1 This appeal is against sentences of six weeks’ imprisonment that were imposed for offences under s 22(1)(d) read with s 23(1) and punishable under s 22(1)(ii) of the Employment of Foreign Manpower Act (Cap 91A, 2009 Rev Ed) (“EFMA”), for abetment by engaging in a conspiracy to make false declarations in connection to three work pass applications.

2 The Prosecution has argued that this appeal provides the High Court the opportunity to set out a sentencing framework that provides guidance on the correct approach to sentencing s 22(1)(d) offences that will utilise the full sentencing range prescribed under s 22(1)(ii) of the EFMA. Having considered the submissions, I agree that guidance in this regard is due and set out the sentencing framework to be applied for offences under this provision.

3 The appellant pleaded guilty to 18 charges under the EFMA and the Employment Act (Cap 91, 2009 Rev Ed). He also consented for 43 charges under both Acts to be taken into consideration for the purposes of sentencing, of which two were also under s 22(1)(d) read with s 23(1) and punishable under s 22(1)(ii) of the EFMA.

4 The facts are set out in Public Prosecutor v Chiew Kok Chai [2018] SGMC 70 (“GD”) at [4]. At the material time, the appellant and Mr Tan Yock Jeen (“Tan”) jointly managed the operations of Wee Chong Construction (“Wee Chong”) and Wan Fu Builders Pte Ltd (“Wan Fu”). The appellant was a registered director of Wan Fu, and was involved in managing the construction projects and foreign employees of both businesses, deploying employees to different worksites and ensuring the payment of employees’ salaries.

5 The EFMA charges concerned a conspiracy between the appellant and Tan to obtain foreign manpower for Wan Fu, which was not entitled to a foreign manpower quota due to its previous levy defaults. The pair agreed that Tan would apply for work pass applications for three foreign employees to be ostensibly employed by Wee Chong, with the intention that they be employed by Wan Fu instead.

6 Tan duly submitted the three work pass applications to the Work Pass Division (“WPD”) of the Ministry of Manpower (“MOM”) on this basis. The three foreign employees were issued work passes, and the WPD confirmed that it would not have approved the applications but for Wee Chong’s false declarations. The three employees worked solely for Wan Fu as construction workers. Two of the foreign employees worked for Wan Fu for about five months; the third worked for Wan Fu for about six months.

7 The appellant was charged, convicted and sentenced under s 22(1)(d) read with s 23(1) and punishable under s 22(1)(ii) of the EFMA. The relevant provisions are set out for ease of reference. Sections 22(1)(d) and 22(1)(ii) state:

8 Section 23(1) states:

As offences under s 22(1)(d) read with s 23(1) and those under s 22(1)(d) are punished under the same provision, this judgment will deal with the sentencing approach taken towards s 22(1)(d) offences generally.

9 The District Judge held that general and specific deterrence are the primary sentencing principles in relation to offences that undermine the work pass regulatory framework: at [22]. Where employers intentionally make false declarations to MOM to employ foreign workers whom they are otherwise not entitled to employ, so as to meet their business needs, a financial penalty might amount to a mere business cost factored into the business’s balance sheet. A custodial sentence would be more likely to serve the deterrent effect that Parliament had intended: at [26] and [27].

10 The District Judge’s view was “fortified” by the High Court’s sentencing frameworks for offences under s 57(1)(k) of the Immigration Act (Cap 133, 1997 Rev Ed) (“the IA 1997”) and s 182 of the Penal Code (Cap 224, 2008 Rev Ed) (“the Penal Code”): at [28]. In Abu Syeed Chowdhury v Public Prosecutor [2002] 1 SLR(R) 182 (“Chowdhury”) at [26], the High Court held that a custodial sentence should be the applicable norm where a false representation is made under s 57(1) of the IA 1997. Similarly, the High Court in Koh Yong Chiah v Public Prosecutor [2017] 3 SLR 447 (“Koh Yong Chiah”) at [50] held that custodial terms should be imposed as a starting point where offences under s 182 of the Penal Code, which involve the making of false representations to public servants, result in “appreciable harm”.

11 The District Judge ultimately refrained from adopting the Prosecution’s proposed sentencing framework: at [35]. Nonetheless, he agreed that its submitted sentences of six weeks’ imprisonment for each EFMA charge were appropriate and in line with the sentences imposed in Chowdhury, Koh Yong Chiah and five recent s 22(1)(d) cases: at [36]. Two of the three custodial sentences were to run consecutively. In so deciding, he also considered the offence- and offender-specific factors that the Prosecution had raised: at [37].

12 The appellant argued that a fine should have been imposed, in line with the sentencing matrix established by the body of s 22(1)(d) cases. Under this sentencing matrix, fines are generally imposed where a false declaration pertains to the salaries payable to foreign workers; the fact that the offender will employ a foreign worker when he has no intention to do so; or a foreign worker’s educational qualifications. The custodial threshold is crossed in more serious cases involving, inter alia, the declaration of “phantom” workers to boost companies’ foreign worker entitlements and where an element of forgery is involved. The categories of principal factual elements determining the appropriate sentencing starting points were set out: ¹

13 Sentences are adjusted from the applicable starting points to account for the following non-exhaustive sentencing considerations: (a) the materiality, nature and extent of the deception; (b) the role and involvement of the offender in the deception; (c) the consequences of the deception; and (d) offender-specific aggravating and mitigating factors. ² The appellant argued that the present matrix correctly reflects Parliament’s intent: it allows for proportionality in sentencing and for the full sentencing range under s 22(1) to be fully utilised. While cases decided prior to the 2012 EFMA amendments resulted in fines of approximately $4,000, cases decided after 2012 saw the imposition of $8,000 fines. The District Judge erred in concluding that a fine was insufficient, given that he had the latitude to impose fines of up to $20,000. ³

14 Finally, there was no need to deviate from or review the current sentencing practice as there was no evidence that a fine was no longer an effective deterrent. The District Judge also misapplied Chowdhury and Koh Yong Chiah, as s 57(1)(k) of the IA 1997 and s 182 of the Penal Code are not in pari materia with s 22(1)(d) of the EFMA, invoking different culpability considerations and involving different penalties and sentencing frameworks. ⁴

15 The young amicus curiae, Mr Chen Zhida (“the amicus”), was appointed to assist the court on the appropriate sentencing framework for s 22(1)(d) offences, taking into account the relevant sentencing principles and the full sentencing range prescribed under s 22(1)(ii) of the EFMA. He submitted that the legislative intent behind s 22(1)(d) encapsulates deterrence as a sentencing principle. He cited Parliamentary debates concerning s 22(1)(d) and its predecessor provisions, and the three increases of the maximum punishment under the provision.

16 Reviewing the case law, the amicus identified the following non-exhaustive considerations: (a) materiality of the deception; (b) nature and extent of the deception; (c) role and involvement of the offender in the deception; (d) harm caused by the deception; and (e) benefits gained by the offender as a result of the deception. He referred also to the Chowdhury sentencing considerations, which have been cited by the District Court in cases dealing with s 22(1)(d) offences. Although Chowdhury deals with a different offence, s 57(1)(k) of the IA 1997 similarly involves an offender providing false information to obtain a permit. ⁵

17 Furthermore, the upward revisions in the maximum prescribed punishment under s 22(1)(ii) of the EFMA should have resulted in corresponding increases in sentences for s 22(1)(d) offences: see Mehra Radhika v Public Prosecutor [2015] 1 SLR 96 at [27]. This, however, has not been the case: ⁶

(a) From 2010 to 2011, sentences of two months’ imprisonment were imposed if the false information was “material”. Sentences of two weeks’ imprisonment were imposed if the false information was not material: see Public Prosecutor v Soh Tze Chai [2010] SGDC 58 at [21]; Public Prosecutor v Tan Lai Heng [2011] SGDC 368 at [23] and [30]; Public Prosecutor v Franco Ong Kim Huat (Wang Jinfa) [2011] SGDC 269 at [22].

(b) At present, the sentencing benchmark is a fine of $8,000 per charge, ie, 40% of the maximum fine under s 22(1) of the EFMA. Imprisonment terms range from two to three months’ imprisonment: around 8–12% of the maximum length of imprisonment under s 22(1).

18 The sentencing regime under s 22(1)(d) should be reviewed given: (a) the legislative intent of introducing more severe penalties to achieve a stronger deterrent effect; (b) the fact that existing sentencing benchmarks do not sufficiently utilise the available sentencing range; and (c) the present sentencing precedents’ lack of deterrent effect: see Public Prosecutor v GS Engineering & Construction Corp [2017] 3 SLR 682 at [51], [55] and [57]. In light of the above, the custodial threshold should be found to be crossed as a starting point. Any deception of a public institution which frustrates the aims of the EFMA cannot be condoned: Lim Kopi Pte Ltd v Public Prosecutor [2010] 2 SLR 413 (“Lim Kopi”) at [10] and [11]. This is supported by the consequences of s 22(1)(d) offences, which invariably carry the potential to cause serious harm to a large group of people, including honest employers placed on an uneven playing field, local workers deprived of job opportunities, and foreign workers who may be exploited. ⁷

19 The length of the sentence should be calibrated using the “two-step sentencing bands” approach utilised in Ng Kean Meng Terence v Public Prosecutor [2017] 2 SLR 449 (“Terence Ng”) at [35] to [39]. At the first stage, the court identifies the manner and mode by which the offence was committed and the harm caused by the offence. The amicus proposed sentencing bands for individual offenders (imprisonment terms) and corporate offenders (fines): ⁸

At the second stage, the court calibrates the appropriate sentence based on the aggravating and mitigating factors personal to the offender, ie, the “offender-specific” factors.

20 The Prosecution submitted that the dominant sentencing considerations for offences under s 22(1)(d) of the EFMA are general and specific deterrence, especially given the legislative history of the provision. ⁹

21 The Prosecution argued that a sentencing framework for s 22(1)(d) offences is necessary to rationalise past inconsistent sentencing practices and to provide guidance as to when the custodial threshold is crossed and how the entire sentencing spectrum is to be considered. A consistent sentencing practice is also desirable given the prevalence of false declarations offences. From 2016 to 2017, at least 134 natural persons were convicted under s 22(1)(d) of the EFMA for collectively making 494 false declarations in connection with work pass applications or renewals processed by the Controller of Work Passes (“the Controller”). ¹⁰

22 The Prosecution agreed with the amicus that the custodial threshold should be found to be crossed once an offender has been convicted of an offence under s 22(1)(d). This sentencing norm gives due weight to Parliament’s intent to deter circumventions of the work pass framework. Second, s 22(1)(d) offences pose high potential harm, cause actual harm to the integrity of the work pass framework and result in investigative resources being put towards addressing offending conduct (see Koh Yong Chiah at [44(c)] and [51(c)]). A fine would not be sufficiently deterrent: the common thread underlying false declarations in connection to work pass applications is the offender’s desire to obtain pecuniary benefits or a willingness to pay to legalise a stay in Singapore.

23 Moreover, a consistent position should be taken for sentencing for s 22(1)(d) work pass offences and offences of making false declarations to immigration authorities under s 57(1)(k) of the Immigration Act (Cap 133, 2008 Rev Ed) (“the Immigration Act”), which generally carry custodial sentences. Failing to impose custodial sentences for work pass offences would create a legal loophole in Singapore’s immigration policy, encouraging persons to legitimise a foreigner’s stay in Singapore through the work pass framework instead of through the immigration framework.

24 Finally, the courts have viewed false declaration offences seriously. In Idya Nurhazlyn bte Ahmad Khir v Public Prosecutor and another appeal [2014] 1 SLR 756 (“Idya Nurhazlyn”) at [37], the High Court held that a custodial sentence should be the starting point for false declarations under s 39(1) of the Passports Act (Cap 220, 2008 Rev Ed), in view of the maximum sentence available and Parliament’s intention “for such offences to be dealt with seriously”. ¹¹

25 The Prosecution also recommended the use of the “two-step sentencing bands” approach. Cases would be sorted into bands depending on the seriousness of the offence, with reference made to offence-specific factors, eg, the materiality of the falsehood on the mind of the decision-maker and the nature and extent of the deception (Chowdhury at [28] and [29]); and whether the work pass framework was exploited for nefarious purposes. The following sentencing categories for the sentencing of natural offenders were proposed: ¹²

At the second stage, the sentence is adjusted based on offender-specific aggravating and mitigating factors, eg, whether the offender was remorseful.

26 The Prosecution applied its proposed sentencing framework to demonstrate that the appellant’s sentence of six weeks’ imprisonment per s 22(1)(d) charge was not manifestly excessive. ¹³

27 The issues before me in this appeal are threefold:

(a) What is the appropriate sentencing framework for offences under s 22(1)(d) of the EFMA?

(b) Does the doctrine of prospective overruling preclude the application of the sentencing framework in this case?

(c) What is the appropriate sentence to impose on the appellant?

28 I accept the Prosecution’s observation that the sentences imposed for offences under s 22(1)(d) of the EFMA have been inconsistent. Recent sentencing trends appear to run along two lines, with courts either meting out fines in the range of $8,000 or imposing short custodial terms. The Prosecution highlighted two categories of cases that display these sentencing patterns. ¹⁴

29 The first category of cases involves false declarations of salary.

(a) In Public Prosecutor v Shokkanarayanan Ramakrishnan [2012] SGDC 127, an employment agent was convicted of four charges under s 22(1)(d) read with s 23(1) of the EFMA for abetting a sole proprietor in a carpentry firm (“Lau”) to falsely declare in S Pass applications that four foreign employees would be paid a monthly salary of $1,800 to $2,000, when he knew they would only be paid $800 to $900. As an employment agent aware that the grant of S Passes required minimum monthly salaries of $1,800 to $2,000, his falsehood in this regard was material. He had also masterminded the scheme, having approached Lau to advise Lau and his wife to commit the offence; prepared and submitted the applications after Lau signed on them; caused 73 fake salary vouchers to be prepared to pre-empt investigation; and subsequently contacted Lau to persuade him to conceal the truth (at [55]). In sentencing the offender to two weeks’ imprisonment per charge, the District Judge noted that the “sentencing norm” for s 22(1)(d) offences was a “custodial term” (at [60]).

(b) Conversely, fines in the range of $8,000 were imposed in 40 unreported cases decided in 2016 and 2017. ¹⁵ In Public Prosecutor v Son Mi Jun Magistrate’s Arrest Case No 907771 to 907792 of 2017, the offender made false statements on behalf of her company that seven foreign employees would be paid salaries of $4,500 to $4,580 when she knew that they would be paid less. The employees were only informed that they would receive salaries lower than promised after they arrived in Singapore, and faced the threat of being sent to their home countries if they did not agree to the lower salaries. They were eventually paid sums ranging from $1,700 to $3,800. The offender was sentenced to fines of $9,000 (one month’s imprisonment in default) for the s 22(1)(d) offences.

30 In the second category are cases where an offender falsely declares his field of employment.

(a) In Public Prosecutor v Nicanora Reyes Puyawan Magistrate’s Arrest Case Nos 904451 to 904453 of 2016, the offender falsely declared in work permit renewal forms that she would be employed as a domestic worker when she had no intention to work in that capacity. She instead intended to use the work pass granted to legalise her employment in Singapore as a freelance pub hostess, and used the work pass on that basis for two years. She was sentenced to two months’ imprisonment per s 22(1)(d) charge.

(b) In Public Prosecutor v Vergara Jerrilyn Tigno Magistrate’s Arrest Case No 909614 of 2018, the offender falsely declared that she would be employed as a domestic worker. She used the work pass granted to legalise her stay in Singapore to continue her relationship with a Singaporean citizen. She eventually started an online page selling various items. Her offence was discovered three years and nine months later. She was sentenced to six weeks’ imprisonment.

(c)  In Public Prosecutor v Florevic Vallera Go Magistrate’s Arrest Case No 907367 of 2016, the offender falsely declared that she would be employed as a domestic worker. She used her work pass to legalise her stay in Singapore for two years and two months, and worked as a freelance salesperson during that period. She was sentenced to a fine of $8,000 (four weeks’ imprisonment in default).

31 The above review shows that sentences imposed for s 22(1)(d) offences have not been entirely consistent. In particular, the wide sentencing disparity in cases involving similar fact patterns is difficult, if not impossible, to rationalise. It is with this in mind that I turn to the legislative intent that should be upheld when courts approach sentencing for these offences.

32 Section 22 of the EFMA is part of an overall regime that regulates the employment of foreign workers in Singapore. It is an instrument of social policy. Economic and business concerns are but one set of factors considered in the framework established by the EFMA; social and immigration concerns are also important considerations in this regime. This emerges from a consideration of the Parliamentary speeches introducing the EFMA and its predecessors. The legislative intent of the EFMA was articulated by the then-Acting Minister for Manpower Mr Tan Chuan-Jin (“Mr Tan”) during the second reading of the Employment of Foreign Manpower (Amendment) Bill (Bill No 22 of 2012):

(Singapore Parliamentary Debates, Official Report (11 September 2012) vol 89 (“the 2012 Parliamentary Debates”).)

33 Similar concerns were reflected in the then-Minister for Manpower Dr Ng Eng Hen’s speech at the second reading of the Employment of Foreign Workers (Amendment) Bill (Bill No 17 of 2007), where he noted as follows:

(Singapore Parliamentary Debates, Official Report (22 May 2007) vol 83 at cols 929 to 931.)

34 Returning to the 2012 Parliamentary Debates, Mr Tan’s speech is useful as it highlights a number of concerns that must be taken as influencing the legislative intent behind the EFMA. Firstly, it was contemplated that employers would try to get around the controls under the EFMA framework:

35 Secondly, Mr Tan recognised that EFMA contraventions hurt Singaporeans, resulting in the need to protect the integrity of the work pass framework. The 2012 EFMA amendments sought to establish a calibrated approach which enhanced deterrence by, inter alia, increasing penalties to be commensurate with the potential profits to be gained from abuses of the system:

36 It is thus apparent, as submitted by the amicus citing Lim Kopi at [10] and [11], that the EFMA aims to protect the work pass framework by imposing deterrent sentences, with offences of deception justifying stiffer penalties. Although Lim Kopi was concerned with the sentencing of a corporate offender for charges under s 22(1)(d) of the Employment of Foreign Manpower Act (Cap 91A, 1997 Rev Ed) for making bogus hires, Chao Hick Tin JA’s observations continue to apply to similar offences of deception under the current EFMA.

37 Parliament’s intent to deter offences of deception through stiff sentences is also reflected by the legislative history of s 22(1)(d). As traced by the amicus and the Prosecution, three increases in the maximum punishment under s 22(1)(d) and its predecessor provisions have occurred over the past 40 years.

38 The equivalent of s 22(1)(d) was first introduced as s 14(1)(i) in the amendments to the Regulation of Employment Act 1965 (Act No 12 of 1965) (“REA”). Section 14(1) of the REA imposed punishment of a fine of up to $1,000, imprisonment of a term not exceeding six months or both. The maximum fine under s 14(1) of the REA was subsequently increased to $5,000 when the REA was repealed and re-enacted as the Employment of Foreign Workers Act 1990 (Act No 21 of 1990) (“EFWA”), with s 18(1)(d) replacing s 14(1)(i).

39 The EFWA was in turn replaced by the Employment of Foreign Workers (Amendment) Act 2007 (No 30 of 2007). In 2007, s 18(1)(d) of the EFWA was replaced by s 22(1)(d), and the maximum punishment under s 22(1)(ii) for offences under s 22(1)(d) was increased to a maximum fine of $15,000 or 12 months’ imprisonment or both. The most recent amendments in 2012 again increased the maximum punishment under s 22(1)(ii) to a fine not exceeding $20,000, imprisonment for a term not exceeding two years or both.

40 Parliament’s intent to deter those who try to circumvent the rules on the hiring of foreign manpower through the increase in the maximum punishment was expressed by Mr Tan in the 2012 Parliamentary Debates. In particular, the stated aim was to increase EFMA penalties to reflect the advantages obtained from such contraventions, with reference made to similar offences under the Immigration Act and the Penal Code:

That correspondence to equivalent provisions in other Acts lies at the base of the approaches taken by the Prosecution and the amicus. I agree that it is appropriate to take into account similar provisions dealing with the provision of false information to public authorities in devising the appropriate sentencing framework under s 22(1)(d) of the EFMA, and deal with this further below.

41 I note that the appellant also invoked portions of the 2012 Parliamentary Debates to argue that Parliament intended that EFMA penalties should be calibrated to reflect the commercial circumstances that offenders may find themselves to be in. Mr Zainudin Nordin (“Mr Zainudin”), Member of Parliament (“MP”) for Bishan-Toa Payoh, highlighted the need for the EFMA amendments to account for the individual circumstances of errant employers. In particular, Mr Zainudin described a situation where a contractor submitted a bid for a construction project but could not hire foreign employees until the bid was approved. Where declining the project would risk his company’s viability and affect his Singaporean employees, one “could imagine that the contractor would be tempted to use other ways to get his workers.” Mr Zainudin asked that the authorities consider each individual employer’s circumstances carefully, as “not every employer who breaks the law is an evil opportunist”. ¹⁶

42 The first issue to be taken with this line of argument is that it is not the whole of the debates in Parliament that guides the interpretation of statutory provisions. It is trite that the court shall prefer the interpretation of a provision of a written law that would promote the purpose or object underlying the written law: s 9A(1) of the Interpretation Act (Cap 1, 2002 Rev Ed). To this end, s 9A(2), read with s 9A(1) and subject to s 9A(4), permits non-statutory material to be considered to confirm the meaning of a provision, or to ascertain its meaning where the provision is ambiguous or obscure, or where its ordinary meaning would lead to an absurd or unreasonable result. Sections 9A(3)(c) and 9A(3)(d) allow for Parliamentary debates to be considered as part of this analysis:

43 The Court of Appeal in Tan Cheng Bock v Attorney-General [2017] 2 SLR 850 at [51] and [52] clarified the scope of s 9A and the relevance of Parliamentary debates in assisting in the interpretation of statutes:

44 It is clear from this guidance that the courts may refer to Parliamentary debates in order to determine the purpose or object of a statute or particular provision. But not all speeches would serve this function. Where the speech relied upon is by a MP, there should be some indication that the position taken in the speech was adopted by the Government or the Minister moving the Bill. In this case, Mr Zainudin’s speech does not assist the appellant; there is no indication that Parliament adopted the position that the appellant put forward to be the purpose of the EFMA. Seen in context, Mr Zainudin’s speech did not express a view as to the purpose or object of the EFMA or its provisions. Rather, he sought to raise his concerns that the 2012 amendments might result in penalties that are too harsh. This was reflected in Mr Tan’s response speech:

45 It is worthwhile to bear in mind that determining the intention of Parliament does not mean examining the subjective intention of those involved in the drafting or the Parliamentary debates. In the words of Professor Andrew Burrows (Andrew Burrows, The Hamlyn Lectures: Thinking About Statutes: Interpretation, Interaction, Improvement (Cambridge University Press, 2018) at p 15):

Professor Burrows advocates avoiding references to Parliamentary intention, preferring a focus on the purpose of the legislation, but the use of Parliamentary intention is perhaps too ingrained. What the excerpt does underline is that while speeches of individual MPs are made as part of the process of debate and deliberation, these are not generally relevant in statutory interpretation by the courts unless they lead to or encapsulate the purpose of the statute, through their adoption in an amendment to the Bill or by the speeches of the Minister moving the Bill.

46 It is on this basis that I disagree with the appellant that the purpose of the EFMA amendments was to allow for flexibility and leniency in the sentencing of errant employers. Instead, the purpose of the EFMA amendments is clearly set out in Mr Tan’s response speech as well as the legislative history set out above: the amendments are targeted at increasing the deterrent effect of penalties for employers who undermine the work pass framework.

47 For completeness, I address the appellant’s argument that the District Judge erred in failing to consider that his actions were borne out of the commercial pressures inherent to the construction industry. He made much of the circumstances that he found himself in, citing previous defaults by Wan Fu which prevented it from further hiring foreigners and how he subsequently became caught in a spiral of financial difficulties. He further argued that he had in fact shown a readiness to be responsible by covering payment defaults by others and by personally paying approximately 300 of his workers. ¹⁷

48 In my view, none of these facts went to the question of the appropriate starting point for his sentence. Employers who breach the law would presumably either be those unperturbed by offending, or, more likely, those whose financial circumstances give rise to the temptation to contravene the law. In other words, financial pressures and financial incentives are the likely background to the commission of most offences of this type, and would not be a reason for a more lenient sentencing regime.

49 It follows from my analysis above that the predominant sentencing consideration for an offence under s 22(1)(d) is deterrence. In this regard, I endorse Chao JA’s statements in Lim Kopi at [10] and [11]: deterrence is necessary to prevent the very object of the EFMA from being flagrantly undermined. Any deception of public institutions which frustrates the aims of the EFMA should not be condoned.

50 To this end, the appellant argued that a fine would have a sufficient deterrent effect. While it is true that fines may be sufficiently deterrent in some circumstances, it does not follow that what would be deterrent in one situation would similarly be so for another. Where there is a significant wider interest to be protected, and where economic benefits may give rise to incentives to breach the law, a fine would not generally be enough to deter would-be offenders.

51 In this case, breaching the work pass system would generally bring some economic or financial advantage to the errant employer, who profits from not paying the true costs of hiring foreign manpower. Additionally, the societal interest to be protected, namely, the proper regulation of foreign manpower in Singapore’s labour market and the protection of local workers and honest employers, requires a heavy response outweighing any likely economic benefit from the breach. I also take the view, as observed in the GD at [26], that the payment of a financial penalty in the form of a fine may encourage potential offenders to treat contraventions to be mere business costs.

52 In these circumstances, I conclude that a custodial sentence should be the norm for offences under s 22(1)(d) of the EFMA. A fine would generally not be sufficient punishment unless substantial mitigating factors are present.

53 The impact of contraventions of s 22(1)(d) is perhaps more diffused than for other offences. But the impact on society and the frustration of policy goals remain: breaches of s 22(1)(d) have knock-on effects on immigration policy and the employment of foreigners. These factors point to an interest in retribution as a sentencing principle, independent from deterrence. As the High Court recognised in Angliss Singapore Pte Ltd v Public Prosecutor [2006] 4 SLR(R) 653 at [21], the public interest may necessitate a custodial sentence where the offence is serious and where retribution therefore applies. This reasoning applies here: retribution also features as a sentencing consideration for s 22(1)(d) offences, further justifying the imposition of a custodial sentence as a starting point.

54 Parties considered the sentencing frameworks set out for offences under s 57(1)(k) of the Immigration Act and s 182 of the Penal Code. Section 57(1)(k) read with s 57(1)(vi) of the Immigration Act states:

Section 182 of the Penal Code states:

55 I accept the appellant’s argument that these provisions are not in para materia with s 22(1)(d) of the EFMA. Indeed, Sundaresh Menon CJ in Idya Nurhazlyn at [31] cautioned against referring to sentences meted out for ostensibly similar offences under other provisions which carry their own considerations. Care must be taken to ensure that the offences are analogous in terms of both policy and punishment: at [30], citing Chan Sek Keong CJ in Luong Thi Trang Hoang Kathleen v Public Prosecutor [2010] 1 SLR 707 at [14].

56 In the present case, I take the view that I am entitled to consider the sentencing approaches and frameworks established in relation to s 57(1)(k) of the Immigration Act and s 182 of the Penal Code. These offences are similar in terms of policy: they are all concerned with the making of false statements to public authorities. Parliament also considered as much, having referred to these offences as informing the policy considerations behind the penalties for EFMA contraventions (see above at [40]). Menon CJ also noted in Idya Nurhazlyn at [33] that the Chowdhury framework may be usefully applied to offences under other statutes involving false statements being made to a public authority, so as to assess the seriousness of the particular offence and where in the sentencing range the case should fall.

57 As will be demonstrated, imprisonment is generally imposed where an offence implicates immigration policy (Chowhdury at [26]) and where the giving of false information to public authorities caused or had the potential to cause appreciable harm (Koh Yong Chiah at [62]). Fines are imposed for these offences only exceptionally when special mitigating circumstances are present: see Chowhdury at [27]; Koh Yong Chiah at [55].

58 The sentencing approach towards offences under s 57(1)(k) read with s 57(1)(iv) of the IA 1997 was set out by Yong Pung How CJ in Chowdhury, which involved an appellant charged with obtaining an employment pass by making a false statement.

59 Yong CJ at [25] cited the legislative history of s 57 and noted that the maximum punishment prescribed under s 57(1)(iv) was doubled in 1995 due to concerns over the increase in offences of false representation. The Minister for Home Affairs’ speech during the second reading of the Amendment Bill also reflected Parliament’s intention to take a tougher stance against such offences to stem illegal immigration in the wake of the then-economic downturn. Deterrence was necessary: false representations made under s 57(1) implicated Singapore’s immigration policies and the welfare of its citizen employees. The imposition of custodial sentences would send a firm signal deterring offences of gaining entrance to Singapore by deception. Fines would only exceptionally be warranted; to economic migrants, fines might just constitute the cost of breaking the law for personal profit: at [26]. Yong CJ went on to explain the four considerations guiding the sentencing for offences of false representation (at [28] to [31], and reproduced below at [67] and [68]).

60 In Koh Yong Chiah, the High Court refrained from doing more than giving broad guidance as to the type of cases under s 182 that generally attract a custodial sentence as a starting point, ie, cases where appreciable harm may be caused: at [50] and [52]. Examples of appreciable harm resulting in custodial terms include false allegations resulting in police reports against innocent parties which create the risk of arrest and embarrassment, or cases where false information causes a significant wastage of public resources: at [54].

61 The High Court identified non-exhaustive factors that affect the degree and harm of s 182 offences, which can be used to determine if the starting point should be departed from and/or what the appropriate quantum of fine or imprisonment term should be: at [56]. Factors relevant in assessing the level of culpability include (at [43]):

Factors relevant to assessing the level of harm caused by the offence include (at [44]):

62 Based on the above considerations, a custodial sentence should be the starting point for offences under s 22(1)(d) of the EFMA, given the legislative objectives of the EFMA in maintaining the integrity of the Singaporean workforce, the resultant need for deterrence to prevent circumventions of the work pass framework, and the seriousness and prevalence of such offences.

63 I agree with the amicus and the Prosecution that the length of the sentence should be calibrated using the “two-step sentencing bands” approach. Offences under s 22(1)(d) may involve a wide variety of factual circumstances, and the identification of “principal factual elements” may prove unduly restrictive. To this end, I adopt the Prosecution’s proposed sentencing bands, as reproduced at [25] above, which define the range of sentences which may usually be imposed for a case engaging certain offence-specific factors. In preferring the Prosecution’s sentencing framework to the amicus’s, I make three observations.

64 First, the main difference between the two sentencing frameworks lay in the range of sentences falling into each sentencing band. The amicus’s framework utilised a linear distribution of sentences: each band spanned a range of eight months’ imprisonment. ¹⁸ Bands 1 to 3 under the Prosecution’s framework encompassed sentencing ranges with widths of less than five months, ten months and nine months’ imprisonment respectively. The Prosecution explained in its oral submissions that shorter sentences are sufficient in Band 1 cases where only one or very few offence-specific factors are present. I am persuaded by this reasoning, with the qualification that there may come a time after a corpus of precedents has been built up following the present case when the courts may consider the need to recalibrate the appropriate sentencing ranges for each band.

65 Second, the Prosecution limited its sentencing framework to s 22(1)(d) false declaration offences that involve work pass applications or renewals. ¹⁹ Section 22(1)(d) covers the giving of false information “in connection with any application … for any … purpose under this Act”. The text of the provision does not limit it to applications in connection with work pass applications or renewals. As the policy considerations identified in the present case may not apply so keenly in other scenarios, I agree with the Prosecution that the sentencing approach in non-work pass cases should be left open to a future court to decide.

66 Finally, the Prosecution’s sentencing framework only covered natural persons. ²⁰ In my view, it is presently unnecessary to revisit the approach towards corporate offenders, which has already been set out in Lim Kopi.

67 Moving to the sentencing framework proper, the court is to consider the following non-exhaustive offence-specific factors at the first stage:

(a) the materiality of the false representation on the mind of the decision-maker (Chowdhury at [28]) – the greater the impact of the falsehood in inducing the grant of the application, the more severe the sentence imposed;

(b) the nature, sophistication and extent of the deception (Chowdhury at [29]) – more severe punishment is merited if the applicant went to greater lengths to deceive or if he acted in conscious defiance of public authorities;

(c) the consequences of the deception (Chowdhury at [30]) – the court may consider the extent to which harm was caused to foreign workers by way of exploitation, the wastage of resources by public authorities in uncovering the deception, whether a potentially better-qualified applicant was deprived of the job opportunity, or whether the offender put others at risk of adverse consequences by performing a job without the requisite skills;

(d) whether a transnational element was present and/or whether the offence was committed as part of a criminal syndicate’s operations;

(e) the specific role played by the offender, and, relatedly, the number of people involved in the furnishing of false information;

(f) whether the offender obtained gains (financial or otherwise) from the commission of the offence; and

(g) the motive of the offender in circumventing the work pass framework, eg, for vice or criminal activities.

Once the gravity of the offence has been ascertained based on these factors, the court places the offence within an appropriate band.

68 At the second stage, the court is to take into account the “offender-specific factors”, ie, the personal mitigating factors applicable to the offender (Chowdhury at [31]). This could relate to his character, personal attributes, expression of remorse, and cooperation with the authorities. It is envisioned that a fine might be appropriate where strong personal mitigating factors are present.

69 Finally, I agree with the Prosecution ²¹ that it would be appropriate for the court to consider imposing confiscatory fines in addition to an imprisonment term to disgorge at least some of the profits the offender may have made from his illegal behaviour: Koh Jaw Hung v Public Prosecutor [2019] 3 SLR 516 at [43], citing Poh Boon Kiat v Public Prosecutor [2014] 4 SLR 892 at [77] to [78]. While these two cases concerned vice-related offences under the Women’s Charter (Cap 353, 2009 Rev Ed), the same principle that an offender should not be allowed to profit from his illegal behaviour applies to s 22(1)(d) cases as well.

70 The appellant argued that even if a new sentencing framework is adopted, the doctrine of prospective overruling would apply such that he should not be sentenced under this new sentencing framework. ²² The Prosecution disagreed, citing Adri Anton Kalangie v Public Prosecutor [2018] 2 SLR 557 (“Adri Anton Kalangie”) at [39], [40] and [43], which emphasised that judicial pronouncements are by default retroactive in nature, and that the court’s discretion to restrict the retroactive effects of their pronouncements should only be exceptionally invoked where it is necessary to avoid serious and demonstrable injustice. The Prosecution argued that this high threshold had not been met in the present case. ²³

71 The High Court in Public Prosecutor v Hue An Li [2014] 4 SLR 661 at [124] and [125], cited in Adri Anton Kalangie at [32] and [33], set out four factors that guide the exercise of the appellant courts’ discretion to restrict the retroactive effect of their pronouncements: (a) the extent to which the law or legal principle concerned is entrenched, (b) the extent of the change to the law, (c) the extent to which the change to the law is foreseeable, and (d) the extent of reliance on the law or legal principle concerned. No one factor is preponderant over any other, and no one factor must necessarily be established before prospective overruling can be invoked in a particular case.

72 The appellant submitted that all four factors militating the invocation of the doctrine of prospective overruling are present. I disagree: contrary to what was argued by the appellant, there was hitherto no settled understanding of the law that had been entrenched, that the present sentencing framework constitutes an unforeseeable change thereof, or that could have been relied upon. As I explained above at [28] to [31], the courts have not previously taken a consistent line as to sentencing. Indeed, the Prosecution pointed towards cases where custodial terms have been imposed for offences under s 22(1)(d) of the EFMA. ²⁴

73 Regardless, the present case involved facts which warranted a stern response by the law: (a) Tan’s false representation to the WPD was material; (b) the deception by Tan and the appellant was consciously and deliberately planned and difficult to detect; (c) the deception was maintained for five to six months for each foreign employee; and (d) Wan Fu gained an economic advantage through the deception, and was able to hire foreign employees at a lower cost as a result. It is further relevant that Wan Fu was specifically precluded from hiring foreign employees due to its previous episodes of levy defaults; the appellant and Tan’s deceptions were thus calculated specifically to circumvent the work pass framework to obtain foreign manpower they knew they were not entitled to. In these circumstances, the custodial threshold would have been crossed even on the prevailing sentencing precedents.

74 Applying the sentencing framework which I have endorsed, and considering the factors raised above at [73], I am satisfied that the sentences imposed by the District Judge are appropriate.

75 As the Prosecution argued, this was a case that fell within the middle range of Band 1, with four offence-specific factors being present. An indicative starting point of two months’ imprisonment applies. A slight sentencing discount is warranted, given the appellant’s plea of guilty and cooperation with the authorities, but aggravating weight is to be attributed to the two similar charges taken into consideration for the purposes of sentencing.

76 In these circumstances, sentences of six weeks’ imprisonment for each charge are consistent with the sentencing framework articulated. It is also appropriate for two of the three imprisonment terms to run consecutively for a total of 12 weeks’ imprisonment in total.

77 For these reasons, I dismiss the appeal and uphold the District Judge’s decision to impose an aggregate sentence of a term of imprisonment of 12 weeks’ imprisonment for the three charges under s 22(1)(d) of the EFMA, read with s 23(1) and punishable under s 22(1)(ii) of the EFMA. I note my appreciation to the parties and the amicus for their helpful submissions.