1 The appellant was an untraced 34-year-old Filipino female who was charged with seven cheating, theft and obstruction of justice offences. She pleaded guilty in the court below to four proceeded charges and consented to the remaining three charges being taken into consideration (“TIC”) for sentencing. Accordingly, the District Judge (“DJ”) convicted her on all four charges and sentenced her to a global sentence of 21 months and eight weeks’ imprisonment, backdated to the date of first arrest, 9 October 2023, to account for the periods of police custody and remand, excluding the bail period.

2 The present appeal was concerned solely with the sentence imposed for the cheating charge under s 420 of the Penal Code 1871 (2020 Rev Ed) (“PC”) amalgamated under s 124(4) of the Criminal Procedure Code 2010 (2020 Rev Ed) (“CPC”), read with s 124(8)(a)(ii) of the CPC (ie, DAC-917863-2025).

3 I allowed the appeal and substituted the initial sentence of 18 months’ imprisonment with a sentence of eight months’ imprisonment. Briefly, I found that the DJ erred in his application of the sentencing framework for amalgamated charges set out in Prakash s/o Mathivanan v Public Prosecutor [2025] 4 SLR 1386 (“Prakash” with reference to the case or “Prakash Framework” with reference to the framework). The present appeal illustrated the importance of proper identification and application of the base offence in Stage 1 of the Prakash Framework by the sentencing court. That did not happen in this case. I therefore approached the question of sentence afresh, working through each stage of the Prakash Framework in turn, and arrived at a final sentence of eight months’ imprisonment, which I considered to be fair, proportionate and just.

4 The appellant pleaded guilty to the following four charges below:

(a) DAC-916661-2023: One charge of cheating under s 417 of the PC amalgamated under s 124(4) of the CPC, read with s 124(8)(a)(ii) of the CPC, for presenting altered PayNow screenshot receipts to dishonestly induce a seafood restaurant to deliver food orders totalling $3,891.75, without actual payment, on 23 occasions between May 2022 and April 2023 (“First Cheating Charge”).

(b) DAC-923562-2024: One charge of theft in dwelling used for the custody of property under s 380 of the PC, for concealing 31 items valued at $738.76 in her bags at a supermarket while paying for only four items at a self-service kiosk, on 27 November 2024 (“Theft Charge”).

(c) DAC-917857-2025: One charge of obstructing the course of justice under s 204A of the PC, by deleting certain WhatsApp messages between her and the representative of a complainant-company (“Obstruction of Justice Charge”).

(d) DAC-917863-2025: One charge of cheating under s 420 of the PC amalgamated under s 124(4) of the CPC, read with s 124(8)(a)(ii) of the CPC, for presenting altered PayNow screenshot receipts to dishonestly induce an Italian restaurant to deliver food orders totalling $6,167.52, without actual payment, on 24 occasions between 15 June 2024 and 28 May 2025 (“Second Cheating Charge”).

5 The following three cheating charges were TIC for the purposes of sentencing (“TIC Charges”):

(a) one charge of cheating under s 417 of the PC amalgamated under s 124(4) of the CPC, read with s 124(8)(a)(ii) of the CPC;

(b) one charge of attempted cheating under s 420 of the PC read with s 511(2)(e) of the PC; and

(c) one charge of attempted cheating under s 420A of the PC.

6 The appellant admitted to placing 24 separate food orders with a restaurant (“Restaurant”) between 15 June 2024 and 28 May 2025.

7 The appellant’s modus operandi, which was identical to that employed in the First Cheating Charge, was as follows:

(a) The appellant used WhatsApp to place food orders with the Restaurant.

(b) A screenshot was taken of the completed PayNow transfer of the exact order amount to the appellant’s own bank account.

(c) A mobile application called “PhotoGrid” was used to edit the screenshots by replacing the original recipient’s details (ie, the appellant’s own bank account) with those of the Restaurant’s bank account.

(d) The edited screenshot was sent on WhatsApp to the Restaurant as proof of payment.

8 By the time the Restaurant discovered the deception on 25 May 2025, the appellant had cheated it of a total sum of $6,167.52, being the value of the food she had ordered. Subsequently, on 6 November 2025, the appellant made full restitution with the assistance of her sister.

9 Pertinently, the entirety of the appellant’s offending conduct in relation to the Second Cheating Charge occurred while she was on bail.

(a) For the period of offending between 15 June 2024 and 27 November 2024, the appellant was on police bail from 9 October 2023, having been arrested on the same day, 9 October 2023, in connection with the First Cheating Charge.

(b) For the period of offending between 5 December 2024 and 28 May 2025, the appellant was released on court bail on 5 December 2024, having been arrested on 27 November 2024 in connection with the Theft Charge.

10 Before I summarise the DJ’s decision and the parties’ cases, it would be helpful to first set out the three-stage sentencing framework for charges amalgamated under s 124(4) of the CPC as set out in Prakash at [42]–[46] and [129].

(a) First, the appropriate sentence for a single incident of the offence by reference to offence-specific factors and any relevant sentencing framework or benchmark (“Stage 1”) has to be determined. This is described as the base offence. Offence-specific factors include the manner and mode by which the offence was committed as well as the harm that was caused to the victim: Prakash at [42].

(b) Second, the appropriate starting point for the amalgamated offence must be arrived at, based on a holistic assessment of an offender’s culpability over the entire course of conduct (“Stage 2”). In addition to the culpability-related factors taken into consideration in Stage 1, the duration and frequency of the offending are relevant considerations, as the repetition of criminal behaviour over a prolonged period invariably demonstrates greater recalcitrance and constitutes an aggravating factor: Prakash at [45(a)] and [129(b)].

(c) Third, consideration should be given to whether adjustments to the starting point sentence are warranted on the basis of offender-specific aggravating and mitigating factors (“Stage 3”). These include aggravating factors such as relevant antecedents indicating a lack of remorse, and mitigating factors such as a timely plea of guilty and genuine remorse on the offender’s part: Prakash at [46] and [129(c)].

11 The DJ’s decision is set out in Public Prosecutor v Santos-Tumalip Maria Monalyn Bagaporo [2026] SGDC 159 (“GD”). The parties’ submissions on sentence before the DJ and the DJ’s sentencing decision are summarised in the table below: GD at [15]–[16].

12 In determining the appropriate sentence for the Second Cheating Charge, the DJ applied the Prakash Framework.

13 The DJ accepted the Prosecution’s reliance on Public Prosecutor v Musadiq s/o Abbas [2019] SGDC 254 (“Musadiq”) as the principal comparator for determining the appropriate starting sentence in Stage 1 of the Prakash Framework: GD at [28]. Finding the facts, methodology and degree of sophistication to be closely similar to those in Musadiq, the DJ arrived at a starting sentence of 15 months’ imprisonment: GD at [28]–[29].

14 Finally, the DJ identified three aggravating factors warranting an uplift of three months (GD at [28]–[29]): (a) the amount cheated in the present case ($6,167.52) was higher than that in Musadiq ($4,070); (b) the offence under the Second Cheating Charge was committed while the appellant was out on court bail; and (c) there were three similar TIC cheating charges in the present case, compared to only one in Musadiq.

15 The DJ therefore imposed a final imprisonment term of 18 months.

16 On appeal, the appellant submitted that the sentence of 18 months’ imprisonment for the Second Cheating Charge was manifestly excessive and should be reduced to seven months’ imprisonment, on four grounds.

17 First, the DJ erred in relying solely on Musadiq as the principal sentencing precedent. Musadiq had no precedential value in the application of the Prakash Framework. Moreover, Musadiq should be distinguished. The fundamental error was in the DJ’s conclusion that Musadiq was materially similar to the present case. On the contrary, Musadiq involved a materially more sophisticated modus operandi as it featured “three layers of fraudulent behaviour”, was “inherently more difficult to detect”, and involved the misuse of personal data and the solicitation of financial documents under false pretences, all of which were absent in the present case. Moreover, the three-month uplift from the Musadiq starting point of 15 months was manifestly excessive. It represented a 20% increase over the sentence in Musadiq despite no actual financial loss being incurred given that full restitution was made, and the offending in Musadiq was of an “aggravated, premeditated and multi-layered nature” which was absent in the present case. Accordingly, a sentence lower than that imposed in Musadiq should be imposed.

18 Second, the DJ erred in his application of the Prakash Framework and in failing to consider “existing relevant precedent cases”. Unlike the First Cheating Charge, the DJ’s analysis of the Second Cheating Charge did not track the three-stage approach under the Prakash Framework. The DJ determined the appropriate sentence by relying solely on Musadiq, a case which pre-dated the Prakash Framework. If the Prakash Framework had been correctly applied, a final sentence of seven months’ imprisonment was just and reasonable, for the following reasons.

(a) At Stage 1, the base offence under s 420 of the PC offence attracted a sentence of four to five months’ imprisonment on a claim trial basis. The sentencing precedents indicated a sentencing range of five to six months’ imprisonment for a s 420 charge involving a similar quantum. However, the offence-specific factors featured in the sentencing precedents, such as economic harm, abuse of trust, a high degree of sophistication and fraudulence, and a sizeable number of victims, were absent in the present case, warranting a downward calibration from the five to six months sentencing range.

(b) At Stage 2, the appellant’s culpability over the entire course of conduct warranted an uplift from the Stage 1 base offence sentence to five to six months’ imprisonment. The present case involved the misuse of financial instruments threatening the conduct of commerce, as well as an offending duration and frequency comparable to the sentencing precedents. However, these factors were mitigated by the lower economic harm and lesser degree of sophistication and difficulty of detection.

(c) Finally, at Stage 3, the offender-specific aggravating factors warranted an uplift from the Stage 2 sentence to nine to ten months’ imprisonment. These factors included the commission of the offence while on bail and the presence of the three similar TIC Charges. However, a reduction to seven months was appropriate after applying a 30% guilty plea discount and giving due weight to the fact that full restitution had been made.

19 Third, the DJ failed to accord sufficient weight to the appellant’s remorse as evidenced by her making full restitution.

20 Finally, the DJ erred in treating the TIC Charges as an aggravating factor to uplift the sentences for both the First Cheating Charge and Second Cheating Charge. The appellant contended the TIC Charges were impermissibly considered at three separate stages of the DJ’s sentencing analysis: first, as an aggravating factor in calibrating the sentence of the First Cheating Charge; second, as an aggravating factor in calibrating the sentence for the Second Cheating Charge; and third, as a reflection of the appellant’s overall criminality in determining whether the aggregate sentence was crushing or disproportionate. This amounted to impermissible double-counting. Instead, little to no aggravating weight ought to have been accorded on the same TIC Charges when determining the sentence for the Second Cheating Charge.

21 The Prosecution submitted that the DJ’s sentence of 18 months’ imprisonment for the Second Cheating Charge was not manifestly excessive and should be upheld.

22 First, the Prosecution defended the DJ’s reliance on Musadiq. The harm-culpability assessment should be based on the totality of the offender’s conduct and not the number of discrete deceptive steps or “layers” that were employed. In this regard, the appellant’s modus operandi was a “deliberate, multi-step process of fabrication” that was comparable in sophistication and concealment to Musadiq.

23 Second, the appellant’s allegation that the DJ failed to consider existing sentencing precedents was erroneous as they had never been placed before the DJ.

24 Third, the DJ correctly applied the Prakash Framework. The DJ correctly identified the relevant sentencing factors such as quantum of loss and level of sophistication, anchored the starting point at 15 months’ imprisonment on the basis of Musadiq at Stage 2, and thereafter correctly applied a three-month uplift for offender-specific aggravating factors at Stage 3.

25 On the question of double counting, the Prosecution submitted that there was no strict prohibition against relying on the same TIC charges in the sentencing analysis for more than one proceeded charge, and that it was appropriate to do so in the present case given that the TIC Charges were relevant to both cheating charges. To hold otherwise would lead to the absurd result of understating the appellant’s overall criminality.

26 The appellant, in its written submissions, did not pursue all the grounds set out in the Petition of Appeal and expressly abandoned some of them.

(a) First, the appellant did not seek to disturb the individual sentences imposed for the First Cheating Charge, Theft Charge and Obstruction of Justice Charge.

(b) Second, the appellant no longer pursued the ground that the DJ erred in finding that the aggregate sentence was not crushing and in ordering the sentences for the Theft Charge, Obstruction of Justice Charge and Second Cheating Charge to run consecutively. The appellant accepted that those sentences should be run consecutively.

(c) Third, the appellant did not seriously pursue in oral submissions the ground that the DJ failed to place sufficient weight on the appellant’s mental condition.

(d) Finally, the appellant made no submissions on the ground that the DJ erred in finding that a “sentence of 18 months’ imprisonment on the full spectrum of a maximum of 10 years was reasonable and did not breach the principle of fairness”.

Therefore, in determining the appeal, I proceeded on the basis that these grounds were abandoned.

27 The primary issue on appeal was therefore whether the sentence of 18 months’ imprisonment imposed for the Second Cheating Charge was manifestly excessive. The determination of this issue turned on whether the DJ erred in his application of the Prakash Framework. In this regard, the following sub-issues arose for consideration:

(a) whether the DJ erred in his reliance on Musadiq;

(b) whether the DJ erred in failing to accord sufficient weight to the full restitution made by the appellant; and

(c) whether the DJ erred in double counting the TIC Charges.

28 Briefly, in my view, the DJ erred in principle by failing to apply each stage of the Prakash Framework sequentially. Instead, the DJ effectively collapsed Stages 1 and 2 into a single exercise, deriving the appropriate sentence by reference to Musadiq alone.

29 The appellant rightly submitted that in relying solely on Musadiq to arrive at the appropriate starting sentence, ie, the base offence, the DJ failed to follow the stages in the Prakash Framework. In particular, the DJ’s sentencing analysis of the Second Cheating Charge (see GD at [27]–[29]) appeared to have collapsed the first two stages of the Prakash Framework. Two observations supported this:

(a) First, the DJ stopped short of determining the appropriate sentence for a single incident of the offence (ie, the base offence): GD at [19] and [27]. Indeed, the Prosecution’s submissions supported the conclusion that the DJ approached the sentencing analysis on this wrong footing. In suggesting that the DJ identified s 417 of the PC as the base offence, the Prosecution accepted that the foundation of the DJ’s analysis was built on the incorrect base offence. A s 417 charge was not a s 420 charge. Section 417 is for cheating simpliciter while s 420 is an aggravated form of cheating that specifically targets instances of cheating that result in the dishonest inducement of the delivery, alteration, or destruction of property. The difference in criminality is recognised in the higher tariff an offence under s 420 carries in terms of a higher maximum imprisonment. To treat one as the sentencing anchor for the other was a fundamental error and a failure to apply Stage 1 of the Prakash Framework correctly.

(b) Second, without addressing Stage 1 of the Prakash Framework by determining the sentence for the base offence, the DJ proceeded to analogise the present case to Musadiq, a case which concerned an amalgamated cheating charge. As I put to the Prosecution at the hearing, the point was a simple but fundamental one. Stage 1 required the base offence to be drawn from a single incident. An amalgamated charge, by its very nature, was the antithesis of that as it was a creature of multiple incidents. To use an amalgamated charge as a Stage 1 reference was to conflate the building block with the building itself.

30 Indeed, there were several principled and practical difficulties with the DJ’s approach of bypassing Stages 1 and 2 of the Prakash Framework by directly relying on a sentencing precedent involving an amalgamated charge.

31 To take this approach is to skip the essential question of what the course of conduct in an amalgamated offence adds to the seriousness of the base offence. Not only is the maximum sentence for an amalgamated charge double the prescribed punishment for the relevant base offence (s 124(8)(a) of the CPC), but this adjustment also confers on the court the discretion to impose a punishment that “better takes into account the offender’s criminality for more than one offence” [emphasis in original]: Prakash at [36]–[40]. The observations in Lian Hoe Heng v Public Prosecutor [2026] 1 SLR 709 at [93]–[96] on the effect of amalgamation and s 124(8)(a) of the CPC are relevant:

(a) First, while s 124(8)(a) of the CPC doubled the sentencing limit for amalgamated charges to reflect the higher criminality and harm occasioned by an offending course of conduct across multiple instances, it did not operate as a mechanical rule requiring the court to impose double the sentence that would had been imposed for a single-incident offence causing the same harm.

(b) Second, the rationale underlying s 124(8)(a) was to ensure parity between an accused convicted of an amalgamated charge and one who was hypothetically convicted separately for each incident, thereby counteracting the risk that an accused might receive an undue discount for multiple offending if the sentencing limit were capped at the maximum for a single offence.

(c) Third, in calibrating the sentence for an amalgamated charge, the court must accord adequate aggravating weight to the fact that the accused “engaged in a course of criminal conduct rather than a single incident”, as any failure to do so would result in precisely the undue discount that s 124(8)(a) of the CPC was designed to prevent.

32 Stage 2 of the Prakash Framework required an assessment of the starting sentence for the amalgamated offence to be founded on a holistic assessment rather than reducing the exercise to a comparison of incident numbers or aggregate outcomes. This was in my view precisely why the court in Prakash at [42] and [129(a)] directed that the sentence of the base offence at Stage 1 should be determined by reference to “the relevant sentencing framework or benchmark for the base offence (or analogise to sentencing precedents where no framework or benchmark is applicable) for individual base offences”. On this basis, having established the relevant sentencing conditions, range and starting point for the base offence, the court considers at Stage 2 how the course of offending should be situated within the doubled punishment limit in s 124(8)(a)(ii) of the CPC. This approach best gave effect to the enhanced harm and culpability arising from the offender’s course of conduct, assessed by reference to the aggregate harm caused, as well as the frequency and duration of the offending, as a result of amalgamation. It also avoids the risk that an undue discount is accorded to an offender who has committed multiple offences.

33 Without first identifying the base offence’s punishment range and the offence-specific factors relevant to it, the court would have no principled basis for situating the offending conduct on the sentencing spectrum under Stage 2. In practical terms, existing sentencing frameworks and benchmarks developed by the courts are calibrated for individual offences rather than amalgamated offences. Without Stage 1 anchoring the analysis in the existing base offence jurisprudence, the court would be commencing the sentencing analysis from the wrong starting point.

34 Further, to the extent the DJ collapsed Stages 1 and 2 of the Prakash Framework into a single undifferentiated step, this approach was expressly rejected in Prakash itself. Musadiq at [18] quoted Menon CJ in Gan Chai Bee Anne v Public Prosecutor [2019] 4 SLR 838 (“Gan Chai Bee Anne”) at [23] to suggest that “the sentencing judge would likely not be faulted if he approached sentencing for that charge in a single step, focusing solely on the totality of the accused’s criminal conduct” [emphasis added]. However, Prakash rejected this single-step approach by proposing a three-step framework where identifying the relevant sentence for the base offence’s punishment range is a necessary preliminary step. 

35 Ultimately, each stage of the Prakash Framework performs a distinct and irreplaceable analytical function. From the above, Stage 1 anchored the analysis in established base offence jurisprudence, while Stage 2 holistically calibrated the sentence for the amalgamated offence to ensure adequate aggravating weight is given to the course of conduct. Thus, Musadiq could not serve as a Stage 1 reference precedent, as it was itself an amalgamated charge rather than a single-incident offence.

36 For completeness, this was not a case where each offence forming the amalgamated charge was “part and parcel of a course of conduct directed at securing a particular illicit outcome” in the manner contemplated in JEH v Public Prosecutor [2026] 3 SLR 1401 (“JEH”) at [50]. JEH concerned a single amalgamated charge under s 182 of the PC for providing false information to a public servant, namely the vice principal of a primary school, on at least five occasions between 6 August 2024 and 7 October 2024. Menon CJ held that each instance of providing false information to the vice principal of a primary school, regarding the residential address of the offender’s daughter, was inextricably linked to a single overarching objective of securing the continued wrongful enrolment of the daughter at that school. Each lie was, in that sense, a necessary step toward a singular illicit end. In contrast, the Second Cheating Charge involved discrete instances of cheating by means of a doctored PayNow screenshot receipt. No single instance was essential or necessary to any other. There was no common objective that undergirded each incident of offending. The appellant was not working toward a single goal. Instead, she was simply repeating the same opportunistic fraud, one transaction at a time.

37 The DJ therefore erred in the application of the Prakash Framework by relying solely on Musadiq. Accordingly, the appropriate sentence for the Second Cheating Charge had to be considered afresh on the basis of the correct facts and principles: see Kavitha d/o Mailvaganam v Public Prosecutor [2017] 4 SLR 1349 at [15].

38 During oral submissions, following questions posed by the Court, the Prosecution did not substantively contest whether: (a) the starting point adopted by the DJ for the s 420 base offence was consistent with the sentencing precedents cited by the appellant; and (b) the final sentence arrived at by the DJ could be rationalised if the Prakash Framework was correctly applied. In the absence of relevant sentencing precedents from the Prosecution, I relied on the sentencing precedents cited by the appellant’s counsel in determining afresh the appropriate sentence.

39 Having considered the relevant sentencing factors and the sentencing precedents cited by the appellant, I found an indicative sentence of four months’ imprisonment appropriate for the base offence.

40 The prescribed punishment for an offence under s 420 of the PC is imprisonment for a term which may extend to ten years: s 420(2) of the PC.

41 Economic harm lies at the heart of financial and property offences. The greater the value misappropriated, the heavier the sentence (Gan Chai Bee Anne at [42], citing Public Prosecutor v Fernando Payagala Waduge Malitha Kumar [2007] 2 SLR(R) 334 (“Fernando”) at [47] and Lim Ying Ying Luciana v Public Prosecutor [2016] 4 SLR 1220 at [69]), and rightly so. The quantum of economic harm is a direct proxy for two critical sentencing considerations (Gan Chai Bee Anne at [42]): (a) the criminal benefit reaped by the offender; and (b) the harm visited upon the victim. That said, sentencing for cheating offences is not a simple arithmetic exercise. The sentences for cheating offences should not bear a “relationship of linear proportionality with the sums involved” and must, instead, involve a holistic assessment of the culpability of the accused and the circumstances of the case: Public Prosecutor v Tan Cheng Yew [2013] 1 SLR 1095 at [184].

42 In Soh Chee Wen v Public Prosecutor [2026] 1 SLR 214 at [45], the Court of Appeal set out a non-exhaustive list of aggravating factors for cheating charges:

(a) The value of the property involved will ordinarily be a relevant sentencing consideration, as it reflects both the degree of criminal benefit received by the offender and the extent of harm caused to the victim.

(b) The extent of harm caused is a relevant but not decisive factor. The fact that minimal or no loss was ultimately suffered, whether due to fortuitous circumstances or the subsequent recovery of the property by the authorities, does not diminish the gravity of the offence.

(c) The difficulty of detecting the offence, including any steps taken by the offender to conceal his offence.

(d) The efforts at preventing a victim from seeking or accessing help is aggravating as it enhances the culpability of the offender and increases the risk that losses may not be recovered.

(e) The length of offending, particularly when it is accompanied by sustained and highly sophisticated planning aimed at concealing and maintaining the offending.

(f) The harm that has been caused to Singapore’s reputation as a financial centre given Singapore’s standing both as a financial hub and a preferred centre of commerce.

43 In the present case, as there was no applicable sentencing framework and/or benchmark for non-credit card cheating offences (cf, Fernando at [55] and [75]), regard should be given to sentencing precedents: Prakash at [42].

44 I therefore turned to the sentencing precedents cited by the appellant. Those sentencing precedents indicated that a loss range of approximately $2,500 to $9,500 for a single s 420 offence broadly attracted a range of three to six months’ imprisonment. I found that the appellant’s offending fell at the lower end of the range.

45 I began by setting out the salient facts of the sentencing precedents relied on by the appellant.

46 In Public Prosecutor v Chua Peng Jun [2020] SGDC 166 (“Chua Peng Jun”), the accused was charged with, amongst others, a total of 19 cheating charges under s 420. The accused was an untraced sales executive of a car rental company. He was alleged to have deceived customers into making balance payments for cars by falsely representing that their vehicles were ready for collection. The accused was convicted after trial. The sentencing outcome was as follows:

(a) The court accorded weight to the following aggravating factors: (i) no restitution was made (at [543(c)]); (ii) the scheme demonstrated careful planning and concealment, as evidenced by the issuance of fabricated sales agreements and invoices bearing the company’s letterhead and stamp (at [543(e)]); and (iii) the accused, acting as a representative of the company, had abused the position of trust reposed in him by customers (at [543(f)]).

(b) The sentences imposed for the s 420 offences which involved amounts closest to the present case were as follows:

(i) At the high end, for the charges involving sums of $11,924, $10,000, $9,677 and $9,400, a sentence of five months’ imprisonment was imposed.

(ii) At the medium end, for the charges involving sums of $7,567, a sentence of four months’ imprisonment was imposed.

(iii) At the lower end, for the charges involving sums of $5,321 and $4,211, a sentence of three months’ imprisonment was imposed.

The above reflected a sentencing range of three to five months’ imprisonment for the sums of approximately between $4,000 and $7,500.

47 In Public Prosecutor v Tan Peng Khoon [2015] SGDC 94 (“Tan Peng Khoon”), the accused was charged with, amongst others, a total of two cheating charges under s 420. The accused was untraced insurance agent who became friends with the victim. He was alleged to have deceived the victim to deliver a cheque of $2,018 and $6,500 on the false basis that the funds were needed to terminate the victim’s life insurance policy and obtain a policy loan from her insurer. The accused was convicted after trial. The sentencing approach was as follows:

(a) The court accorded weight to the following aggravating factors: (i) as the victim’s insurance agent, the accused abused the trust placed in him by the victim (at [216]); (ii) the victim was a vulnerable elderly illiterate Mandarin-speaking widow who did not complete her primary school education (at [218]); (iii) there was a high degree of planning and premeditation as an elaborate scheme was devised to obtain the proceeds of insurance cheques payable only to the victim, which included manipulating the victim into adding the accused as a joint account holder, impersonating the victim’s son in a call on the insurer’s customer service hotline, and deliberately circumventing the checks and balances of one financial institution by exploiting the services of another (at [219]–[220]); and (iv) the absence of any restitution (at [221]).

(b) The accused was sentenced to respectively three months’ imprisonment and six months’ imprisonment in respect of the sums of $2,018 and $6,500.

48 In Public Prosecutor v Mohamad Jamal Abul Hashem [2013] SGDC 3 (“Jamal”), the accused was charged with a total of nine cheating charges under s 420. The accused was alleged to have deceived Bangladeshi nationals into believing that he was providing a service of arranging for work permits to work in Singapore. The accused was convicted after trial. The accused had one relevant antecedent, namely a prior conviction for criminal breach of trust under s 406 of the PC. The sentencing approach was as follows:

(a) The court accorded weight to the following aggravating factors: (i) the accused was not a first offender (at [179]); (ii) the accused displayed considerable premeditation and planning in that he deliberately preyed on fellow countrymen who were vulnerable and desperate for employment in Singapore (at [176]); (iii) the accused’s methods of deception were systematic and calculated, including the production of fabricated in-principle approval documents purportedly issued by the Ministry of Manpower to induce payment, using partial repayments and fresh documents to sustain the victims’ trust, and appealing to religion and kinship to extract further sums (at [176]); (iv) the accused’s exploitation of the victims’ desperation to recover their losses by inducing them to give him further business, thereby compounding the harm (at [177]); and (v) the accused’s repeated threats to victims that they would recover nothing if they reported him to the police, which successfully suppressed any police reports until the scheme unravelled through his own actions (at [177]–[178]).

(b) The sentences imposed for the s 420 offences involving amounts closest to the present case were as follows:

(i) At the high end, for the charges involving sums of $9,300, a sentence of six months’ imprisonment was imposed.

(ii) At the low end, for the charges involving sums between $3,000 to $4,400, a sentence of four months’ imprisonment was imposed.

The above reflected a sentencing range of four to six months’ imprisonment for sums approximately between $3,000 and $9,300.

49 From these sentencing precedents, for a single s 420 PC offence, a range of loss of approximately between $2,500 and $9,500 broadly attracted an indicative sentence of between three and six months’ imprisonment, with the precise sentence calibrated according to the aggravating factors present in each case. In my judgment, the appellant’s culpability was low and placed her in the lower end of the range of sentence.

50 While the exploitation of a nationwide digital payment service in this case was an aggravating feature that struck at the integrity of infrastructure that underpins public trust in Singapore’s digital economy, this was tempered by the unsophisticated nature of the scheme perpetrated by the appellant. I did not accept the Prosecution’s characterisation of the appellant’s conduct as a “deliberate, multi-step process of fabrication that made her deception harder to detect”. The fact was that there was only one layer of deception as I will explain.

(a) The appellant’s modus operandi was rudimentary, confined to the editing of PayNow screenshots on her phone.

(b) Far from concealing her tracks, the simplicity of her method and her use of her mobile phone rendered her readily detectable. Her number, identity and address were known. Therefore, the appellant’s deception was limited to the fact of payment.

(c) Unlike Chua Peng Jun, the appellant repeatedly employed the same modus operandi on the same victim. Each further transaction provided another opportunity for the Restaurant to verify the purported payment. This might be contrasted to the carefully insulated schemes in Chua Peng Jun that targeted a larger pool of unconnected victims, which minimised the risk of detection.

(d) Furthermore, the appellant’s conduct lacked the aggravating features of suppression, threats or abuse of trust seen in Jamal and Tan Peng Khoon.

51 Therefore, I found that the quantum of loss in the present case of $6,167.52 in the present case attracted an indicative sentence of four months’ imprisonment for the base offence that sufficiently accounted for the exploitation of a nationwide digital payment service widely used by businesses and individuals. In my view, general deterrence should feature prominently, so as to deter potential offenders and to safeguard the integrity of Singapore’s digital payment infrastructure and its standing as a financial hub.

52 I assessed the aggregate harm and culpability of the appellant over the entire course of conduct to be low to moderate and imposed an uplift of two months under Stage 2. This yielded an indicative sentence of six months’ imprisonment for the amalgamated Second Cheating Charge.

53 I began by setting out the sentencing precedents cited by the appellant, which served to illustrate the range of uplifts that would be appropriate under Stage 2 of the Prakash Framework.

(a) In Public Prosecutor v Muhammad Haekal Bin Abdul Latif [2026] SGDC 96 (“Haekal”), the accused was charged with a single amalgamated s 420 charge. The accused was a food delivery driver who systematically deceived a delivery platform into issuing vouchers and refunds by falsely representing that he was legitimately entitled to them. At Stage 2, an uplift of four months was applied to arrive at a sentence of 18 months’ imprisonment. The court found the accused’s culpability to be high for the following reasons (at [71]).

(i) The accused had carefully and deliberately planned his scheme, drawing on techniques gleaned from online sources that instructed others on how to defraud delivery platforms for quick financial gain. He specifically targeted the victim’s platform having identified it as one that issued refunds readily with minimal scrutiny (at [30(a)]).

(ii) The accused admitted that he created numerous e-mail addresses to create various accounts with the delivery platform in order to allay suspicion when seeking refunds and to avoid detection.

(iii) The accused offended at a relentless pace, averaging nearly three offences per day (ie, 246 occasions over approximately two months).

(b) In Public Prosecutor v Chai Song Lin [2026] SGDC 4 (“Chai Song Lin”), the accused was charged with a single amalgamated s 420 charge. The accused, an assistant manager at a foreign domestic worker agency, exploited his client-facing role to divert customer payments into his personal bank account by manipulating his PayNow display name to impersonate the company. He did so by providing clients with QR codes linked to his own account, thereby deceiving clients into believing they were paying the company directly. At Stage 2, the court found that a low-to-moderate uplift of five and a half months was warranted and arrived at an indicative sentence of 29.5 months’ imprisonment. The following factors were considered:

(i) The accused committed the offence on 72 occasions over a five-month period, which reflected persistence in furtherance of a criminal enterprise (at [43]).

(ii) The overall harm was high as the outstanding loss after restitution of $10,000 was $78,692.40 (at [44]).

54 In Haekal, there were 246 instances of offending over two months with a total loss of $23,039.43. In Chai Song Lin, there were 72 instances of offending over five months with a total loss of $78,692.40. The uplifts were four months and five and a half months respectively. The present case involved a significantly lower quantum of loss, less frequent offending and minimal premeditation when compared to Haekal and Chai Song Lin, which would ordinarily justify a lower uplift.

(a) The level of premeditation, planning and difficulty of detection was comparatively low. The appellant employed a repetitive and unsophisticated modus operandi against the same victim, making no meaningful attempt to conceal her conduct or evade detection. This stood in stark contrast to Haekal, where the accused had attended online lessons on how to defraud delivery platforms, and went further to create multiple accounts using many email addresses to frustrate detection.

(b) The appellant offended on 24 occasions over approximately 11 months. While the extended duration of offending was aggravating, the relatively low frequency of offending, when compared to the above-mentioned cases, suggested a lower degree of recalcitrance and overall culpability than the offenders in those cases.

(c) The amount involved in the present case was significantly lower than in Haekal and Chai Song Lin. Moreover, the appellant made full restitution, substantially reducing the economic harm suffered by the Restaurant. In contrast, the accused in Chai Song Lin made only partial restitution amounting to a fraction of the total loss (see [53(b)(ii)] above), while the accused in Haekal made none (Haekal at [75(a)]).

55 I therefore found a two-month uplift to be appropriate.

56 I found a further uplift of five months was warranted under Stage 3, bringing the sentence to 11 months’ imprisonment before the plea of guilt discount. The DJ’s uplift of three months’ imprisonment was, in my view, insufficient to adequately reflect the degree of the appellant’s recalcitrance.

57 At this stage, adjustments might be made on the basis of aggravating factors such as relevant antecedents indicating a lack of remorse, and mitigating factors such as a timely plea of guilt and the offender’s genuine remorse: Prakash at [46] and [129(c)].

58 To begin with, limited mitigatory weight should be placed on the appellant’s full restitution at this stage of the Prakash Framework.

59 Restitution carries two independent mitigatory effects on sentence in that it demonstrates the offender’s remorse and ameliorates the harm occasioned to the victim: Gan Chai Bee Anne at [62]–[63]. Against this, full restitution was relevant at Stage 2 and Stage 3 of the Prakash Framework.

(a) At Stage 2, full restitution affected the aggregate harm caused by the offender’s course of conduct. In this regard, the harm-reduction function of restitution operated to lower the aggregate harm that would otherwise have been occasioned had no restitution been made: cf, Prakash at [65] and [67].

(b) At Stage 3, full restitution was an offender-specific mitigating factor to the extent that the voluntary restitution reflected true remorse. Indeed, the court considers at Stage 3 mitigating factors such as “the genuine remorse of the offender”, which warranted a downward adjustment from the starting point: Prakash at [46]. Crucially, the appellant’s full restitution must be “carefully scrutinised to see whether they constitute[d] evidence of genuine, heartfelt remorse” (Chen Weixiong Jerriek v Public Prosecutor [2003] 2 SLR(R) 334 (“Chen Weixiong Jerriek”) at [23]), having regard to its voluntariness, timeliness and genuineness (Soong Hee Sin v Public Prosecutor [2001] 1 SLR(R) 475 at [9]).

60 That said, where the harm-reducing effect of restitution was fully accounted for at the Stage 2 assessment of aggregate harm, ie, the reduction of actual harm caused to the victims, which was the case here (see [54(c)] above), the court should be cautious against attaching it further mitigating weight at Stage 3 on the same basis, lest it amount to double-counting.

61 I should point out that the mitigatory effect of the appellant’s remorse, as evidenced by her restitution, was substantially outweighed by two serious aggravating factors. First, the appellant’s commission of similar offences while on bail reflected a manifested proclivity for offending that warranted a stiff sentencing response in the interests of deterrence. Second, the three related TIC Charges which warranted significant weight.

62 The appellant’s contention that the bail aggravating factor should carry limited weight because only some of the 24 instances of offending occurred while she was on bail was without merit.

63 At the outset, the factual premise of the appellant’s submission was inaccurate. The entirety of the offending conduct was committed while the appellant was on bail (see [9] above).

64 In any event, the mere fact of committing any offence while on bail was in itself an aggravating factor, regardless of whether it applied to all or only some of the instances.

65 Conceptually, the commission of an offence while on bail engaged the sentencing principles of retribution and deterrence in both its specific and general dimensions: Public Prosecutor v Loqmanul Hakim bin Buang [2007] 4 SLR(R) 753 (“Loqmanul”) at [57]–[61]. It reflected both the abuse of trust reposed in the accused and the manifested proclivity for reoffending.

(a) The starting point is that bail is not merely an administrative arrangement to secure the accused’s attendance in court. Instead, the grant of bail is premised on “a calculated assessment on the part of the courts (or the police, in the case of police bail), incorporating both a belief and trust that the alleged offender would not abuse his liberty to reoffend against society”: Loqmanul at [54]. The accused’s culpability was enhanced because she “exploited the trust that has been reposed in [her] by the State”: Loqmanul at [54]. Necessarily, this trust was binary: either the accused respected the conditional liberty granted to her, or she did not. It followed that even a single transgression, ie, committing one offence while on bail, was more than sufficient to justify an uplift in sentence. The very act of reoffending while on bail was a manifest abuse of the trust reposed in the accused by the court or the police. It was no mitigation to say that only some of the reoffending occurred while on bail.

(b) Specific deterrence was also engaged. An accused who committed even a single further infraction on bail “manifest[ed] a lack of any desire to cease unlawful activity” notwithstanding being given “an opportunity to show contrition”. This therefore warranted a stiffer sentence to “precipitate introspection and self-examination of their conduct”: Loqmanul at [59].

(c) General deterrence was equally engaged. The observations of Street CJ in R v Richards [1981] 2 NSWLR 464 at 465, endorsed in Loqmanul at [60] are apposite:

This signal must be sent whenever an accused offends on bail, regardless of whether the offender committed all, some or only one of the charged offences during the bail period.

66 My view was fortified by the statutory framework for bail. Section 94(1)(c) of the CPC imposes the following default condition on every grant of bail:

The phrase “must not commit any offence” underscored the view that a single transgression on bail was a breach of the accused’s undertaking to be law-abiding.

67 The appellant’s attempts to highlight her “sincere remorse” rang a little hollow. The appellant’s reoffending while on bail demonstrated that she was not truly remorseful, held the law in disregard and had little compunction about repeating similar offences: Chen Weixiong Jerriek at [23]; Loo Pei Xiang Alan v Public Prosecutor [2015] 5 SLR 500 at [29].

68 I was of the view that significant weight should be placed on the TIC Charges, which notably included an additional amalgamated offence under s 420 of the PC. The appellant’s contention that uplifting the sentences of the First and Second Cheating Charges on the basis of the TIC Charges constituted impermissible double counting was misconceived. The authorities relied on by the appellant only prohibited the cross-stage reuse of a sentencing factor within a single charge and not its application to different charges.

69 The appellant misconstrued Muhammad Sutarno bin Nasir v Public Prosecutor [2018] 2 SLR 647 (“Sutarno”) at [17] as standing for the broad proposition that the same TIC charges should not be relied upon to uplift the sentences for more than one proceeded charge. Menon CJ’s observation in Sutarno at [17] should be read in context:

70 The final sentence in the passage above stood for a far narrower proposition. Read in context, the final sentence above served only to emphasise that TIC charges irrelevant to a particular proceeded charge ought not to be relied upon to uplift the sentence, in circumstances where the same TIC charges had already been applied to uplift the sentence for a similar proceeded charge. Doing so could amount to double counting.

71  Public Prosecutor v Raveen Balakrishnan [2018] 5 SLR 799 (“Raveen”) is relevant. At [87] of Raveen, it was stated that a sentencing factor should not be double counted at two different sentencing stages. The complaint there concerned cross-stage, not cross-charge, reuse of aggravating sentencing factors. The illustrations in Raveen at [87] are instructive on this point: a factor already taken into account in ordering sentences to run consecutively should not be relied upon again to uplift the length of an individual sentence. This instance of double counting did not occur in the present case or in the DJ’s reasoning. The DJ’s decision to order consecutive sentences was driven by the distinct nature of the offences, which protected different legal interests, were committed on different occasions and were directed against different victims: GD at [35].

72 In any event, the appellant’s double-counting objection was misconceived as the DJ did not take the two cheating charges into account when deciding whether to run the two charges concurrently or consecutively, and in any event, if there was any double counting it was not consequential as the two cheating charges were directed to run concurrently. The net effect of the uplift was only reflected once in the aggregate sentence, through the Second Cheating Charge, which ran consecutively with the sentences for the Theft Charge and Obstruction of Justice Charge.

73 Significant weight should therefore be attached to the presence of the TIC Charges for the purpose of the Stage 3 analysis in the present case.

74 For the reasons above, a further uplift of five months was therefore warranted. This yielded an aggregate sentence of 11 months’ imprisonment at Stage 3, before applying the plea of guilt discount.

75 The appellant pleaded guilty within stage 1 of the Sentencing Advisory Panel’s Guidelines on Reduction in Sentences for Guilty Pleas (“PG Guidelines”). I therefore accorded her the full 30% reduction in sentence available under the PG Guidelines. Applying a 30% reduction to the sentence of 11 months’ imprisonment and rounding up to the nearest whole month, I arrived at a final sentence of eight months’ imprisonment, which was commensurate with the appellant’s overall criminality.

76 For the reasons above, the appellant’s appeal against sentence was allowed. I would take this opportunity to record the court’s appreciation for the submissions and assistance of Mr Sujesh Anandan from Pro Bono SG, counsel for the appellant, which were beneficial in my deliberations.