life, given that there were gaps between the periods he sent the messages (the messages were sent on 23 February 2018, and 4 and 6 March 2018 only) and harassed the victim in person (on 23 February 2018, and 4, 6, 8, 12 and 13 March 2018 only).

(b) Degree of intrusion was high (three points): The degree of intrusion by the appellant was high. On 23 February 2018, after having coerced the victim to allow him to spend the night at her place by inflicting injuries on himself and blaming the victim for such injuries, the appellant refused to leave the victim’s residence, and threatened to hurt himself if she made him leave the house. He could only be removed after the police arrived and arrested him for attempted suicide as he had threatened to kill himself rather than end the relationship with the victim (“the 23 February incident”). ²² While his acts of harassment after 23 February 2018 amounted to merely sending messages or loitering around the victim’s vicinity, viewed in light of the 23 February incident, which would have made the victim fearful of the appellant’s presence, they do not detract from the high degree of intrusion by the appellant into the victim’s life.

(c) Threats to self and not to the victim, unlike most other cases (one point): Finally, the appellant had issued threats to hurt himself. To this effect, he hit himself on the face, and threatened to kill himself rather than end the relationship with the victim. This made the victim feel afraid, such that she had to call the police for assistance. ²³

76 With the aggregate of six points from the offence-specific factors, the appellant’s conduct falls within the lowest end of Band 2, and just crosses the custodial threshold. The indicative starting sentence is therefore 0.8 months’ imprisonment.

77 The District Judge had opined that the custodial threshold was crossed because, unlike the offender in PP v Lim Seow Kiat, where a $5,000 fine was awarded, the appellant had issued threats to the victim (GD at [19]). This is not the case. The custodial threshold is crossed because of the aggregate of more than five points in this case. Indeed, the points-based sentencing framework herein emphasises the importance of considering the intensity of each offence-specific factor; not all threats are equally aggravating. As the threats to the victim in this case never extended to hurting the victim or jeopardising her safety or wellbeing, they were of the less egregious variety, for which one point would suffice. Hence, had the aggregate number of points been between one to five points only, the custodial threshold would not have been crossed notwithstanding the appellant’s threats.

78 In relation to the offender-specific factors, the appellant had pleaded guilty, for which mitigating weight was to be afforded. While the appellant was traced, his antecedents were dated (with the latest conviction being in 2004) and unrelated, with most of them being for drug offences. I therefore do not regard his antecedents as a significant aggravating factor.

79 As this is a borderline case that just crosses into Band 2, and in the light of his plea of guilt, I am prepared to discount the appellant’s sentence and place him on the highest end of Band 1. In the circumstances, I find that the maximum fine of $5,000 (in default two weeks’ imprisonment) would suffice for this case. I therefore allow the appellant’s appeal against sentence to that effect.

80 The sentencing framework as proposed is intended only as a guide for judges to arrive at the sentence that best fits the case before them. By focussing on one factor at a time, the sentencing framework enables the judge to better assess the seriousness of the offence for the purpose of sentencing, without being overwhelmed by the task of having to consider a whole host of offence-specific factors, each carrying different weight, at the same time. The process of considering the degree of aggravation of each offence-specific factor also ensures consistency and clarity in sentencing, as judges must specify their reasons for ascribing a particular weight to each factor. The risks of double-counting against an offender is also thereby mitigated.

81 Having said that, it must be borne in mind that the sentencing framework is not a rigid formula by which judges may surrender the need to apply their judicial discretion. They have to carefully evaluate and weigh a myriad of considerations to determine the most just and appropriate outcome in the case before them.

82 Ultimately, sentencing is a fluid exercise with very concrete consequences for the accused at hand. Careful consideration of where justice may lie in each case is of primary importance.

83 Here, Lord Woolf CJ’s reminder in R v Millberry [2003] 1 WLR 546 at 556 (cited by Chao Hick Tin JA, “The Art of Sentencing – An Appellate Court’s Perspective”, speech at Sentencing Conference 2014: Trends, Tools & Technology (9 October 2014) at para 13) is deserving of restatement: